MOLECULAR CELL BIOLOGY

·. .

.
' .
ABOUT THE AUTHORS

HARVEY LOOISH is Professor of Biology and Professor of Bioengineering at the Massachusetts Institute ofTechnology and a
Founding Member of the Whitehead Institute for Biomedical Research. Dr. Lodish is also a member of the National Academy of
Sciences and the American Academy of Arts and Sciences and was President (2004) of the American Society for Cell Biology.
He is well known for his work on cell-membrane physiology, particularly the biosynthesis of many cell-surface proteins, and
orrtl re lluning and functional analysis of several cell-surface receptor proterns, such as the erythropoietin and TGF ·13 receptors.
His 1aboratory also studies hematopoietic stem cells and has identified novel proteins that support their proliferation. Dr. Lodish
teaches undergraduate and graduate courses in cell biology and biotechnology. Photo credit: John Soares/Whrtehead Institute

ARNOLD BERK holds the UCLA Presidential Chair in Molecular Cell Biology in the Department of Mrcrobiology, Immunology,
and Molecular Genetics and is a member of the Molecular Biology Institute at the University of California, Los Angeles. Dr. Berk
is also a fellow of the American Academy of Arts and Sciences. He is one of the original discoverers of RNA splicing and of
mechanisms for gene control in viruses. His laboratory studies the molecular interactions that regulate transcription initiation
in mammalian cells, focusing n particular on adenovirus regulatory proteins. He teaches an adva'nced undergraduate course
in cell biology of the nucleus and a graduate course in brochemistry.

CHRIS A. KAISER s Professor and Head of the Department of Biology at the Massachusetts Institute ofTechnology. His
1aboratory uses genetic and cell biological methods to understand the basic processes of how newly synthesized membrane
and secretory proteins are folded and stored in the compartments of the secretory pathway. Dr. Kaiser is recognized as a top
undergraduate educator at MIT, where he has taught genetics to undergraduates for many years. '

MONTY KRIEGER rs the Whitehead Professor 1n the Department of Biology at the Massachusetts lnstrtute ofTechnology and
a Senror Associate Member of the Broad lnstrtute of MIT and Harvard. Dr Krieger is also a member of the National Academy
of Sciences. For his innovative teachrng of undergraduate biology and human physrology as well as graduate cell· biology
courses, he has received numerous awards. His laboratory has made contributions to our understanding of membrane traf·
ticking through the Golgi apparatus and has cloned and characterized receptor proteins important for pathogen recognrtion
and the movement of cholesterol into and out of cells, including the HDL receptor

ANTHONY BRETSCHER is Professor of Cell Biology at Cornell University and a member of the Weill Institute for Cell and
Molecular Brology. t i s laboratory is well known for identifying and characterizing new components of the actrn cytoskeleton
and elucidating the biological functions of those components in relation to cell polarity and membrane traffic. For this work,
his laboratory exploits biochemical, genetic, and cell biological approaches in two model systems, vertebrate epithelial cells
and the budding yeast. Dr Bretscher teaches cell biology to undergraduates at Cornel University.

HID DE PLOEGH is Professor of Biology at the Massachusetts Institute of Technology and a member of the Whrtehead
nst<tute for Bromed ical Research. One of the world's leading researchers rn immune system behavior, Dr. Ploegh studies the
various tactics that viruses employ to evade our rmmune responses and the ways our immune system distinguishes friend
from foe. Dr. Ploegh teaches immunology to undergraduate students at Harvard University and MIT

ANGELIKA AMON 1 Professor of Biology at the Massachusetts Institute ofTechnology, a member of the Koch Institute for
lntegrat: •e Cancer Re•;earch, and Investigator at the Howard Hughes Medical Institute She is also a member of the National
Academy of Sciences. Her laboratory studies the molecular mechanisms that govern chromosome segregation during mitosrs
and meiosrs and the consequences-aneuploidy-when these mechanisms fail during normal cell proliferation and cancer
development. Dr. Amon teaches undergraduate and graduate courses n cell biology and genetics.
MOLECULAR CELL
BIOLOGY
SEVENTH EDITION

Harvey Lodish
Arnold Berk
Chris A. Kaiser
Monty Krieger
Anthony Bretscher
Hidde Ploegh
Angelika Amon
Matthew P. Scott

II
W. H. Freeman and Company
New York
 PUBLISHER: Katherine Ahr Parker
ACQUISITIONS EDITOR: Beth McHenry
DEVELOPMENTAL EDITORS: Matthew Tontonoz, Erica Pantages Frost, Erica Champion
AssiSTANT EDITOR: Marni Rolfes
AssoCIATE DIRECTOR OF MARKETING: Debbie Clare
SENIOR PROJECT EDITOR: Mary Louise Byrd
TExT DESIGNER: Marsha Cohen
PAGl:. MAKEUP: Aptara®, Inc.
CoVER DESIGN: Victoria Tomaselli
ILLUSTRATION CooRDINATOR: Susan Timmins
ILLUSTRATIONS: Network Graphics, Erica Beade, H. Adam Steinberg
PHOTO EDITOR: Cecilia Varas
PHOTO RESEARCHER: Christina Micek ..
PRODUCTION MANAGER: Julia DeRosa
MEDIA AND SuPPLEMENTS EDITOR: Beth McHenry
MEDIA DEVELOPERS: Biostudio, Inc., Sumanas, lnc.
CoMPOSITION: Aptara®, Inc.
MANUFACTURING: RR Donnelley & Sons Company

·.
Library of Congress Control Number: 2012932495

ISBN-13: 978-1-4292-3413-9
ISBN-10: l-4292-3413-X
© 2000, 2004, 2008, 2013 by W. H. Freeman and Company
All rights reserved.

Printed in the United States of America

First printing

W. H. freeman and Company

41 Madison Avenue, New York, NY 10010
Houndmills, Basingstoke
RG21 6XS, England

www.whfreeman.com
 To our students and to our teachers,
from whom we continue to learn, and to our families,
for their support, encouragement, and love

'·
l .'
 PREFACE

n writing the seventh edition of Molecular Cell Biology with simplified overview figures, to help students navigate

I we have incorporated many of the spectacular advances

made over the past four years in biomedical science,
driven in part by new experimental technologies that have
the complexity of signaling pathways.
• "The Eukaryotic Cell Cycle" (Chapter 19) now begins with
the concepts of"START" (a cell's commitment to entering the
revolutionized many fields. Fast techniques for sequencing
cell cycle starting with DNA synthesis) and then progresses
DNA and RNA, for example, have uncovered many novel
through the cycle stages. The chapter focuses on yeast and
noncoding RNAs that regulate gene expression and identi-
mammals and uses general names for cell cycle components
fied hundreds of human genes that affect diseases such as
when possible to improve student understanding.
diabetes, osteoporosis, and cancer. Genomics has also led to
many novel insights into the evolution of life forms and the • "Stem Cells, Cell Asymmetry, and Cell Death" (Chapter 21)
functions of individual members of multiprotein families. now incorporates developmental topics, including new cover-
Exploring the most current developments in the field is al- age of induced pluripotent stem (iPS) cells.
ways a priority in writing a new edition, but it is also impor-
tant to us to communicate the basics of cell biology clearly.
To this end, in addition to introducing new discoveries and
technologies, we have streamlined and reorganized several
chapters to clarify processes and concepts for students.

New Co-Author, Angelika Amon

The new edition of MCB introduces a new member to our
author team, respected researcher and teacher Angelika
Amon of the MassachusettS Institute of Technology. Her
laboratory uses the budding yeast S. cerevisiae and mouse
and cell culture models to gain a detailed molecular under-
standing of the regulatory circuits that control chromosome
segregation and the effects of aneuploidy on cell physiology.
Dr. Amon also teaches undergraduate and graduate courses
in Cell Biology and Genetics. FIGURE 9-22 In this mouse fibroblast, FRET has been used to reveal
that the interaction between an active regulatory protein (Rae) and its
binding partner is localized to the front of the migrating cell.
Revised, Cutting Edge Content
The seventh edition of Mr;>lecular Cell Biology includes new
and improved chapters:
• "Molecules, Cells and Evolution" (Chapter l) now frames Increased Clarity, Improved Pedagogy
cell biology in the light of evolution: this perspective explains As experienced teachers of both undergraduate and graduate
why scientists pick particular unicellular and multicellular students, we are always striving to improve student under-
"model" organisms to study specific genes and proteins that standing. In this seventh edition, perennially confusing topics,
are important for cellular function . such as cellular energetics, cell signaling, and immunology,
have been streamlined and revised to improve student un-
• ''Culturing, Visualizing, and Perturbing Cells" (Chapter 9)
derst:mding. Each figure was reconsidered and, if possible,
has been rewritten to include cuttmg edge methods includ-
simplified to highlight key lessons. Heavily revised end-of-
ing FRAP, FRET, siRNA, and chemical biology, making it a
.. state-of-the-art methods chapter.
chapter materials include 30% new questions, including ad-
ditional Analyze the Data problems to give students further
• "Signal Transduction and G Protein-Coupled Receptors" practice at interpreting experimental evidence. The result is a
and "Signaling Pathways that Control Gene Expression" balance of state-of-the-art currency and experimental focus
(Chapters 15 and 16) have been reorganized and ill ustrated with attention to clarity, organization, and pedagogy.

vii
 (a) Amphitelic attachment (b) Merotelic attachment • Assembly of the multiprotein T-cell receptor complex
(Ch. 10 )
@Cohesins
• Structure of the Na /K+ ATPase (Ch. 11 )

===~~
• Structure and mechanism of the multidrug transporter
ABCB1 (MDR1) (Ch. 11)
§ Miocot"b"'"' • Structure and function of the cystic fibrosis transmem-
brane regulator (CITR) (Ch. 11 )

v
Sister chromatids
• The role of an anion antiporter in bone resorption (Ch. 11)
• Structures of complex I and II as well as the mechanism of
(c) Syntelic attachment (d) Monotelic attachment electron flow and proton pumping in the electron transport
chain (Ch. 12)
• Generation and inactivation of toxic reactive oxygen spe-
cies (ROS ) (Ch. 12)
• The mechanism of proton flow 'through the half-channels
of ATP Synthase (Ch. 12)
• Tail-anchored membrane proteins (Ch. 13)
• How modifications of N-linked oligosaccharides are used
FIGURE 19-25 Stable and unstable chromosome attachments. to monitor protein folding and quality control (Ch. 13)
• The mechanism of formation of multivesicular endosomes
involving ubiquitination and ESCRT (Ch. 14)

New Discoveries, New Methodologies • Advances in our understanding of autophagy as a mecha-

nism for recycling organelles and proteins (Ch. 14 )
• Covalent regulation of protein activity by ubiquitination/ • Affinity purification techniques for studying signal trans-
deubiquitination (Ch. 3) duction proteins (Ch. 15)
• Molecular chaperones including the Hsp90 family of pro- • Structure of the (3-adrenergic receptOr in the inactive and
teins (Ch. 3) active states and with its associated trimeric G protein, G,,
• Mammahan protein synthesis and the roles of polymer- (Ch. 15)
ases delta (lagging strand) and epsilon (leading strand) in
• Activation of EGF receptor by EGF via the formation of
eukaryotic DNA: synthesis (Ch. 4) an asymmetric kinase domain dimer (Ch. 16)
• Non-radioactive probes (for in-situ hybridization, for
• Hedgehog signaling in vertebrates involving primary cilia
example) (Ch. 5)
(Ch. 16)
• Quantitative PCR (and RT-PCR) and high-throughput
• NF-KB signaling pathway and polyubiquitin scaffolds
DNA sequencing (Ch. 5)
(Ch. 16 )
• DNA fingerprinting using microsatellites and PCR (Ch. 6)
• Integration of signals in fat cell differentiation via PPAR-y
• Personal genome seq uencing and the 1000 Genome Proj- (Ch . 16)
ect (Ch. 6)
• Mechanism of Arp2/3 nucleation of actin filaments (Ch. 17)
• Epigenetic mechanisms of transcriptional regulation (Ch. 7)
• The dynamics of microfilaments during endocytosis and
• Transcriptional regulation by non-coding RNAs (e.g., Xist
the role of endocytic membrane recycling during cell migra-
in X-chromosome inactivation, siRNA-directed heterochro-
tion (Ch. 17)
matin formation in fission yeast and DNA methylation in
plants) (Ch. 7) • lntraflagellar transport and the function of primary cilia
(Ch. 18 )
• Fluorescent mRNA labeling to follow mRNA localization
in live cells (Ch. 8) • Plant mitosis and cytokinesis (Ch. 18)
• Structure and function of the nuclear pore complex (Chs. 8 • + TIPs as regulators of microtubule (+)end function (Ch. 18 )
and 13) • Proteins involved in mitotic spindle formation and kineto-
• Additional coverage of FRAP, FRET, and siRNA tech- chore attachment to microtubules (Ch. 19)
niques (Ch. 9) • Elastic fibers that permit many tissues to undergo repeated
• Lipid droplets and their formation (Ch. 10) stretching and recoiling (Ch. 20)

viii PREFACE
 • Extracellular matrix remodelling and degradation by ma- learning. Many of these applications hinge on a derailed un-
trix metalloproteinases (Ch. 20) derstandi ng of multiprotein complexes in cells-complexes
• Stem cells in the intestinal epithelium (Ch . 21) that catalyze cell movements; regulate DNA tramcripuon,
replication, and repair; coordinate metabolism; and connect
• Regu lation of gene expression in embryonic stem (ES} cells to other cells and to proteins and carbohydrates in their
cells (Ch. 21) extracellular environment.
• Generation of induced pluripotent stem (iPS) cells (Ch. 21) The following is a list of new medical examples.
• Advances in our understanding of regulated cell death • Cholesterol transport a nd atherosclerosis as an illustra-
(Ch. 21) tion ot the hydrophobic effect (Ch. 2)
• Structure of the nicotinic acetylcholine receptor (Ch. 22) • Use of genetically engineered corn with high lysine content
• Molecular model of the MEC-4 touch receptor complex to promote the growth of livestock as an ill ustration of im-
in C. e/egans (Ch. 22) portance of essential ami n o acids (Ch. 2)
• Synapse formation in neuromuscular junctions (Ch. 22) • Poliovirus and HIV-1 as examples of vtruses that infect
• Toll -like receptors (TLRs) a nd the inflammasome (Ch. 23) only certain cell types due to tissue-specific cell surface re-
ceptors (Ch. 4 )
• Epigenetics and cancer (Ch. 24)
• HPV vaccine and its abi lity to protect against common
types of HPV, and the development of cervical cancer (Ch. 4)
• Huntington's disease as an example of a microsatellite ex-
pansion disease (Ch. 6)
• Potential t reatment o f cystic fibrosis using small molecules
that wou ld allow the m utant protein to traffic normally to
the cell surface (Ch. 11)
• Role o f genetic defects in ClC-7, a chloride ion channel, m
the hereditary bone disease osteopetrosis (Ch . 11)
• Mitochondrial diseases such as Charcot-Marie-Tooth dis-
ease and Miller syndrome (Ch. 12)
• Use of ligand-binding domains of cell-surface receptors as
therapeutic drugs, such as the extracellular domain ofT Fa
receptor to treat arthritis and other inflammatory conditions
(Ch. 15)
• Role of H edgehog (Hh ) signaling in human cancers includ-
ing medulloblastomas and rhabdomyosarcomas (Ch. 16)
• Role of B-Raf kinase in melanoma and use of selective
inhibitors of B-Raf in cancer treatment (Ch. 16)
• Defects in a regulator of dynein as a cause of lissencephaly
(Ch. 18 )
Cells being born in the developing cerebellum. • Elastic fiber p rotein fibrill in 1 and Marfan's Syndrome
'
(Ch. 20)
• Use of iPS cells in uncovering the molecular basis of ALS
Medical Relevance (Ch. 2 1)
Many advances in basic cellular and molecular biology have
• Variations in human sense of smell (Ch. 22)
led to new treatments for cancer and othe r significant human
diseases. These medica l examples are woven throughou t the • Microarray ana lysis of breast cancer tumors as a way to
chapters w here appropriate to give students an apprecia- distinguish gene expression patterns and individualize treat-
tion for the clinical applications of the basic science they are ment (Ch. 24)

·.

PREFACE lx
 MEDIA AND SUPPLEMENTS

For Students For Instructors

•·NEW"· BioPortal for Molecular Cell Biology A robust *NEW* BioPortal for Molecular Cell Biology In addition
teaching and learning tool with all of the study and quizzing to all student resources (including NEW LearningCurve
resources available through the Companion Web Site (listed quizzing tool) and a dynamic eBook, BioPonal also in-
below) as well as a fully-interactive eBook. BioPortal also cludes tools for instructors. Robust gradebook and assign-
includes NEW LearningCurve, a self-paced adaptive quiz- ment features allow instructors to assign any materials to
zing tool for students. With questions tailored to their target their students and monitor their progress throughout the
difficulty level and an engaging scoring system, Learning- semester. Visit http://courses.bfwpub.com for more infor-
Curve encourages students to incorporate content from the mation.
text into their study routine and provides them with a study
plan upon completion. Companion Web Site www.whfreeman.com/lodish7e
All the student resources, plus:
Companion Web Site www.whfreeman.com/lodish7e
• All figures and tables from the book in ]PEG and Power-
• Podcasts narrated by the authors give students a deeper Point formats, which instructors can edit and project section
understanding of key figures in the text and a sense of the by section, allowing students to follow underlying concepts.
thrill of discovery. Optimized for lecture-hall presentation, including enhanced
• More than 125 animations and research videos show the colors, enlarged labels, and boldface type.
dynamic nature of key cellular processes and important ex- • Test Bank in editable Microsoft Word format now fea-
perimental techniques. turing new and revised questions for every chapter. The test
• Classic Experiment essays focus on classic groundbreak- bank is written by Brian Storrie of the University of Arkan-
ing experiments and explore the investigative process. sas for Medical Sciences and Eric A. Wong, Richard Walker,
• Online Quizzing is provided, including multiple-choice Glenda Gillaspy, and Jill Sible of Virginia Polytechnic Insti-
and short answer questions. tute and Stare University and revised by Cindy Klevickis of
James Madison University and Greg M. Kelly of the Univer-
Student Solutions Manual (ISBN: 1-4641-0230-9), written sity of Ontario.
by Brian Storrie of the University of Arkansas for Medical • Additional Analyze the Data problems are available in
Sciences, Eric A. Wong, Richard Walker, Glenda Gillaspy, PDF format.
and Jill Sible of Virginia Polytechnic Institute and State • Lecture-ready Personal Response System "clicker" ques-
University and updated by Tom Huxford of San Diego tions are available as Ylicrosoft Word files and Microsoft
State University, Stephanie Bingham of Barry University, PowerPoint slides.
Brian Sato of University of California-Irvine, Steve Amato
of Johns Hopkins University, Greg Kelly of University of Instructor's Resource CD-ROM (ISBN: 1-4292-0126-6)
Western Ontario, Tom Keller of Florida State University, includes all the instructor's resources from the Web site,
and Elizabeth Good of University of Illinois-Urbana- including all the illustrations from the text, animations,
Champaign, contains complete worked-out solutions to all videos, te~t bank files, clicker questions, and the solutions
the end-of-chapter problems in the textbook. manual files.

eBook (ISBN: 1-4641-0229-5) This customizable eBook Overhead Transparency Set (ISBN: 1-4292-0477-X)
fully integrates the complete contents of the text and its contains 250 key illustrations from the text, optimized for
interactive media in a format that features a variety of help- lecture-hall presentation.
ful study tools, including full-text searching, note-taking,
bookmarking, highlighting, and more. Easily accessible
on any Internet-connected computer via a standard Web
browser, the eBook enables students to take an active ap-
proach to their learning in an intuitive, easy-to-use format.
Visit http://ebooks.bfwpub.com to learn more.

·.
x PREFACE
 ACKNOWLEDGMENTS

In updating, revising and rewriting this book, we were given Topher Gee, University of North Carolina, Charlotte
.~. · invaluable help by many colleagues. We thank the follow- Mary Gehring, Massachusetts Institute of Technology
ing people who generously gave of their time and expertise Elizabeth Good, University of Illinois, Urbana-Champaign
by making contributions to specific chapters in their areas
David Goodenough, Harvard Medical School
of interest, providing us with detailed information about
their courses, or by reading and commenting on one or more Mark Grimes, University of Montana, Missoula
chapters: Lawrence I. Grossman, Wayne State University
Michael Grunstein, University of California, Los Angeles,
David Agard, University of California, San Francisco
School of Medicine
Ravi Allada, Northwestern University
Barry M. Gumbiner, University of Virginia
Stephen Amato, Boston College
Yanlin Guo, University of Southern Mississippi
James M. Anderson, National Institutes of Health and Uni-
Leah Haimo, University of Califorma, Riverside
versity of North Carolina, Chapel Hill
Craig Hart, Louisiana State University
Kenneth Balazovich, University of Michigan, Ann Arbor
Michael Hemann, Massachusetts Institute of Technology
Amit Banerjee, Wayne State University
Chris Hill, University of Utah
Amy Bejsovec, Duke University
H. Robert Horvitz, Massachusetts Institute ofTechnology
Andrew Bendall, University of Guelph, Ridgetown
Tim C. Huffaker, Cornell University
Stephanie Bingham, Barry University, Dwayne 0. Andreas
School of Law Tom Huxford, San Diego State Uniuersity
Doug Black, Howard Hughes Medical Institute and Univer- Richard Hynes, Massachusetts Institute of Technology and
sity of California, Los Angeles Howard Hughes Medica/Institute
Heidi Blank, Massachusetts Institute of Technology Naohiro Kato, Louisiana State University
Jonathan Bogan, Yale University School of Medicine Amy E. Keating, Massachusetts Institute ofTechnology
Laurie Boyer, Massachusetts Institute of Technology Thomas Keller, Florida State Univemty, Panama C1ty
William J. Brown, Cornell University Greg Kelly, University of Western Ontario
Steve Burden, New York Univers1ty Leung Kim, Florida International University, Biscayne Bay
Monique Cadrin, Urziversite du Quebec a Trois-Rivieres Gwendolyn M. Kine brew, j ohn Carroll University
Steven A. Carr, Broad Institute of Harvard and Massachusetts Ashwini Kucknoor, Lamar UniLJersity
Institute ofTechnology Mark Lazzaro, College of Charleston
Paul Chang, Massachusetts lnstitute ofTechnology Maureen Leupold, Genesee Community College, Batavia
Kuang Yu Chen, Rutgers, The State University of New Robert Levine, McGill University
Jersey, Camden Fang Ju Lin, Coastal Carolina Uniuersity
Orna Cohen-Fix, National Institutes of Health Susan Lindquist, Massachusetts Institute ofTechology
Ronald Cooper, University of California, Los Angeles Song-Tao Liu, University of Toledo, Scott Park
David Daleke, lndiana State University Elizabeth Lord, University of California, Rit•erside
Elizabeth De Stasio, Lawrence University Charles Mallery, University of Miami
Linda DeVeaux, Idaho State University C. William McCurdy, University of California, Davis, and
Richard Dickerson, University of California, Los Angeles Lawrence Berkeley National Laboratory
Patrick DiMario, Louisiana State University David McNabb, University of Arkansas
Glenn Dorsam, North Dakota State University James McNew, Rice University
William Dowhan, University of Texas, Houston Raka Mitra, Carleton College
Janet Duerr, Ohio University Ivana Mladenovic, Simon Fraser University
Robert H . Fillingame, Uniuersity of Wisconsin Medical Vamsi K. Mootha, MassachHsetts General Hospital, Boston
School Roderick Morgan, Grand Valley State University
Gerry Fink, Massachusetts Institute ofTechnology Dana Nayduch, Georgia Southern University
David Foster, City University of New York, Hunter College Brent Nielsen, Brigham Young University
Gail Fraizer, Kent State Uniuersity, East Liverpool Terry Orr-Weaver, Massachusetts Institute ofTeclmology
Margaret T. Fuller, Stanford University School of Medicine Rekha Patel, University of South Carolina, Lancaster

PREFACE xi
David Paul, Harvard Medical School Victoria Tomaselli, Christina Micek, Bill O'Neal, Marni
Debra Pires, University of California, Los A11geles Rolfes, Beth McHenry, Susan Timmins, Cecilia Varas, and
Nicholas Quintyne, Florida Atlantic University, JufJiter Julia DeRosa for their labor and for their willingness to
work overtime to produce a book that excels in every way.
Alex Rich, Massachusetts Institute ofTechnology
ln particular, we would like to acknowledge the talent and
Edmund Rucker, University of Kentucky commitment of our text editors, Matthew Tontonoz, Erica
Brian Sa to, University of California, Irvine Pantages Frost, and Erica Champion. They are remarkable
Robert Sauer, Massachusetts Institute ofTechology editors. Thank you for all you've done in this edition.
Thomas Schwartz, Massachusetts Institute ofTeclmology We arc also indebted to I I. Adam Steinberg for his peda-
Gowri Selvan, University of California, Irvine gogical insight and his development of beautiful molecular
models and illustrations.
Jiahai Shi, Whitehead Institute for Biomedical Research
We would like to acknowledge those whose direct con-
Daniel Simmons, University of Delaware tributions to previous editions continue to influence in this
Stephen T. Smale, University of California, Los Angeles edition; especially Ruth Steyn.
Paul Teesdale-Spittle, Victoria University of Wellington Thanks to our own staff: Sally Bittancourt, Diane Bush,
Fernando Tenjo, Virginia Commonwealth University Mary Anne Donovan, Carol Eng, James Evans, George
Andrei Tokmakoff, Massachusetts Institute of Technology Kokkinogenis, Julie Knight, Gui~ky Waller, Nicki Watson,
and Rob Welsh.
Harald Vaessin, Ohio State Universit)~ Columbus
Finally, special thanks to our families for inspiring us
Peter van der Geer, San Diego State University and for granting us the time it takes to work on such a book
Volker M. Vogt, Cornell University and to our mentors and advisers for encouraging us in our
Michael B. Yaffe, Massachusetts Institute ofTechnology studies and teaching us much of what we know: (Harvey
Jing Zhang, University of Wisconsin Lodish) my wife, Pamela; my children and grandchildren
Heidi and Eric Steinert and Emma and Andrew Steinert;
We would also like to express our gratitude and appre- Martin Lodish, Kristin Schardt, and Sophia, Joshua, and
ciation to Leah Haimo of the University of California, River- Tobias Lodish; and Stephanie Lodish, Bruce Peabody, and
side, for her development of new Analyze the Data problems, Isaac and Violet Peabody; mentors Norton Zinder and Sydney
to Cindy Klevickis of james Madison University and Greg Brenner; and also David Baltimore and Jim Darnell for col-
.'vi. Kelly of the University of Ontario for their authorship of laborating on the first editions of this book; (Arnold Berk)
excellent new Review the Concepts problems and Test Bank my wife Sally, jerry Berk, Shirley Berk, Angelina Smith, David
questions, and to Jill Sible of Virginia Polytechnic Institute Clayton, and Phil Sharp; (Chris A. Kaiser) my wife Kathy
and State University for her revision of the Online Quizzing O'Neill; (Monty Krieger) my wife Nancy Krieger, parents
problems. We are also grateful to Lisa Rezcnde of the Uni- I. Jay Krieger and Mildred Krieger, and children jonathan
versity of Arizona for her development of the Classic Experi- Krieger and Joshua Krieger; my mentors Robert Stroud,
ments and Podcasts. Michael Brown, and joseph Goldstein; (Anthony Bretscher)
This edition would not have been possible without the my wife Janice and daughters Heidi and Erika, and advisers
careful and committed collaboration of our publishing part- A. Dale Kaiser and Klaus Weber; (Hidde Ploegh) my wife
ners at W. H. Freeman and Company. We thank Kate Ahr Anne Mahon; (Angelika Amon) my husband Johannes Weis,
Parker, Mary Louise Byrd, Debbie Clare, Marsha Cohen, Theresa and Clara Weis, Gerry Fink and Frank Solomon.

xli PREFAC E
CONTENTS IN BRIEF

Part I Chemical and Molecular Foundations

1 Molecules, Cells, and Evolution 1
2 Chemical Foundations 23
3 Protein Structure and Function 59

Part II Genetics and Molecular Biology

4 Basic Molecular Genetic Mechanisms 115
5 Molecular Genetic Techniques 171
6 Genes, Genomics, and Chromosomes 223

7 Transcriptional Control of Gene Expression 279

8 Post-transcriptional Gene Control 345

Part Ill Cell Structure and Function

9 Culturing, Visualizing, and Perturbing Cells 397
10 Biomembrane Structure 443
11 Transmembrane Transport of Ions and Small Molecules 473
12 Cellular Energetics 517
13 Moving Proteins into Membranes and Organelles 577
14 Vesicular Traffic, Secretion, and Endocytosis 627
15 Signal Transduction and G Protein-Coupled Receptors 673
16 Signaling Pathways That Control Gene Activity 721
17 Cell Organization and Movement 1: Microfilaments 773
18 Cell Org~nization and Movement II: Microtubules and Intermediate Filaments 821
19 The Eukaryotic Cell Cycle 873

Part IV Cell Growth and Development

20 Integrating Cells Into Tissues 925
21 Stem Cells, Cell Asymmetry, and Cell Death 977
22 Nerve Cells 1019
23 Immunology 1059
24 Cancer 1113

xiii
.,,
CONTENTS

Preface vii Genetic Diseases Elucidate Important Aspects of Cell Function 22

The Following Chapters Present Much Experimental Data
Part I Chemical and Molecular That Explains How We Know What We Know About Cell
Structure and Function 22
Foundations

1 Molecules, Cells, and Evolution 1 2 Chemical Foundations 23

1.1 The Molecules of Life 4 2.1 Covalent Bonds and Noncovalent

Prot eins Give Cells Structure and Perform Most Cellular Tasks 6 Interactions 24
Nucleic Acids Carry Coded Information for Making The Electronic Structure of an Atom Determines
Prot eins at t he Right Time and Place 7 t he Number and Geometry of Covalent Bonds
It Can Make 2S
Phospholipids Are t he Conserved Building Blocks
of All Cellular Membranes 10 Elect rons May Be Shared Equally or Unequally
in Covalent Bonds 26
Covalent Bonds Are Much Stronger and More Stable
1.2 Genomes, Cell Architecture, Than Noncovalent Interactions 28
and Cell Function 10
Ionic Interactions Are Attractions Between Oppositely
Prokaryotes Comprise True Bacteria and Archaea 10 Charged Ions 28
Escherichia coli Is Widely Used in Biological Resea rch 13 Hyd rogen Bonds Are Noncovalent Interactions
All Eukaryotic Cells Have Many of the Same Organelles That Determine the Water Solubility
and Other Subcellular Structures 13 of Uncharged Molecules 28

Cellular DNA Is Packaged Within Chromosomes 15 Van der Waals Interactions Are Weak Attractive
Interactions Caused by Transient Dipoles 30
All Eukaryotic Cell s Utilize a Simil ar Cycle to Regulat e
Their Div ision 1S The Hydrophobic Effect Causes Nonpolar Molecules
to Adhere to One Another 31
Molecular Complementarity Due to Noncovalent
1.3 Cells into Tissues: Unicellular and Metazoan
Interactions Leads to a Lock-and-Key Fit Between
Organisms Used for Molecular Cell Biology Biomolecu les 32
Investigations 16
Single-Celled Eukaryotes Are Used to Study
2.2 Chemical Building Blocks of Cells 33
Fu ndamental Aspects of Eukaryot ic Cell Structure
and Fu nction 16 Am ino Acids Differing Only in Their Side Chains
, Compose Proteins 33
Mutations in Yeast Led to the Identification of Key Cell
Cycle Prot eins 17 Five Different Nucleotides Are Used to Build Nucleic Acids 36

Multicellularity Requires Cell-Cell and Cell Matrix Adhesions 17 Monosaccharides Covalently Assemble into Linear
and Branched Polysaccharides 37
Tissues Are Organized into Orga ns 18
Phospholipids Associate Noncovalently to Form
Body Plan and Rud iment ary Tissues Form Early
t he Basic Bilayer Struct ure of Biomembranes 40
in Embryonic Development 18
Invertebrates, Fish, and Other Organisms Serve
as Experiment al Systems for Study of Human 2.3 Chemical Reactions and Chemica l
Development 19 Equilibrium 43
Mice Are Frequen tly Used to Generate Models A Chemical Reaction Is in Equilibrium When t he Rates
of Human Disease 20 of the Forward and Reverse Reactions Are Equal 43
Viruses Are Cellular Parasites That Are Widely The Equilibrium Constant Reflects the Extent of
Employed in Molecular Cell Biology Resea rch 21 a Chemical Reaction 44

XV
Chemical Reactions in Cells Are at Steady State 44 Folding of Proteins in Vivo Is Promoted by Chaperones 72
Dissociation Const ants of Binding Reactions Reflect Alternatively Folded Proteins Are Implicated in Diseases 76
the Affinity of Interacting Molecules 44
Biological Fluids Have Characteristic pH Values 45 3.3 Protein Binding and Enzyme Catalysis 77
Hydrogen Ions Are Released by Acids and Taken Up by Bases 46 Specific Binding of Ligands Underlies the Functions
of Most Proteins 77
Buffers Maintain the pH of Intracellular and Extracellular
Fluids 47 Enzymes Are Highly Efficient and Specific Catalysts 78
An Enzyme's Active Site Binds Substrates and Cdrries
2.4 Biochemical Energetics 48 Out Catalysis 79
Several Forms of Energy Are Important in Biological Serine Proteases Demonstrate How an Enzyme's Active
Systems 48 Site Works 80
Cells Can Transform One Type of Energy into Another 49 Enzymes in a Common Pathway Are Often Physically
Associated with One Another 84
The Change in Free Energy Determines If a Chemica l
Reaction Will Occur Spontaneously 49
3.4 Regulating Protein Function 85
The _l(jO' of a Reaction Can Be Calculated from Its Keq 51
Regu lated Synthesis and Degradation 'of Proteins
The Rate of a Reaction Depends on the Activation Is a Fundamental Property of Cells 85
Energy Necessary to Energize the Reactants
The Proteasome Is a Molecular Machine Used to Degrade
into a Transition State 51
Proteins 85
Life Depends on the Coupling of Unfavorable Chemical
Ubiquitin Marks Cytosolic Proteins for Degradation
Reactions with Energetically Favorable Ones 52
in Proteasomes 87
Hydrolysis of ATP Releases Substantial Free Energy
Noncovalent Binding Permits Allosteric, or Cooperative,
and Drives Many Cellular Processes 52
Regulation of Proteins 88
ATP Is Generated During Photosynthesis and Respiration 54
Noncovalent Binding of Calcium and GTP Are Widely
NAD and FAD Couple Many Biological Oxidation Used as Allosteric Switches to Control Protein Activity 88
and Reduction Reactions 54
Phosphorylation and Dephosphorylation Covalently
Regulate Protein Activity 90
Ubiquitination and Deubiquitination Covalently Regulate
3 Protein Structure and Function 59 Protein Activity 90
Proteolytic Cleavage Irreversibly Activates or Inactivates
3.1 Hierarchical Structure of Proteins 61 Some Proteins 92
The Primary Structure of a Protein Is Its Linear Higher-Orde r Regu lation Includes Control of Protein
Arrangement of Amino Acids 61 Location and Concentration 92
Secondary Structures Are the Core Elements of Protein
Architecture 62 3.5 Purifying, Detecting, and Characterizing
Tertiary Structure Is the Overall Folding of a Polypeptide Proteins 93
Chain 64 Centrifugation Can Separate Particles and Molecules
Different Ways of Depicting the Conformation of Proteins That Differ in Mass or Density 93
Convey Different Types of Information 64 Electrophoresis Separates Molecules on the Basis
Structural Motifs Are Regular Combinations of Secondary of Their Charge-to-Mass Ratio 94
Structures 65 Liquid Chromatography Resolves Proteins by Mass, Charge,
Domains Are Modules of Tertiary Structure 67 or Binding Affinity 96
Multiple Polypeptides Assemble into Quaternary Structures Highly Specific Enzyme and Antibody Assays Can Detect
and Supramolecular Complexes 68 Individual Proteins 97
Members of Protein Families Have a Common Evolutionary Radioisotopes Are Indispensable Tools for Detecting
Ancestor 69 Biological Molecules 99
Mass Spectrometry Can Determine the Mass and Sequ ence
3.2 Protein Folding 70 of Proteins 101
Planar Peptide Bonds Limit the Shapes into Which Proteins Protein Primary Structure Can Be Determined
Can Fold 71 by Chemical Methods and from Gene Sequences 104
The Amino Acid Sequence of a Protein Determines Protein Conformation Is Determined by Sophisticated
How It Will Fold 71 Physical Methods 104

xvi CONTENTS
3.6 Proteomics 106 Translation Is Terminated by Release Factors
When a Stop Codon Is Reached 142
Proteomics Is the Study of All or a Large Subset
of Proteins in a Biological System 106 Polysomes and Rapid Ribosome Recycling Increase
the Efficiency ofTranslation 142
Advanced Techniq ues in Mass Spectrometry
Are Critical to Proteomic Analysis 108 GTPase-Superfamily Proteins Function in Several
Quality Control Steps ofTranslation 143
Nonsense Mutations Cause Premature Termination
Part II Genetics and Molecular Biology of Protein Synthesis 143

4.5 DNA Replication 145

4 Basic Molecular Genetic Mechanisms 115
DNA Polymerases Require a Primer to Initiate Replication 145

4.1 Structure of Nucleic Acids 117 Duplex DNA Is Unwound, and Daughter Strands
Are Formed at the DNA Replication Fork 145
A Nucleic Acid Strand Is a Linear Polymer
with End-to-End Directionalit y 117 Several Proteins Participate in DNA Replication 147

Native DNA Is a Double Helix of Complementary DNA Replication Occurs Bidirectionally from Each Origin 149
Antipara llel St ra nds 118
4.6 DNA Repair and Recombination 151
DNA Can Undergo Reversible Strand Separation 120
DNA Polymerases Introduce Copying Errors
Torsional Stress in DNA Is Relieved by Enzymes 121 and Also Correct Them 151
Different Types of RNA Exhibit Various Conformations Chemical and Radiat ion Damage to DNA Can Lead
Related to Their Functions 122 to Mutations 151
High-Fidelity DNA Excision Repair Systems Recognize
4.2 Transcription of Protein-Coding Genes and Repair Damage 152
and Formation of Functional mRNA 124 Base Excision Repairs T-G Mismatches and Damaged Bases 153
A Template DNA Strand Is Transcribed into Mismatch Excision Repairs Other Mismatches
a Complementary RNA Chain by RNA Polymerase 124 and Small Insertions and Deletions 153
Organization of Genes Differs in Prokaryotic Nucleotide Excision Repairs Chemical Adducts
and Eukaryotic DNA 126 t hat Distort Norma l DNA Shape 154
Eukaryotic Precursor mRNAs Are Processed Two Systems Utilize Recombination to Repair
to Form Funct ional mRNAs 128 Double-Strand Breaks in DNA 155
Alternative RNA Splicing Increases the Number of Proteins Homologous Recombination Can Repair DNA Damage
Expressed from a Single Eukaryotic Gene 129 and Generate Genetic Diversity 156

4.7 Viruses: Parasites of the Cellular

4.3 The Decoding of mRNA by tRNAs 131
Genetic System 160
Messenger RNA Carries Information from DNA
Most Viral Host Ranges Are Narrow 160
in a Three-Letter Genetic Code 131
Viral Capsids Are Regular Arrays of One or a Few Types
The Folded St ructure of tRNA Promotes Its Decoding
of Protein 160
Functions 133
Viruses Can Be Cloned and Counted in Plaque Assays 160
Nonstandard Base Pai ring Often Occurs Between
Codons and Anticodons 134 Lytic Viral Growth Cycles Lead to Death of Host Cells 161
Amino Acids Become Activated When Covalently Viral DNA Is Integrated into the Host-Cell Genome
Linked to tRNAs 135 in Some Non lyt ic Vira l Growth Cycles 164

4.4 Stepwise Synthesis of Proteins

5 Molecular Genetic Techniques 171
on Ribosomes 136
Ribosomes are Protein-Synthesizing Machines 136 5.1 Genetic Analysis of Mutations to Identify
Methionyl-tRNA,Met Recognizes the AUG Start Codon 137 and Study Genes 172
EukaryoticTranslat ion Initiat ion Usually Occurs Recessive and Dominant Mut ant All eles Generally Have
at t he Fi rst AUG Closest t o the 5' End of an mRNA 137 Opposite Effects on Gene Function 172
During Chain Elongation Each Incoming Aminoacyl-tRNA Segregation of Mutations in Breeding Experiments Reveals
Moves Through Th ree Ribosomal Sites 140 Their Dominance or Recessivity 173

CONTENTS xvll
Conditional Mutations Can Be Used t o Study Essential Linkage Studies Can Map Disease Genes with a Resolution
Genes in Yeast 175 of About 1 Centimorgan 208
Recessive Lethal Mutations in Diploids Can Be Ident ified Further Analysis Is Needed to Locate a Disease Gene
by Inbreeding and Maintained in Heterozygotes 176 in Cloned DNA 209
Complementation Tests Determine Whether Different Many Inherited Diseases Result from Multiple
Recessive Mutations Are in the Same Gene 177 Genetic Defects 210
Double Mutants Are Useful in Assessing the Order
in Which Proteins Function 178 5.5 Inactivating the Function of Specific
Genetic Suppression and Synthetic Lethality Can Reveal Genes in Eukaryotes 212
Interacting or Redundant Proteins 179 Normal Yeast Genes Can Be Replaced with Mutant Alleles
Genes Can Be Ident ified by Their Map Position on the by Homologous Recombination 212
Chromosome 180 Transcription of Genes Ligated to a Regulated Promoter
Can Be Controlled Experimentally 213
5.2 DNA Cloning and Characterization 182 Specific Genes Can Be Permanently Inactivated
Restriction Enzymes and DNA Ligases Allow Insertion in the Germ Line of Mice 213
of DNA Fragments into Cloning Vectors 183 Somatic Cell Recombination Can Inactivate Genes
E. coli Plasmid Vectors Are Suitable for Cloning Isolated in Specific Tissues 214
DNA Fragments 184 Dominant-Negative Alleles Can Functionally Inhibit
eDNA Libraries Represent the Sequences of Protein-Coding Some Genes 215
Genes 185 RNA Interference Ca uses Gene Inactivation by Destroying
cDNAs Prepared by Reverse Transcription of Cellular the Corresponding mRNA 216
mRNAs Can Be Cloned to Generate eDNA Libraries 186
DNA Libraries Can Be Screened by Hybridization
to an Oligonucleotide Probe 188
6 Genes, Genomics, and Chromosomes 223
Yeast Genomic Libraries Can Be Constructed with Shuttle
6.1 Eukaryotic Gene Structure 225
Vectors and Screened by Functional Complementation 188
Most Eukaryotic Genes Contain lntrons and Produce
Gel Electrophoresis Allows Separation of Vector DNA
mRNAs Encoding Single Proteins 225
from Cloned Fragments 191
Simple and Complex Transcription Units Are Found
The Polymerase Chain Reaction Amplifies a Specific
in Eukaryotic Genomes 225
DNA Sequence f rom a Complex Mixture 192
Protein-Coding Genes May Be Solitary or Belong
Cloned DNA Molecules Are Sequenced Rapidly
to a Gene Family 227
by Methods Based on PCR 195
Heavily Used Gene Products Are Encoded by Multiple
Copies of Genes 229
5.3 Using Cloned DNA Fragments to Study
Gene Expression 198 Nonprotein-Coding Genes Encode Functional RNAs 230

Hybridization Techniques Permit Detection of Specific DNA

Fragments and mRNAs 198 6.2 Chromosomal Organization of Genes
and Noncoding DNA 231
DNA Microarrays Can Be Used to Evaluate the Expression
of Many Genes at One Time 199 Genomes of Many Organisms Contain Nonfunctional DNA 231

Cluster Analysis of Multiple Expression Experi ments Most Simple-Sequence DNAs Are Concentrated in Specific
Identifies Co-regulated Genes 200 Chromosomal Locations 232

E. coli Expression Systems Can Produce Large Quantities DNA Fingerprinting Depends on Differences in Length
of Proteins from Cloned Genes 201 of Simple-Sequence DNAs 233

Plasmid Expression Vectors Can Be Designed Unclassified Spacer DNA Occupies a Significant
for Use in Animal Cells 203 Portion of the Genome 233

5.4 Locating and Identifying Human 6.3 Transposable (Mobile) DNA Elements 234
Disease Genes 206 Movement of Mobile Elements Involves a DNA or an RNA
Intermediate 235
Monogenic Diseases Show One ofThree Patterns
of Inheritance 206 DNA Transposons Are Present in Prokaryotes
and Eukaryotes 236
DNA Polymorphisms Are Used as Markers for
Linkage-Mapping of Human Mutations 207 LTR Retrotransposons Behave Like Intracellular Retroviruses 238

xviii CONTE NTS

 Non-LTR Retrotransposons Transpose by a Distinct Interphase Polytene Chromosomes Arise
Mechanism 240 by DNA Amplification 269
Other Retroposed RNAs Are Found in Genomic DNA 243 Three Functional Elements Are Required for Replication
Mobile DNA Elements Have Significantly Influenced
and Stable Inheritance of Chromosomes 270
Evolution 243 Centromere Sequences Vary Greatly in Length
and Complexity 271
6.4 Organelle DNAs 245 Addition ofTelomeric Sequences byTelomerase Prevents
Shortening of Chromosomes 273
Mitochondria Contain Multiple mtDNA Molecules 245
mtDNA Is Inherited Cytoplasmically 246
The Size, Structure, and Coding Capacity of mtDNA Vary
Considerably Between Organisms 246
7 Transcriptional Control
Products of Mitochondrial Genes Are Not Exported 248
of Gene Expression 279
Mitochondria Evolved from a Single Endosymbiotic Event
7.1 Control of Gene Expression
Involving a Rickettsia-like Bacterium 249
in Bacteria 282
Mitochondrial Genetic Codes Differ from the Standard
Transcription Initiation by Bacterial RNA Polymerase
Nuclear Code 249
Requires Association with a Sigma Factor 282
Mutations in Mitochondrial DNA Cause Several Genetic
Initiation of lac Operon Transcription Can Be Repressed
Diseases in Humans 250
and Activated 282
Chloroplasts Contain Large DNAs Often Encoding More
Small Molecules Regulate Expression of Many
Than a Hundred Proteins 251
Bacterial Genes via DNA-Binding Repressors
and Activators 284
6.5 Genomics: Genome-wide Analysis
Transcription Initiation from Some Promoters
of Gene Structure and Expression 252 Requires Alternative Sigma Factors 285
Stored Sequences Suggest Functions of Newly Identified 54
Transcription by a -RNA Polymerase Is Controlled
Genes and Proteins 252 by Activators That Bind Far from the Promoter 285
Comparison of Related Sequences from Different Species Many Bacterial Responses Are Controlled
Can Give Clues to Evolutionary Relationships Among by Two-Component Regulatory Systems 285
Proteins 253
Control ofTranscription Elongation 286
Genes Can Be Identified Within Genomic
DNA Sequences 253
The Number of Protein-Coding Genes in an Organism's 7.2 Overview of Eukaryotic Gene Control 288
Genome Is Not Directly Related to Its Biological Regulatory Elements in Eukaryotic DNA Are Found
Complexity 254 Both Close to and Many Kilobases Away
from Transcription Start Sites 289
6.6 Structural Organization Three Eukaryotic RNA Polymerases Catalyze Formation
.. of Eukaryotic Chromosomes 256 of Different RNAs 290
Chromatin Exists in Extended and Condensed Forms 256 The Largest Subunit in RNA Polymerase II Has an Essential
Carboxyl-Terminal Repeat 293
Modifications of Histone Tails Control Chromatin
Condensation and Functton 258
Nonhistone Proteins Organize Long Chromatin Loops 263 7.3 RNA Polymerase II Promoters and
Additional Nonhistone Proteins Regulate Transcription General Transcription Factors 295
and Replication 265 RNA Polymerase II Initiates Transcription at DNA Sequences
Corresponding to the 5' Cap of mRNAs 295
6.7 Morphology and Functional Elements The TATA Box, Initiators, and CpG Islands Function
of Eukaryotic Chromosomes 266 as Promoters in Eukaryotic DNA 295
Chromosome Number, Size, and Shape at MPtaphase General Transcription Factors Position RNA Polymerdse II
Are Species-Specific 266 at Start Sites and Assist in Initiation 297
During Metaphase, Chromosomes Can Be Distinguished In Vivo Transcription Initiation by RNA Polymerase II
by Banding Patterns and Chromosome Painting 267 Requires Additional Proteins 301
Chromosome Painting and DNA Sequencing Elongation Factors Regulate the Initial Stages of
Reveal t he Evolution of Chromosomes 268 Transcription in the Promoter-Proximal Region 301

CONTENTS xlx
7.4 Regulatory Sequences in Protein-Coding Noncoding RNAs Direct Epigenetic Repression
Genes and the Proteins Through Which in Metazoans 331
They Function 302 Plants and Fission Yeast Use Short RNA-Directed
Methylation of Histones and DNA 333
Promoter-Proximal Elements Help Regulate
Eukaryotic Genes 302
7.8 Other Eukaryotic Transcription Systems 336
Distant Enhancers Often Stimulate Transcription
by RNA Polymerase II 303 Transcription Initiation by Poll and Pollllls Analogous
to That by Pol II 336
Most Eukaryotic Genes Are Regu lated by Multiple
Transcription-Control Elements 304 Mitochondrial and Chloroplast DNAs Are Transcribed
by Organelle-Specific RNA Polymerases 338
Footprinting and Gel-Shift Assays Detect Protein-DNA
Interactions 305
Activators Promote Transcription and Are Composed
of Distinct Functional Domains 305
8 Post-transcriptional Gene Control 345
Repressors Inhibit Transcription and Are the Functional 8.1 Processing of Eukaryotic Pre-mRNA 348
Converse of Activators 307
The 5 ' Cap Is Added to Nascent RNAs £hortly After
DNA-Binding Domains Can Be Classified into Numerous Transcription Initiation 348
Structural Types 308
A Diverse Set of Proteins with Conserved RNA-Binding
Structurally Diverse Activation and Repression Domains Domains Associate with Pre-mRNAs 349
Regulate Transcription 311
Splicing Occurs at Short, Conserved Sequences
Transcription Factor Interactions Increase Gene-Control in Pre-mRNAs via Two Transesterification Reactions 351
Options 312
During Splicing, snRNAs Base-Pair with Pre-mRNA 352
Multiprotein Complexes Form on Enhancers 314
Spliceosomes, Assembled from snRNPs and a Pre-mRNA,
Carry Out Splicing 353
7.5 Molecular Mechanisms of Transcription
Repression and Activation 315 Chain Elongation by RNA Polymerase Ills Coupled
to the Presence of RNA-Processing Factors 356
Formation of Heterochromatin Silences Gene Expression
at Telomeres, Near Centromeres, and in Other Regions 315 SR Proteins Contribute to Exon Definition in Long
Pre-mRNAs 356
Repressors Can Direct Histone Deacetylation
at Specific Genes 318 Self-Splicing Group II lntrons Provide Clues to
the Evolution of snRNAs 357
Activators Can Direct Histone Acetylation at Specific Genes 318
3' Cleavage and Polyadenylation of Pre-mRNAs
Chromatin-Remodeling Factors Help Activate or Repress Are Tightly Coupled 358
Transcription 319
Nuclear Exonucleases Degrade RNA That Is Processed
The Mediator Complex Forms a Molecular Bridge Between Out of Pre-mRNAs 359
Activation Domains and Pol II 320
The Yeast Two-Hybrid System 321 8.2 Regulation of Pre-mRNA Processing 360
Alternative Splicing Generates Transcripts with Different
7.6 Regulation ofTranscription-Factor Activity 323 Combinations of Exons 361
All Nuclear Receptors Share a Common Domain Structure 324
A Cascade of Regulated RNA Splicing Controls
Nuclear-Receptor Response Elements Contain Inverted Drosophila Sexual Differentiation 361
or Direct Repeats 324
Splicing Repressors and Activators Control Splicing
Hormone Binding to a Nuclear Receptor Regulates at Alternative Sites 362
Its Activity as a Transcription Factor 325
RNA Editing Alters the Sequences of Some Pre-mRNAs 364
Metazoans Regulate the Pol II Transition from Initiation
to Elongation 325 8.3 Transport of mRNA Across the Nuclear
Pol II Termination Is Also Regulated 326 Envelope 365
Macromolecu les Exit and Enter the Nucleus Through
7.7 Epigenetic Regulation of Transcription 327 Nuclear Pore Complexes 365
Epigenetic Repression by DNA Methylation 327 Pre-mRNAs in Spliceosomes Are Not Exported from
Histone Methylation at Other Specific Lysines Are Linked the Nucleus 367
to Epigenetic Mechanisms of Gene Repression 328 HIV Rev Protein Regulates the Transport of Unspliced
Epigenetic Control by Polycomb and Trithorax Complexes 330 Viral mRNAs 368

xx CONTENTS
8.4 Cytoplasmic Mechanisms Imaging Subcellular Details Often Requires That
of Post-transcriptional Control 370 the Samples Be Fixed, Sectioned, and Stained 408
Micro RNAs Repress Translation of Specific mRNAs 371 Fluorescence Microscopy Can Localize and Quantify
Specific Molecules in Live Cells 408
RNA Interference Induces Degradation of Precisely 2
Complementary mRNAs 373 Determination of Intracellular Ca and H Levels
with ion-Sensitive Fluorescent Dyes 409
Cytoplasmic Polyadenylation Promotes Translation
of Some mRNAs 374 Immunofluorescence Microscopy Can Detect Specific
Proteins in Fixed Cells 409
Degradation of mRNAs in the Cytoplasm Occurs
by Several Mechanisms 375 Tagging w ith Fluorescent Proteins Allows the Visualization
of Specific Proteins in Living Cells 411
Protein Synthesis Can Be Globally Regulated 376
Deconvolution and Confocal Microscopy Enhance
Sequence-Specific RNA-Binding Proteins Control Visualization ofThree-Dimensional Fluorescent Objects 411
Specific mRNA Translation 379
TIRF Microscopy Provides Exceptional Imaging in One
Surveillance Mechanisms Prevent Translation of Improperly Focal Plane 415
Processed mRNAs 380
FRAP Reveals the Dynamics of Cellular Components 415
Localization of mRNAs Permits Production of Proteins
at Specific Regions Within the Cytoplasm 380 FRET Measures Distance Between Chromophores 416
Super-Resolution Microscopy Can Localize Proteins
8.5 Processing of rRNA and tRNA 384 to Nanometer Accuracy 418

Pre-rRNA Genes Function as Nucleolar Organizers

and Are Similar in All Eukaryotes 384 9.3 Electron Microscopy: High-Resolution
Small Nucleolar RNAs Assist in Processing Pre-rRNAs 385 Imaging 419
Self-Splicing Group llntrons Were the First Examples Single Molecules or Structures Can Be Imaged After a Negative
of Catalytic RNA 389 Stain or Metal Shadowing 419
Pre-tRNAs Undergo Extensive Modification Cells and Tissues Are Cut into Thin Sections for Viewing
in the Nucleus 390 by Electron Microscopy 420
Nuclear Bodies Are Functionally Specialized lmmunoelectron Microscopy Localizes Proteins
Nuclear Domains 391 at the Ultrastructural Level 421
Cryoelectron Microscopy Allows Visualization
of Specimens Without Fixation or Staining 421
Part Ill Cell Structure and Function Scanning Electron Microscopy of Metal-Coated
Specimens Reveals Surface Features 423
9 Culturing, Visualizing,
and Perturbing Cells 397 9.4 Isolation and Characterization of Cell
Organelles 424
9.1 Growing Cells in Culture 398 Organelles of the Eukaryotic Cell 424
Culture of Animal Cells Requires Nutrient-Rich Media Disruption of Cells Releases Their Organelles and Other
and Special Solid Surfaces 398 Contents 427
Primary Cell Cultures and C€11 Strains Have a Finite Life Span 399 Centrifugation Can Separate Many Types of Organelles 427
Transformed Cells Can Grow Indefinitely in Culture 400 Organelle-Specific Antibodies Are Useful in Preparing
Flow Cytometry Separates Different Cell Types 400 Highly Purified Organelles 429
Growth of Cells in Two-Dimensional and Three-Dimensional Proteomics Reveals the Protein Composition of Organelles 430
Culture Mimics the In Vivo Environment 401
Hybrid Cells Called Hybridomas Produce Abundant 9.5 Perturbing Specific Cell Functions 430
Monoclonal Antibodies 402 Drugs Are Commonly Used in Cell Biology 430
Chemical Screens Can Identify New Specific Drugs 430
9.2 Light Microscopy: Exploring Cell Structure Small Interfering RNAs (siRNAs) Can Knock Down
and Visualizing Proteins Within Cells 404 Expression of Specific Proteins 432
The Resolution of the Light Microscope Is About 0.2 J..l.m 404 Genomic Screens Using siRNA in the Nematode C. elegans 434
Phase-Contrast and Differential-Interference-Contrast
Microscopy Visualize Unstained Living Cells 405 CLASSIC EXPERIMENT 9.1 Separating Organelles 441

CONTENTS xxi
10 Biomembrane Structure 443 11 Transmembrane Transport of Ions
and Small Molecules 473
10.1 The Lipid Bilayer: Composition
and Structural Organization 445 11.1 Overview ofTransmembrane Transport 474
Phospholipids Spontaneously Form Bilayers 445 Only Gases and Small Uncharged Molecules Cross
Phospholipid Bilayers Form a Sealed Com partment Membranes by Sim ple Diffusion 474
Surrounding an Internal Aqueous Space 446 Th ree Main Classes of Membrane Proteins Tran sport
Biomembranes Contain Three Principal Classes Molecules and Ions Across Biomem branes 475
of Lipids 448
Most Lipids and Many Proteins Are Latera lly Mobile 11.2 Facilitated Transport of Glucose and Water 477
in Biomembranes 450 Unipo rt Transport Is Faster and More Specific than
Lipid Composition Influences t he Physical Properties Simple Diffusion 477
of Membranes 452 The Low Km of the GLUT1 Uniporter Enables It to Transport
Lipid Composition Is Different in t he Exoplasmic Glucose into Most Mammalian Cells 478
and Cytosolic Leaflets 453 The Human Genome Encodes a Family'Of Sugar -
Cholesterol and Sphingolipids Cluster with Specific Transporting GLUT Prot eins 479
Proteins in Membrane Microdoma ins 454 Transport Proteins Can Be Studied Using Artificial
Cells Store Excess Lipids in Lipid Droplets 455 Membranes and Recombinant Cells 480
Osmotic Pressure Ca uses Water to Move Across
Membranes 480
1 0.2 Membrane Proteins: Structure
Aquaporins Increase t he Water Permeability of Cell
and Basic Functions 455
Membranes 481
Proteins Interact with Membranes in Three Different
Ways 456
11.3 AlP-Powered Pumps and the Intracellular
Most Transmembrane Prot eins Have Membrane-
Ionic Environment 483
Spanning a Helices 456
There are Four Main Classes of ATP-Powered Pumps 483
Multiple !3 Strands in Porins Form Membrane-Spanning
"Barrels" 460 ATP-Powered lon Pumps Generate and Maint ain Ionic
Gradients Across Cellular Membranes 485
Covalently Attached Lipids Anchor Some Protei ns
2
to Membranes 460 Muscle Relaxation Depends on Ca ATPases
That Pump Ca 2 from the Cytosol int o
All Transmembrane Proteins and Glycolipids Are
the Sarcoplasmic Reticulum 486
Asymmetrically Oriented in t he Bi layer 461
2
The Mechanism of Action of the Ca + Pump Is Known
Lipid-Binding Motifs Help Target Peripheral Proteins
in Detail 486
to the Membrane 462
Ca lmodu lin Regulat es the Plasma Membrane Pumps
Proteins Can Be Removed from Membranes by Detergents
That Cont rol Cytosolic Ca 2 + Concentrations 487
or High-Salt Solut ions 462
Na • / K ATPase Maint ains the Intracellular Na + and K+
Concentrations in Animal Cells 489
10.3 Phospholipids, Sphingolipids,
V-Ciass W ATPases Maintain the Acidity of Lysosomes
and Cholesterol: Synthesis and Vacuoles 490
and Intracellular Movement 464 ABC Proteins Export a Wide Variety of Drugs and Toxins
Fatty Acids Are Assembled from Two-Carbon Build ing from the Ce ll 49 1
Blocks by Several Important Enzymes 465
Certain ABC Proteins "Flip" Phospholipids and Ot her
Small Cytosolic Proteins Facilitate Movement of Fatty Acids 465 Lipid-Soluble Substrates from One Membrane
Fatty Acids Are Incorporated into Phospholipids Primarily Leaflet to the Other 492
on the ER Membrane 465 The ABC Cystic Fibrosis Transmembrane Regulator (CFTR)
Flippascs Move Phospholipids from One MembrdrH:~
Is a Chloride Channel, Not a Pump 494
Leaflet to the Opposite Leaflet 467
Cholesterol Is Synthesized by Enzymes in the Cytosol 11.4 Nongated lon Channels and the Resting
and ER Membrane 467 Membrane Potential 495
Cholesterol and Phospholipids Are Transported Between Selective Movement of Ions Creates a Transmem brane
Organell es by Several Mechanisms 468 Electric Gradient 495

Jodi CONTENTS
The Resting Membrane Potential in Animal Cells 12.2 Mitochondria and the Citric Acid Cycle 524
Depends Largely on the Outward Flow
Mitochondria Are Dynamic Organelles with Two
of K+ Ions Through Open K Channels 497
Structurally and Functionally Distinct Membranes 524
Ion Channels Are Selective for Certain Ions by Virtue
In the First Part of Stage II, Pyruvate Is Converted
of a Molecular "Selectivity Filter" 497
to Acetyl CoA and High-Energy Electrons 526
Patch Clamps Permit Measurement of lon Movements
In the Second Part of Stage II, the Citric Acid Cycle
Through Single Channels 499
Oxidizes the Acetyl Group in Acetyl CoA to C0 2
Novellon Channels Can Be Characterized by a Combination and Generates High-Energy Electrons 527
of Oocyte Expression and Patch Clamping 501
Transporters in the Inner Mitochondrial Membrane
Help Maintain Appropriate Cytosolic and Matrix
11.5 Cotransport by Symporters Concentrations of NAD and NADH 529
and Antiporters 502 Mitochondrial Oxidation of Fatty Acids Generates ATP 529
Na Entry into Mammalian Cells Is Thermodynamically Peroxisomal Oxidation of Fatty Acids Generates No ATP 531
Favored 502
Na "-Linked Symporters Enable Animal Cells to Import 12.3 The Electron Transport Chain and
Glucose and Amino Acids Against High Concentration
Generation of the Proton-Motive Force 532
Gradients 502
Oxidation of NADH and FADH 2 Releases a Significant
A Bacterial Na / Amino Acid Symporter Reveals How
Amount of Energy 532
Symport Works 504
Electron Transport in Mitochondria Is Coupled to Proton
A Na +-Linked Ca2 + Anti porter Regulates the Strength
Pumping 533
of Cardiac Muscle Contraction 504
Electrons Flow "Downhill" Through a Series of Electron
Several Cotransporters Regulate Cytosolic pH 505 Carriers 534
An Anion Anti porter Is Essential for Transport of C02
Four Large Multiprotein Complexes Couple Electron
by Red Blood Cells 506 Transport to Proton Pumping Across the Mitochondrial
Numerous Transport Proteins Enable Plant Vacuoles Inner Membrane 535
to Accumulate Metabolites and Ions 507
Reduction Potentials of Electron Carriers in the Electron
Transport Chain Favor Electron Flow from NADH to 0 2 539
11.6 Transcellular Transport 508 The Multiprotein Complexes of the Electron Transport
Multiple Transport Proteins Are Needed to Move Glucose Chain Assemble into Supercomplexes 540
and Amino Acids Across Epithelia 508 Reactive Oxygen Species (ROS) Are Toxic By-products
Simple Rehydration Therapy Depends on the Osmotic of Electron Transport That Can Damage Cells 541
Gradient Created by Absorption of Glucose and Na + 509 Experiments Using Purified Electron Transport Chain
Parietal Cells Acidify the Stomach Contents While Complexes Established the Stoichiometry of Proton
Maintaining a Neutral Cytosolic pH 509 Pumping 542
Bone Resorption Requires Coordinated Function The Proton-Motive Force in Mitochondria Is Due Largely
of a V-Ciass Proton Pump and a Specific to a Voltage Gradient Across the Inner Membrane 542
Chloride Channel Protein 510
12.4 Harnessing the Proton-Motive Force
CLASSiij C EXPERIIMENT 1,1.1 Stumbling upon to Synthesize ATP 544
Active Transport 515 The Mechanism of ATP Synthesis Is Shared Among
Bacteria, Mitochondria, and Chloroplasts 544
ATP Synthase Comprises F0 and F1 Multi protein
Complexes 546
12 Cellular Energetics 517 Rotation of the F1 -y Subunit, Driven by Proton
Movement Through F0, Powers ATP Synthesis 547
12.1 First Step of Harvesting Energy
Multiple Protons Must Pass Through ATP Synthase
from Glucose: Glycolysis 519 to Synthesize One ATP 549
During Glycolysis (Stage 1), Cytosolic Enzymes Convert F0 c Ring Rotation Is Driven by Protons Flowing
Glucose to Pyruvate 520 Through Transmembrane Channels 549
The Rate of Glycolysis Is Adjusted to Meet the Cell's ATP-ADP Exchange Across the Inner Mitochondrial
Need for ATP 520 Membrane Is Powered by the Proton-Motive
Glucose Is Fermented When Oxygen Is Scarce 522 Force 550

CONTENTS xxlii
Rate of Mitochondrial Oxidation Normally Depends A Hydrophobic N-Terminal Signal Sequence
on ADP Levels 551 Targets Nascent Secretory Proteins
Brown-Fat Mitochondria Use t he Proton-Motive Force to the ER 580
to Generate Heat 551 Cotranslational Translocation Is Initiated by
Two GTP-Hydrolyzing Proteins 582
12.5 Photosynthesis and Light-Absorbing
Passage of Growing Polypeptides Through the Translocon
Pigments 552 Is Driven by Translation 583
Thylakoid Membranes in Chloroplasts Are the Sites ATP Hydrolysis Powers Post-translational Translociltion
of Photosynthesis in Plants 553 of Some Secretory Proteins in Yeast 584
Three of the Four Stages in Photosynthesis Occur
Only During Illumination 553 13.2 Insertion of Membrane Proteins
Each Photon of Light Has a Defined Amount of Energy 555 into the ER 587
Photosystems Comprise a Reaction Center Several Topological Classes of Integral Membrane
and Associated Light-Harvesting Complexes 555 Proteins Are Synthesized on the ER 587
Photoelectron Transport from Energized Reaction-Center Internal Stop-Transfer and Signal-Anchor Sequences
Chlorophyll a Produces a Charge Separation 556 Determine Topology of Single-PaSS'Proteins 588
Internal Antenna and Light-Harvesting Complexes Multipass Proteins Have Multiple Internal Topogenic
Increase the Efficiency of Photosynthesis 557 Sequences 591
A Phospholipid Anchor Tethers Some Cell-Surface Proteins
12.6 Molecular Analysis of Photosystems 559 to the Membrane 592
The Single Photosystem of Purple Bacteria Generates The Topology of a Membrane Protein Often Can Be
a Proton-Motive Force but No 0 2 559 Deduced from Its Sequence 592
Chloroplasts Contain Two Functionally and Spatially
Distinct Photosystems 561 13.3 Protein Modifications, Folding,
Linear Electron Flow Through Both Plant Photosystems, and Quality Control in the ER 594
PSI I and PSI, Generates a Proton-Motive Force, 0 2,
A Preformed N-Linked Oligosaccharide Is Added
and NADPH 561
to Many Proteins in the Rough ER 595
An Oxygen-Evolving Complex Is Located on the Luminal
Oligosaccharide Side Chains May Promote Folding
Surface of the PSII Reaction Center 562
and Stability of Glycoproteins 596
Multiple Mechanisms Protect Cells Against Damage
Disulfide Bonds Are Formed and Rearranged by Proteins
from Reactive Oxygen Species During
in the ER Lumen 596
Photoelectron Transport 563
Chaperones and Other ER Proteins Facilitate Folding
Cyclic Electron Flow Through PSI Generates
a Proton-Motive Force but No NADPH or 0 2 564
and Assembly of Proteins 598
Improperly Folded Proteins in the ER Induce Expression
Relative Activities of Photosystems I and II Are Regulated 565
of Protein-Folding Catalysts 599
12.7 C02 Metabolism During Photosynthesis 567 Unassembled or Misfolded Proteins in the ER Are Often
Transported to the Cytosol for Degradation 600
Rubisco Fixes C02 in the Chloroplast Stroma 567
Synthesis of Sucrose Using Fixed C0 2 Is Completed
13.4 Targeting of Proteins to Mitochondria
in the Cytosol 567
and Chloroplasts 601
Light and Rubisco Activase Stimulate C02 Fixation 569
Amphipathic N-Terminal Signal Sequences Direct Proteins
Photorespiration Competes with Carbon Fixation to the Mitochondrial Matrix 603
and Is Reduced in C4 Plants 569
Mitochondrial Protein Import Requires Outer-Membrane
Receptors and Translocons in Both Membranes 603
13 Moving Proteins into Membranes Studies with Chimeric Proteins Demonstrate Important
and Organelles 577 Features of Mitochondrial Import 605
Three Energy Inputs Are Needed to Import Proteins
13.1 Targeting Proteins to and Across into Mitochondria 606
the ER Membrane 579 Multiple Signals and Pathways Target Proteins
Pulse-Labeling Experiments with Purified ER Membranes to Submitochondrial Compartments 606
Demonstrated That Secreted Proteins Cross Targeting of Chloroplast Stromal Proteins Is Similar
the ER Membrane 579 to Import of Mitochondrial Matrix Proteins 610

xxiv CONTENTS
 Proteins Are Targeted to Thylakoids by Mechanisms Anterograde Transport Through the Golgi Occurs
Related to Translocation Across the Bacterial by Cisternal Maturation 643
Cytoplasmic Membrane 610
14.4 Later Stages of the Secretory Pathway 646
13.5 Targeting of Peroxisomal Proteins 612 Vesicles Coated with Clathrin and/ or Adapter Proteins
Cytosolic Receptor Targets Proteins with an SKL Sequence Mediate Transport from the trans-Golgi 646
at the (-Terminus into the Peroxisomal Matrix 612
Dynamin Is Required for Pinching Off of Clathrin Vesicles 647
Peroxisomal Membrane and Matrix Proteins
Mannose 6-Phosphate Residues Target )oluble Proteins
Are Incorporated by Different Pathways 613 648
to Lysosomes

13.6 Transport into and out of the Nucleus 615 Study of Lysosomal Storage Diseases Revealed Key
Components of the Lysosomal Sorting Pathway 649
Large and Small Molecules Enter and Leave the Nucleus
615 Protein Aggregation in the trans-Golgi May Function
via Nuclear Pore Complexes
in Sorting Proteins to Regulated Secretory Vesicles 651
- Nuclear Transport Receptors Escort Proteins Containing
Nuclear-Localization Signals into the Nucleus 617 Some Proteins Undergo Proteolytic Processing After
Leaving the trans-Golgi 651
A Second Type of Nuclear Transport Receptors Escort
Several Pathways Sort Membrane Proteins to the Apical
Proteins Containing Nuclear-Export Signals
out of the Nucleus 619 or Basolateral Region of Polarized Cells 652

Most mRNAs Are Exported from the Nucleus

619
14.5 Receptor-Mediated Endocytosis 654
by a Ran-Independent Mechanism
Cells Take Up Lipids from the Blood in the Form
of Large, Well-Defined Lipoprotein Complexes 656
14 Vesicular Traffic, Secretion, Receptors for Low-Density Lipoprotein and Other
and Endocytosis 627 Ligands Contain Sorting Signals That Target
Them for Endocytosis 657
14.1 Techniques for Studying the Secretory The Acidic pH of Late Endosomes Causes Most
Pathway 629 Receptor-Ligand Complexes to Dissociate 658
Transport of a Protein Through the Secretory Pathway The Endocytic Pathway Delivers Iron to Cells
Can Be Assayed in Living Cells 629 Without Dissociation of the Receptor-Transferrin
Complex in Endosomes 659
Yeast Mutants Define Major Stages and Many
Components in Vesicular Transport 632
14.6 Directing Membrane Proteins and
Cell-Free Transport Assays Allow Dissection of Individual
Steps in Vesicular Transport 633
Cytosolic Materials to the Lysosome 661
Multivesicular Endosomes Segregate Membrane Proteins
14.2 Molecular Mechanisms of Vesicle Destined for the Lysosomal Membrane from Proteins
Destined for Lysosomal Degradation 661
Budding and Fusion 634
Retroviruses Bud from the Plasma Membrane by a Process
Assembly of a Protein Coat Drives Vesicle Formation
Similar to Formation of Multivesicular Endosomes 663
and Selection of Cargo Molecules 634
The Autophagic Pathway Delivers Cytosolic Proteins
A Conserved Set of GTPase Switch Proteins Controls
or Entire Organelles to Lysosomes 664
Assembly of Different Vesicle Coats 635
Targeting Sequences on Cargo Proteins Make Specific
CLASSIC EXPERIMENT 14.1 Following a Protein
Molecular Contacts with Coat Proteins 636
Out of the Cell 671
Rab GTPases Control Docking of Vesicles on Target
Membranes 638
Paired Sets of SNARE Proteins Mediate Fusion of
Vesicles with Target Membranes 639 15 Signal Transduction and G Protein-
Dissociation of SNARE Complexes After Membrane Coupled Receptors 673
Fu~ion Is Driven by ATP Hydrolysis 639
15.1 Signal Transduction: From Extracellular
14.3 Early Stages of the Secretory Pathway 640 Signal to Cellular Response 675
COP II Vesicles Mediate Transport from the ER to the Golgi 640 Signaling Molecules Can Act Locally or at a Distance 675
COP I Vesicles Mediate Retrograde Transport Within Binding of Signaling Molecules Activates Receptors
the Golgi and from the Golgi to the ER 642 on Target Cells 676

CONTENTS XXV
Protein Kinases and Phosphatases Are Employed in Structural Studies Established How Gu, · GTP Binds
Virtually All Signaling Pathways 677 to and Activates Adenylyl Cyclase 700
GTP-Binding Proteins Are Frequently Used in Signal cAMP Activates Protein Kinase A by Releasing Inhibitory
Transduction as On/Off Switches 678 Subunits 701
Intracellular "Second Messengers" Transmit and Amplify Glycogen Metabolism Is Regulated by Hormone-Induced
Signals from Many Receptors 679 Activation of Protein Kinase A 701
cAMP-Mediated Activation of Protein Kinase A Produces
1 5.2 Studying Cell-Surface Receptors Diverse Responses in Different Cell Types 702
and Signal Transduction Proteins 681 Signal Amplification Occurs in the cAMP-Protein Kinase
The Dissociation Constant Is a Measure of the Affinity A Pathway 703
of a Receptor for Its Ligand 681 CREB Links cAMP and Protein Kinase A to Activation
Binding Assays Are Used to Detect Receptors and of Gene Transcription 703
Determine Their Affinity and Specificity for Ligands 682 Anchoring Proteins Localize Effects of cAMP to Specific
Maximal Cellular Response to a Signaling Molecule Regions of the Cell 704
Usually Does Not Require Activation of All Receptors 683 Multiple Mechanisms Down-Regulate ~gnaling
Sensitivity of a Cell to External Signals Is Determined from the GPCR/cAMP/PKA Pathway 705
by the Number of Surface Receptors and Their
Affinity for Ligand 684 15.6 G Protein-Coupled Receptors That Trigger
Receptors Can Be Purified by Affinity Techniques 685 Elevations in Cytosolic Ca2+ 707
lmmunoprecipitation Assays and Affinity Techniques Activated Phospholipase C Generates Two Key Second Messengers
Can Be Used to Study the Activity of Signal Derived from the Membrane Lipid Phosphatidylinositol 708
Transduction Proteins 685 The Ca 2• -Calmodulin Complex Mediates Many
Cellular Responses to External Signals 711
1 5.3 G Protein-Coupled Receptors: Signal-Induced Relaxation of Vascular Smooth Muscle
Structure and Mechanism 687 Is Mediated by a Ca2+ -Nitric Oxide-cGMP-Activated
Protein Kinase G Pathway 711
All G Protein- Coupled Receptors Share the Same
2
Basic Structure 687 Integration of Ca and cAMP Second Messengers
Regulates Glycogenolysis 711
Ligand-Activated G Protein-Coupled Receptors Catalyze
Exchange of GTP for GDP on the a Subunit
of a Trimeric G Protein 689 Cli ,SSIC EXPE:RI MENT 15.1 The Infancy
Different G Proteins Are Activated by Different GPCRs of Signal Transduction-GTP Stimulation
and In Turn Regulate Different Effector Proteins 691 of cAMP Synthesis 719

15.4 G Protein-Coupled Receptors

That Regulate lon Channels 693 16 Signaling Pathways That Control
Acetylcholine Receptors in the Heart Muscle Activate Gene Expression 721
a G Protein That Opens K Channels 693
Light Activates G Protein-Coupled Rhodopsins in Rod 16.1 Receptors That Activate Protein
Cells of the Eye 694 Tyrosine Kinases 723
Activation of Rhodopsin by Light Leads to Closing Numerous Factors Regulating Cell Division
of cGMP-Gated Cation Channels 695 and Metabolism Are Ligands for Receptor
Signal Amplification Makes the Rhodopsin Signal Tyrosine Kinases 723
Transduction Pathway Exquisitely Sensitive 696 Binding of Ligand Promotes Dimerization of an RTK
Rapid Termination of the Rhodopsin Signal Transduction and Leads to Activation of Its Intrinsic Kinase 724
Pathway Is Essential for Acute Vision 696 Homo- and Hetero-oligomers of Epidermal Growth
Rod Cells Adapt to Varying Levels of Ambient Light Factor Receptors Bind Members of the Epidermal
by Intracellular Trafficking of Arrestin and Transducin 698 Growth Factor Superfamily 726
Cytokines Influence Development of Many Cell Types 728
15.5 G Protein-Coupled Receptors That Binding of a Cytokine to Its Receptor Activates
Activate or Inhibit Adenylyl Cyclase 699 a Tightly Bound JAK Protein Tyrosine Kinase 728
Adenylyl Cyclase Is Stimulated and Inhibited by Different Phosphotyrosine Residues Are Binding Surfaces
Receptor-Ligand Complexes 699 for Multiple Proteins with Conserved Domains 730

xxvi CONTENTS
SH2 Domains in Action: JAK Kinases Activate STAT Hedgehog Signaling Relieves Repression ofTarget Genes 753
Transcription Factors 730 Hedgehog Signaling in Vertebrates Involves Primary Cilia 755
Multiple Mechanisms Down-Regulate Signaling Degradation of an Inhibitor Protein Activates the NF-KB
from RTKs and Cytokine Receptors 731 Transcription Factor 757
Polyubiquitin Chains Serve as Scaffolds linking Receptors
16.2 The Ras/MAP Kinase Pathway 734 to Downstream Proteins in the NF-KB Pathway 759
Ras, a GTPase Switch Protein, Operates Downstream
of Most RTKs and Cytokine Receptors 735 16.6 Signaling Pathways Controlled by Protein
Genetic Studies in Drosophila Identified Key Cleavage: Notch/Delta, SREBP 760
Signal-Transducing Proteins in the Ras/ MAP
On Binding Delta, the Notch Receptor Is Cleaved,
Kinase Pathway 735
Releasing a Component Transcription Factor 760
Receptor Tyrosine Kinases and JAK Kinases Are linked
Matrix Metalloproteases Catalyze Cleavage of Many
to Ras by Adapter Proteins 737
Signaling Proteins from the Cell Surface 761
Binding of Sos to Inactive Ras Causes a Conformational
Inappropriate Cleavage of Amyloid Precursor Protein
Change That Triggers an Exchange of GTP for GDP 738
Can Lead to Alzheimer's Disease 762
Signals Pass from Activated Ras to a Cascade of Protein
Regulated Intra membrane Proteolysis of SREBP Releases
Kinases, Ending with MAP Kinase 738
a Transcription Factor That Acts to Maintain
Phosphorylation of MAP Kinase Results in a Conformational Phospholipid and Cholesterol Levels 762
Change That Enhances Its Catalytic Activity
and Promotes Kinase Dimerization 740
16.7 Integration of Cellular Responses
MAP Kinase Regulates the Activity of Many Transcription to Multiple Signaling Pathways 765
Factors Controlling Early Response Genes 741
Insulin and Glucagon Work Together to Maintain
G Protein-Coupled Receptors Transmit Signals a Stable Blood Glucose Level 765
to MAP Kinase in Yeast Mating Pathways 742
Multiple Signal Transduction Pathways Interact
Scaffold Proteins Separate Multiple MAP Kinase to Regulate Adipocyte Differentiation Through
Pathways in Eukaryotic Cells 744 PPAR-y, the Master Transcriptional Regulator 767

16.3 Phosphoinositide Signaling Pathways 745

Phospholipase C-y Is Activated by Some RTKs 17 Cell Organization and Movement 1:
and Cytokine Receptors 745
Microfilaments 773
Recruitment of Pl-3 Kinase to Activated Receptors Leads
to Synthesis ofThree Phosphorylated
17.1 Microfilaments and Actin Structures 776
Phosphatidylinositols 745
Actin Is Ancient, Abundant, and Highly Conserved 776
Accumulation of PI 3-Phosphates in the Plasma
Membrane Leads to Activation of Several Kinases 746 G-Actin Monomers Assemble into Long, Helical
F-Actin Polymers 777
Activated Protein Kinase B Induces Many Cellular Responses 747
F-Actin Has Structural and Functional Polarity 778
The Pl-3 Kinase Pathway Is Negatively Regulated by PTEN
Phosphatase 747
17.2 Dynamics of Actin Filaments 779
16.4 Receptor Serine Kinases That Activate Actin Polymerization in Vitro Proceeds in Three Steps 779

Smads 748 Actin Filaments Grow Faster at (+ )Ends Than at (- ) Ends 779

Three Separate TGF-~ Receptor Proteins Participate Actin FilamentTreadmilling Is Accelerated by Profilin
in Binding TGF-13 and Activating Signal Transduction 748 and Cofilin 782

Activated TGF-13 Receptors Phosphorylate Smad Thymosin-134 Provides a Reservoir of Actin

Transcription Factors 749 for Polymerization 782

Negative Feedback Loops Regulate TGF-13/Smad Signaling 751 Capping Proteins Block Assembly and Disassembly
at Actin Filament Ends 783
16.5 Signaling Pathways Controlled by
Ubiquitination: Wnt, Hedgehog, 17.3 Mechanisms of Actin Filament Assembly 784
and NF-KB 752 Formins Assemble Unbranched Filaments 784
Wnt Signaling Triggers Release of a Transcription Factor The Arp2/ 3 Complex Nucleates Branched Filament
from a Cytosolic Protein Complex 752 Assembly 785

CONTENTS XXVII
Intracellular Movements Can Be Powered by Actin
Polymerization 787 18 Cell Organization and Movement II:
Microfilaments Function in Endocytosis 788 Microtubules and Intermediate
Toxins That Perturb the Pool of Actin Monomers Filaments 821
Are Useful for Studying Actin Dynamics 789
----
18.1 Microtubule Structure and Organization 822
17.4 Organization of Actin-Based Cellular Microtubule Walls Are Polarized Structures Built
Structures 790 from al3 Tubulin Dimers 822
Cross-Linking Proteins Organize Actin Filaments Microtubules Are Assembled from MTOCs
into Bundles or Networks 790 to Generate Diverse Organizations 824

Adaptor Proteins Link Actin Filaments to Membranes 791

18.2 Microtubule Dynamics 827
17.5 Myosins: Actin-Based Motor Proteins 793 Individual Microtubules Exhibit Dynamic Instability 827
Myosins Have Head, Neck, and Tai l Domains with Distinct Localized Assembly and "Search-and-Capture"
Functions 794 Help Organize Microt ubules 829
Myosins Make Up a Large Family of Mechanochemical Drugs Affecting Tubulin Polymerization Are Useful
Motor Proteins 796 Experimentally and in Treatment of Diseases 829
Conformational Changes in the Myosin Head Couple ATP
Hydrolysis to Movement 797 18.3 Regulation of Microtubule Structure
Myosin Heads Take Discrete Steps Along Actin Fi laments 799 and Dynamics 830
Myosin V Walks Hand over Hand down an Actin Filament 799 Microtubules Are Stabilized by Side-Binding Proteins 830
+ TIPs Regulate the Properties and Functions of the
17.6 Myosin-Powered Movements 801 M icrotubule ( + ) End 831

Myosin Thick Filaments and Actin Th in Filaments Other End-Binding Proteins Regu late Microtubule
in Skeletal Muscle Slide Past One Another During Disassembly 831
Contraction 801
Skeletal Muscle Is Structured by Stabilizing 18.4 Kinesins and Dyneins: Microtubule-Based
and Scaffolding Proteins 802 Motor Proteins 833
Contraction of Skeletal Muscle Is Regulated Organelles in Axons Are Transported Along Microtubules
by Ca 2 + and Actin-Binding Proteins 802 in Both Directions 833
Actin and Myosin II Form Contractile Bundles Kinesin-1 Powers Anterograde Transport
in Nonmuscle Cells 804 ofVesicles Down AxonsToward the( + ) End
of Microtubules 834
Myosin-Dependent Mechanisms Regu late Contraction
in Smooth Muscle and Nonmuscle Cells 804 Kines ins Form a Large Protein Family with Diverse
Functions 836
Myosin-V-Bound Vesicles Are Carried Along Actin
Filaments 805 Kinesin-1 Is a Highly Processive Motor 837
Dynein Motors Transport Organelles Toward the (- ) End
17.7 Cell Migration: Mechanism, Signaling, of Microtubules 837
and Chemotaxis 808 Kinesins and Dyneins Cooperate in the Transport
Cell Migration Coordinates Force Generation with of Organelles Throughout the Cell 841
Cell Adhesion and Membrane Recycling 808 Tubulin Modifications Distinguish Different Microtubules
The Small GTP-Binding Proteins Cdc42, Rae, and Their Accessibilit y to Motors 842
and Rho Control Actin Organization 810
Cell Migration Involves the Coordinate Regulation 18.5 Cilia and Flagella : Microtubule-Based
of Cdc42, Rae, and Rho 812 Surface Structures 844
Migrating Cells Are Steered by Chemotactic Molecules 813 Eukaryotic Cilia and Flagella Contain Long Doublet
Chemotactic Gradients Induce Altered Phosphoinositide Microtubules Bridged by Dynein Motors 845
Levels Between the Front and Back of a Cell 814 Ciliary and Flagellar Beating Are Produced by Controlled
Sliding of Outer Doublet Microtubules 845
CL• SS C EXPERIMENT 17.1 Looking at Muscle lntraflagellar Transport Moves Material up and
Contraction 819 down Cilia and Flagella 846

xxviii CONT EN TS
Primary Cilia Are Sensory Organelles on Interphase Cells 847 Cyclin-Dependent Kinases Control the Eukaryotic Cell Cycle 876
Defects in Primary Cilia Underlie Many Diseases 848 Several Key Principles Govern the Cell Cycle 876

18.6 Mitosis 849 19.2 Model Organisms and Methods

Centrosomes Duplicate Early in the Cell Cycle to Study the Cell Cycle 877
in Preparation for Mitosis 849 Budding and Fission Yeast Are Powerful Systems
Mitosis Can Be Divided into Six Phases 849 for Genetic Analysis of t he Cell Cycle 877

The Mit otic Spindle Contains Three Classes of M1crotubules 851 Frog Oocytes and Early Embryos Facilitate Biochemical
Characterization of the Cell Cycle Engine 878
Microtubule Dynamics Increase Dramatically in Mitosis 851
Fruit Flies Reveal the Interplay Between Development
Mitotic Asters Are Pushed Apart by Ki nesin-5 and the Cell Cycle 880
and Oriented by Dynein 852
The Study ofTissue Culture Cells Uncovers Cell Cycle
Chromosomes Are Captured and Oriented During Regulation in Mammals 881
Prometaphase 852
Researchers Use Multiple Tools to Study the Cell Cycle 881
Duplicated Chromosomes Are Aligned by Motors
and Microtubule Dynamics 855
19.3 Regulation of CDK Activity 883
The Chromosom al Passenger Complex Regulates
Cyclin-Dependent Kinases Are Small Protein Kinases
Microtubule Attachment at Kinetochores 855
That Require a Regulatory Cyclin Subunit
Anaphase A Moves Chromosomes to Poles for Their Activit y 884
by Microtubule Shortening 856
Cyclins Determine the Activity of CDKs 885
Anaphase B Separates Poles by the Combined Action
Cyclin Levels Are Primarily Regulated by Protein
of Kinesins and Dynein 857
Degradation 887
Additional Mechanisms Cont ribute to Spindle Formation 858
CDKs Are Regulated by Activating and Inhibitory
Cytokinesis Split s t he Duplicated Cell in Two 858 Phosphorylation 888
Plant Cells Reorganize Their Microtubules and Build CDK Inhibitors Control Cyclin-CDK Activity 888
a New Cell Wall in Mitosis 859
Special CDK Alleles Led to the Discovery of CDK Functions 889

18.7 Intermediate Filaments 860 19.4 Commitment to the Cell Cycle

Intermediate Filaments Are Assembled from and DNA Replication 890
Subunit Dimers 861
Cells Are Irreversibly Committed to Cell Division at a Cell
Intermediate Filament Proteins Are Expressed Cycle Point Called START 890
in a Tissue-Specific Manner 862
The E2F Transcription Factor and Its Regulator Rb Control
Intermediate Filaments Are Dynamic 863 the G1-S Phase Transition in Metazoans 891
Defects in Lamins and Keratins Cause Many Diseases 863 Extracellular Signals Govern Ce ll Cycle Entry 892
Degradation of an S Phase CDK Inhibitor Triggers
18.8 Coordination and Cooperation Between DNA Replication 892
Cytoskeletal Elements 865
Replication at Each Origin Is Initiated On ce and Only
Intermediate Filament -Associated Proteins Contribute Once During the Cell Cycle 894
,
to Cellular Organization 865
Duplicated DNA Strands Become Linked During Replication 896
Microfilaments and Microtubules Cooperate to
Transport Melanosomes 865 19.5 Entry into Mitosis 897
Cdc42 Coordinates Microtubules and Microfilaments Precipitous Activation of Mitotic CDKs Initiates Mitosis 897
During Cell Migration 866
Mitotic CDKs Promote Nuclear Envelope Breakdown 898
Advancement of Neu ral Growth Cones Is Coordinated
by Microfilaments and Microtubules 866 Mitotic CDKs Promote Mitotic Spindle Formation 899
Chromosome Condensation Filcilitates Chromosome
Segregation 901
19 The Eukaryotic Cell Cycle 873
19.6 Completion of Mitosis: Chromosome
19.1 Overview of the Cell Cycle and Its Control 875 Segregation and Exit from Mitosis 903
The Cell Cycle Is an Ordered Series of Events Lead ing Separase-Mediated Cleavage of Cohesins Initiates
to Cell Replication 875 Chromosome Segregation 903

CONTENTS xxix
The APC/C Activates Separase Through Securin 20.2 Cell-Cell and Ceii-ECM Junctions
Ubiquitinylation 903 and Their Adhesion Molecules 933
Mitotic CDK Inactivation Triggers Exit from Mitosis 904 Epithelial Cells Have Distinct Apica l, Lateral,
Cytokinesis Creates Two Daughter Cells 905 and Basal Surfaces 933
Three Types of Junctions Mediate Many Cell-Cell
and Ceii-ECM Interactions 934
19.7 Surveillance Mechanisms in Cell
Cycle Regulation 906 Cadherins Mediate Cell-Cell Adhesions in Adherens
Junctions and Desmosomes 935
Checkpoint Pathways Establish Dependencies
and Prevent Errors in the Cell Cycle 907 lntegrins Mediate Ceii-ECM Adhesions, Including
Those in Epithelial Cell Hemidesmosomes 939
The Growth Checkpoint Pat hway Ensures That Cells
Only Enter the Cell Cycle After Sufficient Tight Junctions Seal Off Body Cavities and Restrict
Macromolecule Biosynthesis 907 Diffusion of Membrane Components 940

The DNA Damage Response Halts Cell Cycle Progression Gap Junctions Composed of Connexi ns Allow Small
When DNA Is Compromised 908 Molecules to Pass Directly Between Adjacent Cells 943

The Spindle Assembly Checkpoint Pathway Prevents

Chromosome Segregation Until Chromosomes 20.3 The Extracellular Matrix 1: The Basal Lamina 945
Are Accurately Attached to the Mitotic Spindle 910
The Basal Lamina Provides a Foundation for Assembly
The Spindle Position Checkpoint Pathway Ensures of Cells into Tissues 946
That the Nucleus Is Accurately Partitioned
Lamin in, a Multi-ad hesive Matrix Protein, Helps Cross-link
Between Two Daughter Cells 912
Components of the Basal Lamina 947
Sheet-Forming Type IV Collagen Is a Major Structural
19.8 Meiosis: A Special Type of Cell Division 913 Component of the Basal Lamina 947
Extracellular and Intracellular Cues Regulate Entry Perlecan, a Proteoglycan, Cross-links Components
into Meiosis 913 of the Basal Lamina and Cell-Surface Receptors 950
Several Key Features Distinguish Meiosis
from Mitosis 915 20.4 The Extracellular Matrix II: Connective Tissue 951
Recombination and a Meiosis-Specific Cohesin Fibrillar Collagens Are the Major Fibrous Proteins
Subunit Are Necessary for the Specialized in the ECM of Connective Tissues 951
Chromosome Segregation in Meiosis I 915
Fibrillar Collagen Is Secreted and Assembled into Fibrils
Co-orienting Sister Kinetochores Is Critical for Meiosis I Outside the Cell 952
Chromosome Segregation 918
Type I and II Collagens Associate with Nonfibrillar
DNA Replication Is Inhibited Between Collagens to Form Diverse Structures 953
the Two Meiotic Divisions 918
Proteoglycans and Their Constituent GAGs Play Diverse
Roles in the ECM 954
CLASSIC EXPERIMENT 19.1 Cell Biology Emerging
Hyaluronan Resists Compression, Facilit ates Cell Migrat ion,
from the Sea: The Discovery of Cyclins 923
and Gives Cartilage Its Gel-like Properties 956
Fibronectins Interconnect Cells and Matrix, Influencing
Cell Shape, Differentiation, and Movement 957
Part IV Cell Growth and Development
Elastic Fibers Permit Many Tissues to Undergo Repeated
Stretching and Recoiling 959

20 Integrating Cells Into Tissues 925 Metalloproteases Remodel and Degrade the Extracellular
Matrix 960

20.1 Cell-Cell and Cell-Matrix Adhesion:

An Overview 927 20.5 Adhesive Interactions in Motile
Cell-Adhesion MoiPcules Bind to One Another and Nonmotile Cells 961
and to Intracellular Proteins 927 lntegrins Relay Signals between Cells and Their
The Extracellular Matrix Participates in Adhesion, Three-Dimensional Environment 961
Signaling, and Other Functions 929 Regulation of lntegrin-Mediated Adhesion and Signaling
The Evolution of Multifaceted Adhesion Molecules Controls Cell Movement 962
Made Possible the Evolution of Diverse Connections Between the ECM and Cytoskeleton
Animal Tissues 932 Are Defective in Muscular Dystrophy 964

XXX CONTENTS
lgCAMs Mediate Cell-Cell Adhesion in Neuronal The Par Proteins and Other Polarity Complexes
and Other Tissues 965 Are Involved in Epithelial-Cell Polarity 1001
Leukocyte Movement into Tissues Is Orchestrated by a The Planar Cell Polarity Pathway Orients Cells within
Precisely Timed Sequence of Adhesive Interactions 965 an Epithelium 1002
The Par Proteins Are Also Involved in Asymmetric Cell
20.6 Plant Tissues 967 Division of Stem Cells 1004
The Plant Cell Wallis a Laminate of Cellulose Fibrils
in a Matrix of Glycoproteins 968 21.4 Cell Death and Its Regulation 1006
Loosening of the Cell Wall Permits Plant Cell Growth 969 Programmed Cell Death Occurs Through Apoptosis 1007
Plasmodesmata Directly Connect the Cytosols Evolutionarily Conserved Proteins Participate
of Adjacent Cells in Higher Plants 969 in the Apoptotic Pathway 1007
Only a Few Adhesive Molecules Have Been Identified Caspases Amplify the Initial Apoptotic Signal
in Plants 970 and Destroy Key Cellular Proteins 1009
Neurotrophins Promote Survival of Neurons 1010
Mitochondria Play a Central Role in Regulation
21 Stem Cells, Cell Asymmetry, of Apoptosis in Vertebrate Cells 1011
and Cell Death 977 The Pro-apoptotic Proteins Bax and Bak Form Pores
in the Outer Mitochondrial Membrane 1013
21.1 Early Metazoan Development
Release of Cytochrome c and SMAC/ DIABLO Proteins
and Embryonic Stem Cells 979 from Mitochondria Leads to Formation
Fertilization Unifies the Genome 979 of the Apoptosome and Caspase Activation 1013

Cleavage of the Mammalian Embryo Leads to the First Trophic Factors Induce Inactivation of Bad,
Differentiation Events 979 a Pro-apoptotic BH3-0nly Protein 1013
The Inner Cell Mass Is the Source of Embryonic Vertebrate Apoptosis Is Regulated by BH3-0nly
Stem (ES) Cells 981 Pro-Apoptotic Proteins That Are Activated
by Environmental Stresses 1014
Multiple Factors Control the Pluripotency of ES Cells 983
Tumor Necrosis Factor and Related Death Signals
Animal Cloning Shows That Differentiation Can Be Reversed 984
Promote Cell Murder by Activating Caspases 1015
Somatic Cells Can Generate Induced Pluripotent
Stem (iPS) Cells 984

21.2 Stem Cells and Niches in Multicellular

22 Nerve Cells 1019
Organisms 986 22.1 Neurons and Glia: Building Blocks
Stem Cells Give Rise to Both Stem Cells of the Nervous System 1020
and Differentiating Cells 986
Information Flows Through Neurons from Dendrites
Stem Cells for Different Tissues Occupy Sustaining Niches 986 to Axons 1020
Germ-Line Stem Cells Produce Sperm and Oocytes 987 Information Moves Along Axons as Pulses of lon
Intestinal Stem Cells Continuously Generate All of the Flow Called Action Potentials 1021
Cells of the Intestinal Epithj:!lium 988 Information Flows Between Neurons via Synapses 1022
Neural Stem Cells Form Nerve and Glial Cells The Nervous System Uses Signaling Circuits Composed
in the Central Nervous System 991 of Multiple Neurons 1022
Hematopoietic Stem Cells Form All Blood Cells 993 Glial Cells Form Myelin Sheaths and Support Neurons 1023
Meristems Are Niches for Stem Cells in Plants 995
22.2 Voltage-Gated lon Channels and the
21.3 Mechanisms of Cell Polarity and Propagation of Action Potentials 1025
Asymmetric Cell Division 997 The Magnitude of the Action Potential Is Close to EN•
Cell Polarization and Asymmetry Before Cell Division and Is Caused by Na Influx Through Open
Follow a Common Hierarchy 998 Na Channels 1025
Polarized Membrane Traffic Allows Yeast to Grow Sequential Opening and Closing of Voltage-Gated
Asymmetrically During Mating 998 Na and K Channels Generate Action Potentials 1025
The Par Proteins Direct Cell Asymmetry in the Nematode Action Potentials Are Propagated Unidirectionally
Embryo 998 Without Diminution 1029

CONTENTS xxxl
Nerve Cells Can Conduct Many Action Potentials Mechanical and Chemical Boundaries Form a First Layer
in the Absence of ATP 1029 of Defense Against Pathogens 1062
Voltage-Sensing 54 a Helices Move in Response Innate Immunity Provides a Second Line of Defense After
to Membrane Depolarization 1030 Mechanical and Chemical Barriers Are Crossed 1062
Movement of the Channel-Inactivating Segment Inflammation Is a Complex Response to Injury That
into the Open Pore Blocks ion Flow 1032 Encompasses Both Innate and Adaptive Immunity 1065
Myelination Increases the Velocity of Impulse Conduction 1032 Adaptive Immunity, the Third Line of Defense,
Action Potentials "Jump" from Node to Node Exhibits Specificity 1066
in Myelinated Axons 1033
23.2 Immunoglobulins: Structure and Function 1068
Two Types of Glia Produce Myelin Sheaths 1033
Immunoglobulins Have a Conserved Structure Consisting
22.3 Communication at Synapses 1036 of Heavy and Light Chains 1068

Formation of Synapses Requires Assembly of Multiple Immunoglobulin lsotypes Exist, Each

Presynaptic and Postsynaptic Structures 1037 with Different Functions 1068

Neurotransmitters Are Transport ed into Synaptic Each B Cell Produces a Unique, Clonally Distributed
,
Vesicles by H"'" -Linked Anti port Proteins 1038 Immunoglobulin 1069

Synaptic Vesicles Loaded with Neurotransmitter Immunoglobulin Domains Have a Characteristic Fold
Are Localized near the Plasma Membrane 1039 Composed ofTwo 13 Sheets Stabilized by
2
a Disulfide Bond 1071
Influx of Ca Triggers Release of Neurotransmitters 1040
An Immunoglobulin's Constant Region Determines
A Calcium-Binding Protein Regu lates Fusion of Synaptic
Its Functional Properties 1072
Vesicles with the Plasma Membrane 1041
Fly Mutants Lacking Dynamin Cannot Recycle Synaptic 23.3 Generation of Antibody Diversity
Vesicles 1042 and B-Cell Development 1073
Signaling at Synapses Is Terminated by Degradation A Functional Light-Chain Gene Requires Assembly ofV
or Reuptake of Neurotransmitters 1042 and J Gene Segments 1074
Opening of Acetylcholine-Gated Cation Channels Rearrangement of the Heavy-Chain Locus Involves V, D,
Leads to Muscle Contraction 1043 and J Gene Segments 1075
All Five Subunits in the Nicotinic Acetylcholine Receptor Somatic Hypermutation Allows the Generation and
Contribute to the ion Channel 1044 Selection of Antibodies with Improved Affinities 1077
Nerve Cells Make an Ali-or-None Decision to Generate B-Cell Development Requires Input from a Pre-B-Cell
an Action Potential 1045 Receptor 1077
Gap Junctions Allow Certain Neurons to Communicate During an Adaptive Response, B Cells Switch from Making
Directly 1045 Membrane-Bound lg to Making Secreted lg 1079
B Cells Can Switch the lsotype of Immunoglobulin
22.4 Sensing the Environment: Touch, Pain,
They Make 1080
Taste, and Smell 1047
Mechanoreceptors Are Gated Cation Channels 1047 23.4 The MHC and Antigen Presentation 1081
Pain Receptors Are Also Gated Cation Channels 1048 The MHC Determines the Ability ofTwo Unrelated
Individuals of the Same Species to Accept
Five Primary Tastes Are Sensed by Subsets of Cells
or Reject Grafts 1081
in Each Taste Bud 1048
The Killing Activity of Cytotoxic T Cells Is Antigen Specific
A Plethora of Receptors Detect Odors 1050
and MHC Restricted 1082
Each Olfactory Receptor Neuron Expresses a Single
T Cells wit h Different Functional Properties Are Guided
Type of Odorant Receptor 1051
by Two Distinct Classes of MHC Molecules 1082
MHC Molecules Bind Peptide Antigens and Interact
23 Immunology 1059 with the T-Cell Receptor 1084
Antigen Presentation Is the Process by Which
23.1 Overview of Host Defenses 1061 Protein Fragments Are Complexed with MHC
Pathogens Enter the Body Through Different Routes Products and Posted to the Cell Surface 1086
and Replicate at Different Sites 1061 Class I MHC Pathway Presents Cytosolic Antigens 1087
Leukocytes Circulate Throughout the Body and Take Up Class II MHC Pathway Presents Antigens Delivered
Residence in Tissues and Lymph Nodes 1061 to the Endocytic Pathway 1089

xxxii CONTENTS
23.5 T Cells, T-Cell Receptors, and T-Cell Successive Oncogenic Mutations Can Be Traced
1092 in Colon Cancers 1120
Development
Cancer Cells Differ from Normal Cells in Fundamental Ways 1122
The Structure of the T-Cell Receptor Resembles
the F(ab) Portion of an Immunoglobulin 1093 DNA Microarray Analysis of Expression Patterns
Can Reveal Subtle Differences Between Tumor Cells 1123
TCR Genes Are Rearranged in a Manner Similar
to Immunoglobulin Genes 1093
24.2 The Genetic Basis of Cancer 1124
T-Cell Receptors Are Very Diverse, with Many ofTheir
Gain-of-Function Mutations Convert Proto-oncogenes
Variable Residues Encoded in the Junctions
1095 into Oncogenes 1125
Between V, D, and J Gene Segments
Signaling via Antigen-Specific Receptors Triggers Cancer-Causing Viruses Contain Oncogenes
1095 or Activate Cellular Proto-oncogenes 1127
Proliferation and Differentiation ofT and B Cells
T Cells Capable of Recognizing MHC Molecules Develop Loss-of-Function Mutations in Tumor-Suppressor
Through a Process of Positive and Negative Selection 1097 Genes Are Oncogenic 1128

T Cells Require Two Types of Signal for Full Activation 1098 Inherited Mutations in Tumor -Suppressor Genes
Increase Cancer Risk 1128
Cytotoxic T Cells Carry the CDS Co-receptor and
Are Specialized for Killing 1099 Epigenetic Changes Can Contribute to Tumorigenesis 1129

T Cells Produce an Array of Cytokines That Provide 24.3 Cancer and Misregulation of Growth
Signals to Other Immune Cells 1099
Regulatory Pathways 1131
CD4 T Cells Are Divided into Three Major Classes Based
Mouse Models of Human Cancer Teach Us
on Their Cytokine Production and Expression
About Disease Initiation and Progression 1131
of Surface Markers 1100
Oncogenic Receptors Can Promote Proliferation
Leukocytes Move in Response to Chemotactic Cues
in the Absence of External Growth Factors 1132
Provided by Chemokines 1101
Viral Activators of Growth-Factor Receptors Act
23.6 Collaboration of Immune-System Cells as Oncoproteins 1133
in the Adaptive Respon se 1102 Many Oncogenes Encode Constitutively Active
Signal Transduction Proteins 1134
Toll-Like Receptors Perceive a Variety of Pathogen-Derived
Macromolecular Patterns 1102 Inappropriate Production of Nuclear Transcription
Engagement ofToil-Like Receptors Leads to Activation Factors Can Induce Transformation 1136
of Antigen-Presenting Cells 1104 Aberrations in Signaling Pathways That Control
Development Are Associated with Many Cancers 1137
Production of High-Affinity Antibodies Requires
Collaboration Between B and T Cells 1104 Molecular Cell Biology Is Changing How Cancer Is Treated 1138
Vaccines Elicit Prot~ctive Immunity Against a Variety
of Pathogens 1105 24.4 Cancer and Mutation of Cell Division
and Checkpoint Regulators 1140
CLASSIC EXPERIMENT 23.1 Two Genes Become One: Mutations That Promote Unregulated Passage
Somatic Rearrangement of Immunoglobulin Genes 1111 from G1 to S Phase Are Oncogenic 1140
Loss of p53 Abolishes the DNA Damage Checkpoint 1141
Apoptot ic Genes Can Function as Proto-oncogenes
or Tumor-Suppressor Genes 1143
24 Cancer 1113 Micro-RNAs Are a New Class of Oncogenic Factors 1143

24.1 Tumor Cells and the Onset of Cancer 1114 24.5 Carcinogens and Caretaker Genes
Metastatic Tumor Cells Are Invasive and Can Spread 1115 in Cancer 1144
Cancers Usually Originate in Proliferating Cells 1116 Carcinogens Induce Cancer by Damaging DNA 1144
Local Environment Impacts Heterogeneous Tumor Some Carcinogens Have Been Linked to Specific Cancers 1145
Formation by Cancer Stem Cells 1117 1146
Loss of DNA-Repair Systems Can Lead to Cancer
Tumor Growth Requires Formation of New Blood Vessels 1117 Telomerase Expression Contributes to Immortalization
Specific Mutations Transform Cultured Cells of Cancer Cells 1148
into Tumor Cells 1118
GLOSSARY G-1
A Multi-hit Model of Cancer Induction Is Supported
by Several Lines of Evidence 1119 INDEX 1-1

CONTENTS xxxiii
I·

I .
 CHAPTER

Molecules, Cells,
and Evolution
Cultured mouse embryonic fibrob lasts stained for three proteins that
form the cytoskeleton. [Courtesy of Ana M. Pasapera, Clare M. Waterman)

Nothing in biology makes sense except in the light of evolution. bacteria and protozoans visible only under·a microscope to
mu lticell ular animals of all ki nds. Yet the bewildering array
- Theodosius Dobzhansky
of outward biological forms overlies a powerful uniformity:
(essay in The American Biology
thanks to our common ancestry, all biological systems are
Teacher 35: 125-129, 1973)
composed of the same types of chemical molecules and em-

B
iology is a science fundamenta lly different from phys- ploy similar princ1ples of organization at the cellular level.
ics or chemistry, which deal with unchanging proper- Although the basic kinds of biological molecules have been
tics of matter that can be described by mathematical conserved during the billions of years of evolution, the pat-
equations. Biological systems of course follow the rules of terns in which they arc assembled to form functioning cells
chemistry and physics, but biology is a historical science, as and organisms have undergone considerable change.
the forms and structures of the living world today are the We now know that genes, which chemically are com-
resu lts of bi lliom of years of euolution. Through evolution, posed of deoxyribonucleic acid (DNA), ultimately define bio-
all organisms are re lated in a family tree extending from logica l struct ure and maintain the integration of cellular
primitive single-celled organisms that lived in the distant function. Many genes encode proteins, the primary molecules
past to the diverse plants, animals, and microorganisms of that make up cell structures and carry out cellular activities.
the present era (Figure 1-1, Table 1-1). The great insight of Alterations in the structure and organization of genes, or mu-
Charles Darwin (Figure 1-2) was the principle of natural se- tations, provide the random variation that can alter biologi-
lection: organisms vary randomly and compete within their cal structure and function. While the vast majority of random
envi ronment for resources: Only those that survive to repro- mutations have no observable effect on a gene's or protein's
duce are able to pass down their genetic traits. function, many are deleterious, and only a few confer an evo-
At firs t glance, the biological universe appears amazingly lutionary advantage. In all organisms mutations in D"'A are
diverse-from tiny ferns to tall fir trees, from single-celled constantly occurring, allowing over time the small alterations

OUTLINE

1.1 The Molecules of Life 4 1.3 Cells into Tissues: Unicellular and Metazoan
Organisms Used for Molecular Cell Biology
1.2 Genomes, Cell Architecture, and Cell Function 10 Investigations 16
 An imals EUKARVOTA
Plants Fungi
M icrosporidia
Slim e
mo lds
Entamoeba
BACTERIA
Low G + C gram- Apicomplexa
. hG C {e.g., Plasmodium)
H tg + gram- positives ARCHAEA A
positives I Euglena
Euryarchaeota
o/e purples : Kinetoplasta
Korarchaeota {e.g., Trypanosoma)
Crena rchaeota 1
a purples I

y/p purples / Para basalia

-/' {e.g., Trichomonas)
Spirochaet es ,;
,; M etamonda
Fusobacteria
Flexibacter/
Th ermotogales ," " " {e.g., Giardia)

Bacteroides
Cyanobacteria
Thermus

Aquifex

• • Presumed last common ancestor of

eukaryotes and archaebacteria

•
The Hominidae {great apes)
Presumed last common ancestor of
all extant organisms

Rhesus
macaque Gibbon Su m atran
Macaca Nom ascus ora ngutan Gori ll a Human Bonobo Chimpanzee
mulatta leucogenys Pongo abelii Gorilla gorilla Homo sapiens Pan paniscus Pan troglodytes
0.996
-1 Myr ago

FIGURE 1-1 All living organisms descended from a common fossil record generally agree well with t hose based on molecular data.
ancestral cell. (a) All organisms from simple bacteria to complex (b) Evolution of great apes, a small ape, and an Old World monkey with
mammals probably evolved from a com mon single-celled ancestor. respect to humans, as estimated from the divergence among thei r
This fami ly tree depicts the evolutionary relations among the three genomic DNA sequences. Whole genome DNA sequences were
major lineages of organ isms. The structure of the tree was initially aligned, and the average nucleotide divergence in unique DNA
ascertained from morphological criteria: creatures that look alike were sequences was estimated. Estimates of the times different species
put close toqether. More recently t he seq uences of DNA and proteins diverged from each other, calculated in million s of years (Myr). il r P
found in organisms have been examined as more information-rich indicated at each node;- 1 Myr implies approximately 1 Myr or less.
criteria for assigning relationships. Th e g reater the similarities in t hese [Part (a) adapted from J. R. Brown, 2005, "Universal tree of life," in Encyclopedia of
macromolecular sequences, the more closely related organisms are Life SCiences, Wiley lnterScience (online). Part (b) adapted from D.P. Locke et al.,
thought to be. The trees based on morphological comparisons and t he 2011, Nature 469:529.]

2 CHAPTER 1 • Molecules, Cells, and Evolution

 Timeline for Evolution of Life on Earth, as Determined from the Fossil Record

4600 million years ago The planet Earth forms from material revolving around the young Sun.

-.1900-2500 million years ago Cells resembling prokaryotes appear. These first organisms are chemoautotrophs: they use
carbon dioxide as a carbon source and oxidize inorganic materials to extract energy.

3 sao million years ago Lifetime of the last universal ancestor; the split between bacteria and archaea occurs.

2700 million years ago Photosynthesizing cyanobacteria evolve; they usc water as a reducing agent, thereby
producing oxygen as a waste product.

I 850 million years ago Unicellular eukaryotic cells appear.

1200 million years ago Simple multicellular organisms evolve, mostly consisting of cell colonies of limited complexity.

580-500 million years ago Most modern phyla of an1mals begin to appear in the fossil record during the Cambrian
explosion.

535 million years ago Major diversification of living things in the oceans: chordates, arthropods (e.g., trilobites,
crustaceans), echinoderms, mollusks, brachiopods, foraminifers, radiolarians, ere.

485 million years ago first vertebrates with true bones (jawless fishes) evolve.

434 million years ago First primitive plants arise on land.

225 million years ago Earliest dinosaurs (prosauropods) and teleost fishes appear.

220 million years ago Gymnosperm forests dominate the land; herbivores grow to huge sizes.

215 million years ago First mammals evolve.

65.5 million years ago The Cretaceous-Tertiary extinction event eradicates about half of all animal species,
including all of the dinosaurs.

6.5 million years ago first hominids evolve.

2 million years ago First members of the genus Homo appear.

350 thousand years ago Neanderthals appear.

200 thousand years ago Anatomic<llly modern humans appear in Africa.

30 thousand years ago Extinction of Neanderthals.

in cellular structures and functions that may prove to be ad- organisms plays a fundamental role in this process because it
vantageous. Entirely new structures rarely are created; more allows these changes to come about by small alterations in
often, old structures are adapted to new circumstances. More previously evolved cells, giving them new abilities. The result
rapid change is possible by rearranging or multiplying previ- is that closely related organisms have very similar genes, pro-
ously evolved components rather than hy waiting for a teins, and cellular organization.
wholly new approach to emerge. For instance, in a particular Living systems, including the human body, consist of such
organism one gene may randomly become duplicated; one closely interrelated elements that no single element can be
copy of the gene and its encoded protein may retain their fully appreciated in isolation from the others. Organisms
original function while over time the second copy of the gene contain organs, organs are composed of tissues, tissues con-
mutates such that its protein takes on a slightly different or sist of cells, and cells are formed from molecules (Figure 1-3 ).
even a totally new function. The cellular organization of The unity of living systems is coordinated by many levels of

CHAPTER 1 • Molecules, Cells, and Evolution 3

 In the third section we discuss the formation of tissues from
individual cells, and the diverse types of unicellular and mul-
ticellular organisms used in investigations of molecular cell
biology.
One focus of this chapter is DNA, as we now have the
complete sequence of the genomes of over a hundred organ-
isms and these have provided considerable insight into the
evolution of genes and organisms. Recent studies, for in-
stance, indicate that human and chimpanzee genomes are
about 99 percent identical in sequence and that the ancestors of
these species likely diverged from a common ape-like organism
between 4.5 and 6 million years ago (see Figure 1-lb). This
conclusion is consistent with the fossil record (see Table 1-1 ).
Biologists use evolution as a research tool: if a gene and its
protein have been conserved in, say, all metazoans (multicel-
lular animals) but are not found in unicellular organisms, the
protein likely has an important 'function in all metazoans
and thus can be studied in whatever metazoan organism is
most suitable for the investigation. Interwoven in the second
and third sections of chis chapter are discussions of the rea-
sons scientists pick particular unicellular and multicellular
"model" organisms to study specific genes and proteins that
are important for cellular function.

FIGURE 1-2 Charles Darwin (1 809-1 882). Four years after his
epic voyage on HMS Beagle, Darwin had already begun formulating 1.1 The Molecules of Life
in private notebooks his concept of natural selection, which would
be published in Origin ofSpecies (1859). [Walt AndersonNisuals While large polymers are the focus of molecular cell biology,
Unlimited, Inc.) small molecules are the stage upon which all cellular pro-
cesses are set. Water, inorganic ions, and a wide array of
relatively small organic molecules (Figu re 1-4) account for
75 to 80 percent of living matter by weight, and water ac-
counts for about 75 percent of a cell's volume. These small
interrelationship: molecules carry messages from organ to molecules, including water, serve as substrates for many of
organ and cell to cell and tissues are delineated and inte- the reactions that take place inside the cell, including energy
grated with other tissues by molecules secreted by cells. metabolism and cell signaling. Cells acquire these small mol-
Generally all the levels into which we fragment biological ecules in different ways. Ions, water, and many small organic
systems interconnect. molecules are imported into the cell (Chapter 11 ); other
To learn about biological systems, however, we must ex- small molecules are synthesized within the cell, often by a
amine one small portion of a living system at a time. The series of chemical reactions.
biology of cells is a logical starting point because an organ- Even in the structures of many small molecules, such as
ism can be viewed as consisting of interacting cells, which sugars, vitamins, and amino acids, we sec the footprint of
are the closest thing to an autonomous biological unit that evolution. For example, all amino acids save glycine have an
exists. The last common ancestor of all life on earth was a asymmetric carbon atom, yet only the L-stereoisomer, never
cell, and at the cellular level all life is remarkably similar. All the o-stereoisomer, is incorporated into proteins. Similarly,
cells use the same molecular building blocks, similar meth- only the o-stereoisomer of glucose is invariably found in
ods for the storage, maintenance, and expression of genetic cells, never the mirror image L-stereoisomer (see Figure 1-4).
information, and similar processes of energy metabolism, At an early stage of biological evolution, our common cellu-
molecular transport, signaling, development, and structure. lar ancestor evolved the ability to catalyze reactions with one
In this chapter, we introduce the common features of stereoisomer instead of the other. How these selections hap-
cells. We begin with a brief discussion of the principal small pened is unknown, hut now these choices are locked in place.
molecules and macromolecules found in biological systems. An important and universally conserved small molecule
Next we discuss the fundamental aspects of cell structure is adenosine triphosphate (ATP), which stores readil y
and function that are conserved in present-day organisms, available chemical energy in two of its chemical bonds
and the use of prokaryotic organisms (single-celled organ- (Figure 1-5). When one of these energy-rich bonds in ATP is
Isms without a nucleus) to study the basic molecules of life. broken, forming ADP (adenosine diphosphate), the released

4 CHAPTER 1 • Molecules, Cells, and Evolution

 (a) (b)

Basal lamina
lll!!lli!ili~..d} ~oose connective
ttssue
2011m
l-J
·.· (d)

(c)

lh.r.~-_._, Cytoskeleta I
proteins
(_
Cell-surface
receptor

Multiadhesive protein Hemidesmosome

FIGURE 1-3 Living systems such as the human body consist of (c) Tissues are formed through subcellular adhesion structures
closely interrelated elements. (a) The surface of our hand is covered (desmosomes and hemidesmosomes) that join cells to each other and
by a living organ, skin, that is comprised of several layers of tissue. (b) An to an underlying layer of supporting fibers. (d) At the heart of cellular
outer covering of hard, dead skin cells protects the body from injury, adhesion are its structural components: phospholipid molecules that
infection, and dehydration. This layer is constantly renewed by living make up the cell surface membrane, and large protein molecules.
epidermal cells, which also give rise to hair and fur in animals. Deeper Protein molecules that traverse the cell membrane often form strong
layers of muscle and connective tissue give skin its tone and firmness. bonds with internal and external fibers made of multiple proteins.

energy can be harnessed to power an energy-requiring pro-

cess such as muscle contraction or protein biosynthesis. To
obtain energy for making ATP, all cells break down food
molecules. For instance, when sugar is degraded to carbon
dioxide and water, the energy stored in the sugar molecule's
chemical bonds is released and much of it can be "captured"
in the energy-rich bonds in ATP (Chapter 12). Bacterial,
Sodium Water
plant, and animal cells can all make ATP by this process. In
addition, plants and a few other organisms can harvest en-

+ 0 ergy from sunlight to form ATP in photosynthesis.

~
Other small molecules (e.g., hormones and growth fac-
tors) act as signals that direct the activities of cells (Chapters
~-~ ~~ 15 and 16), and nerve cells communicate with one another
• L-senne
• o-senne•• by releasing and sensing certain small signaling molecules
(Chapter 22). The powerful effect on our body of a frighten-
L-glucose o-glucose ing event comes from the instantaneous flooding of the body
FIGURE 1-4 Some of the many small molecules found in cells. with the small-molecule hormone adrenaline, which mobi-
Only the L-forms of amino acids such as serine are incorporated into lizes the "fight or flight" response.
proteins, not their o-mirror images; only the o-form of glucose, not its Certain small molecules (monomers) can be joined to form
L-mirror image, can be metabolized to carbon dioxide and water. polymers, also called macromolecules, through repetition of a

1.1 The Molecules of Life 5

 NH2 NH 2

N~c"c--N~ N~c"c -- N~
ATP I II /H ADP I II /H
HC~N...--c--..N HC ~N...--C--..N

o=-~-o-~-o-cH o I
0 0 0
Light (photosynthesis) or

6- l a- ~
compounds with high
potential energy (respiration)

HO OH

High-energy bonds High-energy bond

ATP ADP+ P,

Synthesis of Synthesis of other Cellular movements, Transport of Generation of an Heat

cellular macro- cellular constituents including muscle con- molecules against electric potential
molecules (DNA, (such as membrane traction, crawling move- a concentration across a membrane
RNA, proteins, phospholipids and ments of entire cells, gradient (importa nt for nerve
polysaccharides) certain required and movement of function)
metabolites) chromosomes during
mitosis
FIGURE 1-5 Adenosine triphosphate (ATP) is the most common (P,) by photosynthesis in plants, and by the breakdown of sugars and
molecule used by cells to capture and transfer energy. ATP is fats in most cel ls. The energy released by the splitting (hydrolysis) of P,
formed from adenosine diphosphate (ADP) and inorganic phosphate from ATP drives many cellular processes.

single type of covalent chemical-linkage reaction (see Figure 2-1 ). such as DNA and RNA. Cytoskeletal proteins serve as struc-
Cells produce three types of large m acromolecules: polysaccha- tural components of a cell, for example by forming an inter-
rides, proteins, and nucleic acids (figure 1-6). Sugars, for exam- nal skeleton; others power the movement of subcellular
ple, are the monomers used to fo rm polysaccharides. Different structures such as chromosomes, and even of whole cells, by
polymers of o-glucose form the cellulose component of plant cell using energy stored in the chemical bonds of ATP (Chapters
walls and glycogen, a storage form of glucose found in liver and 17 and 1 8). Other proteins bind adj acent cells together or
muscle. The cell is careful to provide the appropriate mix of small form parts of the extracellular matrix (see Figure 1-3). Pro-
molecules needed as precursors for synthesis of macromolecules. teins can be sensors that change shape as temperature, ion
concentrations, or other properties of the cell change. Many
proteins that are embedded in the cell surface (plasma) mem-
Proteins Give Cells Structure and Perform
brane import and export a vari ety of small molecules and ions
Most Cellular Tasks (Chapter 11 ). Some proteins, such as insulin, are hormones;
Proteins, the workhorses of the cell, are the most abundant others arc hormone receptors that bind their ta rget proteins
and functionally versatile of the cellular macromolecules. and then generate a signal that regulates a specific aspect of
Cells string together 20 different amino acids in a linear cell function. Other important classes of proteins bind to spe-
chain to form proteins (see Figure 2-14), which commonly cific segments of DNA, turning genes on or off (Chapter 7). In
range in length from 100 to 1000 ami no acids. During its fact, much of molecular cell biology consists of studying the
polymerization a linear chain of amino acids folds into a com- function of specific proteins in specific cell types.
plex shape, conferring a distinctive three-dimensional struc- How can 20 am ino acids form all the different proteins
ture and function on each protein (see Figure 1-6). Humans needed to perform these varied tasks? It seems impossible at
obtain amino acids either by synthesizing them from other first glance. But if a "typical" protein is abuut 400 ami no
molecules or by breaking down proteins that we eat. acids long, there are 20 400 possible different am ino acid se-
Proteins have a variety of functions in the cell. Many quences. Even assumi ng that many of th ese would be
proteins arc enzym es, wh ich accelerate (catalyze) chemical functionally equivalent, unstable, or otherwise discountable,
rcacttons involving small molecules or macromolecules the number of possible proteins is ast ronomical.
(Chapter 3). Certain proteins catalyze steps in the synthcsi!> of Next we might ask how many protein molecules a cell
proteins; others catalyze synthesis of other macromolecules needs to operate and maintain itself. To estimate this number,

6 CHAPTER 1 • Molecules, Cells, and Evolution

 DNA molecule RNA molecule

Adenylate
kinase
Insulin

Glutamine synthetase Hemoglobin Immunoglobulin Lipid bilayer

FIGURE 1-6 Models of some representative proteins drawn to a bonds. The illustrated proteins include enzymes {glutamine synthetase
common scale and compared with a small portion of a lipid bilayer and adenyl ate kinase), an antibody (immunoglobulin), a hormone
sheet, a DNA molecule, and an RNA molecule. Each protein has a (insu lin), and the blood's oxygen carrier (hemoglobin).
defined three-dimensional shape held together by numerous chemical

let's take a typical euka ryotic cell (a cell containing a nucleus), transfer of genetically determined characteristics from one
such as a hepatocyte (liver cell). This cell, roughly a cube 15 generation to the next.
f.lm (0.00 15 em) on a side, has a volume of 3.4 X 10 - 9 cm 3 (or DNA strands are composed of monomers called nucleo-
milliliters, ml ). Assum ing a cell densiry of 1.03 g/ml, the cell tides; these often are referred to as bases because their struc-
wou ld weigh 3.5 X 10 9 g. Since protein accounts for approx- tures contain cyclic organic bases (Chapter 4). Four different
imately 20 percent of a cell 's weight, the total weight of cel-
lular protein is 7 X 10 10 g. The average p rotein has a
molecular weight of 52,700 g/mol; we can calculate the total
number of p rotein molecules per liver cell as about 7.9 X 109
from the total protein weight and Avogadro's number, t he
number of molecules per mole of any chemical compound
(6.02 X 1023 ) . To carry this calculation one step further, con-
sider that a liver cell contains about 10,000 different proteins;
thus each cell would on average contain close to a mi llion
molecules of each type of protein. In fact, the abundance of
different proteins varies widely, from the quite rare insulin-
binding receptor protein (20,000 molecules per cell ) to the
abundant structural protein actin (5 X l0 8 molecules per cell).
Every cell closely regulates the level of each protein such that
each is present in the appr0priare quantity for irs cellular func-
tions, as we deta il in Chapters 7 and 8.

Nucleic Acids Carry Coded Information for

Making Proteins at the Right Time and Place
T he macromolecule that garners the most public attention is
deoxyribonucleic acid (DNA), whose func tional properties
make it the cell's "master molecule." T he three-dimensional
structure of DNA, first proposed by James D. Watson and FIGURE 1-7 James D. Watson (left) and Francis H. C. Crick (right)
Francis H. C. Crick about 60 years ago (Figure 1-7), consists with the double-helical model of DNA they constructed in
of two long helical strands that are coiled around a common 1952- 1953. Their model ultimately proved correct in all its essential
axis to form a do uble helix (Figure 1-8). The double-helical aspects. [A. Barrington Brown/Science Photo Researcher. From J.D. Watson,
structure of DNA, one of nature's most magnificent con- 1968, The Double Helix, Atheneum, Copyright 1968, p. 21 5; Courtesy of A. C.
structions, is critical to the phenomenon of heredity, the Barrington Brown.]

1.1 The Molecules of Life 7

FIGURE 1-8 DNA consists of two complementary Nucleotide (T)
strands wound around each other to form a double
helix. The double helix is stabilized by weak hydrogen
bonds between the A and T bases and between the
C and G bases. During replication, the two strands are
unwound and used as templates to produce
complementary strands. The outcome is two copies
of the origina l double helix, each containing one
of the original strands and one new daughter
(complementary) strand.

nucleotides, abbreviated A, T , C, and G, a re joined to form helix has a simple construction: w herever one strand has an A,
a DNA strand, with the base parts project ing inward from the other strand has aT, and each Cis matched with a G (see
the backbone of the strand. Two strands bind togeth er via th e Figure 1-8). T hi s complementa ry ma tchi ng of the two stra n d~
bases, a nd twist to form a dou b le helix. Each DNA double is so strong t hat if complementary stra nds are sepa rated, they
.·
li§:b¥1 Genome Sizes of Organisms Used in Molecular Cell Biology Research That Have Been Completely Sequenced

Bacteria Base pairs (millions) Encoded proteins Chromosomes Reference

Mycoplasma genitalum 0.58 482 A

Helicobacter pylori 1.67 1,587 A

Haemoplnlus mfluenza 1.83 1,737 A

Escherichia coli 4.64 4 ,289 l A

Bacillus subtilis 4.22 4,245 A

Archaea

Methanococcu~ jannaschii 1.74 1,785 3 A

Sulfolobus solfat,mcus 2.99 2,960 A

Eukaryotes

Saccharomyces cerevisiae 12. 16 5,885 16 B

Drosophila melanogaster 168 1 3,~8 1 4 c

Caenorhabditis elegans 100 20,424 6 D

Danio rerio 1505 19,929 25 c

Gallus gallus (ch1cken ) 1050 14,923 39 c
Mus m usculus 3421 22,085 20 c
Homo sapiens 3279 21,077 23 c
Arabidopsis thaliana 135 27,4 16 5 E

Table courtesy of Dr. han Lew mer. SOLR( t'l: A, hrrp://cmr.jcvi.org/cgi-bm/C\1R/shared/Genomc~ .cgi; B, http://www.yca~tgenomc.org/; C,
http://uswest. ensembl.org/info/abom/specics.hrml; D. hrrp://wikl.wormbase.org/index.php/WS222; E, http://www.arabidopsis.oqq'porrals/gcnAnnotation/
gene_,tru.:rural_annorarion/annorarion_data.j~p.

8 CHAPTER 1 • Molecules, Cells, and Evolution

will spontaneously zip back together under the right salt
concentration and temperature conditions. Such nucleic acid
hybridization is extremely useful for detecting one strand by
using the other, as we learn in Chapter 5. II 0
Activation DNA
The genetic information carried by DNA resides in its
sequence, the linear order of nucleotides along a strand. Spe- Start !
cific segments of DNA, termed genes, carry instructions for II
Transcription
making specific proteins. Commonly genes contain two
parts: the coding region specifies the amino acid sequence of
pre-m RNA --~=·
a protein; the regulatory region binds specific proteins and
controls when and in which cells the protein is made. II Nucleus
Most bacteria have a few thousand genes; yeasts and other Processing
unicellular eukaryotes have about 5000. Humans and other
metazoans have between 13,000 and 23,000, while many
mRNA
!,
plants like Arabidopsis have more (Table 1-2). As we discuss
later in this chapter, many bacterial genes encode proteins that
are conserved throughout all living organisms. These catalyze Protein
reactions that occur universally, such as the metabolism of II __.....g.,,. Cytoplasm
glucose and synthesis of nucleic acids and proteins. Studies on Translat ion
bacterial cells have yielded profound insights into these basic
life processes. Similarly, many genes in unicellular eukaryotes
O Transcription Q RNA Ribosome
such as yeasts encode proteins that are conserved throughout
all eukaryotes; we will see how yeasts have been used to study
I factor
~
polymerase
Transcribed region of DNA
processes such as cell division that have yielded profound ~ Nontranscribed region of DNA
insights into human diseases such as cancer. - - - - Protein-coding region of (lNA
Cells use two processes in series to convert the coded in- Noncoding region of RNA
formation in DNA into proteins (Figure 1-9). In the first, Amino acid chain
called transcription, the coding region of a gene is copied into
a single-stranded ribonucleic acid (RNA) whose sequence is FIGURE 1-9 The i nformation coded in DNA is converted i nt o t he
the same as one of the two in the double-stranded DNA. A amino acid sequences of pr oteins by a multistep process. Step 0 :
large enzyme, RNA polymerase, catalyzes the linkage of nu- Transcription factors bind to the regulatory regions of the specific
cleotides into an RNA chain using DNA as a template. In genes they control and activate them. Step f): Following assembly of a
eukaryotic cells, the initial RNA product is processed into a multi protein initiation complex bound to the DNA, RNA polymerase
begins transcription of an activated gene at a specific location, the start
smaller messenger RNA (mRNA) molecule, which moves out
site. The polymerase moves along the DNA linking nucleotides into a
of the nucleus to the cytoplasm. Here the ribosome, an enor-
single-stranded pre-mRNA transcript using one of the DNA strands as
mously complex molecular machine composed of both RNA
a template. Step il: The transcript is processed to remove noncoding
and protein, carries out the second process, called transla- sequences. Step 19: In a eukaryotic cell, the mature messenger RNA
tion. During translation, the ribosome assembles and links (mRNA) moves to the cytoplasm, where it is bound by ribosomes that
together amino acids in the precise order dictated by the read its sequence and assemble a protein by chemically linking amino
mRNA sequence according to the nearly universal genetic acids into a linear chain.
code. We examine the cell components that carry out tran-
scription and translation in detail in Chapter 4.
In addition to its role in transferring information from though they do not code for proteins, serve equally impor-
nucleus to cytoplasm, RNA can serve as a framework for tant purposes in the cell (Chapter 6 ). Sma ll micro RNAs,
building a molecular machine. For example, the ribosome is 20-25 nucleotides long, are abundant in metazoan cells and
built of four RNA chains that bind to more than 50 proteins bind to and repress the activity of target mRNAs. By some
to make a remarkably precise and efficient mRNA reader estimates these small RNAs may indirectly regulate the activ-
and protein synthesizer. While most chemical reactions in ity of most or all genes, though the mechanisms and ubiquity
cells are catalyzed by proteins, a few, such as the formation of this type of regulation arc still being explored (Chapter 8).
of the peptide bonds that connect amino acids in proteins, Several long noncoding RNAs bind to DNA or chromosomal
are catalyzed by RNA molecules. proteins and ~o affect chromosome structure and RNA syn-
Well before the entire human genome was sequenced it thesis, processing, and stability. However, we know the func-
was apparent that only about 5 percent of human DNA tion of only very few of these abundant noncoding RNAs.
codes for protein, and for many years most of the human All organisms must control when and where their genes
genome was considered "junk DNA"! However, in recent can be transcribed. Nearly all the cells in our bodies contain
years we've learned that much of the so-called junk DNA is the full set of human genes, but in each cell type only some
actually copied into thousands of RNA molecules that, of these genes are active, or turned on, and used to make

1.1 The Molecules of Life 9

proteins. For instance, liver cells produce some proteins that membrane in each cell also contains groups of proteins that
are not produced by kidney cells, and vice versa. Moreover, allow specific ions and small molecules to cross. Other mem-
many cells respond to external signals or changes in external brane proteins serve to attach the cell to other cells or to poly-
conditions by turning specific genes on or off, the reby adapt- mers tha t surround it; still others give the cell its shape or
ing their repertoire of proteins to meet current needs. Such allow its shape to change. We will learn more about mem-
control of gene activity depends on DNA-binding proteins branes and how molecules cross them in Chapters 10 and 11.
called transcription factors, which bind to specific sequences New cells are always derived from parental cells by cell
of DNA and act as switches, either activating or repressing division. We've seen that the synthesis of new DNA molecules
transcription of particular genes (see Figure 1-9 and Chapter 7). is templated by the two strands of the parental DNA such
Transcription factors often work as multiprotein complexes, that each daughter DNA molecule has the same sequence as
with each protein contributing its own DNA-binding speci- the parental one. In parallel, membranes arc made by incor-
ficity to selecting the regulated genes. poration of li pids and proteins into existing membranes in
the parental cell, and these are divided between daughter
cells by fission. Thus membrane synthesis, similarly to DNA
Phospholipids Are the Conserved Building
synthesis, is also templated by a parental structure.
Blocks of All Cellular Membranes
In essence, any cell is simply a compartment with a watery
interior that is separated from the external environment by a 1.2 Genomes, Cell Architecture,
surface membrane, the plasma membrane, which prevents
and Cell Function
the free flow of molecules in and out. In addition, eukaryotic
cells have extensive internal membranes that further subdi- The biological universe consists of two types of cells-pro-
vide the cell into multiple subcompartments, the organelles. karyotic and eukaryotic. Prokaryotic cells such as bacteria
In all organisms cellular membranes are composed pri- consist of a single closed compartment that is surrounded by
marily of a bilayer (two layers) of phospholipid molecules. the plasma membrane, lack a defined nuCleus, and have a
These bipartite molecules have a "water-loving" (hydrophilic) relatively simple internal organization (Figure 1-11 ). Eu-
end and a "water-hating" (hydrophobic) end. The two phos- karyotic cells, unlike prokaryotic cells, contain a defined
pholipid layers of a membrane are oriented with all the hydro- membrane-bound nucleus and extensive internal membranes
philic ends directed toward the inner and outer surfaces of the that enclose the o rganelles (Figure 1-12 ). The region of the
membrane and the hydrophobic ends buried within its inte- cell lyi ng between the plasma membrane and the nucleus is
rior (Figure 1-10). Smaller amounts of other lipids, such as the cytoplasm, comprising the cytosol (water, dissolved ions,
cholesterol, are inserted into the phospholipid framework. small molecules, and proteins) and the organelles. Eukary-
Phospholipid membranes are impermeable to water, all ions, otes include four kingdoms: the plants, anima ls, fungi, and
and virtually all hydrophilic small molecules. Thus each protists. Prokaryotes comprise the fifth and sixth kingdoms:
the eubacteria (true bacteria ) and archaea.
Hydrophilic Genome sequencing has provided profound insights into
Cholesterol head group the function and evolution of both conserved and noncon-
served genes and proteins found in multiple organisms. In the
following section we describe some basic structural and func-
tional features of prokaryotic and eukaryotic cells and relate
these to insights provided from their genome sequences. We
emphasize the conserved proteins found in multiple diverse
species and explain why scientists have chosen several of
these species as model organisms, systems in which the study
of specific aspects of cellular function and deve lopment can
serve as a model for other species (Figure 1-13 ).
Water/ Transmembrane proteins
FIGURE 1- 1 0 The watery interior of cells is surrounded by Prokaryotes Comprise True Bacteria
t he plasma membrane, a two-layered shell of phospholipids. and Archaea
The phospholipid molecules are oriented with their hydrophobic fatty
acyl chains (black squiggly lines) facing inward and their hydrophilic
In recent years, detailed analysis of the DNA sequences from
head groups (white spheres) faciny outward. Thus both sides of the
a variety of prokaryotic organisms has revealed two d istinct
membrane are lined by head groups, mainly charged phosphates, kingdoms: the eubacteria, often simply called "bacteria,"
adjacent to the watery spaces inside and outside the cell. All biological and the archaea. Eubacteria, a numerous type of prokaryote,
membranes have the same basic phospholipid bilayer structure. arc single-celled organisms; included are the cyanobacteria,
Cholesterol (red) and various proteins are embedded in the bilayer. or blue-green algae, which can be unicellular or filamentous
The interior space is actually much larger relative to the volume of the chains of cells. Figure 1-1 I illustrates the general structure of
plasma membrane than is depicted here. a typical bacterial cell; archaeal cells have a similar structure.

10 CHAPTER 1 • Molecules, Cells, and Evolution

.•

Nucleoid

Outer membrane Inner (plasma) Nucleoid

membrane 0.5 f.Lm

FIGURE 1-11 Prokaryotic cells are have a relatively simple membrane. (Right) This artist's drawing shows the nucleoid (blue)
structure. (Left) Electron micrograph of a thin section of Escherichia and a magnification of the layers that surround the cytoplasm. Most
coli, a common intestinal bacterium. The nucleoid, consisti ng of the of the cell is composed of water, proteins, ions, and other molecules
bacterial DNA, is not enclosed within a membrane. E. coli and other that are too small to be depicted in the scale of this drawing. [Electron
gram-negative bacteria are surrounded by two membranes separated micrograph courtesy of I. D. J. Burdett and R. G. E. Murray.lllustration by
by the peri plasmic space. The thin cell wall is adjacent to the inner D. Goodsell.)

Bacterial cells are commonly 1-2 !J-ill in size and consist of a proteins. Many proteins are precisely localized within the
single closed compartment containing the cytoplasm and cytosol or in the plasma membrane, indicating the presence
bounded by the plasma membrane. Although bacterial cells of an elaborate internal organization.
do not have a defined nucleus, the single circular DNA ge- Bacterial cells possess a cell wall, which lies adjacent to
nome is extensively folded and condensed into the central the external side of the plasma membrane. The cell wall is
region of the cell. In contrast, most ribosomes are found in composed of layers of peptidoglycan, a complex of proteins
the DNA-free region of the cell. Some bacteria also have an and oligosaccharides; it helps protect the cell and maintain
invagination of the cell membrane, called a mesosome, its shape. Some bacteria (e.g., E. coli) have a thin inner cell
which is associated with synthesis of DNA and secretion of wall and an outer membrane separated from the inner cell

(b)
(a)
Nuclear membrane

Peroxisome
Lysosome Lysosome

11-Lm Rough endoplasmic

Endoplasmic reticulum
L...:...._j reticulum
FIGURE 1-12 Eukaryotic cells have a complex internal structure continuous with the rough endoplasmic reticulum, a factory for
with many membrane-limited organelles. (a) EIPctron micrograph assembling secreted and membrane proteins. Golgl vesicles process
and (b) diagram of a plasma cell, a type of white blood cell that secretes and modify secreted and membrane proteins, mitochondria generate
antibodies. A single membrane (the plasma membrane) surrounds the energy, lysosomes digest cell materials to recycle them, peroxisomes
cell and the cell interior contains many membrane-limited compart- process molecules using oxygen, and secretory vesicles carry cell
ments, or organelles. The defining characteristic of eukaryotic cells is materials to the surface to release them. [From P. C. Cross and K. L. Mercer,
segregation of the cellular DNA within a defined nucleus, which is 1993, Cell and Tissue Ultrastructure: A Functional Perspective, W. H. Freeman
bounded by a double membrane. The outer nuclear membrane is and Company.)

1.2 Genomes, Cell Architecture, and Cell Function 11

(;) PODCAST: Common Experimental Organisms
(a) (b)

Viruses Bacteria

Proteins involved in DNA, RNA, Proteins involved in DNA, RNA,

protein synthesis protein synthesis,
Gene regulation metabolism
Cancer and control of cell Gene regulation
proliferation Targets for new antibiotics
Transport of protems and Cell cycle
organelles inside cells Signaling
Infection and immunity
Possible gene therapy approaches

(c) (d)
Yeast (Saccharomyces cerevisiae) Roundworm ( Caenorhabditis
elegans)
Control of cell cycle and cell division
Protein secretion and membrane Development of the body plan
biogenesis Cell lineage
Function of the cytoskeleton Formation and function of the
Cell differentiation nervous system
Aging Control of programmed cell death
Gene regulation and chromosome Cell proliferation and cancer genes
structure Aging
Behavior
Gene regulation and chromosome
structure •
(e) (f)

Fruit fly (Drosophila melanogaster) Zebrafish

Development of the body plan Development of vertebrate body

Generation of differentiated cell tissues
lineages Formation and function of brain and
Formation of the nervous system, nervous system
heart, and musculature Birth defects
Programmed cell death Cancer
Genetic control of behavior
Cancer genes and control of cell
pro I iteration
Control of cell polarization
Effects of drugs, alcohol, pesticides
(g) (h)

Mice, including cultured cells Plant (Arabidopsis thaliana)

Development of body tissues Development and patterning of

Function of mammalian immune tissues
system Genetics of cell biology
Formation and function of brain Agricultural applications
and nervous system Physiology
Models of cancers and other Gene regulation
human diseases Immunity
Gene regulation and inheritance Infectious disease
Infectious disease

FIGURE 1-13 Each experimental organism used in cell biology musculus) (g) are evolutionarily the closest to humans and have provided
has advantages for certain types of studies. Viruses (a) and bacteria models for studying numerous human genetic and infectious diseases.
(b) have small genomes amenable to genetic dissection. Many insights The mustard-family weed Arabidopsis thaliana has been used for genetic
into gene control initially came from studies with these organisms. The screens to identify genes involved in nearly every aspect of plant life.
yeast Saccharomyces cerevisiae (c) has the cellular organization of a Genome sequencing is completed for many viruses and bacterial species,
eukaryote but is a relatively simple single-celled orgdni~m that is easy to the yeast 5. cerevisfae, the roundworm C. elegans, the frUit fly D. mefanogos-
grow and to manipulate genetically. In the nematode worm Caenorhabditis ter, humans, mice, zebrafish, and the plant A. thalia no. Other organisms,
elegans (d), which has a small number of cells arranged in a nearly particularly frogs, sea urchins, chickens, and slime molds, have also had
identical way in every worm, the formation of each individual cell can be their genomes sequenced and continue to be immensely valuable for cell
traced. The fruit fly Drosophila melanogaster (e), first used to discover the biology research. Increasingly, a wide variety of other species are used,
properties of chromosomes, has been especially valuable in identifying especially for studies of evolution of cells and mechanisms. [Part (a) Visuals
genes that control embryonic development. Many of these genes are Unlimited, Inc. Part (b) Kari Lountmaa/Science Photo Library/Photo Researchers, Inc.
evolutionarily conserved in humans. The zebrafish Dania rerio (f) is used for Part (c) Scimat/Photo Researchers, Inc. Part (d) Photo Researchers, Inc. Part (e) Darwin
rapid genetic screens to identify genes that control vertebrate develop- Dale/Photo Researchers, Inc. Part (f) lnge SpenceNisuals Unlimited, Inc. Part (g) J. M.
ment and organogenesis. Of the experimental animal systems, mice (Mus Labat/Jancana/Visuals Unlimited, Inc. Part (h) Darwin Dale/Photo Researchers, Inc.]
 wall by the periplasmic space. Such bacteria are not stained and function to membrane proteins in certain mammalian
by the Gram technique and thus are classified as gram-negative. brain cells that import small nerve-to-nerve signaling mole-
Other bacteria (e.g., Bacillus polymyxa) that have a thicker cules called neurotransmitters (Chapters ll and 22).
cell wall and no outer membrane take the Gram stain and
'• thus arc classified as gram-positive.
All Eukaryotic Cells Have Many of the Same
Working on the assumption that similar organisms diverged
more recently from a common ancestor than did dissimilar Organelles and Other Subcellular Structures
one'>, researchers have developed the evolutionary lineage tree Eukaryotes comprise all members of the plant and animal
shown in Figure 1-Ia. According to this tree, the archaea and kingdoms, as well as fungi (e.g., yeasts, mu~hrooms, molds )
the cukaryotes diverged from bacteria more than a billion years and protozoans (proto, primitive; zoan, animal), which arc
before they diverged from each other (Table 1-1). In addition to exclusively unicellular. Eukaryotic cells are commonly about
DNA sequence distinctions that define the three groups of or- 10-100 JJ-m across, generally much larger than bacteria. A
ganisms, archaeal cell membranes have chemical properties that typical human fibroblast, a connective tissue cell, is about 15
differ dramatically from those of bacteria and eukaryotes. JJ-m across with a volume and dry weight some thousands of
Man) archacans grow in unusual, often extreme, environ- rimes those of an E. coli cell. An amoeba, a ~ingle-celled pro-
ments that may resemble the ancient conditions that existed tozoan, can have a cell diameter of approximately 0.5 mm,
when life first appeared on earth. For instance, halophiles more than thirty times that of a fibroblast.
("salt lovers") require high concentrations of salt to survive, Eukaryotic cells, like prokaryotic cells, are surrounded
and thermoacidophilcs ("heat and acid lovers" ) grow in hot by a plasma membrane. However, unlike prokaryotic cells,
(80 °C} sulfur springs, where a pH of less than 2 is common. most cukaryotic cells (the human red blood cell is an excep-
Still other archaeans live in oxygen-free milieus and generate tion) also contain extensive internal membranes that enclose
methane (CH4 ) by combining water with carbon dioxide. specific subcellular compartments, the organelles, and sepa-
rate them from the rest of the cytoplasm, the regwn of the
cell lying outside the nucleus (see Figure 1-12). Many organ-
Escherichia coli Is Widely Used
elles are surrounded by a single phospholipid membrane, but
in Biological Research the nucleus, mitochondrion, and chlorophist are enclosed b)
The bacterial lineage includes 1-..scherichia coli, a favorite ex- two membranes. Each type of organelle contains a collection
perimental organism which in nature is common in soil and of specific proteins, including enzymes that catalyze requisite
animal intestines. E. coli and several other bacteria have a chemical reactions. The membranes defining these subcellu-
number of advantages as experimental organisms. They lar compartments control their internal ionic composition so
grow rapidly in a simple and inexpensive medium containing that it generally differs from that of the surrounding cytosol
glucose and salts, in which they can synthesize all necessary as well as that of the other organelles.
amino acids, lipids, vitamins, and other essential small mol- The largest organelle in a eukaryotic cell is generally the
ecules. Like all bacteria, E. coli possesses elegant mecha- nucleus, which houses most of the cellular DNA. In animal
nisms for controlling gene activity that are now well and plant cells, most ATP is produced by large multi protein
understood. Over time, workers have developed powerful "molecular machines" located in the organelles termed mito-
systems for genetic analysis of this organism. These systems chondria. Plants carry out photosynthesis in chloroplasts,
are facilitated by the small size of bacterial genomes, the ease organelles that contain molecular machines for synthesizing
of obtaining mutants, the availability of techniques for trans- ATP from ADP and phosphate, similar to those found in mi-
ferring genes into bacteria, an enormous wealth of knowl- tochondria. Similar molecular machines for generating ATP
edge about bacterial gene control and protein functions, and are located in the plasma membrane of bacterial cells. Both
the relative simplicity of mapping genes relative to one an- mitochondria and chloroplasts are thought to have origi-
other in the bacterial gooome. In Chapter 5 we see how E. nated as bacteria that took up residence inside eukaryotic
coli is used in recombinant DNA research. cells and then became welcome collaborators (Chapter 12 ).
Bacteria such as E. coli that grow in environments as di- Over time many of the bacterial genes "migrated" to the cell
verse as the soil and the human gut have about 4000 genes nucleus and became incorporated into the cell's nuclear ge-
encoding about the same number of proteins (see Table 1-2). nome. Both mitochondria and chloroplasts contain small ge-
Parasitic bacteria such as the Mycoplasma species acquire nomes that encode a few of the essential organelle proteins;
amino acids and other nutrients from their host cells, and the sequences of these DNAs reveal their bacterial origins.
lack the genes for enzymes that catalyze reactions in the syn- Cells need to break down worn-out or obsolete parts
thesis of amino acids and certain lipids. Many bacterial genes into small molecules that can he discarded or recycled. In
encoding proteins essential for DNA, RNA, protein synthe- animals this housekeeping task is assigned in part to lyso-
sis, and membrane function are conserved in all organisms, somes, organelles filled with degradative enzymes. The inte-
and much of our knowledge of these important cellular pro- rior of a lysosome has a pH of about 5.0, much more acidic
cesses was uncovered first in E. coli. For example, certain than that of the surrounding cytosol. This aids in the break-
E. coli cell membrane proteins that import amino acids across down of materials by lysosomal enzymes, which can func-
the plasma membrane are closely related in sequence, structure, tion at such a low pH. To create the low-pH environment,

1.2 Genomes, Cell Architecture, and Cell Function 13

 Microtubules Microfilaments Intermediate filaments
FIGURE 1-14 The three types of cytoskeletal filaments have or green). Visualization of the stained cell in a fluorescence microscope
characteristic distributions within mammalian cells. Three views of reveals the location of filaments bound to a particular dye-antibody
the same cell. A cultured fibroblast was permeabilized and then treated preparation. In this case, microtubules are stained blue; microfilaments,
with three different antibody preparations. Each antibody binds specifi- red; and intermediate filaments, green. All three fiber systems
cally to the protein monomers forming one type of filament and is contribute to the shape and movements of cells. [Courtesy ofV. Small.]
chemically linked to a differently colored fluorescent dye (blue, red,

proteins located in the lysosomal membrane pump hydrogen (a) (b)

ions into the lysosome using energy supplied from ATP S phase
(Chapter 11). Plants and fungi contain a vacuole that also
has a low-pH interior and stores certain salts and nutrients.
Peroxisomes are another type of small organelle, found in
virtually all eukaryotic cells, that is specialized for breaking
down the lipid components of membranes.
The cytoplasm of eukaryotic cells contains an array of Chromosome Sister chromatid pair
fibrous proteins collectively called the cytoskeleton (C hap-
ters 17 and 18). Three classes of fibers compose the cytoskel- (c)
eton: microtubules (20 nm in diameter), built of polymers of
the protein tubulin; microfilaments (7 nm in diameter), built
of the protein actin; and intermediate fi laments (10 nm in
diameter), built of one or more rod-shaped protein subunits
(Figure 1-14 ). The cytoskeleton gives the cell strength and
rigidity, thereby helping to maintain cell shape. Cytoskeletal
fibers also control movement of structures within the cell;
for example, some cytoskeletal fibers connect to organelles
or provide tracks along which organelles and chromosomes
move; other fibers play key roles in cell motility. Thus the
cytoskeleton is important for "organizing" the cell.
The rigid cell wall, composed of cellulose and other poly-
mers, that surrounds plant cells contributes to their strength
and rigidity. Fungi are also surrounded by a cell wall, but its
composition differs from that of bacterial or plant cell walls.
Each organelle membrane and each space in the interior
of an organelle has a unique set of proteins that enable it to
carry out its specific functions. For cells to work properly,
the numerous proteins composing the various working com- FIGURE 1-15 Individual chromosomes can be seen in cells
partments must be transported from where they are made to during cell division (mitosis). (a) During the 5 phase of the cell cycle
(see Figure 1-16) chromosomes are duplicated and the daughter "sister
their proper locations (Chapters 17 and 18 ). Some proteins
chromatids," each with a complete copy of the chromosomal DNA,
are made on ribosomes that are free in the cytoplasm; from
remain attached at the centromere. (b) During the actual cell division
there, some proteins are moved into the nucleus while others
process (mitosis) the chromosomal DNA becomes highly compacted
arc directed into mitochondria, 1..hlorupla~r~, or peroxi- and the pairs of sister chromatids can be seen in the electron microscope
somes, depending on their specific functions. Proteins to be as depicted here. (c) Light microscopic image of a chromosomal spread
secreted from the cell and most membrane proteins, in con- from a cultured human male lymphoid cell arrested in the metaphase
trast, are made on ribosomes associated with the endoplasmic stage of mitosis by treatment with the microtubule-depolymerizing
reticulum (ER). This organelle produces, processes, and ships drug colcemid. There is a single copy of the duplicated X andY
out both proteins and lipids. Most protein chains produced on chromosomes and two of each of the others. [Part (b) courtesy of Medical
the ER move to the Golgi complex, where they are further RF/The Medical File/Peter Arnold Inc. Part (c) courtesy of Tatyana Pyntikova.]

14 CHAPTER 1 • Molecules, Cells, and Evolution

..
modified before being forwarded to their final destinations. (Figure 1-16 ). The chromosomes and the DNA they carry
Proteins that travel in this way contain short sequences of are duplicated during the S (synthesis) phase. The repli-
amino acids or attached sugar chains (oligosaccharides) that cated chromosomes separate during the M (mitotic) phase,
serve as addresses for directing them to their correct destina- with each daughter cell getting a copy of each chromosome
tions. These addresses work because they are recognized and during cell division. The M and S phases are separated by
bound by other proteins that do the sorting and shipping in two gap stages, the G 1 phase and the G2 phase, during
various cell compartments. which mRNAs and proteins are made and the cell increases
in size.
In single-celled organisms, both daughter cells often
Cellular DNA Is Packaged Within Chromosomes (though not always) resemble the parent cell. In multicellular
In most prokaryotic cells, most or all of the genetic informa- organisms, when many types of cells divide the daughters look
tion resides in a single circular DNA molecule about a milli- a lot like the parent cell-liver cells, for instance, divide to
meter in length; this molecule lies, folded back on itself many generate two liver cells with the same characteristics and func-
times, in the central region of the micrometer-sized cell (Fig- tions as their parent, as do insulin-producing cells in the pan-
ure l-11). ln contrast, DNA in the nuclei of eukaryotic cells creas. In contrast, stem cells and certain other undifferentiated
is distributed among multiple long linear structures called cells can generate multiple types of differentiated descendant
chromosomes. The length and number of chromosomes arc cells; these cells often divide such that the two daughter cells
the same in all cells of an organism, but vary among different are different. Such asymmetric cell division is critical to the
types of organisms (see Table 1-2). Each chromosome com- generation of different cell types in the body (Chapter 21).
prises a single DNA molecule associated with numerous pro- Often one daughter resembles its parent in that it remains un-
teins, and the total DNA in the chromosomes of an organism differentiated and retains its ability to give rise to multiple
is referred ro as i ts geno me. Chromosomes, which stain in- types of differentiated cells. The other daughter divides many
tensely with basic dyes, are visible in light and electron micro- times and each of the daughter cells differentiates into a spe-
scopes only during cell division, when the DNA becomes ci fie type of cell.
tightly compacted (Figure 1-15). Although the large genomic Under optimal conditions some bacteria, such as E. colz,
DNA molecule in prokaryotes is associated with proteins and can divide to form two daughter cells once every 30 minutes.
often is referred to as a chromosome, the arrangement of Most eukaryotic cells take considerably longer to grow and
DNA within a bacterial chromosome differs greatly from that divide, though cell divisions in the early Drosophila embryo
within the chromosomes of eukaryotic cells. require only 7 minutes. Moreover, the cell cycle in eukary-
otes normally is highly regulated (Chapter 19). This tight
control prevents imbalanced, excessive growth of cells and
All Eukaryotic Cells Utilize a Similar Cycle
tissues if essential nutrients or certain hormonal signals are
to Regulate Their Division lacking. Some highly specialized cells in adult animals, such
Unicellu lar eukaryotcs, animals, and plants use essentially as nerve cells and striated muscle cells, divide rarel} if at all.
the same cell cycle, a series of events that prepares a cell to However, an organism usually replaces worn out cells or
divide, and the actual division process, called mitosis. The makes more cells in response to a new need, as exemplified
eukaryotic cell cycle commonly is represented as four stages by the growth of muscle in response to exercise or damage.

g OVERVIEW ANIMATION: Life Cycle of a Cell

FIGURE 1- 16 During growth, all eukaryotic cells continually Nondividing
p rogress through the four stages of the cell cycle, generating new cells
Resting
daughter cells. In proliferating human cells, the four phases of the cell
G, ~
cells
cycle proceed successively, taking from 10-20 hours depending on cell
type and developmental state. Yeasts divide much faster. During RNA and
interphase, which consists of the G1, S, and G2 phases, the cell roughly protein
doubles its mass. Replication of DNA during the S phase leaves the cell synthesis
with four copies of each type of chromosome. In the mitotic (M) phase,
the chromosomes are evenly partitioned into two daughter cells, and
the cytoplasm divides roughly in ha lf in most cases. Under certain
DNA
conditions, such as starvation or when a tissue has reached its final size,
replication
cells will stop cycling and remain in a waiting state called G0 • Most cells
in G0 can reenter the cycle if conditions change.

1.2 Genomes, Cell Architecture, and Cell Function 15

Another example is the formation of additional red blood (a)
cells when a person ascends to a higher altitude and needs
more capacity to capture oxygen. The fundamental defect in
cancer is loss of the ability to control the growth and division
of cells. In Chapter 24 we examine the molecular and cellular
events that lead to inappropriate, uncontrolled proliferation
of cells.
Budding (S. cerevisiae)
(b)
1.3 Cells into Tissues: Unicellular and Mating between haploid
D
Metazoan Organisms Used for Molecular
Cell Biology Investigations
cells of opposite mating
type
a a /I'
@@ a Vegetative growth
~ of diploid cells

Our current understanding of the molecular functioning of

cells largely rests on studies of just a few types of organisms,
~ Dlploldooll•(ola) ~B"d

I()~
termed model organisms. Because of the evolutionary conser-
vation of genes, proteins, organelles, cell types, and so forth,
discoveries about biological structures and functions ob-
tained with one experimental organism often apply to others.
@
:£@@
Thus researchers generally conduct studies with the organism
that is most suitable for rapidly and completely answering the
question being posed, knowing that the results obtained in
B
~~g::r~tive \
ofhaploid \
~
' !~
one organism are likely to be broadly applicable.
As we have seen, bacteria are excellent models for studies
cells @~
~
0
'-.
~
Fourhaploid
ascospores
@.() @
within ascus
of several cellular functions, but they lack the organelles
found in eukaryotes. Unicellular eukaryotes such as yeasts
arc used to study many fundamental aspects of eukaryotic
Ascus rupt~
IJ spores germin;te
@
e
• ~IJ
0
Starvation ca~ses
asc~s ~ormat10n,
meiOSIS
cell structure and function. Multicellular, or metazoan, mod-
els are required to study more complex tissue and organ sys- FIGURE 1 -17 The yeast Saccharomyces cerevisiae can grow as
tems and development. As we will sec in this section, several both haploids and diploids and can reproduce sexually and
asexually. (a) Scanning electron m icrograph of the budding yeast
eukaryoric model organisms are in wide use to understand
Saccharomyces cerevisiae. These cells grow by an unusual type of
these complex cell systems and mechanisms.
mitosis termed mitotic budding. One daughter nucleus remains in the
"mother" cell; the other daughter nucleus is transported into the bud,
Single-Celled Eukaryotes Are Used to Study which grows in size and soon is released as a new cell. After each bud
Fundamental Aspects of Eukaryotic Cell cell breaks free, a scar is left at the budding site, so the number of
previous buds on the mother cell can be counted. The orange-colored
Structure and Function cells are bacteria. (b) Haploid yeast cells can have different mating
One group of single-celled eukaryotes, the yeasts, has proven types, called a and a; both types contain a single copy of each yeast
exceptionally useful in molecular and generic analysis of cu- chromosome, half the usual number, and grow by mitotic budding.
karyotic cell formation and function. Yeasts and their multi- Two haploid cells that differ in mating type, one a and one a, can fuse
cellular cousins, the molds, which collectively constitute the together to form an a/a diploid cell that contains two copies of each
fungi, have an important ecological role in breaking down chromosome; diploid cells also multiply by mitotic budding. Under
starvation conditions, diploid cells undergo meiosis, a special type of
plant and animal remains for reuse. They also make numer-
cell division, to form haploid ascospores. Rupture of an ascus releases
ous antibiotics and are used in the manufacture of bread,
four haploid spores, which can germinate into haploid a and a cells.
beer, and wine.
These also can multiply asexually. [Part (a) M. AbbeyNisualsUnlimited, Inc.]
The common yeast used to make bread and beer, Saccha-
romyces cerevisiae, appears frequently in this book because it
has proven to be an extremely useful experimental organism.
C)•de and catalyze DNA replication and transcription. S. cere-
Homologs of many of the approximately 6000 different pro-
uzsiae (Figure l-l7a) and other yeasts offer many advantages
teins expressed in an S. cerevisiae cell (Ta hiP 1-2) arc found
to molecular and cellular biologists:
in most if not all eukaryotes and arc important for cell divi-
sion or for the functioning of individual eukaryotic organ- • Vast numbers of yeast cells can be grown easily and cheaply
elles. Much of what we know of the proteins in the endoplasmic in culture from a single cell; such cell clones all have the
reticulum and Golgi apparatus that promote protein secre- same genes and the same biochemical properties. Individual
tion was elucidated first in yeasts. Yeasts were also essential proteins or mulriprotein complexes can be purified from
for the identification of many proteins that regulate the cell large amounts of cells and then studied in detail.

16 CHAPTER 1 • Molecules, Cells, and Evolutton

• Yeast cells can grow by mitosis both as haploids (contain- like yeasts since there is only one copy of each gene, and a
ing one copy of each chromosome) and as diploids (contain- mutation in it will immediately have a consequence.
ing two copies of each chromosome); this makes isolating By analyzing the effects of numerous different temperature-
and characterizing mutations in genes encoding essential cell sensitive mutations that altered division of haploid yeast cells,
proteins relatively straightforward. geneticists discovered most of the genes necessary for cell divi-
sion without knowing anything, initially, about which proteins
• Yeasts, like many organisms, have a sexual cycle that al-
they encode or how these proteins participate in the process.
lows exchange of genes between cells. Under starvation con-
The great power of genetics is to reveal the existence and rele-
ditions, diploid cells undergo meiosis, a special type of cell
vance of proteins without prior knowledge of their biochemi-
division, to form haploid daughter cells, which are of two
cal identity or molecular function. Eventually these
types, a and a cells. Haploid cells can also grow by mitosis.
"mutation-defined" genes were isolated and replicated (cloned)
If haploid a and a cells encounter each other they can fuse,
with recombinant DNA techniques discussed in Chapter 5.
forming an ala diploid cell that contains two copies of each
With the isolated genes in hand, the encoded proteins could be
chromosome (Figure l-17b).
produced in the test tube or in engineered bacteria or cultured
With the use of a single species like S. cerevisiae as a cells. Then biochemists could investigate whether the proteins
model organism, results from studies carried out by tens of associate with other proteins or DNA or catalyze particular
thousands of scientists worldwide, using multiple experi- chemical reactions during cell division (Chapter 19).
mental techniques, can be combined to yield a deeper level of Most of these yeast cell cycle genes are found 111 human
understanding of a single type of cell. As we will see manr cells as well, and the encoded proteins have similar amino
times in this book, conclusions based on studies of S. cerevl- acid sequences. Proteins from different organisms, but with
siae are often generally true for all eukaryotes and form the similar amino acid sequences, are said to be homologous, and
basis for exploring the evolution of more complex processes may have the same or similar functions. Remarkably, it has
in multicellular animals and plants. been shown that a human cell cycle protein, when expressed
in a mutant yeast defective in the homologous yeast protein,
is able to "rescue the defect" of the mutant yeast (that is, to
allow the cell to grow normally), thus demonstrating the pro-
Mutations in Yeast led to the Identification
tein's ability to function in a very different type of eukaryotic
of Key Cell Cycle Proteins cell. This experimental result, which garnered a Nobel Prize
Biochemical studies can tell us much about an individual pro- for Paul Nurse, was especially notable because the common
tein, but they cannot prove that it is required for cell division ancestor cell of present day yeasts and humans is thought to
or any other cell process. The importance of a protein is dem- have lived over a billion years ago. Clearly the eukaryotic cell
onstrated most firmly if a mutation that prevents its synthesis cycle and the genes and proteins that catalyze it evolved early
or makes it nonfunctional adversely affects the process under in biological evolution and have remained quite constant
study. A diploid organism generally carries two versions (al- over a very long period of evolutionary time. Importantly,
leles) of each gene, one derived from each parent. There are subsequent studies showed that mutations in many yeast cell
important exceptions, such as the genes on the X and Y chro- cycle proteins that allow uncontrolled cell growth also fre-
mosomes in mares of some species, including our own. quently occur in human cancers (Chapter 24), again attesting
In a classical genetics approach, scientists isolate and to the important conserved functions of these proteins in all
characterize mutants that lack the ability to do something a eukaryotes.
normal organism can do. Often large genetic "screens" are
done to look for many different mutant individuals (e.g.,
Multicellularity Requires Cell-Cell
fruit flies, yeast cells) that are unable to complete a certain
process, such as cell div..ision or muscle formation. Muta- and Cell Matrix Adhesions
tions usually are produced by treatment with a mutagen, a The evolution of multicellular organisms most likely began
chemical or physical agent that promotes mutations in a when cells remained associated in small colonies after divi-
largely random fashion. But how can we isolate and main- sion instead of separating into individual cells. A few pro-
tain mutant organisms or cells that are defective in some karyotes and several unicellular eukaryotes, such as many
process, such as cell division, that is necessary for survival? fungi and slime molds, exhibit such rudimentary social be-
One way is to isolate organisms with a temperature-sen- havior. The full flowering of multicellularity, however, oc-
sitive mutation. These mutants are able to grow at the per- curred in eukaryotic organisms whose cells became
missive temperature, but not at another, usually higher differentiated and organized into groups, or tissuPs, in
temperature, the nonpermtsstve temperature. Normal cells which the cells performed a specialized, common function.
can grow at either temperature. In most cases, a tempera- Metazoans-be they invertebrates like the fruit fly Drosoph-
ture-sensitive mutant produces an altered protein that works ila melanogaster and the roundworm Caenorhabditis ele-
at the permissive temperature but unfolds and is nonfunc- gans, or vertebrates such as mice and humans-contain
tional at the nonpermissive temperature. Temperature-sensi- between 13,000 and 23,000 protein-coding genes, about
tive screens arc most readily done with haploid organisms three to four times that of a yeast (Table 1-2). Many of these

1.3 Cells into Tissues: Unicellular and Metazoan Organisms Used for Molecular Cell BIOlogy Investigations 17
genes are conserved among the metazoans and essential for
the formation and function of specific tissues and organs.
Animal cells arc often "glued" together in to a chain, a
ball, or a sheet by cell-adh esion proteins (often ca lled cell
adhesion nmlecules, or CAMs) on their surface (Figure 1-3).
Some CAMs bind cel ls to one another; other types bind cells
to the extracellular matrix, forming a cohesive unit. In ani-
mals, the matrix cushions cells and allows nutrients to dif-
fuse toward them and waste products to diffuse away. A
specialized, especially tough matrix called the basailamina,
comprised of multiple proteins such as collagen and polysac-
charides, forms a supporting layer underlying cell sheets and
prevents the cell aggregates from ripping apart. The cells of
higher plants are encased in a network of chambers formed
by the interlocking cell walls surrounding the cells, and are
connected by cytoplasmic bridges called plasmo desmata.

Tissues Are Organized into Organs

FIGURE 1- 18 All organs are organized arrangements of various
The specialized groups of differentiated cells form tissues, tissues, as illustrated i n this cross section of a small artery
which are themselves the major components of organs. For (arteriole). Blood flows through the vessel lumen (Lu), which is lined
example, the lumen of a blood vessel is lined with a sheetlike by a thin sheet of endothelial cells (EC) forming the endothelium (TI)
layer of endothelial cells, or endoth elium, which prevents and by the underlying basal lamina. This tissue adheres to the
blood cells from leaking out (Figure 1-18). A layer of smooth overlying layer of smooth muscle tissue (TM); contraction of the muscle
muscle tissue encircles the endothelium and basal lamina and layer controls blood flow through the vessel. A fibrillar layer of
contracts to limit blood flow. During times of fright, con- connective tissue (TA) surrounds the vessel and connects it to other
striction of smaller peripheral vessels forces more blood to tissues. Dr. Richard Kessel & Dr. Randy KardonNisuals Unlimited, Inc.
the vital organs. The muscle layer of a blood vessel is
wrapped in an outer layer of connective nssue, a network of
fibers and cells that encase and protect the vessel walls from
stretching and rupture. This hierarchy of tissues is copied in
Body Plan and Rudimentary Tissues Form
other blood vessels, which differ mainly in the thickness of
the layers. The wall of a major artery must withstand much Early in Embryonic Development
stress and is therefore thicker than a minor vessel. The strat- The human body consists of some 100 trillion cells, yet it de-
egy of grouping and layering different tissues is used to build velops from a single cell, the zygote, resulting from fusion of a
other complex organs as well. In each case the function of sperm and an egg. The early stages in the development of an
the organ is determined by the specific functions of Jts com- embryo are characterized by rapid cell d ivision (Figure 1-19)
ponent tissues, and each type of cell in a tissue produces the and the differentiation of cells into tissues. The embryonic
specific groups of proteins that enable the tissue to carry out body plan, the spatial pattern of cell types (tissues) and body
irs functions. parts, emerges from two influences: a program of genes that

0 VIDEO: Early Embryonic Development ·.

(a) (b) (c)

FIGURE 1-19 The first few cell divi sions of a fertilized egg set the embryo is surrounded by supporting membranes. The corresponding
stage for all subsequent development. A developing mouse embryo steps in human development occur during the first few days after
is shown at the (a) two-cell, (b) four· cell, and (c) eight-cel l stages. The fertilization. [Claude Edelmann/ Photo Researchers, Inc.]

18 CHAPTER 1 • Molecules, Cells, and Evolution

 (a) FIGURE 1-20 Similar genes, conserved during evolution,
regulate early developmental processes in diverse animals.
(a) Urbilateria is the presumed ancestor of all protostomes and
Urbilateria deuterostomes that existed about 600 million years ago. The positions

foomilli~
of the nerve chord (violet), surface ectoderm (mainly skin, white), and
endoderm (main ly digestive tract and organs, light green) are shown.
/ yearsago ~ (b) Highly conserved proteins called Hox proteins are found in both
protostomes and deuterostomes and determine the identity of body
r·r ~
«=-==
~ . . . .-
,Y , segments during embryonic development. Hox genes are found in
l t - - ... ~ clusters on the chromosomes of most or all animals, and encode
Protostome Deuterostome related transcription factors that control the activities of other genes. In
many animals Hox genes direct the development of different segments
along the head-to-tail axis, as indicated by corresponding colors. Each
gene is activated (transcriptionally) in a specific region along the

• • •
(b)
head-to-tail axis and controls the growth of tissues there. For example,
in the mouse, a deuterostome, the Hox genes are responsible for
Genes the distinctive shapes of vertebrae. Mutations affecting Hox genes in
the fruit fly, a protostome, cause body parts to form in the wrong
locations, such as legs in lieu of antennae on the head. In both
...
y organisms these genes provide a head-to-tail address and serve to
direct the formation of structures in the appropriate places.

Fly Mammal
(protostome) (deuterostome)

specify the pattern of the body, and local cell interactions Remarkably, many patterning genes that are often called
that induce different parts of the program. "master transcription factors," are highly conserved in both
With only a few exceptions, most animals display axial protostomes and deuterostomes (Figure 1-20b). This conser-
symmetry; that is, their left and right sides mirror each other. vation of body plan reflects evolutionary pressure to pre-
This most basic of patterns is encoded in the genome. Devel- serve the commonalities in the molecular and cellular
opmental biologists have divided bilaterally symmetric animal mechanisms controlling development in different organisms.
phyla into two large groups depending on where th e mouth Fly eyes and human eyes are very different in structure,
and anus form in the early embryo. Protostomes develop a functio n, and nerve connections. Nonetheless, the so-called
mouth close to a transient opening in the earl y embryo (the "master regulator genes" that initiate eye development-
.· blastopore) and have a ventral nerve chord; protostomes in- eyeless in the fly and Pax6 in the human-encode highly re-
clude all worms, insects, and mollusks. The deuterostomes • lated transcription factors that regulate the activities of other
develop an anus close to this transient opening in the embryo genes and are descended from the same ancestral gene. Mu-
and have a dorsal central nervous system; these include echi- tations in the eyeless or Pax6 genes cause major defects in
noderms (such as sea stars and sea urchins) and vertebrates. eye formation (Figure 1-21 ).
The bodies of both protostomes and deuterostomes are di-
vided into discrete segments that form early in embryonic de- Invertebrates, Fish, and Other Organisms
velopmen t. Protostomes and deuterostomes likely evolved
Serve as Experimental Systems for Study
from a common ancestor, termed Urbilateria, that lived ap-
proximately 600 million years ago (figure l-20a). of Human Development
Patterning genes specify the general organization of an Studies of cells in specialized tissues make use of animal and
organism, beginning with the major body axes-anterior- plant model organisms. Nerve cells and muscle cells, for in-
posterior, dorsal-ventral, and left-right-and ending with stance, traditionally were studied in mammals or in creatures
body segments such as the head, chest, abdomen, and tail. with especially large or accessible cells, such as the giant neu-
The conservation of axial symmetry from the simplest ral cells of the squid and sea hare or the flight muscles of
worms to mammals is explained by the presence of con- birds. More recently, muscle and nerve development have
served patterning genes in their genomes. Some patterning been extensively studied in fruit flies (Drosophila melano-
genes encode proteins that control expression of other genes; gaster ), roundworms (Caenorhabditis elegans), and zebra-
other patterning genes encode proteins that are important in fish (Dania rerio), in which mutants in muscle and nerve
cell adhesion or in cell signaling. This broad repertoire of formation or function can be readily isolated (Figure 1-13 ).
patterning genes permits the integration and coordination of Organisms with large-celled embryos that develop out-
events in different parts of the developing embryo and gives side the mother (e.g., frogs, sea urchins, fish, and chickens)
each segment in the body its unique identity. are extremely useful for tracing the fates of cells as they form

1.3 Cells into Tissues: Unicellular and Metazoan Organ1sms Used for Molecular Cell Biology Investigations 19
(a) (b) versions of human genetic diseases. Inactivating particular
genes by introducing short pieces of interfering RNA allows
quick tests of gene functions possible in many organisms.

Mice Are Frequently Used to Generate

Models of Human Disease
Mice have one enormous advantage over other experimental
organisms: they are tht: du~est to humans of any anima l for
w hich powe rful genetic approaches are feasible. Mice and
humans have shared living structures for millennia, have
similar immune systems, and are subject to infection by
many of the same pathogens. Both organisms contain about
(c) (d) the same number of genes, and about 99 percent of mouse
protein-coding genes have homologs in the human , and vice
versa. Over 90 percent of mouse and h uman genomes can be
partitioned into regions of conserved synteny-that is, DNA
segments that have the same order of unique DNA sequences
and genes along a segment of a chromosome. T his means
that the gene order in the most recent common ancestor of
humans and mice has been conserved in both species (Figure
1-22). This conserved synteny is consistent with archeologi-
cal and other evidence that huma ns and mice descended
FIGURE 1-21 Similar genes, conserved during evolution, from a common mammalian evolutionary ancestor that
regulate organ development in diverse animals. (a) Development likely lived about 75 mi llion years ago. Of course mice are
of the large compound eyes in fruit flies requires a gene called eyeless not people; relative to humans, mice have expanded families
(named for the mutant phenotype). (b) Flies with inactivated eyeless of genes related to immunity, reproduction, and olfaction,
genes lack eyes. (c) Normal human eyes require the gene Pax6, the likely reflecting the differences between the human and
homolog of eyeless. (d) People lacking ad equate Pax6 function have mouse lifestyle.
the genetic disease aniridia, a lack of irises in the eyes. Pax6 and eyeless In Chapter 5 we will learn about the experi mental utility
encode highly related transcription factors that regulate the activities
of mouse embryonic stem (ES) cells, lines of cells derived
of other genes and are descended from the same ancestral gene.
from early mouse embryos that can be grown in culture in an
[Parts (a) and (b) Andreas Hefti, Interdepartmental Electron Microscopy (I EM)
undifferentiated state. Using techniques of recombinant
Biocenter of the University of Basel. Part (c) \1:1 Simon Fraser/Photo Researchers, Inc.
DNA, scientists can introduce specific muta tions into the
Part (d) Visuals Unlimitid.]
mouse genome that mimic the corresponding mutations in
human disease. For example, patients with a certain type of
different tissues and for making extracts for biochemical cancer accumulate inactivating mutations in a key cell cycle
studies. For instance, a key protein in regulating mitosis was regu latory protein, and the ana logous mutation can be intro-
first identified in studies with frog and sea urchi n embryos duced into the corresponding mouse gene. These gene-
and subsequently purified from their extracts (Chapter 20). altered ES cells can be injected into an early mouse embryo,
Using recombinant DNA techniques, researchers can en- which is then implanted into a pseudopregnant fema le mouse
gineer specific genes to contain mutations that inactivate or (a mouse treated with hormones to trigger physiological
increase production of their encoded proteins. Such genes can changes needed for pregnancy). If th e mice that develop
be introduced into the embryos of worms, flies, frogs, sea from the injected ES cells exhibit a disease sim ilar to the
urchins, chickens, mice, a variety of plants, and other o rgan- human cancer, then the link between the disease and muta-
isms, permitting the effects of these mutations to be assessed. tions in a particular gene or genes is supported. Once mouse
This approach is being used extensively to produce mouse models of a human disease are available, further studies on

FIGURE 1-22 Conservation of synteny

between human and mouse. Shown is a typical
510,000 base pair (bp) segment of mouse
chromosome 12 that shares common ancestry
with a 600,000 bp section of human chromosome
14. Blue lines connect the reciprocal unique DNA
sequences m the two genomes. Mb, 1 million
base pairs. [After Mouse Genome Sequencing
Consortium, 2002, Nature 420:520.] 59.9 60.5 (Mb)

20 CHAPTER 1 • Molecules, Cells, and Evolution

(a) T4 bacteriophage (b) Tobacco mosaic virus

50 nm

(c) Adenovirus

50 nm
l___j

FIGURE 1-23 Viruses must infect a host cell to grow and mottling of the leaves of infected tobacco plants and stunts their
reproduce. These electron micrographs illustrate some of the growth. (c) Adenovirus causes eye and respiratory tract infections
structural variety exhibited by viruses. (a) T4 bacteriophage (bracket) in humans. This virus has an outer membranous envelope from
attaches to an E. cofi bacterial cell via a tail structure and injects its DNA, which long glycoprotein spikes protrude. [Part (a) from A. Levine, 1991,
localized in the head, into the cell. Viruses that infect bacteria are called Viruses, Scientific American Library, p. 20. Part (b) courtesy of R. C. Valentine.
bacteriophages, or simply phages. (b) Tobacco mosaic virus causes a Part (c) courtesy of Robley C. Williams, University of California.)

the molecular defects causing the disease can be done and Consider the adenoviruses, which cause eye and respira-
new treatments can be rested, thereby minimizing human ex- tory tract infections in humans. Human adenoviruses have a
posure to untested treatments. genome of only approximately 35,000 base pairs-about 2
percent the size of a bacterial genome-and encode about 30
proteins, about half of which are conserved among adenovi-
Viruses Are Cellular Parasites That Are Widely
ruses that infect different species. These conserved proteins
Employed in Molecular Cell Biology Research comprise structural proteins that form parts of the mature
Virus-caused diseases are numerous and all roo familiar, includ- virus particle (virion) and proteins that catalyze steps in viral
ing chickenpox, influenza, some types of pneumonia, polio, DNA replication. Late in adenovirus infection of human cells,
measles, rabies, hepatitis, the common cold, and many others. the cell becomes a virtual factory for producing just a few
Viral infections in plants (e.g., dwarf mosaic virus in corn) have viral proteins: about half of the non-ribosomal RNAs synthe-
a major economic impact on crop production . Almost all vi- sized are viral mRNAs, and most of the proteins produced are
ruses have a rather limited host range, infecting only certain viral. In the 1970s-before recombinant DNA techniques
bacteria, plants, or animals (Figure 1-23). Viruses arc much were developed-this permitted experiments on adenovirus
smaller than cells, on the order of 100 nanometers (nm ) in di- mRNA synthesis that demonstrated that mature mRNAs had
ameter. A virus is typically composed of a protein coat that undergone splicing, or removal of noncoding sequences (see
encloses a core containing the genetic material, which can be Figure 1-9). Only later was splicing shown to be a fundamen-
either DNA or RNA and carries the information for producing tal part of biogenesis of virtually all eukaryoric mR!\As.
more viruses (Chapter 4). The coat protects a virus from the A different type of virus, vesicular stomatitis virus, makes
environment and allows it to stick to, or to enter, specific host a single glycoprotein (a protein with attached carbohydrate
cells. In some viruses, the protein coat is surrounded by an chain) that is transported to the plasma membrane and then
outer membrane-like envelope that is formed from the plasma forms part of the membrane coat of this virus. Studies of this
membrane of the infected cell (figure 14-34). protein (Figures 14-2 and 14-3 ) elucidated many aspects of
Because viruses ca nnot grow or reproduce on their own, the biogenesis of membrane glycoproteins that were later
a virus must infect a host cell and take over its internal ma- shown to apply to all cellular glycoproteins.
chinery to synthesize viral proteins. All viruses usc cellular Even today viruses are useful in many aspects of molecu-
ribosomes to synthesize viral proteins; most DNA viruses use lar cell biology. Many methods for generically manipulating
cellular enzymes for replication of their DNA and for tran- cells depend on using viruses to convey DNA molecules into
scription of their DNA into mRNA. Thus studies of virus cells. To do this, the portion of the viral generic material that
DNA replication and RNA synthesis are informative of the encodes proteins that are potentially harmful is replaced
corresponding cellular processes. When newly made viruses with other genetic material, including human genes; adeno-
are released by budding from the cell membrane or when the virus is often employed for this purpose. The altered "iruses,
infected cell bursts, the cycle starts anew. or vectors, still can enter cells toting the introduced genes

1.3 Cells into Tissues: Unicellular and Metazoan Organisms Used for Molecular Cell Biology Investigations 21
FIGURE 1-24 The dystrophin glycoprotein Agrin . . Perlecan
Lammm
complex (DGC) in skeletal muscle cells.
Dystrophin-the protein defective in Duchenne Collagen and ot her
- Basal lamina
muscular dystrophy-links the actin cytoskeleton fibrous proteins
to the multi protein sarcoglycan complex in the
plasma membrane. Other protei ns in th e complex
bind to components ofthe basal lamina, such as
laminin, that in turn bind to the collagen fibers Extracellular space
that give the basal lamina strength and rigidity.
Thus dystrophin is an important member of a
group of proteins that links the muscle cell and proteins
its internal actin cytoskeleton with the surrounding
basal lamina. [Adapted from S.J. Winder, 2001, Plasma membrane
Trends Biochem. Sci. 26:118, and D. E. Michele and Cytosol
K. P. Campbell, 2003, J. Bioi. Chern. 278:15457.) T he protein defective
in Du chenne muscular
dyst rophy

Actin

with them {Chapter 5). One day, diseases caused by defec- The Following Chapters Present Much
tive genes may be treated by using viral vectors to introduce Experimental Data That Explains How We Know
a normal copy of a defective gene into patients. Current re-
What We Know About Cell Structure and Function
search is dedicated to overcoming the considerable obstacles
to this gene therapy approach, such as getting the introduced In subsequent chapters of this book we discuss cellu lar pro-
genes to work in the right cells at the correct times. cesses in much greater detail. We begin (Chapter 2) with a .·,
discussion of the chemical nature of the building blocks of
cells and the basic chemical processes required to understand
the macromolecular processes d iscussed in subsequent chap-
Genetic Diseases Elucidate Important
ters. We go on to discuss the structure and function of pro-
Aspects of Cell Function teins (Chapter 3) and how the info rmation for thei r synthesis
Many genetic diseases are caused by mutations in a single pro- is encoded in DNA {Chapter 4). Chapter 5 describes ma ny of
tein; studies on humans with these diseases have shed light on the techniques used to study genes, gene expression, and pro-
the normal function of the protein. As an example, consider tein function. Gene and chromosome structure and the regu-
Duchenne muscular dystrophy {OMD), the most common lation of gene expression are covered in Chapters 6, 7, and 8.
type of hereditary muscle-wasting diseases, collectively called Chapter 9 discusses many of the techniques biologists use to
muscular dystrophies. DMD is an X chromosome-linked dis- culture and fractionate cells and to visualize specific proteins
order, affecting 1 in 3300 boys, that results in cardiac or respi- and structures within cells. Biomembrane structure and trans-
ratory failure, usually in the late teens o r early nventics. The port of ions and small molecules across membranes are the
first clue to understanding the molecular basis of this disease topics of Chapters 10 and 11, and Chapter 12 discusses cel-
came from the discovery that people with OMD carry muta- lular energetics and the func tions of mitOchondria and chlo-
tions in the gene encoding a protein named dystrophin. This roplasts. Membrane biogenesis, protein secretion, and protein
very large protein was later found to be a cyrosolic adapter trafficking- the sorting of proteins tO their correct subcellu-
protein, binding to actin filaments that are part of the cyto- lar destinations-are the topics of Chapters 13 and 14. Chap-
skeleton (see Figure 1-14) and to a complex of muscle plasma ters I 5 and 16 discuss the many types of signals and signal
membrane proteins termed the sarcoglycan complex {Figure receptors used by cells to communicate and regulate their ac-
1-24). The resulting large multiprotein assemblage, the dys- tivities. The cytoskeleton and cell movements are discussed in
trophin glycoprotein complex {DGC), links the extracellular Chapters 17 and 18. Chapter 19 discusses the cell cycle and
matrix protein laminin to the cytoskeleton within muscle and how cell division is regulated. The interactions among cells
other types of cells . Mutations in dystrophin, other DGC and between cells and the extracellula r matrix that enable
components, or laminin can disrupt the DGC-mediated link formation of tissues and organs are detai led in Chapter 20.
between the exterior and interior of muscle cells and cause Later chapters of the book discuss important types of special-
muscle weakness and eventual death. T he first step in identi- ized cells-stem cells (Chapter 2 1), nerve cells {Cha pter 22),
fying the entire dystrophin glycoprotein complex involved and cells of the immune system {Chapter 23 ). Chapter 24
cloning the dystrophin-encoding gene using DNA from normal discusses cancer and the multiple ways in which cell growth
individuals and patients with Duchenne muscular dystrophy. and differentiation can be altered by mutations.

22 CHAPTER 1 • Molecules, Cells, and Evolution

 CHAPTER

Chemical Foundations

Polarized light microscopic image of crystals of cholesterol. Cholesterol

is a water-insoluble molecule that plays a critical structural role in many
memb ranes of anima l cells and is a precursor for the synthesis of steroid
hormones, bile acids, and vitamin D. Excess deposition of cholesterol in
artery walls is a key step in clogging of the arteries, a major cause of
heart attacks and strokes. [Courtesy of National High Magnetic Field
Laboratory/The Florida State University.]

he life of a cell depends on t housands of chemical in- water control the chemistry of life. Life first arose in a wa-

T teractions and reactions exquisitely coordinated with

o ne another in time and space and under the influence
of the cell's genetic instructions and its environment. By un-
tery envi ronment. Constituting 70-80 percent by weight of
most cells, water is the most abundant molecule in biological
systems. It is within this aqueous milieu that small molecules
derstanding at a molecular level these interactions and reac- and ions, which make up about 7 percent of the weight of
tions, we ca n begin to answer fu ndamental questions about living matter, combine into the larger macromolecules and
cell ular life: How does a cell extract nutrients and informa- macromolecu lar assemblies that make up a cell's machinery
tion from its environment? How does a cell convert the en- and architecture and so the remaining mass of organisms.
ergy stored in n utrients into the work of movement or These small molecules include amino acids (the building
metabolism? How does a cell transform nutrients into the blocks of proteins), nucleotides (the building blocks of DNA
cellular components required for its survival? How does a and RNA), lipids (the building blocks of biomembranes),
cell link itself to other cells to form a tissue? How do cells and sugars (the building blocks of complex carbohydrates).
communicate with o ne another so that a complex, efficiently Many of the cell's biomolecules (e.g., sugars) readily dis-
functioni ng organ ism can develop and thrive? One of the solve in water; these molecules are called hydrophilic ("water
goals of Molecular Cell Biology is to answer these and other liking"). Others (e.g., cholesterol) are oily, fatlike substances
questions about the structure and function of cells and or- that shun water; these are said to be h ydrop hobic ("water
ganisms in terms of the properties of individual molecu les fearing") . Still other biomolecules (e.g., phospholipids) con-
and ions. tain both hydrophilic and hydrophobic regions; these mole-
For example, the properties of one such molecule, water, cules are said to be amphipathic ("both liking"). Phospholipids
have controlled and contin ue to control the evolution, struc- are used to bui ld the fie xi ble membranes that enclose cells
ture, and function of all cells. An understanding o f biology is and their in ternal orga nelles. The smooth functioning of
not possible without appreciating how the properties of cells, tissues, and organisms depends on all these molecules,

OUTLINE

2.1 Covalent Bonds and Noncovalent Interactions 24 2.3 Chemical Reactions and Chemical Equilibrium 43

2.2 Chemical Building Blocks of Cells 33 2.4 Biochemical Energetics 48

 {a) Molecular complementarity (b) Chemical building blocks
Protein A

Polymerization
Noncovalent
) interactions

Protein B

Small molecule Macromolecule

subunits

{c) Chemical equilibrium (d) Chemical bond energy

"High-energy"
phosphoanhydride .._-----•
bonds

Adenosine
triphosphate
{ATP)

FIGURE 2-1 Chemistry of life: four key concepts. (a) Molecular chemicals between starting reagents (left) and the products of the
complementarity lies at the heart of all biomolecular interactions, reactions (right) depends on the rate constant s of the forward (k1, upper
as when two proteins with complementary shapes and chemical arrow) and reverse (k,. lower arrow) reactions. The ratio of these, Keq.
properties come together to form a tightly bound complex. (b) Small provides an informative measure of the relative amounts of products
molecules serve as building blocks for larger structures. For example, and reactants that will be present at equilibrium. (d) In many cases, the
to generate the information-carrying macromolecule DNA, four sma ll source of energy for chemical reactions in cells is the hydrolysis of the
nucleotide building blocks are covalently linked into long strings molecule ATP. This energy is released when a high-energy phosphoan-
(polymers), which then wrap around each other to form the double hydride bond linking the~ and-y phosphates in the ATP molecule (red)
helix. (c) Chemical reactions are reversible, and the distribution of the is broken by the addition of a water molecule, forming ADP and P,.

from the smallest to the largest. Indeed, the chemistry of the biological systems, we end the chapter with basic principles
simple proton (H ·) can be as important to the survival of a of biochemical energetics, including the central role of ATP
human cell as that of each gigantic DNA molecule (the mass (adenosine triphosphate) in capturing and transferring en-
of the DNA molecule in human chromosome I is 8.6 X 10 10 ergy in cellular metabolism.
times that of a proton!). The chemical interactions of all of
these molecules, large and small, with water and with one
another, define the nature of life. 2.1 Covalent Bonds and Noncovalent
Luckily, although many t ypes of biomolecu les interact
Interactions
and react in numerous and complex pathways to form func-
tional cells and organisms, a relatively small number of Strong and weak attractive forces between ato ms are the
chemical principles arc necessary to understand cellular pro- "glue" that holds individual molecules together and permits
cesses at the molecular level (Figure 2-1 ). In this chapter we interactions between different mo lecules. When two atoms
review these key principles, some of which you already know share a ~ingl e pa ir uf deurons, the result is a covalent
well. We begin with the covalent bonds that connect atoms bond-a type of strong force that holds a toms together in
into molecules and the noncovalcnt interactions that stabi- molecules. Sharing of multiple pairs of electrons results in
lize groups of atoms within and between molecules. We then multiple covalent bonds (e.g., "double" or "triple" bonds).
consider the basic chemical building blocks of macromole- The weak attractive forces of noncovalent interactions a re
cules and macromolecular assemblies. After reviewing those eq uall y important in determining the properties and func-
aspects of chemical eq uilibrium that are most relevant to tions of biomolecules such as proteins, nucleic acids,

24 CHAPTER 2 • Chemical Foundations

 significance for the shapes and flexibility of biomolecules
such as phospholipids, proteins, and nucleic acids.
Carbon can also bond to four rather than three atoms. As
.· illustrated by methane (CH 4 ), when carbon is bonded to four
other atoms, the angle between any two bonds is 109.5° and
the positions of bonded atoms define the four points of a
tetrahedron (Figure 2-3b). This geometry defines the struc-
H tures of many biomolecules. A carbon (or any other) atom
bonded to four dissimilar atoms or groups in a nonplanar
Methane configuration is said to be asymmetric. The tetrahedral orien-
• H
tation of bonds formed by an asymmetric carbon atom can
be arranged in three-dimensional space in two different ways,
FIGURE 2-2 Covalent bonds form by the sharing of electrons. producing molecules that are mirror images of each other, a
Covalent bonds, the strong forces that hold atoms together into
property called chirality (from the Greek word cheir, mean-
molecules, form when atoms share electrons from their outermost
ing "hand") (Figure 2-4). Such molecules are called optical
electron orbitals. Each atom forms a defined number and geometry
isomers, or stereoisomers. Many molecules in cells contain at
of covalent bonds.
least one asymmetric carbon atom, often called a chiral car-
bon atom. The different stereoisomers of a molecule usually
have completely different biological activities because the ar-
carbohydrates, and lipids. In this section, we first review rangement of atoms within their structures, and thus their
covalent bonds and then disc uss the four major types of non- ability to interact with other molecules, differs.
covalent interactions: ionic bonds, hydrogen bonds, Vander
Waals interactions, and the hydrophobic effect.

(a) Formaldehyde
The Electronic Structure of an Atom
Determines the Number and Geometry
of Covalent Bonds It Can Make
Hydrogen, oxygen, carbon, nitrogen, phosphorus, and sul-
fur are the most abundant elements in biological molecules.
These atoms, which ra rely exist as isolated entities, readily
form covalent bonds, using electrons in the outermost elec-
tron orbitals surrounding their nuclei (Figure 2-2). As a rule,
(b) Methane
each type of atom forms a characteristic number of covalent
bonds with other atoms, with a well-defined geometry deter- H
mined by the atpm's size and by both the distribution of I
H-C-H
electrons around the nucleus and the number of electrons I
that it can share. In some cases, the number of stable cova- H
lent bonds an atom can make is fixed; carbon, for example,
always forms four covalent bonds. In other cases, different Chemical Ball-and ·stick Space-filling
numbers of stable covalent bonds arc possible; for example, structure model model

sulfur can form two, four, or six stable covalent bonds. FIGURE 2-3 Geometry of bonds when carbon is covalently
All the biological building blocks are organized around linked to three or four other atoms. (a) A carbon atom can be
the carbon atom, which forms four covalent bonds. In these bonded to three atoms, as in formaldehyde (CH10). In this case, the
organic biomolecules, each carbon usually bonds to three or carbon-bonding electrons participate in two single bonds and one
four other atoms. (Carbon can also bond to two o t her double bond, which all lie in the same plane. Unlike atoms connected
atoms, as in the linear molecule carbon dioxide, C02 , which by a single bond, which usually can rotate freely about the bond axis,
those connected by a double bond cannot. (b) When a carbon atom
has two carbon-oxygen double bonds (O=C=O); however,
forms four single bonds, as in methane (CH4). the bonded atoms (all H
such bond arrangements of carbon are not found in biologi-
in this case) are oriented in space in the form of a tetrahedron. The
cal building blocks.) As illustrated in Figure 2-3a for formal-
letter representation on the left clearly indicates the atomic composi-
dehyde, carbon ca n bond to three atom~, all in a common tion of the molecule and the bond ing pattern. The ball-and-stick model
.· plane. The carbon atom forms two single bonds with two in the center illustrates the geometric arrangement of the atoms and
atoms and one double bond (two shared electron pairs) with bonds, but the diameters o f the balls representing the atoms and their
the third a tom. In t he absence of other constraints, atoms nonbonding electrons are unrealistically small compared with the
joined by a single bond generally can rotate freely about the bond lengths. The sizes of the electron clouds in the space-filling
bond axis, whereas those connected by a double bond cannot. model on the right more accurately represent the structure in three
The rigid planarity imposed by double bonds has enormous dimensions.

2.1 Covalent Bonds and Noncovalent Interactions 25

 Bonding Properties of Atoms Most
Abundant in Biomolecules

Atom and Usual Number Typical

Outer Electrons of Covalent Bonds Bond Geometry

H H H

·0· 2 ·a· ·.
/ "
·S· 2, 4, or 6
o isomer L isomer
·N· 3 or 4

II
FIGURE 2-4 Stereoisomers. Many molecules in cells contain at least ·P · 5 ,..- P,
one asymmetric carbon atom. The tetrahedral orientation of bonds
formed by an asymmetric carbon atom can be arranged in three- I
dimensional space in two different ways, producing molecules that
C· 4 ,..-c,
are mirror images, or stereoisomers, of each other. Shown here is the
common structure of an amino acid, with its central asymmetric carbon
and four attached groups, including the R group, discussed in Section
2.2. Amino acids can exist in two mirror-image forms, designated Land ATP, contains three phosphate groups (see Section 2.4). A
o. Although the chemical properties of such stereoisomers are summary of common covalent linkages and functional
identical, their biological activities are distinct. Only Lamino acids are groups, which confer distinctive chemica 1 properties to the
found in proteins. molecules of which they are a part, is provided in Table 2-2.

Some drugs are mixtures of the stereoisomers of small

Electrons May Be Shared Equally or Unequally
molecules in which only one stereoisomer has the bio-
logical activity of interest. The use of a pure single stereoiso- in Covalent Bonds
mer of the chemical in place of the mixture may result in a The extent of an atom's ability to attract an electron is called
more potent drug with reduced side effects. For example, one its electronegativity. In a bond between atoms with identical
stereoisomer of the antidepressant drug citalopram (Celexa ) or similar electronegativities, the bonding electrons are es-
is 170 times more potent than the other. Some stereoisomers sentially shared equally between the two atoms, as is the case
have very different activities. Darvon is a pain reliever, for most carbon-carbon single bonds (C-C) and carbon-
whereas its stereoisomer, Novrad (Darvon spelled back- hydrogen single bonds (C- H). Such bonds are called non-
ward ), is a cough suppressant. One stereoisomer of ketamine polar. In many molecules, the bonded atoms have different
is an anesthetic, whereas the other causes hallucinations. • elecrronegativities, resulting in unequal shanng of electrons.
The bond between them is said to be polar.
The typical number of covalent bonds formed hy other One end of a polar bond has a partial negative charge
atoms common to biomolecules is shown in Table 2-1. A (8 ), and the other end has a partial positive charge (8-'-). Jn an
hydrogen atom forms only one covalent bond. An atom of 0 - H bond, for example, the greater electronegativity of the
oxygen usually forms only two covalent bonds but has two oxygen atom relative to hydrogen results in the electrons
additional pairs of electrons that can participate in noncova- spending more time around the oxygen atom than the hydro-
lent interactions. Sulfur forms two covalent bonds in hydro- gen. Thus the 0-H bond possesses an electric dipole, a posi-
gen sulfide (H 2S) but also can accommodate six covalent tive charge separated from an equal but opposite negative
bonds, as in sulfuric acid (H2 S04 ) and its sulfate derivatives. charge. The amount of 8 charge on the oxygen atom of a
Nitrogen and phosphorus each have five electrons to share. 0 - H dipole is approximately 25 percent of that of an elec-
In ammonia (NHl), the nitrogen atom forms three covalent tron, with an equivalent and opposite 8 charge on the H
bonds; the pair of electrons around the atom not involved in atom. A common quantitative measure of the extent of charge
a covalent bond can take part in noncovalent interactions. In separation, or strength, of a dipole is called the dipole mo-
the ammonium ion (NH 4 +), nitrogen forms four covalent ment, IJ., which for a chemical bond is the product of the par-
bonds, which have a tetrahedral geometry. Phosphorus com- tial charge on each atom and the distance between the two
monly forms five covalent bonds, as in phosphoric acid atoms. ror a molecule with multiple dipoles, the amount of
(H lP04 ) and its phosphate derivatives, which form the back- charge separation for the molecule as a whole depends in part
bone of nucleic acids. Phosphate groups covalently attached on the dipole moments of all of its individual chemical bonds
to proteins play a key role in regulating the activity of many and in part on the geometry of the molecule (relative orienta-
proteins, and the central molecule in cellular energetics, tions of the individual dipole moments). Consider the example

26 CHAPTER 2 • Chemical Foundations

 Common Functional Groups and Linkages in Biomolecules

Functio nal Groups

0 0 0
-OH I I
Hydroxyl -C-R - c- -C-O
Acyl carbonyl Carboxyl
(alcohol)
(triacylglycarol) (ketone) (carboxylic acid)

0
II 0 0
+
- o- P o-
- SH -NH 2 or -NH 3 1 -0-P 0 - P-
Sulfhydryl
a-
Amino
Phosphate
o- o-
(Thiol) (amines) Pyrophosphate
(phosphorylated
molecule) (diphosphate)

Linkages

0 I 0
I I -c o c- I
-C-0-C- -N-C-
1
Ether
Ester Amide

of water (H 20), which has two 0-H bonds and thus two moment and the electronic properties of the oxygen and hy-
individual bond dipole moments. If water were a linear mol- drogen atoms allow water to form electrostatic, noncovalent
ecu le with the two bonds on exact opposite sides of the 0 interactions with other waters and with other molecules.
atom, the two dipoles on each end of the molecule would be These interactions play a critical role in almost every biochem-
identical in strength but would be onented in opposite direc- ical interaction in cells and organisms and will be discussed
tions. The two dipole moments would cancel each other and shortly.
the dipole moment of molecule as a whole would be zero. Another important example of polarity is the O=P dou-
However, because water is a V-shaped molecule, with the in- ble bond in H,P04 • In the structure of H 1P0 4 shown on the
dividual dipoles of its two 0-H bonds both pointing toward left below, lines represent single and double bonds and non-
the oxygen, one e'n d of the water molecule (the end with the bonding electrons arc shown as pairs of dots:
oxygen atom) has a partial negative charge and the other end
(the one with the two hydrogen atoms) has a partial positive
H H
charge. As a consequence, the molecule as a whole is a dipole I
with a well-defined dipole moment (Figure 2-5). This dipole 0 0
I .. . I+ ..
H- 0 - P-O - H ~ H-O- P 0 H
· II ·· ·· I
.o o· -

Because of the polarity of the O=P double bond, the H 3P0 4

-I Dipole
moment
can also be represented by the structure on the right, in
which one of the electrons from the P=O double bond has
accumu lated around the 0 atom, giving it a negative
+ charge and leaving the P atom with a positive charge.
These charges are importaut in noncovalent interactiOns.
FIGURE 2-5 The dipole nature of a water molecule. The symbol B Neither of these two models precisely describes the elec-
represents a partial charge (a weaker charge than the one on an electron tronic state of H 3P0 4 . The actual structure can be consid-
or a proton). Because of the difference in the electronegativities of Hand ered to be an intermediate, or hybrid, between these two
0, each of the polar H-Q bonds in water is a dipole. The sizes and representations, as indicated by the double-headed arrow
directions of the dipoles of each of the bonds determine the net distance between them. Such intermediate structures are called res-
and amount of charge separation, or dipole moment, of the molecule. onance hybrids .

2.1 Covalent Bonds and Noncovalent Interactions 27

FIGURE 2-6 Relative energies of covalent Noncovalent interactions Covalent bonds
bonds and noncovalent interactions. Bond
energies are defined as the energy required to
break a particular type of linkage. Shown here are van der Hydrogen
the energies required to break a variety of linkages, Waals bonds
arranged on a log scale. Covalent bonds, including Thermal Hydrolysis of ATP
those for single (C-C) and double (C=C) phosphoanhydride bond c-c c= c
carbon-carbon bonds, are one to two powers of
10 stronger than noncovalent interactions. The
latter are somewhat greater than the thermal 1 1 1
energy of the environment at normal room 0.24 2.4 24 240
temperature (25 °C). Many biological processes are kcal/mol
coupled to the energy released during hydrolysis
of a phosphoanhydride bond in ATP.
Increasing bond strength

Covalent Bonds Are Much Stronger and More contributed by the sodium atom is completely transferred to the
Stable Than Noncovalent Interactions chlorine atom. (Figure 2-7a). Unlike covalent bonds, ionic inter-
actions do not have fixed or specific geometric orientations be-
Covalent bonds are considered ro be strong because the en- cause the electrostatic field around an ion-its attraction for an
ergies required to break them are much greater than the ther- opposite charge--is uniform in all directions. In solid NaCI, op-
mal energy available at room temperature (25 oq or body positely charged ions pack tightly together in an alternating pat-
temperature (37 °C} . As a comequence, they are stable at tern, forming the highly ordered crystalline array typical of salt
these temperatures. For example, the thermal energy at crystals (Figure 2-7b). The energy required to break an ionic
25 oc is approximately 0.6 kilocalorie per mole (kcal/mol ), interaction depends on the distance between the ions and the
whereas the energy required to break the c-c bond in eth- electrical properties of the environment of the ions.
ane is about 140 times larger (Figure 2-6). Consequently, at W hen solid salts dissolve in water, the ions separate from
room temperature (25 °C}, fewer than 1 in 10 12 ethane mol- one another and are stabilized by their interactions with water
ecules is broken into a pair of ·CH1 molecules, each contain- molecules. In aqueous solutions, simple ions of biological sig-
ing an unpaired, nonbonding electron (called a radical ). nificance, such as Na .. , K\ Cal+, Mi .. ,and Cl- , are hydrated,
CO\·alent single bonds in biological molecules have ener- surrounded by a stable shell of water molecules held in place
gies similar to the energy of the C-C bond in ethane. Be- by ionic interactions between the central ion and the oppo-
cause more electrons are shared between atoms in double sitely charged end of the water dipole (Figure 2-7c). Most ionic
bonds, they require more energy to break than single bonds. compounds dissolve readily in water because the energy of hy-
For instance, it takes 84 kcal/mol to break a single C-0 dration, the energy released when ions tightly bind water mol-
bond but 170 !<cal/mol to break a C= O double bond. The ecules and spread out in an aqueous solution, is greater than
most common doub le bonds in biologica l molecules arc the lattice energy that stabilizes the crysta I structure. Parts or
C=O, C=N, C=C, and P= O. all of the aqueous hydration shell must be removed from ions
In contrast, the energy required to break noncovalent inter- when they directly interact with proteins. For example, water
actions is only l-5 kcal/mol, much less than the bond energies of hvdration is lost w hen ions pass through protein pores in the
of covalent bonds (see Figure 2-6). Indeed, noncovalent interac- cell ~embrane during nerve cond uction.
tions are weak enough that they are constantly being formed The relative strength of the interaction between two op-
and broken at room temperature. Although these interactions positely charged ions, A and c-, depends on the concentra-
are weak and have a transient existence at physiological tem- tion of other ions in a solution. The higher the concentration
peratures (25-37 °C}, multiple noncovalent interactions can, as of other ions (e.g., Na + and Cl ), the more opportunities A
we will see, act together to produce highly stable and specific and c+haYe tO interact ionically with these Other ions and
associations between different parts of a large molecule or be- thus the lower the energy required to break the interaction
tween different macromolecules. Protein-protein and protein- between A and c-. As a result, increasing the concentra-
nucleic acid interactions are good examples of noncovalent tions of salts such as NaCl in a solution of biological mole-
interactions. Below, we review the four main types of noncova- cules can weaken and even disrupt th e ionic interactions
lent interactions and then consider their roles in the binding of holding the biomolecules together.
biomolecules to one another and to other molecules.

Ionic Interactions Are Attractions Between Hydrogen Bonds Are Noncovalent

Oppositely Charged Ions Interactions That Determine the Water
Ionic interactions res ult from the attraction of a positively Solubility of Uncharged Molecules
charged ion-a cation-for a negatively charged ion-an anion. A hydrogen bond is the interaction of a partially positively
In sodium chloride (NaCI), for example, the bonding electron charged h ydrogen atom in a molecular dipole, such as water,

28 CHAPTER 2 • Chemical Foundations

 (a) (b) (c)

·.
+ H2 0
--- Cl
Cl dissolving

- - - - Cl Cl Crystallizing

Donation of electron

FIGURE 2-7 Electrostatic interactions of oppositely charged ions balance each other. (c) When the crystals are dissolved in water, the ions
of salt (NaCI) in crystals and in aqueous solution. (a) In crystalline separate and their charges, no longer balanced by immediately adjacent
table salt, sodium atoms are positively charged ions (Na +)due to the ions of opposite charge, are stabilized by interactions with polar water.
loss of one electron each, whereas chloride atoms are correspondingly Water molecules and the ions are held together by electrostatic
negatively charged (CI ) by gaining one electron each. (b) In solid interactions between the charges on the ion and the partial charges on
form, ionic compounds form neatly ordered arrays, or crystals, of tightly the water's oxygen and hydrogen atoms. In aqueous solutions, all ions
packed ions in which the positive and negatively charged ions counter- are surrounded by a hydration shell of water molecules.

with unpaired electrons from another arom, either in the acceptor "pulls" the hydrogen away from the donor. An im-
same or a different molecule. Normally, a hydrogen atom portant feature of all hydrogen bonds is directionality. In the
forms a covalent bond with only one other atom. However, strongest hydrogen bonds, the donor atom, the hydrogen
a hydrogen atom covalently bonded to an electronegative atom, and the acceptor atom all lie in a straight line. Nonlin-
donor atom D may form an additional weak association, the ear hydrogen bonds are weaker than linear ones; still, mul-
hydrogen bond, with an acceptor atom A, which must have tiple nonlinear hydrogen bonds help to stabilize the
a nonbonding pair of electrons available for the interaction: three-dimensional structures of many proteins.
Hydrogen bonds are both longer and weaker than covalent
bonds between the same atoms. In water, for example, the dis-
H8 + · .. < A"-
'---..--' tance between the nuclei of the hydrogen and oxygen atoms of
Hydrogen bond
adjacent, hydrogen-bonded molecules is about 0.27 nm, about
twice the length of the covalent 0-H bonds within a single
The length of the covalent D-H bond is a bit longer than it water molecule (Figure 2-Sa). The strength of a hydrogen bond
would be if there were no hydrogen bond because the between water molecules (approximately 5 kcal/mol) is much

(a) (b) (c) :0 - H· ..

I
H

H
:o
"

H H H H :Q-H · II
I I I I C- N - ·H 0
: O- H ·o - H :Q H· . O-H · I I
H H
H H
I I
H H H :Q-H· :N-CH3 H 0: 0
I I I I II
·H- 0 : ·H- O : : Q- H · H H - C-0-
Water-water Alcohol-water Amine-water Peptide group- water Ester group-water

FIGURE 2-8 Hydrogen bonding of water with itself and with network of hydrogen-bonded molecules. (b) Water also can form
other compounds. Each pair of nonbonding outer electrons in an hydrogen bonds with alcohols and amines, accounting for the high
oxygen or a nitrogen atom can accept a hydrogen atom in a hydrogen solubility of these compounds. (c) The peptide group and ester group,
bond. The hydroxyl and the amino groups can also form hydrogen which are present in many biomolecules, commonly participate in
bonds with water. (a) In liquid water, each water molecule forms hydrogen bonds with water or polar groups in other molecules.
transient hydrogen bonds with several others, creating a dynamic

2.1 Covalent Bonds and Noncovalent Interactions 29

FIGURE 2 -9 Distribution of bonding and outer nonbonding
electrons in t he peptide group. Shown here is a peptide bond linking
two amino acids within a protein called crambin. No protein has been
structurally characterized at higher resolution than crambin. The black
lines represent the covalent bonds between atoms. The red (negative)
and blue (positive) lines represent contours of charge determined
using x-ray crystallography and computational methods. The greater
the number of contour lines, the higher the charge. The high density
of red contour lines between atoms represents the cova lent bonds
(shared electron pairs). The two sets of red contour li nes emanating
from the oxygen (0) and not falling on a covalent bond (black line)
represent the two pairs o f non bonded electrons on the oxygen that are
available to participate in hydrogen bonding. The high density of blue
contour lines near the hydrogen (H) bonded to nitrogen (N) represents
a partial positive charge, indicating that this H can act as a donor in
hydrogen bonding. [From C. Jelsch et al., 2000, Proc. Nat'/. Acad. Sci. USA
97:3171. Courtesy of M. M. Teeter.]

weaker than a covalent 0-H bond (roughly 110 kcal/mol), the electrons of the other. This perturbation generates a
although it is greater than that for many other hydrogen transient dipole in the second atom, and the two dipoles will
bonds in biological molecules (1-2 kcaVmol ). Extensive in- attract each other weakly (Figure 2-1 0). Similarly, a polar
termolecular hydrogen bonding between water molecules ac- covalent bond in one molecule will attract an oppositely ori-
counts for many of its key properties, including its unusually ented dipole in another.
high melting and boiling points and its ability to dissolve Van der Waals interacti ons, involving either transiently
many other molecules. induced or permanent electric dipoles, occur in all types of
The solubility of uncharged substances in an aqueous en- molecules, both polar and nonpolar. ln particular, van der
vironment depends largely on their abili ty to form hydrogen Waals interactions are responsible for the cohesion between
bonds with water. For instance, the hydroxyl group (-OH) nonpolar molecules such as heptane, CH 3-(CH 2 ) 5-CH3 ,
in an alcohol (XCH 2 0H) and the amino group (-NH 2 ) in that cannot form hydrogen bonds or ionic interactions with
amines (XCH 2NH 2 ) can form several hydrogen bonds with each other. The strength of van der Waals interactions
water, allowing these molecules to dissolve in water to high
concentrations (Figure 2-8b). In general, molecules with
polar bonds that easily form hydrogen bonds with water, as
well as charged molecu les and ions that interact with the
dipole in water, can readily dissolve in water; that is, they
are hydrophilic. Many biological molecules contain, in addi-
tion to hydroxyl and amino groups, peptide and ester groups,
which form hydrogen bonds with water via otherwise non -
bonded electrons on their carbonyl oxygens (Figure 2-8c).
X-ray crystallography combined with computational analy-
sis permits an accurate depiction of the distribution of the
outermost unbonded electrons of atoms as well as the elec- Covalent van der Waals
radius radius
trons in covalent bonds, as illustrated in Figure 2-9 . (0.062 nm) (0.14 nm)
FIGURE 2-10 Two oxygen molecules in van der Waals contact. In
Van der Waals Interactions Are Weak Attractive this model, red indicates negative charge and blue indicates positive
Interactions Caused by Transient Dipoles charge. Transient dipoles in the electron clouds of all atoms give rise to
weak attractive forces, called van der Waals interactions. Each type of
When any two atoms approach each other closely, they ere atom has a characteristic van der Waals radius at which van der Waals
ate a weak, nonspecific attractive force called a van der interactions with other atoms are optimal. Because atoms repel one
Waals interaction. These nonspecific interactions result from another if they are close enough together for t heir outer electrons to
the momentary random fluctuations in the distribution of overlap without being shared in a covalent bond, the van der Waals
the electrons of any atom, which give rise to a transient un- radius is a measure of the size of the electron cloud surrounding an
equal distribution of electrons. If two noncovalently bonded atom. The covalent radius indicated here is for the double bond of
atoms are close enough, electrons of one atom will perturb 0=0; the sing le-bond covalent rad ius of oxygen is slightly longer.

30 CHAPTER 2 • Chemical Foundations

 decreases rapidly with increasing distance; thus these nonco- Waters released into bulk
valent interactions can form only when atoms are quite close Nonpolar Highly ordered solution

~0 0
substance water molecules
tO one another. However, if atoms get too close together, the
negative charges of their electrons create a repulsive force.
When the van der Waals attraction between two atoms ex- 0 0
actly balances the repulsion between their two electron
clouds, the atoms arc said to be in van der Waals contact.
The strength of the van der Waals interaction is about 1
kcal/mol, weaker than typical hydrogen bond:. and only
slightly higher than the average thermal energy of molecules
at 25 °C. Thus multiple van der Waals interactions, a van der
Waals interaction together with other noncovalent interac-
tions, or both are required to form stable attractions within
and between molecules.
oo
The Hydrophobic Effect Causes Nonpolar Lower entropy Higher entropy

Molecules to Adhere to One Another FIGURE 2-11 Schematic depiction of the hydrophobic effect.
Cages of water molecules that form around nonpolar molecules rn
Because nonpolar molecules do not contain charged groups, solution are more ordered than water molecules in the surrounding
do not possess a dipole moment, or become hydrated, they bulk liquid. Aggregation of nonpolar molecules reduces the number of
.· are insoluble or almost insoluble in water; that is, they are water molecules involved in highly ordered cages, resulting in a higher-
hydrophobic. The covalent bonds between two carbon entropy, more energetically favorable state (right) compared with the
atoms and between carbon and hydrogen atoms are the most unaggregated state (left).
common nonpolar bonds in biological systems. Hydrocar-
bons-molecules made up only of carbon and hydrogen-
are virtually insoluble in water. Large triacylglycerols (also
known as triglycerides), which make up animal fats and veg- hydrophobic effect and van der Waals interactions is a very
etable oils, also are insoluble in water. As we will see later, powerful tendency for hydrophobic molecules to interact with
the major part of these molecules consists of long hydrocar- one another, not with water. Simply put, like dissolves like.
bon chains. After being shaken in water, triacylglycerols Polar molecules dissolve in polar solvents such as water; non-
form a separate phase. A familiar example is the separation polar molecules dissolve in nonpolar solvents such as hexane.
of oil from the water-based vinegar in an oil-and-vinegar
salad dressing. One well-known hydrophobic molecule is cholesterol
Nonpolar molecules or nonpolar parts of molecules tend (see the structure in Section 2.2). Cholesterol, as well
to aggregate in water owing to a phenomenon called the hy- as triglycerides and other poorly water-soluble molecules, is
drophobic effec~. Because water molecules cannot form hy- called a lipid. Unlike hydrophilic molecules such as glucose
drogen bonds with nonpolar substances, they tend to form or amino acids, lipids cannot readily dissolve in the blood,
"cages" of relatively rigid hydrogen-bonded pentagons and which is an aqueous circulatory system for transporting mol-
hexagons around nonpolar molecules (Figure 2-11, left). ecules and cells throughout the body. Instead, lipids such as
This state is energetically unfavorable because it decreases cholesterol must be packaged into hydrophilic carriers that
the entropy, or randomness, of the population of water mol- can themselves dissolve in the blood and be transported
ecules. (The role of entropy in chemical systems is discussed throughout the body. There can be hundreds to thousands of
in Section 2.4.) If nonpolat molecules in an aqueous environ- lipids packed into the center, or core, of each carrier. The
ment aggregate with their hydrophobic surfaces facing each hydrophobic core is surrounded by amphipathic molecules
other, the net hydrophobic surface area exposed to water is that have hydrophilic parts that interact with water and hy-
reduced (Figure 2-11, right). As a consequence, less water is drophobic parts that interact with each other and the core.
needed to form the cages surrounding the nonpolar mole- The packaging of these lipids into special carriers, called li-
cules, entropy increases relative to the unaggregated state, poproteins (discussed in Chapter 14), permits their efficient
and an energetically more favorable state is reached. In a transport in blood and is reminiscent of the containerization
sense, then, water squeezes the nonpolar molecules into of cargo for efficient long-distance transport via cargo ships,
spontaneously forming aggregates. Rather than constituting trains and trucks. High-densit) lipoprotein (HDL) and low-
an attractive force such as in hydrogen bonds, the hydropho- density lipoprotein (LDL) are two such lipoprotein carriers
bic effect results from an avoidance of an unstable state- that are associated with either reduced or increased heart
that is, extensive water cages around individual nonpolar disease, respectively, and are therefore often referred to as
molecules. "good" and "bad" cholesterol. Actually, the cholesterol
Nonpolar molecules can also associate, albeit weakly, molecules and their derivatives that are carried by both HDL
through van der Waals interactions. The net result of the and LDL arc essentially identical and in themselves neither

2.1 Covalent Bonds and Noncovalent Interactions 31

"good" nor "bad." However, HDL and LDL have different holds the two chains of DNA together in a double helix (see
effects on cells, and as a consequence LDL contributes to Chapter 4). Similar interactions underlie the association of
and HDL protects from the clogging of the arteries (known groups of molecules into multimolecular assemblies, or com-
as atherosclerosis) and consequent heart disease and stroke. plexes, leading, for example, to the fo rmation of muscle fi-
Thus LDL is known as "bad" cholesterol. • bers, to the gluelike associations between cells in solid
tissues, and to numerous other cellular structures.
Depending on the number and strength of the noncova-
Molecular Complementarity Due lent interactions between the two molecules and on their en-
to Noncovalent Interactions Leads to a vironment, their binding may be tight or loose and, as a
Lock-and-Key Fit Between Biomolecules consequence, either long lasting or transient. The higher the
Both inside and outside cells, ions and molecules constantly affinity of two molecules for each other, the better the mo-
collide. The higher the concentration of any two types of lecular "fit" between them, the more noncovalcnt interac-
molecule, the more likely they are to encounter each other. tions can form, and the tighter they can bind together. An
When two molecules encounter each other, they most likely important quantitative measure of affinity is the binding dis-
will simply bounce apart because the noncovalent interac- sociation constant Kd, described later.
tions that would bind them together are weak and have a As we discuss in Chapter 3, nearly all the chemical reac-
transient existence at physiological temperatures. However, tions that occur in cells also depend'on the binding properties
molecules that exhibit molecular complementarity, a lock- of enzymes. These proteins not only speed up, or catalyze,
and-key kind of fit between their shapes, charges, or other reactions but also do so with a high degree of specificity, a
physical properties, can form multiple noncovalent interac- reflection of their ability to bind tightly to only one or a few
tions at close range. When two such structurally complemen- related molecules. The specificity of intermolecular interac-
tary molecules bump into each other, these multiple tions and reactions, which depends on molecular comple-
interactions cause them to stick together, or bind. mentarity, is essential for many processes crirical to life.
rigure 2- 12 illustrates how multiple, different weak in-
teractions can cause two proteins to bind together tightly.
Almost any other arrangement of the same groups on the KEY CONCEPTS of Section 2.1
two surfaces wou ld not allow the molecules to bind so
Covalent Bonds and Noncovalent Interactions
tightly. Such molecular complementarity between regions
within a protein molecule allow it to fold into a unique Covalent bonds consist of pairs of electrons shared by two
three-dimensional shape (see Chapter 3); it is also what atoms. Covalent bonds arrange the atoms of a molecule into
a specific geometry.
Covalent bonds are stable in biological systems because the
relatively high energies required to break them (50-200 kcal/
mol) are much larger than the thermal kinetic energy available
at room (25 °C) or body (37 oq temperatures.
Many molecules in cells contain at least one asymmetric
carbon atom, which is bonded to four dissimilar atoms. Such
molecules can exist as optical isomers (mirror images), des-
ignated o and L (see Figure 2-4), which have different bio-
logical activities. In biological systems, nearly all sugars are
o isomers, whereas nearly all amino acids are L isomers.
• Electrons may be shared equally or unequally in covalent

H3 C- r
ll bonds. Atoms that differ in electronegativity form polar cova-
r_., lent bonds, in which the bonding electrons arc distributed un-
Protein A Protein B Protein A Protein C equally. One end of a polar bond has a partial positive charge
St able complex less stable complex and the other end has a partial negative charge (see figure 2-5 ).
FIGURE 2-12 Molecular complementarity permits tight protein Noncovalenr interactions between atoms are considerably
bondi ng via multiple noncovalent interactions. The complementary weaker than covalent bonds, with energies ranging from
shapes, charges, polarity, and hydrophobicity of two protein surfaces about 1-5 kca l/mol (see Figure 2-6).
permit multiple weak interactions, which in combination produce a
Four main types of noncovalent interactions occur in bio-
strong interaction and tight binding. Because deviations from molecular
complementarity substantially weaken binding, a particular surface
logical systems: ionic bonds, hydrogen bonds, van der Waals
region of any given biomolecule usually can bind tightly to only one or interactions, and interactions due to the hydrophobic effect.
a very limited number of other molecules. The complementarity of the Ionic bonds result from the electrostatic attraction be-
two protein molecules on the left permits them to bind much more tween the positive and negative charges of ions. In aqueous
tightly than the two noncomplementary proteins on the right.

32 CHAPTER 2 • Chemical Foundati ons

 hydrolysis. These bonds linking the monomers together are
solutions, all cations and anions are surrounded by a shell normally stable under normal biological conditions (e.g.,
of bound water molecules (see Figure 2-7c). Increasing the 37 oc, neutral pH), and so these biopolymers are stable and
salt (e.g., NaCI) concentration of a solution can weaken the can perform a wide variety of jobs in cells, such as store infor-
relative strength of and even break the ionic bonds between mation, catalyze chemical reactions, serve as structural ele-
biomolecules. ments in defining cell shape and movement, and many others.
• In a hydrogen bond, a hydrogen atom covalently bonded Macromolecular structures can also be assembled using
to an electronegative atom associates with an acceptor noncovalent interactions. The two-ply, or "bilayer" structure
atom whu:.t: nonbonding electrons attract the hydrogen of cellular membranes is built up by the noncovalent assem-
(see Figure 2-8). bly of many thousands of small molecules called phospholip-
ids (see Figure 2-13). In this chapter, we focus on the chemical
• Weak and relatively nonspecific van der Waals interac-
building blocks making up cells-amino acids, nucleotides,
tions result from the attraction between transient dipoles as-
sugars, and phospholipids. The structure, function, and as-
sociated with all molecules. They form when two atoms ap-
sembly of proteins, nucleic acids, polysaccharides, and bio-
proach each other closely (see Figure 2-10).
membranes are discussed in subsequent chapters.
• In an aqueous environment, nonpolar molecules or non-
polar parts of larger molecules are driven together by the
hydrophobic effect, thereby reducing the extent of their di- Amino Acids Differing Only in Their Side
rect contact with water molecules (see Figure 2-11). Chains Compose Proteins
• Molecular complementarity is the lock-and-key fit be- The monomeric building blocks of proteins are 20 amino
tween molecules whose shapes, charges, and other physical acids, which-when incorporated into a protein polymer-are
properties are complementary. Multiple noncovalenr inter- sometimes called residues. All amino acids have a characteris-
actions can form between complementary molecules, caus- tic structure consisting of a central alpha carbon atom (Co.)
•. ' ing them to bi nd tightly (see Figure 2-12), but not between bonded to four different chemical groups: an amino (NH2 )
molecules that are not complementary. group, a carboxyl or carboxylic acid (COOH) group (hence
the name amino acid ), a hydrogen (H) atom, and one variable
• The high degree of binding specificity that results from
group, called a side chain or R group. Because the a carbon in
molecular complementarity is one of the features that under-
all amino acids except glycine is asymmetric, these molecules
lies intermolecular interactions in biology and thus is essen-
can exist in two mirror-image forms called by convention the
tial for many processes critical to life.
o (dextro) and the L (levo) isomers (see Figure 2-4). The two
isomers cannot be interconverted (one made identical to the
other) without breaking and then re-forming a chemical bond
in one of them. With rare exceptions, only the L forms of
amino acids are found in proteins. However, o amino acids are
2.2 Chemical Building Blocks of Cells prevalent in bacterial cell walls and other microbial products.
A common theme in biology is the construction of large To understand the three-dimensional structures and
macromolecu les and macromolecular structures out of functions of proteins, discussed in detail in Chapter 3, you
smaller molecular subunits. Often these subunits are similar must be familiar with some of the distinctive properties of
or identical. The three main types of biological macromole- amino acids, which are determined in part by their side
cules-proteins, n ucleic acids, and polysaccharides-are all chains. You need not memorize the detailed structure of
polymers composed of multiple covalently linked building each type of side chain to understand how proteins work
block small molecules, or monomers (Figure 2-13). Proteins because amino acids can be classified into several broad cat-
are linear polymers containing around 10 to several thou- egories based on the size, shape, charge, hydrophobicity (a
sand amino acids linked by peptide bonds. Nucleic acids are measure of water solubility), and chemical reactivity of the
linear polymers containing hundreds to millions of nuclco- side chains (Figure 2-14). However, you should be familiar
tides linked by phosphodiester bonds. Polysaccharides are with the general properties of each category.
linear or branched polymers of monosaccharides (sugars) Amino acids with nonpolar side chains are called hydro-
such as glucose linked by glycosidic bonds. Although the ac- phobic and are poorly soluble in water. The larger the non-
tual mechanisms of covalent bond formation between mono- polar side chain, the more hydrophobic the amino acid. The
mers arc complex and will be discussed later, the formation side chains of alanine, valine, leucine, and isoleucine are lin-
of a covalent bond between two monomer molecules usually ear or branc..htJ hydrocarbons that do not form a ring and
involves the net loss of a hydrogen (H) from one monomer are therefore called aliphatic amino acids. These amino acids
and a hydroxyl (OH) from the other monomer-or the net are all nonpolar, as is methionine, which is similar except it
loss of one water molecule-and can therefore be thought of contains one sulfur atom. Phenylalanine, tyrosine, and tryp-
as a dehydration reaction. The breakdown, or cleavage, of tophan have large, hydrophobic, aromatic rings in their side
this bond in the polymer that releases a monomeric subunit chains. In later chapters, we will see in detail how hydrophobic
involves the reverse reaction or the addition of water, called side chains under the influence of the hydrophobic effect

2.2 Chemical Building Blocks of Cell s 33

 MONOMERS POLYMERS

H 0 H 0 H H 0 H H 0 H H 0 H H 0
I II I II I I II I I II I I II I I II
H2 N- C -C- OH
I
R
+ H>-N-C-C- OH
I
R
I
R1
l
H - N-C-C~N-c-c~N-C-C~N-C-C-O H

I
R2
I
R3
I
R 4

Ami no acid peptide bond

Polypeptide

B
~

0-*
, Base
5
HO-f-0 _ _ .
1 II 3'
o- . +
HO- P-0 5' OH
3'
I
HO o-
Nucleotide

glycosidic bond

+
_ _ ~o" \\ H~
o ~n~o~,..,/ OH
HO OH OH 0
Monosaccharide Polysaccharide

Hydrophilic
} head group

Hydrophobic
fatty acyl
tails
Phospholipid bilayer

Phospholipid

FIGURE 2-13 Overview of the cell's principal chemical building polysaccharides from monosaccharides (sugars). Each monomer is
blocks. (Top) The three major types o f b iological macromolecules are covalently linked into t he polymer by a reaction whose net result is loss
each assembled by the polymerization of multiple small molecules of a water molecu le (dehydration). (Bottom) In contrast. phospholip id
(monomers) of a particular type: proteins from amino acids (see monomers noncovalently assemble into a bilayer structure, w h ich
Chapter 3), nucleic acids from nucleotides (see Chapter 4), and forms the basis of all cellular membranes (see Chapter 10).

often pack in the interior of proteins or line the surfaces of shift from being positive ly charged to uncharged depending
proteins that are embedded within hydrophobic regions of on small changes in the acidity of its environment:
biomembranes.
Amino acids with polar side chains are called hydrophilic; I
CH 2
the most hydrophilic of these amino acids is the subset with w--H
side chains that are charged (ionized) at the pH typical of c-- \
biological fluids (-7), both inside and outside the cell (see
II ~C- H
/ c ..._ N
Section 2.3 ). Arginine and lysine have positively charged side H
chains and are called basic amino acids; aspartic acid and pH 7.8
glutamic acid have negatively charged side chains due to the
carboxylic acid groups in their side chains (their charged The activities of man y proteins arc modu lated by shifts in
forms are called aspartate and glutamate) and are called environmenta l acidity (pH) through protonation or depro-
acidic. A fifth amino acid, histidme, has a side chain contain- tonation of histidine !.ide chains. Asparagine and glutamine
ing a ring with two nttrogens, called imidazole, which can are uncharged but have polar side chains containing amide

34 CH APTER 2 • Chemical Foundations

 HYDROPHOBIC AMINO ACIDS

coo- coo- coo- coo- coo- coo- coo- coo-

I I I I I I
+H3 N- C- H +H 3 N C- H +H 3 N C-H +H 3 N-C-H +H3 N-C H +H 3 N - C H +H 3 N C- H +H 3 N - C- H

CH~ CH H C- CH 3 CH 2 CH 2 CH 2
"CH
c5 Qn
H3C / I
3 CH 2 CH CH 2 C= CH

CH3
H3C/ "CH 3 s
\
NH

CH 3 OH

Alanine Valine Isoleucine leucine Methionine Phenylalanine Tyrosine Tryptophan

(Ala orA) (Val orV) (lie or I) (leu or l) (Met or M) {Phe or F) (Tyr orY) (Trp orW)

HYDROPHILIC AMINO ACIDS

Acidic amino acids Polar amino acids with uncharged R groups

Basic amino acids coo- coo- coo-

l I
+H3 N- C- H +H 3 N - C- H +H3 N - C- H
coo- coo- coo-
l
+H 3 N- C- H +H 3 N - C- H
l
+H 3 N- C- H
l CH 2 ~Hl H c ON
I
CH 2 CH 2
coo OH CH
CHl
I I Aspartate Serine Threonine
CH 2 CH 2 C ~ NH (Asp or D) (Ser or S) (Thr orT)
I
CH 2
I
CH 2
I
I.
C-N+
:cH coo- coo-
I
coo-
CH 2 NH H H +H3 N-C - H +H 3 N- C- H +H3 N C- H
I I
. ' NH:f C=NHi
I
CH 2 CH;: CH;:
NH I
CH 2
/
c~ CH 2
lysine Arginine Histidine I Hz".J 0
(lys or K) (Arg orR) (His or H) coo- /
c::,..
H N 0
Glutamate Asparagine Glutamine
SPECIAL AMINO ACIDS (Giu or E) (Asn or N) (Gin or Q)

coo- coo- coo- FIGURE 2-14 The 20 common amino acids used to build
I I
+H 3 N - C-H +H 3 N - C - H c1/ H proteins. The side chain (R group; red) determines the characteristic
+H 2 N / CH 2 properties of each amino acid and is the basis for grouping amino acids
CH 2 H
I H.. c - CH 2 into three main categories: hydrophobic, hydrophilic, and special.
SH Shown are the ionized forms that exist at the pH (= 7) of the cytosol.
Cysteine Glycine Proline In parentheses are the three-letter and one-letter abbreviations for
(Cys or C) (Giy or G) (Pro or P)
.. each amino acid .

groups with extensive hydrogen-bonding capacities. Simi-

larly, serine and threonine are uncharged but have polar hy- I I
N- H N-H
droxyl groups, which also participate in hydrogen bo nds
with other polar molecules. H- C CH 2 SH + HS-CH 2C-H
Lastly, cysteine, glycine, and proline exhibit special roles in I I
C= O C= O
proteins because of the unique properties of their side chains. I I
The side chain of cysteine contains a reactive sulfhydryl group
(-SH). On release of a proton (H ), a sulfhydryl is converted H
into a thiolate anion (S- ). Thiolatc anions can play important I
H- N N- H
roles in catalysis, notably in certain enzymes that destroy pro- I I
teins (proteases). In proteins, each of two adjacent sulfhydryl H- C - CH 2 - S-S-CH C- H
groups can be oxidized, each releasing a proton and an electron, O= C C= O
to form a covalent disulfide bond (-S- S-): I

2.2 Chemical Building Blocks of Cells 35

Disulfide bonds serve to "cross-link" regions within a single
H
polypeptide chain (intramolecular) or between two separate A cetyl lysine N- CH2 -CH 2 -CH 2 -CH 2 -TH-coo-
chains (intermolecular). Disulfide bonds stabilize the folded
NH3+
structure of some proteins. The smallest amino acid, glycine,
0
has a single hydrogen atom as its R group. Its small size al-
lows it to fit into tight spaces. Unlike the other common Phosphoserine "0-P O-CH 2 CH-COO-
amino acids, the side chain of proline (pronounced pro-leen) 0 NH/
bends around to form a ring by covalently bonding to the
nitrogen atom in the amino group attached to the Ca. As a OH
result, proline is very rigid and the amino group is not avail- 3-Hydroxy proline
able for typical hydrogen bonding. The presence of proline
in a protein creates a fixed kink in the polymer chain, limit-
ing how it can fold in the region of the proline residue.
Some amino acids are more abundant in proteins than oth- CH - coo-
HC=C - CH 2
ers. Cysteine, tryptophan, and methionine are not common 3-Methylhistidine I 1
amino acids: together they constitute approximately 5 percent H3 C- N, c <"N NH/
of the ammo acids in a typical protein. Four amino acids- H
leucine, serine, lysine, and glutamic acid-arc the most abun-
-ooc
dant amino acids, totaling 32 percent of all the amino acid -y-Carboxyglut amate CH-CH - CH - COO-
residues in a typical protein. However, the amino acid compo- / 2 I
sitions of proteins can vary widely from these values. -ooc NH +
3

HO~~O\
Humans and other mammals can synthesize I I of the 0-GicNAc-threonine
20 amino acids. The other ni ne are called essential CH 3
amino acids and must be included in the diet to permit nor-
mal protein production. These are phenylalanine, valine,
HO~ NH
I
0 - CH - CH-COO-
I NH/
threonine, tryptophan, isoleucine, methionine, leucine, lysine, O "' C
\
and histidine. Adequate provision of these essential amino CH
acids in feed is key to the livestock ind ustry. Indeed, a geneti-
FIGURE 2 - 15 Common modifications of amino acid side chains
cally engineered corn with high lysine content is now in use
in proteins. These modified residues and numerous others are formed
as an "enhanced" feed to promote the growth of animals. • by addition of various chemical groups (red) to the amino acid side
chains during or after synthesis of a polypeptide chain.
Although cells use the 20 amino acids shown in Figure
2-14 in the initial synthesis of proteins, analysis of cellular
proteins reveals that they contain upward of 100 different residues in membrane receptors, and the -y carboxylation of
amino acids. The difference is due to the chemical modifica- glutamate in blood-clotting factors such as prothrombin.
tions of some of amino acids after they are incorporated into Deamiclation of Asn and Gin to the corresponding acids,
protein by the addition of acetyl groups (CH~CO) and a va- Asp and Glu, is also a common occurrence.
riety of other chemical groups (Figure 2-15). An important Acetylation, addition of an acetyl group, to the amino
modification is the addition of a phosphate (P0 4 ) to hy- group of the N-termina l residue, is the most common form
droxyl groups in serine, threonine, and tyrosine residues, a of amino acid chemical modification, affecting an estimated
process known as phosphorylation. We will encounter nu- 80 percent of all proteins:
merous examples of proteins whose activity is regulated by
0 R 0
reversible phosphorylation and dephosphorylation. Phos- I
phorylation of nitrogen in the side chain of histidine is well CH3 C N-C-C-
known in bacteria, fungi, and plants but less studied- 1 I
H H
perhaps because of the relative instability of phosphorylated Acetylated N -tenninus
histidine-and apparently rare in mammals. The side chains
of asparagine, serine, and threonine are sites for glycosyl- This modification may play an important role in controlling
ation, the attachment of linea r and branched carbohydrate the life span of proteins within cells because many nonacety-
chains. Many secreted proteins and membrane proteins con- lated proteins are rapidly degtadt:J.
tain glycosylated residues, and the reversible modification of
hydroxyl groups on specific serines and threonines by a
Five Different Nucleotides Are Used
sugar called N-acetylglucosamine also regu lates protein ac-
tivities. Other amino acid modifications found in selected to Build Nucleic Acids
proteins include the hydroxylation of proline and lysine resi- Two types of chemically similar nucleic acids, DNA (deoxyri-
dues in collagen (sec Chapter 19), the methylation of histidine bonucleic acid) and RNA (ribonucleic acid), are the principal

36 CHAPTER 2 • Chemical Foundations

(a) (b) PURINES
s·

H~r:O~H
N NH 2
eCH
'"""
9 / N~c'c ...-N'-
'N i)W,t'~ 1 II CH
HC~N,.....c -.... ~
0 OH OH
II 5' Ribose
-o-P-0 -CH H H
I 2
Adenine (A) Guanine (G)
o- 4• H 5'

Phosphate 2 C
HO 0 0.
OH
4 H H ,. PYRIMIDINES
OH H H 0 0 NH 2
Ribose 3' 2

OH H c c c
Adenosine 5'-monophosphate HN;' ' ' CH HN ' c - CH3 N?" ' CH
(AMP) 2-0eoxyribose
1"c , N,.....cH ,PC'- ./CH /Jc,
{7 I
,..... CH
FIGURE 2-16 Common structure of nucleotides. (a) Adenosine 0 0 N 0
5' -monophosphate (AMP), a nucleotide present in RNA. By convention,
H H H
the carbon atoms of the pentose sugar in nucleotides are numbered
Uracil (U) Thymine (T) Cytosine (C)
with primes. In natural nucleotides, the 1' carbon is joined by a 13
linkage to the base (in this case adenine); both the base (blue) and the FIGURE 2 -17 Chemical structures of the princi pal bases in
phosphate on the 5' hydroxyl (red) extend above the plane of the nucleic acids. In nucleic acids and nucleotides, nitrogen 9 of purines
sugar ring. (b) Ribose and deoxyribose, the pentoses in RNA and DNA, and nitrogen 1 of pyrimidines (red) are bonded to the 1' carbon of
respectively. ribose or deoxyribose. U is only found in RNA, and Tis only found in
DNA. Both RNA and DNA contain A, G, and C.

genetic-information-carrying molecules of the cell. The mono- phosphate (see Figure 2-16a); nucleoside diphosphates con-
mers from which DNA and RNA polymers arc built, called tam a pyrophosphate group:
nucleotides, all have a common structure: a phosphate group
linked by a phosphoester bond to a penrose (five-carbon)
sugar that in turn is linked to a nitrogen- and carbon-containing
0 0
ring structure commonly referred to as a base (Figure 2- 16a). II II
In RNA, the penrose is ribose; in DNA, it is deoxyribose that - o-P- 0 - P- 0 -
1
at position 2' has a proton rather than a hydroxyl group (Fig- o- o-
ure 2-16b). (We describe the structures of sugars in more de- Pyrophosphate
tail below.) The bases adenine, guanine, and cytosine (Figure
2-17) are found in both DNA and RNA; thymine is found
only in DNA, and uracil is found only in RNA.
Adenine and guanine arc purines, which contain a pair of and nucleoside triphosphates have a third phosphate.
fused rings; cytosine, thymine, and uracil are pyrimidines, Table 2 -3 lists the names of the nucleosides and nucleotides
which conta in a single ring (see Figure 2-17). The bases in nucleic acids and the various forms of nucleoside phos-
are often abbreviated A, G, C, T, and U, respectively; these phates. The nucleoside triphosphates are used in the synthe-
same single-letter abbreviations are also commonly used to sis of nucleic acids, which we cover in Chapter 4. Among
denote the entire nuclcotitles in nucleic acid polymers. In their other functions in the cell, GTP participates in intracel-
nucleotides, the I ' carbon atom of the sugar (ribose or de- lular signaling and acts as an energy reservoir, particularly in
oxyribo!>e) is attached to the nitrogen at position 9 of a pu- protein synthesis, and ATP, discussed later in this chapter, is
rine (N 9 ) or at position 1 of a pyrimidine (N Jl. The acidic the most widely used biological energy carrier.
character of nucleotides is due to the phosphate group,
which under normal intracellular conditions releases hydro-
gen ions (H -.- ), leaving the phosphate negatively charged (see
Monosaccharides Covalently Assemble into
Figure 2-16a). Most nucleic acids in cells are associated with
proteins, which form ionic interactions with the negatively Linear and Branched Polysaccharides
charged phosphates. The building blocks of the polysaccharides are the simple
Cells and extracellular fluids in organisms contain small sugars, or monosaccharides. Monosaccharides are carbohy-
concentrations of nu cleosides, combinations of a base and a drates, which are literally covalently bonded combinations
sugar without a phosphate. Nucleotides are nucleosides that of carbon and water in a one-to-one ratio (CH20),., where n
have one, two, or three phosphate groups esterified at the 5' equals 3, 4, 5, 6, or 7. Hexoses (11 = 6) and pentoscs (n = 5 )
hydroxyl. Nucleoside monophosphates have a single esterified arc the most common monosaccharides. All monosaccharides

2.2 Chemical Building Blocks of Cells 37

FfJ:i!fii Terminology of Nucleosides and Nucleotides

Purines Pyrimidines
--- --- -- - - - -- -- -
Uracil (U)
Bases Adenine (A) Guanine(G) Cytosine (C) Thymine (T)
--- ----- - r- - - -- --
( in RNA Adenosine Guanosine Cytidine Uri dine
Nucleos1dcs
in DNA Deoxyadenosine Deoxyguanosine Deoxycytidine Deoxythymidine

( in RNA Adenylate Guanylate Cytidylate Uridylate

NucleotJdes
in DNA Deoxyadenylate Deoxyguan>•late Deoxycytidylate Deoxythymidylate

Nucleoside monophosphates AMP GMP CMP UMP

'
Nucleos1de diphosphates ADP GDP CDP UDP

Nucleoside triphosphates ATP GTP CTP UTP

Dcoxynucleoside mono-,
di-, and triphosphates dAMP, etc. dGMP, etc. dCMP, etc. dTMP, etc.

contain hydroxyl (-OH) groups and either an a ldehyde or (a)

a keto group: 6
CH 2 0H
H rf_o 6
CH 2 0H
5 12
H H
0 0 HC~OH
0 H
1
H-C-OH
13 4
I I I · OH H HO - C-H
-C- C- H - C- C-C - H OH HO OH
3 2 H-C~ OH 3 2
Aldehyde Ket o H OH Is H OH
H-C - OH
o-Giucofuranose o·Giucopyranose
(rare I 6 (commonI
CH 2 0H
~1any biologically important sugars are hexoses, including
o-Giucose
glucose, mannose, and galactose (Figure 2-18 ). Mannose is
identical to glucose except that the orientation of the groups (b)
H'- 1 # 0 H, 1# 0
bonded to carbon 2 is reversed. Similarly, galactose, another c~ c~
hexose, differs from glucose only in the orientation of the
groups attached to carbon 4. lnterconversion of glucose and H O-C~ H H-C~OH
13
mannose or galactose requires the breaking and making of HO-C - H HO C~H
covalent bonds; such reactions are carried out by enzymes 14 14
H- C- OH HO- C-H
called epimerases.
o-Giucose (C6 H 12 0 6 ) is the principal external source of H - C~OH H-C~ OH
energy for most cells in complex multicellular organisms and 6 I sl
CH 20H CH2 0H
can exist in three different forms: a linear structure and two o-Man nose o-Galactose
different hemiacetal ring structures (Figure 2-18a ). If the al-
FIGURE 2 -18 Chemical structures of hexoses. All hexoses have the
dehyde group on carbon 1 combines w ith the hydroxyl group
same chemical formula (C6 H120 6 ) and contain an aldehyde or a kPto
on carbon 5, the resulting hemiacetal, o-glu\..opyranu~t:, con- group. (a) The ring forms of o-Giucose are generated from the linear
tains a six-member ring. In the a anomer of o-glucopyranose, molecule by reaction of the aldehyde at carbon 1 with the hydroxyl
the hydroxyl group attached to carbon 1 points "down- on carbon 5 or carbon 4. The three forms are readily interconvertible,
ward" from the ring as shown in Figure 2-l8a; in the 13 ano- although the pyranose form (right) predominates in biological systems.
mer, this hydroxyl points "upward." In aqueous solution, (b) In o-mannose and o-galactose, the configuration of the H (green)
the a and 13 anomers readily interconvert spontaneously; at and OH (blue) bound to one carbon atom differs from that in glucose.
equilibrium there is about one-third a anomer and two-thirds These sugars, like glucose, exist primarily as pyranoses (six-member rings).

38 CHAPTER 2 • Chemical Foundations

~,with very little of the open-chain form. Because enzymes Both glycogen and starch are composed of the a anomer of
can distinguish between the a and ~ anomcrs of o-glucose, glucose. In contrast, cellulo se, the major constituent of plant
these forms have distinct biological roles. Condensation of cell walls that confers stiffness to many plant structures (see
the hydroxyl group on carbon 4 of the linear glucose with its Chapter 19), is an unbranched polymer of the ~ anomer of
aldehyde group results in the formation of o-glucofuranose, glucose. Human digestive enzymes can hydrolyze the a glyco-
a hemiacetal containing a five-member ring. Although all sidic bonds in starch but not the ~ glycosidic bonds in cellu-
three forms of o-glucose exist in biological systems, the py- lose. Many species of plants, bacteria, and molds produce
ranose (six-member ring) form is by far the most abundant. cellulose-degrading enzymes. Cows and termites can break
The pyranose ring in Figure 2-18a is depicted as planar. down cellulose hPcause they harbor cellulose-degrading bacte-
In fact, because of the tetrahedral geometry around carbon ria in their gut. Bacterial cell walls consist of peptidoglycan, a
atoms, the most stable conformation of a pyranose ring has a polysaccharide chain cross-linked by peptide cross-bridges,
nonplanar, chairlike shape. In this conformation, each bond which confers rigidity and cell shape. Human tear and gastro-
from a ring carbon to a nonring atom (e.g., H or 0) is either intestinal fluids contains lysozyme, an enzyme capable of hy-
nearly perpendicular to the ring, referred to as axial (a), or drolyzing peptidoglycan in the bacterial cell wall.
nearly in the plane of the ring, referred to as equatorial (c): The enzymes that make the glycosidic bonds linking mono-
saccharides into polysaccharides are specific for the a or ~
anomer of one sugar and a particular hydroxyl group on the
a
other. In principle, any two sugar molecules can be linked in a
e --~0\ variety of ways because each monosaccharide has multiple hy-

~e
e :
droxyl groups that can pa rticipate in the formation of glyco-
sidic bonds. Furthermore, any one monosaccharide has the
a a potential of being linked to more than two other monosaccha-
Pyranoses a-n-Glucopyranose rides, thus generating a branch point and nonlinear polymers.
Glycosidic bonds are usually formed between the growing
D isaccharides, formed from two monosaccharides, are polysaccharide chain and a covalently modified form of a
the simplest polysaccharides. The disaccharide lactose, com- monosaccharide. Such modifications include a phosphate (e.g.,
posed of galactose and g lucose, is the major sugar in milk; glucose 6-phosphare) or a nucleotide (e.g., UDP-galactose):
the disaccharide sucrose, composed of glucose and fructose,
6 6
is a principal product of plant photosynthesis and is refined CH 2 -0PO/ -
into common table sugar (Figure 2-19).
Larger polysaccharides, containing dozens to hundreds of H H
1
OQ
H,OH
C", ~ ~
monosaccharide units, can fu nction as reservoirs for glucose,
as structural components, or as adhesives that help hold cells HO 0H H O-P- 0 - P- 0 - U ridine
I I
together in tissues. The most common storage carbohydrate in H OH H OH o- o- - - - - '
animal cells is glycogen, a very long, highly branched polymer Glucose 6-phosphat e UDP-gal act ose
of glucose. As much as 10 percent by weight of the liver can
be glycogen. The·primary storage carbohydrate in plant cells, T he epimerase enzymes that interconvert different monosac-
starch, also is a glucose polymer. It occu rs in an unbranched charides often do so using the nucleotide sugars rather than
form (amylose) and lightly branched form (amylopectin). the unsubstituted sugars.

~
CH20~ OH
I
0 4
OH H
l H

H OH FIGURE 2-19 Formation of

Galactose Glucose Lactose the disaccharides lactose and
sucrose. ln any glycosidic linkage,
t he anomeric carbon of one sugar
molecule (in either the u or f3
conformation) is linked to a
hydroxyl oxygen on another sugar
- 0- molecule. The linkages are named
accord ingly; thus lactose contains
a [3(1---+ 4) bond, and sucrose
Glucose Fruct ose Sucrose contains an a(l ---+ 2) bond.

2.2 Chemical Building Blocks of Cells 39

 Fatty acid chains

Hydrophilic head
c
0 ()
'C
Hydrophobic tail ()

Glycerol
PHOSPHATIDYLCHOLINE Choline

FIGURE 2 -20 Phosphatidylcholine, a typical phosphoglyceride. of the fatty acyl side chains in a phosphoglyceride may be saturated
All phosphoglycerides are amphipathic phospholipids, having a or unsaturated. In phosphatidic acid (red), the simplest phospholipid,
hydrophobic tail (yellow) and a hydrophilic head (blue) in which the phosphate is not linked to an alcohol.
glycerol is linked via a phosphate group to an alcohol. Either or both

Many complex polysaccharides contain modified sugars phospholi pids are listed in Table 2-4. Fatty acids often are
that are covalently linked to various small groups, particu- designated by the abbreYiation Gx:y, where xis the number
larly amino, sulfate, and acetyl groups. Such modifications of carbons in the chain andy is the number of double bonds.
are abundant in glycosami noglycans, major polysaccharide Fatty acids containing 12 or more carbon atoms are nearly
components of the extracellular matrix that we describe in insoluble in aqueous solutions because of their long hydro-
Chapter 19. phobic hyd rocarbon chains.
Fatty acids in which all the carbon-carbon bonds are
Phospholipids Associate Noncovalently single bonds, that is, the fatty acids have no carbon-carbon
double bonds, are said to be saturated; those wi th at least
to Form the Basic Bilayer Structure
one ca rbon-carbon double bond are called unsaturated. Un-
of Biomembranes saturated fatty acids with more than one carbon-carbon
Biomembranes are large, flexible sheets with a two-ply, or double bond are referred to as p ol y unsatura ted. Two
bilayer, structure that serve as the boundaries of cells and "essential" polyunsaturated fatty acids, linoleic acid (Cl8:2)
their intracellular organelles and form the outer surfaces of and linolen ic acid (C l 8:3), cannot be synthesized by mam-
some viruses. Membranes literally define what is a cell (the mals and must be supplied in their diet. Mammals can syn-
outer membrane and the contents within the membrane) and thesize other common fatty acids.
what is not (the extracellular space outside the membrane). In phospholipids, fatty acids are covalently attached to
Unlike proteins, nucleic acids, and polysaccharides, mem- another molecule by a type of dehydration reaction called
branes are ass~mbled by the noncoualent association of their esterification, in which the OH from the carboxyl group of
component building blocks. The primary building blocks of the fatty acid and an H from a hydroxyl group on the other
all biomembranes are phospho li pids, whose physical proper- molecule arc lost. In the combined molecule formed by this
ties are responsible for the formation of the sheet-like bi layer reaction, the part deri\ed from the fatty acid is called an acyl
structure of membranes. In addition to phospholipids, bio- group, or fatty acyl group. This is illustrated by the most
membranes can contain a variety of other molecules, includ- common forms of phospholipids: p hosphoglyceridcs, which
ing cholesterol, glycolipids, and proteins. The structure and contain two acyl groups attached to two of the hydroxyl
functions of biomembranes \Viii be described in detail in groups of glycerol (see figure 2-20).
Chapter 10. Here we will focus on the phospholipids in bio- In phosphoglycerides, one hydroxyl group of the glycerol
membrancs. is esterified to phosphate while the other two normally are
To understand the structure of phospholipids, we have to esterified to fatty acids. The simplest phospholipid, phospha-
understand each of their component parts and how they are tidic acid, contains only these components. Phospholipids
assembled. Phospholipids consist of two long-chain, nonpo- such as phosphatidic acids are not only membrane building
lar fatty acid groups linked (usually by an ester bond) to blocks but arc also important signaling mo lecules. Lyso-
small, highly polar groups, including a phosphate and a phosphatidic acid, in which the acyl chain at the 2 position
short organic molecule, such as glycerol (trihydroxy propa- has been removed, is relatively water soluble and can be a
nol) (Figure 2-20). potent inducer of cell division (ca lled a mitogen ). In most
Fatty acid~ cnn~ist of a hydrocarbon (acyl) chain at- phospholipids found in membranes, the phospha te group i~
tached to a carboxyl group (-COOH). Like glucose, fatty also esterified to a hydroxyl group on another hydrophilic
acids are an important energy source for many cells (see compound. In phosphatidylcholine, for example, choline is
Chapter 12). They differ in length, although the predomi- attached to the phosphate (see Figure 2-20). The negative
nant fatty acids in cells have an even number of carbon charge on the phosphate as well as the charged o r polar
atoms, usually 14, 16, L8, or 20. The major fatty acids in groups esterified to it can interact strongly with water. The

40 CHAPTER 2 • Chem1cal Foundations

Fi¥ifH' Fatty Acids That Predominate in Phospholipids

Common Name of Acid (ionized form in parentheses) Abbreviation Chemical Formula

Saturated Fatty Acids

Myristic (myristate) C14:0

Palmitic (palmitate) Cl6:0

Stearic (stearate) Cl8:0

Unsaturated Fatty Acids

---- -- ---- - -- -- - - - - --
Oleic (oleate) C18: 1

Linoleic (linoleate) Cl8:2

Arachidonic (arachidonate) C20:4 CH3 (CH2 )4 (CH=CHCH2 ) 3 CH=CH(CH2 ) 3COOH

phosphate and its associated esterified group, the "head" hydrophobic and hydrophilic regions are called amphipa-
group of a phospholipid, is hydrophilic, whereas the fatty thic. In Chapter 1 0, we will see how the amphipathic proper-
acyl chains, the "tails," are hydrophobic. Other common ties of phospholipids are responsible for the assembly of
phosphoglycerides and associated head groups are shown in phospholipids into sheet-like bilayer biome1pbranes in which
Table 2-5. Molecules such as phospholipids that have both the fatty acyl tails point into the center of the sheet and the
head groups point outward toward the aqueous environ-
ment (see Figure 2-13 ).
Fatty acyl groups also can be covalently linked into other
Common Phosphoglycerides
fatty molecules, including triacylglycerols, or triglycerides,
and Head Groups
which contain three acyl groups esterfied to glycerol:
Common Phosphoglycerides Head Group
- -- -- - - - - -----

Phosphatidylch9line
H3 C- (CH 2 );-C
Choline
0
II
H H3 C (CH ), C
I /H
Triacylglycerol
Phosphatidylethanolamine "'~N~
0 H

Ethanolamine
and covalently attached to the very hydrophobic molecule
cholesterol, an alcohol, to form cholesteryl esters:

Phosphatidylserine

Serine

OH OH
, HO~OH
0
Phospharidylinositol ~0H
1 3

Inositol HO
Cholesterol

2.2 Chemical Building Blocks of Cells 41

 0

0
Cholesteryl ester

Triglyccrides and cholesteryl esters are extremely water- Unsaturated fatty acids or fatty acyl chains with the cis double
insoluble molecules in which fatty acids and cholesterol are bond kink cannot pack as closely together as saturated fatty
either stored or transported . Triglycerides arc the storage acyl chains. Thus, vegetable oils, composed of triglycerides
form of fatty acids in the fat cells of adipose tissue and are with unsaturated fatty acyl groups, usually are liquid at room
the principal components of dietary fats. Cholesteryl esters temperature. Vegetable and similar oils are partially hydroge-
and triglycerides are transported between tissues through the nated to convert some of their unsaturated fatty acyl chains to
bloodstream in specialized carr iers called lipoproteins (sec saturated fatty acyl cha ins. As a consequence, the hydroge-
Chapter 14). nated vegetable oil can be moldcd'into solid sticks of marga-
r ine. A by-product of the hydrogenation reaction is the
We saw above that the fatty acids making up phospho- conversion of some of the fatty acyl chains into trans fatty
lipids (both phosphoglycerides and triglycerides) can acids, popularly called "trans fats." The "trans fats," found in
be either sa turated or unsaturated. An important conse- partially hydrogenated ma rgarine and other food products, arc
quence of the carbon-carbon double bond (C=C) in an not natural. Saturated and trans fatty acids have similar physi-
unsaturated fatty acid is that two stereoisomeric configura- cal properties; for example, they tend tO be solids at room tem-
tions, cis and tram., are possible around each of these bonds: perature. Their consumption, relative to the consumption of
unsaturated fats, is associated with increased plasma choles-
terol levels and is discouraged by some nutritionists. •

Cis Trans
KEY CONCEPTS of Section 2.2
a cis double bond introduces a rigid kink in the otherwise flex-
ible straight acyl chain of a saturated fatty acid (Figure 2-21). Chemical Building Blocks of Cells
In general, the unsaturated fatty acids in biological systems • Macromolecules are polymers of monomer subunits linked
contain only cis. double bonds. Saturated fatty acids without together by covalent bonds via dehydration reactio ns. Three
the kink can pack together tightly and so have higher melting major types of macromolecules are fo und in cells: proteins,
points than unsaturated fatty acids. The main fatty molecules composed of amino acids linked by peptide bonds; nucleic
in butter are triglycerides with saturated fatty acy l chains, acids, composed of n ucleotides linked by phosphodiester
which is why butter is usually solid at room temperature.

H3C H
\/
H/ \ / H
c
H/ \ / H
H/ \
c / H
c
H/ \ / H
/c\ / H
H C
H/ \ / H
H H H H H H H H H H H H H H 0 H/ c\ H H H H H H H O
Hc
I I I I I I I I I I I I I I /c""' I I I I I I I ~
3 -c-c-c-c-c-c-c-c-c-c-c-c-c-c-c H c-c-c-c-c-c-c-c-c
I I I I I I I I I I I I I I 0 I I I I I I I I o
H H H H H H H H H H H H H H H H H H H H H H
Palmitate Oleate
(ionized form of palmitic acid) (ionized form of oleic acid)
FIGURE 2-21 The effect of a double bond on the shape offatty hydrocarbon chain is often linear; the cis double bond in oleate creates
acids. Shown are chemical structures of the ionized form of palmitic a rigid kink in the hydrocarbon chain. [After L. Stryer, 1994, Biochemistry,
acid, a saturated fatty acid with 16 C atoms, and oleic acid, an 4th ed., W. H. Freeman and Company, p. 265.]
unsaturated one with 18 C atoms. In saturated fatty acids, the

42 CHAPTER 2 • Chemica l Foundations

 bonds; and polysaccharides, composed of monosaccharides
(suga rs) linked by glycosidic bonds (see Figure 2-13). Phos- Rate of forward reaction
pholipids, the fourth major chemical building block, assem- (decreases as the concentration of reactants decreases)
ble noncovalently into biomembranes. /

• Differences in the size, shape, charge, hydrophobicity, and

reactivity of the side chains of the 20 common amino acids Chemical equilibrium
(forward and reverse rates are
determine the chemical and structural properties of proteins equal, no change in concentration
(see Figure 2-14). of reactant~ i:lnd products)
·. • T he bases in the nucleotides composing DNA and RNA
arc carbon- and nitrogen-containing rings attached to a pen-
~ Rate of reverse reaction
(increases as the concentration of products increases)
rose suga r. They form two groups: the purines-adenine (A)
and guanine (G)-and the pyrimidines--cytosine (C), thy-
mine (T), and uracil (U) (see Figure 2-17). A, G, T, and Care When reactants are first mixed,
found in DNA, and A, G, U, and C are found in RNA.
• Glucose and other hexoses can exist in three forms: an open-
V initial concentration of products= 0

Time-
chain linear structure, a six-member (pyranose) ring, and a FIGURE 2- 22 Time dependence of the rates of a chemical
five-member (furanose) ring (see Figure 2-18). In biological reaction. The forward and reverse rates of a reaction depend in part
systems, the pyranose form of o-glucose predominates. on the initial concentrations of reactants and products. The net
• Glycosidic bonds arc formed between either the a or the 13 forward reaction rate slows as the concentration of reactants decreases,
anomer of one sugar and a hydroxyl group on another whereas the net reverse reaction rate increases as the concentration of
sugar, leading to formation of disaccharides and other poly- products increases. At equilibrium, the rates of the forward and reverse
reactions are equal and the concentrations of reactants and products
saccharides (see Figure 2-19).
remain constant.
• Phospholipids are amphipathic molecules with a hydro-
phobic tail (often two fatty acyl chains) connected by a small
organic molecule (often glycerol) to a hydrophilic head (see
Figure 2-20).
A Chemical Reaction Is in Equilibrium
• The long hydrocarbon chain of a fatty acid may be satu-
rated (containing no carbon-carbon double bond) or unsatu- When the Rates of the Forward and Reverse
rated (containing one or more double bonds). Fatty sub- Reactions Are Equal
stances such as butter that have primarily saturated fatty When reactants first mix together-before any products
acyl chains tend to be solid at room temperature, whereas have been formed-the rate of the forward reaction to form
unsaturated fats with cis double bonds have kinked chains products is determined in part by their initial concentrations,
that cannot pack closely together and so tend to be liquids at which determine the likelihood of reactants bumping into
room temperature. one another and reacting (Figure 2-22). As the reaction
products accumulate, the concentration of each reactant de-
creases and so does the forward reaction rate. ;v{eanwhile,
some of the product molecules begin to participate in the
reverse reaction, which re-forms the reactants. The ability of
2.3 Chemical Reactions and Chemical a reaction to go "backward" is ca lled microscopic revers-
ilnlity. The reverse reaction is slow at first but speeds up as
Equilibrium
the concentration of product increases. Eventually, the rates
We now shift our discussion to chemical reactions in which of the forward and reverse reactions become equal, so that
bonds, primarily covalent bonds in reactant chemicals, are the concentrations of reactants and products stop changing.
broken and new bonds are formed to generate reaction The system is then said to be in chemical equilibrium (plural:
products. At any one time, several hundred different kinds equilibria).
of chemical reactions are occurring simultaneously in every The ratio of the concentrations of the products to reac-
cell , and many chemicals can, in principle, undergo multi- tants when they reach equilibrium, called the equilibrium
p le chemical reactions. Both the extent to which reactions constant, K.q, is a fixed value. Thus Keq provides a measure
can proceed and the rate at which they take place deter- of the extent to which a reaction occurs by the time it reaches
mine the chemica l composition of cells. In this section, we equilibrium. The rate of a chemical reaction can be increased
discuss the concepts of equilibrium and steady state as well by a catalyst, but a catalyst does not change the equilibrium
as dissociation constants and pH. In Section 2.4, we dis- constant (see Section 2.4 ). A catalyst accelerates the making
cuss how energy influences the extents and rates of chemi- and breaking of covalent bonds but itself is not permanently
ca l reactions. changed during a reaction.

2.3 Chemical Reactions and Chemical Equilibrium 43

The Equilibrium Constant Reflects (a)Test tube equilibrium concentrations
the Extent of a Chemical Reaction ;...88
For any chemical reaction, K.q depends on the nature of the
AAA :;:::::==== 888
888
reactants and products, the temperature, and the pressure
(particularly in reactions involving gases). Under standard (b) Intracellular steady-state concentrations
physical conditions (25 °C and 1 atm pressure for biological
systems), the Keq is always the same for a given reaction, AA
-----"- 8 8 ~ ______,_ cc
......-- 888 ..,...-- cc
whether or not a catalyst is present.
For the general reaction with three reactants and three FIGURE 2-23 Comparison of reactions at equilibrium and steady
products state. (a) In the test tube, a biochemical reaction (A--+ B) eventually
will reach equilibrium, in which the rates of the forward and reverse
aA + bB + cC ~ zZ + yY + xX (2-1) reactions are equal (as indicated by the reaction arrows of equal
length). (b) In metabolic pathways within cells, the product B
where capital letters represent particular molecules or atoms commonly would be consumed, in this example by conversion
and lowercase letters represent the number of each in the to C. A pathway of linked reactions is at steady state when the rate of
reaction formula; the equilibrium constant is given by formation of the intermediates (e.g., B) equals their rate of consump-
tion. As indicated by the unequallengt~ of the arrows, the individual
(2-2) reversible reactions constituting a metabolic pathway do not reach
equilibrium. Moreover, the concentrations of the intermediates at
where brackets denote the concentrations of the molecules at steady state can differ from what they would be at equilibrium.
equilibrium. The rate of the forward reaction (left to right in
Equation 2- 1) is pumped our of the cell. In this more complex situation, the
Rateforward = kf[A ]a[B ]b[C]c original reaction can never reach equilibrium because some
of the products do not have a chance to be converted back to
where k1 is the rate constant for the forward reaction. Similarly, reactants. Nevertheless, in such non-equilibrium conditions
the rate of the reverse reaction (right to left in Equation 2-1) is the rate of formation of a substance can be equal to the rare
Ratercvcrse = k,[X]'[YY[Z ]' of its consumption, and as a consequence the concentration
of the substance remains constant over time. In such circum-
where k, is the rate constant for the reverse reaction. It is im- stances, the system of linked reactions for producing and con-
portant to remember that the forward and reverse rates of a suming that substance is said to be in a steady state (Figure
reaction can change because of changes in reactant or product 2-23). One consequence of such linked reactions is that they
concentrations, yet at the same time the forward and reverse prevent the accumulation of excess intermediates, protecting
rate constants do not change; hence the name "constant." cells from the harmful effects of intermediates that are toxic
Confusing rates and rate constants is a common error. At equi- at high concentrations. When the concentration of a product
librium the forward and reverse rates are equal, so Raterorward/ of an ongoing reaction is not changing over time, it might be
Raterever;.e = 1. By rearranging these equations, we can express a consequence of a state of equilibrium or it might be a con-
the equilibrium constant as the ratio of the rate constants: sequence of a steady stare. In biological systems when me-
kt tabolite concentrations, such as blood glucose levels, are nor
K.q = k (2-3) changing with time-a condition called homeostasis-it is a
r
consequence of a steady state rather than equilibrium.
The concept of Keq is particularly helpful when we want to
think about the energy that is released or absorbed when a Dissociation Constants of Binding Reactions
chemical reaction occurs. We will discuss this in consider- Reflect the Affinity of Interacting Molecules
able detail in Section 2.4.
The concept of equilibrium also applies to the binding of one
molecule to another. Many important cellular processes de-
Chemical Reactions in Cells Are at Steady State pend on such binding "reactions," which involve the making
Under appropriate conditions and given sufficient time, a and breaking of various noncovalent interactions rather than
single biochemical reaction carried out in a test tube eventu- covalent bonds, as discussed above. A common example is
ally will reach equilibrium and the concentration of reactants the binding of a ligand (e.g., the hormone insulin or adrena-
and products does not change with time because the rates of line) to its receptor on the surface of a cell, which triggers an
the forward and reverse reactions are equal. Within cells, intracellular signaling pathway (see Chapter 15 ). Another
however, many reactions are linked in pathways in which a example is the binding of a protein to a specific sequence of
product of one reaction has alternative fates to simply recon- base pairs in a molecule of DNA, which frequently causes
verting via a reverse reaction to the reactants and thus ulti- the expression of a nearby gene to increase or decrease (see
mately reaching equilibrium. For example, the product of one Chapter 7). If the equilibrium constant for a binding reac-
reaction might serve as a reactant in another, or it might be tion is known, the stability of the resulting complex can be

44 CHAPTER 2 • Chemical Foundations

FIGURE 2 -24 Macromolecules can have distinct binding sites for Multiligand binding macromolecule (e.g., protein)
multiple ligands. A large macromolecule (e.g., a protein, blue) with
three distinct binding sites (A-C) is shown; each binding site exhibits
molecular complementarity to three different binding partners (ligands
A-C) with distinct dissociation constants (KdA-cl·

Ligand B
(e.g., small
predil:tc:u. To illustrate the general approach for determining molecule)
the concentration of noncovalently associated complexes,
we will calculate the extent to which a protein (P) is bound
to DNA (D), forming a protein-DNA complex (PD):
P+ D ~ PD
Most commonly, binding reactions are described in terms of
the dissociation constan t Kd, which is the reciprocal of the
equilibrium constant. For this binding reaction, the dissocia-
tion constant is given by
[P][ DJ
K - (2-4)
d- [ PO ]

It is worth noting that in such a binding reaction, when half

of the DNA is bound to the protein ([PO] = [D]), the concen- Biological Fluids Have Characteristic pH Values
tration of Pis equal to the Kd. The lower the K0 , the lower the
concentration of P needed to bind to half of D. In other The solvent inside cells and in all extracellular fluids is water.
words, the lower the Kd, the tighter the binding (the higher An important characteristic of anr aqueous solution is the con-
the affinity) of P for D. Typical reactions in which a protein centration of positively charged hydrogen ions (H+) and nega-
binds to a specific DNA sequence have a Kd of 10 10 M, tively charged hydroxyl ions (OH ). Because these ions are the
where M symbolizes molarity, or moles per liter (moi/L). To dissociation products of H 2 0, they arc constituents of all li ving
relate the magnitude of this dissociation constant to the in- systems, and they arc liberated by many reactions that take
tracellular ratio of bound to unbound DNA, let's consider place between organic molecules within cells. These ions also
the simple example of a bacterial cell having a volume of can be transported into or out of celh., as when highly acidic
1.5 X 10 15 Land containing 1 molecule of DNA and 10 gastric juice is secreted by cells lining the walls of the stomach.
molecules of the DNA-binding protein P. In this case, given a When a water molecule dissociates, one of its polar H- 0
Kd of 10 10 M and the total concentration of the Pin the cell bonds breaks. The resulting hydrogen ion, often referred to
(- 111 X 10- 10 M, 100-fold higher than the Kd), 99 percent as a proton, has a short lifetime as a free ion and quickly
of t he time this specific DNA sequence will have a molecule combines with a water molecule to form a hydronium ion
of protein bound to it and 1 percent of the time it will not, (H 3 0 ~). For convenience, we refer to the concentration of
even though the cell contains only 10 molecules of the pro- hydrogen ions in a solution, [H +I, even though this really
tein! Clearly, P and D have a high affinity for each other and represents the concentration of hydronium ions, [H10 ].
bind tightly, as reflected by the low value of the dissociation Dissociation of H 2 0 generates one OH ion along with each
constant for their bindillg reaction. For protein-protein and H.._. The dissociation of water is a reversible reaction:
protein-DNA bind ing, Kd values of ::510 9 M (nanomolar) H20 ~ H + OH
are considered to be tight, -10 6 M (micromolar) modestly
tight, and - 10 3 M (millimolar) relatively weak. At 25 °C, [H+][OH l = 10 14 M 2 , so that in pure water,
A large biological macromolecule, such as a protein, can IH. ] = [OH-] = 10 ~ M.
have multiple binding surfaces for binding several molecules The concentration of hydrogen ions in a solution is ex-
simultaneously (Figure 2-24). In some cases, these binding pressed conventionally as its pH, defined as the negative log
reactions are independent, with their own distinct K..1 values of the hydrogen ion concentration. The pH of pure water at
that are constant. In other cases, binding of a molecule at 25 oc is 7:
om: :,itc: on a macromolecu le can change the three-dimen- 1 1
sional shape of a distant site, thus altering the binding inter- pH= - log[H ] = log[ :;-= log 7 = 7
H ] 10
actions of that distant site with some other molecule. This is an
important mechanism by which one molecule can alter, and It is important to keep in mind that a I unit difference in pH
thus regu late, the binding activity of another. We examine represents a tenfold difference in the concentration of pro-
this regulatory mechanism in more detail in Chapter 3. tons. On the pH scale, 7.0 is considered neutral: pH values

2.3 Chemical React1ons and Chemical Equilibrium 45

 Increasingly basic Hydrogen Ions Are Released by Acids
(lower H concentration)
and Taken Up by Bases
In general, an acid is any molecule, ion, or chemical group
pH scale
that tends to release a hydrogen ion (H ), such as hydrochlo-
ric acid (HCI ) or the carboxyl group (-COOH ), which
f - - - - - - + -- - 1 4 Sodium hydroxide (1 N) tends to dissociate to form the negatively charged carboxyl-
~--------~--13 ate ion (-COO - ). Likewise, a base is any molecule, ion, or
Household bleach chemical group that readily combines with a H , such as the
f -- - - - -+---12
Ammonia (1 N)
~--------~---11
hydroxyl ion (OH-); ammonia (NH 3), which forms an am-
monium ion (N H 4 ~);or the amino group (-NH 1 ).

:------~-- ~} Seawater
1
When acid is added to an aqueous solution, the [H +l in-
creases and the pH goes down. Conversely, when a base is
8 ____/ Interior of cell added to a solution, the IH +] decreases and the pH goes up.
7
Fertilized egg Because [H~][OH ] = 10 14 M 2 , any increase in fH j is cou-
- - - - Unfertilized egg pled with a commensurate decrease in [OH- ] and vice versa.
6 Urine
Many biological molecules contain both acidic and basic
5 groups. For example, in neutral solutions (pH = 7.0), many
Interior of the lysosome
4 amino acids exist predominantly in the doubly ionized form,
3 Grapefruit juice in which the carboxyl group has lost a proton and the amino
group has accepted one:
2
Gastric juice NH +
0 Hydrochloric acid (1 N)
I 3
H-c-coo-
R

where R represents the uncharged side chain. Such a mol-

Increasingly acidic
(greater H concentration)
ecule, containing an equal number of positive and nega-
tive ions, is called a zwitterion. Zwitterions, having no net
FIGURE 2-25 pH values of common solutions. The pH of an
charge, are neutral. At extreme pH values, only one of these
aqueous solution is the negative log ofthe hydrogen ion concentra·
two ionizable groups of an amino acid will be charged.
tion. The pH values for most intracellular and extracellular biological
The dissociation reaction for an acid (or acid group in a
fluids are near 7 and are carefully regulated to permit the proper
functioning of cells•. organelles, and cellular secretions.
larger molecule) HA can be written as HA ~ H + + A .
The equilibrium constant for this reaction, denoted Ka (the
subscript a stands for "acid"), is defined as Ka = [H+lfA ]/
below 7.0 indicate acidic solutions (higher [H ' ]),and values [HA]. Taking the logarithm of both sides and rearranging
above 7.0 indicate basic, or alkaline, solutions (Figure 2-25). the result yields a very useful relation between the equilib-
For instance, gastric juice, which is rich in hydrochloric acid rium constant and pH:
(HCI), has a pH of about 1. Its [H+] is roughly a millionfold
greater than that of cytoplasm, with a pH of about 7.2. [A ]
pH= pK. + log [HA] (2-5)
Although the cytosol of cells normally has a pH of about
7.2, the interior of certain organelles in eukaryotic cells (see
Chapter 9 ) can have a much lower pH. Lysosomes, for ex- where pK3 equals - log K,.
ample, have a pH of about 4.5. The many degradative en- From this expression, commonly known as the Henderson-
zymes withm lysosomes function optimally in an acidic Hasselbalch equation, it can be seen that the pK, of any acid
environment, whereas their action is inhibited in the near is equal to the pH at which half the molecules are dissoci-
neutral environment of the cytoplasm. As this example ated and half are neutral (u ndissociated). This is because
illustrates, maintenance of a specific pH is essential for when fA ] = fHAj, then log ([A ]/[HA]) = 0, and thus
proper functioning of some cellular structures. On the other pK~ = pH. The Henderson-Hasselbalch equation allows us
hand, dramatic shifts in cellular pH may play an important to calculate the degree of dissociation of an acid if both the
role in controlling cellular activity. For example, the pH of pH of the solution and the pKa of the acid are known. Ex-
the cytoplasm of an unfertilized egg of the sea urchin, an perimentally, by measuring the fA J and fHA] as a function
aquatic animal, is 6.6. Within 1 minute of fertilization, how- of the solution's pH, one can calculate the pK" of the acid
ever, the pH rises to 7.2; that is, the [H+] decreases to about and thus the equilibrium constant K , for the dissociation
one-fourth its original value, a change necessary for subse- reaction (figure 2-26). Knowing the pK, of a molecule not
quent growth and division of the egg. only provides an important description of its properties but

46 CHAPTER 2 • Chemical Foundations

 H 2 C03 ~ HC03 - + W 8

H2 C0 3
CH 3COOH ~ CH 3 COO + H
"~
::;
'()
100
0) (.)
6
(.) Q)
·c:o
_gE
~ Q)
0)-
(.) 0) 50
-c :X:
0 0
_.o c. 4
c ~
Q) 0)
o.:?
a;.o
a..~
0 00 2 4 6 7.4 8
2
pH
FIGURE 2-26 The relationship between pH, pK., and the
dissociation of an acid. As the pH of a solution of carbonic acid rises
from 0 to 8.5, the percentage of the compound in the undissociated,
or non-ionized, form ( H2 C03 ) decreases from 100 percent and that of 0 0 .2 0.4 0.6 0.8 1.0
the ionized form increases from 0 percent. When the pH (6.4) is equal Fraction of dissociated CH 3COOH
to the acid's pK., half of the carbonic acid has ionized. When the Added OW ---7
pH rises to above 8, virtually all of the acid has ionized to the
F IGURE 2 - 27 The titration curve of the buffer acet ic acid
bicarbonate form (HC03 - ).
(CH 3COOH). The pK. for the dissociation of acetic acid to hydrogen and
acetate ions is 4.75. At this pH, half the acid molecules are dissociated.
Because pH is measured on a logarithmic scale, the solution changes
from 91 percent CH 3COOH at pH 3.75 to 9 percent CH3COOH at pH
also allows us to exploit these properties to manipulate the
5.75. The acid has maximum buffering capacity in this pH range.
acidity of an aqueous solution and to understand how bio-
logical systems control this critical characteristic of their
aqueous fluids.
depends on the concentration of the buffer and the relation-
ship between its pK. value and the pH, which is expressed by
the Henderson-Hasselbalch equation.
Buffers Maintain the pH of Intracellular
The titration curve for acetic acid shown in Figure 2-27
and Extracellular Fluids illustrates the effect of pH on the fraction of molecules in the
A living, actively metabolizing cell must maintain a constant un-ionized (HA ) and ionized forms (A ). At one pH unit
pH in the cytoplasm of about 7.2-7.4 despite the metabolic below the pK. of an acid, 91 percent of the molecules are in
production of many acids, such as lactic acid and carbon the HA form; at one pH unit above the pK., 91 percent are
dioxide; the latter reacts with water to form carbonic acid in the A form. At pH values more than one unit above or
(H 2 C0 3 ). Cells nave a reservoir of weak bases and weak below the pK, the buffering capacity of weak acids and
acids, called buffers, which ensure that the cell's cytoplasmic bases declines rapidly. In other words, the addition of the
pH remains relatively constant despite small fluctuations in same number of moles of acid to a solution containing a
the amounts of H+ or OH being generated by metabolism mixture of HA and A- that is at a pH near the pK. will cause
or by the uptake or secretion of molecules and ions by the less of a pH change than it would if the HA and A- were not
cell. Buffers do this by "soaking up" excess H -+ or OH - present or if the pH were far from the pK3 value.
when these ions are adde.d to the cell or are produced by All biological systems contain one or more buffers. Phos-
.· metabolism .
If additional acid (or base) is added to a buffered solu-
phate ions, the ionized forms of phosphoric acid, are present
in considerable quantities in cells and are an important fac-
tion whose pH is equal to the pK. of the buffer ([HAJ = tor in maintaining, or buffering, the pH of the cytoplasm.
!A- )), the pH of the solution changes, but it changes less Phosphoric acid (H 3 P04 ) has three protons that are capable
than it would if the buffer had not been present. This is be- of dissociating, but they do not dissociate simultaneously.
cause protons released by the added acid arc taken up by the Loss of each proton can be described by a discrete dissocia-
ionized form of the buffer (A-); likewise, hydroxyl ions gen- tion reaction and pK3 , as shown in Figure 2-28. The titration
erated by the addition of base are neutralized by protons curve for phosphoric acid shows that the pK. for the disso-
released by the undissociated buffer (HA). The capacity of a ciation of the second proton is 7.2. Thus at pH 7.2, about
substance to release hydrogen ions or take them up depends 50 percent of cellular phosphate is H 2P04 and about 50 per-
partly on the extent to which the substance has already taken cent is HPO/ - according to the Henderson-Hasselbalch
up or released protons, which in turn depends on the pH of equation. For this reason, phosphate is an excellent buffer at
the solution relative to the pK. of the substance. The ability pH values around 7.2, the approximate pH of the cytoplasm
of a buffer to minimize changes in pH, its buffering capacity, of cells, and at pH 7.4, the pH of human blood.

2.3 Chemical Reactions and Chemical Equilibrium 47

 14
formed between the molecules (e.g., ligand-receptor or pro-
12 tein-DNA complexes).
• The pH is the negative logarithm of the concentration of
10 hydrogen ions (-log [H ]). The pH of the cytoplasm is nor-
mally about 7.2-7.4, whereas the interior of lysosomes has a
8 pH of about 4.5.
pK3 = 7.2
I
a. • Acids release protons (H · ), and bases bind them. In bio-
6 logical molecules, the carboxyl (-COOH) and phosphoryl
groups (-H 2P04 ) are the most common acidic groups; the
amino group (-NH2 ) is the most common basic group.
4
• Buffers are mixtures of a weak acid (HA) and its correspond-
2 ing base form (A-), which minimize the change in pH of a solu-
tion when acid or base is added. Biological systems use various
buffers to maintain their pH within a very narrow range.
0
Added OH ------7
FIGURE 2-28 The titration curve of phosphoric acid (H3 P04 ),
a common buffer in biolog ical syste ms. This biologically ubiquitous
molecule has three hydrogen atoms that dissociate at different 2.4 Biochemical Energetics
pH values; thus phosphoric acid has three pK. values, as noted on the The transformation of energy, its storage, and its use are
graph. The shaded areas denote the pH ranges-within one pH unit of
central to the economy of the cell. Energy may be defined as
the three pK. values-where the buffering capacity of phosphoric acid
the ability to do work, a concept as applicable to cells as to
is high. In these regions, the addition of acid (or base) will cause
automobile engines and electric power plants. The energy
relatively small changes in the pH.
stored within chemical bonds can be harnessed to support
chemical work and the physical movements of cells. In this
section, we will review how energy influences the extents of
KEY CONCI=PTS of Sectiol' 2.3 chemical reactions, a discipline ca lled chemical thermody-
namics, and the rates of chemical reactions, a discipline
Chemical Reactions and Chemical Equilibrium
called chemical kinetics.
A chemical reaction is at equilibrium when the rate of the
forward reaction is equal ro the rate of the reverse reaction,
and thus there is no net change in the concentration of the Several Forms of Energy Are Important
reactants or products. in Biological Systems
• The equilibrium constant Kcq of a reaction reflects the ra- There are two principal forms of energy: kinetic and poten-
tio of products to reactants at equilibrium and thus is a mea- tial. Kinetic energy is the energy of movement-the motion
sure of the extent of the reaction and the relative stabilities of molecules, for example. Potential energy is stored energy-
of the reactants and products. the energy stored in covalent bonds, for example. Potential
• The K<q depends on the temperature, pressure, and chemi- energy plays a particularly important role in the energy
cal properties of the reactants and products but is indepen- economy of cells.
dent of the reaction rate and of the initial concentrations of Thermal energy, or heat, is a form of kinetic energy-the
reactants and products. energy of the motion of molecules. For heat to do work, it
must flow from a region of higher temperature-where the
For any reaction, the equilibrium constant Ke4 equals the
average speed of molecular motion is greater-to one of
ratio of the forward rate constant to the reverse rate con-
lower temperature. Although diffe rences in temperature can
stant (k 1/k,). The rates of conversion of reactants to products
exist between the internal and external environments of
and vice versa depend on the rate constants and the concen-
cells, these thermal gradients do not usually serve as the
trations of the reactants or products.
source of energy for cellular activities. The thermal energy in
• Within cells, the linked reactions in metabolic pathways warm-blooded animals, which have evolved a mechanism
general ly are at steady state, not equilibrium, at which rate for thermoregulation, is used chiefly to maintain constant
of formation ot the intermediates equals their rate of con- organismic temperatures. This is an important homeostatic
sumption (see Figure 2-23) and thus the concentrations of function because the rates of many cellular activities are tem-
the intermediates are not changing. perature dependent. For example, cooling mammalian cells
• The dissociation constant Kd for the noncovalent binding from their normal body temperature of 37 oc to 4 oc can
of two molecules is a measure of the stability of the complex virtually "freeze" or stop many cellular processes (e.g., intra-
cellular membrane movements).

48 CHAPTER 2 • Chemical Foundations

 Radiant energy is the kinetic energy of photons, or waves conversions are very important in biology. In photosynthesis,
of light, and is critical to biology. Radiant energy can be for example, the radiant energy of light is transformed into
converted to thermal energy, for instance when light is ab- the chemical potential energy of the covalent bonds between
sorbed by molecules and the energy is converted to molecu- the atoms in a sucrose or starch molecule. In muscles and
lar motion. Radiant energy absorbed by molecules can also nerves, chemical potential energy stored in covalent bonds is
change the electronic structure of the molecules, moving transformed, respectively, into the kinetic energy of muscle
electrons into higher-energy orbitals, whence it can later be contraction and the electric energy of nerve transmission. In
recovered to perform work. For example, during photosyn- all cells, potential energy-released by breaking certain chem-
thesis, light energy absorbed by pigment molecules such as ical bonds-is used to gener:ue potential energy in the form of
chlorophyll is subsequently converted into the energy of concentration and electric potential gradients. Similarly, en-
chemical bonds (see Chapter 12 ). ergy stored in chemical concentration gradients or electric po-
Mechanical energy, a major form of kinetic energy in bi- tential gradients is used to synthesize chemical bonds or to
ology, usually results from the conversion of stored chemical transport molecules from one side of a membrane to another to
energy. For example, changes in the lengths of cytoskeletal generate a concentration gradient. The latter process occurs
filaments generate forces that push or pull on membranes during the transport of nutrients such as glucose into certain
and organelles (see Chapters 17 and 18 ). cells and transport of many waste products out of cells.
Electric energy-the energy of moving electrons or other Because all forms of energy are interconvertible, they can
charged particles-is yet another major form of kinetic en- be expressed in the same units of measurement. Although
ergy, one with particular importance tO membrane function, the standard unit of energy is the joule, biochemists have
such as in electrically active neurons (see Chapter 22). traditionally used an alternative unit, the calorie ( 1 joule =
Several forms of potential energy are biologically signifi- 0.239 calorie). A calorie is the amount of energy required to
cant. Central to biology is chemical potential energy, the en- raise the temperature of one gram of water by 1 oc. Through-
ergy stored in the bonds connecting atoms in molecules. out this book, we use the kilocalorie to measure energy
Indeed, most of the biochemical reactions described in this changes (1 kcal = 1000 cal). When you read or hear about
book involve the making or breaking of at least one covalent the "Calories" in food (note the capital C), the reference is
chemical bond. We recognize this energy when chemicals un- almost always to kilocalories as defined hete.
dergo energy-releasing reactions. For example, the high poten-
tial energy in the covalent bonds of glucose can be released by The Change in Free Energy Determines If a
controlled enzymatic combustion in cells (see Chapter 12 ).
Chemical Reaction Will Occur Spontaneously
This energy is harnessed by the cell to do many kinds of work.
A second biologically important form of potential energy is Chemical reactions can be divided into two types depending
the energy in a concentration gradient. When the concentration on whether energy is absorbed or released in the process. In
of a substance on one side of a barrier, such as a membrane, is an exergonic ("energy-releasing") reaction, the products
different from that on the other side, a concentration gradient contain less energy than the reactants. Exergonic reactions
exists. All cells form concentration gradients between their inte- take place spontaneously. The liberated energy is usually re-
rior and the external fluids by selectively exchanging nutrients, leased as heat (the energy of molecular motion), and gener-
waste products, tmd ions with their surroundings. Also, organ- ally results in a rise in temperature, such as in the oxidation
elles within cells (e.g., mitochondria, lysosomes) frequently con- (burning) of wood. In an endergonic ("energy-absorbing")
tain different concentrations of ions and other molecules; the reaction, the products contain more energy than the reac-
concentration of protons within a lysosome, as we saw in the tants and energy is absorbed during the reaction. If there is
last section, is about 500 times that of the cytoplasm. no external source of energy to drive an endergonic reaction,
A third form of potential energy in cells is an electric it cannot take place. Endergonic reactions are responsible
potential-the energy o~ charge separation. For instance, for the ability of instant cold packs often used to treat inju-
there is a gradient of electric charge of - 200,000 volts per ries to rapidly cool below room temperature. Crushing the
em across the plasma membrane of virtually all cells. We pack mixes the reagents, initiating the reaction.
discuss how concentration gradients and the potential differ- A fundamentally important concept in understanding if a
ence across cell membranes are generated and maintained in reaction is exergonic or endergonic, and therefore if it occurs
Chapter 11 and how they are converted to chemical poten- spontaneously or not, is free energy, G, named after J. W.
tial energy in Chapter 12. Gibbs. Gibbs, who received the first PhD in engineering in
America in 1863, showed that "all systems change in such a
way that free energy [G] is minimized." In other words, a
Cells Can Transform One Type chemical reaction occurs spontaneously when the free energy
of Energy into Another of the products is lower than the free energy of the reactants.
According to the first law of thermodynamics, energy is nei- In the case of a chemical reaction, reactants~ products,
ther created nor destroyed but can be converted from one the free-energy change, llG, is given by
form to another. (In nuclear reactions, mass is converted to
energy, but this is irrelevant to biological systems. ) Energy LlG = Gproduct> - G reacrants

2.4 Biochemical Energetics 49

(a) (b) entropy lead to a lower .lG. That is, if temperature remains
constant, a reaction proceeds spontaneously only if the free-
Exergonlc Endergonic
energy change, .lG, in the following equation is negative:
Reactants

i
(!)
\ D.G<O
i
(!) Products
.lG = ilH- T .lS

In an exothermic ("heat-releasing") chemica l reaction, tJ.H is

(2-6 )

>- >-
e>
Q)
c:
Q)
Q)
Q)

U:
\ Products
0>
Q;
c:
Q)
Q)
Q)

U:
llG>O
I negative. ln an endothermic ("heat-absorbing") reaction,
:lH is positive. The combined effects of the changes in the
enthalpy and entropy determine if the llG for a reaction is
positive or negative and thus if the reaction occurs spontane-
ously. An exothermic reaction (.lH < 0), in which entropy
Reactants
increases (tJ.S > 0), occurs spontaneously (~G < 0). An en-
dothermic reaction (.lH > 0 ) will occur spontaneously if tiS
increases enough so that the TtJ.S term can overcome the
positive .lH.
Progress of reaction~ Progress of reaction~
Many biological reactions lead to an increase in order
FIGURE 2-29 Changes in the free energy (.iG) of exergonic and and thus a decrease in entropy (t{S < 0). An obvious exam-
endergonic reactions. (a) In exergonic reactions, the free energy of ple is the reaction that links amino acids to form a protein.
the products is lower than that of the reactants. Consequently, these A solution of protein molecules has a lower entropy than
reactions occur spontaneously and energy is released as the reactions does a solution of the same amino acids unlinked becau~e
proceed. (b) In endergonic reactions, the free energy of the products is the free movement of any amino acid in a protein is more
greater than that of the reactants and these reactions do not occur restricted (greater order) when it is bound into a long chain
spontaneously. An external source of energy must be supplied if the than when it is not. Thus when cells synthesize polymers
reactants are to be converted into products.
such as proteins from their constituent monomers, the po-
lymerizing reaction will only be spontaneous if the cells can
efficiently transfer energy to both generate the bonds that
The relation of .lG to the direction of any chemical reaction hold the monomers together and overcome the loss in en-
can be summarized in three statements: tropy that accompanies polymerization. Often cells accom-
plish this feat by ''coupling" such synthetic, entropy-lowering
• If llG is negative, the forward reaction will tend to occur
reactions with independent reactions that have a very high ly
spontaneously and energy usually will be released as the reac-
negative .1G (see below). In this way, cells can convert
tion takes place (exergonic reaction) (Figure 2-29). A reaction
sources of energy in their environment into the building of
with a negative llG is called thermodynamically favorable.
highly organized structures and metabolic pathways that are
• If .lG is positive, the forward reaction will not occur sponta- essential for life.
neously; energy will have to be added to the system in order to The actual change in free energy ~ G during a reaction is
force the reactants to become products (endergonic reaction). influenced by temperature, pressure, and the initial concentra-
tions of reactants and products and usually differs from the
• If .1G is zero, both forward and reverse reactions occur at standard free-energy change .1G 0 ' . Most biological reactions-
equal rates and there will be no spontaneous net conversion of
like others that take place in aqueous solutions-also are af-
reactants to products, or vice versa; the system is at equilibrium.
fected by the pH of the solution. We can estimate free-energy
By convention, the standard free-energy change of a reac- changes for temperatures and initial concentrations that differ
tion .1G 0 ' is the value of the change in free energy under the from the standard conditions by using the equation
conditions of 298 K (25 °C), 1 atm pressure, pH 7.0 (as in
pure water), and initial concentrations of 1 M for all reac- [products ]
.lG = ..lGo' + RT lnQ = .lG 0 ' + RT In (2-7)
tants and products except protons, which are kept at 10 ~ M [reactants]
(pH 7.0). Most biological reactions differ from standard
conditions, particularly in the concentrations of reactants, where R is the gas constant of 1.987 cal/(degree·mol), Tis
which are normally less than 1 M . the temperature (in degrees Kelvin), and Q is the initial ratio
The free energy of a chemical system can be defined as of products to reactants. For a reaction A + B ~ C, 111
G = H- TS, where His the bond energy, or enthalpy, of which two molecules combine to form a third, Q in Equa-
the system; Tis its temperature 111 degrees Kelvin (K); and tion 2 -7 equals [CI/IAilBI. In this case, an increase in the
S is the entropy, a measure of its randomness or disorder. initial concentration of either IAl or [Bl will result in a larger
According to the second law of thermodynamics, the natural negative value for llG and thus drive the reaction toward
tendency of any system is to become more disordered-that spontaneous formation of C.
is, for entropy to increase. A reaction can occur spontane- Regardless of the ..lG0 ' for a particular biochemical reac-
ously only if the combined effects of changes in enthalpy and tion, it will proceed spontaneously within cells only if llG is

50 CHAPTER 2 • Chemical Foundations

negative, given the intracellular concentrations of reactants products on the extent to which a reaction will occur
and products. For example, the conversion of glyceraldehyde spontaneously.
3-phosphate (G3P) to dihydroxyacetone phosphate (DHAP),
two intermediates in the breakdown of glucose,
The Rate of a Reaction Depends on the
G3P~DHAP
Activation Energy Necessary to Energize
the Reactants into a Transition State
has a 1G0 ' of -1840 cal/mol. If the initial concentrations of As a chemical reaction proceeds, reactants approach each
G3P and DHAP are equal, then ~G = .l G 0 ' because RT In other; some bonds begin tu form whi le others begin to break.
1 = 0; in this situation, the reversible reaction G3P ~ DHAP One way to think of the state of the molecules during this
wi ll proceed spontaneously in the direction of DHAP forma- transition is that there are strains in the electronic configura-
tion until equ ili brium is reached. However, if the initial tions of the atoms and their bonds. The collection of atoms
[DHAPl is 0. 1 M and the initial [GJP] is 0.001 M, with moves from the relatively stable state of the reactants to this
other conditions standard, then Q in Equation 2-7 equals transient, intermediate, and higher-energy state during the
0. 1/0.001 = 100, giving a ~G of +887 cal/mol. Under these cou rse of the reaction (Figure 2-30). The state during a
conditions, the reaction will proceed in the direction of for- chemical reaction at wh ich the system is at its highest energy
mation of G3 P. level is called the transition state, and the collection of reac-
The ~G for a reaction is independent of the reaction rate. tants in that state is called the transition-state intermediate.
Indeed, under usual physiological conditions, few if any of The energy needed to excite the reactants to this higher-
the biochemical reactions needed to sustain life would occur energy state is called t he activation energy of the reaction.
without some mechanism for increasing reaction rates. As The activation energy is usually represented by ~ct, analo
we describe below and in more detail in Chapter 3, the rates gous to the representation of the change in Gibbs free energy
of reactions in biological systems are usually determined by (.lG) already discussed. From the transition state, the collec-
the activity of enzymes, the protein catalysts that accelerate tion of atoms can either release energy as the reaction prod-
the formation of products from reactants wi thout altering the ucts are formed or release energy as the atom~ go "backward"
va lue of ~G. and re-form the original reactants. The velocity (V) at which
products arc generated from reactants during the reaction
The ~Go' of a Reaction Can Be Calculated under a given set of conditions (temperature, pressure, reac-
from Its Keq tant concentratiom) will depend on the concentration of

A chemical mixture at eq uilibrium is in a stable state of min-

imal free energy. for a system at equilibrium (~G = 0, Q =
Kcq), we can wri te T------- Transition state
(u ncata lyzed)

.l Go' = - 2.3RT log Kcq = -1362 log K,q (2-8) \ G~ncat.

(!)
Transition state
under standard Conditions (note the change to base 10 loga- >
Cl (catalyzed)
Qj
rithms). Thus if we determine the concentrations of reactants c:
Q)
and products at equilibrium (i.e., the K.q), we can calculate Q)

the value of ~ G '. For example, the Kcq for the interconver-
0 ~
u..
sion of glycera ldehyde 3-phosphate to dihydroxyacetone
phosphate (GJP ~ DHAP) is 22.2 under standard condi-
tions. Substituting this value into Equation 2 -8, we can eas- Products
ily calculate the ~G ' for this reaction as - 1840 cal/mol.
0

By rearranging Equation 2-8 and taking the anti loga- Progress of reaction ~
rithm, we obtain
FIGURE 2-30 Activation energy of uncatalyzed and catalyzed
K = 10 ( ~(;" '12.J RT ) chemical reactions. This hypothetical reaction pathway (blue) depicts
cq (2-9)
the changes in free energy, G, as a reaction proceeds. A reaction will
take place spontaneously if the free energy (G) of the products is less
From this expression, it is clear that if ..lG 0 ' is negative, th e than that of the reactants (~G < 0). However, all chemical reactions
exponent will be positive and hence K cq w ill be greater proceed through one (shown here) or more high-energy transition
than 1. T herdore, at equilibrium there wtll be more products states, and the rate of a reaction is inversely proportional to the
than reactants; in other words, the formation of products activation energy (~G*), which is the difference in free energy between
from reactants is favored. Conversely, if .l G 0 ' is positive, the reactants and the transition state. In a catalyzed reaction (red), the
the exponent w ill be negative and Keq w ill be less than 1. free energies of the reactants and products are unchanged but the free
The relationship between Kcq and uG 0 ' further emphasizes energy of the transition state is lowered, thus increasing the velocity of
the influence of relative free energies of reactants and the reaction.

2.4 Biochemical Energetics 51

material in the transition state, which in turn will depend on Hydrolysis of ATP Releases Substantial Free
the activation energy and the characteristic rate constant (v) Energy and Drives Many Cellular Processes
at which the transition state is converted to products. The
higher the activation energy, the lower the fraction of reac- In almost all organisms, the nucleoside triphosphate adeno-
tants that reach the transition state and the slower the over- sine triphosphate, or ATP (Figure 2-31 ), is the most impor-
all rate of the reaction. The relationship between the tant molecule for capturing, transiently storing, and
concentration of reactants, v, and Vis subsequently transferring energy to perform work (e.g., bio-
synthesis, mechanical motion ). Commonly referred to as a
V = v [reactants] X 1 0 p.c·Jl.JRT) cell's energy "currency," ATP is a type of usable energy th::tr
cells can "spend" in order to power their activities. The sto-
From this equation, we can see that lowering the activation ried history of ATP begins with its discovery in 1929 appar-
e~tly simultaneously by Kurt Lohmann, who was working
energy-that is, decreasing the free energy of the transition
state .lGt-leads to an acceleration of the overall reaction wtth the great biochemist Otto Meyerhof in Germany and
rate V. A reduction in .lGt of 1.36 kcal/molleads to a ten- who published first, and by Cyrus Fiske and Yellagaprada
fold increase in the rate of the reaction, whereas a 2. 72 kcal/ SubbaRow in the United States. Muscle contractions were
mol reduction increases the rate l 00-fold. Thus relatively shown to depend on ATP in the 1930s. The proposal that
small changes in .lGt can lead to large changes in the overall ATP is the main intermediary for the transfer of energy in
rate of the reaction. cells is credited to Fritz Lipmann around 1941. Many Nobel
Catalysts such as enzymes (discussed further in Chapter 3) Prizes have been awarded for the study of ATP and its role
accelerate reaction rates by lowering the relative energy of the in cellular energy metabolism, and its importance in under-
transition state and so the activation energy required to reach it standing molecular cell biology cannot be overstated.
(see Figure 2-30). The relative energies of reactants and prod- The useful energy in an ATP molecule is contained in
ucts will determine if a reaction is thermodynamically favorable phosphoanhydride bonds, which are covalent bonds formed
(negative .lG), whereas the activation energy will determine
how rapidly products form-that is, its reaction kinetics. Ther- NH 2
modynamically favorable reactions will not occur at apprecia-
N
c'-c ..- N
ble rates if the activation energies are too high.
Phosphoanhydrid< bond• HC
I II \H
1
~ ,....c ..._ N
0 0 0 N
Life Depends on the Coupling of Unfavorable
Chemical Reactions with Energetically
0-~-0
I
~-0
I
~-0-~CO
I
1 1 2

0 0 0 H H
Favorable Ones H H
Many processes in cells are energetically unfavorable (6G > 0) Adenosine triphosphate
(ATPl HO OH
and will not proceed spontaneously. Examples include the
synthesis of DNA from nucleotides and transport of a sub-
stance across the plasma membrane from a lower to a higher
concentration. Cells can carry out an energy-requiring, or
endergonic, reaction (.lG 1 > 0) by coupling it to an energy-
releasing, or exergonic, reaction (6G 2 < 0) if the sum of the
two reactions has an overall net negative .lG.
Suppose, for example, that the reaction A~ B + X has
a .lG of + 5 kcal/mol and that the reaction X~ Y + Z has 0
II
a 6G of - 10 kcal/mol: 0 P- 0 - H +
o-
(1 ) A~B+X .lG = + 5 kcal!mol
(2) X~Y+Z ~G = -10 kcal!mol
Inorganic phosphate Adenosine diphosphate
Sum: A~ B +Y +Z .:lG 0
' = -5 kcal/mol (P;l IADP)

FIGURE 2-31 Hydrolysis of adenosine triphosphate (ATP).

In the absence of the second reaction, there would be much The two phosphoanhydride bonds (red) in ATP (lOp), which link the
more A than Bat equilibrium. However, because the conver- three phosphate groups, each have a .lG" of about - 7.3 kcal/ mol for
sion of X toY + Z is such a favorable reaction, it will pull hydrolysis. Hydrolysis of the terminal phosphoanhydride bond by the
the first process toward the formation of B and the con- addition of water results in the release of phosphate and generation of
sumption of A. Energetically unfavorable reactions in cells ADP. Hydrolysis of the phosphoanhydride bonds of ATP, especially the
often are coupled to the energy-releasing hydrolysis of ATP, terminal one, is the source of energy that drives many energy-requiring
as we discuss next. reactions in biological systems.

52 CHAPTER 2 • Chemical Foundations

...
from the condensation of two molecules of phosphate by the requires less energy, and less energy is released when this
loss of water: bond is hydrolyzed.
Cells have evolved protein-mediated mechanisms for
0 0 transferring the free energy released by hydrolysis of phos-
II II phoanhydride bonds to other molecules, thereby driving re-
o- - p- OH + HO - P- 0 - ;:=
actions that would otherwise be energetically unfavorable.
For example, if the ~C for the reaction B + C ~ D is posi-
0 0 tive but less than the .lG for hydrolysis of ATP, the reaction
II II can he driven to the right by coupling it to hydrolysis of the
o--p o - P-o- + H2 o
I I terminal phosphoanhydride bond in ATP. In one common
o- o- mechanism of such energy coupling, some of the energy
stored in this phosphoanhydride bond is transferred to one
As shown in Figure 2-31, an ATP molecule has two key phos-
of the reactants by breaking the bond in ATP and forming a
.· phoanhydride (also called phosphodiester) bonds. Forming
covalent bond between the released phosphate group and
these bonds in ATP requires an input of energy. When these
one of the reactants. The phosphorylated intermediate gen-
bonds are hydrolyzed, or broken by the addition of water,
erated in this way can then react with C to form D + P, in a
that energy is released. Hydrolysis of a phosphoanhydride
reaction that has a negative ..lC:
bond (represented by the symbol - ) in each of the following
reactions has a highly negative ..lGor of about - 7.3 kcal/mol:
B + Ap- p- p ~ B- p + Ap- p
Ap- p-p + H2 0 ~ Ap-p + P, + H + B- p +C ~ D+ P
(ATP) (ADP)
The overall reaction
Ap- p-p + H 20 ~ Ap + PP; + H~

(ATP) (AMP) B + C + ATP ~ D + ADP + P;

Ap-p + H 20 ~ Ap + P, + H+
is energetically favorable {.lG < 0 ).
(ADP) (AMP) An alternative mechanism of energy coupling is to usc
the energy released by ATP hydrolysis to change the confor-
In these reactions that occur in biological systems, P, stands mation of the molecule to an "energy-rich" stressed state. In
for inorganic phosphate (POl ) and PP, for inorganic pyro- turn, the energy stored as conformational stress can be re-
phosphate, rwo phosphate groups linked by a phosphoanhy- leased as the molecule "relaxes" back into its unstressed
dride bond. As the top two reactions show, the removal of a conformation. If this relaxation process can be mechanisti-
phosphate or a pyrophosphate group from ATP leaves ade- cally coupled to another reaction, the released energy can be
nosine diphosphate (ADP) or adenosine monophosphate harnessed to drive important cellular processes.
(AMP), respectively. As with many biosynthetic reactions, transport of mole-
A phosphoanhydride bond or other "high-energy bond" cules into or out of the cell often has a positive ..lG and thus
(commonly denoted by-) is not intrinsically different from requires an input of energy to proceed. Such simple trans-
other covalent bonds. High-energy bonds simply release sub- port reactions do not directly involve the making or break-
stantial amounts of energy when hydrolyzed. For instance, ing of covalent bonds; thus the ~cor is 0. In the case of a
the ~co r for hydrolysis of a phosphoanhydride bond in ATP substance moving into a cell, Equation 2-7 becomes
( -7.3 kcal/mol) is more than three times the ..lGor for hydro-
lysis of the phosphoester bond (red) in glycerol 3-phosphate [Cm]
(- 2.2 kcallmol ): ' ..lC = RT ln -
C [1 (2-10)
our,

0 OH
II where [Cml is the initial concentration of the substance in-
HO - P 0 - CH - CH CH 2 0H side the cell and [Courl is its concentration outside the cell.
I 2
o- We can see from Equation 2-10 that ~G is positive for trans-
Glycerol 3-phosphate port of a substance into a cell against its concentration gradi-
ent (when [Cml > [C001 ] ) ; the energy to drive such '"uphill"
A principal reason for this difference is that ATP and its hy- transport often is supplied by the hydrolysis of ATP. C:on-
droly~is products ADP and P; are h1ghly charged at neutral versely, when a substance moves down its concentration gra-
pH. During synthesis of ATP, a large input of energy is re- dient (ICourl > jC, 11 ]), ..lG is negative. Such "downhill"
quired to force the negative charges in ADP and P, together. transport releases energy that can be coupled to an energy-
Conversely, this energy is released when ATP is hydrolyzed requiring reaction, say, the movement of another substance
to ADP and P,. In comparison, formation of the phospho- uphill across a membrane or the synthesis of ATP itself (see
ester bond between an uncharged hydroxyl in glycerol and P, Chapters 11 and 12).

2.4 Biochemical Energetics 53

 ·.

ATP Is Generated During Photosynthesis NAD+ and FAD Couple Many Biological
and Respiration Oxidation and Reduction Reactions
ATP is contmually being hydrolyzed to provide energy for In many chemical reactions, electrons are transferred from
many cellular activities. Some estimates suggest that hu- one atom or molecule to another; this transfer may or may
mans daily hydrolyze a mass of ATP equal to their entire not accompany the formation of new chemical bonds or the
body weight. Clearly, to continue functioning, cells must release of energy that can be coupled to other reactions. The
constantly replenish their ATP supply. Constantly replen- loss of electrons from an atom or a molecule is called oxida-
ishing ATP requires that cells continually obtain energy tion, and the gain of electrons by an atom or a molecule is
from their environment. For nearly all cells, the ultimate called reduction. An example of oxidation is the removal of
source of energy used to make ATP is sunlight. Some organ- electrons from the sulfhydryl groups of two cysteines to
isms can use sunlight directly. Through the process of pho- form a disulfide bond, described above in Section 2.2. Be-
tosynthesis, plants, algae, and certain photosynthetic cause electrons are neither created nor destroyed in a chemi-
bacteria trap the energ} of sunlight and use it to synthesize cal reaction, if one atom or mo lecule is oxidized, another
ATP from ADP and P,. Much of the ATP produced in pho- must be reduced. For example, oxygen draws electrons from
tosynthesis is hydrolyzed to provide energy for the conver- fe 2 (ferrous) ions to form FeP (ferric) ions, a reaction that
sion of carbon dioxide to six-carbon sugars, a process called occurs as part of the process by ,wh ich carbohydrates arc
carbon fixation: degraded in mitochondria. Each oxygen atom receives two
electrons, one from each of two Fe2 ions:

The sugars made during photosynthesis are a source of food, Thus Fe2 ~ is oxidized and 0 2 is reduced. Such reactions in
and thus energy, for the plants or other photosynthetic or- which one molecule is reduced and another oxidized often are
ganisms making them and for the non-photosynthetic organ- referred to as redox reactions. Oxygen is an electron acceptor
isms, such as animals, that either consume the plants directly in many redox reactions in cells under aerobic conditions.
or indirectly by eating other animals that have eaten the Many biologically important oxidation and reduction
plants. In this way sunlight is the direct or indirect source of reactions involve the removal or the addition of hydrogen
energy for most organisms (see Chapter 12). atoms (protons plus electrons) rather than the transfer of
In plants, animals, and nearly all other organisms, the
isolated electrons on their own. The oxidation of succinate
free energy in sugars and other molecules derived from food to fumarate, which also occurs in mitochondria, is an ex-
is released in the processes of glycolysis and cellular respira-
ample (Figure 2-32). Protons are soluble in aqueous solu-
tion. During cellular respiration, energy-rich molecules in tions (as H 30 ), but electrons are not and must be transferred
food (e.g., glucose) arc oxidized to carbon dioxide and directly from one atom or molecule to another without a
water. The complete oxidation of glucose, water-dissolved intermediate. In this type of oxidation reac-
tion, electrons often are transferred to small electron-carrying
molecules, sometimes referred to as coenzymes. The most
common of these electron carriers are NAD+ (nicotinamide
has a ~G ' of -686 kcal/mol and is the reverse of photosyn-
0 adenine dinucleotide), which is reduced to NADH, and FAD
thetic carbon fixation. Cells employ an elaborate set of protein- (flavin adenine dinucleotide), which is reduced to FADH 2
mediated reactions to couple the oxidation of 1 molecule of (Figure 2-33 ). The reduced forms of these coenzymes can
glucose to the synthesis of as many as 30 molecules of ATP transfer protons and electrons to other molecules, thereby
from 30 molecules of ADP. This oxygen-dependent (aerobic) reducing them.
degradation (catabolism) of glucose is the major pathway
for generating ATP in all animal cells, non-photosynthetic 0 0
plant cells, and many bacterial cells. Catabolism of fatty II
c-o C-0
acids can also be an important source of ATP. We discuss I I
the mechanisms of photosynthesis and cellular respiration in H C-H C-H
)
Chapter 12. H C- H \1 \1 C-H
Although light energy captured in photosynthesis is the 2e 2H
C-0 C-0
primary source of chemical energy for cells, it i~ nor the only I II
source. Certain microorganisms that live in or around deep 0 0
ocean vents, where adequate sunlight is unavailable, derive Succinate Fumarate
the energy for converting ADP and Pi into ATP from the FIGURE 2-32 Conversion of succinate to fumarate. In this
oxidation of reduced inorganic compounds. These reduced oxidation reaction, which occurs in mitochondria as part of the citric
compounds originate deep in the earth and are released at acid cycle, succinate loses two electrons and two protons. These are
the vents. transferred to FAD, reducing it to FADH 2•

54 CHAPTER 2 • Chemical Foundations

 (a) (b)

Oxidized: NAD+ Reduced: NADH

HO 6-
H" 0
6- N
I
Riuose
II

Nicotinam ide
+ H+
C-NH2 + 2 9 _ ~
~
I I
..
N
I
Ribose
II
C-NH2

Ribitol
I I I
2P 2P 2P
I I . I
Adenosine Adenosine Adenosine

NADH FAD+ 2 H+ + 2 e- ;:::::::=: FADH 2

FIGURE 2 -3 3 The electron-carrying coenzymes NAD+ and FAD. is released into solution. (b) FAD (flavin adenine dinucleotide) is
(a) NAD (nicotinamide adenine dinucleotide) is reduced to NADH by reduced to FADH 2 by the addition of two electrons and two protons, as
the addition of two electrons and one proton simultaneously. In many occurs when succinate is converted to fumarate (see Figure 2-32).1n
biological redox reactions, a pair of hydrogen atoms (two protons and this two-step reaction, addition of one electron together with one
two electrons) are removed from a molecule. In some cases, one of the proton first generates a short-lived semiquinone intermediate (not
protons and both electrons are transferred to NAD ; the other proton shown), which then accepts a second electron and proton.

To describe redox reactions, such as the reaction of fer- to, or reduce, a molecule with a more positive reduction po-
rous ion (Fe2 ) and oxygen (0 2 ), it is easiest to divide them tential. In this type of reaction, the change in electric poten-
into two half-reactions: tial .lE is the sum of the reduction and oxidation potentials
for the two half-reactions. The .lE for a redox reaction is
Oxidatio11 of Fe 2 +: 2 Fe2 + ~ 2 Feh + 2 e related to the change in free energy .lG by the following ex-
Reduction of 0 2 : 2 e + 1/2 0 2 ~ 0 2 - pression:

In this case, the reduced oxygen (0 2 - ) readily reacts with llG (cal/mol)= -n (23,064) ~E (volts) (2-11)
two protons to form one water molecule (H 20). The readi-
ness with which an atom or a molecule gains an electron is where n is the number of electrons transferred. Note that a
its reduction potential E. The tendency to lose electrons, the redox reaction with a positive .lE value will have a negative
oxidation potential, has the same magnitude bur opposite .1G and thus will tend to proceed spontaneously from left to
sign as the reduction potential for the reverse reaction. right.
Reduction potentials are measured in volts (V) from an
arbitrary zero point set at the reduction potential of the fol-
lowing half-reaction under standard conditions (25 °C, 1 atm,
and reactants at 1 M): KEY CONCEPTS of Section 2.4
Biochemical Energetics
• The change in free energy, ~G, is the most useful measure
oxidarion
for predicting the potential of chemical reactions to occur
spontaneously in biological systems. Chemical reactions
The value of E for a molecule or an atom under standard
tend to proceed spontaneously in the direction for which llG
conditions is its standard reduction potential, E' o. A mole-
is negative. The magnitude of ~G is independent of the reac-
cule or an ion with a positive E' 0 has a higher affinity for
tion rate. A reaction with a negative tiC is called thermody-
electrons than the H + ion does under standard conditions.
namically favorable.
Conversely, a molecule or ion with a negative £' 0 has a
0
lower affinity for electrons than the H + ion does under stan- • The chemical free-energy change, .1G equals -2.3 RT ',

dard conditions. Like the values of .lG0 ' , standard reduction log Keq· Thus the value of .tiG0 ' can be calculated from the
potentials may differ somewhat from those found under the experimentally determined concentrations of reactant~ and
conditions in a cell because the concentrations of reactants products at equilibrium.
in a cell are not 1 M. • The rate of a reaction depends on the activation energy
In a redox reaction, electrons move spontaneously to- needed to energize reactants to a transition state. Catalysts
ward atoms or molecules having more positive reduction such as enzymes speed up reactions by lowering the activa-
potentials. In other words, a molecule having a more nega- tion energy of the transition state.
tive reduction potential can transfer electrons spontaneously

2.4 Biochemical Energetics ss

• A chemical reaction having a positive ..lG can proceed if it Review the Concepts
is coupled with a reaction having a negative ~G of larger
1. The gecko is a reptile with an amazing ability to climb
magnitude. smooth surfaces, including glass. Recent discoveries indicate
• Many otherwise energetically unfavorable cellular pro- that geckos stick to smooth surfaces via van der Waals inter-
cesses arc driven by the hydrolysis of phosphoanhydride actions between septae on their feet and the smooth surface.
bonds in ATP (see Figure 2-31). How is this method of stickiness advantageous over covalent
• Directly or indirectly, light energy captured hy photosyn- interactions? Given that van der Waals forces are among the
thesis in plants, algae, and photosynthetic bacteria is the ul- weakest molecular interactions, how can the gecko's feet
timate source of chemical energy for nearly all cells on earth. stick so effectively?
2. The K+ channel is an example of a transmembrane pro-
• An oxidation reaction (loss of electrons) is always coupled
tein (a protein that spans the phospholipid bilayer of the
with a reduction reaction (gain of electrons).
plasma membrane). What types of amino acids are likely to
• Biological oxidation and reduction reactions often are be found (a) lining the channel through which K passes,
coupled by electron-carrying coenzymes such as NAD and (b) in contact with the hydrophobic core of the phospholipid
FAD (see Figure 2-33). bilayer containing fatty acyl groups, (c) in the cytosolic do-
Oxidation-reduction reactions with a positive .lE have a main of the protein, and (d ) in th~ extracellular domain of
negative .lG and thus tend to proceed spontaneously. the protein?
3. V-M-Y-F-E-N: This is the single-letter amino acid ab-
breviation for a peptide. What is the net charge of this pep-
tide at pH 7.0? An enzyme called a protein tyrosine kinase
Key Terms can attach phosphates to the hydroxyl groups of tyrosine.
What is the net charge of the peptide at pH 7.0 after it has
acid 46 hydrophilic 23 been phosphorylated by a tyrosine kinase? What is the
adenosine triphosphate hydrophobic 23 likely source of phosphate utilized by the kinase for this
(ATP) 52 hydrophobic effect 31 reaction?
o: carbon atom (Cu) 33 ionic interactions 28 4. Disulfide bonds help to stabilize the three-dimensional
amino acid 33 molecular structure of proteins. What amino acids are involved in the
amphipathic 23 complementarity 32 formation of disulfide bonds? Does the formation of a disul-
base 46 monomer 33 fide bond increase or decrease entropy (..15)?
buffer 47 monosaccharide 37 5. In the 1960s, the drug thalidomide was prescribed to
pregnant women to treat morning sickness. However, tha-
catalyst 43 noncovalent
lidomide caused severe limh defect~ in the children of some
chemical potential interactions 24
women who took the drug, and its use for morning sickness
energy 49 nucleoside 37 was discontinued. It is now known that thalidomide was ad-
covalent bond 24 nucleotide 37 ministered as a mixture of two stereoisomeric compounds,
dehydration reaction 33 oxidation 54 one of which relieved morning sickness and the other of
dipole 26 pH45 which was responsible for the birth defects. What are stereo-
dissociation constant phosphoanhydride isomers? Why might two such closely related compounds
(KJ) 45 bond 52 have such different physiologic effects?
disulfide bond 35 phosphoglyceride 40 6. Name the compound shown below.
endergonic 49 phospholipid bilayer 40
endothermic 50 polar 26 0
energy coupling 53 polymer 33 .·.
C N (

enthalpy (H) 50 redox reaction 54 HN;-"6 5 c -- 7~

2 4
1 sCH
entropy (S) 50 reduction 54 /c~ 3 c ..._ ~
equilibrium constant saturated 40 H2N N I
(Keq) 43 steady state 44
0 0 0 I
exergonic 49 stereoisomer 25 II II II 5' I
exothermic 50 -o-P- 0 0-P- 0 0-P- 0
transition state 51
fatty acids 40 unsaturated 40 4 ·C H o H ,.
FU2
~G (free-energy change) 49 H H
van der Waals 3' 2'

hydrogen bond 28 interaction 30 OH OH

56 CHAPTER 2 • Chemical Foundations

Is this nucleotide a component of DNA, RNA, or both? biologically active. What particular amino acid undergoes this
Name one other function of this compound. modification, and what is the biological relevance? Warfarin,
7. The chemical basis of blood-group specificity resides in a derivative of coumarin, which is present in many plants, in-
the carbohydrates displayed on the surface of red blood cells. hibits -y-carhoxylation of this amino acid and was used in the
Carbohydrates have the potential for great structural diver- past as a rat poison. At present, it is also used clinically in hu-
sity. Indeed, the structural complexity of the oligosaccharides mans. What patients might be prescribed warfarin and why?
that can be formed from four sugars is greater than that for
oligopeptides from four amino acids. What properties of car-
bohydrates make this great structural diversity possible? Analyze the Data
8. Calculate the pH of 1 L of pure water at equilibrium. How 1. During much of the "Age of Enlightenment" in eigh-
will the pH change after 0.008 moles of the strong base NaOH teenth-century Europe, scientists toiled under the belief that
are dissolved in the water? Now, calculate the pH of a 50 mM living things and the inanimate world were fundamentally
aqueous solution of the weak acid 3-(N-morpholino) propane- distinct forms of matter. Then in 1828, Friedrich Wohler
1-sulfonic acid (MOPS) in which 61% of the solute is in its showed that he could synthesize urea, a well-known waste
weak acid form and 39% is in the form of MOPS conjugate product of animals, from the minerals silver isocyanate and
base (the pK3 for MOPS is 7.20). What is the final pH after ammonium chloride. "I can make urea without kidneys!" he
0.008 moles of NaOH are added to 1 L of this MOPS buffer? is said to have remarked. Of Wohler's discovery the preemi-
9. Ammonia (NH 3) is a weak base that under acidic condi- nent chemist Justus von Liebig wrote in 1837 that the "pro-
tions becomes protonated to the ammonium ion in the fol- duction of urea without the assistance of vital functions ...
lowing reaction: must be considered one of the discoveries with which a new
era in science has commenced." Slightly more than 100 years
later, Stanley Miller discharged sparks into a mixture of
H 20, CH 4, NH1, and H 2 in an effort to simulate the chemi-
NH, freely permeates biological membranes, including those cal conditions of an ancient reducing earth a.tmosphere (the
of lysosomes. The lysosome is a subcellular organelle with a sparks mimicked lightning striking a primordial sea or
pH of about 4.5-5.0; the pH of cytoplasm is - 7.0. What is "soup") and identified many biomolecules in the resulting
the effect on the pH of the fluid content of lysosomes when mixture, including amino acids and carbohydrates. What do
cells are exposed to ammonia? Note: Protonated ammonia these experiments suggest about the nature of biomolecules
does not diffuse freely across membranes. and the relationship between organic (living) and inorganic
10. Consider the binding reaction L + R--+ LR, where L is (nonliving) matter? What do they suggest about the evolu-
a ligand and R is its receptor. When 1 X 1o->M L is added tion of life? What do they indicate about the value of chem-
to a solution containing 5 X 10 1 M R, 90 percent of the L istry in understanding living things?
binds to form LR. What is the K cq of this reaction? How will 2. The graph below illustrates the effect that the addition of
the Keq be affected by the addition of a protein that facilitates a strong base such as sodium hydroxide has on the pH of an
(catalyzes) this binding reaction? What is the dissociation aqueous 0.1 M solution of an amino acid. Assume that prior
equilibrium constant Kd? to the addition of any OH , the entire dissolved amino acid
11. What is the ionization state of phosphoric acid in the
cytoplasm? Why is phosphoric acid such a physiologically
14
important compound?
12. The ~G ' for the reaction X + Y --+ XY is -1000 call
0

12
mol. What is the ~Gat 25 oc (298 Kelvin) starting with 0.01
M each X, Y, and XY? Su'ggest two ways one could make
10
this reaction energetically favorable.
13. According to health experts, saturated fatty acids, which
come from animal fats, are a major factor contributing to coro- 8
I
nary heart disease. What distinguishes a saturated fatty acid a.
from an unsaturated fatty acid, and to what does the term satu- 6
rated refer? Recently, trans unsaturated fatty acids, or trans
fats, which raise total cholesterol levels in the body, have also 4
heen implicated in heart disease. How docs the cis stereoi5umer
differ from the trans configuration, and what effect does the cis 2
configuration have on the structure of the fatty acid chain?
14. Chemical modifications to amino acids contribute to the
diversity and function of proteins. For instance, -y-carboxyl- 0.05
ation of specific amino acids is required to make some proteins

Analyze the Data 57

sample is in its fully protonated form. The addition of OH- Davenport, H. W. 1974. ABC of Acid-Base Chemistry, 6th ed.
causes the expected steep increase in solution pH until, be- Umversity of Ch1cago Press.
tween roughly 0.03-0.07 M NaOH, the solution pH remains Eisenberg, D., and D. Crothers. 1979. Physical Chemistry with
AfJ/Ilzcations to the Life Sciences. Benjamin-Cummings.
almost constant at a pH of approximately 1.8. What causes
Guyton, A. C., and J. E. Hall. 2000. Textbook of Medical
the resistance to change pH in this range? What are solutions Physiology, lOth ed. Saunders.
that resist changes in pH called? What organic chemical group Hill, T. J. 1977. Free Energy Transduction in Biology. Academic
on the amino acid is most likely responsible for this phenom- Press.
enon at pH 1.8? Additional base causes the pH to again in- Klotz, l. M. 1978. l:nergy Changes m Rinchn111cnl Renct1ons.
crease rapidly until the base concentration reaches 0.15 M and Academic Press.
0.25 M, at which points the solution pH hovers around values Murray, R. K., et al. 1999. Harper's Biochemistry, 25th ed. Lange.
of 6 and 9.3, respectively. What is the significance of these pH Nicholls, D. G., and S. .J. Ferguson. 1992. Bioenergetics 2.
values? Which amino acid do you suspect is being titrated? Academic Press.
Oxtoby, D., H. Gillis, and N. Nachtrieb. 2003. Prmciples of
Modern Chemistry, 5th ed. Saunders.
Sharon, N. 1980. Carbohydrates. Sc1. Am. 243(5):90-116.
References Tanford, C. 1980. The Hydrophobic Effect: Formation of
M1ce/les and 81ological Membranes, 2d ed. Wiley.
Alberty, R. A., and R. J. Silbey. 2005. PhysiCal Chemistry, 4th Tinoco, 1., K. Sauer, and J. Wang.'2001 . Phys1cal Chemlstry-
ed. Wiley. Prmoples and Applicatwns 111 Biolog1cal Sciences, 4th ed. Prent1ce
Atkins, 1'., and J. de Paula. 2005. The Elements of Physical Hall.
Chemistry, 4th ed. W. H. Freeman and Company. Van Holde, K., W. Johnson, and P. Ho. 1998. Principles of
Berg, J. M., ]. L. Tymoczko, and L. Stryer. 2007. Biochemistry, Phys1ca/ B1ochem1stry. Prentice Hall.
6th ed. W. H. Freeman and Company. Voet, D., and j. Voer. 2004. Bzochem1stry, 3d ed. Wiley.
Cantor, P. R., and C. R. Schimmel. 1980. Bwphysical Chemis- Wood, W. B., et al. 1981. B1ochem1stry: A Problems Approach,
try. W. H. Freeman and Company. 2d ed. Benjamin-Cummings.

. .

58 CHAPTER 2 • Chemical Foundations

 CHAPTER

Protein Structure
and Function

Molecular model of the proteasome from the heat- and acid-loving

archaeon T. acidophilium, represented using both solvent-accessible
surfaces (bottom) and ribbons (top). Proteasomes are protein-digesting
molecular machines, comprising a middle catalytic core (red, beige,
and gray), where degradation takes place, and two regulatory subunit
caps (yellow and black), which recognize proteins that have been
tagged for destruction by the addition of ubiquitin molecules. [Ramon
Andrade 30ciencia/Science Photo Library.]

P
roteins, which are polymers of amino acids, come in understanding how cells work. Much of this textbook IS de-
many sizes and shapes. Their three-dimensional diver- voted to examining how proteins act together to allow cells
sity principally reflects variations in their lengths and to live and function properly.
amino acid sequences. In general, the linear, unbranched Although their structures arc diverse, most indivtdual
polymer of amino' acids composing any protein will fold into proteins can be grouped into one of a few broad functional
only one or a few closely related three-dimensional shapes- classes . Structural proteins, for example, determine the
called conformations. The conformation of a protein together shapes of cells and their extracellular environments and
with the distinctive chemical properties of its amino acid side serve as guide wires or rails to direct the intracellular move-
chains determines irs function. Because of their many differ- ment of molecules and organelles. They usually are formed
ent shapes and chemical properties, proteins can perform a by the assembly of multiple protein subunits into very large,
dazzling array of distinct runctions inside and outside cells long structures. Scaffold proteins bring other proteins to-
that either are essential for life or provide selective evolution- gether into ordered arrays to perform specific functions
ary advantage to the cell or organism that contains them. It more efficiently than if those proteins were not assembled
is, therefore, not surprising that characterizing the structures together. Enzymes are proteins that catalyze chemical reac-
and activities of proteins is a fundamental prerequisite for tions. Membrane transport proteins permit the flow of ions

OUTLINE

3.1 Hierarchical Structure of Proteins 61 3.5 Purifying, Detecting, and Characterizing

Proteins 93
3.2 Protein Folding 70
3.6 Proteomics 106
3.3 Protein Binding and Enzyme Catalysis 77

3.4 Regulating Protein Function 85

and molecules across cellular membranes. Regulatory pro- (a) MOLECULAR STRUCTURE
teins act as signals, sensors, and switches to control the Primary (sequence)
activities of cells by altering the functions of other proteins
and genes. Regulatory proteins include signaling proteins,
Secondary (loca l folding)
such as hormones and cell-surface receptors that transmit
extracellular signals to the cell interior. Motor proteins are
responsible for moving other proteins, organelles, cells- Tertiary Supra molecular
even whole organisms. Any one protein can be a member of (overall conformation) (large-scale assembly)
more than one protein class, as is the case with some cell-
surface signaling receptors that are both enzymes and regula- Quaternary
(multimeric structure)
tor proteins because they transmit signals from outside to
inside cells by catalyzing chemical reactions. To accomplish
efficiently their diverse missions, some proteins assemble (b)
into large complexes, often called molecular machines.
Regulation Signa ling
How do proteins perform so many diverse functions? ... .·
They do this by exploiting a few simple activities. Most fun- @
damentally, proteins bind-to one another, to other macro- (9 )
molecules such as DNA, and to small molecules and ions. In
many cases such binding induces a conformational change in
the protein and thus influences its activity. Binding is based Structure

~
on molecular complementarity between a protein and its
binding partner, as described in Chapter 2. A second key
activity is enzymatic catalysis. Appropriate folding of a pro- Movement Catalysis A
tein will place some amino acid side chains and carboxyl and
amino groups of its backbone into positions that permit the
catalysts of covalent bond rearrangements. A third activity
=--c- /

involves folding into a channel or pore within a membrane

through which molecules and ions flow. Although these are FIGURE 3-1 Overview of protein structure and function.
(a) Proteins have a hierarchical structure. A polypeptide's linear
·.
especially crucial protein activities, they are not the only
sequence of amino acids linked by peptide bonds (primary structure)
ones. For example, fish that live in frigid waters-the Ant-
folds into local helices or sheets (secondary structure) that pack into a
arctic borchs and Arctic cods-have antifreeze proteins in
complex three-dimensional shape (tertiary structure). Some individual
their circulatory systems to prevent water crystall ization .
polypeptides associate into multichain complexes (quaternary
A complete understanding of how proteins permit cells
structure), which in some cases can be very large, consisting of tens to
to live and thrive requires the identification and character- hundreds of subunits (supramolecular complexes). (b) Proteins perform
ization of all the proteins used by a cell. In a sense, molecular numerous functions, including organizing in three-dimensional space
cell biologists want to compile a complete protein "parts the genome, organelles, the cytoplasm, protein complexes, and
list" and construct a "user's manual" that describes how membranes (structure); controlling protein activity (regu lation);
these proteins work. Compiling a comprehensive inventory monitoring the environment and transmitting information (signaling);
of proteins has become feasible in recent years with the se- moving small molecules and ions across membranes (transport);
quencing of entire genomes-complete sets of genes-of catalyzing chemical reactions (via enzymes); and generating force for
more and more organisms. From a computer ana lysis of movement (via motor proteins). These functions and others arise from
genome sequences, researchers can deduce the amino acid specific binding interactions and conformational changes in the
sequences and approximate number of the encoded proteins structure of a properly folded protein.
(see Chapter 5). The term proteome was coined to refer to
the entire protein complement of an organism. The human In this chapter, we begin our study of how the structure
genome contains some 20,000-23,000 genes that encode of a protein gives rise to its function, a theme that recurs
proteins. However, variations in mRNA production, such as throughout this book (Figure 3-1 ). The first section exam-
alternative splicing (see Chapter 8), and more than 100 types ines how linear cha ins of amino acid building blocks are
of protein modifications may generate hundreds of thou- arranged in a three-dimensional structural hierarchy. The
sands of distinct human proteins. By comparing the se- next section discusses how proteins fold into these struc-
quences and structures of proteins of unknown function to tures. We then turn to protein function, focusing on enzymes,
those of known function, scientists can often deduce much the special class of proteins that catalyze chemical reactions.
about what these proteins do. In the past, characterization of Various mechanisms that cells use to control the activities
protein function by genetic, biochemical, or physiological and life spans of proteins are covered next. The chapter con-
methods often preceded the identification of particular pro- cludes with a discussion of commonly used techniques for
tems. In the modern genomic and proteomic era, a protein is identifying, isolating, and characterizing proteins, including
usually identified prior to determining its function. a discussion of the burgeoning field of proteomics.

60 CHAPTER 3 • Protein Structure and Function

 3.1 Hierarchical Structure of Proteins The Primary Structure of a Protein Is Its linear
Arrangement of Amino Acids
A protein chain folds into a distinct three-dimensional shape
that is stabilized primarily by noncovalent interactions be- As discussed in Chapter 2, proteins are polymers constructed
tween regions in the linear sequence of amino acids. A key out of 20 different types of amino acids. Individual amino
concept in understanding how proteins work is that function acids are linked together in linear, unbranched chains by co-
is derived from three-dimensional structure, and three- valent amide bonds, called peptide bonds. Peptide bond for-
dimensional structure is determined by both a protein's amino mation between the amino group of one amino acid and the
acid sequence and intramolr>cular noncovalent interactions. carboxyl group of another results in the net relea~e of a
Principles relating biological structure and function initially water molecule and thus is a form of dehydranon reaction
were formulated by the biologists Johann von Goethe (1749- (Figure 3-3a). The repeated amide N, a carbon (Cu), car-
1832), Ernst Haeckel ( J 834-1919), and D'Arcy Thompson bonyl C, and oxygen atoms of each amino acid residue form
(1860-1948), whose work has been widely influential in biol-
ogy and beyond. Indeed, their ideas greatly influenced the
(a)
school of "organic" architecture pioneered in the early twen-
tieth century that is epitomized by the dicta "form follows
function" (Louis Sulli van) and "form is function" (Frank
Lloyd Wright). Here we consider the architecture of proteins
at four levels of organization: primary, secondary, tertiary,
and quaternary (Figure 3-2).

H 0 H 0
I I I II
+H3 N- C"- C- N - Cr,-C - 0
(a) Primary structure (b) Secondary structure I I
-Aia-Giu-Val- Thr-Asp-Pro-Giy- R1 1
H R2
Peptide
bond
(b)
H O H H H O HO
I I I I II I II
+H3 N - Cu-C- N - Crt-1 - r..J - C"- C- N - Cu- c - o -

1 h, H h2 0 h3 H h4 I
Amino end Carboxyl end
•. ' (N-terminus) (C-terminus)
Domain

(c)
(d) Quaternary structure

Peptide
bond

FIGURE 3-3 Structure of a polypeptide. (a) Individual amino

acids are linked together by peptide bonds, which form via reactions
.· FIGURE 3-2 Four levels of protein hierarchy. (a) The linear that result in a loss of water (dehydration). R1, R2, etc., represent the
sequence of amino acids linked together by peptide bonds is the side chains ("R groups") of amino acids. (b) Linear polymers of
primary structure. (b) Folding of the polypeptidt! chain into local a peptide-bond-linked amino acids are called polypeptides, which have a
helices or 13 sheets represents secondary structure. (c) Secondary free amino end (N-terminus) and a free carboxyl end (C-terminus). (c) A
structural elements together with various loops and turns in a single ball-and-stick model shows peptide bonds (yellow) linking the amino
polypeptide chain pack into a larger independently stable structure, nitrogen atom (blue) of one amino acid (aa) with the carbonyl carbon
which may include distinct domains; this is tertiary structure. (d) Some atom (gray) of an adjacent one in the chain. The R groups (green)
?roteins consist of more than one polypeptide associated together in a extend from the a carbon atoms (black) of the amino acids. These side
quaternary structure. chains largely determine the distinct properties of individual proteins.

3.1 Hierarchical St ructure of Proteins 61

the backbone of a protein molecule from which the various highly flexible parts of a polypeptide chain that have no
side-chain groups prOJeCt (Figure 3-3b, c). As a consequence fixed three-dimensional structure. In an average protein, 60
of the peptide linkage, the backbone exhibits directionality, percent of the polypeptide chain exists as a helices and 13
usually referred to as an N-to-C orientation, because all the sheets; the remainder of the molecule is in irregular struc-
amino groups are located on the same side of the Cu atoms. tures, coils and turns. Thus a helices and 13 sheets arc the
Thus one end of a protein has a free (unlinked ) amino group major internal supportive elements in most proteins. In this
(the N-termmus), and the other end has a free carboxyl group section, we explore the shapes of secondary structures and
(the C-terminus). The sequence of a protein chain is conven- the forces that favor their formation. In later sections, we
tionally written with its N-terminal amino acid on the left and examine how arrays of secondary structure fold together into
its C-terminal amino acid on the right, and the amino acids larger, more complex arrangements called tertiary structure.
are numbered sequentially starting from the amino terminus.
The primary structu re of a protein is simply the linear The a Helix In a polypeptide segment folded into an a helix,
covalent arrangement, or sequence, of the amino acid resi- the backbone forms a spiral structure in which the carbonyl
dues that compose it. The first primary structure of a protein oxygen atom of each peptide bond is hydrogen-bonded to
determined was that of insulin in the early 1950s and today the amide hydrogen atom of the amino acid four residues
the number of known sequences exceeds 10 m illion and is farther along the chain in the direction of the C-terminus
growing daily. Many terms are used to denote the chains (figure 3-4). Within an a helix, a-ll the backbone amino and
formed by the polymerization of amino acids. A short chain carboxyl groups are hydrogen-bonded to one another except
of amino acids linked by peptide bonds and having a defined at the very beginning and end of the helix. This periodic
sequence is called an oligopep tide, or just peptide; longer
chains are referred to as polypeptides. Peptides generally
Amino terminus
contain fewer than 20-30 amino acid residues, whereas
polypeptides are often 200-500 residues long. The longest
protein described to date is the muscle protein titin with
>35,000 residues. We generally reserve the term protein for
a polypeptide (or complex of polypeptides) that has a well-
defined three-dimensional structure.
The size of a protein or a polypeptide is expressed either as
its mass in daltons (a dalton is 1 atomic mass unit) or as its
molecular weight (MW), which is a dimensionless number
equal to the mass in daltons. For example, a 10,000-MW pro-
tein has a mass of 10,000 daltons (Da), or 10 kilodaltons
(k Da ). Later in this chapter, we will consider different methods
for measuring .the sizes and other physical characteristics of
proteins. The proteins encoded by the yeast genome have an
average molecular weight of 52,728 and contain, on average,
466 amino acid residues. The average molecular weight of
amino acids in proteins is 113, taking into account their aver-
age relative abundances. This value can be used to estimate the
number of residues in a protein from its molecular weight or,
conversely, its molecular weight from the number of residues. 3.6 residues/turn

Secondary Structures Are the Core Elements

of Protein Architecture
The second level in the hierarchy of protein structure is sec-
ondary structure. Secondary structures are stable spatial ar-
rangements of segments of a polypeptide chain held together
1
by hydrogen bonds between backbone amide and carbonyl
groups and often involving repeating structural patterns. A
single polypeptide may contain multiple types of secondary Carboxyl terminus
structure in various portions of the chain, depending on its FIGURE 3-4 The a helix, a common secondary structure in
sequence. The principal secondary structures are the alpha (a) prot eins. The polypeptide backbone (seen as a ribbon) is folded into
helix, the beta ( ~ ) sheet, and a short U-shaped beta(~) turn. a spiral that is held in place by hydrogen bonds between backbone
Parts of the polypeptide that don't form these structures but oxygen and hydrogen atoms. Only hydrogens involved in bonding
nevertheless have a well-defined, stable shape are said to are shown. The outer surface of the helix is covered by the side-chain
have an irregular structure. The term random coil applies to R groups (green).

62 CHAPTER 3 • Protein Structure and Function

arrangement of bonds confers an amino-to-carboxy-termi- (a) Top view
na l directionality on the helix because all the hydrogen bond
acceptors (i.e., the carbonyl groups) have the same orienta-
tion (pointing in the downward direction in Figure 3-4), re-
sulting in a structure in which there is a complete turn of the
spiral every 3.6 residues. An a helix 36 amino acids long has
10 turns of the helix and is 5.4 nm long (0.54 nm/turn).
The stable arrangement of hydrogen-bonded amino acids
in the a helix holds the backbone in a straight, rodlike cylinder
from which the side chains point outward. The relative hydro-
phobic or hydrophilic quality of a particular helix within a
protein is determined entirely by the characteristics of the side
chains. In water-soluble proteins, hydrophilic helices tend to
be found on the outside surfaces, where they can interact with
the aqueous environment, whereas hydrophobic helices tend to Amino Carboxyl
be buried within the core of the folded protein. The amino acid terminus terminus
proline is usually not found in a helices because the covalent
bonding of its amino group with a carbon in the side chain (b) Side view
prevents its participation in stabilizing the backbone through
normal hydrogen bonding. While the classic a helix is the most
intrinsically stable and most common helical form in proteins,
there are variations, such as more tightly or loosely t\VISted
helices. For example, in a specialized helix called a coiled coil
(described several sections farther on), the helix is more tightly FIGURE 3 -5 The fl sheet, another common secondary structure
wound (3.5 residues and 0.51 nm per turn). in proteins. (a) Top view of a simple three-stranded 13 sheet with
antiparallell3 strands, as indicated by the arrows that represent the
The p Sheet Another type of secondary structure, the 13 sheet, N-to-C orientations of the chains. The stabilizing hydrogen bonds
consists of laterally packed 13 strands. Each 13 strand is a short between the 13 strands are indicated by green dashed lines. (b) Side
(5- to 8-residue), nearly fully extended polypeptide segment. view of a 13 sheet. The projection of the R groups (green) above and
Unlike in the a helix, where hydrogen bonds occur between below the plane of the sheet is obvious in this view. The fixed bond
the amino and carboxyl groups in the backbone between angles in the polypeptide backbone produce a pleated contour.
nearly adjacent residues, hydrogen bonds in the 13 sheet occur
between backbone atoms in separate, but adjacent, [3 strands direction of the polypeptide backbone, often toward the pro-
and are oriented perpendicularly to the chains of backbone tein's interior. These short, U-shaped secondary structures
atoms (Figure 3-5a). These distinct 13 strands may be either are often stabilized by a hydrogen bond between their end
within a single polypeptide chain, with short or long loops residues (Figure 3-6). Glycine and proline are commonly
between the [3 strand segments, or on different polypeptide present in rums. The lack of a large side chain in glycine and
chains in a protein composed of multiple polypeptides. Figure
3-5b shows how two or more [3 strands align into adjacent
rows, forming a nearly two-dimensional [3 pleated sheet (or
simply pleated sheet), in which hydrogen bonds within the
plane of the sheet hold the [3 strands together as the side chains
stick out above and below the plane. Like a helices, [3 strands
have a directionality defined by the orientation of the peptide
bond. Therefore, in a pleated sheet, adjacent 13 strands can be
oriented in the same (parallel) or alternating opposite (anti-
parallel) directions with respect to each other. In Figure 3-5a,
you can see that the N-to-C orientations of the chains, indi-
cated by arrows, alternate directions between adjacent chains,
signifying an antiparallel sheet. In some proteins, [3 sheets
form the floor of a binding pocket or a hydrophobic core; in
prorems embedded in membranes the [3 sheets curve around
and form a hydrophilic central pore through which ions and FIGURE 3- 6 Structure of a fl t urn. Composed of four residues, 13
small molecules may flow (see Chapter 11 ). turns reverse the direction of a polypeptide chain (180° U-turn). The C"
carbons of the first and fourth residues are usually < 0.7 nm apart, and
PTurns Composed of four residues, [3 turns are located on those residues are often linked by a hydrogen bond.l3 turns facilitate
the surface of a protein, forming sharp bends that reverse the the folding of long polypeptides into compact structures.

3.1 Hierarchical Structure of Proteins 63

@ OVERVIEW ANIMATION: Oil Drop Model of Protein Structure

FIGURE 3-7 Oil drop model of protein folding.

The hydrophobic residues (blue) of a polypeptide chain
tend to cluster together, somewhat like an oil drop, on the Core
inside, or core, of a folded protein, driven away from the Folding
aqueous surroundings by the hydrophobic effect (see Water
Chapter 2). Charged and uncharged polar side chains (red) Unfolding
appear on the protein's surface. where they can form
stabilizing interactions with surrounding water and ions. Surface

Unfolded protein Folded protein

the presence of a built-in bend in proline allow the polypeptide protein conformation because of the relatively hydrophobic,
backbone to fold into a tight U shape. 13 turns help large pro- or "oily," core of a protein (Figure 3-7). Uncharged hydro-
teins to fold into highly compact structures. There are six types philic polar side chains are found •on both the surface and
of well-defined turns, their detailed structures depending on the inner core of proteins.
arrangement of H-bonding interactions. A polypeptide back- Proteins usually fall into one of three broad structural
bone also may contain longer bends, or loops. In contrast with categories, based on their tertiary structure: globular pro-
tight 13 turns, which exhibit just a few well-defined conforma- teins, fibrous proteins, and integral membrane proteins.
tions, longer loops can have many different conformations. Globular proteins are generally water-soluble, compactly
folded structures, often but not exclusively spheroidal, that
comprise a mixture of secondary structures (sec the struc-
Tertiary Structure Is the Overall Folding ture of myoglobin, below). fibrous proteins are large, elon-
of a Polypeptide Chain gated, often !.tiff molecules. Some fibrous proteins are
Tertiary structure refers to the overall conformation of a poly- composed of a long polypeptide chain comprising many
peptide chain-that is, the three-dimensional arrangement of tandem copies of a short amino acid sequence that forms a
all its amino acid residues. In contrast with secondary struc- single repeating secondary structure (see the structure of
tures, which are stabilized only by hydrogen bonds, tertiary collagen, the most abundant protein in mammals, in Fig-
structure is primarily stabilized by hydrophobic interactions ure 20-24 ). Other fibrous proteins are composed of repeat-
between nonpolar side chains, together with hydrogen bonds ing globular protein subunits, such as the helical array of
involving polar side chains and backbone amino and carboxyl G-actin protein monomers that forms the F-actin microfila-
groups. These stabilizing forces compactly hold together cle- ments (see Chapter 17). Fibrous proteins, which often
ments of secondary structure---a helices, 13 strands, turns, and aggregate into large multiprotein fibers that do not readily
coils. Because the stabilizing interactions are often weak, how- dissolve in water, usually play a structural role or partici-
ever, the tertiary structure of a protein is not rigidly fixed but pate in cellular movements. Integral membrane proteins arc
undergoes continual, minute fluctuations, and some segments embedded within the phospholipid bilayer of the mem-
within the tertiary structure of a protein can be so very mobile branes that enclose cells and organelles (see Chapter 10).
they are considered to be disordered-that is, lacking well- The three broad categories of proteins noted here arc not
defined, stable, three-dimensional structure. This variation in mutually exclusive-some proteins arc made up of combina-
structure has important consequences for the function and tions of two or even all three categories.
regulation of proteins.
Chemical properties of amino acid side chains help de- Different Ways of Depicting the Conformation
fine ternary structure. Disulfide bonds between the side
of Proteins Convey Different Types
chains of cysteine residues in some proteins covalently link
regions of proteins, thus restricting the proteins' flexibility of Information
and increasing the stability of their tertiary structures. Amino The simplest way to represent three-dimensional protein
actds with charged hydrophilic polar side chains tend to be structure is to trace the course of the backbone atoms, some-
on the outer surfaces of proteins; by interacting with water, times only the C.. atoms, with a solid line (called a Ca trace,
they help to make proteins soluble in aqueous solutions and Figure 3-Sa); the most complex rnuud ~huws every arom
can form noncovalent interactions with other water-soluble (Figure 3-Sb). The former shows the overall fold of the poly-
molecules, including other proteins. In contrast, amino acids peptide chain without consideration of the amino acid side
with hydrophobic nonpolar side chains arc usually seques- chains; the latter, a ball-and-stick model (with balls repre-
tered away from the water-facing surfaces of a protein, in senting atoms and sticks representing bonds), details the
many cases forming a water-insoluble central core. This interactions between side-chain atoms, including those that
observation led to what's known as the "oil drop model" of stabilize the protein's conformation and interact with other

64 CHAPTER 3 • Protein Structure and Function

 (a) Cu backbone trace (b) Ball and stick FIGURE 3-8 Four ways to visualize protein
structure. Shown here are four distinct methods
for representing the structure of a protein called
ras, a monomeric (single polypeptide chain)
protein that binds to guanosine diphosphate
(GDP, depicted in blue). (a) The Co. backbone trace
demonstrates how the polypeptide is tightly
packed into a small volume. (b) A ball-and-stick
representation reveals thP location of all atoms.
(c) A ribbon representation emphasizes how (3
strands (light blue) and a helices (red) are organized
in the protein. Note the turns and loops connecting
pairs of helices and strands. (d) A model of the
water-accessible surface reveals the numerous
lumps, bumps, and crevices on the protein surface.
Regions of positive charge are shaded purple;
(c) Ribbons (d) Solvent-accessible surface regions of negative charge are shaded red.

molecules as well as the atoms of the backbone. Even though

Structural Motifs Are Regular Combinations
both views are useful, the elements of secondary structure
are not always easily discerned in them. Another type of rep- of Secondary Structures
resentation uses common shorthand symbols for depicting A particular combination of two or more secondary struc-
secondary structure-for example, coiled ribbons or solid tures that form a distinct three-dimensional structure is
cylinders for a h~lices, flat ribbons or arrows for 13 strands, called a structural motif when it appears in multiple pro-
and flexible thin strands for !3 turns, coils, and loops (Figure teins. A structural motif is often, but not always, associated
3-Sc). In a variation of the basic ribbon diagram, ball-and- with a specific function. Any particular structural motif will
stick or space-filling models of all or only a subset of side frequently perform a common function in different proteins,
chains can be attached to the backbone ribbon. In this way, such as binding to a particular ion or small molecule, for
side chains that are of interest can be visualized in the con- example, calcium or ATP.
text of the secondary structure that is especially clearly rep- One common structural motif is the a helix-based coiled
resented by the ribbons. coil, or heptad repeat. Many proteins, including fibrous pro-
However, none of these three ways of representing pro- teins and DNA-regulating proteins called transcription fac-
tein structure conveys much information about the protein tors (see Chapter 7), assemble into dimers or trimers by
surface, which is of interest because it is where other molecules using a coiled-coil motif, in which a helices from two, three,
usually bind to a protein. Computer analysis can identify the or even four separate polypeptide chains coil about one
surface atoms that are in contact with the watery environ- another-resulting in a coil of coils; hence the name (Fig-
ment. On this water-accessible surface, regions having a ure 3-9a). The individual helices bind tightly to one another
common chemical character, such as, hydrophobicity or because each helix has a strip of aliphatic (hydrophobic, but
hydrophilicity, and charge characteristics, such ::~~positive not aromatic) side chains (leucine valine, t:Lc) running along
(basic) or negative (acidic) side chains, can be indicated by one side of the helix that interacts with a similar strip in the
·. coloring (Figure 3-Sd). Such models reveal the topography adjacent helix, thus sequestering the hydrophobic groups
of the protein surface and the distribution of charge, both away from water and stabilizing the assembly of multiple
important features of binding sites, as well as clefts in the independent helices. These hydrophobic strips are generated
surface where small molecules bind. This view represents a along only one side of the helix because the primary struc-
protein as it is "seen" by another molecule. ture of each helix is composed of repeating seven-amino-acid

3.1 Hierarchical Structure of Proteins 65

 (a) Coiled-coil motif (b) EFhand/helix-loop-helix motif (c) Zinc-finger motif
Ca2+
N N

..

FIGURE 3-9 Motifs of protein secondary structure. (a) The parallel calcium-binding and DNA-binding regulatory proteins. In calcium-
two-stranded coiled-coil motif {left) is characterized by two ex heIices binding proteins such as calmodulin, oxygen atoms' from five residues
wound around each other. Helix packing is stabilized by interactions in the acidic glutamate- and aspartate-rich loop and one water
between hydrophobic side chains (red and blue) present at regular molecule form ionic bonds with a Ca 2• ion. (c) The zinc-finger motif is
intervals along each strand and found along the seam of the inter- present in many DNA-binding proteins that help regulate transcription.
twined helices. Each ex helix exhibits a characteristic heptad repeat A Zn 2 ion is held between a pair of 13 strands (blue) and a single ex
sequence with a hydrophobic residue often, but not always, at helix (red) by a pair of cysteine residues and a pair of histidine residues.
positions 1 and 4, as indicated. The coiled-coil nature of this structural The two invariant cysteine residues are usually at positions 3 and 6, and
motif is more apparent in long coiled coils containing many such the two invariant histidine residues are at positions 20 and 24 in this
motifs (right). (b) An EF hand, a type of helix-loop-helix motif, consists 25-residue motif. [See A. Lew1t-Bentley and S. Rety, 2000, Curr. Opin. Struc. Bioi.
of two helices connected by a short loop in a specific conformation. 10:637-643; S. A. Wolfe, L. Nekludova, and C. 0. Pabo, 2000, Ann. Rev. Biophys.
This structural motif is common to many proteins, including many Biomol. Struc. 29:1 83-21 2.]

units, called heptads, in which the side chains of the first and gene activity (see Chapter 7). Yet another structural motif
fourth residues are aliphatic and the other side chains are commonly found in proteins that bind RNA o r DNA is the
often hydrophilic (Figure 3-9a). Because hydrophilic side zinc finger, which contains three secondary structures-an
chains extend from one side of the helix and hydrophobic ~helix and two~ strands with an antiparallel orientation-
side chains extend from the opposite side, the overall helical that form a fingerlike bundle held together by a zinc ion
structure is amphipathic. Because leucine frequently appears (Figure 3-9c).
in the fourth positions and the hydrophobic side chains The relationship between the primary structure of a poly-
merge together like the teeth of a zipper, these structural mo- peptide chain and the structural motifs into which it folds is
tifs are also called leucine zippers. not always straightforward. The amino acid sequences re-
Many other structural motifs contain ~ helices. A com- sponsible for any given structural motif may be very similar
mon calcium-binding motif called th e EF hand contains two to one another. In other words, a common sequence motif
short helices connected by a loop (Figure 3-9b). This struc- can result in a common structural motif. This is the case for
tural motif, one of several helix-turn-helix structural motifs, the heptad repeats that form coiled coils. However, it is pos-
is found in more than 100 proteins and is used for sensing sible for seemingly unrelated amino acid sequences to fold
the calcium levels in cells. The binding of a Ca 2 ~ ion to oxy- into a common structu ral motif, so it is not always possible
gen atoms in conserved residues in the luup depends on the to predict which amino acids sequences wtll fold mto a given
concentration of Ca 2 and often induces a conformational structural motif. Conversely, it is possible that a commonly
change in the protein, altering its activity. Thus calcium occurring sequence motif does not fold into a well-defined
concentrations can directly control proteins' structures and structural motif. Sometimes short sequence motifs that have
functions. Somewhat different helix-turn-helix and basic an unusual abundance of a particular amino acid, for exam-
helix-loop-helix (bHLH) structural motifs are used for pro- ple, proline or aspartate or glutamate, are called "domains";
tein binding to DNA and consequently the regulation of however, these and other short contiguous segments are

·.
66 CHAPTER 3 • Protein Structure and Function
 more appropriately called sequence motifs than domains, parts of a protein are responsible for particular activities ex-
which has a distinct meaning that is defined below. hibited by the protein. Indeed, functional domains are often
We will encounter numerous additional motifs in later dis- also associated with corresponding structural domains.
cussions of other proteins in this and other chapters. The pres- A structural domai11 is a region - 40 or more amino actds
·.· ence of the same structural motif in different proteins with in length, arranged in a single, stable, and distinct structure
similar functions clearly indicates that these useful combina- often comprising one or more secondary structures. Struc-
tions of secondary structures have been conserved in evolution. tural domains often can fold into their characteristic struc-
tures independently of the rest of the protein 111 which they
are embedded. As a consequence, distmct structural domains
Domains Are Modules of Tertiary Structure can be linked together-sometimes by short or long spacers-
Distinct regions of protein structure often are referred to as to form a large, multidomain protein. Each of the polypeptide
domains. There are three main classes of protei n domains: chains in the trimeric flu virus hemagglutinin, for example,
functional, structural, and topological. A functional domain contains a globular domain and a fibrous domain (Figure
is a region of a protein that exhibits a particular activity char- 3-1 Oa). Like structural motifs (composed of secondary struc-
acteristic of the protein, usually even when isolated from the tures), structural domains are incorporated as modules into
rest of the protein. For instance, a particular region of a pro- different proteins. The modular approach to protein archi-
tein may be responsible for its catalytic activity (e.g., a kinase tecture is particularly easy to recognize in large proteins,
domain that covalently adds a phosphate group to another which tend to be mosaics of different domains that confer
molecule) or binding ability (e.g., a DNA-binding domain or distinct activities and thus can perform different functions
a membrane-binding domain). Functional domains are often simultaneously. As many as 75 percent of the proteins in
identified experimentally by wh ittling down a protein to its eukaryotes have multiple structural domains. Structural do-
smallest active fragment with the aid of proteases, enzymes mains frequently are also functional domains in that they
that cleave one or more peptide bonds in a target polypep- can have an activity independent of the rest of the protein.
tide. Alternatively, the DNA encoding a protein can be mod- The epidermal growth factor (EGr) domain is a structural
ified so that when the modified DNA is used to generate a domain present in several proteins (Figure ;3-11 ). EGr is a
protein, only a particular region, or domain, of the full-length small, soluble peptide hormone that binds to cells in the em-
protein is made. Thus it is possible to determine if specific bryo and in skin and connective tissue in adults, causing them

(a)

Globular
domain
FIGURE 3 - 10 Tertiary and quaternary levels of
structure. The protein pictured here, hemagglutinin
(HAl, is found on the surface of the influenza virus.
This long, multimeric molecule has three identical
subunits, each composed of two polypeptide chains,
HA 1 and HA2 • (a) Tertiary structure of each HA subunit
comprises the folding of its helices and strands into a
compact structure that is 13.5 nm long and divided
into two domains. The membrane-distal domain
(silver) is folded into a globular conformation. The
membrane-proximal domain (gold) has a fibrous,
Fibrous
domain stem like conformation owing to the alignment of
two long a helices (cylinders) of HA 2 with 13 strands
in HA 1• Short turns and longer loops, often at the
surface of the molecule, connect the helices and
HA1
strands in each chain. (b) Quaternary structure of HA
is stabilized by lateral interactions between the long
helices (cylinders) in the fibrous domains of the three
N subunits (gold, blue, and green), forming a triple-
I
External '-1 I stranded coiled-coil stalk. Each of the distal globular
I
Viral I domains in HA binds sialic acid (red) on the surface
\
membrane \
I of target cells. Like many membrane proteins, HA
Internal I
I contains several covalently linked carbohydrate
c chains (not shown).

3.1 Hierarchical Structure of Proteins 67

 EGF it is possible to use that information to search for similar
precursor domains in other proteins and to suggest potentially similar
functions for those domains in those proteins.
~ Neu
Multiple Polypeptides Assemble into Quaternary
EGF Structures and Supra molecular Complexes
Multimeric proteins consist of two or more polypeptide
OOOOO TPA chains, which in this context are referred to as subunits. A
fourth level of structural o rganization, quaternary structure,
describes the number (stoichiometry) and relative positio ns
FIGURE 3 -1 1 Modular nature of protein domains. Epidermal
growth factor (EGF) is generated by proteolytic cleavage of a precursor
of the subunits in mu ltimeric proteins. Flu virus hemaggluti-
protein containing multiple EGF domains (green) and a membrane-
nin, for example, is a trimer of three identical subunits
spanning domain (blue). The EGF domain is also present in the Neu (homotrimer) held together by noncovalent bonds (see Fig-
protein and in tissue plasminogen activator (TPA). These proteins also ure 3-10b). Other multimeric proteins can be composed of
contain other widely distributed domains, indicated by shape and color. various numbers of identical (homomeric) or different (hetero-
[Adapted from I. D. Campbell and P. Bork, 1993. Curr. Opin. Struc. Bioi. 3:385.) meric) subunits. Hemoglobin, the oxygen-carryi ng molecule
in blood, is an example of a heteromeric multimeric protein.
It has two copies each of two different polypeptide chains
to divide. It is generated by proteolytic cleavage (breaking of (discussed below). Often, the individual monomer subunits
a peptide bond) between repeated EGF domains in the EGF of a multimeric protein cannot function normally unless they
precursor protein, w hich is anchored in the cell membrane by are assembled into the multimeric protein. In some cases,
a membrane-spanning domain. EGF domains with sequences assembly into a multimeric protein permits .proteins that act
similar to, but not identical to, those in the EGF peptide hor- seq uentially in a pathway to increase their efficiency of op-
mone are present in other proteins and can be liberated by eration owing to their juxtaposition in space, a phenomenon
proteolysis. These proteins include tissue plasminogen activa- referred to as ''metabolic coupling. " Classic examples of this
tor (TPA), a protease that is used to dissolve blood clots in coupling are fatty acid synthases, the enzymes in fungi that
heart attack victims; Neu protein, which takes part in embry- synthesize fatty acids, and the polyketide synthases, the large
onic differentiation; and Notch protein, a receptor protein in multiprotein complexes in bacteria that synthesize a diverse
the plasma membrane that functions in developmentally im- set of pharmacologically relevant molecules called polyketides,
portant signaling (see Chapter 16 ). Besides the EGF domain, including the antibiotic erythromycin.
these proteins ha ve other domains in common with other The highest level in th e hierarchy of protein structure is
proteins. For example, TPA possesses a trypsin domain, a the association of proteins into supramolecular complexes.
functional domain in some proteases. It is estimated that T ypically, such structures are very la rge, in some cases ex-
there are about 1000 different types of structural domains in ceeding 1 MDa in mass, approaching 30-300 nm in size,
all proteins. Some of these are not very common, whereas and containing tens to hundreds of polypeptide chains and
others are found in many different proteins. Indeed, by some sometimes o ther biopolymers such as nucleic acids. The
estimates only nine major types of structural domains ac- capsid that encases the nucleic acids of the viral genome is
cou nt for as much as a third of all the structural domains in an example of a supramolecular complex with a structural
all proteins. Structural domains can be recognized in proteins function. The bundles of cytoskeletal filaments that support
whose structures have been determined by x-ray crystallogra- and give shape to the plasma membrane are another exam-
phy or nuclear magnetic resonance (NMR) analysis or in im- ple. Other supramolec ula r complexes act as molecular ma-
ages captured by electron microscopy. chines, carrying out the most complex cellular processes by
Regions of proteins that are defined by their distinctive integrating multiple proteins, each with distinct functions,
spatial relationships to the rest of the protein are topological into one large assembly. For example, a transcriptional ma-
domains. For example, some proteins associated with cell- chine is re spo nsible for sy nthesizi ng messenger RNA
surface membranes can have a part extending inward into (mRNA ) using a DNA template. This transcriptional ma-
the cytop lasm (cytoplasmic domain), a part embedded chine, the operational details of which are discussed in
within the phospholipid bilayer membrane (me mbrane- Chapter 4, consists of RNA polymerase, itself a multimeric
spanning domain), and a part extending outward into the protein, and at least 50 additional components, including
t:xtracellular space (extracellular domain). Each of these can general transcription tactors, promoter- binding protein s,
comprise one or more structural and functional domains. helicase, and other protein complexes (Figure 3-12 ). Ribo-
In Chapter 6 we consider the mechanism by which the somes, also discussed in Chapter 4, are complex multiprotein
gene segments that correspond to domains became shuffled and multi-nucleic acid machines that synthesize proteins.
in the course of evolution, resulting in their appearance in One of the most complex multiprotein assemblies is the
many proteins. Once a functional, structural, or topological nuclear pore, a structure that allows communication and pas-
domain has been identified and characterized in one protein, sage of macromolecules between nucleoplasm and cytoplasm

68 CHAPTER 3 • Protein Structure and Function

 General transcription factors comparative approach is very powerful, caution must always

,.
n ."-' be exercised when attributing to one protein, or a part of a
protein, a similar function or structure to another based
+ ~o + only on amino acid sequence similarities. There are exam-
RNA polymerase ples in which proteins with similar overall structures display
different functions and cases in which functionally unre-
lated proteins with dissimilar amino acid sequences never-
DNA theless have very similar folded tertiary structures, as will be
f1'-'.E"~.E':f'S0t:=].~(l{'~~~("f.~ explained below. Neverrhele:.~, in many cases such compar-
Promoter
isons provide important insights into protein structure and
! function.
The molecular revolution in biology during the last de-
cades of the twentieth century created a new scheme of bio-
logical classification based on simi larities and differences in
Transcription the amino acid sequences of proteins. Proteins that have a
preinitiation common ancestor are referred to as homologs. The main
complex
evidence for homology among proteins, and hence for their
common ancestry, is similarity in their sequences, which is
often also reflected in similar structures. We can describe
FIGURE 3-12 A macromolecular machine: the transcription- homologous proteins as belonging to a "family" and can
initiation complex. The core RNA polymerase, general transcription trace their lineage from comparisons of their sequences.
factors, a mediator complex containing about 20 subunits, and other Generally, more closely related proteins will exhibit greater
protein complexes not depicted here assemble at a promoter in DNA. sequence similarity than more distantly related proteins be-
The polymerase carries out transcription of DNA; the associated cause, over evolutionary time, mutationt; accumulate in the
proteins are required for initial binding of polymerase to a specific
genes encoding these proteins. The folded thr~e-dimensional
promoter. The multiple components function together as a machine.
structures of homologous proteins can be similar even if parts
of their primary structure show little evidence of sequence
homology. Initially, proteins with relatively high sequence
(see Chapter 14 ). It i5 composed of multiple copies of about
similarities (> 50 percent exact matches, or "identities") and
30 distinct proteins and forms an assembly with an esti-
related functions or structures were defined as an evolution-
mated mas~ of around 50 megadaltons. The fatty acid syn-
thases and polyketide synthases referred to above are also arily related family, while a superfamily encompassed two or
more families in which the interfamily sequences matched
supramolecular machines.
less well (-30-40 percent identities) than within one family.
It is generally thought that proteins with 30 percent sequence
identity are likely to have similar three-dimensional struc-
Members of Protein Families Have a Common
tures; however, such high sequence identity is not required
Evolutionary Ancestor for proteins to share simi Ia r structures. Recently, revised
Studies of myoglobin and hemoglobin, the oxygen-carrying definitions of family and superfamily have been proposed, in
proteins in muscle and red blood cells, respectively, provided which a family comprises proteins with a clear evolutionary
early evidence that a protein's function derives from its relationship (>30 percent identity or additional structural
three-dimensional structure, which in turn is specified by and functional information showing common descent but
amino acid sequence. X-ray crystallographic analysis showed < 30 percent identity), while a superfamily comprises pro-
that the three-dimensional structures of myoglobin (a mono- teins with only a probable common evolutionary origin-for
mer) and the a and 13 subunits of hemoglobin (a a 2132 tetramer) example, lower percent sequence identities but one or more
arc remarkably simi lar. Sequencing of myoglobin and the common motifs or domains.
hemoglobin subunits revealed that many identical or chemi- The kinship among homologous proteins is most easily
cal ly similar residues are found in equivalent positions visualized by a tree diagram based on sequence analyses. For
throughout the primary structures of both proteins. A muta- example, the amino acid sequences of globins-the protems
tion in the gene encoding the 13 chain that results in the sub- hemoglobin and myoglobin and their relatives from bacteria,
stitution of a va line for a glutamic acid disturbs the folding plants, and animals-suggest that they evolved from an an-
and function of hemoglobin and causes sickle-cell anemia . cestral monomeric, oxygen-binding protein (Figure 3 13).
~imilar comparisons between other proteins conclusively With the passage of time, the gene for this ancestral protein
confirmed the relation between the amino acid sequence, slowly changed, initially diverging into lineages leading to
three-dimensiona l structure, and function of proteins. Use animal and plant globins. Subsequent changes gave rise to
of sequence comparisons to deduce protein function has myoglobin, the monomeric oxygen-storing protein in mus-
expanded substantially in recent years as the genomes of cle, and to the a and 13 subunits of the tetrameric hemoglo-
more and more organisms have been sequenced. While this bin molecule (a 2l32 ) of the vertebrate circulatory system.

3.1 Hierarchical Structure of Proteins 69

 Vertebrate
IHEMOGLO.BIN l
~ ......-- ---.

Hemoglobin

Ancestral
oxygen-binding
protein ~subunit Myoglobin leghemoglobin
of hemoglobin

FIGURE 3 -1 3 Evolution ofthe globin protein family. Left: A duplication gave rise to the a and ()subunits of hetnoglobin.
primitive monomeric oxygen-binding globin is thought to be the Right: Hemoglobin is a tetra mer of two a and two() subunits. The
ancestor of modern-day blood hemoglobins, muscle myoglobins, and structural similarity of these subunits with leghemoglobin and
plant leghemoglobins. Sequence comparisons have revealed that myoglobin, both of which are monomers, is evident. A heme molecule
evolution of the globin proteins parallels the evolution of animals and (red) noncovalently associated with each globin polypeptide is directly
plants. Major junctions occurred with the divergence of plant glob ins responsible for oxygen-binding in these proteins. [Adapted from
from animal globins and of myoglobin from hemoglobin. later gene R. C. Hardison, 1996, Proc. Nat'/ Acad. Sci. USA 93:5675.)

KEY CO CEPTS of Section 3.1 • Proteins often contain distinct domains, independently
folded regions with characteristic structu raJ, functional, and
Hierarchical Structure of Proteins topological properties (see Figure 3-10).
Proteins are linear polymers of amino acids linked together • The incorporation of domains as modules in different pro-
by peptide bonds. A protein can have a single polypeptide teins in the course of evolution has generated diversity in
chain or multiple polypeptide chains. The primary structure protein structure and function.
of a polypeptide chain is the sequence of covalently linked
amino acids that compose the chain. Various, mosrly nonco-
valent interactions between amino acids in the linear se-
The number and organization of individual polypeptide sub-
units in multimeric proteins define their quaternary structure.
...
quence stabilize a protein's specific folded three-dimensional • Cells contain large supramolecular assemblies, sometimes
structure, or conformation. called molecular machines, in which all the necessary par-
The a helix, ~ strand and sheet, and 13 turn are the most ticipants in complex cellular processes (e.g., DNA, RNA,
prevalent elements of protein secondary structure. Secondary and protein synthesis; photosynthesis; signal transduction)
structures are stabilized by hydrogen bonds between atoms are bound together.
of the peptide backbone (see Figures 3-4 through 3-6). • Homologous proteins are proteins that evolved from a
Protein tertiary structure results from hydrophobic inter- common ancestor and thus have similar sequences, struc-
actions between nonpolar side groups and hydrogen bonds tures, and functions. They can be classified into families and
and ionic interactions involving polar side groups and the superfamilies.
polypeptide backbone. These interactiom -;rabilize folding of
the protein, including its secondary structural elements, into
an overall three-dimensional arrangement.
3.2 Protein Folding
Certain combinations of secondary structures give rise to dif-
ferent structural motifs, which are found in a variety of proteins As noted above, when it comes to the architecture of pro-
and are often associated with specific functions (see Figure 3-9). teins, "form follows function." Thus it is essential that when
a polypeptide is synthesized with its particular amino acid

70 CHAPTER 3 • Protein Structure and Function

 sequence, it folds into the proper three-dimensional confor- there is no rotation possible about the peptide bond itself. As a
mation with the appropriate secondary, tertiary, and possi- consequence, the only flexibility in a polypeptide chain back-
bly quaternary structure if it is to fulfill its biological role bone, allowing it to twist and turn-and thus fold into different
within or outside cells . How is a protein with a proper se- three-dimensional shapes-is rotation of the fixed planes of ad-
quence generated? A polypeptide chain is synthesized by a jacent peptide bonds with respect to one another about two
complex process called translation, which occurs in the cyto- bonds: theCa - amino nitrogen bond (rotational angle called <I> )
plasm on a large protein-nucleic acid complex called a ribo- and the C.,-carbonyl carbon bond (rotational angle called '1').
some. During translation, a sequence of messenger RNA Yet a further constraint on the potential conformations
(mRNA) serves as a template from which the assembly of a that a polypeptide backbone chain can adopt is the fact that
corresponding amino acid sequence is directed. The mRNA only a limited number of <I> and 'I' angles are possible be-
is initially generated by a process called transcription, cause for most <I> and 'I' angles, the backbone or side cham
whereby a nucleotide sequence in DNA is converted, by atoms would come too close to one another and thus the as-
transcriptional machinery in the nucleus, into a sequence of sociated conformation would be highly unstable or even
mRNA. The intricacies of transcription and translation are physically impossible to achieve.
considered in Chapter 4. Here we describe the key determi-
nants of the proper folding of a newly formed or forming
The Amino Acid Sequence of a Protein
(nascent) polypeptide chain as it emerges from the ribosome.
Determines How It Will Fold
Planar Peptide Bonds Limit the Shapes While the constraints of backbone bond angles seem very
into Which Proteins Can Fold restrictive, any polypeptide chain containing only a few resi-
dues could, in principle, still fold into many conformations.
A critical structural feature of polypeptides that limits how for example, if the <I> and 'I' angles were limited to only eight
the chain can fold is the planar peptide bond. Figure 3-3 il- combinations, an 11-residue-long peptide would potentially
·. lustrates the amide group in peptide bonds in a polypeptide have 8" conformations; for even a small polypeptide of only
chain. Because the peptide bond itself behaves partially like 10 residues, that's about 8.6 million possible conformations!
a double bond, In general, however, any particular protein adopts only one
or just a few very closely related conformations called the
native state; for the vast majority of proteins, the native state is
the most stably folded form of the molecule and the one that
permits it to function normally. In thermodynamic terms,
the native state is usually the conformation with the lowest
free energy (G) (see Chapter 2).
the carbonyl carbon and amide nitrogen and those atoms di- What features of proteins limit their folding from very
rectly bonded to them must all lie in a fixed plane (Figure 3-14); many potential conformations to just one? The properties
of the side chains (e.g., size, hydrophobicity, ability to form
hydrogen and ionic bonds), together with their particular
sequence along the polypeptide backbone, impose key re-
strictions. For example, a large side chain such as that of
tryptophan might sterically block one region of the chain
from packing closely against another region, whereas a side
chain with a positive charge such as arginine might attract a
segment of the polypeptide that has a complementary nega-
tively charged side chain (e.g., aspartic acid ). Another ex-
ample we have already discussed is the effect of the aliphatic
side chains in heptad repeats in promoting the association of
helices and the consequent formation of coiled coils. Thus a
polypeptide's primary structure determines its secondary,
tertiary, and quaternary structures.
The initial evidence that the information necessary for a
FIGURE 3 -14 Rotation between planar peptide groups in
protein to fold properly is encoded in its amino acid se-
proteins. Rotation about the ("-amino nitrogen bond (the <I> angle)
quence came from in vitro studies on the refolding of puri-
and the ( " -carbonyl carbon bond (the 'I! angle) permits polypeptide fied proteins, especially the Nobel Prize-winning studies in
backbones, in principle, to adopt a very large number of potential the 1960s by Christian Anfinsen of the refolding of ribonu-
conformations. However, steric restraints due to the structure of the clease A, an enzyme that cleaves R:--JA. Others had previ-
polypeptide backbone and the properties of the amino acid side ously shown that various chemical and physical perturbations
chains dramatically restrict the potential conformations that any can disrupt the weak noncovalent interactions that stabilize
given protein can assume. the native conformation of a protein, leading to the loss of

3.2 Protein Folding 71

its normal tertiary structure. The process by which a pro- (a)
tein's structure (and thi~ can include secondary as well as
tertiary structure) is disrupted is called denaturation. Denatur-
ation can be induced by thermal energy from heat, extremes
of pH that alter the charges on amino acid side chains, and
exposure to denaturants such as urea or guanidine hydro-
chloride at concentrations of 6-8 M, all of which disrupt (b)
structure-stabilizing noncovalent interactions. Treatment
with reducing agents, such as 13-mcrcaptuethanul, that break
disulfide bonds can further destabilize disulfide-containing
proteins. Under such unfolding or denaturing conditions, a
population of uniformly folded molecules is destabilized and
converted into a collection of many unfolded, or denatured,
molecules that have many different non-native and biologi-
cally inactive conformations. As we have seen, a large number
(c)
of possible non-native conformations exist (e.g., 8" - 1).
The spontaneous unfolding of proteins under denaturing
conditions is not surprising, given the substantial increase in
entropy that accompanies the denatured protein assuming
many non-native conformations. What is striking, however,
is that when a pure sample of a single type of unfolded pro-
tein in a test tube is shifted back very carefully to normal
conditions (body temperature, normal pH levels, reduction
in the concentration of denaturants), some denatured poly-
peptides can spontaneously refold into their native, biologi- (d)
cally active states as in Anfinsen's experiments. This kind of
refolding experiment, as well as studies that show synthetic
proteins made chemically can fold properly, showed that the
information contained in a protein's primary structure can
be sufficient to direct correct refolding. Newly synthesized
proteins appear to fold into their proper conformations just
as denatured proteins do. The observed similarity in the
folded, three-dimensional structures of proteins with similar
amino acid sequences, noted in Section 3.1, provided addi-
tional evidence that the primary sequence also determines
protein folding in vivo. lt appears that formation of second-
ary structures and structural motifs occurs early in the fold- FIGURE 3- 15 Hypothetical protein-folding pathway. Folding
ing process, followed by assembly of more complex structural of a monomeric protein follows the structural hierarchy of primary
domains, which then associate into more complex tertiary (a)-+ secondary (b-d)-+ tertiary (e) structure. Formation of small
and quaternary structures (Figure 3-15). structural motifs (c) appears to precede formation of domains
(d) and the final tertiary structure (e).

Folding of Proteins in Vivo Is Promoted

by Chaperones such help, cells might waste much energy in the synthesis of
The conditions of refolding of a purified, denatured protein improperly folded, nonfunctional proteins, which would
1n a test tube differ markedly from the conditions under have to be destroyed to prevent their disrupting cell func-
which a newly synthesized polypeptide folds in a cell. The tion. Cells clearly have such mechanisms since more than 95
presence of other biomolecules, including many other pro- percent of the proteins present within cells have been shown
teins at very high concentration (- 300 mg/ml in mammalian to be in their native conformations. The explanation for the
cells), some of which are themselves nascent and in the pro- cell\ remarkable efficiency in promoting proper protein
cess of folding, can potentially inrcrfere with the autono folding is that cells make a set of proteins, called chaperones,
mous, spontaneous folding of a protein. Furthermore, that facilitate proper folding of nascent proteins.
although protein folding into the native state can occur in The importance of chaperones is highlighted b} the ob-
vitro, this does not happen for all unfolded molecules in a servations that many are evolutionarily conserved. Chaper-
timely fashion. Given such impediments, cells require a ones are found in all organisms from bacteria to humans,
faster, more efficient mechanism for folding proteins into and some arc homologs with high sequence similarity that
their correct shapes than sequence alone provides. Without use almost identical mechanisms to assist protein folding.

72 CHAPTER 3 • Protein Structure and Func tion

 Chaperones use ATP binding, ATP hydrolysis to ADP, and transiently binds to exposed hydrophobic regions of an in-
exchange of a new ATP molecule for the ADP to induce a completely folded or partially denatured target protein and
series of conformational changes essential for their function. then rapidly releases this substrate as long as ATP is bound
Chaperones can fold newly made proteins into functional (step 0 in Figure 3-16a). Hydrolysis of the bound ATP causes
conformations, refold misfolded or unfolded proteins into the molecular chaperone to assume a closed form that binds
functional conformations, disassemble potentially toxic pro- its substrate protein much more tightly and this tighter bind-
tein aggregates that form due to protein misfolding, and as- ing appears to facilitate the target protein's folding, in part by
semble and dismantle large multiprotein complexes. preventing it from aggregating with other unfolded proteins.
Chaperones, which in eukaryotes are located in every cellular The exchange of ATP for the protein-bound ADP causes a
compartment and organelle, bind to the target proteins whose conformational change in the chaperone that releases the tar-
folding they will assist. There are several different classes of get protein. If the target is now properly folded, It cannot
chaperones with distinct structures that all use ATP binding rebind to an Hsp70. If it remains at least partially unfolded,
and hydrolysis in a variety of ways to facilitate folding. These it can bind again to give it another chance to fold properly.
include (1) enhancing the binding of protein substrates and Additional proteins, such as the co-chaperone Hsp40 in
(2) switching the conformation of the chaperones. The ATP- eukaryotes (Dna] in bacteria), help increase efficiency of
dependent conformational switch is used (1) to optimize fold- Hsp70-mediated folding of many proteins by stimulating-
ing after one substrate is folded, (2) to return the chaperone together with the binding of substrate-the rate of hydrolysis
to its initial state so that it is available to help fold another of ATP by Hsp70/DnaK by 100- to 1000-fold (see step 6 in
molecule, and (3) to set the time permitted for refolding, Figure 3-16a). Members of four different families of nucleo-
which can be determined by the rate of ATP hydrolysis. tide exchange factors (e.g., GrpE in bacteria; BAG, HspBP,
Two general families of chaperones have been identified: and HspllO families in eukaryotes) also interact with the
Hsp70/DnaK, promoting the exchange of ATP for ADP. Mul-
• Molecular chaperones, which bind to a short segment of a
tiple molecular chaperones are thought to bind all nascent
protein substrate and stabilize unfolded or partly folded pro-
polypeptide chains as they are being synthesized on ribosomes.
teins, thereby preventing these proteins from aggregating
ln bacteria, 85 percent of the proteins are released from their
and being degraded;
chaperones and proceed to fold normally; ari even higher per-
• Chaperonins, which form small folding chambers into centage of proteins in eukaryotes follow this pathway.
which all or part of an unfolded protein can be sequestered, The Hsp70 protein family is not the only class of molecu-
giving it time and an appropriate environment to fold properly. lar chaperones. Another distinct class of molecular chaper-
ones is the Hsp90 family. Hsp90 family members are present
One reason that chaperones are needed for intracellular pro-
in all organisms except archaea. In eukaryotes there are dis-
tein folding is that they help prevent aggregation of unfolded
tinct Hsp90s located in different organelles and Hsp90 is one
proteins. Unfolded and partly folded proteins tend to aggre-
of the most abundant proteins in the cytosol (1-2 percent of
gate into large, often water-insoluble masses, from which it
total protein) . Although the range of protein substrates of
is extremely difficult for a protein to dissociate and then fold
Hsp90 chaperones is not as broad as for some other chaper-
into its proper conformation. ln part this aggregation is due
ones, the Hsp90s are critically important in cells. They help
to the exposure pf hydrophobic side chains that have not yet
cells cope with denatured proteins generated by stress (e.g.,
had a chance to be buried in the inner core of the fo lded
heat shock) and they ensure that some of their substrates,
protein. These exposed hydrophobic side chains on different
usually called "clients," can be converted from an inactive to
molecules will stick to one another, owing to the hydropho-
an active state or otherwise held in a functional conforma-
bic effect (see Chapter 2) and thus promote aggregation. The
tion. In some cases the Hsp90s form a relatively stable com-
risk for such aggregation is especially high for newly synthe-
plex with their clients until an appropriate signal causes their
sized proteins that have not yet completed their proper fold-
dissociation from the client, freeing the client to perform
ing. Chaperones prevent aggregation by binding to the target
some regulated function in the cells. These clients include
polypeptide or sequestering it from other partially or fully
transcription factors, such as the receptors for the steroid
unfolded proteins, rhus giving the nascent protein time to
hormones estrogen or testosterone that regulate sexual devel-
fold properly.
opment and function by controlling the activities of many
genes (see Chapter 7). Another type of Hsp90 client is en-
Molecular Chaperones The heat-shock protein Hsp70 in the zymes called kinases, which control the activities of many
cytosol and its homologs (Hsp70 in the mitochondrial ma- proteins by phosphorylation (see Chapters 15 and 16). It is
trix, BiP in the endoplasmic reticulum, and DnaK in bacteria) estimated that as many as 20 percent of all proteins in yeast
are molecular chaperone~. They were first identified by their are directly or indirectly influenced by the activities of Hsp90.
rapid appearance after a cell has been stressed by hear shock Unlike monomeric Hsp70, Hsp90 functions as a dimer in
(Hsp stands for "heat-shock protein"). Hsp70 and its homo- a cycle in which ATP binding, hydrolysis, and ADP release
logs are the major chaperones in all organisms. When bound are coupled to major conformational changes and to binding,
' 0
to ATP, the monomeric Hsp70 protein assumes an open form activation, and release of clients (Figure 3-16b ). Although
in which an exposed hydrophobic substrate binding pocket much about the mechanism of Hsp90 remains to be learned,

3.2 Protein Folding 73

0 FOCUS ANIMATION: Chaperone-Mediated Folding

FIGURE 3 · 16 Molecular chaperone- (a) Ribosome

mediat ed protein folding. (a) Hsp70. Many
proteins fold into their proper three-
Unfolded protein
dimensional structures with the assistance Nucleotide-binding domain
of Hsp70-like proteins. These molecular
chaperones transiently bind to a nascent
polypeptide as it emerges from a ribosome or
Or _.; U
to proteins that have otherwise unfolded. In
D
Rapid
the Hsp70 cycle, an unfolded protein
substrate binds in rapid equilibrium to the
Protein
open conformation of the substrate-binding
binding
domain (SBD, orange) of the monomeric
Hsp70, to which an ATP (red oval) is bound in
the nucleotide-binding domain (NBD, blue)
®J~1 . ;, Hsp70
ATPase cycle
lDnaJ/Hsp40

(step 0 ). The substrate binding pocket is l~ P; fJ

shown as a green patch on the substrate-
release EJ ,
binding domain. Co-chaperone accessory GrpE/BAG1
proteins (DnaJ/Hsp40) stimulate the hydroly-
sis of ATP to ADP (blue oval) and conforma-
~"ATP ADP7
tional change in Hsp70, resulting in the closed
form, in which the substrate is locked into the
SBD; here proper folding is facilitated (step F1
). Exchange of ATP for the bound ADP,
stimulated by other accessory co-chaperone (b) Nucleotide-binding domain Substrate-binding domain
proteins (GrpE/BAG 1), converts the Hsp70
back to the open form (step ill, releasing the
properly folded substrate (step D). (b) Hsp90.
Hsp90 proteins are dimers, whose monomers
contains an N-terminal NBD domain (blue), a
central substrate (client) binding domain (SBD,
orange), and a (-terminal dimerization domain
(gray). The Hsp90 cycle begins when there is no
nucleotide bound to the NBD and the dimer is
in a very flexible, open (Y-shaped) configura-
tion that can bind·substrates (step 0 ). Rapid
ATP binding leads to a slow conformational Hsp90
ATPase cycle
change in which the NBDs dimerize and the
SBDs move together into a closed conforma-
tion (step f) ). ATP hydrolysis results in folding
ofthe client and client protein release (steps il Highly
and ~ ). The ADP-bound form of Hsp90 can compact
adopt several conformations, including a highly
compact form. Release of ADP regenerates the
initial state, which can then interact with
additional clients (step ~ ). [Part (b) modified from
M. Taipale, D. F.Jarosz, and S. Lindquist. 2010, Nat.
Rev. Mol. Cei/Bial. 1 1 (7):515-528.] Closed

it is clear that clients bind to the "open" conformation, that clients (client specificity). Co-chaperones also ca n help coordi-
ATP binding leads to interaction of the ATP binding domains nate the acri virico; of Hsp90 and Hsp70. For example, Hsp70
and formation of a "closed" conformation, and t hat hydroly- can help begin the folding of a client that is then banded off
sis of ATP plays an important role in activating some client by a co-chaperone to Hsp90 for add itional processing. Hsp90
proteins and their subsequent release from t he Hsp90. We activity can also be influenced by its covalent modification by
also know that there are at least 20 co-chaperones that can small molecules. finally, Hsp90s can help cells recognize mis-
have profound effects on the activ ity of Hsp90, including folded proteins t hat are unable to refold and fac ilitate their
ATPasc activity and determining which proteins w ill be degradation by mechanisms discussed la ter in t his chapter.

74 CHAPTER 3 • Protein Structure and Function

 Thus, as part of the quality-control system in cells, chaper- interact with a homoheptameric co-chaperone "lid." The
ones can help properly fold proteins or faci litate the destruc- bacteria group I chaperonin, known as GroEL/GroF.S, is
tion of those that cannot fold properly. shown in Figure 3-l?a. In the bacterium E. coli, GroEL is
thought to participate in the folding of about 10 percent of
Chaperonins The proper folding of a large variety of newly all proteins. Group II chaperonins, which are found in the
synthesized proteins also requires the assistance of another cytosol of eukaryotic cells (e.g., TriC in mammals) and in
class of proteins, the chaperonins, also called Hsp60s. These archaea, can have eight to nine either homomeric or hetero-
huge cylindrical supramolecular assemblies are formed from meric subunits in each ring, and the "lid" function is incor-
two rings of oligomers. There are two distinct groups of porated in those subunits themselves-no <;ep;Jr<lte lid 1s
chaperonins that d1ffer somewhat in their structures, de- needed. It appears that ATP hydrolysis triggers the closing of
tailed molecular mechanisms, and locations. Group I chap- the lid of group II chaperonins.
eronins, found in prokaryotes, chloroplasts, and mitochondria, Figure 3-17b illustrates the GroEL/GroES cycle of protein
are composed of two rings, each having seven subunits that folding. A partly folded or misfolded polypeptide <60 kD in

(;} FOCUS ANIMATION: GroEL ATPase Cycle

(a) (b) Unfolded protein Unfolded protein Folding within

Ribosome binding in chamber in upper chamber upper chamber

Partially folded ,J ADP GroES l

GroES or misfolded + t
protein P;

GroEL
D
GroEL -~-~'""'-'~
T.J.--1
ATP
--J-.- (
ATP ~
.fJ
( ~
GroES---"""'

Two
independent
l Slow step
P;~
I EJ

~~
folding
chambers

Recycle
to refold

Incompletely
folded protein
~
~ ADP
or \ \
Properly
folded protein

FIGURE 3-17 Chaperonin-mediated protein folding. Proper GroEL rings that take place control the binding of the GroES lid that
folding of some proteins depends on chaperonins such as the prokary- seals the chamber (step f) ). The polypeptide remains encased in the
otic group I chaperonin GroEL. (a) GroEL is a barrel-shaped complex chamber capped by the lid, where it can undergo folding until ATP
of 14 identical -60,000-MW subunits arranged in two stacked rings hydrolysis, the slowest, rate-limiting step in the cycle (t112 - 10 s) (step II
(blue and red) of seven subunits each that form two distinct internal ), induces binding of ATP and a different GroES to the other ring
polypeptide folding chambers. Homoheptameric (1 0,000-MW subunits) (transient intermediate shown in brackets). This then causes the GroES
lids, GroES (yellow), can bind to either end of the barrel and seal the lid and ADP bound to the peptide-containing ring to be released,
chamber on that side. (b) The GroEL-GroES folding cycle. A partly folded opening the chamber and permitting the folded prote>in to diffuse out
or misfolded polypeptide enters one of the folding chambers (step 0 ). of the chamber (step [)). If the polypeptide folded properly, it can
The second chamber is blocked by a GroES lid. Each ring of seven GroEL proceed to function in the cell. If it remains partially folded or misfolded,
subunits binds seven ATPs, hydrolyzes them, and releases the ADPs in a it can rebind to an unoccupied GroEL and the cycle can be repeated.
set order coordinated with GroES binding and release and polypeptide [Part (a) modified from David L. Nelson and Michael M. 2000, Cox, Lehninger:
binding, folding, and release. The major conformational changes in the Principles of Biochemistry, 3d ed., W. H. Freeman and Co.]
.•

3.2 Protein Folding 75

mass is captured by hydrophobic residues near the entrance (a) (b)
of the GroEL chamber and enters one of the folding cham-
bers (upper chamber in Figure 3-17b). The second chamber is
blocked by a GroES lid. Each of the 14 subunits of GroEL
can bind ATP, hydrolyze it, and subsequently release ADP.
These reactions are concerted for each set of seven subunits
in a single ring and lead to major conformational changes.
These changes control both the binding of the GroES lid that
seals the chamber and the environment of the chamber in
which polypeptide fo lding takes place. The polypeptide
remains encased in the chamber capped by the lid. There it
can undergo folding until ATP hydrolysis in that chamber, the
slowest, rate-limiting step in the cycle (t 112 - 10 s), induces bind-
ing of ATP and a different GroES to the other ring. This then
causes the GroES lid and ADP bound to the peptide contain-
ing ring to be released, opening t he chamber and permitting
the folded protein to diffuse out of the chamber. If the poly-
peptide folded properly, it can proceed to function in the cell.
If it remains partially folded or misfolded, it can rebind to an
unoccupied GroEL and the cycle can be repeated. There is a
reciprocal relationship between the two rings in one GroEL
complex. The capping of one chamber by GroES to permit
sequestered substrate folding in that chamber is accompanied 20 f.l.m 100 nm
by the release of substrate polypeptide from the chamber of t____:___j
the second ring (simultaneous binding, fo lding, and release FIGURE 3-18 Alzheimer's disease is characterized by the
from the second chamber is not illustrated in Figure 3-17b). formation of insoluble plaques composed of amyloid protein.
There IS a striking similarity between the capped-barrel de- (a) At low resolution, an amyloid plaque in the brain of an Alzheimer's
sign of GroEL!GroES, in which proteins arc sequestered for patient appears as a tangle of filaments. (b) The regular structure of
folding, and the structure of the 26S proteasome that partici- filaments from plaques is revealed in the atomic force microscope.
pates in protein degradation (discussed in Section 3.4). In ad- Proteolysis of the naturally occurring amyloid precursor protein yields
dition, a group of proteins that arc part of the AAA family a short fragment, called ~-amyloid protein, that for unknown reasons
of ATPases are composed of hexameric rings with a central changes from an a-hel ical to a ~-sheet conformation. This alternative
structure aggregates into the highly stable filaments (amyloid) found
pore into which substrates can enter for folding or unfolding
in plaques. Similar pathologic changes in other proteins cause other
or in some cases proteolysis; examples of these will be dis-
degenerative diseases. [Courtesy of K. Kosik.]
cussed in Chapter 13.

Alternatively Folded Proteins Are filaments composing these structures derive from abundant
Implicated in Diseases natural proteins such as amyloid precursor protein, which is
embedded in the plasma membrane; Tau, a microtubule-
Recent evidence suggests that a protein may fold into
binding protein; and prion protein, an "infectious" protein.
an alternative three-dimensional structure as the result
Influenced by unknown causes, these a helix-containing pro-
of mutations, inappropriate cova lent modifications made
teins or their proteolytic fragments fold into alternative 13
after the protein is synthesized, or other as yet unidentified
sheet-containing structures that polymerize into very stable
reasons. Such "misfolding" not only leads to a loss of the
filaments. Whether the extracellular deposits of these fila-
normal function of the protein but often marks it for proteo-
ments or the soluble alternatively folded proteins are roxie to
lytic degradation. However, when degradation isn't com-
the cell is unclear. •
plete or doesn't keep pace with misfolding, the subsequent
accumulation of the misfolded protein or its proteolytic frag-
ments contributes to certain degenerative diseases character-
ized by the presence of insoluble, disordered aggregates of KFY CONCEPTS of section 3.2
twisted-together protein, or plaques, in various organs, in-
cludmg the liver and brain. Protein Folding
Some neurodcgenerative diseases, including Alzheimer's The primary structure (amino acid sequence) of a protein
disease and Parkinson's disease in humans and transmissible determines its three-dimensional structure, which determines
spongiform encephalopathy ("mad cow" disease) in cows and its function. In short, function derives from structure; struc- .·
sheep, are marked by the formation of tangled filamentous ture derives from sequence.
plaques in a deteriorating brain (Figure 3-18). The amyloid

76 CHAPTER 3 • Protein Structure and Function

 inverse of the equilibrium constant K e<J for the binding reac-
• Because protein function derives from protein structure, tion, is the most common quantitative measure of affinity
newly synthesized proteins must fold into the correct shape (see Chapter 2 ). The stronger the interaction between a pro-
to function properly. tein and ligand, the lower the value of KJ. Both the specific-
The planar structure of the peptide bond limits the number ity and the affinity of a protein for a ligand depend on the
of conformations a polypeptide can have (see figure 3-14 ). structure of the ligand-binding site. l-or high-affinity and
highly specific interactions to take place, the shape and
The amino acid sequence of a protein dictates its folding
chemical properties of the binding site must be complemen-
into a specific three-dimensional conformation, the native
t:=~ry to that of the ligand molecule, a property termed mo-
stare. Proteins will unfold, or denature, if trealeu under con-
lecular complementarity. As we saw in Chapter 2, molecular
ditions that disrupt the noncovalent interactions stabilizing
complementarity allows molecules to form multiple nonco-
their three-dimensional structures.
valent interactions at close range and thus stick together.
Protein folding in vivo occurs with assistance from ATP- One of the best-studied examples of protein-ligand bind-
dependent chaperones. Chaperones can influence proteins in ing, involving high affinity and exquisite specificity, is that
several ways, including preventing misfolding and aggrega- of antibodies binding to antigens. Antibodies arc proteins
tion, facilitating proper folding, and maintaining an appro- that circulate in the blood and arc made by the immune sys-
priate, stable structure required for subsequent protein activity tem (see Chapter 23) in response to antigens, which are usu-
(see Figure 3-16). ally macromolecules present in infectious agents (e.g., a
• There are two broad classes of chaperones: (1 ) molecular bacterium or a virus) or other foreign substances (e.g., pro-
chaperones, which bind to a short segment of a protein sub- teins or polysaccharides in pollens). Different antibodies are
strate, and (2) chaperonins, which form folding chambers generated in response to different antigens, and these anti-
into which all or part of an unfolded protein can be seques- bodies have the remarkable characteristic of binding specifi-
tered, giving it time and an appropriate environment to fold cally to ("recognizing") a part of the antigen, called an
properly. epitope, which initially induced the production of the anti-
body, and not to other molecules. Antibodies act as specific
• Some neurodegencrative diseases are caused by aggregates of
sensors for antigens, forming antibody-antigen complexes
proteins that are stably folded in an alternative conformation.
that initiate a cascade of protective reactions in cells of the
immune system.
All antibodies are Y-shaped molecules formed from two
identical longer, or heavy, chains and two identical shorter,
3.3 Protein Binding and Enzyme Catalysis or light, chains (Figure 3 - 19a). Each arm of an antibody
molecule contains a single light chain linked to a hcav}
Proteins perform an extraordinarily diverse array of activi- chain by a disulfide bond. Near the end of each arm arc six
ties both inside and outside cells, yet most of these diverse highly variable loops, called complementarity-determining
functions are based on the ability of proteins to engage in a regions (CDRs), which form the antigen-binding sites. The
common activity: binding. Proteins bind to themselves, to sequences of the six loops are highly variable among anti-
other macromolecules, to small molecu le!>, and to ions. In bodies, generating unique complementary ligand-binding
this section, we describe some key features of protein bind- sites that make them specific for different cpitopes (Figure
ing and then turn to look at one group of proteins, enzymes, 3-19b) . The intimate contact between these two surfaces,
,. in greater detail. The activities of the other functional classes stabilized by numerous noncovalent interactions, is respon-
of proteins (structural, scaffold, transport, regu Ia tory, sible for the extremely precise binding specificity exhibited
motor) will be described in later chapters. by an antibody.
The specificity of antibodies is so precise that they can
distinguish between the cells of individual members of a spe-
Specific Binding of Ligands Underlies
cies and in some cases between proteins that differ by only a
the Functions of Most Proteins single amino acid, or even between proteins with identical
The molecule to which a protein binds is called its ligand. In sequences and only different post-translational modifica-
some cases, ligand binding causes a change in the shape of a tions. Because of their specificity and the ease with which
protein . Such conformational changes arc integral to the they can be produced (see Chapter 23), antibodies are highly
mechanism of action of many proteins and are important in useful reagents used in many of the experiments discussed in
regulating protein activity. subsequent chapters.
Two properties of a protein characterize how it binds li- We will see many examples of protein-ligand binding
gands. Specificity refers to the ability of a protein to bind one throughout this book, including hormones binding to recep-
molecule or a very small group of molecules in preference to tors (see Chapter 15), regulatory molecules binding to DNA
all other molecules. Affimty refers to the tightness or strength (see Chapter 7), and cell-adhesion molecules binding to ex-
of binding, usually expressed as the dissociation constant tracellular matrix (see Chapter 20), to name just a few. Here
(KJ). The Kd for a protein-ligand complex, which is the we focus on how the binding of one class of proteins, enzymes,

3.3 Protein Binding and Enzyme Catalysi s 77

 specific catalyst, usually an enzyme. Another form of cata-
lytic macromolecule in cells is made from RNA. These RNAs
are called ribozymes (sec Chapter 4).
Thousands of different types of enzymes, each of which
catalyzes a single chemical reaction or set of closely related
reactions, have been identified. Certain enzymes are found in
the majority of cells because they catalyze the synthesis of
common cellular products (e.g., proteins, nucleic acids, and
phu~pholipids) or take part in harvesting energy from nutn-
ents (e.g., by the conversion of glucose and oxygen into car-
bon dioxide and water during cellular respiration). Other
enzymes are present only in a particular type of cell because
they catalyze chemical reactions unique to that cell type
(e.g., the enzymes in nerve cells that convert tyrosine into
dopamine, a neurotransmitter). Although most enzymes are
located within cells, some are secreted and function at extra-
cellular sites such as the blood, tb.e digestive tract, or even
outside the organism (e.g., toxic enzymes in the venom of
poisonous snakes).
Like all catalysts (see Chapter 2), enzymes increase the
rate of a reaction but do not affect the extent of a reaction,
which is determined by the change in free energy llG between
reactants and products and are not themselves permanently
changed as a consequence of the reaction tney catalyze. En-
zymes increase the reaction rate by lowering the energy of the
transition state and therefore the activation energy required
to reach it (Figure 3-20). In the test tube, catalysts such as
charcoal and platinum facilitate reactions but usually only at
high temperatures or pressures, at extremes of high or low
pH, or in organic solvents. Within cells, however, enzymes

FIGURE 3-19 Protein-ligand binding of antibodies. (a) Ribbon

model of an antibody. Every antibody molecule of the immunoglobulin Transi t ion state

I
(uncat alyzed)
lgG class consists of two identica l heavy chains (light and dark red) and
two identical light chains (blue) covalently linked by disulfide bonds.
The inset shows a diagram of the overall structure containing the two ~nca•
(.!)
heavy and two light chains. (b) The hand-in-glove fit between an Transition state
antibody and the site to which it binds (epitope) on its target antigen- >
Cl (catalyzed)
in this case, chicken egg-white lysozyme. Regions where the two Q;
c
Ql
molecules make contact are shown as surfaces. The antibody contacts Ql
Ql
the antigen with residues from all its complementarity-determining U:
regions (CDRs). In this view, the molecular complementarity of the
antigen and antibody is especially apparent where "fingers" extending
from the antigen surface are opposed to "clefts" in the antibody surface.
Products

Progress of reaction ~
to their ligands results in the catalysis of the chemical reac-
tions essential for the survival and function of cells. FIGURE 3-20 Effect of an enzyme on the activation energy of
a chemical reaction. This hypothetical reaction pathway depicts the
changes in free energy, G, as a reaction proceeds. A reaction will take
Enzymes Are Highly Efficient p lace spontaneously only if the total G of the products is less than that
and Specific Catalysts of the reactants (negative .lG). However, all chemical reactions proceed
through one or more high-energy transition states, and the rate of a
The subset of proteins that catalyze chem ical reactions, the reaction is inversely proportional to the activation energy (~G*), which
making and breaking of covalent bonds, is called enzymes, is the difference in free energy between the reactants and the transition
and enzymes' ligands are called substrates. Enzymes make state (highest point along the pathway). Enzymes and other catalysts
up a large and very important class of proteins-indeed, al- accelerate the rate of a reaction by reducing the free energy of the
most every chemical reaction in the cell is cata lyzed by a transition state and thus JoG'.

78 CHAPTER 3 • Protein Structure and Function

must function effectively in an aqueous environment at 37 oc the precise molecular complementarity between its substrate-
and 1 atmosphere pressure and at physiologic pH values, binding site and the substrate. Usually only one or a few sub-
usually 6.5-7.5 but sometimes lower. Remarkably, enzymes strates can fit precisely into a binding site.
exhibit immense catalytic power, accelerating the rates of re- The idea that substrates might bind to enzymes in the
actions 106-10 12 times that of the corresponding uncatalyzed manner of a key fitting into a lock was suggested first by
reactions under otherwise similar conditions. Emil Fischer in 1894. In 1913 Leonor Michaelis and Maud
Leonora Menten provided crucial evidence supporting this
hypothesis. They showed that the rate of an enzymatic reac-
An Enzyme's Active Site Binds Substrates tion wa!> proportional to the substrate concentration at low
and Carries Out Catalysis substrate concentrations, but that as the substrate concentra-
Certain amino acids of an enzyme are particularly important in tions increased, the rate reached a maximal velocity, Ymm
determining its specificity and catalytic power. In the native and became substrate concentration independent, with the
conformation of an enzyme, critically important amino acids value of Ymax being directly proportional to the amount of
(which usually come from different parts of the linear sequence enzyme present in the reaction mixture (Figure 3-22).
of the polypeptide) are brought into proximity, forming a cleft
in the surface called the active site (Figure 3-21) . An active site
usually makes up o nly a small part of the total protein, with (a)
the remaining part involved in folding of the polypeptide, regu- c:
lation of the active site, and interactions with other molecules. .g2 Vmax
An active site consists of two functionally important re- ~ ·c
Q) ::J
... Q)
2.0 -----------------------------------
gions: the substrate-binding site, which recognizes and binds b.~
the substrate or substrates, and the catalytic site, which car- c: "' 1.5 [E!'"' 1.0 unit
.g~
ries out the chemical reaction once the substrate has bound. "'-
§~ 1.0
The catalytic groups in the catalytic site are amino acid side o-
- u
chains and backbone carbonyl and amino groups. In some 0 .g 0.5
enzymes, the catalytic and substrate-binding sites overlap; in
others, the two regions are structurally distinct.
- ...
~
Q) 0
roo.
0~----~----------------------------------
lEI ~ 0.25 unit

The substrate-binding site is responsible for the remark-

able specificity of enzymes-their ability to act selectively on Concentration of substrate lSI
one substrate or a small number of chemically similar sub-
strates. The alteration of the structure of an enzyme's substrate (b)
by only one or a few atoms, or a subtle change in the geome- Vmax
try (e.g., stereochemistry) of the substrate, can result in a vari- 1.0 --------- --------------------------
c:
ant molecule that is no longer a substrate of the enzyme. As 0
·~
0.8
with the specificity of antibodies for antigens described above,
the specificity of enzymes for substrates is a consequence of
-
"'~
0
Q)
0.6

0.4 I
I
s·
(a) Catalytic site
"'
~
0.2 I

Km
I
for S I
~m for

OL--L----~-------------------------------
Concentration of substrate ((Sl or [S'J)
FIGURE 3 - 22 Km and Vmu for an enzyme-catalyzed reaction .
Km and Vmax are determined from analysis of the dependence of the initial
reaction velocity on substrate concentration. The shape of these hypo-
thetical kinetic curves is characteristic of a simple enzyme-catalyzed
reaction in which one substrate (5) is converted into product (P). The initial
velocity is measured immediately after addition of enzyme to substrate
Binding pocket before the substrate concentration changes appreciably. (a) Plots of the
initial velocity at two different concentrations of enzyme [E) as a function
of substrate concentration [5). The [5] that yields a half-maximal reaction
rate is the Michaelis constant Krrv a measure of the affinity of E for turning
FIGURE 3-21 Active site ofthe enzyme trypsin. (a) An enzyme's S into P. Quadrupling the enzyme concentration causes a proportional
active site is composed of a binding pocket, which binds specifically to increase in the reaction rate, and so the maximal velocity Vmax is quadru-
a substrate, and a catalytic site, which carries out catalysis. (b) A surface pled; the Krrv however, is unaltered. (b) Plots of the initial velocity versus
representation of the serine protease trypsin. Active site clefts containing substrate concentration with a substrate 5 for which the enzyme has a
the catalytic site (side chains of the catalytic triad 5er-195, Asp-1 02, and high affinity and with a substrate 5' for which the enzyme has a lower
His-57 shown as stick figures) and the substrate side chain specificity affinity. Note that the Vmax is the same with both substrates because (E] is
binding pocket are clearly visible. [Part (b) courtesy of P. Teesdale-Spittle.] the same but that Km is higher for 5', the low-affinity substrate.

3.3 Protein Binding and Enzyme Catalysis 79

 where the Michaelis constant, Km, a measure of the affinity of
an enzyme for its substrate, is the substrate concentration that
yields a half-maximal reaction rare (i.e., 1/2 Ymax in Figure
3-22). The Km is somewhat similar in nature, but not identical,
to the dissociation constant, Kd (see Chapter 2). The smaller
Catalytic site the value of Km, the more effective the enzyme is at making
product from dilute solutions of substrate and the lower the
Enzyme substrate concentration needed to reach half-maximal velociry.
The smaller the Kd, the lower the ligand concentration needed
to reach SO percent of binding. The concentrations of the vari-
ous small molecules in a cell vary widely, as do the Km values
..
j! Binding pocket
for the different enzymes that act on them. A good rule of
thumb is that the intracellular concentration of a substrate is
approximately the same as, or somewhat greater than, the Krn
value of the enzyme to which it binds.
Enzyme-substrate The rates of reaction at substrate saturation vary enor-
complex
mously among enzymes. The maximum number of sub-
strate molecules converted to product at a single enzyme
active site per second is called the turnover number, which
can be less than l for very slow enzymes. The turnover num-
ber for carbonic anhydrase, one of the fastest enzymes, is 6 X
105 molecules/s.
Many enzymes catalyze the conversion of substrates to
products by dividing the process into multiple, discrete chem-
ical reactions that involve multiple, distinct enzyme substrate
complexes (ES, ES' , ES", etc. ) generated prior to the final re-
Enzyme lease of the products:

E +S ~ ES ~ ES' ~ ES" ~ ... E + P

FIGURE 3-23 Schematic model of an enzyme's reaction The energy profiles for such multistep reactions involve mul-
mechanism. Enzyme kinetics suggest that enzymes (E) bind substrate tiple hills and valleys (Figure 3-24), and methods have been
molecules (5) through a fixed and limited number of sites on the developed to trap the intermediates in such reactions to learn
enzymes (the active sites). The bound species is known as an enzyme- more about the details of how enzymes catalyze reactions.
substrate (ES) complex. The ES complex is in equilibrium with the
unbound enzyme and substrate and is an intermediate step in the
conversion of substrate to products (P). Serine Proteases Demonstrate How
an Enzyme's Active Site Works
Serine proteases, a large family of protein-cleaving, or pro-
teolytic, enzymes, are used throughout the biological world-
They deduced that this saturation at high substrate con- to digest meals (the pancreatic enzymes trypsin, chymotrypsin,
centrations was due to the binding of substrate molecules (S) and elastase), to control blood clotting (the enzyme throm-
to a fixed and limited number of sites on the enzymes (E), bin), even to help silk moths chew their way out of their co-
and they called the bound species the enzyme-substrate (ES) coons (cocoonase). This class of enzymes usefully illustrates
complex. They proposed that the ES complex is in equilib- how an enzyme's substrate-binding site and catalytic site
rium with the unbound enzyme and substrate and is an inter- cooperate in multistep reactions to convert substrates to
mediate step in the ultimately irreversible conversion of products. Here we will consider how trypsin and its two evo-
substrate to product (P) (Figure 3-23): lutionarily closely related pancreatic proteases, chymotrypsin
and elastase, catalyze cleavage of a peptide bond:
E+S~ ES~E +P

and that the rate V0 of formation of product at a particular /:p

rs
substrate concentration J is given by what is now called the P-C-
1 ""0
+ ·.
Michaelis-Menten equation:

(3-1)
where in the polypeptide substrate P 1 is the part of the pro-
tein on theN-terminal side of the peptide bond and P2 is the

80 CHAPTER 3 • Protein Structure and Function

 (a) (b)

Transition X

Enzyme-transition
(.!)
state complex
> EX
~
~ Enzyme+
a> substrate

-
~ \
·.· Product

Substrate ES

Progress of reaction~ Progress of reaction~

S~X"'-+P E + S ~ ES ~ EX - E+P
FIGURE 3-24 Fr ee-energy reaction profiles of uncatalyzed and multiple discrete steps, in this case the initial formation of an ES
multistep enzyme-catalyzed reactions. (a) The free-energy reaction complex followed by conversion via a single transition state (EX*) to
profile of a hypothetical simple uncatalyzed reaction converting the free enzyme (E) and P. The activation energy for each of these steps
substrate (S) to product (P) via a single high-energy transition state. is significantly less than the activation energy for the uncatalyzed
(b) Many enzymes catalyze such reactions by dividing the process into reaction; thus the enzyme dramatically enhances the reaction rate.

portion on the C-terminal side. We first consider how serine are two key binding interactions. First, the substrate (black
proteases bind specifically to their substrates and then show polypeptide backbone) and enzyme (blue polypeptide
in detail how catalysis takes place. backbone) form hydrogen bonds that resemble a 13 sheet.
Figure 3-25a shows how a substrate polypeptide binds to Second, a key side chain of the substrate .that determines
the substrate-binding site in the active site of trypsin. There \.\hich peptide in the substrate is to be cleaved extends into
the enzyme's side-chain-specificity binding pocket, at the
(a) bottom of which resides the negatively charged side chain of
the enzyme's Asp-189. Trypsin has a marked preference for
hydrolyzing substrates at the carboxyl (C=O) side of an
Pept ide bond amino acid with a long positively charged side chain (a rgi-
to be cleaved nine or lysine) because the side chain is stabilized in the en-
Subst rate peptide
0 R1 H 0 Oxyanion
zyme's specificity binding pocket by the negative Asp-189.

~~~N0~
hole Slight differences in the structures of otherwise similar spec-
ificity pockets help explain the differing substrate specificities of
, 0 , R2 , two related serine proteases: chymotrypsin prefers large aro-
6 ' 6 I
Arginine side
chain (R3 ) in matic groups (as in Phe, Tyr, Trp), and elastase prefers the
J-....~N NH substrate small side chains of Gly and Ala (Figure 3-25b) . The un-

HN~NH
N charged Ser-189 in chymotrypsin allows large, uncharged,
H
0 hydrophobic side chains to bind stably in the pocket. The
Bindmg site ~~
·· -.<? Guanidinium
FIGURE 3-25 Substrate b i ndi ng in t he active site of tryp sin-like
Side-chain-specifictty group
serine p roteases. (a) The active site of trypsin (purple and blue molecule)
binding pocket
with a bound substrate (black molecule). The substrate forms a two-
stranded 13 sheet with the binding site, and the side chain of an arginine
(b) (R3) in the substrate is bound in the side-chain-specificity binding pocket.
Its positively charged guanidinium group is stabilized by the negative
charge on the side chain of the enzyme's Asp-189. This binding aligns the
peptide bond of the arginine appropriately for hydrolysis catalyzed by
the enzyme's active-site catalytic triad (side chains of Ser-195, His-57, and
Asp-1 02). (b) The amino acids lining thP side-chain-specificity binding
pocket determine its shape and charge and thus its binding properties.
Trypsin accommodates the positively charged side chains of arginine
and lysine; chymotrypsin, large, hydrophobic side chains such as
S"r-18fl phenylalanine; and elastase, small side chains such as glycine and
alanine. [Part (a) modified from J. J. Perona and C. 5. Craik. 1997, J. Bioi. Chern.
Trypsin Chymotrypsin Elastase 272(48):29987-29990.]

3.3 Protein Binding and Enzyme Catalysis 81

specificity of elastase is influenced hy the replacement of gly- on the carbonyl carbon in the substrate. This attack initially
cines in the sides of the pocket in trypsin with the branched forms an unstable transition state with four groups attached to
aliphatic side chains of va lines (Va l-216 and Val-190) that th is carbon {tetrahedral intermediate). Breaking of the C-N
obstruct the pocket (Figure 3-25b). As a consequence, large peptide bond then releases one part of the substrate protein
side chains in substrates are prevented from fitting into the (NH3-P2 ), while the other parr remaim covalently attached to
pocket of elastase, whereas substrates with the sho rt alanine the enzyme via an ester bond to the serine's oxygen, forming a
or g lycine side chains at this position can bind well and be relatively stable acyl enzyme intermediate. The subsequent
subject to subsequent cleavage. replacement of this oxygen by one from water, in a reaction
In the catalytic site, all three enzymes usc the hydroxyl involving another unsLable tetrahedral intermediate, leads to
group on the side chain of a serine in position 195 to catalyze release of the fi nal product (P 1-COOH ). The tetrahedral
the hydrolysis of peptide bonds in protein substrates. A cata- intermediate transition states are partially stabilized by hydro-
lytic triad fo rmed by the three side chains of Ser-195, His-57, gen bonding from the enzyme's backbone amino groups in
and Asp-l02 participates in wha t is essentially a two-step hy- what is called the oxyanion hole. The large family of serine
drolysiS reaction. Figure 3-26 show~ how the catalytic t riad proteases and related enzymes with an active site serine illus-
cooperates in breaking the peptide bond, with Asp- I 02 and trates how an efficient reaction mechanism is used over and
His-57 supporting the attack of the hydroxyl oxygen of Ser-195 over by distinct enzymes to catalyze similar reactions.

(a) ES complex (b) Tetrahedral intermediate (transition state) (c) Acyl enzyme (ES' complex)

Asp···· His+ H Ser

P2._N
H
)o/
/.{) •• -·
/
c-o '•,, Oxyanion
P1 ' hole

(f) EP complex (e) Tetrahedral intermediate (transition state) (d) Acyl enzyme (ES' complex)

Asp-- - ":\a~·:
/
c-o-'•,,
__ Oxyani on
p
1 ' hole

His-57 side chain

HN~N ~HN~N H
\ I -.;:----- \ + /
Active Inactive (low pH)

FIGURE 3-26 Mechanism of serine-protease-mediated release of one of the reaction products (NH 2-P 2). and formation of
hydrolysis of peptide bonds. The catalytic triad of Ser-195, His-57, the acyl enzyme (ES' complex). (d) An oxygen from a solvent water
and Asp-1 02 in the active sites of serine proteases employs a multistep molecule then attacks the carbonyl carbon of the acyl enzyme. (e) This
mechanism to hydrolyze peptide bonds in target protein s. (a) After a attack results in the formation of a second tetrahedral intermediate.
polypeptide substrate binds to the active site (see Figure 3-25), forming (f) Additional electron movements result in the breaking of the
an ES complex, the hydroxyl oxygen of Ser-195 attacks the carbonyl Ser-195-substrate bond (formation of the EP complex) and release of
carbon of the substrate's targeted peptide bond (yellow). Movements the final reaction product (P 1-COOH). The side chain of His-57, which
of electrons are indicated by arrows. (b) This attack results in the is held in the proper orientation by hydrogen bonding to the side cha in
formation of a transition state called the tetrahedral intermediate, in of Asp-1 02, facilitates catalysis by withdrawing and donating protons
which the negative charge on the substrate's oxygen is stabilized by throughout the reaction (inset). If the pH is too low and the side chain
hydrogen bonds formed with the enzyme's oxyanion hole. (c) Addi- of His-57 is protonated, it cannot participate in catalysis and the
tional electron movements result in the breaking of the peptide bond, enzyme is inactive.

82 CHAPTER 3 • Protein Structure and Function

 Lysosomal enzyme The pH sensitivity of an enzyme's activity can be due to
Chymotrypsin changes in the ionization of catalytic groups, groups that
c participate directly in substrate binding, or groups that influ-
">
·~
u
ence the conformation of the protein. Pancreatic proteases
"'
Ql evolved to function in the neutral or slightly basic conditions
E in the intestines; hence their pH optima arc -8. Proteases
>
N
c
Ql
and other hydrolytic enzymes that function in acidic condi-
Ql tions must employ a different catalytic mechanism. This is
.~
-ro the case for enzymes within the sromach (pH - 1) such as the
w protease pepsin or those within lysosomes (pH -4.5), which
0:

2 3 4 5 6 7 8 9 10 play a key role in degrading macromolecules within cells (see

pH lysosomal enzyme data in Figure 3-27) . Indeed, lysosomal
hydrolases, which degrade a wide variety of biomolecules
FIGURE 3 -27 pH dependence of enzyme activity.lonizable
(proteins, lipids, etc.), are relatively inactive at the pH in the
(pH-titratable) groups in the active sites or elsewhere in enzymes often
cytosol (-7), helping protect a cell from self-digestion if
must be either protonated or deprotonated to permit proper substrate
binding or catalysis or to permit the enzyme to adopt the correct
these enzymes escape the confines of the membrane-bound
conformation. Measurement of enzyme activity as a function of pH can
lysosome.
be used to identify the pK;s of these groups. The pancreatic serine One key feature of enzymatic catalysis not seen in serine
proteases, such as chymotrypsin (right curve), exhibit maximum activity proteases but found in many other enzymes is a cofactor or
around pH 8 because oftitration of the active site His-57 (required for prosthetic group. This "helper" group is a nonpolypeptide
catalysis, pK. -6.8) and of the amino terminus of the protein (required small molecule or ion (e.g., iron, zinc, copper, manganese)
for proper conformation, pK. -9). Many lysosomal hydrolases have that is bound in the active site and plays an essential role in
evolved to exhibit a lower pH optimum (- 4.5, left curve) to match the the reaction mechanism. Small organic prosthetic groups in
low internal pH in lysosomes in which they function. [Adapted from enzymes are also called coenzymes. Some of these are chemi-
P. Lozano, T. De Diego, and J. L. lborra, 1997 Eur. J. Biochem. 248{1 ):80-85, and cally modified during the reaction and thus need to be re-
W. A. Judice et at., 2004, Eur. J. Biochem. 271 (5):1046-1053.) placed or regenerated after each reaction; others are not.
Examples include NAD (nicotinamide adenine dinucleotide),
FAD (flavin adenine d inucleotide) (see Figure 2-33), and heme
The serine protease mechanism points out several general groups that bind oxygen in hemoglobin or transfer electrons
key features of enzymatic catalysis: (1) enzyme catalytic sites in some cytochromes (see Figure 12-14). Thus the chemistry
have evolved to stabilize the binding of a transition state, catalyzed by enzymes is not restricted by the limited num-
thus lowering the activation energy and accelerating the ber of types of amino acids in polypeptide chains. Many
overall reaction; (2) multiple side chains, together with the vitamins-for example, the B vitamins, thiamine (Btl, ribofla-
polypeptide backbone, carefully organized in three dimen- vin (B 2 ), niacin (B,), and pyridoxine (B 6 ), and vitamin C-
sions, work together to chemically transform substrate into which cannot be synthesized in mammalian cells, function as
product, often by multistep reactions; and (3) acid-base ca- or are used to generate coenzymes. That is why supplements
talysis mediated by one or more amino acid side chains is of vitamins must be added to the liquid medium in which
often used by enzymes, as when the imidazole group of His- mammalian cells are grown in the laboratory (see Chapter 9).
57 in serine proteases acts as a base to remove the hydrogen Small molecules that can bind to active sites and disrupt
from Ser-195's hydroxyl group. As a consequence, often catalytic reactions are called enzyme inhibitors. Such inhibi-
only a particular ionization state (protonated or nonproton- tors are useful tools for studying the roles of enzymes in cells
ated) of one or more amino acid side chains in the catalytic and whole organisms. Inhibitors that directly bind to an en-
site is compatible with cat.alysis, and thus rhe enzyme's activ- zyme's binding sire and thus compete directly with the nor-
ity is pH dependent. mal substrates' ability to bind substrate are called competitive
For example, the imidazole of His-57 in serine proteases, inhibitors. Alternatively, inhibitors can interfere with en-
whose pK, is -6.8, can help the Ser-195 hydroxyl attack the zyme activity in other ways, for example by binding to some
substrate only if it is not protonatcd. Thus the activity of the other site on the enzyme and changing its conformation; this
protease is low at pH <6.8, and the shape of the pH activity is called noncompetitive inhibition. Inhibitors complement
profile in the pH range 4-8 matches the titration of the His-57 the use of mutations in genes and a technique called RNA
side chain, which is governed by the Henderson-Hasselbalch interference (RNAi) for probing an enzyme's function in
equation, with an inflection near pH 6.8 (see chymotrypsin cells (see Chapter 5). In all three approaches the cellular con-
data in figure 3-27 and see Chapter 2). The activity drops at sequences of disrupting an enzyme's activity can be used to
higher pH values, generating a bell-shaped curve, because deduce the normal function of the enzyme. The same ap-
the proper folding of the protein is disrupted when the proaches can be used to study the functions of non-enzymatic
amino group at the protein's amino terminus (pK,, -9) is macromolecules. However, interpreting results of inhibitor
deprotonated; the conformation near the active site changes studies can be complicated if, as is often the case, the inhibi-
as a consequence. tors block the activity of more than one protein.

3.3 Protein Binding and Enzyme Catalysis 83

~ Small-molecule inhibition of protein activity is the basis In the simplest such mechanism, polypeptides with dif-
H for most drugs and also for chemical warfare agents. ferent catalytic activities cluster closely together as subunits
Aspirin inhibits enzymes called cyclooxygenases, whose prod- of a multimeric enzyme or assemble on a common "scaf-
ucts can cause pain. Sarin and other nerve gases react with the fold" that holds them together (Figure 3-28b). This arrange-
active serine hydroxyl groups of both serine proteases and a ment allows the products of one reaction to be channeled
related enzyme, acetylcholine esterase, which is a key enzyme directly to the next enzyme in the pathway. In some cases,
in regulating nerve conduction (see Chapter 22). • independent proteins have been fused together at the genetic
level to create a single multidomain, multifunctional enzyme
(Figure 3-28c). Metabolic coupling usually involves large
Enzymes in a Common Pathway Are Often multiprotein complexes, as described earlier in this chapter.
Physically Associated with One Another
Enzymes taking part in a common metabolic process (e.g.,
the degradation of glucose to pyruvate during glycolysis; see
Chapter 12) are generally located in the same cellular com-
partment, be it the cytosol, at a membrane, or within a par- KEY CONCEPTS of Section 3.3
ticular organelle. Within this compartment, products from Protein Binding and Enzyme Catalysis
one reaction can move by diffusion to the next enzyme in the
pathway. However, diffusion entails random movement and • A protein's function depends ~n its abil ity to bind other
can be a slow, relatively inefficient process for moving mol- molecules known as ligands. For example, antibodies bind
ecules between enzymes (Figure 3-28a). To overcome this to a group of ligands known as antigens and enzymes bind
impediment, cells have evolved mechanisms for bringing en- to reactants called substrates that will be converted by chem-
zymes in a common pathway into close proximity, a process ical reactions into products.
called metabolic coupling. • The specificity of a protein for a particular ligand refers to
the preferential binding of one or a few 'closely related li-
gands. The affinity of a protein for a particular ligand refers
(a) Reactants to the strength of binding, usually expressed as the dissocia-
I tion constant Kd.
Products • Proteins are able to bind to ligands because of molecular
complementarity between the ligand-binding sites and the
corresponding ligands.
Enzymes are catalytic proteins that accelerate the rate of
cellular reactions by lowering the activation energy and sta-
bilizing transition-state intermediates (see Figure 3-20). . ·.
• An enzyme's active site, which is usually only a small part
of the protein, comprises two functional parts: a substrate-
binding site and a catalytic site. The substrate-binding site is
OR responsible for the exquisite specificity of enzymes owing to
its molecular complementarity with the substrate and the
Products transition state.
(c) • The initial binding of substrates (S) to enzymes (E) results
Reactants Products
in the formation of an enzyme-substrate complex (ES),
'\.,.. which then undergoes one or more reactions catalyzed by
the catalytic groups in the active site until the products (P)
are formed.
FIGURE 3-28 Assembly of enzymes into efficient multienzyme • From plots of reaction rate versus substrate concentration,
complexes. In the hypothetical reaction pathways illustrated here, the two characteristic parameters of an enzyme can be deter-
initial reactants are converted into final products by the sequential mined: the Michaelis constant, K11, a rough measure of the
action of three enzymes: A, B, and C. (a) When the enzymes are free in
enzyme's affinity for converting substrate into product, and
solution or even constrained within the same cellular compartment,
the maximal velocity, Vrna" a measure of its catalytic power
the intermediates in the reaction sequence must diffuse from one
(see Figure 3-22).
enzyme to the next, an inherently slow process. (b) Diffusion is greatly
reduced or eliminated when the individual enzymes associate into • The rates of enzyme-catalyzed reactions vary enormously,
muitisubunit complexes, either by themselves or with the aid of a with the turnover numbers (number of substrate molecules
scaffold protein. (c) The closest integration of different catalytic converted to products at a single active site at substrate satu-
activities occurs when the enzymes are fused at the genetic level, ration) ranging between <1 to 6 X 105 molecules/s.
becoming domains in a single polypeptide chain.

84 CHAPTER 3 • Protein Structure and Function

 All three types of regulation play essential roles in the lives
Many enzymes catalyze the conversion of substrates to and functions of cells. In this section, we first discuss mecha-
product!. by dividing the process into multiple discrete chem- nisms for regulating the amount of a protein and then turn
ical reactions that involve multiple distinct enzyme substrate to noncovalent and CO\'alent interactions that regulate protein
complexes (£5', ES" etc.) . activity.
Serine proteases hydrolyze peptide bonds in protein sub-
strates using as catalytic groups the side chains of Ser-195,
Regulated Synthesis and Degradation
His-57, and Asp-102. Amino acids lining the specificity
binding pocket in the binding site of serine protcases deter- of Proteins Is a Fundamental Property of Cells
mine the residue in a protein substrate whose peptide bond The rate of synthesis of proteins is determined by the rate at
will be hydrolyzed and account for differences in the speci- which the DNA encoding the protein is convened to mRNA
ficity of trypsin, chymotrypsin, and elastase. (transcription), the steady-state amount of the active mRNA
Enzymes often use acid-base catalysis mediated by one or in the cell, and the rate at which the mRNA is converted into
more amino acid side chains, such as the imidazole group of newly synthesized protein (translation). These important
His-57 in serine proteases, to catalyze reactions. The pH de- pathways arc described in detail in Chapter 4.
pendence of protonation of catalytic groups (pK,) is often The life span of intracellular proteins varies from as short
reflected in the pH-rate profile of the enzyme's activity. as a few minutes for mitotic cyclins, which help regulate pas-
sage through the mitotic stage of cell division (see Chapter
• In some enzymes, nonpolypeptide small molecules or ions, 19), to as long as the age of an organism for proteins in the
called cofactors or prosthetic groups, can bind to the active lens of the eye. Protein life span is controlled primarily by
site and play an essential role in enzymatic catalysis. Small regulated protein degradation.
organic prosthetic groups in enzymes are also called coen- There are two especially important roles for protein degra-
zymes; vitamins, which cannot be synthesized in higher animal dation. First, degradation removes proteins that are potentially
cells, function as or are used to generate coenzymes. toxic, improperly folded or assembled, or damaged-including
• Enzymes in a common pathway are located within specific the products of mutated genes and prot~ins damaged by
cell compartments and may be further associated as domains chemically active cell metabolites or stress (e.g., heat shock).
of a monomeric protein, subunits of a multimeric protein, or Despite the existence of chaperone-mediated protein folding,
components of a protein complex assembled on a common some of the newly made proteins are rapidly degraded be-
·. scaffold (sec Figure 3-28). cause they arc misfolded. This might occur due to failure of
timely engagement of the necessary chaperones to guide their
folding or their defective assembly into complexes. Most
other proteins are degraded more slowly, about 1-2 percent
3.4 Regulating Protein Function degradation per hour in mammalian cells. Second, the con-
trolled destruction of otherwise normal proteins provides a
Most processes in cells do not take place independently of powerful mechanism for maintaining the appropriate levels
one another or at a constant rate. The activities of all pro- of the proteins and their activities and for permitting rapid
teins and other 'biomolecules are regulated to integrate th eir changes in these levels to help the cells respond to changing
functions for optimal performance for survival. For example, conditions.
the catalytic activity of enzymes is regulated so that the Eukaryotic cells have several pathways for degrading
amount of reaction product is just sufficient to meet the proteins. One major pathway is degradation by enzymes
needs of the cell. As a result, the steady-state concentrations within lysosomes, membrane-limited organelles whose acidic
of substrates and products will vary depending on cellular interior (pH - 4.5) is filled with a host of hydrolytic en-
conditions. Regulation of nonenzymatic proteins-the open- zymes. Lysosomal degradation is directed primarily toward
ing or closing of membrane channels or the assembly of a aged or defective organelles of the cell-a process called au-
macromolecular complex, for example-is also essential. tophagy (see Chapter 14)-and toward extracellular pro-
In general, there are three ways to regulate protein activ- teins taken up by the cell. Lysosomes will be discussed at
ity. First, cells can increase or decrease the steady-state level length in later chapters. Here we will focus on another im-
of the protein by altering its rate of synthesis, its rate of deg- portant pathway: cytoplasmic protein degradation by pro-
radation, or both. Second, cells can change the intrinsic ac- teasomes, which can account for up to 90 percent of the
tivity, as distinct from the amount, of the protein. For protein degradation in mammalian cells.
example, through noncovalent and covalent interactions, cells
can change the affinity of substrate binding, or the fraction of
The Proteasome Is a Molecular Machine
time the protein is in an active versus inactive conformation.
Third, there can be a change in location or concentration Used to Degrade Proteins
with in the cell of the protein itself, the target of the protein's Proteasomes are very large, protein-degrading macromolec-
activity (e.g., an enzyme's substrate), or some other molecule ular machines that influence many different cellular func-
required for the protein's activity (e.g., an enzyme's cofactor). tions, including the cell cycle (see Chapter 19), transcription,

3.4 Regulating Protein Function 85

~ Overview Animation: The Proteosome

(b)
-NH2 FIGURE 3-29 Ubiquitin- and proteasome-mediated
proteolysis. (a) Left: The 265 proteasome has a
E1 E2 Cytosolic
Ub cylindrical structure with a 195 cap (blue) at each end of
AMP I I target protein
C=O a 205 core consisting of four stacked heptameric rings

J.~o ~ 10
+ ATP + PP; E2 E1
l (-110 Adiameter X 160 Along) each containing either
E1 \ jl Ub \ J. l•opoptld• o. (outer rings) or~ (inner rings) subunits (yellow). Right:
D fJ
~E2 bond
Cutaway view of 205 core showing inner thdmbers.

-NHJ~-Ub
Proteolysis of ubiquitin-tagged proteins occurs within
0
El = ubiquitin-actlvating enzyme the inner chamber of the core formed by the 13 rings.
E2 = ubiquitin-conjugating enzyme (b) Proteins are targeted for proteasomal degradation
E3 = ubiquitin ligase by polyubiquitination. Enzyme E1 is activated by attach-
Ub = ubiquitin 0 ment of a ubiquitin (Ub) molecule (step 0 ) and then
transfers this Ub molecule to a cysteine residue in E2
~ (step 0 ). Ubiquitin ligase (E3) transfers the bound Ub
EJ 1 Steps 1, 2, 3
(a)
!(n times) molecule on E2 to the side-chain -NH 2 of a lysine

k
residue in a target proteili, forming an isopeptide bond
(step 0 ). Additional Ub molecules are added to the
Ub -~Ub~Ub n + 1 Ub-modified target protein via isopeptide bonds to the
previously added Ub by repeating steps 0--11. forming
a polyubiquitin chain (step [)). The polyubiquitinylated
target is recognized by the proteasome cap, which uses
ATP~ Iil deubiquitinase enzymes to remove the Ub groups and
ADP ..-/! ATP hydrolysis to drive unfolding and transfer of the
unfolded protein into the proteolysis chamber in the core,
Release
Recognition from which the short peptide digestion fragments are
Ub later released (step 111) [Part (a) from W. Baumeister et at., 1998,
Ub Ub Ub
Ub Ub Ub Ce//92:357, courtesy of W. Baumeister and modified from M.
"---" Bochtler et al.,1999, Ann. Rev. Biophys. Biomol. Struct. 28:295-317.]
ATP) ............_ Unfolding

Peptides

..,. J\ ~Discharge

DNA repair (see Chapter 4), programmed cell death or them into the inner chamber of the proteasome's catalytic
apoptosis (see Chapter 21 ), recognition and response to in- core. Genetic studies in yeast have shown that cells cannot
fection by foreign organisms (see Chapter 23), and removal survive without functional proteasomes, thus demonstrating
of misfolded proteins. There are approximately 30,000 pro- their importance. Furthermore, proper proteasomal activity
teasomes in a typical mammalian celL is so important that cells will expend as much as 30 percent
Proteasomes consist of -50 protein subunits and have a of the energy needed to synthesize a protein to degrade it in
mass of 2-2.4 X 10 6 Da_ Proteasomes have a cylindrical, a proteasomc.
barrel-like catalytic core called the 205 proteasome (where S The 205 proteasomal catalytic core comprises two inner
is a Svedberg unit based on the sedimentation properties of rings of seven 13 subunits each, with three proteolytic active
the particle and is proportional to its size), which is approx- sites per ring facing toward the inner chamber of the
imately 14.8 nm tall and 11.3 nm in diameter. Bound to the -1. 7-nm-diamcter barrel and two outer rings of seven ex
ends of this core arc either one or two 195 cap complexes subuni ts each that control substrate access (Figure 3-29a).
(Figure 3-29a) that regulate the activity of the 20S car::~ lyric Proteasomes can degrade most proteins thoroughly because
core. When the core and one or two caps are combined, they the three active sites in each 13 subunit ring can cleave after
are both referred to as the 265 complex, even though the hydrophobic residues, acidic residues, or basic residues.
two-cap-containing complex is larger (30$). A 19$ cap has Polypeptide substrates must enter the chamber via a regu-
16-18 protein subunits, six of which can hydrolyze ATP lated - L3-nm-diameter aperture at the center of the outer ex
(i.e., they are AAA-type ATPases) to provide the energy subunit rings_ In the 265 proteasome, the opening of the ap-
needed to unfold protein substrates and selecti vely transfer erture, which is narrow and often allows the entry of only

86 CHAPTER 3 • Protein Structure and Function

unfolded proteins, is controlled by ATPases in the 19$ cap Subsequent ligase reactions covalently attach the C-terminal
that also participate in selectively binding and unfolding the glycine of an additional ubiquitin via an isopeptide bond
substrate (Figure 3-29b, bottom right ). The short peptide to the side chain of lysine 48 of the previously added
products of proteasomal digestion (2-24 residues long) exit ubiquitin to generate a polyubiquitin chain covalently
the chamber and are further degraded rapidly by cytosolic attached to the target protein.
peptidases, eventually being converted to individual ("free")
Following attachment of four or more ubiquitins in the
am ino acids. One researcher has quipped that a proteasome
polyubiquitin chain, the 19S regulatory cap of the 265 pro-
is a "cellular chamber of doom" in which proteins suffer a
teasome (sometimes with tht> help of accessory proteins) rec
"death by a thousand cuts."
ognizes the polyubiquitin-labeled proteins and unfolds and
transports them into the proteasome for degradation (see
Inhibitors of proteasome function can be used thera-
Figure 3-29b). As a polyubiquitinated substrate is unfolded
peutically. Because of the global importance of protea-
and passed into the core of the proteasome, enzymes called
somes for cells, continuous, complete inhibition of
deubiquitinases (Dubs) hydrolyze the bonds between indi-
proteasomes k ills cells. However, partial proteasome inhibi-
vidual ubiquitins and between the targeted protein and ubiq-
tion for short intervals has been introduced as an approach
uitin, recycling the ubiquitins for additional rounds of
to cancer chemotherapy . To survive and grow, cells nor-
protein modification (see Figure 3-29b). Analysis of the
mally require the robust activity of a regulatory protein
human genome sequence indicates the presence of -90
called NFkB (see C hapter 16) as well as other, similar "pro-
Dubs, about 80 percent of which use cysteine in a catalytic
survival" proteins. In turn, NFkB can function fully and pro-
triad, similar to serine proteases described earlier. In some
mote survival only when its inhibitor, IkB, is disengaged and
Dubs, zinc is a key participant in the catalytic reactions.
degraded by proteasomes (see Chapter 16). Partial inhibition
of proteasomal activity by a small-molecule inhihitor drug
results in increased levels of IkB and, consequently, reduced Specificity of Degradation Targeting of specific proteins for
NFkB activity (tha t is, loss of pro-survival activity) . Cells proteasomal degradation is primarily achieved through the
subsequently die hy apoptosis. Because at least some types of substrate specificity of the E3 ligase. As a testament to their
tumor cells are more sensitive to being killed by proteasome importance, there are mor e than 600 predicted ubiquitin
inhibitors than normal cells are, controlled administration of ligase genes in the human genome. The many E3 ligases in
proteasome inhibitors, at levels that kill the cancer cells but mammalian cells ensure that the wide variety of proteins to
not norma l cells, has proven to be an effective therapy for at be polyubiquitinatcd can be modified when necessary. Some
least one type of lethal cancer, multiple myeloma. • E3 ligases are associated with chaperones that recognize un-
folded or misfolded proteins; for example, the E3 ligase
CHIP is a co-chaperone for Hsp70. These and other proteins
Ubiquitin Marks Cytosolic Proteins (co-c haperones, escort factors, adaptors) can mediate E3-
for Degradation in Proteasomes ligase-catalyzed polyubiquitination of these dysfunctional
proteins that cannot be readily refolded properly and their
If proteasomes are to rapidly degrade only those proteins that delivery to proteasomes for degradation. In such cases, the
are either defective or scheduled to be removed, they must be chaperone-ubiquitination-proteasome system works in con-
able to distinguish between those proteins that need to be de- cert for protein quality control.
graded and those that don't. Cells mark proteins that should In addition to quality control, the ubiquitin-proteasome
be degraded by covalently attaching to them a linear chain of system can be used to regulate the activity of important cel-
multiple copies of a 76-residue polypeptide called ubiquitin lular proteins. An example is the regulated degradation of
that is highly conserved from yeast to humans. This "polyu- proteins called cyclins, which control the cell cycle (see Chap-
biquitin tail" serves as a c~llular "kiss of death," marking the ter 19 ). Cyclins contain the internal sequence Arg-X-X-Leu-
protein for destruction in the proteasome. The ubiquitination Gly-X-Ile-Gly-Asp/Asn (X can be any amino acid), which is
process (Figure 3-29b) involves three distinct steps: recognized by specific ubiquitinating enzyme complexes. At
1. Activation of ubiquitin-activating enzyme (E 1) hy the ad- a specific time in the cell cycle, each cyclin is phosphorylated
dition of a ubiquitin molecule, a reaction that requires ATP; by a cyclin kinase. This phosphorylation is thought to cause
a conformational change that exposes the recognition se-
2. Transfer of this ubiquitin molecule to a cysteine residue quence to the ubiquitinating enzymes, leading to polyubiqui-
in ubiquitin-conjugating enzyme (£2); tination and proteasomal degradation.
3. Formation of a covalent bond between the carboxyl
group of the C-terminal glycine 76 of the uhiquitin bound Other Functions of Ubiquitin and Ubiquitin-Rel ated Mole-
to E2 and the amino group of the side chain of a lysine cules There are several close relatives of ubiquitin that employ
residue in the target protein, a reaction catalyzed by ubiquitin- similar E1-, E2-, and E3-dependent mechanisms of activation
protein ligase (E3). This type of bond is called an isopeptide and transfer to acceptor substrates. These ubiquitin-like modi-
bond because it covalently links a side chain amino group, fications control processes as diverse as nuclear import, regu-
rather than the a amino group, to the carboxyl group. lated by the ubiquitin-like modifier Sumo, and autophagy,

3.4 Regulating Protein Function 87

regulated by the ubiquitin-like modifier Atg8/LC3 (see Chap-
ter 14 ). The attachment of ubiquitin to a target protein can be
used for purposes other than to mark the protein for degrada-
tion, as we will discuss later in this chapter, and some of these c::
involve polyubiquitin linkages other than those via Lys48. 0
·~
As is the case for ubiquitination itself, deubiquitination is
2
involved in processes other than proteasome-mediated pro- C1)
(f)
tein degradation. Large-scale, mass-spectroscopy-based '#
"proteomic" methods described later in t his chapter together
with sophisticated computational approaches have suggested
that Dubs, which are often bound in multiprotein com-
plexes, are involved in an extraordinarily wide range of cell
0 20 40 60 80 100
processes. These vary from cell division and cell cycle con-
trol (see Chapter 19) to membrane trafficking (see Chapter
J 4) to cell signaling pathways (see Chapters J 5 and 16).
i p0 2 (torr)

p0 2 in capillaries
i
p0 2 in alveoli
of active muscles of lungs '.

EXPERIMENTAL FIGURE 3-30 Hemoglobin binds oxygen

Noncovalent Binding Permits Allosteric,
cooperatively. Each tetra me ric hemoglobin protein has four oxygen-
or Cooperative, Regulation of Proteins binding sites; at saturation all the sites are loaded with oxygen. The
In addition to regulating the amount of a protein, cells can oxygen concentration in tissues is common ly measured as the partial
also regulate the intrinstc activity of a protein. One of the pressure (p0 1). P50 is the p02 at which half the oxygen-binding sites
most important mechanisms for regulating protein function at a given hemoglobin concentration are occupied; it is somewhat
is through allosteric interactions. Broadly speaking, allostery analogous to the Km for an enzymatic reaction. The large change in the
amount of oxygen bound over a small range of p01 values permits
(from the Greek "other shape") refers to any change in a
efficient unloading of oxygen in peripheral tissues such as muscle. The
protein's tertiary or quaternary structure, or in both, in-
sigmoidal shape of a plot of percent saturation versus ligand concen-
duced by the noncovalent binding of a ligand. When a ligand
tration is indicative of cooperative b inding, in which the binding of one
binds to one site (A) in a protein and induces a conforma- oxygen molecule allosterically influences the binding of subsequent
tional change and associated change in activity of a different oxygens. In the absence of cooperative binding, a binding curve is a
site (B), the ligand is called an allosteric effector of the pro- hyperbola, similar to the curves in Figure 3-22. [Adapted from L. Stryer,
tein, while site A is called an allosteric binding site, and the 1995, Biochemistry, 4th ed., W. H. Freeman and Company.]
protein is called an allosteric protein. By definition, allosteric
proteins have multiple binding sites, at least one for the al-
losteric effector and at least one for other molecules with conformational change whose effect spreads to the other
which the protein interacts. The allosteric change in activity subunits, lowering the Kd (increasing the affinity) for the
can be positive or negative; that is, it can induce an increase binding of additional oxygen molecules to the remaining
or a decrease in protein activity. Negative allostery often in- hemes and yielding a sigmoidal oxygen-bind ing curve (Fig-
volves the end product of a multistep biochemical pathway ure 3-30). Because of the sigmoidal shape of the oxygen-
binding to and reducing the activity of an enzyme that cata- saturation curve, it takes only a fourfold increase in oxygen
lyzes an early, rate-controlling step for that pathway. In this concentration for the percent saturation of the oxygen-binding
way excessive buildup of the product is prevented. This kind of sites in hemoglobin to go from 10 to 90 percent. Conversely,
regulation of a metabolic pathway is also called end-product if there were no cooperati\ ity and the shape of the curve was
mhibition or feedback rnhibition. Allosteric regulation is typical of that for Michaelis-Menten-type, or noncooperative,
particularly prevalent in multimcric enzymes and other pro- binding, it would take an 81-fold increase in oxygen concen-
teins where conformational changes in one subunit arc tration to accomplish the same increase in loading of its
transmitted to an adjacent subunit. binding sites in hemoglobin. This cooperativity permits hemo-
Cooperativity is a term often used synonymously with globin to take up oxygen ,·cry efficiently in the lungs, where the
allostery and usually refers to the influence (positive or nega- oxygen concentration is high and unload it in tissues, where the
tive) that the binding of a ligand at one site has on the bind- concentration is low. Thus cooperativity amplifies the sensitiv-
ing of another molecule of the same type of ligand at a it) of a system to concentration changes in its ligands, provid-
different site. Hemoglobin presents a classic example of pos- ing in many cases selective evolutionary advantage.
itive cooperative binding in thal the binding of a single li-
gand, molecular oxygen (0 2 ), increases the affinity of the Noncovalent Binding of Calcium and GTP
binding of the next oxygen molecule. Each of the four sub-
Are Widely Used as Allosteric Switches
units in hemoglobin contains one heme molecule. The heme
groups are the oxygen-binding components of hemoglobin to Control Protein Activity
(see Figure 3-13). The binding of oxygen to the heme mole- Unlike oxygen, which causes graded allosteric changes in the
cule in one of the four hemoglobin subunits induces a local activity of hemoglobin, some other allosteric effectors act as

88 CHAPTER 3 • Protein Structure and Function

switches, turning the activity of many different proteins on (a) Calmodulin without calcium
or off by binding to them noncovalently. Two important al-
losteric switches that we will encounter many times through-
out this book, especially in the context of cell signa li ng
pathways (see Chapters 15 and 16), are Ca 1 + and GTP.

Ca2+ /Calmodulin-Mediated Switching The concentration of

Ca 2 ... free in the cytosol (not bound to molecules other than
water) is kept very low ( -1 o-7 M) by specialized mt:mbrane
transport proteins that continually pump excess CaH out of
the cytosol. However, as we learn in Chapter 11, the cyto-
solic Ca 1 + concentration can increase from 10- to 100-fold (b) Ca2•fcalmodulin bound to target peptide
when Ca1 + -permeable channels in the cell-surface membranes
open and allow extracellular Ca2+ to flow into the cell. This
rise in cytosolic Ca 1 + is sensed by specialized Ca 1 -binding
proteins, which alter cellular behavior by turning the activi-
ties of other proteins on or off. The importance of extracel-
lular Ca 2 + for cell activity was first documented by S. Ringer
in 1883, when he discovered that i~olated rat hearts sus-
pended in an NaCl solution made with "hard" (Ca 2 + -rich)
London tap water contracted beautifully, whereas they beat
poorly and stopped quickly if distilled water was used.
Many of the Ca 2 +-binding proteins bind Cal+ using the EF
hand/helix-loop-helix structural motif discussed earlier (see
Figure 3-9b). A well-studied EF hand protein, calmodulin, is
found in all eukaryotic cells and may exist as an individual FIGURE 3-31 Conformational changes induced by Ca2+ binding
monomeric protein or as a subunit of a multimeric protein. A to calmodulin. Calmodulin is a widely distributed cytosolic protein
that contains four Ca 2 -binding sites, one in each of its EF hands. Each
dumbbell-shaped molecule, calmodulin contains four Ca2+-
EF hand has a helix-loop-helix motif. At cytosolic Ca 2 concentrations
binding EF hands with K/s of-10- 6 M. The binding of Ca 1 +
above about 5 X 10 ·7 M, binding of Ca2+ to calmodulin changes the
to calmodulin causes a conformational change that permits
protein's conformation from the dumbbell-shaped, unbound form
Ca 1 .. /calmodulin to bind to conserved sequences in various (a) to one in which hydrophobic side chains become more exposed to
target proteins (Figure 3-31 ), thereby switching their activities solvent. The resulting Ca 2 -calmodulin can wrap around exposed
on or off. Calmodulin and similar EF hand proteins thus func- helices of various target proteins (b), thereby altering their activity.
tion as switch proteins, acting in concert with changes in Ca 1 -
levels to modulate the activity of other proteins.
proteins to which they bind and (2) an inactive ("off") form
Switching Mediated by Guanine-Nucleotide-Binding Pro- with bound GOP. The switch is turned on, that is, the con-
teins Another group of intracellular switch proteins consti- formation of the protein changes from inactive to active,
tutes the GTPase superfamily. As the name suggests, these when a GTP molecule replaces a bound GDP in the inactive
proteins are enzymes, GTPases, that can hydrolyze GTP conformation. The switch is turned off when the relatively
(guanosine triphosphate) to GDP (guanosine diphosphate). slow GTPase activity of the protein hydrolyzes bound GTP,
They include the monomeric Ras protein, whose structure is converting it to GOP and leading the conformation to change
shown in Figure 3-8 with b6und GDP shown in blue, and the to the inactive form. The amount of time any given GTPase
G" subunit of the trimeric G proteins, both discussed at switch remains in the active, GTP-bound form depends on
length in Chapters 15 and 16. Both Ras and Gu can bind to how long the GTP remains bound before it is converteJ to
the plasma membrane, function in cell signaling, and play a GDP. This time depends on the rate of the GTPase activity.
key role in cell proliferation and differentiation. Other mem- Thus the GTPase activity acts as a timer to control this
bers of the GTPase superfamily function in protein synthesis, switch. Cells contain a variety of proteins that can modulate
the transport of proteins between the nucleus and the cyto- the baseline (or intrinsic) rate of GTPase activity for any
plasm, the formation of coated vesicles and their fusion with given GTPase switch and so can control how long the switch
target membranes, and rearrangements of the actin cytoskel- remains on. Cells also have spel.ifil: proteins whose function
eton. The Hsp70 chaperone protein we encountered earlier is to regulate the conversion of inactive GTPases to active
is an example of an ATP/ADP switch, similar in many re- ones-that is, turn the switch on-by mediating the replace-
spects to a GTP/GOP switch. ment of bound GDP with a GTP. These are called GTP ex-
All the GTPase switch proteins exist in two forms, or change factors, or GEFs. We examine the role of various
conformations (Figure 3-32): (1) an active ("on") form with GTPase switch proteins in regulating intracellular signaling
bound GTP that can influence the activity of specific target and other processes in several later chapters.

3.4 Regulating Protein Function 89

FIGURE 3- 32 The GTPase switch. GTPases are enzymes Active ("on")
that bind to and hydrolyze GTP to GOP. When bound to GTP, the
GTPase protein adopts its active or "on" conformation and can
interact with target proteins to regulate their activities; when the
bound GTP is hydrolyzed to GOP by the intrinsic GTPase activity
of the protein, the GTPase with GOP bound assumes an inactive
or "off' conformation. The GTPase switch can be turned back on
when another protein, called a GEF (guanine nucleotide
exchange factor), mediates the replacement of the bound GOP
with a GTP molecule from the surrounding fluid. The binding
of the active form of the GTPases to its targets is a form of
noncovalent regulation. Various proteins can influence the rates
p
of GTP hydrolysis (i.e., inactivator proteins) and exchange of GOP
for GTP (GEFs).

Phosphorylation and Dephosphorylation conformational change that can significantly alter ligand
Covalently Regulate Protein Activity binding or other features of the protein, causing an increase
or decrease in its activity. In addition, several conserved pro-
In addition to exploiting the noncovalcnt regulators de- tein domains specifically bind to phosphorylated peptides.
scribed above, cells can use covalent modifications to regu- Thus phosphorylation can mediate the formation of protein
late the intrinsic activity of a protein. One of the most complexes that can generate or extinguish a wide variety of
common covalent mechanisms for regulating protein activity cellular activities, discussed in many subsequent chapters.
is phosphorylation, the reversible addition of phosphate Nearly 3 percent of all yeast proteins a~e protein kinascs
groups to hydroxyl groups on the side chains of serine, thre- or phosphatascs, indicating the importance of phosphoryla-
onine, or tyrosine residues. Phosphorylation is catalyzed by tion and dephosphorylation reactions even in simple cells.
enzymes called protein kinases, while the removal of phos- All classes of proteins-including structural proteins, scaf-
phates, known as dephosphorylation, is catalyzed by phos- folds, enzymes, membrane channels, and signaling mole-
phatases. The counteracting activities of kinases and cules-have members regulated by kinase/phosphatase
phosphatascs provide cells with a "switch" that can turn on modifications. Different protein kinases and phosphatases
or turn off the function of various proteins (Figure 3-33 ). arc specific for different target proteins and so can regulate
Sometimes phosphorylation sites are masked transiently by distinct cellular pathways, as discussed in later chapters.
the reversible covalent modification with the sugar N-acetyl- Some kinases have many targets and so a single kinase can
glucosamine as an additional means of regulation. Phos- serve to integrate the activities of many targets simultane-
phorylation changes a protein's charge and can lead to a ously. Frequently, the target of the kinase (and phosphatase)
is yet another kinase or phosphatase, creating a cascade ef-
fect. There are many examples of such kinase cascades,
which permit amplification of a signal and many levels of
Active fine-tuning control (see Chapters 15 and 16).

Ubiquitination and Deubiquitination Covalently

ADP H20
Regulate Protein Activity
Protein Both ubiquitin and ubiquitin-like proteins (of which there arc
Protein
kinase phosphatase
more than a dozen in humans) can be covalently linked to a
target protein in a regulated fashion, analogousl y to phos-
phorylation. Deubiquitinases can reverse ubiquitination,
ATP P, analogously to the action of phosphatases. These ubiquitin
modifications are structurally far more complex than rela-
tively simple phosphorylation, however, and so can mediate
.·
Inactive many distinct interactions between the ubiquitinated protein
FIGURE 3-33 Regulation of protein activity by phosphorylation and other cellular proteins. Ubiquinitatiun ~:an involve at-
and dephosphorylation. The cyclic phosphorylation and dephosphor- tachment of a single ubiquiti n to a protein (monoubiquitina-
ylation of a protein is a common cellular mechanism for regulating tion ), addition of multiple, single ubiquitin molecules to
protein activity. In this example, the target protein is active (top) when different sites on one target protein (multiubiquitination), or
phosphorylated and inactive (bottom) when dephosphorylated; some addition of a polymeric chain of ubiquitins to a protein (poly-
proteins have the opposite response to phosphorylation. ubiquitination ). An additional source of variation is that

90 CHAPTER 3 • Protem Structure and Function

·.
lnter-ubiquitin isopeptide

:::~:::::nl lsopepf"d, f~ys,~~:iquitins ~I~~JJ N~Hz ,

[Signaling (e.g. immunity)] bond Lys33-NH2 0 Lys48- NH
0 Lys 48- NH2 II Lys13-N~---)>
II ys13- NHr--0-C-Giy I
Lys-NH - O-C~ Giy76 's
'"::::>-Lys,-NH2

l ys 48-Giy 76
f~Lys,- NH:A2 ~LYS33-NH,
0 ~ys., - N~·---)>
[Proteasomal degradation] I Lys33- NH2 II _':::::>-Lys - NH 2
Lys - NH~ O-C-tGiy
r-------....
63
0 76
II _r Lys83
..
- NHzr
- -
Lys-NH- 0-C Gly, 6 j ~ys, - NH2
Lys33- NH
0 Lys 48-NH2
Lys33-Giy76 s,-NH2 II Lys83 NH/
IT-lymphocyte control] Lys - N 0-C-Giy
0 33 ...;;,;..:.a_ -
I II ys 48- NHY ~
Lys-NH- 0-C-Giy YSu-NH2 lys, -N ·-)>
__!!... _ - Lys33-3/N
H2
0 '-:::::::::-Lys - NH
~Lys,- Ntil- O-C~Giy
II '::::>-Lv•m~NH,'
76
Lys11 -Giy76
[Cell division) Lys33-NH, j
.!!!.----
~ LyS48- NH, t lsopept1de
Target protein Lys-NH· O-C-.Giy 76 ys63 -3 bond

FIGURE 3·34 Determination of polyubiquitin function by the The lysine used for the isopeptide bonds determines the function of
lysine used for inter-ubiquitin isopeptide bonds. Different ubiquitin the polyubiquitination. For example, polyubiquitins with Lys48:Giy76
ligases catalyze polyubiquitination of distinct target (substrate) isopeptide bonds direct the target to proteasomes for degradation.
proteins (colored ovals) using distinct lysine side chains of ubiquitin Those with Lys63, Lys33, and Lysll influence signaling, T-lymphocyte
molecules (purple) to generate the inter-ubiquitin isopeptide linkages control, and cell division, respectively. lsopeptide bonds involving
(blue) with Gly76 of the adjacent ubiquitin. Dotted blue arrows ubiquitin's Lys6, Lys27, and Lys29 and bonds using its N-terminal amino
represent additional ubiquitins in the chain that are not shown. group (not shown) can also be used to generate polyubiquitin chains.

different amino groups in the ubiquitin can be used to form an role. With this great structural diversity, it is not surprising
isopetide bond wjrh the carboxy terminal Gly76 in an adja- that cells usc ubiquitination and deubiquitination to control
cent ubiquitin in a polyubiquitin chain. All seven lysine resi- many different cellular functions.
dues in ubiquitin (Lys6, Lys l l, Lys27, Lys29, Lys33, Lys48, We have already seen how polyubiquitination via Lys48
and Lys63) and theN-terminal amino group of ubiquitin can residues is used to tag proteins for proteasomal degradation.
participate in inter-ubiquitin linkages. Different ubiquitin Ubiquitination unrelated to protein degradation also can
ligases exhibit specificity for both the target (substrate) to be control diverse cell functions, including repair of damaged
ubiquitinated a nd the lysine side chains on the ubiquitins that DNA, metabolism, messenger RNA synthesis (transcrip-
participate in the inter-ubi'quitin isopeptide linkages (Lys63 tion), defense against pathogens, cell division/cell cycle pro-
or Lys 48, etc.) (figure 3-34). These multiple forms of ubiq- gression, cell signaling path ways, trafficking of proteins
uitination resu lt in the generation of a wide variety of recog- within a cell, and programmed cell death (apoptosis). The
nition surfaces that can participate in many protein-protein lysine used to form the inter-ubiquitin isopeptide bonds can
interactions with the hundreds of proteins (> 200 in humans) vary depending on the cellular system that is regulated (see
that contain more than a dozen distinct ubiquitin-binding do- Figure 3-34). For example, polyubiquitination with Lys63
mains (UBD). In addition, any given polyubiquitin chain has linkages is used in many cellular identification and signaling
the potential to bind simultaneously more than one UBD- systems, such as recognition of the presence of intracellular
containing protein, leading to the formation of uhiquitination- viral RNA and the consequent induction uf a protective im-
dependent multiprotein complexes. Some deubiquitinases can mune response. Lysll-linked polyubiquitin chaim regulate
remove an intact polyubiquitin chain from a modified protein cell division. Lys33-linked chains help suppress the activity
("anchored" chain) and th us generate a polyubiquitin chain of receptors on specialized white blood cells, called T lym-
not covalently linked to another protein ("unanchored" phocytes (see Chapter 23 ), and so control the activity and
chain). Even these unanchored chains may serve a regulatory function of the lymphocytes that bear these receptors.

3.4 Regulat1ng Protein Function 91

Proteolytic Cleavage Irreversibly Activates neously in different parts of a cell. We describe the mecha-
or Inactivates Some Proteins nisms that cells use to direct various proteins to different
compartments in Chapters 13 and 14.
Unlike phosphorylation and ubiquitination, which are revers-
ible, the activation or inactivation of protein function by pro-
teolytic cleavage is an irreversible mechanism for regulating
protein activity. For example, many polypeptide hormones, KEY CONCEPTS of Section 3.4
such as insulin, are synthesized as longer precursors, and prior
Regulating Protein Function
to secretion from cells some of their peptitle:: bonds must be
hydrolyzed for them to fold properly. In some cases, a single • Proteins may be regulated at the level of protein synthesis,
long precursor prohormone polypeptide can be cleaved into protein degradation, or the intrinsic activity of proteins
several distinct active hormones. To prevent the pancreatic through noncovalent or covalent interactions.
serine proteases from inappropriately digesting proteins be- • The life span of intracellular proteins is largely determined
fore they reach the small intestines, they are synthesized as by their susceptibility to proteolytic degradation.
zymogens, inactive precursor enzymes. Cleavage of a peptide
• Many proteins are marked for destruction with a polyu-
bond near the N-terminus of trypsinogen (the zymogen of
biquitin tag by ubiquitin ligases and then degraded within
trypsin) by a highly specific protease in the small intestine gen-
proteasomes, large cylindrical co~plexes with multiple pro-
erates a new N-terminal residue (IIe-16), whose amino group
tease active sites in their interior chambers (see Figure 3-29).
can form an ionic bond with the carboxylic acid side chain of
an internal aspartic acid. This causes a conformation change • Ubiquitination of proteins is reversible due to the activity
that opens the substrate-binding site, activating the enzyme. of deubiquitinating enzymes.
The active trypsin can then activate trypsinogen, chymotryp- • In allostery, the noncovalent binding of one ligand mole-
sinogen, anti other zymogens. Similar but more elaborate pro- cule, the allosteric effector, induces a conf01;mational change
tease cascades (one protease activating inactive precursors of that alters a protein's activity or affinity for other ligands. The
others) that can amplify an initial signal play important roles allosteric effector can he identical in structure to or different
in several systems, such as the blood-clotting cascade and the from the other ligands, whose binding it affects. The allosteric
complement system (see Chapter 23 ). The importance of care- effector can be an activator or an inhibitor.
fully regulating such systems is clear-inappropriate clotting
• In multimeric proteins, such as hemoglobin, that bind mul-
could fatally clog the circulatory system, while insufficient
tiple identical ligand molecules (e.g., oxygen), the binding of
clotting could lead to uncontrolled bleeding.
one ligand molecule may increase or decrease the binding
An unusual and rare type of proteolytic processing,
affinity for subsequent ligand molecules. This type of allo-
termed protein self-splicing, takes place in bacteria and some
stery is known as cooperativity (see figure 3-30).
eukaryotes. This process is analogous to editing film: an in-
ternal segment 9f a polypeptide is removed and the ends of • Several allosteric mechanisms act as switches, turning pro-
the polypeptide are rejoined (ligated). Unlike other forms of tein activity on and off in a reversible fashion.
proteolytic processing, protein self-splicing is an autocata- • Two classes of intracellular switch proteins regulate a va-
lytic process, which proceeds by itself without the participa- riety of cellular processes: (1) Ca 2 -binding proteins (e.g.,
tion of other enzymes. The excised peptide appears to calmodulin) and (2) members of the GTPase superfamily
eliminate itself from the protein by a mechanism similar to (e.g., Ras), which cycle between active GTP-bound and inac-
that used in the processing of some RNA molecules (see tive GDP-bound forms (see Figure 3-32).
Chapter 8). In vertebrate cells, the processing of some pro-
• The phosphorylation and dephosphorylation of hydroxyl
teins includes self-cleavage, but the subsequent ligation step
groups on serine, threonine, or tyrosine residue side chains
is absent. One such protein is Hedgehog, a membrane-bound
by protein kinases and phosphatases provide reversible on/
signaling molecule that is critical to a number of develop-
off regulation of numerous proteins (see Figure 3-33).
mental processes (see Chapter 16).
• Variations in the nature of the covalent attachment of
ubiquitin to proteins (mono-, multi-, and polyubiquitination
Higher-Order Regulation Includes Control involving a variety of linkages between the ubiquitin mono-
of Protein Location and Concentration mers) are involved in a wide variety of cellular functions
other than proteasome-mediated degradation, such as changes
All the regulatory mechanisms heretofore described affect a
in the location or activity of proteins (see Figure 3-34).
protein locally at its site of auiun, turning its activity on or
off. Normal functioning of a cell, however, also requires the • Many types of covalent and noncovalent regulation are
segregation of proteins to particular compartments such as reversible, but some forms of regulation, such as proteolytic
the mitochondria, nucleus, and lysosomes. In regard to en- cleavage, are irreversible.
zymes, compartmentation not only provides an opportunity • Higher-order regulation includes compartmentation of
for controlling the delivery of substrate or the exit of product proteins and control of protein concentration.
but also permits competing reactions to take place simulta-

92 CHAPTER 3 • Protein Structure and Function

3.5 Purifying, Detecting, and the size of an American football) that holds the samples in
tubes at rates as high as 2500 revolutions per second!
Characterizing Proteins Centrifugation is used for two basic purposes: ( 1) as a
A protein often must be purified before its srrucrure and the preparative technique to separate one type of material from
mechanism of its action can be studied in detail. However, others and (2) as an analytical technique to measure physical
because proteins vary in size, shape, oligomerization state, properties (e.g., molecular weight, density, shape, and equi-
charge, and water solubility, no single method can be used to librium binding constants) of macromolecules. The sedimen-
isolate all proteins. To isolate one particular protein from the tation constant, s, of a protein is a measure of its sedimentation
estimated I 0,000 different proteins in a particular type of cell rate. The sedimemation constant is commonly expressed in
is a daunting task that requires methods both for separating svedbergs (5), where a typical, large protein complex is
proteins and for detecting the presence of specific proteins. about 3-55, a proreasome is 265, and a eukaryotic ribosome
Any molecule, whether protein, carbohydrate, or nucleic is 80S.
acid, can be separated, or resolved, from other molecules on
the basis of their differences in one or more physical or chemi- Differential Centrifugation The most common initial step in
cal characteristics. The larger and more numerous the differ- protein purification from cells or tissues is the separation of
ences between two proteins, the easier and more efficient their water-soluble proteins from insoluble cellular material by
separation. The three most widely used characteristics for sep- differential centrifugation. A starting mixture, commonly a
arating proteins are size, defined as either length or mass; net cell homogenate (mechanically broken cells), is poured into
electrical charge; and binding affinity for specific ligands. In a tube and spun at a rotor speed and for a period of time that
this section, we briefly outline several important techniques for forces cell organelles such as nuclei and large unbroken cells
separating proteins; these separation techniques arc also useful or large cell fragments to collect as a pellet at the bottom; the
for the separation of nucleic acids and other biomolecules. soluble proteins remain in the supernatant (Figure 3-35a).
(Specialized methods for removing membrane proteins from The supernatant fraction then is poured off, and either it or
membranes are described in Chapter 10 after the unique prop- the pellet can be subjected to other purification methods to
erties of these proteins are discussed.) We then consider the usc separate the many different proteins that they .contain.
of radioactive compounds for tracking biological activity. Fi-
nally, we consider several techniques for characterizing a pro- Rate-Zonal Centrifugation On the basis of difference~ in
tein's mass, sequence, and three-dimensional structure. their masses, water-soluble proteins can be separated by cen-
trifugation through a solution of increasing density called a
density gradient. A concentrated sucrose sol uti on is com-
monly used to form density gradients. When a protein mix-
Centrifugation Can Separate Particles and
ture is layered on top of a sucrose gradient in a rube and
Molecules That Differ in Mass or Density subjected to centrifugation, each protein in the mixture mi-
The first step in a typical protein purification scheme is cen- grates down the tube at a rate controlled by the factors that
trifugation. The principle behind centrifugation is that two affect the sedimentation constant. All the proteins start from
particles in suspension (cells, cell fragments, organelles, or a thin zone at the top of the tube and separate into bands
molecules) with different masses or densities will settle to the (actually, disks) of proteins of different masses. In this sepa-
bottom of a tube at different rates. Remember, mass is the ration technique, called rate-zonal centrifugation, samples
weight of a sample (measured in daltons or molecular weight arc centrifuged just long enough to separate the molecules of
units), whereas density is the ratio of its mass to volume interest into discrete bands (Figure 3-3Sb). If a sample is cen-
(often expressed as grams/liter because of the methods used trifuged for too short a time, the different protein molecule~
to measure density). Proteins vary greatly in mass but not in will not separate sufficiently. If a sample is centrifuged much
density. Unless a protein hao an attached lipid or carbohy- longer than necessary, all the proteins will end up in a pellet
drate, its density will nor vary by more than 15 percent from at the bottom of the tube.
1.37 g/cm\ the average protein density. Heavier or more Although the sedimentation rate is strongly influenced by
dense molecules settle, or sediment, more quickly than particle mass, rate-zonal centrifugation is seldom effective in
lighter or less dense molecules. determining precise molecular weights because variations in
A centrifuge speeds sedimentation by subjecting particles shape also affect sedimentation rate. The exact effects of
in suspension to centrifugal forces as great as 1 million rimes shape are hard to assess, especially for proteins or other mol-
the force of gravity, g, which can sediment particles as small ecules, such as single-stranded nucleic acid molecules, that
as 10 kDa. Modern ultracentrifuges achieve these forces by can assume' many complex shapes. Nevertheless, rate-zonal
reaching speeds ot 150,000 revolutions per minute (rpm) or centrifugation has proved to be a practical method for separat-
greater. However, small particles with masses of 5 kDa or ing many different types of polymers and particles. A second
less will not sediment uniformly even at such remarkably density-gradient technique, called equilibrium density-gradient
high rotation rates. The extraordinary technical achieve- centrifugation, is used mainly to separate DNA, lipoproteins
'llents of modern ultracentrifuges can be appreciated by con- that carry lipids through the circulatory system, or organelles
sidering that they can rotate a several-pound rotor (about (see Figure 9-35 ).

3.5 Purify1ng, Detecting, and Characterizing Proteins 93

EXPERIMENTAL FIGURE 3-35 (a) Differential centrifugation (b) Rate-zonal centrifugation
Centrifugation techniques separate
particles that differ in mass or
D Sample is poured into tube D Sample is layered on top of density gradient

density. (a) In differential centrifuga-

tion, a cell homogenate or other Larger particle -----~
mixture is spun long enough to Smaller particle---.c1;!~
sediment the larger particles (e.g., cell
organelles, cells), which collect as
a pellet at the bottom of the tube
(step fl ). The smaller particles (e.g., Low density
soluble proteins, nucleic acids) remain (low sucrose
concentration)
in the liquid supernatant, which can be
transferred to another tube (step 10). Centrifuge
(b) In rate-zonal centrifugation, Particles settle High density
according to (high sucrose
a mixture is spun (step 0 ) just long concentration)
mass
enough to separate molecules that
differ in mass but may be similar in Centrifuge
shape and density (e.g., globular Particles settle
Centrifugal ~
according to
proteins, RNA molecules) into discrete force mass
zones within a density gradient
common ly formed by a concentrated Cen trifugal force ~
sucrose solution. Fractions are
removed from the bottom of the tube
and subjected to testing (assayed).

Stop centrifuge
Stop centrifuge Collect fractions
Decant liquid and do assay
into container

u
~

Supernatant Pellet

Decreasing mass of particles

Electrophoresis Separates Molecules on the congealed gelati n found in desserts) rather than in a liquid
the Basis of Their Charge-to-Mass Ratio solution. Electrophoretic separation of proteins is most com-
monly performed in polyacrylamide gels. When a m ixture of
Electrophoresis is a technique for separating molecules in a proteins is placed in a gel and an electric current is applied,
mixture under the influence of an applied electric field and is smaller proteins migrate faster through the gel than do larger
one of the most frequen tl y used techniques to study proteins proteins because the gel acts as a sieve, with smaller species
and nucleic acids. Dissolved molecules in an electric field able to maneuver more rapidly through the pores in the gel
move, or migrate, at a speed determined by their charge-to- than larger species. The shape of a molecule can also influence
mass (charge:mass) ratio and the physical properties of the me- its rate of migration (long asymmetric molecules migrate more
dium through which they migrate. For example, if two slowly than spherical ones of the same mass).
molecules have the same mass and shape, the one with the Gels are cast into flat, relatively th in slabs between a pair
greater net charge will move faster toward an electrode of of glass plates by polymerizing a solution of acry lam ide
the opposite polarity. monomers into polyacrylamide chains and simultaneously
cross-linking the chains into a semisolid matrix. The pore
SDS-Polyacrylamide Gel Electrophoresis Because many pro- size of a gel can be varied by adjusting the concentrations of
teins or nucleic acids that differ in size and shape have nearly polyacrylamide and the cross-linking reagent. T he rate ar
identical charge:mass ratios, electrophoresis of these macro- wh ich a protein moves through a gel is influenced by the
molecules in solution results in little or no separation of mol- gel's pore size and the strength of the electric field. By suit-
ecules of different lengths. H owever, successful separation of able ad justment of these parameters, proteins of widely vary-
proteins and nucleic acids can be accomplished by electropho- ing sizes can be resolved (separated from one another) by
resis in various gels (semisolid suspensions in water similar to polyacrylamide gel electrophoresis (PAGE).

94 CHAPTER 3 • Protein Structure and Function

@ TECHNIQUE ANIMATION: SDS Gel Electrophoresis

EXPERIMENTAL FIGURE 3-36 50S-polyacrylam ide gel (a) Denature sample with
electr ophoresis (SDS-PAGE) sepa rates proteins primarily on the
basis of their masses- (a) Initial treatment with SDS, a negatively
charged detergent, dissociates multimeric proteins and denatures all
D
1.
sodium dodecylsulfate

the polypeptide chains (step 0 ). During electrophoresis, the 50S-

~~ SDS-coated
protein complexes migrate through the polyacrylamide gel (step f)). .. \:::t; ~- _. i- proteins
f"'C' ..

l
Small complexes are able to move through the pores faster than larger
ones. Thus the protein~ ~epa rate Into bands according to their sizes as ~ ~
they migrate. The separated protein bands are visualized by staining
with a dye (step ID). (b) Example of SDS-PAGE separation of all the Load samples
fJ
Place mixture of proteins on gel;
apply electric field
proteins in a whole-cell lysate (detergent solubilized cells): {left) the here ~
many separate stained proteins, appearing almost as a continuum;
(right) a protein purified from the lysate by a single step of antibody-
affinity chromatography. The proteins were visualized by staining with Cross-linked
Wells
a silver-based dye. [Part (b) modified from B. Uu and M. Krieger, 2002, J. Bioi. +---- polyacrylamide

1o;"";::~f
Chem. 277(37):34125-34135.]
Partially {
separ.ated
protems m;g,Uoo

In the most powerful technique for resolving protein mix-

tures, proteins are exposed to the ionic detergent SDS (sodium L----------------" e
dodecylsulfate) before and during gel electrophoresis (Figure
3-36). SDS denatures proteins, in part because it binds to hy-
drophobic side chains, destabilizing the hydrophobic interac-
tions in the core of a protein that contribute to its stable
II lsto;" to.;'"'""
separated bands
(b)

~
1!1
(lJ
e
0.
't:J
CZJ
.s
1!1

conformation. (SDS treatment is usually combined with heat- -"=:

Decreasing :::: ·;:
ing, often in the presence of reducing agents that break disul- size
kDa rJ~
fide bonds.) Under these conditions, multimeric proteins
dissociate into their subunits, and typically the amount of SDS
that binds to the protein is proportional to the length of the
polypeptide chain and relatively independent of the sequence.
c§
~
-- 200
116
97
66
Two proteins of similar size will bind the same absolute quan-
tity of SDS, whereas a protein twice that size will bind twice the 45
amount of SDS. Denaturation of a complex protein mixture 31
with SDS in combination with heat usually forces each poly-
peptide chain into an extended conformation and imparts on
each of the proteins in the mixture a constant mass/charge
ratio because the dodecylsulfate is negatively charged. As the relatively large differences in mass but cannot readily resolve
SDS-bound proteins move through the polyacrylamide gel, proteins having similar masses (e.g., a 41-kDa protein versus
they are separated according to size by the sieving action of a 42-kDa protein). To separate proteins of similar masses, an-
the gel. SDS treatment thus eliminates the effect of differences other physical characteristic must be exploited. Most com-
in shape in native structures; therefore, chain length, which monly, this characteristic is electric charge, which is determined
corresponds to mass, is the principal determinant of the mi- by the pH and the relative number of the protein's positively
gration rate of proteins in SDS-fJolyacrylamide electrophore- and negatively charged groups, which is in turn dependent on
sis (SDS-PAGE). Even chains that differ in molecular weight the pK;s of the ionizable groups (see Chapter 2). Two unre-
by less than 10 percent can be resolved by this technique. lated proteins having similar masses are unlikely to have iden-
Moreover, the molecular weight of a protein can be estimated tical net charges because their sequences, and thus the number
by comparing the distance that it migrates through a gel with of acidic and basic residues, are different.
the distances that proteins of known molecular weight mi- In two-dimensional electrophoresis, proteins are sepa-
grate (there is roughly a linear relationship benveen migration rated sequentially, first by their charges and then by their
distance and the log of the molecular weight). Proteins within masses (Figure 3-37a) . In the first step, a cell or tissue extract
the gels can be extracted for further analysis (e.g., identifica- is fully denatured by high concentrations (8 M) of urea (and
tion by methods described below). sometimes SDS) and then layered on a gel strip that contains
urea, which removes any bound SDS, and a continuous pH
Two-Di mensional Gel Electro phoresis Electrophoresis of all gradient. The pH gradient is formed by ampholytes, a mix-
cellular proteins by SDS-PAGE can separate proteins having ture of polyanionic and polycationic small molecules, that

3.5 Purifying, Detecting, and Characterizing Proteins 95

 (a) Protein pH 4.0 (b) lsoelectric focu sing (D ) ~
mixture
--
-
•
Separate lsoelectric
in first D
1
focusing (IEF}
dimension I') 66 II
by charge 0
~

X
"'
·;;;
!!
....
~
0
pH 10.0 .t:
Cl
·a;
..
~ 43
.. ..
Q.
0
'tj

l
Apply first gel Q)
El
to top of second ~
:::1
u .• Qj
(/J

--..
Cll 30 c(/J
0
~
.:\.
4IL •
•
pH 4.0 I I I I
••
I I I I pH 10.0
16
• • -· ~
Separate
in second
II
•
•
SDS
electrophoresis
•

-
...
, ... -
dimension
by size
•
•
1 4.2 5.9
pi
7.4

EXPERIMENTAL FIGURE 3-37 Two-dimensional gel electro- into spots by mass (step IDJ. (b) In this two-dimension'al gel of a protein
phoresis separates prot eins on the basis of charge and mass. (a} In extract from cultured cells, each spot represents a single polypeptide.
this technique, proteins are first separated into bands on the basis of their Polypeptides can be detected by dyes, as here, or by other techniques
charges by isoelectric focusing (step 0 }. The resulting gel strip is applied such as autoradiography. Each polypeptide is characterized by its
to an 50S-polyacrylamide gel (step FJ), and the proteins are separated isoelectric point (pi} and molecular weight. [Part (b) courtesy of J. Celis.]

arc cast into the gel. When an electric field is applied to the entiated cells or in normal and cancer cells because as many as
gel, ampholytes will migrate, so that ampholytes with an ex- 1000 proteins can be resolved as individual spots simultane- /
cess of negative charges will migrate toward the anode, ously. Unfortunately, memhrane proteins separate relatively
where they establish an acidic pH (many protons), while am- poorly using this technique. Sophisticated methods have been
pholytes with a(l excess positive charge will migrate toward developed to permit the comparison of complex patterns of
the cathode, where they establish an alkaline pH. The care- proteins in two-dimensional gels from related, but distinct,
ful choice of the mixture of ampholytes and the preparation samples (e.g., tissue from a normal versus a mutant individual)
of the gel allows the construction of stable pH gradients to permit identification of differences in the types or amounts of
anywhere from pH 3 to pH 10. A charged protein placed proteins in the samples (see Section 3.6, on proteomics, below).
onto one end of such a gel will migrate through the gradient Sophisticated mass spectrometry methods, described below, are
until it reaches its isoelcctric point (pi), the pH at which the often used in place of two-dimensional gel electrophoresis to
net charge of the protein is zero. With no net charge, the identify the protein components of a complex sample.
protein will not migrate under the influence of the electric
field. This technique, called isoelectric focusing (T EF), can
Liquid Chromatography Resolves Proteins
resolve proteins that differ by only one charge unit. Proteins
that have been separated on an IEF gel can then be separated by Mass, Charge, or Binding Affinity
in a second dimension on the hasis of their molecular A third common technique for separating mixtures of pro-
weights. To accomplish this separation, the IEF gel strip is teins or fragments of proteins, as well as other molecules, is
placed lengthwise on one outside edge of a sheet-like (two- based on the principle that molecules dissolved in a solution
dimensional, or slab) polyacrylamide gel, this time saturated can differentially interact (bind and dissociate) with a par-
with SDS to confer on each separated protein a more or less ticular solid surface, depending on the physical and chemical
co nsr;:~nr m;:~ss:charge ratio. When an electric field is im- properties of the molecule and the surface. If the solution is
posed, the proteins will migrate from the IEF gel into the allowed to flow across the surface, then molecules that inter-
SDS slab gel and then separate according to their masses. act frequently with the surface will spend more time bound to
The sequential resolution of proteins by charge and mass the surface and thus flow past the surface more slowly than
can achieve excellent separation of cellular proteins (Figure molecules that interact infrequently with it. In this technique,
3-37b). For example, two-dimensional gels have been very use- called liquid chromatography (LC), the sample is placed on
ful in comparing the proteomes in undifferentiated and differ- top of a tightly packed column of spherical beads held within

96 CHAPTER 3 • Protein Structure and Function

a glass or plastic cylinder. The sample then flows down the tography. In this technique, ligand or other molecules that
column, usually driven by gravitational or hydrostatic forces bind to the protein of interest are covalently attached to the
alone or with the assistance of a pump, and small aliquots of beads used to form the column. Ligands can be enzyme
fluid flowing out of the column, called fractions, are collected substrates, inhibitors or their analogs, or other small mol-
sequentially for subsequent analysis for the presence of the ecules that bind to specific proteins. In a widely used form
proteins of interest. The nature of the beads in the column of this technique-antibody-affinity, or immunoaffimty,
determines whether the separation of proteins depends on chromatography-the attached molecule is an antibody spe-
differences in mass, charge, or binding affinity. cific for the desired protein (Figure 3-38c). (We discuss anti-
bodies as tools to study proteins next; see abo Chapter 23,
Gel Filtration Chromatography Proteins that differ in mass which describes how antibodies are made.)
can be separated on a column composed of porous beads An affinity column in principle will retain only those
made from polyacrylamide, dextran (a bacterial polysaccha- proteins that bind the molecule attached to the beads; the
ride), or agarose (a seaweed derivative)-a technique called remaining proteins, regardless of their charges or masses,
gel filtration chromatography. Although proteins flow will pass through the column because they do not bind.
around the spherical beads in gel filtration chromatography, However, if a retained protein is in turn bound to other mol-
they spend some time within the large depressions that cover ecules, forming a complex, then the entire complex is re-
a bead's surface. Because smaller proteins can penetrate into tained on the column. The proteins bound to the affinity
these depressions more readily than larger proteins can, they column are then eluted by adding an excess of a soluble form
travel through a gel filtration column more slowly than larger of the ligand, by exposure of bound materials to detergents,
proteins (Figure 3-38a). (In contrast, proteins migrate or by changing the salt concentration or pH such that the
through the pores in an electrophoretic gel; thus smaller pro- binding to the molecule on the column is disrupted. The abil-
teins move faster than larger ones.) The total volume of liquid ity of this technique to separate particular proteins depends
required to elute (or separate and remove) a protein from a on the selection of appropriate binding partners that bind
gel filtration column depends on its mass: the smaller the more tightly to the protein of interest than to other proteins.
mass, the more time it is trapped on the beads, the greater the
elution volume. By use of proteins of known mass as stan-
Highly Specific Enzyme and Antibody Assays
dards to calibrate the column, the elution volume can be used
to estimate the mass of a protein in a mixture. A protein's Can Detect Individual Proteins
shape as well as its mass can influence the elution volume. The purification of a protein, or any other molecule, requires
a specific assay that can detect the presence of the molecule
!on-Exchange Chromatography In ion-exchange chroma- of interest as it is separated from other molecules (e.g., in
tography, a second type of liquid chromatography, proteins column or density-gradient fraction s or gel bands or spots).
are separated on the basis of differences in their charges. An assay capitalizes on some highly distinctive characteristic
This technique makes use of specially modified beads whose of a protein: the ability to bind a particular ligand, to cata-
surfaces are covered by amino groups or carboxyl groups lyze a particular reaction, or to be recognized by a specific
and thus carry either a positive charge (NH3 -) or a negative antibody. An assay must also be simple and fast to minimize
charge (COO ) a"t neutral pH. errors and the possibility that the protein of interest becomes
The proteins in a mixture carry various net charges at denatured or degraded while the assay is performed. The
any given pH. When a solution of a protein mixture flows goal of any purification scheme is to isolate sufficient
through a column of positively charged beads, only proteins amounts of a given protein for study; thus a useful assay
with a net negative charge (acidic proteins) adhere to the must also be sensitive enough that only a small proportion of
beads; neutral and positively charged (basic) proteins flow the available material is consumed by it. Many common pro-
unimpeded through the column (Figure 3-38b). The acidic tein assays require just 10 9 to 10- 12 g of material.
proteins are then eluted selectively from the column by pass-
ing a solution of increasing concentrations of salt (a salt gra- Chromogenic and light-Emitting Enzyme Reactions Many
dient) through the column. At low salt concentrations, assays are tailored to detect some functional aspect of a pro-
protein molecules and beads arc attracted by their opposite tein. For example, enzymatic activity assays are based on the
charges. At higher salt concentrations, negative salt ions ability to detect the loss of substrate or the formation of prod-
bind to the positively charged beads, displacing the nega- uct. Some enzymatic assays utilize chromogenic substrates,
tively charged proteins. In a gradient of increasing salt con- which change color in the course of the reaction. (Some sub-
centration, weakly bound proteins, those with relatively low strates are naturally chromogenic; if they ~re not, they can be
charge, are eluted first and highly charged proteins are eluted linked to a chromogenic molecule.) Because of the specificity
last. Similarly, a negatively charged column can be used to of an enzyme for its substrate, only samples that contain the
retain and fractionate basic (positively charged) proteins. enzyme will change color in the presence of a chromogenic
substrate; the rate of the reaction provides a measure of the
Affinity Chromatography The ability of proteins to bind quantity of enzyme present. Enzymes that catalyze chromo-
specifically to other molecules is the basis of affinity chroma- genic reactions can also be fused or chemically linked to an

3.5 Purifying, Detecting, and Characterizing Proteins 97

 (a) Gel filtration chromatography (c) Antibody-affinity chromatography

~......- Large protein

r:""" Small protein

e Protein ~
Layer
sample
on
column
1 Add buffer
to wash
proteins
:
I
Collect
fractions
recognized
by antibody
~~
Elute
___.
with
pH3
through

t~
buffer
column

y
il ~fl
)
.. ~Eluted
Polymer gel bead 3 2 1 fractions Antibody
2

Eluted
(b) ion-exchange chromatography fractions

Anions
retained Elute negatively
by beads
charged protein
Layer with salt solution
Cations
sample
elute
(NaCI) .a fi
on
column out

Eluted
fractions

~ /11\
~Jf~·;&l~l
Positively
charged
gel bead 3 2 fractions
~·
4 3
..,.
2

lX: ERIMENTAL FIGURE 3-38 Three commonly used liquid

0
proteins having the opposite charge bind to the beads more or less
chromatographic techniques separate proteins on the basis of mass, tightly, depending on their structures. Bound proteins-in this case,
charge, or affinity for a specific binding partner. (a) Gel filtration negatively charged-are subsequently eluted by passing a salt gradient
chromatography separates proteins that differ in size. A mixture of (usually of NaCI or KCI) through the column. As the ions bind to the
proteins is carefully layered on the top of a cyl inder packed with porous beads, they displace the proteins (more tightly bound proteins require
beads. Smaller proteins travel through the column more slowly than higher salt concentration in order to be released). (c) In antibody-affinity
larger proteins. Thus different proteins emerging in the eluate flowing chromatography, a mixture of proteins is passed through a column
out of the bottom of the column at different times (different elution packed with beads to which a specific antibody is covalently attached.
volumes) can be collected in separate tubes, called fractions. Only protein with high affinity for the antibody is retained by the
(b) ion-exchange chromatography separates proteins that differ in net column; all the nonbinding proteins flow through. After the column is
charge in columns packed with special beads that carry either a positive washed, the bound protein is eluted with an acidic solution or some
charge (shown here) or a negative charge. Proteins having the same net other solution that disrupts the antigen-antibody complexes; the
charge as the beads are repelled and flow through the column, whereas released protein then flows out of the column and is collected.

antibody and used to "report" the presence or location of an recognize a protein antigen of interest can be genera ted and
antigen to which the antibody binds (see below). used to detect the presence of the protein, either in a complex
mixture of other proteins (finding a needle in a haystack, as
Antibody Assays As noted earlier, antibodies have the dis- it were) or in a partially purified preparation of a particular
tinctive characteristic of binding tightly and specifically to an- protein. An antibody molecule will generally only bind tightly
ngens. As a consequence, preparations of antibod ies that to a small parr of a target molecule (antigen) that exhibits

98 CHAPTER 3 • Protein Structure and Function

 molecular complementarity with the antibody. This antibody- protein, with a lower limit of detection of -4-10 ng. Silver
binding region of the target is called the antibody's cognate staining or fluorescence staining are more sensitive (lower
epitope, or simply epitope. Thus the presence of the antigen limit of -1 ng). Coomassie and other stains can also be used
that contains an epitope can be visualized by labeling the an- to visualize proteins after transfer to membranes; however,
tibody with an enzyme, a fluorescent molecule, or radioactive the most common method to visualize proteins in these mem-
isotopes. For example, luciferase, an enzyme present in fire- branes is immunoblotting, also called Western blotting.
flies and some bacteria, can be linked to an antibody. In the Immunoblotting combines the resolving power of gel
presence of ATP and the substrate luciferin, luciferase cata- electrophoresis and the specificity of antibodies. This multi-
lyzes a light-emitting reaction. In either ca~e, after the anti- step procedure is commonly used co ~t:parate proteins and
body binds to the protein of interest (the antigen) and unbound then identify a specific protein of interest. As shown in
antibody is washed away, substrates of the linked enzyme are figure 3-39, two different antibodies are used in this method,
added and the appearance of color or emitted light is moni- one that is specific for the desired protein and a second that
tored. The intensity is proportional to the amount of enzyme- binds to the first and is linked to an enzyme or other mole-
linked antibody, and thus antigen, in the sample. A variation cule that permits detection of the first antibody (and thus the
of this technique, particularly useful in detecting specific pro- protein of interest to which it binds). Enzymes to which the
teins within living cells, makes use of green fluorescent protein second antibody is attached can either generate a visible col-
(GFP), a naturally fluorescent protein found in jellyfish (see ored product or, by a process called chemiluminescence,
Figure 9-15). Alternatively, after the first antibody that is not produce light that can readily be recorded by film or a sensi-
otherwise labeled binds to its target protein, a second, labeled tive detector. The two different antibodies, sometimes called
antibody that can recognize the first antibody is used to bind a "sandwich," are used to amplify the signals and improve
to the complex of the first antibody and its target. This com- sensitivtty. If an antibody is not available but the gene en-
bination of two antibodies permits very high sensitivity in the coding the protein is available and can be used to express the
detection of a target protein because the labeled second anti- protein, recombinant DNA methods (see Chapter 5 ) can in-
body is often a mixture of antibodies that bind to multiple corporate a small peptide epitope (epitope tagging) into the
sites on the first antibody. It is important to remember that an normal sequence of the protein that can be detected by a
antibody binds to or recognizes its cognate epitope on a target commercially available antibody to that epitope.
antigen. If that epitope is altered, for example by partial un-
folding or post-translational modifications, or is blocked Radioisotopes Are Indispensable Tools
when the antigen protein is bound to some other molecule, the
for Detecting Biological Molecules
ability of the antibody to bind may be reduced or completely
lost. Thus the absent of antibody binding does not necessarily A sensitive method for tracking a protein or other biological
mean the antigen is not present in a sample, only that the epi- molecule is by detecting the radioactivity emitted from ra-
tope is not present or accessible for binding. dioisotopes introduced into the molecule. At least one atom
To generate the antibodies, the intact protein or a frag- in a radiolabeled molecule is present in a radioactive form,
ment of the protein is injected into an animal (usually a rabbit, called a radioisotope.
·.
mouse, or goat). Sometimes a short synthetic peptide of
Radioisotopes Useful in Biological Research Hundreds of
10-15 residues based on the sequence of the protein is used as
biological molecules (e.g., amino acids, nucleosides, and nu-
the antigen to induce antibody formation. A synthetic peptide,
merous metabolic intermediates) labeled with various radio-
when coupled to a large protein carrier, can induce an animal
isotopes are commercially available. These preparations vary
to produce antibodies that bind to that part (the epirope) of the
considerably in their specific activity, which is the amount of
full-size, natural protein. Biosynthetically or chemically artach-
radioactivity per unit of material, measured in disintegra-
ing the epirope to an unrelated protein is called epitope tagging.
tions per minute (dpm) per millimole. The specific activity of
As we'll see throughout thio book, antibodies generated using
a labeled compound depends on the probability of decay of
either peptide epiropes or intact proteins are extremely versatile
the radioisotope, determined by its half-life, which is the
reagents for isolating, detecting, and characterizing proteins.
time required for half the atoms to undergo radioactive
decay. In general, the shorter the half-life of a radioisotope,
Detecting Proteins in Gels Proteins embedded within a one- the higher its specific activity (Table 3-1 ).
or two-dimensional gel usually are not visible. The two gen- The specific activity of a labeled compound must be high
eral approaches for detecting proteins in gels are either to enough that sufficient radioactivity is incorporated into mol-
label or stain the proteins while they are still within the gel or ecules to be accurately detected. For example, methionine
to elecrrnphoretica lly transfer the proteins to a membrane auJ t:ysteine labeled with sulfur-35 e5 ~) are widely used to
made of nitrocellulose or polyvinylidene difluoride and then biosynthetically label cellular proteins because preparations
detect them. Proteins within gels are usually stained with an of these amino acids with high specific activities (>10 15
organic dye or a silver-based stain, both detected with normal dpm/mmol) are available. Likewise, commercial prepara-
visible light, or with a fluorescent dye that requires special- tions of 3 H-Iabeled nucleic acid precursors have much higher
ized detection equipment. Coomassie blue is the most com- specific activities than those of the corresponding 14 C-Iabeled
monly used organic dye, typically used to detect -1000 ng of preparations. In most experiments, the former are preferable

3.5 Purifying, Detecting, and Characterizing Proteins 99

0 TECHNIQUE ANIMATION: lmmunoblotting

IJ Electrophoresis and transfer Antibody detection IJ Chromogenic det ection

50S-polyacrylamide gel Membrane Incubate with React with substrate

Ab 1 (y); Incubate with enzyme- for Abrlinked enzyme
wash excess linked Ab2 ("();
wash excess
:XPERIMENTAL FIGURE 3-39 Western blotting (immunoblot- membrane is washed to remove unbound Ab 1• Step ID: The membrane
ting) combines several techniques to resolve and detect a specific is incubated with a second antibody (Ab2) that specifically recognizes
protein. Step 0 :After a protein mixture has been electrophoresed and binds to the first Ab1 • This second antibody is covalently linked to
through an SDS gel, the separated bands (or spots, for a two-dimensional either an enzyme (e.g., alkaline phosphatase, which can catalyze a
gel) are transferred (blotted) from the gel onto a porous membrane chromogenic reaction), a radioactive isotope, or some other substance
from which it is not readily removed. Step f): The membrane is flooded whose presence can be detected with great sensitivity. Step EJ:Finally,
with a solution of antibody (Ab 1) specific for the desired protein and the location and amount of bound Ab 2 are detected (e.g., by a deep-
allowed to incubate for a while. Only the membrane-bound band purple precipitate from chromogenic reaction). permitting the
containing this protein binds the antibody, forming a layer of antibody electrophoretic mobility (and therefore the mass) of the desired protein
molecules (whose position cannot be seen at this point). Then the to be determined, as well as its quantity (based on band intensity).

because they allow RNA or DNA to be adequately labeled beled substrates. The presence of such radioactive atoms is
after a shorter time of incorporation or require a smaller cell indicated with the isotope in brackets (no hyphen) as a prefix
sample. Various phosphate-containing compounds in which (e.g., [ 1H]Ieucine). In contrast, labeling almost all biomole-
the phosphorus atom is the radioisotope phosphorus-32 are cules (e.g., protein or nucleic acid) with the radioisotope
readily available. Because of their high specific activity, 32P- iodine-125 ( 125 1) requires the covalent addition of 125I to a
labeled nucleoti.des are routinely used to label nucleic acids mo lecule that normally does not have iodine as part of its
in cell-free systems. structure. Because this labeling procedure modifies the chem-
Labeled compounds in which a radioisotope replaces ical structure, the biological activity of the labeled molecule
atoms normally present in the molecule have virtually the may differ somewhat from that of the nonlabeled fo rm. The
same chemical properties as the corresponding nonlabeled presence of such radioactive atoms is indicated with the iso-
compounds. Enzymes, for instance, generally cannot distin- tope as a prefix with a hyphen (no bracket) (e.g., 125I-trypsin).
guish between substrates labeled in this way and their nonla- Standard methods for labeling proteins with 125 I result in co-
valent attachment of the 125 I primarily to the aromatic rings
of tyrosine side chains (mono- and diiodotyrosine). Nonra-
Radioisotopes Commonly Used dioactive isotopes find increasing use in cell biology, espe-
in Biological Research cially in nuclear magnetic resonance studies and in mass
spectroscopy applications, as will be explained below.
Isotope Half-Life
la be ling Experi ments a nd Detectio n of Ra d iolabe led Mole-
Phosphorus-32 14.3 days
cules Whether labeled compounds are detected by autoradi-
ography, a semiquantitative visual assay, or their radioactivity
Iodine-125 60.4 days
is measured in an appropriate "counter, .. a highly quantita-
tive assay that can determine the amount of a radiolabeled
Sulfur-35 87.5 days
compound in a sample, depends on the nature of the ex-
Tritium (hydrogen-3) 12.4 years
periment. In some experiments, both types of detection are
used.
Carbon-14 5730.4 years In one use of autoradiography, a tissue, cell, or cell con-
stituent is labeled with a radioactive molecule, unassociated

100 CHAPTER 3 • Protein Structure and Function

.· radioactive material is washed away, and the structure of the specific anti body to the protein of interest can be used to
sample is stabilized either by chemically cross-linking the precipitate the protein away from the other proteins in the
macromolecules in the sample ("fixation") or by freezing it. sample (immunoprecipitation). The precipitate is then solu-
The sample is then overlaid with a photographic emulsion bilized under denaturing conditions, for example, in an SDS-
sensitive to radiation. Development of the emulsion yields containing buffer, to separate the antibody from the protein,
small silver grains whose distribution corresponds to that of and the sample is analyzed by SDS-PAGE followed by auto-
the radioactive material and is usually detected by micros- radiography. In this type of experiment, a fluorescent com-
copy. Autoradiographic studies of whole cells were crucial in pound that is activated by the radiation ("scintillator") may
determining the intracellular sites where various macromol- be infu!>ed into rhe gel on completion of the electrophoretic
ecules are synthesized and the subsequent movements of separation so that the light emitted can be used to detect the
these macromolecules within cells. Various techniques em- presence of the labeled protein, either using film or a two-
ploying fluorescent microscopy, which we describe in Chap- dimensional electronic detector. This method is particularly
ter 9, have largely supplanted autoradiography for studies of useful for weak f3 emitters such as 3H.
this type. However, autoradiography is sometimes used in Pulse-chase experiments are particularly useful for tracing
various assays for detecting specific isolated DNA or RNA changes in the intracellular location of proteins or the modi-
sequences at specific tissue locations (see Chapter 5) in a tech- fication of a protein or metabolite over time. In this experi-
nique referred to as in situ hybridization. mental protocol, a cell sample is exposed to a radiolabcled
Quantitative measurements of the amount of radioactiv- compound that can be incorporated or otherwise attached to
ity in a labeled material are performed with several different a cellular molecule of interest-the "pulse"-for a brief pe-
instruments. A Geiger counter measures ions produced in a riod. The pulse ends when the unincorporated radtoactive
gas by the f3 particles or "' rays emitted from a radioisotope. molecules are washed away and the cells arc exposed to a
These instruments are mostly handheld devices used to mon- vast excess of the identical, but unlabeled compound to dilute
itor radioactivity in the laboratory to protect investigators the radioactivity of any remaining, but unincorporated radio-
from excess exposure. In a scintillation counter, a radiola- active compound. This procedure prevents further incorpora-
beled sample is mixed with a liquid containing a fluorescent tion of significant amounts of radiolabel after the "pulse"
compound that emits a flash of light when it absorbs the period and initiates the "chase" period (Figure 3-40). Sam-
energy of the f3 particlesor"' rays released in the decay of the ples taken periodically during the chase period are assayed to
radioisotope; a phototube in the instrument detects and determine the location or chemical form of the radiolabel as
counts these light flashes. Phosphorimagers are used to de- a function of time. Often, pulse-chase experiments, in which
tect radioactivity using a two-dimensional array detector, the protein is detected by autoradiography after immunopre-
storing digital data on the number of decays in disintegra- cipitation and SDS-PAGE, are used to follow the rate of syn-
tions per minute per small pixel of surface area . These in- thesis, modification, and degradation of proteins by adding
struments, which can be thought of as a kind of reusable radioactive amino acid precursors during the pulse and then
electronic film, are commonly used to quantitate radioactive detecting the amounts and characteristics of the radioactive
molecules separated by gel electrophoresis and are replacing protein during the chase. One can thus observe postsynthetic
photographic emulsions for this purpose. modifications of the protein that change its electrophoretic
A combination of labeling and biochemical techniques mobility and the rate of degradation of a specific protein,
and of visual and quantitative detection methods is often which is detected as the loss of signal with increasing time of
employed in labeling experiments. For instance, to identify chase. A classic use of the pulse-chase technique was in stud-
the major proteins synthesized by a particular cell type, a ies to elucidate the pathway traversed by secreted proteins
sample of the cells is incubated with a radioactive amino from their site of synthesis in the endoplasmic reticulum to
e
acid (e.g., 5Sjmethionine) for a few minutes, during which the cell surface (see Chapter 14).
time the labeled amino add enters the cells and mixes with
the cellular pool of unlabeled amino acids and some of the
Mass Spectrometry Can Determine the Mass
labeled amino acid is biosynthetically incorporated into
newly synthesized protein. Subsequently unincorporated ra- and Sequence of Proteins
dioactive amino acid is washed away from the cells. The cells Mass spectrometry (MS) is a powerful techmque for charac-
are harvested; the mixture of cellular proteins is extracted terizing proteins, especially for determining the mass of a
from the cells (for example, by a detergent solution) and protein or fragments of a protein. With such information Ill
then separated by any of the commonly used methods to re- hand, it is also possible to determine part of or all of the
solve complex protein mixtures into individual components. protein's sequence. This method permits the very highly ac-
Gel electrophoresis in combination with autoradiography or curate direct determination of the ratio of the mass (m) of a
phosphorimager analysis is often the method of choice. The charged molecule (molecular ion) to its charge (z), or m/z.
radioactive bands in the gel correspond to newly synthesized Techniques are then used to deduce the absolute mass of the
proteins, which have incorporated the radiolabeled amino molecular ion. There arc four key features of all mass spec-
acid. To detect a specific protein of interest rather than the trometers. The first is an ion source, from which charge, usu-
entire ensemble of biosynthetically radiolabeled proteins, a ally in the form of protons, is transferred to the peptide or

3.5 Purifying, Detecting, and Characterizmg Proteins 101

 (a)
I Pulse(h )
Chase (h ) o 1.s
0.5
1 2 14 1s 8 112 24
1
I
Normal m- -.- - --- - - + +
protein 0 0
p- !ri.

(b)

m-
Mutant
protein
p
• I ?_...,,.. Lightest ions
arrive at
Precursor protein (p) converted to ..__ _ _ _ _ _ ___. detector first
mature protein (m) by posttranslational
T ime
ca rbohyd rate addition
EXPERIMENTAL FIGURE 3-41 Molecular mass can be deter-
E.XPER 'IIENTA FIGURE 3-40 Pulse-chase experiments can
mined by matrix-assisted laser desorption/ionization t ime-of-
track the pathway of protein modification or movement within
flight (MALDI-TOF) mass spectrometry. In a MALDI-TOF mass
cells. (a) To follow the fate of a specific newly synthesized protein in a
spectrometer, pulses of light from a laser ionize a protein or peptide
cell, cells were incubated with [35S]met hionine for 0.5 h (the pulse) to
mixture that is absorbed on a metal target (step 0 ). An electric field
label all newly synthesized proteins, and the radioactive amino acid
accelerates the ions in the sample toward the detector (steps fJ and IJ).
not incorporated into the cells was then washed away. The cells were
The time to the detector is proportional to the square root of the
further incubated (the chase) for varying times up to 24 hours, and
mass-to-charge (m/z) ratio. For ions having the same charge, the
samples from each time of chase were subjected to immunoprecipita-
smaller ions move faster (shorter time to the detector). The molecular
tion to isolate one specific protein (here the low-density lipoprotein
weight is calculated using the time of flight of a standard.
receptor). SDS-PAGE of the immunoprecipitates followed by autoradi-
ography permitted visualization of t he one specific protein, which is
initially synthesized as a small precursor {p) and then rapidly modified
(ES). In MALDI (Figure 3-41) the peptide or protein sample
to a larger mature form (m) by addition of carbohydrates. About half of
is mixed with a low-molecular-weight, UV-absorbing or-
the labeled protein was converted from p tom during the pulse; the
ganic acid (the matrix) and then dried on a metal target.
rest was converted after 0.5 hour of chase. The protein remains stable
for 6-8 hours before it begins to be degraded (indicated by reduced
Energy from a laser ionizes and vaporizes the sample, pro-
band intensity). (b) The same experiment was performed in cells in
ducing singly charged molecular ions from the constituent
which a mutant form of the protein is made. The mutant p form cannot molecules. In ES (Figure 3-42a), the sample of peptides or
be properly converted to them form, and it is more quickly degraded proteins in solution is converted into a fine mist of tiny drop-
than the normal protein. [Adapted from K. F. Kozarsky, H. A. Brush, and lets by spraying thr ough a narrow capillary at atmospheric
M. Krieger, 1986,J. Cell Bioi. 102(5):1567-1575.] pressure. The droplets are formed in the presence of a high
electric field, rendering them highly charged . The droplets
evaporate in their short flight (mm) to the entrance of the
protein molecules. The formation of these ions occurs in the mass spectrometer's ana lyzer, forming multiply charged ions
presence of a high electric field that then directs the charged from the peptides and proteins. The gaseous ions are sam-
molecular ions into the second key component, the mass pled into the analyzer region of the MS, where they are then
analyzer. The mass analyzer, which is always in a high vac- accelerated by electric fields and separated by the mass ana-
uum chamber, physically separates the ions on the basis of lyzer on the basis of their mlz.
their differing mass-to-charge (mlz) ratios. The mass-sepa - The two most frequently used mass analyzers are time-
rated ions are subsequently directed to strike a detector, the of-flight (TOF) instruments and ion traps. TOF instru-
third key component, which provides a measure of the rela- ments exploit the fact that the time it takes an ion to pass
tive abundances of each of the ions in the sample. The fourth through the length of the analyzer before reaching the de-
essential component is a computerized data system that is tector is proportional to the square root of mlz (sma ller
used to calibrate the instrument; acquire, store, and process ions move faster than larger ones with the same charge; see
the resulting data; and often direct the instrument automati- Figure 3-41). In ion-trap analyzers, tunable electric fields
cally to collect additional specific types of data from the are used to capture, or "trap," ions with a specific mlz and
sample, based on the initial observations. This type of auto- to sequentially pa~~ the trapped ions out of the analyzer
mated feedback is used for the tandem MS (MS/MS) pep- onto the detector (see Figure 3-42a). By varying the electric
tide-sequencing methods described below. fields, ions with a wide range of mlz values can be exam-
The two most frequently used methods of generating ined one by one, producing a mass spectrum, which is a
ions of proteins and protein fragments are (1) matrix-assisted graph of mlz (x axis) versus relative abundance (y axis)
laser desorption/ionization (MALDI) and (2) electrospray (Figure 3-42b, top panel).

102 CHAPTER 3 • Protein Structure and Funct1on

 (a) Electrospray
needle Atmosphere
I 3-5~
Liquid ~ ===C=======;';:===- ---'
Mass Detector
~~--~--~ analyzer
Droplets Ions
containing Mass spectrometer
solvated ions

Electrospray ionization

(b) 568.65
100
(/) 852.49
c 90

-
.Q
0
<I)
u
c
80
70
836.47
co 60
"c::::l
50
..0
co
40
<I)
30
·~co 20
Qi 426.25 525.36
a: 10 932.43
0
600 900 1300 1400

1199.53

100
(/) 880.46
c 90 FIIVGYVDDTQFVR

-.Q
0
<I)
u
c
80
70
693.26 792.35
979.49

co 60
1142.53 1298.60
"c
::::l
50 706.62 1497.46
..0
co 40
<I)
30 650 ·44 765.40 1251.46
.~
ro
Qi
a:
0
300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600
m /z
EXPERIMENTAL FIGURE 3-42 Molecular mass of proteins and fragmentation into smaller ions that are then analyzed and detected.
peptides can be determined by electrospray ionization ion-trap The MS/ MS spectrum (also called the product-ion spectrum) provides
mass spectrometryz. (a) Electrospray (E$) ionization converts proteins detailed structural information about the parent ion, including
and peptides in a solution into highly charged gaseous ions by passing sequence information for peptides. Here the ion with an m! z of 836.47
the solution through a needle (forming the droplets) that has a high w as selected and fragmented and the m/z mass spectrum of the
voltage across it (charging the droplets). Evaporation of the solvent product ions measured. Note there is no longer an ion with an m/z of
produces gaseous ions that enter a mass spectrometer. The ions are 836.47 present because it was fragmented. From the varying sizes of
analyzed by an ion-trap mass analyzer that then directs ions to the the product ions, the understanding that peptide bonds are often
detector. (b) Top panel: Mass spectrum of a mixture of three major and broken in such experiments, the known m/z values for individual amino
several minor peptides is presented as the relative abundance of the acid fragments, and database information, the sequence of the
ions striking the detector (y axi~) as a function of the mass-to-charge peptide, FIIVGYVDDTQFVR, can be deduced. [Part (a) based on a figure
(m/z) ratio (x axis). Bottom panel: In an MS/ MS instrument such as the from S. Carr; part (b), unpublished data from S. Carr.)
ion trap shown in part (a), a specific peptide ion can be selected for

3. 5 Purifying, Detecting, and Charac terizing Prote in s 103

 In tandem, or MS/MS, instruments, any given parent ion Protein Primary Structure Can Be Determined by
in the original mass spectrum (Figure 3-42b, top panel) can Chemical Methods and from Gene Sequences
be mass-selected, broken into smaller ions by collision with
an inert gas, and then the mlz and relative abundances of the The classic method for determining the amino acid sequence
resulting fragment ions measured (Figu re 3-42b, bottom of a protein is Edman degradation. In this procedure, the
panel), all within the same machine in about 0.1 s per se- free amino group of theN-terminal amino acid of a polypep-
lected parent ion. This second round of fragmentation and tide is labeled, and the labeled amino acid is then cleaved
analysis permits the sequences of short peptides ( < 25 amino from the polypeptide and identified by high-pressure liquid
::~cid~) ro be determined because collisional fragmentation
chromatography. The polypeptide is left one residue shorter,
occurs primarily at peptide bonds, so the differences in with a new amino acid at the N-terminus. The cycle is re-
masses between ions correspond to the in-chain masses of peated on the ever-shortening polypeptide until all the resi-
the individual amino acids, permitting deduction of the se- dues have been identified.
quence in conjunction with database sequence information Before about 1985, biologists commonly used the Edman
(Figure 3-42b, bottom panel). chemical procedure for determining protein sequences. Now,
Mass spectrometry is highly sensi tive, able to detect as however, complete protein sequences usually arc determined
little as 1 X 10- 16 mol (100 attomoles) of a peptide or 10 X primarily by analysis of genome sequences. The complete ge-
10 15 mol (10 femtomoles) of a protein of 200,000 MW. nomes of several organisms hav~ already been sequenced,
Errors in mass measurement accuracy are dependent on the and the database of genome sequences from humans and nu-
specific mass analyzer used but are typically -0.01 percent merous model organisms is expanding rapidly. As discussed
for peptides and 0.05-0.1 percent for proteins. As described in Chapter 5, the sequences of proteins can be deduced from
in Section 3.6, below, it is possible to use MS to analyze DNA sequences that are predicted to encode proteins.
complex mixtures of proteins as well as purified proteins. A powerful approach for determining the primary struc-
Most commonly, protein samples arc digested by proteases, ture of an isolated protein combines MS and the use of se-
and the peptide digestion products are subjected to analysis. quence databases. First, the peptide "mass finge rprint" of the
An especially powerfully application of MS is to take a com- protein is obtained by MS. A peptide mass fingerprint is the
plex mixture of proteins from a biological specimen, digest it list of the molecular weights of peptides that are generated
with trypsin or other proteases, partially separate the com- from the protein by digestion with a specific protease, such as
ponents using liquid chromatography (LC), and then trans- trypsin. The molecular weights of the parent protein and its
fer the solution flowing out of the chromatographic column proteolytic fragments are then used to search genome data-
directly into an ES tandem mass spectrometer. This tech- bases for any similar-size protein with identical or similar
nique, called LC-MS/MS, permits the nearly continuous peptide mass maps. Mass spectrometry can also be used to
ana lySIS of a very complex mixture of proteins. directly sequence peptides using MS/MS, as described above.
The abundances of ions determined by mass spectrom-
etry in any given sample arc relative, not absolute, values.
Protein Conformation Is Determined
Therefore, if orie wants to use MS to compare the amounts
of a particular protein in two different samples (e.g., from a by Sophisticated Physical Methods
normal versus a mutant organism), it is necessary to have an In this chapter, we have emphasized that protein function is
internal standard in the samples whose amounts do not dif- dependent on protein structure. Thus, to figure out exactly
fer between the two samples. One then determines the how a protein works, its three-dimensional structure must
amounts of the protein of interest relative to that of the be determined. Determining a protein's conformation re-
standard in each sample. This permits quantitatively accu- quires sophisticated physical methods and complex ana lyses
rate inter-sample comparisons of protein levels. An alterna- of the experimental data. We brietly describe three methods
tive approach involves simultaneously comparing the used to generate three-dimensional models of proteins.
amounts of proteins from two different cell or tissue sam-
ples that are mixed together. To do this, investigators first X-ray Crystallography The use of x-ray crystallography to
incubate one of the samples with amino acids containing determine the three-dimensional structures of proteins was
"heavy" isotope atoms. These are biosynthetically incorpo- pioneered by Max Perutz and john Kendrew in the 1950s. In
rated into all of the proteins of that sample. Proteins from this technique, beams of x-rays are passed through a protein
the two samples a re then mixed together and analyzed by crystal in which millions of protein molecules are precisely
mass spectrometry. Proteins and pcptides derived from the aligned with one another in a rigid crystalline array. The
"heavy" sample can be distinguished in the mass spectrom- wavelength~ of x-rays are about 0.1-0.2 nm, short enough
eter from those from the other, " light," sample because of to determine the positions of individual atoms in the protein .
their higher masses. Thus a direct comparison of the relative The electrons in the atoms of the crystal scatter the x-rays,
amounts of each protein in each sample can be made. When which produce a diffraction pattern of discrete spots when
the samples are cells grown in the laboratory, the method they are intercepted by photographic fi lm or an electronic de-
is called stable isotope labeling with amino acids in cell tector (Figure 3-43). Such patterns are extremely complex-
culture (SILAC). composed of as many as 25,000 diffraction spots, or reflections,

104 CHAPTER 3 • Protein Structure and Function

 (a) of electrons (called the electron density map). With the three-
dimensional electron density map in hand, one then "fits" a
molecular model of the protein to match the electron density,
·. and it is these models that one sees in the various diagrams of
X-ray proteins throughout this book (e.g., Figure 3-8). The process
beam is analogous to reconstructing the precise shape of a rock
from the ripples that it creates when thrown into a pond.
Crystal-- Although sometimes the structures of parts of the protein
cannot be clearly defined, using x-ray cry:.tallography, re-
searchers are systematically determining the structures of rep-
resentative types of most proteins. To date, the detatled
three-dimensional structures of more than 18,000 proteins
have been established using x-ray crystallography. These
structures can be found in the Research Collaboratory for
Structural Bioinformatics Protein Data Bank (http://www.
rcsb.orglpdb/home/home.do), each with its own PDB entry.

(b) Cryoelectron Microscopy Although some proteins readily

crystallize, obtaining crystals of others-particularly large
multisubunit proteins and membrane-associated proteins-
requires a time-consuming, often robot-assisted trial-and-
error effort to find just the right conditions, if they can be
found at all. (Growing crystals suitable for structural studies
is as much an art as a science.) There are several ways to
determine the structures of such difficult-to-crystallize pro-
teins. One is cryoelectron microscopy. In this technique, a
protein sample is rapidly frozen in liquid helium to preserve
its structure and then examined in the frozen, hydrated state
in a cryoelectron microscope. Pictures of the protein are
taken at various angles and recorded on film using a low
dose of electrons to prevent radiation-induced damage to the
structure. Sophisticated computer programs analyze the im-
ages and reconstruct the protein's structure in three dimen-
sions. Recent advances in cryoelectron microscopy permit
researchers to generate molecular models that can help pro-
vide insight into how the protein functions. The use of cryo-
electron microscopy and other types of electron microscopy
EXPERIMENTAL FIGURE 3-43 X-ray crystallography provides for visualizing cell structures is discussed in Chapter 9.
diffraction data from which the three-dimensional structure of
a protein can be determined. (a) Basic components of an x-ray NMR Spectroscopy The three-dimensional structures of small
crystallographic determination. When a narrow beam of x-rays strikes proteins containing as many as 200 amino acids can be studied
a crystal. part of it passes straight through and the rest is scattered routinely with nuclear magnetic resonance (NMR) spectros-
(diffracted) in various directions. The intensity of the diffracted waves, copy. Specialized approaches can be used to extend the size
which form periodic arrangements of diffraction spots, is recorded range to somewhat larger proteins. In this technique, a concen-
on an x-ray film or with a solid-state electronic detector. (b) X-ray trated protein solution is placed in a magnetic field, and the
diffraction pattern for a protein crystal collected on a solid-state effects of different radio frequencies on the nuclear spin states
detector. From complex analyses of patterns of spots like this one, the of different atoms are measured. The spin state of any atom is
location of the atoms in a protein can be determined. [Part (a) adapted influenced by neighboring atoms in adjacent residues, with
from L. Stryer, 1995, Biochemistry, 4th ed., W. H. Freeman and Company, p. 64;
closely spaced residues having a greater influence than distant
part (b) courtesy of J. Berger.]
residues. From the magnitude of the effect, the distances be-
tween residues can be calculated hy a triangulation-like pro-
whose measured intensities vary depending on the distribu- cess; these distances are then used to generate a model of the
tion of the electrons, which is, in turn, determined by the three-dimensional structure of the protein. An important
atomic structure and three-dimensional conformation of the distinction between x-ray crystallography and NMR spec-
protein. Elaborate calculations and modifications of the pro- troscopy is that the former method directly determines the
tein (such as the binding of heavy metals) must be made to locations of the atoms while the later directly determines the
interpret the diffraction pattern and calculate the distribution distances between the atoms.

3.5 Purifying , Detecting, and Characterizing Proteins 105

 Although NMR does not require the crystallization of a
protein, a definite advantage, this technique is limited to • Isotopes, both radioactive and "heavy" or " light" nonra-
proteins smaller than about 20 kDa. However, NMR analy- dioactive, play a key role in the study of proteins and other
sis can provide information about the ability of a protein to biomolecules. They can be incorporated into mo lecules
adopt a set of closely related, but not exactly identical, con- without changing the chemical composition of the molecule
formations and to move between these conformations (pro- or as add-on tags. They can be used to help detect the syn-
tein dynamics). This is a common feature of proteins, which thesis, location, processing, and stability of proteins.
are not absolutely rigid structures but can "breathe" or ex- • Autoradiography is a semiquantitative technique for de-
hibit ~light variations in the relative positions of their con- tecting radioactively labeled molecules in cells, tissues, or
stituent atoms. In some cases these variations can have electrophoretic gels.
functional significance, for example in how proteins bind to
Pulse-chase labeling can determine the intracellular fate of
one another. NMR structural analysis has been particularly
proteins and other metabolites (see figure 3-40).
useful in studying isolated protein domains, which can often
be obtained as stable structures and tend to be small enough Mass spectrometry is a very sensitive and highly precise
for this technique. To date, there arc more than 5000 NMR- method of detecting, identifying, and characterizing proteins
determined protein structures available in the Protein Data and peptides.
Bank (http://www.rcsb.org/pdb/home/home.do). Three-dimensional structures of proteins are obtained by
x-ray crystallography, cryoelectron microscopy, and NMR
spectroscopy. X-ray crystallography provides the most detailed
structures but requires protein crystallization. Cryoelectron
KEY CO 1CEPTS of Section 3.5 microscopy is most useful for large protein complexes, which
are difficult to crystallize. Only relatively small proteins are
Purifying, Detecting, and Characterizing Proteins amenable to NMR analysis.
Proteins can be separated from other cell components and
from one another on the basis of differences in their physical
and chemical properties.
Various assays are used to detect and quantify proteins. 3.6 Proteomics
Some assays use a light-producing reaction to generate a
For most of the twentieth century, the study of proteins was
readily detected signal. Other assays produce an amplified
restricted primarily to the analysis of individual proteins. for
colored signal with enzymes and chromogenic substrates.
example, one would study an enzyme by determining its enzy-
Centrifugation separates proteins on the basis of their matic activity (substrates, products, rate of reaction, require-
rates of sedimentation, which are influenced by their masses ment for cofactors, pH, etc.), its structure, and its mechanism
and shapes (see.Figure 3-35). of action. In some cases, the relationships between a few en-
Electrophoresis separates proteins on the basis of their rates zymes that participate in a metabolic pathway might also be
of movement in an applied electric field. SDS-polyacrylamide studied. On a broader scale, the localization and activity of an
gel electrophoresis (SDS-PAGE) can resolve polypeptide enzyme wou ld be examined in the context of a cell or tissue.
chains differing in molecular weight by 10 percent or less The effects of mutations, diseases, or drugs on the expression
(see Figure 3-36). Two-dimensional gel electrophoresis pro- and activity of the enzyme might also be the subject of inves-
vides additional resolution by separating proteins first by tigation. This multipronged approach provided deep insight
charge (first dimension) and then by mass (second dimension ). into the function and mechanisms of action of individual pro-
teins or relatively small numbers of interacting proteins. How-
Liquid chromatography separates proteins on the basis of
ever, such a one-by-one approach to studying proteins does
their rates of movement through a column packed with
not readily provide a global picture of what is happening in
spherical beads. Proteins differing in mass arc resolved on
the proteome of a cell, tissue, or entire organism.
gel filtration columns; those differing in charge, on ion-exchange
columns; and those differing in ligand-binding properties, on
affinity columns, including antibody-based affinity chroma- Proteomics Is the Study of All or a Large Subset
tography (see Figure 3-38). of Proteins in a Biological System
Antibodies are powerful reagents used to detect, quantify, The advent of genomics (sequencing of genomic DNA and
and isolate proteins. its associated technologies, such a~ !>irnulraneous analysis of
lmmunoblotting, also called Western blotting, is a fre- the levels of all mRNAs in cells and tissues) clearly showed
quently used method to study specific proteins that exploits that a global, or systems, approach to biology could provide
the high specificity and sensitivity of protein detection by unique and highly valuable insights. Many scientists recog-
antibodies and the high-resolution separation of proteins by nized that a global analysis of the proteins in biological sys-
SDS-PAGE (see Figure 3-39). tems had the potential for equally valuable contributions to
our understanding. Thus a new field was born-proteomics.

106 CHAPTER 3 • Protein Structure and Function

G) PODCAST: Use of Mass Spectrometry in Cell Biology

Electrospray ionization mass spectrometer

1st MS Collision 2nd MS Detector
fragmentation
Complex
mixture of
peptides

~(?"-.~.
?? )~?~.??
?{?i???~ 4 3 2 1

?? ? .... LC
separation
into fractions

i Protease
(e.g., trypsin)
of less
complex
mixtures

Select mass of a Sequence

Protein
from e
a fraction
mixture in
biological
sample Repeat for other peptides in the fraction

Repeat for multiple fractions from the LC outflow to sequence most peptides
in the starting complex peptide mixture

~
Compare results with databases to identify proteins in the original
biological sample

EXPERIMENTAL FIGURE 3 44 LC-MS/ MS is used to identify electrospray ionization into a tandem mass spectrometer. The
the proteins in a complex biological sample. A complex mixture fractions are then sequentially subjected to multiple cycles of MS/MS
of proteins in a biological sample (e.g., isolated preparation of Golgi until masses and sequences of many of the peptides are determined
organelles) is digested with a protease; the mixture of resulting and used to identify the proteins in the original biological sample
peptides is fractionated by liquid chromatography (LC) into multiple, through comparison with protein databases. [Based on a figure provided
less complex, fractions, which are slowly but continuously injected by by 5. Carr.]

Protcomics is the systematic study of the amounts, modifica- • Which proteins are present in large multiprotein complexes,
tions, interactions, localization, and functions of all or sub- and which proteins are in each complex? What arc the func-
sets of proteins at the whole-organism, tissue, cellular, and tions of these complexes, and how do they interact?
subcellular levels.
• When the state (e.g., growth rate, stage of cell cycle, dif-
A number of broad questions are addressed in proteomic
ferentiation, stress level) of a cell changes, do the proteins in
studies:
the cell or secreted from the cell change in a characteristic
(fingerprint-like) fashion? Which proteins change and how
• In a given sample (whole organism, tissue, cell, subcellular
(relative amounts, modifications, splice forms, etc.)? (This is
compartment), what fraction of the whole proteome is ex-
a form of protein expression profiling that complements the
pressed (i.e., which proteins are present)?
Lranscriptional (mRNA) profJ!ing discussed in Chapter 7.)
• Of those proteins present in the sample, what are their
• Can such fingerprint-like changes be used for diagnostic
relative abundances?
purposes? For example, do certain cancers or heart disease
• What are the relative amounts of the different splice forms cause characteristic changes in blood proteins? Can the pro-
and chemically modified forms (e.g., phosphorylated, meth- teomic fingerprint help determine if a given cancer is resistant
ylated, fatty acylated) of the proteins? or sensitive to a particular chemotherapeutic drug? Proteomic

3.6 Proteomics 107

EXPERIMENTAL FIGURE 3-45 Densit y-gradient centrifugation (a)

~
and LC-MS/ MS can be used to identify many of the proteins in
....
organelles. (a) The cells in liver tissue were mechanically broken to 0 QJ
Vl :l
release the organelles, and the organelles were partially separated by -Vl

Mitoc~ondria1~Early
-(f)
QJ.-
density-gradient centrifugation. The locations of the organelles-which u ...
were spread out through the gradient and somewhat overlapped with endosomes
one another-were determined using immunoblotting with antibodies Golgi
that recognize previously identified, organelle-specific proteins. c:
+J•Q
Fractions from the gradient were subjected to proteolysis and LC-MS/
MS to identify the peptides, and hence the proteins, in each fraction.
Comparisons with the locations of the organelles in the gradient (called
c: ...
"'"'
·"O:l
~~
- Ol

t9C:QJ
(__. !It-,;]
protein correlation profiling) permitted assignment of many individual
proteins to one or more organelles (organelle proteome identification).
(,)

! lmmunoblotting
(b) The hierarchical breakdown of data derived from the procedures in
F1 ATP synthase
part (a). Note that not all proteins identified could be assigned to
organelles and some proteins were assigned to more than one
organelle. [From L. J. Foster et al., 2006, Cell 125(1):187-199.)
6c:.,.
:l 0
E:O
E
Early endosome
antigen 1
1, 2-a-mannosidase
--
Pr~teolysis and LC-MS/MS
fingerprints can also be the starting point for studies of the
mechanisms underlying the change of state. Proteins (and
l Localization of proteins

other biomolecules) that show changes that are diagnostic of

a particular state are called biomarkers.
• Can changes in the proteome help define targets for drugs
or suggest mechanisms by which that drug might induce
l'lf:/Y:'\
0.0 L.....""--~=----"""'-.;_---'"---"-

roxie side effects? If so, it might be possible to engineer mod- (b)

32 gradient fractions
ified versions of the drug with fewer side effects.
~
These are just a few of the questions that can be ad- 22,260 peptides
dressed using proteomics. The methods used to answer these
questions are as diverse as the questions themselves, and ~
2,197 proteins
their numbers are growing rapidly.
~
1,500 proteins quantified
Advanced Techniques in Mass Spectrometry ~
Are Critical to Proteomic Analysis 1.404 proteins localized---+ Nucleus 196

Advances in proteomics technologies (e.g., mass spectrome- ~

Cytosol 488 +- 1,258 cytoplasmic proteins
try ) profoundly affect the types of questions that can be
practically studied. For many years, two-dimensional gel ~
electrophoresis has allowed researchers to separate, display,

.- - /!
Mitochondrion + - - Cytoplasmic _.... Proteosome
and characterize a complex mixture of proteins (see Figure 297 organelles 50
3-37). The spots on a two-dimensional gel can be excised, the ~ 968 -----..
·.
protein fragmented by proteolysis (e.g., by trypsin digestion), Endoplasmic
\~ ~
~ Re<my::~ane
9
and the fragments identified by MS. An alternative to this reticulum Plasma
two-dimensional gel method is high-throughput LC-MSIMS. 229
/ ' '
Figure 3-44 outlines the general LC-MS/MS approach, in
Golgi
which a complex mixture of proteins is digested with a pro- 67 endosomes
tease; the myriad resulting peptides are fractionated by LC 326
ER/Golgi Early
into multiple, less complex fractions, which are slowly but vesicles endosomes
continuously injected by electrospray ionization into a tan- 220 76
dem mass spectrometer. The fractions are then sequentially
subjected to multiple cycles of MS/MS until sequences of
many of the peptides are determined and used to identify release the organelles, and the organelles were partially sepa-
from databases the proteins in the original biological sample. rated by density-gradient centrifugation. The locations of
An example of the use of LC-MS/MS to identify many of the organelles in the gradient were determined using immu-
the proteins in each organelle is seen in Figure 3-45. Cells from noblotting with antibodies that recognize previously identified,
murine (mouse ) liver tissue were mechanically broken to organelle-specific proteins. Fractions from the gradient were

108 CHAPTER 3 • Protein Structure and Function

 subjected to LC-MS/MS to identify the proteins in each frac- lenging problem requires supercomputers or large clusters of
tion, and the distributions in the gradient of many individual computers working in synchrony. Currently, only the struc-
proteins were compared with the distributions of the organ- tures of very small domains containing 100 residues or fewer
elles. This permitted assignment of many individual proteins to can be predicted at a low resolution. However, continued
one or more organelles (organelle proteome proteome profil- developments in computing and models of protein folding,
ing). More recently, a combination of organelle purification, combined with large-scale efforts to solve the structures of
MS, biochemical localization and computational methods has all protein structural motifs by x-ray crystallography, will
been used to show that at least 1000 distinct proteins are local- a !low the prediction of the structures of larger proteins.
ized in the mitochondria of humans and mice. With an exponentially expanding database of ~tructurally
Proteomics combined with molecular genetics methods defined motifs, domains, and proteins, scientists will be able
are currently being used to identify all protein complexes in a to identify the motifs in an unknown protein, match the
eukaryotic cell, the yeast Saccharomyces cerevisiae. Approxi- motif to the sequence, and use this to predict the three-
mately 500 complexes have been identified, with an average dimensional structure of the entire protein.
of 4.9 distinct proteins per complex, and these in turn are New combined approaches will also help in determining
involved in at least 400 complex-to-complex interactions. high-resolution structures of molecular machines. Although
Such systematic proteomic studies are providing new insights these very large macromolecular assemblies usually are dif-
into the organization of proteins within cells and how pro- ficult to crystalltze and thus to solve by x-ray crystallogra-
teins work together tO permit cells to live and function. phy, they can be imaged in a cryoelectron microscope at
hquid helium temperatures and high electron energies. From
millions of individual "particles," each representing a random
view of the protein complex, the three-dimensional structure
..· KEY CONCEPTS of Section 3.6 can be built. Because subunits of the complex may already
Proteomics be solved by crystallography, a composite structure consist-
ing of the x-ray-derived subunit structures fit to the EM-de-
• Proteomics is the systematic study of the amounts (a nd nved model will be generated.
changes in the amounts), modifications, interactions, local- Methods for rapid structure determination combined
ization, and functions of all or subsets of all proteins in bio- with identification of novel substrates and inhibitors will
logical systems at the whole-organism, tissue, cellular, and help determine the structures of enzyme-substrate complexes
subcellular levels. and transition states and thus help provide detailed mforma-
• Proteomics provides insights into the fundamental organi- tion regarding the mechanisms of enzyme catalysis. Mem-
zation of proteins within cells and how this organization is brane proteins, because of the specialized environment in
influenced by the state of the cells (e.g., differentiation into which they reside and their solubility characteristics, remain
distinct cell types; response to stress, disease, and drugs). challenging, although progress in this area is accelerating.
• A wide variety of methods are used for proteomic analy- Although our understanding of chaperone structure and ac-
ses, including two-dimensional gel electrophoresis, density- tivity continues to grow exponentially, a number of critical
gradient centri{ugation, and mass spectroscopy (MALDI- questions remain a mystery. We do not understand precisely
TOF and LC-MS/MS). how cells make the distinction between unfolded and misfolded
versus properly folded. Clearly, the exposure of hydrophobic
• Proteomics has helped begin to identify the proteomes of side chains plays a role, but what are the other determinants of
organelles ("organelle proteome profiling") and the organi- this key recognition process? How is the decision made to turn
zation of individual proteins into multiprotein complexes from trying to refold a protein to degrading it?
that interact in a complex network to support life and cel- The rapid development of new technologies can be ex-
lular function (see Figure,3-45). pected to help solve some of the still outstanding problems in
proteomics. It is becoming possible to identify and character-
ize intact proteins as large as 30-70 kDa in complex mix-
tures using MS techniques without first digesting the samples
Perspectives for the Future
into peptides-a method called the "top-down" approach,
Impressive expansion of the computational power of com- in contrast to starting with fragments of the protein (" bot-
puters is at the core of advances in determining the three- tom-up" approach). An ongoing problem in proteomic anal-
dimensional structures of proteins. For example, vacuum ysis of complex mixtures is that it is difficult to detect and
tube computers running on programs punched on cards identify protein fragments from <;::~mples whose concentra-
were used to solve the first protein structures on the basis of tions in the sample differ by more than 1000-fold: some
x-ray crystallography, a process that at the time took years samples, such as blood plasma, contain proteins whose con-
but can now be accomplished in a matter of days and in centrations vary over a 10 11 -fold range. Routine analysis of
some cases hours. In the future, researchers aim to predict specimens with such diverse concentrations should dramati-
the structures of proteins using only amino acid sequences cally improve the mechanistic and diagnostic value of blood
deduced from gene sequences. This computationally chal- plasma proteomics.

Perspectives for the Future 109

 Which enzyme better stabilizes the transition state? Which
Key Terms enzyme functions as a better catalyst?
a helix 62 liquid chromatography 96
activation energy 78 motif 65
active site 79 peptide bond 61
allostery 88 phosphorylation 90
amyloid filament 76 polypeptide 62 (!)
autoradiography 100 primary structure 62
> r-------------~~~-\-------­
El - No enzyme
13 sheet 62 proteasome 85 ~ r-----------,f~~~r4~------ -E1
Q)
13 turn 62 protein 62 Q) - E1
~
chaperone 72 proteome 60 LL

conformation 59 proteomics 106

cooperativity 88 quaternary structure 68
domain 67 rate-zonal centrifugation 93
electrophoresis 94 secondary structure 62 E+ S ES
enzyme 78 tertiary structure 64 Progress of reaction ~
homology 69 ubiquitin 87
kinase 90 Vmax 79 5. A healthy adaptive immune system can raise antibodies that
Km 80 Western blotting 99 recognize and bind with high affinity to almost any stable mol-
ligand 19 x-ray crystallography 104 ecule. The molecule to which an antibody binds is known as
"antigen." Antibodies have been exploited by enterprisi ng sci-
entists to generate valuable tools for research, diagnosis, and
therapy. One clever application is the generation of antibodies
Review the Concepts that function like enzymes to catalyze complicated chemical
reactions. If you wished to produce such a "catalytic" anti-
1. The three-dimensional structure of a protein is determined
body, what would you suggest using as the antigen? Should it
by its primary, secondary, and tertiary structures. Define the
be the substrate of the reaction? The product? Something else?
pnmary, secondary, and tertiary structures. What are some
of the common secondary structures? What are the forces 6. Proteins are degraded in cells. What is ubiquitin, and
that hold together the secondary and tertiary structures? what role does it play in tagging proteins for degradation?
What is the role of proteasomes in protein degradation?
2. Proper folding of proteins is essential for their biological
How might proreasome inhibitors serve as chemotherapeutic
activity. In genetal, the functional conformation of a protein
(cancer-treating) agents?
is the conformation with lowest energy. This means that if an
unfolded protein is allowed to reach equilibrium, it should 7. The function of proteins can be regu lated in a number of
assemble automatically into its native, functioning folded ways. What is cooperativity, and how does it influence pro-
state. Why then is there a need for molecular chaperones and tein function? Describe how protein phosphorylation and
chaperonins in cells? What different roles do molecular chap- proteolytic cleavage can modulate protein function.
erones and chaperonins play in the folding of proteins? 8. A number of techniques can separate proteins on the
3. Enzymes catalyze chemical reactions. What constitutes rhe basis of their differences in mass. Describe the use of two of
active sire of an enzyme? What are the turnover number (k,a,), these techniques, centrifugation and gel electrophoresis. The
the Michaelis constant (K111 ), and the maximal velocity ( V1113 , ) blood proteins transferrin (MW 76 kDa) and lysozyme (MW
of an enzyme? The k,,t for carbonic anhydrase is 5 X 10 5 15 kDa) can be separated by rate-zonal centrifugation or
molecules/s. This is a "rate constant" bur nor a "rate." What SDS-polyacrylamide gel electrophoresis. Which of the two
is the difference? By what concentration wou ld you multiply proteins will sediment faster during centrifugation? Which
this rate constant in order to determine an actual rate of prod- will migrate faster during electrophoresis?
uct formation (V)? Under what circumstances would this rare 9. Chromatography is an analytical method used to separate
become equal to the maximal velocity (V ma.xl of the enzyme? proteins. Describe the principles for separating proteins by
4. The following reaction coordinate diagram charts the en- gel filtration, ion-exchange, and affinity chromatography.
ergy of a substrate molecule (S) as it passes through a transi- 10. Various methods have been developed for detecting pro-
tion state (Xt) on its way to becoming a stable product (P) teins. Describe how radioisotopes and autoradiography can
alone or in the presence of one of two different enzymes (El be used for labeling and detecting proteins. How does West-
and E2). How does the addition of either enzyme affect the ern blotting detect proteins?
change in Gibb's free energy (.1G) for the reaction? Which of 11. Physical methods are often used to determine protein con-
the two enzymes binds with greater affinity to the substrate? formation. Describe how x-ray crystallography, cryoelectron

110 CHAPTER 3 • Protein Structure and Function

 microscopy, and NMR spectroscopy can be used to determine control and treated cells. In the following example, only a
the shape of proteins. What are the advantages and disadvan- few protein spots are shown for simplicity. Proteins are sep-
tages of each method? Which is better for small proteins? arated in the first dimension on the basis of charge by iso-
Large proteins? Huge macromolecular assemblies? electric focusing (pH 4-1 0) and then separated by size by
12. Mass spectrometry is a powerful tool in proteomics. What SDS-polyacrylamide gel electrophoresis. Proteins are de-
are the four key features of a mass spectrometer? Describe tected with a stain such as Coomassie blue and assigned
briefly how MALDI and two-dimensional polyacrylamide gel numbers for identification.
electrophoresis (2D-PAGE) could be used to identify a protein a. Cells arc treated with a drug ("1 Drug") or left un-
expressed in cancer cells but nor in normal healthy cells. treated ("Control"), and then proteins are extracted and
separated by two-dimensional gel electrophoresis. The
stained gels are shown below. What do you conclude
Analyze the Data about the effect of the drug on the steady-state levels of
proteins 1-7?
1. Beautiful models of macromolecules such as protems and

.. nucleic acids are generated from files of atomic coordinates

obtained usually from x-ray diffraction of crystallized samples 4
Control
pH 10 4
+Drug
pH 10
or NMR analysis of the molecules in solution. The Protein 1

•
2
Data Bank (PDB) is a publicly accessible repository of macro-
molecular atomic coordinate files that can be accessed online 3
• • •
• 5 •
•
at http://www.rcsb.org. Access the PDB and familiarize your- 4
•
self with its homepage. How many molecular structures does it 6 • 7
. +

contain today? What is the "Molecule of the Month"? Down- + • • + • • +

load a coordinate file for the serine protease chymotrypsin by

typing the accession code "lACB" into the search window. b. You suspect that the drug may be inducing a protein
This will take you to a page describing the x-ray crystal struc- kinase and so repeat the experiment in part (a) in the pres-
ture of a complex between bovine alpha-chymotrypsin and the ence of 32 P-labeled inorganic phosphate. In. this experiment
small pseudosubstrate inhibitor protein eglin-c. When and in the two-dimensional gels are exposed to x-ray film to detect
what journal was the study reporting this structural model the presence of 32 P-Iabeled proteins. The x-ray films are
published? Click on the "Download File" link, select "PDB shown below. What do you conclude from this experiment
File (Text)", and download the file "lACB.pdb". This is an about the effect of the drug on proteins 1-7?
atomic coordinate (.pdb) file that specifies the relative posi-
tions for each atom in this protein complex as determined ex- Control +Drug
perimentally by x-ray crystallography. Open the file in a text 4 pH 10 4 pH 10
viewer or word processor and look at its format. The first sev-
eral hundred lines contain background information including
the names of the molecules, their natural sources, how they
• •
were prepared for the experiment, statistical analysis of the •
model quality, and bibliographic information. Eventually, you
will arrive at a long list of lines that each begin with "ATOM".
These are the coordinates, listed by atom number, atom type,
•
+L-----'---- •
+'--------'

amino acid rype, and chain number. Each "ATOM" line ends
c. To determine the cellular localization of proteins 1-7,
with five numbers representing the atomic position on an x, y,
the cells from part (a) were separated into nuclear and cyto-
z axis, its "occupancy," ansJ its "thermal factor." Close the file
plasmic fractions by differential centrifugation. Two-
and download software for viewing the molecular model.
dimensional gels were run, and the stained gels are shown
There are many, such as RasMol, iMol, Swiss-PDB Viewer,
below. What do you conclude about the cellular localization
and PyMol, that are available for download in a free format
of proteins 1-7?
for educational purposes. Open the lACB.pdb file and rwirl it
around in the viewer. Can you identify the protease? The in-
Control
hibitor protein? Can you find the enzyme's active site? What
other observations can you make about serine proteases from Nuclear Cytoplasmic
the model of this inactivated complex? 4 pH 10 4 pH 10

2. Proteomics involves the global analysis of protein ex-

•
pression. In one approach, all the proteins in control cells ••
and treated cells are extracted and subsequently separated
using two-dimensional gel electrophoresis. Typically, hun- •
•
• •
dreds or thousands of protein spots are resolved and the
steady-state levels of each protein are compared between ·-
+L__ _ _ _ __
+L--------'
•

Analyze the Data 111

 + Drug Gough, J. 2006. Genomic scale sub-fam1ly assignment of protein
domains. Nucl. Acids Res. 34(13):3625-3633.
Nuclear Cytoplasmic
Koonin, E. V., Y. I. Wolf, and G. P. Karev. 2002. The structure
4 pH 10 4 pH 10
of the protein universe and genome evolution. Nature 420:218-223.

• • Lesk, A. M. 2001. Introduction to Protem Architecture. Oxford .

Levitt M. 2009. Nature of the protein universe. Proc. Nat/.

•.
Acad. Sci. USA 106(27):11079-11084 .

+ • •
•
+
• •
Lin, Z., and H. S. Rye. 2006. GroEL-mediated protein folding:
makmg the impossible, possible. Cnt. Rev. Biochem. Mol. Bioi.
41(4):211-239.
.·
Orengo, C. A., D. T. Jones, and J. M. Thornton. 1994. Protein
superfamilies and domain superfolds. Nature 372:631-634.
d. Summarize the overall properties of proteins 1-7, Patthy, L. 1999. Protem Evolutzon. Blackwell Science.
combining the data from parts (a), (b), and (c). Describe how Rocher, J.-C., and P. T. Landsbury. 2000. Amyloid fibrillogen-
you could determine the identity of any one of the proteins. esis: themes and variations. Curr. Opm. Struc. Bioi. 10:60-68.
Taipale, .\<I., D. F. Jarosz, and S. Lindquist. 2010. HSP90 at the
hub of protein homeostasis: emerging mechanistic insights. Nat.
Reu. Mol. Cell Bioi. 11(7):515-528.
References Vogel, C., and C. Chothia. 2006. ~rotein family expansions and
biolog1cal complexity. PLoS Comput. Bzol. 2(5):e48.
General References Yaffe, M. B. 2006. "Bits" and pieces. Sci STKE. 2006(340):pe28.
Berg, .J. ~1., J. [. Tymoczko, and L Srryer. 2007. Biochemistry, Young, J. C., et al. 2004. Pathways of chaperone-mediated
6th ed. W. H. Freeman and Company. protein folding in the cytosol. Nat. Rev. Mol. Cell Bioi. 5:781-791.
Nelson, D. L., and M. Yl.. Cox. 2005. l.ehmnger Princzples of Wlodarski, T., and B. Zagrovic. 2009. Conformational selection
Biochemzstry, 4th cd. W. H. freeman and Company. and induced fit mechanism underlie specificity 111 noncovalent
interactions with ubiquinn. Proc. Nat/. Acad. Scr. USA
Web Sites 106(46):19346-19351. .
Entry site into proteins, structures, genomes, and taxonomy:
http://www.ncbi.nlm.nih.gov/Entrez/ Protein Binding and Enzyme Catalysis
The prorein 3-D structure database: hnp://www.rcsb.org/
Dressler, D. H., and H. Potter. 1991. Dzscovering Enzymes.
Strucrural classifications of proteins: http:/lscop.berkcley .edu/
Scientific American Library.
Sites containing general information about proteins: http://
Fersht, A. 1999. En;:;yme Structure and Mechanism, 3d ed.
www .expasy .ch/; hrrp://www. proweb.org/; http://scop. berkeley .edu/
W. H. Freeman and Company.
mrro.html
Jeffery, C. J. 2004. Molecular mechanisms for multitasking:
PROSITE database of protein families and domains: hrrp://
recent crystal structures of moonlighting prorems. Curr. Opin.
www.e'l.pasy .org/prosite/
Struc. Bzol. 14(6):663-668.
Domam orgamzation of proteins and large collection of multiple
Marnett, A. B., and C. S. Crmk. 2005. Papa's got a brand new
sequence alignments: http://www.sanger.ac.uk/Software/Pfam/;
tag: advances in identification of proteases and their substrates.
http://people.crysr. bbk.ac. uk/ ·ubcg 16zicpn/elmovies.html
Trends Bioteclmol. 23(2):59-64.
MiroC:arra: An Inventor} of Mammalian ,\l1tochondnal
Polgar, L. 2005. The catalytic mad of serme peptidases. Cell
Genes:hrrp://www.broadinsmute.org/pubs/MiroCarra/index.html
Mol. Life Sci. 62( 19-20):2161-2172.
Hierarchical Structure of Proteins and Protein Folding Radisky, E. S., et al. 2006. Insights into the serine protease
Branden, C., and J. Tooze. 1999. flztroductum to Protein mechanism from atomic resolution structures of trypsin reaction
Structure. Garland. intermediates. Proc. Nat'/ Acad. Scz. USA 103(18):6835-6840.
Broadley, S. A., and F. U. Hartl. 2009. The role of molecular Schenone, M, B. C. Furie, and B. Furie. 2004. The blood
chaperones in human misfolding diseases. FEBS Lett. coagulation cascade. Curr. Opin. Hematol. 11 (4):272-277.
583( 16):264..,-2653. Schramm, V. L. 2005. Enzymatic transition stares and transmon
Brodsky, J. L., and G. Chiosis. 2006. Hsp70 molecular state analogues. Curr. Opm. Struc. Bioi. 15(6):604-613.
chaperones: emerging roles m human disease and identification of
\mall molecule modulators. Curr. Top. Med. Chem. 6(11):1215- Regulating Protein Function
1225. Bellelli, A., eta!. 2006. The allosteric properties of hemoglobin:
Bubu, B, J. Weissman, and A. Horwich. 2006. Yl.olecular insights from natural and sire directed mutants. Curr. Prot. Pep. Sci.
chaperones and protem qualit} control. Ce/1125(3):443-451. 7(1):17-45.
Cohen, F. E. 1999. Protein misfolding and prion disea,es.}. Bochtler, .\-1., et al. 1999. The proteasome. Ann. Reu. Bzophys.
Mol. Bwl. 293:313-320. Biomol. Struct. 28:295-3 t 7.
Coulson, A. f., and J. .\1oult. 2002. A unifold, mesofold, and Burack, \XI. R., and A. S. Shaw. 2000. Signal transduction:
'uperfold model of protein fold use. Protems 46:61- ~1. hangmg on a scaffold. Curr. Opin. Cell Bioi. 12:211-216.
Daggett, V, and A. R. Fershr. 2003. ls there a unifying mecha- Gallastegui, N., and M. Groll. 2010. The 26S proreasome:
nism for protein folding? Trends Biochem. Sci. 28( 1): 18-25. assembly and function of a destructive machine. Trends Biochem.
Dobson, C. M. 1999. Prorem misfolding, evolution, and disease. Sci. 35( 11 ):634-642.
Trends Bzochem. Scz. 24:329-332. Glen, R., et al. 2008. Regulatory monoubiquitination of
Gimona, M. 2006. Protem linguistics-a grammar for modular phosphoenolpyruvate carboxylase in germinating castor oil seeds
protein assembly? Nat. Ret<. Mol. Cell Bzol. 7(1):68-73. }. Bml. Chem. 283:29650-2965-.

112 CHAPTER 3 • Protein Structure and Function

 Glickman, .\1. H., and A. Ciechanover. 2002. The ubiquitin- Zolk, 0., C. Schenke, and A. Sarikas. 2006. The ubiquirin-
proteasome proteolytic pathway: d estruction for the sake of proreasome system: focus on the heart. Cardtovasc. Res. 70(3):410-
construction. Physio/. Reu. 82(2):373-428. 421.
Goldberg, A. L. 2003. Protem degradation and protection
against misfolded or damaged pro teins. Nature 426:895- 899. Purifying, Detecting, and Characterizing Proteins
Goldberg, A. L, S. J. Elledge, and J. W. Harper. 200 I. The cellular Domon, B., and R. Aehersold. 2006. Mass spectrometry and
chamber of doom. Sci. Am. 284(1):68-73. protein analy&is. Science 312(5771 ):212-217.
Groll, M., and R. Huber. 2005. Purification, crystallization, and Encarnacion, S., et at. 2005. Comparative proreomics using 2-D
x-ray analysis of the yeast 20S proreasome. Meth. Enzymol. gel electrophoresis and mass spectrometry as tools to dissect
398:329-336. snmulons and regulons in bacteria with sequenced or parnally
Hallmg, D. B, P. Aracena-Parks, and S. L. Hamilton. 2006. sequenced genomes. Bwl. Proc. On/me 7:117-135.
Regulation of voltage-gated Ca 2 I channels by calmodulin. Set STKF Hames, B. D. A Practtcal Approach. Oxford University Press.
2005 (3l5):rel5. A methods series that descnbes protein purification methods and
Horovitz, A., et al. 2001. Review: allostery in chaperonins. assays.
}. Struc. Bioi. 135:104-114. O'Connell, M. R., R. GamsJaeger, and J.P. Mackay. 2009. The
Huang, H., er al. 2010. K33-linked polyubiqumnarion ofT cell structural analysis of protein-protein interactions by J\'~1R spectros-
receptor-~ regulates proteolysis-independent T cell signaling. copy. Proteomtcs 9(23):5224-5232.
Immunity. 33(1 ):60-70. Parton, W. F. 2002. Derecnon technologies m proreome analysis.
Katz, E.]., M. Isasa, and B. Crosas. 2010. A new map to ]. Chromatogr. B. Ana/yt. Techno/. Biomed. Life Set. 771( 1-2):3-31.
understand deubiquitination. Biochem. Soc. Trans. 38(pr. 1):21-28. White, I. R., et at. 2004. A statistical comparison of silver and
Kern, D., and E. R. Zuiderweg. 2003. The role of dynamics m SYPRO Ruby staining for proreomic analysis. Flectrophorests
allosteric regulation. Curr. Opin. Struc. Btol. 13(6):748-757. 25 ( 17):3048-3054.
Kisselev, A. F., A. Callard, and A. L. Goldberg. 2006. Impor-
Proteomics
tance of the different proteolytic sires of the proreasome and the
efficacy of inhibitors varies with the protein substrate.}. Bwl. Calvo, S. 1::.., and V. K. Moorha. 2010. The mirochondnal
Chem. 281(13):8582-8590. proteome and human disease. Annu. Rev. Genomics Hum. Genet.
Lane, K. T., and L. S. Beese. 2006. Thematic review series: lipid 11:25-34.
posrrranslational modifications. Structural biology of protein Foster, L.]., et at. 2006. A mammalian organelle map hy
farnesyltransferase and geranylgeranyltransferase type I.]. Lipid protein correlation profiling. Cell 125(1 ): 187-199.
Res. 47(4 ):681-699. Fu, Q., and J. E. Van Eyk. 2006. Proreomics and heart disease:
Lim, W. A. 2002. The modular logic of signaling proteins: identifying biomarkers of clinical utility. Expert Rev. Proteomics
building allosteric switches from simple binding domains. Curr. 3(2):237-249.
Opin. Struc. Bioi. 12:61-68. Gavin, A. C., er al. 2006. Proreome survey reveals modulanty
Marrin, C., and Y. Zhang. 2005. The diverse funcnons of of rhe yeast cell machinery. Nature 440(7084):63 1-636.
histone lysine methylation. Nat. Ret•. Mol. Cell Bw/. 6(11):838-849. Keltic, J. F., er at. 2010. The emerging process of Top Down mass
Rabl,]., er at. 2008. Mechanism of gate opening in the 205 spectrometry for protein analysis: biomarkers, protein-therapeutics,
protcasome by the proteasomal ATPases. Mol. Cell. 30(3):360-368. and achie1ing high throughput. Mol. Biosyst. 6(9):1532-1539.
Rechsteiner, M., and C. P. Hill. 2005. Mobilizing the proteo- Kislinger, T., er al. 2006. Global survey of organ and organelle
lytic machine: cell biological roles of proteasome activators and protein expression in mouse: combined proreomic and transcrip-
inhibitors. Trends Cell Bioi. 15( l ):27-33 . romic profiling. Cell 125(1 ): 173-186.
Sawyer, T. K. 1 er al. 2005. Protein phosphorylation and signal Kolker, E., R. Higdon, and J. ~1. Hogan. 2006. Protein
transduction modulation: chemistry perspectives for small-molecule identification and expression analysis using mass spectrometry.
drug discovery. M ed. Chern. 1(3) :293-3 19. Trends Microbial. 14(5):229-235.
Sowa, M. E., er al. 2009. Defining the human deubiquirinating Krogan, N. ]., er at. 2006. Global la ndscape of protein com-
plexes in the yeast Saccharomyces cerevisiae. Nature
·. enzyme interaction landscape. Cell 138(2):389-403.
440(7084):637-64 3.
Xm, Z., and D. R. Storm. 2005. The role of calmodulin as a
stgnal mregraror for synaptic plasnciry. Nat. Rev. Neurosci. Ong, S. E., and M. Mann. 2005. Mass spectrometry-based
6(4):267-276. proteomics turns quantitative. Nat. Chern. Btol. 1(5):252-262.
I

Yap, K. L., er al. ·1999. Diversity of conformational stares and Rifai , N., M.A. Gillette, and S. A. Carr. 2006. Protein bio-
changes within the EF-hand protein superfamily. Protems 37:499-507. marker discovery and validation: the long and uncertain path ro
Zeng, W., et al. 20 I 0. Reconstitution of rhe RIG-I pathway clinical utility. Nat. Biotech. 24(8):971-983.
reveals a signaling role of unanchored polyubiquinn chains in innate Walther, T. C., and .\1. Mann. 2010. ·"lass spectrometry-based
immunity. Cell 141(2):315-330. proreomics m cell biology.]. Cell Bw/. 190(4):491-500.
Zhou, P. 2006. REGgamma: a shortcut to destruction. Cell Zhou, M., and C. V. Robinson. When proreomics meers
124(2):256-257. structural biology. Trends Biochem. Sci. 35:522-539.

References 113
-·
 CHAPTER

Basic Molecular Genetic

Mechanisms

Colored t ransmission electron micrograph of one ribosomal RNA

transcription unit from a Xenopus oocyte. Transcription proceeds
from left to right, with nascent ribosomal ribonucleoprotein complexes
(rRNPs) growing in length as each successive RNA polymerase I molecule
moves along the DNA template at the center. In this preparation each
rRNP is oriented either above or below the central strand of DNA being
transcribed, so that the overall shape is similar to a feather. In the
nucleolus of a living cell, the nascent rRNPs extend in all directions, like
a bottle brush. [Professor Oscar L. Miller/Science Photo Library.]

he extraordinary versatility of proteins as molecular ma- Deoxyribonucleic acid (DNA) is an informational mole-

T chines and switches, cellular catalysts, and components

of cellular structures was described in Chapter 3. In this
chapter we consider how proteins are made, as well as other
cule that contains in the sequence of its nucleotides the infor-
mation required to build all the proteins of an organism, and
hence the cells and tissues of that organism. It is ideally
cellular processes that are critical for the survival of an organ- suited to perform this function on a molecular level. Chemi-
ism and its descendants. Our focus will be on the vital mole- cally, it is extraordinarily stable under most terrestrial condi-
cules known as nucleic acids, and how they ultimately are tions, as exemplified by the ability to recover DNA sequence
responsible for governing all cellular function. As introduced in from bones and tissues that are tens of thousands of years
Chapter 2, nucleic acids arc linear polymers of four types of old. Because of this, and because of repair mechanisms that
nucleotides (see Figures 2-] 3, 2-] 6, and 2-17). These macro- operate in living cells, the long polymers that make up a
9
molecules (1) contain in the precise sequence of their nucleo- DNA molecule can be up to 10 nucleotides long. Virtually
tides the information for determining the amino acid sequence all the information required for the development of a fertil-
and hence the strucrure and function of all the proteins of a cell, ized human egg into an adult made of trillions of cells with
(2) are critical functional components of the cellular macromo- specialized functions can be stored in the sequence of the four
lecular factories that select and align amino acids in the correct types of nucleotides that makes up the ""'3 X 109 base pairs
order as a polypeptide cha~n is being synthesized, (3 ) catalyze a in the human genome. Because of the principles of base pair-
number of fundamental chemical reactions in cells, including ing discussed in the following, the information is readily
formation of peptide bonds between amino acids during pro- copied with an error rate of < J in 109 nucleotides per gen-
tein synthesis, and (4) regulate the expression of genes. eration. The exact replication of this information in any

OUTLINE

4.1 St ructure of Nucleic Acids 117 4.5 DNA Replication 145

4.2 Transcription of Protein-Coding Genes 4.6 DNA Repair and Recombination 151
and Formation of Funct ional mRNA 124
4.7 Viruses: Parasites of the Cellular Genetic System 160
4.3 The Decoding of mRNA by tRNAs 131

4.4 Stepwise Synthesis of Proteins on Ribosomes 136

species assures its genetic continuity from generation to gen- normal development of organisms as complex as prokary-
eration and is critical to the normal development of an indi- otes and eukaryotes. This is achieved by chemical processes
vidual. DNA fulfills these functions so well that it is the that operate with extraordinary accuracy coupled with mul-
vessel for genetic information in all forms of life known (ex- tiple layers of checkpoint or surveillance mechanisms that
cluding RNA viruses, which are limited to extremely short test whether critical steps in these processes have occurred •,
genomes because of the relative instability of RNA com- correctly before the next step is initiated. The highly regu-
pared to DNA, as we will see}. The discovery that virtually lated expression of genes necessary for the development of a
all forms of life use DNA to encode their genetic informa- multicellular organism requires integrating information from
tion, and also use nearly the identical genetk code, implies signals sent by distant cells in the developing organism, as
that all forms of li fe descended from a common ancestor well as from neighboring cells, and an intrinsic developmen-
based on the storage of information in nucleic acid sequence. tal program determined by earlier steps in embryogenesis
This information is accessed and replicated by specific base taken by that cell's progenitors. All of this regulation is de-
pairing between nucleotides. The information stored in DNA pendent on control sequences in the DNA that function with
is arranged in hereditary units, known as genes, that control proteins called transcription factors to coordinate the expres-
identifiable traits of an organism. In the process of transcrip- sion of every gene. RNA sequences we discuss in Chapter 8
tion, the information stored in DNA is copied into ribonu- that regulate RNA processing and translation also are en-
cleic acid (RNA}, which has three distinct roles in protein coded in DNA originally. Nucleic acids function as the
synthesis. "brains and central nervous system" of the cell, while pro-
Portions of the DNA nucleotide sequence are copied into teins carry out the functions they specify.
messenger RNA (mRNA} molecules that direct the synthesis In this chapter, we first review the structures and proper-
of a specific protein. The nucleotide sequence of an mRNA ties of DNA and RNA, and explore how the different charac-
molecule contains information that specifies the correct teristics of each type of nucleic acid make them suited for their
order of amino acids during the synthesis of a protein. The respective functions in the cell. In the next several sections we
remarkably accurate, stepwise assembly of amino acids into discuss the basic processes summarized in Figure 4-1: tran-
proteins occurs by translation of mRNA. In this process, the scription of DNA into RNA precursors, processing of these
nucleotide sequence of an mRNA molecule is "read" by a precursors to make functional RNA molecules, translation of
second type of RNA called transfer RNA (tRNA} with the mRNAs into proteins, and the replication of DNA. Proteins
aid of a third type of RNA, ribosomal RNA (rRNA}, and regulate cell structure and most of the biochemical reactions
their associated proteins. As the correct amino acids are in cells, so we first consider how the amino acid sequences of
brought into sequence by tRNAs, they are linked by peptide proteins, which determines their three-dimensional structures
bonds to make proteins. RNA synthesis is ca lled transcrip- and hence their functions, is encoded in DNA and translated.
tion because the nucleotide sequence "language" of DNA is After outlining functions of mRNA, tRNA, and rRNA in pro-
precisely copied, or transcribed, into the nucleotide sequence tein synthesis, we present a detailed description of the compo-
of an RNA mol<:;cule. Protein synthesis is referred to as trans- nents and biochemical steps in translation. Understanding
lation because the nucleotide sequence "language" of DNA these processes gives us a deep appreciation of the need to
and RNA is translated into the amino acid sequence "lan- copy the nucleotide sequence of DNA precisely. Consequently,
guage" of proteins. we next consider the molecular problems involved in DNA
Discovery of the structure of DNA in 1953 and subse- replication and the complex cellular machinery for ensuring
quent elucidation of how DNA directs synthesis of RNA, accurate copying of the genetic material. Along the way, we
which then directs assembly of proteins-the so-called cen- compare these processes in prokaryotes and eukaryotes. The
tral dogma-were monumental achievements marking the next section describes how damage to DNA is repaired, and
early days of molecular biology. However, the simplified how regions of different DNA molecules are exchanged in the
representation of the central dogma as DNA ~ RNA ~ process of recombination to generate new combinations of
protein does not reflect the role of proteins in the synthesis traits in the individual organisms of a species. The final sec-
of nucleic acids. Moreover, as discussed here for bacteria tion of the chapter presents basic information about viruses,
and in later chapters for eukaryotes, proteins are largely re- parasites that exploit the cellular machinery for DNA replica-
sponsible for regulating gene expression, the entire process tion, transcription, and protein synthesis. In addition to being
whereby the information encoded in DNA is decoded into significant pathogens, viruses are important model organisms
proteins in the correct cells at the correct times in develop- for studying these cellular mechanisms of macromolecular
ment. As a consequence, hemoglobin is expressed only in synthesis and other cellular processes. Viruses have relatively
cells in the bone marrow (reticulocytes) destined to develop simple structures compared to cells, and small genomes that
into circulating red blood cells (erythrocytes), and develop- made them tractable for historic early studies of these basic
ing neurons make the proper synapses (connections) with cellular processes. Viruses continue to teach important lessons
10 11 other developing neurons in the human brain. The fun- in molecular cell biology today and have been adapted as ex-
damental molecular genetic processes of DNA replication, perimental tools for introducing any desired genes into cells,
transcription, and translation must be carried out with tools that are currently being tested for their effectiveness in
extraordinary fidelity, speed, and accurate regulation for the human gene therapy.

116 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 RNA
virus

0 · /
("'\ .....___AfJ RNA
~ processing

m RNA
Protein
Amino acids
Translation
• factors

FIGURE 4 - 1 Overview of four basic molecular genetic processes. cytoplasm. During translation (!D), the four-base code of the mRNA is
In this chapter we cover the three processes that lead to production decoded into the 20-amino acid language of proteins. Ribosomes, the
of prot eins (0-IJ) and the process for replicating DNA (fl ). Because macromolecular machines that translate the mRNA code, are composed
viruses utilize host-cell machinery, they have been important models for of two subunits assembled in the nucleolus from ribosomal RNAs
studying these processes. During transcription of a protein-coding gene (rRNAs) and multiple proteins {left). After transport to the cytoplasm,
by RNA polymerase (0 ), the four-base DNA code specifying the amino ribosomal subunits associate with an mRNA and carry out protein
acid sequence of a protein is copied, or transcribed, into a precursor synthesis with the help of transfer RNAs (tRNAs) and various translation
messenger RNA (pre-mRNA) by the polymerization of ribonucleoside factors. During DNA replication (fl ), which occurs only in cells preparing
triphosphate monqmers (rNTPs). Removal of noncoding sequences and to divide, deoxyribonucleoside triphosphate monomers (dNTPs) are
other modifications to the pre-mRNA (fJ), collectively known as RNA polymerized to yield two identical copies of each chromosomal DNA
processing, produce a functional mRNA, which is transported to the molecule. Each daughter cell receives one of the identical copies.

4.1 Structure of Nucleic Acids A Nucleic Acid Strand Is a Linear Polymer

DNA and RNA are chemically very similar. The primary with End-to-End Directionality
structures of both are linear polymers composed of mono- In all organisms, DNA and R1'1A are each comprised of only
mers called nucleotides. Both function primarily as informa- four different nucleotides. Recall from Chapter 2 that all nucle-
tional molecules, carrying information in the exact sequence otides consist of an organic base linked to a five-carbon sugar
of their nucleotides. Cellular RNAs range in length from less that has a phosphate group attached to carbon 5. In RNA, the
than one hundred to many thousands of nucleotides. Cellu- sugar is ribose; in DNA, deoxyribose (see Figure 2-16). The
lar DNA molecules can be as long as several hundred million nucleotides used in synthesis of DNA and RNA contain five
nucleotides. These large DNA units in association with pro- different bases. The bases adenine (A) and guanine (G) are pu-
teins can be stained with dyes and visualized in the light mi- rines, which contain a pair of fused rings; the bases cytosine
croscope as chromosomes, so named because of their (C), thymine (T), and uracil (U) are pyrimidines, which contain
stainability. Though chemically similar, DNA and RNA ex- a single ring (see Figure 2-17). Three of these bases-A, G, and
hibit some very important differences. For example, RNA C-are found in both DNA and RNA; however, T is found
can also function as a catalytic molecule. As we will see, it is only in DNA, and U only in RNA. (Note that the single-letter
the different and unique properties of DNA and RNA that abbreviations for these bases are also commonly used to denote
makes them each suited for their specific roles in the cell. the entire nucleotides in nucleic acid polymers.)

4.1 Structure of Nucleic Acids 117

(a) (b) The linear sequence of nucleotides linked by phosphodi-
0
ester bonds constitutes the primary structure of nucleic acids.
5'end 0 - P= O C A G Like polypeptides, polynucleotides can twist and fold into
three-dimensional conformations stabilized by noncovalent
0
13' 13' 13' bonds. Although the primary structures of DNA and RNA
H 2C~s
0 C p~OH are generally similar, their three-dimensional conformations
H H 5' 5' 5'
are quite different. These structural differences are critical to
H H the different functions of the two types of nucleic acids.
3'
0 H
Phospho- Native DNA Is a Double Helix of Complementary
diester 0-P=O 5' C-A-G 3'
bond I Antiparallel Strands
0
The modern era of molecular biology began in 1953 when
2
H Co
0 sA James D. Watson and Francis H. C. Crick proposed that
DNA has a double-helical structure. Their proposal was
H H based on analysis of x-ray diffraction patterns of DNA fibers
3
0 H generated by Rosalind Franklin and Maurice Wilkins, which
Phospho· I showed that the structure was helical, and analyses of the
diester 0-P=O base composition of DNA from multiple organisms by Erwin
bond I
0 Chargaff and colleagues. Chargaff's studies revealed that
while the base composition (percent of A, T, G, and C) varies
H,~ greatly between distantly related organisms, in all organisms
the percent of A always equals the percent ofT, and the per-
cent of G always equals the percent of C. Based on these
3' end OH H discoveries and the structures of the four nucleotides, Watson
FIGURE 4-2 Chemical directionality of a nucleic acid strand. and Crick performed careful molecular model building, pro-
Shown here are alternative representations of a single strand of DNA posing a double helix, with A always hydrogen-bonded toT
containing only three bases: cytosine (C), adenine (A), and guanine (G). and G always hydrogen-bonded to C at the axis of the dou-
(a) The chemical structure shows a hydroxyl group at the 3' end and a ble helix. The Watson and Crick model proved correct and
phosphate group at the 5' end. Note also that two phosphoester paved the way for our modern understanding of how DNA
bonds link adjacent nucleotides; this two-bond linkage commonly is functions as the genetic material. Today, our most accurate
referred to as a phosphodiester bond. (b) In the "stick" diagram (top), the models for DNA structure come from high-resolution x-ray
sugars are indicated as vertical lines and the phosphodiester bonds as diffraction studies of crystals of DNA, made possible by the
slanting lines; the bases are denoted by their single-letter abbrevia· chemical synthesis of large amounts of short DNA mole-
tions. In the simple"st representation (bottom), only the bases are cules of uniform length and sequence that are amenable to
indicated. By convention, a polynucleotide sequence is always written crystallization (Figure 4-3a).
in the 5'~3' direction (left to right) unless otherwise indicated.
DNA consists of two associated polynucleotide strands
that wind together to form a double helix. The two sugar-
phosphate backbones are on the outside of the double helix,
and the bases project into the interior. The adjoining bases in
A single nucleic acid strand has a backbone composed of each strand stack on top of one another in parallel planes
repeating penrose-phosphate units from which the purine (Figure 4-3a). The orientation of the two strands is antipar·
and pyrimidine bases extend as side groups. Like a polypep- a/lei; that is, their 5'~3 ' directions are opposite. The strands
tide, a nucleic acid strand has an end-to-end chemical orien- are held in precise register by formation of base pairs between
tation: the .5' end has a hydroxyl or phosphate group on the 5' the two strands: A is paired with T through two hydrogen
carbon of its terminal sugar; the 3' end usually has a hydroxyl bonds; G is paired with C through three hydrogen bonds
group on the 3' carbon of its terminal sugar (Figure 4-2). This (Figure 4-3b). This base-pair complementarity is a conse-
directionality, plus the fact that synthesis proceeds 5' to 3', quence of the size, shape, and chemical composition of the
has given rise to the convention that polynucleotide sequences bases. The presence of thousands of such hydrogen bonds in
are written and read in the 5'~3' direction (from left to a DNA molecule contributes greatly to the stability of the
right); for example, the sequence AUG is assumed to be douhlt> helix. Hydrophobic and van der Waals interactions
(5')AUG(3'). As we will see, the 5'~3 ' directionality of a between the stacked adjacent base pairs further stabilize the
nucleic acid strand is an important property of the molecule. double-helical structure.
The chemical linkage between adjacent nucleotides, com- In natural DNA, A always hydrogen bonds with T, and
monly called a phosphodiester bond, actually consists of two G with C, forming A·T and G·C base pairs as shown in Fig-
phosphoester bonds, one on the 5' side of the phosphate and ure 4-3b. These associations, always between a larger purine
another on the 3' side. and a smaller pyrimidine, are often called Watson-Crick

118 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 {a) {b) FIGURE 4 -3 The DNA double helix. (a) Space-filling
model of B DNA, the most common form of DNA in cells.
The bases {light shades) project inward from the
sugar-phosphate backbones (dark red and blue) of each
strand, but their edges are accessible through major and
minor grooves. Arrows indicate the 5'~3' direction of
each strand. Hydrogen bonds between the bases are in
the center of the structure. The major and minor grooves
·. are lined by potential hydrogen bond donors and
acceptors {highlighted in yellow). {b) Chemical structure
of DNA double helix. This extended schematic shows the
two sugar-phosphate backbones and hydrogen bonding
between the Watson-Crick base pairs, A·T and G·C.
Major
groove [Part (a) adapted from R. Winget al., 1980, Nature 287:755.
Part (b) adapted from R. E. Dickerson, 1983, Sci. Am. 249:94.)

base pairs. Two polynucleotide strands, or regions thereof, Under laboratory conditions in which most of the water is
in which all the nucleotides form such base pairs are said to removed from DNA, the crystallographic structure of DNA
be complementary. However, in theory and in synthetic DNAs, changes to the A form, which is wider and shorter than B-
other base pairs can form. For example, guanine (a purine) form DNA, with a wider and deeper major groove and a more
cou ld theoreticall y form hydrogen bonds with thymine (a narrow and shallow minor groove (Figure 4-4). RNA-DNA
pyrimidine), causing only a minor distortion in the helix. and RNA-RNA helices exist in this form in cells and in vitro.
The space available in the helix also would allow pairing
between the two pyrimidines cytosine and thymine. Although (b) A DNA
{a) B DNA
the nonstandard G·T and C·T base pairs are normally not
found in DNA, G·U base pairs are quite common in double-
helical regions that form within otherwise single-stranded
RNA. Nonstandard base pairs do not occur naturally in du-
plex DNA because the DNA copying enzyme, which is de-
scribed later in th is chapter, does not permit them.
Most DNA in cells is a right-handed helix. The x-ray
diffraction pattern of DNA indicates that the stacked bases
arc regularly spaced 0.34 nm apart along the helix axis.
The helix makes a complete turn every 3.4 to 3.6 nm, de-
pending on the sequence; thus there are about 10-10.5
base pairs per turn. This is referred to as the B form of
DNA, the normal form present in most DNA stretches in
cells. On the outside of B-form DNA, the spaces between
the intertwined strands form two helical grooves of differ-
FIGURE 4·4 Comparison of A- and B-Form DNA. The sugar-
ent widths described as the rnaior groove and the minor phosphate backbones of the two strands, which are on the outside of
groove (see Figure 4-3a). As a consequence, the atoms on both structures, are shown in red and blue; the bases (lighter shades)
the edges of each base within these grooves are accessible are oriented inward. (a) The B form of DNA has = 10.5 base pairs per
from outside the helix, forming two types of binding sur- helical turn. Adjacent stacked base pairs are 0.34 nm apart. (b) The
faces. DNA-binding proteins can "read" the sequence of more compact A form of DNA has 11 base pairs per turn with a much
bases in duplex DNA by contacting atoms in either the deeper major groove and much more shallow minor groove than
major or the minor grooves. Bform DNA.

4.1 Structure of Nucl e1c Acids 119

 Important modifications in the structure of standard B- TATA box-binding protein
form DNA come about as a resu lt of protein binding to spe-
cific DNA sequences. Although the multitude of hydrogen
and hydrophobic bonds between the bases provides stability
to DNA, the double helix is flexible about its long axis. Unlike
the ex helix in proteins (see Figure 3-4 ), there are no hydrogen
bonds parallel to the axis of the DNA helix. This property al-
lows DNA to bend when comp lexed with a DNA-binding
protem (hgure 4-5). Bending of DNA is critical to the dense
packing of DNA in chromatin, the protein-DNA complex in
which nuclear DNA occurs in eukaryotic cells (Chapter 6).
Why did DNA evolve to be the carrier of genetic informa-
tion in cells as opposed to RNA? The hydrogen at the 2' posi-
tion in the deoxyribose of DNA makes it a far more stable
molecule than RNA, which instead has a hydroxyl group at
the 2' position of ribose (see Figure 2-16 ). The 2'-hydroxyl
FIGURE 4-5 Protein interaction can bend DNA. The conserved
groups in RNA participate in the slow, OH -catalyzed hy-
( -terminal domain of the TATA box-binding protein (TBP) binds to the
drolysis of phosphodiester bonds at neutral pH (Figure 4-6).
minor groove of specific DNA sequences rich in A and T, untwisting
The absence of 2'-hydroxyl groups in DNA prevents this pro-
and sharply bending the double helix. Transcription of most eukaryotic
cess. Therefore, the presence of deoxyribose in DNA makes it genes requires participation of TBP. [Adapted from D. B. Nikolov and S. K.
a more stable molecule-a characteristic that is critical to its Burley, 1997, Proc. Nat'/ Acad. Sci. USA 9 4:15.)
function in the long-term storage of genetic information.

DNA Can Undergo Reversible Strand Separation we describe the cellular mechanisms that separate and subse-
During replication and transcription of DNA, the strands of quently reassociate DNA strands during replication and tran-
the double helix must separate to allow the internal edges of scription. Here we discuss fundamental factors that influence
the bases to pair with the bases of the nucleotides being po- the separation and reassociation of DNA strands. These
lymerized into new polynucleotide chains. In later sections, properties of DNA were elucidated by in vitro experiments.

I I I
0-P = O - o - P= O -o-P=O
0 0 0
0-P=O I I I
I

~t" ~"'" ~'""

0 2', 3 ' cyclic H20
monophosphate
or

~"""
derivative
H H H H
o, /0 0 OH OH 0
p~ 0 - P=O 0-P=O
0 0-H OH 0 0 I
OH OH
+
I
0 - P- O 3' m onophosphat e 2' monophosphat e
I
0 OH

-o-P=O 0 P= O
I I
0 0

FIGURE 4 -6 Base-catalyzed hydrolysis of RNA. The 2 '-hydroxyl phosphodiester bond hydrolysis cannot occur in DNA, which lacks
group in RNA can act as a nucleophile, attacking the phosphodiester 2' -hydroxyl groups. [Adapted from Nelson et al., Lehninger Principles of
bond. The 2',3' cyclic monophosphate derivative is further hydrolyzed Biochemistry, 4th ed., W. H. Freeman and Company.]
to a mixture of 2' and 3' monophosphates. This mechanism of

120 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 (a) (b)

1.0 100
....
.<::
-~ ~ 80
E
c:
"'a.
6~
<;

-
(!) 60
N
0.75 0
0 Q)
c: Ol
40
0 ~
·a c:
Q)
<.l
0 Qi
V)
.D c.. 20
<(

0.5 0
75 80 85 90 70 80 90 100 110
Temperature (•C) Tm (°C)
EXPERIMENTAL FIGURE 4 · 7 G·C content of DNA affects which half the bases in a double-stranded DNA sample have denatured
melting temperature. The temperature at which DNA denatures is denoted Tm (for "temperature of melting"). Light absorption by
increases with the proportion of G·C pairs. (a) Melting of doubled- single-stranded DNA changes much less as the temperature is
stranded DNA can be monitored by the absorption of ultraviolet light increased. (b) The Tm is a function of the G·C content of the DNA; the
at 260 nm. As regions of double-stranded DNA unpair, the absorption higher the G-t C percentage, the greater the Tm·
of light by those regions increases almost twofold. The temperature at

The unwinding and separation of DNA strands, referred again repelling each other because of the similar charge. In
to as denaturation, or "melting," can be induced experimen- cells, pH and temperature are, for the most parr, maintamed.
tally by increasing the temperature of a solution of DNA. As These features of DNA separation are most useful for ma-
the thermal energy increases, the resulting increase in mo- nipulating DNA in a laboratory setting.
lecular motion eventually breaks the hydrogen bonds and The single-stranded DNA molecules that result from
other forces that stabilize the double helix; the strands then denaturation form random coils without an organized
separate, driven apart by the electrostatic repulsion of the structu re. Lowering the temperature, increasing the ion
negatively charged deoxyribose-phosphate backbone of each concentration, or neutralizing the pH causes the two com-
strand. Near the denaturation temperature, a small increase plementary strands to reassociate into a perfect double helix.
in temperature causes a rapid, nearly simultaneous loss of The extent of such renaturation is dependent on time, the
the multiple weak interactions holding the strands together DNA concentration, and the ionic concentration. Two DNA
along the entire length of the DNA molecules. Because the strands that are not related in sequence will remain as ran-
stacked base pairs in duplex DNA absorb less ultraviolet dom coils and will not renature; most importantly, they will
(UV) light than the unstacked bases in single-stranded DNA, not inhibit complementary DNA partner strands from find-
this leads to an abrupt increase in the absorption of UV light, ing each other and renaturing. Denaturation and renatur-
a phenomenon known as hyperchromicity (Figure 4-7a). ation of DNA are the basis of nucleic acid hybridization, a
The melting temperature (TmJ at which DNA strands will powerful technique used to study the relatedness of two
separate depends on several factors. Molecules that contain DNA samples and to detect and isolate specific DNA mole-
a grea ter proportion of G·C pairs require higher tempera- cules in a mixture containing numerous different DNA se-
tures to denature because the three hydrogen bonds in G·C quences (see Figure 5-16).
pairs make these base pairs more stable than A·T pairs,
which have only two hydrogen bonds. Indeed, the percent-
age of G·C base pairs in a DNA sample can be estimated Torsional Stress in DNA Is Relieved by Enzymes
from its Tm (Figure 4-7b). The ion concentration also influ- Many bacterial genomic DNAs and many viral DNAs are
ences the Tm because the negatively charged phosphate circular molecules. Circular DNA molecules also occur in
groups in the two strands are shielded by positively charged mitochondria, which are present in almost all eukaryotic
ions. When the ion concentration is low, this shielding is cells, and in chloroplasts, which are present in plants and
decreased, thus increasing the repulsive forces between the some unicellular eukaryotcs.
strands and reducing the Trn. Agents that destabilize hydro- Each of the two strands in a circular DNA molecule
gen bonds, such as formamide or urea, also lower the T m· forms a closed structure without free ends. Localized un-
Finally, extremes of pH denature DNA at low temperature. winding of a circular DNA molecule, which occurs during
At low (acid) pH, the bases become protonated and thus DNA replication, induces torsional stress into the remaining
, positively charged, repelling each other. At high (alkaline) portion of the molecule because the ends of the strands are
pH, the bases lose protons and become negatively charged, not free to rotate. As a result, the DNA molecule twists back

4.1 Structure of Nucleic Acids 121

 :XPERIMENTAL FIGURE 4-8 (a) Supercoiled (b) Relaxed circle
Topoisomerase I relieves torsional stress on
DNA. (a) Electron micrograph of SV40 viral DNA.
When the circular DNA of the SV40 virus is isolated
and separated from its associated prot ein, the
DNA duplex is underwound and assumes the
supercoiled configuration. (b) If a supercoiled DNA
is nicked (i.e., one strand cleaved), the strands can
rewind, leading to loss of a supercoil. Topoisomer-
ase I catalyzes this reaction and also reseals the
broken ends. All the supercoils in isolated SV40
DNA can be removed by the sequential action
of this enzyme, producing the relaxed-circle
conformation. For clarity, the shapes of the
molecules at the bottom have been simplified.

on itself, like a twisted rubber band, forming supercoils (Fig- The presence of thymine rather than uracil in DNA is impor-
ure 4-8a). In other words, when part of the DNA helix is tant to the long-term stability of DNA because of its func-
underwound, left-handed supercoils are introduced into the tion in DNA repair (see Section 4. 7). As noted earlier, the
circular DNA molecule, as in Figure 4-Sa. Bacterial and eu- hydroxyl group on the 2' C of ribose makes RNA more
karyotic cells, however, contain topoisomerase J, which can chemically labile than DNA. As a result of this lability, RNA
relieve any torsional stress that develops in cellular DNA is cleaved into mononucleotides by alka line solution (see
molecules during replication or other processes. This enzyme Figure 4-6), whereas DNA is not. The 2' -C hydroxyl of
binds to DNA at random sites and breaks a phosphodiester RNA a lso provides a chemically reactive group that takes
bond in one strand. Such a one-strand break in DNA is part in RNA-mediated catalysis. Like DNA, RNA is a long
called a nrck. The broken end then winds around the uncut polynucleotide that can be double-stranded or single-
strand, leading to loss of supercoils (Figure 4-8b). Finally, stranded, linear or circular. It can also participate in a hybrid
the same enzyme joins (ligates) the two ends of the broken helix composed of one RNA strand and one DNA strand. As
strand. Another type of enzyme, topoisomerase II, makes discussed above, RNA-RNA and RNA-DNA double helices
breaks in both strands of a double-stranded DNA and then have a compact conformation like the A form of DNA (see
religates them. As a result, topoisomerase II can both relieve Figure 4-4b).
torsional stress and link together two circular DNA mole- Unlike DNA, which exists primarily as a very long double
cules as in the links of a chain. helix, most cellular RNAs are single-stranded and exhibit a
Although eukaryotic nuclear DNA is linear, long loops variety of conformations (Figure 4-9). Differences in the
of DNA are fixed in place within chromosomes (Chapter 6). sizes and conformations of the various types of RNA permit
Thus torsional stress and the consequent formation of super- them ro carry out specific functions in a cell. The simplest
coils also could occur during replication of nuclear DNA. As secondary structures in single-stranded RNAs are formed by
in bacterial cells, abundant topoisomerase I in eukaryotic pairing of complemen tary bases. "Hairpins" are formed by
nuclei relieves any torsional stress in nuclear DNA that pairing of bases within =5-10 nucleotides of each other, and
would develop in the absence of this enzyme. "stem-loops" by pairing of bases that are separated by >10
to several hundred nucleotidcs. These simple folds can coop-
crate to form more complicateJ tertiary structures, one of
Different Types of RNA Exhibit Various
which is termed a "pseudoknot."
Conformations Related to Their Functions As discussed in detail later, tRNA molecules adopt a well-
The primary structure of RNA is generally similar ro that of defined three-dimensional architecture in solution that is cru-
DNA with two exceptions: the sugar component of RNA, cial in protein synthesis. Larger rRNA molecules also have
ribose, has a hydroxyl group at the 2' position (see Figure locally well-defined three-dimensional structures, with more
2-16b), and thymine in DNA is replaced by uracil in RNA. flexible linkers in between. Secondary and tertiary structures

122 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 (a) Secondary structure during formation of the majority of functional mRNA mol-
ecules in multicellular eukaryotes, and also occurs in single-
celled eukaryotes such as yeast, bacteria, and archaea.
Remarkably, some RNAs carry out self-splicing, with the
catalytic activity residing in the sequence that is removed.
The mechanisms of splicing and self-splicing are discussed in
stem region detail in Chapter 8. As noted later in this chapter, rRNA
plays a catalytic role in the formation of peptide bonds dur-
Stem-loop ing protein synthesis.
In this chapter, we focus on the functions of mRNA,
(b) Tertiary structure tRNA, and rRNA in gene expression. ln later chapters we
will encounter other RNAs, often associated with proteins,
C G C 3' that participate in other cell functions.
U CG
c UA
Loop u GC
1 Stem AU
u
2 CG
u UA
u UA
U UA A KEY CONCEPTS of Section 4.1
Loop GC A
2 UA A Structure of Nucleic Acids
CG C • Deoxyribonucleic acid (DNA), the genetic material, car-
GC
GC A ries information to specify the amino acid sequences of pro-
5'
5' GC A teins. It is transcribed into several types of ribonucleic acid
Pseudoknot cA
(RNA), including messenger RNA (mRNA), transfer RNA
FIGURE 4-9 RNA secondary and tertiary structures. (a) Stem-loops,
(tRNA), and ribosomal RNA (rRNA), which function in
hairpins, and other secondary structures can form by base pairing
protein synthesis (see Figure 4-1 ).
between distant complementary segments of an RNA molecule. In
stem-loops, the single-stranded loop between the base-paired helical • All DNAs and most RNAs are long, unbranched polymers of
stem may be hundreds or even thousands of nucleotides long, whereas nucleotides, which consist of a phosphorylated penrose linked
in hairpins, the short turn may contain as few as four nucleotides. to an organic base, either a purine or a pyrimidine.
(b) Pseudo knots, one type of RNA tertiary structure, are formed by
• The purines adenine (A) and guanine (G) and the pyrimi-
interaction of secondary loops through base pairing between
complementary bases. The structure shown forms the core domain of
dine cytosine (C) are present in both DNA and RNA. The
the human telomerase RNA. Left: Secondary-structure diagram with pyrimidine thymine (T) present in DNA is replaced by the
base-paired nucleotides in green and blue and single-stranded pyrimidine uracil (U) in RNA.
regions in red. Middle: Sequence of the telomerase RNA core domain, • Adjacent nucleotides in a polynucleotide are linked by
colored to correspohd to the secondary structure diagram at the left. phosphodiester bonds. The entire strand has a chemical di-
Right: Diagram of the telomerase core domain structure determined by rectionality with 5' and 3' ends (see Figure 4-2).
2D-NMR, showing bases-paired bases only and a tube for the sugar
phosphate backbone, colored to correspond to the diagrams at left. • Natural DNA (B DNA) contains two complementary an-
[Part (b) middle and right adapted from C. A. Theimer et al., 2005, Mol. Ce// 17:671.] tiparallel polynucleotide strands wound together into a regu-
lar right-handed double helix with the bases on the inside
and the two sugar-phosphate backbones on the outside (see
Figure 4-3). Base pairing between the strands and hydropho-
bic interactions benveen adjacent base pairs stacked perpen-
also have been recognized in mRNA, particularly near the dicular to the helix axis stabilize this native structure.
ends of molecules. C learly, then, RNA molecules are like pro-
• The bases in nucleic acids can interact via hydrogen bonds.
reins in that they have structured domains connected by less
structured, flexible stretches. The standard Watson-Crick base pairs are G·C, AT (in
DNA), and G·C, A·U (in RNA). Base pairing stabilizes the
The folded domains of RNA molecules not only are
structurally analogous to the o: helices and p strands found native three-dimensional structures of DNA and RNA.
in proteins, but in some cases also have catalytic capacities. • Binding of protein to DNA c::tn deform its helical structure,
Such catalync RNAs are called ribozymes. Although ribo- causing local bending or unwinding of the DNA molecule.
zymes usually are associated with proteins that stabilize the • Heat causes the DNA strands to separate (denature). The
ribozyme structure, it is the RNA that acts as a catalyst. Some melting temperature Tm of DNA increases with the percent-
ribozymes can catalyze splicing, a remarkable process in age of G·C base pairs. Under suitable conditions, separated
which an internal RNA sequence is cut and removed, and complementary nucleic acid strands will renature.
the two resulting chains then ligated. This process occurs

4.1 Structure of Nucleic Acids 123

 complementary RNA chain. Bases in the template DNA strand
Circular DNA molecules can be twisted on themselves, base-pair with complementary incoming rNTPs, which then
forming supercoi ls (see Figure 4-8). Enzymes called topoi- are joined in a polymerization reaction catalyzed by RNA
somerases can relieve torsional stress and remove supercoils polymerase. Polymerization involves a nucleophilic attack
from circular DNA molecules. Long linear DNA can also by the 3' oxygen in the growing RNA chain on the a phos-
experience torsional stress because long loops are fixed in phate of the next nucleotide precursor to be added, resulting
place within chromosomes. in formation of a phosphodiester bond and release of pyro-
Cellular RNAs are single-stranded polynucleotidcs, some phosphate (PP,). As a consequence of this mechanism, RNA
of which form well-defined secondary and tertiary structures molecules arc always synthesized in the 5 '--+3' direction
(see Figure 4-9). Some RNAs, called ribozymes, have cata- (Figure 4-10a).
lytic activity. The energetics of the polymerization reaction strongly
favors addition of ribonucleotides to the growi ng RNA chain
because the high-energy bond between the a and 13 phosphates
of rNTP monomers is replaced by the lower-energy phos-
phod iester bond between nucleotides. The equilibrium for
4.2 Transcription of Protein-Coding Genes the reaction is driven further toward chain elongation by
pyrophosphatase, an enzyme that" catalyzes cleavage of the
and Formation of Functional mRNA
released PP; into two molecules of inorganic phosphate. Like
The simplest definition of a gene is a " unit of DNA that con- the two strands in DNA, the template DNA strand and
tains the information to specify synthesis of a single polypep- the growing RNA strand that is base-paired to it have op-
tide chain or functional RNA (such as a tRNA)." The DNA posite 5'--+3' directionality.
molecules of small viruses contain only a few genes, whereas By convention, the site on the DNA at which RNA poly-
the single DNA molecule in each of the chromosomes of merase begins transcription is numbered +1 (Figure 4-10b).
higher animals and plants may contain several thousand Downstream denotes the direction in which 'a template DNA
genes. The vast majority of genes carry information to build strand is transcribed; upstream denotes the opposite direc-
protein molecules, and it is the RNA copies of such protein- tion. Nucleotide positions in the DNA sequence downstream
coding genes that constitute the mRNA molecules of cells. from a start site are indicated by a positive ( +) sign; those
During synthesis of RNA, the four-base language of DNA upstream, by a negative (-) sign. Because RNA is synthe-
containing A, G, C, and T is simply copied, or transcribed, sized 5 '--+3 ', RNA polymerase moves down the template
into the four-base language of RNA, which is identical ex- DNA strand in a 3'--+5' direction. The newly synthesized
cept that U replaces T. In contrast, during protein synthesis, RNA is complementary to the template DNA strand; there-
the four-base language of DNA and RNA is translated into fore it is identical to the nontemplate DNA strand, with ura-
the 20-amino acid language of proteins. In this section, we cil in place of thymine.
focus on formati9n of functional mRNAs from protein-coding
genes (see Figure 4-1, step 0 ). A similar process yields the Stages in Tra nscription To carry out transcription, RNA
precursors of rRNAs and tRNAs encoded by rRNA and polymerase performs several distinct functions, as depicted in
tRNA genes; these precursors are then further modified to Figure 4-11. During transcription initiation, RNA polymerase,
yield functional rRNAs and tRNAs (Chapter 8). Similarly, with the help of initiation factors discussed later, recognizes
thousands of micro RNAs (miRNAs) that regulate transla- and binds to a specific site, called a promoter, in double-
tion of specific target mRNAs arc transcribed into precur- stranded DNA (step 0 ). After binding, RNA polymerase and
sors by RNA polymerases and processed in to functional the initiation factors separate the DNA strands to make the
miRNAs (Chapter 8). Other non-protein-coding (or simply bases in the temp late strand available for base pairing with
"noncoding") RNAs help to regu late transcription of spe- the bases of the ribonucleoside triphosphates that it will po-
cific protein-coding genes. Regulation of transcription al- lymerize. RNA polymerases and initiation factors then melt
lows distinct sets of genes to be expressed in the multiple 12-14 base pairs of DNA around the transcription start site,
different types of cells that make up a multicellular organ- which is located on the template strand within the promoter
ism. It also allows different amounts of mRNA to be tran- region (step If)). This allows the template strand to enter the
scribed from different genes, resulting in differences in the active site of the enzyme that catalyzes phosphodicster bond
amounts of the encoded proteins in a cell. Regulation of formation between ribonucleotide triphosphates tha t arc
transcription is addressed in Chapter 7. complementary to the promoter template strand at the start
site of transcription. The 12- to 14-base-pair region of melted
DNA in the polymerase is known as the transcription bubble.
A Template DNA Strand Is Transcribed into a Transcription initiation is considered complete when the first
Complementary RNA Chain by RNA Polymerase two ribonucleotides of an RNA chain are linked by a phos-
During transcription of DNA, one DNA strand acts as a phodiester bond (step 10).
template, determining the order in which ribonucleoside tri- After several ribonucleotides have been polymerized,
phosphate (rNTP) monomers are polymerized to form a RNA polymerase dissociates from the promoter DNA and

124 CHAPTER 4 • Basic Molecular Genetic Mechanisms

(a) general t ranscription factors. During the stage of strand
3' elongation, RNA polymerase moves along the template
5 ' --+-3' RNA
strand growth DNA one base at a time, opening the double-stranded DNA
in front of its d irection of movement and guiding the strands
A 0" together so that they hybridize at the upstream end of the
) 5' transcription bubble (Figure 4-1 1, step 19). One ribonucleo-
tide at a time is added to the 3' end of the growing (nascent)
RNA chain during strand elongation by the polymerase. The
OH enzyme maimain~ a melted region of approximate!) 14 base
pairs in the transcription bu bble. Approximately eight nucle-
<>tides at the 3' end of the growing RNA strand remain base-
paired to the template DNA strand in the transcription
bubble. The elongation complex, comprising RNA poly-
D merase, template DNA, and the growing (nascent) RNA
N
A strand, is extraordinarily stable. For example, RNA poly-
t
OH 0 merase transcribes the longest known mammalian gene, con-
e I
0 - P=O taining about 2 million base pairs, without dissociating from
m
p I

0 ~~
I G
a
t
e H
FIGURE 4-10 RNA is synthesized S'-t3 '. (a) Polymerization
s
t 3' H of ribonucleotides by RNA polymerase during transcription. The
r
OH OH
ribonucleotide to be added at the 3' end of a growing RNA strand is
a \ Polyme~ization specified by base pairing between the next base in the template DNA
n
~ I ~ ll y
d 0 0 strand and the complementary incoming ribonucleo~ide triphosphate
.. u (rNTP). A phosphodiester bond is formed when RNA polymerase
0-P-0-P-0-P-o-
1 I catalyzes a reaction between the 3' 0 of the growing strand and the
6 0 0 a phosphate of a correctly base-paired rNTP. RNA strands always are
synthesized in the 5'-t3 ' direction and are opposite in polarity to their
OH template DNA strands. (b) Conventions for describing RNA transcription.
Incoming rNTP Top: The DNA nucleotide where RNA polymerase begins transcription is
designated + 1. The direction the polymerase travels on the DNA is
"downstream," and bases are marked w ith positive numbers. The
opposite direction is "upstream," and bases are noted with negative
numbers. Some important gene features lie upstream of the transcrip-
tion start site, including the promoter sequence that localizes RNA
polymerase to the gene. (Bottom) The DNA strand that is being
transcribed is the template strand; its complement, the nontemplate
strand. The RNA being synthesized is complementary to the template
strand, and therefore identical with the nontemplate strand sequence,
5' except with uracil in place of thymine. (Part (b) adapted from Griffiths et al.,
Modern Genetic Analysis, 2d ed., W. H. Freeman and Company.]

(b) Transcription
Promoter )

+1
Upstream Downstream

Nontemplate strand 5' CTGCCATTGTCAGACATGTATACCCCGTACGTCTTCCCGAGCGAAAACGATCTGCGCTGC 3'}

DNA
Template strand 3' GACGGTAACAGTCTGTACATATGGGGCATGCAGAAGGGCTCGCTTTTGCTAGACGCGACG 5'

~
5' OlJG!i&!iiJUilBiliiCii.SSIRDJIDI!IDilllilllllm&ll&eiSI!DHlmiDiliiiDIIliBUilll!l!lllllii1U.MMIIIU'iiiii!JI!Di$111rmat'U1£1 3' Primary RNA
transcript

4.2 Transcription of Prote1n -Coding Genes and Formation of Functional mRNA 125
0 FOCUS ANIMATION: Basic Transcriptional Mechanism

FIGURE 4-11 Three stages in transcription. During .·

RNA polymerase Stop site
initiation of transcription, RNA polymerase forms a
on template
transcription bubble and begins polymerization of INITIATION strand
ribonucleotides (rNTPs) at the start site, which is located
within the promoter region. Once a DNA region has been D Polymerase binds to
transcribed, the separated strands reassociate into a promoter sequence ' I J<n.o""""' no. \ '<fH'" 'lW.,._ 1\;n[':-J]n..,~~~~.£ 5'
double helix. The nascent RNA is displaced from its in duplex DNA. 3'
template strand except at its 3' end. T he 5' end of the ' Closed complex'
RNA strand exits the RNA polymerase through a channel Promoter
in the enzyme. Termination occurs when the polymerase
encounters a specific termination sequence (stop site). fJ Polymerase melts
duplex DNA near ~~~l\.on~~~~ 5'
See the text for details. For simplicity, the diagram
transcription start site, 3'
depicts transcription of four turns of the DNA helix forming a transcription
encoding = 40 nucleotides of RNA. Most RNAs are bubble. "Open
considerably longer, requiring transcription of a longer complex"
region of DNA.
II Polymerase catalyzes
~~ a~ra~·~~r a ·~~·~~ 5'
phosphodiester linkage
of two initial rNTPs. 3'

ELONGATION

II Polymerase advances s·~~.n~...#'JF.~'n'llll!

3' - 5' down template 3'
strand, melting duplex
DNA and adding rNTPs
to growing RNA. 5' hybrid region

TERMINATION 5'~>n>fJI:~f.lr«urJ~:c!\nn ~.ncr>e...fl~,nr<l\lln.f 5.

3' 3'
II At transcription stop site,
polymerase releases
completed RNA and
dissociates from DNA.
Completed
RNA strand

the DNA template or releasing the nascent RNA. RNA syn- subunit (w) that is not essential for transcription or cell viabiliry
thesis occurs at a rate of about 1000 nucleotides per minute but that stabilizes the enzyme and assists in the assembly of
at 37 °C, so the elongation complex must remain intact for its subunits. Archaeal and eukaryotic RNA polymerases have
more than 24 hours to assure continuous RNA synthesis of several additional small subunits associated with this core
the pre-mRNA from this very long gene. complex, which we describe in Chapter 7. Schematic dia-
During transcription termination, the final stage in RNA grams of the transcription process generally show RNA poly-
synthesis, the completed RNA molecule is released from the merase bound to an unbent DNA molecule, as in Figure 4-11.
RNA polymerase and the polymerase dissociates from the However, x-ray crystallography and other studies of an elon-
template DNA (Figure 4-1 1, step Ia). Once it is released, an gating bacterial RNA polymerase indicate that the DNA
RNA polymerase is free to transcribe the same gene again or bends at the transcription bubble (Figure 4-12).
another gene.
Organization of Genes Differs in Prokaryotic
Structu re o f RNA Polymerases The RNA polymerases of
bacteria, archaea, and eukaryotic cells are fundamentally and Eukaryotic DNA
similar in structure and function. Bacterial RNA polymer- Having outlined the process of transcription, we now briefly
ases are composed of two related large subunits ([3' and (3), consider the large-scale arrangement of information in DNA
two copies of a smaller subunit (a), and one copy of a fifth and how this arrangement dictates the requirements for

126 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 RNA synthesis so that information transfer goes smoothly.
(a)
In recent years, sequencing of the entire genomes from mul-
tiple organisms has revealed not only large variations in the
number of protein-coding genes but also differences in their
organization in bacteria and eukaryotes.
The most common arrangement of protein-codmg genes
in all bacteria has a powerful and appealing logic: genes
encoding proteins that function together, for example, the
enzymes required tu ~ynthesize the amino acid tryptophan,
are most often found in a contiguous array in the DNA. Such
an arrangement of genes in a functional group is called an
operon, because it operates as a unit from a single promoter.
Transcription of an operon produces a continuous strand of
mRNA that carries the message for a related series of proteins
(Figure 4-13a). Each section of the mRNA represents the unit
(or gene) that encodes one of the proteins in the series. This
arrangement results in the coordinate expression of all the
genes in the operon. Every time an RNA polymerase mole-
cule initiates transcription at the promoter of the operon, all
(b) ~subu nit"-. the genes of the operon are transcribed and translated. In
prokaryotic DNA the genes are closely packed with very fe""
noncoding gaps, and the DNA is transcribed directly into
mRNA. Because DNA is not sequestered in a nucleus in pro-
karyotes, ribosomes have immediate access to the translation
start sites in the mRNA as they emerge from the surface of
the RNA polymerase. Consequently, translation of the
mRNA begins even while the 3' end of the mRNA is still
being synthesized at the active site of the RNA polymerase.
This economic clustering of genes devoted to a single
metabolic function does not occur in eukaryotes, even sim-
ple ones such as yeasts, which can be metabolically similar to
bacteria. Rather, eukaryotic genes encoding proteins that
function together are most often physically separated in the
DNA; indeed, such genes usually are located on different
chromosomes. Each gene is transcribed from its own pro-
moter, producing one mRNA, which generally is translated
FIGURE 4- 12 Bacterial RNA polymerase. This structure corre- to yield a single polypeptide (Figure 4-13b).
sponds to the polymerase molecule in the elongation phase (step 19) When researchers first compared the nucleotide sequences
of Figure 4-11. In these diagrams, transcription is proceeding in the of eukaryotic mRNAs from multicellular organisms with the
leftward direction. Arrows indicate where downstream DNA enters DNA sequences encoding them, they were surprised to find
the polymerase and upstream DNA exits at an angle from the that the uninterrupted protein-coding sequence of a given
downstream DNA; the coding strand is red, the noncoding strand mRNA was discontinuous in its corresponding section of
blue; nascent RNA green. The litNA polymerase 13' subunit is gold, ~ is DNA. They concluded that the eukaryotic gene existed in
light gray, and the ex subunit visible from this angle brown. The upper pieces of coding sequence, the exons, separated by non-protein-
diagram is a space-filling model of the elongation complex viewed
·. from an angle that emphasizes the bend in the DNA as it passes
coding segments, the introns. This astonishing finding im-
plied that introns are removed from the long initial primary
through the polymerase. The elongation complex is rotated in the
transcript-the RNA copy of the entire transcribed DNA
bottom diagram as shown, and proteins are made largely transparent
sequence-and the remaining exons are spliced together to
to reveal the structure of the transcription bubble inside the
polymerase that is not visible in the space-filling model. Nucleotides
produce eukaryotic mRNAs.
complementary to the template DNA are added to the 3' end of the
Although introns are common in multicellular eukary-
nascent RNA strand (at the left). ThP newly synthesized nascent otes, they :~re extremely rare in bacteria and archaea and
RNA exits the polymerase at the bottom through a channel formed uncommon in many unicellular eukaryotes such as baker's
between the~ and 13' subunits. Thew subunit and the other yeast. However, introns are present in the DNA of viruses
ex subunit are visible from this angle. [Courtesy of Seth Darst; see that infect eukaryotic cells. indeed, the presence of introns
N. Korzheva et al., 2000, Science 289:619-625, and N. Opalka et al., 2003, Cell was first discovered in such viruses, whose DNA is tran-
114:335-345.] scribed by host-cell enzymes.

4.2 Transcription of Protein-Coding Genes and Formation of Functional mRNA 127

 (a) Prokaryotes (b) Eukaryotes

Yeast chromosomes
kb TRP1 TRP4

1550 IV ==~p~====~p~==
E. coli genome TRP2
580 v ====CF==
.t.
TRP5

8 kb 910 VII ===~

t TRP3
Start site
for trp mRNA
synthesis
680 XI 0~====
+
l Transcription
I Transcription and
t RN'A processing
trp mRNA 5'
t
Start sites for
protein synthesis
3' trp
mRNAs
t- ----t
2
t
3
t
4

t
5

~ Translation ~Translation
E
- o
Proteins
- c Proteins
- - -- - 2 3 4 5
-
-
B

-A

FI GURE 4-13 Gene organization in prokaryotes and eukaryotes.

(a) The tryptophan (trp) operon is a continuous segment of the E. coli
genome parallels the sequential function of the encoded proteins in
the tryptophan pathway. (b) The five genes encoding the enzymes
chromosome, containing five genes (blue) that encode the enzymes required for tryptophan synthesis in yeast (Saccharomyces cerevisiae)
necessary for the stepwise synthesis of tryptophan. The entire operon are carried on four different chromosomes. Each gene is transcribed
is transcribed from one promoter into one long continuous trp mRNA from its own promoter to yield a primary t ranscript that is processed
(red). Translation of this mRNA begins at five different start sites, into a functional mRNA encoding a single protein. The lengths of the
yielding five proteins (green). The order of the genes in the bacterial various chromosomes are given in kilobases (10 3 bases).

Eukaryotic Precursor mRNAs Are Processed

that is connected to the terminal nucleotide of the RNA by an
to Form Functional mRNAs unusual5',5' triphosphate linkage (Figure 4-14). The cap pro-
In bacterial cells, which have no nuclei, translation of an tects an mRNA from enzymatic degradation and assists in its
mRNA into protein can begin from the 5' end of the mRNA export to the cytoplasm. The cap also is bound by a protein
even while the 3' end is still being synthesized by RNA poly- factor required to begin translation in the cytoplasm.
merase. In other words, transcription and translation occur Processing at the 3' end of a pre-mRNA involves cleavage
concurrently in bacteria. In eukaryotic cells, however, not by an endonuclease to yield a free 3'-hydroxyl group to which
only is the site of RNA synthesis-the nucleus-separated a string of adenylic acid residues is added one at a time by an
from the site of translation-the cytoplasm-but also the enzyme called poly(A ) polymerase. The resulting poly(A) tail
primary transcripts of protein-coding genes are precursor contains 100-250 bases, being shorter in yeasts and inverte-
mRNAs (pre-mRNAs) that must undergo several modifica- brates than in vertebrates. Poly(A) polymerase is part of a
tions, collectively termed RNA processing, to yield a func- complex of proteins that can locate and cleave a tra nscript at
tional mRNA (see Figure 4-1, step f)). This mRNA then a specific site and then add the correct n umber of A residues,
must be exported to the cytoplasm before it can be trans- in a process that does not require a template.
lated into protein. Thus transcription and translation cannot Another step in the processing of many different eukary-
occur concurrently in eukaryotic cells. otic mRNA molecules is RNA splicing: the internal cleavage of
All eukaryotic pre-mRNAs initiall y are modified at the two a transcript to excise the introns and stitch together the coding
ends, and these modifications are retained in mRNAs. As the cxons. Figure 4-15 summarizes the basic steps in cukaryotic
5' end of a nascent RNA chain emerges from the surface of mRNA processing, using the 13-globin gene as an example. We
RNA polymerase, it is immediately acted on by several en- examine the cell ular machinery for carrying out processing of
zymes that together synthesize the 5' cap, a 7-methylguanylate mRNA, as well as tRNA and rRNA, in Chapter 8.

128 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 FIGURE 4-14 Structure of the 5' methylated cap. The distinguish
ing chemical features of the 5' methylated cap on eukaryotic mRNA are
7-Methylguanylate (1) the s·~s· linkage of 7-methylguanylate to the initial nucleotide of
the mRNA molecule and (2) the methyl group on the 2' hydroxyl of the
ribose of the first nucleotide (base 1). Both these features occur in all
animal cells and in cells of higher plants; yeasts lack the methyl group
on nucleotide 1. The ribose of the second nucleotide (base 2) also is
methylated in vertebrates. [See A. J. Shatkin, 1976, Ce//9:645.]

The functional eukaryoric mRNAs produced by RNA

.· processing retain noncoding regions, referred to as 5' and J '
o - F>- o 5' ~ 5' linkage untranslated regions (UTRs), at each end. In mammalian
0 mRNAs, the 5' UTR may be a hundred or more nucleorides
0 - P= O long, and the 3' UTR may be several kilobases in length.
Bacterial mRNAs also usually have 5' and 3' UTRs, but
0
s·l these are much shorter than those in eukaryotic mRNAs,
C~H
0 Basel generally containing fewer than 10 nucleotides.
4' 1
H H H H
3 2' Alternative RNA Splicing Increases
/ o o~cH3 the Number of Proteins Expressed
0 - P=O from a Single Eukaryotic Gene
cl In contrast to bacterial and archaeal genes, the vast majority
I
C~H
0 Base2 of genes in higher, multicellular eukaryotes contain multiple
H H introns. As noted in Chapter 3, many proteins from higher
H H eukaryotes have a multidomain tertiary structure (see Fig-
/0 O- CH 3 ure 3-11). lndi\idual repeated protein domains often are
0-P=O

Q} OVERVIEW ANIMATION: Life Cycle of an mRNA

FIGURE 4-15 Overview of RNA processing. RNA processing 31 32 105 106 147
~-Globin
produces functional mRNA in eukaryotes. The 13-globin gene contains genomic
three protein-coding exons (constituting the coding region, red) and DNA t t

!
Start site for Addition of Poly( A)
two intervening noncoding introns (blue). The introns interrupt the RNA synthesis m 7 Gppp cap site
protein-coding sequence between the codons for amino acids 31 and
32 and 1OS and 106. Transcript ion of eukaryotic protein-coding genes
starts before the sequence that encodes the first amino acid and
Primary s· ..
RNA
----------------== 3' cleavage and
3'

--
extends beyond the sequence encoding the last amino acid, resulting transcript

1
addition of
in noncoding regions (gray) at the ends of the primary transcript. These poly(A) tail
untranslated regions (UTRs) areJetained during processing. The 5' cap
7
(m Gppp) is added during formation ofthe primary RNA transcript, Exon
..--------------<Aln /
-
which extends beyond the poly(A) site. After cleavage at the poly(A)
site and addition of multiple A residues to the 3' end, splicing removes lntron Poly( A)
the introns and joins the exons. The small numbers refer to positions in
the 147-amino acid sequence of 13-globin. UTR __.
() 1 n _ - . ( tail
•
m 7 Gppp cap .
1. - - - - IAln
n () .
Jl K
lntron excision,
exon ligation

~-Globin
mRNA
..---------:Aln 147

4.2 Transcription of Protein-Coding Genes and Formation of Functional mRNA 129

 ·.

Fibroblast 5' 3'

fibronectin mRNA

Hepatocyte
fibronectin mRNA 5' 3'

FIGURE 4-16 Alternative splicing. The = 75-kb fibronectin gene in fibroblasts includes the EIIIA and ElliS exons, whereas these exons
(top) contains multiple exons; splicing of fibronectin varies by cell type. are spliced out of fibronectin mRNA in hepatocytes.ln this diagram,
The ElliS and EIIIA exons (green) encode binding domains for specific introns (black lines) are not drawn to scale; most of them are much
proteins on the surface of fibroblasts. The fibronectin mRNA produced longer than any of the exons.

encoded by one exon or a small number of exons that code Clearl y, alternative RNA splicing-greatly expands the num-
for identical or nearly identical amino acid sequences. Such ber of proteins encoded by the genomes of higher, multicel-
repeated cxons are thought to ha vc evolved by the accidental lular organisms.
multiple duplication of a length of DNA lying between two
sites in adjacent introns, resulting in insertion of a string of
repeated cxons, separated by introns, between the original
two introns. The presence of multiple introns in many eu- KEY CONCEPTS of Section 4.2
karyotic genes permits expression of multiple, related pro-
Transcription of Protein-Coding Genes and Formation
teins from a single gene by means of alternative splicing. In
of Functional mRNA
higher eukaryotes, alternative splicing is an important mech-
anism for production of different forms of a protein, called Transcription of DNA is carried out by RNA polymerase,
isoforms, by different types of cells. which adds one ribonucleotide at a time to the 3' end of a
fib ronccrin, a multidomain protein found in mammals, growing RNA chain (sec Figure 4-11). The sequence of the
provides a good example of alternative splicing (figure 4-16). template DNA strand determines the order in which ribo-
Fibronectin is a long, adhesive protein secreted into the extra- nucleotides are polymerized ro form an RNA chain.
cellular space that can bind other proteins together. What During transcription initiation, RNA polymerase binds to
and where it binds depends on which domains are spliced a specific site in DNA {the promoter), locally melts the double-
together. The fibronectin gene contains numerous exons, stranded DNA to reveal the unpaired template strand, and
grouped into several regions corresponding to specific do- polymerizes the first two nucleotides complementary to the
mains of the protein. Fibroblasts produce fibronectin mRNAs template strand. The melted region of 12-14 base pairs is
that contain exons ElliA and EIIIB; these exons encode am ino known as the transcription bubble.
acid sequences that bind tightly to proteins in the fibroblast
• During strand elongation, RNA polymerase moves down
plasma membrane. Consequently, this fibronccrin isoform
the DNA, melting the DNA ahead of the polymerase so that
adheres fibroblasts to the extracellular matrix. Alternative
the template strand can enter the active sire of the enzyme
splicmg of rhc fibronectm primary transcript in hepatocyres,
and allowing the complementary strands of the region just
the major type of cell in the liver, yields mRNAs that lack the
transcribed to reanneal behind it. The transcription bubble
EIIIA and EIIIB exons. As a result, the fibronectin secreted by
moves with the polymerase as the enzyme adds ribonucleo-
hepatocytes into the blood does nor adhere tightly to fibro-
tides complementary to the template strand to the 3' end of
blasts or most other cell types, allowing it to circulate. During
the growing RNA chain.
formation of blood clots, however, the fibrin-binding do-
mains of hepatocyte fibroncctin binds to fibrin, one of the When RNA polymerase reaches a termination sequence in
principal constituents of clots. The bound fibronectin then the DNA, the enzyme stops transcription, leading to release
interacts with integrins on the membranes of passing plate- of the completed RNA and dissociation of the enzyme from
lets, thereby expanding the clot by addition of platelets. the template DNA.
More than 20 different isoforms of fibronectin have been • In prokaryotic DNA, several protein-coding genes com-
identified, each encoded by a different, alternatively spliced monly are clustered into a functional region, an operon, which
mRNA composed of a unique combination of fibronectin is transcribed from a single promoter into one mRNA encod-
gene exons. Sequencing of large numbers of mRNAs isolated ing multiple proteins with related functions (see Figure 4-13a).
from various tissues and comparison of their sequences with Translation of a bacterial mRNA can begin before synthesis
genomic DNA has revealed that nearly 90 percent of all of the mRNA is complete.
human genes are expressed as alternatively spliced mRNAs.

130 CHAPTER 4 • Basic Molecular Genetic Mechani sms

 aartRNA7
In eukaryotic DNA, each protein-coding gene is transcribed arriving 1;1
from its own promoter. The initial primary transcript very H2N- c - R
often contains noncoding regions (introns) interspersed among 7

coding regions (exons). C==o

I
0
• Eukaryotic primary transcripts must undergo RNA pro-
cessing to yield functional RNAs. During processing, the
ends of nearly all primary transcripts from protein-coding
genes are modified by addition of a 5' cap and 3' poly(A) tail. tRNA4
Transcripts from genes containing introns undergo splicing, leaving
the removal of the introns and joining of the exons (see
Figure 4-15).
mANA
• The individual domains of multidomain proteins found in
higher eukaryotes are often encoded by individual exons or
a small number of exons. Distinct isoforms of such proteins
aa 1
often are expressed in specific cell types as the result of alter-
native splicing of exons. Movement of ribosome ---~

FIGURE 4 - 1 7 The three roles of RNA in protein synthesis.

Messenger RNA (mRNA) is translated into protein by the joint action
of transfer RNA (tRNA) and the ribosome, which is composed of
numerous proteins and three (bacterial) or four (eukaryotic) ribosomal
RNA (rRNA) molecules (not shown). Note the base pairing between
4.3 The Decoding of mRNA by tRNAs tRNA anticodons and complementary codons in the mRNA. Formation
Although DNA stores the information for protein synthesis of a peptide bond between the amino-group N on the incoming
and mRNA conveys the instructions encoded in DNA, most aa-tRNA and the carboxy-terminal Con the growing. protein chain
biological activities are carried out by proteins. As we saw in (green) is catalyzed by one of the rRNAs. aa = amino acid; R - side
group. [Adapted from A. J. F. Griffiths et al., 1999, Modern Genetic Analysis,
Chapter 3, the linear order of amino acids in each protein
W. H. Freeman and Company.]
determines its three-dimensional structure and activity. For
this reason, assembly of amino acids in their correct order, as
encoded in DNA, is critical to production of functional pro-
teins and hence the proper functioning of cells and organisms. tRNAs and various accessory proteins necessary for protein
Translation is the whole process by which the nucleotide synthesis. Ribosomes are composed of a large and a small
sequence of an mRNA is used as a template to join the subunit, each of which contains its own rRNA molecule or
amino acids in a polypeptide chain in the correct order (see molecules.
Figure 4-1, step D ). In eukaryotic cells, protein synthesis
These three types of RNA participate in the synthesis of
occurs in the cytoplasm, where three types of RNA mole-
proteins in all organisms. In this section, we focus on the de-
cules come together to perform different but cooperative
coding of mRNA by tRNA adaptors, and how the structure of
functions (Figure 4-17):
each of these RNAs relates to its specific task. How they work
1. Messenger RNA (mRNA) carries the genetic informa- together with rRNA, ribosomes, and other protein factors to
tion transcribed from DNA in a linear form. The mRNA is synthesize proteins is detailed in the following section. Because
read in sets of three-nucleotide sequences, called codons, translation is essential for protein synthesis, the two processes
each of which specifies a particular amino acid. commonly are referred to interchangeably. However, the
polypeptide chains resulting from translation undergo post-
2. Transfer RNA (tRNA) is the key to deciphering the
translational folding and often other changes (e.g., chemical
codons in mRNA. Each type of amino acid has its own
modifications, association with other chains) that are required
subset of tRNAs, which bind the amino acid and carry it
for production of mature, functional proteins (Chapter 3).
to the growing end of a polypeptide chain when the next
codon in the mRNA calls for it. The correct tRNA with its
attached amino acid is selected at each step because each Messenger RNA Carries Information from DNA
specific tRNA molecule contains a three-nucleotide se- in a Three-letter Genetic Code
quence, an anticodon, that can base-pair with its comple-
As noted above, the genetic code used by cells is a triplet
mentary codon in the mRNA.
code, with every three-nucleotide sequence, or codon, being
3. Ribosomal RNA (rRNA) associates with a set of proteins "read" from a specified starting point in the mRNA. Of the
to form ribosomes. These complex structures, which physi- 64 possible codons in the genetic code, 61 specify individual
cally move along an mRNA molecule, catalyze the assem- amino acids and three are stop codons. Table 4-1 shows that
bly of amino acids into polypeptide chains. They also bind most amino acids are encoded by more than one codon.

4.3 The Decoding of mRNA by tRNAs 131

 The Genetic Code (Codons to Amino Acids) '
- --
Second Position

u c A G

Phe Ser Tyr Cys u

Phe Ser Tyr Cys c
u
Leu Ser Stop Stop A
Leu Ser Stop Trp G

leu Pro His Arg u ....

:0
w
c: leu Pro H is Arg c J
:::;·

~
c 0..
-o
leu Pro Gin Arg A, 0
c:
.·v;.
0
leu (Met)* Pro Gin Arg G
VI
;:+
o·
:I
0
......
0.. w
VI lie Thr Asn Ser u m
:I
u:::
lie Thr Asn Ser c .3:
A
lie Thr lys Arg A
Met (Start) Thr l ys Arg G

Val Ala Asp Gly u

Val Ala Asp Gly c
G
Val Ala Glu Gly A
Val (:\1et)* Ala Glu Gly G

• AUG is the most common imtiator codon; GUG usuallr codes for valine and CUG for leucine, bur,
rarely, these codons can a lso code for methionine to imtiate a protein chain.

Only two--methionine and tryptophan-have a single codon; linear sequence of amino aci ds in a polypeptide cha in and
at the other extreme, leucine, serine, and arginin e are each also signals where synthesis of the chain starts and stops.
specified by six different codons. The different codons for a Because the gene tic code is a non-overlapping triplet
given amino acid are said to be synonymous. The code itself code without divisions between codons, a particular mRNA
is termed degenerate, meaning that a particular amino acid theo retically could be translated in three different read ing
can be specified by multiple codons. frames. Indeed , some mRNAs have been shown to contain
Synthesis of all polypeptide chains in prokaryotic and eu- overlapping information that can be translated in different
karyotic cells begins with the amino acid methionine. In bac- reading frames, yielding different polypeptides (Figure 4-18).
teria, a specialized form of methionine is used with a formyl The vast majority of mRNAs, however, can be read in o nly
group linked to its amino group. In most mRNAs, the start o ne frame because stop codons encountered in the other t wo
(initiator) codon specifying this amino-terminal methionine possible reading frames terminate translation before a func-
is AUG. In a few bacterial mRNAs, GUG is used as the ini- tional protein is produced. Very rarely, another unusual cod-
tiator codon, and CUG occasionally is used as an initiator ing arra ngement occurs because of frame-shifting. In this
codon for methionine in eukaryotes. The three codons UAA, case the protein-synthesizing machinery may read four nu-
UGA, and UAG do not specify amino acid<> hnt constitute cleotides as one amino acid and then continue reading trip
stop (termination) codons that mark the carboxyl terminus lets, or it may back up one base and read all succeeding
of polypeptide chains in almost all cells. The sequence of triplets in the new frame until termination of the chain oc-
codons that runs from a specific start codon to a stop codon curs. Only a few dozen such instances are known.
is called a reading frame. This precise linear array of ribo- The meaning of each codon is th e same in most known
nucleotides in groups of three in mRNA specifies the precise organisms-a strong argument that li fe on earth evolved

132 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Frame 1 The Folded Structure of tRNA Promotes
5 ' -- ,GCU , UGU UUA , CGA , AUU AA- mRNA Its Decoding Functions
Polypeptide 1 Translation, or decoding, of the four-nucleotide language of
DNA and mRNA into the 20-amino acid language of pro-
Frame 2
teins requires tRNAs and enzymes called ammoacyl-tRNA
5' -G,CUU , GUU , UAC .. GAA, UUA A - mRNA sy11thetases. To participate in protein synthesis, a tRNA mol-
Polypept ide 2 ecule must become chemically linked to a particular amino
acid via a high-ent>rgy bond, forming an aminoacyl-tRNA
Frame 3 (Figure 4-1 9). The anticodon in the tRNA then base-pairs
5 ' -- GC UUG UUU ACG AAU UAA- mRNA with a codon in mRNA so that the activated amino acid can
be added to the growing polypeptide chain (see Figure 4-17).
Polypeptide 3
Some 30-40 different tRNAs have been identified in bac-
FIGURE 4- 18 Multiple reading frames in an mRNA sequence. terial cells and as many as 50-100 in animal and plant cells.
If translation of the mRNA sequence shown begins at three different Thus the number of tRNAs in most cells is more than the
upstream start sites (not shown), then three overlapping reading number of amino acids used in protein synthesis (20 ) and
frames are possible. In this example, the codons are shifted one base also differs from the number of amino acid codons in the
to the right in the middle frame and two bases to the right in the third
genetic code (61). Consequently, many amino acids have
frame, which ends in a stop codon. As a result, the same nucleotide
more than one tRNA to which they can attach (explaining
sequence specifies different amino acids during translation. Although
how there can be more tRNAs than amino acids); in addi-
regions of sequence that are translated in two of the three possible
tion, many tRNAs can pair with more than one codon (ex-
reading frames are rare, there are examples in both prokaryotes and
eukaryotes, and especially in their viruses, where the same sequence
plaining how there can be more codons than tRNAs).
is used in two alternative mRNAs expressed from the same region of The function of tRNA molecules, which are 70-80 nucle-
DNA, and the sequence is read in one reading frame in one mRNA otides long, depends on their precise three-dimensional
and in an alternative reading frame in the other mRNA. There are even structures. In solution, all tRNA molecules fold into a simi-
a few instances where the same short sequence is read in all three lar stem-loop arrangement that resembles a Cloverleaf when
possible reading frames. drawn in two dimensions (Figure 4-20a). The four stems are
short double helices stabihzed by Watson-Crick base pair-
ing; three of the four stems have loops containing seven or
eight bases at their ends, while the remaining, unlooped stem
only once. In fact, the genetic code shown in Table 4-1 is contains the free 3' and 5' ends of the chain. The three nu-
known as the universal code. However, the genetic code has cleotides composing the anticodon are located at the center
been found to differ for a few codons in many mitochondria, of the midd le loop, in an accessible position that facilitates
in ciliated protozoans, and in Acetabularia, a single-celled codon-anticodon base pairing. In all tRNAs, the 3' end of
plant. As shown in Table 4-2, most of these changes involve the un looped amino acid acceptor stem has the sequence
reading of normal stop codons as amino acids, not an ex- CCA, which in most cases is added after synthesis and pro-
change of one amino acid for another. These exceptions to cessing of the tRNA are complete. Several bases in most
the universal code probably were later evolutionary develop- tR:::-JAs also are modified after transcription, creating non-
ments; that is, at no single time was the code immutably standard nucleotides such as inosine, dihydrouridine, and
fixed, although massive changes were not tolerated once a pseudouridine. As we will see shortly, some of these modi-
general code began to function early in evolution. fied bases are known to play an important role in protein

1§:1!jE Known Deviations from the Universal Genetic Code

Codo n Universal Code Unusual Code* Occurrence

UGA Stop Trp Mycoplasma, Sp1roplasma, mirochondria of many speCies

CUG Leu Thr .Mirochondria in yeam

l iAA, UAG Stop Gin Acetabulana, Tetrahymena, ParameciUm. etc.

UGA Stop Cys Euplotes

• Found in nuclear genes of rhe hsred organisms and m mitochondnal genes as mdicated.
~OURCE: S. Osawa eta!., 1992, Microbial. Rev. 56:129.

4.3 The Decoding of mRNA by tRNAs 133

 Amino acid (Phe) High-energy
ester bond
H
I
u/
I
H
I
0
I

l
H2N-C-C-0 H,N-C-C-0

II ~H, ~H, j
Linkage of
Phe to tRNAPhe
lo
6 Phe-tRNAPhe binds
to the UUU codon 6 Net result:
Phe is selected
by its codon

?"S
ATP AMP
+ PP;
AAA
Aminoacyl-
tRNA synthetase
specific for Phe
tRNA specific for
Phe (tRNAPhe)
Aminoacyl-tRNA s·- - •3'
mRNA

FIGURE 4-1 9 Decoding nucleic acid sequence into ami no acid terminal adenosine in the corresponding tRNA. Step H : A three-base
seq uence. The process for translating nucleic acid sequences in mRNA sequence in the tRNA (the anticodon) then base-pairs with a codon in
into amino acid sequences in proteins involves two steps. Step O : An the mRNA specifying the attached amino acid. If an error occurs in
aminoacyl-tRNA synthetase first couples a specific amino acid, via a either step, the wrong amino acid may be incorporated into a
high·energy ester bond (yellow}, to either the 2' or 3 ' hydroxyl of the polypeptide chain. Phe = phenylalanine.

(a) 3' synthesis. Viewed in three dimensions, the folded tRNA mol-
ecule has an L shape with the anticodon loop and acceptor
@ = dihydrouridine stem forming the ends of the two arms (Figure 4-20b).
Q)= inosine
(f)= ribothymidine
Nonstandard Base Pairing Often Occurs
8 = pseudouridine
m =methyl group
Between Codons and Anticodons
If perfect Watson-Crick base pairing were demanded between
T'i'CG
codons and anticodons, cells would have to contain at least 61
different types of tRNAs, one for each codon that specifies an
amino acid. As noted above, however, many cells contain
fewer than 61 tRNAs. The explanation for the smaller number
lies in the capability of a single tRNA anticodon to recognize
more than one, but not necessarily every, codon corresponding
to a given amino acid. This broader recognition can occur be-
cause of nonstandard pairing between bases in the so-called
wobble position: that is, the third (3') base in an mRNA codon
and the corresponding first (5') base in its tRNA anticodon.
: Anticodon
The first and second bases of a codon almost always form
standard Watson-Crick base pairs with the third and second
3 2 1 bases, respectively, of the corresponding anticodon, bur four
'--- !!'~'::!~---' nonstandard interactions can occur between bases in the
(b) wobble position. Particularly important is the G·U base pair,

Acceptor stem
FIGURE 4 -20 Structure of tRNAs. (a) Although the exact nucleotide
sequence varies among tRNAs, they all fold into four base-paired stems
and three loops. The CCA sequence at the 3' end also is found in all
tRNAs. Attachment of an amino acid to the 3' A yields an aminoacyl-
tRNA. Some of the A, C, G, and U residues are modified post-
transcriptionally in most tRNAs (see key}. Dihydrouridine (D) is nearly
always present in the D loop; likewise, ribothymidine (T} and pseudo-
uridine ('i') are almost always present in the T'I'CG loop. Yeast alanine
tRNA, represented here, also contains other modified bases. The triplet
at the tip of the anticodon loop base-pairs with the corresponding
codon in mRNA. (b) Three-dimensional model of the generalized
backbone of all tRNAs. Note the L shape of the molecule. [Part (a) see
R. W. Holly et al., 1965, Science 147:1462. Part (b) adapted from J. G. Arnez and
D. Moras, 1997, Trends Biochem. Sci. 22:211.]

134 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 tRNA acid. For example, four of the six codons for leucine (CUA,
3' CUC, CUU, and UUA) are all recognized by the same tRNA
5' with the anticodon 3'-GAI-5'; the inosine in the wobble po-
sition forms nonstandard base pairs with the third base in
the four codons. In the case of the UUA codon, a nonstan-
If these bases are in
first, or wobble, position of dard G·U pair also forms between position 3 of the antico-
anticodon don and position 1 of the codon.
c A G u
12 G u c A then the tRNA may Amino Acids Become Activated When
5' mRNA 3' u G A recognize codons in
Covalently Linked to tRNAs
u mRNA having these
bases in th1 a position Recognition of the codon or codons specifying a given amino
acid by a particular tRNA is actually the second step in decod-
If these bases are in ing the genetic message. The first step, attachment of the ap-
, or wobble, position propriate amino acid to a tRNA, is catalyzed by a specific
5' mRNA 3' of codon of an mRNA aminoacyl-tRNA synthetase. Each of the 20 different synthe-
12 c A G u tases recognizes one amino acid and all its compatible, or cog-
G u c A then the codon may nate, tRNAs. These coupling enzymes link an amino acid to the
I I u G be recognized by a free 2' or 3' hydroxyl of the adenosine at the 3' terminus of
I tRNA having these tRNA molecules by an ATP-requiring reaction. In this reaction,
bases in first position
of anticodon the amino acid is linked to the tRNA by a high-energy bond
and thus is said to be activated. The energy of this bond subse-
quently drives formation of the peptide bonds linking adjacent
amino acids in a growing polypeptide chain. The equilibrium of
3'
the aminoacylation reaction is driven further t.oward activation
tRNA
of the amino acid by hydrolysis of the high-energy phosphoan-
FIGURE 4 -21 Nonstandard base pairing at t he wobble position . hydride bond in the released pyrophosphate (sec Figure 4-19).
The base in the third (or wobble) position of an mRNA codon often Aminoacyl-tRNA synthetases recognize their cognate
forms a nonstandard base pair with the base in the first (or wobble) tRNAs by interacting primarily with the anitcodon loop and
position of a tRNA anticodon. Wobble pairing allows a tRNA to
acceptor stem, although interactions with other regions of a
recognize more than one mRNA codon (top); conversely, it allows a
tRNA also contribute to recognition in some cases. Also, spe-
codon to be recognized by more than one kind of tRNA (bottom),
cific bases in incorrect tRNAs that are structurally similar to a
although each tRNA will bear the same amino acid. Note that a tRNA
with I (inosine) in the wobble position can "read" (become paired with)
cognate tRNA will inhibit charging of the incorrect tRNA.
three different codons, and a tRNA with G or U in the wobble position Thus, recognition of the correct tRNA depend~ on both posi-
can read two codons. Although A is theoretically possible in the tive interactions and the absence of negative interactions. Still,
wobble position of ~he anticodon, it is almost never found in nature. because some amino acids are so similar structurally, amino-
acyl-tRNA synthetases sometimes make mistakes. These are
corrected, however, by the enzymes themselves, which have a
which structurally fits almost as well as the standard G·C pair. proofreading activity that checks the fit in their amino acid-
Thus, a given anticodon in tRNA with Gin the first (wobble) binding pocket. If the wrong amino acid becomes attached to
position can base-pair with the two corresponding codons a tRNA, the bound synthetase catalyzes removal of the amino
that have either pyrimidine (Cor U) in the third position (Fig- acid from the tRNA. This crucial function helps guarantee
ure 4-21 ). For example, the phenylalanine codons UUU and that a tRNA delivers the correct amino acid to the protein-
UUC (5'--*3') are both recognized by the tRNA that has synthesizing machinery. The overall error rate for translation
GAA (5'--*3') as the anticodon . In fact, any two codons of in E. coli is very low, approximately 1 per 50,000 codons,
the type NNPyr (N = any base; Pyr = pyrimidine) encode a evidence of both the fidelity of tRNA recognition and the im-
single amino acid and are decoded by a single tRl\'A with G portance of proofreading by aminoacyl-tRNA synthetases.
in the first (wobble) position of the anticodon.
Although adenine rarely is found in the anticodon wobble
position, many tRNAs in plants and animals contain inosine KEY CONCEPTS of Section 4.3
(1), a deaminated product of adenine, at this position. Ino-
The Decoding of mRNA by tRNAs
sine can form nonstandard base pairs with A, C, and U. A
tRNA with inosine in the wobble position thus can recognize • Genetic information is transcribed from DNA into mRNA
the corresponding mRNA codons with A, C, or U in the in the form of an overlapping, degenerate triplet code.
third (wobble) position (see Figure 4-21). For this reason, • Each amino acid is encoded by one or more three-nucleotide
inosine-containing tRNAs are heavily employed in transla- sequences (codons) in mRNA. Each codon specifies one amino
tion of the synonymous codons that specify a single amino

4.3 The Decoding of mRNA by tRNAs 135

 amino acid polymerization would be very slow. The effi-
acid, but most amino acids are encoded by mul tiple codom ciency of translation is greatly increased by the binding of the
(see Table 4-1). mRNA and the individual aminoacyl-tRNAs within a ribo-
• The AUG codon for methionine is the most common start some. The ribosome, the most abundant RNA-protein com-
codon, specifying the amino acid at the NHrterminus of a plex in the cell, directs elongation of a polypeptide at a rate
protein chain. T hree codons (UAA, UAG, UGA) function as of three to five amino acids added per second. Small proteins
stop codons and specify no amino acids. of 100-200 amino acids are therefore made in a minute or
less. On the other hand, it takes 2-3 hours to make the larg-
A reading frame, the unin terrupted sequence of codons in
est known protein, titin, wh1ch IS found in muscle and con-
mRNA from a specific start codon to a stop codon, is trans-
tains about 30,000 amino acid residues. The cellular machine
lated into the linear sequence of amino acids in a polypeptide
that accomplishes this task must be precise and persistent.
chain.
With the aid of the electron microscope, ribosomes were
Decoding of the nucleotide sequence in mRNA into the first discovered as small, discrete, RNA-rich particles in cells
amino acid sequence of proteins depends on tRNAs and that secrete large amounts of protein. However, their role in
aminoacyl-tRNA synthetases. protein synthesis was not recognized unti l reasonably pure
All tRNAs have a similar three-dimensional structure that ribosome preparations were obtained. In vitro radiolabeling
includes an acceptor arm for attachment of a specific amino experiments with such preparati011s showed that radioactive
acid and a stem-loop with a thr ee-base anticodon sequence amino acids were first incorporated into growing polypeptide
at its ends (see Figure 4-20). The anticodon can base-pair chains that were associated with ribosomes before appearing
with its corresponding codon in mRNA. in finished chains.
Though there are differences between the ribosomes of
Because of nonstandard interactions, a tRNA may base-
prokaryotes and eukaryotes, the great structural and func-
pair with more than one mRNA codon; conversely, a par-
tional similarities between ribosomes from all species reflects
ticular codon may base-pair with multiple tRNAs. In each
the common evolutionary origin of the most basic constitu-
case, however, only the proper amino acid is inserted into a
ents of living cells. A ribosome is composed of three (in bac-
growing polypeptide chain.
teria) or four (in eukaryotes) different rRNA molecules and
Each of the 20 aminoacyl-tRNA synthetases recognizes a as many as 83 proteins, organized into a large subunit and a
single amino acid and covalently links it to a cognate tRNA, small subunit (Figure 4-22). The ribosomal subunits and the
forming an aminoacyl-tRNA (see Figure 4-19). This reaction rRNA molecules are commonly designated in Svedberg units
activates the amino acid, so it can participate in peptide (S), a measure of the sedimentation rare of macromolecules
bond formation. centrifuged under standard conditions-essentially, a mea-
sure of size. The small ribosomal subunit contains a single
rRNA molecule, referred to as small rRNA. The large sub-
unit contains a molecule of large rRNA and one molecule of
SS rRNA, plus an additional molecule of 5.8S rRNA in ver-
4.4 Stepwise Synthesis of Proteins tebrates. The lengths of the rRNA molecules, the quantity of
proteins in each subunit, and consequently the sizes of the
on Ribosomes
subunits differ between bacterial and eukaryoric cells. The
The previous sections have introduced two of the major par- assembled ribosome is 70S in bacteria and 80S in vertebrates.
ticipants in protein synthesis-mRNA and aminoacylated The sequences of the small and large rRNAs from several
tRNA. Here we first describe the third key player in protein thousand organisms are now known. Although the primary
synthesis-the rRNA-containing ribosome-before taking a nucleotide sequences of these rRNAs vary considerably, the
detailed look at how all three components are brought to- same parts of each type of rRNA theoretically can form
gether to carry out the biochemical events leading to formation base-paired stem-loops, which would generate a similar
of polypeptide chains on ribosomes. Similar to transcription, three-dimensional structure for each rRNA in all organisms.
the complex process of translation can be divided into three The three-dimensional structures of bacterial rRNAs have
stages-initiation, elongation, and termination-which we been determined by x-ray crystallography of isolated 50S
consider in order. We focus our description on translation in and 30S subunits and entire 70S ribosomes (Figure 4-23 ).
eukaryotic cells, but the mechanism of translation is funda- The multiple, much smaller ribosomal proteins for the most
mentally the same in all cells. part arc associated with the surface of the rRNAs. Although
the number of protein molecules in ribosomes greatly ex-
ceeds the number of RNA molecules, RNA constitutes about
Ribosomes are Protein-Synthesizing
60 percent of the mass of a ribosome. The sites where tRNAs
Machines are bound by ribosomes are known as the A site, the P s1te,
If the many components that participate in translating mRNA and the E site. As we will see shortly, tRNAs move between
had to interact in free solution, the likelihood of simultane- these sires as protein synthesis takes place. Crystal structures
ous collisions occurring would be so low that the rate of of the yeast 80S ribosome, a protozoan 405 subunit, as well

136 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Assembled
rRNA Proteins Subunits ribosomes

+ Total: 31
23S 5S
(2900 rNTs) (120 rNTs)

+ Total: 21
16S
(1500 rNTs)

2ss~y.ss + Total: 50

28S: 5.8S 5S
(4800 rNTs, 160 rNTs) (120 rNTs)

+ Total: 33

18S
(1900 rNTs) 40S

FIGURE 4- 22 Prokaryotic and eukaryotic ribosome components. molecules (235 and 165 rRNA in bacteria; 285 and 185 rRNA in
In all cells, each ribosome consists of a large and a small subunit. The vertebrates) and a 55 rRNA. The large subunit of vertebrate ribosomes
two subunits contain rRNAs (red) of different lengths, as well as a also contains a 5.85 rRNA base-paired to the 285 rRNA. The number
different set of proteins. All ribosomes contain two major rRNA of ribonucleotides (rNTs) in each rRNA type is indicated.

as cryo-electron microscopy of plant ribosomes have been frame for the entire mRNA. Both prokaryotes and eukary-
reported recently. They are generally similar to bacterial ribo- otes contain two different methionine tRNAs: tRNA,~Iet can
somes but are about 50 percent larger because of eukaryotic- initiate protein synthesis, and tRNA Met can incorporate methio-
specific insertions of RNA segments into regions of the nine only into a growing protein chain. The same aminoacyl-
bacterial rRNAs as well as the presence of a larger number tRNA synthetase (MetRS) charges both tRNAs with
of proteins (see Figure 4-22). Basic aspects of protein synthe- methionine. But only Met-tRNA,Met (i.e., activated methio-
sis are thought to be similar, although initiation of transla- nine attached to tRNA,~Ict) can bind at the appropriate site on
tion in eukaryotes, discussed later, is more complex and the small ribosomal subunit, the P site, to begin synthesis of
subject to additional mechanisms of regulation. a polypeptide chain. The regular Met-tRNA \let and all other
During translation, a ribosome moves along an mRNA charged tRNAs bind only to the A site, as described later.
chain, interacting with various protein factors and tRNAs and tRNAs enter the exit orE site after transferring their cova-
undergoing large conformational changes. Despite the com- lent!} bound amino acid to the growing polypeptide chain.
plexity of the ribosome, great progress has been made in deter-
mining the overall structure of bacterial ribosomes and how Eukaryotic Translation Initiation Usually
they function in protein synthesis. More than 50 years after the
initial discovery of ribosomes, their overall structure and func-
Occurs at the First AUG Closest
tion during protein synthesis are finally becoming clear. to the 5 ' End of an mRNA
During the first stage of translation, the small and large ribo-
Methionyl-tRNA,Met Recognizes <;omal subunits assemble around an mRNA that has an a<.:ti-
vated initiator tRNA correctly positioned at the start codon
the AUG Start Codon in the ribosomal P site. in eukaryotes, this process IS mediated
As noted earlier, the AUG codon for methionine functions as by a special set of proteins known as eukaryotic translation
the start codon in the vast majority of mRNAs. A critical as- initiation factors (eiFs). As each individual component joins
pect of translation initiation is to begin protein synthesis at the complex, it is guided by interactions with specific eiFs.
the start codon, thereby establishing the correct reading Several of these initiation factors bind GTP, and the hydrolysis

4.4 Stepwise Synthesis of Proteins on Ribosome s 137

0 VIDEO: Rotating 3-0 Model of a Bacterial Ribosome

50S

30S

FIGURE 4-23 Structure of E. coli 70S ribosome as determined colored light purple and dark purple, respectively; and the SS rRNA is
by x-ray crystallography. Model of the ribosome viewed along the colored dark blue. The positions of the ribosomal A, P, and E sites are
interface between the large (50S) and small (305) subunits. The 165 indicated. Note that the ribosomal proteins are located primarily on
rRNA and proteins in the small subunit are colored light green and dark the surface of the ribosome, and the rRNAs on the inside, lining the
green, respectively; the 235 rRNA and proteins in the large subunit are A, P, and E sites. [From B.S. Schuwirth et al., 2005, Science 310:827.]

of GTP to GDP functions as a proofreading switch that only progress has been made in the past few years in understand-
allows subsequent steps to proceed if the preceding step has ing translation initiation in vertebrates.
occurred correctly. Before GTP hydrolysis, the complex is un- The current model for initiation of translation in verte-
stable, allowing a second attempt at complex formation until brates is depicted in Figure 4-24. Large and small ribosomal
the correct complex assembles, resulting in GTP hydrolysis subunits released from a previous round of trans lation are
and stabilization of the appropriate complex. Considerable kept apart by the binding of e!Fs l, lA, and 3 to the small

138 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Met
e1F2 ternary
complex Met
formation 405
D
>>----+ (?}GTP:r-435 complex format1on
@GTP 0
eiF4 complex { ~ ~e\ }
2 GTP
5 435preinitiation
complex
@

5 .• _5?_____ Attachment to mRNA

AUG

5' to 3' scanning

FIGURE 4-24 Initiation of translation in eukaryotes.

The current model of eukaryotic initiation involves eight
steps. Step 0: An eiF2 ternary complex forms when
eiF2-GTP binds a tRNA,M••. Step fJ:When a ribosome
dissociates at the termination of translation, the 40S subunit
is bound by eiFl, eiFl A, and eiF3. A 43S preinitiation
complex forms when this associates with an eiF2 ternary Initiation codon recognition,
complex and e1F5. Step 10: An mRNA is activated when a hydrolysis of e1F2-bound GTP
and P, release
multiple-subunit eiF4 complex binds: subunit eiF4E binds to
Met
the 5' cap structure and subunit eiF4G binds multiple copies
of the poly(A)-binding protein (PABP) bound to the mRNA
poly(A) tail. Subunit e1F4A RNA helicase activity unwinds any
RNA secondary structure at the 5' end of the mRNA. eiF4B,
485 Initiation
which stimulates ei~4A helicase activity, also joins this complex
}
circular complex in which both the mRNA 5' cap and poly(A)
tail are associated w ith the e1F4 complex. Step D:The 43S
preinitiation complex binds an eiF4-mRNA complex.
Step r!l: The eiF4A RNA helicase unwinds RNA secondary
structure as the 40S complex scans in the 5'~3' direction
until it recognizes the initiation codon. Step rl'!: Recognition
of the initiation codon causes eiF5 to stimulate hydrolysis of
eiF2-bound GTP. This switches the conformation of the
scanning complex to a 48S initiation complex with the
anticodon oftRNA,Met base-paired to the initiator AUG in the
40S P site. Step 6 :The 60S subunit joins the 40S subunit,
leading to release of most the earlier-acting eiFs as
eiF5B-GTP binds to eiF1A in the ribosomal A site. The
released e1F4 complex and e1F4B associate with the cap and Hydrolysis of eiF5B-bound GTP
PABP as shown in step 1D to prepare for interaction with and release of eiF5B and eiF1A
another 43S preinitiation complex. For simplicity, this is not
® ®GOP
shown. Step(;): Correct association of the 40S and 60S
subunit results in hydrolysis of eiF5B-bound GTP, release
of eiF5B-GDP and eiFl A, and formation of the 80S initiation
80S Initiation
complex with tRNA,M••base-paired to the initiation codon in complex
the ribosomal P site [Adapted from R. J. Jackson et al .. 201 o,
Nar. Rev. Mol. Cell Bioi. 1 0:113.)

4.4 Stepwise Synthesis of Proteins on Ribosomes 139

40S subunit (Figure 4-24, top). The first step of translation cellular mRNAs contain an internal ribosome entry site (IRES)
initiation IS formation of a 43S preinitiation complex. This located far downstream of the 5' end. It is thought that cel-
preinitiation complex is formed when the 405 subunit with lular IRESs form RNA structures that interact with a com-
eiFs 1, 1A, and 3 associates with elF5 and a ternary complex plex of eiF4A and eiF4G that then associates with e!F3
consisting of the Met-tRNAi:\ler and eiF2 bound to GTP (Fig- bound to 405 subunits with eiF1 and eiFIA. This assembly
ure 4-24, steps 0 and f)). The initiation factor eiF2 alter- then binds an eiF2 ternary complex to assemble an initiation
nates between association with GTP and GDP; it can only complex directly on a neighboring AUG codon. In addition,
bind Met-tRNA,Mcr when it is associated with GTP. Cells can translation of some positive-stranded viral RNAs, which
regulate protein synthesis by phosphorylating a serine resi- lack a 5' cap, is initiated at viral IRI:~ sequences. These fall
due on the eiF2 bound to GDP; the phosphorylated complex into different classes depending on how many of the stan-
is unable to exchange the bound GDP for GTP and cannot dard eiFs are required for initiation. In the case of cricket
bind Met-tRNAi~'k', thus inhibiting protein synthesis. paralysis virus, the =200-nt-long IRES folds into a complex
The mRNA to be translated is bound by the multiple- structure that interacts directly with the 40S ribosome and
subunit eiF4 complex, which interacts with both the 5'-cap leads to initiation without any of the elFs or even the initia-
structure and the poly(A)-binding protein (PABP) bound in tor Met-tRNA,"•'!
multiple copies to the mRNA poly(A) tail. Both interactions In bacteria, binding of the small ribosomal subunit to an
are required for translation of most mRNAs. This results in initiation site occurs by a differtmt mechanism that allows
formation of a circular complex (Figure 4-24, step 11). The eiF4 initiation at internal sites in the polycistronic mRNAs tran-
cap-binding complex consists of several subunits with different scribed from operons. In bacterial mRNAs, an =6-base
functions. The eiF4E subunit binds the 5'-cap structure on sequence complementary to the 3' end of the small rRNA
mRNAs (Figure 4-14). The large eiF4G subunit interacts with precedes the AUG start codon by 4-7 nucleotides. Base pair-
PABP bound to the mRNA poly(A) tail, and also forms a scaf- ing between this sequence in the mRNA, called the Shine-
fold to which the other eiF4 subunits bind. The mRNA-eJF4 Dalgarno sequence after its discoverers, and the small rRNA
complex then associates with the preinitiation complex places the small ribosomal subunit in the proper position for
through an interaction between e!F4G and elF3 (step 19). initiation. Initiation factors comparable to eiFlA, eiF2, eiF3,
The imtiation complex then slides along, or scans, the and f-Met-tRNA,\1"' then associate with the small subunit,
associated mRNA as the helicase activity of elF4A, stimu- followed by association of the large subunit to form the
lated by eiF4B, uses energy from ATP hydrolysis to unwind complete bacterial ribosome.
RNA secondary structure (step A). Scanning stops when the
tRNAi:\ler anticodon recognizes the start codon, which is the During Chain Elongation Each Incoming
first AUG downstream from the 5' end in most eukaryotic
Aminoacyl-tRNA Moves Through Three
mRNAs. Recognition of the start codon leads to hydrolysis
of the GTP associated with elf2, an irreversible step that Ribosomal Sites
prevents furthe_r scanning, resulting in formation of the 485 The correctly positioned ribosome-Met-tRNAi\ler complex
mitiation complex (step fa). This commitment to the correct is now ready to begin the task of stepwise addition of amino
initiation codon is facilitated by eiF5, an eJF2 GTPase- acids by the in-frame translation of the mRNA. As is the case
activating protein (GAP, see Figure 3-32). Selection of the with initiation, a set of special proteins, termed translation
mitiating AUG is facilitated by specific surrounding nucleo- elongation factors (EFs), are required to carry out this pro-
tides called the Kozak sequence, for Marilyn Kozak, who de- cess of chain elongation. The key steps in elongation are entry
fined it: (5') ACCMK!G (3'). The A preceding the AUG (in of each succeeding aminoacyl-tRNA with a tRNA comple-
hold) and the G immediately following it are the most impor- mentary to the next codon, formation of a peptide bond, and
tant nucleotides affecting translation initiation efficiency. the movement, or translocation, of the ribosome one codon
Association of the large (60S) subunit is mediated by at a time along the mRNA.
e!F5B bound to GTP, and results in displacement of many of At the completion of translation initiation, as noted al-
the initiation factors (step fJ). Correct association between ready, Met-tRNA,:-.1" is bound to the P site on the assembled
the ribosomal subunits results in hydrolysis of the eiF5B- 80S ribosome (Figure 4-25, top). This region of the ribo-
bound GTP to GDP and release of e1F5B-GDP and elF1A some is called the P site because the tRNA chemically linked
(step[;)), completing formation of an 80S initiation complex. to the growing polypeptide chain is located here. The second
Coupling the ribosome subunit joining reaction to GTP hy- aminoacyl-tRNA is brought into the ribosome as a ternary
drolysis by e!F5B allows the initiation process to continue complex in association with EFlec-GTP and becomes bound
only when the subunit interaction ha~ occurred correctly. It to the A site, so named because it is where aminoacylatcd
also makes this an irreversible step, so that the ribosomal tRNAs bind (step 0). EFlcx·GTP bound to various aminoacyl-
subunits do not dissociate until the entire mRNA is trans- tRNAs diffuse into the A site, but the next step in translation
lated and protein synthesis is terminated. proceeds only when the tRNA anticodon base-pairs with the
The eukaryotic protein-synthesizing machinery begins second codon in the coding region. When that occurs prop-
translation of most cellular mRNAs within about 100 nucle- erly, the GTP in the associated EF1cx·GTP is hydrolyzed. The
otides of the 5'-capped end as just described. However, some hydrolysis of GTP promotes a conformational change in

140 CHAPTER 4 • Basic Molecular Genetic Mechanisms

0 FOCUS ANIMATION: Protein Synthesis

FIGURE 4 -25 Pe ptidyl chain elongation in eukaryotes. Once the

80S ribosome with Met-tRNA1Met in the ribosome P site is assembled (top),
a ternary complex bearing the second amino acid (aa 2) coded by the
mRNA binds to the A site (step O J. Following a conformational change
in the ribosome induced by hydrolysis of GTP in EFl a ·GTP (step f) ), the
large rRNA catalyzes peptide bond formation between Met and aa 2
(step 10). Hydrolysis of GTP in EF2·GTP causes another conformational

~•GTP
rhange in the ribosome that results in its translocation one codon along 80S ribosome
the mRNA and shifts the unacylated tRNA,Met to theE site and the tRNA
with the bound peptide to the P site (step B J. The cycle can begin again
with binding of a ternary complex bearing aa 3 to the now open A site. In Entry of next 6 4)
the second and subsequent elongation cycles, the tRNA at the E site is
ejected during step f) as a result of the conformational change induced
aa-tRNA at
A site ~•GTP
by hydrolysis of GTP in EF 1a ·GTP. [Adapted from K. H. Nierhaus et al., 2000, in
R. A. Garrett et al., eds., The Ribosome: Structure, Function, Antibiotics, and Cellular
Interactions, ASM Press, p. 319.]

EFlex that leads to release of the resulting EFlex·GDP com-

GTP hydrolysis,
plex and tight binding of the aminoacyl-tRNA in the A site ribosome
(step f)). This conformational change also positions the ami- conformational
noacylated 3' end of the tRNA in the A site in close proxim- change
ity to the 3' end of the Met-tRNA/''cr in the P site. GTP
hydrolysis, and hence tight binding, does not occur if the
anticodon of the incoming aminoacyl-tRNA cannot base-
pai r with t he codon at t he A site. In this case, the ternary
complex diffuses away, leaving an empty A site that can as-
sociate with other am inoacyl-tRNA-Eflex·GTP complexes
until a correctly base-paired tRNA is bound. Thus, GTP hy-
Peptide bond
drolysis by EF1 ex is another proofreading step that allows formation
protein synthesis to proceed only when the correct amino
acylated tRNA is bound to the A site. This phenomenon
contributes to the fidelity of protein synthesis.
With the initiating Met-tRNA,:-.Icr at the P site and the sec-
ond aminoacyl-tRNA tightly bound at the A site, the ex-amino
group of the second amino acid reacts with the "activated"
(ester-linked) methionine on the initiator tRNA, forming a

r
peptide bond (Figure 4-25, step 11; see Figure 4-17) . T his
peptidyltransferase reaction is catalyzed by the large rRNA, EF2•GTP
Ribosome
which precisely orients t he interacting atoms, permitting the
translocation II
reaction to proceed. The Z' -hydroxyl of the terminal A of the EF2•GDP + P;

peptidyl-tRNA in the P site also participates in catalysis. The

catalytic abi lity of the large rRNA in bacteria has been dem-
onstrated by carefully removing the vast majority of the pro-
tein from large ribosomal subunits. The nea rly pure bacterial
235 rRNA can cat alyze a peptidyltransferase reaction be-
tween ana logs of aminoacylated-tRNA and peptidyl-tRNA.
Further support for the catalytic role of la rge rRNA in pro-
tein synthesis came from high-resolution crysta llographic
sLUJies showing that no protetns lie near the sire of peptide
bond synthesis in the crysta l structure of the bacterial large eukaryotic EF2·GTP. Once translocation has occurred cor-
subunit. rectly, the bound GTP is hydrolyzed, another irreversible pro
Following peptide bond synthesis, the ribosome translo- cess that prevents the ribosome from moving along the RNA
cates along the mRNA a distance equal to one codon. This in the wrong direction or from translocating an incorrect
translocation step is monitored by hydrolysis of the GTP in number of nucleotides. As a resu lt of conformational changes

4.4 Stepwise Synthesis of Proteins on Ribosomes 141

in the ribosome that accompany proper translocation and the tRNAs between the A, P, and E sites as the ribosome translo-
resulting GTP hydrolysis by EF2, tRNAiMcr, now without its cates along the mRNA one three-nucleotide codon at a time.
activated methionine, is moved to the E (exit) site on the ribo-
some; concurrently, the second tRNA, now covalently bound
to a dipeptide (a peptidyl-tRNA), is moved to the P site (Fig- Translation Is Terminated by Release Factors
ure 4-25, step 19). Translocation thus returns the ribosome When a Stop Codon Is Reached
conformation to a state in which the A site is open and able The final stages of translation, like initiation and elongation,
to accept another aminoacylated tRNA complexed with require highly specific molecular signals that decide the fate
Ef 1u · GTP, bt:ginning another cycle of chain elongation. of the mRNA-ribosome-tRNA-peptidyl complex. Two types
Repetition of the elongation cycle depicted in Figure 4-25 of specific protein release factors (RFs) have been discovered.
adds amino acids one at a time to the C-terminus of the Eukaryotic eRFl, whose shape is similar to that of tRNAs,
growing polypeptide as directed by the mRNA sequence, acts by binding to the ribosomal A site and recognizing stop
until a stop codon is encountered. In subsequent cycles, the codons directly. Like some of the initiation and elongation
conformational change that occurs in step lfJ ejects the un- factors discussed previously, the second eukaryotic release
acylated tRNA from the E sire. As the nascent polypeptide factor, eRF3, is a GTP-binding protein. The eRF3·GTP acts
chain becomes longer, it threads through a channel in the large in concert with eRFl to promote cleavage of the peptidyl-
ribosomal subunit, exiting at a position opposite the side that tRNA, thus releasing the completed protein chain (Figure
interacts with the small subunit (Figure 4-26). 4-27). Bacteria have two release factors (RFl and RF2) that
In the absence of the ribosome, the three-base-pair RNA- are functionally analogous to eRFl and a GTP-binding factor
RNA hybrid between the tRNA anticodons and the mRNA (RF3) that is analogous to eRF3. Once again, the eRF3 GTPase
codons in the A and P sites would not be stable; RNA-RNA monitors the correct recognition of a stop codon by eRFl.
duplexes between separate RNA molecules must be consider- The peptidyl-tRNA bond of the tRNA in the P site is not
ably longer to be stable under physiological conditions. How- cleaved, terminating translation, until one of the three stop
ever, multiple interactions between the large and small codons is correctly recognized by eRFl, another example of
rRNAs and general domains of tRNAs (e.g., the D and a proofreading step in protein synthesis.
T'l'CG loops, see Figure 4-20) stabilize the tRNAs in the A Release of the completed protein leaves a free tRNA in the
and P sites, while other RNA-RNA interactions sense correct P site and the mRNA still associated with the 80S ribosome
codon-anticodon base pairing, assuring that the genetic code with eRFl and eRF3-GDP bound in the A site. In eukaryotes,
is read properly. Then, interactions between rRNAs and the ribosome recycling occurs when this post-termination com-
general domains of all tRNAs result in the movement of the plex is bound by a protein called ABCEl, which uses energy
from ATP hydrolysis to separate the subunits and release the
mRNA. Initiation factors eiFl, eiFlA, and e1F3 are also re-
quired and load onto the 405 subunit, making it ready for
Polypeptide another round of initiation (Figure 4-24, top). In reality, a
free mRNA is never released as diagrammed in Figure 4-27
for simplicity. Rather, the mRNA has other ribosomes as-
sociated with it in various stages of elongation, PABP bound
to the poly A tail, and eiF4 complex associated with the 5'-cap,
ready to associate with another 43S preinitiation complex
(Figure 4-24 ).

Polysomes and Rapid Ribosome Recycling

Increase the Efficiency of Translation
Translation of a single eukaryotic mRNA molecule to yield a
typical-sized protein takes one to two minutes. Two phenomena
significantly increase the overall rate at which cells can synthe-
size a protein: the simultaneous translation of a single mRNA
molecule by multiple ribosomes, and rapid recycling of ribo-
3' somal subunits after they disengage from the 3' end of an
FIGURE 4·26 Model of E. coli 70S ribosome bound to an mRNA
mRNA. Simultaneous translation of an mRNA by multiple
with a nascent polypeptide chain in the exit tunnel. The model is ribosomes is readily observable in electron micrographs and
based on cryo-EM studies. Three tRNAs are superimposed on the by sedimentation analysis, revealing mRNA attached to mul-
A (pink), P (green), and E (yellow) sites. The nascent polypeptide chain tiple ribosomes bearing nascent growing polypeptide chains.
is buried in a tunnel in the large ribosomal subunit that begins close These structures, referred to as polyribosomes or polysomes,
to the acceptor stem of the tRNA in the P site. [See I. S. Gabashvili et al., were seen to be circular in electron micrographs of some tissues.
:woo, Cell 100:537; courtesy of J. Frank.] Subsequent studies with purified initiation factors explained

142 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 tail and the eiF4G subunit of eiF4. Since the elF4E subunit of
.. • eiF4 binds to the cap structure on the 5' end of an mRNA,
the two ends of an mRNA molecule are bridged by the inter-
vening proteins, forming a "circular" mRNA (Figure 4-28a).
Because the two ends of a polysome are relatively close to-
gether, ribosomal subunits that disengage from the 3' end are
3' positioned near the 5' end, facilitating re-initiation by the in-
teraction of the 405 subunit and its associated initiation fac-
tors wirh eiF4 bound to the 5' cap. The circular pathway
depicted in figure 4-28b is thought to enhance ribosome re-
cycling and thus increase the efficiency of protein synthesis.

GTPase-Superfamily Proteins Function in Several

Quality Control Steps of Translation
We can now see that one or more GTP-binding proteins par-
ticipate in each stage of translation. These proteins belong to
3'
the GTPase superfamily of switch proteins that cycle between
a GTP-bound active form and GOP-bound inactive form (see
Figure 3-32). Hydrolysis of the bound GTP causes a confor-
Peptidyl-tRNA
cleavage
~P, mational change in the GTPase itself and other associated
proteins that are critical to various complex molecular pro-
cesses. In translation initiation, for instance, hydrolysis of
eiF2·GTP to eiF2·GDP prevents further scanning of the
mRNA once the start site is encountered and allows binding
of the large ribosomal subunit to the small .subunit (see Fig-
Post-termination ure 4-24, step fa ). Similarly, hydrolysis of EFlo:·GTP to
complex Eflo:·GDP during chain elongation occurs only when the A
3' site is occupied by a charged tRNA with an anticodon that
base-pairs with the codon in the A site. GTP hydrolysis causes
a conformational change in EFlo: resulting in release of its
bound tRNA, allowing the aminoacylated 3' end of the
charged tRNA to move into the position required for peptide
ATP bond formation (Figure 4-26, step f) ). Hydrolysis of
8 -GDP EF2·GTP to EF2·GDP leads to correct translocation of the
ADP+ P1 ribosome along the mRNA, (see Figure 4-26, step a ), and
ABCE1
eRF hydrolysis of eRF3·GTP to eRF3·GDP assures correct termi-
nation of translation (Figure 4-27). Since hydrolysis of the

'b
high energy !3--y phosphodiester bond of GTP is irreversible,
coupling of these steps in protein synthesis to GTP hydrolysis
prevents them from going in the reverse direction.

5' 3'
Nonsense Mutations Cause Premature
FIGURE 4-27 Termination oftranslation in eukaryotes. When a Termination of Protein Synthesis
ribosome bearing a nascent protein chain reaches a stop codon (UAA,
UGA, UAG), release factor eRF 1 enters the A site together with One kind of mutation that can inactivate a gene in any organ-
eRF3·GTP. Hydrolysis of t he bound GTP is accompanied by cleavage of ism is a base-pair change that converts a codon normally en-
the peptide chain from the tRNA in the P site and ejection of the tRNA coding an amino acid into a stop codon, e.g., UAC (encoding
in theE site, forming a post-termination complex. The ribosomal tyrosine)~ UAG (stop ). When this occurs early in the read-
subunit s are separated by the action of the ABCEl ATPase together ing frame, the resulting truncated protein usually is nonfunc-
with eiFl, eiFlA, and e1F3. The 405 subunit is released bound to these tional. Such mutations are ~.:ailed nonsense mutations because
eiFs, ready to initiate another cycle of translation (see Figure 4-24). when the genetic code equating each triplet codon sequence
with a single amino acid was being deciphered by researchers,
the circular shape of polyribosomes and suggested the mecha- the three stop codons were found to not encode any amino
nism by wh ich ribosomes recycle efficiently. acid-they did not "make sense."
These studies revealed that multiple copies of the poly(A)- In genetic studies with the bacterium E. coli, it was discov-
binding protein (PABP) interact with both an mRNA poly(A) ered that the effect of a nonsense mutation can be suppressed

4.4 Stepwise Synthesis of Proteins on Ribosomes 143

{a) both the original nonsense mutation and the second mutation
in the anticodon of the tRNA Tvr gene consequently can insert
a tyrosine at the position of the mutant stop codon, allowing
protein synthesis to continue past the original nonsense mu-
tation. This mechanism of suppression is not highly efficient,
so translation of normal mRNAs with a UAG stop codon
terminates at the normal position in most instances. If enough
of the protein encoded by the original gene with the nonsense
mutatiuu i~ produced to provide its essential functions, the ef-
fect of the first mutation is said to be supfJTessed by the second
mutation in the anticodon of the rRNA gene.
This mechanism of nonsense suppression is a powerful
tool in generic studies in bacteria. For example, mutant bac-
terial viruses can be isolated that cannot grow in normal cells
but can grow in cells expressing a nonsense-suppressing tRNA
because the mutant virus has a nonsense mutation in an es-
sential gene. Such mutant viruses grown on the nonsense-
suppressing cells can then be used in experiments to analyze
the function of the mutant gene by infecting normal cells that
{b) ----------- ... do nor suppress the mutation and analyzing what step in the
viral life cycle is defective in the absence of the mutant protein.

KEY CONCEPTS of Section 4 .4

Stepwise Synthesis of Proteins on Ribosomes
• Both prokaryotic and eukaryotic ribosomes-the large
ribonucleoprotein complexes on which translation occurs-
consist of a small and a large subunit (see Figure 4-22). Each
subunit contains numerous different proteins and one major
rRNA molecule (small or large). The large subunit also con-
tains one accessory 55 rRNA in bacteria and two accessory
rRNAs in eukaryotes (55 and 5.85 in vertebrates).
EXPERIMENTAL FIGURE 4-28 Circular structure of mRNA • Analogous rRNAs from many different species fold into
increases translation efficiency. Eukaryotic mRNA forms a circular quite similar three-dimensional structures containing numer-
structure owing to interactions of three proteins. (a) In the presence of ous stem-loops and binding sires for proteins, mRNA, and
purified poly(A)-binding protein (PABP), eiF4E, and eiF4G, eukaryotic tRNAs. Much smaller ribosomal proteins are associated
mRNAs form circular structures, visible in this force-field electron with the periphery of the rRNAs.
micrograph. In these structures, protein-protein and protein-mRNA
interactions form a bridge between the 5' and 3' ends of the mRNA. • Of the two methionine tRNAs found in all cells, only one
(b) Model of protein synthesis on circular polysomes and recycling of (rRNA;I\Icr) functions in initiation of translation.
ribosomal subunits. Multiple individual ribosomes can simultaneously Each stage of translation-initiation, chain elongation,
translate a eukaryotic mRNA, shown here in circular form stabilized by
and termination-requires specific protein factors including
interactions between proteins bound at the 3' and 5' ends. When a
GTP-binding proteins that hydrolyze their bound GTP to
ribosome completes translation and dissociates from the 3' end, the
GOP when a step has been completed successfully.
separated subunits can rapidly find the nearby 5' cap (m 7G) and
PABP-bound poly(A) tail and initiate another round of synthesis. • During initiation, the ribosomal subunits assemble near the
[Part (a) courtesy of A. Sachs.] translation start site in an mRNA molecule with the tRNA
carrying the amino-terminal methionine (Met-tRNA,'1e') base-
paired with the start codon (Figure 4-24).
by a second mutation in a tRNA gene. This occurs when the Chain elongation entails a repetitive four-step cycle: loose
sequence encoding the anticodon in the tRNA gene is changed binding of an incoming aminoacyl-tRNA to the A site on
to a triplet that is complementary to the original stop codon, the ribosome; tight binding of the correct aminoacyl-tRNA
e.g., a mutation in tRNA r) r that changes its anticodon from to the A site accompanied by release of the previously used
GUA to CUA, which can base-pair with the UAG stop codon. tRNA from the E site; transfer of the growing peptidyl chain
The mutant tRNA can still be recognized by the tyrosine to the incoming amino acid catalyzed by large rRNA; and
amino acyl synthetase and coupled to tyrosine. Cells with

144 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 DNA Polymerases Require a Primer
translocation of the ribosome ro the next codon, thereby mov-
to Initiate Replication
ing the peptidyl-tRNA in the A site to the P site and the now
unacylated tRNA in the P site to theE site (see Figure 4-25). Analogous to RNA, DNA is synthesized from deoxynucleoside
5' -triphosphate precursors (dNTPs). Also like RNA synthesis,
In each cycle of chain elongation, the ribosome undergoes
DNA synthesis always proceeds in the 5'---+3' direction because
two conformational changes monitored by GTP-binding pro-
chain growth results from formation of a phosphoester bond
teins. The first (EF lo:) permits tight binding of the incoming
between the 3' oxygen of a growing strand and the o: phosphate
aminoacyl-tRNA to the A site and ejection of a tRNA from
of a dNTP (see Figure 4-10a ). As discussed earlier, an RNA
theE site, and the second (EF2) leads to translocation .
polymerase can find an appropriate transcription start site on
Termination of translation is carried out by two types of duplex DNA and initiate the synthesis of an RNA complemen-
termination factors: those that recognize stop codons and tary to the template DNA strand (see Figure 4-11 ). ln contrast,
those that promote hydrolysis of peptidyl-tRNA (see Fig- DNA polymcrases cannot initiate chain synthesis de novo; in-
ure 4-27). Once again, correct recognition of a stop codon stead, they require a short, preexisting RNA or DNA strand,
is monitored by a GTPase (eRF3). called a primer, to begin chain growth. With a primer base-
• The efficiency of protein synthesis is increased by the simul- paired to the template strand, a DNA polymerase adds deoxy-
taneous translation of a single mRNA by multiple ribosomes, nucleotides to the free hydroxyl group at the 3' end of the
forming a polyribosome, or simply polysome. In eukaryotic primer as directed by the sequence of the template strand:
cells, protein-mediated interactions bring the two ends of a
polyribosome close rogether, thereby promoting the rapid Primer
recycling of ribosomal subunits, which further increases the
efficiency of protein synthesis (see Figure 4-28b). 5' \ 3'
3' 1111111 II I I I I II I 5'
/
Template strand

4.5 DNA Replication When RNA is the primer, the daughter strand that IS formed
Now that we have seen how genetic information encoded in is RNA at the 5' end and DNA at the 3' end.
the nucleotide sequence of DNA is translated into the pro-
teins that perform most cell functions, we can appreciate the
Duplex DNA Is Unwound, and Daughter Strands
necessity for precisely copying DNA sequences during DNA
replication, in preparation for cell division (see Figure 4-1, Are Formed at the DNA Replication Fork
step 19). The regular pairing of bases in the double-helical In order for duplex DNA to function as a template during
DNA structure suggested to Watson and Crick that new replication, the two intertwined strands must be unwound,
DNA strands are synthesized by using the existing (parental) or melted, to make the bases available for pairing with the
strands as templates in the formation of new, daughter base~ of the dNTPs that are polymerized into the newly
strands complementary to the parental strands. synthesized daughter strands. This unwinding of the paren-
This base-pairing template model theoretically could tal DNA strands is by specific helicases, beginning at unique
proceed either by a conservative or a semiconseruative mech- segments in a DNA molecule called replication origins, or
anism. In a conservative mechanism, the two daughter strands simply origins. The nucleotide sequences of origins from
would form a new double-stranded (duplex) DNA molecule different organisms vary greatly, although they usually
and the parental duplex wou ld remain intact. In a semicon- contain A·T-rich sequences. Once helicases have unwound
servative mechanism, the parental strands are permanently the parental DNA at an origin, a specialized RNA poly-
separated and each forms a duplex molecule with the daugh- merase called primase forms a short RNA primer comple-
ter strand base-paired to it. Definitive evidence that duplex mentary to the unwound template strands. The primer, still
DNA is replicated by a semiconservative mechanism came base-paired to irs complementary DNA strand, is then
from a now classic experiment conducted by M. Meselson elongated by a DNA polymerase, thereby forming a new
and W. F. Stahl, outlined in Figure 4-29. daughter strand.
Copying of a DNA template strand into a complementary The DNA region at which all these proteins come to-
strand thus is a common feature of DNA replication, transcrip- geth er to carry out synthesis of daughter strands is called the
tion of DNA into RNA, and, as we will see later in this chapter, replication fork. As replication proceeds, the replication fork
DNA repair and recombination. In all cases, the information in and associated proteins move away from the origin. As
the template in the form of the specific sequence of nucleotides noted earlier, local unwinding of duplex DNA produces tor-
is preserved. In some viruses, single-stranded RNA molecules sional stress, which is relieved by ropoisomerase I. In order
function as templates for synthesis of complementary RNA or for DNA polymerases to move along and copy a duplex
DNA strands. However, the vast preponderance of RNA and DNA, helicase must sequentially unwind the duplex and
DNA in cells is synthesized from preexisting duplex DNA. topoisomerase must remove the supercoils that form.

4.5 DNA Replication 145

 (a) Predicted results (b) Actual results

Conservative mechanism Semiconservative mechanism Density--+ Density--+

Generation

0
Parental strands
synthesized in 15N
0.3

0.7
H H

New
~ 1.0

1.1
Old
After first
doubling in ' 4N 1.5

1.9

l\1\ 2.5

3.0

4.1

After second 0 and 1.9

doubling in mixed

0 and 4.1
mixed
H H L L L L l L H H H-· H-H . H-L HH

EXPERIMENTAL FIGURE 4 ·29 The Meselson-Stahl experiment. with either mechanism. (b) Actual banding patterns of DNA subjected .·
This experiment showed that DNA replicates by a semiconservative to equilibrium density-gradient centrifugation before and after shifting
15 14
mechanism. E. coli cells initially were grown in a medium containing N-Iabeled E. coli cells to N-containing medium. DNA bands were
ammonium salts prepared with "heavy" nitrogen CN) until all the
5
visualized under UV light and photographed. The traces on the left are
cellular DNA was labeled. After the cells were transferred to a medium a measure of the density of the photographic signal, and hence the
containing the normal "light" isotope C4 N), samples were removed DNA concentration, along the length of the centrifuge cells from left
periodically from the cultures and the DNA in each sample was to right. The number of generations (far left) following the shift to
14
analyzed by equilibrium density-gradient centrifugation, a procedure N-containing medium was determined by counting the concentra-
that separates macromolecules on the basis of their density. This tion of E. coli cells in the culture. This value corresponds to the number
technique can separate heavy-heavy (H-H), light-light (L-L), and of DNA replication cycles that had occurred at the time each sample
heavy-light (H-L) duplexes into distinct bands. (a) Expected composi- was taken. After one generation of growth, all the extracted DNA had
15
tion of daughter duplex molecules synthesized from N-Iabeled DNA the density of H-L DNA. After 1.9 generations, approximately half the
14
after E. coli cells are shifted to N-containing medium if DNA replica- DNA had the density of H-L DNA; the other half had the density of
tion occurs by a conservative or semiconservative mechanism. Parental L-L DNA. With additional generations, a larger and larger fraction of the
heavy (H) strands are in red; light (L) strands synthesized after shift to extracted DNA consisted of L-L duplexes; H-H duplexes never
14
N-containing medium are in blue. Note that the conservative appeared. These results match the predicted pattern for the semicon-
mechanism never generates H-L DNA and that the semiconservative servative replication mechanism depicted in (a). The bottom two ..
mechanism never generates H-H DNA but does generate H-L DNA centrifuge cells contained mixtures of H-H DNA and DNA isolated at 1.9
during the first and subsequent doublings. With additional and 4.1 generations in order to clearly show the positions of H-H, H-L,
replication cycles, the 15 N-Iabeled (H) strands from the original DNA are and L-L DNA in the density gradient. [Part (b) from M. Meselson and
diluted, so that the vast bulk of the DNA would consist of L-L duplexes F. W. Stahl, 1958, Proc. Nat'/ Acad. Sci. USA 44:671.]

A major complication in the operation of a DNA replica- fork (figure 4-30). The problem comes in synthesis of the
tion fork arises from two properties: the two strands of the other daughter strand, called tht> lagging strand.
parental DNA duplex are antiparallel, and DNA polymer- Because growth of the lagging strand must occur in the
ases (like RNA polymerases) can add nucleotides to the 5'~3' direction, copying of its template strand must some-
growing new strands only in the 5'~3' direction. Synthesis how occur in the opposrte direction from the movement of
of one daughter strand, called the leading strand, can pro- the replication fork. A cell accomplishes this feat by synthe-
ceed continuously from a single RNA primer in the 5'~3' sizing a new primer every few hundred hases or so on the
direction, the same direction as movement of the replication second parental strand, as more of the strand is exposed by

146 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 (;) ANIMATION: Nucleotide Polymerization by DNA Polymerase

FIGURE 4 -30 Leading-st rand and lagging-strand DNA 5'

synthesis. Nucleotides are added by a DNA polymerase to each 3'
growing daughter strand in the 5'~3' direction (indicated by
arrowheads). The leading strand is synthesized continuously from a
single RNA primer (red) at its 5' end. The lagging strand is
synthesized discontinuously from multiple RNA primers that are
formed periodically as each new region of the parental duplex is
unwound. Elongation of these primers initidlly produces Okazaki 3' ~~~~~~~~,·~----- Short RNA primer
fragments. As each growing fragment approaches the previous 5'
primer, the primer is removed and the fragments are ligated.
Repetition of this process eventually results in synthesis of the Direction of fork
entire lagging strand. movement

5'
3'

unwinding. Each of these primers, base-paired to their tem- unwind the parental strands at a replication fork. Primers
plate strand, is elongated in the 5'~3 ' direction, forming for leading and lagging daughter-strand DNA are synthe-
discontinuous segments called Okazaki fragments after sized by a complex of prinzase, which synthesizes a short
their discoverer Reiji Okazaki (see Figure 4-30). The RNA RNA primer of "'=' 10 nucleotides, and DNA polymerase a
primer of each Okazaki fragment is removed and replaced (Pol a), which extends the RNA primer wi.th deoxynucleo-
by DNA chain growth from the neighboring Okazaki tides for another "'='20 nucleotides, forming a mixed RNA -
fragment; finally, an enzyme called DNA ligase joins the DNA primer.
adjacent fragmen ts. The primer is extended into daughter-strand DNA by
DNA polymerase 8 (Pol 8) and DNA polymerase € (Pol €),
which are less likely to make errors during copying of the
template strand than is Pol a because of their proofreading
Several Proteins Participate
mechanism (see Section 4.7). Recent results indicate that
in DNA Replication during the replication of cellular DNA, Pol 8 synthesizes
Detailed understanding of the eukaryotic proteins that par- lagging-strand DNA, while PolE synthesizes most of the
ticipate in DNA replication initially came largely from stud- length of the leading strand. Pol 8 and Pol € each form a
ies with small viral DNAs, particularly SV40 DNA, the complex with Rfc (replication factor C) and PCNA (prolif-
circular genome of a small virus that infects monkeys. Virus- erating cell nuclear antigen), which displaces the primase-
infected cells replicate large numbers of the simple viral ge- Pol a complex following primer synthesis. As illustrated in
nome in a short period of time, making them ideal model Figure 4-3lb, PCNA is a homotrimeric protein that has a
systems for studying basic aspects of DNA replication. Be- central hole through which the daughter duplex DNA passes,
cause simple viruses like SV40 depend largely on the DNA thereby preventing the PCNA-Rfc-Pol8 and PCNA-Rfc-Pol E
replication machinery of their host cells (in this case monkey complexes from dissociating from the template. As such,
cells), they offer a unique opportunity to study DNA replica- PCNA is known as a sliding clamp that enables Pol 8 and Pol E
tion of multiple identical small DNA molecules by cellular to remain stably associated with a single template strand for
proteins. Figure 4-31 depicts the multiple proteins that coor- thousands of nucleotides. Rfc functions to open the PCNA
dinate copying of SV40 DNA at a replication fork. The as- ring so that it can encircle the short region of double-
sembled proteins at a replication fork further illustrate the stranded DNA synthesized by Pol a. Consequently, Rfc is
concept of molecular machines introduced in Chapter 3. often called a clamp loader.
These multicomponent complexes permit the cell to carry After parental DNA is separated into single-stranded tem-
out an ordered sequence of events that accomplishes essen- plates at the replication fork, it is hound by multiple copies of
tial cell functions. RPA (replication protein A), a heterotrimeric protein (Figure
The molecular machine that replicates SV40 DNA con- 4-31c). Binding of RPA maintains the template in a uniform
tains only one viral protein. All other proteins involved in conformation optimal for copying by DNA polymerases.
SV40 DNA replication are provided by the host cell. This Bound RPA proteins arc dislodged from the parental strands
viral protein, large T-antigen, forms a hexameric replicative by Pol a, Pol 8, and Pol £ as they synthesize the complemen-
helicase, a protein that uses energy from ATP hydrolysis to tary strands base-paired with the parental strands.

4.5 DNA Replication 147

I '
G) FOCUS ANIMATION: Coordination of Leading- and Lagging-Strand Synthesis

(a) SV40 DNA replication fork

3'
Lagging strand 1!1 Pol8
5'
Rfc
PCNA
I 5'

(b) PCNA

Leading strand

5'

FIGURE 4-31 Model of an SV40 DNA replication fork. (a) A representing PCNA bound to DNA in (a). (c) The large subunit of RPA
hexamer of large T-antigen (0 ), a viral protein, functions as a helicase contains two domains that bind single-stranded DNA. On the left, the
to unwind the parental DNA strands. Single-strand regions of the structure determined for the two DNA-binding domains of the large
parental template unwound by large T-antigen are bound by multiple subunit bound to single-stranded DNA is shown with the DNA
copies of the heterotrimeric protein RPA (fJ). The leading strand is backbone (white backbone w ith blue bases) parallel to the plane of the
synthesized by a complex of DNA polymerase£ (Pol£), PCNA, and page. Note that the single DNA strand is extended with the bases
Rfc (i)). Primers for lagging-strand synthesis (red, RNA; light blue, DNA) exposed, an optimal conformation for replication by a DNA poly-
are synthesized by a complex of DNA polymerase et (Pol et) and primase merase. On the right, the view is down the length of the single DNA
(fi)). The 3' end of each primer synthesized by Pol et-primase is then strand, revea ling how RPA f3 strands wrap around the DNA. The
bound by a PCNA-Rfc-Pol 8 complex, which proceeds to extend the diagram at the bottom shows the icon representing RPA bound to DNA
primer and synthesize most of each Okazaki fragment (Iii). (b) The in part (a). [Part (a) adapted from 5. J. Flint et al., 2000, Virology: Molecular
three subunits of PCNA, shown in different colors, form a circular Biology, Pathogenesis, and Control, ASM Press. Part (b) adapted from J. M. Gulbis
structure w ith a central hole through which double-stranded DNA et al., 1996, Ce/1 8 7:297. Part (c) adapted from A. Bochkarev et al., 1997, Nature
passes. A d iagram of DNA is shown in the center of a ribbon model of 385:176.]
the PCNA trimer. The diagram at the upper left shows the icon

Several other eukaryotic proteins that function in DNA of the parental strands (see Figure 4-8a). Ribonuclease Hand
replication are not depicted in Figure 4-31. For example, FEN I remove the ribonuclcotides at the 5' ends of Okazaki
topoisomerase I associates with the parental DNA ahead of fragments; these are replaced by deoxynucleotides added by
the replicative helicase, i.e., to the left ofT-antigen in Figure DNA polymerase o as it extends the upstream Okazaki frag-
4-31, to remove torsional stress introduced by the unwinding ment. Successive Okazaki fragments are coupled by DNA

148 CHAPTER 4 • Basic Molecular Genetic Mechanisms

0 FOCUS ANIMATION: Bidirectional Replication of DNA
EXPERIMENTAL FIGURE 4·32 Bidirectional replication in EcoRI
SV40 DNA. Electron microscopy of replicating SV40 DNA indicates or7\~~t~triction
bidirectional growth of DNA strands from an origin. The replicating
viral DNA from SV40-infected cells was cut by the restriction
enzyme EcoRI, which recognizes one site in the circular DNA. This
u EcoRI

was done to provide a landmark for a specific sequence in the SV40

Circular viral
genome: the EcoRI recognition sequence is now easily recognized chromosome
as the ends of linear DNA molecules visualized by electron
microscopy. Electron micrographs of EcoRI-cut replicating SV40
DNA molecules showed a collection of cut molecules with
Replication
increasingly longer replication "bubbles," whose centers are a
constant distance from each end of the cut molecules. This finding
~ bubble
is consistent with chain growth in two directions from a common
origin located at the center of a bubble, as illustrated in the
corresponding diagrams. [See G. C. Fa reed et al .. 1972, J. Virol. 1 0:484;
photographs courtesy of N. P. Salzman.]

ligase through standard 5'-+3' phosphoester bonds. Other

specialized Dl\A polymerases are involved in repair of mis-
matches and damaged lesions in DNA (see Section 4.7).

DNA Replication Occurs Bidirectionally

from Each Origin I
As indicated in Figures 4-30 and 4-31, both parental DNA
strands that arc exposed by local unwinding at a replication
fork are copied into a daughter strand. In theory, DNA rep-
lication from a single origin could involve one replication
fork that moves in one direction. Alternatively, two replica-
tion forks might assemble at a single origin and then move in
K>-----i J
opposite directions, leading to bidirectional growth of both
daughter strands. Several types of experiments, including the
one shown in Figure 4-32, provided early evidence in sup-
port of bidirectional strand growth. resulting single-stranded DNA. Synthesis of primers and
The general consensus is that all prokaryotic and eukary- subsequent steps in replication of cellular DNA are thought
otic cells employ a bidirectional mechanism of DNA replica- to be analogous to those in SV40 DNA replication (see Fig-
tion. In the case of SV40 DNA, replication is initiated by ures 4-31 and 4-33 ).
binding of two large T-antigen hexameric helicases to the Replication of cellular DNA and other events leading to
single SV40 origin and ::t.ssembly of other proteins to form proliferation of cells are tightly regulated, so that the appro-
two replication forks. These then move away from the SV40 priate numbers of cells constituting each tissue are produced
origin in opposite directions, with leading- and lagging-strand during development and throughout the life of an organism.
synthesis occurring at both forks. As shown in Figure 4-33, Control of the initiation step is the primary mechanism for
the left replication fork extends DNA synthesis in the left- regulating cellular DNA replication. Activation of MCM he-
ward direction; similarly, the right replication fork extends licase activity, which is required to initiate cellular DNA rep-
DNA synthesis in the rightward direction. lication, is regulated by specific protein kinases called
Unlike SV40 DNA, eukaryotic chromosomal DNA mol- S-phase cyclin-dcpendent kinases. Other cyclin-dependent
ecules contain multiple replication origins separated by tens kinases regulate additional aspects of cell proliferation, in-
to hundreds of kilobases. A six-subunit protein called ORC, cluding the complex process of mitosis by which a eukary-
for origin recognition complex, binds to each origin and otic cell divides into two daughter cells. Mitosis and another
associates with other proteins required to load cellular hex- specialized type of cell division called meiosis, which gener-
americ helicases composed of six homologous MCM pro- ates haploid sperm and egg cells, are discussed in Chapter 5.
teins. Two opposed MCM helicases separate the parental We discuss the various regulatory mechanisms that deter-
strands at an origin, with RPA proteins binding to the mine the rate of cell division in Chapter 20.

4.5 DNA Replication 149

0 FOCUS ANIMATION: Coordination of Leading- and Lagging-Strand Synthesis

FIGURE 4 -33 Bidirectional mechanism of DNA replication.

The left replication fork here is comparable to the replication fork
diagrammed in Figure 4-31, which also shows proteins other than large
T-antigen. Top: Two large T-antigen hexameric helicases first bind at
the replication origin in opposite orientations. Step 0 : Using energy D 1
Unwinding
provided from ATP hydrolysis, the helicases move in opposite
directions, unwinding thP parental DNA and generating single-strand
templates that are bound by RPA proteins. Step H:Primase-Pol a
complexes synthesize short primers (red) base-paired to each of the
separated parental strands. Step ID:PCNA-Rfc-Pol &/e complexes
replace the primase-Pol a complexes and extend the short primers,
fJ 1Leading-strand primer synthesis

generating the leading strands (dark green) at each replication fork.

Step f) : The helicases further unwind the parental strands, and
RPA proteins bind to the newly exposed single-strand regions.
Step g ; PCNA-Rfc-Pol & complexes extend the leading strands further.
Step l'i!: Primase- Pol a complexes synthesize primers for lagging-
Ell Leadjng-strand extension

strand synthesis at each replication fork. Step lfJ: PCNA-Rfc-Pol &

complexes displace the primase-Pol a complexes and extend the
lagging-strand Okazaki fragments (light green), which eventually are
ligated to the 5' ends of the leading strands. The position where
ligation occurs is represented by a circle. Replication continues by
further unwinding of the parental strands and synthesis of leading
ID 1
Unwinding

and lagging strands as in Steps fJ- IfJ. Although depicted as individual

steps for clarity, unwinding and synthesis of leading and lagging
strands occur concurrently.

-Ill Leading-strand extension

KEY CONCEPTS of Section 4.5

DNA Replication
Each strand in a parental duplex DNA acts as a template
for synthesis of ·a daughter strand and remains base-paired
to the new strand, fo rming a daughter duplex (semiconser-
1!1 1
Lagging-strand primer synthesis

vative mechanism). New strands are formed in the 5'~3 '

direction.
Replication begins at a sequence called an origin. Each
eukaryotic chromosomal DNA molecule contains multiple
replication origins.
DNA polymerases, unlike RNA polymerases, cannot un-
111 Lagging-strand extension

wind the strands of duplex DNA and cannot initiate synthe-

sis of new strands complementary to the template strands.
• At a replication fork, one daughter strand (the leading
strand) is elongated continuously. The other daughter strand
(the lagging strand) is formed as a series of discontinuous
Okazaki fragments from primers synthesized every few hun-
dred nucleotides (Figure 4-30).
The ribonucleotides at the 5' end of each Obzaki frag-
ment are removed and replaced by elongation of the 3' end by multiple copies of a single-stranded DNA-binding pro-
of the next Okazaki fragment. Finally, adjacent Okazaki tein, RPA. Primase synthesizes a short RNA primer, which
fragments are joined by DNA ligase. remains base-paired to the template DNA. This initially is
Hclicases use energy from ATP hydrolysis to separate the extended at the 3' end by DNA polymerase a (Pol a ), result-
parental (template) DNA strands which are initially bound ing in a short (5')RNA-(3 ' )DNA daughter strand.

150 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 integrity can be compromised, and then discuss the repair
• Most of the DNA in eukaryotic cells is synthesized by Pol mechanisms that cells have evolved to ensure the fidelity of
Sand Pol£, which take over from Pol a and continue elonga- this very important molecule.
tion of the daughter strand in the 5'~3' direction. PolS syn-
thesizes most of the length of the lagging strand, while Pol £
synthesizes the leading strand. Pol Sand Pol£ remain stably DNA Polymerases Introduce Copying
associated with the template by binding to Rfc protein, Errors and Also Correct Them
which in turn binds to PCNA, a trimeric protein that encir- The first line of defense in preventing mutations is DNA poly-
cles the daughter duplex DNA, functioning as a sliding clamp merase itself. Occasionally, when repliLative DNA polymer-
(see Figure 4-31 ). ases progress along the template DNA, an incorrect nucleotide
• DNA replication generally occurs by a bidirectional mech- is added to the growing 3' end of the daughter strand. E. coli
anism in which two replication forks form at an origin and DNA polymerases, for instance, introduce about 1 incorrect
move in opposite directions, with both template strands be- nucleotide per 104 (ten thousand ) polymerized nucleotides.
ing copied at each fork (see Figure 4-33). Yet the measured mutation rate in bacterial cells is much
lower: about 1 mistake in 109 (one billion) nucleorides incor-
• Synthesis of eukaryotic DNA in vivo is regulated by control-
porated into a growing strand. This remarkable accuracy is
ling the activity of the MCM helicases that initiate DNA repli-
largely due to proofreading by E. coli DNA polymerases. Eu-
cation at multiple origins spaced along chromosomal DNA.
karyotic Pol S and Pol £ employ a similar mechanism.
Proofreading depends on a .3'~5' exonuclease actiuity of
some DNA polymerases. When an incorrect base is incorpo-
rated during DNA synthesis, base pairing between the 3' nu-
cleotide of rhe nascent strand and the template strand does nor
occur. As a result, the polymerase pauses, then transfers the 3'
4.6 DNA Repair and Recombination end of the growing cham to the exonuclease site, where the
Damage to DNA is unavoidable and arises in many ways. incorrect mispaired base is removed (Figure 4-34 ). Then the
DNA damage can be caused by spontaneous cleavage of 3' end is transferred back ro the polymerase site, where this
chemical bonds in DNA, by environmental agents such as region is copied correctly. All three E. coli DNA polymerases
ultraviolet and ionizing radiation, and by reaction with gena- have proofreading activity, as do the two eukaryotic DNA
toxic chemicals that are by-products of normal cellular me- polymerases, S and £, used for replication of most chromo-
tabolism or occur in the environment. A change in the normal somal DNA in animal cells. It seems likely that proofreading IS
DNA sequence, called a mutation, can occur during replica- indispensable for all cells to avoid excessive mutations.
tion when a DNA polymerase inserts the wrong nucleotide as
it reads a damaged template. Mutations also occur at a low
Chemical and Radiation Damage to DNA
frequency as the result of copying errors introduced by DNA
polymerases when they replicate an undamaged template. If Can Lead to Mutations
such mutations were left uncorrected, cells might accumulate DNA is continually subjected to a barrage of damaging
so many mutations that they could no longer function prop- chemical reactions; estimates of the number of DNA damage
erly. In addition, the DNA in germ cells might incur too many events in a single human cell range from 104 to 106 per day!
mutations for viable offspring to be formed. Thus the preven- Even if DNA were not exposed to damaging chemicals, cer-
tion of DNA sequence errors in all types of cells is important tam aspects of DNA structure are inherently unstable. For
for survival, and several cellular mechanisms for repairing example, the bond connecting a purine base to deoxyribose
damaged DNA and correcting sequence errors have evolved. is prone to hydrolysis at a low rate under physiological con-
One of these mechanisms £or repairing double-stranded DNA ditions, leaving a sugar without an attached base. Thus cod-
breaks, by the process of recombination, is also used by eu- ing information is lost, and this can lead to a mutation
karyotic cells to generate new combinations of maternal and during DNA replication. Normal cellular reactions, includ-
paternal genes on each chromosome through the exchange of ing the movement of electrons along the electron-transport
segments of the chromosomes during the production of germ chain in mitochondria and lipid oxidation in perox1somes,
cells (e.g., sperm and eggs). produce several chemicals that react with and damage DNA,
Significantly, defects in DNA repair mechanisms and including hydroxyl radicals and superoxide (0 2 ). These too
cancer are closely related. When repair mechanisms are com- can cause mutations, including those that lead to cancers.
promised, mutations accumulate in the cell's DNA. If these Many spontaneous mutatiom are point mutations, which
mutations affect genes that are normally involved in the involve a change in a single base pair in the DNA sequence.
careful regulation of cell division, cells can begin ro divide This can introduce a stop codon, causing a nonsense muta-
uncontrollably, leading to tumor formation, and cancer. tion as discussed earlier, or a change in the amino acid se-
Chapter 25 outlines in detail how cancer arises from defects quence of an encoded protein, called a missense mutation.
in DNA repair. We will encounter a few examples in this Silent mutations do not change the amino acid sequence (e.g.,
section as well, as we first consider the ways in which DNA GAG to GAA; both encode glutamine). Point mutations can

4.6 DNA Repair and Recombination 151

 ,;, rl 5'
Thumb

I
Fingers
5'
Thumb

- Growin g \ I
Pol
3'
......... Palm
strand
\5'
Pol { P•Im
\\1. ~1
5'

\ill' . ~~
II ~ I \ t> I I I 3' ~~\1'
-3'
' 1 111 1
3'
Template
Exo strand Exo

FIGURE 4-34 Proofreading by DNA polymerase. All DNA base at t he 3' end ca uses melting of the newly formed end of the
polymerases have a sim ilar three-dimensional structure, w hich duplex. As a result, t he polymerase pauses, and the 3' end of the
resembles a half-opened right hand. The "fingers" bind the single- growing strand is transferred to the 3 '~s· exonuclease site (Exo)
stranded segment of the template stran d, and t he polym erase catalytic about 3 nm away, w here t he m ispaired base and probably other bases
activity (Pol) lies in the junction bet ween the fingers and palm. As long are removed. Su bsequently, the 3' end flips back into the polymerase
as the correct nucleotides are added to the 3' end of th e growing site and elongation resumes. [Adapted fro,;, C. M. Joyce and T. T. Steitz, 1995,
strand, it remains in the polymerase site. Incorporation of an incorrect J. Bacterial. 177:6321, and 5. Bell and T. Baker, 1998, Ce/1 92:295.)

also occur in a non-protein-coding DNA sequence that func- T as template to fo rm a U·A o r T·A base pair, thus creating a
tions in the regulation of a gene's transcription, as discussed permanent change to the DNA sequence (Figure 4-35) .
in Chapter 7. One of the most frequent point muta tions
comes from deamination of a cytosine (C) base, which con-
High-Fidelity DNA Excision Repair Systems
verts it into a uracil (U) base. In addition, the common modi-
fied base 5-methyl cytosine forms t hymi ne w hen it is Recognize and Repair Damage
deaminated. If these alteratio ns are not corrected before the In additi on to proofreading, cells have other repa ir systems
DNA is replicated, the cell will use the strand containing U or for preventing mu tations due to copying errors and exposure

NH 2 0
I II
c c
N<"' 'C-CH 3 Deami nation HN/ 'C-CH3
I II I II
0 ""'c' N /c
<"'C'- /CH
0 N
I I
2-Deoxyribose 2-Deoxyribose

S·M ethylcytosine Thymine

5' 3' 5' 3' 5' 3' 5' 3'

Deamination Replication

D fJ
Base-excision
repair
3' 5' 3' 5' 3' 5' 3' 5'
Wild-type Mutant Wild-type
DNA DNA DNA

FIGURE 4-35 Deamination leads to point mutations. A spontane- w ill lead to a perma nent change in seq uence (i.e., a m utation)
ous point mutation can form by deamination of 5-methylcytosine (C) following DNA replication (step fl ). After one round of replicat ion, one
to form thymine tn. If the resulting T·G base pair is not restored to t he daughter DNA molecule will have the mutant T·A base pair and th e
normal C·G base pair by base excision-repai r mechanisms (step 0 ), it other w ill have the wi ld-type C·G base pai r.

152 CHAPTER 4 • Basic Molecular Genetic Mechanisms

to chemicals and radiation. Several DNA excision-repair sys- 5' 3'

tems that normally operate with a high degree of accuracy I I+

have been well studied. These systems were first elucidated I IG
T J II
3'
through a combination of genetic and biochemical studies in 5'
E. coli. Homologs of the key bacterial proteins exist in eukary- ~!DNA glycosylase
otes from yeast to humans, indicating that these error-free
mechanisms arose early in evolution to protect DNA integrity. II I I•
Each of these systems functions in a similar manner-a seg- I I I I9
l~EI
ment of the damaged DNA strand is excised, and the gap is
filled by DNA polymerase and ligase using the complemen- endonuclease
tary DNA strand as template.
We will now turn to a closer look at some of the mecha- ...
""TI....,II"'""'r"-r-1 '-T1....,1.....,....1rl
nisms of DNA repair, ranging from repair of single base muta-
tions to repair of DNA broken across both strands. Some of I I I9I I I I
these accomplish their repairs with great accuracy; others are
less precise. • 1I!! lyase
(part of DNA Pol~)
Base Excision Repairs T-G Mismatches I I II DNA Pol~ c~ I I
and Damaged Bases I I GI
II DNA ligase 9I I
In humans, the most common type of point mutation is a C II Repaired
wild-type DNA
toT, which is caused by deamination of 5-methyl C toT (see
Figure 4-35). The conceptual problem with base excision re- FIGURE 4-36 Base excision repair of a T·G mismatch. A DNA
glycosylase specific for G·T mismatches, usually formed by deamina-
pair in this case is determining which is the normal and
tion of 5-mC residues (see Figure 4-35), flips the thymine base out of
which is the mutant DNA strand, and repairing the latter so
the helix and then cuts it away from the sugar-phosphate DNA
that it is properly base-paired with the normal strand. But
backbone (step 0 ), leaving just the deoxyribose (black dot). An
since a G·T mismatch is almost invariably caused by chemi- endonuclease specific for the resultant baseless site (apurinic endo-
cal conversion of C to U or 5-methyl C toT, the repair sys- nuclease I, APEl) then cuts the DNA backbone (step f) ), and the
tem evolved to remove the T and replace it with a C. deoxyribose phosphate is removed by an endonuclease, apurinic lyase
The G·T mismatch is recognized by a DNA glycosylase (AP lyase), associated with DNA polymerase (3, a specialized DNA
that flips the thymine base out of the helix and then hydro- polymerase used in repair (step IJ). The gap is then filled in by DNA
lyzes the bond that connects it to the sugar-phosphate DNA Pol f3 and sealed by DNA ligase (step ~ ), restoring the original G·C
backbone. Following this initial incision, an apurinic (AP) base pair. [Adapted from 0. Scharer, 2003, Angewandte Chemie 42:2946.)
endonuclease cuts the DNA strand near the abasic site. The
deoxyribose phosphate lacking the base is then removed and
replaced with a C by a specialized repair DNA polymerase
Mismatch Excision Repairs Other Mismatches
that reads the G ln the template strand (Figure 4-36).
As mentioned earlier, this repair must take place prior to and Small Insertions and Deletions
DNA replication, because the incorrect base in this pair, T, oc- Another process, also conserved from bacteria to humans,
curs naturally in normal DNA. Consequently, it would be able principally eliminates base-pair mismatches and insertions or
to engage in normal Watson-Crick base pairing during replica- deletions of one or a few nucleotides that are accidentally in-
tion, generating a stable point mutation that is now unable to troduced by DNA polymerases during replication. As with
be recognized by repair meahanisms (see Figure 4-35, step f)). base excision repair of aT in a T-G mismatch, the conceptual
Human cells contain a battery of glycosylases, each of problem with mismatch excision repair is determining which
which is specific for a different set of chemically modified is the normal and which is the mutant DNA strand, and re-
DNA bases. For example, one removes 8-oxyguanine, an oxi- pairing the latter. How this happens in human cells is not
dized form of guanine, allowing its replacement by an undam- known with certainty. It is thought that the proteins that bind
aged G, and others remove bases modified by alkylating to the mismatched segment of DNA distinguish the template
agents. The resulting nucleotide lacking a base is then replaced and daughter strands; then the mispaired segment of the
by the repair mechanism discussed above. A similar mecha- daughter strand-the one with the replication error-is ex-
nism also functions in the repair of lesions resulting from cised and repairrd to become an exact complement of the tem-
depurmation, the loss of a guanine or adenine base from DNA plate strand (Figure 4-37). In contrast to base excision repair,
resulting from hydrolysis of the glycosylic bond between de- mismatch excision repair occurs after DNA replication.
oxyribose and the base. Depurination occurs spontaneously
and is fairly common in mammals. The resulting abasic sites, Predisposition to a coLon cancer known as hereditary
if left unrepaired, generate mutations during DNA replication nonpolyposis colorectal cancer results from an inherited
because they cannot specify the appropriate paired base. loss-of-function mutation in one copy of either the MLHJ or the

4.6 DNA Repair and Recombination 153

 0 H

+
N
\ / c
Deoxyribose -

I
N
'\_C===C/
H
"
"-
C= O

® O CH 3
~ H
C- N
\
Deoxyribose - N / " C= O

MLH1 endonuclease,
PMS2
I '\_C- = C/
H " cH 3
Two thymine residues
DNA helicase
uv

II
DNA exonuclease
1
irradiation

5
.
I I I I II I AI IIIIIIII . 3
\ 0~ H
~C- N

I 3 I I I I I s·
3
I I I I I 5· Deoxyribose - N<H >c=O
c- c
Gap repair by DNA
polymerase and ligase
a
U
® I I ~CH3
\ 0~ H
\ ~e-N'.
Deoxyribose - N( >c.= O
5
IIIIIII I I IA I I I I I I I I 3. c- c
3
•

~ ~-.I.~..I....~..1.....1---'--..:...1....:..1_1_--=.1 ~ I Ll_l I 5'

I
'
H ~CH
3

Thymine-thymine dim er residue

FIGURE 4-3 7 M ismatch excision repair in human cells. The
mismatch excision-repair pathway corrects errors introduced during FIGURE 4-38 Formation of thymine-thymine dimers. The most
replication. A complex of the MSH2 and MSH6 proteins (bacterial MutS common type of DNA damage caused by UV irradiation, thymine-
homologs 1 and 6) binds to a mispaired segment of DNA in such a way thymine dimers, can be repaired by an excision-repair mechanism.
as to distinguish between the template and newly synthesized
daughter strands (step 0 ). This triggers binding of MLHl and PMS2
(both homologs of bacterial Mutl). The resulting DNA-protein complex
for bulges or other irregularities in the shape of the double
then binds an endonuclease that cuts the newly synthesized daughter
helix. For example, this mechanism repairs thymine-thymine
strand. Next a DNPi helicase unwinds the helix, and an exonuclease
dimers, a common type of damage caused by UV light (Fig-
removes several nucleotides from the cut end of the daughter strand,
including the mismatched base (step f) ). Finally, as with base excision
ure 4-38); these dimers interfere with both replication and
repair, the gap is then filled in by a DNA polymerase (Pol&, in this case) transcription of DNA.
and sealed by DNA ligase (step 10).
Figure 4-39 illustrates how the nucleotide excision-
repair system repairs damaged DNA. Some 30 pro-
MSH2 gene. The MSH2 and MLH 1 proteins are essenrial for teins are involved in this repair process, the first of which
DNA mismatch repair (see Figure 4-37). Cells with at least one were identified through a study of the defects in DNA repair
functional copy of each of these genes exhibit normal mismatch in cultured cells from individuals with xeroderma pigmento-
repair. However, tumor cells frequently arise from those cells sum, a hereditary disease associated with a predisposition to
that have experienced a random mutation in the second copy; cancer. Individuals with this disease frequently develop the
when both copies of one gene are not functional, the mismatch skin cancers called melanomas and squamous cell carcino-
repair system is lost. Inactivating mutations in these genes are mas if their skin is exposed to the UV rays in sunlight. Cells
also common in non inherited forms of colon cancer. • of affected patients lack a functional nucleotide excision-
repair system. Mutations in any of at least seven different
genes, called XP-A through X P-G, lead to inactivation of
Nucleotide Excision Repairs Chemical Adducts this repair system and cause xeroderma pigmentosum; all .,
that Distort Normal DNA Shape produce the same phenotype anJ have the same conse-
Cells use nucleotide excision repair to fix DNA regions con- quences. The roles of most of these XP proteins in nucleotide
taining chemically modified bases, often called chemical ad- excision repair are now well understood (see Figure 4-39). •
ducts, that distort the normal shape of DNA locally. A key
to this type of repair is the ability of certain proteins to slide Remarkably, five polypeptide subunits of TFIIH, a general
along the surface of a double-stranded DNA molecule looking transcription factor required for transcription of all genes

154 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Q. (see Figure 7-16) are also required for nucleotide excision
5'
I I I I I T T I I I I I I 3'
repair in eukaryotic cells. Two of these subunits have homol-
3'
I I I I ~ ~~ ~ I I I I I 5 ogy to helicases, as shown in Figure 4-39. In transcription, the
helicase activity of TFIIH unwinds the DNA helix at the start

Dl'"''''' d•m•g•
recogn1t1on
site, allowing RNA polymerase to initiate (see Figure 7-16). It
appears that nature has used a similar protein assembly in
two different cellular processes that require helicase activity.
The use of shared subunits in transcription and DNA re-
pair may help explain the observation that DNA damage in
5' 3'
higher eukaryotes is repaired at a much faster rate in regions
of the genome being actively transcribed than in nonrran-
3 5
scribed regions-so-called transcription-coupled repair.
Since only a small fraction of the genome is transcribed tn

fJ l
Opoolog of DNA do"blo holl>
any one cell in higher eukaryotes, transcription-coupled re-
pair efficiently directs repair efforts to the most critical re-
gions. In this system, if an RNA polymerase becomes stalled
at a lesion on DNA (e.g., a thymine-thymine dimer), a small
protein, CSB, is recruited to the RNA polymerase; this trig-
gers opening of the DNA helix at that point, recruitment of
TFIIH, and the reactions of steps fJ through 19 depicted in
5' Figure 4-39.
3'

Two Systems Utilize Recombination to Repair

Double-Strand Breaks in DNA
XP-F and XP-G Ionizing radiation (e.g., x- an d-y-radiation) and some anti-
endonucleases cancer drugs cause double-strand breaks in DNA. These are
particularly severe lesions because incorrect rejoining of dou-
ble strands of DNA can lead to gross chromosomal rear-
rangements that can affect the functioning of genes. For
example, incorrect joining could create a "hybrid" gene that
XP-F cut XP-G cut
codes for theN-terminal portion of one amino acid sequence
5
I II " 3

I I I
5
/
1"1'13 fused to the C-terminal portion of a completely different
3
I I I I I I 5•
amino acid sequence; or a chromosomal rearrangement could

Ill DNA polymerase

DNA ligase
bring the promoter of one gene into close proximity to the
coding region of another gene, changing the level or cell type in
which that gene is expressed.
Two systems have evolved to repair double-strand breaks:
5
. I I I t! I 3' homologous recombination, discussed in the next section,
AA and nonhomologous end-ioinmg (N HE]), which is error-
3 I I I I I

prone, since several nucleotides are invariably lost at the

Wild-type DNA
, point of repair.
FIGURE 4-39 Nucleotide excision repair in human cells. A DNA
lesion that causes distortion of the double helix, such as a thymine Error-Pron e Repa ir by Nonhomologous End-Joining The
dimer, is initially recognized by a complex of the XP-C (xeroderma
predominant mechanism for repairing double-strand breaks
pigmentosum C protein) and 238 proteins (step 0 ). This complex then
in multicellular organisms involves rejoining the nonhomol-
recruits transcript ion factor TFIIH, whose helicase subunits, powered
ogous ends of two DNA molecules. Even if the joined DNA
by ATP hydrolysis, partially unwind the double helix. XP-G and RPA
proteins then bind to the complex and further unwind and stabilize the
fragments come from the same chromosome, the repair pro-
helix until a bubble of =25 bases is formed (step f) ). Then XP-G (now cess results in loss of several base pairs at the joining point
acting as an endonuclease) and XP-F, a second endonuclease, cut the (Figure 4-40). Formation of such a possibly mutagenic dele-
damaged strand at points 24-32 bases apart on each side of the lesion tion is one example of how repair of DNA damage can intro-
(step 10). This releases t h e DNA fragment with the damaged bases, duce mutations.
which is degraded to mononucleotides. Finally the gap is filled by DNA Since movement of DNA within the protein-dense nucleus
polymerase exactly as in DNA replication, and the remaining nick is is fairly minimal, the correct ends are generally rejoined to-
sealed by DNA ligase (step 19). [Adapted from J. Hoeijmakers, 2001, Nature gether, albeit with loss of base pairs. Occasionally, however,
411 :366, and 0. Scharer, 2003, Angewandte Chemie 42:2946.] broken ends from different chromosomes are joined together,

4.6 DNA Repair and Recombination 155

 Double-strand break a mutation in one allele of either the BRCA-1 or the BRCA-
~rr~~1-1rr1T1~1 ~~~~-~~----~~ 2 genes that encode proteins participating in this repair pro-
........................1...1....1.. . j I I I"
cess. Loss or inactivation of the second allele inhibits the

II
r ~
DNA-PK

KU80/KU70
heterodimer
homologous recombination repair pathway and thus tends to
induce cancer in mammary or ovarian epithelial cells. Yeasts
can repair double-strand breaks induced by -y-irradiation.
Isolation and analysis of radiation-sensitive (RAD) mutants
that are deficient in this repair system facilitated study of the
process. Virtually all the yeast Rad proteins have homologs
in the human genome, and the h uman and yeast proteins
function in an essentially identical fashion.
A variety of DNA lesions that are not repa ired by mecha-
nisms discussed earlier can be repaired by mecha nisms in
which the damaged sequence is copied from an undamaged
copy of the same or highl y homologous DNA sequence on
II IISXB IIII the homologous chromosome of diplo id organisms or the

·1
sister chromosome following DNA replication in haploid
and diploid organisms. These mechanisms involve an ex-
change of strands between separate DNA molecules and
Ug"'
hence are referred to as DNA recombinatio n.
In addition to providing a mechanism for DNA repair,
IIIIII similar recombination mechanisms generate genetic diversity
IIIIII among the individuals of a species by causing the exchange
FIGURE 4-40 Nonhomologous end-jo ining. When sister chromatids of large regions of chromosomes between the maternal and
are not available to help repair double-strand breaks, nucleotide paternal pair of homologous chromosomes during meiosis,
sequences are butted together that were not apposed in the unbroken the special type of cellular division that generates germ cells
DNA. These DNA ends are usually from the same chromosome locus, and (spe rm and eggs) (Figure 5-3) . In fact, the exchange of re-
when linked together, several base pairs are lost. Occasionally, ends from
gions of homologous chromosomes, called crossing-over, is
different chromosomes are accidentally joined together. A complex of
required for proper segregation of chromosomes during the
two proteins, Ku and DNA-dependent protein kinase (DNA-PI<), binds to
first meiotic cell division. Meiosis and the consequences of
the ends of a double-strand break (step 0 ). After formation of a synapse,
the ends are further processed by nucleases, resulting in removal of a
generating new combinations of maternal and paternal genes
few bases (step f)), and the two double-stranded molecules are ligated on one chromosome by recombination are discussed furt her
together (step IJ). As a result, the double-strand break is repaired, but in Chapter 5. The mechanisms leading to proper segregation
several base pairs at the site of the break are removed. [Adapted from of chromosomes during meiosis are discussed in Chapter 20.
G. Chu, 1997, J. Bioi. Chern. 272:24097; M. Lieber et al., 1997, Curr. Opin. Genet. Here we will focus on the molecular mechanisms of DNA
Deve/. 7:99; and D. van Gant et al., 2001, Nature Rev. Genet. 2:1 %.] recombination, highlighting the exchange of DNA strands
between two recombining DNA molecules.

Repair of a Collapsed Replication Fork An example of recom-

leading to translocation of pieces of DNA from one chromo- binational DNA repair is the repair of a "collapsed" replica-
some to another. Such translocations may generate chimeric tion fork. If a break in the phosphodiester backbone of one
genes that can have drastic effects on normal cell function, DNA strand is not repaired before a replication fork passes,
such as uncontrollable cell growth, which is the hallmark of the replicated portions of the daugh ter chromosomes be-
cancer (see Figure 6-42). The devastating effects of double- come separated when the replication helicase reaches the
strand breaks make these the "most unkindest cuts of all," to break in the parental DNA strand because there are no cova-
borrow a phrase from Shakespeare's julius Caesar. lent bonds between the two fragments of the parental strand
on either side of the nick. This process is called replication
fork collapse (figure 4-41, step 0 ). If it is not repaired, it is
generally lethal to at least one daughter cell following ceil
Homologous Recombination Can Repair DNA
division, because of the loss of genetic information between
Damage and Generate Genetic Diversity the break and the end of the chromosome. The re<.:umbina-
At one time homologous recombination was thought to be a tion process that repairs t he resulting doub le-stranded
minor repair process in human cells. This changed when it break and regenerates a replication fork involves multiple
was realized that several human cancers arc potentiated by enzymes and other proteins, only some of which are men-
inherited mutations in genes essential for homologous tioned here.
recombination repa ir (see Table 25-1 ). For example, some The first step in the repair of the double-strand break is ·.
women with an inherited susceptibility to breast cancer have exonucleolytic digestion of the strand with its 5' end at the

156 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Break in phosphodiester backbone parent chromosome (light blue), as shown in Figure 4-41,
~ I 5
step f). A critical protein required for the next step is RecA
in bacteria, or the homologous Rad51 in S. cerev1siae and
::::::~~~·~~----------1--.---
. R-e-pl-ic-a-ti_o_n_f_o r-k_c_o_ll_a_ps_e-- 3
. other eukaryotes. Multiple RecA/Rad51 molecules bind to
the single-stranded DNA and catalyze its hybridization to a
perfectly or nearly perfectly complementary sequence 111 an-
other, homologous, double-stranded DNA molecule. The
complementary strand of this target double-stranded DNA
========---------
fJ
3
5' (dark blue) is displaced a'> a single-stranded loop of DNA

1 5'-exonuclease acts on over the region of hybridization to the invading strand (Fig-
broken end. Other daughter ure 4-41, step IJ). This RecA/Rad5 I catalyzed invasion of a
strand (pink) ligated to duplex DNA by a single-stranded complement of one of the
repaired parental strand
(light blue) in unbroken strands is key to the recombination process. Since no base
chromosome. pairs are lost or gained in this process, called strand inva-

=======------.;;;;:....-
11 RecA- or Rad51-mediated
5'
3
sion, it does not require an input of energy.
Next, the hybrid region between target DNA and the in-

-------------~
-1-r::)'--+
_ ____.,
1 strand invasion
/ , ._ _ _ _ _ _ _ _ 3'5
,
vading strand is extended in the direction away from the
break by proteins that use energy from ATP hydrolysis. This
process is called branch migration (Figure 4-41, step 0 ) be-
cause the position where the target DNA strand crosses from
Holliday structure
'\~-,
I ,....
1 IJ Branch migration
one complementary strand (dark blue) to its complement in
the broken DNA molecule (dark red), i.e. the pink diagonal
line after step IJ, is called a branch in the DNA structure. In
:- -1 5' this diagram, the diagonal lines represent only one phospho-
~------------------ 3'
===-
'~
I
._j__.J
I diester bond. Molecular modeling and other studies show
that the first base on either side of the branch is base-paired
11 Cut strands at crossover
1 (arrows)

5'
to a complementary nucleotide. As this branch migrates to
the left, the number of base pairs remains constant; one new
base pair formed with the (red) invading strand IS matched

= ==----.J ,------------------- 3' by the loss of one base pair with the parental (blue) strand.

1Iii Ligate ends

When the region of hybrid extends beyond the 5' end of
rhe broken strand (light blue), the single-stranded parental
DNA strand generated (light blue) base pairs with the com-
plementary region of the other parental strand (dark blue)
that becomes single-stranded as the branch migrates to the
left (Figure 4-41, step 0 ). The resulting structure is called a

111 Rebuild replication fork and

continue replication
Holliday structure, after Robin Holliday, the geneticist who
first proposed it as an intermediate in genetic recombination.
Again, the diagonal lines in the diagram following step 0
===::::..::..__....!!;ll/_______ 5'
3' represent single phosphodiester bonds, and all bases in the
Holliday structure are base-paired to complementary bases
FIGURE 4 -41 Recombinational repair of a collapsed replication in the parental strands. Cleavage of the phosphodiester bonds
fork. Parental strands are light ~nd dark blue. The leading daughter that cross over from one parental strand to the other (step H)
strand is dark red, and the lagging daughter strand pink. Diagonal lines and ligation of the 5' and 3' ends base paired to the same
in step IJ and beyond represent a single phosphodiester bond from parental strands (step l'il) result in the generation of a struc-
the DNA strand of the corresponding color. Small black arrows
ture similar to a replication fork. Rebinding of rep lication
following step B represent cleavage of the phosphodiester bonds at
fork proteins results in extension of the leading strand past
the crossover of DNA strands in the Holliday structure. See http://
the point of the original strand break and re-initiation of
www.sheffield.ac.uk/mbb/ruva for an animation of branch migration
lagging-strand synthesis (step f.,i), thus regenerating a repli-
catalyzed by E. coli proteins RuvA and RuvB. See the text for a discus-
sion. [Adapted from D. L. Nelson and M. M. Cox, 2005, Lehninger Principles of
cation fork. The overall process allows the ligated upper
Biochemistry, 4th ed., W. H. Freeman and Company.] strand in the lower molecule following step fJ to serve as
template for extension of the leading strand in step 6 .

broken end of DNA, leaving the strand with a 3' end at the Double-Stra nded DNA Break Re pair by Homo logous Recom-
break single-stranded (Figure 4-41, step f)). The lagging na- bination A similar mechanism called homologous recombi-
scent strand (pink) base paired to the unbroken parent nation can repair a double-strand break in a chromosome
strand (dark blue) is ligated to the unreplicated portion of the and can also exchange large segments of two double-stranded

4.6 DNA Repair and RecombinatiOn 157

 1 D Ends digested by exonucleases, leaving
3 ' single-stranded ends
)

1 lfJ RecA- or Rad51-mediated

strand invasion

I( X
ll 3' end of invading strand is extend'ed by

I
DNA polymerase until the displaced
single-strand (dark blue) base-pairs with
the other 3 ' single strand generat ed
initially (pink)

~ ...... X
111 3' end is extended by DNA polymerase

.. ······>-
/.._...... X

~:======~~-~
-···~
- ~X~======~
1 1!1 Cleav}phosphodiester bonds indicated
with arrows and ligate alternative ends

~----------~--------~~·~·~·~·~·~
··------------~====~----~====~:
FIGURE 4-42 Double-strand DNA break repair by homologous Note that in the diagram of the upper DNA molecule the strand with its
recombination. For simplicity, each DNA double helix is represented by 3' end at the right is on the top, while in the diagram of the lower DNA
two parallel lines with the polarities of the strands indicated by arrow- molecule this strand is drawn on the bottom. See the text for discussion.
heads at their 3' ends. The upper molecule has a double-strand break. [Adapted from T. L. Orr-Weaver and J. W. Szostak, 1985, Microbiol. Rev. 49:33.]

DNA molecules (Figure 4-42). First, the broken ends of the displacing the parental strand as an enlarging single-stranded
DNA molecule are resected by exonucleases that leave a single loop of DNA (dark blue) (step IJ). When the loop extends to
stranded region of DNA with a 3' end (step 0 ). RecA in bac- a sequence that is complementary to the other broken end of
teria and Rad51 in eukaryotes then catalyzes strand invasion DNA (the fragment o n the left following step 0 ), the com-
of one of these 3' ends into the homologous region of the plementary sequences base-pair (diagram following step 1)).
homologous chromosome as discussed above for repair of a This 3' end is then extended by a DNA polymerase using the
collapsed replication fork (step f)). The 3' end of the invad- displaced single-stranded loop of parental DNA (dark blue)
ing DNA strand is then extended by a DNA polymerase, as template (step IJ).

158 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 D

5' 3' 5' 3'

3' ~ 5' ---+ 3'+--, r-- 5·
3' + - - - " ' - 5' 5'______J ~ 3 '
5' 3' 3'
cG
I

FIGURE 4-43 Alternative resolution of a Holliday structure.

Diagonal and vertical lines represent a single phosphodiester bond.
l
ligating the ends as indicated regenerates the original chromosomes.
Cutting the strands as shown in f) and religating as shown at the
It is simplest to diag ram the process by rotating the diagram of the bottom generates recombinant chromosomes. See http:// engels.
bottom molecule 180° so that the top and bottom molecules have genetics.wisc.edu/Holliday/ holliday3D.html for a three-dimensional
the same strand orientations. Cutting the bonds as shown in 0 and animation ofthe Holliday structure and its resolution.

Next, the new 3' ends are ligated (step ~) to the exonu-
clease-digested 5' ends. This generates two Holliday struc- KEY CONCEPTS of Sec ion 4.6
tures in the pa ired molecules (step ~). Branch migration of DNA Repair and Recombination
these Holliday structures can occur in either direction (not
Changes in the DNA sequence result from copying errors
diagrammed). finally, cleavage of the strands at the posi-
and the effects of various physical and chemical agents.
tions shown by the arrows, and ligation of the alternative 5'
and 3' ends at each cleaved Holliday structure generates two Many copying errors that occur during DNA replication
recombinant chromosomes that contain the DNA of one pa- are corrected by the proofreading function of DNA polymer-
rental DNA molecule on one side of the break point (pink ases that can recognize incorrect (mispaired) bases at the 3 '
and red strands), and the DNA of the other parental DNA end of the growing strand and then remove them by an in-
molecu le on the other side of the initial break point (light herent 3'~5' exonuclease activity (see figure 4-34).
and dark blue) (step m ). The region in the immediate vicinity Eukaryotic cells have three excision-repair systems for cor-
of the initial break point forms a heteroduplex, in which one recting mispaired bases and for removing UV-induced thymine
strand from one parent is base-paired ro the complementary thymine dimers or large chemical adducts from DNA. Base
strand of the other parent (pink or red strand base-paired to excision repair, mismatch repair, and nucleotide excision re-
the light or dark blue strand). Base-pair mismatches between pair operate with high accuracy and generally do not introduce
the two parental strands are usually repaired by repair mech- errors.
anisms discussed above ro generate a complementary base
Repair of double-strand breaks by the nonhomologous
pair. In the process, sequence differences between the two
end-joining pathway can link segments of DNA from different
parents are lost, a process referred to as gene conversion.
chromosomes, possibly forming an oncogenic translocation.
Figure 4-43 diagrams .how cleavage of one or the other
The repair mechanism also produces a small deletion, even
pair of strands at the four-way strand junction in the Holliday
when segments from the same chromosome arc jomcd.
structure generates parental or recombinant molecules. This
process, called resolution of the Holiday structure, separates Error-free repair of double-strand breaks in DNA is ac-
DNA molecu les initiall y joined by RecA/Rad51-catalyzed complished by homologous recombination using the undam-
strand invasion. Each Holliday structure in the intermediate aged sister chromatid as template.
following Figure 4-42, step ~' can be cleaved and religated in lnherited defects in the nucleotide excision-repair pathway,
the two possible ways shown by the two sets of small black as in individuals with xeroderma pigmentosum, predispose
arrows in Figure 4-43. Consequently, there arc four possible them to skin cancer. Inherited colon cancer frequentl y i!l a!l-
products of the recombinanon process shown in Figure 4-42. sociated with mutant forms of proteins essential for the mis-
Two of these regenerate the parental chromosomes [with the match repair pathway. Defects in repair by homologous re-
exception of the heteroduplex region at the break point that is combination are associated with inheritance of one mutant
repaired into the sequence of one parent or the other (gene allele of the BRCA-1 or BRCA-2 gene and result in predis-
conversion)!. The other two possible products generate re- position to breast and uterine cancer.
combinant chromsomes as shown in Figure 4-42.

4.6 DNA Repair and Recombi nation 159

4.7 Vi ruses: Parasites of the Cellular viral gene. Because of this structure, a virus is able to encode
all the information for making a relatively large capsid in a
Genetic System small number of genes. This efficient use of genetic informa-
Viruses are obligate, intracellular parasites. They cannot re- tion is important, since only a limited amount of DNA or
produce by themselves and must commandeer a host cell's RNA, and therefore a limited number of genes, can fit into a
machinery to synthesize viral proteins and in some cases to virion capsid. A capsid plus the enclosed nucleic acid is called
replicate the viral genome. RNA viruses, which usually rep- a nucleocapsid.
licate in the host-cell cytoplasm, have an RNA genome, and Nature has found two basic ways of arranging the mul-
DNA viruses, which commonly replicale in Lhe host-cell nu- tiple capsid protein subunits and the viral genume into a nu-
cleus, have a DNA genome (see Figure 4-1 ). Viral genomes cleocapsid. In some viruses, multiple copies of a single coat
may be single- or double-stranded, depending on the specific protein form a helical structure that encloses and protects
type of virus. The entire infectious virus particle, called a the viral RNA or DNA, which runs in a helical groove within
virion, consists of the nucleic acid and an outer shell of pro- the protein tube. Viruses with such a helical nucleocapsid, such
tein that both protects the viral nucleic acid and functions in as tobacco mosaic virus, have a rodlike shape (Figure 4-44a).
the process of host-cell infection. The simplest viruses con- The other major structural type is based on the icosahedron,
tain only enough RNA or DNA to encode four proteins; the a solid, approximately spherical object built of 20 identical
most complex can encode =200 proteins. In addition to faces, each of which is an equilateral triangle (Figure 4-44b).
their obvious importance as causes of disease, viruses are During infection, some icosahedral viruses interact with host
extremely useful as research tools in the study of basic bio- cell-surface receptors via clefts in between the capsid sub-
logical processes, such as those discussed in this chapter. un its, others interact via long fiberlike proteins extending
from the nucleocapsid.
In many DNA bacteriophages, the viral DNA is located
Most Viral Host Ranges Are Narrow within an icosahedral "head" that is attached to a rodlike
The surface of a virion contains many copies of one type of "tail." During infection, viral proteins at the tip of the tail bind
protein that binds specifically to multiple copies of a recep- to host-cell receptors, and then the viral DNA passes down the
tor protein on a host cell. This interaction determines the tail into the cytoplasm of the host cell (Figure 4-44c).
host range-the group of cell types that a virus can infect- In some viruses, the symmetrically arranged nucleocap-
and begins the infection process. Most viruses have a rather sid is covered by an external membrane, or envelope, which
limited host range. consists mainly of a phospholipid bilayer but also contains
A virus that infects only bacteria is called a bacteriophage, one or two types of virus-encoded glycoproteins (Figure
or simply a phage. Viruses that infect animal or plant cells are 4-44d). The phospholipids in the viral envelope are similar
referred to generally as animal viruses or plant viruses. A few to those in the plasma membrane of an infected host cell.
viruses can grow in both plants or animals and the insects The viral envelope is, in fact, derived by budd ing from that
that feed on them. The highly mobile insects serve as vectors membrane, but contains mainly viral glycoproteins, as we
for transferring. such viruses between susceptible animal or will discuss shortly.
plant hosts. Wide host ranges are also characteristic of some
strictly animal viruses, such as vesicular stomatitis virus,
which grows in insect vectors and in many different types of
Viruses Can Be Cloned and Counted ...
mammals. Most animal viruses, however, do not cross phyla, in Plaque Assays
and some (e.g., poliovirus) infect only closely related species The number of infectious viral particles in a sample can be
such as primates. The host-cell range of some animal viruses quantified by a plaque assay. This assay is performed by
is further restricted to a limited number of cell types because culturing a dilute sample of viral particles on a plate covered
only these cells have appropriate surface receptors to which with host cells and then counting the number of local le-
the virions can attach. One example is poliovirus, which only sions, called plaques, that develop (Figure 4-45). A plaque
infects cells in the intestine and, unfortunately, motor neu- develops on the plate wherever a single virion initially infects
rons in the spinal chord, causing paralysis. Another is H IV-1, a single cell. The virus replicates in this initial host cell and
discussed further below, which infects cells essential for the then lyses {ruptures) the cell, releasing many progeny virions
immune response called CD4+ T-lymphocytes, causing AIDS that infect the neighboring cells on the plate. After a few
(see Chapter 23), and cerrain neurons and other cells of the such cycles of infection, enough cells are lysed to produce
central nervous system called glial cells. a visible clear area, or plaque, in the layer of remaining
uninfected cells.
Since all t he progeny virions in a plaque are derived from
Viral Capsids Are Regular Arrays of One
a single parental virus, they constitute a virus clone. This
or a Few Types of Protein type of plaque assay is in standard use for bacterial and ani-
The nucleic acid of a virion is enclosed within a protein coat, mal viruses. Plant viruses can be assayed similarly by count-
or capsid, composed of multiple copies of one protein or a ing local lesions on plant leaves inoculated with viruses.
few different proteins, each of which is encoded by a single Analysis of viral mutants, which are commonly isolated by

160 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 (a) 50 nm (b) 10 nm
l._.l

Poliovirus

Tobacco mosaic virus (d) 50 nm

(c) 50 nm
L_____j

·.

Avian influenza virus

FIGURE 4 - 44 Virion structures. (a) Helical tobacco mosaic virus.

(b) Small icosahedral virus. An icosahedron is composed of 20 equilateral
triangular faces. The example shown is poliovirus. In poliovirus, each
face is built from three capsomeres, outlined in red. The numbers show
how five capsomeres associate at the 12 vertices of the icosahedron.
(c) Bacteriophage T4. (d) Influenza virus, an example of an enveloped
virus. [Part (a): 0. Bradfute, Peter Arnold/Science Photo Library; Part (b) courtesy
ofT. 5. Baker; Part (c): Dept. of Microbiology, Biozentrum/Science Photo Library;
Bacteriophage T4 Part (d) :James Cavallini/Photo Researchers, Inc.]

plaque assays, has contributed extensively to our current un- 2. Penetration-Viral genome crosses the plasma mem-
derstanding of molecular cellular processes. brane. For some viruses, viral proteins packaged inside the
capsid also enter the host cell.
3. Replication-Viral mRNAs arc produced with the aid
Lytic Viral Growth Cycles Lead
of the host-cell transcription machinery (DNA viruses) or
to Death of Host Cells by viral enzymes (RNA viruses). For both types of viruses,
Although details vary anwng different types of viruses, those viral mRNAs are translated by the host-cell translation
that exhibit a lytic cycle of growth proceed through the machinery. Production of multiple copies of the viral
following general stages: genome is carried out either by viral proteins alone or with
the help of host-cell proteins.
1. Adsorption-Virion interacts with a host cell by binding
of multiple copies of capsid protein to specific receptors on 4. Assembly-Viral proteins and replicated genomes asso-
the cell surface. ciate to form progeny virions.

4.7 Viruses: Parasites of the Cellular Genetic System 161

 (a)
Confluent layer of susceptible host cells
growing on surface of a plate
~ I Free virion

--j
d. #I ~
..../_ E. coli
Lysis and
release ~ Adsorption~ chromosome

'".;,i"'r.L --]
l
Add dilute suspension containing vi rus;
after infection, cover layer of cel ls
w ith C~yar; i ncubate

~l'"~ Expression
of vir~ I early
protems
~

~~ ~
Each plaque represents cell lysis initiated by one viral
particle (agar restricts movement so that virus can
· ~ .sr ·
-
• .tr

,
\&_c..
~ l
infect only contiguous cells)
"'- / vira l
(b) ~ proteins
Replication of viral DNA
Expression of v iral late protein s

FIGURE 4-46 Lytic replication cycle of a nonenveloped, bacterial

virus. E. coli bacteriophage T4 has a double-stranded DNA genome
and lacks a membrane envelope. After viral coat proteins at the tip of
the tail in T4 interact with specific receptor proteins on the exterior
of the host cell, the viral genome is injected into the host (step 0 ).
Host-cell enzymes then transcri be viral "early" genes into mRNAs and
subsequently translate these into viral "early" proteins (step fJ). The
early proteins replicate the viral DNA and induce expression of viral
"late" proteins by host-cell enzymes (step D ). The viral late proteins
include capsid and assembly proteins and enzymes that degrade the
host-cell DNA, supplying nucleotides for synthesis of more viral DNA.
Plaque Progeny virions are assembled in the cell (step C l and released (step liJ)
EXPERIMENTAL: FIGURE 4-45 Plaque assay. The plaque assay when viral proteins lyse the cell. Newly liberated viruses initiate
determines the number of infectious particles in a viral suspension. another cycle of infection in other host cells.
(a) Each lesion, or plaque, which develops where a single virion initially
infected a single cell, constitutes a pure viral clone. (b) Plaques on a
lawn of Pseudomonas fluorescens bacteria made by bacteriophage <1>51.
Processing of the viral RNA primary transcript by host-cell
[Part (b) Courtesy of Dr. Pierre ROSSI, Ecole Polytechnique federale de Lausanne
enzymes yields viral mRNA, which is transported to the cyto-
(LBE-EPFL).]
plasm and translated into viral proteins by host-cell ribosomes,
tRNA, and translation factors. The viral proteins are then
5. Release-Infected cell either ruptures suddenly (lysis), transported back into the nucleus, where some of them either
releasing all the newly formed virions at once, or disinte- replicate the viral DNA directly or direct cellular proteins to
grates gradually, with slow release of virions. Both cases replicate the viral DNA, as in the case of SV40 discussed in an
lead to the death of the infected cell. earlier section. Assembly of the capsid proteins with the newly
replicated viral DNA occurs in the nucleus, yielding thousands
Figure 4-46 illustrates the lytic cycle for T4 bacteriophage, a to hundreds of thousands of progeny virions.
nonenveloped D A virus that infects E. coli. Viral capsid Most plant and animal viruses with an RNA genome do
proteins generally are made in large amounts because many not require nuclear functions for lytic replication. In some of
copies of them are required for the assembly of each progeny these viruses, a virus-encoded enzyme that enters the host
virion. In each infected cell, about 100 200 T4 progeny viri- during penetra tiun transcribes the genomic RNA into
ons are produced and released by lysis. mRNAs in the cell cytoplasm. The mRNA is directly trans-
The lytic cycle is somewhat more complicated for DNA lated into viral proteins by the host-cell translation machin-
viruses that infect eukaryotic cells. In most such viruses, the ery. One or more of these proteins then produces additional
DNA genome is transported (with some associated proteins) copies of the viral RNA genome. Finally, progeny genomes
into the cell nucleus. Once inside the nucleus, the viral DNA are assembled with newly synthesized capsid proteins into
is transcnbed into RNA by the host's transcription machinery. progeny virions in the cytoplasm.

162 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Rabies virus
~\ft''¥}'"--
Lipid bilayer,
~ >'.l<; ~ ~ Nucleocapsid protein
~ "" Matrix protein
Genomic RNA . :lS
~ .).. Receptor-binding glycoprotein
">':' '1( • Viral RNA polymerase
~,.V.}.-~"'
1 ~ Adsorption
Budding ~ ~
J.~
't-~'t-'1-'Y-'i't-'t'Jt~ }~_,,..,;t;~l'"''*' ~i~ Virus receptor Cell membrane

~ r../ ~ i A ssociation ~~~ ~ /--- L

Fusion,. iE at membrane Cytosol

~
t- =
fJJ't; ~ Endocytosis

I I\
Progeny capsid
assem bly
Golgi

.---
---
--
r! nu~=~~~;~~
Endosome

Transport (
synthesis
.l-
~ __....... Viral
~-"" mRNA Replication

~
synthesis '--
Transcription
Release

FIGURE 4-47 Lytic replication cycle of an enveloped animal of the endoplasmic reticulum (ER) as it is synthesized on ER-bound
virus. Rabies virus is an enveloped virus with a single-stranded RNA ribosomes (step 6 ). Carbohydrate is added to the large folded domain
genome. The structural components of this virus are depicted at the inside the ER lumen and is modified as the membrane and the
top. After a virion adsorbs to multiple copies of a specific host associated glycoprotein pass through the Golgi apparatus (step [l)).
membrane protein (step 0 ), the cell engulfs it in an endosome (step f)). Vesicles with mature glycoprotein fuse with the host plasma membrane,
A cellular protein in the endosome membrane pumps H ~ ions from the depositing viral glycoprotein on the cell surface with the large
cytosol into the endosome interior. The resulting decrease in en do- receptor-binding domain outside the cell (step li)). Meanwhile, other
somal pH induces a conformational change in the viral glycoprotein, viral mRNAs are translated on host-cell ribosomes into nucleocapsid
leading to fusion of the viral envelope with the endosomallipid bilayer protein, matrix protein, and viral RNA polymerase (step IE). These
membrane and release of the nucleocapsid into the cytosol (steps IJ proteins are assembled with replicated viral genomic RNA (bright red)
and 0 ). Viral RNA polymerase uses ribonucleoside triphosphates in the into progeny nucleocapsids (step ill), which then associate with the
cytosol to replicate the viral RNA genome (step ~ ) and to synthesize cytosolic domain of viral transmembrane glycoproteins in the plasma
viral mRNAs (step 1!1). One of the viral mRNAs encodes the viral membrane (step iE). The plasma membrane is folded around the
transmembrane glycoprotein, which is inserted into the membrane nucleocapsid, forming a "bud" that eventually is released (step ill).

After the synthesis of hundreds to hundreds of thousands for nonenveloped viruses. To illustrate lytic replication of
of new virions has been completed, depending on the type of enveloped viruses, we consider the rabies virus, whose nu-
virus and host cell, most infected bacterial cells and some cleocapsid consists of a single-stranded RNA genome sur-
infected plant and animal cells arc lysed, releasing all the vi- rounded by multiple copies of nucleocapsid protein. Like
rions at once. In many plant and animal viral infections, other lytic RNA viruses, rabies virions are replicated in the
however, no discrete lytic event occurs; rather, the dead host cytoplasm and do not require host-cell nuclear enzymes. As
cell releases the virions as it gradually disintegrates. shown in Figure 4-47, a rabies virion is adsorbed by endocy-
As noted previously, enveloped animal viruses arc sur- tosis, and release of progeny virions occurs by budding from
rounded by an outer phospholipid bilayer derived from the the host-cell plasma membrane. Budding virions are clearly
plasma membrane of host cells and containing abundant visible in electron micrographs of infected cells, as illustrated
viral glycoproteins. The processes of adsorption and release in Figure 4-48. Many tens of thousands of progeny vmons
of enveloped viruses differ substantially from these processes bud from an infected host cell before it dies.

4.7 Viruses: Parasites of the Cellular Genetic System 163

 more common transcription of DNA into RNA. In the retro-
viral life cycle (Figure 4-49), a viral enzyme called reverse
transcriptase initially copies the viral RNA genome into sin-
gle-stranded DNA complementary to the virion RNA; the
same enzyme then catalyzes synthesis of a complementary
DNA strand. (Th is complex reaction is detailed in Chapter 6
when we consider closely related intracellular parasites
called retrotransposons.) The resulting double-stranded
DNA is integrated into the chromosomal DNA of the in-
fected cell. Finally, the integrated DNA, called a provirus, is
transcribed by the cell's own machinery into RNA, which
either is translated into viral proteins or is packaged within
virion coat proteins to form progeny virions that are released
by budding from the host-cell membrane. Because most ret-
roviruses do not kill their host cells, infected cells can repli-
cate, prod ucing daughter cells with integrated proviral DNA.
These daughter cells continue {o transcribe the proviral
DNA and bud progeny virions.
Some retroviruses contain cancer-causing genes (onco- ·.
genes), and cells infected by such retroviruses are oncogeni-
call y transformed into tumor cells. Studies of oncogenic
retroviruses (mostly viruses of birds and mice) have revealed
EX PER MEN"~" AL r GURE 4-48 Release of progeny virions by
a great deal about the processes that lead to transformation
budding. Progeny virions of enveloped viruses are released by
budding from infected cells. In this transmission electron micrograph
of a normal cell into a cancer cell (Chapter 24).
of a cell infected with measles virus, virion buds are clearly visible
protruding from the cell surface. Measles virus is an enveloped RNA
virus with a helical nucleocapsid, like rabies virus, and replicates as Among the known human retroviruses are human T-cell
illustrated in Figure 4-47. [From A. Levine, 1991, Viruses, Scientific American lymphotrophic virus (HTLV), which causes a form of
Library, p. 22.] leukemia, and human immunodeficiency virus (HIV-1), which
causes acquired immune deficiency syndrome (AIDS). Both
of these viruses can infect only specific cell types, primarily
Viral DNA Is Integrated into the Host-Cell certain cells of the immune system and, in the case of HIV-1,
some central nervous system neurons and glial cells. Only
Genome in Some Nonlytic Viral
these cells have cell-surface receptors that interact with viral
Growth Cycles envelope proteins, accounting for the host-cell specificity of
Some bacterial viruses, called temperate phages, can estab- these viruses. Unlike most other retroviruses, HIV-1 eventu-
lish a nonlytic association with their host cells that does not ally kills its host cells. The eventual death of large numbers
kill the cell. For example, when bacteriophage "11. infects £. of immune-system cells results in the defective immune re-
coli, most of the time it causes a lytic infection. Occasionally, sponse characteristic of AIDS.
however, the viral DNA is integrated into the host-cell chro- Some DNA viruses also can integrate into a host-cell chro-
mosome rather than being replicated. The integrated viral mosome. One example is the human papillomaviruses {HPVs),
DNA, called a prophage, is replicated as part of the cell's which most commonly cause warts and other benign skin
DNA from one host-cell generation to the next. This phe- lesions. The genomes of certain HPV serotypes, however,
nomenon is referred to as lysogeny. If the host cell suffers occasionally integrate into the chromosomal DNA of infected
extens1ve damage to its DNA from ultraviolet light, the pro- cervical epithelial cells, initiating development of cervical can-
phage DNA is activated, leading to its excision from the cer. Routine Pap smears can detect cells in the early stages of
host-cell chromosome, entrance into the lytic cycle, and sub- the transformation process initiated by HPV integration, per-
sequent production and release of progeny virions before the mitting effective treatment before cancer develops. A vaccine
host cell dies. for the types of HPV associated with cervical cancer has been
The genomes of a number of animal viruses also can in- developed and can protect against the initial infection by these
tegrate mro the host-cell genome. Amuug the most impor- viruses, and consequently, agamst development of cervical
tant arc the retroviruses, which are enveloped viruses with a cancer. However, once an individual is infected with these
genome consisting of two identical strands of RNA. These HPVs, the "window of opportunity" is missed, and the vac-
viruses are known as retroviruses because their RNA ge- cine does not protect against the development of cancer. Be-
nome acts as a template for formation of a DNA molecule- cause the vaccine is not 100 percent effective, even vaccinated
the opposite flow of genetic information compared with the women should have regular Pap smears. •

164 CHAPTER 4 • Basic Molecular Ge netic Mechanisms

 ~ OVERVIEW ANIMATION: Life Cycle of a Retrovirus

Host-cell
chromosomal DNA

Nucleoca psid

Reverse
transcription
J
T
Provirus
Transport to
nucleus and ,( ~
integration ~
Viral DNA

FIGURE 4 -49 Retroviral life cycle. Retroviruses have a genome of into one of many possible sites in the host-cell chromosomal DNA. For
two identical copies of single-stranded RNA and an outer envelope. simplicity, only one host-cell chromosome is depicted. Step D: The
Step 0 : After viral glycoproteins in the envelope interact with a specific integrated viral DNA (provirus) is transcribed by the host-cell RNA
host-cell membrane protein, the retroviral envelope fuses directly with polymerase, generating mRNAs (dark red) and genomic RNA molecules
the plasma membrane, allowing entry of the nucleocapsid into the (bright red). The host-cell machinery translates the viral mRNAs into
cytoplasm of the cell. Step f): Viral reverse transcriptase and other glycoproteins and nucleocapsid proteins. Step 1!'1: Progeny vi rions then
proteins copy the viral ssRNA genome into a double-stranded DNA. assemble and are released by budding as illustrated in Figure 4-47.
Step ID :The viral dsDNA is transported into the nucleus and integrated

KEY CONCEPTS of Section 4.7 Lytic viral infection entails adsorption, penetration, syn-
thesis of viral proteins and progeny genomes (replication),
Viruses: Parasites of the Cellular Genetic System assembly of progeny virions, and release of hundreds to thou-
• Viruses are small parasites that can replicate only in host sands of virions, leading to death of the host cell (see Figure
cells. Viral genomes may be either DNA (DNA viruses) or 4-46 ). Release of enveloped viruses occurs by budding
RNA (RNA viruses) and either single- or double-stranded. through the host-cell plasma membrane (see Figure 4-47).
The capsid, which surrounds the viral genome, is com- • Nonlytic infection occurs when the viral genome is inte-
posed of multiple copies 6f one or a small number of virus- grated into the host-cell DNA and generally does not lead to
encoded proteins. Some viruses also have an outer envelope, cell death.
which is similar to the plasma membrane hut contains viral
transmembrane proteins. • Retroviruses are enveloped animal viruses containing a
• Most animal and plant DNA viruses require host-cell nu- single-stranded RNA genome. After a host cell is penetrated,
reverse transcriptase, a viral enzyme carried in the virion,
·. clear enzymes to carry out transcription of the viral genome
into mRNA and production of progeny genomes. In con- converts the viral RNA genome into doubl e-stranded
trast, most RNA viruses encode enzymes that can transcribe DNA, which integrates into chromosomal DNA (see fig-
the RNA genome inLU viral mRNA and produce new copies ure 4-49).
of the RNA genome. Unlike infection by other retroviruses, HIV infection even-
• Host-cell ribosomes, tRNAs, and translation factors are tually kills host cells, causing the defects in the immune re-
used in the synthesis of all viral proteins in infected cells . sponse characteristic of AIDS.

4.7 Viruses: Parasites of the Cellular Genetic System 165

• Tumor viruses, which contain oncogenes, may have an RNA Key Terms
genome (e.g., human T-cell lymphotrophic virus) or a DNA anticodon 131 Okazaki fragments 147
genome (e.g., human papillomaviruses). In the case of these vi-
codon 131 phosphodiester bond 118
ruses, integration of the viral genome into a host-cell chromo-
some can cause transformation of the cell into a tumor cell. complementary 119 polyribosome 142
crossing over 156 primary transcript 127
deamination 152 primer 145
depurination 153 promoter 124
DNA end-joining 155 reading frame 132
Perspectives for the Future
DNA polymerase 145 recombination 116
The basic cellular molecular genetic processes discussed in double helix 118 replication fork 145
this chapter form the foundation of contemporary molecu-
excision-repair systems I 53 retroviruses 164
lar cell biology. Our current understanding of these pro-
cesses is grounded in a wealth of experimental results and is exon 127 reverse transcriptasc 164
not likely to change. However, the depth of our understand- gene conversion 159 ribosomal RNA (rRNA )
ing will continue to increase as additional details of the genetic code 13 I t16
structures and interactions of the macromolecular machines Holliday structure 157 ribosome 131
involved are uncovered. The determination in recent years RNA polymerase 124
homologous
of the three-dimensional structures of RNA polymerases,
recombination 157 thymine-thymine
ribosomal subunits, and DNA replication proteins has al-
intron 127 dimer 154
lowed researchers to design ever more penetrating experi-
isoform 130 transcription 116
mental approaches for revealing how these macromolecules
operate at the molecular level. The detailed level of under- lagging strand 146 transfer RNA (tRNA ) 116
standing currently being developed may allow the design of leading strand 146 translation 116
new and more effective drugs for treating illnesses of hu- messenger RNA viral envelope 160
mans, crops, and livestock. For example, the recent high- (mRNA) 116 Watson-Crick base pairs 118
resolution structures of ribosomes are providing insights
mutation 151
into the mechanism by which antibiotics inhibit bacterial
protein synthesis without affecting the function of mamma-
lian ribosomes. This new knowledge may allow the design
of even more effective antibiotics. Similarly, detailed under-
standing of the mechanisms regulating transcription of spe-
cific human geoes may lead to therapeutic strategies that Review the Concepts
can reduce or prevent inappropriate immune responses that 1. What are Watson-Crick base pairs? Why are they impor-
lead to multiple sclerosis and arthritis, the inappropriate cell tant?
division that is the hallmark of cancer, and other pathologi-
2. Preparing plasmid (double-stranded, circular) DNA for
cal processes.
sequencing involves annealing a complementary, short, single-
Much of current biological research is focused on dis-
stranded oligonucleotide DNA primer to one strand of the
covering how molecular interactions endow cells with deci-
plasmid template. This is routinely accomplished by heating
sion-making capacity and their special properties. For this
the plasmid DNA and primer to 90 oc and then slowly
reason, several of the following chapters describe current
bringing the temperature down to 25 oc. Why does this
knowledge about how such interactions regulate transcrip-
protocol work?
tion and protein synthesis in multicellular organisms and
how such regulation endows cells with the capacity to perform 3. What difference between RNA and DNA helps to explain
their specialized functions. Other chapters deal with how the greater stability of DNA? What implications does this
protein-protein interactions underlie the construction of spe- have for the function of DNA?
cialized organelles in cells, and how they determine cell 4. What are the major differences in the synthesis and struc-
shape and movement. The rapid advances in molecular cell ture of prokaryotic and eukaryotic mRNAs?
biology in recent years hold promise that in the not too dis- 5. While investigating the function of a specific growth fac-
tant future we will understand more deeply how the regula tor receptor gene from human~, rc~t::archers found that two
tion of specialized cell function, shape, and mobility, coupled types of proteins are synthesized from this gene. A larger
with regulated cell replication and cell death (apoptosis), protein containing a membrane-spanning domain functions
lead to the growth of complex organisms such as flowering to recognize growth factors at the cell surface, stimulating a
plants and human beings. specific downstream signaling pathway. In contrast, a related,

166 CHAPTER 4 • Basic Molecular Genetic Mechanisms

smaller protein is secreted from the cell and functions to bind 15. You have learned about the events surrounding DNA
available growth factor circulating in the blood, thus inhibit- replication and the central dogma. Identify the steps associ-
ing the downstream signaling pathway. Speculate on how ated with these processes that will be adversely affected in
the cell synthesizes these disparate proteins. the following scenarios.
6. The transcription of many bacterial genes relies on func· a. Helicases unwind the DNA, but stabilizing proteins
tiona! groups called operons, such as the tryptophan operon are mutated and cannot bind to the DNA.
(Figure 4-13a). What is an operon? What advantages are b. The mRNA molecule forms a hairpin loop on itself
there to having genes arranged in an operon, compared with v1a complementary base pairing in an area spanning the
the arrangement in eubryotes? AUG start sire.
7. How would a mutation in the poly(A)-binding protein c. The cell is unable to produce functional tRNA,mer.
gene affect translation? How would an electron micrograph 16. Use the key provided below to determine the amino acid
of polyribosomes from such a mutant differ from the normal sequence of the polypeptide produced from the following
pattern? DNA sequence. lntron sequences are highlighted. Note: Not
8. What characteristic of DNA results in the requirement all amino acids in the key will be used.
that some DNA synthesis is discontinuous? How are Oka- 5'TTCTAAACGCATGAAGCACCGTCTCAGAGCCAGTGA3'
zaki fragments and DNA ligase utilized by the cell? 3'AAGATTTGCGTACTTCGTGGCAGAGTCTCGGTCACT5'
9. Eukaryotes have repair systems that prevent mutations
due to copying errors and exposure to mutagens. What are
Direction of DNA unwinding
the three excision-repair systems found in eukaryotcs, and
which one is responsible for correcting thymine-thymine di-
mers that form as a result of UV light damage to DNA? Asn = AAU Cys = TCG Gly = CAG H1s CAU Lys ' AAC.
.\let "' AUG Phe - L UC Scr = ACC Tyr = UAC Val ' GUC; GUA
10. DNA-repair systems are responsible for maintaining ge-
nomic fideliry in normal cells despite the high frequency with
which mutational events occur. What type of DNA mutation
is generated by (a) UV irradiation and (b) ionizing radiation?
5'
Describe the system responsible for repairing each of these
types of mutations in mammalian cells. Postulate why a loss 3' .............................,..~~JlJ.;Llw_
of function in one or more DNA-repair systems typifies many
cancers.
11. What is the name given to the process that can repair 17. a. Look at the figure above. Explain why it is necessary
DNA damage and generate genetic diversity? Briefly describe for Okazaki fragments to be formed as the lagging strand is
the similarities and differences of the two processes. produced (instead of a continuous strand).
12. The genome of a retrovirus can integrate into the host- b. If the DNA polymerase in the figure above could only
cell genome. What gene is unique to retroviruses, and why is bind to the lower template strand, under what conditions(s)
the protein encoded by this gene absolutely necessary for would it be able to produce a leading strand?
maintaining the retrovirallife cycle? A number of retrovi- 18. The DNA repair systems preferentially target the newly
ruses can infect certain human cells. List two of them, briefly synthesized strand. Why is this important?
describe the medical implications resulting from these infec- 19. Identify the specific types of point mutations below (you
tions, and describe why only certain cells are infected. are viewing the direct DNA version of the RNA sequence).
13. a. Which of the following DNA strands, the top or bot-
tom, would serve as a template for RNA transcription if the Original sequence: 5' AUG TCA GGA CGT CAC TCA GCT 3'
DNA molecule were to ul'lwind in the indicated direction? Mutation A: 5' AUG TCA GGA CGT CAC TGA GCT 3'
Mutation B: 5' AUA TCA GGA CGT CAC TCA GCT 3'

5' ACGGACTGTACCGCTGAAGTCATGGACGCTCGA 3' 20. a. Detail the key differences between lytic and nonlytic
3 'TGCCTGACATGGCGACTTCAGTACCTGCGAGCT 5' viral infection and provide an example of each.
b. Which of the following processes occurs in both lytic
and nonlytic viral infections?
Direction of DNA unwinding (i) Infected cell ruptures to release viral particles.
(ii) Viral mRNAs are transcribed by the host-cell trans-
b. What would be the resulting RNA sequence (written lation machinery.
5' to 3')? (iii) Viral proteins and nucleic acids are packaged to pro-
14. Contrast prokaryotic and euka ryotic gene characteristics. duce virions.

Review the Concepts 167

Analyze the Data c c c c A
-3 u u A A c
Protein synthesis in eukaryotes normally begins at the first u u c c c
AUG codon in the mRNA. Sometimes, however, the ribo-
c c A
A A A A
somes do not begin protein synthesis at this first AUG but AUG 1 {A u u u u
scan past it (leaky scanning), and protein synthesis begins
preCAT ~ G G G G
instead at an internal AUG. In order to understand what +-4 G G A
fearures of an mRNA affect efficiency of initiation at the first
preCAT--. - --~- -
AUG, studie~ have been undertaken in which the synthesis of CAT-._.
chloramphenicol acetyltransferase was examined. Transla- Lane 1 2 3 4 5
tion of its message can give rise to a protein referred to as
preCAT or give rise to a slightly smaller protein, CAT (see
b. What are some additional alterations to this message,
M . Kozak, 2005, Gene 361:13). The two proteins differ in
other than those shown in the figure, that would further elu-
that CAT lacks several amino acids found at theN-terminus
cidate the importance of the ACCAUGG sequence as an op-
of preCAT. CAT is not derived by cleavage of preCAT but,
timal context for synthesis of preCAT rather than CAT?
instead, by initiation of translation of the mRNA at an inter-
How would you further examine :vhether A at the (-3) po-
nal AUG:
sition and Gat the (+4) position are the most important
nucleotides to provide context for the AUG start?
c. A mutation causing a severe blood disease has been
found in a single family (see T. Matthes et al., 2004, Blood
preCAT CAT
Start Start Stop
104:2181 ). The mutation, shown in red in the figure below,
has been mapped to the 5 '-untranslated region of the gene
t t t encoding hepcidin and has been found ra alter the gene's
AUG AUG UAA AAAAn
mRNA. The shaded regions indicate the coding sequence of
2
the normal and mutant genes. No hepcidin is produced from
the altered mRNA, and lack of hepcidin results in the dis-
ease. Can you provide a reasonable explanation for the lack
of synthesis of hepcidin in the family members who have
a. Results from a number of studies have given rise to
inherited this mutation? What can you deduce about the im-
the hypothesis that the sequence (-3)ACCAUGG(+4), in
portance of the context in which the start site for initiation
which the start codon AUG is shown in boldface, provides
of protein synthesis occurs in this case?
an optimal context for initiation of protein synthesi~ and en-
sures that ribosomes do nor scan past this first AUG to begin
initiation instea.d at a downstream AUG. In the numbering Start hepcidin
scheme used here, the A of the AUG initiation is designated t
..... GCAG UGGGACAGCCAGACAGACGGCACG -'~ 1 GGCACUG..... Normal
(11); bases 5' of this are given negative numbers [so that the
first base of this sequence is (-3)1, and bases 3' to the (+1) !
..... GCA/ ,UGGGACAGCCAGACAGACGGCACGAUGGCACU........ Mutant
A are given positive numbers lso that the last base of this
sequence is (+4)]. To test the hypothesis that the start site
sequence (-3)ACCAUGG(+4) prevents leaky scanning, the
chloramphenicol acctyltransferase mRNA sequence was modi- References
fied and the resulting effects on translation assessed. In the
Structure of Nucleic Acids
following figure, the sequence (red) surrounding the first
Arnorr, S. 2006. Historical article: DNA polymorph1sm and the
AUG codon (black) of the mRNA that gives rise to the syn-
early history of the double helix. Trends Bwcllem. Set. 31:349-354.
thesis of preCAT is shown above lane 3. Modification of this Berger, J. :V1., and j. C. Wang. 1996. Recent development~ in
message is shown above the other gel lanes (altered nucleotides DNA ropoisomerase II structure and mechanism. Curr. Opin. Struc.
are in blue), and the completed proteins generated from each Bioi. 6:84-90.
modified message appear as bands on the 50S-polyacrylamide Cech, T. R. 2009. Evolution of h1ological catalysis: nbozyme to
gel below. The intensity of each band is an indication of tbe RNP enzyme. Cold Spring Harbor Symp. Quant. Rio/. 74:11-16.
amount of that protein synthesized. Ana lyze the alterations Dickerson, R. E. 1992. DNA Structure from A to Z. Methods
to the wild-type sequence, and describe how the)' affect l:n;:,ymol. 211:6..,-111.
Kornberg, A., and T. A. Baker. 2005. DNA Replrcatwn.
translation. Are the positions of some nucleotides more im-
University Science, chap. I. A good summary of the principles of
portant than others? Do the data shown in this figure pro- DNA structure.
vide support for the hypothesis that the context in which the L11ley, D. M. 2005. Structure, foldmg and mechanisms of
first AUG is present affects efficiency of translation from this ribozymes. Curr. Opin. Struc. Bioi. 15:313-323.
site? ls ACCAUGG an optimal context for initiation from Vicens, Q., and T. R. Cech. 2005. Atomic level architecture of
the first AUG? group I introm revealed. Trends Btocllem. Set. 31:4 1-51.

168 CHAPTER 4 • Basic Molecular Genetic Mechanisms

 Wang, J. C. 1980. Superhelical DNA. Trends Btochem. Set. Gai, D., Y. P. Chang, and X. ). Chen. 20 I 0. Origm D:t\A
5:219-221. melting and unwinding in DNA replication. Curr. Opm. Struc. Btal.
Wigley, D. B. 1995. Structure and mechanism of DNA topo- 20(6):756-762.
isomerases. Ann. Rev. Biophys. Biomol. Struc. 24:185-208. Kornberg, A., and T. A. Baker. 2005. DNA Repltcation.
University Science.
Transcription of Protein-Coding Genes and Formation Langston, L. D., C. Indiani, and M. O'Donnell. 2009. Whither
of Functional mRNA rhe replisome: emergmg perspectives on the dynamic narure of the
Brenner, S., F. Jacob, and M. Meselson. 1961. An unstable DNA replication machinery. Cell Cycle 8:2686-2691.
intermediate carrying information from genes w nbosomes for Langston, L. D., and M. O'Donnell. 2006. DNA replication:
protein synthesis. Nature 190:576-581. keep moving and don't mmd the gap. Mol. Cells 23:155-160.
Brueckner F., J. Orriz, and P. Cramer 2009. A mov1e of the Schoeffler, A. j., and j. M. Berger. 2008. DNA ropoisomerases:
RNA polymerase nucleotide addition cycle. Curr. Opin. Struc. Bioi. harnessing and constraining energy to govern chromosome topology.
19:294-299. Quart. Rev. Blophys. 41:41-101.
Murakami, K. S., and S. A. Darst. 2003. Bacterial RNA Stillman, B. 2008. DNA polymerases ar the replication fork in
polymerases: rhe whole story. Curr. Opm. Struc. 810/. 13:31-39. eukaryores. Cell 30:259-260.
Okamoto K., Y. Sugino, and M. Nomura. 1962. Synthesis and
turnover of phage messenger RNA in E. coli infected with bacterio-
phage T4 in the presence of chloromycerin. J. Mol. Bioi. 5:527-534. DNA Repair and Recombination
Steitz, T. A. 2006. Visualizing polynucleotide polymerase Andressoo, J. 0., and .J. H. Hoeijmakers. 2005. Transcnption-
machines at work. EMBO j . 25:3458-3468. coupled repair and premature aging. Mutat. Res. 577:179-194.
Barnes, D. E., and T. Lindahl. 2004. Repmr and genenc
The Decoding of mRNA by tRNAs consequences of endogenous DNA base damage in mammalian cells.
Alexander, R. W., and P. Schimmel. 2001. Domain-domain Ann. Ret•. Genet. 38:445-4""'6.
communication in aminoacyl-tRNA syntherases. Prog. Nucl. Aetd Bell, C. E. 2005. Strucrure and mechanism of F.schericlna culi

.... Res. Mol. Bioi. 69:317-349.

Hatfield, D. l.., and V. X Gladyshev. 2002. How selenium has
RecA ATPase. Mol. !vlicrobiol. 58:358-366.
Friedberg, E. C., er al. 2006. DNA repair: from molecular
altered our understanding of the generic code. Mol. Cell Bioi. mechanism to human disease. DNA Repair 5:986-996.
22:3565-3576. Haber, j. E. 2000. Partners and pathways repa1ring a double-
Hoagland, M . B., et al. 1958. A soluble ribonucleic acid strand break. Trends Genet. 16:259-264.
intermediate in protein synthesis. J. Btal. Chern. 231:241-257. Jiricny, J. 2006. The multifaceted m1smarch-repa1r system. Nat.
lbba, M., and D. Soli. 2004. Aminoacyl-rRNAs: setting the Rev. Mol. Cell Bioi. 7:335-346.
limits of the genetic code. Genes Dev. 18:731-738. Khuu, P. A., et al. 2006. The stacked-X DNA Holhday JUnction
Khorana, G. H., er al. 1966. Polynucleonde syntheSIS and the and protem recognmon. J. Mol. Recog. 19:234-242.
genetic code. Cold Spring Harbor Symp. Quant. 810/. 31:39-49. Lilley, D. vl., and R. M. Clegg. 1993. The structure of rhe
Nakanishi, K., and 0. Nureki. 2005. Recent progress of structural four-way JUncnon in DNA. Annu. Rev. B10phys. Bl()mol. Struc.
biology of tRNA processing and modification. Mol. Cells 19:157-166. 22:299-328.
Nirenberg, M., er al. 1966. The RNA code in protein syntheSIS. Mirchandani, K. D., and A. D. D'Andrea. 2006. The Fancom
Cold Spring Harbor Symp. Quant. Btal. 31:11-24. anemia/BRCA pathway: a coordinator of cross-link repa1r. F.xp.
Rich, A., and S.-H. Kim. 1978. The three-dimensional structure Cell Res. 312:2647-2653.
of transfer RNA. Sci. Am. 240(1):52-62 (offprint 1377). Mitchell, J. R., J. H. Hoeijmakers, and L. J. Niedernhofcr. 2003.
DIVIde and conquer: nucleotide excision repatr battles cancer and
Stepwise Synthesis of Proteins on Ribosomes aging. Curr. Opin. Cell Bio/.15:232-240.
Belousoff, M. J., er al. 2010. Ancient machinery embedded in Orr-Weaver, T. L., and J. W. S10stak. 1985. Fungal recombina-
the contemporary ribosome. B10chem. Soc. Trans. 38:422-427. tion. Microbiol. Rev. 49:33-58.
Frank, j., and R. L. Gonzalez, Jr. 2010. Structure and dynamics Shin, D. S., er al. 2004. Structure and function of the double
of a processive Brownian motor: the translarmg ribosome. Amtu. strand break repair machinery. DNA Repatr 3:863-873.
Rev. Biochem. 79:381-412. Wood, R. D., M. Mitchell, and T. l.mdahl. Human DNA repa1r
Jackson, R.j., er al. 2010. The mechanism of eukaryotic genes. Mutat. Res. 577:275-283.
translation initiation and prin~iples of its regulation. Nat. Rev. Mol. Yoshida, K., andY. Miki. 2004. Role of BRCA land BRCA2 as
Cell Bioi. 11:113-127. regulators of DNA repair, transcnprion, and cell cycle m response ro
Korostelev, A., D. N. Ermolenko, and H. F. Noller. 2008. DNA damage. Cancer Sci. 95:866-871.
Structural dynamics of the ribosome. Cttrr. Opin. Chem. Bioi.
12:674-683.
Livingstone, M., et al. 2010. Mechanisms governing the control Viruses: Parasites of the Cellular Genetic System
of mRNA translation. Phys. B10/. 7(2):021001. Flint, S. ]., er al. 2000. Principles of Virology: Molewlar
Sarnow, P., R. C. Cevallos, and F. Jan. 2005. Takeover of host Biology, Pathogenests, and Control. AS.'vt Press.
nbosomes by divergent IRES clements. Biochem. Soc. Trans. Hull, R. 2002. Mathews· Plant Virology. Academic Press.
33:1479-1482. Klug, A. ~~~~. I he tobacco mosaic virus particle: strucrure and
Steitz, T. A. 2008. A structural understanding of the dynamtc assembly. Pht!. Trans. R. Soc. Lond. B Btol. Set. 354:531-535.
ribosome machine. Nat. Rev. Mol. Cell Bioi. 9:242-253. Knipe, D. M., and P.M. Howley, eds. 2001. he/ds Vrrology.
L1ppincorr Williams & Wilkms.
DNA Replication Kornberg, A., and T. A. Baker. 1992. DNA Replicatton, 2d ed.
DePamphilis, M. L., ed. 2006. DNA Replicatton and Human W. H. freeman and Company. Good summaq of bacrenophage
Disease. Cold Spnng Harbor Laboratory Press. molecular biology.

References 169
L
 CH APTER

Molecular Genetic
Techniques
The Planarian is a freshwater flatworm with an amazing capacity for
regeneration. If the head and tail of an adult Planarian are sliced off, the
worm will readily regenerate these structures (as shown in the upper
left frame). The role of specific genes in the regeneration process can
be tested by shutting off gene expression by RNA interference (RNAi)
before cutting off the head and tail. The remaining eight frames show
the variety of regeneration defects that can be observed after RNAi of
d ifferent genes responsible for regeneration. The genes inhibited by
RNAi are from left to right starting at the upper center panel are:
smad4, (3-catenin-1, carcinoma antigen, POU2/ 3, rootletin, Novel.
tolloid, and piwi. [Courtesy of Peter Reddien/MIT, Whitehead Institute.]

n the field of molecular cell biology, reduced to irs most where and when the encoded protein is expressed in an organ-

I basic clements, we seek an understanding of the biological

behavior of cells in terms of the underlying chemical and
molecular mechanisms. Often the investigation of a new mo-
ism. The second strategy follows essentially the same steps as the
classical approach bur in reverse order, beginning with isolation
of an interesting protein or its identification based on analysis of
lecular process focuses on the function of a particular protein. an organism's genomic sequence. Once the corresponding gene
There arc three fundamental questions that cell biologists usu- has been isolated, the gene can be altered and then reinserted into
ally ask about a newly discovered protein: what is the function an organism. In both strategies, by examining the phenotypic
of the protein in the context of a living cell, what is the bio- consequences of mutations t hat inactivate a particular gene, ge-
chemical fu nction of the p urified protein, and where is the pro- neticists are able to connect knowledge about the sequence,
tein located? To answer these questions, investigators employ structure, and biochemical activity of the encoded protein to its
three molecular genetic tools: the gene t hat encodes the protein, function in the context of a living cell or multicellular organism.
a mutant cell line or organism that lacks the functional protein, An important component in both strategies for studying
and a source of the purified protein for biochemical studies. In a protein and its biological function is isolation of the corre-
this chapter, we consider various aspects of two basic experi- sponding gene. Thus we discuss va rious techniques by which
mental strategies for obtaining all three tools (Figure 5-1 ). researchers can isolate, sequence, and manipulate specific
The first strategy, often referred to as classical genetics, be- regions of an organism's DNA. Next we introduce a variety
gins with isolation of a mutant that appears to be defective in of techn iques that are commonly used to analyze where and
some process of interest. Gtneric methods then are used to iden- when a particular gene is expressed and where in the cell its
tify and isolate the affected gene. The isolated gene can be ma- protein is localized . In some cases, knowledge of protein
nipulated to produce large quantities of the p rotein for function can lead to significant medical advances, and the
biochemical experiments and to design probes for studies of first step in developing treatments for an inherited disease is

OUTLINE

5.1 Genetic Analysis of Mutations to Identify 5.4 Locating and Identifying Human
and Study Genes 172 Disease Genes 206

5.2 DNA Cloning and Characterization 182 5.5 Inactivating the Function of Specific
Genes in Eukaryotes 212
5.3 Using Cloned DNA Fragments to Study
Gene Expression 198
FIGURE 5-1 Overview of t w o strategies for Mutant organism/ cell
relating the function, location, and structure Comparison of mutant and
of gene products. A mutant organism is the wild-type function
starting point for the classical genetic strategy
(green arrows). The reverse strategy (orange
Gene inactivation
arrows) usually begins with identification of a
protein-coding sequence by analysis of genome-
sequence databases. In both strategies, the actual
gene is isolated either from a DNA library or by Cloned gene
specific amplification of the gene 5equence from DNA sequencing

t
genomic DNA. Once a cloned gene is isolated, it
can be used to produce the encoded protein in Database search to identify
bacterial or eukaryotic expression systems. protein-coding sequence
Alternatively, a cloned gene can be inactivated by Expression in cultured ' PCR isolati on of corresponding
cells gene
one of various techniques and used to generate
mutant cells or organisms.
Protein
Localization
Biochemical studies
Determination of structure

to identify and isolate the affected gene, which we describe to designate a standard genotype for use as a reference in
here. Finally, we discuss techniques that abolish normal protein breeding experiments. Thus the normal, nonmutant allele
function in order to analyze the role of the protein in the cell. will usually be designated as the w ild type,. Because of the
enormous natura ll y occurring allelic variation that exists in
human populations, the term wild type usually denotes an
allele that is present at a much higher frequency than any of
5.1 Genetic Analysis of Mutations
the other possible alternatives.
t o Identify and Study Genes Geneticists draw an important distinction between the
genotype and the phenotype of an organism. The phenotype
As described in Chapter 4, the information encoded in the
refers to all the physical attributes or traits of an individual
DNA sequence of genes specifies the sequence-and there-
that are the consequence of a given genotype. In practice, how-
fore the structure and function-of every protein molecule in
ever, the term phenotype often is used to denote the physical
a cell. The power of genetics as a tool for studying cells and
consequences that result from just the alleles that are under
organisms lies in the ability of researchers to selectively alter
experimental study. Readily observable phenotypic character-
every copy of just one type of protein in a cell by making a
istics are critical in the genetic analysis of mutations.
change in the gene for that protein . Genetic ana lyses of mutants
defective in a particular process can reveal (a) new genes required
for the process to occur, (b) the order in which gene products Recessive and Dominant Mutant Alleles
act in the process, and (c) whether the proteins encoded by
Generally Have Opposite Effects
different genes interact with one another. Before seeing how
genetic studies of this type can provide insights into the on Gene Function
mechanism of a complicated cellular or developmental pro- A fundamental genetic difference between experimental organ-
cess, we first explain some basic genetic terms used through- isms is whether their cells carry a single set of chromosomes
out our discussion. or two copies of each chromosome. The former are referred
The different forms, or variants, of a gene are referred to to as haploid; the latter as diploid. Complex multicellular or-
as alleles. Geneticists commonly refer to the numerous natu- ganisms (e.g., fruit flies, mice, humans) are diploid, whereas
rally occurring genetic variants that exist in populations, many simple unicellular organisms arc haploid. Some organ-
particularly human populations, as alleles. The term muta- isms, notably the yeast Saccharomyces cerevisiae, ca n exist in
tion usually is reserved for instances in wh ich an a llele is either haploid or dip loid states. The normal cells of some
known to have been newly formed, such as after treatment organisms, both plants and animals, carry more than two
of an experimental organism with a mutagen, an agent that copies of each ch romosome and are thus designated poly-
causes a heritable change in the DNA sequence. ploid. Moreover, cancer cells begin as diploid cel ls but
Strictly speaking, the particular set of alleles for all the through the process of transformation into cancer cells can
genes carried by an individ ual is its genotype. However, this gain extra copies of one or more chromosomes and are thus
term also is used in a more restricted sense to denote just the designated as aneuploid. However, our discussion of genetic
alleles of the particular gene or genes under examination. techniques and ana lysis relates to diploid organisms, includ-
For experimental organisms, the term wild type often is used ing diploid yeasts.

172 CHAPTERs • Molecular Genetic Techniques

 T
I
DIPLOID ~ ~ ~ ~ ~
GENOTYPE :::::e::=::::J: :::::e:=::::::F ~ :::::e::=::::J: ~

DIPLOID
Wild type Mutant Mutant Wild type Mutant
PHENOTYPE

FIGURE 5-2 Effects of dominant and recessive mutant alleles on of a recessive allele must be present to cause a mutant phenotype.
phenotype in diploid organisms. A single copy of a dominant allele Recessive mutations usually cause a loss of function; dominant
is sufficient to produce a mutant phenotype, whereas both copies mutations usually cause a gain of function or an altPrE>d function.

Although many different alleles of a gene might occur in recessive for the trait of sickle-cell disease. On the other hand,
different organisms in a population, any individual diploid heterozygous (Hb'!Hbd) individuals are more resistant to ma-
organism will carry two copies of each gene and thus at most laria than homozygous (Hbd!Hba) individuals, revealing that
can have two different alleles. An individual with two different Hbs is dominant for the trait of malaria resistance. •
alleles is heterozygous for a gene, whereas an individual that
carries nvo identical alleles is homozygous for a gene. A reces- A commonly used agent for inducing mutations (muta-
sive mutant allele is defined as one in which both alleles must be genesis) in experimental organisms is ethylmethane sulfonate
mutant in order for the mutant phenotype to be observed; that (EMS). Although this mutagen can alter DNA sequences in
is, the individual must be homozygous for the mutant allele to several ways, o ne of its most common effects is to chemically
show the mutant phenotype. In contrast, the phenotypic conse- modify guanine bases in DNA, u ltimately leading to the con-
quences of a dominant mutant allele can be observed in a het- version of a GC base pair into an A-T base pair. Such an altera-
erozygous individual carryi ng one mutant and one wi ld-type tion in the sequence of a gene, which involves only a single
allele (Figure 5-2). base pair, is known as a point mutation. A silent point muta-
Whether a mutant allele is recessive or dominant provides tion causes no change in the amino acid sequence or activity
valuable information about the function of the affected gene of a gene's encoded protein. However, observable pheno-
and the nature of the causative mutation. Recessive al leles typic consequences due to changes in a protein's activity can
usually result from a mutation that inactivates the affected arise from point mutations that result in substitution of one
gene, leading to a partial or complete loss of function. Such amino acid for another (missense mutation), introduction of
recessive mutations may remove part of the gene or the entire a premature stop codon (nonsense mutation ), or a change in
gene from the chromosome, disrupt expression of the gene, the reading frame of a gene (frameshift mutation ). Because
or alter the structure of the encoded protein, thereby altering alterations in the DNA sequence leading to a decrease in
its function. Conversely, dominant alleles are often the conse- protein activity arc much more li kely than alterations lead-
quence of a mutation that causes some kind of gain of func- ing to an increase or qualitative change in protein activity,
tion. Such dominant mutations may increase the activity of mutagenesis usually produces many more recessive muta-
the encoded protein, confer a new function on it, or lead to tions than dominant mutations.
its inappropriate"spatial or tempora l pattern of expression.
Dominant mutations in certain genes, however, are as- Segregation of Mutations in Breeding
sociated with a loss of function. For instance, some genes are
Experiments Reveals Their Dominance
haplo-insufficient, in that removing o r inactivating one of
·. the two alleles of such a gene leads to a mutant phenotype or Recessivity
because not enough gene product is made. In other rare in- Geneticists exploit the normal life cycle of an organism to
stances, a dominant mutation in o ne allele may lead to a test for the dominance or recessivity of alleles. To see how
structural change in the protein that interferes with the func- this is done, we need first to review the type of cell division
tion of the wild-type protein encoded by the other a llele. that g ives rise to gametes (sperm and egg cells in higher
This type of mutation, referred to as a dominant-negative, plants and animals). Whereas the body (somatic) cells of
produces a phenotype similar ro that obtained from a loss- most multicellular organisms divide by mitosis, the germ
of-function mutation. cells that give rise to gametes undergo meiosis. Like somatic
cells, premeiotic germ cells are diploid, containing two ho-
8 Some alleles can exhibit both recessive and dominant mologs of each morphological type of chromosome. The two
H properties. In such cases, statements about whether an homologs constituting each pai r of homologous chromo-
allele is dominant or recess1ve must specify the phenotype. For somes are descended from different parents, and thus their
example, the allele of the hemoglobin gene in humans desig- genes may exist in different allelic forms. Figure 5-3 depicts
nated Hb 5 has more than one phenotypic consequence. Indi- the major events in mitotic and meiotic cell division. In mi-
viduals who a re homozygous for this allele (Hb5/Hb 5 ) have the tosis, DNA replication is a lways followed by cell division,
debilitating disease sickle-cell anemia, hut heterozygous indi- yielding two diploid daughter cells. In meiosis, one round of
viduals (Hb'/Hb'' ) do not have the disease. Therefore, Hbs is DNA replication is followed by two separate cell divisions,

5.1 Genetic Analysis of Mutations to Identify and Study Genes 173

~ FOCUS ANIMATION: Mitosis

MITOTIC CELL Paternal MEIOTIC CELL Paternal

DIVISION homolog DIVISION homolog

Maternal Maternal
homolog homolog

Somatic cell (2n) Premeiotic cell (2n)

! DNA replication ! DNA replication

@) Replicated
chromosomes
(4n)
Repl icated
chromosomes
(4n)

1 Homologous chromosomes
! 1- align; synapsis and crossing over

::;~~·;:,~
@ .
Mitotic
apparatus .,
'ii)
0
'iii
:!!
Me..ph"' I

{cell division)

Daughter cells
in metaphase II
(2nl

· Daughter cells (2n)

! !
0 0
=.,
'ii)
0
'iii
:!!

cg§@ ~@
1n 1n 1n 1n

FIGURE 5-3 Comparison of mitosis and meiosis. Both somatic of each morphological type then goes into each daughter cell. The
cells and premeiotic germ cells have two copies of each chromosome resulting cells undergo a second division without intervening DNA
(2n), one maternal and one paternal. In mitosis, the replicated replication, with the sister chromatids of each morphological type
chromosomes, each composed of two sister chromatids, align at the being apportioned to the daughter cells. In the second meiotic division
cell center in such a way that both daughter cells receive a maternal the alignment of chromatids and their equal segregation into daughter
and paternal homolog of each morphological type of chromosome. cells is the same as in mitotic division. The alignment of pairs of
During the first meiotic division, however, each replicated chromosome homologous chromosomes in metaphase I is random w ith respect to
pairs with its homologous partner at the cell center; this pairing off other chromosome pairs, resulting in a mix of paternally and materna lly
is referred to as synapsis, and crossing over between homologous derived chromosomes in each daughter cell.
chromosomes is evident at this stage. One replicated chromosome

174 CHAPTER 5 • Molecular Genet ic Techniques

(a) Segregation of dominant mutation yielding four haploid ( 1n) cells that contain only one chromo-
Mutant
some of each homologous pair. The apportionment, or segre-
gation, of the replicated homologous chromosomes to
® daughter cells during the first meiotic division is random; that
is, maternally and paternally derived homologs segregate in-
! dependently, yielding daughter cells with different mixes of
Gametes
G) GJ paternal and maternal chromosomes.
As a way to avoid unwanted complexity, geneticists usu-

~/
ally strive to begin breeding experiments with strains that

a are homozygous for the genes under examination. In such

First filial
generation, F1 : trtte-hreeding strains, every individual will receive the same
all offspring have
mutant phenotype allele from each parent and therefore the composition of

/~
alleles will not change from one generation to the next.
When a true-breeding mutant strain is mated to a true-
breeding wild-type strain, all the first filial (F d progeny will
Gametes ~ orG0 ~ orG0 be heterozygous (Figure 5-4). If the F 1 progeny exhibit the

Second filial
~ / mutant trait, then the mutant allele is dominant; if the F 1
progeny exhibit the wild-type trait, then the mutant is reces-
generation, F2 :
3/4 of offspring have
sive. Further crossing between F 1 individuals will also reveal
m utant phenotype
"-y----1
different patterns of inheritance according to whether the
Mutant Normal mutation is dominant or recessive. When F1 individuals that
are heterozygous for a dominant allele are crossed among
themselves, three-fourths of the resulting F 2 progeny will ex-
hibit the mutant trait. In contrast, when F1 individuals that
are heterozygous for a recessive allele are crossed among
themselves, only one-fourth of the resulting F2 progeny will
exhibit the mutant trait.
As noted earlier, the yeastS. cerevisiae, an important ex-
(b) Segregation of recessive mutation perimental organism, can exist in either a haplmd or a dip-
loid state. In these unicellular eukaryotes, crosses between

s!
Mutant Wild type
haploid cells can determine whether a mutant allele is domi-
8 nant or recessive. Haploid yeast cells, which carry one copy
of each chromosome, can be of two different mating types
! known as a and a. Haploid cells of opposite mating type can
Gametes
® cv mate to produce ala diploids, which carry two copies of each
chromosome. If a new mutation with an observable pheno-

~/
type is isolated in a haploid strain, the mutant strain can be
First filial mated to a wild-type strain of the opposite mating type to

B
generation, F1 _ produce ala diploids that arc heterozygous for the mutant al-
no offspring hiiVe
mut ant phenotype lele. If these diploids exhibit the mutant trait, then the mutant

/~
allele is dominant, but if the diploids appear as wild type, then
the mutant allele is recessive. When ala diploids are placed
under starvation conditions, the cells undergo meiosis, giving
Gametes G) or~ ~or~ rise to a tetrad of four haploid spores, two of type a and two
Second filial ~ / of type a. Sporulation of a heterozygous diploid cell yields
two spores carrying the mutant allele and two carrying the
generation, F2 :
1/4 of offspring have wild-type allele (Figure 5-5). Under appropriate conditions,
mutant phenotype yeast spores will germinate, producing vegetative haploid
Mutant Normal strains of both mating types.
FIGURE 5-4 Segregation patterns of dominant and recessive
mutations in crosses between true-breeding strains of diploid
organisms. All the offspring in the first (F 1) generation are heterozy- Conditional Mutations Can Be Used
gous. If the mutant allele is dominant, the F1 offspring will exhibit the to Study Essential Genes in Yeast
mutant phenotype, as in part (a). If the mutant allele is recessive, the F1
offspring will exhibit the wild-type phenotype, as in part (b). Crossing The procedures used to identify and isolate mutants, referred
of the F1 heterozygotes among themselves also produces different to as genetic screens, depend on whether the experimental
segregation rat ios for dominant and recessive mutant alleles in the F2 organism is haploid or diploid and, if the latter, whether the
generation. mutation is recessive or dominant. Genes that encode pro-

5.1 Genetic Analysis of Mutations to Identify and Study Genes 175

 (a) Wild type Mutant rant phenotype is observed is called nonpermissive; a
(type a) (type a) permissive temperature is one at which the mutant phenotype
Haploid cells of
opposite mating type 8 0 is not observed even t hough the mutant allele is present. Thus
mutant strains can be ma intained at a permissive temperature ·.
~ /
and then subcultured at a nonpermissive temperature for
analysis of the mutant phenotype.
An examp le of a particularly importan t screen for
Diploid cells:
will not exhibit mutant
phenotype if mutation
is recessive
8
(type a/a)
temperature-sensitive muta nts in the yeastS. cerevisiae comes
from the studies of L. H. Hartwell and colleagues in the late
1960s and early 1970s. They set out to identify genes impor-

1 Spowl.tloc
tant in regulation of the cell cycle during which a cell synthe-
sizes proteins, replicates its DNA, and then undergoes mitotic
cell division, with each daughter cell receiving a copy of each
chromosome. Exponential growth of a single yeast cell for
Haploid spores in tetrad: 0 20-30 cell div isions forms a visible yeast colony on solid
2 wil l be mutant
2 will be wild type ®00 agar medium. Since mutants with a complete block in the
cell cycle would not be able to form a colony, conditional
mutants were required to study mutations that affect this
(b)

• • • basic cell process. To screen for such mutants, the research-

ers first identified muragenized yeast cells that could grow

• • • normally at 23 °C but that could not form a colony when

placed at 36 oc (Figure 5-6a) .
Once temperature-sensitive mutants were isolated, fur-

••• •
FIGURE 5-5 Segregation of alleles in yeast. (a) Haploid Saccharo-
ther analysis revealed that some indeed were defective in cell
division. In S. cerevisiae, cell d ivision occurs through a bud-
ding process, and the size of the bud, which is easily visual-
myces cells of opposite mating type {i.e., one of mating type a and one ized by light microscopy, indicates a cell's position in the cell
of mating type a) can mate to produce an a/a diploid. If one haploid cycle. Each of the mutants that could not grow at 36 oc was
carries a dominant wild-type allele and the other carries a recessive examined by microscopy after several hours at the nonpermis-
mutant allele of the same gene, the resulting heterozygous diploid sive temperature. Examination of many different temperature-
will express the dominant trait. Under certain conditions, a diploid cel l sensitive mutants revealed that about 1 percent exhibited a
will form a tetrad of four haploid spores. Two of the spores in t he tetrad d istinct block in the cell cycle. These mutants were therefore
will express the recessive t rait and two will express the dominant trait. designated cdc (cell division cycle) mutants. Importantly,
(b) If the mutant phenotype is not viable under restrictive growth
these yeast mutants did not simply fa il to grow, as they might
conditions, each tetrad, shown here as four spores separated vertically
if they carried a mutation affecting general cellular metabo-
and grown into colonies on restrictive media, consists of two viable
lism. Rather, at the nonpermissive temperature, the mutants
and two nonviable spores. !Part (b) from B. Senger et al., 1998,
EMBO J 1 7:2196]
of interest grew normally for part of the cell cycle b ut then
arrested at a particular stage of the cel l cycle, so that many
cells at this stage were seen (Figure 5-6b ). Most cdc muta-
tions in yeast are recessive; that is, when haploid cdc strains
are mated ro wild-type haploids, the resulting heterozygous
reins essential for life are among the most interesting and diploids are neither temperature sensitive nor defective in cell
important ones to study. Since phenotypic expression of mu- division.
tations in essential genes leads to death of the individual,
clever genetic screens arc needed to isolate and maintain or-
ganisms with a lethal mutation. Recessive Lethal Mutations in Diploids Can
In haploid yeast cells, essential genes can be studied
Be Identified by Inbreeding and Maintained
through the use of conditional mutations. Among the most
common conditiona l mutations are temperature-sensitive mu- in Heterozygotes
tations, which can be isolated in bacteria and lower eukary- In diploid organisms, phenotypes resulting from recessive mu-
otes but not in warm-blooded eukaryote<;. For instance, a tations can be observed only in individuals homozygous fur
single missense mutation may cause the resulting mutant pro- the mutant alleles. Since mutagenesis in a diploid organism
tein to have reduced thermal stability such that the protein is typically changes on ly one al lele of a gene, yielding heterozy-
fully functiona l at one temperature (e.g., 23 oq bur begins to gous mutants, genetic screens must include inbreeding steps to
denature and is inactive at another temperature (e.g., 36 °C), generate progeny that are homozygous for the mutant alleles.
whereas the normal protein wou ld be fu lly stable and func- The geneticist H. Muller developed a general and efficient
tional at both temperatures. A temperature at which the mu- procedure for carrying out such inbreedi ng experiments in the

176 CHAPTER 5 • Molecular Genetic Techniques

 fJ Incubate at 23 oc for 5 h EXPERIMENTAL FIGURE 5-6 Haploid yeasts carrying
temperature-sensitive lethal mutations are maintained at
D -1
permissive temperature and analyzed at non permissive tempera-
Add mutagen;
distribute into
tJLJuJ~LuL
u I u u u u D D
ture. (a) Genetic screen for temperature-sensitive cell-division cycle
(cdc) mutants in yeast. Yeasts that grow and form colonies at 23 oc
smaller aliquots D I D D D D I (permissive temperature) but not at 36 oc (nonpermissive temperature)
{7-{7-{7-{7-1{7-{7-{7- may carry a lethal mutation that blocks cell division. (b) Assay of
Yeast in liquid temperature-sensitive colonies for blocks at specific stages in the cell
culture I Plate out cycle. Shown here are micrographs of wild-type yeast and two different
EJ1 individual temperaturt!-~ensitive mutants after rncubation at the non permissive

+---~~-·__!aliquots
Colonies
temperature for 6 h. Wild-type cells, which continue to grow, can be
seen with all different sizes of buds, reflecting different stages of the
cell cycle. In contrast, cells in the lower two micrographs exhibit a block
Incubate at a specific stage in the cell cycle. The cdc28 mutants arrest at a point
at 23 C
before emergence of a new bud and therefore appear as unbudded
Temperature-sensitive cells. The cdc7 mutants, which arrest just before separation of the
for growth; growth at 23 C, mother cell and bud (emerging daughter cell), appear as cells with
no growth at 36 °C
large buds. [Part (a) see L. H. Hartwell, 1967, J. Bacterial. 93:1662; part (b) from
L. M. Hereford and L. H. Hartwell, 1974, J. Mol. Bioi. 84:445.]

36 cc
fruit fly Drosophila. Recessi vc lethal mutations in Drosoplnla
(b) and other diploid organisms can be maintained in heterozy-
Wild type
gous individuals and their phenotypic consequences analyzed
in homozygotcs.
The Muller approach was used to great effect by C.
Niisslein-Volhard and E. Wieschaus, who systematically
screened for recessive lethal mutations affecting embryogene-
sis in Drosophila. Dead homozygous embryos carrying reces-
sive lethal mutations identified by this screen were examined
under the microscope for specific morphological defects in the
embryos. Current understanding of the molecular mecha-
nisms underlying development of multicellular organisms is
based, in large part, on the detailed picture of embryonic de-
velopment revealed by characterization of these Drosophrla
cdc28 mutants
mutants.

<":\ ~
~
• Complementation Tests Determine Whether
t:~ ~ 0 Different Recessive Mutations Are
·~~
~
in the Same Gene
~,
• In the genetic approach to studying a particular cellular
, ("
·-' 0
process, researchers often isolate multiple recessive muta-
tions that produce the same phenotype. A common test for
determining whether these mutations are in the same gene
cdc7 mutants or in different genes exploits the phenomenon of genetic
complementation, that is, the restoration of the wild-type
phenotype by mating of two different mutants. If two re-
cessive mutations, a and b, are in the same gene, then a
diploid organism heterozygous for both mutations (i.e.,
carrying one a allele and one b allele) will exhibit the mu-
tant phenotype because neither allele provides a functional
copy of the gene. In contrast, if mutations a and b arc in
separate genes, then heterozygotes carrying a single copy of
each mutant allele will not exhibit the mutant phenotype
because a wild-type allele of each gene will also be present.

5.1 Genetic Analysis of Mutations to Identify and Study Genes 177

In this case, the mutations are said to comfJlement each division is regulated in organisms ranging from yeast to
other. Complementation analysis cannot be performed on humans.
dominant mutants because the phenotype conferred by the
mutant allele is displayed even in the presence of a wild-
Double Mutants Are Useful in Assessing
type allele of the gene.
Complementation ana lysis of a set of mutants exhibiting the Order in Which Proteins Function
the same phenotype can distinguish the individual genes in Based on careful analysis of mutant phenotypes associated
a set of functionally related genes, all of which must func- with a particular cel lular process, researchers often can de-
tion to produce a given p henotypic trait. for example, the duce the order in which a set of genes and their protein prod-
screen for cdc mutations in Saccharomyces described previ- ucts function. Two general types of processes are amenable
ously yielded many recessive temperature-sensitive mutants to such analysis: (a) biosynrhetic pathways in which a pre-
that appeared arrested at the same cell cycle stage. To deter- cursor material is converted via one or more intermediates to
mine how many genes were affected by these mutations, a fina l product and (b) signaling pathways that regulate
Hartwell and his colleagues performed complementation other processes and involve the flow of information rather
tests on all of the pair-wise combinations of cdc mutants than chemical intermediates.
following the general protocol outlined in Figure 5-7. These
tests identified more t han 20 different CDC genes. The sub- Ordering of Biosynthetic Pathways A simple example of the
sequent molecular characterization of the CDC genes and first type of process is the biosynthesis of a metabolite such as
their encoded proteins, as described in detail in Chapter 20, the amino acid tryptophan in bacteria. In this case, each of the
has provided a framework for understanding how cell enzymes required for synthesis of tryptophan catalyzes the

Mutant Mutant Mut-ant Mutant

e
(type a ) (type a) (type a) (type a)
Mate haploids of
opposite mating types
and carrying different
recessive temperature-
8 8 8
sensitive cdc mutations
\I cdcXJcdcY
\I cdcXJcdcZ
(type a/a) (type a/a)
Pl ate and incubate

1 at permissive
temperature 1
Test resulting diploids
for a temperature-
sensitive cdc phenotype

XPERIMl!ll TAL FIGURE S-7 Comple-

at nonpermissive
temp erature Q ·
p--
·~- ... _ ____
~.
~-
---~~ _
~

mentation analysis determines whether

recessive mutations are in the same or
36 c Growth 36 oc No growth

different genes. Complementation tests in

yeast are performed by mating haploid a and a
cells carrying different recessive mutations to
produce diploid cells. In the analysis of cdc
PHENOTYPE: @
Wild type Mutant
mutations, pairs of different haploid temperature-
sensitive cdc strains were systematically mated INTERPRETATION: Growth indicates that Absence of growth
m utations cdcX and cdcY indicates that mutations
ilnd the resulting diploids tested for growth at
the permissive and nonpermissive temperatures.
are in different genes cdcX and cdcZ are in the
same gene
..
In this hypothetical example, the cdcX and cdcY
mutants complement each other and thus have
==a:=:J=
mutations in different genes, whereas the cdcX
~
and cdcl mutants have mutations in the Respective wild-type alleles Both alleles nonfun ctional
p rovide normal function
same gene.

178 CHAPTERs • Molecular Genetic Techniques

conversion of one of the intermediates in the pathway to the (a) Analysis of a biosynthetic pathway
next. In E. coli, the genes encoding these enzymes lie adjacent A mutation in A accumulates intermediate 1.
tO one another in the genome, constituting the trp operon (sec A mutation in B accumulates intermediate 2.
Figure 4-13a). The order of action of the different genes for PHENOTYPE OF
these enzymes, hence the order of the biochemical reactions in DOUBLE MUTANT: A double mutation in A and B accumulates
the pathway, initially was deduced from the types of interme- intermediate 1.
diate compounds that accumulated in each mutant. In the case INTERPRETATION: The reaction catalyzed by A precedes the
reaction catalyzed by B.
of complex synthetic pathways, however, phenotypic analysis
of mutants defective in a single step may give amhigunus results ITJ A 0 B 0 _j
that do not permit conclusive ordering of the steps. Double
mutants defective in two steps in the pathway are particularly (b) Analysis of a signaling pathway
useful in ordering such pathways (Figure 5-8a).
In Chapter 14, we discuss the classic use of the double- A mutation in gives repressed reporter expression.
mutant strategy to help elucidate the secretory pathway. In A mutation in B gives constitutive reporter expression.
this pathway, proteins to be secreted from the cell move PHENOTYPE OF
from their site of synthesis on the rough endoplasmic reticu- DOUBLE MUTANT: A double mutation in A and B gives
repressed reporter expression.
lum (ER) to the Golgi complex, then to secretory vesicles,
and finally to the cell surface. INTERPRETATION: A positively regulates reporter expression
and is negatively regulated by B.

Ordering of Signaling Pathways As we learn in later chapters, B --o---1

expression of many cukaryotic genes is regulated by signal-
ing pathways that are initiated by extracellular hormones,
growth factors, or other signals . Such signaling pathways
may include numerous components, and double-mutant PHENOTYPE OF
ana lysis often can provide insight into the functions and in- DOUBLE MUTANT: A double mutation in A and B gives
teractions of these components. The only prerequisite for constitutive reporter exp'ression.
obtaining useful information from this type of analysis is INTERPRETATION: B negatively regulates reporter expression
that the two mutations must have opposite effects on the and is negatively regulated by A.
output of the same regulated pathway. Most commonly, one
mutation represses expression of a particular reporter gene
even w hen the signal is present, while another mutation re-
sults in reporter gene expression even when the signal is ab-
sent (i.e., constitutive expression) . As illustrated in Figure FIGURE 5-8 Analysis of double mutants often can order the
5-8b, two simple regulatory mechanisms are consistent with steps in biosynthetic or signaling pathways. When mutations in
such single mutants, but the double-mutant phenotype can two different genes affect the same cellular process but have distinctly
distinguish between them. This genera l approach has en- different phenotypes, the phenotype of the double mutant can often
ab led geneticists to delineate many of the key steps in a vari- reveal the order in which the two genes must function. (a) In the case
ety of different regulatory pathways, setting the stage for of mutations that affect the same biosynthetic pathway, a double
more specific biochemical assays. mutant will accumulate the intermediate immediately preceding
the step catalyzed by the protein that acts earlier in the wild-type
Note that this technique differs from the complementation
organism. (b) Double-mutant analysis of a signaling pathway is
analysis just described in that both dominant and recessive
possible if two mutations have opposite effects on expression of a
mutants can be subjected to double-mutant analysis. When
reporter gene. In this case, the observed phenotype of the double
two recessive mutations are tested, the double mutant created mutant provides information about the order in which the proteins
must be homozygous for both mutations. Furthermore, domi- act and whether they are positive or negative regulators.
nant mutants can be subjected to double-mutant analysis.

Genetic Suppression and Synthetic

Suppressor Mutations The first type of analysis is based on
lethality Can Reveal Interacting
genetic suppression. To understand this phenomenon, sup-
or Redundant Proteins pose that point mutations lead to structural changes in one
Two other types of genetic ana lysis can provide additional protein (A) that disrupt ire; ability to associate with another
clues about how proteins that function in the same cellular protein (B) involved in the same cellular process. Similarly,
process may interact with one another in the living cell. Both mutations in protein B lead to small structural changes that
of these methods, which arc applicable in many experimental inhibit its abil ity to interact with protein A. Assume, further-
organisms, involve t he use of double mutants in which the more, that the normal functioning of proteins A and B de-
phenotypic effects of one mutation are changed by the pres- pends on their interacting. In theory, a specific structural
ence of a second mutation. change in protein A might be suppressed by compensatory

5.1 Genetic Analysis of Mutations to Identify and Study Gen es 179

(a) Suppression that of suppression. In this case, the deleterious effect of one
mutation is greatly exacerbated (rather than suppressed) by a
Genotype AB aB Ab ab
second mutation in a related gene. One situation in which
Phenotype Wild type Mutant Mutant Suppressed
such synthetic lethal mutations can occur is illustrated in Fig-

@) @) @)
mutant
ure 5-9b. In this example, a heterodimeric protein is partially,
INTERPRETATION@
but not completely, inactivated by mutations in either one of
the nonidentical subunits. However, in double mutants carry-
ing specific mutations in the genes encoding both subunits,
little interaction betwt:t:n subunits occurs, resulting in severe
phenotypic effects. Synthetic lethal mutations also can reveal
(b) Synthetic lethality 1 nonessential genes whose encoded proteins function in re-
dundant pathways for producing an essential cell component.
Genotype AB aB Ab ab
As depicted in Figure 5-9c, if either pathway alone is inacti-

@ ® (@)J
Phenotype Wild type Partial Partial Severe
defect defect
vated by a mutation, the other pathway will be able to supply
defect
the needed product. However, if both pathways are inacti-
vated at the same time, the essential product cannot be syn-
INTERPRETATION®
thesized and the double mutants w.ill be nonviable.

Genes Can Be Identified by Their Map

(c) Synthetic lethality 2 Position on the Chromosome
AB aB Ab ab
The preceding discussion of ge netic analysis illustrates how
Genotype
a geneticist can gain insight into gene function by observing
Phenotype Wild type Wi ld type Wild type Mutant the phenotypic effects produced by joining together different
combinations of mutant alleles in the same cell o r organism.
INTERPRETATION Precursor Precur sor Precursor
For example, combinations of different a lleles of the same
A! !B gene in a diploid can be used to determine w hether a muta-
tion is dominant or recessive or whether two different reces-
Product •
sive mutations are in the same gene. Furthermore, combinations
FIGURE 5-9 Mutations that result in genetic suppression or
of mutations in different genes can be used to determine the
synthetic lethality reveal interacting or redundant proteins. order of gene function in a pathway or to identify functiona l
(a) Observation that double mutants with two defective proteins relationships between genes such as suppression and syn-
(A and B) have a wild-type phenotype but that single mutants give a thetic enha ncement. Generally speaking, all these methods
mutant phenotype .indicates that the function of each protein depends can be viewed as analytica l tests based on gene function . We
on interaction with the other. (b) Observation that double mutants will now consider a fu ndamentally different type of genetic
have a more severe phenotypic defect than single mutants also is analysis based on gene position. Studies designed to deter-
evidence that two proteins (e.g., subunits of a heterodimer) must mine the position of a gene on a chromosome, often referred
interact to function normally. (c) Observation that a double mutant is to as genetic mapping studies, can be used to identify the gene
nonviable but that the corresponding single mutants have the affected by a particular mutation or to determine w hether
wild-type phenotype indicates that two proteins function in redundant two mutations are in the same gene.
pathways to produce an essential product.
In many organisms, genetic mapping studies rely on ex-
changes of genetic information that occur during meiosis. As
changes in protein B, allowing the mutant proteins to inter- shown in Figure 5-lOa, genetic recombination takes place
act. In the rare cases in w hich such suppressor mutations before the first meiotic cell division in germ cells, when the
occur, strains carrying both mutant alleles would be normal, replicated chromosomes of each homologous pair align with
whereas strains carrying only one or the other mutant allele each other. At this tim e, homologous DNA sequences on
would have a mutant phenotype (Figure 5-9a). maternally and paternally derived chromatids can exchange
The observation of genetic suppression in yeast strains with each other, a process known as crossing over. We now
carrying a mutant actin allele (actl-1) and a second muta- know that the resulting crossovers between homologous
tion (sac6) in another gene provided early evidence fo r a di- chromosomes provide structural links that are important for
rect interaction in vivo between the proteins encoded by the the proper ~egrega tion of pairs of homologous chromatids to
two genes. Later biochemical studies showed that these two opposite poles during the first meiotic cell division (for dis-
proteins-Actl and Sac6-do indeed interact in the con- cussion see Chapter 19).
struction of functional actin structures withi n the cell. Consider two d ifferent mutations, one inherited from
each parent, that are located close to each other on the same
Synthetic Lethal Mutations Another phenomenon, called chromosome. Two different types of gametes can be pro-
synthetic lethality, produces a phenotypic effect opposite to duced accord ing to whether a crossover occurs between the

180 CHAPTERs • Molecular Genetic Techniques

(a) Replicated (b) Consider two linked genes A and B with recessive alleles a and b.

mlJl-mz chromosomes
(4n) Cross of two mutants to construct a doubly heterozygous strain:
i Synapsis and
crossing over
AlA bib x ala BIB

~
I
.! I Metaphase I
A b

i a B

~ ~ Anaphase I Cross of double heterozygote to test strain:

A b
X
a b

)~
A a B a b

.. ( i ( ) ~ Anaphase II
A
a
b
b
a
a
B
b
A
a
B
b
a
a
b
b

i i i i Parental types Recombinant types

:0
;[,\
1n

Parental
gamete
' 1n
m1 iJ':\

Recombinant gametes
m2
1n 1n

Parental
gamete
Genetic distance between A and B can be determined from
frequency of parental and recombinant gametes:

Genetic distance in eM= 100 x

recombinant gametes
t t t
o a 1 game es

FIGURE 5- 1 0 Recombination during meiosis can be used to map likely they are to be separated by recombination and the greater the
the position of genes. (a) Shown is an individual that carries two proportion of recombinant gametes produced. (b) In a typical mapping
mutations, designated m 1 (yellow) and m2 (green), that are on the experiment, a strain that is heterozygous for two different genes is
maternal and paternal versions of the same chromosome. If crossing const ructed. The frequency of parental or recombinant gametes
over occurs at an interval between m 1 and m2 before the first meiotic produced by this strain can be determined from the phenotypes of the
division, then two recombinant gametes are produced; one carries progeny in a testcross to a homozygous recessive strain. The genetic
both m 1 and m2, whereas the other carries neither mutation. The map distance in centimorgans (eM) is given as the percent of the
longer the distance between two mutations on a chromatid, the more gametes that are recombinant.

mutations during meiosis. If no crossover occurs between steps . In the first step, a strain is constructed that carries a
them, gametes known as parental types, which contain either different mutation at each position, or locus. In the second
o ne or the other mutation, w ill be produced. In contrast, if a step, the progeny of this strain are assessed to determine the
crossover occurs between the two mutations, gametes known relative frequency of inheritance of parental or recombinant
as recombinant types will be produced. In this example, re- types. A typical way to determine the frequency of recombi-
combinant chromosomes would contain either both muta- nation between two genes is to cross one diploid parent het-
tions or neither of them. The sites of recombination occur erozygous at each of the genetic loci to another parent
more or less at random along the length of chromosomes; homozygous for each gene. For such a cross, the proportion
thus the closer together two genes are, the less likely that of recombinant progeny is readi ly determined because re-
recombination will occur, between them during meiosis. In combinant phenotypes will differ from the parental pheno-
other words, the less frequently recombination occurs be- types. By convention, one genetic map unit is defined as the
tween two genes on the same chromosome, the closer to- distance between two positions along a chromosome that
gether they are. Two genes that are on the same chromosome results in one recombinant individual in 100 total progeny.
and that arc sufficiently close together such t hat there are The distance corresponding to this 1 percent recombmation
significantly fewer recombinant gametes produced than pa- frequency is called a centimorgan (eM) in honor of Sturte-
rental gametes are considered to be genetically linked. vant's mentor, Morgan (Figure 5-1 Ob).
The techn ique of recombinational mapping was devised in A complete discussion of the methods of genetic mapping
191 1 by A. Sturtevant while he was an undergraduate work- experiment<; is beyond the scope of this introductory discu:.-
ing 111 the laboratory ofT. H. Morgan at Columbia Univer- sion; however, two features of measuring distances by recom-
sity. O riginally used in studies on Drosophila, this technique bination mapping need particular emphasis. First, the
is still used roday to assess the distance between two genetic frequency of genetic exchange between two loci is strictly
loci o n the same chromosome in many experimental organ- proportional to the physical distance in base pairs separating
ism~. A typical experiment designed to determine the map them only for loci that are relatively close together (say, less
distance between two genetic positions wou ld involve two than about 10 eM). For linked loci that are farther apart than

5.1 Genetic Analysis of Mutations to Identify and Study Genes 181

this, a distance measured by the frequency of genetic ex-
change tends to underestimate the physical distance because Dominant and recessive mutations exhibit characteristic
of the possibility of two or more crossovers occurring within segregation patterns in genetic crosses (see Figure 5-4).
an interval. In the limiting case in which the number of re- • In haploid yeast, temperature-sensitive mutations are par-
combinant types will equal the number of parental types, the ticu larly usefu l for identifying and studying genes essential
two loci under consideration could be far apart on the same to survival.
chromosome or they could be on different chromosomes, and
The number of functionally related genes involved in a
in such cases the loci are considered to be unlinked.
process can be defined by complementation analysis (sec
A second important concept needed for interpretation of
Figure 5-7).
genetic mapping experiments in different types of organisms
is that although genetic distance is defined in the same way • The order in which genes function in a signaling pathway
for different organisms, the relationship between recombina- can be deduced from the phenotype of double mutants de-
tion frequency (i.e., genetic map distance) and physical dis- fective in two steps in the affected process.
tance varies between organisms. For example, a 1 percent Functionally significant interactions between proteins can
recombination frequency (i.e., a genetic distance of 1 eM) be deduced from the phenotypic effects of allele-specific sup-
represents a physical distance of about 2.8 kilobases in yeast pressor mutations or synthetic lethal mutations.
compared with a distance of about 400 kilobases in Dro-
Genetic mapping experiments ~ake use of crossing over
sophila and about 780 kilobases in humans.
between homologous chromosomes during meiosis to mea-
One of the chief uses of genetic mapping studies is to lo-
sure the genetic distance between two different mutations on
cate the gene that is affected by a mutation of interest. The
the same chromosome.
presence of many different already mapped genetic traits, or
genetic markers, distributed along the length of a chromo-
some permits the position of an unmapped mutation to be
determined by assessing its segregation with respect to these
marker genes during meiosis. Thus the more markers that 5.2 DNA Cloning and Characterization
are available, the more precisely a mutation can be mapped. Detailed studies of the structure and function of a gene at the
In Section 5.4, we will see how the genes affected in inherited molecular level require large quantities of the individual
human diseases can be identified using such methods. A sec- gene in pure form. A variety of techniques, often referred to
ond general use of mapping experiments is to determine as recombinant DNA technology, are used in DNA cloning,
whether two different mutations are in the same gene. If two wh ich permits researchers to prepare large numbers of identi-
mutations arc in the same gene, they will exhibit tight link- cal DNA molecules. Recombinant DNA is simply any DNA
age in mapping experiments, but if they are in different molecule composed of sequences derived from different
genes, the} will usually be unlinked or exhibit weak linkage. sources.
The key to cloning a DNA fragment of interest is to link
it to a vector DNA molecule that can replicate within a host
cell. After a single recombinant DNA molecule, composed of
a vector plus an inserted DNA fragment, is introduced into a
KEY CONCEPTS of Section 5.1 host cell, the inserted DNA is replicated along with the vec-
tor, generating a large number of identical DNA molecules.
Genetic Analysis of Mutations to Identify
The basic scheme can be summarized as follows:
and Study Genes
Diploid organisms carry two copies (alleles) of each gene, Vector + DNA fragment
whereas haploid organisms carry only one copy. j_
Recessive mutations lead to a loss of function, which is Recombinant DNA
masked if a normal allele of the gene is present. For the mu-
j_
tant phenotype to occur, both alleles must carry the mutation.
Replication of recombinant DNA within host cells
Dominant mutations lead to a mutant phenotype in the
presence of a normal allele of the gene. The phenotypes as- j_
sociated with dominant mutations often represent a gain of Isolation, sequencing, and manipulation
function but in the case of some genes result from <1 lm~ of of purified DNA fragment
function.
In meiom, a diploid cell undergoes one DNA replication Although investigators have devised numerous experimental
and two cell divisions, yielding four haploid cells in which variations, this flow diagram indicates the essential steps in
maternal and paternal chromosomes and their associated al- DNA cloning. In this section, we first describe methods for
leles are randomly assorted (see Figure 5-3). isolating a specific sequence of DNA from a sea of other DNA
sequences. This process often involves cutting the genome into

182 CHAPTERs • Molecular Genetic Techniques

 fragments and then placing each fragment in a vector so that For each restriction enzyme, bacteria also produce a mod-
the entire collection can be propagated as recombinant mole- ification enzyme, which protects a host bacterium's own
cules in separate host cells. W hile many different types of vec- DNA from cleavage by modifying the host DNA at or near
tors exist, our discussion will mainly focus on plasmid vectors each potential cleavage site. The modification enzyme adds a
in E. coli host cells, which are commonly used. Various tech- methyl group to one or two bases, usually within the restric-
niques can then be employed to identify the sequence of inter- tion site. When a methyl group is present there, the restriction
est from this collection of DNA fragments, known as a DNA endonuclease is prevented from cutting the DNA. Together
library. Once a specific DNA fragment is isolated, it is typi- with the restriction endonuclease, the methylating enzyme
cally characterized hy determining the exact sequence of nu- forms a restriction-modification system that protects the host
cleotides in the molecu le. We end with a discussion of the DNA while it destroys incoming foreign DNA (e.g., bacterio-
polymerase chain reaction (PCR). This powerful and versatile phage DNA or DNA taken up during transformation) by
technique can be used in many ways to generate large quanti- cleaving it at all the restriction sites in the DNA.
ties of a specific sequence and otherwise manipulate DNA in Many restriction enzymes make staggered cuts in the two
the laboratory. The various uses of cloned DNA fragments are DNA strands at their recognition site, generating fragments
discussed in subsequent sections. that have a single-stranded "tail" at both ends, sticky ends
(see Figure 5-11 ). The tails on the fragments generated at a
given restriction site are complementary to those on all other
Restriction Enzymes and DNA Ligases fragments generated by the same restriction enzyme. At room
Allow Insertion of DNA Fragments temperature, these single-stranded regions can transiently
into Cloning Vectors base-pair with those on other DNA fragments generated with
A major objective of DNA cloning is to obtain discrete, the same restriction enzyme. A few restriction enzymes, such
small regions of an organism's DNA that constitute specific as Alul and Smai, cleave both DNA strands at the same point
genes. In addition, only relatively small DNA molecules can wtthin the restriction site, generating fragments with "blunt"
be cloned in any of the available vectors. For these reasons, (flush) ends in which all the nucleotides at the fragment ends
the very long DNA molecules that compose an organism's are base-paired to nucleotides in the complementary strand.
genome must be cleaved into fragments that can be inserted The DNA isolated from an individual organism has a spe
into the vector DNA. Two types of enzymes-restrictio n en- cific sequence, which purely by chance will contain a specific
·. · zymes and DNA ligases-faci litate production of such re- set of restriction sites. Thus a given restriction enzyme will
combinant DNA molecules. cut the DNA from a particular source into a reproducible set
of fragments called restriction fragments . The frequency with
Cutting DNA Molecules into Small Fragments Restriction en- which a restriction enzyme cuts DNA, and thus the average
zymes are endonucleases produced by bacteria that typically size of the resulting restriction fragments, depends largely on
recognize specific 4- to 8-bp sequences, called restriction the length of the recognition site. For example, a restriction
sites, and then cleave both DNA strands at this site. Restric- enzyme that recognizes a 4-bp site will cleave DNA an aver-
tion sites commonly are short fJalindromic sequences; that is, age of once every 4\ or 256, base pairs, whereas an enzyme
the restriction-site sequence is the same on each DNA strand that recognizes an 8-bp sequence will cleave DNA an average
when read in the 5' to 3' direction (Figure 5- 11 ). of once every 4H base pairs (-65 kbp). Restriction enzymes
have been purified from several hundred different species of
bacteria, allowing DNA molecules to be cut at a large num-
ber of different sequences corresponding to the recognition
£coAl
sites of these enzymes (Table 5-1).

Inserting DNA Fragments into Vectors DNA fragments with

either sticky ends or blunt ends can be inserted into vector
T DNA with the aid of DNA ligases. During normal DNA rep-
Cl,.v•g•l :oR I lication, DNA ligase catalyzes the end-to-end joining (liga-
tion) of short fragments of DNA called Okazaki fragments.
Sticky ends
For purposes of DNA cloning, purified DNA ligase is used to
covalently join the ends of a restriction fragment and vector

s· ~~~ G, UTT'c ~~~ 3'

DNA that have complementary ends (Figure 5-12). The vec-
tor DNA and restriction fragment are covalently ligated to-
3' C: C T T A A GC 5' gether through the standard 3' to 5' phosphodiester bonds
FIGURE 5-11 Cleavage of DNA by the restriction enzyme EcoRI. of DNA. In addition to ligating complementary sticky ends,
This restriction enzyme from E. coli makes staggered cuts at the specific the DNA ligase from bacteriophage T 4 can ligate any two
6·bp palindromic sequence shown, yielding fragments with single· blunt DNA ends. However, blunt-end ligation is inherently
stranded, complementary "sticky" ends. Many other restriction inefficient and requires a higher concentration of both DNA
enzymes also produce fragments with sticky ends. and DNA ligase than does ligation of sticky ends.

5.2 DNA Cloning and Characterization 183

8 §='!¥11
Enzyme
Selected Restriction Enzymes and Their Recognition Sequences

Source Microorganism Recognition Site* Ends Produced

-1-
BamHl Bacillus amyloltque(aciens -G-G-A-T-C-C- Sticky
-C-C-T-A-G-G-
j

-1-
Sau3A Staphylococcus aureus -G-A-T-C- Sticky
-C-T-A-G-
i

-1-
EcoRI Escherichia coil -G-A-A-T-T-C- Sticky
-C-T-T-A-A-G
i

-1-
Hind III Haemophilus influem;ae -A-A-G-C-T-T- Sticky
-T-T-C-G-A-A-
i

!
Smal Serratia marcescens -C-C-C-G-G-G- Blunt
-G-G-G-C-C-C-
i

!
Notl Nocardia otitidis-cauiarum -G-C-G-G-C-C-G-C- Sticky
-C-G-C-C-G-G-C-G-
i
• ~1any of rhese recognition sequences are included in a common polylmker ~cqucnce (see Figure 5-13 ).

inserted (Figure 5-13 ). Host-cell enzymes replicate a plasmid

E. coli Plasmid Vectors Are Suitable for Cloning
beginning at the replication origin (ORI), a specific DNA
Isolated DNA Fragments sequence of 50-100 base pairs. Once DNA replication is ini-
Plasmids are circular, double-stranded DNA (dsDNA) mol- tiated at the ORI, it continues around the circular plasmid
ecules that are separate from a cell's chromosomal DNA. regardless of its nucleotide sequence. Thus any DNA se-
These extrachromosomal DNAs, which occur naturally in quence inserted into such a plasmid is replicated along with
bacteria and in lower eukaryotic cells (e.g., yeast), exist in a the rest of the plasmid DNA.
parasitic or symbiotic relationship with their host cell. Like figure .5-14 outlines the general procedure for cloning a
the host-cell chromosomal DNA, plasmid DNA is duplicated DNA fragment using E. coli plasmid vectors. When E. coli
before every cell division. During cell division, copies of the cells are mixed with recombinant vector DNA under certain
plasmid Dl"A segregate to each daughter cell, ensuring con- conditions, a small fraction of the cells will take up the plas-
tinued propagation of the plasmid through successive gen- mid DNA, a process known a<, transformation. Typically,
erations of the host cell. 1 cell in about 10,000 incorporates a single plasmid DNA
The plasmids most commonly used in recombinant DNA molecule and thus becomes transformed. After plasmid vec-
technology are those that replicate in E. coli. Investigators tors are incubated with E. coli, those cells that take up the
have engineered these plasrnids to optimize their use as vectors plasmid can be easily selected from the much larger number
in DNA cloning. For instance, removal of unneeded portions of cells. for instance, if the plasmid carries a gene that con-
from naturally occurring£. coli plasmids yields plasmid vec- fers resistance to the antibwtic ampicillin, transformed cells
tors - 1.2-3 kb in circumferential length that contain three can be selected by growing them in an ampicillin-containing
regions essential for DNA cloning: a replication origin; a medium.
marker that permits selection, usually a drug-resistance gene; DNA fragments from a few base pairs up to - 10 kb can
and a region in which exogenous DNA fragments can be be inserted into plasmid vectors. When a recombinant plasmid

184 CHAPTERs • Molecular Genetic Techniques

 Genomic DNA fragments the colony grows. In this way, the initial fragment of DNA is
(a) replicated in the colony of cells into a large number of identi-
p AATT-t ::l 3' cal copies. Since all the cells in a colony arise from a single
Vector DNA OH-t 15' transformed parental cell, they constitute a clone of cells,
(a·) (b) and the initia l fragment of DNA inserted into the parental
5·----0H + P-C G - t v « ~
HO I
J 31
15'
plasmid is referred to as cloned DNA or a DNA clone.
3' TT AA P The versatility of an E. coli plasmid vector is increased by
(c) the add ition of a poly/in ker, a synthetically generated se-
P- A G C T -1 .::J 3'
HO I 1 5'
quen ce containing one copy of several different restriction
sites that are not present elsewhere in the plasmid sequence
(see Figure 5-13 ). When such a vector is treated with a re-

l
Complementary
ends base-pair striction enzyme that recognizes a restriction site in the
polylinker, the vector is cut only once within the polylinker.
OH P Subsequently, any DNA fragment of appropriate length pro-
duced with the same restriction enzyme can be inserted into
(a') \ A/ ATT •> (a)
5' • --.: ~""" ' 3' Unpai red genomic the cut plasmid with DNA ligase. Plasmids containing a
3' Cl====::::~~-f +>~ \!====::J 5• + fragments and polylinker permit a researcher to use the same plasmid vector
when cloning DNA fragments generated with different re-
p HO striction enzymes, which simplifies experimental procedures.
2 ATP For some purposes, such as the isolation and manipula-
. ' tion of large segments of the human genome, it is desirable to
T4 DNA ligase clone DNA segments as large as several megabases II mega-
base (Mb) = 1 million nucleotides]. For this purpose special-
2AMP+2 PP;
ized plasmid vectors known as BA Cs (bacterial artificial
chromosomes) have been developed. One type of BAC uses a
(a') (a)
5· --• A A TT--c::::::::=::J 3'
I I I I
3' Cl====::::JI-TTAA-c==~ 5'
replication origin derived from an endogenous plasmid of E.
coli known as the F factor. The F factor and cloning vectors
derived from it can be stably maintained at a single copy per
FIGURE 5-12 Ligation of restriction fragments with complemen- E. coli cell even when they contain inserted sequences of up
tary sticky ends. In this example, vector DNA cut with EcoRI is mixed
to about 2 Mb. Production of BAC libraries requires special
with a sample containing restriction fragments produced by cleaving
methods for the isolation, ligation, and transformation of
genomic DNA with several different restriction enzymes. The short
large segments of DNA because segments of DNA larger than
base sequences composing the sticky ends of each fragment type are
shown. The sticky end on the cut vector DNA (a') base-pairs only with
about 20 kb are highly vulnerable to mechanical breakage by
the complementary sticky ends on the EcoRI fragment (a) in the even standard manipulations such as pipetting.
genomic sample. The adjacent 3' hydroxyl and 5' phosphate groups
(red) on the base-paired fragments then are covalently joined (ligated)
by T4 DNA ligase. eDNA Libraries Represent the Sequences
of Protein-Coding Genes
with an inserted DNA fragment transforms an £. coli cell, all A collection of DNA molecules each cloned into a vector mol-
the antibiotic-resistant progeny cells that arise from the ini- ecule is known as a DNA library. When genomic DNA from
tial transformed cell will contain plasmids with the same in- a particular organism is the source of the starting DNA, the
serted DNA. The inserted DNA is replicated along with the set of clones that collectively represent all the DNA sequences
rest of the plasmid DNA and segregates to daughter cells as in the genome is known as a genomic library. Such genomic

Psrt
Sa~
FIGURE S-13 Basic components of a plasmid cloning vector that can replicate
Xbai within an E. coli cell. Plasmid vectors contain a selectable gene such as amp', which
BamHI encodes the enzyme ~-lactamase and confers resistance to ampicillin. Exogenous DNA
Smai
can be inserted into the bracketed region without disturbing the ability of the plasmid to
Kpnl
replicate or express the amp' gene. Plasmid vectors also contain a replication origin (ORI)
sequence where DNA replication is initiated by host-cell enzymes. Inclusion of a synthetic
polylinker containing the recognition sequences for several different restriction enzymes
Polylinker Plasmid increases the versatility of a plasmid vector. The vector is designed so that each site in the
cloning vector polylinker is unique on the plasmid.

5.2 DNA Cloning and Characterization 185

~ TECHNIQUE ANIMATION: Plasmid Cloning

EXPERIMENTAL FIGURE S-14 DNA cloning in a plasmid vector

~
permits amplification of a DNA fragment. A fragment of DNA to be
cloned is first inserted into a plasmid vector containing an ampicillin-
1
+ DNA fragment
resistance gene (amp ), such as that shown in Figure 5-13. Only the few to be cloned
cells transformed by incorporation of a plasmid molecule will survive
on ampicillin-containing medium. In transformed cells, the plasmid
Enzymatically insert

1
0
DNA replicates and segregates into daughter cells, resulting in
DNA into plasmid vector
formation of an ampicillin-resistant colony.

Recombinant
plasmid
libraries are ideal for representing the genetic content of rela-
tively simple organisms such as bacteria or yeast but present 8(\)

certain experimental difficulties for higher eukaryotes. First,

Mix E. coli with plasmids
the genes from such organisms usually contain extensive in- in pPesence of CaCI2; heat-pulse
tron sequences and therefore can be too large to be inserted
intact into plasmid vectors. As a result, the sequences of in-
dividual genes are broken apart and carried in more than
one clone. Moreover, the presence of introns and long inter-
E. coli
chromosome l Culture on nutrient agar
plates containing ampicillin

genic regions in genomic.: DNA often makes it difficult to

identify the important parts of a gene that actually encode
protein sequences. For example, only abour 1.5 percent of
~-~
the human genome actually represents protein-coding gene Transformed cell Cells that do not
sequences. Thus for many studies, cellular mRNAs, which survives take up pl asmid die
on ampici llin plates
lack the noncoding regions present in genomic DNA, are a
more useful starting material for generating a DNA library.
In this approach, DNA copies of mRNAs, called comple-
mentary DNAs (cDNAs), are synthesized and cloned into
1 Plasmid replication

plasmid vectors. A large collection of the resulting eDNA

clones, representing all the mRNAs expressed in a cell type,
is called a eDNA library.

cDNAs Prepared by Reverse Transcription

of Cellular mRNAs Can Be Cloned
to Generate eDNA Libraries
The first step in preparing a eDNA library is to isolate the
total mRNA from the cell type or tissue of interest. Because
of their poly(A) tails, mRNAs arc easily separated from the
much more prevalent rRNAs and tRNAs present in a cell
extract by use of a column to which short strings of thymi-
dylate (oligo-dTs) are linked to the matrix. The general pro-
cedure for preparing a eDNA library from a mixture of Colony of cells, each containing copies
of the same recombinant plasmid
cellular mRNAs is outlined in Figure 5-15. The enzyme re-
verse transcriptase, which is found in retroviruses, is used to
synthesize a strand of DNA complementary to each mRKA
molecule, starting from an oligo-dT primer (steps 0 and f)).
The resu lting cDNA-mRNA hybrid molecules are converted To prepare double-stranded cDNAs for cloning, short
in ~evcral steps to double-stranded eDNA molecules corre- double-stranded DNA molecu les containing the recognition
sponding to all the mRNA molecules in the original prepara- site for a particular restriction enzyme are ligated to both
tion (steps D-~). Each double-stranded eDNA contains an ends of the c.:DNAs using DNA ligase from bacteriophage T4
oligo-dC.oligo-dG double-stranded region at one end and an (Figure 5-15, step Q ). As noted earlier, this ligase can join
oligo-dToligo-dA double-stranded region at the other end. "blunt-ended" double-stranded DNA molecules lacking
Methylation of the eDNA protects it from subsequent re- sticky ends. The resulting molecules are then treated with the
striction enzyme cleavage (step Gl). restriction enzyme specific for the attached linker, generating

186 CHAPTERs • Molecular Genetic Techniques

 m R NA5 ' ~ 3 ' FIGURE 5-15 A eDNA library contains representative copi es
3' poly(A) tail of cellular mRNA sequences. A mixture of mRNAs is the starting point
. . Hybridize m RNA with for preparing recombinant plasmid clones each containing a eDNA.
Oilgo-dT pnmer oligo-dT primer Transforming E. coli with the recombinant plasm ids generates a set
T T T T 5' of eDNA clones representing all the cellular mRNAs. See the text for
a step-by-step discussion.
~

fJ! T T T T 5'
T ranscribe RNA into eDNA

~~2'~Am 3 '
..__ _ ___, T T T T 5'

II! Remove RNA with alkali

Add poly(dG) tail

eDNA
n!
Single-stranded 3' G G G G L - - -- - ' T T T T 5'

Hybridize with
.,. oligo-de primer

5'~~~
3' GGGG
,---~----, TTTT 5'
r.ll!
U
Synthesize complementa ry
strand

Double-stran ded 5 '~~~~~~~~~~3 '

eDNA 3' G GGG TTTT 5'

I'll!
CH
3
Protect eDNA by
met hylat ion at EcoRI sites

5'~~~~~3
I TT.TT 5''
3''GGGGI
I
CH 3

EcoRI linker ~ Ligate eDNA to rest riction

D G .i'l A l"T C D site linkers
D C JAA G D

co
.~ G O

rim! Cleave with EcoRI

l\ Cl nz=:: G
[11!
I
1
G O GGGG..__ __,T T T T O C 1 Cut with EcoRI
Sticky end
~ Ugote to pl,mid
AAT C ~~~
G

ii!]l Transfo rm E. coli

Select for amp'
Individual
·. clones Plasmid with
sticky ends

amp'

5.2 DNA Cloning and Characterization 187

eDNA molecules with sticky ends (step li!l). In a separate must be attached to a solid support. A replica of the petri
procedure, plasmid DNA is treated with the same restriction dish containing a large number of individual c. coli clones is
enzyme to produce the appropriate sticky ends (step IIDJ). reproduced on the surface of a nitrocellulose membrane. The
The vector and the collection of cDNAs, all containing DNA on the membrane is denatured, and the membrane is
complementary sticky ends, then are mixed and joined cova- then incubated in a solution containing a probe specific for
lently by DNA ligase (Figure 5-15, step P,J). The resulting the recombinant DNA containing the fragment of interest
DNA molecules are transformed into E. coli cells to generate that is labeled either radioactively or fluorescenrly. Under
individual clones; each clone carrying a eDNA derived from hybridization conditions (near neutral pH, 40-65 °C,
a smgle mRNA. 0.3-0.6 M NaCI), this labeled probe hybridizes to any com-
Because different genes are transcribed at very different plementary nucleic acid strands bound to the membrane.
rates, eDNA clones corresponding to abundantly transcribed Any excess probe that does not hybridize is washed away,
genes will be represented many times in a eDNA library, and the labeled hybrids arc detected by autoradiography or
whereas cDNAs corresponding to infrequently transcribed by fluorescent imaging of the filter. This technique can be
genes will be extremely rare or not present at all. This prop- used to screen both genomic and eDNA libraries but is most
erty is advantageous if an investigator is interested in a gene commonly used to isolate specific cDNAs.
that is transcribed at a high rate in a particular cell type. In Clearly, identification of specific clones by the mem-
this case, a eDNA library prepared from mRNAs expressed brane-hybridization technique depends on the availability
in that cell type will be enriched in the eDNA of interest, of complementary radiolabeled probes. For an oligonucle-
facilitating isolation of clones carrying that eDNA from the otide to be useful as a probe, it must be long enough for its
library. However, to have a reasonable chance of including sequence to occur uniquel y in the clone of interest and not
clones corresponding to slowly transcribed genes, mam- in any other clones. For most purposes, this condition is
malian eDNA libraries must contain 10b-LO- individual satisfied by oligonucleotides containing about 20 nucleo-
recombinant clones. tides. This is because a specific 20-nucleqtide sequence oc-
curs once in every 4 20 ( - 10 12 ) nucleotides. Since all genomes
are much smaller (-3 X 109 nucleotides for humans), a
specific 20-nucleotide sequence in a genome usually occurs
DNA Libraries Can Be Screened by Hybridization
only once. With automated instruments now available, re-
to an Oligonucleotide Probe searchers can program the chemical synthesis of oligonu-
Both genomic and eDNA libraries of various organisms con- cleotides of specific sequence up to about 100 nucleotides
tain hundreds of thousands to upward of a million individual long. Longer probes can be prepared by the polymerase
clones in the case of higher eukaryotes. Two general ap- chain reaction (PCR), a widely used technique for amplify-
proaches are available for screening libraries to identify clones ing specific DNA sequences that is described later in this
carrying a gene or other DNA region of interest: (1) detection chapter.
with oligonucleotide probes that bind to the clone of interest How might an investigator design an oligonucleotide
and (2) detection based on expression of the encoded protein. probe to identify a clone encoding a particular protein? lt
Here we describe the first method; an example of the second helps if all or a part of the amino acid sequence of the pro-
method is presented in the next section. tein is known. Thanks to the availability of the complete
The basis for screening with oligonucleotide probes is genomic sequences for humans and many other organisms,
hybridization, the ability of complementary single-stranded including all of the model organisms such as the mouse,
DNA or RNA molecules to associate (hybridize) specifically Drosophila, and the roundworm Caenorhabditis elegans, a
w1th each other via base pairing. As discussed in Chapter 4, researcher can use an appropriate computer program to
double-stranded (duplex) DNA can be denatured (melted) search the genomic sequence database for the coding se-
into single strands by heating in a dilute salt solution. If the quence that corresponds to the amino acid sequence of the
temperature then is lowered and the ion concentration raised, protein under study. If a match is found, then a single,
complementary single strands will reassociate (hybridize) into unique DNA probe based on this known genomic sequence
duplexes. In a mixture of nucleic acids, only complementary will hybridize perfectly with the clone encoding the protein
single strands (or strands containing complementary regions) of interest.
will reassociate; moreover, the extent of their reassociation is
virtually unaffected by the presence of noncomplementary
strands. As we will see later in this chapter, the ability to Yeast Genomic Libraries Can Be Constructed
identify a particular DNA or RNA sequemx within a highly
with Shuttle Vectors and Screened
complex mixture of molecules through nucleic acid hybrid-
ization is the basis for many techniques employed to study by Functional Complementation
gene expression. In some cases, a DNA library can be screened for the ability to
The steps involved in screening an E. coli plasmid eDNA express a functional protein that complements a recessive muta-
library are depicted in Figure 5-16. First, the DNA to be screened tion. Such a screening strategy would be an efficient way to

188 CHAPTER 5 • Molecular Genetic Techntques

'·

Individual colonies Bound single-stranded DNA

Master plate of I Filter

E. coli colo nies I
I

!
I
I
I
I Incubate with labeled DNA(- )

l Place nitrocellulose fi lter on plate

to pick up cells from each colony

I
I
I
I
I
I
I

Hybridized
Nitrocellulose filter I
complementary DNAs
I
I
I

.· I
I
I

l Wash away labeled DNA that does

I
I

l---
lncubate f ilter in alkaline I not hybridize to DNA bound to filter
I
so lution to lyse cells and I
denature released plasmid DNA /
/

/
Hybridize with labeled probe /
/
/

~ ///

~'"· -~ ~ ~, ,."'/
\ • -
..........
- . t:::..J.J_
_____.., ----_-_-.:.:- -
...... ...... ......
/
---
! Perform autoradiography
Perform autoradiography

Signal appears over

' plasmid DNA that is
\
• ) complementary
/
to probe

(b)

....
.
t
. ' .. . .
: ..
• :. o:..·.

- Jc:PERIMENTAL FIGURE 5 16 eDNA libraries can be screened (although for ease of comparison, it is not shown reversed here).
with a radiolabeled probe to identify a clone of interest. The Aligning the autoradiogram with the original petri dish will locate the
appearance of a spot on the autoradiogram indicates the presence of corresponding clone from which £.coli cells can be recovered. (b) This
a recombinant clone containing DNA complementary to t he probe. autoradiogram shows five colonies (arrows) of f. coli containing the
The position of the spot on the autoradiogram is the mirror image of desired eDNA. [Part (b) from H. Fromm and N_-H_ Chua, 1992, Plant. Mol. Bioi.
the position of that particu lar clone on the origina l petri dish Rep. 10:199.)

isolate a cloned gene that corresponds to an interesting recessive Libraries constructed for the purpo~e of screening among
mutation iJentified in an experimental organism. To illustrate yeast gene sequences usually are constructed from genomic
this method, referred to as fu nctional complementation, we de- DNA rather than eDNA. Because Saccharomyces genes do
scribe how yeast genes cloned in special E. coli plasmids can be nor contain mu ltiple introns, they arc sufficiently compact
introduced into mutant yeast cells to identify the wild-type gene that the entire sequence of a gene can be included in a ge-
f that is defective in the mutant strain. nomic DNA fragment inserted into a plasmid vector. To

5.2 DNA Cloning and Characterization 189

 ·.

construct a plasmid genomic library that is to be screened cells carry a plasmid-borne copy of the wild-type URAJ gene,
by functional complementation in yeast cells, the plasmid they can be selected by their ability to grow in the absence of
vector must be capable of replication in both E. coli cells uracil. Typically, about 20 petri dishes, each containing about
and yeast cells. This type of vector, capable of propagation 500 yeast transformants, are sufficient to represent the entire
in two different hosts, is called a shuttle vector. The struc- yeast genome. This collection of yeast transformants can be
ture of a typical yeast shuttle vector is shown in Figure maintained at 23 oc, a temperature permissive for growth of
5-1 7a. This vector contains the basic elements that permit the cdc28 mutant. The entire collection on 20 plates is then
cloning of DNA fragments in E. coli. In addition, the shuttle transferred to replica plates, which are placed at 36 °C, a
vector contains an autonomously rephcattng sequence nonpermissive temperature for cdc mutants. Yeast colonies
(ARS), which functions as an origin for DNA replication in that carry recombinant plasmids expressing a wild-type copy
yeast; a yeast centromere (called CEN), which allows faith-
ful segregation of the plasmid during yeast cell division; and
a yeast gene encoding an enzyme for uracil synthesis (URA3 ),
which serves as a selectable marker in an appropriate yeast Polylinker
(a)
mutant.
To increase the probability that all regions of the yeast
genome are successfully cloned and represented in the plas-
mid library, the genomic DNA usually is only partially di-
gested to yield overlapping restriction fragments of -10 kb. Shuttle vector
These fragments are then ligated into the shuttle vector in
which the polylinker has been cleaved with a restriction
enzyme that produces sticky ends complementary to those
on the yeast DNA fragments (Figure 5-17b). Because the
I 0-kb restriction fragments of yeast DNA are incorporated
into the shuttle vectors randomly, at least 10' E. coli colo- CEN
nics, each containing a particular recombinant shuttle vec-
(b)
tor, arc necessary to ensure that each region of yeast DNA

0
has a high probability of being represented in the library at
least once.
Figure 5-18 outlines how such a yeast genomic libraf}-
can be screened to isolate the wi ld-type gene corresponding Yeast genomic DNA
to one of the temperature-sensitive cdc mutations mentioned Shuttle vector
earlier in this chapter. The starting yeast strain is a double
mutant that requires uracil for growth due to a ura3 muta-
tion and is temperature sensitive due to a cdc28 mutation
!Cut with BamHI
j Partially digest
with Sau3A

identified by its phenotype (see hgure 5-6). Recombinant ~ "VJilO!lOl~

plasmids isolated from the yeast genomic library are mixed
with yeast cells under conditions that promote transforma-
tion of the cells with foreign DNA. Since transformed yeast

EXPERIMENTAL FIGURE 5-17 A yeast genomic library can be

constructed in a plasmid shuttle vector that can replicate in yeast
and E. coli. (a) Components of a typical plasmid shuttle vector for
cloning Saccharomyces genes. The presence of a yeast origin of DNA
replication (ARS) and a yeast centromere (CEN) allows stable replication
and segregation in yeast. Also included is a yeast selectable marker
0000 Transform E. coli
such as URA3, which allows a ura3 mutant to grow on medium lacking Screen for ampicillin resistance
uraciL Finally, the vector contains sequences for replication and
selection in E. coli (ORI and amp') and a polylinker for easy insertion of
yeast DNA fragments. (b) Typical protocol for constructing a yeast
genomic library. Partial digestion of total yeast genomic DNA with
Sau3A is adjusted to generate fragments with an average size of about
Isolate and pool recombinant

!
10 kb. The vector is prepared to accept the genomic fragments by
plasmids from 105 transformed
digestion with Bam HI, which produces the same sticky ends as Sau3A.
E. coli colonies
Each transformed clone of E. coli that grows after selection for
ampicillin resistance contains a single type of yeast DNA fragment. Assay yeast genomic library by functional complementation

190 CHAPTERs • Molecular Genetic Techniques

 Library of yeast genomic DNA
carrying URA3 selective marker

Temperature-sensitive
cdc-mutant yeast;
ura3- (requires uracil)
0000
Transform yeast by treatment with
LiOAC, PEG, and heat shock

Plate and incubate at

Only colonies
1 permissive temperature
on medium lacking uracil
Only colonies carrying
a wild-type CDC gene
are able to grow

carrying a
URA3 marker
are able to
grow
Replica-plate and
incubate at nonpermissive
temperature
Q 36 C
.'

EXPERIMENTAL FIGURE 5 18 Screening of a yeast genomic mutant yeast cells under conditions that promote transformation. The
library by functional complementation can identify clones ca rrying relatively few transformed yeast cells, which contain recombinant
the normal form of a mutant yeast gene. In this example, a wild-type plasmid DNA, can grow in the absence of uracil at 23 C. When
CDC gene is isolated by complementation of a cdc yeast mutant. The transformed yeast colonies are replica-plated and placed at 36 C
Saccharomyces strain used for screening the yeast library carries ura3 (a non permissive temperature), only clones carrying a library plasmid
and a temperature-sensitive cdc mutation. This mutant strain is grown that contains the wild-type copy of the CDC gene will survive. LiOAC -
and maintained at a permissive temperature (23 °(}. Pooled recombi- lithium acetate; PEG polyethylene glycol.
nant plasmids prepared as shown in Figure 5-17 are incubated with the

of the CDC28 gene will be able tO grow at 36 oc.

Once tem- nucleotides can be separated electrophoretically on poly-
perature-resistant yeast colonies have been identified, plas- acrylamide gels, and larger molecules from about 200 nucle-
mid DNA can be extracted from the cultured yeast cells and otides to more than 20 kb on agarose gels.
analyzed by subcloning and DNA sequencing, topics we take A common method for visualizing separated DNA bands
up next. on a gel is to incubate the gel in a solution containing the fluo-
rescent dye ethidium bromide. This planar molecule binds to
DNA by intercalating between the base pairs. Binding con-
centrates ethidium in the DNA and also increases 1ts intrinsic
Gel Electrophoresis Allows Separation of Vector
fluorescence. As a result, when the gel is illuminated with
DNA from Cloned Fragments ultraviolet light, the regions of the gel containing DNA fluo-
In order to manipulate or sequence a cloned DNA fragment, it resce much more brightly than the regions of the gel without
sometimes must first be separated from the vector DNA. This DNA.
can be accomplished by cu.rting the recombinant DNA clone Once a cloned DNA fragment, especially a long one, has
with the same restriction enzyme used to produce the recombi- been separated from vector DNA, it often is treated with
nant vectors originally. The cloned DNA and vector DNA then various restriction enzymes to yield smaller fragments. After
are subjected to gel electrophoresis, a powerful method for separation by gel electrophoresis, all or some of these
separating DNA molecules of different size (Figure 5-19). smaller fragments can be ligated individually into a plasmid
Near neutral pH, DNA molecules carry a large negative vector and cloned in £. coli by the usual procedure. This
charge and therefore move toward the positive electrode process, known as subcloning, is an important step in rear-
during gel electrophoresis. Because the gel matrix restricts ranging parts of genes into useful new configurations. For
random diffusion of the molecules, molecules of the same instance, an investigator who wants to change rhe condi-
length migrate together as a band whose width equals that of tions under which a gene is expressed might use subcloning
the well into which the original DNA mixture was placed at to replace the normal promoter associated with a cloned
the start of the electrophoretic run. Smaller molecules move gene with a DNA segment containing a different promoter.
through the gel matrix more readily than larger molecules, Subcloning also can be used to obtain cloned DNA frag-
so that molecules of different length migrate as distinct ments that are of an appropriate length for determining the
bands. Smaller DNA molecules from about 10 to 2000 nucleotide sequence.

5.2 DNA Cloning and Characterization 191

 (a} DNA restriction fragments (b)

~~

~ Place mixture in the well of

1
an agarose or polyacrylamide
gel. Apply electric field

EXPERIMENTAL FIGURE 5-19 Gel electrophoresis separates

DNA molecules of d ifferent lengths. (a} A gel is prepared by pouring
a liquid containing either melted agarose or unpolymerized acrylamide
between two glass plates a few millimeters apart. As the agarose
solidifies or the acrylamide polymerizes into polyacrylamide, a gel
matrix (orange ovals} forms consisting of long, tangled chains of
polymers. The dimensions of the interconnecting channels, or pores,
depend on the concentration of the agarose or acrylamide used to
form the gel. The separated bands can be visualized by autoradiogra-
phy (if the fragments are radiolabeled} or by addition of a fluorescent
dye (e.g., ethidium bromide} that binds to DNA. (b) A photograph of
a gel stained with ethidium bromide (EtBr}. EtBr binds to DNA and
fluoresces under UV light. The bands in the far left and far right lanes
are known as DNA ladders-DNA fragments of known size that serve
as a reference for determining the length of the DNA fragments in the
experimental sample. [Part (b) Science Photo Library.]
Molecules move through pores

1 in gel at a rate inversely

proportional to their chain length

The Polymerase Chain Reaction Amplifies a

Specific DNA Sequence from a Complex Mixture
If the nucleotide sequences at the ends of a particular DNA
region are known, the intervening fragment can be amplified
directly by the polymerase chain reaction (PCR). Here we de-
scribe the basic PCR technique and three situations in which
it is used.
The PCR depends on the ability to alternately denature
(melt) double-stranded DNA molecules and hybridize com-
plementary single strands in a controlled fashion. As out-
+ lined in Figure 5-20, a typical PCR procedure begins by
heat-denaturation at 95 oc of a DNA sample into single

l
Subject to autoradiography strands. Next, two synthetic oligonucleotides complemen-
or incubate with fluorescent dye
tary to the 3' ends of the target DNA segment of interest are
added in great excess to the denatured DNA, and the tem-
perature is lowered to 50-60 °C. These specific oligonucle- t .,
otides, which are at a very high concentration, will hybridize
with their complementary sequences in the DNA sample,
whereas the long strands of the sample DNA remain apart
because of their low concentr::nion. Tht> hybridi?ed oligonu-
Signals cleotides then serve as primers for DNA chain synthesis in
corresponding the presence of deoxynucleotides (dNTPs) and a temperature-
to DNA bands resistant DNA polymerase such as that from Thermus aquat-
icus (a bacterium that lives in hot springs). This enzyme,
called Taq polymerase, can remain active even after being
heated to 95 °C and can extend the primers at temperatures
up to 72 oc. When synthesis is complete, the whole mixture

192 CHAPTERs • Molecular Genetic Techniques

 @ TECHNIQUE ANIMATION: Polymerase Chain Reaction

FIGURE 5-20 The polymerase chain reaction (PCR) is w idely used

to amplify DNA regions of known sequences. To amplify a specific
Cycle 1 1 Denaturation of DNA
region of DNA, an investigator will chemically synthesize two different t Annealing of primers
oligonucleotide primers complementary to sequences of approximately !lg 'tS
18 bases flanking the region of interest (designated as light blue and
dark blue bars). The complete reaction is composed of a complex
mixture of double-stranded DNA (usually genomic DNA containing the ~ Elongation of primer:;
target sequ~nce of interest), a stoichiometric excess of both primers,
the four deoxynucleoside triphosphates, and a heat-stable DNA
polymerase known as Taq polymerase. During each PCR cycle, the
reaction mixture is first heated to separate the strands and then cooled
Cycle 2 1 Denaturation of DNA
to allow the primers to bind to complementary sequences flanking the t Annealing of primers
region to be amplified. Taq polymerase then extends each primer from
its 3' end, generating newly synthesized strands that extend in the 3'
direction to the 5' end of the template strand. During the third cycle,
two double-stranded DNA molecules are generated equal in length to
the sequence of the region to be amplified. In each successive cycle the
target segment, which will anneal to the primers, is duplicated and will
eventually vastly outnumber all other DNA segments in the reaction ~ Elongation of primers
mixture. Successive PCR cycles can be automated by cycling the
reaction for timed intervals at high temperature for DNA melting and
F s..;s

at a defined lower temperature for the annealing and elongation parts

of the cycle. A reaction that cycles 20 times will amplify the specific
target sequence 1-million-fold.
I Denaturation of DNA
Cycle 3
t Annealing of primers

is then heated to 95 oc to denature the newly formed DNA

duplexes. After the temperature is lowered again, another
cycle of synthesis takes place because excess primer is still
present. Repeated cycles of denaturation (heating) followed
by hybridization and synthesis (cooling) quickly amplify the
sequence of interest. At each cycle, the number of copies of
the sequence between the primer sites is doubled; therefore, ~ Elongation of primers
the desired sequence increases exponentially-about a million-
fold after 20 cycles-whereas all other sequences in the orig-
inal DNA sample remain unamplified. * ;;;;;,;;; ;; 7
I 4th &u¥hiSS%§SW±WH'S' }1

Direct Isolation of a Specific Segment of Genomic DNA For W <, •A'"""~ >f o<§¥1

organisms in which all or most of the genome has been

sequenced, PCR amplificat.jon starting with the total genomic
DNA often is the easiest way to obtain a specific DNA region
of interest for cloning. In this application, the two oligonu-
cleotide primers are designed to hybridize to sequences ~
flanking the genomic region of interest and to include Cycles 4, 5, 6, et c.
sequences that are recognized by specific restriction enzymes
(Figure 5-21). After amplification of the desired target
sequence for about 20 PCR cycles, cleavage with the appro-
priate restriction enzymes produces sticky ends that allow
efficient llgation of the fragment into a plasmid vector Note that this method does not involve cloning of large
cleaved by the same restriction enzymes in the polylinker. numbers of restriction fragments derived from genomic
The resulting recombinant plasmids, all carrying the identi- DNA and their subsequent screening to identify the specific
cal genomic DNA segment, can then be cloned in E. coli fragment of interest. In effect, the PCR method inverts this
cells. With certain refinements of the PCR, even DNA seg- traditional approach and so avoids its most tedious aspects.
ments greater than 10 kb in length can be amplified and The PCR method is useful for isolating gene sequences to be
cloned in this way. manipulated in a variety of useful ways described later. In
.. 5.2 DNA Cloning and Characterization 193
 Region to be amplified EXPERIMENTAL FIGURE 5·21 A specific target region in total
genomic DNA can be amplified by PCR for use in cloning. Each
primer for PCR is complementary to one end of the target sequence
3'
S' IGGATCC -3' and includes the recognition sequence for a restriction enzyme that
Primer 1 INA sy'1thesis does not have a site within the target region. In this example, primer 1
contains a 8amHI sequence, whereas primer 2 contains a Hind Ill

3' --c:c===========o- -
\G GAT C C M! :; f%£f'f <§£@§§ $%4£ 3'
sequence. (Note that for clarity, in any round amplification of only one
of the two strands is shown, the one in brackets.) After amplification,
the target segments are treated with appropriate restriction enzymes,

e
3 ' I ITT c G A Al5 '
DNA synthe generating fragments with sticky ends. These can be incorporated into
Primer 2
2 complementary plasmid vectors and cloned in E. coli by the usual
IT)IG[JG[KA}T]C~c
£11•E":ii:%l'!hi:::%!ii:'fi:1$~@!l;l1lli:'Nffim¥~"'"ilff¥!ili@~~-l---- 3 '
procedure (see Figure 5-13).

3'ICCTAGGI I I ITTCGAAI59
S' IGGATCC -3'
Primer 1 II'IA syntllesis
Round 3 To ca rry our a quantitative RT-PCR react ion, the amount of
J. J. ,
dou ble-stranded DNA sequence p roduced by each ampl ifica-
3' lccrAGGI ITTCGAAI5 . I . d · d h IT · f · 1
s' IGfG Arc c Wl§'§h@i¢1§@4#'\(01 1 • AfA G c r rl3' non eye e IS etermme as t e amp 1 tcanon o a parttcu ar
mRNA sequence proceeds. By extrapola tion from these
8amHI site
. _ Hindlll site amounts, an esti mate of the amount of sta rting mRNA se-
Contmu e for= 20 · d ..Suc h quantita · RT -PC I''- carne
·d
PCR cycles quence can b e o b tame · tive

enzymes
j
Cut wit h restriction out on tissues or whole organisms using primers targeted to
genes of interest provide one of the most accurate means to
3' IGI I I ITTCGAI5' follow changes in gene expression.
5' IG AT c e M % ,~YW M AI3 '

Sticky end Sticky end

Pre paration of Probes Earlier we mentioned that o ligonucle-
j Ligate with plasmid vector
with sticky ends
otide probes for hybridization assays can be chemically synthe-
sized. Preparation of such probes by PCR amplification requires
chemical synthesis of only two relatively short primers cor-
responding to the two ends of the target sequence. The start-
ing sample for PCR amplification of the target sequence can
be a preparation of genomic DNA or a preparation of eDNA
synthesized from the total cellular mRNA. To generate a ra-
diolabeled product from PCR, 32 P-Iabeled dNTPs are in-
cluded during the last several amp li fication cycles or a
fluorescently labeled product can be obtained by using fluo-
rescently labeled dNTPs during the last amplification cycles.
Because probes prepared by PCR are relatively long and
have many radioactive or fluorescent nucleotides incorpo-
addttion, the PCR method can be used to isolate gene se-
rated into them, these probes usually give a stronger and
quences from mutant organisms to determine how they dif-
more specific signal than chemically synthesized probes.
fer from the wild type.
A variation on the PCR method allows PCR amplifica-
tion of a specific eDNA sequence fr om cellular mRNAs. This Tagging of Ge nes by Inse rtion Mutations Another useful ap-
method, known as reverse transcriptase-PCR (RT-PCR ), be- plication of the PCR is to amplify a "tagged" gene from the
gins with the same procedure described previously for isola- genomic DNA of a mutant strain. This approach is a simpler
tion of cD]';'A from a collection of cellular mRNAs. Typically, method for identifying genes associated with a particular
an oligo-dT primer, which will hybridize to the 3' poly(A) mutant phenotype than screening of a library by functional
tail of the mRNA, is used as the primer for the first strand of complementation (see Figure 5-18).
eDNA synthesis by reverse transcriptase. A specific eDNA The key to this use of the PCR is the ability to produce
can then be isolated from this complex mixture of cDNAs by mutations by insert ion of a known DNA sequence into the
PCR amplification using two oligonucleotide primers de- genome uf an experimemal urganism. Such insertion muta-
signed to march sequences at the 5' and 3' ends of the corre- tions can be generated by use of mobi le DNA elements,
sponding mRNA. As described previously, these primers which can move (or transpose) from one chromosomal site
could be designed to include restriction sires to facilitate the to another. As discussed in more detail in Chapter 6, these
insertion of amplified eDNA into a suitable p lasmid vector. DNA sequences occur naturally in the genomes of most or-
RT-PCR can be performed so that the starting amount of ganisms and may give rise to loss-of-function mutations if
a particular cellular mRNA can be determined accurately. they transpose into a protein-coding region.

194 CHAPTER 5 • Molecular Genetic Techniques

 Transposon FIGURE 5-22 The genomic sequence at the
insertion site of a transposon is revealed by PCR

Restriction sites: t t t amplification and sequencing. To obtain the DNA

sequence of the insertion site of a P-element transpo-
son it is necessary to PCR-amplify the junction

1 Cut wit h
restriction enzyme
between known transposon sequences and unknown
flanking chromosomal sequences. One method to
achieve this is to cleave genomic DNA with a restric
tion enzyme that cleaves once within the transposon
sequence. Ligation of the resulting restriction
fragments will generate circular DNA molecules. By
using appropriately designed DNA primers that match
transposon sequences, it is possible to PCR-amplify
the desired junction fragment. Finally, a DNA
sequencing reaction (see Figures 5-23 and 5-24) is

1
Ligate
to circularize performed using the PCR-amplified fragment as a
template and an oligonucleotide primer that matches
sequences near the end of the transposon to obtain
the sequence of the junction between the transposon
PCR primers
and chromosome.

Sequencing 1 PCR amplification

with primers to transposon

.....
primer

For example, researchers have modified a Drosophila mobile DNA clements or viruses with sequenced genomes
mobile DNA element, known as the P element, to optimize that can insert randomly into the genome.
its use in the experimental generation of insertion mutations.
Once it has been demonstrated that insertion of a P element
Cloned DNA Molecules Are Sequenced Rapidly
causes a mutation with an interesting phenotype, the genomic
sequences adjacent to the insertion site can be amplified by a by Methods Based on PCR
variation of the standard PCR protocol that uses synthetic The complete characterization of any cloned DNA fragment
primers complementary to the known P-element sequence requires determination of its nucleotide sequence. The tech-
but that allows unknown neighboring sequences to be am- nology used to determine the sequence of a DNA segment
plified. One such method, depicted in Figure 5-22, begins by represents one of the most rapidly developing fields in molec-
cleaving Drosophila genomic DNA containing a P-element ular biology. In the 1970s, F. Sanger and his colleagues devel-
insertion with a restriction enzyme that cleaves once within oped the chain-termination procedure, which served as the
the P-element DNA. The collection of cleaved Dl\A frag- basis for most DNA sequencing methods for the next 30 years.
ments treated with DNA ligase yields circular molecules, The idea behind this method is to synthesize from the DNA
some of w hich will contain P-element DNA. The chromo- fragment to be sequenced a set of daughter strands that are
somal region flanking the P element can then be amplified by labeled at one end and terminate at one of the four nucleo-
PCR using primers that match P-element sequences and are tidcs. Separation of the truncated daughter strands by gel elec-
elongated in opposite directions. The seq uence of the result- trophoresis, which can resolve strands that differ in length by
ing amplified fragment can then be determined using a third one nucleotide, can then reveal the length of all strands ending
DNA primer. The crucial sequence for identifying the site of in G, A, T, or C. From these collections of strands of different
P-clement insertion is the junction between the end of the lengths, the nucleotide sequence of the original DNA frag-
P-element and genomic seq uences. Overall, this approach ment can be established. The Sanger method has undergone
avoids the cloning of large numbers of DNA fragments and many refinements and now can be fully automated, but be-
their screening to detect a cloned DNA corresponding to a cause each new DNA sequence requires a separate individual
mutated gene of interest. sequencing reaction, the overall rate by which new DNA se-
Similar methods have been applied to other organisms quences can be produced by this method is limited by the total
for which insertion mutations can be generated using either number of reactions that can be performed at one time.

5.2 DNA Cloning and Characterization 195

 0 Cut one DNA strand,
denature, and wash,
leaving single-strand

fJ Add new primer, C-

then fluorescently
labeled dNTPs;
-T-
one dNTP binds, A-
wash away excess -G-

!
IJ Fluorescent imaging c liJI Repeat until
to determine which DNA strand
dNTP bound is replicated

~ and wash

!
c
IJ Chemically remove
bound fluorophore
and wash

~ DNA synthesis
~o?
II In ij
10
x PCR / ._
•
- •
-...

•
• •
••

'
r• ' ... •
•
••
I

·.

E KPERIMENTAL FIGURE 5-23 Generation of clusters of

.. •
•
.•• •
identical DNA molecules attached to a solid support. A large
•
collection of DNA molecules to be sequenced is ligated to double-
••
stranded linkers, which become attached to each end of the fragment.
The DNA is then amplified by PCR using primers matching the
sequences of the linkers that are covalently attached to a solid
substrate. Ten cycles of amplification yield about 1000 identical copies
•.•
'·
•
•
•
••
• .....• • • ... • •
-
•

EXPERIMENTAL FIGURE 5-24 Using fluorescent-tagged

of the DNA fragment localized in a small cluster, which is attached at deoxyribonucleotide triphosphates for sequence determination.
both ends to the solid substrate. These reactions are optimized to The reaction begins by cleaving one strand of the clustered DNA. After
produce as many as 3 x 109 discrete, non-overlapping clusters that are melting, a single DNA strand remains attached to the flow cell. A
ready to be sequenced. synthetic oligodeoxynucleotide is used as the primer for the polymer-
ization reaction that contains dNTPs, each fluorescently tagged with a
different color. The fluorescent tag is designed to block the 3 ' OH group
A breakthrough in sequencing technology occurred when on the dNTP so that once the fluorescent dNTP has been incorporated,
methods were devised to allow a single sequencing instrument further elongation is not possible. Because DNA polymerase w ill
incorporate the same fluorescent dNTP into each of the - 1000 DNA
to carry out billons of sequencing reactions simultaneously by
copies in a cluster, the entire cluster will be uniformly labeled with the
localizing them in tiny clusters on the surface of a solid sub-
same fluorescent color, which can be imaged in a special microscope.
strate. Since 2007, when the so-called next generation sequenc-
Once all of the clusters have been imaged, the fluorescent tags are
ers became commercially available, the capacity for new removed by a chemical reaction that leaves a new primer terminus
sequence production increased enormously and since then has available for the next cycle of fluorescently labeled dNTP addition. A
been doubling every year. In one popular sequencing method, typical sequencing reaction may carry out 100 polymerization cycles,
billions of different DNA strands to be sequenced are prepared allowing 100 bases of sequence for each cluster to be determined. Thus
by ligating double-stranded linkers to their ends (Figure 5-23 ). a total sequencing reaction ofthis type may generate as much as 3 X 10
11

Next the DNA fragments are amplified by PCR using primers bases of sequence information in about two days. [Open WetWare user
that match the linker sequences. This PCR amplification reaction Andrea Loehr (http://openwetware.org/wiki/ User:Andrea_Loehr).]

196 CHAPTERs • Molecular Genetic Techniques

differs from the standard PCR amplification shown in Figure Unknown genome of interest
5-20 in that the primers used are covalently attached to a solid
substrate. Thus as the PCR amplification proceeds, one end of
each daughter DNA strand is covalently linked to the substrate
and at the end of the amplification - 1,000 identical PCR Create aligned Create random

-- - --
library of eDNA library of eDNA
products are linked to the surface in in a tight cluster.
These clusters can then be sequenced by using a special
microscope to image fluorescently labeled deoxyribonucleo-
---
-----
--- --- --
tides (dNTPs) as they are incorporated by DNA polymerase
one at a rime into a growing DNA chain (Figure 5-24). First, l l

--- -
one strand is cut and washed out, leaving a single-stranded Sequence ordered Sequence unordered
fragments fragments
DNA template. Then sequencing is carried our on the - 1,000
identical templates in clusters, one nucleotide at a time. All
!!!!I~ -~ - 1!!!!!1
I!!!! 11!!!!!!1~
four dNTPs are fluorescently labeled and added to the se-
quencing reaction. After they are allowed to anneal, the sub-
-
e!!
1!!!!!!1
- er ~-~
strate is imaged and the color of each cluster is recorded. Next l
Read sequence in order
l
Align sequenced
the fluorescent tag is chemically removed and a new dNTP is
dictated by clone map clones by computer
allowed to bind. This cycle is repeated about 100 times, re-
sulting in billions of -100 nucleotide sequences. ""':L:Il!""" ~

In order to sequence a long continuous region of genomic

DNA or even the entire genome of an organism, researchers
usually employ one of the strategies outlined in Figure 5-25.
The first method requires the isolation of a collection of cloned Genomic sequence
DNA fragments whose sequences overlap. Once the sequence
EXPERIMENTAL FIGURE 5 -25 Two strategies for assembling
of one of these fragments is determined, oligonucleotides based whole genome sequences. One method (left) depends on isolating
on that sequence can be chemically synthesized for use as prim- and assembling a set of cloned DNA segments that span the genome.
ers in sequencing the adjacent overlapping fragments. In this This can be done by matching cloned segments by hybridization or by
way, the sequence of a long stretch of DNA is determined incre- alignment of restriction-site maps. The DNA sequence of the ordered
mentally by sequencing of the overlapping cloned DNA frag- clones can then be assembled into a complete genomic sequence. The
ments that compose it. A second method, which is called whole alternative method (right) depends on the relative ease of automated
genome shotgun sequencing, bypasses the time-consuming step DNA sequencing and bypasses the laborious step of ordering the
of isolating an ordered collection of DNA segments that span library. By sequencing enough random library clones so that each
the genome. This method involves simply sequencing random segment of the genome is represented from 3 to 10 times, it is possible
clones from a genomic library. A total number of clones are to reconstruct the genomic sequence by computer alignment of the
chosen for sequencing so that on average each segment of the very large number of sequence fragments.
genome is sequenced about 10 times. This degree of coverage
ensures that each segment of the genome is sequenced more
than once. The entire genomic sequence is then assembled using
a computer algorithm that aligns all the sequences, using their
regions of overlap. Whole genome shotgun sequencing is the fragments that often have self-complementary single-stranded
fastest and most cost-effective method for sequencing long tails (sticky ends).
stretches of DNA, and most gcnomes, including the human ge- • Two restriction fragments with complementary ends can
nome, have been sequenced by this method. be joined with DNA ligase to form a recombinant DNA mol-
ecule (see Figure 5-12).
• E. coli cloning vectors are small circular DNA molecules
(plasmids) that include three functional regions: an origin of
KEY CONCEPTS of Section 5.2
replication, a drug-resistance gene, and a site where a DNA
DNA Cloning and Characterization fragment can be inserted. Transformed cells carrying a vector
• In DNA cloning, recombinant DNA molecules are formed grow into colonies on the selection medium (see Figure 5-13 ).
in vitro by inserring DNA fragments into vector DNA mole- • A eDNA library is a ~f't of eDNA clones prepared from the
cules. The recombinant DNA molecules are then introduced mRNAs isolated from a particular type of tissue. A genomic
into host cells, where they replicate, producing large numbers library is a set of clones carrying restriction fragments pro-
of recombinant DNA molecules. duced by cleavage of the entire genome.
• Restriction enzymes (endonucleases) typically cut DNA at • In eDNA cloning, expressed mRNAs arc revcrse-tran~cribcd
specific 4- to 8-bp palindromic sequences, producing defined into complementary DNAs, or cDNAs. By a series of reactions,

5.2 DNA Cloning and Characterization 197

 ration by gel electrophoresis and hybridization with a comple-
single-stranded cDNAs are converted into double-stranded mentary DNA probe that is either radioactively or fluorescently
DNAs, which can then be ligated into a plasmid vector (see labeled. A third method involves hybridizing labeled probes
Figure 5-15). directly onto a prepared tissue sample. We will encounter
• A particular cloned DNA fragment wit hin a library can be references to all three of these techniques, which have nu-
detected by hybridization to a radiolabeled oligonucleotide merous applications, in other chapters.
whose sequence is complementary to a part of the fragment
(sec Figure 5-16). Southern Blotting The first hybridization technique to detect
DNA fragments of a specific sequence is known as Southern
Shuttle vectors that replicate in both yeast and E. coli can
blotting, after its originator E. !vi. Southern. This technique
be used to construct a yeast genomic library. Specific genes
is capable of detecting a single specific restriction fragment
can be isolated by their ability to complement the corre-
in the highly complex mixture of fragments produced by
sponding mutant genes in yeast cells (see Figure 5-17).
cleavage of the entire human genome with a restriction en-
Long cloned DNA fragments can be cleaved with restriction zyme. When such a complex mixture is subjected to gel elec-
enzymes, producing smaller fragments that are then separated trophoresis, so many different fragments of nearly the same
by gel electrophoresis and subcloned into plasmid vectors prior length are present it is not possible to resolve any particular
to sequencing or experimental manipulation. DNA fragments as a discrete band on the gel. Nevertheless,
• The polymerase chain reaction (PCR) permits exponential it is possible to identify a particular fragment migrating as a
amplification of a specific segment of DNA from just a single hand on the gel by its ability to hybridize to a specific DNA
initial template DNA molecule if the sequence flanking the probe. To accomplish this, the restriction fragments present
DNA region to be amplified is known (see Figure 5-20). in the gel arc denatured with alkali and transferred onto a
nitrocellulose filter or nylon membrane by blotting (Figure
PCR is a highly versatile method that can be programmed
5-26). This procedure preserves the distribution of the frag-
to amplify a specific genomic DNA sequence, a eDNA, or a
ments in the gel, creating a replica of the gel on the filter.
sequence at the junction between a transposable element and
(The blot is used because probes do not readily diffuse into
flanking chromosomal sequences.
the original gel.) The filter then is incubated under hybrid-
• DNA fragments up to about 100 nucleotides long are se- ization conditions with a specific labeled DNA probe, which
quenced by generating clusters of identical molecules by PCR usually is generated from a cloned restriction fragment. The
and imaging fluorescently labeled nucleotide precursors in- DNA restriction fragment that is complementary to the
corporated by DNA polymerase (see Figures 5-23 and 5-24). probe hybridizes, and its location on the filter can be re-
• Whole genome sequences can be assembled from the se- vealed by autoradiography for a radiolabelcd probe or by
quences of a large number of overlapping clones from a ge- fluorescent imaging for a fluorescently labeled probe . Al-
nomic library (see Figure 5-25). though PCR is most commonly used to detect the presence
of a particular sequence in a complex mixture, Southern
blotting is still useful for reconstructing the relationship be-
tween genomic sequences that are too far apart to be ampli-
fied by PCR in a single reaction.
5.3 Using Cloned DNA Fragments
Northern Blotting One of the most basic ways to characterize
to Study Gene Expression
a cloned gene is to determine when and where in an organism
In the last section, we described the basic techniques for using the gene is expressed. Expression of a particular gene can be
recombinant DNA technology to isolate specific DNA clones, followed by assaying for the corresponding mRNA by North-
and ways in which the clones can be further characterized. ern blotting, named, in a play on words, after the related
Now we consider how an isolated DNA clone can be used to method of Southern blotting. An RNA sample, often the total
study gene expression. We discuss several widely used general cellular RNA, is denatured by treatment with an agent such
techniques that rely on nucleic acid hybridization to elucidate as formaldehyde that disrupts the hydrogen bonds between
when and where genes are expressed, as well as methods for base pairs, ensuring that all the RNA molecules have an un-
generating large quantities of protein and otherwise manipu- folded, linear conformation. The individual RNAs are sepa-
lating amino acid sequences to determine their expression pat- rated according to size by gel electrophoresis and transferred
terns, structure, and function. More specific applications of all to a nitrocellulose filter to which the extended denatured
these basic techniques are examined in the following sections. RNAs adhere. As in Southern blotting, the filter then is ex-
posed to a labeled DNA probe that is complementary to the
gene of interest; finally, the labeled filter is subjected to auto-
Hybridization Techniques Permit Detection
radiography. Because the amount of a specific RNA in a
of Specific DNA Fragments and mRNAs sample can be estimated from a Northern blot, the procedure
Two very sensitive methods for detecting a particular DNA is widely used to compare the amounts of a particular mRNA
or RNA sequence within a complex mixture combine sepa- in cells under different conditions (Figure 5 -27).

198 CHAPTERs • Molecular Genetic Techniques

 ~ DNA

1
Cleave with
restriction enzymes
Gel Nitrocellulose Autoradiogram

Filter Nitrocellulose Hybridize with

paper Gel labeled DNA or
RNA probe

Capillary action transfers

DNA from gel to nitrocellulose
EXPERIMENTAL FIGURE 5-26 Sout hern blot technique can DNA fragments. Only fragments that hybridize to a labeled probe will
detect a specific DNA fragment in a complex mixture of restrict ion give a signal on an autoradiogram. A similar technique called Northern
fragments. The diagram depicts three different restriction fragments blotting detects specific mRNAs within a mixture. [See E. M. Southern,
in the gel, but the procedure can be applied to a mixture of millions of 1975, J. Mol. Bioi. 98:508.]

In Situ Hybridization Northern blotting requires extracting subjected to in situ hybridization to detect the mRNA encoded
the mRNA from a cell or mixture of cells, which means that by a particular gene. This technique allows gene transcription
the cells arc removed from their normal location within an to be monitored in both time and space (Figure 5-28).
organism or tissue. As a result, the location of a cell and its
relation to its neighbors is lost. To retain such positional infor-
DNA Microarrays Can Be Used to Evaluate
mation in precise studies of gene expression, a whole or sec-
tioned tissue or even a whole permcabilized embryo may be the Expression of Many Genes at One Time
Monitoring the expression of thousands of genes simultane-
ously is possible with DNA microarray analysis, another
UN 48 h 96 h technique based on the concept of nucleic acid hybridization.
• A DNA microarray consists of an organized array of thou-
sands of individual, closely packed gene-specific sequences
attached to the surface of a glass microscope slide. By cou-
pling microarray analysis with the results from genome se-
5kb - quencing projects, researchers can analyze the global patterns
of gene expression of an organism during specific physiologi-
cal responses or developmental processes.

Preparation of DNA Microarrays In one method for prepar-

ing microarrays, an = 1-kb part of the coding reg10n of each
gene analyzed is individually amplified by the PCR. A ro-
2 kb- botic device is used to apply each amplified DNA sample to
the surface of a glass microscope slide, which then is chemi-
cally processed to permanently attach the DNA sequences to
.. 1 kb - the glass surface and to denature them. A typical array might
contain =6000 spots of DNA in a 2 X 2-cm grid.
- 0.65 kb- J3·globin In an alternative method, multiple DNA oligonucleotides,
-mANA
usually at least 20 nucleotides in length, are synthesized from
an initial nucleotide that is covalently bound to the surface of
EXPERIMENTAL FIGURE 5-27 Northern blot analysis reveals
a glass slide. The synthesis of an oligonucleotide of specific
increased expression of Jl-globin mRNA in differentiated erythro-
leukemia cells. The total mANA in Pxtracts of erythroleukemia cells
sequence can be programmed in a small region on the surface
that were growing but uninduced and in cells induced to stop growing of the slide. Several oligonucleotide sequences from a single
and allowed to differentiate for 48 hours or 96 hours was analyzed gene are thus synthesized in neighboring regions of the slide to
by Northern blotting for 13-globin mANA. The density of a band is analyze expression of that gene. With this method, oligonu-
proportional to the amount of mRNA present. The 13-globin mRNA is cleotides representing thousands of genes can be produced on
barely detectable in uninduced cells (UN lane) but increases more than a single glass slide. Because the methods for constructing
1000-fold by 96 hours after differentiation is induced. [Courtesy of L. Kole.] these arrays of synthetic oligonucleotides were adapted from

5.3 Using Cloned DNA Fragments to Study Gene Expression 199

 .·

(a) (b) (c)

Head

..

EXPERIMENTAL FIGURE 5-28 In situ hybridization can detect has been removed, substrate for the reporter enzyme is added. A
activity of spe cific ge nes in whole and sectioned embryos. The colored precipitate forms where the probe has hybridized to the mRNA
specimen is permeabilized by treatment with detergent and a protease being detected. (a) A whole mouse embryo at about 10 days of
to expose the mRNA to the probe. A DNA or RNA probe, specific for the development probed for Sonic hedgehog mRNA. The stain marks the
mRNA of interest, is made with nucleotide analogs containing chemical notochord (red arrow), a rod of mesoderm running along the future
groups that can be recognized by antibodies. After the permeabilized spinal cord. (b) A section of a mouse embryo similar to that in part (a).
specimen has been incubated with the probe under conditions that The dorsal/ventral axis of the neural tube (ND can be seen, with the
promote hybridization, the excess probe is removed with a series of Sonic hedgehog-expressing notochord (red arrow) below it and the
washes. The specimen is then incubated in a solution containing an endoderm (blue arrow) still farther ventral. (c) A whole Drosophila
antibody that binds to the probe. This antibody is covalently joined to embryo probed for an mRNA produced during trachea development.
a reporter enzyme (e.g., horseradish peroxidase or alkaline phospha- The repeating pattern of body segments is visible. Anterior (head) is up;
tase) that produces a colored reaction product. After excess antibody ventral is to the left. [Courtesy of L. Milenkovic and M.P. Scott.]

methods for manufacturing microscopic integrated circuits are not transcribed under these growth conditions give no
used in computers, these types of o ligonucleotide microarrays detectable signal. Genes that are transcribed at the same level
arc often called DNA chips. under both conditions will hybridize equally to both red- and
green-labeled eDNA preparations. Microarray analysis of gene
Using Microarrays to Compare Gene Expression Under Dif- expression in fibroblasts showed that transcription of about
ferent Cond it ions The initial step in a microarray expression 500 of the 8600 genes examined changed substantially after
study is to prepare fluorescently labeled cDNAs correspond- addition of serum.
ing to the mRNAs expressed by the cells under study. When
the eDNA preparation is applied to a microarray, spots repre-
Cluster Analysis of Multiple Expression
senting genes that are expressed will hybridize under appro-
priate conditions to their complementary cDNAs in the Experiments Identifies Co-regulated Genes
labeled probe mix and can subsequently be detected in a scan- firm conclusions rarely can be drawn from a single microar-
ning laser microscope. ray experiment about whether genes that exhibit similar
Figure 5-29 depicts how this method can be applied to changes in expression are co-regulated and hence likely to be
examine the changes in gene expression observed after starved closely related functionally. For example, many of the ob-
human fibroblasts are transferred to a rich, serum-containing served differences in gene expression just described in fibro-
growth medium. In this type of experiment, the separate blasts could be indirect consequences of the many different
eDNA preparations from starved and serum-grown fibro- changes in cell physiology that occur when cells arc trans-
blasts are labeled with differently colored fluorescent dyes. A ferred from one med ium to another. In other words, genes
DNA array comprising 8600 mammalian genes then is incu- that appear to be co-regulated in a single m icroarray expres-
bated with a mixture containing equal amounts of the two sion experiment may undergo changes in expression for very
eDNA preparations under hybridiza tion conditions. After d ifferent reasons and may actually have very different bio-
unhybridized eDNA is washed away, the intensity of green logical fum:tions. ;\ solution to this problem is to combine
and red fluorescence at each D A spot is measured using a the information from a set of expression array expcnments
fluorescence microscope and stored in computer files under to find genes that are similarly regulated under a variety of
the name of each gene according to its known position on conditions or over a period of time.
the slide. The relative intensities of red and green fluores- This more informative use of multiple expression array
cence signals at each spot are a measure of the relative level experiments is illustrated by examining the relative expression
of expression of that gene in response to serum. Genes that of the 8600 genes at different times after serum addition,

2 00 CHAPTERs • Molecular Genetic Techniques

t,;) TECHNIQUE ANIMATION:Synthesizing an Oligonucleotide Array
TECHNIQUE ANIMATION: Screening for Patterns of Gene Expression

EXPERIMENTAL FIGURE 5 29 DNA microarray analysis can (a) Fibroblasts Fibroblasts

reveal differences in gene expression in fibroblasts under different without serum with serum added
experimental conditions. (a) In this example, eDNA prepared from
mRNA isolated from fibroblasts either starved for serum or after serum
addition is labeled with different fluorescent dyes. A microarray
composed o f DNA spots representing 8600 mammalian genes is
exposed to an equal mixture of the two eDNA preparations under
hybridization cond itions. The ratio of the intensities of red and green
! Isolate total mANA !
fluorescence over each spot, detected with a scanning confocal laser
microscope, indicates the relative expression of each gene in response
Green dye~ Reverse-transcribe
to eDNA labeled with
to serum. (b) A micrograph of a small segment of an actual DNA a fluorescent dye
microarray. Each spot in this 16 X 16 array contains DNA from a
different gene hybridized to control and experimental eDNA samples
labeled with red and green fluorescent dyes. (A yellow spot indicates
equal hybridization of green and red fluorescence, indicating no
change in gene expression.) [Part (b) Alfred Pasieka/Photo Researchers. Inc.]
cDNAs hybridized to
DNAs for a single gene
! Hybridize to DNA
m1croarray

generating more than 10 4 individual pieces of data. A com-

! Wash

puter program, related to the one used to determine the re-

latedness of different protein sequences, can organize these
! Measure green and red
flu orescence over each spot

data and cluster genes that show similar expression over the
time cou rse after serum addition. Remarkably, such cluster
analysis groups sets of genes whose encoded proteins par-
ticipate in a common cellular process, such as cholesterol bio-
synthesis or the cell cycle (Figure 5-30).

In the furure, rnicroarray analysis will be a powerful diag-

nostic tool in medicine. For instance, particular sets of
I-
mRNAs have been found tO distinguish tumors w ith a poor A If a spot is green, expr ession of that gene decreases in
prognosis from those with a good prognosis. Previously in- cells after serum addition
distinguishable disease variations are now detectable. Analysis
If a spot is red, expression of that gene increases in cells
of tumor biopsies for these distinguishing mRNAs will help after serum addition
physicians to select the most appropriate treatment. As more
patterns of gene expression characteristic of various diseased (b)
tissues are recognized, the diagnostic use of DNA microarrays
will be extended to other conditions. •

E. coli Expression Systems Can Produce Large

Quantities of Protein~ from Cloned Genes
Many protein hormones and other signaling or regula-
tory proteins are normally expressed at very low con-
centrations, precluding their isolation and purification in large
quantities by standard biochemical techniques. Widespread
therapeutic use of such proteins, as well as basic research on
their strucrure and functions, depends on efficient procedures
for producing them in large amounts at reasonable cost. Re-
combinant DNA teL:hniques that turn E. coli cells into facto-
ries fo r synthesizing low-abundance proteins now are used to
commerciall y produce granulocyte-colony-stimulating factor
(G-CSF), insulin , growth hormone, and other human pro-
teins with therapeutic uses. For example, G-CSF stimulates
the production of granulocytes, the phagocytic white blood

5.3 Usrng Cloned DNA Fragments to Study Gene Expression 201

 Each column represents a different gene at
times after addition of serum

Ql
E
i=

A B c D
EXPERIMENTAL F GURE 5-30 Cluster analysis of data from expression. The "tree" diagram at the top shows how the expression
multiple microarray expression experiments can identify co- patterns for individual genes can be organized in a hiera rchical fashion
regulated genes. The expression of 8600 mammalian genes was to group together the genes with the greatest similarity in their
detected by microarray analysis at time intervals over a 24-hour period patterns of expression over time. Five clusters of coordinately
after serum-starved fibroblasts were provided with serum. The cluster regulated genes were identified in this experiment, as indicated by the
diagram shown here is based on a computer algorithm that groups bars at the bottom. Each cluster contains multiple genes whose
genes showing similar changes in expression compared with a encoded proteins function in a particular cellular process: cholesterol
serum-starved control sample over time. Each column of colored boxes biosynthesis (A), the cell cycle (8), the immediate-early response (C),
represents a single gene, and each row represents a time point. A red signaling and angiogenesis (D), and wound healing and tissue
box indicates an increase in expression relative to the control; a green remodeling (E). [Courtesy of Michael B. Eisen, Lawrence Berkeley National
box, a decrease in expression; and a black box, no significant change in Laboratory.]

cel ls critical to defense against bacterial infections. Adminis-

tration of G-CSF to cancer patients helps offset the reduction
1n gran ulocyte production caused by chemotherapeutic
(a)

agents, thereby protecting patients against serious infection

while they arc receiving chemotherapy. •

The first step in producing large amounts of a low-abundance

protein is to obtain a eDNA clone encoding the full-length
protein by methods discussed previously. The second step is to - IPTG + IPTG
engineer plasmid vectors that will express large amounts of
the encoded protein when it is inserted into E. coli cells. The
~
(b)
key to designing such expression vectors is inclusion of a pro-
moter, a DNA sequence from which transcription of the
J ))eDNA
G-CSF

c
eDNA can begin. Consider, for example, the relatively simple
system for expressing G-CSF shown in Figure 5-31. In this Plasmid

~~JacZ
expression
vector

PlERIME'NT F GURE 5· 31 Some eukaryotic proteins can / ) gene

be produced in E. coli cells from plasmid vectors containing the lac
promoter. (a) The plasmid expression vector contains a fragment of
the E. coli chromosome containing the lac promoter and the neighbor-
ing lacZ gene. In the presence of the Ia ctose analog IPTG, RNA
1 Transform
E. coli

n
polymerase normally transcribes the lacZ gene, producing lacZ mRNA,
8

C6
~G~SF
r~moter ) -G-CSF-+~
which is translated into the encoded protein, 13-galactosidase. (b) The
lacZ gene can be cut out of the expression vector with restriction
enzymes and replaced by a cloned eDNA, in this case one encoding ~ mRNA ~
granulocyte-colony-stimulating factor (G-CSF). When the resulting eDNA G-CSF
protem
plasmid is transformed into E. coli cells, addition of IPTG and subse-
quent transcription from the lac promoter produce G-CSF mRNA,
........._
which is translated into G-CSF protein. - IPTG + IPTG

202 CHAPTER 5 • Molecular Genetic Techniques

 case, G-CSF is expressed in E. coli transformed with plasmid to the yeast shuttle vectors described previously. For use in
•. ' mammalian cells, plasmid vectors are engineered also to
vectors that contain the lac promoter adjacent to the cloned
eDNA encoding G-CSF. Transcription from the lac pro- carry an origin of replication derived from a virus that infects
moter occurs at high rates only when lactose, or a lactose mammalian cells, a strong promoter recognized by mamma-
analog such as isopropylthiogalactoside (lPTG), is added to lian RNA polymerase, and the cloned eDNA encodtng the
the culture medium. Even larger quantities of a desired pro- protein to be expressed adjacent to the promoter (figure
tein can be produced in more complicated E. coli expression 5-32a). Once such a plasmid vector enters a mammalian cell,
systems. the viral origin of replication allows it to replicate efficiently,
To aid in purification of a eukaryotic protein produced generating numerous plasmids from which the protein is ex-
in an E. coli expression system, resea rchers often modify the pressed. However, during cell division such plasmids are not
eDNA encoding the recombinant protein to facilitate its ~ep­ faithfully segregated into both daughter cells and in time a
aration from endogenous E. coli proteins. A commonly used substantial fraction of the cells in a culture will not contain a
modification of this t ype is to add a short nucleotide se- plasmid, hence the name transient transfection.
quence to the end of the eDNA, so that the expressed protein
wi ll have six histidine residues at the C-terminus. Proteins Stable Transfection (Transformation) If an introduced vector
modified in this way bind tightly to an affinity matrix that integrates into the genome of the host cell, the genome is per-
.· contains chelated nickel atoms, w hereas most E. coli pro- manently altered and the cell is said to be transformed. Integra-
teins will not bind to such a matrix. The bound proteins can tion most likely is accomplished by mammalian enzymes that
be released from t he nickel atoms by decreasing the pH of normally function in DNA repair and recombination. A com-
the surrounding medium. In most cases, this procedure yields monly used selectable marker is the gene for neomycin phos-
a pure recombinant protein that is functional, since addition photransferase (designated neo'), which confers resistance to a
of short amino acid seq uences to either the C-terminus or the toxic compound chemically related to neomycin known as
N-terminus of a protein usually does not interfere with the G-418. The basic procedure for expressing a cloned eDNA by
protein's biochemical activity. stable transfection is outlined in Figure 5-32b. Only those cells
that have integrated the expression vector into the host chro-
mosome will survive and give rise to a clone ·in the presence of
Plasmid Expression Vectors Can Be a high concentration of G-418. Because integration occurs at
Designed for Use in Animal Cells random sites in the genome, individual transformed clones re-
sistant to G-41 8 will differ in their rates of transcribing the
Although bacterial expression systems can be used success- inserted eDNA. Therefore, the stable transfectants usually are
fully to create large q uantities of some proteins, bacteria screened to identify those that produce the protein of interest
cannot be used in all cases. Many experiments to examine at the highest levels.
the function of a protein in an appropriate cellular context
require expression of a genetically modified protein in cul- Retroviral Expression Systems Researchers have exploited
tured animal cells. Genes are cloned into specialized eukary- the basic mechanisms used by viruses for introduction of ge-
otic expression vectors a nd are introduced into cultured netic material into anima l cells and subsequent insertion into
animal cells by a process called transfection. Two common chromosomal DNA to greatly increase the efficiency by
methods for transfecting animal cells differ in whether the which a modified gene can be stably expressed in animal
recombinant vector DNA is or is not integrated into the cells. An example of one such viral expression is derived
host-cell genomic DNA. from a class of retroviruses known as lentiviruses. As shown
In both methods, cultured animal cells must be treated to in Figure 5-33, three different plasmids, introduced into cells
facilitate their initial uptake of the recombinant plasmid vec- by transient transfection, are used to produce recombinant
tor. T his can be done by exposing cells to a preparation of lentivirus particles suitable for efficient introduction of a
lipids that penetrate the'plasma membrane, increasing its cloned gene into target animal cells. The first plasmid,
permeabili ty to DNA. Alternatively, subjecting cells to a known as the vector plasmid, contains a cloned gene of in-
brief electric shock of several thousand volts, a technique terest next to a selectable marker such as neo' flanked by
known as e/ectroporation, makes them transiently perme- lentivirus LTR sequences. As described in Chapter 6, viral
able to DNA. Usually the plasmid DNA is added in sufficient LTR sequences direct synthesis of a viral RNA molecule that
concentration to ensure that a large proportion of the cul- on introduction into a virally infected target cell can be cop-
tured cells will receive at least one copy of the plasmid DNA. ied into DNA by reverse transcription and then integrated
Researchers have a lso harnessed viruses for their use in the into chromosomal DNA. A ~econd plasmid known as the
laboratory; viruses can be modified to coma in DNA of inter- packaging plasmid carries all the viral genes, except for the
est, which is then introduced into host cells by simply infect- major vira l envelope protein, necessary for packaging LTR-
ing them with the recombinant virus. containing viral RNA into a functionallentivirus particle.
The final plasmid allows expression of a viral envelope pro-
Transient Transfection The simplest of the two expression tein that when incorporated into a recombinant lentivirus
methods, called transiellt transfection, employs a vector similar wi ll allow the resulting hybrid virus particles to infect a

5.3 Using Cloned DNA Fragments to Study Gene Expression 203

 (a) Transient t ransfection EXPERIMENTAL FIGURE 5 · 3 2 Transient and stable transfec-
eDNA tion with specially designed plasmid vectors permits expression of
cloned genes in cultured animal cells. Both methods employ plasmid
vectors that contain the usual elements-OR!, selectable marker (e.g.,
Viral origin of amp1, and polylinker-that permit propagation in E. coli and insertion
replication
of a cloned eDNA with an adjacent animal promot er. For simplicity,
these elements are not depicted. (a) In transient transfection, the
plasmid vector contains an origin of replication for a vi rus that can
Transfect cultured replicate in the cultured animal cells. Since the vector is not incorpo-

1 cells by lipid treatment

or electroporation
rated into the genome of the cultured cells, production of the
eDNA-encoded protein continues only for a limited time. (b) In stable
transfection, the vector carries a selectable marker such as neo', which
confers resistance to G-4 18. The relatively few transfected animal cells
that integrate the exogenous DNA into their genomes are selected on
medium containing G-4 18. Because the vector is integrated into the
genome, these stably transfected, or transformed, cells will continue to
Protein is expressed from eDNA in plasmid DNA produce the eDNA-encoded protein as long as the culture is main-
tained. See the text for discussion.
(b) Stable transfection (transformation)

eDNA

cells with a stably integrated cloned gene and neo' marker

can be selected for by resistance to G-418. Many of the tech-
niques for inactivating the function of specific genes (sec Sec-
tion 5 .5) require that an entire population 'of cultured cells
be genetically modified simultaneously. Engineered lentivi-

l
Transfect cultured
cells by lipid treatment ruses are particularly useful for such experiments because
or electroporation they infect cells with very high efficiency such that every cell
in a population will receive at least one copy of the lentivirus-

!Select for G-418 resistance

G-418-resistant
clones
borne plasmid.

Gene a nd Protein Tagging Expression vectors can provide a

way to study the expression and intracellular localization of
eukaryotic proteins. This method often relies on using a re-
porter protein such as green fluorescent protein (CFP) that
can conveniently be detected in cells. Here we describe two
ways to create a hybrid gene that connects expression of the
reporter protein to the protein of interest- When the hybrid
gene is reintroduced into cells, either by transfection with a
plasmid expression vector containing the modified gene or
Protein is expressed from eDNA integrated by creation of a transgenic animal as described in Section
into host chromosome 5.5, the expression of the reporter protein can be used to
determine where and when a gene is expressed. This method
provides similar data to in situ hybridization experiments
described previously but often with greater resolution and
sensitivity.
desired target cell type. A common envelope protei n used in Figure 5-34 illustrates the usc of two different types of
thts context is the glycoprotein of the vesicular stomatitis GFP-tagging experiments to study the expression of an odor-
virus (VSV-G protein), which can readily replace the normal ant receptor protein in C. elegans. When the promoter for
lentivirus envelope protein on the surface of completed virus the odorant receptor is linked directly to the coding sequence
particles and will allow the resulting virus particles to infect of GFP in a configuration usually known as a promoter-
a wide variety of mammalian cell types, including hemato- {llsion, GFP is expressed in specific neurons, filling the cyto-
poietic stem cells, neurons, and muscle and liver cells. After plasm of those neurons. In contrast, when the hybrid gene is
cell infection, the cloned gene flanked by viral LTR se- constructed by linking GFP to the coding sequence of the
quences is reverse-transcribed into DNA, which is trans- receptor, the resulting protein-fusion can be localized by
ported into the nucleus and then integrated into the host GFP fluorescence at the distal cilia in sensory neurons, the
genome. If necessary, as in the case for stable transfection, site at which the receptor protein is normally located.

204 CHAPTERs • Molecular Genetic Techniques

 EXPERIMENTAL FIGURE 5 .. 33 Retroviral
vectors can be used for efficient integration of
cloned genes into the mammalian genome.
See the text for discussion.

Retroviral RNA

~ Cloned
Gene for
viral coat
protein

Viral
origin of
replication

An alternative to GFP tagging for detecting the intracel- antibody. Such a short peptide that can be bound by an
lular location of a protein is to modify the gene of interest by antibody is called an epitope; hence this mechod is known as
fusing it with a short DNA sequence that encodes a short epitope tagging. After transfection with a plasmid expression
stretch of amino acids recognized by a known monoclonal vector containing the modified gene, the expressed epitope-
tagged form of the protein can be detected by immunofluo-
rescence labeling of the cells with the monoclonal antibody
(a) Promoter-fusion; ODR10 promoter fused to GFP specific for the epitope. The choice of whether to use a short
epitope or GFP to rag a given protein often depends on what
types of modification a cloned gene can tolerate and still
remain functional.

EXPERIMENTAL FIGURE 5-34 Gene and protein tagging facili-

tate cellular localization of proteins expressed from cloned genes.
(b) Protein-fusion; ODR10-GFP fusion protein
In this experiment, the gene encoding a chemical odorant receptor,
Odrl 0, of C. elegans was fused to the gene sequence for green
fluorescent protein (GFP). (a) A promoter-fusion was generated by
linking GFP to the promoter and the first four amino acid codons of
Odrl 0. This protein is expressed in the cytoplasm of specific sensory
neurons in the head of C. elegans. Note that the cell body (dashed
arrow) and sensory dendrites (solid arrow) are fluorescently labeled.
(b) A protein-fusion was construded by linking GFP to the end of the
full-length Odrl 0 coding sequence. In this case, the Odrl 0-GFP fusion
protein is targeted to the membrane at the tip of the sensory neurons
and is only apparent at the distal end of the sensory cilia. The
observed distribution can be inferred to reflect the normallocat1on
of Odrl 0 protein in specific neurons. Because the promoter-fusion
shown in (a) lacks the OdrlO localization sequences, the expressed
GFP fills the entire cell cytoplasm rather than being localized just to
distal tip of the sensory cilia. [P. Sengupta et al., 1996, Ce/184:899 (derived
from Figures 4 and 5).)

5.3 Using Cloned DNA Fragments to Study Gene Expression 205

 defective hemoglobin molecules, causing the sickle-like de-
KEY CONCEPTS of Section 5.3 formation of red blood cells in individuals who have inher-
ited two copies of the Hb' allele for sickle-cell hemoglobin.
Using Cloned DNA Fragments
Most often, however, the genes responsible for inherited
to Study Gene Expression
diseases must be found w ithout any prior knowledge or rea-
• Southern blotting can detect a single, specific DNA frag- sonable hypotheses about the nature of the affected gene or its
ment within a complex mixture by combining gel electro- encoded protein. In this section, we will see how human ge-
phoresis, transfer (blotting) of the separ ated bands to a filter, neticists can find the gene responsible for an inherited disease
and hybridization with a complementary radiolabeled DNA by following da: ~t::grt::gation of the disease in families. The
probe (see Figure 5-26). The similar technique of Northern segregation of the disease can be correlated with the segrega-
blotting detects a specific RNA within a mixture. tion of many other genetic markers, eventually leading to
• The presence and d istribution of specific mRNAs can be identification of the chromosoma l posit ion of t he affected
detected in living cells by in situ hybridization. gene. This information, along with knowledge of the sequence
of the human genome, can ultimately allow the affected gene
• DNA m1croarray analysis simultaneously detects the rela-
and the disease-causing mutations to be pinpointed.
tive level of expression of thousands of genes in different
types of cells or in the same cells under different conditions
(see Figure 5-29).
Cluster analysis of the data from multiple microarray expres- Monogenic Diseases Show One of Three
sion experiments can identify genes that are similarly regulated Patterns of Inheritance
under various conditions. Such co-regulated genes commonly Human genetic diseases that result from mutation in one spe-
encode proteins that have biologically related functions. cific gene are referred to as monogenic diseases and display
• Expression vectors derived from plasmids allow the pro- different inheritance patterns depending on the nature and
duction of abundant amounts of a protein from a cloned gene. chromosomal location of the alleles that cause them. One
characteristic pattern is that exhibited by a dominant allele in
Eukaryotic expression vectors can be used to express cloned
an autosome (that is, one of the 22 human chromosomes that
genes in yeast or mammalian cells. An important application
is not a sex chromosome). Because an autosomal dominant
of these methods is t he tagging of proteins with GFP or an
allele is expressed in the heterozygote, usually at least one of
epitope for antibody detection.
the parents of an affected individual will also have the dis-
ease. It is often the case that the diseases caused by dominant
alleles appear later in life after the reproductive age. If this
were not the case, natural selection would have eliminated
5.4 Locating and Identifying Human the allele during human evolution. An example of an autoso-
mal dominant disease is Huntington's disease, a neural de-
Disease Genes generative disease that generally strikes in mid- to late life. If
Inherited human diseases are the phenot ypic con- either parent carries a mutant HD allele, each of his or her
sequence of defective human genes. Table 5-2 lists children (regardless of sex) has a 50 percent chance of inherit-
several of the most commonly occurring inherited diseases. ing the mutant allele and being affected (Figure 5-35a).
Although a "disease" gene may result from a new mutation A recessive allele in an autosome exhibits a quite different
that arose in the preceding generation, most cases of inher- segregation pattern. For an autosomal recessive allele, both
ited diseases are caused by preexisting mutant a lleles that parents must be heterozygous carriers of the allele in order
have been passed from one generation to the next for many for their children to be at nsk of being affected with the dis-
generations. • ease. Each child of heterozygous parents has a 25 percent
chance of receiving both recessive alleles and thus being . ~ .
;

The typical first step in deciphering the underlying cause for affected, a 50 percent chance of receivi ng one normal and
any inherited human d1sease is to identify the affected gene one mutant allele and thus being a carrier, and a 25 percent
and its encoded protein. Comparison of the sequences of a chance of receiving two normal alleles. A clear example of an
disease gene and its product with those of genes and proteins autosomal recessive disease is cystic fibrosis, which results
whose sequence and function are known can provide clues from a defective chloride-channel gene known as CFTR (Fig-
to the molecular and cellular cause of the disease. Histori- ure 5-35b). Related individuals (e.g., first or second cousins)
cally, researchers have used whatever phenotypic clues might have ;J relatively high probability of being carriers for the
be relevant to make guesses about the molecular basis of in- same recessive alleles. Thus children born to related parents
herited diseases. An early example of successful guesswork are much more likely than those born to unrelated parents to
was the hypothesis that sickle-cell anemia, known to be a be homozygous for, and therefore affected by, an autosomal
disease of blood cells, might be caused by defective hemoglo- recessive disorder.
bin. This idea led to identification of a specific amino acid The th ird common pattern of inheritance is that of an
substitution in hemoglobin that causes polymerization of the X-lmked recessive allele. A recessive allele on the X chromo-

2 06 CHAPTER 5 • Molecular Genetic Techniques

 Common Inherited Human Diseases

Disease Molecular and Cellular Defect Incidence

Autosomal Recessive

Sickle-cell anemia Abnormal hemoglobin causes deformation of 1/625 of sub-Saharan African origin
red blood cells, w hich can become lodged in
capillaries; also confers resistance to m::~l::~ r ia.

Cystic fibrosis Defective chloride channel (CFTR) in epithelial cells 1/2500 of European origin
leads to excessive mucus in lungs.

Phenylketonuria (PKU) Defective enzyme in phenylalanine metabolism 1/10,000 of European origin

(tyrosine hydroxylase) results in excess phenylalanine
leading to mental retardation, unless restricted by diet.

Tay-Sachs disease Defective hexosaminidase enzyme leads to accumulation 1/1000 eastern European Jews
of excess sphingolipids in the lysosomes of neurons,
·. impairing neural development.

Autosomal Dominant

Huntington's disease Defective neural protein (hunnngtin) may assemble into l/10,000 of European origm
aggregates, causing damage to neural tissue.

Hypercholesterolemia Defective LDL receptor leads to excessive cholesterol in 1/122 French Canadians
blood and early heart attacks.

X-Linked Recessive

Duchenne muscular Defective cytoskeletal protein dystrophin leads to 1/3500 males

dystrophy (DMD) im paired muscle function.

Hemophilia A Defective blood clotting factor VIII leads to 1-2110,000 males

uncontrolled bleeding.

some will most often be expressed in males, who receive only markers using the basic principle of genetic linkage as de-
one X chromosome from their mother, but not in females, scribed in Section 5.1. The presence of many different al-
who receive an X chromosome from both their mother and ready mapped genetic tra its, or markers, distributed along
their father. This leads to a distinctive sex-linked segregation the length of a chromosome facilitates the mapping of a new
pattern where the disease is exhibited much more frequently mutation: one can assess its possible linkage to these marker
in males than in females . •For example, Duchenne muscular genes in appropriate crosses. The more markers that are
dystrophy (DMD), a muscle degenerative disease that spe- available, the more precisely a mutation can be mapped. The
cifically affects males, is caused by a recessive allele on the X density of gen etic markers needed for a high-resolution
chromosome. DMD exhibits the typical sex-linked segrega- human genetic map is about one marker every 5 cemimor-
tion pattern in which mothers who are heterozygous and gans (eM) (as discussed previously, one genetic map unit, or
therefore phenotypically normal can act as carriers, trans- centimorgan, is defined as the distance between two posi-
mitting the DMD allele and therefore the disease to 50 per- tions along a c h romosome that results in one recombinant
cent of their male progeny (Figure 5-35c). individual in 100 progeny). Thus a high-resolution genetic
map requires 25 or so genetic markers of known position
spread along the length of each human chromosome.
DNA Polymorph isms Are Used as Markers
In the experimental organisms commonly used in genetic
for Linkage-Mapping of Human Mutations studies, numerous markers with easily detectable phenotypes
Once the mode of inheritance has been determined, the next are available for genetic mapping of mutations. This is not
step in determining the position of a disease alle le is to the case for mapping genes whose mutant alleles are associated
genetically map its position with respect to known genetic with inherited diseases in humans. However, recombinant

5.4 Locating and Identifying Human Disease Genes 207

 Grandparents Grandpa rents
(a) Autosomal X
dominant:
Huntington's
disease
Males and females
A+ lA+
Not affected

(b) Autosomal cr ACFTR;A+ X ACFTR;A+ S'

recessive: Carrier Carrier
cystic fibrosis [XPERIMENTAL FIGURE 5-36 Single-Nucleotide Polymor-
phisms (SNPs) can be followed like genetic markers. Hypothetical
Males and females pedigree based on SNP analysis of the DNA from a region of a
ACFTR IACFTR ACFTR;A+ A+ IACFTR A+lA+ chromosome. In this family, t he SNP exists as an A, T, or C nucleotide.
Affected Carrier Carrier Noncarrier Each individual has two alleles: some contain an A on both chromo-
somes, and others are heterozygous at t his site. Circles indicate
(c) X-linked recessive: cf x+ /Y X XDMD/X+ S' females; squares indicate males. Blue indicates unaffected individuals;
Duchenne muscular Carrier orange indicates individuals with the tra}t. Analysis reveals that the trait
dystrophy ~ segregates wit h a Cat t he SNP.
Males Females
XOM0 tY x+tY
Affected Unaffected Carrier Noncarrier
number of repetitions of a one-, two-, or three-base sequence.
Such polymorphisms, known as simple-sequence repeats
FIGURE 5- 35 Three common inheritance patterns for human (SSRs) or microsatellites, presumably are formed by recom-
genetic d iseases. Wild-type autosomal (A) and sex chromosomes bination or a slippage mechanism of either the template or
(X andY) are indicated by superscript plus signs. (a) In an autosomal
newly synthesized strands during DNA repl ication. A useful
dominant disorder such as Huntington's disease, only one mutant
property of SSRs is that different individuals will often have
allele is needed to confer the disease. If either parent is heterozygous
for the mutant HD allele, his or her children have a SO percent chance
different numbers of repeats. The existence of multiple ver- ..
of inheriting the mutant allele and getting the disease. (b) In an
sions of an SSR makes it more likely to produce an informa-
autosomal recessive disorder such as cystic fibrosis, two mutant alleles tive segregation pattern in a given pedigree and therefore to
must be present to confer the disease. Both parents must be heterozy- be of more general use in mapping the positions of disease
gous carriers of the mutant CFTR gene for their children to be at risk of genes. These polymorphisms can be detected by PCR ampl i-
being affected or being carriers. (c) An X-linked recessive disease such fication and DNA sequencing.
as Duchenne muscular dystrophy is caused by a recessive mutation on
the X chromosome and exhibits the typical sex-linked segregation
pattern. Males bor~ to mothers heterozygous for a mutant DMD allele
Linkage Studies Can Map Disease Genes
have a SO percent chance of inheriting the mutant allele and being with a Resolution of About 1 Centimorgan
affected. Females born to heterozygous mothers have a SO percent Without going into all the technical considerations, let's see
chance of being carriers.
how the allele conferring a particular dominant trait (e.g.,
familial hypercholesterolemia) might be mapped. The first
step is to obtain DNA samples from all the members of a
DNA technology has made available a wealth of useful family contain ing individuals that exhibit the disease. The
DNA-based molecular markers. Because most of the human DNA from each affected and unaffected individual then is
genome does not code for protein, a large amount of sequence ana lyzed to determine the identity of a large number of
variation exists between individuals. Indeed, it has been esti- known DNA polymorph isms (either SSR or SNP markers can
mated that nucleotide differences between unrelated individu- be used). The segregation pattern of each DNA polymor-
als can he detected on an average of every 10 3 nucleotides. If phism within the fami ly is t hen compared with the segrega-
these variations in DNA sequence, referred to as DNA poly- tion of the d isease under study to find those polymorphisms
morfJiusms, can be followed from one generation to the next, that tend to segregate along with the disease. Finally, com-
they can serve as genetic markers for linkage studies. Cur- puter analysis of the segregation data is used to calculate the
rently, a panel of as many as 104 different known polymor- likelihood of linkage between each DNA polymorphism and
phisms whose locations have been mapped in the human the disease-causing allele.
genome is used for genetic linkage studies in humans. In practice, segregation data are collected from different
Single-nucleotide polymorphisms (SNPs) constitute the families exhibiting the same disease and pooled. The more
most abundant type and are therefore useful for constructing families exhibiting a particular disease that can be examined,
genetic maps of maximum resolution (Figure 5-36). Another the greater the statistical significance of evidence for linkage
useful type of DNA polymorphism consrsts of a variable that can be obtained and the greater the precision with which

208 CHAPTERs • Molecular Genetic Techniques

Generation
1
. II}
New ~utatton
tn parttcular---
haplotype m
Chromosome
with different
haplotype
some conserved through many generations (Figure 5-37). By
assessing the distribution of specific markers in all the affected
individuals in a population, geneticists can identify DNA
markers tightly associated with the disease, thus localizing the
~ Meiotic recombination disease-associated gene to a relatively small region. Under
ideal circumstances, linkage disequilibrium studies can im-
Generation
prove the resolution of mapping studies to less than 0.1 cen-
2
1111 timorgan. The resolving power of this method comes from the
ability to determine whether a polymorphism and the disease
allele were ever separated by a meiotic recombination event at
~ any time since the disease allele first appeared on the ancestral
chromosome-in some cases this can amount to finding mark-
! ers that are so closely linked to the disease gene that even after
hundreds of meioses, they have never been separated by
! recombination.

Generation
10
1111111111111111
FIGURE 5 - 37 Linkage-disequilibrium studies of human popula-
Further Analysis Is Needed to Locate
a Disease Gene in Cloned DNA
tions can be used to map genes at high resolution. A new disease Although linkage mapping can usually locate a human dis-
mutation will arise in the context of an ancestral chromosome among a ease gene to a region containing about 10 5 base pairs, as
set of polymorphisms known as the haplotype (indicated by red shading). many as 10 different genes may be located in a region of this
After many generations, chromosomes that carry the disease mutation size. The ultimate objective of a mapping study is to locate
will also carry segments of the ancestral haplotype that have not been the gene within a cloned segment of DNA and then to deter-
separated from the disease mutation by recombination. The blue mine the nucleotide sequence of this fragment. The relative
segments of these chromosomes represent general haplotypes derived scales of a chromosomal genetic map and physical maps cor-
from the general population and not from the ancestral haplotype in
responding to ordered sets of plasmid clones and the nucleo-
which the mutation originally arose. This phenomenon is known as
tide sequence are shown in Figure 5-38.
linkage disequilibrium. The position of the disease mutation can be
One strategy for further localizing a disease gene within
located by scanning chromosomes containing the disease mutation for
the genome is to identify mRNA encoded by DNA in the re-
highly conserved polymorph isms corresponding to the ancestral
haplotype. gion of the gene under study. Comparison of gene expression
in tissues from normal and affected individuals may suggest
tissues in which a particular disease gene normally is expressed.
l-or instance, a mutation that phenotypically affects muscle,
the distance can be measured between a linked DNA polymor- but no other tissue, might be in a gene that is expressed only in
phism and a disease allele. Most family studies have a maxi- muscle tissue. The expression of mRNA in both normal and
mum of about l 00 individuals in which linkage between a affected individuals generally is determined by Northern blot-
disease gene and a panel of DNA polymorphisms can be tested. ting or in situ hybridization of labeled DNA or RNA to tissue
This number of individuals sets the practical upper limit on the sections. Northern blots, in situ hybridization, or microarray
resolution of such a mapping study to about 1 centimorgan, or experiments permit comparison of both the level of expres-
a physical distance of about 7.5 X 105 base pairs. sion and the size of mRNAs in mutant and wild-type tissues.
A phenomenon called Jinkage disequilibrium is the basis Although the sensitivity of in situ hybridization is lower than
for an alternative strategy, which often can afford a higher that of Northern blot analysis, it can be very helpful in identi-
degree of resolution in mapping studies. This approach de- fying an mRNA that is expressed at low levels in a g1ven tissue
pends on the particular circumstance in which a genetic disease bur at very high levels in a subclass of cells within that tissue.
commonly found in a particular population results from a An mRNA that is altered or missing in various individuals af-
single mutation that occurred many generations in the past. fected with a disease compared with wild-type individuals
The DNA polymorphisms carried by this ancestral chromo- would be an excellent candidate for encoding the protein
some are collectively known as the haplotype of that chromo- whose disrupted function causes that disease.
some. As the disease allele is passed from one generation to the In many cases, point mutations that give rise to disease-
next, only the polymorphism~ that are closest to the disease causing alleles may result in no detectable change in the level
gene will not be separated from it by recombination. After of expression or electrophoretic mobility of mRNAs. Thus if
many generations, the region that contains the disease gene comparison of the mRNAs expressed in normal and affected
will be evident because this will be the only region of the chro- individuals reveals no detectable differences 111 the candidate
mosome that will carry the haplotype of the ancestral chromo- mRNAs, a search for point mutations in the DNA regions

5.4 Locating and Identifying Human Disease Genes 209

 Polymorphic

I markers -
~~/clones
Plasmid or BAC
G

II ~
II ~
,, i
I ~
750kb ,, ~
1
1 I.
II ~
II ~
II ~
I ¥
LEVEL OF
RESOLUTION: Cytogenetic Linkage Physical Sequence ·.
map map map map
METHOD OF
DETECTION: Chromosome Linkage to single- Hybridization DNA
banding pattern nucleotide poly- to plasmid sequencing
Fluorescence in m orphisms SNPs and clones
situ hybridization simple sequence
(FISH) repeats SSRs

FIGURE 5-38 The relationship between the genetic and physical mapped relative to DNA-based genetic markers. Local segments of the
maps of a humart chromosome. The diagram depicts a human chromosome can be analyzed at the level of DNA sequences identified
chromosome ana lyzed at different levels of detail. The chromosome as by Southern hybridization or PCR. Finally, important genetic differ-
a whole can be viewed in the light microscope when it is in a con- ences can be most precisely defined by differences in the nucleotide
densed state that occurs at metaphase, and the approximate location sequence of the chromosomal DNA. [Adapted from L. Hartwell et al., 2003,
of specific sequences can be determined by fluorescence in situ Genetics: From Genes to Genomes, 2d ed., McGraw-Hill.]
hybridization (FISH). At the next level of detail, genetic traits can be

encoding the mRNAs is undertaken. Now that highly efficient

Many Inherited Diseases Result from Multiple
methods for sequencing DNA are available, researchers fre-
quently determine the sequence of candidate regions of DNA Genetic Defects
isolated from affected individuals to identify point mutations. Most of the inherited human diseases that are now under-
The overall strategy is to search for a coding sequence that stood at the molecular level are monogenetic diseases; that
consistently shows possibly deleterious alterations in DNA is, a clearly discernible disease state is produced by a defect
from individuals that exhibit the disease. A limitation of this in a single gene. Monogenic diseases caused by mutation in
approach is that the region near the affected gene may carry one specific gene exhibit one of the characteristic inheritance
naturally occurring polymorphisms unrelated to the gene of patterns shown in Figure 5-35. T he genes associated with
interest. Such polymorphisms, not functionall y related to the most of the common monogenic diseases have already been
disease, can lead to misidentification of the DNA fragment mapped using DNA-based markers as described previously.
carrying the gene of interest. For this reason, the more mu- However, many other inherited diseases show more com-
tant alleles a\·ailable for analysis, the more like!) that a gene pl icated patterns of inheritance, making the identification of
will be correctly identified. the underlying genetic cause much more difficult. One type of

210 CHAPTERs • Molecular Genetic Techniques

 added complexity that is frequently encountered is genetic GWAS. Genomic sequencing and other methods can then be
heterogeneity. In such cases, mutatiom. in any one of multiple used to identify possible disease causing mutations in these re-
different genes can cause the same disease. For example, reti- gions. Although GW AS can be a powerful tool to identify can-
nitis pigmentosa, which is characterized by degeneration of didate disease genes, much further work is needed to determine
the retina usually leading to blindness, can be caused by muta- how an individual carrying a particular mutation might be pre-
tions in any one of more than 60 different genes. In human disposed to the disease.
linkage studies, data from multiple families usually must be Models of human disease in experimental organisms may
combined to determine whether a statistically significant link- also contribute to unraveling the genetics of complex traits
age exists between a disease gene and known molecular mark- such as obesity or diahetP'>. For instance, large-scale con
ers. Genetic heterogeneity such as that exhibited by retinitis trolled breeding experiments in mice can identify mouse
pigmcntosa can confound such an approach because any sta- genes associated with diseases analogous to those in humans.
tistical trend in the mapping data from one family tends to be The human orthologs of the mouse genes identified in such
canceled out by the data obtained from another family with studies would be likely candidates for involvement in the
an unrelated causative gene. corresponding human disease. DNA from human popula-
Human geneticists used two different approaches to tions then could be examined to determine if particular al-
identify the many genes associated with retinitis pigmentosa. leles of the candidate genes show a tendency to be present in
The first approach relied on mapping studies in exception- individuals affected with the disease but absent from unaf-
ally large single families that contained a sufficient number fected individuals. This "candidate gene" approach is cur-
of affected individuals to provide statistically significant evi- rently being used intensively to search for genes that may
dence for linkage between known DNA polymorphisms and contribute to the major polygenic diseases in humans.
a single causative gene. The genes identified in such studies
showed that several of the mutations that cause retinitis pig-
mentosa lie within genes that encode abundant proteins of
KEY CONCEPTS of Section 5.4
the retina. Following up on this clue, geneticists concentrated
their attention on those genes that are highly expressed in the Locating and Identifying Human Disease Genes
retina when screening other individuals with retinitis pig-
Inherited diseases and other traits in humans show three
mentosa. This approach of using additional information to
major patterns of inheritance: autosomal dominant, autoso-
focus screening efforts on a subset of candidate genes led to
mal recessive, and X-linked recessive (see Figure 5-35).
identification of additional rare causative mutations in many
different genes encoding retinal proteins. Genes for human diseases and other traits can be mapped by
A further complication in the genetic dissection of human determining their co-segregation during meiosis with markers
diseases is posed by diabetes, heart disease, obesity, predis- whose locations in the genome are known. The closer a gene is
position to cancer, and a variety of mental disorders that to a particular marker, the more likely they are to co-segregate.
have at least some heritable properties. These and many Mapping of human genes with great precision requires
other diseases can be considered to be polygenic diseases in thousands of molecular markers distributed along the chromo-
the sense that alleles of multiple genes, acting together within somes. The most useful markers are differences in the DNA
.. an individual, contribute to both the occurrence and the se- sequence (polymorphisms) between individuals in noncoding
verity of disease. How to systematically map complex poly- regions of the genome.
genic traits in humans is one of the most important and
• DNA polymorphisms useful in mapping human genes in-
challenging problems in human genetics today.
clude single-nucleotide polymorphisms (SNPs ) and simple
One of the most promising methods to study diseases that
sequence-repeats (SSRs).
exhibit genetic heterogeneity or are polygenic is to seek a sta-
tistical correlation betweep inheritance of a particular region • Linkage mapping often can locate a human disease gene to
of a chromosome and the propensity to have a disease using a a chromosomal region that includes as many as 10 genes. To
procedure known as a genome-wide association study (GWAS). identify the gene of interest within this candidate region typi-
The identification of disease-causing genes by GWAS relics on cally requires expression analysis and comparison of DNA
the phenomenon of linkage disequilibrium described previ- sequences between wild-type and disease-affected individuals.
ously. If an allele that causes or even predisposes an individual • Some inherited diseases can result from mutations in dif-
to a disease has occurred relatively recently during human evo- ferent genes in different individuals (genetic heterogeneity).
lution, the disease-causing allele will tend to remain associated The occurrence and severity of other diseases depend on the
with the particular set of DNA-based markers in the neighbor- presence of mutant alleles of multiple genes in the same indi-
hood of its chromu~omallocation. By examining a large num- viduals (polygenic traits). Mapping of the genes associated
ber of DNA markers in populations of individuals with a with such diseases can be achieved by finding a statistical
particular disease as well as in control populations of individu- correlation between the disease and a particular chromo-
als without the disease, chromosomal regions that tend to be somal location in a genome-wide association study.
correlated with occurrence of the disease can be identified with

5.4 Locating and Identifying Human Disease Genes 211

5.5 Inactivating the Function of Specific (a) 20-nt flanking 20-nt flanking
sequence sequence

Genes in Eukaryotes
The elucidation of DNA and protein sequences in recent
years has led to identification of many genes, using sequence o--c::J
DNA'' •tl '" Primer 2
patterns in genomic DNA and the sequence similarity of the
encoded proteins with proteins of known function. As dis- r=::;==lc:1k~an~M~1X~:]:1=== ~:
Primer 1
cussed in Chapter 6, the general functions of proteins identi-
c::J-ooCl
fied by sequence searches may be predicted by analogy with
known proteins. However, the precise in vivo roles of such 1 PCR

''new" proteins may be unclear in the absence of mutant ~ Disruption

forms of the corresponding genes. In this section, we describe ~construct
several ways for disrupting the normal function of a specific
gene in the genome of an organism. Analysis of the resulting
mutant phenotype often helps reveal the in vivo function of
the normal gene and its encoded protein.
Three basic approaches underlie these gene-inactivation
techniques: (1) replacing a normal gene with other sequences,
(b)

Diploid
cell
e· .
=• It"'$!•=
=
,

(2) introducing an allele whose encoded protein inhibits func- o --._j Transform diploid cells with
~ disruption construct
tioning of the expressed normal protein, and (3) promoting

~Homologous
destructiOn of the mRNA expressed from a gene. The normal
endogenous gene is modified in techniques based on the first
approach but is not modified in the other approaches. ~ recombi~ation

Normal Yeast Genes Can Be Replaced

l Select for G-418 resistance

with Mutant Alleles by Homologous

Recombination
Modifying the genome of the yeast S. cerevisiae is particu-
larly easy for two reasons: yeast cells readily take up exoge-
nous DNA under certain conditions, and the introduced
9
If the disrupted
DNA is efficiently exchanged for the homologous chromo- gene is essential,
somal site in the recipient cell. This specific, targeted recom- these spores
bination of identical stretches of DNA allows any gene in
Four
haploid
f will be nonviable

yeast chromosomes to be replaced with a mutant allele. (As spores

we discuss in Section 5.1, recombination between homolo-
gous chromosomes also occurs naturally during meiosis.)
In one popular method for disrupting yeast genes in this
fashion, the PCR is used to generate a disruption construct EXPERIMENTA FIGURE 5-39 Homologous recombination
containing a selectable marker that subsequently is transfected with transfected disruption constructs can inactivate specific
into yeast cells. As shown in Figure 5-39a, primers for PCR target ge nes in yeast. (a) A suitable construct for disrupting a
target gene can be prepared by the PCR. The two primers designed
amplification of the selectable marker are designed to include
for this purpose each contain a sequence of about 20 nucleotides
about 20 nucleotides identical with sequences flanking the
(nt) that is homologous to one end of the target yeast gene as well as
yeast gene to be replaced. The resulting amplified construct
sequences needed to amplify a segment of DNA carrying a select-
comprises the selectable marker (e.g., the kanMX gene, which able marker gene such as kanMX, which confers resistance to G-4 18.
like neo' confers resistance to G-418) flanked by about 20 (b) When recipient diploid Saccharomyces cells are transformed with
base pairs that match the ends of the target yeast gene. Trans- the gene-disruption construct, homologous recombination between
formed diploid yeast cells in which one of the two copies of the ends of the construct and the corresponding chromosomal
the target endogenous gene has been replaced by the disrup- sequences will integrate the kanMX gene into the chromosome,
tion construct are identified by their resistance to G-418 or replacing the target gene sequence. The recombinant diploid cPII~
other selectable phenotype. These heterozygous diploid yeast will grow on a medium containing G-4 18, whereas nontransformed
cells generally grow normally regardless of the function of the cells will not. If the target gene is essential for viability, half the
target gene, but half the haploid spores derived from these haploid spores that form after sporulation of recombinant diploid
cells will carry only the disrupted allele (Figure 5-39b). If a cells will be nonviable.
gene is essential for viability, then spores carrying a disrupted
allele will not survive.

2 12 CHAPTERs • Molecular Genetic Techniques .·

 Disruption of yeast genes by this method is proving par- study provided the first evidence for the unexpected role of
ticularly useful in assessing the role of proteins identified by Hsc70 protein in translocation of secretory proteins into the
analysis of the entire genomic DNA sequence (see Chapter 6). ER, a process examined in detail in Chapter 13.
A large consortium of scientists has replaced each of the ap-
proximately 6000 genes identified by this analysis with the
kanMX disruption construct and determined which gene dis-
Specific Genes Can Be Permanently Inactivated
ruptions lead to nonviable haploid spores. These analyses have
shown that about 4500 of the 6000 yeast genes are notre- in the Germ line of Mice
quired for viability, an unexpectedly lnrge number of appar- ~1any of the methods for disrupting genes 10 yeast can be
ently nonessential genes. In some cases, disruption of a applied to genes of higher eukaryotes. These altered genes
particular gene may give rise to subtle defects that do not com- can be introduced into the germ line via homologous recom-
promise the viability of yeast cells growing under laboratory bination to produce animals with a gene knockout, or sim-
conditions. Alternatively, cells carrying a disrupted gene rna) ply "knockout." Knockout mice in which a specific gene 1s
be viable because of operation of backup or compensatory disrupted are a powerful experimental system for studying
pathways. To investigate this possibility, yeast geneticists mammalian development, behavior, and physiology. They
currently are searching for synthetic lethal mutations that also arc useful in studying the molecular basis of certain
might reveal nonessential genes with redundant functions human genetic diseases.
(see Figure 5-9c). Gene-targeted knockout mice are generated by a two-
stage procedure. In the first stage, a DNA construct con tam-
ing a disrupted allele of a particular target gene is introduced
Transcription of Genes Ligated to a Regulated
into embryonic stem (ES) cells. These cells, which are de-
Promoter Can Be Controlled Experimentally rived from the blastocyst, can be grown in culture through
Although disruption of an essential gene required for cell many generations (see Figure 21- ~). In a small fraction of
growth will yield nonviable spores, this method provides lit- transfected cells, the introduced DNA undergoes homolo-
tle information about what the encoded protein actually does gous recomhination with the target gene, although recombi-
in cells. To learn more about how a specific gene contributes nation at nonhomologous chromosomal sires occurs much
to cell growth and viability, investigators must be able to se- more frequently. To select for cells in which homologous
lectively inactivate the gene in a population of growing cells. gene-targeted insertion occurs, the recombinant DNA con-
One method for doing this employs a regulated promoter to struct introduced into ES cells needs to include two select-
selectively shut off transcription of an essential gene. able marker genes (Figure 5-40). One of these genes (neo'),
A useful promoter for this purpose is the yeast GALl which confers G-418 resistance, is inserted within the target
promoter, which is active in cells grown on galactose but gene (X), thereby disrupting it. The other selectable gene, the
completely inactive in cells grown on glucose. In this ap- thymidine kinase gene from herpes simplex virus (tkHS 1 ), is
proach, the coding sequence of ~n essential gene (X) ligated inserted into the construct outside the target-gene sequence.
to the GALl promoter is inserted into a yeast shuttle vector ES cells that undergo recombmation between the DNA con-
(see figure 5-17a). The recombinant vector then is introduced struct and the homologous site on the chromosome will con-
into haploid yeast cells in which gene X has been disrupted. tain ne(/ but will not incorporate tkHS\'. Because tkHsv
Haploid cells that arc transformed will grow on galactose confers sensitivity to the cytotoxic nucleotide analog ganci-
medium since the normal copy of gene X on the vector is clovir, the desired recombinant ES cells can be selected by
expressed in the presence of galactose. When the cells are their ability to survive in the presence of both G-418 and
transferred to a glucose-containing medium, gene X no lon- ganciclovir. In these cells, one allele of gene X will be
ger is transcribed; as the cells divide, the amount of the en- disrupted.
coded protein X will decline, eventually reaching a state of In the second stage in production of knockout mice, ES
depletion that mimics a complete loss-of-function mutation. cells heterol.}'gous for a knockout mutation in gene X are
The observed changes in the phenotype of these cells after the injected into a recipient wild-type mouse blastocyst, which
~hift to glucose medium may suggest which cell processes de- subsequently is transferred into a surrogate pseudopregnant
pend on the protein encoded by the essential gene X. female mouse (figure 5-41 ). The resulting progeny will be
In an early application of this method, researchers ex- chimeras, containing tissues derived from both the trans-
plored the function of cytosolic Hsc70 genes in yeast. Hap- planted ES cells and the host cells. If the ES cell5 also are
loid cells with a disruption in all four redundant Hsc70 genes homozygous for a visible marker trait (e.g., coat color), then
were nonviable unless the cells carried a vector containing a chimeric progeny in which the ES cells survived and prolifer-
copy of the Hsc70 gene that could he expressed from the ated can be identified easily. Chimeric mice are then mated
GAL 1 promoter on galactose medium. On transfer to glu- with mice homozygous for another allele of the marker trait
cose, the vector-carrying cells eventually stopped growing to determine if the knockout mutation is incorporated into the
because of insufficient Hsc70 activity. Careful examination germ line. Finally, mating of mice, each heterozygous for the
of these dying cells revealed that their secretory proteins knockout allele, will produce progeny homozygous for
could no longer enter the endoplasmic reticulum (ER). This the knockout mutation.

5.5 Inactivating the Function of Specific Genes in Eukaryotes 213

E)(PERIM :NTAL FIGURE 5-40 Isolation of mouse ES cells with (a) Formation of ES cells carrying a knockout mutation
a gene-targeted di sruption is the first stage in production of neo' tkHSV
knockout mice. (a) When exogenous DNA is introduced into embry-
onic stem (ES) cells, random insertion via nonhomologous recombina- ~
Gene X replacement construct
tion occurs much more frequently than gene-targeted insertion via
homologous recombination. Recombinant cells in which one allele of
gene X (orange and white) is disrupted can be obtained by using a
Homologo~~n
recombina/
..?- ES
cells
ES
cells
~ ~~~homologous
~ombination
recombinant vector that carries gene X disrupted with neo' (green),
which confers resistance to G-418, and, outside the region of homol-
ogy, t0sv (yellow), the thymidine kinase gene from herpes simplex ==={IJJ==O=
X X
virus. The viral thymidine kinase, unlike the endogenous mouse
enzyme, can convert the nucleotide analog ganciclovir into the
====ti:II==
'---v---' ES·cell DNA ~DNA
monophosphate form; this is then modified to the triphosphate form, Gene X Other genes
which inhibits cellular DNA replicat ion in ES cells. Thus ganciclovir is
~ene-.targeted
cytotoxic for recombinant ES cells carrying the t05v gene. Nonhomolo-
gous insertion includes the t05v gene, whereas homologous insertion
does not; therefore, only cells with nonhomologous insertion are
sensitive to ganciclovir. (b) Recombinant cells are selected by treat-
l msertton
l Random
insertion

ment with G-418 since cells that fail to pick up DNA or integrate it into Mutation in gene X No mutation in gene X
their genome are sensitive to this cytotoxic compound. The surviving Cells are resistant to Cells are resistant to
recombinant cells are treated with ganciclovir. On ly cells with a G-418 and ganciclovir G-418 but sensitive
targeted disruption in gene X, and therefore lacking t he t05v gene and to ganciclovir
its accompanying cytotoxicity, will survive. (Sees. L. Mansour et al., 1988,
Nature 336:348.]
(b) Positive and negative selection of recombinant ES cells

0 O 0 Q) 0- Nonrecombinant cells
Development of knockout mice that mimic certain Recombinants -(!) 0
with random 0 O 0 0- Recombinants with
human diseases can be illustrated by cystic fibrosis. By insertion Q) gene-targeted insertion
methods discussed in Section 5 .4, the recessive mutation that Q)O
ca uses this disease eventually was shown to be located in a

l
Treat with G-418
gene known as CFTR, which encodes a chloride channel. (positive selection)
Using the cloned wild-type human CFTR gene, researchers
isolated the homologous mouse gene and subsequently intro-
duced mutations in it. The gene-knockout technique was
then used ro produce homozygous mutant mice, which
showed symptoms (i.e., a phenotype), including disturbances
to the functioning of epithelial cells, similar to those of hu-
mans with cystic fibrosis. These knockout mice are currently
being used as a model system for studying this genetic dis-
ease and developing effective therapies. •
0
l Treat with ganciclovir
(negative selection)

0
0 0 0
0 0
Somatic Cell Recombination Can Inactivate
Genes in Specific Tissues ES cells with targeted disruption in gene X

Investigators often are interested in examining the effects of

knockout mutations in a particular tissue of the mouse, at a
specific stage in development, or both. H owever, mice carry- cells. An essential feature of th is techniq ue is that expression
ing a germ-line knockout may have defects in numerous tis- of Cre is controlled by a cell-type-specific promoter. In loxP-
sues or die before the developmental stage of interest. To Cre mice generated by the procedure depicted in Figure 5-42,
address this problem, mouse geneticists have devised a clever inactivation of the gene of interest (X) occurs only in cells in
technique to inactivate target genes in specific type<; of <;o- which rhe promoter controlling the ere gene is active.
matic cells or at particular times during development. An early application of t his technique provided strong
This technique employs site-specific DNA recombination evidence t hat a particular neurotransmitter receptor is im-
sites (called loxP sites) and the enzyme Cre that catalyzes re- portant for learning and memory. Previous pharmacological
combination between them. The loxP-Cre recombination sys- and physiological studies had indicated that normal learning
tem is derived from bacteriophage Pl, but this site-specific requires the NMDA class of glutamate receptors in t he hip-
recombination system also functions when placed in mouse pocampus, a region of the brain. Bur mice in which the gene

214 CHA PTER 5 • Molecular Genetic Techniques

~ VIDEO: Microinjection of ES Cells into a Blastocyst

EXPERIMENTAL FIGURE 5·41 ES cells heterozygous for a Inject ES cells into blastocoel
disrupted gene are used to produce gene-targeted knockout mice. D cavity of early embryos
Step 0 : Embryonic stem (ES) cells heterozygous for a knockout Brown mouse
mutation in a gene of interest (X) and homozygous for a dominant (AlA, x-tx+>
allele of a marker gene (here brown coat color, A) are transplanted into Black mouse
the blastocoel cavity of 4.5-day embryos that are homozygous for a (ala, X+/X+)
recessive allele of the marker (here black coat color, a). Step H: The 4.5-day blastocyst
early embryos then are implanL~d into a pseudopregnant female.
Those progeny containing ES-derived cells are chimeras, indicated by o lsurgically transfer embryos
into pseudopregnant female
their mixed black and brown coats. Step II: Chimeric mice then are
backcrossed to black mice; brown progeny from this mating have
ES-derived cells in their germ line. Steps [1- l'it Analysis of DNA isolated ea
from a small amount of tail tissue can identify brown mice heterozy-
gous for the knockout allele.lntercrossing of these mice produces Foster mother
some individuals homozygous for the disrupted allele, that is, knockout
mice. [Adapted from M. R. Capecchi, 1989, Trends Genet. 5:70.] Possible
progeny

• 1•
l

Chimeric Black
encoding an NMDA receptor subunit was knocked out died Select chimeric mice for
neonatally, precluding analysis of the receptor's role in crosses to wild-type black mice
learning. Following the protocol in Figure 5-42, researchers
generated mice in which the receptor subunit gene was inac-
tivated in the hippocampus but expressed in other tissues .
These mice survived to adulthood and showed learning and
memory defects, confirming a role for these receptors in the Possible germ cells: All germ cells:
A/X+; A/ X-; a/X+ a/X+
ability of mice to encode their experiences into memory.
111 ES-cell-derived progeny
will be brown
Dominant-Negative Alleles Can Functionally
Inhibit Some Genes
In diploid organisms, as noted in Section 5.1, the phenotypic
Ala, x+;x+ Ala, X ;x+ ala, x +;x +
effect of a recessive allele is expressed only in homozygous
individuals, whereas dominant alleles are expressed in het- Progeny from ES-cell-derived
erozygotes. Thus an individual must carry two copies of a germ cells
recessive allele but only one copy of a dominant allele to ex-
hibit the corresponding phenotypes. We have seen how strains
a! Screen brown progeny DNA
to identify X IX+ heterozygotes
of mice that are homozygous for a given recessive knockout
mutation can be produced by crossing individuals that are D! Mate X I X+ heterozygotes
heterozygous for the same knockout mutation (see Figure
5-41). For experiments with cultured animal cells, however, it D! Screen progeny DNA to identify
X ;x- homozygotes
is usually difficult to disrupt both copies of a gene in order to
produce a mutant phenotype. Moreover, the difficulty in pro-
ducing strains with both copies of a gene mutated is often
compounded by the presence of related genes of similar func- Knockout mouse
tion that must also be inactivated in order to reveal an observ-
able phenotype.
For certain genes, the difficulties in producing homozy-
gous knockout mutants can be avoided by use of an allele
carrying a dominant-negative mutation. These alleles are ge- Useful dominant-negative alleles have been identified for
netically dominant; that is, they produce a mutant phenotype a variety of genes and can be introduced into cultured cells
even in cells carrying a wild-type copy of the gene. However, by transfection or into the germ line of mice or other organ-
un like other types of dominant alleles, dominant-negative isms. In both cases, the introduced gene is integrated into
alleles produce a phenotype equivalent to that of a loss-of- the genome by nonhomologous recombination. Such ran -
function mutation. domly inserted genes are called transgenes; the cells or

5.5 Inactivating the Function of Specific Genes tn Eukaryotes 215

 loxP Cre
mouse mouse
X
All cells carry endogenous gene Heterozygous for gene X knock-
X with loxP sites flanki ng exon 2 out; all cells carry ere gene

Cell-type-specific
loxP loxP
promoter

Cells not expressing Cre Cells expressing Cre

~
Gene function is normal

loxP-Cre
mouse
All cells carry one copy of loxP-
modified gene X, one copy of
gene X knockout, and ere genes
Gene funct ion is disrupted

EXPERIMENTAl FIGURE 5-42 The loxP-Cre recombination function of other genes. In the loxP-Cre mice that result from crossing,
system can knock out genes in specific cell types. A /oxP site (purple) Cre protein is produced only in those cells in which.the promoter is
is inserted on each side of the essential exon 2 of the target gene X active. Thus these are the only cells in which recombinat ion between
(blue) by homologous recombination, producing a loxP mouse. Since the /oxP sites catalyzed by Cre occurs, leading to deletion of exon 2. Since
the /oxP sites are in introns, they do not disrupt the function of X. The the other allele is a constitutive gene X knockout, deletion between the
Cre mouse carries one gene X knockout allele and an introduced ere /oxP sites results in complete loss of function of gene X in all cells
gene (orange) from bacteriophage Pl linked to a cell-type-specific expressing Cre. By using different promoters, researchers can study the
promoter (yellow). The ere gene is incorporated into the mouse effects of knocking out gene X in various types of cells.
genome by nonhomologous recombination and does not affect the

organisms carrY.ing them are referred to as transgenic. Trans-

RNA Interference Causes Gene Inactivation
genes carrying a dominant-negative allele usually are engi-
neered so that the allele is controlled by a regulated by Destroying the Corresponding mRNA
promoter, allowing expression of the mutant protein in dif- The phenomenon known as RNA interference (RNAi) is per-
ferent tissues at different times. As noted above, the random haps the most straightforward method to inhibit the func-
mtegration of exogenous DNA via nonhomol ogou~ recom- tion of specific genes. This approach is technically simpler
bination occurs at a much higher frequency than insertion than the methods described above for disrupting genes. First
via homologous recombination. Because of this phenome- observed in the roundworm C. elegans, RNAi refers to the
non, the production of transgenic mice is an efficient and ability of double-stranded RNA to block expression of its
straight forward process (Figure 5-43 ). corresponding single-stra nded mRNA but not that of
Among the genes that can be functionally inactivated b} mRNAs with a different sequence.
introduction of a dominant-negative allele are those encod- As described in Chapter 8, the phenomenon of RNAi
ing small (monomeric) GTP-binding proteins belonging to rests on the general ability of eukaryotic cel ls to cleave double-
the GTPase superfamily. As we will examine in several later stranded RNA into short (23-nt) double-stranded segments
chapters, these proteins (e.g., Ras, Rae, and Rab) act as in- known as small inhibitory Rr-\A (siRNA ). The RNA endo-
tracellular switches. Conversion of the small GTPases from nuclease that catalyzes this reaction, known as Dicer, is
an inactive GDP-bound state to an active GTP-bound state found in all metazoans but not in simpler eukaryotes such as
depends on their inrer:Jcting with a corresponding guanine yeast. The siRNA molecules, in turn, can caul.e dea vage of
nucleotide exchange factor (GEF). A mutant small GTPase mRNA molecules of matching sequence, in a reaction cata-
that permanently binds to the GEF protein will block con- lyzed by a protein complex known as RJSC. RISC mediates
version of endogenous wild-type small GTPases to the active recognition and hybridization between one strand of the
GTP-bound state, thereby inhibiting them from performing siRNA and its complementary sequence on the target
their switching function (Figure S-44 ). mRNA; subsequently, specific nucleases in the RISC complex
·.

216 CHAPTERs • Molecular Genetic Techniques

 (i) TECHNIQUE ANIMATION: Creating a Transgenic Mouse
VIDEO ANIMATION: DNA Injected into a Pronucleus of a Mouse Zygote
EXPERIMENTAL FIGURE 5 43 Transgenic mice are produced Inject foreign DNA
by random integration of a foreign gene into the mouse germ line. into one of the pronuclei
·· ' Foreign DNA injected into one of the two pronuclei (the male and
female haploid nuclei contributed by the parents) has a good chance
of being randomly integrated into the chromosomes of the diploid Pronuclei
zygote. Because a transgene is integrated into the recipient genome
by nonhomologous recombination, it does not disrupt endogenous
genes. [SeeR. L. Brinster et al., 1981, Ce// 27:223.] Fertilized mouse egg prior
to fusion of male and
female pronuclei

Transfer injected eggs

cleave the mRNA-siRNA hybrid. This model accounts for
the specificity of RNAi, since it depends on base pairing, and
for its potency in silencing gene function, since the comple-
1 into foster mother

mentary mRNA is permanently destroyed by nucleolytic

degradation. The normal function of both Dicer and RISC is
to allow fo r gene regulation by small endogenous RNA mol-
ecules known as micro-RNAs, or miRNAs.
Researchers exploit the micro-RNA pathway for inten-
tional silencing of a gene of interest by using either of two
1
general methods for generating siRNAs of defined sequence.
In the first method, a double-stranded RNA corresponding to
the target-gene sequence is produced by in vitro transcription About 10- 30% of offspring will contain
foreign DNA in chromosomes of
of both sense and antisense copies of this sequence (Figure 5-45a). all their tissues and germ line

Breed mice expressing

(a) Cells expressing only

wild-type alleles of a
small GTPase
1
@
foreign DNA to propagate
DNA in germ line

(3oP
Wild type
--
(b) Cells expressing both
This dsRNA is injected into the gonad of an adult worm,
wild-type alleles and a
dominant-negative allele where it is converted to si RNA by Dicer in the developing
embryos. In conjunction with the RISC complex, the siRNA
molecules cause the corresponding mRNA molecules to be de-
Dominant-negative stroyed rapidly. The resulting worms display a phenotype
mutant similar to the one that would result from disruption of the cor-
responding gene itself. In some cases, entry of just a few mol-
FIGURE 5-44 Inactivation of the function of a wild-type GTPase ecules of a particular dsRNA into a cell is sufficient to inactivate
by the action of a dominant-negative mutant allele. (a) Small
many copies of the corresponding mRNA. Figure 5-45b illus-
(monomeric) GTPases (purple) are activated by their interaction with
trates the ability of an injected dsRNA to interfere with pro-
a guanine nucleotide exchange factor (GEF), which catalyzes the
duction of the corresponding endogenous mRNA in C. elegans
exchange of GOP for GTP. (b) Introduction of a dominant-negative
allele of a small GTPase gene into cultured cells or transgenic animals
embryos. In this experiment, the mRNA levels in embryos
leads to expre~~ion of a mutant GTPase that binds to and inactivates
were determined by in situ hybridization, as describeJ earlier,
the GEF. As a result, endogenous wild-type copies of the same small using a fluorescently labeled probe.
GTPase are trapped in the inactive GOP-bound state. A single The second method is to produce a specific double-
dominant-negative allele thus causes a loss-of-function phenotype stranded RNA in vivo. An efficient way to do this is to express
in heterozygotes similar to that seen in homozygotes carrying two a synthetic gene that is designed to contain tandem segments
recessive loss-of-function alleles. of both sense and antisense seq uences corresponding to the

5.5 Inactivating the Function of Specific Genes in Eukaryotes 217

@ PODCAST: RNA Interference

EXPERIMENTAL FIGURE S-45 RNA interference (RNAi) can (a) In vitro production of double-stranded RNA
functionally inactivate genes in C. efegans and other organisms. (a) In
vitro production of double-stranded RNA (dsRNA) for RNAi of a specific
target gene. The coding sequence of the gene, derived from either a
eDNA clone or a segment of genomic DNA, is placed in two orientations
in a plasm id vector adjacent to a strong promoter. Transcription of both
constructs in vitro using RNA polymera<;P ;md ribonucleoside triphos-
phates yields many RNA copies in the sense o rientation (identical with
the mRNA sequence) or complementary antisense orientation. Under
suitable conditions, these complementary RNA molecules w ill hybridize
to form dsRNA. When the dsRNA is injected into cells, it is cleaved by
Dicer into siRNAs. (b) Inhibition of mex3 RNA expression in worm
embryos by RNAi (see the text for the mechanism). (Left) Expression of
(b )
mex3 RNA in embryos was assayed by in situ hybridization with a probe
specific for this mRNA, that is, linked to an enzyme t hat produces a
colored (purple) product. (Right) The embryo derived from a worm
injected with double-stranded mex3 mRNA produces little or no
endogenous mex3 mRNA, as indicated by the absence of color. Each
four-cell-stage embryo is '--SO fl.m in len gth. (c) In vivo production of
double-stranded RNA occurs via an engineered plasmid introduced
directly into cells. The synthetic gene construct is a tandem arrangement
of both sense and antisense sequences of the target gene. When it is Non injected Injected
transcribed, double-stranded small hairp in RNA forms (shRNA). The
shRNA is cleaved by Dicer to form siRNA. [Part (b) from A. Fire et al., 1998, (c) In vivo production of double-stranded RNA
Nature 39 1:806.)
Sense transcript

target gene (figure 5-45c). When this gene is transcribed, a

double-stranded RNA "hairpin" structure forms, known as
small hairpin RNA, or shRNA. The shRNA wi ll then be
cleaved by Dicer to form siRNA molecules. The lentiviral ex-
pression vectors are particularly useful for introducing syn-
thetic genes for the expression of shRNA const ructs into
animal cells.
.
siRNA
~

Both RNAi methods lend themselves to systematic stud-

ies to inactivate each of the known genes in an organism and
to observe what goes wrong. For example, in initial studies
with C. elegans, R A interference with 16,700 genes (about
86 percent of the genome) yielded 1722 vis ibly abnormal I<FY CONCEPTS of Section 5.5
phenotypes. The genes whose functional inactivation causes Inactivating the Function of Specific Genes
particular abnormal phenotypes can be grouped into sets; in Eukaryotes
each member of a set presumabl) controls the same signals
Once a gene has been cloned, important clues about its
or events. The regulatory relations between the genes in the
norma l function in vivo can be deduced from the observed
set-for example, the genes that control m uscle development-
phenotypic effects of mutating the gene.
can then be worked out.
Other organisms in which RNAi -mediated gene inactiva- Genes can be disrupted in yeast by inserting a selectable
tion has been successful include Drosophila, many kinds of marker gene into one allele of a wild-type gene via homologous
plants, zebra fish, the frog Xenopus, and mice and are now recombination, producing a heterozygous mutant. When such
the subjects of large-scale RNAi screens. For example, lenti- a heterozygote is sporulated, disruption of an essential gene will
viral vectors have been designed to inactivate by RNAi more produce nvo nonviable haploid spores (see Figure 5-39).
than 10,000 different genes expressed in cultured mamma- A yeast gene can be inactivated in a controlled manner by
lian cells. The function of the inactivated genes can be in- us ing the GALl promoter to shut off transcription of a gene
ferred from defects in growth or morphology of cell clones when cells are transferred to gl ucose medium.
transfected with lentiviral vectors.

218 CHAPTER 5 • Molecular Genetic Techniques

 Already sets of vectors for RNAi inactivation of most de-
• In mice, modified genes can be incorporated into the germ fined genes in the nematode C. elegans allow efficient genetic
line at their original genomic location by homologous re- screens to be performed in this multicellular organism. These
combination, producing knockouts (see Figures 5-40 and methods are now being applied to large collections of genes
5-41 ). Mouse knockouts can provide models for human ge- in cultured mammalian cells, and in the near future, either
netic diseases such as cystic fibrosis. R!':Ai or knockout methods wi ll have been used to inacti-
• The loxP-Cre recombination system permits production vate every gene in the mouse.
of mice in which a gene is knocked out in a specific tissue. In the past, a scientist might spend many years studying
only a single gene, but nowadays s~.oicnti!>ts commonly study
• In the prouucrion of transgenic cells or organisms, exoge-
whole sets of genes at once. For example, with DNA microar-
nous DNA is integrated into the host genome by nonhomol-
rays the level of expression of all genes in an organism can be
ogous recombination (see Figure 5-43). Introduction of a
measured almost as easily as the expression of a single gene.
dominant-negative allele in this way can functionally inacti-
One of the great challenges facing geneticists in the twenty-
vate a gene without altering its sequence.
I first century will be to exploit the vast amount of available
1,
• In many organisms, including the roundworm C. elegans, data on the function and regulation of individual genes to
double-stranded RNA triggers destruction of the all the understand how groups of genes are organized to form com-
mRNA molecules with the same sequence (see Figure 5-45). plex biochemical pathways and regulatory networks.
This phenomenon, known as RNAi (RNA interference),
provides a specific and potent means of functionally inacti-
vating genes without altering their structure.
Key Terms
alleles 172 phenotype 172
clone 185 plasmids 184
Perspectives for the Future complementary DNAs point mutation 173
As the examples in this chapter and throughout the book il- (cDNAs) 186 polymerase chain reaction
lustrate, genetic analysis is the foundation of our under- DNA cloning 182 (PCR) 192
standing of many fundamental processes in cell biology. By DNA library 185 probes 188
examining the phenotypic consequences of mutations that DNA microarray 199 recessive 173
inactivate a particular gene, geneticists are able to connect dominant 173 recombinant DNA 182
knowledge about the sequence, structure, and biochemical
functional complementation recombination 180
activity of the encoded protein to its function in the context
189 restriction enzymes 183
of a living cell or multicellular organism. The classical ap-
proach to making these connections in both humans and gene knockout 213 RNA interference (RNAi)
simpler, experimentally accessible organisms has been to genomics 222 216
identify new mutations of interest based on their phenotypes genotype 172 segregation 175
and then ro isolate the affected gene and its protein product. heterozygous 173 Southern blotting 198
Although scientists continue to use this classical genetic homozygous 173 temperatu re-sensi ti ve
approach to dissect fundamental cellular processes and bio- mutations 176
hybridization 188
chemical pathways, the availability of complete genomic se-
in situ hybridization 199 transfection 203
quence information for most of the common experimental
organisms has fundamentally changed the way genetic ex- linkage 182 transformation 184
periments are conducted.. Using various computational mutagen 172 transgenes 215
methods, scientists have identified the protein-coding gene mutation 172 vector 182
sequences in most experimental organisms, including E. coli, Northern blotting 198 wild type 172
yeast, C. elegans, Drosophila, Arabidopsis, mouse, and hu-
mans. The gene sequences, in turn, reveal the primary amino
acid seq uence of the encoded protein products, providing us
with a nearl y complete list of the proteins found in each of Review the Concepts
the major experimental organisms.
The approach taken by most researchers has thus shifted 1. Genetic mutations can provide insights into the mecha-
from discovering new genes and proteins to discovering the nisms of complex cellular or developmental processes. How
functions of genes and proteins whose sequences are already might your analysis of a genetic mutation be different depend-
known. Once an interesting gene has been identified, ge- ing on whether a particular mutation is recessive or dominant?
nomic sequence information greatly speeds subsequent ge- 2. What is a temperature-sensitive mutation? Why are
netic manipulations of the gene, including its designed temperature-sensitive mutations useful for uncovering the func-
inactivation, so that more can be learned about its function. tion of a gene?

Review the Concept 219

3. Describe how complementation analysis can be used to 12. In determining the identity of the protein that corre-
reveal whether two mutations are in the same or in different sponds to a newly discovered gene, it often helps to know
genes. Explain why complementation analysis will not work the pattern of tissue expression for that gene. For example,
with dominant mutations. researchers have found that a gene called SERPINA6 is ex-
4. Jane has isolated a mutant strain of yeast that forms red pressed in the liver, kidney, and pancreas but not in other
colonies instead of the normal white when grown on a plate. tissues. What techniques might researchers use to find out
To determine the mutant gene, she decides to use functional which tissues express a particular gene?
complementation with a DNA library containing a lysine se- 13. DNA polymorphisms can be used as DNA markers. De-
lection marker. lu addition to the unknown gene mutation, scribe the d1fferences between SNP and SSR polymorphisms.
the yeast are lacking the gene required to synthesize the How can these markers be used for DNA-mapping studies?
amino acids leucine and lysine. What media will Jane grow 14. How can linkage-disequilibrium mapping sometimes
her yeast on to ensure that they have acquired the library provide a much higher resolution of gene location than clas-
plasmids? How will she know when a library plasmid has sical linkage mapping?
complemented her yeast mutation? 15. Genetic linkage studies can usually only roughly locate
5. Restriction enzymes and DNA ligase play essential roles the chromosomal position of a "disease" gene. How can ex-
in DNA cloning. How is it that a bacterium that produces a pression analysis and DNA sequence analysis help locate a
restriction enzyme does not cut its own DNA? Describe disease gene within the region identified by linkage mapping?
some general features of restriction-enzyme sites. What are 16. The ability to selectively modify the genome in the
the three types of DNA ends that can be generated after cut- mouse has revolutionized mouse genetics. Outline the proce-
ting DNA with restriction enzymes? What reaction is cata- dure for generating a knockout mouse at a specific genetic
lyzed by DNA ligase? locus. How can the loxP-Cre system be used to conditionally
6. Bacterial plasmids often serve as cloning vectors. De- knock out a gene? What is an important medical application
scribe the essential features of a plasmid vector. What are the of knockout mice? .·
advantages and applications of plasmids as cloning vectors? 17. Two methods for functionally inactivating a gene with-
7. A DNA library is a collection of clones, each containing a out altering the gene sequence are by dominant-negative mu-
different fragment of DNA, inserted into a cloning vector. tations and RNA interference (RNAi). Describe how each
What is the difference between a eDNA and a genomic DNA method can inhibit expression of a gene.
library? You would like to clone gene X, a gene expressed
only in neurons, into a vector using a library as the source of
insert. If you have the following libraries at your disposal
(genomic library from skin cells, eDNA library from skin Analyze the Data
cells, genomic library from neurons, eDNA library from neu-
1. A culture of yeast that requires uracil for growth (ura3
rons), which could you use and why?
was mutagenized, and two mutant colonies, X andY, have
8. In 1993, Ka'ry Mullis won the Nobel Prize in Chemistry
been isolated. Mating type a cells of mutant X arc mated with
for his invention of the PCR process. Describe the three steps
mating type a cells of mutant Y to form diploid cells. The
in each cycle of a PCR reaction. Why was the discovery of a parental (ura3 ), X, Y, and diploid cells are streaked onto
thermostable DNA polymerase (e.g., Taq polymerase) so im- agar plates containing uracil and incubated at 23 oc or 32 °C.
portant for the development of PCR?
Cell growth was monitored by the formation of colonies on
~ ..
9. Southern and Northern blotting are powerful tools in the culture plates as shown in the figure below.
molecular biology based on hybridization of nucleic acids.
How are these techniques the same? How do they differ? Give
Denotes growth of cells
some specific applications for each blotting technique.
10. A number of foreign proteins have been expressed in
bacterial and mammalian cells. Describe the essential fea-
tures of a recombinant plasmid that are required for expres-
sion of a foreign gene. How can you modify the foreign
protein to facilitate its purification? What is the advantage
of expressing a protein in mammalian cells versus bacteria?
11. Northern blotting, RT-PCR, and microarrays can be used
to analyze gene expression. A lab studies yeast cells, Lumpar-
ing their growth in two different sugars, glucose and galac-
Growth at 23 C Growth at 32 oc
tose. One student is comparing expression of the gene HMG2
under the different conditions. Which technique(s) could he a. What can be deduced about mutants X and Y from
use and why? Another student wants to compare expression the data provided?
of all the genes on chromosome 4, of which there are approx- b. A wild-type yeast eDNA library, prepared in a plasmid
imately 800. What technique(s) could she use and why? that contains the wild-type URA3 + gene, is used to transform X

220 CHAPTERs • Molecular Genetic Techniques

cells, which are then cultured as indicated. Each black spot e. Haploid offspring of the diploid cells from part (a)
below represents a single clone growing on a petri plate. What above are generated. XY double mutants constitute 1/4 of
are the molecular differences between the clones growing on the these offspring. Haploid X cells, Y cells, and XY cells in liq-
two plates? How can these results be used to identify the plas- uid culture are synchronized at a stage just prior to budding
mid that contains a wild-type copy of gene X? Based on these and then are shifted from 2 3 oc to 32 oc. Examination of the
results, how can the identity of gene X be uncovered? cells 24 hours later reveals that X cell s are arrested with
small buds, Y cells are arrested with large buds, and XY cells
are arrested with small buds. What is the relationship be-
tween X andY?

Growth at 23 oc on Growth at 32 oc on
media lacking uracil media lacking uracil References
Genetic Analysis of Mutations to Identify and Study Genes
c. DNA is extracted from the parental cells, from X cells, Adams, A. E. .\1., D. Botstem, and D. B. Drubm. 1989. A yeast
and from Y cells. PCR primers are used to amplify the gene actin-binding protein is encoded by sac6, a gene found by suppres-
encoding X in both the parental and the X cells. The primers sion of an actin muration. Sctence 243:231.
are complementary to regions of DNA just external to the Griffiths, A. G. F., et al. 2000. An lntroduct10n to Genetic
gene encoding X. The PCR results are shown in the gel at the Analysts, 7th ed. W. H. heeman and Compa ny.
right. What can be deduced about the mutation in the X Guarente, L. 1993. Synthetic enhancement in gene mteraction: a
gene from these data? genetic tool comes of age. Trends Genet. 9:362-366.
Hartwell, L. H. 1967. Macromolecular synthesis of temperature-

-
sensitive mutants of yeast.}. Bactenof. 93:1662.
Hartwell, l.. H . 19 7 4. Genetic contro l of the cell di,islOn cycle
Ill yeast. Sctence 183:46.

- - --
(\;iisslein-Volhard, C., and E. Wieschaus. 1980. Mutations
affecting segment number and polarity in Drosophila. Nature
287:795-801.
Simon, M.A., er al. 1991. Rasl and a putative guanine
nucleotide exchange fac tor perform crucial steps in s1gnahng by the
sevenless protein tyrosine kinase. Ce//67:701-716.
Tong, A. H., et al. 2001. Systematic genenc analysis with
Parental X Y Parental X ordered arrays of yeast deletion mutants. Science 294:2364- 2368.
Southern PCR DNA Cloning and Characterization
Ausubel, F. M., ct al. 2002. Current Protocols 111 \!lo/ecular
d. A construct of the wild-type gene X is engineered to 810/og)'. Wiley.
encode a fu sion protein in which green fluorescent protein Gu bler, U., and B. J. Hoffman. 1983. A simple and ,·ery efficient
(GFP) is present a.r theN-terminus (GFP-X) or the C-tcrminus method for generating cO!'.' A hhranes. Gene 25:263-289.
Han, J. H., C. Strarowa, and W. ]. Rutter. 1987. Isolation of
(X-GFP ) of protein X. Both constructs, present on a URAT'-
full-length putative rat lysophospholipasc eDNA usmg improved
plasmid, are used to transform X cells grown in the absence methods for mRNA isolation and cDKA clomng. Bwchem.
of uracil. The transformants are then monitored for growth 26:161 7-1632.
at 32 °C, shown below at the left. At the right are typical ltakura, 1\.. , ]. J. Ross1, and R. B. Wallace. 1984. Synthesis and
fluorescent images of X-GFP a nd GFP-X cells grown at usc of synthetic oligonucleotides. Ann. Rev. Bwchem. 53:323-356.
23 oc in which green denotes the presence of green fluores- :\1aniatis, T., er al. 1978. The isolation of structu ral genes from
cent protein. What is a rea§onable explanation for growth of libraries of euea ryotlc DNA. Cel/15:687-701.
GfP-X but not X-GFP cells at 32 °C? :-\asmyth, K. A., and S. I. Reed. 1980. Isolation of genes by
complementation m yea~r: molecular cloning of a cell-cycle gene.
Proc. Nat'/ Acad. Set. USA 77:2119-2 123.
N"'l'u~ Nathans, D., and H. 0 . Smith. 1975. Restriction endonuclcases
in the analysis and restructuring of DNA molecules. Ann. Reu.
GFP-X cell
B10chem. 44:2 73-293.
Roberts, R. j ., and D. Macelis. 199-. REBASE-resmcrion
enzymes and methylases. Nucl. Acids Res. 25:248-262. Information
o n ::~ccess ing a continuously updated database on restriction and

~
modification enzymes ar http://www.neb.com/rebasc.
X-GFP"II
Using Cloned DNA Fragments to Study Gene Expression
Andrews, A. T. 1986. Electrophorests, 2d ed. Oxford University
X-GFP Press.
Growth at 32 C GFP localization in cells Erlich, H ., ed. 1992. PCR Technolog)': Prmciples and Appftca-
grown at 23 oc ttons for DNA Amplification. W. H. Freeman and Company.

References 221
 Pelltcer, A., ~1. Wigler, R. Axel, and S. Silverstein. 19~8. The Hastbacka, T., et al. 1994. The diastrophtc dysplasta gene
transfer and stable integration of the HSV thymidine kinase gene encodes a novel sulfate transporter: positional cloning by fine-structure
rnto mouse cells. Ce/141:133-141. linkage disequilibrium mapping. Ce/178:1073.
Saiki, R. K., ct al. 1988. Pnmer-directed enzymatic amplifica- Orita, M., et al. 1989. Rapid and sensitiYc detection of point
tion of Dt--:A with a thermostable DNA polymera~e. Scrence mutations and D;\JA polymorphisms using the polymerase chain
239:487-491. reaction. Genomrcs 5:874.
Sanger, F. 1981. Determination of nucleotide sequences in Tabor, H. K., N.J. Risch, and R. M. Myers. 2002. Opinion:
DNA. Scrence 214:1205-1210. cand1dare-gene approaches for studying complex genetic traits:
Souza, L. M., et al. 1986. Recombinant human granulocyte- practical considerations. Nat. Rev. Genet. 3:391-397.
colony stimulating factor: effcLl' u11 uurrnal and leukemic myeloid
Inactivating the Function of Specific Genes in Eukaryotes
cells. Science 232:61-65.
Capecchi, ~1. R. 1989. Altenng the genome b)' homologous
Wahl, G. M., J. L. Meinkoth, and A. R. Kimmel. 1987.
recombination. Scrence 244:1288-1292.
Northern and Southern blots. Meth. Enzymol. 152:572-581.
Deshaies, R. J., et al. 1988. A subfamily of stress proteins
Wallace, R. B., et al. 1981. The use of synthetic oligonucle-
facilitates translocation of secretory and mitochondrial precursor
otides as hybridization probes. II: Hybridization of oligonucleotides
polypeptides. Nature 332:8Q0-805.
of mixed sequence to rabbit (3-globin DNA. Nucl. Acids Res.
Fire, A., et al. 1998. Potent and specific genetic interference by
9:879-887.
double-stranded R~A in Caenorhabditis elegans. Nature 391:806--811.
Locating and Identifying Human Disease Genes Gu, H., et al. 1994. Deletion of a pKA polymerase beta gene
Botstein, D., et al. 1980. Construction of a generic linkage map segment rn T cells using cell type-spectfic gene targeting. Science
in man using restriction fragment length polymorphisms. Am.]. Genet. 265:103-106.
32:314-331. Zamore, P. D., et al. 2000. RNAi: double-stranded RNA directs
Donis-Keller, H., et al. 1987. A generic linkage map of the the ATP-dependenr cleavage of mRNA at 21 ro 23 nucleotide
human genome. Ce/151:319-337. intervals. Ce/1101:25-33.
Hartwell, L., et al. 2006. Genetics: From Genes to Genomes. Zimmer, A. 1992. Mampulating the genome by homologous
McGraw-Hill. recombmation in embryomc stem cells. Ann. Reu. Neurosct. 15:115.

. '

222 CHAPTERs • Molecular Genetic Techniques

 f , ;~
CHAPTER

'' •(. e• J ~ ~

I~ ~ ~ •: ll t 1f
1 -
2 3 4 5 6

.. ··'
"'
l i )f &·

'' ·~
._a 'a>
10 11 12

.
7 8 9

II . .
13
~ ~·

14 •• 15
~~
16
II 17
,,
18
~

Genes, Genomics,
II t •. (1 and Chromosomes
19 20 21 22 X

These brightly colored RxFISH-painted chromosomes are both

beautiful and useful in revealing chromosome anomalies and in
comparing karyotypes of different species.[~ Department of Clinical
Cytogenetics, Addenbrookes Hospital/Photo Researchers, Inc]

n previous chapters we learned how the structure and gene sequences often provide insight into possible functions

I composition of proteins allow them to perform a wide

variety of cellular functions. We a lso examined another
vital component of cells, the nucleic acids, a nd the process
of newly identified genes. Comparisons of genome sequence
and organization between species also help us understand
the evolution of organisms.
by which info rmation encoded in t he sequence of DNA is Surprisingly, DNA sequencing revealed that a large por-
translated into protein. In this chapter, our focus aga in is on tion of the genomes of higher eukaryotes does not encode
DNA and proteins as we consider the characteristics of eu- mRNAs or any other RNAs required by the organism. Re-
karyot ic n uclear and organellar genomes: the features of markably, such noncoding DNA constitutes = 98.5 percent
genes and the other DNA sequence that comprise the ge- of human chromosomal DNA! The noncoding DNA in mul-
nome, and how this DNA is structured and organized by ticellular organisms contains many regions that are similar
proteins w ithin the cell. but not identical. Variations within some stretches of this
By the beginning of the twenty-first century, molecular repetitious DNA between individuals are so great that every
biologists had completed seq uencing the entire genomes of person can be distinguished by a DNA "fingerprint" based
hundreds of viruses, scores of bacteria, and one unicellular on these sequence variations. Moreover, some repetitious
eukaryote, the budding yeastS. cerevisiae. By now, the vast DNA sequences are not found in the same positions in the
majority of t he genome sequence is also known for the fis- genomes of different individuals of the same species. At one
sion yeast S. pombe, the simple plant A. thaliana, rice, and time, all noncoding DNA was collectively termed "junk
mu ltiple multicellular animals (metazoans) including the DNA" and was considered to serve no purpose. '~/e now
rou ndworm C. elegans, the fruit fly D. melanogaster, mice, understand the evolutionary basis of all this extra DNA,
humans and at least one representative each of the =35 meta- and t he variation in location of certain sequences between
zoan phyla. Detailed analysis of these sequencing data has individuals. Cellular genomes harbor transposable (mobile)
revea led insights into evorution, genome organization, and DNA elemen ts that can copy themselves and move through-
gene fu nction. It has allowed researchers to identify previ- out the genome. Although transposable DNA clements seem
ously unknown genes and to estimate the total number of to have little function in the life cycle of an individual or-
protein-coding genes in each genome. Comparisons between ganism, over evolutionary time they have shaped our gcnomcs

OUTLINE

6.1 Eukaryotic Gene Structure 225 6.5 Genomics: Genome-wide Analysis of Gene
Structure and Expressi on 252
6.2 Chromosomal Organization of Genes
and Noncoding DNA 231 6.6 Structural Organization of Eukaryotic
Chromosomes 256
6.3 Transposable (Mobile) DNA Elements 234
6.7 Morphology and Functional Elements
6.4 Organelle DNAs 245 of Eukaryotic Chromosomes 266
and contributed to the rapid evolution of multicellular has revealed that these organelles evolved from intracellular
organisms. eubacteria that developed symbiotic relationships with ancient
In higher eukaryotes, DNA regions encoding proteins or eukaryotic cells.
functional RNAs-that is, genes-lie amidst this expanse of The sheer length of cellular DNA is a sign ificant prob-
apparently nonfunctional DNA. In addition to the nonfunc- lem with w hich cells must contend. The DNA in a single
tional DNA between genes, noncoding introns are common human cell, which measures about 2 meters in total length,
within genes of multicellular plants and animals. Sequencing must be contained within cells with diameters of less than
of the same protein-coding gene in a variety of eukaryotic spe- lO J..lm, a compaction ratio of greater than 10' to 1. In rela-
cies has shown that evolutionary pressure selects tor mainte- tive terms, if a cell were 1 centimeter in diameter (about the
nance of relatively similar sequences in the coding regions, size of a pea), the length of DNA packed into its nucleus
or exons. In contrast, wide sequence variation, even includ- would be about 2 kilometers (1.2 miles)! Specialized eu-
ing total loss, occurs among introns, suggesting that most karyotic proteins associated with nuclear DNA exquisitely
intron sequences have little functional significance. How- fold and organize the DNA so that it fits into nuclei. And
ever, as we shall see, although most of the DNA sequence of yet, at the same time, any given portion of this highly com-
introns is not functional, the existence of introns has favored pacted DNA can be accessed readily for transcription, DNA
the evolution of multidomain proteins that are common in replication, and repair of DNA damage without the long
higher cukaryotes. It also allowed the rapid evolution of pro- DNA molecules becoming tangled or broken. Furthermore,
teins with new combinations of functional domains. In addi- the integnty of DNA must be maintained during the process
tion, short noncoding RNAs called siRNAs and miRNAs of cell division when it is partitioned into daughter cells. In
that regulate translation and mRNA stability, and long non- eukaryotes, the complex of DNA and the proteins that or-
coding RNAs that may regulate transcription through their ganize it, called chromatin, can be visualized as individual ·.
influence on chromatin structure, can be processed from chromosomes during mitosis. As we wi ll sec in this and the
some introns (Chapters 7 and 8}. following chapter, the o rganization of DNA into chromatin
Mitochondria and chloroplasts also contain DNA that allows a mechanism for regulation of gene expression that
encodes proteins essential to the function of these vital or- is not available in bacteria.
ganelles. We shall see that mitochondrial and chloroplast In the first five sections of this chapter, we provide an
DNAs are evolutionary remnants of the origins of these or- overview of the landscape of eukaryotic genes and genomes.
ganelles. Comparison of DNA sequences between different First we discuss the structure of eukaryotic genes and the com-
classes of bacteria and mitochondrial and chloroplast genomes plexities that arise in higher organisms from the processing of

Higher-ord~
chromatin
~
folding
Loops of
30-nm fiber

FIGURE 6-1 Overview of the structure

of genes and chromosomes. DNA of higher
eukaryotes consists of unique and repeated "Beads on a string"
sequences. Only "" 1.5 percent of human DNA
encodes proteins and functional RNAs and the Mitochondrion
regulatory sequences that control their expres- ~
sion; the remainder is merely spacer DNA between
genes and introns within genes. Much of this DNA,
= 45 percent in humans, is derived from mobile
DNA elements, genetic symbiotes that have
contributed to the evolution of contemporary
genomes. Each chromosome consists of a single, Nucleosome
long molecule of DNA up to =280Mb in humans,
organized into increasing levels of condensation
by the histone and nonhistone proteins with Major Types of DNA Sequence
which it is intricately complexed. Much smaller Single-copy genes Simple-sequence DNA
DNA molecules are localized in mitochondria and Gene families Transposable DNA elements
Tandemly repeated genes Spacer DNA
chloroplasts. lntrons

224 CHAPTER 6 • Genes, Genomics, and Chromosomes

 mRNA precursors into alternatively spliced mRNAs. Next we Most Eukaryotic Genes Contain lntrons and
discuss the main classes of eukaryotic DNA, including the spe-
Produce mRNAs Encoding Single Proteins
cial properties of transposable DNA elements and how they
shaped contemporary genomes. We then consider organelle As discussed in Chapter 4, many bacterial mRNAs (e.g., the
DNA and how it differs from nuclear DNA. This background mRNA encoded by the trp operon) include the coding region
prepares us to discuss genomics, computer-based methods for for several proteins that function together in a biological
analyzing and interpreting vast amounts of sequence data. process. Such mRNAs are said to be polycistronic. (A cistron
The final two sections of the chapter address how DNA is is a genetic unit encoding a single polypeptide.) In contrast,
physically org::1nin•d in eukaryotic cells. We consider the most eukaryotic mRNA~ are monocistronic; that is, each
packaging of DNA and histone proteins into compact com- mRNA molecule encodes a single protein. This difference
plexes called nucleosomes that are the fundamental building between polycistronic and monocistronic mRNAs correlates
blocks of chromatin, the large-scale structure of chromo- with a fundamental difference in their translation.
somes, and the functional elements required for chromosome Within a bacterial polycistronic mRNA, a ribosome-
duplication and segregation. Figure 6-1 provides an overview binding site is located near the start site for each of the
of these interrelated subjects. The understanding of genes, ge- protein-coding regions, or cistrons, in the mRNA. Transla-
nomics, and chromosomes gained in this chapter will prepare tion initiation can begin at any of these multiple internal sites,
us to explore current knowledge about how the synthesis and producing multiple proteins (see Figure 4-13a). In most eu-
concentration of each protein and functional RNA in a cell is karyotic mRNAs, however, the 5'-cap structure directs ribo-
regulated in the following two chapters. some binding, and translation begins at the closest AUG start
codon (see Figure 4-13b). As a result, translation begins only
at this site. In many cases, the primary transcripts of eukary-
otic protein-coding genes are processed into a single type of
6.1 Eukaryotic Gene Structure mRNA, which is translated to give a single type of polypep-
In molecular terms, a gene commonly is defined as the entire tide (see Figure 4-15).
nucleic acid sequence that is necessary for the synthesis of a Unlike bacterial and yeast genes, which generally lack in-
functional gene product (polypeptide or RNA). According to trons, most genes in multicellular animals an·d plants contain
this definition, a gene includes more than the nucleotides en- introns, which are removed during RNA processing in the
coding an amino acid sequence or a functional RNA, referred nucleus before the fully processed mRNA is exported to the
to as the coding region. A gene also includes all the DNA se- cytosol for translation. In many cases, the introns in a gene are
quences required for synthesis of a particular RNA transcript, considerably longer than the exons. The median intron length
no matter where those sequences are located in relation to the in human genes is 3.3 kb. Some, however, are much longer: the
coding region. For example, in eukaryotic genes, transcrip- longest known human intron is 17,106 bp and lies within the
tion-control regions known as en hancers can lie 50 kb or titin gene, encoding a structural protein in muscle cells. In com-
more from the coding region. As we learned in Chapter 4, parison, most human exons contain only 50-200 base pairs.
other critical noncoding regions in eukaryotic genes include The typical human gene encoding an average-size protein is
the promoter, a~ well as sequences that specify 3' cleavage ""'50,000 bp long, but more than 95 percent of that sequence
and polyadenylation, known as poly(A) sites, and splicing of consists of introns and flanking noncoding 5' and 3' regions.
primary RNA transcripts, known as splice sites (see Figure Many large proteins in higher organisms that have re-
4-15). Mutations in these sequences, which control transcrip- peated domains are encoded by genes consisting of repeats
tion initiation and RNA processing, affect the normal expres- of similar exons separated by introns of variable length. An
sion and function of RNAs, producing distinct phenotypes in example of this is fibronectin, a component of the extracel-
mutant organisms. We examine these various control ele- lular matrix. The fibronectin gene contains multiple copies
ments of genes in greater detail in Chapters 7 and 8. of five types of exons (see figure 4-16). Such genes evolved
Although most genes are transcribed into mRNAs, which by tandem duplication of the DNA encoding the repeated
encode proteins, some DNA sequences are transcribed into exon, probably by unequal crossing over during meiosis as
RNAs that do not encode proteins (e.g., tRNAs and rRNAs, shown in Figure 6-2a.
described in Chapter 4, and miRNAs and siRNAs that regu-
late mRNA translation and stability, discussed in Chapter 8).
Because the DNA that encodes tRNAs rRNAs miRNAs Simple and Complex Transcription Units
and siRNAs can cause specific phenot~pes wh~n they ar~ Are Found in Eukaryotic Genomes
mutated, these DNA regions generally are referred to as The cluster of gPnes that forms a bacterial operon comprises
tRNA, rRNA, miRNA, and siRNA genes, even though the a single transcription unit that is transcribed from a specific
final products of these genes are RNA molecules and not promoter in the DNA sequence to a termination site, pro-
proteins. ducing a single primary transcript. In other words, genes and
In this section, we will examine the structure of genes in transcription units often are distinguishable in prokaryotes,
.· bacteria and eukaryotes and discuss how their respective since a single transcription unit contains several genes when
gene structures influence gene expression and evolution. they are part of an operon. In contrast, most eukaryotic genes

6.1 Eukaryotic Gene Structure 225

(al Exon duplication

L1 Exon 1 Exon 2 Exon 3

Parental
--- =:a::=::=========-====::~::::::~a:::::::=--­
chromosomes
{ L1 Exon 1 Exon 2
X
Exon 3

l Recombination (unequal crossing overl

L1 Exon 1 Exon 2 Exon 3 Exon 3

Recombinant {
chromosomes +
Exon 1 Exon 2
--- =-==:c=:===:~:==-=---
(bl Gene duplication
13-globin gene
L1 ~
Parental { - -- ===-•==~UUU===::a:=::::;=
chromosomes

13-globin gene l Recombination (unequal crossing overl

[}-globin gene

Recombinant { =::
chromosomes
===•-=====~ I
+
~F===~•====---
FIGURE 6 - 2 Exon and Gene Duplication. (a) Exon duplication results generate one recombinant chromosome where the gene now has four
from unequal crossing over during meiosis. Each parental chromosome exons (two copies of exon 3) and one chromosome where the gene is
(top) contains one ancestral gene containing three exons and two introns. missing exon 3. (b) The same process can generate duplications of entire
Homologous noncoding L1 repeated sequences lie 5' and 3' of the gene, genes. Each parental chromosome (top) contains one ancestralj3-globin
and also in the intron between exons 2 and 3. As we shall see later in the gene. After the u nequal recombination between L1 sequences, subse-
chapter, L1 sequences have been repeatedly transposed to new sites in quent independent mutations in the resulting duplicated genes could
the genome over th'e course of human evolution, so that all chromo- lead to slight changes in sequence that might result in slightly different
somes are peppered with them. The parental chromosomes are shown functional properties of the encoded proteins. Unequal crossing over also
displaced relative to each other, so that the L1 sequences are aligned. can result from rare recombinations between unrelated sequences.
Homologous recombination between L1 sequences as shown would [Part (b) see D. H. A. Fitch et al., 1991, Proc. Nat'/ Acad. Sci. USA 88:7396.]

are expressed from separate transcription units, and each mRNA. Multi ple mRNAs can arise from a primary tra n-
mRNA is translated into a single protein. script in three ways, as shown in Figure 6-3b.
Eukaryotic transcription units, however, are classified Examples of all three types of a lternative RNA process-
into two types, depend ing on the fate of the primary tran- ing occur in the genes that regu late sexual differentiation in
script. The primary transcript produced from a s1mple tran- Drosophila (sec Figure 8-16 ). Commonly, one mRNA is
scription unit, such as the one encoding [3-globin (Figure 4-15), produced from a complex transcrip ti on uni t in some cell
is processed to yield a single type of mRNA, encoding a sin- types, and a different mRNA is made in other cell types. For
gle protein. Mutations in exons, introns, and transcription- example, alternative splicing of the primary fibronectin tran-
control regions all may influence expression of the protein script in fibroblasts and hepatocytes determines whether o r
encoded by a simple transcription unit (Figure 6-3a). In hu- not the secreted protein includes domains that adhere to cell
mans, simple transcriprion units such as [3-globin are rare. surfaces (sec figure 4-16). The phenomenon of alternative
Approximately 90 percent of human transcription units arc splicing greatly expands the number of proteins encoded in
complex. The primary RNA transcript can be processed in the genomes of higher organisms. It is estimated that = 90
more than one way, leading to formation of mRNAs con- percent of human genes are contained within complex tran-
taining different exons. Each alternate mRNA, however, is ~cription units that give rise to alternatively spliced mRNAs
monocistronic, being translated into a single polypeptide, encod ing proteins w ith distinct functions, as for the fibr o-
with translation usually initiating at the first AUG in the blast and hepatocyte forms of fibronectin.

226 CHAPTER 6 • Genes, Genomics, and Chromosomes

(a) Simple transcription unit FIGURE 6-3 Simple and complex eukaryotic transcription units.
(a) A simple transcription unit includes a region that encodes one protein,
extending from the 5' cap site to the 3' poly(A) site, and associated
control regions. lntrons lie between exons (light blue rectangles) and are
removed during processing of the primary transcripts (dashed red lines);
Gene thus they do not occur in the fundional monocistronic mRNA. Mutations
in a transcription-control region (a, b) may reduce or prevent transcrip-
Control regions tion, thus red ucing or eliminating synthesis of the encoded protein. A
..-"' .............. .
,,-
mutation within an exon (c) may result in an abnormal protein with
mRNA 5' c:::=:l..c::::::::J. ..t::=:J 3' diminished activity. A mutation w ithin an intron (d) that introduces a
new splice site results in an abnormally spliced mRNA encoding a
nonfunctional protein. (b) Complex transcription units produce primary
(b) Complex transcription units transcripts that can be p rocessed in alternative ways. (Top) If a primary
Cap site transcript contains alternative splice sites, it can be processed into
a
1 c
Poly(A)
mRNAs with the same 5' and 3' exons but different internal exons.
(Middle) If a primary transcript has two poly(A) sites, it can be processed
Gene into mRNAs with alternative 3' exons. (Bo ttom) If alternative promoters
Exon 1 Exon 2 Exon 3 Exon 4 (for g) are active in different cell types, mRNA, , produced in a cell type in
which f is activated, has a different first exon (1A) than mRNA2 has, which
mRNA1 ---t::=i . VIlllll1.. ......
, - -....] 3' is produced in a cell type in which g is activated (and where exon 1B is
used). Mutations in control reg ions (a and b) and those designated c
or
5' ~--- •••••••t:=::f_...•·· ··-..._t::::::::=:::::JI 3' within exons shared by the alternative mRNAs affect the proteins
encoded by both alternatively processed mRNAs. In contrast, mutations
(designated d and e) w ithin exons unique to one of the alternatively
processed mRNAs affed only the protein translated from that mRNA.
Poly(A) For genes that are transcribed from different promoters in different cell

Gene
a
1 types (bottom), mutations in different control regions (f and g) affect
expression only in the cell type in which that control region is adive.

Exon 1 Exon 2
ment each other in a genetic complementation test, even
Vl!11TilPijl 3' though they occu r in the sam e gene. This is because a chro-
or mosome with mutation d can express a normal protein en-
- -~---- - ................. coded by mRNA 2 and a chromosome with mutation e can
5'c=:J ""VJ;.,..,'l!"'z"'lln/Znl!.'"B'
express a normal protein encoded by mRNA 1• Both mRNAs
produced from this gene would be present in a diploid cell
carrying both mutations, generating both protein products
and hence a wild-type phenotype. However, a chromosome
Gene with mutation c in an exon common to both mRNAs would
not complement either muta tion d or e. In other words, mu-
tation c would be in the same complementation g roups as
-- ·-. c:::::::r·· ··...., :~~----., 3' mutations d and e, even though d and e themselves would not
be in the same complementation group! Given these compli-
or ,... .."',.
cations with the genetic definition of a gene, the genomic
, 5' Wlli1.· ·•.t:::::::f,. •·...r=------.1 3'
defin ition outlined at the beginning of this section is com-
monly used. In the case of protein-coding genes, a gene is the
DNA sequence transcribed into a pre-mRNA precursor,
The relationship between a mutation and a gene is not eq uiva lent to a transcription unit, p lus any other regulatory
always straightforward w hen it comes to complex transcrip- elements requi red for synthesis of the primary transcript. The
tion units. A mutation in the control region or in an exon various proteins encoded by the alternatively spliced mRNAs
shared by alternatively spliced mRNAs will affect a ll th e al- expressed from one gene are called isoforms.
ternative proteins encoded by a giYen complex transcription
un it. On the other hand, mutations in an exon present in
Protein-Coding Genes May Be Solitary
only one of the alternati ve mRNAs will affect only the pro-
tein encoded by that mRNA. As explained in Chapter 5, ge- or Belong to a Gene Family
netic complementation tests commonl y are used to determine The nucleotide sequences within chromosomal DNA can be
if two mutations are in the same or d ifferent genes (see Fig- classified on the basis of structure and function, as shown in
ure 5-7). However, in the complex transcription unit shown Table 6-1. We will examine the properties of each class, begin-
in Figure 6-Jb (middle), mutations d and e would comple- ning with protein-coding genes, which comprise two groups.

6.1 Eukaryotic Gene Structure 227

1§:1!#1' Major Classes of Nuclear Eukaryotic DNA and Their Representation in the Human Genome

Class Length Copy Number in Human Genome Fraction of Human Genome (%)

Protem-coding genes 0.5-2200 kb =25,000 =55* (11.8)t

Tandemly repeated genes

U2 snRr-.;A = 20 <0.001
rRNAs = 300 0.4

Repetitious DNA
Simple-sequence DNA 1-500 bp Variable =6
Interspersed repeats (mobile DNA elements)
DNA transposons 2-3 kb 300,000 3
6-11 kb
..
LTR retrotransposons 440,000 8
Non-LTR retrotransposons
LINEs 6-8 kb 860,000 21
SINEs 100-400bp 1,600,000 13
Processed pseudogenes Variable 1-=100 =0.4

Unclassified spacer DNA§ Variable n.a. =25

' Complete transcripnon units mcluding mtrons.

1
Transcription units not including introns. Protein-coding regions (exons) totall.l% of the genome.
'Length of each repeat in a tandemlr repeated sequence.
'Sequences between transcription umts that are not repeated m the genome; n.a. = not applicable.
souRCE: International Human Genome Sequencing Consortium, 2001, Nawre 409:860 and 2004, Nature 431:9 3'1.

In multicellular organisms, roughly 25-50 percent of the identical a-globin polypeptides (encoded by another gene
protein-coding genes are represented only once in the haploid family) and four small heme groups to form a hemoglobin
genome and thus are termed solitary genes. A well-studied molecule (see Figure 3-13 ). All the hemoglobins formed from
example of a solitary protein-coding gene is the chicken lyso- the different [3-like globins carry oxygen in the blood, but
zyme gene. The 15-kb DNA sequence encoding chicken lyso- they exhibit somewhat different properties that are suited to
zyme constitutes a simple transcription unit containing four their specific functions in human physiology. For example,
exons and three introns. The flanking regions, extending for hemoglobins containing either the A.., or G'Y polypeptides are
about 20 kb upstream and downstream from the transcrip- expressed only during fetal life. Because these fetal hemoglo-
tion unit, do not encode any detectable mRNAs. Lysozyme, bins have a higher affinity for oxygen than adult hemoglobins,
an enzyme that cleaves the polysaccharides in bacterial cell they can effectively extract oxygen from the maternal circula-
walls, is an abundant component of chicken egg-white pro- tion in the placenta. The lower oxygen affinity of adult he-
tein and also is found in human tears. Its activity helps to moglobins, which are expressed after birth, permits better
keep the surface of the eye and the chicken egg sterile. release of oxygen to the tissues, especiaJly muscles, which
Duplicated genes constitute the second group of protein- have a high demand for oxygen during exercise.
coding genes. These are genes with close but nonidentical The different [3-globin genes arose by duplication of an
sequences that often are located within 5-50 kb of one an- ancestral gene, most likely as the resu lt of an "unequal cross-
other. A set of duplicated genes that encodes proteins w ith over" during meiotic recombination in a developing germ cell
similar but nonidentical amino acid sequences is called a (egg or sperm, see Figure 6-2b). Over evolutionary time the
gene family; the encoded, closely related, homologous pro- two copies of the gene that resulted accumulated ra ndom
teins constitute a protein family. A few protein families, such mutations; beneficial mutations that conferred some refine-
as protein kinases, vertebrate immunoglobulins, and olfac- ment in the basic oxygen-carrying function of hemoglobin
tory receptors, include hundreds of members. Most protein were retained by natural selection, resulting in sequence drift.
families, however, include from just a few to 30 or so mem- Repeated gene duplications and subsequent sequence drift
bers; common examples are cytoskclctal proteins, the myo- are thought to have generated the contemporary globin-like
sin heavy chain, and the a- and !3-globins in vertebrates. genes observed in humans and other mammals today.
The genes encoding the [3-like globins are a good example Two regions in the human [3-like globin gene cluster con-
of a gene family. As shown in Figure 6-4a, the [3-like globin tain nonfunctional sequences, called pseudogenes, similar to
gene family contains five functional genes designated f3, o, A.y, those of the functional [3-like globin genes (see Figure 6-4a).
G'Y, and e; the encoded polypeptides are similarly designated. Sequence analysis shows that these pseudogenes have the same
Two identical [3-like globin polypeptides combine with two apparent exon-intron structure as the functional [3-like globin

228 CHAPTER 6 • Genes, Genomics, and Chromosomes

 (a) Human ~- globi n gene cluster (chromosom e 11 )

I0Exon D Pseudogene "' Alusite I

(b ) S. cerevisiae (chromosome Ill)

I0 Open reading frame I tRNA gene

I

FIGURE 6-4 Comparison of gene density in hi gher and lower repeated sequence that is abundant in the human genome. (b) In the
eukaryotes. (a) In the diagram of the ~-globin gene cluster on diagram of yeast DNA from chromosome Ill, the green boxes indicate
human chromosome 11, the green boxes represent exons of ~-globin­ open reading frames. Most of these potential protein-coding
related genes. Exons spliced together to form one mRNA are sequences are functional genes without introns. Note the much
connected by caret-like spikes. The human ~-globin gene cluster higher proportion of noncoding-to-coding sequences in the human
contains two pseudogenes (white); these regions are related to the DNA than in the yeast DNA. [Part (a), see F. S. Collins and S. M. Weissman,
functional globin-type genes but are not transcribed. Each red arrow 1984, Prog. Nuc/. Acid Res. Mol. Bioi. 31 :315. Part (b), seeS. G. Oliver et al., 1992,
indicates the location of an Alu sequence, an =300-bp noncoding Nature 357:28.]

genes, suggesting that they also arose by duplication of the genes of the a- and 13-globin gene clusters found in mammals
same ancestral gene. However, there was little selective pres- today.
sure to maintain the function of these genes. Consequently, Several different gene families encode the. various proteins
sequence drift during evolution generated sequences that either that make up the cytoskeleton. These proteins are present in
terminate translation or block mRNA processing, rendering varying amounts in almost all cells. In vertebrates, the major
such regions nonfunctional. Because such pseudogenes are not cytoskeletal proteins are the actins, tubulins, and intermediate
deleterious, they remain in the genome and mark the location filament proteins such as the keratins, discussed in Chapters
of a gene duplication that occurred in one of our ancestors. 17, 18, and 20. We examine the origin of one such family, the
Duplications of segments of a chromosome (called seg- rubulin family, in Section 6 .5. Although the physiological ra-
mental duplication) occurred fairly often during the evolution tionale for the cytoskeletal protein families is not as obvious
of multicellular plants and animals. As a result, a large frac- as it is for the globins, the different members of a family prob-
tion of the genes in these organisms today has been dupli- ably have similar but subtly different functions suited to the
cated, allowing the process of sequence drift to generate gene particular type of cell in which they are expressed.
families and pse1,1dogenes. The extent of sequence divergence
between duplicated copies of the genome and characterization of
Heavily Used Gene Products Are Encoded
the homologous genome sequences in related organisms allows
an estimate of the time in evolutionary histOry when the duplica- by Multiple Copies of Genes
tion occurred. For example, the human fetal -y-globin genes (G'Y In vertebrates and invertebrates, the genes encoding ribosomal
and A'Y) evolved following the duplication of a 5.5-kb region RNAs and some other nonprotein-coding RNAs such as those
·. in the 13-globin locus that included the single -y-globin gene in involved in RNA splicing occur as tandemly repeated arrays.
the common ancestor of' caterrhine primates (Old World These are distinguished from the duplicated genes of gene
monkeys, apes, and humans) and platyrrhine primates (New families in that the multiple tandemly repeated genes encode
World monkeys) about 50 million years ago. identical or nearly identical proteins or functional RNAs.
Although members of gene families that arose relatively Most often, copies of a sequence appear one after the other, in
recently in evolution, such as genes of the human 13-globin a head-to-tail fashion, over a long stretch of DNA. Within a
locus, are often found near each other on the same chromo- tandem array of rRNA genes, each copy is nearly exactly like
some, members of gene families may also be found on differ- all the others. Although the transcribed portions of rRNA
ent chromosomes in the same organism. This is the case for genes are the same in a given individual, the nontranscribed
the human a-globin genes, which were separated from the spacer regions between the transcribed regions can vary.
13-globin genes by an ancient chromosomal translocation. Both These tandemly repeated RNA genes are needed to meet
the a- and 13-globin genes evolved from a single ancestral the great cellular demand for their transcripts. To understand
globin gene that was d uplicated (see Figure 6-2b) to generate why, consider that a fixed maximal number of RNA copies
the predecessors of the contemporary a- and 13-globin genes can be produced from a single gene during one cell generation
in mammals. Both the primordial a- and 13-globin genes then when the gene is fully loaded with RNA polymerase mole-
underwent further duplications to generate the different cules. If more RNA is required than can be transcribed from

6.1 Eukaryotic Gene Structure 229

one gene, multiple copies of the gene are necessary. For ex- Nonprotein-Coding Genes Encode
ample, during early embryonic development in humans, Functional RNAs
many embryonic cells have a doubling time of =24 hours and
contain 5-10 million ribosomes. To produce enough rRNA In addition to rRNA and tRNA genes, there are hundreds of
to form this many ribosomes, an embryonic human cell needs additional genes that arc transcribed into nonprotein-coding
at least 100 copies of the large and small subunit rRNA RNAs, some with various known function~, and many
genes, and most of these must be close to maximally active whose functions are not yet known. for example, small nu-
for the cell to divide every 24 hours. That is, multiple RNA clear RNAs (snRNAs) function in RNA splicing, and small
polymerases must be transcnbing each rRNA gene at the nucleolar RNAs (snoRNAs) function in rRNA processing
same time (see Figure 8-36). Indeed, all eukaryote~, including and base modification in the nucleolus. The RNase P RNA
yeasts, contain 100 or more copies of the genes encoding SS functions in tRNA processing, and a large fami ly (=1000) of
rRNA and the large and small subunit rRNAs. short micro RNAs (miRNAs) regulates the translation and
Multiple copies of tRNA genes and the genes encoding the stability of specific mRNAs. The functions of these nonpro-
histone proteins also occur. As we will see later in this chap- tein-coding RNAs are discussed in Chapter 8. An RNA
ter, histones bind and organize nuclear DNA. Just as the cell found in telomerase (see Figure 6-47) functions in maintain-
requires multiple rRNA and tRNA genes to support efficient ing the sequence at the ends of chromosomes, and the 7SL
translation, multiple copies of the histone genes are required RNA functions in the import of secreted proteins and most
to produce sufficient histone protein to bind the large amount membrane proteins into the endoplasmic reticulum (Chapter 13 ).
of nuclear DNA produced in each round of replication. While These and other nonprotein-coding RNAs encoded in th e
tRNA and histone genes often occur in clusters, they generally human genome and their functions, when known, are listed
do not occur in tandem arrays in the human genome. in Table 6-2.

1 §=1!3!1
RNA
Known Nonprotein-Coding RNAs and Their Functions

Number of Genes in Human Genome Function

rR!\As ==300 Protem synthesis

tRNAs ==500 Protein synthesis

snRNAs ==80 mRNA splicing

U7 snRNA Histone mRNA 3' processing

snoRNAs == 85 Pre-rRNA processing and rRNA modification

miRNAs ==1000 Regulation of gene expression

Xist X-chromosome inactivation

7SK 1 Transcription control

RNase P RNA tRNA 5' processing

7SL RKA 3 Protein secretion (component of signal recognition particle, SRP)

RNase MRP RNA rRNA processing

Telomerase RNA Template for addition of telomeres

Vault Rl\'As 3 Components of Vault ribonucleoproteins ( RNPs ), function unknown

hYI, hYJ, hY4, hY 5 ==30 Components of ribonuclcoproteins (RNPs), function unknown

Hl9 Unknown

~OuRn: International Human Genome Sequencing Consortium, 200 I, Nature 409:860, and P.D. Zamore and B. Haley, 2005, Sc 1cnce 309:1519.

230 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Ia tory function. For example, yeasts, fruit flies, chickens, and
KEY CONCEPTS of Section 6.1 humans have successively more D~A in their haploid chro-
mosome sets (12; 180; 1300; and 3300 Mb, respectively), m
Eukaryotic Gene Structure
keeping with what we perceive to be the increasing complex-
• In molecular terms, a gene is the entire DNA sequence re- ity of these organisms. Yet the vertebrates with the greatest
quired for synthesis of a functional protein or RNA mole- amount of DNA per cell are amphibians, which arc surely
cule. In addition to the coding regions (exons), a gene in- less complex than humans in their structure and behavior.
cludes control regions and sometimes introns. Even more surprising, the unicellular protozoan species
• A simple eukaryotic transcription unit produces a single Amoeba dubia has 200 times more DNA per cell than hu-
monocistronic mRNA, which is translated into a single protein. mans. Many plant species also have considerably more DNA
per cell than humans have. For example, tulips have 10 times
• A complex eukaryotic transcription unit is transcribed into a
as much DNA per cell as humans. The DNA content per cell
primary transcript that can be processed into two or more dif-
also varies considerably between closely related species. All
ferent monocistronic mRNAs depending on the choice of splice
insects or all amphibians would appear to be similarly com-
sites or polyadenylation sites. A complex transcription unit
plex, but the amount of haploid DNA in species within each
with alternate promoters or alternate polyadenylation sites also
of these phylogenetic classes varies by a factor of 100.
generates two or more different mRNAs (see Figure 6-3b).
Detailed sequencing and identification of exons in chro-
• Many complex transcription units (e.g., the fibronectin mosomal DNA have provided direct evidence that the ge-
gene) express one mRNA in one cell type and an alternate nomes of higher eukaryotes contain large amounts of
mRNA in a different cell type. noncoding DNA. For instance, only a small portion of the
About half the protein-coding genes in vertebrate genomic 13-globin gene cluster of humans, about 80 kb long, encodes
DNA are solitary genes, each occurring only once in the hap- protein (see Figure 6-4a). In contrast, a typical 80-kb stretch
loid genome. The remainder are duplicated genes, which of DNA from the yeast 5. cerevisiae, a single-celled eukary-
arose by duplication of an ancestral gene and subsequent ote, contains many closely spaced protein-coding sequences
independent mutations (see Figure 6-2b). The proteins en- without introns and relatively much less noncoding DNA
coded by a gene family have homologous but nonidentical (see Figure 6-4b). Moreover, compared witli other regions of
amino acid sequences and exhibit similar but slightly differ- vertebrate DNA, the 13-globin gene cluster is unusually rich
ent properties. in protein-coding sequences, and the introns in globin genes
are considerably shorter than those in most human genes.
In invertebrates and vertebrates, rRNAs are encoded by
Globin proteins comprise >50 percent of the total protein in
multiple copies of genes located in tandem arrays in genomic
developing red blood cells (reticulocytes) and the globin genes
DNA. Multiple copies of tRNA and histone genes also occur,
are expressed at maximum rate, i.e., a new RNA polymerase
often in clusters, but not generally in tandem arrays.
initiates transcr iption as soon as the previous polymerase
• Many genes also encode functional RNAs that are not transcribes far enough from the promoter to allow a second
translated into protein but nonetheless perform significant RNA polymerase to bind and initiate. Consequently, there
functions, such as rRNA, tRNA, and snRNAs. Among these has been selective press ure to evolve small introns in the glo-
are micro RNAs; possibly up to 1000 in humans, whose bio- bin genes that are compatible with the required high rate of
logical significance in regulating gene expression has only globin mRNA transcription and processing. However, the
recently been appreciated. vast majority of human genes are expressed at much lower
levels requiring production of one encoded mRNA on a time
scale of only tens of minutes or hours. Consequently, there
has been little selective pressure to reduce the sizes of introns
6.2 Chromosomat Organization of Genes in most human genes.
The density of genes varies greatly in different regions of
and Noncoding DNA human chromosomal DNA, from "gene-rich" regions, such
Having reviewed the relationship between transcription as the 13-globin cluster, to large gene-poor "gene deserts." Of
units and genes, we now consider the organization of genes the 96 percent of human genomic DNA that has been se-
on chromosomes and the relationship of noncoding DNA quenced, only "'=' 1.5 percent corresponds to protein-coding
sequences to coding sequences. sequences (exons). We learned in the previous section that
the intron sequences of most human genes are significantly
longer than the exon sequences. Approxim::~rely one-third of
human genomic DNA is thought to be transcribed into pre-
Genomes of Many Organisms Contain
mRNA precursors or nonprotein-coding RNAs in one cell or
Nonfunctional DNA another, but some 95 percent of this sequence is intronic,
Comparisons of the total chromosomal DNA per cell in vari- and thus removed by RNA splicing. This amounts to a large
ous species first suggested that much of the DNA in certain fraction of the total genome. The remaining two-thirds of
organisms does not encode RNA or have any apparent regu- human DNA is noncoding DNA between genes as well as

6 .2 Chromosomal Organization of Genes and Noncoding DNA 231

regions of repeated DNA sequences that make up the centro- with just the function or expression of the encoded gene. But
meres and telomeres of the h u man chromosomes. Conse- in the more common types of diseases associated with ex-
quently, =98.5 percent of human DNA is noncoding. panded microsatellite repeats, such as Huntington disease and
Different selective pressures during evolution may ac- myotonic dystrophy type 1, the expanded repeats behave like
count, at least in part, for the r emarkable difference in the dominant mutations. In some microsatellite-repeat diseases,
amount of nonfunctional DNA in different organisms. For triplet repeats occur within a coding region, resulting in the
example, many microorganisms must compete with other spe- formation of long polymers of a single amino acid that may
cies of microorganisms in the same environment for limited aggregate over time in long-lived neuronal cells, eventually in-
amounts of available nutrients, and metabolic economy rhus terfering with norma l cellular function. For example, expan-
is a critical characteristic. Since synthesis of nonfu nctional sion of a CAG repeat in the first exon of t he Huntington gene
(i.e., noncoding) DNA requires time, nutrients, and energy, leads to synthesis of long stretches of polyglutamine that over
presumably there was selective pressu re to lose nonfu nctional several decades form roxie aggregates resulting in neuronal cell
DNA during the evolution of many rapidly growing microor- death in H untington disease.
ganisms such as the yeast S. cerevisiae. On the other hand, Pathogenic expanded repeats can also occur in the non-
natural selection in vertebrates depends largely on their be- coding regio ns of some genes, where they are though t to
havior. The energy invested in DNA synthesis is trivial com-
pared with the metabolic energy required for t he movement of
(a) Normal replication
muscles; thus there may have been little selective pressure to
n
eliminate nonfunctional DNA in vertebrates. Also, the rephca- 5' 3'
tion time of cells in most multicellular organisms is much lon- 3' f------5'
ger than in rapid ly growing microorganisms, so there was n
little selective pressure to eliminate nonfunctional DNA in
order to permit rapid cellular replication. (b) Backw ard slippage

Most Simple-Sequence DNAs Are Concentrated

in Specific Chromosomal Locations
5' 3'
Beside~ duplicated protein-coding genes and tandemly repeated
3' 1 - - - - - - - 5'
genes, eukaryotic cells contain multiple copies of other DNA n
sequences in the genome, generally referred to as repeti tious
DNA (see Table 6-1). Ofrhe rwo ma in types of repetitious DNA,
the less prevalent is simple-sequence DNA or satellite DNA, n+ 1
which constitutes about 6 percent of the human genome and is 5' 3'
composed of perfect or nearly perfect repeats of relatively short 3' f - - - - - - 5'
sequences. The .more common type of repetitious DNA, col- n
lectively called interspersed repeats, is composed of much lon-
ger sequences. These sequences, consisting of several types of
1 Second replication

transposable elements, are discussed in Section 6.3 .

(c) Second replication
The length of each repeat in simple-sequence DNA can
range from I to 500 base pairs. Simple-sequence DNAs in
n+1
which the repeats contain 1-13 base pairs are often called 3'
microsatellites. Most microsatellite DNA has a repeat length 3' 5'
of 1-4 base pairs and usually occurs in tandem repeats of n+ 1
!50 repeats or fewer. Microsatellites arc thought to have +
originated by "backward slippage" of a daughter strand on n
5' 3'
its template strand du ring DNA replication so that the same
short sequence is copied twice (Figure 6-5). 3' 5'
n
FIGURE 6-5 Generation of microsatellite repeats by backward
Microsatellites occasionally occur within transcription
slippage of the nascent daughter strand during DNA replication. lf,
units. Some individuals are born with a larger number of
during replication (a), the nascent daughter strand "slips" backward
repeats in specific genes than are observed in rlw general popu- relative to the template strand by one repeat, one new copy of the
lation, presumably because of daughter-strand slippage during repeat is added to the daughter strand when DNA replication
DNA replication in a germ cell from which they developed. continues (b). An extra copy of the repeat forms a single-stranded loop
Such expanded microsatellites have been found to cause at in the daughter strand of the daughter duplex DNA molecule. If this
least 14 different types of neuromuscu lar diseases, depending single-stranded loop is not removed by DNA repair proteins before the .·
on the gene in which they occur. In some cases expanded micro- next round of DNA repl ication (c), the extra copy of the repeat is added
satellites behave like a recessive mutation because they interfere to one of the double-stranded daughter DNA molecules.

232 CHAPTER 6 • Genes, Genomics, and Chromosomes

function as dominant mutations because they interfere with the proper segregation of chromosomes to daughter cells dur-
RNA processing of a subset of mRNAs in the muscle cells ing mitosis. Simple-sequence DNA is also found in long tan-
and neurons where the affected genes are expressed. For ex- dem repeats at the ends of chromosomes, the telomeres, where
ample, in patients with myotonic dystrophy type 1, tran - they function to maintain chromosome ends and prevent their
scripts of the DMPK gene contain between 50 and 1500 joining to the ends of other DNA molecules, as discussed fur-
repeats of the sequence CUG in the 3' untranslated region, ther in the last section of this chapter.
compared to 5 to 34 repeats in normal individuals. The ex-
tended stretch of CUG repeats in affected individuals is DNA Fingerprinting Depends on Differences
thought to form long RNA hairpins (see Figure 4-9), which in length of Simple-Sequence DNAs
bind and sequester nuclear RNA-binding proteins that nor-
Within a species, the nucleotide sequences of the repeat units
ma lly regulate alternative RNA splicing of a subset of pre-
composing simple-sequence DNA tandem arrays are high!)
mRNAs essential for muscle and nerve cell function. •
conserved among individuals. In contrast, the number of re-
peats, and thus the length of simple-sequence tandem arrays
Most simple-sequence satellite DNA is composed of re-
containing the same repeat unit, is quite variable among in-
peats of 14-500 base pairs in tandem arrays of 20-100 kb. In
dividuals. These differences in length are thought to result
situ hybridization studies with metaphase chromosomes have
from unequal crossing over within regions of simple-sequence
localized these simple-sequence DNAs to specific chromo-
DNA during meiosis. As a consequence of this unequal cross-
somal regions. Much of this DNA lies near centromeres, the
ing over, the lengths of some tandem arrays are unique in
discrete chromosomal regions that attach to spindle microtu-
each individual.
bules during mitosis and meiosis (Figure 6-6). Experiments in
In humans and other mammals, some of the simple-
the fission yeast Schizosaccharomyces pombe indicate that
sequence DNA exists in relatively short 1- to 5-kb regions •
these sequences are required to form a specialized chromatin
made up of 20-50 repeat units, each containing "= 14-100
structure called centromeric heterochromatin, necessary for
base pairs. These regions are called minisatellites, in contrast
to microsatellites made up of tandem repeats of 1-13 base
pairs. Even slight differences in the total lengths of various

... minisatellites from different individuals can be detected by

Southern blotting (see Figure 5-26). This technique was ex-
# ploited in the first application of DNA fingerprinting, which
,
~ was developed to detect DNA polymorphmns (i.e., differ-
ences in sequence between individuals of the same species,
• • .. •
~ ·~
Figure 6-7). Today, the far more sensitive technique of poly-

... ... .. t.
merase chain reaction (PCR, Figure 5-20) is generally used in
forensic genetic testing. Microsatellite DNA sequences of short

.. "" • ... ' • .,

..
randem repeats of four bases in =30-50 copies are usually
analyzed today. The exact number of repeats generally varies
on the two homologous chromosomes of an individual
where they occur (one inherited from the mother and one

• f' 0
••• from the father) and on theY chromosome in males. A mix-
ture of pairs of PCR primers that hybridize to unique sequences
• ~
. flanking 13 of these short tandem repeats and a Y-chromosome
short tandem repeat arc used to amplify DNA in a sample.
ft The resulting mixture of PCR product lengths is unique in
the human population, except for identical twins. The use of
PCR methods allows analysis of minute amounts of DNA,
EXPERIMENTAL FIGURE 6-6 Simple-sequence DNA is localized and individuals can be distinguished more precisely and reli-
at the centromere in mouse chromosomes. Purified simple-sequence ably than by conventional fingerprinting.
DNA from mouse cells was copied in vitro using E. coli DNA polymerase I
and fluorescently labeled dNTPs to generate a fluorescently labeled DNA Unclassified Spacer DNA Occupies
"probe" for mouse simple-sequence DNA. Chromosomes from cultured a Significant Portion of the Genome
mouse cells were fixed and denatured on a microscope slide, and then
the chromosomal DNA was hybridized in situ to the labeled probe {light
As Table 6.1 shows, =25 percent of human DNA lies be-
blue). The slide was also stained with DAPI, a DNA-binding dye, to tween transcription units and is not repeated anywhere else in
visualize the full length of the chromosomes (dark blue). Fluorescence the genome. Much of this DNA probably arose from ancient
microscopy shows that the simple-sequence probe hybridizes primarily transposable elements that have accumulated so many muta-
to one end of the telocentric mouse chromosomes (i.e., chromosomes in tions over evolutionary time that they can no longer be recog-
which the centromeres are located near one end). [Courtesy of Sabine Mal, nized as having arisen from this source (see Section 6.3).
Ph.D., Manitoba Institute of Cell Biology, Canada.] Transcription-control regions on the order of 50-200 base

6.2 Chromosomal Organization of Genes and Non coding DNA 233

 (a) Paternity (b) Criminal identification
determination • In contrast, vertebrate and higher plant genomes contain
many sequences that do not code for RNAs or have any reg-
M C F1 F2
c:
ulatory function. Much of this nonfunctional DNA is com-
--==-+- Q)

E posed of repeated sequences. In humans, only about 1.5 per-

E
·;:;
'() cent of total DNA (the exons) actually encodes proteins or
Q)
<.J a. functional RNAs.
> en
• Variation in the amount of nonfunctional DNA in the ge-

--
- ..- nomes of various species is largely responsible for the lack of
-
- -- -
a consistent relationship between the amount of DNA in the
.._
--- -
5
- -
=-
--
haploid chromosomes of an animal or plant and its phyloge-
netic complexity.

- - --
-== • Eukaryotic genomic DNA consists of three major classes
of sequences: genes encoding proteins and functional RNAs,
- - - --
---=
____ -=
-= -
repetitious DNA, and spacer DNA (see Table 6-1 ).

--
.._ Simple-sequence DNA, short ~equences repeated in long
--=
---

--=-
+-- -- =
- -~
---
tandem arrays, is preferentially located in centromeres, tela-
meres, and specific locations within the arms of particular

--- a- --=--
---
chromosomes.

----
-==· - - --
FIGURE 6-7 Distinguishing individuals by DNA fingerpri nting.
(a) In this analysis of paternity, several minisatellite repeat lengths were
The length of a particular simple-sequence tandem array is
quite variable between individuals in a species, probably be-
cause of unequal crossing over during meiosis. Differences in
the lengths of some simple-sequence tandem arrays form the
determined by Southern blot analysis of restriction enzyme-digested basis for DNA fingerprinting (see Figure 6.7) .
genomic DNA and hybridization with a probe for a sequence shared by
several minisatellite sequences. This generated hypervariable multiband
patterns for each individual called "DNA fingerprints." Lane M shows the
pattern of restriction fragment bands using the mother's DNA; C, using
the child's DNA; and Fl and F2 using DNA from two potential fathers.
The child has minisatellite repeat lengths inherited from either the
6.3 Transposable {Mobile) DNA Elements
mother or Fl, demonstrating that Fl is the father. Arrows indicate Interspersed repeats, the second type of repetitious DNA in
restriction fragments from Fl , but not F2, found in the child's DNA. (b) In eukaryotic genomes, is composed of a very large number of
these "DNA fingerprints" of a specimen isolated from a rape victim and copies of relatively few sequence families (see Table 6-1).
three men suspected of the crime, it is clear that minisatellite repeat Also known as moderately repeated DNA, or intermediate-
lengths in the specimen match those of suspect 1. The victim's DNA was
repeat DNA, these sequences are interspersed throughout
included in the analysis to ensure that the specimen DNA was not
mammalian genomes and make up = 25-50 percent of mam-
contaminated with DNA from the victim. [From T. Strachan and A. P. Read,
malian DNA (=45 percent of human DNA).
Human Molecular Genetics 2. 1999, John Wiley & Sons.)
Because interspersed repeats have the unique ability to
"move" in the genome, they are collectively referred to as
pairs in length that help to regulate transcription from distant transposable DNA elements or mobile DNA elements (we use
promoters also occur in these long stretches of unclassified these terms interchangeably). Although transposable DNA
spacer DNA. In some cases, sequences of this seemingly non- elements originally were discovered in eukaryotes, they also
functional DNA are nonetheless conserved during evolution, are found in prokaryotes. The process by which these se-
indicating that they may perform a significant function that is quences are copied and inserted into a new site in the genome
not yet understood. For example, they may contribute to the is transposition. Transposable DNA elements are essentially
structures of chromosomes discussed in Section 6. 7. molecular symbiotes that in most cases appear to have no
specific function in the biology of their host organisms, but
exist only to maintain themselves. For this reason, Francis
KI=Y CONCEPTS of Section 6.2 Crick referred to them as "selfish DNA."
When rransposition occurs in germ cells, the transposed
Chromosomal Organization of Genes sequences at their new sites are passed on to succeeding gen-
and Noncoding DNA erations. In this way, mobile elements have multiplied and
In the genomes of prokaryotes and most lower eukaryotes, slowly accumulated in eukaryotic genomes over evolutionary
which contam few nonfunctional sequences, coding regions time. Since mobile clements are eliminated from eukaryotic
are densely arrayed along the genomic DNA. genomes very slowly, the} now constitute a significant por-
tion of the gcnomes of many eukaryotes.

234 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Not only are mobile elements the source for much of the Movement of Mobile Elements Involves
DNA in our genomes, they also provided a second mechanism, a DNA or an RNA Intermediate
in addition to meiotic recombination, for bringing about
chromosomal DNA rearrangements during evolution (see Barbara McClintock discovered the first mobile elements
Figure 6-2). One reason for this is that during transposition while doing classical genetic experiments in maize (corn) dur-
of a particular mobile element, adjacent DNA sometimes also ing the 1940s. She characterized genetic entities that could
is mobilized (see Figure 6-19 later in the chapter). T ransposi- move into and back out of genes, changing the phenotype of
tions occur rarely: in humans, about one new germ-line trans- corn kernels. Her theories were very controversial until simi-
position for every eight individuals. Since 98 ..5 percent of our lar mobile elements were discovered in bacteria, where they
DNA is noncoding, most transpositions have no deleterious were characterized as specific DNA sequences, and the mo-
effects. But over time they played an essential parr in the evo- lecular basis of their transposition was deciphered.
lution of genes having multiple exons and of genes whose As research on mobile elements progressed, they were
expression is restricted to specific cell types or developmental found to fall into two categories: (1) those that transpose di-
periods. In other words, although transposable elements rectly as DNA and (2) those that transpose via an RNA inter-
probably evolved as cellular symbiotes, they have had an im- mediate transcribed from the mobile element by an RNA
portant function in the evolution of complex, multicellular polymerase and then converted back into double-stranded
organisms. DNA by a reverse transcriptase (Figure 6-8). Mobile elements
Transposition also may occur within a somatic cell; in this that transpose directly as DNA are generally referred to as
case the transposed sequence is transmitted only to the daugh- DNA transposons, or simply transposons. Eukaryotic DNA
ter cells derived from that cell. In rare cases, such somatic-cell transposons excise themselves from one place in the genome,
transposition may lead to a somatic-cell mutation with detri- leaving that site and moving to another. Mobile elements that
mental phenotypic effects, for example, the inactivation of a transpose to new sites in the genome via an RNA intermediate
tumor-suppressor gene (Chapter 24 ). In this section, we first are called retrotransposons. Rctrotransposons make an RNA
describe the structure and transposition mechanisms of the copy of themselves and introduce this new copy into another
major types of transposable DNA elements and then consider site in the genome, while also remaining at their origmalloca-
their likely role in evolution. tion. The movement of retrotransposons is·analogous to the

(a) (b)

DNA transposon Retrotransposon

~ ~

Donor DNA Donor DNA

Flanking

Former
DNA
I~~t
~ymo"~
transposon
site Donor DNA
t ~
RNA intermediate
Donor DNA

Insertion
DNA
- - -_..----- intermediates ~
'==
! Reverse
transcriptase

Insertion
site site FIGURE 6-8 Two major classes of mobile

t elements. (a) Eukaryotic DNA transposons

(orange) move via a DNA intermediate, which is
Target Target excised from the donor site. (b) Retrotransposons
DNA DNA (green) are first transcribed into an RNA
molecule, which then is rever~e-transcribed into
double-stranded DNA. In both cases, the
double-stranded DNA intermediate is integrated
into the target-site DNA to complete movement.
Thus DNA transposons move by a cut-and-paste
mechanism, whereas retrotransposons move by
a copy-and-paste mechanism.

6.3 Transposable (Mobile) DNA Elements 235

infectious process of retroviruses (Figure 4-49). Indeed, retrovi- IS element (z 1- 2 kb)
ruses can be thought of as retrotransposons that evolved genes
5' 3'
encoding viral coats, thus allowing them to transpose betvveen 3' 5'
cells. Retrotransposons can be further classified on the basis
of their specific mechanism of transposition. To summarize, Inverted rep eat Protein-coding Target-site
( ~50 bp) region direct repeat
DNA transposons can be thought of as transposing by a "cut- (5-11 bp)
and-paste" mechanism, w hile retrotransposons move by a
"copy-and-paste" mechanism in which the copy is an RNA FIGURE 6-9 General structure of bacterial IS elements. The
relatively lilrge central region of an IS element, which encodes one or
intermediate.
two enzymes requi red for transposition, is flanked by an inverted
repeat at each end. The sequences of the inverted repeats are nearly
DNA Transposons Are Present in Prokaryotes identical, but they are oriented in opposite directions. The sequence
is characteristic of a particular IS element. The 5' and 3' short direct
and Eukaryotes
(as opposed to inverted) repeats are not transposed with the insertion
Most mobile elements in bacteria transpose directly as DNA. element; rather, they are insertion-site sequences that become
In contrast, most mobile elements in eukaryotes are ret- duplicated, with one copy at each end, during insertion of a mobile
rotransposons, but eukaryotic DNA transposons also occur. element. The length of t he direct repeats is constant for a given IS
Indeed, the original mobile elements discovered by Barbara element, but their sequence depends oM the site of insertion and there-
McClintock are DNA transposons. fore varies with each transposition of the IS element. Arrows indicate
sequence orientation. The regions in this diagram are not to scale; the
Bacte rial Insertion Sequences The first molecular under- coding region makes up most of the length of an IS element.
standing of mobile elements came from the study of certain
E. coli mutations caused by the spontaneous insertion of a
DNA sequence, = 1-2 kb long, into the middle of a gene. on the IS element, immediately adjacent to both ends of the
These inserted stretches of DNA arc called insertion sequences, inserted element. The length of the direct repeat is character-
or IS elements. So far, more than 20 different IS elements have istic of each type of IS element, but its sequence depends o n
been found in E. coli and other bacteria. the target site where a particular copy of the IS element in-
Transposition of an IS element is a very rare event, occur- serted. When the sequence of a mutated gene containing an IS
ring in only one in 105-10 cells per generation, depending on
7
element is compared with the wild-type gene sequence, only
the IS element. Often, transpositions inactivate essential genes, one copy of the short d irect-repeat sequence is found in the
killing the host cell and the IS elements it carries. Therefore, wild-type gene. Duplication of this target-site sequence to cre-
higher rates of transposition would probably result in too ate the second direct repeat adjacent to an IS element occurs
great a mutation rate for the host organism to survive. How- during the insertion process.
ever, since IS elements transpose more or less randomly, some As depicted in Figure 6-10, transposition of an IS element
transposed seq'-:lences enter nonessential regions of the ge- occurs by a "cut-and-paste" mechanism_ Tra nsposase per-
nome (e.g., regions between genes), allowing the cell to sur- forms three functions in this process: it ( 1) precisely excises
vive. At a very low rate of transposition, most host cells the IS element in the donor DNA, (2) makes staggered cuts in
survive and therefore propagate the symbiotic IS element. IS a short sequence in the target DNA, and (3) ligates the 3' ter-
elements also can insert into plasmids or lysogenic viruses, mini of the IS element to the 5' ends of the cut donor DNA.
and thus be transferred to other cells. In this way, IS elements Finally, a host-cell DNA polymerase fills in the single-stranded
can transpose into the chromosomes of virgin cells. gaps, generating the short direct repeats that flank IS elements,
The general structure of IS elements is diagrammed in and DNA ligase joins the free ends.
Figure 6-9. An inverted repeat of = 50 base pairs is invari-
ably present at each end of an insertion sequence. In an in- Eukaryotic DNA Transposons McClintock's original discov-
verted repeat, the 5' ~3' sequence on one strand is repeated ery of mobile elements came from observation of spontaneous
on the other strand, such as mutations in maize that affect production of enzymes required
to make anthocyanin, a purple pigment in maize kernels. Mu-
) tant kernels are white, and wild-type kernels are purple. One
5' GAGC;----GCTC 3' class of these mutations is revertible at high frequency,
3' CTCG CGAG 5' whereas a second class of mutations does not revert unless
they occur in the presence of the first class of mutations .
.\1cCiinrock called the agent responsible for the fir~t class of
Between the inverted repeats is a region that encodes mutations the activator (A c) element and those responsible for
transposase, an enzyme required for transposition of the IS the second class dissociation (Ds) elements because they also
element to a new site. The transposase is expressed very rarely, tended to be associated with chromosome breaks.
accounting for the very low frequency of transposition. An Many years after McClintock's pioneering discoveries,
important hallmark of IS elements is the presence of a short cloning and sequencing revealed that Ac elements are equiv-
direct-refJeat sequence, containing 5-11 base pairs, depending alent to bacterial IS elements. Like IS elements, they contain

236 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Donor DNA Target DNA transposed because they retain the inverted terminal repeats
recognized by the transposase.
5'
( ) 3' 5' ) 3'

'
1510 ]1111111 Since McClintock's early work on mobile clements m corn,
3' ( 5' 3'( 5'
1' I "' transposons have been identified in other eukaryotes. For in-
9-bp target site

II
1 T"O<PO"" m'ke< bluot-•nd•d
cuts in donor DNA and staggered
cuts in target DNA
stance, approximately half of all t he spontaneous mutations
observed in Drosophila are due to the insertion of mobile ele-
ments. Although most of the mobile elements in Drosophila
function as retrotransposons, at least one-the P element-
functions as a DNA transposon, moving by a mechanism simi-
5' 3' lar to that used by bacter ial insertion sequences. Current
1510
3' 1 1 5'
methods for constructing transgenic Drosophila depend on en-
gineered, high-level expression of the P-element transposase
5'---~ 3' 5 ' rrr111mlllm111---~) 3'
111111111 5• and use of the P-element inverted terminal repeats as targets for
3. ( 3· 5•
"----.r---1 transposition, as discussed in Chapter 5 (see Figure 5-22).
Unpaired
bases
DNA transposition by the cut-and-paste mechanism can
result in an increase in the copy number of a transposon if it
Transposase ligates 1510 occurs during the S phase of the cell cycle, when DNA syn-
fJ to 5' single-stranded ends
thesis occurs. An increase in the copy number happens when

5' 3'
.;.(--~>.1111111111
1 of target DNA

1570
5'
1111111111 <
3'
)
the donor DNA is from one of the two daughter DNA mol-
eCL.Jes in a region of a chromosome that has replicated and
the target DNA is in the region that has not yet replicated.
3' 5' 3' 5' When DNA replication is complete at the end of the S phase,
the target DNA in its new location is also replicated, resulting
in a net increase in the total number of these transposons in
Cellular DNA polymerase extends
IJ 3' cut ends and ligase join s the cell (Figure 6-11 ). When such a transposition occurs dur-

1 extended 3' ends to 1510 5' ends

)
ing the S phase preceding meiosis, one of die four germ cells
produced contains the extra copy of the transposon. Repeti-
tion of this process over evolutionary time has resulted in the
Ill~ 157o :JIIIIIII accumu lation of large numbers of DNA transposons in the
genomes of some organisms. Human DNA contains about
9-bp target-site 300,000 copies of full-length and deleted DNA transposons,
direct repeats
FIGURE 6-1 0 Model for transposition of bacterial insertion
sequences. Step 0 :Transposase, which is encoded by the IS element One copy of
(IS10 in this example), cleaves both strands of the donor DNA next to 111111 11 r 11111111111111111111 transposon
before S phase
the inverted repeats (da rk red), excising the ISIO element. At a largely
random target site;transposase makes staggered cuts in the target
DNA. In the case of IS 10, the two cuts are 9 bp apart. Step fl: Ligation
of the 3' ends of the excised IS element to the staggered sites in the
target DNA also is catalyzed by transposase. Step i): The 9-bp gaps of
sing le-stranded DNA left in the resulting intermediate are filled in by
5 phase: DNA
a cel lular DNA polymerase; finally, cellular DNA ligase forms the 3'--tS'
1111111 replication and
phosphodiester bonds between the 3' ends of the extended target DNA transposition
DNA strands and the 5' ends af the IS 10 strands. This process results in
duplication of the target-site sequence on each side of the inserted IS
element. Note that the length of the target site and IS 70 are not to
sca le. [See H. W. Benjamin and N. Kleckner, 1989, Ce// 59:373; and 1992, Proc.
Nat'/ A cad. Sci. USA 89:4648.]
After S phase, one
1111111111111111111 1111 : 111111 daughter molecule
has two copies of
inverted terminal r epeat sequences that flank t he coding re- 111 11 11 11 1111111 :: 1111111 the transposon
gion for a transposase, which recognizes the terminal repeats FIGURE 6-11 Mechanism for increasing DNA-transposon copy
and catalyzes transpositio n to a new site in DNA. Ds elements number. lf a DNA transposon that transposes by a cut-and-paste
are deleted forms of the Ac element in which a portion of the mechanism (see Figure 6-1 0) transposes during S phase from a region
sequence encoding transposase is missing. Because it does not of the chromosome that has replicated to a region that has not yet
encode a functional transposase, a Ds element cannot move replicated, then, when chromosomal replication is completed, one of
by itself. However, in plants that carry the Ac element and the two daughter chromosomes will have a net increase of one
thus express a functional transposase, Ds elements can be transposon insertion.

6.3 Transposable (Mobi le) DNA Elements 237

amounting to = 3 percent of human DNA. As we wi ll see LTR Coding region LTR Host-cell DNA
shortly, this mechanism can lead to the transposition of ge-

~
Integrated
nomic DNA as well as the transposon itself. retroviral
DNA t
LTR Retrotransposons Behave Like
Intracellular Retroviruses
The genomes of all eukaryotes studied, from yeast to hu-
Start site

1 Poly(A) site

RNA polym"'" II

mans, contain rctrorransposons, mobile DNA elements that Primary v ./VVVVV' ~ 3'
tra nscript 5'
transpose through an RNA intermediate utilizing a reverse
RNA-processing enzymes
transcriptase (see Figure 6-Sb ). These mobile elements are
divided into two major categories, those containing and
those lacking long terminal repeat s {LTRs). LTR retrotrans-
posons, which we discuss here, are common in yeast (e.g., Ty
Retroviral
RNA
1
fVVVVV'
Po ly(A) polymerase

(A)n
genome R-U5 U3-R
elements ) and in Drosophila (e.g., copia elements). Although
less abundant in mammals than non-LTR retrotransposons, FIGURE 6-13 Generation of retroviral genomic RNA from
LTR retrotransposons nonethe less constitute =8 percent of integrated retrovlral DNA. The left LTFYdirects cellular RNA poly-
human genomic DNA. In mammals, retrotransposons lack- merase to initiate transcription at the first nucleotide of the left R
ing LTRs are the most common type of mobile element; these region. The resulting primary transcript extends beyond the right LTR.
The right LTR, now present in the RNA primary transcript, directs
are described in the next section.
cellular enzymes to cleave the primary transcript at the last nucleotide
The general structure of L TR retrotransposons found in
of the right R region and to add a poly(A) tail, yielding a retroviral RNA
eukaryotes is depicted in Figure 6-12. In addition to short 5'
genome with the structure shown at the top of Figure 6-14. A similar
and 3' direct repeats typical of al l transposons, these ret- mechanism is thought to generate the RNA intermediate during
rotransposons arc marked by the presence of LTRs flanking transposition of retrotransposons. The short direct-repeat sequences
the central protein-coding region. These long direct terminal (black) of target-site DNA are generated during integration of the
repeats, containing =250-600 base pairs depending on the retroviral DNA into the host-cell genome.
type of LTR retrotransposon, are characteristic of integrated
retroviral DNA and arc critical to the life cycle of retrovi-
ruses. In addition to sharing L TRs with retroviruses, L TR viral LTR functions as a promoter that directs host-cell RNA
retrorransposons encode all the proteins of the most common polymerase to initiate transcription at the 5' nucleotide of the
type of retroviruses, except for the envelope proteins. Lack- R sequence. After the entire downstream retroviral DNA has
ing these envelope proteins, L TR rerrotransposons cannot been transcribed, the RNA sequence corresponding to the
bud from their host cell and infect other cells; however, they rightward LTR directs host-cell RNA-processing enzymes to
can transpose tq new sites in the DNA of their host cell. Be- cleave the primary transcript and add a poly(A) tai l at the 3'
cause of their clear relationship with retroviruses, this class of end of the R sequence. The resulting retroviral RNA genome,
retrotransposons is often called retrovirus-like elements. which lacks a complete LTR, exits the nucleus and is pack-
A key step in the retrovirallife cycle is formation of retro- aged into a virion that buds from the host cell.
viral genomic RNA from integrated retroviral DNA {see Fig- After a retrovirus infects a cell , reverse transcription of
ure 4 -49 ). This process serves as a model for generation of its RNA genome by the retrovirus-encoded reverse transcrip-
the RNA intermediate during transposition of L TR ret- tase yields a double -stranded DNA containing complete
rotransposons. As depicted in Figure 6-13, the leftward retro- LTRs (Figure 6-14). This DNA synthesis rakes place in the

LTR retrot ransposon (=<6-11 kb) FIGURE 6- 14 Model for reverse transcription of retroviral
genomic RNA into DNA. In this model, a complicated series of nine

~: ==
~/t----'~~'---_..lf= ~: events generates a double-stranded DNA copy of the single-stranded
RNA genome of a retrovirus. The genomic RNA is packaged in the virion
LTR Protein-coding Target-site with a retrovirus-specific cellular tRNA hybridized to a complementary
(250-600 bp) region direct repeat sequence near its 5' end called the primer-binding site (PBS). The
(5-10 bp)
retroviral RNA has a short direct-repeat terminal sequence (R) at each
FIGURE 6-12 General st ructure of eukaryotic LTR retrotra nspo- end. The overall reaction is carried out by reverse transcriptase, which
sons. The central protein-coding region i~ flanked by two long term mal catalyzes polymerization of deoxyribonucleotides. RNaseH digests the
repeats (LTRs), which are element-specific direct repeats. Like other RNA strand in a DNA-RNA hybrid. The entire process yields a double-
mobile elements, integrated retrotransposons have short target-site stranded DNA molecule that is longer than the template RNA and has
direct repeats at each end. Note that the different regions are not a long terminal repeat (LTR) at each end. The different regions are not
drawn to scale. The protein-coding region constitutes 80 percent or shown to scale. The PBS and Rregions are actually much shorter than
more of a retrotransposon and encodes reverse transcriptase, the US and U3 regions, and the central coding region is very much
integrase, and other retroviral proteins. longer than the other regions. [See E. Gilboa et al., 1979, Cell 18:93.]

238 CHAPTER 6 • Genes, Genomics, and Chromosomes

 ~ FOCUS ANIMATION: Retroviral Reverse Transcription

Coding region
Retrovi ral
genomic 5' '----'-'--'--="---'--'--'::..:::._---'-------,-------L~'-.1--!.!...-l
RNA

D t RNA extended
,--- -"--.
DNA ~
to form DNA copy
U3
of 5' end of
genomic RNA

fJ RNA of DNA-RNA 3'1 R

hybrid digested

IJ First jump: DNA hybridized

with remaining RNA R sequence
5' Ll_P~B~S::....._-L---:------....L~~~!-- (A)n3'

'· a DNA st rand extended 3.' I

-----.----____:...1-----..~~
PBS us
A,.
3' U3 R"
from 3' end . II I I I I I I II
5 PBS U3 R (A)n 3'

ll Most hybrid RNA digested

~_:.,__
! ~~___...~5

llJ 3' end of second DNA strand 3' I PBS U3 R U5

synthesized I I I I I I I I I
5' U3 R U5 PBS 3'
1 1
IJ tRNA in DNA-RNA l'lybrid 3' 1 PBS U3 R U5 j s•
digested
5' I U3 R U5 PBS 13'

liJ Second jump 3' I. I I I I

PBS
1 U3 R U5 15'
5' 1 U3 R Us PBS 3'

mBo th strands 3' II IU3 R US PBS

1
completed by . I I I I I I I
synth esis from 5' U3 R U5 PBS
3' ends
LTR LTR
Retroviral DNA

6 .3 Tran sposable (Mobil e) DNA Elem ent s 239

 cytosol. The double-stranded DNA with an LTR at each end
is then transported into the nucleus in a complex with inte-
grase, another enzyme encoded by retroviruses. Retroviral
integrases are closely related to the transposases encoded by
TymRNA ~
DNA transposons and use a similar mechanism to insert the
r - 1 Galactose-sensitive lntron from
1.-l promoter D another gene
double-stranded retroviral DNA into the host-cell genome.
In this process, short direct repeats of the target-site sequence
are generated at either end of the inserted viral DNA se-
quence. Although the mechanism of reverse transcnptlon is
Experiment 1 Experiment 2 complex, it is a critical aspect of the retrovirus life cycle. The
process generates the complete 5' LTR that functions as a
promoter for initiation of transcription precisely at the 5'

0'-----y----1
Gal-responsive Ty

Transform yeast cells;

'-----y----1
Gal-responsive Ty
with unrelated
added intron
nucleotide of the R sequence, while the complete 3' LTR
functions as a poly(A) site leading to polyadenylation pre-
cisely at the 3'nucleotide of the R sequence. Consequently,
no nucleotides are lost from an LTR retrotransposon as it
undergoes successive rounds of insertion, transcription, re-
verse transcription, and re-insertion at a new site.
As noted above, LTR retrotransposons encode reverse
grow in galactose- and

1 nongalactose-containing media

Results in galactose-
!
Results in galactose-
transcriptase and integrase. By analogy with retroviruses,
these mobile elements move by a "copy-and-paste" mechanism
whereby reverse transcriptase converts an RNA copy of a
donor element into DNA, which is inserted into a target site
containing medium containing medium
1. Ty mRNA synthesis increased 1. Ty mRNAs lack intron by integrase. The experiments depicted in Figure 6-15 pro-
2. Transposition of Ty elements 2. Transposed Ty elements vided strong evidence for the role of an RNA intermediate in
increased lack intron
transposition of Ty elements in yeast.
The most common LTR retrotransposons in humans are
~
Primary transcript called ER Vs, for endogenous retroviruses. Most of the 443,000
lANA ERV-related DNA sequences in the human genome consist
• splicing
only of isolated LTRs. These are derived from full-length pro-
~
viral DNA by homologous recombination between the two
Transposed Ty Ty mRNA
J I Reverse LTRs, resulting in deletion of the internal retroviral sequences.
~transcription Isolated LTRs such as these cannot be transposed to a new
position in the genome, but recombination between homolo-
Transposed Ty gous LTRs at different positions in the genome have likely con-
tributed to the chromosomal DNA rearrangements leading to
EXPERIMENTAL FIGURE 6- !i The yeast Ty element gene and exon duplications, the evolution of proteins with new
transposes through an RNA intermediate. When yeast cells are combinations of exons, and, as we will see in Chapter 7, the
transformed with a Ty-containing plasmid, the Ty element can evolution of complex control of gene expression.
transpose to new sites, although normally this occurs at a low rate.
Using the elements diagrammed at the top, researchers engineered
two different recombinant plasmid vectors containing recombinant
Non-LTR Retrotransposons Transpose
Ty elements adjacent to a galactose-sensitive promoter. These plasm ids by a Distinct Mechanism
were transformed into yeast cells, which were grown in a galactose- The most abundant mobile clements in mammals are ret-
containing and a nongalactose medium. In experiment 1, growth of
rotransposons that lack LTRs, sometimes called non viral
cells in galactose-containing medium resulted in many more transposi-
retrotransposons. These moderately repeated DNA se-
tions than in nongalactose medium, indicating that transcription into an
quences form two classes in mammalian genomes: long in-
mRNA intermediate is required forTy transposition. In experiment 2, an
terspersed elements (LINEs) and short interspersed elements
intron from an unrelated yeast gene was inserted into the putative
protein-coding region of the recombinant galactose-responsive Ty
(SINEs). In humans, full-length LINEs are ""6 kbp long, and
element. The observed absence of the intron in transposed Ty elements
SINEs are "'='300 bp long (see Table 6-1 ). Repeated sequences
is strong evidence that transposition involves an mRNA intermediate with the characteristics of LINEs have been observed in pro-
from which the intron was removed by RNA splicing, as depicted in the tozoans, insects, and plants, but for unknown reasons they
box on the right. In contrast, eukaryotic DNA transposons, like the Ac are particularly abundant in the genomes of mammals.
element of maize, contain introns within the transposase gene, SINEs also are found primarily in mammalian DNA. Large
indicating that they do not transpose via an RNA intermediate. [See J. numbers of LINEs and SINEs in higher eukaryotes have ac-
Boeke et al., 1985, Ce//40:491.] cumulated over evolutionary time by repeated copying of

240 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Long interspersed element (LINE) (=6 kb) element containing the inserted intron had transposed to
new sites in the hamster genome through an RNA intermedi-
~: ==•L;-r·_j'·~i6l;IIJj!all•. r ate that underwent RNA splicing to remove the intron. •

AfT-rich Protein-coding Target-site Since LINEs do not contain LTRs, their mechanism of
region region direct repeat
transposition through an RNA intermediate differs from
FIGURE 6-16 General structure of a LINE, a non-LTR retrotrans- that of LTR retrotransposons. ORFl and ORF2 proteins are
poson. Mammalian DNA carries two classes of non-LTR retrotranspo- translated from a LINE RNA. In vitro studies indicate that
sons, LINEs and SINEs. LINE structure is depicted here. The length of the transcription by RNA polymerase is directed b} promoter
target-site direct repeats varies among copies of the element at
sequences at the left end of integrated LINE DNA. LINE
different sites in the genome. Although the full-length L1 sequence
RNA is polyadenylated by the same post-transcriptional
is = 6 kb long, variable amounts of the left end are absent at over
90 percent of the sites where this mobile element is found. The shorter
mechanism that polyadenylates other mRNAs. The LINE
open reading frame (ORF 1), = 1 kb in length, encodes an RNA-binding
RNA then is exported into the cytosol, where it is translated
protein. The longer ORF2, ""4 kb in length, encodes a bifunctional into ORFl and ORF2 proteins. Multiple copies of ORFl
protein with reverse transcriptase and DNA endonuclease activity. Note protein then bind to the LINE RNA, and ORF2 protein
that LINEs lack the long terminal repeats found in LTR retrotransposons. binds ro the poly(A) tail. The LINE RNA is then transported
back into the nucleus as a complex with ORFland ORF2
proteins, and is reverse-transcribed into LINE DNA in the
sequences at a few positions in the genome and insertion of nucleus by ORF2. The mechanism involves staggered cleav-
the copies into new positions. age of cellular DNA at the insertion site, followed by prim-
ing of reverse transcription by the resulting cleaved cellular
LINEs Human DNA contains three major families of LINE DNA as detailed in Figure 6-17. The complete process re-
sequences that are similar in their mechanism of transposi- sults in insertion of a copy of the original LINE retrotrans-
tion but differ in their sequences: Ll, L2, and L3. Only mem- poson into a new site in chromosomal DNA. A short direct
bers of the Ll family transpose in the contemporary human repeat is generated at the insertion site bec<~;use of the initial
genome. Apparently there are no remaining functional cop- staggered cleavage of the two chromosomal DNA strands.
ies of L2 or L3. LINE sequences are present at = 900,000 As noted already, the DNA form of an LTR retrotrans-
sites in the human genome, accounting for a staggering 21 poson is synthesized from its RNA form in the cytosol using
percent of total human DNA. The general structure of a a cellular tRNA as a primer for reverse transcription of the
complete LINE is diagrammed in Figure 6-16. LINEs usually first strand of DNA (see Figure 6- 14 ). The resulting double-
are flanked by short direct repeats, the hallmark of mobile stranded DNA with long terminal repeats is then transported
elements, and contain two long open reading frames (ORFs). into the nucleus, where it is integrated into chromosomal
ORF1, = 1 kb long, encodes an RNA-binding protein. DNA by a retrotransposon-encoded integrase. In contrast,
ORF2, = 4 kb long, encodes a protein that has a long region the DNA form of a non-LTR retrotransposon is synthesized
of homology with the reverse transcriptases of retroviruses in the nucleus. The synthesis of the first strand of the non-
and LTR retrorransposons, but also exhibits DNA endonu- LTR retroviral DNA by ORF2, a reverse transcriptase, is
clease activity. · primed by the 3' end of cleaved chromosomal DNA, which
base-pairs with the poly(A) tail of the non-L TR retroviral
B Evidence for the mobility of L 1 elements first came RNA (see Figure 6-17, step 111). Since its synthesis is primed
H from analysis of DNA cloned from humans with cer- by the cut end of a cleaved chromosome, and synthesis of the
tain genetic diseases such as hemophilia and myotonic dys- other strand of the non-LTR retrotransposon DNA is primed
trophy. DNA from these patients was found to carry by the 3' end of chromosomal DNA on the other side of the
mutations resulting from ,insertion of an L1 clement into a initial cut in chromosomal DNA (step [4J), the mechanism of
gene, whereas no such element occurred within this gene in synthesis results in integration of the non-LTR retrotranspo-
either parent. About 1 in 600 mutations that cause signifi- son DNA. There is no need for an integrase to insert the
cant disease in humans are due toLl transpositions or SINE non-LTR retrotransposon DNA.
transpositions that are catalyzed by Ll -encoded proteins. The vast majority of LINEs in the human genome are
Later experiments similar to those just described with yeast truncated at their 5' end, suggesting that reverse transcription
Ty elements (see Figure 6-15) confirmed that L1 elements terminated before completion and the resulting fragments ex-
transpose through an RNA intermediate. In these experiments, tending variable distances from the poly(A) tail were inserted.
an intron was introduced into a cloned mouse L1 element, Because of this shortening, the average size of UNE elements
and the recombinant L 1 element was stably transformed into is only about 900 base pairs, even though the full-length se-
cultured hamster cells. After several cell doublings, a PCR- quence is = 6 kb long. Truncated LINE elements, once
amplified fragment corresponding to the Ll element but formed, probably are not further transposed because they
lacking the inserted intron was detected in the cells. This lack a promoter for formation of the RNA intermediate in
finding strongly suggests that, over time, the recombinant Ll transposition. In addition to the fact that most Ll insertions

6.3 Transposable (Mobile) DNA Elements 241

FIGURE 6-1 7 Proposed mechanism of LINE reverse transcript ion ORF2 p rotein
and integration. Only ORF2 protein is represented. Newly synthesized Chromosomal DNA
LINE DNA is shown in black. ORFl and ORF2 proteins, produced by 5' 3'
translation of LINE RNA in the cytoplasm, bind to LINE RNA and 3' 5'
transport it into the nucleus. Step 0 : In the nucleus, ORF2 makes
staggered cuts in AT-rich target-site DNA, generating the DNA 3'-0H
ends indicated by blue arrowheads. Step f): The 3' end of the T-rich
DNA strand hybridizes to the poly(A) tail of the LINE RNA and primes
DNA synthesis by ORF2. Step D :ORF2 extends the DNA strand using D 1 Nicking
the LINE RNA as template. Steps B and lit: When synthesis of the LINE Nick site Nick site
DNA bottom strand reaches the 5 ' end of the LINE RNA template, ORF2
extends the newly synthesized LINE DNA using as template the 5' 3'
3' 5'
top-strand cellular DNA generated by the initial ORF2 staggered
cleavage. Step [;J: A cellular DNA polymerase extends the 3' end of the
top strand generated by the in itial ORF2 staggered cut, using the newly
synthesized bottom-strand LINE DNA as template. The LINE RNA is
digested as the DNA polymerase extends the upper-strand DNA, just
as occurs during removal of lagging-strand primer RNA during cellular fJ 1 Priming of reve rse tra}1scri ption
DNA synthesis (Figure 4·33). Step 6 :The 3 ' end of the newly synthe- by chromosomal DNA
sized DNA strands are ligated to the 5' ends of the cellular DNA strands
as in lagging-strand cellular DNA synthesis. [Adapted from D. D. Luan et al., 5' 3'
1993, Cell 72:595.] 3' 5'

are truncated, nearl y all the full-length elements contain stop

codons and frameshift mutations in ORF1 and ORF2; these
lEI 1 Reve rse transcription
of LIN E RNA by ORF2
mutations probably have accumulated in most LINE se-
quences over evolutionary time. As a result, per-
5' 3'
cent of the LINE sequences in the human genome are 3' TGA 5'
full-length, with intact open reading frames for ORF1 and
ORF2, representing =60-100 in total number.

SINEs The second most abundant class of mobile elements

in the human genome, SINEs constitute = 13 percent of total
human DNA. Varying in length from about 100 to 400 base
a1
pairs, these retrotransposons do not encode protein, bur
most contain a 3' Aff-rich sequence similar to that in LINEs.
SINEs are transcribed by the same nuclear RNA polymerase
5'
3' - 3'
5'

that transcribes genes encoding tRNAs, SS rRNAs, and

other small stable RNAs. Most li kely, the ORF1 and ORF2 I!J 1 Copying of chromosoma l
proteins expressed from full-length LI:t\Es mediate reverse DN A by O RF 2
transcnption and integration of SINEs by the mechanism de-
picted in hgure 6-17. Consequently, SINEs can be viewed as 5' AAA 3'
parasites of the LINE symbiotes, competing with LINE RNAs 3' TTTATGA 5'
for binding and reverse transcription/integration by LINE
encoded ORF I and ORF2.
SINI:.s occur at about 1.6 million sites in the human ge- Ill 1 1nsertion completed by
nome. Of these, =1.1 mi llion are Alu elements, so named cell u lar enzymes
because most of them contain a single recognition site for the 3'
5' AAATACT~ ./'\,/'AAA 3'
restriction enzyme Alul. Alu elements exhibit considerable
3' ~TTTATGA~TTTATGA 5'
sequence homology with and probably evolved from 7SL
RNA, a cytosolic RNA in ;1 ribonucleoprotein complex
called the signal recognition particle. This abundant cyto-
solic ribonucleoprotein particle aids in targeting certain
polypeptides to the membranes of the endoplasmic reticulum
ol 3'
5'
(Chapter 13). Alu elements arc scattered throughout the 3' 5'
human genome at sites w here their insertion has not dis-
rupted gene expression: between genes, within introns, and

242 CHAPTER 6 • Genes, Genomics, and Chromosomes

in the 3' untranslated regions of some mRNAs. For instance, Mobile DNA Elements Have Significantly
nine Alu elements are located within the human [3-globin Influenced Evolution
gene cluster (see Figure 6-4a). Of the one new germ-line non-
LTR retrotransposition that is estimated to occur in about Although mobile DNA elements appear to have no direct func-
every eight individuals, = 40 percent involve Ll elements tion other than to maintain their own existence, their presence
and 60 percent involve SINEs, of which =90 percent are Alu has had a profound impact on the evolution of modern-da)'
elements. organisms. As mentioned earlier, about half the spontaneous
Similar to other mobile elements, most SINEs have ac- mutations in Drosophila result from insertion of a mobile
cumulated mutatiom from the time of their insertion in the DNA element into or near a transcription unit. In mammab,
germ line of an ancient ancestor of modern humans. Like mobile elements cause a much smaller proportion of sponta-
LINEs, many SINEs also are truncated at their 5' end. neous mutations:= J 0 percent in mice, and only 0.1-0.2 per-
cent in humans. Still, mobile elements have been found in
mutant al leles associated with several human genetic dis-
eases. For example, insertions into the clotting factor IX
Other Retroposed RNAs Are Found
gene cause hemophilia, and insertions into the gene encoding
in Genomic DNA the muscle protein dystrophin lead to Duchenne muscular
In add ition to the mobile elements listed in Table 6-1, DNA dystrophy. The genes encoding factor IX and dystrophin are
copies of a wide variety of mRNAs appear to have integrated both on the X chromosome. Because the male genome has
into chromosomal DNA. Since these sequences lack introns only one copy of the X chromosome, transposition inser-
and do not have flanking sequences similar to those of the tions into these genes predominantly affect males.
functional gene copies, they clearly are not simply duplicated In lineages leading to higher eukaryotes, homologous re-
genes that have drifted into nonfunctionality and become combination between mobile DNA clements dispersed
pseudogenes, as discussed earlier (Figure 6-4a). Instead, throughout ancestral genomes may have generated gene du-
these DNA segments appear to be retrotransposed copies of plications and other DNA rearrangements during evolution
spliced and polyadenylated mRNA. Compared with normal (see Figure 6-2b). For instance, cloning and sequencing of the
genes encoding mRNAs, these inserted segments generally [3-globin gene cluster from various primate species has pro-
contain multiple mutations, which are thought to have ac- vided strong evidence that the human G-y and A-y genes arose
cumulated since their mRNAs were first reverse-transcribed from an unequal homologous crossover between two L1 se-
and randomly integrated into the genome of a germ cell in an quences flanking an ancestral globin gene. Subsequent diver-
ancient ancestor. These nonfunctional genomic copies of mRNAs gence of such duplicated genes could lead to acquisition of
are referred to as processed pseudogenes. Most processed pseu- distinct, beneficial functions associated with each member of
dogenes are flanked by short direct repeats, supporting the hy- a gene family. Unequal crossing over between mobile elements
pothesis that they were generated by rare retrotransposition located within introns of a particular gene could lead to the
events involving cellular mRNAs. duplication of exons within that gene (see Figure 6-2a). This
Other interspersed repeats representing partial or mutant process most likely influenced the evolution of genes that con-
copies of genes encoding small nuclear RNAs (snRNAs) and tain multiple copies of simi lar exons encoding similar protein
tRNAs are found in mammalian genomes. Like processed domains, such as the fibronectin gene (see Figure 4-16).
pseudogenes derived from mRNAs, these nonfunctional cop- Some evidence suggests that during the evolution of higher
ies of small RNA genes are flanked by short direct repeats eukaryotes, recombination between mobile DNA elements
and most likely result from rare retrotransposition events (e.g., Alu elements) in introns of two separate genes also oc-
that have accumulated through the course of evolution. En- curred, generating new genes made from novel combinations
zymes expressed from a LINE are thought to have carried of preexisting exons (Figure 6-18). This evolutionary pro-
out all these retrotransppsition events involving mRNAs, cess, termed exon shuffling, may have occurred during evo-
snRNAs, and tRNAs. lution of the genes encoding tissue plasminogen activator,

Alu Alu

': ~:::
Gene 1

Gene2 D FIGURE 6-18 Exon shuffling via recombina-

Alu Alu
tion between homologous interspersed

1 Double crossover
between A/u elements
repeats. Recombination between interspersed
repeats in the introns of separate genes produces

==*·~·-=====*
• ==~0~~·
+
~·~ .. ~=
transcription units with a new combination of
exons. In the example shown here, a double
crossover between two sets of Alu repeats results
in an exchange of exons between the two genes.

6.3 Transposable (Mobile) DNA Elements 243

FIGURE 6 - 19 Ex on shuffling by transposi- (a) DNA transposons

~-~E£ ~~==~·~
tion . (a) Transposition of an exon flanked by
homologous DNA transposons into an intron on Gene 1
a second gene. As we saw in Figure 6-10, step D .
~ Transposase excision from gene 1
transposase can recognize and cleave the DNA
at the ends of the transposon inverted repeats. .----.r::~=-a=CJ
In gene 1, if the transposase cleaves at the left
end of the transposon on the left and at the right lnsert•on site
end ofthe transposon on the right, it can
transpose all the intervening DNA, including the Gene2
t
exon from gene 1, to a new site in an intron of
gene 2. The net result is an insertion of the exon
~ Transposase insertion into gene 2
from gene 1 into gene 2. (b) Integration of an
exon into another gene via LINE transposition.
Some LINEs have weak poly(A) signals. If such a
LINE is in the 3'-most intron of gene 1, during Weak poly(AI Gene's poly(A)
transposition its transcription may continue (b) signal signal
beyond its own poly(A) signals and ext end into
the 3' exon, transcribing the cleavage and Gene 1 LINE
t' t
polyadenylation signals of gene 1 itself. This RNA - II 3'exon
can then be reverse-transcribed and integrated
by the LINE ORF2 protein (Figure 6-17) into an
intron on gene 2, introducing a new 3' exon
(from gene 1) into gene 2.
~AAAA
! Transcription and polyadenylation
at end of downstream exon

nsertton site

Gene2

! ORF2 reverse transcription

and insertion

the Neu receptor, and epidermal growth factor, which all con- produced in modern organisms, as we discuss in the next
tain an EGf domain (see Figure 3-11 ). In this case, ex on chapter.
shuffling presumably resulted in insertion of an EGF domain- These considerations suggest that the early view of mobile
encoding exon into an intron of the ancestral form of each of DNA elements as completely selfish m olecular parasites
these genes. misses the mark. Rather, they have contributed profoundly to
Both DNA transposons and LINE retrotransposons have the evolution of higher organisms by promoting (1) the gen-
been shown to occasionally carry unrelated fla nking se- eration of gene families via gene duplica tion, (2) the creation
quences when they transpose to new sites by the mechanisms of new genes via shuffling of preexisting exons, and (3) for-
diagrammed in Figure 6-19. These mechanisms likely also mation of more complex regulatory regions that provide mul-
contributed to exon shuffling during the evolution of con- tifaceted control of gene expression. Today, researchers are
temporary genes. attempting to harness transposition mechanisms for inserting
In addition to causing changes in coding sequences in therapeutic genes into patients as a form of gene therapy.
the genome, recombination between mobile elements and
transposition of DNA adjacent to DNA transposons and B A process analogous to that shown in Figure 6-19a is
retrotransposons likely played a significant ro le in the evo- H largely responsib le for the rapid spr ead of antibiotic
lution of regulatory sequences that control gene expression. resistance among pathogenic bacteria, a major problem in
As noted earlier, eukaryotic genes have transcription-control modern medicine. Bacterial genes encoding enzymes that inac-
regions called enhancers that can operate over distances of tivate antibiotics (drug resistance genes) have been flanked by
rcns of thousands of base pairs. Transcription of many insertion sequences generating drug re!>i!>tam:e transposons.
genes is controlled through the combined effects of several The widespread use of antibiotics in medicine, often unnec-
enhancer elements. Insertion of mobile elements near such essarily in the treatment of viral infections where they have
transcription-control regions probably contributed to the no effect, and to prevent infections of healthy agricultural
evolution of new combinations of en hancer sequences. animals, has led to the selection of such drug resistance
These in turn control which specific genes are expressed in transposons that have inserted into conjugating plasmids.
particular cell types and the amount of the encoded protein Conjugating plasm ids encode proteins that result in the

244 CHAPTER 6 • Genes, Genomics, and Chromosomes

'·

replication and transfer of the plasmid to related bacteria

through a complex macromolecular tube called a pilus. These distinguishes them from pscudogenes, which arose by sequence
plasmids, called R factors (for drug resistance), can contain drift of duplicated genes.
multiple drug resistance genes introduced by transposition o Mobile DNA elements most likely influenced evolution
and selected in environments where antibiotics are used to significantly by serving as recombination sites and by mobi-
sterilize surfaces, such as hospitals. These have led to the lizing adjacent DNA sequences.
rapid spread of resistance to multiple antibiotics between
pathogenic bacteria. Coping with the spread of R factors is a
major challenge for modern medicine. •

6.4 Organelle DNAs

Although the vast majority of DNA in most eukaryotes is
.· KEY CONCEPTS of Section 6.3 found in the nucleus, some DNA is present within the mito-
chondria of animals, plants, and fungi, and within the chlo-
Transposable (Mobile) DNA Elements roplasts of plants. These organelles are the main cellular
• Transposable DNA elements are moderately repeated se- sites for ATP formation, during oxidative phosphorylation
quences interspersed at multiple sites throughout the genomes in mitochondria and photosynthesis in chloroplasts (Chap-
of higher eukaryotes. They are present less frequently in pro- ter 12). Many lines of evidence indicate that mitochondria
karyotic genomes. and chloroplasts evolved from eubacteria that were engulfed
into ancestral cells containing a eukaryotic nucleus, forming
o DNA transposons move to new sites directly as DNA; ret-
endosymbiotes (Figure 6-20). Over evolutionary time, most
rotransposons are first transcribed into an RNA copy of the
of the bacterial genes were lost from organellar DNAs.
element, which then is reverse-transcribed into DNA (see
Some, such as genes encoding proteins involved in nucleo-
Figure 6-8).
tide, lipid, and amino acid biosynthesis, were lost because
• A common feature of all mobile elements is the presence their functions were provided by genes in the nucleus of the
of short direct repeats flanking the sequence. host cell. Other genes encoding components of the present-
o Enzymes encoded by transposons themselves catalyze in- day organelles were transferred to the nucleus. However,
sertion of these sequences at new sites in genomic DNA. mitochondria and chloroplasts in today's eukaryotes retain
DNAs encoding some proteins essential for organellar func-
• Although DNA transposons, similar in structure to bac-
tion, as well as the ribosomal and transfer RNAs required
terial IS elements, occur in eukaryotes (e.g., the Drosophila
for synthesis of these proteins. Thus eukaryotic cells have
P element), retrotransposons generally are much more abun-
multiple genetic systems: a predominant nuclear system and
dant, especially in vertebrates.
secondary systems with their own DNA, ribosomes, and
o LTR retrotransposons are flanked by long terminal re- tRNAs in mitochondria and chloroplasts.
peats (LTRs), similar to those in retroviral DNA; like retro-
viruses, they e11code reverse transcriptase and integrase.
They move in the genome by being transcribed into RNA, Mitochondria Contain Multiple
which then undergoes reverse transcription in the cytosol,
mtDNA Molecules
nuclear import of the resulting DNA with LTRs, and inte-
gration into a host-cell chromosome (see Figure 6-14). Individual mitochondria are large enough to be seen under
the light microscope, and even the mitochondrial DNA
• Non-LTR retrotransposons, including long interspersed
(mtDNA) can be detected by fluorescence microscopy. The
elements (LINEs) and short interspersed elements (SINEs),
mtDNA is located in the interior of the mitochondrion, the
lack LTRs and have an Aff-rich stretch at one end. They
region known as the matrix (sec Figure 12-6). As judged b}
arc thought to move by a nonviral retrotransposition
the number of yellow fluorescent "dots" of mtDNA, a Eu-
mechanism mediated by LINE-encoded proteins involving
glena gracilis cell contains at least 30 mtDNA molecules
priming of reverse transcription by chromosomal DNA (see
(Figure 6-21 ).
Figure 6-17).
Replication of mtDNA and division of the mitochondrial
• SINE sequences exhibit extensive homology with small cellu- network can be followed in living cells using time-lapse mi-
lar RNAs and are transcribed by the same RNA polymerase. croscopy. Such studies show that, in most organisms, mtDNA
Alu elements, the most common SINEs in humans, are =300-bp replicates throughout interphase. At mitosis, each daughter
sn1uences found scattered throughout the human genome. cell receives approximately the same number of mitochon-
• Some interspersed repeats are derived from cellular RNAs dria, but since there is no mechanism for appornoning ex-
that were reverse-transcribed and inserted into genomic actly equa l numbers of mitochondria to the daughter cells,
DNA at some time in evolutionary history. Processed pseu- some cells contain more mtDNA than others. By isolating
dogenes derived from mRNAs lack introns, a feature that mitochondria from cells and analyzing the DNA extracted
from them, it can be seen that each mitochondrion contains

6.4 Organelle DNAs 245

 ·.

Endocytosis of bacterium
_,.-- Endocytosis of bacterium
Ancestral
capable of oxidative cell capable of photosynthesis
pho~
sphorylation
ATP synthase ~ ATP/synthase Bacterial
Bacterial ""' ~ genome
genome ~ \ ./?, '

(~ ~ Chlo.opl"t
Mitochondrial Mitochondrial Stroma Thylakoid genome
matrix genome membrane
FIGURE 6-20 Endosymbiont hypothesis model of mitochondria portion of the protein once facing the extracellular space now faces the
and chloroplast evolution. Endocytosis of a bacterium by an ancestral intermembrane space. Budding of vesides from the inner chloroplast
eukaryotic cell would generate an organelle with two membranes, the membrane, such as occurs during development of chloroplasts in
outer membrane derived from the eukaryotic plasma membrane and contemporary plants, would generate the thylakoid membranes of
the inner one from the bacterial membrane. Proteins localized to the chloroplasts. The organellar DNAs are indicated.
ancestral bacterial membrane retain their orientation, such that the

multiple mtDNA molecules. Thus the total amount of mtDNA Is Inherited Cytoplasmically
mtDNA in a cell depends on the number of mitochondria,
Studies of mutants in yeasts and other single-celled organ-
the size of the mtDNA, and the number of mtDNA mole-
isms first indicated that mitochondria exhibit cytoplasmic
cules per mitochondrion. Each of these parameters varies
inheritance and thus must contain their own genetic system
greatly between different cell types.
(Figure 6-22) . For instance, petite yeast mutants exhibit
structurally abnormal mitochondria and are incapable of
oxidative phosphorylation. As a result, petite cells grow more
slowly than wild-type yeasts and form smaller colon ies. Ge-
netic crosses between d ifferent (haploid) yeast strains showed
that the petite mutation does not segregate with any known
nuclear gene or chromosome. In later studies, most petite
mutants were found to contain deletions of mtDNA.
In the mating by fusion of haploid yeast cells, both par-
ents contribute equally to the cytoplasm of the resulting dip-
loid; thus inheritance of mitochondria is bipa rental (see
Figure 6-22a ). In mammals and most other multicellular or-
ganisms, however, the sperm contributes little (if any) cyto-
plasm to the zygote, and virtually all the mitochondria in the
embryo are derived from those in the egg, not the sperm.
Studies in mice have shown that 99.99 percent of mtDNA is
maternally in herited, but a small part (0.01 percent) is inher-
ited from the male parent. In higher plants, mtDNA is inher-
ited exclusively in a unipa rental fashion through the fema le
parent (egg), not the male (pollen).

XP RIMENTAL F GURE 6-21 Dual staining reveals the The Size, Structure, and Coding Capacity of
multiple mitochondrial DNA molecules in a growing Euglena mtDNA Vary Considerably Between Organisms
gracilis cell. Cells were treated with a mixture of two dyes: ethidium
bromide, which binds to DNA and emits a red fluorescence, and DiOC6,
Surprisingly, the size of the mtDNA, the number and nature
which is incorporated specifically into mitochondria and emits a green of the proteins it encodes, and even the mitochondrial ge-
fluorescence. Thus the nucleus emits a red fluorescence, and areas rich netic code itself vary greatly between different organisms.
in mitochondrial DNA fluoresce yellow- a combination of red DNA and The mtDNAs of most mu lticellular an imals are = 16-kb cir-
green mitochondrial fluorescence. [From Y. Hayashi and K. Ueda, 1989, cular molecules that encode inrron-less genes compactly ar-
J. Cell Sci. 93:565.] ranged on both DNA strands. Vertebrate mtDNAs encode

2 46 CHAPTER 6 • Genes, Genomics, and Chromosomes

(a) Haploid parents with FIGURE 6-22 Cytoplasmic inheritance of an mt DNA petite
wild-type nuclea r genes mutation in yeast. Petite-strain mitochondria are defective in
Normal "Petite" oxidative phosphorylation owing to a deletion in mtDNA. (a) Haploid
mitochondrion mitochondrion cells fuse to produce a diploid cell that undergoes meiosis, during
which random segregation of parental chromosomes and mitochon-
dria containing mtDNA occurs. Note that alleles for genes in nuclear
DNA (represented by large and small nuclear chromosomes colored
red and blue) segregate 2:2 duri ng meiosis (see Figure 5-S). In contrast,
Mating by since yeast normally contain ~so mtDNA molecule~ per cell, all

1 cell fusion products of meiosis usually contain both normal and petite mtDNAs
and are capable of respiration. (b) As these haploid cells grow and
divide mitotically, the cytoplasm (including the mitochondria) is
randomly distributed to the daughter cells. Occasionally, a cell is
generated that contains only defective petite mtDNA and yields a
petite colony. Thus formation of such petite cells is independent of any
Diploid
nuclear genetic marker.
zygote

the two rRNAs found in mitochondrial ribosomes, the 22

Meiosis: random distribution
tRNAs used to translate mitochondrial mRNAs, and 13 pro-

1
of mitochondria to
daughter cells teins involved in electron transport and ATP synthesis
(Chapter 12). The smallest mitochondrial genomes known
arc in Plasmodium, single-celled obligate intracellular para-
sires that cause malaria in humans. Plasmodmm mtDNAs
are kb, encoding five proteins and· the mitochon-
drial rRNAs.
The mitochondrial genomes from a number of different
All haploid cells respiratory-proficient metazoan organisms (i.e., multicellular animals) have now
been c loned and sequenced, and mtDNAs from all these
sources encode essential mitochondrial proteins (Figure 6-23).
All proteins encoded by mtDNA are synthesized on mitO-
chondrial ribosomes. Most mitochondrially synthesized
polypeptides identified thus far are subunits of multimeric
complexes used in electron transport, ATP synthesis, or in-
sertion of proteins into the inner mitochondrial membrane
or intermembrane space. However, most of the proteins lo-
calized in mitochondria, such as those involved in the pro-
cesses listed at the top of Figure 6-23, are encoded by nuclear
genes, synthesized on cytosolic ribosomes, and imported inro
the organelle by processes discussed in Chapter 13.
In contrast to metazoan mtDNAs, plant mtDNAs arc
many times larger, and most of the DNA does not encode
protein. For instance, the mtDNA in the important model
plant Arabidopsis thaliana is 366,924 base pairs, and the
largest known mtDNA is :=:.2 Mb, found in cucurbit plants
(e.g., melon and cucumber). Most plant mtDNA consists of
long introns, pseudogenes, mobile DNA elements restricted
to the mitochondrial compartment, and pieces of foreign
(chloroplast, nuclear, and viral) DNA that were probably
inserted into plant mitochondrial genomes during their evo-
lution. Duplicated sequences also contribute to the greater
length of plant mtDNAs.
Differences in the number of genes encoded by the
mtDNA from various organisms most likely reflect the
movement of DNA between mitochondria and the nucleus
Respiratory-proficient Petite Respiratory· during evolution. Direct evidence for this movement comes
proficient from the observation that several proteins encoded by mtDNA

6.4 Organelle DNAs 247

 Lipid metabolism Carbohydrate metabolism Ubiquinone synthesis Chaperones
Nucleotide metabolism Heme synthesis Co-factor synthesis Signa ling pathways
Amino acid metabolism Fe-S synthesis Proteases DNA repair, replication, etc.

~ Heme
~ lyase

RNA
poly m erase

TIM '""''"""'({? .z4 ~ ~'\j ~"' '""''"""

Soc'""''"""~ • f{fffn ~ Tat translocase

Complex [g [@ 0 'W3
Complex
Complex
V
I Complex Complex
II Ill \
Cytochrome c
IV

FIGURE 6 - 23 Proteins encoded in mitochondrial DNA and their protein import and export, and insertion of protein-s into the inner
involvement in mitochondrial processes. Only the mitochondrial membrane (see Chapter 13). RNase Pis a ribozyme that processes the
matrix and inner membrane are depicted. Most mitochondrial 5' end of tRNAs (discussed in Chapter S).lt should be noted that the
components are encoded by the nucleus (blue); those highlighted in majority of eukaryotes have a mu ltisubunit Complex I as depicted, with
pink are encoded by mtDNA in some eukaryotes but by the nuclear three subunits invariantly encoded by mtDNA. However, in a few
genome in other eukaryotes, whereas a small portion are invariably organisms (Saccharomyces, Schizosaccharomyces, and Plasmodium).
specified by mtDNA (orange). Mitochondrial processes that have this complex is replaced by a nucleus-encoded, single-polypeptide
exclusively nucleus-encoded components are listed at the top. enzyme. For more details on mitochondrial metabolism and transport,
Complexes 1-V are involved in electron transport and oxidative see Chapters 12 and 13. [Adapted from G. Burger et al., 2003, Trends Genet.
.·
phosphorylation. TIM, Sec, Tat, and Oxa1 translocases are involved in 19:709.)

in some species are encoded by nuclear DNA in other, closely by which processed pseudogenes are generated in the nuclear
related species. The most striking example of this phenome- genome from nucleus-encoded mRNAs.
non involves the cox Il gene, which encodes subunit 2 of In addition to the large differences in the sizes of mtDNAs
cytochrome c oxidase, which constitutes complex IV in the in different eukaryotes, the structure of the mtDNA also var-
mitochondrial electron-transport chain (see Figure 12-16 ). ies greatly. As mentioned above, mtDNA in most animals is
This gene is found in mtDNA in all mu lticellular plants stud- a circular molecule = 16 kb. However, the mtDNA of many
ied except for certain re lated species of legumes, including organisms such as the protist Tetrahymena exists as linear
the mung bean and the soybean, in which the cox 1I gene is head-to-tail concatemers of repeating sequence. In the most
nuclear. The cox II gene is completely missing from mung extreme examples, the mtDNA of the protist Amoebidium
bean mtDNA, but a defective cox II pseudogene that has ac- parasiticum is composed of several hundred distinct short
cumulated many mutations can still be recognized in soy- linear molecules. And the mtDNA of Trypanosoma is com-
bean mtDKA. prised of multiple maxicircles concatenated (interlocked) to .,
Many RNA transcripts of plant mitochondrial genes are thousands of minicircles encoding guide RNAs involved in
edited, mainly by the enzyme-catalyzed conversion of se- editing the sequence of the mitochondrial mRNAs encoded
lected C residues to U, and occasionally U to C. (RNA edit- in the maxicircles.
ing is discussed in Chapter 8.) The nuclear cox II gene of
mung bean corresponds more closely to the edited cox ll
Products of Mitochondrial Genes
RNA transcripts than to the mitochondrial cox II genes
found in other legumes. These observations are strong evi- Are Not Exported
dence that the cox II gene moved from the mitochondrion to As far as is known, all RNA transcripts of mtDNA and their
the nucleus during mung bean evolution by a process that translation products remain in the mitochondrion in which
involved an RNA intermediate. Presumably this movement they are produced, and all mtDNA-encoded proteins are
involved a reverse-transcription mechanism similar to that synthesized on mitochondrial ribosomes. Mitochondrial

2 48 CHAPTER 6 • Genes, Genomics, and Chromosomes

DNA encodes the rR NAs that form mitochondrial ribo- evolving into an intracellular symbiote. The mtDNA with
somes, although most of the ribosomal proteins are imported the largest number of encoded genes so far found is in the
from the cytosol. In anima ls and fungi, all the tRNAs used protist species Reclinomonas americana. All other mtDNAs
for protein synthesis in mitochondria also are encoded by have a subset of the R. amencana genes, strongly implying
mtDNAs. However, in plants and many protozoans, most th at they evolved from a common ancestor with R. ameri-
mitochondrial tRNAs are encoded by the nuclear DNA and cana, losing different groups of mitochondrial genes by dele-
imported into the mitochondrion. tion and/or transfer to the nucleus over time.
In organisms whose mtDNA includes only a limited num-
B Reflecting the bacterial ancestry of mitochondria, mi- ber of genes, the same set of mitochondrial genes is retained,
H tochondrial ribosomes resemble bacterial ribosomes independent of the phyla that include these organisms (see
and differ from eukaryotic cytosolic ribosomes in their RNA Figure 6-23, orange proteins). One hypothesis for why these
and protein compositions, their size, and their sensitivity to genes were never successfully transferred to the nuclear ge-
certain antibiotics (see Figure 4-22). For instance, chloram- nome is that their encoded polypeptides are too hydrophobic
phenicol blocks protein synthesis by bacterial and mitochon- to cross the outer mitochondrial membrane, and therefore
drial ribosomes from most o rganisms, but cycloheximide, would not be imported back into the mitochondria if they
which inhibits protein synthesis on eukaryotic cytosolic ri- were synthesized in the cytosol. Similarly, the large size of
bosomes, does not affect mitochondrial ribosomes. This rRNAs may interfere with their transport from the nucleus
sensitivity of mitochondrial ri bosomes to the important through the cytosol into mitochondria. Alternatively, these
aminoglycoside class of antibiotics that includes chloram- genes may not have been transferred to the nucleus during
phenicol is the main cause of the toxicity that these antibiotics evolution because regulation of their expression in response
can cause. • to conditions within individual mitochondria may be advan-
tageous. If these genes were located in the nucleus, conditions
within each mitochondrion could not influence the expres-
Mitochondria Evolved from a Single sion of proteins found in that mitochondrion.
Endosymbiotic Event Involving
a Rickettsia-like Bacterium
Mitochondrial Genetic Codes Differ
Analysis of the mtDNA sequences from various eukaryotes,
including single-celled protists that diverged from other eu- from the Standard Nuclear Code
karyotes early in evolution, provides strong support for the The genetic code used in animal and fungal mitochondria is
idea that the mitochondrion had a single origin. Mitochon- different from the standard code used in all prokaryotic and
dria most likely arose from a bacterial symbiote whose clos- eukaryotic nuclear genes; remarkably, the code even differs
est con temporary relatives are in the Rickettsiaceae group. in mitochondria from different species (Table 6-3 ). Why and
Bacteria in this gro up are obligate intracellular parasites. how these differences arose during evolution is mysterious.
Thus, the ancestor of the mitochondrion probably also had UGA, for example, is normally a stop codon, but is read as
an intracellular lifestyle, putting it in a good location for tryptophan by human and fungal mitochondrial translation

1 1 1 Alterations in the Standard Genetic Code in Mitochondria

6H #1 Mitochondria
- --
Codon Standard Code* Mammals Drosophila Neurospora Yeasts Plants
--- -- --
UGA Stop Trp Trp Trp Trp Stop

AGA,AGG Arg Stop Ser Arg Arg Arg

AUA lie Met Met lie Met lie

AUU Ile Met \Iter Met Mer Ile

CUU,CUC,CUA,CUG Leu Leu Leu Leu Thr Leu

*For nuclear-encoded proteins.

sOURGS: S. Anderson er al., 1981, Nature 290:45~; P. Borst, m Jntematzonal Cell Bwlog)' 1980-1981, H. G. Schweiger, ed., Spnngcr-Verlag, p. 239;
C. Brettenberger and U. L. RaJ Bhandary, 19!!5, Trends Biochem. Sci. 10:4 7 8-483; V. K. Eckenrode and C. S. Levmgs, 1986, In Vztro Cell Det•. Bioi.
22:169-176; J. M. Gualber er al., 1989, Nature 341:660-662; and P. 5. Covello and ;VI. W. Gray, 1989, Nature 341:662-666.

6.4 Organelle DNAs 249

systems; however, in plant mitochondria, UGA is still recog- (a) Wi pe mouse Homozygous mutant
nized as a stop codon. AGA and AGG, the standard nuclear
codons for arginine, also code for arginine in fungal and
plant mtDNA, but they are stop codons in mammalian
mtDNA and serine codons in Drosophila mtDNA.

As shown in Table 6-3, plant mitochondria appear to

utilize the standard genetic code. However, compari-
sons of the amino acid sequences of plant mitochondrial pro-
tcim with the nucleotide sequences of plant mtDNAs
suggested that CGG could code for either arginine (the "stan- (b) 100
dard" amino acid) or tryptophan. This apparent nonspecific- 90 ~!!!1!!!!!!1!!!.-•,~~""'\-,~=:::;.--- Wild type
ity of the plant mitochondrial code is explained by editing of 80 ..__ _ Heterozygous
mitochondrial RNA transcripts, which can convert cytosine "'
> 70
residues to uracil residues. If a CGG sequence is edited to ·~
;;,
60
UGG, the codon specifies tryptophan, the standard amino "'c:
+'
50
acid for UGG, whereas unedited CGG codons encode the (1) 40
f:
standard arginine. Thus the translation system in plant mito- (1)
30

l
a..
chondria does utilize the standard genetic code. • 20
10
Mutations in Mitochondrial DNA Cause 0
0 100 200 300 400 500 600 700 800 900 1000
Several Genetic Diseases in Humans Age (days)

The severity of disease caused by a mutation in mtDNA de- (PERIMEII AL F GURE 6· 24 Mice with a mitochondrial DNA
pends on the nature of the mutation and on the proportion of polymerase defective for proofreading exhibit premature aging.
mutant and wild-type mtDNAs present in a particular cell A line of "knock-in" mice were prepared by methods discussed in
type. Generally, when mutations in mtDNA are found, cells Chapter 5 with an aspartic acid-to-alanine mutation in the gene
contain mixtures of wild-type and mutant mtDNAs-a condi- encoding mitochondrial DNA polymerase (D257A), inactivating the
tion known as heteroplasmy. Each time a mammalian somatic polymerase's proofreading function. (a) Wild-type and homozygous
or germ-line cell divides, the mutant and wild-type mtDr As mutant mice at 390 days old (13 months). The mutant mouse displays
many of the features of an aged mouse(> 720 days, or 24 months of
segregate randomly into the daughter cells, as occurs in yeast
age). (b) Plot of survival versus time of wild-type (+/+),heterozygous
cells (sec Figure 6-22b). Thus, the mtDNA genotype, which
(D257A/+) and homozygous (D257A/D257A) mice. [From G. C. Kujoth
fluctuates from one generation and from one cell division to et al., 2005, Science 309:481. Part (a) courtesy of Jeff Miller/University of
the next, can drift toward predominantly wild-type or predom- Wisconsin-Madison and Gregory KuJoth, Ph.D.]
inantly mutant mtDNAs. Since all enzymes required for the
replication and growth of mammalian mitochondria, such as
the mitochondrial DNA and RNA polymerases, are encoded in
the nucleus and imported from the cytosol, a mutant mtDNA
should not be at a "replication disadvantage"; mutants that requirement for ATP produced by oxidative phosphoryla-
involve large deletions of mtDNA might even be at a selective tion and tissues that require most or all of the mtDNA in the
advantage in replication, because they can replicate faster. cell to synthesize sufficient amounts of functional mitochon-
Recent research suggests that the accumulation of muta- drial proteins. For instance, Leber's hereditary optic neu-
tions in mtDNA is an important component of aging in mam- ropathy (degeneration of the optic nerve) is caused by a
mals. Mutations in mtDNA have been observed to accumulate missense mutation in the mtDNA gene encoding subunit 4 of
with aging, probably because mammalian mtDNA is not re- the NADH-CoQ reductase (complex 1), a protein required
paired in response to DNA damage. To study this hypothesis, for ATP production by mitochondria (see Figure 12-16).
researchers used gene "knock-in" techniques to replace the Any of several large deletions in mtDNA causes another set
nuclear gene encoding mitochondrial DNA polymerase with of diseases, including chronic progressive external ophthal-
normal proofreading activity (sec Figure 4-34) with a mutant moplegia, characterized by eye defects, and Kearns-Sayre
gene encoding a polymerase defective in proofreading. Muta- syndrome, characterized by eye defects, abnormal heartbeat,
tions in mtDNA accumulated much more rap1dly m homozy- and central nervous system degeneration. A third condition,
gous mutant mice than in wild-type mice, and the mutant causing "ragged" muscle fibers (with improperly assembled
mice aged at a highly accelerated rate (Figure 6-24). mitochondria) and associated uncontrolled jerky move-
ments, is due to a single mutation in the T'I'CG loop of the
With few exceptions, all human cells have mitochon- mitochondrial lysine tRNA. As a result of this mutation, the
dria, yet mutations in mtDNA affect only some tissues. translation of several mitochondrial proteins apparently is
Those most commonly affected are tissues that have a high inhibited. •

250 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Chloroplasts Contain Large DNAs Often plants. The large number of chloroplast DNA molecules per
Encoding More Than a Hundred Proteins cell permits the introduction of thousands of copies of an
engineered gene into each cell, resulting in extraordinarily
fn Like mitochondria, chloroplasts are thought to have high levels of foreign protein production. Chloroplast trans-
.a\t evolved from an ancestral endosymbiotic photosyn- formation has recently led to the engineering of plants that
thetic bacterium (see Figure 6-20). However, the endosymbi- are resistant to bacterial and fungal infections, drought, and
otic event giving rise to chloroplasts occurred more recently herbicides. The level of production of foreign proteins is
(1.2-1.5 billion years ago) than the event leading to the evo- comparable with that achieved with engineered bacteria,
lution of mitochondria ( 1.5-2.2 billion ye<Jrs ago). Conse- makmg it likely that chloroplast transformation will be used
quently, contemporary chloroplast DNAs show less structural for the production of human pharmaceuticals and possibl y
diversity than do mtDNAs. Also similar to mitochondria, for the engineering of food crops containing high levels of all
chloroplasts contain multiple copies of the organellar DNA the amino acids essential to humans. •
and ribosomes, which synthesize some chloroplast-encoded
proteins using the standard genetic code. Like plant mtDNA,
chloroplast DNA is inherited exclusively in a uniparental
fashion through the female parent (egg). Other chloroplast
proteins are encoded by nuclear genes, synthesized on cyto- KEY CONCEPTS of Section 6.4
solic ribosomes, and then incorporated into the organelle
(Chapter 13). • Organelle DNAs
• Mitochondria and ch loroplasts most likely evolved from
In higher plants, chloroplast DNAs are 120-160 kh bacteria that formed a symbiotic relationship with ancestral
long, depending on the species. They initially were thought cells containing a eukaryotic nucleus (see Figure 6-20).
to be circular DNA molecules because in genetically tracta- • Most of the genes originally within mitochondria and
ble organisms such as the model plant protozoan Chlam- chloroplasts were either lost because their functions were re-
ydomonas reinhardtii, the genetic map is circular. However, dundant with nuclear genes or moved to the nuclear genome
recent st udies have revealed that plant chloroplast DNAs over evolutionary time, leaving different gene sets in the or-
are actually long head-to-tail linear concatemers plus re- ganellar DNAs of different organisms (see Figure 6-23).
combination intermediates between these long linear mole-
cules. In these studies, researchers have used techniques that • Animal mtDNAs are circular molecules, reflecting their prob-
minimize mechanical breakage of long DNA molecules dur- able bacterial origin. Plant mtDNAs and chloroplast DNAs gen-
ing isolation and gel electrophoresis, permitting analysis of erally are longer than mtDNAs from other eukaryotes, largely
megabase-size DNA. because they contain more noncoding regions and repetitive
The complete sequences of severa l chloroplast DNAs sequences.
from higher plants have been determined. They contain • All mtDNAs and chloroplast DNAs encode rRNAs and
120-135 genes, 130 in the important model plant Arabidop- some of the proteins involved in mitochondrial or photosyn-
sis thaliana. A. thaliana chloroplast DNA encodes 76 protein- thetic electron transport and ATP synthesis. Most animal
coding genes and 54 genes with RNA products such as rRNAs mtDNAs and chloroplast DNAs also encode the tRNAs nec-
and tRNAs. Chloroplast DNAs encode the subunits of a essary to translate the organcllar mRNAs.
bacterial-like RNA polymerase and express many of their • Because most mtDNA is inherited from egg cells rather
genes from polycistronic operons as in bacteria (see Figure than sperm, mutations in mtDNA exhibit a maternal cyto-
4-13a). Some chloroplast genes contain introns, but these plasmic pattern of inheritance. Similarly, chloroplast DNA is
are similar to the specialized introns found in some bacterial exclusively inherited from the maternal parent.
genes and in mitochondricrl genes from fungi and protozo-
ans, rather than the introns of nuclear genes. As in the evolu- • Mitochondrial ribosomes resemble bacterial ribosomes in
tion of mitochondrial genomes, many genes in the ancestral their structure, sensitivity to chloramphenicol, and resistance
chloroplast endosymbiote that were redundant with nuclear to cycloheximide.
genes have been lost from chloroplast DNA. Also, many • The genetic code of animal and fungal mtDNAs differs
genes essential for chloroplast function have been trans- slightly from that of bacteria and the nuclear genome and
ferred to the nuclea r geno me of plants over evolu tionary varies among different animals and fungi (see Table 6-3). In
time. Recent estimates from sequence analysis of the A. contrast, plant mtDNAs and chloroplast DNAs appear to
thaliana and cyanobacterial genomes indicate that = 4500 conform to the standard genetic code.
genes have been transferred from the original endosymbiote • Several human neuromuscular disorders result from muta-
to the nuclear genome. tions in mtDNA. Patients generally have a mixture of wild-
type and mutant mtDNA in their cells (heteroplasmy): the
iilfl Methods similar to those used for the transformation higher the fraction of mutant mtDNA, the more severe is the
( ~ of yeast cells (Chapter 5) have been developed for sta- mutant phenotype.
bly introducing foreign DNA into the chloroplasts of higher

6.4 Organel le DNAs 251

6.5 Genomics: Genome-wide Analysis structure of the proteins. By comparing the amino acid se-
quence of the protein encoded by a newly cloned gene with
of Gene Structure and Expression the sequences of proteins of known function, an investigator
Using automated DNA sequencing techn iques and computer can look for sequence similarities t hat provide clues to the
algorithms to piece together the sequence data, researchers function of the encoded protein. Because of the degeneracy
have determined vast amounts of DNA sequence including in the genetic code, related proteins invariably exhibit more
nearly the entire genomic sequence of humans and many key sequence similarity than the genes encoding them. For this
experimental organisms. This enormous volume of data, which reason, protein sequences rather than the corresponding
is growing at a rapid pace, has been stored and organized in DNA sequences are usually compared.
two primary data banks: the GenBank at the National Insti- The most widely used computer program for this pur-
tutes of Health, Bethesda, Maryland, and the EMBL Sequence pose is known as BLAST (basic local alignment search tool).
Data Base at the European Molecular Biology Laboratory in The BLAST algorithm divides the "new" protein sequence
Heidelberg, Germany. These databases continuously exchange (known as the query sequence) into shorter segments and
newly reported sequences and make them available to scien- then searches the database for significant matches to any of
tists throughout the world on the Internet. By now, the ge- the stored sequences. The matching program assigns a high
nome sequences have been completely, or nearly completely, score to identically matched amino acids and a lower score
determmed for hundreds of viruses and bacteria, scores of ar- to matches between amino acids that are related (e.g., hydro-
chaea, yeasts (eukaryotes), plants including rice and maize, phobic, polar, positively charged, negatively charged) but
important model multicellular eukaryotes such as the round- not identical. W hen a significant match is found for a seg-
worm C. elegans, the fruit fly Drosophila melanogaster, mice, ment, the BLAST algorithm will search locally to extend the
humans, and representatives of the =35 metazoan phyla. The region of similarity. After searching is completed, the pro-
cost and speed of sequencing a megabase of DNA has fallen so gram ranks the matches between the query protein and vari-
low that the entire genome in cancer cells has been sequenced ous known proteins according to their p-values. This
and compared to the genome in normal cells from the same parameter is a measure of the probability of finding such a
patient in order to determine all the mutations that have ac- degree of similarity between two protein sequences by chance.
cumulated in that patient's tumor cells. This approach may The lower the p-value, the greater is the sequence similarity
reveal genes that are commonly mutated in all cancers, as we ll between two sequences. A p-value less t han about I 0 3 usu-
as genes that are commonly mutated in tumor cells from differ- ally is considered as significant evidence that two proteins
ent patients with the same type of cancer (e.g., breast versus share a common ancestor. Many alternative computer pro- ·.
colon cancer). This approach may eventually lead to highly grams have been developed in addition to BLAST that can
individualized cancer treatments tailored to the specific muta- detect relationships between proteins that are more distantly
tions m the tumor cells of a particular patient. The latest auto- related to each other than can be detected by BLAST. The
mated DNA sequencing techniq ues are so powerful that a development of such methods is currently an active area of
project known as the" l 000 Genomes Project" is currently un- bioinformatics research.
derway with the goal of sequencing most of the genomes of
1000-2000 randomly chosen individuals from all over the To illustrate the power of this approach, we consider
world in order to determine the extent of human genetic varia- the human gene NFJ. Mutations in NFJ are associated
tion as a basis for investigating the relationship between geno- with the inherited disease neurofibromatosis 1, in which mu l-
type and phenotype in humans. Moreover, privately owned tiple tumors develop in the peripheral nervous system, causing
companies have been founded that wi ll sequence much of an large protuberances in the skin. After a eDNA clone of NF1
individual's genome for = $1 00 in order to search for sequence was isolated and sequenced, the deduced sequence of the NFl
variations that may influence the probability of developing protein was checked against al l other protein sequences in
specific diseases. GenBank. A region of Nfl protein was discovered to have
In this section, we examine some of the ways researchers considerable homology to a portion of the yeast protein called
are mining this treasure trove of data to provide insights Ira (Figure 6-25 ). Previous studies had shown that Ira is a
about gene function and evolutionary relationships, to iden- GTPase-activating protein (GAP) that modulates the GTPase
tify new genes whose encoded proteins have never been iso- activity of the monomeric G protein called Ras (see Figure 3-32).
lated, and to determine when and where genes are expressed. As we examine in detail in Chapter 16, GAP and Ras proteins
This use of computers to analyze sequence data has led to normally function to control cell replication and differentia-
the emergence of a new field of biology: bioinformatics. tion in response to signals from neighboring cells. Functional
studies on the normal NF 1 protein, obtained by expression of
the cloned wild-type gene, showed that it did, indeed, regu late
Stored Sequences Suggest Functions of Newly
Ras activity, as suggested by its homo logy with Ira . These
Identified Genes and Proteins findings suggest that patients with neurofibromatosis express
As discussed in Chapter 3, proteins with similar functions a mutant NFl protein in cells of the peripheral nervous sys-
often contain similar amino acid sequences that correspond tem, leading to inappropriate cell division and formation of
to important functional domains in the three-dimensional the tumors characteristic of the disease. •

252 CHAPTER 6 • Genes , Genomics, and Chromosomes

 NF1 841 T R A T F M E V L T K I L Q Q G T E F 0 T L A E T V L A 0 R F E R L V E L V T M M G 0 Q G E L P I A 890
Ira 1500
•
I h I A f L R\ F I 0
• •
V . T
•
NY PVNP 1 K HEM
• •
KM L A I 0 0 F L
••• ~
•
Y I I KN P I
• A
• F 154o
891 MA LA NVVPCSQW O EL A RVLVTL FO S R HLLYO L LWNMFSK E VEL A OSMQTL ~0
CS PA 0 V
•• • • •• ~
•I •
K R ,:, A R S 0 0 I 1594
1547 G S _ ·, . L Y ·, G G F L N A - . T t NA SH I , V T E L L KQ
941 F R G N S L A S K I M T F C F K V Y G A T Y L Q K L L 0 P L L R I V I T S S 0 W Q H V S F E V 0 P T 990
• • • • • • • 0 N K E• •
1595 L H R 'J . C 1\ T R A L SLY T R S R • N K I T R ' V QG 1638

.. . . . . .
991 RLE P SESLE EN ORNL L OMT EK F . . . . FH A I I SS SSE FP PQ L RSV C HCL YQ 1ro6
639 K M K · G S t '. S E K M D L F r t-- Y M T R L I D , , T :; I D D I " I E ' V D I 1 K T I ' N 1685
. . .
1037 V V S Q R F P Q N S I G A V G S AM F L R F I N P A I V S P Y E A G I L 0 K K P P P R I E R G L K L 1086
• •
1686 A A . V N " E Y A Y I -~ '
•
' FV ~ '1 1 G~ · L
• 0 SEN
• • •
T H A H 0 R K P F I T 1734
1087 M S K I L Q S I AN . . . . . . . . H V L F T K E E H M R P F N 0 . . . . F V K S N F 0 A A R R F F 1124
•
1735 L A
••
V I
•
- L . '. G R E N I F K K 0 I
• VS 1
••
FL KTCS
•
K I F N c L S E L C K I P TN N c 1784
1125 L 0 I AS 0 C PT S 0 A V N H S L . . . . . . . . . . . . . S F I S 0 G NV LA L H R L L W N N . 1159
• •
1785 T V N V R E 0 P ~ P I S F 0 Y F L H K F F Y L N E F T R I( E I
•• • • •
N E S K L P G E F S F _ K , T V 1834
1160 . . Q E K I G Q Y L S S N R 0 H K A V G R R P F . . . . 0 K MAT L LAY L G P P E H K P V A 1200
1835 M L N 0 "
• •
L GV , GQ P SME I KN E I P ~ 1 V V EN R E
• k
•• •
Y P S I. Y E F M S R Y A F • K 0 1882

FIGURE 6 · 25 Comparison ofthe regions of human NF1 protein but nonidentical side chains are connected by a blue dot. Amino acid
and 5. cerevisiae Ira protein that show significant sequence numbers in the protein sequences are shown at the left and right ends
similarity. The NFl ana the Ira sequences are shown on the top and of each row. Black dots indicate "gaps" in the protein sequence
bottom lines of each row, respectively, in the one-letter amino acid inserted in order to maximize the alignment of homologous amino
code (see Figure 2-14). Amino acids that are identical in the two acids. The BLAST p-value for these two sequences is 10 28, indicating a
proteins are highlighted in yellow. Amino acids with chemically similar high degree of similarity. (From G. Xu et al., 1990, Ce//62:599.]

Even when a protein shows no significant similarity to All the different members of the tubulin family of genes
other proteins with the BLAST algorithm, it may neverthe- (or proteins) are sufficiently similar in sequence to suggest a
less share a short sequence that is functionally important. common ancestral sequence. Thus all these sequences are
Such short segments recurring in many different proteins, considered to be homologous. More specifically, sequences
referred to as structural motifs, generally have similar func- that presumably diverged as a result of gene duplication
tions. Several such motifs are described in Chapter 3 and il- (e.g., the a- and f3-tubulin sequences) are described as paral-
lustrated in Figure 3-9. To search for these and other motifs ogous. Sequences that arose because of speciation (e.g., the
'·' in a new protein, researchers compare the query protein se- a-tubulin genes in different species) are described as orthol-
quence with a database of known motif sequences. ogous. From the degree of sequence relatedness of the tubu-
lins present in different organisms today, evolutionary
relationships can be deduced, as illustrated in Figure 6-26b.
Comparison of Related Sequences from Different Of the three types of sequence relationships, orthologous
Species Can Give Clues to Evolutionary sequences are the most likely to share the same function.
Relationships Among Proteins
BLAST searches for related protein sequences may reveal that
Genes Can Be Identified Within Genomic
proteins belong to a protein family. Earlier, we considered
gene families in a single organism, using the f3-globin genes in DNA Sequences
humans as an example (see Figure 6-4a). But in a database that The complete genomic sequence of an organism contains within
includes the genome sequences of multiple organisms, protein it the information needed ro deduce the sequence of every pro-
families also can be recognized as being shared among related tein made by the cells of that organism. For organisms such as
organisms. Consider, for example, the tubulin proteins; these bacteria and yeast, whose genomes have few introns and short
are the basic subunits of microtubules, which are important intergenic regions, most protein-coding sequences can be found
components of the cytoskeleton (Chapter 18). According to simply by scanning the genomic sequence for open reading
the simplified scheme in Figure 6-26a, the earliest eukaryotic frames (ORFs) of significant length. An ORF usually is defined
cells are thought to have contained a single tubulin gene that as a stretch of DNA Lontaining at least 100 codons that begins
was duplicated early in evolution; subsequent divergence of the with a start codon and ends with a stop codon. Because the
different copies of the original tubulin gene formed the ances- probability that a random DNA sequence will contain no stop
tral versions of the a- and f3-tubulin genes. As different species codons for 100 codons in a row is very small, most ORFs
diverged from these early cukaryotic cells, each of these gene encode protein.
sequences further diverged, giving rise to the slightly different ORF analysis correctly identifies more than 90 percent of
forms of a-tubulin and f3-tubulin now found in each species. the genes in yeast and bacteria. Some of the very shortest genes,

6.5 Genomics: Genome-wide Analysis of Gene Structure and Expression 253

 Ancestral~
(a) (b) Orthologous
..______

cell~
a-Tubulin (human) -
a-Tubulin (fly)
Gene duplication

1 and divergence m)-f-

st)

Orthologous
Ancestral
tubulin ~-Tubu lin (human)-

~-Tubulin (fly)

m)-

st)

Species 1 Species 2
FIGURE 6 -26 Generation of diverse tubulin sequences during sequences diverged. For example, node 1 represents the duplication
the evolution of eukaryotes. (a) Probable mechanism giving rise to event that gave rise to the a-tubulin and ~-tubulin families, and node 2
the tubulin genes found in existing species. It is possible to deduce represents the divergence of yeast from multicellular species. Braces
that a gene duplication event occurred before speciation because the and arrows indicate, respectively, the orthologous tubulin genes,
a-tubulin sequences from different species (e.g., humans and yeast) which differ as a result of speciation, and the paralogous genes, which
are more alike than are the a-tubulin and ~-tubulin sequences within a differ as a result of gene duplication. This diagram is simplified
species. (b) A phylogenetic tree representing the relationship between somewhat because flies, worms, and humans actually contain multiple
the tubulin sequences. The branch points (nodes), indicated by small a-tubulin and ~-tubulin genes that arose from later gene duplication
numbers, represent common ancestral genes at the time that two events.

however, are missed by this method, and occasionally long have most genes in common, although largely nonfunctional
open reading frames that are not actually genes arise by chance. DNA sequences, such as intergenic regions and introns, will
Both types of mis-assignments can be corrected by more sophis- tend to be very different because these sequences are not
ticated analysis of the sequence and by genetic tests for gene under strong selective pressure. Thus corresponding seg-
function. Of the Saccharomyces genes identified in this manner, ments of the human and mouse genome that exhibit high
about half were. already known by some functional criterion sequence similarity are likely to be functionally important:
such as mutant phenotype. The functions of some of the pro- exons, transcription-control regions, or sequences with other
teins encoded by the remaining putative (suspected) genes iden- functions that are not yet understood.
tified by ORF ana lysis have been assigned based on their
sequence similarity to known proteins in other organisms. The Number of Protein-Coding Genes
Identification of genes in organisms with a more complex
in an Organism's Genome Is Not Directly
genome structure requires more sophisticated algorithms than
searching for open reading frames. Because most genes in Related to Its Biological Complexity
higher eukaryotes are composed of multiple, relatively short The combination of genomic sequencing and gene-finding
exons separated by often quite long noncoding introns, scan- computer algorithms has yielded the complete inventory of
ning for ORFs is a poor method for finding genes. The best protein-coJing genes for a variety of organisms. Figure 6-27
gene-finding algorithms combine all the available data that shows the total number of protein-coding genes in several
might suggest the presence of a gene at a particular genomic eukaryotic genomes that have been completely sequenced.
site. Relevant data include alignment or hybridization of the The functions of about half the proteins encoded in these
query sequence to a full-length eDNA; al ignment to a partial genomes are known or have been predicted on the basis of
eDNA sequence, generally 200-400 bp in length, known as sequence comparisons. One of the surprising features of this
an ex[nessed sequence tag (EST); fitting to models for exon, comparison is that the number of protein-coding genes within
intron, and splice site sequences; and sequence similarity to difft:n:nt organisms does not seem proportional ro our intui-
other organisms. Using these computer-based bioinformatic tive sense of their biological complexity_ For example, the
methods, computational biologists have identified approxi- roundworm C. elegans apparently has more genes than the
mately '"" 19,800 protein coding genes in the human genome. fruit fly Drosophila, which has a much more complex body
A particularly powerful method for identifying human plan and more complex behavior. And humans have fewer
genes is to compare the human genomic sequence with that than one and one-half the number of genes as C. elegans.
of the mouse. Humans and mice are sufficiently related to When it first became apparent that humans have fewer than

254 CHAPTER 6 • Genes, Genomics, and Chromosomes

 Organism Human Arabidopsis (plant) C. elegans (roundworm)

Genes -25,000 25,706 18,266

0 Metabolism

[I DNA replication/modification

• Transcription/translation

• Intracellular signaling

Organism Drosophila (fly) Saccharomyces (yeast) 0 Cell-cell communication

Genes 13,338 -6000
lliJ Protein folding and degradation

• Transport

FIGURE 6-27 Comparison of the number and types of proteins • Multifunctional proteins
encoded in the genomes of different eukaryotes. For each organism,
the area of the entire pie chart represents the total number of
protein-coding genes, all shown at roughly the same scale. In most
L2J Cytoskeleton/structure

cases, the functions of the proteins encoded by about half the genes D Defense and immunity
are still unknown (light b lue). The functions of the remainder are
known or have been predicted by sequence similarity to genes of [2'J Miscellaneous function
known function. [Adapted from International Human Genome Sequencing
Consortium, 2001 , Nature 409:860.]
D Unknown

twice the number of protein-coding genes as the simple round- replication and gene expression, leading to increasing com-
worm, it was difficult to understand how such a small increase plexity of embryological development.
in the number of proteins could generate such a staggering The specific functions of many genes and proteins identi-
difference in complexity. fied by analysis of genomic sequences still have not been
Clearly, simple quantitative differences in the number of determined. As researchers unravel the functions of individ-
genes in the genomes of different organisms are inadequate ual proteins in different organisms and further detail their
for explaining differences in biological complexity. How- interactions with other proteins, the resulting advances will
ever, several phenomena ean generate more complexity in become immediately applicable to all homologous proteins
the expressed proteins of higher eukaryotes than is predicted in other organisms. When the function of every protein is
from their genomes. First, alternative splicing of a pre- known, no doubt, a more sophisticated understanding of the
mRNA can yield multiple functional mRNAs corresponding molecular basis of complex biological systems will emerge.
to a particular gene (Chapter 8). Second, variations in the
post-trans lational modification of some proteins may pro-
duce functional differences. Finally, increased biological KEY CONCEPTS of Section 6.5
complexity results from increased numbers of cells built of
the same kinds of proteins. Larger numbers of cells can inter- Genomics: Genome-wide Analysis of Gene Structure
act in more complex combinations, as in comparing the ce- and Expression
rebral cortex from mouse to man. Similar cells are present in The function of a protein that has not been isolated (a query
both the mouse and human cerebral cortex, but in humans protein) often can be predicted on the basis of similarity of its
more of them make more complex connections. Evolution of amino acid sequence to the sequences of proteins of known
the increasing biological complexity of multicellular organ- function.
isms likely required increasingly complex regulation of cell

6.5 Genomics: Genome -wide Analysis of Gene Structure and Expression 255
 phase cells (those that are not undergoing mitosis). Further
• A computer algorithm known as BLAST rapidly searches folding and compaction of chromatin during mitosis pro-
databases of known protein sequences to find those with sig- duces the visible metaphase chromosomes, whose morphol-
nificant similarity to a query protein. ogy and staining characteristics were detailed by early
Proteins with common functional motifs, which often can cytogeneticists. Although every eukaryotic chromosome in-
be quite short, may not be identified in a typical BLAST search. cludes millions of individual protein molecules, each chro-
Such short sequences may be located by searches of motif mosome contains just one, extremely long, linear DNA
databases. molecule. The longest DNA molecules in human chromo-
somes, for instance, are 2.8 X 108 base pairs, or almost 10
A protein family comprises multiple proteins all derived
em, in length! The structural organization of chromatin al-
from the same ancestral protein. The genes encoding these
lows this vast length of DNA to be compacted into the mi-
proteins, which constitute the corresponding gene family,
croscopic constraints of a cell nucleus (see Figure 6-1 ). Yet
arose by an initial gene duplication event and subsequent
chromatin is organized in such a way that specific DNA se-
divergence during speciation (see Figure 6-26).
quences within the chromatin are readily available for cellular
Related genes and their encoded proteins that derive from processes such as the transcription, replication, repair, and
a gene duplication event are paralogous, such as the a- and recombination of DNA molecules. In this section, we con-
[3-globins that combine in hemoglobin (a 2 [32 ); those that de- sider the properties of chromatin' and irs organization into
rive from mutations that accumulated during speciation are chromosomes. Important features of chromosomes in their
orrhologous. Proteins that are orrhologous usually have a entirety are covered in the next section.
similar function in different organisms, such as the mouse
and human adult [3-globins.
Chromatin Exists in Extended
Open reading frames (ORFs) are regions of genomic DNA
and Condensed Forms
containing at least 100 codons located between a start codon
and stop codon. When the DNA from eukaryotic nuclei is isolated using a
method that preserves native protein-DNA interactions, it is
Computer search of the entire bacterial and yeast genomic
associated with an equal mass of protein in the nucleopro-
sequences for open reading frames (ORFs) correctly identifies
tein complex known as chromatin. Histones, the most abun-
most protein-coding genes. Several types of additional data must
dant proteins in chromatin, constitute a family of small,
be used to identify probable (putative) genes in the genomic se-
basic proteins. The five major types of histone proteins-
quences of humans and other higher eukaryotes because of their
termed Hl, H2A, H2B, HJ, and H4-are rich in positively
more complex gene structure, in which relatively short coding
charged basic amino acids, which interact with the nega-
exons are separated by relatively long, noncoding introns.
tively charged phosphate groups in DNA.
• Analysis of the complete genome sequences for several dif- When chromatin is extracted from nuclei and examined
ferent organisn1s indicates that biological complexity is not in the electron microscope, its appearance depends on the salt
directly related to the number of protein-coding genes (see concentration to which it is exposed. At low salt concentra-
Figure 6-27). tion in the absence of divalent cations such as Mg + l, isolated
chromatin resembles "beads on a string" (Figure 6-28a). In
this extended form, the string is composed of free DNA called
"linker" DNA connecting beadlike structures termed nucleo-
6.6 Structural Organization somes. Composed of DNA and histones, nucleosomes are
about 10 nm in diameter and are the primary structural units
of Eukaryotic Chromosomes
of chromatin. If chromatin is isolated at physiological salt
Now that we have examined the various types of DNA se- concentration, it assumes a more condensed fiberlike form
quences found in eukaryotic genomes and how they are or- that is 30 nm in diameter (figure 6-28b).
ganized within it, we turn to the question of how DNA
molecules as a whole are organized within eukaryotic cells. Structure of Nucleosomes The DNA component of nucleo-
Because the total length of cellular DNA is up to a hundred somes is much less susceptible to nuclease digestion than is
thousand times a cell's diameter, the packing of DNA is cru- the linker DNA between them. If nuclease treatment is care-
cial to cell architecture. It is also essential to prevent the long fully controlled, all the linker DNA can be digested, releasing
DNA molecules from getting knotted or tangled with each individual nucleosomes with their DNA component. A nu-
other during cell division, when they must be precisely segre- cleosome consists of a pruLciu t.:urc with DNA wound around
gated to daughter cells. The task of compacting and organiz- its surface like thread around a spool. The core is an octamer
ing chromosomal DNA is performed by abundant nuclear containing two copies each of histones H2A, H2B, H3, and
proteins called histones. The complex of histones and DNA H4. X-ray crystallography has shown that the octameric his-
is called chromatin. tone core is a roughly disk-shaped structure made of inter-
Chromatin, which is about half DNA and half protein by locking histone subunits (Figure 6-29). Nucleosomes from all
mass, is dispersed throughout much of the nucleus in inter- eukaryotes contain =147 base pairs of DNA wrapped one

256 CHAPTER 6 • Genes, Genomics, and Chromosomes

 (a} (b) EXPERIMENTAL FIGURE 6-28 The extended
and condensed forms of extracted chromatin
have very different appearances in electron
micrographs. (a} Chromatin isolated in low-ionic-
strength buffer has an extended "beads-on-a-
string" appearance. The "beads" are nucleosomes
(10-nm diameter) and the "string" is connecting
(linker) DNA. (b) Chromatin isolated in buffer with
a physiolog ical ionic strength (0.1 5 M KCI) appe;m
as a condensed fiber 30 nm in diameter.
[Part (a) courtesy of S. McKnight and 0. Miller, Jr. Part
(b) courtesy of B. Hamkalo and J. B. Rattner.]

and two-thirds turns around the protein core. The length of nucleosomes assembled from recombinant histones, indicates
the linker DNA is more variable among species, and even that the 30-nm fiber has a "zig-zag ribbon" structure that is
between different cells of one organism, ranging from about wound into a "two-start" helix made from two "strands" of
10 to 90 base pairs. During cell replication, DNA is assem- nucleosomes stacked on top of each other like coins. The two
bled into nucleosomes shortly after the replication fork passes "strands" of stacked nucleosomes are then wound into a
(see Figure 4-33). T his process depends on specific histone double helix similarly to the two strands in a DNA double
chaperones that bind to histones and assemble them together helix, except that the helix is left handed, rather than right
with newly repl icated DNA into nucleosomes. handed as it is in DNA (Figure 6-30). The 30-nm fibers also
include Hl, the fifth major histone. Hl is bound to the DNA
Structure of the 30-nm Fiber When extracted from cells in as it enters and exits the nuclcosome core, but its structure in
isotonic buffe rs (i.e., buffers with the same salt concentration the 30-nm fiber is not known at atomic resolution.
found in cells, ""'0.15 M KCl, 0.004 M MgCI2 ), most chro- The chromatin in chromosomal regions that are not being
matin appears as fibers ""'30 nm in diamete r (see Figure transcribed or replicated exists predominantly in the con-
6-28b). Current research, including x-ray crystallography of densed, 30-nm fiber form and in higher-order folded structures

·. FIGURE 6-29 Structure of the nucleosome based on x-ray are green. TheN-terminal tails of the eight hi stones and the H2A and
crystallography. (a) Nucleosome w ith space-filling model of the H2B (-terminal tails, involved in condensation of the chromatin, are
hist ones. The sugar-phosphate backbones of the DNA strands are not visible beca use they are d isord ered in the crystal. (b) Space-filling
represented as w h ite tu bes to allow better visualization of the hi stones. model of hi stones and DNA (white) viewed from the side of the
Nucleosom e shown from the top (left) and from the sid e (right, rotated nucleosome. [Parts (a) and (b) after K. Luger et al., 1997, Nature 389:251.]
clockwise 90°). H2A su b units are yellow; H2Bs are red; H3s are blue; H4s

6.6 Structural Organization of Eukaryotic Chromosomes 257

whose detailed conformation is not currently understood. The
regions of chromatin actively being transcribed are thought to
assume the extended beads-on-a-string form.

Conservation of Chromatin St ructure The general structure

of chromatin is remarkably similar in the cells of all eukary- Chain of
otes, including fungi, plants, and animals, indicating that the nucleosomes

l
structure of chromatin was optimized early in the evolution of
eukaryotic cells. The amino acid sequences for four histones
(H2A, H2B, H3, and H4) are highly conserved between dis-
tantly related species. For example, the sequences of histone
l !3 from sea urchin tissue and calf thymus differ by only a
single amino acid, and H3 from the garden pea and calf thy-
mus differ only in four amino acids. Apparently, significant
deviations from the histone amino acid sequences were se-
lected against strongly during evolution. The amino acid se-
quence of Hl, however, varies more from organism to
organism than do the sequences of the other major histones.
The similarity in sequence among histones from all eukaryotes
suggests that they fold into very similar three-dimensional con-
formations, which were optimized for histone function early in
L
evolution in a common ancestor of all modern eukaryotes.
Minor histone variants encoded by genes that differ from
the highly conserved major types also exist, particularly in
vertebrates. For example, a special form of H2A, designated
H2AX, is incorporated into nucleosomes in place of H2A in
a small fraction of nucleosomes in all regions of chromatin.
At sites of DNA double-stranded breaks in chromosomal Two-start
DNA, H2AX becomes phosphorylated and participates in helix
the chromosome-repair process, probably by functioning as
a binding site for repair proteins. In the nucleosomes at cen-
tromeres, H3 is replaced by another variant histone called
CENP-A, which participates in the binding of spindle micro-
tubules during mitosis. Most minor histone variants differ
only slightly in sequence from the major histones. These
slight changes in histone sequence may influence the stability
(a) (b)
of the nuclcosome as well as its tendency to fold into the
30-nm fiber and other higher-order structures. FIGURE 6·30 St ructure of the 30-nm chromatin fiber. (a) Model
for the folding of a nucleosomal chain at top into a "zig-zag ribbon" of
nucleosomes that then folds into a two-start helix at bottom. For
simplicity, DNA is not represented in the two-start helix. (b) Model of
Modifications of Histone Tails Control Chromatin the 30-nm fiber based on x·ray crystallography of a tetranucleosome (a
Condensation and Function short stretch of four nucleosomes). [Part (a) adapted from C. L. F. Woodcock
Each of the histone proteins making up the nucleosome core et al., 1984, J. Cell Bioi. 99:42. Part (b) from T. Schalch et al., 2005, Noture 436:138.)
contains a flexible N-terminus of 19-39 residues extending
from the globular structure of the nucleosome; the H2A and
H2B proteins also contain a flexible C-terminus extending from Histone tails are suhject to multiple post-translational
the globular histone octamer core. These termini, called histone modifications such as acetylation, methylation, phosphory-
tails, are represented in the model shown in Figure 6-31a. The lation, and ubiquitination. Figure 6-31 b summarizes the
histone tails are required for chromatin to condense from the types of post-translational modifications observed in human
beads-on-a-string conformation into the 30 nm fiber. For histones. A particular hi:,luiH:: pwrc::in never has all of these
example, recent experiments indicate that the N-terminal modifications simultaneously, but the histones in a single nu-
tails of histone H4, particularly lysine16, are critical for cleosome usually contain several of these modifications simul-
forming the 30-nm fiber. This positively charged lysine inter- taneously. The particular combinations of post-transcriptional
acts with a negative patch at the H2A-H2B interface of the modifications found in different regions of chromatin have
next nucleosome in the stacked nucleosomes of the 30-nm been suggested to constitute a histone code that influences
fiber (see Figure 6-30). chromatin function by creating or removing binding sires for

258 CHAPTER 6 • Genes, Genomics, and Chromosomes

·.
 (a) (b)
~ AcAcAc J!I
SGRGKQGCKARAKAKTRSS ' H2A ~ VLLPKKTESH H KAKG K
120
·.· 1 5 9 13
119

Ac Ac ~c A• Ub
PEPAKSAPAPKKGSKKAVTKA AVSEGTKAVTKYTSSK
5 12 1415 20 120
Me
H4 I I

~ A·~ ~c{f) R~EIAQDF~TDLRFO

MeiMe MeM9~ Ac MeAc Ac MeM'f-1 Me
ARTKQTARKSTGGKAPRKQLATKAARKSAPATGGVKKPH
23 4 8 J91011 14 1718 23 261128 36
9 21 H3
i.7
I
I
H2B' ~
I <f> ~e Ac A• Ac Ac <f) ~e ( H4
I .... ~ H2A
H2A' , • SGRGKGGKGLG KGGAKRHRKVLR DNIQG IT_>.
•..
I
1 3 5 8 12 16 18 20
...... H2B' H2B

'
Ac Phosphorylation

M~~ Me Methylation
'--VKKKARKSAGAAK
261 Ac Acetylation
26
Ub Ubiquitination

FIGURE 6 -31 Post-translational modifications observed on (b) Summary of post-translational modifications observed in human
human histones. (a) Model of a nucleosome viewed from the top with histones. Histone-tail sequences are shown in the on~- letter amino acid
histones shown as ribbon diagrams. This model depicts the lengths code (see Figure 2-14). The main portion of each histone is depicted
of the histone tails (dotted lines), which are not visible in the crystal as an oval. These modifications do not all occur simultaneously on a
structure (see Figure 6-29). The H2A N-terminal tails are at the bottom, single histone molecule. Rather, specific combinations of a few these
and the H2A ( -terminal tails are at the top. The H2B N-terminal tails modifications are observed on any one histone. [Part (a) from K. Luger
are on the right and leh, and (-terminal tails are at the bottom center. and T. J. Richmond, 1998, Curr. Opin. Genet. Devel. 8:140. Part (b) adapted from
Histones H3 and H4 have short (-terminal tails that are not modified. R. Margueron et al., 2005, Curr. Opin. Genet. Oevel. 15: 163.]

chromatin-associated proteins dependent on the specific from chromatin protein, digested to completion with a re-
combinations of these modifications present. Here we de- striction enzyme, and analyzed by Southern blotting. An in-
scribe the most abundant kinds of modifications found in tact gene treated with a restriction enzyme yields fragments
histone tails and how these modifications control chromatin of characteristic sizes. If isolated nuclei are exposed first to
condensation and function. We end with a discussion of a DNase, the gene may be cleaved at random sites within the
special case of chromatin condensation, the inactivation of boundaries of the restriction enzyme cut sites. Consequently,
X chromosomes in female mammals. any Southern blot bands normally seen with that gene will
be lost. This method was first used to show that the tran-
Histone Acetylation Histone-taillysines undergo reversible scriptionally inactive ~-globin gene in nonerythroid cells,
acetylation and deacetylati~n by enzymes that act on specific where it is associated with relatively unacetylatcd histones, is
lysines in the N-termini. In the acetylated form, the positive much more resistant to DNase I than is the active, tran-
charge of the lysine £-amino group is neutralized. As men- scribed ~-globin gene in erythroid precursor cells, where it is
tioned above, lysine 16 in histone H4 is particularly important associated with acetylated histones (Figure 6-32). These re-
for the folding of the 30-nm fiber because it interacts with a sults indicate that the chromatin structure of nontranscribed
negatively charged patch on the surface of the neighboring DNA in hypoacetylated chromatin makes the DNA less ac-
nucleosome in the fiber. Consequently, when H4 lysine 16 is cessible to the small DNase I enzyme (=10 kD) than it is in
acetylated, the chromatin tends to form the less condensed transcribed, hyperacetylated chromatin. This is thought to
"beads-on-a-string" conformation conducive for transcrip- be because chromatin cont::tining the repressed gene is folded
tion and replication. into condensed structures that sterically inhibit access of the
Histone acetylation at other sites in H4 and in other his- associated DNA to the nuclease. In contrast, the transcribed
tones (see Figure 6-31a) is correlated with increased sensitivity gene is associated with a more unfolded form of chromatin
of chromatin DNA to digestion by nucleases. This phenome- that allows better access of the nuclease to the associated
non can be demonstrated by digesting isolated nuclei with DNA. Presumably, the condensed chromatin structure in non-
DNase l. Following digestion, the DNA is completely separated erythroid cells also sterically inhibits access of the promoter

6.6 Structural Organization of Eukaryotic Chromosomes 259

(a) Decondensed chromatin Condensed ch romatin in histones, are required for the full activation of transcrip-
tion of a number of genes. These enzymes are now known to
have other substrates that influence gene expression in addi-
tion to histones. Consequently, they are more ge nerally
known as nuclear lysine acetyl transferases, or KATs because
4.6 kb 4.6 kb K represents lysine in the single-letter code for amino acids
,--- ---"---------- (Figure 2-14 ). Conversely, early genetic studies in yeast indi-
[ Globin I I Globin I cated that complete repression of many yeast genes requires
i
BamHI
\\1,-A
DNase BamHI
i BamHI
i \ltt
DNase BamHI
the action of histone deacetylases (HDACs) that remove ace-
tyl groups of acetylated lysines from histone tails, as dis-
cussed further in Chapter 7.
14-day erythroblast MSB

Other Histone Modifications As shown in Figure 6-31 b, his-

(b) DNA from tone tails in chromatin can undergo a variety of other cova-
14-day DNA
lent modifications at specific amino acids. Lysine £-amino
erythroblasts from
,---------A---------~ MSB groups can be methylated, a process that prevents acetyla-
tion, thus maintaining their posit'Jve charge. Moreover, the
DNase (~tg/ ml) 0 .01 .05 .1 .5 .1 1.5 1.5
N of lysine £-amino groups can be methylated once, twice,
or three times. Arginine side chains can also be methylated.
The 0 in hydroxyl groups (-OH) of serine and threonine
. .. side chains can be reversibly phosphorylated, introducing

• •·:
--. • • -·.... . -.:,- - 4.6 kb
two negative charges. Each of these post-translational modi-
fications contributes to the binding of chro,matin-associated
.... ·. . proteins that participate in the control of chromatin folding

..... .• -...
,. . ... -.. and the ability of DNA and RNA polymerases to replicate or
transcribe the associated DNA. Finally, a single 76-amino-
acid ubiquitin molecule can be reversibly added to a lysine in
the C-terminal tails of H2A and H2B. Recall that addition of
multiple linked ubiquitin molecules to a protein can mark it
for degradation by the proteasome (see Figure 3-29b). In this
case, however, the addition of a single ubiquitin molecule
EXPERIMENTAL FIGURE 6-32 Nontranscribed genes are less does not affect the stability of a histone, although it does
susceptible to DNase I digestion than active genes. Chick embryo influence chromatin structure.
erythroblasts at 14. days actively synthesize globin, whereas undifferen-
As mentioned previously, it is the precise combination of
tiated chicken lymphoblastic leukemia (MSB) cells do not. (a) Nuclei
modified amino acids in histone tails that helps control the
from each type of cell were isolated and exposed to increasing
condensation, or compaction, of chromatin and its ability to
concentrations of DNase I. The nuclear DNA was then extracted and
treated with the restriction enzyme BamHI, which cleaves the DNA
be transcribed, replicated, and repaired. This can be ob-
around the globin sequence and normally releases a 4.6-kb globin served by electron microscopy and light microscopy using
fragment. (b) The DNase 1- and BamHI-digested DNA was subjected dyes that bind DNA. Condensed regions of chromatin known
to Southern blot analysis with a probe of labeled cloned adult globin as heterochromatin stain much more darkly than less con-
DNA, which hybridizes to the 4.6-kb Bam HI fragment. If the globin densed chromatin, known as euchromatin (Figure 6-33a).
gene is susceptible to the initial DNase digestion, it would be cleaved Heterochromatin does not fully decondense following mito-
repeatedly and would not be expected to show this fragment. As seen sis, remaining in a compacted state during inte rphase and
in the Southern blot, the transcriptionally active DNA from the 14-day usually associating with the nuclear envelope, nucleoli, and
globin-synthesizing cells was sensitive to DNase I digestion, indicated additional distinct foci. Heterochromatin includes centro-
by the absence of the 4.6-kb band at higher nuclease concentrations. meres and telomeres of chromosomes, as well as transcrip-
In contrast, the inactive DNA from MSB cells was resistant to digestion. tionally inactive genes. In contrast, areas of euchromatin,
These results suggest that the inactive DNA is in a more condensed
which are in a less compacted state during interphase, stain
form of chromatin in which the globin gene is shielded from DNase
lightly with DNA dyes. Most transcribed regions of DNA
digestion. [See J. Stalder et al., 1980, Ce// 19:973; photograph courtesy of
arc found in euchromatin. Heterochromatin usually con-
H. Weintraub.]
taim histone H3 modified by methylation of lysine 9 or 27,
while euchromatin generally contains histone H3 exten-
and other transcription control sequences in DNA to the pro- sively acetylated on lysine 9 and 14, and to a lesser extent at
teins involved in transcription, contributing to transcriptional other H3 lysines, methylation of lysine 4, and phosphoryla-
repression (Chapter 7). tion of serine I 0 (Figure 6-33b). Other histone tails are also
Genetic stud ies in yeast indicated that histone acetyl specifically modified in euchromatin versus heterochroma-
transferases (HATs), which acetylate specific lysine residues tin. For example, H4 lysine 16 is generally unacetylated in

260 CHAPTER 6 • Genes, Genomics, and Chromosomes

 to DNase 1 digesuon in isolated nuclei (see Figure 6-32). Higher
eukaryotes express a number of proteins containing a so-
called chromodomain that binds to histone tails when they
are methylated at specific lysines. One example ts hetero-
chromatin protein 1 (H P1). In addition to hi stones, HPl is
one of the major proteins associated with heterochromatin.
The HPl chromodomain binds the H3 N-terminal tail only
when it is tri-methylated at lysine 9 (see Figure 6-33b). HPl
also contains a secuml domain called a chromoshadow do-
main because it is frequently found in proteins that contain
a chromodomain. The chromoshadow domain binds to
other chromoshadow domains. Consequently, chromatin
containing H3 tri-methylated at lysine 9 (H3K9Me 3 ) is as-
sembled into a condensed chromatin structure by HPl, al-
though the structure of this chromatin is not well understood
(Figure 6-34a).
In addition to binding to itself, the chromoshadow do-
main also binds the enzyme that methylates H3 lysine 9, an
11-Lm
H3K9 histone methyl transferase (HMT). As a consequence,
nucleosomes adjacent to a region of HP !-containing hetero-
chromatin also become methylated at lysine 9 (Figure 6-34b).
(b) Heterochromatin (inactive/ condensed) This creates a binding site for another HPl that can bind the
Me H3K9 histone methyl transferase, resulting in "spreading"
H3 ARTKQTARKSTGGKAPRKQLATKAARKSAPAT of the heterochromatin structure along the chromosome
9
until a boundary element is encountered th~t blocks further
Me. spreading. Boundary elements so far characterized are gener-
H3 ARTKOTARKSTGGKAPRKQLATKAARKSAPAT ally regions in chromatin where several nonhistone proteins
27
bind to DNA, possibly blocking histone methylation on the
other side of the boundary.
Euchromatin (active/open) Significantly, the model of heterochromatin formation in
Me. Acf11l Ac Ac Ac Figure 6-34b provides an explanation for how heterochro-
I T I
H3 ARTKOTARKSTGGKAPRKOLATKAARKSAPAT matic regions of a chromosome are reestablished following
4 910 14 18 27
DNA replication during the S phase of the cell cycle. When
FIGURE 6-33 Heterochromatin versus euchromatin. (a) In this DNA in heterochromatin is replicated, the histone octamers
electron micrograph of a bone marrow stem cell, the dark-staining that are tri-methylated at H3 lysine 9 become distributed to
areas in the nucleus (N) outside the nucleolus (n) are heterochromatin.
both daughter chromosomes along with an equal number of
The light-staining, whitish areas are euchromatin. (b) The modifications
newly assembled histone octamers. The H3K9 histone
of histone N-terminal tails in heterochromatin and euchromatin differ,
methyl transferase associated with the H3K9 tri-methylated
as illustrated here for h istone H3. Note in particular that histone tails
are generally much more extensively acetyl ated in euchromatin
nucleosomes methylate lysine 9 of the newly assembled nu-
compared with heterochromatin. Heterochromatin is much more
cleosomes, regenerating the heterochromatin in both daugh-
condensed (thus less accessible to proteins) and is much less transcrip- ter chromosomes. Consequently, heterochromatin is marked
tionally active than is euchromatin. [Part (a) P. C. Cross and K. L. Mercer, with an epigenetic code, so called because it does not depend
1993, Cell and nssue Ultrastructure, W. H. Freeman and Company, p. 165. Part on the sequence of bases in DNA that maintains the repres-
(b) adapted from T. Jenuwein and C. D. Allis, 2001, Science 293:1074.] sion of associated genes in replicated daughter cells.
Other protein domains associate with histone-tail modi-
fications typical of euchromatin. For example, the bromodo-
main binds to acety lated histone tails and therefore is
associated with transcriptionally active chromatin. Several
heteroch romatin, allowing it to interact with neighboring proteins involved in stimulating gene transcription contain
nucleosomes and stabilize chromatin folding into the 30-nm bromodomains, such as the largest subunit of TFIID (see
fiber (figure 6-30). Chapter 7). This transcription factor contains two closely
spaced bromodomains that probably help TFIID to associate
Reading the Histone Code The histone code of modified with transcriptionally active chromatin (i.e., euchromatin).
amino acids in the histone tails is "read" by proteins that This protein and other bromodomain-containing proteins
bind to the modified tails and in turn promote condensation also have histone acetylase activity, which helps to maintain
or decondensation of chromatin, forming "closed" or the chromatin in a hyperacetylated state conducive to tran-
"open" chromatin structures, as judged by their sensitivity scription. Consequently, an epigenetic code associated with

6.6 Structural Organization of Eukaryotic Chromosomes 261

(a) euch romatin helps to maintain the transcriptional activity of
genes in euch romatin through successive cell divisions. These
epigenetic codes for heterochromatin and euchromatin help
to maintain the patterns of gene expression established in

l Histone H3K9
methyl transferase
different cell types during early embryonic development as
specific differentiated cell types increase in numbers by cell
division. Importantly, abnorma l alterations in these epigen-
etic codes have been found to contribute to the pathogenic
replication and beha v10r of cancer cells (Chapter 24).
In summary, multiple types of covalent modifications of
histone tails can influence chromatin structure by altering
nucleo~ome-nucleosome interactions and interactions with

l Binding of HP1
chromodomain to H3K9Me3
additional proteins that participate in or regu late processes
such as transcription and DNA replication. The mechanisms
and molecular processes governing ch romatin mod ifications
that regu late transcription are discussed in greater detail in
the next chapter.

X-Chromosome Inactivation in Mammalian Females One

important example of epigenetic gene control through re-
pression by heterochromatin is the random inactivation and

! HP1 oligomerization
condensation of one of the two X chromosomes in female
mammals. Each female mammal has two X chromosomes,
one contributed by the egg from which it developed (X 111 )
and one contributed by the sperm (Xp) . Early during embry- .·
onic development, random inactivation of either the Xm or
the x, chromosome occurs in each somatic cell. In the female
embryo, about half the cells have an inactive X,, and the
other half have an inactive X, . All subsequent daughter cells
maintain the same inactive X chromosomes as thei r parent
cells. As a result, the adult fema le is a mosaic of clones, some
expressing the genes from the X, and the rest expressing the
genes from the X,. This inactivation of one X chromosome
Heterochromatin in female mammals results in dosage compmsaticm, the pro-
cess that ensures that cells of females express the same level
of proteins encoded on the X chromosome as the cells of
males, which have only one X chromosome.
Histones associated with the inactive X chromosome have
post-translational modifications characteristic of other regions
of heterochromatin: hypoacetylation of lysines, di- and tri- ,·.
methylation of histone H3 lysine 9, tri-methylation of H3
lysine 27, and a lack of methylation at histone H3 lysine 4
(see Figure 6-33b). X-chromosome inactivation at an early
stage in embryonic development is controlled by the X-inac-
tivation center, a complex locus on the X chromosome that
determines which of the two X chromosomes will be inacti-
vated and in which cells. The X-inactivation center also con-
FIGURE 6-34 Model for the formation of heterochromatin by
tains the Xist gene, which encodes a remarkable, long,
the binding of HP1 to histone H3 trimethylated at lysine 9. (a) HPl
non-protein-coding RNA that coats only the X chromosome
contributes to the condensation of heterochromatin by binding to
it was transcribed from, thereby triggering silencing of the
histone H3 N-terminal tails tri-methylated at ly~inP 9, followed by
association of the histone-bound HPl. (b) Heterochromatin condensa-
chromosome.
tion can spread along a chromosome because HPl binds a histone
Although the mechanism of X-chromosome inactivation
methyltransferase (HMn that methylates lysine 9 of histone H3. This is not fully understood, it involves several processes includ-
creates a binding site for HPl on the neighboring nucleosome. The ing the action of Polycomb protein complexes that are dis-
spreading process continues until a "boundary element " is encoun- cussed further in Chapter 7. One subunit of the Polycomb
tered. [Part (a) adapted from G. Thiel et al., 2004, Eur. J. Biochem. 271 :2855. complex contains a chromodomain that binds to histone H3
Part (b) adapted from A. J. Bannister et al., 2001 , Noture 41 0:120.) tails when they are tri-methylated at lysine 27. The Polycomb

262 CHAPTER 6 • Genes, Genomics, and Chromosomes

complex also contains a histone methyl transferase specific
. F
for H3 lysine 27. This finding helps to explain how the X-
inactivation process spreads along large regions of the X
chromosome and how it is maintained through DNA replica-
E. .G /Loop of
30-nm
tion, similar to heterochromatization by the binding of HPl D. chromatin
fiber
to histone H3 tails methylated at lysine 9 (see Figure 6-34b).
X-chromosome inactivation is another example of an
A •• B ct
.H
epigenetic process, that is, a process that affects the expres-
sion of specific genes and is inherited by daughter cells but is
II II
not the result of a change in DNA sequence. Instead, the
activity of genes on the X chromosome in female mammals
is controlled by chromatin structure rather than the nucleo-
tide sequence of the underlying DNA. And the inactivated X
chromosome (either X,. or Xp) is maintained as the inactive
chromosome in the progeny of all future cell divisions be- EXPERIMENTAL FIGURE 6-35 Fluorescent-labeled probes
cause the histones are modified in a specific, repressing man- hybridized to interphase chromosomes demonstrate chromatin
ner that is faithfully inherited through each cell division. loops and permit their measurement. In situ hybridization of
mterphase cells was carried out with several different probes specific
for sequences separated by known distances in linear, cloned DNA.
Nonhistone Proteins Organize Long Lettered circles represent probes. Measurement of the distances
Chromatin Loops between different hybridized probes, which could be distinguished by
their color, showed that some sequences (e.g., A and B), separated from
Although histones are the predominant proteins in chroma- one another by millions of base pairs, appear located near one another
tin, less abundant, nonhistone chromatin-associated pro- within nuclei. For some sets of sequences, the measured distances in
teins, and the DNA molecule itself, are also crucial to nuclei between one probe (e.g., C) and sequences successively farther
chromosome structure. Recent results indicate that it is not away initially appear to increase (e.g., D, E, and F) and then appear to
protein alone that gives a metaphase chromosome its struc- decrease (e.g., G and H). [Adapted from H. Yokota et al., 1995, J. Cell Bioi.
ture. Micromechanical studies of large metaphase chromo- 130:1239.]
somes from newts in the presence of proteases or nucleases
indicate that DNA, not protein, is responsible for the me-
chanical integrity of a metaphase chromosome when it is
of the nuclear and cytoplasmic material required for division
pulled from its ends. These results are inconsistent with a
of the fertilized egg into the thousands of differentiated cells
continuous protein scaffold at the chromosome axis. Rather,
required to generate a feeding embryo that can ingest addi-
the integrity of chromosome structure requires the complete
tional nutrients.
chromatin complex of DNA, histone octamers, and nonhis-
In general, SARs/MARs are found between transcription
tone chromatin-associated proteins.
units, and genes are located primarily within the chromatin
In situ hybr~dization experiments with several different
loops. As discussed below, the loops are tethered at their bases
fluorescent-labeled probes to the DNA of one chromosome
by a mechanism that does not break the duplex DNA mole-
in human interphase cells support a model in which chroma-
cule that extends the entire length of the chromosome. Evi-
tin is arranged in large loops. In these experiments, some
dence indicates that SARs/MARs may insulate neighboring
probe sequences separated by millions of base pairs in linear
genes. Some SARs/MARs function as insulators, that is, DNA
DNA appeared reproducibly very close to one another in
sequences of tens to hundreds of base pairs that separate tran-
interphase nuclei from different cells of the same type (Fig-
scription units from each other. Proteins regulating transcrip-
ure 6-35). These closely spaced probe sites are postulated to
tion of one gene cannot influence the transcription of a
lie close to regions of chromatin, called scaffold-associated
neighboring gene that is separated from it by an insulator.
regions (SA Rs) or matrix-attachment regions (MARs), lo-
cated at the bases of the DNA loops. SARs/MARs have been
mapped by digesting histone-depleted chromosomes with re- Ringlike Structure of SMC Protein Complexes The bases of
striction enzymes and then recovering the fragments that re- chromatin loops (see Figure 6-35) in interphase chromosomes
main associated with the digested histone-depleted may be held in place by proteins called ~tructural mainte-
preparation. The measured distances between probes are nance of {;.hromosome proteins, or SMC proteins. These
consistent with chromatin loops ranging in size from 1 mil- nonhistone proteins are critical for maintaining the morpho-
lion to 4 million base pairs in mammalian interphase cells. logical structure of condensed chromosomes during mitosis.
Loops of chromatin are also directly visualized by light mi- In extracts prepared from the large nuclei of Xenopus laevus
croscopy in the acti ve chromatin of growing amphibian oo- (African frog) eggs, chromosomes can be induced to con -
cytes ("lampbrush chromosomes," shown in the figure at the dense as they do in intact cells as they enter the prophase
beginning of Chapter 8). These cells are enormous compared period of mitosis. This condensation fails to occur when one
to most cells(= 1 mm in diameter) because they stockpile all type of SMC protein is depleted from the extract with specific

6.6 Structural Organization of Eukaryotic Chromosomes 263

(a) Hinge (b) Chromatin fibers The hinge domain of one monomer (blue) binds to the hinge
domain domain of a second monomer (red), forming a rough!} U-
shaped dimeric complex. The head domains of the monomers
Coi led-coil have ATPase activity and are linked by members of another
domain small protein family called kleisins. The overall SMC com-
plex is a ring with a diameter large enough to accommodate
Head two 30-nm chromatin fibers (Figure 6-36b), and is capable of
domain linking two circular DNA molecu les in vitro. SMC proteins
arc proposed to form the base of chromatin loops by formmg
Kleisin topologically constrained knots in 30-nm chromatin fibers,
as diagrammed in Figure 6-36c. This can explain why cleav-
age of the DNA at a relatively small number of sites leads to
(c) rapid dissolution of condensed metaphase chromosome
Chromatin loop containing
a transcription unit structure, whereas protease cleavage of proteins has only a
minor effect on chromo~ome structure until most of the pro-
tein is digested: When the DNA is cut anywhere in a long
region of chromatin containing soveral chromatin loops, the

(a)

FIGURE 6- 36 Model of SMC compl exes bound to chromatin.

(a) Model of an SMC protein complex. (b) Model of SMC complex
topologically linking two chromatin fibers represented by cylinders
with the diameter of a nucleosome relative to the dimensions of the
SMC complex. (c) Model for the binding of SMC complexes to the
base of a loop of transcribed chromatin. (Adapted from K. Nasmyth and
C. H. Haering, 2005, Ann. Rev. Btochem. 74:595.]

antibodies. Yeast with mutations in certain SMC proteins

fail to properly assoctate daughter chromatids following
D A replication in the S phase. As a result, chromosomes
do not properly segregate to daughter cells during mitosis.
Related SMC proteins are required for proper segregation of
chromosomes in bacteria and archaea, indicating that this is
an ancient class of proteins vital to chromosome structure
and segregation in all kingdoms of life. (b)
Each SMC monomer contains a hinge region where the
polypeptide folds back on itself, forming a very long, coiled
coil region and bringing theN- and C-termini together so they
can interact to form a globular head domain (Figure 6-36a).

XPERIME 'AL FIGURE 6-37 During interphase, human

chromosomes remain in nonoverlapping territories in the nucleus.
Fixed interphase human fibroblasts were hybridized in situ to fluores-
cently labeled probes specific for sequences along the full length of
human chromosomes 7 (cyan) and 8 (purple). DNA is stained blue with
DAPI. In the diploid cell, each of the two chromosome 7s and two
chromosome 8s is restricted tu d territory or domain within the nucleus,
rather than stretching throughout the entire nucleus. (b) Similar to
(a) except that chromosome paint probes specific for each chromosome
were hybridized to reveal the location of nearly all of the chromosomes
in a fibroblast from a human male. Some of the chromosomes are not
observed in this confocal slice through the nucleus. [Part (a) courtesy of
Drs. I. Solovei and T. Cremer. Part (b) from A. Bolzer et al., 2005, PLOS Bio/ 3:826.]

264 CHAPTER 6 • Genes, Genomics, and Chromosomes

 broken ends can slip through the SMC protein rings, "unty-
ing" the topological knots that constrain the loops of chro-
matin. In contrast, most of the individual rings of SMC
proteins must be broken before the topological constraints
holding the base of the loops together is released.

Interphase Chromosome Territories In the small nuclei of

most cells, individual interphase chromosomes, which are less
condensed than metaphase chromosomes, cannot be resolved
by standard microscopy or electron microscopy. Nonetheless,
the chromatin of one chromosome in interphase cells is not
spread throughout the nucleus. Rather, interphase chromatin
is organized into chromosome territories. As illustrated in Fig-
ure 6-37, in situ hybridization of interphase nuclei with chro-
mosome-specific fluorescent-labeled probes shows that the
probes are visualized within restricted regions of the nucleus
rather than appearing throughout the nucleus. Use of probes
specific for different chromosomes shows that there is little
overlap between chromosomes in interphase nuclei. However,
the precise positions of chromosomes are not reproducible
between cells.

Metaphase Chromosome Structure Condensation of chro-

mosomes during prophase may involve the formation of many
more loops of chromatin, so that the length of each loop is
greatly reduced compared to that in interphase cells. How-
ever, the folding of chromatin in metaphase chromosomes is
not well understood. Microscopic analysis of mammalian
chromosomes as they condense during prophase indicates that
the 30-nm fiber folds into a 100- to 130-nm fiber called a

- - - - - 30nm FIGURE 6 -39 Typical metaphase chromosome. As seen in this

scanning electron micrograph, the chromosome has replicated and
B ----- 100- 130 nm comprises two chromatids, each containing one of two identical DNA
(chromonema fiber) molecules. The centromere, where chromatids are attached at a
constriction, is required for their separation late in mitosis. Special
.·
telomere sequences at the ends function in preventing chromosome
shortening. [Andrew Syred/Photo Researchers, Inc.]
cl- - - - - 200-250 nm
(middle prophase chromatid)
chromonema fiber. As depicted in Figure 6-38, a chromonema
fiber then folds into a structure with a diameter of 200-250 nm
called a middle prophase chromatid, which then folds into the
500- to 750-nm-diamcter chromatids observed during meta-
phase. Ultimately, the full lengths of two associated daughter
chromosomes generated by DNA replication during the previ-
- - 500-750 nm ous S phase of the cell cycle condense into bar-shaped struc-
(metaphase chromatid) tures that in most eukaryotes are linked at the central
constriction called the centromere (Figure 6-39).

Additional Nonhistone Proteins Regulate

Transcription and Replication
FIGURE 6 -38 Model for the folding of the 30-nm chromatin fiber The total mass of the histones associated with DNA in chro-
in a metaphase chromosome. A single chromatid of a metaphase matin is about equal to that of the DNA. Interphase chroma-
chromosome is depicted. [Adapted from N. Kireeva et al., 2004, J. Cell Bioi. tin and metaphase chromosomes also contain small amounts
166:775.] of a complex set of other proteins. For instance, thousands

6.6 Structural Organization of Eukaryotic Chromosomes 265

of different transcription factors are associated with interphase
chromatin. The structure and function of these critical non- moshadow domain of HPl also associates with itself and
histone proteins, which regulate transcription, are examined with the histone methyl transferase rha t methylates H3 ly-
in Chapter 7. Other low-abundance nonhistone pr oteins as- sine 9. These interactions cause condensation of the 30-nm
sociated with chromatin regulate DNA replication during chromatin fiber and spreading of the heterochromatic struc-
the eukaryotic cell cycle (Chapter 20). ture along the chromosome until a boundary element is en-
A few other nonhistone DNA-binding proteins are present countered (see Figure 6-34).
in much larger amounts than the transcription or replication • One X chromosome in nearly every cell of mammalian
factors. Some of these exhibit high mobility during electro- females is highly condensed heterochromatin, resulting in
phoretic separation and thus have been designated HMG repression of expression of nearly all genes on the inactive
(high-mobility group) proteins. When genes encoding the chromosome. This inactivation results in dosage compensa-
most abundant HMG proteins are deleted from yeast cells, tion so that genes on the X chromosome are expressed at the
normal transcription is disturbed in most genes examined. same level in both males and females.
Some HMG proteins have been found to assist in the coopera-
Each eukaryotic chromosome contains a single DNA mol-
tive binding of several transcription factors to specific DNA
ecule packaged into nucleosomes and folded into a 30-nm
sequences that are close to each other, stabilizing multi protein
chromatin fiber, which is associated with a protein scaffold
complexes that regulate transcription of a neighboring gene,
made up in part of structural maintenance of chromosome
as discussed in Chapter 7.
(SMC) proteins at sires between transcription units (see Fig-
ure 6-36c). Additional folding of the scaffold further com-
pacts the structure into the highly condensed form of meta-
KEY CONCEPTS of Section 6.6
phase chromosomes (see Figure 6-38).
Structural Organization of Eukaryotic Chromosomes
In eukaryotic cells, DNA is associated with about an equal
mass of histone proteins in a highly condensed nucleoprotein ·.
complex called chromatin. The building block of chromatin
is the nucleosome, consisting of a histone octamer around 6.7 Morphology and Functional Elements
which is wrapped 147 bp of DNA (see Figure 6-29). of Eukaryotic Chromosomes
The chromatin in transcriptionally inactive regions of DNA Having examined the detailed structural organization of
within cells is thought to exist in a condensed, 30-nm fiber chromosomes in the previous section, we now view them
form and higher-order structures built from it (see Figure 6-30 from a more global perspective. Early microscopic observa-
and 6-38). tions on the number and size of chromosomes and their stain-
The chromatin in transcriptionally active regions of DNA ing patterns led to the discovery of many important general
within cells is thought to exist in an open, extended form. characteristics of chromosome structure. Researchers subse-
quently identified specific chromosomal regions critical to
• The histone H4 rails, particularly H4lysine 16, are required
their replication and segregation to daughter cells during cell
for beads-on-a-string chromatin (the 10-nm chromatin fiber)
division. In this section we discuss these functional elements
to fold into a 30-nm fiber.
of chromosomes and consider how chromosomes evolved
Histone tails can be modified by acetylation, methylation, through rare rearrangements of ancestral chromosomes.
phosphorylation, and monoubiquitinarion (sec Figure 6-31 ).
These modifications inAuence chromatin structure by regu-
lating the binding of histone tails to other, less abundant Chromosome Number, Size, and Shape
chromatin-associated proteins.
at Metaphase Are Species-Specific
The reversible acetylation and deacerylation of lysine resi-
As noted previously, in nondividing cells, individual chromo-
dues in theN-termini of the core histones regulates chroma-
somes are not visible, even with the aid of histologic stains for
tin condensation. Proteins involved in transcription, replica-
DNA (e.g., Feulgen or Giemsa stains) or electron microscopy.
tion, and repair, and enzymes such as DNasel can more
During mitosis and meiosis, however, the chromosomes con-
easily access chromatin with hyperacetylated hisrone tails
dense and become visible in the light microscope. Therefore,
(euchromatin) than chromatin with hypoacetylated histone
almost all cytogenetic work (i.e., studies of chromosome mor-
tails (heterochromatin). phology) has been done with condensed metaphase chromo-
When metaphase chromosomes decondense during inter- somes obtained from dividing cells- either somatic cells in
phase, areas of heterochromatin remain much more con- mitosis or dividing gametes during meiosis.
densed than regions of euchromatin. The condensation of metaphase chromosomes probably
Heterochromatin protein 1 (HPJ) uses a chromodomain results from several orders of folding of 30-nm chromatin
to bind to histone H3 tri-methylated on lysine 9. The chro- fibers (see Figure 6-38). At the time of mitosis, cells have
already progressed through the S phase of the cell cycle and

266 CHAPTER 6 • Genes, Genomics, and Chromosomes

have replicated their DNA. Consequently, the chromosomes Today, the method of chromosome painting greatly sim-
that become visible during metaphase are dupltcated struc- plifies differentiating chromosomes of similar size and shape.
tures. Each metaphase chromosome consists of two sister This technique, a variation of fluorescence in situ hybridiza-
chromatids, which are linked at a constricted region, the tion (FISH), makes use of probes specific for sites scattered
centromere (see Figure 6-39 ). The number, sizes, and shapes along the length of each chromosome. The probes are la-
of the metaphase chromosomes constitute the karyotype, beled with several different fluorescent dyes with distinct
which is distinctive for each species. In most organisms, all excitation and emission wavelengths. Probes specific for
somatic cells have the same karyotype. However, species each chromosome are labeled with a predetermined fraction
that appear quite similar can have very different karyotvpes, of each of the dyes. A frer the probes arc hybridized to chro
tndicating that similar genetic potential can be organized on mosomes and the excess removed, the sample is observed
chromosomes in very different ways. For example, two spe- with a fluorescence microscope in which a detector deter-
cies of small deer-the Indian mumjac and Reeves muntjac- mines the fraction of each dye present at each fluorescing
contain about the same total amount of genomic DNA. In position in the microscopic field. This information is con-
one species, this DNA is organized into 22 pairs of homolo- veyed to a computer, and a special program assigns a false-
gous autosomes and two physically separate sex chromo- color image to each type of chromosome (Figure 6-40, left ).
somes. In contrast, the other species contains the smallest Computer graphics allows the two homologs of each chro-
number of chromosomes of any mammal, only three pairs of mosome to be placed next to each other and named accord-
autosomes; one sex chromosome is physically separate, but ing to their decreasing size. Such a display clearly displays
the other is joined to the end of one autosome. the cell's karyotype. Figure 6-40 shows a normal human
male karyotype. Chromosomal painting is a powerful
During Metaphase, Chromosomes Can method for detecting an abnormal number of chromosomes,
such as chromosome 21 trisomy in patients with Down syn-
Be Distinguished by Banding Patterns
drome, or chromosomal translocations that occur in rare
and Chromosome Painting individuals and in cancer cells (Figure 6-41 ). The use of
Certain dyes selectively stain some regions of metaphase probes with different ratios of fluorescent dyes that hybrid-
chromosomes more intensely than other regions, producing ize to distinct positions along each normal" human chromo-
characteristic banding patterns that are specific for individ- some allows finer structure analysis of the chromosomes that
ual chromosomes. The regularity of chromosomal bands can more readily reveal deletions or duplications of chromo-
serves as useful visible landmarks along the length of each somal regions. The figure at the beginning of the chapter
chromosome and can help to distinguish chromosomes of illustrates the use of such multicolor FISH in analysis of the
similar size and shape. karyotype of a normal human female.

(a) (b)

EXPERIMENTAL FIGURE 6··40 Human chromosomes are using chromosome paint probes. (b) Alignment of these painted
readily identified by chromosome painting. (a) Fluorescence in situ chromosomes by computer graphics to reveal the normal human male
hybridization (FISH) of human chromosomes from a male cell in mitosis karyotype. [Courtesy of M. R. Speicher.]

6.7 Morphology and Functional Elements of Eukaryotic Chromosomes 267

(a) same chromosome paint probes were hybridized to human
metaphase chromosomes, most of the probes hybridized to
Philadelphia the long arm of chromosome 10 (Figure 6-42b). Further,

w
chromosome
when multiple probes from the long arm of human chromo-
some 10 with different fluorescent dye labels were hybridized
to human chromosome 10 and tree shrew metaphase chromo-
somes, tree shrew sequences homologous to each of these
der (22) probes were found along tree shrew chromosome 16 in the
same order that they occur un human chromosome 10.
22 These results indicate that during the evolution of hu-
mans and tree shrews from a common ancestor that lived
=85 million years ago, a long, continuous DNA sequence on
one of the ancestral chromosomes became chromosome 16
in tree shrews, but evolved into the long arm of chromosome
9 10 in humans. The phenomenon of genes occurring in the
same order on a chromosome in two different species is re-
der (9) ferred to as conserved synteny (d~rived from Latin for "on
the same ribbon"). The presence of two or more genes in a
(b)
common chromosomal region in two or more species indi-
cates a conserved syntenic segment.
The relationships between the chromosomes of many pri-
mates have been determined by cross-species hybridizations of
chromosome paint probes as shown for human and tree
shrew in Figure 6-42a, b. From these re'lationships and
higher-resolution analyses of regions of synteny by DNA se-
quencing and other methods, it has been possible to propose
the karyotype of the common ancestor of all primates based
on the minimum number of chromosomal rearrangements
necessary to generate the regions of synteny in chromosomes
of contemporary primates.
Human chromosomes are thought to have derived from a
common primate ancestor with 23 autosomes plus the X and
Y sex chromosomes by several different mechanisms (Fig-
ure 6-42c). Some human chromosomes were derived without
large-scale rearrangements of chromosome structure. Others
·41 Chromosomal translocations
are thought to have evolved by breakage of an ancestral chro-
can be analyzed using chromosome paint probes for FISH.
mosome into two chromosomes or, conversely, by fusion of
Characteristic chromosomal translocations are associated with certain two ancestral chromosomes. Still other human chromosomes
genetic disorders and specific types of cancers. For example, in nearly appear to have been generated by exchanges of parts of the
all patients with chronic myelogenous leukemia, the leukemic cells arms of distinct chromosomes, that is, by reciproca l transloca-
contain the Philadelphia chromosome, a shortened chromosome 22 tion involving two ancestral chromosomes. Analysis of re-
[der (22)], and an abnormally long chromosome 9 [der (9)] ("der" stands gions of conserved synteny between the chromosomes of
for derivative). These result from a translocation between normal many mammals indicates that chromosomal rearrangements
chromosomes 9 and 22. This translocation can be detected by classical such as breakage, fusion, and translocations occurred rarely in
banding analysis diagrammed in (a) and FISH with chromosome paint mammalian evolution, about once every five million years.
probes (b). [Part (a) from J. Kuby, 1997, Immunology, 3d ed., W. H. Freeman and When such chromosomal rearrangements did occur, they very
Company, p. 578. Part (b) courtesy of J. Rowley and R. Espinosa.] likely contributed to the evolution of new species that could
not interbreed with the species from which they evolved.
Chromosomal rearrangements similar to those inferred
Chromosome Painting and DNA Sequencing
for the primate lineage have been inferred for other groups of
Reveal the Evolution of Chromosomes related organism<>, including invertebrate, plant, and fungi lin-
Analysis of chromosomes from different species has provided eages. The excellent agreement between predictions of evolu-
considerable insight about how chromosomes evolved. For tionary relationships based on analysis of syntenic regions of
example, hybridization of chromosome paint probes for chro- chromosomes from organisms with related anatomical struc-
mosome 16 of the tree shrew (Tupaia belangeri) to tree shrew ture (i.e., among mammals, among insects with similar body
metaphase chromosomes revealed the two copies of chromo- organization, among similar plants, etc.) and the evolutionary
some 16, as expected (Figure 6-42a). However, when the relationships based on the fossil record and on the extent of

.·
268 CHAPTER 6 • Genes, Genomics, and Chromosomes
 (a) (c)

I I ~,,~..T'i"o·
2 3 4 5 6
u
X

I~ III~I~
7 8 9 10 11 12 13 14

0~ 0 I
15 16 17
I I
18 19 20 21
I I
22 23
0

Homo sapiens

·1·:1·I·~ · ~ · I
2 3 4 5 6
xn
L
X

::171120 8~ 13 ::~ 150 51

7 8 9 10 11 12 13 14

(b)
5 1 ~: 160170 ~~~tal 21 m•
15 16 17 18 19 20 21 22

FIGURE 6-42 Evolution of primate chromosomes. (a) Chromo-

some paint probes for chromosome 16 of the tree shrew (T. belangeri,
distantly related to humans) were hybridized (yellow) to tree shrew
metaphase chromosomes (red). (b) The same tree shrew chromosome
16 paint probes were hybridized to human metaphase chromosomes.
(c) Proposed evolution of human chromosomes (bottom) from the
chromosomes of the common ancestor of all primates (top). The
proposed common primate ancestor chromosomes are numbered
according to their sizes, with each chromosome represented by a
different color. The human chromosomes are also numbered according
to their relative sizes with colors taken from the colors of the proposed
common primate ancestor chromosomes from which they were
derived. Small numbers to the left of the colored regions of the human
chromosomes indicate the number of the ancestral chromosome from
which the region was derived. Human chromosomes were derived
from the proposed chromosomes of the common primate ancestor
without significant rearrangements (e.g ., human chromosome 1), by
fusion (e.g., human chromosome 2 by fusion of ancestral chromosomes
9 and 11 ), breakage (e.g., human chromosomes 14 and 15 by breakage
of ancestral chromosome 5), or chromosomal translocations (e.g., a
reciprocal translocation between ancestral chromosomes 14 and 21
generated human chromosomes 12 and 22). [Parts (a) and (b) Muller et al.
divergence of DNA sequences for homologous genes is a 1999. Cromosoma 108:393. Part (c) derived from L. Froenicke, 2005, Cytogenet.
strong argument for the validity of evolution as the process Genome Res. 108:1 22.]
that generated the diversity of contemporary organisms.

Interphase Polytene Chromosomes Arise

by DNA Amplification somes are characterized by a large number of reproducible,
The larval salivary glands of Drosophila species and other well-demarcated bands that have been assigned standardized
dipteran insects contain enlarged interphase chromosomes numbers (Figure 6-43a ). The densely staining bands repre-
that are visible in the light microscope. When fixed and sent regions where the chromatin is more condensed, and the
stained with a dye that stains DNA, these polytene chromo- light, interband regiom, where chromatin is less condensed.

6.7 Morphology and Functional Elements of Eukaryotic Chromosomes 269

 them (see Figure 7-13). Insect polytene chromosomes offer
one of the only experimental systems in all of nature where
such immuno-localization studies on decondensed inter-
phase chromosomes are possible.
A generalized amplification of DNA gives rise to the poly-
tene chromosomes found in the salivary glands of Drosoph-
ila. This process, termed polytenization, occurs when the
DNA repeatedly replicates everywhere except at the tela-
meres and centromere, but the daughter chromosomes do not
separate. The result is an enlarged chromosome composed of
many parallel copies of itself, 1024 resulting from ten such
replications in Drosophila melanogaster salivary glands (Fig-
ure 6-43b). The amplification of chromosomal DNA greatly
increases gene copy number, presumably to supply sufficient
mRNA for protein synthesis in the massive salivary gland
cells. The bands in Drosophila polytene chromosomes repre-
sent ""'50,000-100,000 base pairs, and the banding pattern
reveals that the condensation of DNA varies greatly along
these relatively short regions of an interphase chromosome.

(b)
Three Functional Elements Are Required
for Replication and Stable Inheritance
of Chromosomes
Although chromosomes differ in length and number among
species, cytogenetic studies have shown that they all behave
similarly at the time of cell division. Moreover, any eukaryotic
chromosome must contain three functional elements in order
to replicate and segregate correctly: (1) replication origins at
EXPERIMENTAL FIGURE 6-43 Banding on Drosophila
polytene salivary gland chromosomes. (a) In this light micrograph of which DNA polymerases and other proteins initiate synthesis
Drosophila melanogaster larval salivary gland chromosomes, four of DNA (see Figures 4-31 and 4-33); (2) the centromere, the
chromosomes can be observed (X, 2, 3, and 4), with a total of approxi- constricted region required for proper segregation of daughter
mately 5000 distinguishable bands. The banding pattern results from chromosomes; and (3) the two ends, or telomeres. The yeast
reproducible packing of DNA and protein within each amplified site transformation studies depicted in Figure 6-44 demonstrated
along the chromosome. Dark bands are regions of more highly the functions of these three chromosomal elements and estab-
compacted chromatin. The centromeres of all four chromosomes often lished their importance for chromosome function.
appear fused at the chromocenter. The tips of chromosomes 2 and 3 As discussed in Chapter 4, replication of DNA begins
are labeled (L = left arm; R = right arm), as is the tip of the X chromo- from sites that are scattered throughout eukaryotic chromo-
some. (b) The pattern of amplification of one chromosome during five somes. The yeast genome contains many= 100-bp sequences,
replications. Double-stranded DNA is represented by a single line. called autonomously refJlicating sequences (ARSs), that act
Telomere and centromere DNA are not amplified. In salivary gland
as replication origins. The observation that insertion of an
polytene chromosomes, each parental chromosome undergoes "" 10
ARS into a circular plasmid allows the plasmid to replicate
replications (2 10 = 1024 strands). [Part (a) courtesy of J. Gall. Part (b) adapted
in yeast cells provided the first functional identification of
from C. D. Laird et al., 1973, Cold Spring Harbor Symp. Quant. Bioi. 38:311.)
origin sequences in eukaryotic DNA (see Figure 6-44a).
Even though circular ARS-containing plasmids can repli-
cate in yeast cells, only about 5-20 percent of progeny cells
Although the molecular mechanisms that control the forma- contain the plasmid because mitotic segregation of the plas-
tion of bands in polytene chromosomes are not yet under- mids is faulty. However, plasmids that also carry a CEN se-
stood, the highly reproducible banding pattern seen in quence, derived from the centromcres of yeast chromosomes,
Drosophila salivary gland chromosomes provides an ex- segregate equally or neMiy so to both mother and daughter
tremely powerful method for locating specific DNA sequences cells during mitosis (sec Figure 6-44b).
along the lengths of the chromosomes in this species. Chro- If circular plasmids containing an ARS and CEN sequence
mosomal translocations and inversions arc readily detectable are cur once with a restriction enzyme, the resulting linear
in polytene chromosomes, and specific chromosomal pro- plasmids do not produce LEU+ colonies unless they contain
teins can be localized on interphase polytene chromosomes special telomeric (TEL) sequences ligated to their ends (sec
by immunostaining with specific antibodies raised against Figure 6-44c). The first successful experiments involving

270 CHAPTER 6 • Genes, Genomics. and Chromosomes

 Plasmid with Transfected Progeny of transfected cell Conclusion EXPERIMENTAL FIGURE 6·44 Yeast
sequence from leu- cell transfection experiments identify the functional
normal yeast Growth Mitotic chromosomal elements necessary for normal
without segregation
(a) chromosome replication and segregation. In
leucine
these experiments, plasmids containing the LEU
gene from normal yeast cells are constructed and
ARS required introduced into leu cells by transfection. If the
for plasmid plasmid is maintained in the leu cells, they are
replication transformed to LEU by the LEU gene on the
plasmid and can form colonies on medium lacking
leucine. (a) Sequences that allow autonomous
replication (ARS) of a plasmid were identified
Poor In presence because their insertion into a plasmid vector
(5-20% ofARS, containing a cloned LEU gene resulted in a high
of cells plasmid
frequency of transformation to LEU' . However,
have replication occurs,
plasmid) but mitotic even plasmids with ARS exhibit poor segregation
segregation is during mitosis, and therefore do not appear in
faulty each of the daughter cells. (b) When randomly
broken pieces of genomic yeast DNA are inserted
into plasm ids containing ARS and LEU, some of the
(b) subsequently transfected cells produce large
colonies, indicating that a high rate of mitotic
Yes segregation among their plasmids is facilitating the
Good Genomic continuous growth of daughter cells. The DNA
(>90% fragment recovered from plasmids in these large colon ies
of cells CEN required contains yeast centromere (CEN) sequences.
have for good (c) When leu yeast cells are tra~sfected with
Yes plasmid) segregation
linearized plasmids containing LEU, ARS, and CEN,
no colonies grow. Addition of telomere (TEL)
sequences to the ends of the linear DNA gives the
linearized plasmids the ability to replicate as new
(c)
chromosomes that behave very much like a normal
chromosome in both mitosis and meiosis.
Linear plasmid
[See A. W. Murray and J. W. Szostak, 1983, Nature 305:89,
lacking TEL
Restriction is unstable and L. Clarke and J. Carbon, 1985,Ann. Rev. Genet. 19:29.]
enzyme
produces linear
plasmid
Yes Linear plasmids
containing ARS
TEL TEL and CEN behave
- ARS • LEU· CEN - Good like normal
chromosomes if
genomic fragment
Yes
TEL is added
to both ends

transfection of yeast cells with linear plasmids were achieved

Centromere Sequences Vary Greatly
by using the ends of a DNA molecule that was known to
replicate as a linear molecule in the ciliated protozoan Tetra- in Length and Complexity
hymena. During part of the life cycle of Tetrahymena, much Once the yeast centromere regions that confer mitotic segre-
of the nuclear DNA is repeatedly copied in short pieces to gation were cloned, their sequences could be determined and
form a so-cal led macronucleus. One of these repeated frag- compared, revealing three regions (I, ll, and III) that are con-
ments was identified as a dimer of ribosomal DNA, the ends served between the centromeres on different yeast chromo-
of which containcJ a repeated sequence (G 4 T 2 ),. When a sec- somes (Figure 6-45a). Short, fairly well conserved nucleotide
tion of this repeated TEL sequence was ligated to the ends of sequences are present in regions I and Ill. Region II does not
linear yeast plasm ids containing ARS and CEN, replication have a specific sequence, but is A-T rich with a fairly constant
and good segregation of the linear plasmids occurred. This length, probably so that regions land HI lie on the same side
·. first cloning and characterization of telomeres garnered the of a specialized centromere-associated histone octamer. This
Nobel Prize in Medicine and Physiology in 2009. specialized centromere-associated histone octamer contains

6.7 Morphology and Functional Elements of Eukaryotic Chromosomes 271

 (a)
II Ill

A A T
Yeast CEN G T C A C G T G f - - - - - 78-86 bp - - - - l T G T T T C T G N T T T C C G A A A

(b) Ndc80 complex

Domains that associate Domains that associate

with a microtubule with the CBF3 complex

(c)

Centrom.eric [ ~AL~~ ~
chromatin ~~l..'-

CBF3 [
complex Ndc80 complex

Lateral
attachment

Spindle
pole
..---
Lateral to
end-on
conversion
1
End-on

..---

FIGURE 6-45 Kinetochore-microtubule interaction in complexes initially make lateral interactions with the side of a spindle
S. c~revisiae. (a) Sequence of the simple centromeres of S. cerevisiae. microtubule (top) and then associate with the Daml ring, making an
(b) The Ndc80 complexes associate with both the microtubule and the end-on attachment (bottom) to the microtubule. [Part (a) from L. Clarke
CBF3 complex. (c) Diagram of the centromere-associated CBF3 and J. Carbon, 1985, Ann. Rev. Genet. 19:29. Parts (b) and (c) adapted from
complex and its associated Ndc80 complexes that associate with a ring T. U. Tanaka, 2010, EMBOJ. 29:4070.]
of Dam 1 proteins at the end of a spindle microtubule. The Ndc80

the usual histones H2A, H2B, and H4, but a variant form of and subsequently interact with a Daml complex that for ms a
histone H3. Centromeres from all eukaryotes similarly con- ring aro und rh<> <>nd of the microtubule (Figure 6-45c). This
tain nucleosomes with a specialized, centromere-specific form results in an end-on in teraction of the centromere with the
of histone H3, called CENP-A in humans. In S. cerevisiae the spindle microtubule. S. cerevisiae has by far the simplest cen-
CBF3 complex of proteins associates w ith this specialized nu- tromere known in nature.
cleosome. The CBF3 complex in turn associates with multi- In the fission yeast S. pombe, centromeres are = 40-1 00
protein elonga ted Ndc80 complexes (Figure 6-45b) that kb in length and are composed of repeated copies of seq uences
initially make lateral interactions with a spindle microtubule similar to those in S. cerevisiae cenrromeres. Multiple copies ·.

272 CHAPTER 6 • Genes, Genomics, and Chromosomes

 (;} FOCUS ANIMATION: Telomere Replication

FIGURE 6-46 Standard DNA replication leads to loss of DNA lagging strand DNA synthesis
<
at the 5' end of each st rand of a linear DNA molecule. Replication
of the right end of a linear DNA is shown; the same process occurs at
# .... ~
the left end (shown by inverting the figure). As the replication fork
RNA primer / ··• ' · - - -Parent
--- - - - - - - - 3'
strands
·.· approaches the end of the parental DNA molecule, the leading strand
can be synthesized all the way to the end of the parental template ~::::::::::~
~~,----1-----C·h-ro•m•o•s•o•m·e~nd
strand without the loss of deoxyribonucleotides. However, since -'< .
synthesis of thP lagging strand requires RNA primers, the right end Leading strand DNA synthesis
of the lagging daughter DNA strand would remain as ribonucleotides
which are removed and therefore cannot serve as the template for a
;:==:::-:::::;:;;;::::::-::::::;;;:::==:-::
......tJ ......_ ...... 3'
5'
replicative DNA polymerase. Alternative mechanisms must be utilized Polymerase Primer / ~
to prevent successive shortening of the lagging strand with each round ....,. 3'
of replication. [Adapted from the Nobel Assembly at the Karolinska lnst1tute.] 5'

!
of proteins homologous to those that interact with S. cerevi-
siae centromeres bind to these complex S. pombe centromeres
-- \.:.
ligation
;;·...
;=::==:::::=::-::-:-:~==-=-
/ ~
Gap fill-in /
3'

3'
and in turn bind the much longer S. pombe chromosomes to I
I 5'
several microtubules of the mitotic spindle apparatus. In
! Primer
removal

===========;s:--I:-
plants and animals, cemromeres are megabases in length and
are composed of multiple repeats of simple-sequence DNA. In 5'
humans, centromeres contain 2- to 4-megabase arrays of a 3'
Shortened end /

==========::;t:-=3'
171-bp simple-sequence DNA called alphoid DNA that is
bound by nuclcosomes containing the CENP-A histone H3 II 5'
variant, as well as other repeated simple-sequence DNA. Gap not
In higher eukaryotes, a complex protein structure called the filled
kinetochore assembles at ccntromeres and associates with mul-
tiple mitotic spindle fibers during mitosis (see Figure 18-39).
Homologs of many of the centromere-as~ociatcd proteins
found in the yeasts occur in humans and other higher eukary- The need for a specialized region at the ends of eukaryotic
otes. For those yeast proteins where clear homologs arc not chromosomes is apparent when we consider that all known
evident in higher cells based on amino acid sequence compari- DNA polymerases elongate DNA chains at the 3' end, and all
sons (such as the Dam l complex), alternative complexes with require an RNA or DNA primer. As the replication fork ap-
similar properties have been proposed to function at kineto- proaches the end of a linear chromosome, synthesis of the
chores that are bound to multiple spindle microrubules. The leading strand continues to the end of the DNA template
function of the centromere and kinetochore proteins that bind strand, completing one daughter DNA double helix. How-
to it during the segregation of sister chromatids in mitosis and ever, because the lagging-strand template is copied in a dis-
meiosis is described in Chapters 18 and 19. continuous fashion, it cannot be replicated in its entirety
(Figure 6-46). When the final RNA primer is removed, there
Addition of Telomeric Sequences is no upstream strand onto which DNA polymerase can build
to fill the resulting gap. Without some special mechanism, the
by Telomerase Prevents Shortening
daughter DNA strand resulting from lagging-strand synthesis
of Chromosomes would be shortened at each cell division.
Sequencing of telomcres from multiple organisms, including The problem of telomere shortening is solved by an en-
humans, has shown that most arc repetitive oligomers with zyme that adds telomeric (TEL) repeat sequences to the ends
a high G content in the strand with its 3' end at the end of of each chromosome. The enzyme is a protein-RNA complex
the chromosome. The telomere repeat sequence in humans called telomere terminal transferase, or telomerase. Because
and other vertebrates is TTAGGG. These simple sequences the sequence of the telomcrase-associared RNA, as we will
are repeated at the very termini of chromosomes for a total see, serves as the template for addition of deoxyribonucleo-
of a few hundred base pairs in yeasts and protozoans, and a tides to the ends of telomeres, the source of the enzyme and
few thousand base pairs in vertebrates. The 3' end of the G- not the source of the telomeric DNA primer determines the
rich strand extends 12-16 nucleotidcs beyond the 5' end of sequence added. This was proven by transforming Tetrahy-
the complementary C-rich strand. This region is bound by mena with a mutated form of the gene encoding the tclomerase-
specific proteins that protect the ends of linear chromosomes a~sociated RNA. The resulting telomerase added a DNA
from attack by exonucleases. sequence complementary to the mutated RNA sequence to the

6.7 Morphology and Functional Elements of Eukaryotic Chromosomes 273

(;} FOCUS ANIMATION: Telomere Replication

FIGURE 6- 47 Mechanism of action of telomerase. The single-

stranded 3' terminus of a telomere is extended by telomerase,
counteracting the inability of the DNA replication mechanism to
synthesize the extreme terminus of linear DNA. Telomerase elongates
this single-stranded end by a reiterative reverse-transcription mecha-
nism. The action of the telomerase from the protozoan Tetrahymena, )
which adds a T2G4 repeat unit, is depicted; other telomerases add
slightly different sequences. The telomerase contains an RNA template
(red) that base-pairs to the 3' end of the lagging-strand template. The
telomerase catalytic site then adds deoxyribonucleotides TIG (blue)
EI0"9'';"" 1D
using the RNA molecule as a template (step Dl. The strands of the
resulting DNA-RNA duplex are then thought to slip (translocate)
relative to each other so that the TIG sequence at t he 3' end of the )
replicating DNA base-pa irs to the complementary RNA sequence in the
telomerase RNA (step fl ). The 3' end o f the replicating DNA is again
extended by telomerase (step J)). Telomerases can add multiple
repeats by repetition of steps fl and D . DNA polymerase a-primase
T""''o"Uo" 1fJ
can prime synthesis of new Okazaki fragments on this extended
template strand. The net result prevents shortening of the lagging
strand at each cycle of DNA replication [From c.w. Greider and
E. H. Blackburn, 1989, Nature 337:331.]
)
Elo"g'Uo" 1II
ends of telomeric primers. Thus telomerase is a specialized form
of a reverse transcriptase that carries its own internal RNA tem-
plate to direct DNA synthesis. These experiments also earned
)
the Nobel Prize in Physiology and Medicine fo r the discovery
and characterization of the mechanism of telomerase.
Figure 6-4 7 depicts how telomerase, by reverse transcrip-
tion of its associated RNA, elongates the 3' end of the single-
stranded DNA at the end of the G-rich strand mentioned species maintain telomere lengths by t he regu lated insertion
above. Cells frsm1 knockout mice that cannot produce the of non-LTR retrotransposons into telomeres. This is one of
telomerase-associated RNA exhibit no telomerase acti\'ity, the few instances in which a mobi le element has a specific
and their telomeres shorten successively with each cell gen- function in its host organism.
eration. Such mice ca n breed and reproduce norma ll y for
three generations before the long telomere repeats become
substantially eroded. Then, the absence of telomere DNA
results in adverse effects, includ ing fusion of chromosome KEY CONCEPTS of Section 6.7
termini and chromosomal loss. By the fourth generation, the
reproductive potential of these knockout mice declines, and Morphology and Functional Elements
they cannot produce offspring after the sixth generation. of Eukaryotic Chromosomes
• D uri ng metaphase, eukaryotic chromosomes become suf-
The human genes expressing the telomerase protein ficientl y condensed that they can be visualized individually
and the telomerase-associated RNA are active in germ in the light microscope.
cells and stem cells, but are turned off in most cells of adult
• T he ch romosomal karyotype is characteristic of each spe-
tissues that replicate only a limited number of times, or will
cies. Closely related species can have dramatically different
never replicate again (such cells are called postmitotic). How-
ka ryotypes, ind icating that simila r genetic information can
ever, these genes are activated in most human cancer cells,
be organized o n chromosomes in d ifferent ways.
where telomerase is required for th(' mu ltiple cell divisions
necessary to form a tumor. This phenomenon has stimulated • Banding analysis and chromosome painting are used to
a search for inhibitor s of human te lomerase as potential identify the different h uman metaphase chromosomes and to
therapeutic agents for treating cancer. • detect t ranslocations and deletions (see Figure 6-41 ).
• Ana lysis of chromosomal rearrangements and regions of
While telomcrase prevents telomere shorten ing in most conserved synteny between r elated species allows scientists
cukar}·otcs, some organisms use alternative strategies. Drosophila

274 CHAPTER 6 • Genes, Genomics, and Chromosomes

 inserted into new sites in individuals throughout our history.
to make predictions about the evolution of chromosomes Large numbers of these interspersed repeats are polymorphic
(see Figure 6-42c). The evolutionary relationships between within populations, occurring at a particular site in some
organisms indicated by these studies are consistent with pro- individuals and not others. Individuals sharing an insertion
posed evolutionary relationships based on the fossil record at a particular site descended from a common ancestor that
and DNA sequence analysis. developed from an egg or sperm in which that insertion oc-
• The highly reproducible banding patterns of polytene curred. The time elapsed from the initial insertion can be
chromosomes make it possible to visualize chromosomal de- estimated by the differences in sequences of the element that
letions and rearrangements as changes in the normal p:ntern arose from the accumulation of random murarions. Furrher
of bands. analysis of retrotransposon polymorphisms will undoubt-
• Three types of DNA sequences are required for a long lin- edly add immensely to our understanding both of human
ear DNA molecule to function as a chromosome: a replica- migrations since Homo sapiens first evolved, as well as the
history of contemporary populations.
tion origin, called ARS in yeast; a centromere (CEN) se-
quence; and two telomere (TEL) sequences at the ends of the As described in Chapter 5, the Drosophila P-element
DNA (see Figure 6-44). DNA transposon was exploited for the facile stable transfor-
mation of genes into the Drosophila germ line. This has been
Telomerase, a protein-RNA complex, has a special reverse a powerful method for molecular cell biology experimenta-
transcriptase activity that completes replication of telomeres tion in this organism. An active area of current research is to
during DNA synthesis (see Figure 6-47). In the absence of use mammalian transposons and retrotransposons for the
telomerase, the daughter DNA strand resulting from lagging- transformation of human cells for gene therapy. This prom-
strand synthesis would be shortened at each cell division in ises to be an exciting area of medicine in the future treatment
most eukaryotes (see Figure 6-46). of genetic diseases such as sickle-cell anemia and cystic fibro-
sis as well as more common diseases, especially when cou-
pled with the recently developed techniques for generating
pluripotent stem cells (iPS cells) from differentiated cells of a
pediatric or adult patient (Chapter 21). ·

Perspectives for the Future

The human genome sequence is a goldmine for new discov-
eries in molecular cell biology, in identifying new proteins
that may be the basis of effective therapies of human dis-
eases, and for understanding early human history and evolu- Key Terms
tion. However, finding new genes is like finding a needle in bioinformatics 252 matrix-associated regions
a haystack, because only "=' 1.5 percent of the sequence en- centromere 270 (MARs) 263
codes proteins or functional RNA. Identification of genes in monocistronic 225
chromatid 267
bacterial genome sequences is relatively simple because of
chromatin 224 nucleosome 225
the scarcity of introns; simply searching for long open read-
ing frames free of stop codons identifies most genes. In con- cytoplasmic open reading frame (ORf)
trast, the search for human genes is complicated by the inheritance 246 253
structure of human genes, most of which are composed of DNA transposon 235 polytene chromosome 269
multiple, relatively short exons separated by much longer, epigenetic code 261 protein family 228
noncoding introns. Identification of complex transcription euchromatin 260 pseudogene 228
units by analysis of genomic DNA sequences alone is ex- retrotransposon 235
exon shuffling 243
tremely challenging. future improvements in bioinformatic
fluorescence in situ repetitious DNA 223
methods for gene identification, and characterization of
eDNA copies of mRNAs isolated from the hundreds of hybridization (FISH) 267 scaffold-associated regions
human cell types, will likely lead to the discovery of new gene family 228 (SARs) 263
proteins, to a better understanding of biological processes, genomics 225 simple-sequence (sa tellite)
and may lead to applications in medicine and agriculture. heterochromatin 260 DNA 232
We have seen that although most transposons do not func- hi~tones 225 SINEs 240
tion din:<.:tly in cellular processes, they have helped to shape SMC proteins 263
histone code 258
modern gcnomes by promoting gene duplications, exon shuf- telomere 2 70
fling, the generation of new combinations of transcription- karyotype 267
LINEs 240 transcription unit 225
control sequences, and other aspects of contemporary
genomes. The} are also teaching us about our own history long terminal repeats transposable DNA element
and origins, because L l and Alu rctrotransposons have (LTRs) 238 223

Key Terms 275

 13. What is chromosome painting, and how is this tech-
Review the Concepts
nique useful? How can chromosome paint probes be used to
1. Genes can be transcribed into mRNA for protein-coding analyze the evolution of mammalian chromosomes?
genes or into RNA for genes such as ribosomal or transfer 14. Certain organisms contain cells that possess polytene
RNAs. Define a gene. For the following characteristics, state chromosomes. What are polytene chromosomes, where are
whether they apply to (a) continuous, (b) simple, or (c) com- they found, and what function do they serve?
plex transcription units. 15. Replication and segregation of eukaryotic chromosomes • I

(i) Found in eukaryotes require three functional elements: replication origins, a cen-
(ii) Contain introns tromere, and telomeres. How would a chromosome be af-
(iii) Only capable of making a single protein from a fected if it lacked (a) replication origins or (b) a centromere?
gtven gene 16. Describe the problem that occurs during DNA replica-
2. Sequencing of the human genome has revealed much tion at the ends of chromosomes. How are telomeres related
about the organization of genes. Describe the differences be- to this problem?
tween solitary genes, gene families, pseudogenes, and tan-
demly repeated genes.
3. Much of rhe human genome consists of repetitious DNA.
Describe the difference between microsatellite and minisatel- Analyze the Data
lite DNA. How is this repetitious DNA useful for identifying
1. To determine whether gene transfer from an organelle ge-
individuals by the technique of DNA fingerprinting?
nome to the nucleus can be observed in the laboratory, a chlo-
4. Mobile DNA elements that can move or transpose to a roplast transformation vector was constructed that contained
new site directly as DNA are called DNA transposons. De- two selectable antibiotic-resistance markers, each with its own
scribe the mechanism by which a bacterial DNA transposon, .·
promoter: the spectinomycin-resistance gene and the kanamycin-
called an insertion sequence, can transpose. resistance gene (seeS. Stegemann et al., 2003, Proc. Nat'/ Acad.
5. Retrotransposons are a class of mobile elements that Sci. USA 100:8828-8833). The spectinomycin-resistance gene
transpose via an RNA intermediate. Contrast the mechanism was controlled by a chloroplast promoter, yielding a chloro-
of transposition between rctrotransposons that contain long plast-specific selectable marker. Plants grown on spectinomy-
terminal repeats (LTRs) and those that lack LTRs. cin are white unless they express the spectinomycin-resistance
6. Discuss the role that transposons may have played in the gene in the chloroplast. The kanamycin-resistance gene, in-
evolution of modern organisms. What is exon shuffling? What serted into the plasmid adjacent to the spectinomycin-resis-
role do transposons play in the process of exon shuffling? tance gene, was under the control of a strong nuclear
7. Mitochondria and chloroplasts are thought to have promoter. Transgenic, spectinomycin-resistanr tobacco
evolved from symbiotic bacteria present in nucleated cells. plants were selected following transformation with this plas-
What is the experimental evidence from this chapter that mid by identifying green plants grown on medium with spec-
supportS this hypothesis? tinomycin. These plants contain the two antibiotic-resistance
8. What are paralogous and orthologous genes? What are genes inserted into the chloroplast genome by a recombina-
some of the explanations for the finding that humans are a tion event; however, kanamycin resistance is not expressed
because it is under the control of a nuclear promoter. These ...
much more complex organism than the roundworm C. ele-
gans, yet have only fewer than one and a half as many genes spectinomycin-resistant plants were grown for multiple gen-
(25,000 versus 18,000)? erations and used in the following studies.
a. Leaves from the spectinomycin-resistant transgenic
9. The DNA in a cell as~ociates with proteins to form chro-
plants were placed in a plant-regeneration medium contain-
matin. What is a nucleosome? What role do histones play in
ing kanamycin. Some of the leaf cells were resistant to ka-
nucleosomes? How are nucleosomes arranged in condensed
namycin, and grew into kanamycin-resistant plants. Pollen
30-nm fibers?
(paternal) from kanamycin-resistant plants was used to pol-
10. How do chromatin modifications regulate transcrip- linate wild-type (nontransgenic) plants. In tobacco, no chlo-
tion? What modifications are observed in regions of the ge- roplasts are inherited from pollen. The resulting seeds were
nome that are actively being transcribed? What about for germinated on media with and without kanamycin. Half of
regions that are not actively transcribed? the resulting seedlings were kanamycin resistant. When these
11. Describe the general organization of a eukaryotic chro- kanamycin-resistant plants were allowed to self-pollinate,
mosome. What structural role do scaffold-associated regions the offspring exhibited a 3:1 ratio of kana myCJn-reststant to
(SARs ) or matrix attachment regions (MARs) play? Why sensitive phenotypes. What can be deduced from these data
does it make sense that protein-coding genes are not located about the location of the kanamycin-resistance gene?
in these regions? b. To determine whether transfer of the kanam ycin-
12. What is FISH? Briefly describe how it works. How is FISH resistance gene to the nucleus was mediated via DNA or an
used to characterize chromosomal translocations associated RNA intermediate, DNA was extracted from J 0 seedling
with certain genetic disorders and specific types of cancers? plants germinated from seeds produced by a wild-type plant

276 CHAPTER 6 • Genes, Genomics, and Chromosomes

 pollinated with a kanamycin-resistant plant. The I0 seedling expression then progressively decreased as the repeat number
plants, numbered 1-1 0 in the corresponding gel lanes in the increased from 13 to 60. What conclusion can be made based
figure below, consist of 5 kanamycin-resistant (+)and 5 kana- on these experimental results? Explain how Q RT-PCR can be
mycin-sensitive (-) plants. Each DNA sample was subjected used to analyze SDTl gene expression.
to PCR analysis using primers to amplify the kanamycin-re- c. The conclusion from part (b) then led the group to de-
sistance gene (gel at left) or rhe spectinomycin-resistance gene termine whether a cell can adapt to its surroundings by alter-
(gel at right). The lane marked M shows molecular weight ing the number of repeats in promoter satellite DNA. The
markers. What does the correspondence between the presence promoter from the SDTl gene (containing 48 repeats) was
or absence of PCR products generated in the same plant with ;~ttached to the URA3 open reading frame, a gene responsible
both sets of primers suggest about the mode of transfer of the for synthesis of the nucleotide uracil. Cells with this hybrid
kanamycin gene to the nucleus? protein were then placed on media lacking uracil or media
containing uracil. After growth on each media, the number
M 1 2 3 4 5 6 7 8 9 10 M 1 2 3 4 5 6 7 8 9 10 of repeats in the promoter driving URA3 expression was
analyzed. Those that were grown on the media lacking uracil

-- - showed a decrease in the number of repeats in the promoter,

while those grown with uracil still had on average 48 repeats.
What conclusion can be made from these results? Based on
the data in part (b), how many repeats were likely found in
--- ---
the URA3 promoters for cells grown without uracil?
d. The location of the repeat DNA in the promoters was
compared ro the location of nucleosomes, and it was found
that the nucleosome density in a promoter was inversely cor-
- + + + - - + - + - - + ~ + - - + - + - related with the number of repeats at that location of DNA.
Kan resista nee Kan resistance From this, what would you conclude about repetitive DNA
(PCR products using (PCR products using
·. kanamycin-resistance spectinomycin-resistance
and chromatin packaging? What effect would decreasing the
gene primers] gene primers] repeat regions have on histone core binding' at that DNA?

c. When the original transgenic plants, which were select- References

ed on spectinomycin but not on kanamycin, were used to pol-
Eukaryotic Gene Structure
linate wild-type plants, none of the offspring were kanamycin
resistant. What can be deduced from these observations? Black, D. L. 2003. Mecham~ms of alternative pre-messenger
RNA ~pi icing. Ann. Rev. Biochem. 72:291-336.
2. Satellite DNA is a known component of our genome and Davuluri, R. V., er al. 2008. The functional consequences of
can be found in both coding and noncoding DNA. When it alternative promoter use in mammalian genomcs. Trends Genet.
is found in coding DNA, the number of repeats can result in 24:167-177.
altered proteins. Bur the effect of these repeats in noncoding Wang, E. T., et al. 2008. Alternative isoform regulation in
DNA is nor as well understood. To determine whether re- human tissue transcnptomes. Nature 456:4..,0-476.
peats in the promoter region can alter gene expression and Chromosomal Organization of Genes and Noncoding DNA
chromatin compaction, Vinces eta!. (Vinces et al., 2009, Sci- Celnikcr, S. E., and G. M. Rubm. 2003. The Drosophzla
ence 324: 1213-1216) searched the Sacharomyces cereuisiae melanogaster genome. Ann. Ret•. Genomics Hum. Genet. 4:89-11 7.
genome for the presence of repetitive DNA in promoters and Crook, Z. R., and D. llousman. 2011. Hunrmgton's d1sease:
examined how altering the number of repeats affected gene can mice lead the way to treatment? Neuron 69:423-435.
expression and DNA packaging. Feutllet, C., et al. 20 I I. Crop genome sequencmg: lessons and
a. The group first searched for satellite DNA in the ge- rationales. Trends Plant Set. 16:77-88.
nome of various S. cerevisiae strains and found that 25 per- International Human Genome Sequencing Consortium. 2004.
Finishmg the euchromatic sequence of the human genome. Nature
cent of promoters contained at least one repeat region. In ad-
431:931-945.
dition, a single promoter in different strains often contained Giardina, E., A. Spinella, and G. Novelli. 2011. Past, present and
distinct numbers of repeats in each region of satellite DNA. furure of forensic DNA typing. Nanonredicine (L01zd.) 6:257-2..,0.
What is the proposed mechanism by which the number of Hannan, A.]. 2010. TRPing up the genome: tandem repeat
repeats in a given region of satellite DNA can increase? polymorphism~ as dynamic sources of genetic variability m health
b. To determine whether there is a correlation between and d1sease. Drscot•. Med. 10:314-321.
the number of repeats and gene expression, transcription at Jobling, M. A., ,md P. Gill. 2004. Encoded evidence: lJ'\A 111
the SDTl gene was analyzed. The SDTl promoter contains forensic analysis. N,lf. Reu. Genet. 5:739-..,., I.
satellite DNA, and the number of repeats in this region was Lander, E. S., et al. 200 I. Initial sequencing and analysi' of the
human genome. Nature 409:860-921.
modified, ranging between 0 repeats to 60 repeats. It was dis-
Todd, P. J.-., and H. L. Paulson. 20 I 0. RNA-mediated neurode-
covered that SDTl expression, analyzed by quantitative re- generatlon in repeat expansion d1sordcrs. Ann. Neurol. 67:291-300.
verse transcriptase PCR (Q RT-PCR, see Chapter 5), increased Venter, J. C., eta!. 200 I. The sequence of rhe human genome.
when the repeat number increased from 0 to 13 repeats. SDT1 Sczence 291:1304-1351.

References 277
Transposable (Mobi le) DNA Elements Slearor, R. D. 2010. An overview of the current status of
Curc1o, ~1. J., and K. M. Derbyshire. 2003. The ours and ms of eukaryote gene prediction strategies. Gene 461:1 -4.
transposition: from mu to kangaroo. Nat. Rev. Mol. Cell Bwl. Sonah, H., et al. 2011. Genomic resources in horticultural
4:865-8 77 . crops: status, urilit: and challenges. Btotechnol. Adv. 29:1 99-209.
Goodier, J. L., and H. H. Kazazian, Jr. 2008. Rerrorranspo\ons Stratton, .\1. R. 2011. Exploring the genomes of ca ncer cells:
rev1sired: the restraint and rehabilitation of parasites. Cel/135:23-35. progress and p romise. Science 331:1553-1558.
Jones, R. N. 2005. McClintock's conrrollmg elements: the full Venter, J. C. 2011 . Genome-sequencing anniversary. The human
srory. Cytogenet. Genome Res. 109:90-1 03. genome ar 10: successes and challenges. Science 331:546-547.
Lisch, D. 2009. Ep1generic regulation of transposable elements
in plants. Ann. Rev. Plant Btol. 60:43-66. Structural Organization of Eukaryotic Chromosomes
Bannister, A. J., and T. Kouzarides. 2011. Regulation o f
Organelle DNAs chromatin by h istone modifications. Cell Res. 21:381-395.
Bendich, A. J. 2004. Circular chloroplast chromosomes: the Bernstein, B. E., A. Meissner, and E. S. Lander. 2007. The
grand illusion. Plant Cel/16:166 1- 1666. mammalian epigenome. Ce/1 128:669-68 1.
Bonawitz, N.D., D. A. Clayton, and G. S. Shadel. 2006. H orn, P. J., and C. L. Peterson. 2006. Heterochromatin
Initiation and beyond: multiple functions of the human mitochon- assembly: a new twist on an old model. Chromosome Res. 14:83-94.
drial transcription machinery. Mol. Cel/ 24: 8 13-825. Kurdisrani, S. K. 201 1. Hisrone modifications in cancer biology
Chan, D. C. 2006. Mitochondria: dynamic organelles in d isease, and prognosis. Prog. Drug Res. 67:91-106.
aging, and development. Cel/ 125: 1241-1252. Luger, K. 2006. Dynamic nucleos6mes. Chromosome Res.
14:5-16.
Genomics: Genome-wide Analysis of Gene Structure Luger, K., and T.]. Richmond. 1998. The histone rails of the
and Expression nucleosome. Cz1rr. Opin. Genet. Dev. 8:140-146.
BLAST Information can be found at: hnp://blast.ncbi.nlm.mh. ~asmyrh, K., and C. H . Haenng. 2005. The structure and function
gov/Blast.cgi of S.\IC and kleisin complexes. Ann. Rev. B10chem.74:595-648.
1000 Genomes Project Consortium, G.P. 2010. A map of Schalch, T., et al. 2005. X-ray structure of a retranucleosome
human genome vanation from population-scale sequencing. Nature and its implications for the chromatin fibre. Nature 4 36:138-14 1.
467:1061-1073. Woodcock, C. L., and R. P. Ghosh. 2010. Chromatin higher-
Alkan, C., B. P. Coe, and E. E. Eichler. 2011. Genome structural order strucwre and dynamics. Cold Spring Hal'hor Perspect. Bioi.
variation discovery and genotyping. Nat. Rev. Genet. 12:363- 376. 2:a000596.
Chimpanzee Sequencing and Analysis Consortium. 2005. Initial
sequence of the chimpanzee genome and comparison with rhe Morphology and Functional Elements of Eukaryotic Chromosomes
human genome. Nature 437:69-87. Armanios, M., and C. W. Greider. 2005. Telomerasc and cancer
du Plessis, L., N. Skunca, and C. Dessimoz. 201 1. The what, stem cells. Cold Spring Harbor Symp. Quant. Btol. 70:205-208.
where, how and why of gene ontology-a primer for bioinformatl- Belmont, A. S. 2006. Mitotic chromosome structure and
cians. Bne( Btomform. 2011 Feb 17. [Epub ahead of prim] condensation. Curr. Opin. Cell Bioi. 18:632-638.
Ideker, T., J. Dutkowski, and L. H ood. 2011. Boosting signal-to- Blackburn, E. H. 2005. Telomeres and telomera~e: rheir
noise in complex biology: prior knowledge is power. Cel/ 144:860-863. mechanisms of action and the effects of altering their functions.
International Human Genome Sequencing Consortium. 2004. fEBS l.ett. 579:859-862.
Finishing the euchromatic sequence of rhe human genome. Nature Cvetic, C., and J . C. Walter. 2005. EukaryotiC origms of D~A
431:931-945. repilcanon: could you please be more specific? Semm. Cell Dev. Bioi.
Lander, E. S. 20 I I. lmrialtmpacr of the sequencing of the 16:343-353.
human genome. Nature 4 70:187-197. Froenicke, L. 2005. Origins of primate chromosomes as
~1ills, R. 1::.., er al. 2011. Mapping copy number variation b} delmeated by Zoo-FISH and alignments of human and mouse draft
populanon-scale genome sequencmg. Nature 470:59-65. genome sequences. Cytogenet. Genome Res. 108:1 22-138 .
Picardi, E., and G. Pesole. 2010. Computational methods for ab MacAlpine, D. M., and S. P. Bell. 2005. A genomic view of
mitio and comparative gene finding. Meth. Mol. Bioi. 609:269-284. eukaryotic DNA replication. Chromosome Res. 13:309-326.
Ramskold, D., et al. 2009. An abundance of ubiquitously Ohra, S., ct al. 2011. Building mitotic c h romo~omes. Curr.
expressed genes revealed hy tissue transcriptome sequence data. Opm. Cell Bioi. 23:114-1 2 1.
PLoS Comput. B10l. 5:e I 000598. Tanaka, T. U. 2010. Kinetochore-microtubule interactions:
Raney, B. J., et al. 20 I I . ENCODE whole-genome data in the UCSC: steps towards hi-orientation. EMBO }. 29:4070-4082.
genome browser (20 II update). Nucl. Acids Res. 39:D871-D875.

278 CHAPTER 6 • Genes, Genomics, and Chromosomes

 CHAPTER
...
.....' ..
'
, t
,
•
• \... ,,
~ -
- ...
t •:

'
'. '
~
~

•
)

....'
-;
~

"
Transcriptional Control
/ ,
/
' I' ·' of Gene Expression
' ~·
•• '
Drosophila polytene chromosomes stained with antibodies against
a chromatin-remodeling ATPase called Kismet (blue), RNA polymerase
II with low CTD phosphorylation (red), and RNA polymerase II with
high CTD phosphorylation (green). [Courtesy of John Tamkun; see
S. Srinivasan et al., 2005, Development 132:1623.]

n previous chapters we have seen that the properties and develop, as well as how pathological abnormalities of gene
functions of each cell type are determined by the proteins expression occur, it is essential to understand the molecular
it contains. In this and the next chapter, we consider how interactions that control protein production.
the kinds and amounts of the various proteins produced by The basic steps in gene expression, i.e., the enure process
a particular cell type in a multicellular organism are regu- whereby the information encoded in a particular gene is de-
lated. This regulation of gene expression is the fundamental coded into a particular protein, are reviewed in Chapter 4.
process that controls the development of a multicellular or- Synthesis of mRNA requires that an RNA polymerase initiate
ganism such as ourselves from a single fertilized egg cell into transcription (initiation ), polymerize ribonucleoside triphos-
the thousands of cell types from which we are made. When phates complementary to the DNA coding strand (elongation),
gene expression goes awry, cellular properties are altered, a and then terminate transcription (termination ) (see Figure 4-11 ).
process that all too often leads to the development of cancer. In bacteria, ribosomes and translation-initiation factors have
As discussed further in Chapter 25, genes encoding proteins immediate access to newly formed RNA transcripts, which
that restrain cell growth are abnormally repressed in cancer function as mRNA without further modification. In eukary-
cells, whereas genes encoding proteins that promote cell otes, however, the initial RNA transcript is subjected to pro-
growth and replication arc inappropriately activated in can- cessing that yields a functional mRNA (see Figure 4-15). The
cer cells. Abnormalities in gene expression also resu lt in de- mRNA then is transported from its site of s;nthcsis in the
velopmental defects such as cleft palate, tetralogy of bllot (a nucleus to the cytoplasm, where it is translated into protein
serious developmental defect of the heart that can be treated with the aid of ribosomes, tRNAs, and translation factors
surgically), and many others. Regulation of gene expression (see Figures 4-24, 4-25, and 4-27).
also plays a vital role in bacteria and other single-celled mi- Regulation may occur at several of the various steps in gene
croorganisms, where it allows cells to adjust their enzymatic expression outlined above: transcription initiation, elongation,
machinery and structural 'components in response to their RNA processing, mRNA export from the nucleus, and transla-
changing nutritional and p h ys ical environment. Conse- tion into protein. This results in differential production of
quently, to understand how microorganisms respond to their proteins in different cell types or developmental stages or
environment and how multicellular organisms normally in response to external conditions. Although examples of

OUTLINE

7.1 Control of Gene Expression in Bacteria 282 7.5 Molecular Mechanisms of Transcription Repression
and Activation 31 5
7.2 Overview of Eukaryotic Gene Control 288
7.6 Regulation of Transcription-Factor Activity 323
7.3 RNA Polymerase II Promoters and General
Transcription Factors 295 7.7 Epigenetic Regulation of Transcription 327

7.4 Regulatory Sequences in Protein-Coding Genes 7.8 Other Eukaryotic Transcription Systems 336
and the Proteins Through Which They Function 302
regulation at each step m gene expression have been found, neurological processes such as learning and memo ry. When
control of transcription initiation and elongation-the first rwo these regu latory mechanisms contro ll ing transcription fu nction
steps-arc the most important mechanisms fo r determining imp roperly, pathological processes may occur. For example,
w hether most genes are expressed and how much of t he en- reduced activity of the Pax6 gene causes aniridia, failure to de-
coded mRNAs and, consequently, proteins are produced. The velop an iris (Figure 7-la) . Pax6 is a transcription factor that
molecular mechanisms that regulate transcription initiation and normally regulates transcription of genes in volved in eye devel-
elongation are critical to numerous biological phenomena, in- opment. In other organisms, mutations in transcription factors
cluding t he development of a mu lticellular organism from a cause an extra pair of wings ro develop in Drosophila (Figure
single fertilized egg cell as mentioned above, the immune re- 7- 1b), alter the structures of flowers in plants (Figu re 7- l c), and
sponses that protect us from pathogenic microorganisms, and are responsi ble fo r multiple other developmental abnormalities.

(a)

(b)

Haltere

Normal Ubxmutant
(c)

FIGURE 7·1 Phenotypes of mutations in genes encoding t haliana that inactivate both copies of t hree flora l organ-identity gPn P~
transcription factors. (a) A mutation that inactivates one copy of the transform the normal parts of the flower into leafl ike structures. In
Pax6 gene on either the maternal or paternal chromosome 9 resu lts in each case, these mutations affect master regulatory tran scription
failure to develop an iris, or aniridia. (b) Homozygous mutations that factors that reg ulate m ultiple genes, incl uding many genes encoding
prevent expression of the Ubx gene in the t hird thoracic segment of other transcription factors. [Part (a), left, Simon Fraser/Photo Researchers,
Drosophila result in transformation of t he third segment, which Inc.; right, Visuals Unlimited. Part (b) from E. B. Lewis, 197B, Nature 276:565.
norma lly has a balancing organ called a haltere, into a second copy of Part (c) from D. Wiegel and E. M. Meyerowiu, 1994, Ce// 78:203.]
the thoracic segment that develops wings. (c) Mutations in Arabidopsis

280 CHAPTER 7 • Tran scriptional Control of Gen e Expression

 Transcription is a complex process involving many layers as the binding sites for transcription factors (repressors and
of regulation. In this chapter, we focus on the molecular events activators) and how the RNA polymerases responsible for
that determine when transcription of a gene occurs. First, we transcription bind to promoter sequences to initiate the synthe-
consider the mechanisms of gene expression in bacteria, where sis of an RNA molecule complementary to template DNA.
DNA is not hound by histones and packaged into nuclco- Next, we consider how activators and repressors influence
somes. Repressor and activator proteins recognize and bind to transcription through interactions with large, multiprotein
specific regions of DNA to control the transcription of a nearby complexes. Some of these multiprotein complexes modify
gene. The remainder of the chapter focuses on cukaryotic tran- chromatin condensation, altering access of chromosomal DNA
scription regulation and how rhP basic tenets of bacterial regu- to transcriptiuu fa<.:tors and RNA polymerases. Other com-
lation arc applied in more complex ways in higher organisms. plexes influence the rate at which RNA polymerase binds to
These mechanisms also make use of the association of DNA DNA at the site of transcription initiation, as well as the fre-
with histone octamers, forming chromatin structures with quency of initiation. Very recent research has revealed that, in
varying degrees of condensation and post-translational modi- multicellular animals, for many genes, the RNA polymerase
fications such as acetyla tion and methylation to regulate tran- pauses after transcribing a short RNA and transcription regu-
scription. Figure 7-2 provides an overview of transcription lation involves a release of the paused polymerase, allowing it
regulation in metazoans (multicellular animals) and the pro- to transcribe through the rest of the gene. We discuss how
cesses outlined in this chapter. We discuss how specific DNA transcription of specific genes can be specified by particular
sequences function as transcription-control regions by serving combinations of the =2000 transcription factors encoded in

Closed
Gene /chromatin
"Off"

Repressors

Chromatin
~ t Activators

Me Me

Transcriptional
activators
l Activators

Ac

FIGURE 7-2 Overview of eukaryotic t ranscription

control. Inactive genes are assembled into regions of
condensed chromatin that inhibit RNA polymerases and
their associated general transcription factors (GTFs) from
interacting with promoters. Activator proteins bind to
specific DNA sequence-control elements in chromatin
and interact with multiprotein chromatin co-activator
complexes to decondense chromatin and the multisub-
unit mediator to assemble RNA polymerase and general
transcription factors on promoters. Alternatively,
Ac
repressor proteins bind to other control elements to
inhibit initiation by RNA polymerase and interact with
multiprotein co-repressor complexes to condense
chromatin. RNA polymerase initiates transcription but
pauses after transcribing 20-50 nucleotides due to the
act1on of elongation inhibitors. Activators promote the
association of elongation factors that release the

-
Activators elongation inhibitors and allow productive elongation
through the gene. OS IF is the DRB sensitivity-inducing
factor, NELF is the negative elongation factor, and P-TEFb
is a protein kinase comprised of CDK9 and cyclin T. [Adapted
Nascent transcript from S. Malik and R. G. Roeder, 2010, Nat. Rev. Genet. 11:761.]

.·
CHAPTER 7 • Transcriptional Control of Gene Expression 281
the human genome, giving rise to cell-type-specific gene ex- encodes five polypeptides needed in the biosynthesis of trypto-
pression. We will consider the various ways in which the ac- phan (see Figure 4-13). Similarly, the lac operon encodes three
tivities of transcription factors themselves are controlled to proteins required for the metabolism of lactose, a sugar present
ensure that genes are expressed only at the right time and in the in milk. Since a bacterial operon is transcribed from one start
I
right place. We will al!>o discuss recent studies revealing that site into a single mRNA, all the genes wi thin an operon are
RNA-protein complexes in the nucleus can regulate transcrip- coordinately regulated; that is, they are all activated or re-
tion. New methods for sequencing DNA coupled with reverse pressed to the same extent.
transcription of RNA into DNA in vitro have revealed that Transcription of operons, as well as of isolated genes, is
much of the genome of eukaryotes is transcribed into low- controlled by interplay between RNA polymerase and specific
abundance RNAs that do not encode protein, raising the pos- repressor and activator proteins. In order to initiate transcrip-
sibility that transcription control by such noncoding RNAs tion, however, E. coli RNA polymerase must be associated
may be a much more general process than is currently under- with one of a small number of a (sigma) factors. The most
stood. RNA processing and various post-transcriptional mech- common one in eubacterial cells is a ~o. a "'O binds to RNA poly-
anisms for controlling eukaryotic gene expression are covered merase and to promoter DNA sequences, bringing the RNA
in the next chapter. Subsequent chapters, particularly Chapters polymerase enzyme to a promoter. a 70 recognizes and binds to
15, 16, and 21, provide examples of how transcription is regu- both a six-base-pair sequence centered at =- l 0 and a seven-
lated by interactions between cells and how the resulting gene base-pair sequence centered at =-~5 from the + 1 transcrip-
control contributes to the development and function of specific tion start. Consequently, the -10 plus the -35 sequences
types of cells in multicellular organisms. constitute a promoter for E. coli RNA polymerase associated
with a 70 (see figure 4-lOb) . Although the promoter sequences
70
contacted by a are located at -35 and -10, E. coli RNA
7.1 Control of Gene Expression polymerase binds to the promoter region DNA from=- 50 to
= + 20 through interactions with DNA that do not depend on
in Bacteria
the sequence. a 70 also assists the RNA polyn;;erase in separat-
Since the structure and function of a cell are determined by ing the DNA strands at the transcription start site and inserting
the proteins it contains, the control of gene expression is a the coding strand into the active site of the polymerase so that
fundamental aspect of molecular cell biology. Most com- transcription starts at + 1 (see Figure 4-11, step f)). The opti-
monly, the "decision" to transcribe the gene encoding a par- mal a 70-RNA polymerase promoter sequence, determined as
ticular protein is the major mechanism for controlling the "consensus sequence" of multiple strong promoters, is
production of the encoded protein in a cell. By controlling
- 35 reg1on - 10 region
transcription, a cell can regulate which proteins it produces
TIGACAT--15-17 bp--TATAAT
and how rapidly. When transcription of a gene is repressed,
the corresponding mRNA and encoded protein or proteins The consensus sequence has the most commonly occurring
are synthesized at low rates. Conversely, when transcription base at each of the positions in the - 35 and -10 regions. The
of a gene is activated, both the m RNA and encoded protein size of the font indicates the importance of the base at that
or proteins arc produced at much higher rates. position, determined by the influence of mutations of these
In most bacteria and other single-celled organisms, gene bases. The sequence shows the strand of DNA that has the
expression is highly regulated in order to adjust the cell's same 5'~3' orientation as the transcribed RNA (i.e., the
enzymatic machinery and structural components to changes nontemplate strand). However, the a 70-RNA polymerase ini-
in the nutritional and physical environment. Thus, at any tially binds to double-stranded DNA. After the polymerase
given time, a bacterial cell normally synthesizes only those transcribes a few tens of base pairs, a 70 is released. Thus a "'O
proteins of tts entire proteome that are required for survival acts as an initiation factor that is required for transcription
under particular conditions. Here we describe the basic fea- initiation but not for RNA strand elongation once initiation
tures of transcription control in bacteria, using the lac op- has taken place.
eron and the glutamine synthetase gene in 1-•. coli as our
primary examples. Many of the same processes, as well as
Initiation of lac Operon Transcription Can
others, arc involved in eukaryotic transcription control,
which is the subject of the remainder of this chapter. Be Repressed and Activated
When E. coli is in an environment that lacks lactose, synthe-
Transcription Initiation by Bacterial sis of lac mRNA is repressed so that cellular energy is not
wasted synthesizing enzymes the cc::lb <..:amwt use. In an envi-
RNA Polymerase Requires Association
ronment containing both lactose and glucose, E. coli cells
with a Sigma Factor preferentially metabolize glucose, the central molecule of
In E. coli, about half the genes are clustered into operons, each carbohydrate metabolism. Lactose is metabolized at a high
of which encodes enzymes involved in a particular metabolic rate only when lactose is present and glucose is largely de-
pathway or proteins that interact to form one multisubunit pleted from the medium. This metabolic adjustment is
protein. For instance, the trp operon discussed in Chapter 4 achieved by repressing transcription of the lac operon until

282 CHAPTER 7 • Transcriptional Control of Gene Expression

 +1 (transcription start site) FIGURE 7-3 Regulation of transcription from the lac operon of
Promoter T E. coli. (Top) The transcription-control region, composed of - 100 base
lacZ pairs, includes three protein-binding regions: the CAP site, which
CAP site Operator binds catabolite activator protein; the lac promoter, which binds the
E. coli lac transcription-control regions a 70-RNA polymerase complex; and the lac operator, which binds lac
repressor. The lacZ gene encoding the enzyme 13-galactosidase, the
first of three genes in the operon, is shown to the right. (a) In the
absence of lactose, very little lac mRNA is produced because the lac
(a) repressor binds to the operator, inhibiting transcription initiation by
-lactose
a 70- RNA polymerase. (b) In the presence of glucose and lactose, lac
repressor binds lactose and dissociates from the operator, allowing
+glucose
(low cAMP) No mRNA transcription a 70-RNA polymerase to initiate transcription at a low rate. (c) Maximal
transcription of the lac operon occurs in the presence of lactose and
absence of glucose. In this situation, cAMP increases in response to
the low glucose concentration and forms the CAP-cAMP complex,
(b)
which binds to the CAP site, where it interacts with RNA polymerase
+lactose to stimulate the rate of transcription initiation. (d) The tetrameric lac
lacZ
+glucose repressor binds to the primary lac operator (0 1) and one of two
(low cAMP) Low transcription secondary operators (02 or 03) simultaneously. The two structures
are in equilibrium. [Part (d) adapted from B. Muller-Hill, 1998, Curr. Opm.
Microbiof. 1 :145.]

i
(c)

+lactose
- glucose
(highcAMP)
oAMP

~,
" ....,..
~:: : ;
'~ h~
lacZ
H ir
F
operator. As a result, the polymerase can bind ro the promoter
and initiate transcription of the lac operon. J:-Iowever, when
glucose also is present, the rate of transcription initiation (i.e.,
the number of times per minute different RNA polymerase
(d) molecules initiate transcription) is very low, resulting in syn-
thesis of only low levels of lac mRNA and the proteins en-
coded in the lac operon (Figure 7-3b). The frequency of
transcription initiation is low because the - 35 and -10 se-
quences in the lac promoter differ from the ideal a ' 0 -binding
sequences shown previously.
Once glucose is depleted from the medium and the intracel-
lular glucose concentration falls, E. coli cells respond by syn-
thesizing cyclic AMP, or cAMP. As the concentration of cAMP
increases, it binds to a site in each subunit of the dimeric CAP
protein, causing a conformational change that allows the pro-
tein ro bind to the CAP site in the lac transcription-control re-
lactose is present and allowing synthesis of only low levels of gion. The bound CAP-cAMP complex interacts with the
lac mRNA until the cytosolic concentration of glucose falls polymerase bound ro the promoter, greatly stimulating the rate
to low levels. Transcription of the lac operon under different of transcription initiation. This activation leads to synthesis of
conditions is controlled by, lac repressor and catabolite acti- high levels of lac mRNA and subsequently of the enzymes en-
vator protein (CAP) (also called CRP, for catabolite receptor coded by the lac operon (Figure 7-3c).
fJrotein), each of which binds to a specific DNA sequence in In fact, the lac operon is more complex than depicted in
the lac transcription-control region called the operator and the simplified model of Figure 7-3, parts (a)-(c). The tetra-
the CAP site, respectively (Figure 7-3, top). meric lac repressor actually binds to two sites simultane-
For transcription of the lac operon to begin, the a ~o sub- ously, one at the primary operator (/acO 1) that overlaps the
unit of the RNA polymerase must bind to the lac promoter at region of DNA bound by RNA polymerase at the promoter
the -35 and -10 promoter sequences. When no lactose is and at one of two secondary operators centered at +412
present, the lac repressor binds to the lac operator, which (/ac02) and -S2 (/ac03) (Figure 7-3d). The lac repressor
overlaps the transcription start site. Therefore, lac repressor tetramer is a dimer of dimers. Each dimer binds to one op-
bound to the operator site blocks a ~o binding and hence tran- erator. Simultaneous binding of the tetrameric lac repressor
scription initiation by RNA polymerase (Figure 7-3a). When to the primary lac operator 01 and one of the two secondary
lactose is present, it binds to specific binding sites in each sub- operators is possible because DNA is quite flexible, as we saw
unit of the tetrameric lac repressor, causing a conformational in the wrapping of DNA around the surface of a histone
change in the protein that makes it dissociate from the lac octamer in the nucleosomes of eukaryotes (Figure 6-29).

7.1 Control of Gene Expression in Bacteria 283

These secondary operators function to increase the local that support a high rate of transcription initiation have -10
concentration of lac repressor in the micro-vicinity of the and -35 sequences similar to the ideal promoter shown pre-
primary operator where repressor binding blocks RNA poly- viously and are called strong promoters. Those that support
merase binding. Since the equiLibrium of binding reactions a low rate of transcription initiation differ from this ideal
depends on the concentrations of the binding partners, the sequence and are called weak promoters. The lac operon, for
resulting increased local concentration of lac repressor in the instance, has a weak promoter. Its sequence differs from the
vicinity of 01 increases repressor binding to 01. There are ap- consensus strong promoter at several positions. This low in-
proximately 1 0 lac repressor tetramers per E. coli cell. Be- trinsic rate of initiation is further reduced by the lac repressor
cause of binding to 02 and 03, there is nearly always a lac and substantially increased by the cAMP-CAP activator.
repressor tetramer much closer to 01 than would otherwise
be the case if the 10 repressors were diffusing randomly Small Molecules Regulate Expression of Many
through the cell. If both 02 and 03 are mutated so that the
Bacterial Genes via DNA-Binding Repressors
lac repressor no longer binds to them with high affinity, re-
pression at the lac promoter is reduced by a factor of 70. and Activators
Mutation of only 02 or only 03 reduces repression two- Transcription of most E. coli genes is regulated by processes
fold, indicating that either one of these secondary operators similar to those described for the lac operon, although the
provides most of the stimulation of repression. detailed interactions differ at eath promoter. The general
Although the promoters for different E. coli genes exhibit mechanism involves a specific repressor that binds to the op-
considerable homology, their exact sequences differ. The pro- erator region of a gene or operon, thereby blocking tran-
moter sequence determines the intrinsic rate at which an scription initiation. A sma ll-molecu le ligand (or ligands)
RNA polymerase-<J complex initiates transcription of a gene binds to the repressor, controlling irs DNA-binding activity
in the absence of a repressor or activator protein. Promoters and consequently the rate of transcription as appropriate for

Sigma Factors of E. coli

Promoter Consensus

Sigma Factor Promoters Recognized -35 Region -10 Region

Housekeeping genes, most genes in TTGACA TATAAT _)

exponentially replicating cells

Stationary-phase genes and general TTGACA TATAAT

stress response

Induced by unfolded proteins in the TCTCNCCCTIGAA CCCCATNTA

cytoplasm; genes encoding chaperones
that refold unfolded proteins and
protea~e systems leading to the
degradation of unfolded proteins in
the cytoplasm

Activated by unfolded proteins in the GAACTT TCTGA

periplasmic space and cell membrane;
genes encoding proteins that restore
inregnty to the cellular envelope

Genes involved in flagellum assembly CTAAA CCGATAT

Genes required for iron uptake TTGGAAA GTAATG

-24 REGION -12 REGION

Genes for nitrogen metabolism CTGGI"\A TTGCA

and other functions

~OURCES: T. M. Gruber and C. A. Gross, 2003, Ann. Ret•. Microbtol. 57:441, '>. L McKnight and K. R. Yamamoto, ed~ .• Cold Spring Harbor
Laboratory Prc~s; R. L. Gour~e, \X'. Ross, and S. T. Rutherford, 2006, j. Bactenol. 188:4627; U.K. Sharma and D. Ch.mcrp, 2010,
FFMS M1crolnol. Reu. 34:646.

284 CHAPTER 7 • Transcriptional Control of Gene Expression

 the needs of the cell. As for the lac operon, many eubacterial NtrB. In response to low levels of glutamine, NtrB phosphor-
transcription-control regions contain one or more secondary ylates dimeric NtrC, which then binds to an enhancer up-
operators that contribute to the level of repression. stream of the ginA promoter. Enhancer-bound phosphorylated
Specific activator proteins, such as CAP in the lac operon, NtrC then stimulates the a 54 -polymerase bound at the pro-
also control transcription of a subset of bacterial genes that moter to separate the DNA strands and initiate transcription.
have binding sites for the activator. Like CAP, other activators Electron microscopy studies have shown that phosphory-
bind to DNA together with RNA polymerase, ~timulating lated NtrC bound at enhancers and a ; 4 -polymerasc bound at
transcription from a specific promoter. The DNA-binding ac- the promoter directly interact, forming a loop in the DNA be-
tivity of an activator can hC' modulated in response to cellular tween the binding sires (figure 7-4). As d1scussed later in this
needs by binding specific small-molecule ligands (e.g., cAMP) chapter, this activation mechanism resembles the predommant
or by post-translational modifications, such as phosphoryla- mechanism of transcriptional activation in eukaryores.
tion, that alter the conformation of the activator. NtrC has ATPase activity, and ATP hydrolysis is required
for activation of bound <r 54-polymerase by phosphorylated
Transcription Initiation from Some Promoters NtrC. Evidence for this is that mutants with an NtrC defec-
tive in ATP hydrolysis are invariably defective in stimulating
Requires Alternative Sigma Factors the a ' 4 -polymerase to melt the DNA strands at the transcrip-
Most E. coli promoters interact with a "'0-RNA polymerase, tion start site. It is postulated that ATP hydrolysis supplies the
the major initiating form of the bacterial enzyme. Transcrip- energy required for melting the DNA strands. In contrast,
tion of certain groups of genes, however, is initiated by E. coli the a ~ 0 -polymerase does not require ATP hydrolysis to sepa
RNA polymerases containing one of several alternative sigma rate the strands at a starr site.
factors that recognize different consensus promoter sequences
than cr70 does (Table 7-1 ). These alternative a-factors are re-
Many Bacterial Responses Are Controlled
quired for the transcription of sets of genes with related func-
tions such as those involved in the response to heat shock or by Two-Component Regulatory Systems
nutrient deprivation, motility, or sporulation in gram-positive As we have just seen, control of the E. coli gln.A gene depends
eubacteria. In E. coli there are six alternative a-factors in ad- on two proteins, NtrC and NtrB. Such two-component regu-
dition to the major "housekeeping" a-factor, a "'0 • The ge- latory systems control many responses of bacteria to changes
nome of the gram-positive, sporulating bacterium Streptomy- in their environment. At high concentrations of glutamine,
ces coelicolor encodes 63 a-factors, the current record, based glutamine binds to a sensor domain of NtrB, causing a con-
on sequence analysis of hundreds of eubacterial genomes. formational change in the protein that inhibits its histidine
Most are structurally and functionally related to a ~ 0 • But one kinase activity (Figure 7-Sa). At the same time, the regulatory
class is unrelated, represented in E. coli by a 54 • Transcription domain of NtrC blocks the DNA-binding domain from bind-
initiation by RNA polymerases containing a 70 -like factors is ing the ginA enhancers. Under conditions of low glutamine,
regulated by repressors and activators that bind ro DNA near glutamine dissociates from the sensor domain in the NtrB
the region where the polymerase binds, similar to initiation protein, leading to activation of a histidine kinase transmitter
by cr ~0 -RNA poly~erasc itself. domain in NtrB that transfers the -y-phosphatc of ATP to a
histidine residue (H) in the transmitter domain. This phos-
Transcription by o.s 4 -RNA Polymerase Is phohistidine then transfers the phosphate to an aspartic acid
residue (D) in the NtrC protein. This causes a conformational
Controlled by Activators That Bind Far
change in NtrC that unmasks the NtrC DNA-binding do-
from the Promoter main so that it can bind to the ginA enhancers.
The sequence of one E. coli sigma factor, a 54 , is distinctly Many other bacterial responses are regulated by two pro-
different from that of all the a ~ 0 -like factors. Transcription teins with homology to NtrB and NtrC (Figure 7-Sb). In
of genes by RNA polymerases containing a' 4 is regulated each of these regulatory systems, one protein, called a histi-
solely by activators whose binding sites in DNA, referred to dine kinase sensor, contains a latent histidine kinase trans-
as enhancers, generally are located 80-160 base pairs up- mitter domain that is regulated in response to environmental
stream from the start site. Even when enhancers are moved changes detected by a sensor domain. When activated, the
more than a kilobase away from a start site, a ' 4 -activators transmitter domain transfers the-y-phosphate of ATP to a his-
can activate transcription. tidine residue in the transmitter domain. The second protein,
The best-characterized a 54 -activator-the NtrC protein called a resfmnse regulator, contains a receiver domain homol-
(nitrogen regulatory protein C)-stimulates transcription of ogous ro the region of NtrC containing the aspartic acid
the ginA gene. ginA encodes the enzyme glutamine synthe- residue that is phosphorylated by activated NtrB. The response
tase, which synthesizes the amino acid glutamine from glu- regulator contains a second functional domain that is regu-
tamic acid and ammonia. The a' 4-RNA polymerase binds ro lated by phosphorylation of the receiver domain. In many
the ginA promoter but does not melt the DNA strands and cases this domain of the response regulator is a sequence-
initiate transcription until it is activated by NtrC, a dimeric specific DNA-binding domain that binds to related DNA se-
protein. NtrC, in turn, is regulated by a protein kinase called quences and functions either as a repressor, like the lac

7.1 Control of Gene Expression in Bacteria 285

·.
 (a) NtrC dimers a 54- RNA po lyme rase
Pair of phosphorylated
NtrC dimers
\ cr 54 - RNA

ginA
promoter

(b)

NtrC dimers cr 54 - RNA polymerase

EXPERIMENTAL FIGURE 7-4 DNA looping permits interaction (b) Drawing (left) and electron micrograph (right) of the same fragment
of bound NtrC and u 54- RNA polymerase. (a) Drawing (left) and preparation showing NtrC dimers and cr 54-RNA polymerase binding to
electron micrograph (right) of DNA restriction fragment with phosphor- each other with the intervening DNA forming a loop between them.
ylated NtrC dimers binding to the enhancer region near one end and [Micrographs from W. Suet al., 1990, Proc. Nat'/ Acad. Sci. USA 87:5505; courtesy
cr54-RNA polymerase bound to the ginA promoter near the other end. of S. Kustu.]

repressor, or as an activator, like CAP or NtrC, regulating the occurs is further controlled by a process called attenuatron
transcription of specific genes. However, the effector domain when the concentration of charged tRNArrp is sufficient to
can have other functions as well, such as controlling the di- support a high rate of protein synthesis. The first 140 nt of
rection in which the bacterium swims in response to a con- the Trp operon does not encode proteins required for trypto-
centration gradient of nutrients. Although all transmitter phan biosynthesis, but rather consists of a leader sequence as
domains are homologous (as are receiver domains ), the diagrammed in Figure 7-6a. Region 1 of this leader sequence
transmitter domain of a specific sensor protein will phos- contains two successive Trp codons. Region 3 can base-pair
phorylate only the receiver domains of specific response with both regions 2 and 4. A ribosome follows closely be-
regulators, allowing specific responses to different environ- hind the RNA polymerase, initiating translation of the leader
mental changes. Similar two-component histidyl-aspartyl peptide shortly after the 5' end of the Trp leader sequence
phospho-relay regulatory systems are also found in plants. emerges from the RNA polymerase. When the concentration
of tRNA1'P is sufficient to support a high rate of protein
synthesis, the ribosome translates through region 1 into re-
Control of Transcription Elongation gion 2, blocking the ability of region 2 to base-pair with re-
In addition to regulation of transcription initiation by acti- gion 3 as it emerges from rhe surface of the transcribing ·.
vators and repressors, expression of many bacterial operons RNA polymerase (Figure 7-6b, left). Instead, region 3 base-
is controlled by regulation of transcriptional elongation in pairs with region 4 as soon as it emerges from the surface of
the promoter-proximal region. This was first discovered in the polymerase, forming an RNA hairpin (see Figure 4-9a)
studies of Trp operon transcription in E. coli (Figure 4-13). followed hy several uracils, which is a signal for bacterial
Trp operon transcription is repressed by the Trp repressor RNA polymerase to pause transcription and terminate. As a
when the concentration of tryptophan in the cytoplasm is consequence, the remainder of the long Trp operon is not
high. But the low level of transcription initiation that still transcribed, and the cell does not waste energy required for

286 CHAPTER 7 • Transcriptional Control of Gene Expression

(a) Two-component system regulating response to low Gin FIGURE 7 -5 Two-component regulatory systems. At low
NtrB NtrC cytoplasmic concentrations of glutamine, glutamine dissociates from
Regulatory NtrB, resulting in a conformational change that activates a protein
Sensor domain
Ff®Dm domain kinase transmitter domain that transfers an ATP "(·phosphate to a

c:;
High [G inl

cC conserved histidine (H) in the transmitter domain. This phosphate is

then transferred to an aspartic acid (D) in the regulatory domain of
the response regulator NtrC. This converts Ntrc into its activated form,
His kinase DNA-binding
transmitter domain domain which binds the enhancer sites upstream of the ginA promoter
(Figure 7-4). (b) General organization of two-component histidyl-aspartyl
Low [Ginl

--·
phospho-relay regulatory systems in bacteria and plants. [Adapted from
Sensor His kinase DNA-binding
A. H. West and A.M. Stock, 2001, Trends Biochem. Sci. 26:369.]
domain transmitter domain domain

$ .._
~
..L_ATP
~ •
Gin enhancer
its synthesis or for the translation of the encoded proteins
when the concentration of tryptophan is high.
However, when the concentration of tRNArrp is not suffi-
cient to support a high rate of protein synthesis, the ribosome
(b) General two-component signaling system stalls at the rwo successive Trp codons in region 1 (Figure 7-6b,
Sensor Receiver right). As a consequence, region 2 base-pairs with region 3 as
domain domain soon as it emerges from the transcribing RNA polymerase.

~
I
Histidine
kinase
sensor
cC Response
Regulator
This prevents region 3 from base-pairing with region 4, so the
3-4 hairpin does not form and does not cause pausing by RNA
His kinase Effector polymerase or transcription termination. As a result, the pro-
domain
1Stimulus
domain teins required for tryptophan synthesis are translated by ribo-
somes that initiate translation at the start codons for each of
Sensor His kinase
domain domain these proteins in the long polycistronic Trp mRNA.
• .:Ibfa-C • Attenuation of transcription elongation also occurs at
$~ L ATP G Effect?r some operons and single genes encoding enzymes involved in

~ dar the biosynthes is of other amino acids and metabolites

through the function of riboswitches. Riboswitches form
RNA tertiary structures that can bind small molecules when
Response
they are present at sufficiently high concentration. In some

(a) trp leader RNA

Translation
start codon
1 w 1~ 100
s·l.....-..!...------t==r.::=J--c:cr--.J....-cx :r-c=:!:::J uuuuul3·
(b) Transl ation of trp leader

High tryptophan Low tryptophan

Ribosome covers region 2 Ribosome is stalled at trp codons in region 1

2-3 stem-loop
forms
UUUUU 3'
5' RNA polymerase
continues
transcription
'------------- ~
5' transcription
FIGURE 7-6 Transcription control by regulation of RNA At high concentrations of amino-acylated tRNA1'P, formation of the 3-4
polymerase elongation and termination in the E. coli Trp operon. stem-loop followed by a series of Us causes termination of transcrip-
(a) Diagram of the 140-nucleotide trp leader RNA. Colored regions are tion. At low amino-acylated tRNA1 'P, region 3 is sequestered in the 2-3
critical to the control of attenuation. (b) Translation of the trp leader stem-loop and cannot base-pair with region 4. In the absence of the
sequence begins from the 5' end soon after it is synthesized, while stem-loop structure required for termination, transcription of the trp
synthesis of the rest of the polycistronic trp mRNA molecule continues. operon continues. [See C Yanofsky, 1981, Nature 289:751 .]

7.1 Control of Gene Expression in Bacteria 287

cases this results in the formation of hairpin structures that
lead to early termination of transcription as in the Trp op- -y-phosphate of an ATP is transferred first to a histidine in
eron. When the concentration of these small-molecule li- the sensor protein and then to an aspartic acid in a second
gands is lower, the metabolites arc not bound by the RNA protein, the response regulator. The phosphorylated response
and alternative RNA structures form that do not induce regulator then performs a specific function in response to
transcription termination. As discussed below, although the the stimulus, such as binding to DNA regulatory sequences,
mechanism of transcriptional pausing and termination in eu- thereby stimulating or repressing transcription of specific genes
karyotes is tlifferent, regulation of promoter-proximal tran- (sec Figure 7-5).
scriptional pausing and termination has recently been • Transcription in bacteria can also be regu lated by control-
discovered to occur frequently in the regulation of gene ex- ling transcriptional elongation in the promoter-proximal re-
pression in multicellular organisms as well. gion. This can be regulated by ribosome binding to the nascent
mRNA as in the case of the Trp operon (Figure 7-6), or by
riboswitches, RNA tertiary structures that bind small mole-
KEY CO CEPTS of Section 7 1 cules, to determine whether a stem-loop followed by a string
of uracils forms, causing the bacterial RNA polymerase to
Control of Gene Expression in Bacteria pause and terminate transcription.
,
Gene expression in both prokaryotes and eukaryotes is
regulated primarily by mechanisms that control the initia-
tion of transcription. 7.2 Overview of Eukaryotic Gene Control
The first step in the initiation of transcription in E. coli is
In bacteria, gene control serves mainly to allow a single cell
binding of the u subunit complexed with an RNA polymerase
to adjust to changes in irs environment so that its growth and
to a promoter.
division can be optimized. In multicellular organisms, envi-
The nucleotide sequence of a promoter determines its ronmental changes also induce changes in gene expression.
strength, that is, how frequently different RNA polymerase An example is the response to low oxygen (hypoxia) in which
molecules can bind and initiate transcription per minute. a specific set of genes is rapidly induced that helps the cell
Repressors are proteins that bind to operator sequences survive under the hypoxic conditions. These include secreted
that overlap or lie adjacent to promoters. Binding of a re- angiogenic proteins that stimulate the growth and penetra-
pressor to an operator inhibits transcription initiation. tion of new capillaries into the surrounding tissue. However,
the most characteristic and biologically far-reaching purpose
The DNA-binding activity of most bacterial repressors is
of gene control in multicellular organisms is execution of the
modulated by small-molecule ligands. This allows bacterial
genetic program that underlies embryological development.
cells to regulate transcription of specific genes in response to
Generation of the man} different cell types that collectively
changes in the concentration of various nutrients in the envi-
form a multicellular organism depends on the right genes
ronment and metabolites in the cytoplasm.
being activated in the right cells at the right time during the
The lac operon and some other bacterial genes also are reg- tlcvelopmental period.
ulated by activator proteins that bind next to promoters and In most cases, once a developmental step has been taken by
mcrease the rate of transcription initiation by interacting di- a cell, it is nor reversed. Thus these decisions are fundamentally
rectly with RNA polymerase bound to an adjacent promoter. different from the reversible activation and repression of bacte-
70 rial genes in response to environmental conditions. In execut-
The major sigma factor in E. coli is u , but several other,
less abundant sigma factors are also found, each recognizing ing their genetic programs, many differentiated cells (e.g., skin
different consensus promoter sequences or interacting with cells, red blood cells, and antibody-producing cells) march
different activators. down a pathway to final cell death, leaving no progeny behind.
The fixed patterns of gene control leading to differentiation
Transcription initiation by all £. coli RNA polymerases,
serve the needs of the whole organism and not the survival of
except those containing cT 5\ can be regulated by repressors
an individual cell. Despite the differences in the purposes of gene
and activators that bind near the transcription start site (see
control in bacteria and eukaryotcs, two key features of tran-
Figure 7-3).
scription control first discovered in bacteria and described
Genes transcribed by o- 54 -RNA polymerase are regulated in the previous section also apply to eukaryotic cells. First,
by activators that bind to enhancers located ""1 00 base pairs protein-binding regulatory DNA sequences, or control cle-
upstream from the start sire. When the activator and u 54-RNA ments, are associated with gene~. Sel:ond, specific proteins that
polymerase interact, the DNA between their binding sites bind to a gene's regulatory sequences determine where tran-
forms a loop (see Figure 7-4). scription will start and either activate or repress its transcrip-
In two-component regulatory systems, one protein acts as tion. A fundamental difference between transcription control
a sensor, monitoring the level of nutrients or other compo- in bacteria and eukaryotes is a consequence of the association
nents in the environment. Under appropriate conditions, the of eukaryotic chromosomal DNA with histone octamers,
forming nucleosomes that associate into chromatin fibers that

288 CHAPTER 7 • Transcriptional Control of Gene Expression

' .
further associate into chromatin of varying degrees of conden- required for the development of certain regions of the brain
sation (Figures 6-29, 6-30, 6-32, and 6-33). Eukaryotic cells and spinal cord, and the cells in the pancreas that secrete
exploit chromatin structure to regulate transcription, a mecha- hormones such as insulin. As also mentioned earlier, hetero-
nism of transcription control that is not available to bacteria. zygous humans with only one functional Pax6 gene are born
As represented in Figure 7-2, in multicellular eukaryotes, many with aniridia, a lack of irises in the eyes (Figure 7-la). The Pax6
inactive genes are assembled into condensed chromatin, which gene is expressed from at least three alternative promoters that
inhibits the binding of RNA polymerases and general tran- function in different cell types and at different times during
scription factors required for transcription initiation. Activator embryogenesis (Figure 7-7a).
proteins bind to control elements near the transcription start Researchers often analyze gene control regions by prepar-
site of a gene as well as kilo bases away and promote chromatin ing recombinant DNA molecules that can contain a fragment
decondensation, binding of RNA polymerase to the promoter, of DNA to be rested with the coding region for a reporter
and transcriptional elongation through chromatin. Repressor gene that is easy to assay. Typical reporter genes are lucifer-
proteins bind to alternative control elements, causing conden- ase, which generates light that can be assayed with great sen-
sation of chromatin and inhibition of polymerase binding or sitivity and over many orders of magnitude of intensity using
elongation. In this section, we discuss general principles of eu- a luminometer. Other frequently used reporter genes encode
karyotic gene control and point out some similarities and dif- a green fluorescent protein, which can be visualized by fluo-
ferences between bacterial and eukaryotic systems. Subsequent rescence microscopy (see Figures 9-8d and 9-15) and E. coli
sections of this chapter will address specific aspects of eukary- [3-galactosidase, which generates an intensely blue insoluble
otic transcription in greater detail. precipitate when incubated with the colorless soluble lactose
analog X-gal. When transgenic mice are prepared (see fig-
ure 5-43) containing a [3-galactosidase reporter gene fused to
Regulatory Elements in Eukaryotic DNA Are 8 kb of DNA upstream from Pax6 exon 0, [3-galactosidase is
observed in the developing lens, cornea, and pancreas of the
Found Both Close to and Many Kilobases
embryo halfway through gestation (Figure 7-7b). Analysis of
Away from Transcription Start Sites transgenic mice with smaller fragments of DNA from this
Direct measurements of the transcription rates of multiple region allowed the mapping of separate transcription-control
genes in different cell types have shown that regulation of tran- regions regulating transcription in the pancreas and in the
scription, either at the initiation step or during elongation lens and cornea. Transgenic mice with other reporter gene
away from the transcription start site, is the most widespread constructs revealed additional transcription-control regions
form of gene control in eukaryotes, as it is in bacteria. In eu- (Figure 7-7a). These control transcription in the developing
karyotes, as in bacteria, a DNA sequence that specifies where retina and different regions of the developing brain (encepha-
RNA polymerase binds and initiates transcription of a gene is lon). Some of these transcription-control regions are in in-
called a promoter. Transcription from a particular promoter is trans berween exons 4 and 5 and berween exons 7 and 8. For
controlled by DNA-binding proteins that are functionally example, a reporter gene under control of the region labeled
equivalent to bacterial repressors and activators. Recent results retina in Figure 7-7a between exons 4 and 5 led to reporter
suggest that the intrinsic ability of the DNA sequence of a pro- gene expression specifically in the retina (Figure 7-7c).
moter region to associate with histone octamers also influences Control regions for many genes are found several hun-
transcription. Since transcriptional regulatory proteins can dreds of kilobases away from the coding exons of the gene.
often function either to activate or to repress transcription, de- One method for identifying such distant control regions is
pending on their association with other proteins, they are more to compare the sequences of distantly related organisms.
generally called transcription factors. The DNA control ele- Transcription-control regions for a conserved gene are also
ments in eukaryotic genomes that bind transcription factors often conserved and can be recognized in the background of
often are located much farther from the promoter they regulate a nonfunctional sequence that diverges during evolution. For
than is the case in prokaryotic genomes. In some cases, tran- example, there is a human DNA sequence =:::500 kilobases
scription factors that regulate expression of protein-coding downstream of the SALLJ gene that is highly conserved in
genes in higher eukaryotes bind at regulatory sites tens of thou- mice, frogs, chickens, and fish (Figure 7-8a). SALLJ encodes
sands of base pairs either upstream (opposite to the direction a transcription repressor required for normal development
of transcription) or downstream (in the same direction as tran- of the lower intestine, kidneys, limbs, and ears. When trans-
scription) from the promoter. As a result of this arrangement, genic mice were generated containing this conserved DNA se-
transcription of a single gene may be regulated by the binding quence linked to a [3-galactosidase reporter gene (Figure 7-Sb),
of multiple different transcription factors to alternative control the transgenic embryos expressed a very high level of the 13-
elements, directing expressiOn of the same gene in different galactosidase reporter gene specifically in the developing
types of cells and at different times during development. limb buds (Figure 7-Sc). Human patients with deletions in
For example, several separate transcription-control DNA this region of the genome develop with limb abnormalities.
sequences regulate expression of the mammalian gene encod- These results indicate that this conserved region directs tran-
ing the transcription factor Pax6. As mentioned earlier, Pax6 scription of the SALL 1 gene in the developing limb. Presum-
protein is required for development of the eye. Pax6 is also ably, other enhancers control expression of this gene in other

7.2 Overview of Eukaryotic Gene Control 289

 (a) /AAA
II II I I I I I I I I I
0 12 3 4 (l 5 6 7 8 9 10 11 12 13
D
Pancreas
•
Lens and
cornea
Telencephalon •
Retina Retina• • ••
Di- and rhombo-
encephalon

Transcript a

AAA
Transcript b

AAA
Transcript c

5 10 15 20 25 30 kb .·
FIGURE 7-7 Analysis of transcription-control reg ions of the mouse (b) (c)
Pax6 gene in transgenic mice. (a) Three alternative Pax6 promoters
are utilized at distinct times during embryogenesis in different specific
tissues ofthe developing embryo. Transcription-control regions
regulating expression of Pax6 in different tissues are indicated by
colored rectangles. The telencephalon-specific control region in intron
1 between exons 0 and 1 has not been mapped to high resolution. The
other control regions shown are 200-500 base pairs in length.
(b) 13-galactosidase expressed in tissues of a mouse embryo with a
13-galactosidase reporter transgene 10.5 days after fertilization. The
genome of the mouse embryo contained a transgene with 8 kb of DNA
upstream from ex on 0 fused to the 13-galactosidase coding region.
Lens pit (LP) is the tissue that will develop into the lens of the eye.
Expression was also observed in tissue that will develop into the
pancreas (P). (c) 13-galactosidase expression in a 13.5-day embryo with
a 13-galactosidase reporter gene under control of the sequence in part
(a) between exons 4 and 5 marked Retina. Arrow points to nasal and
temporal regions of the developing retina. Pax6 transcription-control some mushrooms (Figure 7-9). RNA polymerase I is insensitive
regions have also been found = 17 kb downstream from the 3' exon in to o:-amanitin, but RNA polymerase II is very sensitive-the
an intron of the neighboring gene. [Part (a) adapted from B. Kammendal drug binds near the active site of the enzyme and inhibits
et al., 1999, Dev. Bioi. 205:79. Parts (b) and (c) courtesy of Peter Gruss.) translocation of the enzyme along the DNA template. RNA
polymerase III has intermediate sensitivity.
Each eukaryotic RNA polymerase catalyzes transcription
of genes encoding different classes of RNA (Table 7-2). RNA
types of cells, where it functions in the normal development polymerase I (Pol 1), located in the nucleolus, transcribes
of ears, the lower intestine, and kidneys. After discussing the genes encoding precursor rRNA (pre-rRNA), which is pro-
proteins that carry out transcription in eukaryotic cells and cessed into 285, 5.85, and 185 rRNAs. RNA polymerase II1
eukaryotic promoters, we will return to a discussion of how (Pollli) transcribes genes encoding tRNAs, 55 rRNA, and an
such distant transcription-control regions, called enhancers, array of small, stable RNAs, including one im·olved in RNA
are thought to function. splicing (U6) and the RNA component of the signal-recognition
particle (SRP) involved in directing nascent proteins to the
endoplasmic reticulum (Chapter 13). RNA polymerase II
(Pol II) transcribes all protein-coding genes: that is, it functions
Three Eukaryotic RNA Polymerases Catalyze
in production of mRNAs. RNA polymerase II also produces
Formation of Different RNAs four of the five small nuclear RNAs that take part in RNA
The nuclei of all eukaryotic cells examined ~o far (e.g., verre- splicing and micro-RNAs (m1RNAs) involved in translation
brate, Drosophila, yeast, and plant cells) contain three differ- control as well as the closely related endogenous small inter-
ent RNA polymerases, designated I, II, and Ill. These enzymes fering RNAs (siRNAs) (see Chapter 8).
are eluted at different salt concentrations during ion-exchange Each of the three eukaryotic RNA polymerases is more
chromatography, reflecting the polymerases' various net complex than E. coli RNA polymerase, although their struc-
charges. The three polymerases also differ in their sensitivity tures arc similar (Figure 7- 1Oa, b). All three contain two large
to o:-amanitin, a poisonous cyclic octapeptide produced by subunits and I 0-14 smaller subunits, some of which are

290 CHAPTER 7 • Transcriptional Control of Gene Expression

(a) Comparative analysis EYPERIME TAL FIGURE '-8 The
human SALL 1 gene enhancer activates
expression of a reporter gene in limb b uds
of the developing mouse embryo.
(a) Graphic representation of the conservation
of DNA sequence in a region of the human
genome (from 50214-50220.5 kb of the
chromosome 16 sequence) ""500 kb down-
stream from the SALL 7 gene encoding a
zinc-finger transcription repressor. A region of
""500 bp of noncoding sequence is conserved
from fish to human. Nine hundred base pairs
including this conserved region were inserted
Frog into a plasmid next to the coding region for
E. coli [3-galactosidase. (b) The plasmid was
microinjected into a pronucleus of a fertilized
mouse egg and implanted in the uterus of a
I Fish pseudo-pregnant mouse to generate a
50217 50219 transgenic mouse embryo with the "reporter
Chromosome 16 (kb) gene" on the injected plasmid incorporated
into its genome (see Figure 5-43). (c) After 11.5
(b) Mouse egg microinjection (c) E11.5 reporter staining days of development, when limb buds
develop, the fixed and permeabilized embryo
was incubated in X-gal, which is converted by
[3-galactosidase into an insoluble, intensely
blue compound. The ""900-bp region of
human DNA contained an enhancer that
stimulated strong transcription of the
Forelimb [3-galactosidase reporter gene in limb buds
bud specifically. [From the VISTA Enhancer Browser,
http://enhancer.lbl.gov. Parts (b) and (c) courtesy of
Hindlimb - -o-:..l Len A. Pennacchio, Joint Genome Institute, Lawrence
bud
Berkeley National Laboratory.]

common betweeh two o r all three of the polymerases. The acterized. In addition, the three-dimensional structure of yeast
best-characterized eu karyotic RNA polymerases are from RNA polymerase II has been determined (Figure 7-lOb, c).
the yeast Saccharomyces cerevisiae. Each of the yeast genes The t h ree nuclear RNA polymerascs from all eukaryotes so
encoding t he polymerase subunits has been subjected to far examined arc very similar to those of yeast. Plants con-
gene-knockout mutations and the resulting phenotypes char- tain two additional nuclear RNA polymerases (RNA poly-
merascs IV and V), which are closely related to their RNA
[NaCI]~
polymerase II but have a unique large subunit and some ad-
ditional unique subunits. These function in transcriptional
Poll

Total
protein
Q)
i
u
c: EXPERIMENTAL FIGURE 7-9 Column chromatography
Q)
Cl) c: separates and identifies the three eukaryotic RNA polymerases,
Q) -

i
Cl)
O.·E
c:
- E
~ ~
"'
each with its own sensitivity to o-amanitin. A protein extract from
the nuclei of cultured eukaryotic cells was passed through a DEAE

i~
C1l l:l
5ephadex column and adsorbed protein eluted (black curve) with a
.r;_
solution of constantly increasing NaCI concentration. Fractions from
c: >
c>-E
·- "' .,e> the eluate were assayed for RNA polymerase activity (red curve). At a
~ <( <(::t
~z
o.. cc
,.:: concentration of 1 f.Lg/ ml, a-amanitin inhibits polymerase II activity but
0 has no effect on polymerases I and Ill (green shading). Polymerase Ill
is inhibited by 10 f.Lg/ml of a-amanitin, whereas polymerase I is
unaffected even at this higher concentration. [SeeR. G. Roeder, 1974,
Fraction number J. Bioi. Chern. 249:241.]

7.2 Overview of Eukaryotic Gene Control 291

*tHtm
Polymerase
Classes of RNA Transcribed by the Three Eukaryotic Nuclear RNA Polymerases and Their Functions

RNA Transcribed RNA Function

RNA pol ymerase I Pre r-RNA (285, 185, 5.85 rRNAs) Ribosome components, protein synthesis

RNA polymerase II mRNA Encodes protein

snRNAs RNA splicing
si RNAs Chromatin-mediated repression, translation control
miRNAs Translation control

RNA polymerase III tRNAs Protein synthesis

SS rRNA Ribosome component, protein synthesis
snRNA U6 RNA splicing
75 RNA Signal-recognition panicle for insertion of pol ypeptides into
the endoplasmic reticulum
Other stable short RNAs Various functio ns, unknown for many ,

repression d irected by nuclea r siRNAs in p lants, discussed su bu nits, respectively (Figure 7-10 ). Each of the eukaryotic
toward the end of this chapter. p olymerases also contains an w-like and two nonidentical
The two large subun its of a ll three eukaryotic RNA poly- a -like subun its (Figure 7- 11 ). The extensive sim ila rity in the
merases (and RNA polymerases IV and V of p lants) are re- structu res of t hese core subunits in RNA polymerases from
lated to each other and are similar to the E. coli 13' and j3 various sources indicates th at this e nzyme arose early in

(a ) Bacterial RNA polymerase (b) Yeast RNA poly merase II (c) Yeast RNA polymerase II

FIGURE 7-10 Comparison of three-dimensional structures of position marked with a red arrow. (RPB is the abbreviation for "RNA
bacterial and eukaryotic RNA polymerases. (a, b) These Ca trace polymerase 8," which is an alternative way of referri ng to RNA
models are based on x-ray cryst allographic analysis of RNA polymerase polymerase II.) DNA entering t he polymerases as they transcribe to the
from the bacterium T. aquaricus and core RNA polymerase II from right is diagrammed. (c) Space-filling model of yeast RNA polymerase II
5. cerevisiae. (a) The five subunits of t he bacterial enzyme are distin- incl uding subu nits 4 and 7. These subunits extend from the core
guished by color. Only the N-terminal domains of the a subunits are portion of the enzyme shown in (b) near the region of the ( -termina l
included in this model. (b) Ten of t he 12 subunits constituting yeast domain of the large subunit. [Part (a) based on crystal structures from
RNA polymerase II are shown in this model. Subunits that are simi lar in G. Zhang et al., 1999, Ce// 98:811 . Part (b) adapted from P. Cramer et al., 200 1,
conformation to those in the bacterial enzyme are shown in the same Science 292:1863. Part (c) from K. J. Armache et al., 2003, Proc. Nat'/ Acad. Sci. USA
colors. The ( -terminal domain of the large subunit RPBl was not 100:6964, and D. A. Bushnell and R. D. Kornberg, 2003, Proc. Nat'/ Acad. Sci. USA
observed in the crystal structure, but it is known to extend from the 100:6969.]

292 CHAPTER 7 • Transcriptional Co ntrol of Gen e Expression

,'

FIGURE 7-11 Schematic representation ofthe subunit structure

E. coli core RNA polymerase (u2 ~jrw)
of the E. coli RNA core polymerase and yeast nuclear RNA polymer-
ases. All three yeast polymerases have five core subunits homologous
to the [3, (3 ', two a, and w subunits of E. coli RNA polymerase. The
largest subunit (RPBl) of RNA polymerase II also contains an essential
(-terminal domain (CTD). RNA polymerases I and Ill contain the same
two nonidentical a-like subunits, whereas RNA polymerase II contains
Eukaryotic RNA polymerases
two other nonidentical a-like subunits. All three polymerases share the
II Ill same w-like subunit and four other common subunits. In ncldition, each
w-and yeast polymerase contains three to seven unique smaller subunits.

~-like subunits

a-like subunits .... C> •o .... C> transcription elongation factor called DSIF, discussed later,

•0 •0 •0
associates with the elongating polymerase, holding the clamp
w-like subunit in its closed conformation. As a consequence, the polymerase
is extraordinarily processive, which is to say that it continues
to polymerize ribonuclcotides until it terminates transcription.
After termination and RNA is released from the exit channel,
Common 0 0 0 the clamp can swing open, releasing the enzyme from the tem-
subunits plate DNA. This can explain how human RNA polymerase II
D D D
• • •
can transcribe the longest human gene encoding dystrophin
(D MD), which is =2 million base pairs in length, without
dissociating and terminating transcription. Since transcription
Additional elongation proceeds at 1-2 kb per minute, transcription of the
enzyme-specific +5 +3 +7
subunits DMD gene requires approximately one day!.
Gene-knockout experiments in yeast indicate that most
of the subunits of the three nuclear RNA polymerases are
essential for cell viability. Disruption of the few polymerase
evolution and was largely conserved. This seems logical for subunit genes that are not absolutely essential for viability
an enzyme catalyzing a process so fundamental as copying (e.g., subunits 4 and 7 of RNA polymerase ll) nevertheless
RNA from DNA. In addition to their core subunits related results in very poorly growing cells. Thus, all the su bunits
to the E. coli RNA polymerase subunits, all three yeast RNA arc necessary for eukaryotic RNA polymerases ro function
polymerases contain four additional sma ll subunits, com- normally. Archaea, like eubacteria, have a single type of
mon to them but not to the bacterial RNA polymerase. Fi- RNA polymerase involved in gene transcription. But the ar-
nally, each eukaryotic nuclear RNA polymerase has several chaeal RNA polymerases, like the eukaryotic nuclear RNA
enzyme-specific subunits that are not present in the other polymerases, have on the order of a dozen subunits. Archaea
two nuclear RNA polymerases (Figure 7-11 ). Three of these also have related general transcription factors, discussed
additional subunits of Pol I and Pol III are homologous to later, consistent with their closer evolutionary relationship
the three additional Pol Il-specific subunits. The other two to eukaryotes than to eubactcria (Figure 1-la).
Pol !-specific subunits are homologous to the Pol II general
transcription factor THIF, discussed later, and the four ad-
The Largest Subunit in RNA Polymerase II Has
ditional subunits of Pol IU are homologous to the Pol II gen-
eral transcription factors TFIIF and TFIIE. an Essential Carboxyl-Terminal Repeat
The clamp domain of RPBI is so designated because it has The carboxyl end of the largest subunit of RNA polymerase II
been observed in two different positions in crystals of the free (RPBl ) contains a stretch of seven amino acids that is nearly
enzyme (Figure 7-12a) and a complex that mimics the elon- precisely repeated multiple times. Neither RNA polymerase I
gating form of the enzyme (Figure 7-12b, c). This domain nor Ill contains these repeating units. This heptapcptide repeat,
rotates on a hinge that is probably open when downstream with a consensus sequence of Tyr-Ser-Pro-Thr-Ser-Pro-Scr, is
DNA (dark blue template strand, cyan nontemplate strand) is known as the carboxyl-terminal domain (CTD) (Figure 7-lOb,
inserted into this region of the polymerase, and then swings extending from the reel arrow). Yeast RNA polymera se II
shut when the enzyme is in its elongation mode. RNA base- contains 26 or more repeats, vertebrate enzymes have 52 re-
paired to the template strand is red in Figure 7-12b and c. It peats, and an intermediate number of repeats occur in RNA
is postulated that when the 8-9 base-pair RNA-DNA hybrid polymerase II from nearly all other eukaryotes. The CTD is
region near the active site (Figure 7-12c) is bound between critical for viability, and at least 10 copies of the repeat must
RBP l and RBP2 and nascent RNA fills the exit channel, the be present for yeast to survive.
clamp is locked in its closed position, anchoring the poly- In vitro experiments with model promoters first showed
merase to the downstream double-stranded DNA. Also, a that RNA polymerase II molecules that initiate transcription

7.2 Overview of Eukaryotic Gene Control 293

(a) Free RNA polymerase II (b) T ranscribing RNA polymerase II (c) Side view
Clamp
domain

·.

,
FIGURE 7- 12 The clamp domain of RPBI. The structures of the free phosphodiester bond formation is shown in green. Wall is the domain of
(a) and transcribing (b) RNA polymerase II differ mainly in the position RPB2 that forces the template DNA entering the jaws of the polymerase
of a clamp domain in RPBl (orange). which swings over the cleft to bend before it exits the polymerase. The bridge a helix shown in
between the jaws of the polymerase during formation of t he transcrib- green extends across the cleft in the polymerase (see Figure 7-1 Ob) and
ing complex, trapping the template DNA strand and transcript. Binding is postulated to bend and straighten as the polymerase translocates
of the clamp domain to the 8- 9-base-pair RNA-DNA hybrid may help one base down the template strand. The nontemplate strand is
couple clamp closure to the presence of RNA, stabilizing the closed, thought to form a flexible single-stranded region a~ove the cleft (not
elongating complex. RNA is shown in red, the template DNA strand in shown) extending from three bases downstream of the template
dark blue, and the downstream nontemplate DNA strand in cyan in this base-paired to the 3' base of the growing RNA and extending to the
model of an elongating complex. (c) The clamp closes over the template strand as it exits the polymerase, where it hybridizes with the
incoming downstream DNA. This model is shown with portions of RBP2 template strand to generate the transcription bubble. [Adapted from
that form one side of the cleft removed so that the nucleic acids can A. L. Gnatt et al., 2001 , Science 292:1876.)
be better visualized. The Mg 2 ion that participates in catalysis of

have an unphosphorylated CTD. Once the polymerase initi- scription. The large chromosomal "puffs" induced at this
ates transcription and begins to move away from the pro- time in development are regions where the genome is very
moter, many of the serine and some tyrosine residues in the actively transcribed. Staining with antibodies specific for
CTD are phosphorylated. Analysis of polytene chromo- the phosphorylated or unphosphorylated CTD demon-
~omes from Drosophila salivary glands prepared just before strated that RNA polymerase II associated with the highly
molting of the larva, a time of active transcription, indicate transcribed puffed regions contains a phosphorylated CTD
that the CTD also is phosphorylated during in vivo tran- (Figure 7-13).

EXPE 1M ENTAL FIGURE 7-13 Antibody staining demon-

strates that the carboxyl-terminal domain (CTD) of RNA poly-
merase II is phosphorylated during in vivo transcription.
Salivary-gland polytene chromosomes were prepared from Drosophila
larvae just before molting. The preparation was treated with a rabbit
antibody specific for phosphorylated CTD and with a goat antibody
specific for unphosphorylated CTD. The preparation then was stained
with fluorescein-labeled anti-goat antibody (green) and rhodamine-
labeled anti-rabbit antibody (red). Thus polymerase molecules with an
unphosphorylated CTD stain green, and those with a phosphorylated
CTD stain red. The moltinq hormone ecdysone induces very high ratP~
of transcription in the puffed regions labeled 74EF and 75B; note that
only phosphorylated CTD is present in these regions. Smaller puffed
regions transcribed at high rates also are visible. Non puffed sites that
stain red (up arrow) or green (horizontal arrow) also are indicated, as is
a site staining both red and green, producing a yellow color (down
arrow). [From J. R. Weeks et al., 1993, Genes Dev. 7:2329; courtesy of J. R. Weeks
and A. L. Greenleaf.)

294 CHAPTER 7 • Transcriptional Control of Gene Expression

 found in eukaryotic protein-coding genes and some tech-
KEY CONCEPTS of Section 7.2 niques used to identify them .
Overview of Eukaryotic Gene Control
• The primary purpose of gene control in mu lticellular organ- RNA Polymerase II Initiates Transcription
isms is the execution of precise developmental decisions so at DNA Sequences Corresponding
that the proper genes are expressed in the proper cells during to the 5' Cap of mRNAs
embryologic development and cellular differentiation.
In vitro transcription experiments using purified RNA poly-
Transcri ptiona 1 control is the primary means of regulating merase II, a protein extract prepared from the nuclei of cul-
gene exp ression in eukaryotes, as it is in bacteria. tured cells, and DNA templates containing sequences
In eukaryotic genomes, DNA transcription-control elements encodmg the 5' ends of mRNAs for a number of abundantly
may be located many kilobases away from the promoter they expressed genes revea led that the transcripts produced al-
regulate. Different control regions can control transcription of ways contained a cap structure at their 5' ends identical with
the same gene in different cell types. tha t present at the 5' end of the spliced mRNA expressed
Eukaryotes contain three types of nuclear RNA polymer- from the gene (see Figure 4-14). In these experiments, the 5' cap
ases. All three contain two large and three smaller core sub- was added to the 5' end of the nascent RNA by enzymes in
units with homology to the 13', ~,a, and w su bunits of E. co/1 the nuclear extract, which can only add a cap to an RNA
RNA polymerase, as well as several additional small sub- that has a 5' tri-or diphosphate. Because a 5' end generated by
units (see Figure 7-11). cleavage of a longer RNA would have a 5' monophosphate, it
would not be capped. Consequently, researchers concluded
• RNA polymerase I synthesizes only pre-rRNA. RNA poly- that the capped nucleotides generated in the in vitro tran-
merase II synthesizes mRNAs, some of the small nuclear RNAs scription reactions must have been the nucleotides with
that participate in mRNA splicing, micro-RNAs (miRNAs) which transcription was initiated. Sequence analysis revealed
that regulate translation of complementary mRNAs, and small that, for a given gene, the sequence at the 5' end of the RNA
interfering RNAs (siRNAs) that regulate the stabiliry of com- transcripts produced in vitro is the same as that at the 5' end
plementary mRNAs. RNA polymerase Ill synthesizes tRNAs, of the mRNAs isolated from cells, confirming that the capped
55 rRNA, and several other relatively short, stable Rl"'lAs (see nucleotide of eukaryotic mRNAs coincides with the tran-
Table 7-2). scription start site. Today, the transcription start site for a
• The carboxyl-terminal doma in (CTD) in the largest sub- newly characterized mRNA generally is determined simply
unit of RNA polymerase II becomes phosphorylated during by identifying the DNA sequence encoding the 5' -capped nu-
transcription initiation and remains phosphorylated as the cleotide of the encoded mRNA.
enzyme transcribes the template.

The TATA Box, Initiators, and CpG Islands

Function as Promoters in Eukaryotic DNA
7.3 RNA Polymerase II Promoters Several different DNA sequences can function as promoters
for RNA polymerase II, directing the polymerase where to
and General Transcription Factors initiate transcription of an RNA complementary to the template
The mechanisms that regulate transcription initiation and strand of a doub le-stranded DNA. These include T ATA
elongation by RNA polymerase II have been studied exten- boxes, initiators, and CpG islands.
sively, because this is the polymerase that transcribes mRNAs.
Transcription initiation and elongation by RNA polymerase TATA Boxes The first genes to be sequenced and studied
11 are the initial biochemical processes required for the ex- through in vitro transcription systems were viral genes and
pression of protei n-coding genes and are the steps in gene cellular protein-coding genes that are very actively tran-
expression that are most freq uently regu lated to determine scribed either at particular times of the cell cycle or in spe-
when and in which cells specific proteins are synthesized. As cific differentiated cell types. In all these highly transcribed
noted in the previous section, expression of eukaryotic protein- genes, a conserved sequence called the TATA box was found
coding genes is regulated by multiple protein-binding DNA :.:26-31 base pairs upstream of the transcription start site
sequences, genericall y referred to as transcription-control re- (Figure 7-14). Mutagenesis studies have shown that a single-
gions. These include promoters, which determine where tran- base change in this nucleotide sequence drastically decreases
scription of the DNA template begins, and other types of in vitro transcription by RNA polymerase II of genes adja-
control elements located near transcription start sites as well cent to a TAT A box. If the base pairs between the TATA
as seq uences located far from the genes they regulate, which box and the normal transcription start site arc deleted, tran-
control the type of cell in which the gene is transcribed and scription of the altered, shortened template begins at a new
how freq uentl y it is transcribed. In this section, we take a site :.:25 base pairs downstream from the TATA box. Conse-
closer look at the properties of various control elements quently, the TATA box acts simi larly to an E. coli promoter

7.3 RNA Polymerase II Promoters and General Transcription Factors 295

 = -37 to -32 = -31 to -26 -2 to +4 +28 to +32 In mammals, most Cs followed by a G that are not as-
sociated with CpG island promoters are methylated at posi-
tion 5 of the pyrimidine ring (5-methyl C, represented C~ 1 ";
see figure 2-17). CG sequences are thought to be underrep-
resented in mammalian genomes because spontaneous de-
lnr OPE
Initiator Downstream core
am ination of 5-methyl C generates thymidine. Over the time
TFIIB
Recognition Drosophila + 1 G T promoter element scale of mammalian evolution, this is thought to have led to
TCAT T C
element the conversion of most CGs to TG by DNA-repair mecha-
GGGCGCC T nisms. As a consequence, the freq uency of CG in t he human
CCA Mammals YYANA YY
genome is o nly 21 percent of the expected frequency if Cs
FIGURE 7-14 Core promoter elements of non-CpG island were randomly followed by a G. However, the Cs in active
promoters in metazoans. The sequence of the strand with the 5' end CpG isla nd promoters are unmethylated . Consequently,
at the left and the 3' end at the right is shown. The most frequently when they deaminate spontaneously, they are converted to U,
observed bases in TATA-box promoters are shown in largerfont. A - l is a base that is recognized by DNA repair enzymes and con-
the base at which transcription starts, Y is a pyrimidine (Cor T), N is any verted back to C. As a result, the frequency of CG sequences
of the four bases. [Adapted from S. T. Smale and J. T. Kadonaga, 2003, Ann. Rev.
in CpG island promoters is close to that expected if C were
Biochem. 72:449.]
followed by any of the other three nucleotides randomly.
CG-rich sequences are bound by histone octamers more
weakly than CG-poor sequences because more energy is
to position RNA polymerase II for transcription initiation required to bend them into the small-diameter loops required
(see Figure 4-12). to wrap around the histone octamer forming a nucleosome
(Figure 6-29) . As a consequence, CpG islands coincide with
Initiator Sequences Instead of a TATA box, some eukaryotic nucleosome-free regions of DNA. Much remains to be
genes contain an alternative promoter element called an initia- learned about the molecular mechanisms that control tran-
tor. Most naturally occurring initiator elements have a cyto- scription from CpG island promoters, but a current hypoth-
sine (C) at the -1 position and an adenine (A) residue at the esis is that the general transcription factors discussed in the
transcription start site ( + 1 ). Directed mutagenesis of mamma- next section can bind to them because CpG islands exclude
lian genes with an initiator-containing promoter revealed that nucleosomes.
the nucleotide sequence immediately surrounding the start site
determines the strength of such promoters. Unlike the con- Divergent Transcription from CpG Island Promoters Another
served TATA box sequence, however, only an extremely de- remarkable feature of CpG islands is that transcription is initi-
generate initiator consensus sequence has been defined: ated in both directions, even though only transcription of the
sense strand yields an mRNA. By a mechanism(s) that remains
(~') Y-Y-A+ 1-N-T/A-Y-Y-Y (3') to be elucidated, most RNA polymerase II molecules tra nscrib-
ing in the "wrong" direction, i.e., transcribing the non-sense
where A 1 is the base at which transcription starts, Y is a py- strand, pause or terminate by =1 kb from the transcription
rimidine (Cor T), N is any of the four bases, and T/A is Tor start site. This was discovered by taking advantage of the sta-
A at position +3. As we shall see after discussing general tran- bility of the elongation complex, presumably conferred by the
scription factors required for RNA polymerase II initiation, RNA polymerase II clamp domain when an RNA-DNA hybrid
other specific DNA sequences designated BRE and DPE can is bound near the active site (Figure 7-l2b, c).
he bound by these proteins and influence promoter strength Nuclei were isolated from cultured human cells and incu-
(Figure 7-14). bated in a buffered solution containing a concentration of
salt and mi ld detergent that removes RNA polymerases ex-
CpG Islands Transcription of genes with promoters contain- cept for those in the process of elongation because of their
ing aT ATA box or initiator element begins at a well-defined stable associa tion with template DNA. Nucleotide triphos-
initiation site. However, transcription of most protein-coding phates were then added with UTP substituted by bromo-
genes in mammals (=60-70 percent) occurs at a lower rate UTP containi ng uracil with a Br atom at the 5 position on
than TATA box and Initiator-containing promoters, and ini- the pyrimidine ring (Figure 2-17). The nuclei were then incu-
tiates at several alternative start sites within regions of bated at 37 oc long enough for "-'100 nucleotides to be po-
=1 00-1000 base pairs that have an unusually high frequency lymerized by the RNA polymerase II (Pol ll) molecules that
of CG sequences. Such genes often encode protein~ that are were in the process of transcription elongation at the time
not required in large numbers (e.g., enzymes involved in the nuclei were isolated. RNA was then isolated and RNA
basic metabolic processes required in all cells, often ca lled containing bromo-U was immunoprecipitated with antibody
"housekeeping genes"). These promoter regions are called specific for RNA labeled w ith bromo-U. Thirty-three nucleo-
CpG islands (where "p" represents the phosphate between tides at the 5' ends of these RNAs were then sequenced by
the C and G nucleotides) because they occur relatively rarely massively parallel DNA sequencing of reverse transcripts,
in the genome sequence of mammals. and the sequences were mapped on the human genome.

296 CHAPTER 7 • Transcriptional Control of Gene Expression

~ ' •t
 ---------------------

100 sequences reads to the left of 0 and blue sequence reads to the
Q)
--+50bp right of 0), indicating that there is a low level of transcription
"'
Cl)
.0
80
from seemingly random sites throughout the genome. These
_Q 60
:.;;< recent discoveries of divergent transcription from CpG island
Q; 40 promoters and low-level transcription of most of the genomes
c.
"'
"'0 20 of eukaryotes have been a great surprise to most researchers.
Cl)

~
0
Q)
(.) Chromatin lmmunoprecipitation The technique of chroma-
c
Q)
:>
-20 tin immunuprecipitation outlined in Figure 7-16a provided
0"
Q)
-40 additional data supporting the occurrence of divergent tran-
C/)
scription from most CpG island promoters in mammals. The
-60
-3 -2 -1 0 2 3 data from this analysis are reported as the number of times a
Distance relative to TSS (kb) specific sequence from this region of the genome was identi-
EXPERIMENTAL FIGURE 7 · 15 Analysis of elongating RNA
fied per million total sequences analyzed (Figure 7-16b). At
polymerase II molecules in human fibroblasts. Nuclei from cultured
divergently transcribed genes, such as the Hsd1 7b 12 gene en-
fibroblasts were isolated and incubated in a buffer with a non ionic coding an enzyme involved in intermediary metabolism, two
detergent that prevents RNA polymerase II from initiating transcrip- peaks of immunoprecipitated Dl'\A were detected, corre-
tion. Treated nuclei were incubated with ATP, CTP, GTP, and Br-UTP for sponding to Pol II transcribing in the sense and antisense di-
5 minutes at 30 °(, a time sufficient to incorporate "" 100 nucleotides. rections. However, Pol II was only detected > 1 kb from the
RNA was then isolated and fragmented to "" 100 nucleotides by start site in the sense direction. The number of counts per
controlled incubation at high pH. Specific RNA oligonucleotides were million from this region of the genome was very low because
ligated to the 5' and 3' ends of the RNA fragments, which were then the gene is transcribed at low frequency. However, the num-
subjected to reverse transcription. The resulting DNA was amplified ber of counts per million at the start-site regions for both
by polymerase chain reaction and subjected to massively parallel DNA sense and antisense transcription was much higher, reflecting
sequencing. The sequences determined were aligned to the transcrip- the fact that Pol II molecules had initiated .transcnption in
tion start sites (TSS) of all known human genes and the number of
both directions at this promoter, but paused before transcrib-
sequence reads per kilobase of total sequenced DNA was plotted
ing >500 base pairs from the start sites in each direction. In
for 10 base-pair intervals of sense transcripts (red) and antisense
contrast, the Rp/6 gene encoding a large ribosomal subunit
transcripts (blue). See text for discussion. [From L. J. Core, J. J. Waterfall,
and J. T. Lis, 2008, Science 322:1845.]
protein that was abundantly transcribed in these proliferating
cells was transcribed almost exclusively in the sense direction.
The number of sequence counts per million > 1 kb down-
Figure 7-15 shows a plot of the number of sequence reads stream from the transcription start site was much higher, re-
per kilobase of total BrU-labeled RNA relative to the major flecting the high rate of transcription of this gene.
transcription start sites (TSS) of all currently known human Transcription start-site-associated RNAs (TSSa RNAs,
protein-coding genes. The results show that approximately red and blue arrowheads at the bottom of Figure 7-16b) rep-
equal numbers of RNA polymerase molecu les transcribed resent sequences of short RNAs isolated from these cells,
most promoters (mostly CpG island promoters) in both the thought to result from degradation of nascent RNAs re-
sense direction (red, plotted upward to indicate transcription leased from paused Pol II molecules that terminate. Note
in the sense direction) and the antisense direction (blue, plot- that they include transcripts of both the sense (blue arrows)
ted downward to represent transcription in the opposite, an- and antisense (red arrows) from the divergently transcribed
tisense direction). A peak of sense transcripts was observed at gene, whereas only sense TSSa RNAs were found for the
""+50 relative to the major transcription start site (TSS), in- unidirectionally transcribed gene. The observation of these
dicating that Poll! pauses'in the +50 to + 250 reg1on before TSSa RNAs from CpG island promoters further support the
elongating further. A peak at -250 to -500 relative to the conclusion that they are transcribed in both directions.
major sense transcription start site of Pol II transcribing in the
antisense direction also was observed, revealing paused RNA
polymerase II molecules at the other end of the nucleosomes- General Transcription Factors Position
free regions in CpG island promoters. Note that the number
RNA Polymerase II at Start Sites
of sequence reads, and therefore the number of elongating
polymerases, is lower for polymerases transcribing in the an- and Assist in Initiation
tisense direction more than 1 kb from the transcription start Initiation by RNA polymerase II requires several initiation
compared to polymerases transcribing more than 1 kb from factors. These initiation factors position Pol II molecules at
the transcription start site in the sense direction. The molecu- transcription start sites and help to separate the DNA strands
lar mechanism(s) accounting for this difference is currently so that the template strand can enter the active site of the
an intense area of investigation. Note that a low number of enzyme. They are called general transcription factors be-
sequence reads was also observed transcribing in the "wrong" cause they are required at most, if not all, promoters tran-
direction upstream of the major transcription start sites (red scribed by RNA polymerase II. These proteins are designated

7.3 RNA Polymerase II Promoters and General Transcript ion Factors 297
 (a)
IJ Treat living cells or tissues with a
membrane permeable cross-linker
such as formaldehyde

fJ Sonicate to shear cellular

Antibody to Pol II
1 chromatin to short fragments
and add antibody to Pol II

' P;ym"'" Eloogot;oo ;oh;Mo• \ ' ~

I 5'

'
Paused
Antibody to Pol II

polymerase-. ~
~~~
.
111
Elongation inhibitor
lmmunoprecipitate to isolate
Pol II cross-linked to DNA

1.a Reverse cross-linking, isolate DNA and subject

to massively paral lel DNA sequencing

~
(b) Bidirectional initiation Unidirectional initiation
c::

]
.2 30f
·-
RNA Pol II~
c
:::J
0 I
,JA, L I J
u 93955000 Chrom Position 93962000
J,,,L
121467000 Ch rom Position 121463000

TSSa RNA Hsd17b12

4
•
f7 //
~
Rp/6 .... - I //
_.)

EXPERIMENTAL FIGURE 7 · 16 Chromatin immunoprecipita- sequencing. (b) Results from DNA sequencing of chromatin from
tion technique. (a) Step 0 : Live cultured cells or tissues are incubated mouse embryonic stem cells immunoprecipitated with antibody to
in 1% formaldehyde to covalently cross-link protein to DNA and RNA polymerase II are shown for a gene that is divergently transcribed
proteins to proteins. Step 6 : The preparation is then subjected to (left) and a gene that is transcribed only in the sense direction (right).
sonication to solubilize and shear chromatin to fragments of 200 to 500 Data are plotted as the number of times a DNA sequence in a SO-base-
base pairs of DNA. Step il: An antibody to a protein of interest, here pair interval was observed per million base pairs sequenced. The region
RNA polymerase II, is added, and DNA covalently linked to the protein encoding the 5' end of the gene is shown below, with exons shown as
of interest is immunoprecipitated. Step EJ: The covalent cross-linking is rectangles and introns as lines. TSSa RNAs (red and blue arrowheads)
then reversed and DNA is isolated. The isolated DNA can be analyzed represent RNAs of = 20-50 nucleotides that were isolated from the
by polymerase chain reaction with primers for a sequence of interest. same cells. Blue indicates RNAs transcribed in the sense direction, and
Alternatively, total recovered DNA can be amplified, labeled by red indicates RNAs transcribed in the antisense direction. [Part (a), see A.
incorporation of a fluorescently labeled nucleotide, and hybridized to a Hecht and M. Grunstein, 1999, Methods Enzymol. 304:399. Part (b) adapted from
microarray (Figure 5-29) or subjected to massively parallel DNA P. B. Rahl et al., 2010, Ce// 141 :432.]

TFIIA, TFIIB, etc., and most are mu ltimeric proteins. The to a promoter and ready to initiate transcription is called a
largest is TFIID, which consists of a single 38-kDa TATA preinitiation complex. Figure 7-17 summarizes the stepwise
box bmding protein (TBP) and 13 TBP-associated factors assembly of the Pol II transcription preinitiation complex in
(TAFs). General transcription factors with similar activities vitro on a promoter containing a TAT A box. The TBP sub-
and homologous sequences are found in all eukaryotes. The unit ofTFIID rather than the intact TFIID complex was used
complex of Pol II and its general transcription factors bound in the studies that revealed the order of general transcription

298 CHAPTER 7 • Tran scriptional Control of Gene Expression

t,;) PODCAST: Assembly of the Pol II Preinitiation Complex

FIGURE 7-17 In vitro assembly of RNA polymerase II preinitia- TATA box

tion complex. The indicated general transcription factors and purified ~n'-.0::.r'I:.E'":._~~~.!"""...~~'.."':.0.
RNA polymerase II (Pol II) bind sequentially to TATA-box DNA to form a
preinitiation complex. ATP hydrolysis then provides the energy for TBP
unwinding of DNA at the start site by a TFIIH subunit. As Pol II initiates
transcription in the resulting open complex, the polymerase moves
away from the promoter and its CTD becomes phosphorylated. In vitro,
the general transcnption factors (except for TBP) dissociate from the
TBP-promoter complex, but it is not yet known which factors remain
associated with promoter regions following each round of transcrip-
tion initiation in vivo.

Polll ~

factor and R1 A polymerase II assembly because it can be TFIIF ~~t~ IJ~~.~

·~ y ~.-
expressed at a high level in E. coli and r eadily purified, while
intact TFIID is difficult to purify from eukaryotic cells. CTD
TBP is the first protein to bind to aTAT A box promoter. All
eukaryotic TBPs analyzed to date have very similar C-terminal
domains of 180 residues. This domain of TBP folds into a
saddle-shaped structure; the two halves of the molecule ex-
hibit an overall dyad symmetry but are not identical. TBP in-
teracts with the minor groove in DNA, bending the helix
considerably (see Figure 4-5). The DNA-binding surface of
TBP is conserved in all eukaryotes, explaining the high conser-
vation of the TATA box promoter element (see Figure 7-14).
Once TBP has bound to the TATA box, TriiB can bind.
TFIIB is a monomeric protein, slightly smaller than TBP. The
C-terminal domain of TFIIB makes contact with both TBP
and DNA on either side of theTAT A box. During transcrip-
tion initiation, its N-terminal domain is inserted into the
RNA exit channel of RNA polymerase II (see Figure 7-1 0 ).
The TFIIB N-terminal domain assists Pol II in melting the TFIIH
DNA strands at the transcription start site and interacts with
the template strand near the Po l II active sire. Following
TFIIB binding, a preformed complex of TFIIF (a heterodimer
of two different subunits in mammals) and Pol II binds, posi-
tioning the polymerase over the start site. Two more general Preinitiation
complex
transcription factors must bind before the DNA duplex can }
_ _;:,!. . .~';':!
be separated to expose the template strand. First to bind is
tetrameric TF!lE comprised of two copies each of two differ-
NTPs+ATP
ent subunits. TFIIE creates a docking site for TFIIH, another
multimeric factor containing 10 different subunits. Binding 1!1
1
~ADP
ofTFIIH completes assembly of the transcription preinitiation
complex in vitro (Figure 7-17). Figure 7-18 shows a current
model for the structure of a preinitiation complex.

~~~ent ';~!:~~
The helicase activity of one of the TFIIH subunits uses
energy from ATP h ydrolysis to help unwind the DNA du- Elongating
plex at the start site, allowing Pol II to form an open com-
Release of fjf:.l' ~ Poll! with
phosphorylated
plex in which the DNA duplex surrounding the start site is
general factors, ' fP\ ~ :-;. CTD
melted and the template strand is bound at the polymerase except TBP { _y ~ 1 1.~
active site. Figure 7-19 shows molecular models based on ~-
x-ray crystallography of the complex of TBP (purple), TFIIB
(red), and Pol II (gold) associated with promoter DNA be-
fore the strands near the transcription starr site are separated

7.3 RNA Polymerase II Promoters and General Transcription Factors 299

G) VIDEO: 30 Model of an RNA Polymerase II Preinitiation Complex

FIGURE 7 - 18 Model for the structure of an RNA polymerase II

preinitiation complex. Yeast RNA polymerase II is shown as a space-
filling model with the direction of transcription to the right. The template
strand of DNA is shown in dark blue and the nontemplate strand in
cyan. The start site of transcription is sh own as a space-filling cyan and
dark blue base pair. TBP and TFIIB are shown as purple and red worm
traces of the polypeptide backbone. St ructures for TFIIE, F, and H have
not been determined to high resolution. Their approximate positions
lying over the DNA in the preinitiation complex are shown by ellipses
for TFIIE (green), TFIIF (violet), and TFIIH (light blue). [Adapted from G. Miller
and 5. Hahn, 2006, Nat. Strucr. Bioi. 13:603.]

(closed complex, Figure 7- 19a) and after the strands are sep-
arated and the template strand enters the Pol II-TFIIB com-
plex, placing the transcription start site ( + 1) at the active
site (open complex, Figure 7-19b). A Mg2 • ion bound at the
active site of Pol II assists in catalysis of phosphodiester
bond synthesis. If all the ribonucleoside triphosphates are
present, Pol II begins tra nscri bing the temp late strand.
As the polymerase transcribes away from the promoter ~-Pro-Ser repeat that comprises the CTD. As we shall discuss
region, the N-terminal domain of TFIIB is released from the further in Chapter 8, the CTD that is multiply phosphory-
RNA exit channel as the 5' end of the nascent RNA enters it. lated on serine 5 is a docking site for the enzymes that form
A subunit of TFIIH phosphorylates the Pol TI CTD mu ltiple the cap structure (Figure 4-14) on the 5' end of RNAs transcribed
times on the serine 5 (underlined) of the Tyr-Ser-Pro-Thr- by RNA polymerase II. In the minimal in vitro transcription

(a) Closed complex (b) Open complex

Downstream
DNA

TFI

·.

.·

FIGURE 7-19 Models for the closed and open complexes of in the closed complex (a), where the strands are initially separated
promoter DNA in complex with TBP, TFIIB, and Pol II based on (point of DNA opening). The Mg 2 ion at the active site is shown as a
x-ray crystallography. Pol II is shown in tan, TBP in purple, TFIIB in red, green sphere. The nontemplate strand of the transcription bubble in
the DNA template strand in dark blue, and the DNA nontemplate the open complex (b) is not visualized in crystal structures of models of
strand in cyan. The base encoding the transcri ption start site ( + 1) is the open complex because it has alternat ive conformations in different
shown as space-filling. The B-linker region ofTFIIB interacts with DNA complexes. [Adapted from D. Kostrewa et al., 2009, Narure 462:323.]

300 CHAPTER 7 • Transcriptional Control of Gene Expression

 assay containing only these general transcription factors and Chromatin immunoprecipitation assays (Figure 7-16)
purified RNA polymerase II, TBP remains bound to the using antibodies to TBP show that it binds in the region be-
TATA box as the polymerase transcribes away from the tween the sense and antisense transcription start sites in CpG
promoter region, but the other general transcription factors island promoters. Consequently, the same general transcrip-
dissociate. tion factors probably are required for initiation from the
weaker CpG island promoters as from promoters containing
Remarkably, the first subunits of TFIIH to be cloned a TATA box. The absence of the promoter clements sum-
from humans were identified because mutations in marized in Figure 7-14 may account for the divergent tran-
them cause defects in the repair of damaged DNA. In normal scription from multiple transcription start sites observed
individuals, when a transcribing RNA polymerase becomes from these promoters, since cues from the DNA sequence
stalled at a region of damaged template DNA, a subcomplex are not present to orient the preinitiation complex. TFIID
composed of several subunits of TFIIH, including the helicase and the other general transcription factors may choose
subunit mentioned above, recognizes the stalled polymerase among alternative, nearly equivalent weak binding sites in
and then associates with other proteins that function with this class of promoters, potentially explaining the low fre-
TFIIH in repairing the damaged DNA region. In patients with quency of transcription initiation as well as the alternative
mutant forms of these TFIIH subunits, such repair of dam- transcription start sites in divergent directions generally ob-
aged DNA in transcriptionally active genes is impaired. As a served from CpG island promoters.
result, affected individuals have extreme skin sensitivity to
sunlight (a common cause of DNA damage is ultraviolet light)
·. Elongation Factors Regulate the Initial Stages of
and exhibit a high incidence of cancer. Consequently, these
subunits of TFIIH serve two functions in the cell, one in the Transcription in the Promoter-Proximal Region
process of transcription initiation and a second function in the In metazoans, at most promoters, Pol II pauses after transcrib-
repair of DNA. Depending on the severity of the defect in ing = 20-50 nucleotides, due to the binding of a five-subunit
TFIIH function, these individuals may suffer from diseases protein called NELF (negative elongation factor). This is fol-
such as xeroderma pigmentosum and Cockayne's syndrome lowed by the binding of a two-subunit elongation factor called
(Chapter 24). • DSIF (DRB sensitivity-inducing factor), so named because an
ATP analog called ORB inhibits further transcription elonga-
tion in its presence. The inhibition of Pol II elongation that re-
sults from NELF binding is relieved when DSIF, NELF, and
In Vivo Transcription Initiation by RNA
serine 2 of the Pol II CTD repeat (Tyr-Ser-Pro-Thr-Ser-Pro-Ser)
Polymerase II Requires Additional Proteins are phosphorylated by a protein kinase with two subunits,
Although the general transcription factors discussed above CDK9-cyclin T, also called P-TEFb, which associates with the
allow RNA polymerase to initiate transcription in vitro, an- Pol II, NELF, DSIF complex. The same elongation factors regu-
other general transcription factor, TFllA, is required for ini- late transcription from CpG island promoters. These factors
tiation by Poll! in vivo. Purified TFIIA forms a complex with that regulate elongation in the promoter-proximal region pro-
TBP and TATA box DNA. X-ray crystallography of this com- vide a mechanism for controlling gene transcription in addition
plex shows that TFllA interacts with the side of TBP that is to the regulation of transcription initiation. This overall strat-
upstream from the direction of transcription on promoters egy for regulating transcription at both the steps of initiation
containing aTATA box. In metazoans (multicellular animals), and elongation in the promoter-proximal region is similar to
TFIIA and TFIID, with its multipleTAf subunits, bind first to the regulation of the Trp operon in E. coli (Figure 7-6),
TATA box DNA, and then the other general transcription although the molecular mechanisms involved are distinct.
factors subsequently bind as indicated in Figure 7-17.
The T AF subunits of ;TFIID function in initiating tran- Transcription of HIV (human immunodeficiency virus),
scription from promoters that lack a TATA box. For instance, the cause of AIDS, is dependent on the activation of
some T AF subunits contact the initiator element in promoters CDK9-cyclin T by a small viral protein called Tat. Cells in-
where it occurs, probably explaining how such sequences can fected with tar mutants produce short viral transcripts =50
replace a TATA box. Additional TFTID TAF subunits can nucleotides long. In contrast, cells infected with wild-type
bind to a consensus sequence AJG-G-AJT-Crr-GINC centered HIV synthesize long viral transcripts that extend throughout
=30 base pairs downstream from the transcription start site the integrated proviral genome (see Figure 4-49 and Fig-
in many genes that lack a TATA box promoter. Because of its ure 6-13). Thus Tat protein functions as an antitermination
position, this regulatory sequence is called the downstream factor, permitting RNA polymerase II to read through a tran-
promoter element (DPE) (1-igure 7-14). The DPE facilitates scriptional block. (Tat is initially made by rare transcripts
transcription ofTATA-less genes that contain it by increasing that fail to terminate when the HIV promoter is transcribed
TFIID binding. Also, an a helix of TFIIB binds to the major at high rate in "activated" T-lymphocytes, one type of white
groove of DNA upstream of the TATA-box (see Figure 7-19), blood cell; see Chapter 23). Tat is a sequence-specific RNA-
and the strongest promoters contain the optimal sequence for binding protein. It binds to the RNA copy of a sequence
this interaction, the BRE shown in Figure 7-14. called TAR, which forms a stem-loop structure ncar the 5' end

7.3 RNA Polymerase II Promoters and General Transcription Factors 301

 5'
In metazoans, NELF associates with Pol II after initiation,
inhibiting elo ngation =50-200 base pairs from the transcrip-
tion start site. Inhibition of elongation is relieved when the
heterodimeric elongation factors DSIF and CDK9-cyclin T (P-
TEFb) associate with the elongation complex and CDK9 phos-
phorylates subunits of NELF, DSlF, and serine 2 of the Pol Il
CTD heptapeptide repeat.

CTD
7.4 Regulatory Sequences in Protein-
FIGURE 7-20 Model of antitermination complex composed of
HIV Tat protein and several cellular proteins. The TAR element in Coding Genes and the Proteins
the HIV transcript contains sequences recognized by Tat and the
Through Which They Function
cellular protein cyclin T. Cyclin T activates and helps position the
protein kinase CDK9 near its substrate, the CTD of RNA polymerase II. As noted in the previous section! expression of eukaryotic
CTD phosphorylation at serine 2 of the Pol II CTD heptad repeat is protein-coding genes is regulated by multiple protein-binding
required for transcription elongation. Cellular proteins DSIF (also called DNA sequences, generically referred to as transcription-control
Spt4/5) and the NELF complex are also involved in regulating Pol II regions. These include promoters and other types of control
elongation, as discussed in the text. [SeeP. Wei et al., 1998, Cell 92:451; elements located near transcription start sites, as well as se-
T. Wad a et al., 1998, Genes Dev. 12:357; and Y. Yamaguchi et al., 1999, Cell 9 7:41 .)
quences located far from the genes they regulate. In this sec-
tion, we take a closer look at the properties of various control
elements found in eukaryotic protein-codi'ng genes and the
of the HIV transcript (Figure 7-20). TAR also binds cyclin T, proteins that bind to them.
holding the CDK9-cyclin T complex close to the polymerase,
where tt efficiently phosphorylates its substrates, resulting in
Promoter-Proximal Elements Help
transcription elongation. Chromatin immunoprecipitation as-
says done after treating cells with specific inhibitors of CDK9 Regulate Eukaryotic Genes
mdicate that the transcription of =30 percent of mammalian Recombinant DNA techniques have been used to systemati-
genes is regulated by controlling the activity of CDK9-cyclin T cally mutate the nucleotide sequences of various eukaryotic
(P-TEFb), although this is probably done most frequently by genes in order to identify transcription-control regions. For ex-
sequence-specific DNA-binding transcription factors rather an ample, linker scanning mutations can pinpoint the sequences
RNA-binding pfotein, as in the case of HIV Tat. • within a regu latory region that fu nction to control transcrip-
tion. In this approach, a set of constructs with contiguous
overlapping mutations are assayed for their effect on expres-
KEY CONCEPTS of Section 7 3 sion of a reporter gene or production of a specific m RNA (Fig-
ure 7-21a). This type of analysis identified promoter-proximal
RNA Polymerase II Promoters and General elements of the thymidine kinase (tk) gene from herpes simplex
Transcription Factors type I virus (HSV-T) . The resu lts demonstrated that the DNA
RNA polymerase II initiates transcription of genes at the region upstream of the HSV tk gene contains three separate
nucleotide in the DNA template that corresponds to the 5' transcription-control sequences: a TATA box in the interval
nucleotide that is capped in the encoded mRNA. from -32 to -16 and two other control elements farther up-
stream (Figure 7-2 1b). Experiments using mutants containing
Transcription of protein-coding genes by Pol ll can be initiated
single-base-pair changes in promoter-proximal control ele-
in vitro by sequential binding of the following in the indicated
ments revealed that they are generally =6-10 base pairs long.
order: TBP, which binds toTATA box DNA; TFIIB; a complex
Recent results indicate that they are found both upstream and
of Polll and THIF; TFUE; and finally, TFITH (see Figure 7-17).
downstream of the transcription start site for human genes at
The hclicase activity of a TFTIH subunit helps to separate the equal frequency. While, strictly speaking, the term promoter
template strands at the starr site in most promoters, a process refers to the DNA sequence that determines where a poly-
that requires hydrolysis of ATP. As Pol II begins transcribing merase initiates transcription, the term is often used to refer
away from the start site, its CTD is phosphorylated on serine 5 to both a promoter and its associated promoter-proximal
of the heptapeptide CTD by another TFIIH subunit. control elements.
In vivo transcription initiation by Pol II also requires TFIIA To test the spacing constraints on control elements in
and, in metazoans, a complete TFIID protein complex, in- the HSV tk promoter region identified by analysis of linker
cluding its multiple TAF subunits as well as the TBP subunit. scanning mutations, researchers prepared and assayed con-
structs containing small deletions and insertions between

302 CHAPTER 7 • Transcriptional Control of Gene Expression

 (a)
Reporter gene
Vector DNA tkmRNA

Mutant
=i~ Control region f £=] +++

no.

=i)=::fl!/!i1 £=] +++

2
=i~ tzl_. £=] +

3
=i~ llfZJ I I
£=] +

4
=i~ tzl!J I
£=] +++

5
=i~ i :~ I
£=] +

6
=i~ i i !WJJ I
£=] +++

7
=i~ i ii t2Z1JJ £=] +++

8
=i~ i i! t?La2 I I I I
£=]
=i~ :: £=]
I I
9 : tz/'//J
I
I
I I I I I I
+++

Control elements • I I

Control region of tk gene

EXPERIMENTAL FIGURE 7·21 Linker scanning mutations is assayed. In the example shown here, the sequence from 120 to
identify transcription-control elements. (a) A region of eukaryotic + 1 of the herpes simplex virus thymidine kinase gene, LS mutations 1,
DNA (tan) that supports high-level expression of a reporter gene (light 4, 6, 7, and 9 have little or no effect on expression of the reporter gene,
purple) is cloned in a plasmid vector as diagrammed at the top. indicating that the regions altered in these mutants contain no control
Overlapping linker scanning (LS) mutations (crosshatch) are introduced elements. Reporter-gene expression is significantly reduced in mutants
from one end of the region being analyzed to the other. These 2, 3, 5, and 8, indicating that control elements (brown) lie in the
mutations result from scrambling the nucleotide sequence in a short intervals shown at the bottom. (b) Analysis of these LS mutations
stretch of the DNA. After the mutant plasmids are transfected identified a TATA box and two promoter-proximal elements (PE-1 and
separately into cultured cells, the activity of the reporter-gene product PE-2). [Part (b), seeS. L. McKnight and R. Kingsbury, 1982, Science 217:316.]

the elements. Changes in spacing between the promoter and common in eukaryotic genomes but fairly rare in bacterial
promoter-proximal control elements of 20 nucleotides or genomes. Procedures such as linker scanning mutagenesis have
fewer had little effect. However, insertions of 30 to 50 base indicated that enhancers, usually on the order of ==:200 base
pairs between the HSV-I tk promoter-proximal elements and pairs, like promoter-proximal elements, are composed of sev-
the TATA box was equivalent to deleting the element. Simi- eral functional sequence elements of '"='6-10 base pairs. As dis-
lar analyses of other eukaryotic promoters have also indi- cussed later, each of these regulatory elements is a binding site
cated that considerable flexibility in the spacing between for a sequence-specific DNA-binding transcription factor.
promoter-proximal clements is generally tolerated, but separa- Analyses of many different eukaryotic cellular enhancers
tions of several tens of base pairs may decrease transcription. have shown that in metazoans, they can occur with equal
probability upstream from a promoter or downstream from
a promoter within an intron, or even downstream from the
Distant Enhancers Often Stimulate
final exon of a gene, as in the case of the Salll gene (see Fig-
Transcription by RNA Polymerase II ure 7-8a}. Many enhancers are cell-type specific. For exam-
As noted earlier, transcription from many eukaryotic pro- ple, an en hancer controlling Pax6 expression in the retina
moters can be stimulated by control elements located thou- was characterized in the intron between exons 4 and 5 (see
sands of base pairs away from the start site. Such long-distance Figure 7-7a), whereas an enhancer controlling Pax6 expres-
transcription-control elements, referred to as enhancers, are sion in the hormone-secreting cells of the pancreas is located

7.4 Regulatory Sequences in Protein-Coding Genes and the Proteins Through Which They Function 303
 (a) Mammalian gene with aTATA box .. ,
===e=J::::t~
up to -200 -30
. ~~•______,-=f~
IIIII lfli.-. .u--- -.==1

+10to
-50 kb or more +50 kb or more

(b) Mammalian CpG-island promoter gene

(c) S. cerevisiae gene +1 • Exon 0 lntron 0 TATA box

• Promoter-proximal D Enhancer;
element yeast UAS
"' 90
~ CpG island

FIGURE 7-22 General organization of control elements that Iian genes. (b) Mammalian CpG-island promoters. Transcription
regulate gene expression in multicellular eukaryotes and yeast. initiates at several sites in both the sense and antisense directions from
(a) Mammalian genes with a TATA-box promoter are regulated by the ends of the CpG-rich region. Transcripts in the sense direction are
promoter-proximal elements and enhancers. Promoter elements elongated and processed into mRNAs by RNA spl icing. They express
shown in Figure 7-14 position RNA polymerase II to initiate transcrip- mRNAs with alternative 5' exons determined by the transcription start
tion at the start site and influence the rate of transcription. Enhancers site. CpG-island promoters contain promoter-proximal control
may be either upstream or downstream and as far away as hundreds of elements. Currently, it is not clear whether they are illso regulated
kilobases from the transcription start site. In some cases, enhancers lie by distant enhancers. (c) Most 5. cerevisiae genes conta in only one
.·.
within introns. Promoter-proximal elements are found upstream and regulatory region, called an upstream activating sequence (UAS}, and
downstream of transcription start sites at equal frequency in mamma- a TATA box, which is =90 base pairs upstream from the start site.

in an ""200-base-pair region upstream of exon 0 (so named contro l elements that can stimulate transcription from
because it was discovered after the exon called "exon 1 ").In distances between these t\Vo extremes.
the important model organism Saccharomyces cerevisiae Figure 7-22a summarizes the locations of transcription-
(budding yeast), genes are closely spaced (Figure 6-4b) and control sequences for a hypothetical mammalian gene with a
few genes contain introns. In this organism, enhancers usu- promoter containing a TAT A box. The start site at which
ally lie within ""200 base pairs upstream of the promoters of transcription initiates encodes the first (5') nucleotide of the
the genes they regulate and are referred to by the term up- first exon of an mRNA, the nucleotide that is capped. In ad-
stream activating sequence (UAS). dition to the TATA box at""- 31 to -26, promoter-proximal
elements, which are relatively short (""6-10 base pairs), are
located within the first ::::.200 base pairs either upstream or
Most Eukaryotic Genes Are Regulated by
downstream of the start site. Enhancers, in contrast, usually
Multiple Transcription-Control Elements are about 50-200 base pairs long and are composed of mul-
Initially, enhancers and promoter-proximal elements were tiple elements of =6-10 base pairs. Enhancers may be lo-
thought to be distinct types of transcription-control clements. cated up to 50 kilobases or more upstream or downstream
However, as more enhancers and promoter-proximal ele- from the start site or within an intron. As for the Pax6 gene,
ments were analyzed, the distinctions bet\veen them became many mammalian genes are controlled by more than one
less clear. For example, both types of element generally can enhancer region that func tion in different types of cells.
stimulate transcription even when inverted, and both types Figure 7-22b summarizes the promoter region of a mam-
often arc cell-type specific. The general consensus now is rhat malian gene with a CpG island promoter. About 60-70 percent
a spectrum of control elements regulates transcription by of mammalian genes are expressed from CpG island promot-
RNA polymerase II. At one extreme are enhancers, which ers, usually at much lower level<; than gcnec; with TATA box
can stimulate transcription from a promoter tens of thou- promoters. Multiple alternative transcription start sites are
sands of base pairs away. At the other extreme are promoter- used, generating mRNAs with alternative 5' ends for the
proximal clements, such as the upstream elements controlling first exon derived from each start site. Transcription occurs
the HSV tk gene, which lose their influence when moved an in both directions, but Pol II molecules transcribing in the
additional 30-50 base pairs farther from the promoter. Re- sense direction are elongated to > 1 kb much more efficiently
searchers have identified a large number of transcription- than transcripts in the antisense direction.

304 CHAPTER 7 • Transcriptional Control of Gene Expression

 The S. cerevisiae genome contains regulatory elements pattern of bands is observed that depends on the DNA se-
called upstream activating sequences (UASs), which function quence and results from cleavage at some phosphodiester
similarly to enhancers and promoter-proximal elements in bonds and not others. However, when increasing amounts of
higher eukaryotes. Most yeast genes contain only one UAS, TBP are incubated with the end-labeled DNA before digestion
which generally lies within a few hundred base pairs of the with DNase I, TBP binds to the TATA box and protects the
start site. In addition, S. cerevisiae genes contain a TATA box region from =- 35 to -20 from digestion when sufficient
=90 base pairs upstream from the transcription start site TBP is added to bind all the labeled DNA molecules. In con-
(Figure 7-22c). trast, increasing amounts of TFIID (lanes 7 and 8) protect
both the TATA box region from DNasc I digc:.tion, as well as
regions near -7, + l to + 5, + 10 to+ 15, and+ 20, producing
Footprinting and Gel-Shift Assays Detect
a different "footprint'' from TBP. Results such as this tell us
Protein-DNA Interactions that other subunits of TFIID (the TBP-associated factors or
The various transcription-control elements found in eukary- T AFs) also bind to the DNA in the region downstream from
otic DNA are binding sites for regulatory proteins generally the TATA box.
called transcription factors. The simplest eukaryotic cells en- The electrophoretrc mobility shift assay (EMSA), also
code hundreds of transcription factors, and the human ge- called the gel-shift or band-shift assay, is more useful than
nome encodes over 2000. The transcription of each gene in the footprinting assay for quantitative analysis of DNA-
the genome is independently regulated by combinations of binding proteins. In general, the electrophoretic mobility of
specific transcription factors that bind to its transcription- a DNA fragment is reduced when it is complexed to protein,
control regions. The number of possible combinations of this causing a shift in the location of the fragment band. This
many transcription factors is astronomical, sufficient to gener- assay can be used to detect a transcription factor in protein
ate unique controls for every gene encoded in the genome. fractions incubated with a radiolabeled DNA fragment con-
[n yeast, Drosophila, and other genetically tractable eu- taining a known control element (Figure 7-24). The more of
karyotes, numerous genes encoding transcriptional activa - the transcription factor that is added to the binding reaction,
tors and repressors have been identified by classical genetic the more labeled probe is shifted to the position of the DNA-
analyses like those described in Chapter 5. However, in mam- protein complex. ·
mals and other vertebrates, which are less amenable to such In the biochemical isolation of a transcription factor, an
genetic analysis, most transcription factors have been de- extract of cell nuclei commonly is subjected sequentially to sev-
tected initially and subsequently purified by biochemical eral types of column chromatography (Chapter 3). Fractions
techniques. ln this approach, a DNA regulatory element that eluted from the columns are assayed by DNase I footprinting
has been identified by the kinds of mutational analyses de- or EMSA using DNA fragments containing an identified regu-
scribed above is used to identify cognate proteins that bind latory element (see Figure 7-21 ). Fractions containing a protein
specifically to it. Two common techniques for detecting such that binds to the regulatory element in these assays probably
cognate proteins are DNase l footprinting and the electro- contain a putative transcription factor. A powerful technique
phoretic mobility shift assay. that is commonly used for the final step in purifying transcrip-
DNase [ footprinting takes advantage of the fact that tion factors is sequence-specific DNA affinity chromatogra-
when a protein Is bound to a region of DNA, it protects that phy, a particular type of affinity chromatography in which
DNA sequence from digestion by nucleases. As illustrated in long DNA strands containing multiple copies of the transcrip-
Figure 7-23a, samples of a DNA fragment that is labeled at tion factor-binding site are coupled to a column matnx.
one end are digested under carefully controlled conditions in Once a transcription factor is isolated and purified, its
the presence and absence of a DNA-binding protein, and then partial amino acid sequence can be determined and used to
denatured, electrophoresed, and the resulting gel subjected to clone the gene or eDNA encoding it, as outlined in Chapter 5.
autoradiography. The region protected by the bound protein The isolated gene can then be used to test the ability of the
appears as a gap, or "footprint," in the array of bands result- encoded protein to activate or repress transcription in an in
ing from digestion in the absence of protein. When footprint- vivo transfection assay (Figure 7-25).
ing is performed with a DNA fragment containing a known
DNA control element, the appearance of a footprint indicates
Activators Promote Transcription and Are
the presence of a transcription factor that binds that control
element in the protein sample being assayed. Footprinting also Composed of Distinct Functional Domains
identifies the specific DNA sequence to which the transcrip- Studies with a yeast transcription activator called GAL4 pro-
tion factor binds. vided early insight into the dom;Jin structure of transcription
For example, DNase I footprinting of the strong adenovi- factors. The gene encoding the GAL4 protein, which pro-
rus late promoter shows a protected region over the TATA motes expression of enzymes needed to metabolize galactose,
box when TBP is added to the labeled DNA before DNase I was identified by complementation analysis of gal4 mutants
digestion (Figure 7-23b). DNase I does not digest all phospho- that cannot form colonies on an agar medium in which galac-
diester bonds in a duplex DNA at equal rate. Consequently, in tose is the only source of carbon and energy (Chapter 5 ).
the absence of added protein (lanes 1, 6, and 9), a particular Directed mutagenesis studies like those described previously

7.4 Regulatory Sequences in Protein-Coding Genes and the Proteins Through Which They Function 305
 (a) (b)
Sample A Sample B
(DNA-binding protein absent) (DNA-binding protein present) bp from
theTSS
Sequence-specific
Protein-binding
binding protein
sequence
5' 3' 5'
~ 3' - --
c:-==·=- -50

•
3' 5'
•
3' <> 5'
11•~- E ·sl
-40

-30

~ <>
~ -20
t- :;:1
• - - -
- ..___ -
,§ - 10
"T1

---..... ........ _____ I -g.

- - - - -- - - 0
~ +1
,.. a
• I
* •t ===•<>•==== - ·- _I
--- -.-10

+20
•
* I •t ===•<>-===== =-.--- . . --...... . ___
..........
___ ___ _
.,: +30

---- ---
...._____ _ +40

•t ===•<>-=====* 1 2 3 4 5 6 7 8 9 lane

EXPERIMENTAL FIGURE 7 -23 DNase I footprinting reveals the the labeled DNA, as in sample B (right), the protein binds to the DNA,
region of a DNA sequence where a transcription factor binds. (a) A thereby protecting a portion of the fragment from digestion. Following
DNA fragment known to contain a control element is labeled at one DNase treatment. the DNA is separated from protein, denatured to
end with 32 P (red dot). Portions of the labeled DNA sample then are separate the strands, and electrophoresed. Autoradiography of the
digested with DNase I in the presence and absence of protein samples resulting gel detects only labeled strands and reveals fragments
containing a sequence-specific DNA-binding protein. DNase I extending from the labeled end to the site of cleavage by DNase
hydrolyzes the phosphodiester bonds of DNA between the 3' oxygen I. Cleavage fragments containing the control sequence show up on
on the deoxyribose of one nucleotide and the 5' phosphate of the next the gel for sample A but are missing in sample B because the bound
nucleotide. A low concentration of DNase I is used so that, on average, cognate protein blocked cleavages within that sequence and thus
each DNA molecule is cleaved just once (vertical arrows). If the protein production of the corresponding fragments. The missing bands on
sample does not contain a cognate DNA-binding protein, the DNA the gel constitute the footprint. (b) Footprints produced by increasing
fragment is cleaved at multiple positions between the labeled and amounts ofTBP (indicated by the triangle) and of TFIID on the strong
unlabeled ends of the original fragment, as in sample A (/eft).lf the adenovirus major late promoter. [Part (b) from Q. Zhou et al., 1992,
protein sample contains a protein that binds to a specific sequence in Genes Dev. 6:1964.]

EXPERIME TAL FIGURE 7·L4 Electrophoretic mobility shift Fraction ON 1 2 3 4 5 6 7 8 9 10 11 12 14 16 18 20 22

assay can be used to detect transcription factors during purification.
In this example, protein fractions separated by column chromatogra-
phy were assayed for their ability to bind to a radiolabeled DNA-fragment Bound
probe containing a known regulatory element. After an aliquot of the probe-+-
protein sample was loaded onto the column (ON) and successive
column fractions (numbers) were incubated with the labeled probe,
the samples were electrophoresed under conditions that do not
disrupt protein-DNA interactions. The free probe not bound to protein
migrated to the bottom of the gel. A protein in the preparation applied
to the column and in fractions 7 and 8 bound to the probe, forming a
DNA-protein complex that migrated more slowly than the free probe. Free
probe ...
These fractions therefore likely contain the regulatory protein being
sought. [From S. Yoshinaga et aL. 1989, J. Bioi. Chem. 264:1 0529.]

306 CHAPTER 7 • Transcriptional Control of Gene Expression

 Gene-encoding Reporter an in vivo assay like that depicted in Figure 7-25. Thus the
protein X ene internal portion of the protein is not required for functioning

e e
of GA LA as a transcription factor. Simi lar experiments with
.· another yeast transcription factor, GCN4, which regulate~
X-binding genes required for synthesis of many amino acids, indicated
site that it contains an =50-amino acid DNA-binding domain at

Y ,.,.;,x its C-terminus and an =20-amino acid activation domain

near the middle of its sequence.
Further evidence for the existence of dtstmct activation
domains in GAL4 and GCN4 came from experiments in
which their activation domains were fused to a DNA-binding
domain from an entirely unrelated E. coli DNA-binding pro-
tein. When these fusion proteins were assayed in vivo, they
Reporter-gene activated transcription of a reporter gene containing the cog-
transcripts nate site fo r the E. coli protein. Thus functional transcription
factors can be constructed from entirely novel combinations
of prokaryotic and eukaryotic elements.
Studies such as these have now been carried out with many
eukaryotic activators. The structural model of eukaryotic acti-
EXPERIMENTA FIGURE .' -25 In vivo transfection assay vators that has emerged from these studies is a modular one
measures transcription activity to evaluate proteins believed to be in which one o r more activation domains are connected to a
transcription factors. The assay system requires two plasm ids. One sequence-specific DNA-binding domain through flexible
plasmid contains the gene encoding the putative transcription factor protein domains (Figure 7-27). In some cases, amino acids in-
(protein X). The second plasmid contains a reporter gene (e.g., cluded in the DNA-binding domain also contribute to tran-
luciferase) and one or more binding sit es for protein X. Both plasmids scriptional activation. As discussed in a later .section, activation
are simultaneously int roduced into cells that lack the gene encoding
domains are thought to function by binding other proteins
protein X. The production of reporter-gene RNA transcripts is mea-
involved in transcription. The presence of flexible domains
sured; alternatively, the activity of the encoded protein can be assayed.
connecting the DNA-binding domains to activation domains
If reporter-gene transcription is greater in the presence of the
X-encoding plasmid than in its absence, then the protein is an
may explain why alterations in the spacing between control
activator; if transcription is less, then it is a repressor. By use of plasmids
elements are so well tolerated in eukaryotic control regions.
encod ing a m utated or rea rranged tran scription factor, important Thus even when the positions of transcription factors bound to
domains of the protein can be identified. DNA are shifted relative to each other, their activation do-
mains may still be able to interact because they are attached to
their DNA-binding domains through flexible protein regions.

Repressors Inhibit Transcription and Are

identified UASs for the genes activated by GAL4. Each of
these UASs was found to contain one or more copies of a re- the Functional Converse of Activators
lated 17-bp sequence called UAScAt· DNase l footprinting Eukaryotic transcription is regu lated by repressors as well as
assays with recombinant GAL4 protein produced in E. coli activators. For example, geneticists have identified muta-
fro m the yeast GAL4 gene showed that GAL4 protein binds tions in yeast that result in continuously high expression of
to UAScAL sequences. When a copy of UASc.11 was cloned certain genes. This type of unregulated, abnormally high ex-
upstream of a TATA box followed by a 13-galacrosidase re pression is called co nsti tutive expression and results from
porter gene, expression of 13-galactosidase was activated in the inactivation of a repressor that normally inhibits the
galactose media in wild-type cells but not in ga/4 mutants. transcription of these genes. Similarly, mutants of Drosophila
These resu lts showed that UAScAL is a transcription-control and Caenorhabditis elegans have been isolated that are de-
clement activated by the GAL4 protein in galactose media. fective in embryonic development because they express genes
A remarkable set of experiments with ga/4 deletion mu- in embryonic cells where those genes are normally repressed.
tants demo nstrated t hat t he GAL4 transcription factor is The mutations in these mutants inactivate repressors, lead-
composed of separable functional domains: an N-terminal ing to abnormal development.
DNA-binding do m ai n, wh ich binds to specific DNA se- Repressor-binding sires in DNA have been identified by
quences, and a C-terminal activation domain, which interacts systematic linker scanning mutation analysis similar to that
with other proteins to stimul ate transcription from a nearby depicted in Figure 7-21. In th is type of analysis, mutation of
promoter (Figure 7-26). When theN-terminal DNA-binding an activator-binding site leads to decreased expression of
doma in of GAL4 was fused directly to various portions of its the linked reporter gene, whereas mutation of a repressor-
own C-termina l region, the resulting truncated proteins re- binding site leads to increased expression of a reporter gene.
tained the ability to stimulate expression of a reporter gene in Repressor proteins that bind such sites can be purified and

7.4 Regulatory Sequences in Protein- Coding Genes and the Proteins Through Which They Function 307
 . KPERIMENTAL FIGURE 7-26 Deletion
mutants of the GAL4 gene in yeast with a UASGAL
r eporter-gene construct demonstrate the separate
f unct ional d omains in an activator. (a) Diagram of UASGAL TATA
box
DNA construct containing a lacZ reporter gene (encod-
ing I)-galactosidase) and TATA box ligated to UASGAL• a
regulatory element that contains several GAL4-binding (b) Wild-type and mutant GAL4 proteins Binding ~-galactosidase

sites. The reporter-gene construct and DNA encoding to UASGAL activity ·.

wild-type or mutant (deleted) GAL4 were simultaneously
introduced into mutant (ga/4) yeast cells, and the activity
Wild-type c + +++
DNA-binding Activation
of I)-galactosidase expressed from lacZ was assayed.
domain domain
Activity will be high if the introduced GAL4 DNA encodes
a functional protein. (b) Schematic diagrams of wild-type 50 881
~----------------------~
GAL4 and various mutant forms. Small numbers refer to
848 + +++
positions in the wild-type sequence. Deletion of SO L---------------------~
amino acids from theN-terminal end destroyed the + +++
N-and L-----------------------_J 823
ability of GAL4 to bind to UASGAL and to stimulate
C-terminal
expression of I)-galactosidase from the reporter gene. + ++
deletion L---------------------~ 792
Proteins with extensive deletions from the (-terminal mutants
+ +
end still bound to UASGAL. These results localize the L-.-------------'1 755
DNA-binding domain to the N-terminal end of GAL4. The
ability to activate I)-galactosidase expression was not [L---~~----__.1 692 +

entirely eliminated unless somewhere between 126 and

189 or more amino acids were deleted from the
0 74 +

C-terminal end. Thus the activation domain lies in the

0 74 L

{
(-terminal region of GAL4. Proteins with internal 684 Jaa1 + +++

deletions (bottom) also were able to stimu late expression Internal

of I)-galactosidase, indicating that the central region of
deletion 0 74 738 c:=:J 881 + +++
mutants
GAL4 is not crucial for its function in this assay. [See J. Ma
and M. Ptashne, 1987, Ce/148:847; I. A. Hope and K. Struhl, 1986,
0 74 768 D aa1 + ++

Ce/146:885; and R. Brent and M. Ptashne, 1985, Ce//43:729.]

Examples assayed using the same biochemical techniques described

earlier for activa tor proteins.
N~CGAL4
Eukaryotic transcription repressors are the functional con-
verse of activators. They can inhibit transcription from a gene
N~CGCN4 they do not norma lly regulate when their cognate binding sites
are placed within tens of base pairs to many kilobases of the
N C GR gene's start site. Like activators, most eukaryotic repressors are
modular proteins that have two functional domains: a DNA-
N C SP1 binding domain and a repression domain. Similar to activation
domains, repression domains continue to function when fused
to another rype of DNA-binding doma in. If bi nding sites for
DNA-binding
domain
this second DNA-binding domain are inserted within a few
hundred base pairs of a promoter, expression of the fusion
Activation protein inhibits transcription from the promoter. Also like ac-
domain
tivation doma ins, repression domains function by interacting
J\/'lv Flexible protein with other proteins, as discussed later in this chapter.
domain

FIGURE 7-27 Schematic d iagrams illustrating the modular DNA-Binding Domains Can Be Classified
structure of eukaryotic transcription activators. Transcription into Numerous Structural Types
factors may contain more than one activation domain but rarely
contain more than one DNA-binding domain. GAL4 and GCN4 are The DNA-binding domains of eukaryotic activators and re-
yeast transcription activators. The glucocorticoid receptor (GR) pressors contain a variety of structural motifs that bind spe-
promotes transcription of target genes when certain hormones are cific DNA sequences. The ability of DNA-binding proteins
bound to the (-terminal activation domain. SP1 binds to GC-rich to bind to specific DNA sequences commonly results from
promoter elements in a large number of mammal ian genes. noncovalent interactions between atom!> in an ex helix in the

308 CHAPTER 7 • Transcriptional Control of Gene Expression

 Here we introduce several common classes of DNA-binding
proteins whose three-dimensional structures have been deter-
mined. In all these examples and many other transcription
factors, at least one o. helix is inserted into a major groo\'e of
DNA. However, some transcription factors contain alterna-
tive structural motifs (e.g., 13 strands and loops, see NFA Tin
Figure 7-32 as an example) that interact with DNA.
Homeodomain Proteins Many eukaryotic transcription fac-
tors that function during development contain a conserved
60-residue DNA-binding motif, called a homeodomain, that
is similar to the helix-turn-helix motif of bacterial repres-
sors. These transcription factors were first identified in Dro-
sophila mutants in which one body part was transformed
into another during development (see Figure 7-lb). The con-
served homeodomain sequence has also been found in verte-
brate transcription factors, including those that have stmilar
master-control functions in human development.
FIGURE 7-28 Interaction of bacteriophage 434 repressor with Zinc-Finger Protein s A number of different eukaryotic pro-
DNA. (a) Ribbon diagram of 434 repressor bound to its specific teins have regions that fold around a central Zn2 ion, produc-
operator DNA. Repressor monomers are in yellow and green. The
ing a compact domain from a relatively short length of the
recognition helices are indicated by asterisks. A space-filling model of
polypeptide chain. Termed a zinc finger, this structural motif
the repressor-operator complex (b) shows how the protein interacts
was first recognized in DNA-binding domains bur now is
intimately with one side of the DNA molecule over a length of 1.5 turns.
[Adapted from A. K. Aggarwal et al., 1988, Science 242:899.]
known to occur also in proteins that do not bind to DNA. Here
we describe two of the several classes of zinc-finger motifs that
have been identified in eukaryotic transcription factors.
The C 2 H 2 zinc finger is the most common DNA-binding
motif encoded in the human genome and the genomes of
DNA-binding domain and atoms on the edges of the bases most other multicellular animals. It is also common in mul-
within a major groove in the DNA. Ionic interactions be- ticellular plants but is not the dominant type of DNA-binding
tween positively charged residues arginine and lysine and domain in plants as it is in animals. This motif has a 23- to
negatively charged phosphates in the sugar phosphate back- 26-residue consensus sequence containing two conserved
bone and, in some cases, interactions with atoms in a DNA cysteine (C) and two conserved histidine (H) residues, whose
minor groove also contribute ro binding. side chains bind one Zn 2 ~ ion (Figure 3-9c). The name "zinc
The principles of specific protein-DNA interactions were finger" was coined because a two-dimensional diagram of
first discovered during the study of bacterial repressors. the structure resembles a finger. When the three-dimensional
Many bacterial repressors are dimeric proteins in which an o. structure was solved, it became clear that the binding of the
helix from each monomer inserts into a major groove in the Zn2 + ion by the two cysteine and two histidine residues folds
DNA helix (Figure 7 -28). This o. helix is referred to as the the relatively short polypeptide sequence into a compact do-
recognition helix or sequence-reading helix, because most of main, which can insert its a helix into the major groove of
the amino acid side chains that contact DNA extend from DNA. Many transcription factors contain multiple C 2 H 2
this helix. The recognition helix that protrudes from the sur- zinc fingers, which interact with successive groups of base
face of bacterial represso's to enter the DNA major groove pairs, within the major groove, as the protein wraps around
and make multiple, specific interactions with atoms in the the DNA double helix (Figure 7-29a).
DNA is usually supported in the protein structure in part by A second type of zinc-finger structure, designated the C4
hyd rophobic interactions with a second o. helix just N-terminal zinc finger (because it has four conserved cysteines in contact
to it. This structural element, which is present in many bac- with the zn2+), is found in =50 human transcription factors.
terial repressors, is called a helix-turn-helix motif. The first members of this class were identified as specific in
Many additional motifs that can present an o. helix to the tracellular high-affinity binding proteins, or "receptors," for
major groove of DNA are found in eukaryotic transcription steroid hormones, leading to the name steroid receptor su-
factors, which often are classified according to the type of perfamily. Because similar intracellular receptors for nonste-
DNA-binding domam they contain. Because most of these roid hormones subsequently were found, these transcription
motifs have characteristic consensus amino acid sequences, facrors are now common ly called nuclear receptors. The
potential transcription factors can be recognized among the characteristic feature of c4 zinc fingers is the presence of two
eDNA sequences from various tissues that have been charac- groups of four critical cysteines, one toward each end of the
terized in humans and o ther species. The human genome, for 55- or 56-residue domain. Although the C 4 zinc finger ini-
instance, encodes =2000 transcription factors. tially was named by analogy with the C2 H 2 zinc finger, the

7.4 Regulatory Sequences in Protein-Coding Genes and the Proteins Through Which They Function 309
 (a) 5' (b)
3'

5' 3'
5'

FIGURE 7-29 Eukaryotic DNA-binding domains that use an a leucine-zipper proteins, basic residues in the extended a-helical
helix to interact with the major groove of specific DNA sequences. regions of the monomers interact with the DNA backbone at adjacent
(a) The GL 1 DNA-binding domain is monomeric and contains five C2H2 major grooves. The coiled-coil dimerization domain is stabilized by
zinc fingers. The a helices are shown as cylinders, the Zn 2 ions as
T hydrophobic interactions between the monomers. (d) In bHLH
spheres. Finger 1 does not interact with DNA, whereas the other four proteins, the DNA-binding helices at the bottom (N-termini of the
fingers do. (b) The glucocorticoid receptor is a homodimeric C4 monomers) are separated by nonhelicalloops from a leucine-zipper-
zinc-finger protein. The a helices are shown as purple ribbons, the like region containing a coiled-coil dimerization domain. [Part (a), see
(3-strands as green arrows, the Zn 2 ions as spheres. Two a helices N. P. Pavletich and C. 0. Pabo, 1993, Science 2 61 :1701. Part (b), see B. F. Luisi et
(darker shade), one in each monomer, interact with the DNA. Like all C4 al., 1991, Nature 352:497. Part (c), see T. E. Ellenberger et al., 1992, Ce// 71 :1223.
zinc-finger homodimers, this transcription factor has twofold rotational Part (d), see A. R. Ferre-D'Amare et al., 1993, Nature 363:38.)
symmetry; the center of symmetry is shown by the yellow ellipse. (c) In

' '

310 CHAPTER 7 • Transcriptional Control of Gene Expression

 three-dimensional structures of proteins containing these Termed a basic helix-loop-helix (bHLH), this motif was pre-
DNA-binding motifs later were found to be quire distinct. A dicted from the amino acid sequences of these proteins, which
particularly important difference between the two is that contain an N-terminal a helix with basic residues that inter-
C2H 2 zinc-finger proteins generally contain three or more act with DNA, a middle loop region, and a C-terminal region
repeating finger units and bind as monomers, whereas C4 with hydrophobic amino acids spaced at intervals character-
zinc-finger proteins generally contain only two finger units istic of an amphipathic a helix. As with basic-zipper proteins,
·. and generally bind to DNA as homodimers or heterodimers. different bHLH proteins can form heterodimers.
Homodimers of C 4 zinc-finger DNA-binding domains have
twofold rotational symmetry (Figure 7-29b). Consequeutly,
homodimeric nuclear receptors bind to consensus DNA se-
Structurally Diverse Activation and Repression
quences that are inverted repeats.
Domains Regulate Transcription
l e ucine-Zipper Proteins Another structural motif present in Experiments with fusion proteins composed of the GAL4
the DNA-binding domains of a large class of transcription DNA-binding domain and random segments of E. coli pro-
factors contains the hydrophobic amino acid leucine at every teins demonstrated that a diverse group of amino acid se-
seventh position in the sequence. These proteins bind to quences can function as activation domains, "'=' 1 percent of
DNA as dimers, and mutagenesis of the leucines showed that all E. coli sequences, even though they evolved to perform
they were required for dimerization. Consequently, the name other functions. Many transcription factors contain activa-
leucine zipper was coined to denote this structural motif. tion domains marked by an unusually high percentage of
The DNA-binding domain of the yeast GCN4 transcrip- particular amino acids. GAL4, GCN4, and most other yeast
tion factor mentioned earlier is a leucine-zipper domain. X-ray transcription factors, for instance, have activation domains
crystallographic analysis of complexes between DNA and the that are rich in acidic amino acids (aspartic and glutamic
GCN4 DNA-binding domain has shown that the dimeric pro- acids). These so-called acidic activation domains generally
tein contains two extended a helices that "grip" the DNA are capable of stimulating transcription in nearly all types of
molecule, much like a pair of scissors, at two adjacent major eukaryotic cells-fungal, animal, and plant cells. Activation
grooves separated by about half a turn of the double helix domains from some Drosophila and mammalian transcrip-
(Figure 7-29c). The portions of the a helices contacting the tiOn factors are glutamine-rich, and some are proline-rich;
DNA include positively charged (basic) residues that interact still others are rich in the closely related amino acids serine
with phosphates in the DNA backbone and additional resi- and threonine, both of which have hydroxyl groups. How-
dues that interact with specific bases in the major groove. ever, some strong activation domains are not particularly
GCN4 forms dimers via hydrophobic interactions be- rich in any specific amino acid.
tween the C-terminal regions of the a helices, forming a Biophysical studies indicate that acidic activation domains
coiled-coil structure. This structure is common in proteins have an unstructured, random-coil conformation. These do-
containing amphipathic a helices in which hydrophobic mains stimulate transcription when they are bound to a pro-
amino acid residues are regularly spaced alternately three or tein co-activator. The interaction with a co-activator causes
four positions apart in the sequence, forming a stripe down the activation domain to assume a more structured a-helical
one side of the a helix. These hydrophobic stripes make up conformation in the activation domain-co-activator complex.
the interacting surfaces between the a-helical monomers in a A well-studied example of a transcription factor with an
coiled-coil dimer (see Figure 3-9a). acidic activation domain is the mammalian CREB protein,
.. Although the first leucine-zipper transcription factors to
be ana lyzed contained leucine residues at every seventh posi-
which is phosphorylated in response to increased levels of
cA~lP. This regulated phosphorylation is required for CREB
tion in the dimerization region, additional DNA-binding to bind to its co-activator CBP (CREB binding protein), re-
proteins containing other hydrophobic amino acids in these sulting in the transcription of genes whose control regions
positions subsequently were identified. Like leucine-zipper contain a CREB-binding site (see Figure 15-32). When the
proteins, they form dimers containing a C-terminal coiled- phosphorylated random coil activation domain of CREB in-
coil dimerization region and an N-terminal DNA-binding teracts with CBP, it undergoes a conformational change to
domain. The term basic zipper (bZIP) now is frequently form two a helices linked by a short loop, which wrap around
used to refer to all proteins with these common structural the interacting domain of CBP (Figure 7-30a).
features. Many basic-zipper transcription factors are het- Some activation domains are larger and more highly struc-
erodimers of two different polypeptide chains, each contain- tured than acidic activation domains. For example, the ligand-
ing one basic-zipper domain. binding domains of nuclear receptor~ function as activation
domains when they bind their specific ligand (Figure 7-30b, c).
Basic Helix-loop-Helix (bHlH ) Proteins The DNA-binding Binding of ligand induces a large conformational change that
domain of another class of dimeric transcription factors con- allows the ligand-binding domain with bound hormone tO in-
tains a structural motif very similar to the basic-zipper motif teract with a short a helix in nuclear-receptor co-activators;
except that a nonhelical loop of the polypeptide chain sepa- the resulting complex then can activate transcription of genes
rates two a-helical regions in each monomer (Figure 7-29d). whose control regions bind the nuclear receptor.

7.4 Regulatory Sequences in Protein-Coding Genes and the Proteins Through Which They Function 311
(a) nuclear-receptor ligand-binding activation domain is a struc-
tured globular domain that interacts with a short a helix in a
co-activator, which probably is a random coil before it is bound.
In both cases, however, specific protein-protein interactions
Domain
ofCBP CRE B
between co-activators and the activation doma ins permit the
activation transcription factors to stimulate gene expression.
doma in Currently, less is known about t he structure of rep ression
domains. The g lobular ligand-binding domains of some nu-
clear receptors function as repression domains in the absence
of their specific hormone ligand. Like activation domains,
repression domains may be relatively short, comprising 15
or fewer am ino acids. Biochemical and genetic studies indi-
cate that repression domains also med iate protein-protein
mteractions and bind to co-repressor proteins, forming a
complex that inhibits transcription initiation by mechanisms
that are discussed later in the chapter.
'

Transcription Factor Interactions Increase

Gene-Control Options
Two types of DNA-binding proteins discussed previously-
basic-zipper proteins and bHLH proteins-often exist in alter-
native heterodimeric combinations of monomers. Other classes
of transcription factors not discussed here also form heterodi-
meric proteins. In some heterodimeric tra nscription factors,
each monomer recognizes the same sequence. In these proteins,
the formation of alternative heterodimers does not increase t he
number of different sites on which the monomers can act, b ut
rather allows t he activation domains associated with each
monomer to be brought together in alternative combinations
FIGURE 7-30 Activation domains may be random coils until they that bind to the same site (Figure 7-3 1a). As we w ill see later,
interact with co-activator proteins or folded protein domains. and in subsequent chapters, the activities of individual tran-
(a) The activation domain of CREB (cyclic AMP response element-binding scription factors can be regu lated by multiple mechanisms.
protein) is activated by phosphorylation at serine 123. It is a random coil Consequently, a single bZIP or bH LH DNA regulatory cle-
until it interacts with a domain of the CBP co-act ivator (shown as a ment in the control region of a gene may elicit different tran-
space-filling surface model with negatively charged regions in red and scriptional responses depending o n which bZIP or bHLH
positively charged regions in blue). When the CREB activation domain monomers that bind to that site are expressed in a particular ..
binds to CBP, it folds into two amphipathic ex helices. Side chains in the cell at a particular time and how their activities are regu lated.
activation domain that interact with the surface of the CBP domain are
In some heterodimeric transcription factors, however,
shown. (b) The ligand-binding activation domain of the estrogen receptor
each monomer has a different DNA-binding specificity. The
is a folded-protein domain. When estrogen is bound to the domain, the
resulting combinatorial possibilities increase t he number of
green ex helix interacts with the ligand, generating a hydrophobic groove
potential DNA sequences that a family of transcription fac-
in the ligand-binding domain (dark brown helices). which binds an
amphipathic a helix in a co-activator subunit (blue). (c) The conformation
tors can bind. Three different factor monomers theoretically
of the estrogen receptor in the absence of hormone is stabilized by could combine to form six homo- and heterodimeric factors,
binding of the estrogen antagonist tamoxifen. In this conformation, the as illustrated in Figure 7-31 b. Fo ur different facror mono-
green helix of the receptor folds into a conformation that interacts with mers could form a total of 10 dimeric factors, five monomers,
the co-activator- binding groove of t he active receptor, sterically blocking 16 dimeric factors, and so forth. In addition, in hibitory fac-
binding of co-activators. [Part (a) from I. Radhakrishnan et al. (1997) Ce// 91 :741, tors are known that bind to some basic-zipper and bHLH
courtesy of Peter Wright. Parts (b) and (c) from A. K Shiau et al., 1998, Ce// 95:927.] monomers, thereby blocking their binding to DNA. When
these inhibitory factors arc expressed, they repress transcnp-
tional activation by the factors with which they interact (Fig-
Thus the acidic activation domain in CREB and the ligand- ure 7-3 1c). The ru les governing the interactions of members
binding activation domains in nuclear receptors represent two of a heterodimeric transcription factor class are complex.
structural extremes. The CREB acidic activation domain is a This combinatorial complexity expands both the number of
random coil that folds into two a helices when it binds to the DNA sites from which these factors can acti\'ate transcrip-
surface of a globular domain in a co-activator. In contrast, the tion and the ways in which they can be regu lated.

312 CHAPTER 7 • Transcriptional Control of Gene Expression

(a) DNA. However, when both NFAT and APl are present,
Factor Factor Factor protein-protein interactions between them stabilize the DNA
A B ~Activation ternary complex composed of NFAT, API, and DNA (Fig-
g ~
2:J.___ domain
DNA-binding
ure 7-32a). Such cooperative DNA binding of various tran-
scription factors results in considerable combinatorial
domain
complexity of transcription control. As a result, the = 2000
transcription factors encoded in the human genome can bind
to DNA through a much larger number of cooperative inter-
actions, resulting in unique transcriptional control for each of

(b)
Factor Factor Factor
A B ~Activation
g ~
~domain
DNA-binding
•
/Inhibitory
factor
domain

gg ~ ~ ~ ~ ~
m
Site 1
::[I]::
Site 2
m
Site 3
:::CIJ:
Site 4
:::CIJ:
Site 5
m
Site 6

(c)

~ ~ ~
Weak NFAT Weak AP1 Cooperative binding
binding site binding site of N~AT and AP1
m
Site 1
m
Site 2
m
Site 3
:::CIJ:
Site 4
m
Site 5
:::CIJ:
Site 6

FIGURE 7· 31 Combinatorial possibilities due to formation of (b)

heterodimeric transcription factors. (a) In some heterodimeric
transcription factors, each monomer recognizes the same DNA
sequence. In the hypothetical example shown, transcription factors A,
B, and C can all interact with one another, creating six different
alternative combinations of activation domains that can all bind at the
same site. Each composite binding site is divided into two half-sites,
and each heterodimeric factor contains the activation domains of its
two constituent monomers. (b) When transcription-factor monomers
recognize different DNA sequences, alternative combinations of the
three factors bind to six different DNA sequences (sites 1- 6). each with
a unique combination of activation domains. (c) Expression of an
inhibitory factor (red) that interacts only with factor A inhibits binding;
hence transcriptional activation at sites 1, 4, and 5 is inhibited, but
activation at sites 2, 3, and 6 is unaffected.

Similar combinatorial transcriptional regulation is FIGURE 7 · 32 Cooperative binding of two unrelated transcrip-
achieved through the interaction of structurally unrelated t ion factors to neighboring sites in a composite control element.
transcription factors bound to closely spaced binding sites in (a) By themselves, both monomeric NFAT and heterodimeric AP1
transcription factors have low affinity for their respective binding sites in
DNA. An example is the interaction of two transcription
the IL -2 promoter-proximal region. Protein-protein interactions between
factors, NFAT and APl, which bind to neighboring sites in
NFAT and AP1 add to the overall stability of the NFAT-AP1-DNA complex,
a composite promoter-proximal element regulating the gene
so that the two proteins bind to the composite site cooperatively.
encoding interleukin-2 (IL-2). Expression of the JL-2 gene is (b) Cooperative DNA btnding by dimeric SRF and monomeric SAP 1 can
critical to the immune response, but abnormal expression of occur when their binding sites are separated by 5 to = 30 bp and when
IL-2 can lead to autoimmune diseases such as rheumatoid the SAP-1 binding site is inverted because the B-box domain of SAP-1 that
arthritis. Neither NFAT nor APl binds to its site in the IL-2 interacts with SRF is connected to the ETS DNA-binding domain of SAP-1
control region in the absence of the other. The affinities of by a flexible linker region of the SAP-1 polypeptide chain (dotted line).
the factors for these particular DNA sequences are too low [(a) See L. Chen et al.. 1998, Nature 392:42; (b) seeM. Hassler and T. J. Richmond,
for the individual factors to form a stable complex with 2001, EMBOJ. 20:3018.]

7.4 Regulatory Sequences in Protein -Coding Genes and the Proteins Through Which They Function 313
the =25,000 human genes. In the case of IL-2, transcription DNA-protein complexes that assemble from transcription fac-
occurs only when both NFAT is activated, resulting in its tors as they bind to their multiple binding sires in an enhancer.
transport from the cytoplasm to the nucleus, and the two Because of the presence of flexible regions connecting the
subunits of API are synthesized. These events are controlled DNA-binding domains and activation or repression domains
by distinct signal transduction pathways (Cha pters 15 and in transcription factors (see Figure 7-27), and the ability of
16), allowing stringent control of IL-2 expression. interacting proteins bound to distant sites to produce loops
Cooperative binding by NFAT and API occurs only when in the DNA between their binding sires (Figure 7-4), consid-
their weak binding sites are positioned quite close to each erable leeway in the spacing between regulatory elements in
other in DNA. The sires must be located at a precise distance transcription-control regions is permissible. This tolerance
from each other for effective binding. The requirements for for variable spacing between binding sites for regulatory
cooperative binding are not so stringent in the case of some transcription factors and promoter-binding sites for the general
other transcription factors and control regions. For example, transcription factors and Pol II probably contributed to rapid
the EGR-1 control region contains a composite binding sire evolution of gene control in eukaryotes. Transposition of
to which the SRF and SAPl transcription factors bind coop- DNA sequences and recombination between repeated se-
eratively (sec Figure 7-32b). Because a SAPl has a long, flexi- quences over evolutionary time likely created new combina-
ble domain that interacts with SRF, the two proteins can bind tions of control clements that were subjected to natural
cooperatively when their individual sites in DNA are sepa- selection and retained if they proved beneficial. The latitude
rated by any distance up to =30 base pairs or are inverted in spacing between regulatory elements probably allowed
relative to each other. many more functional combinations to be subjected to this
evolutionary experimentation than would be the case if con-
straints on the spacing between regulatory elements were
Multiprotein Complexes Form on Enhancers strict, as for most genes in bacteria.
As noted previously, enhancers generally range in length from
about 50 to 200 base pairs and include binding sites for several
transcription factors. Analysis of the =50-bp enhancer that
regulates expression of ~-interferon, an important protein in KEY CONCEPTS of Section 7 4
defense against viral infections in vertebrates, provides a good
example of one of the few examples thus far of the structure of Regulatory Sequences in Protein-Coding Genes
the DNA-binding domains bound to the several transcription and the Proteins Through Which They Function
factor-binding sites that comprise an enhancer (Figure 7-33). Expression of eukaryotic protein-coding genes generally is
The term enhanceosome has been coined to describe such large regulated through multiple protein-binding control regions
that are located close to or distant from the transcription
start sire (Figure 7-22).
Promoters direct binding of RNA polymerase II to DNA,
determine the site of transcription initiation, and influence
transcription rate.
Three principal types of promoter sequences have been
identified in eukaryotic DNA. The TATA box is prevalent in ·.
highly transcribed genes. Initiator promoters are found in
some genes, and CpG islands, the promoters for 60-70 percent
of protein-coding genes in vertebrates, are characteristic of
genes transcribed at a low rate.
Promoter-proximal clements occur within =200 base pairs
of a start sire. Several such elements, containing =6-1 0 base
pairs, may help regulate a particular gene.
Enhancers, which contain multiple short control elements,
may be located from 200 base pairs to tens of kilobases up-
stream or downstream from a promoter, within an intron, or
downstream from the final exon of a gene.
Promoter-proximal elements and enhancers often are cell-
FIGURE 7-33 Model of the enhanceosome that forms on the
J3-interferon enhancer. Two heterodimeric factors, Jun/ATF-2 and
type specific, functioning only in specific differentiated cell types.
pSO/ Rei A (NF KB). and two copies each ofthe monomeric transcription Transcription factors, which stimulate or repress tran-
factors IRF-3 and IRF-7, bind to the six overlapping binding sites in this scription, bind to promoter-proximal regulatory elements
enhancer. [Adapted from D. Penne, T. Manntatis, and S. Hamson, 2007, Cell and enhancers in eukaryotic DNA.
129:1111.]

314 CHAPTER 7 • Transcriptional Control of Gene Expression

 chromatin-mediated transcriptional control, activators and
• Transcription activators and repressors are generally repressors interact with a large multiprotein complex called
modular proteins conta ining a single DNA-binding domain the mediator of transcrifJtion complex, or simply mediator.
and one or a few activation domains (for activators) or re- This complex in turn binds to Pol 11 and directly regulates
pression domains (for repressors). The different domains fre- a~sembly of transcription preinitiation complexes. In addi-
quently arc linked through flexible polypeptide regions (see tion, some activation domains interact with Tl-llD-TAF sub-
Figure 7-27). units or other components of the preinitiation complex,
• Among the most common structural motifs found in the interactions that contribute to prcinitiation complex assem-
DNA-binding domains of cukaryotic transcription factors are bly. Finally, activation domain5 may also inreract with the
the C 2H 2 zinc finger, homeodomain, basic helix-loop-helix elongation factor P-TEFb (CDK9-cyclin T) and other as yet
(bHLH), and basic zipper (leucine zipper). All these and many unknown factors to stimulate Pol II elongation away from
other DNA-binding motifs contain one or more o: helices that the promoter region.
interact with major grooves in their cognate site in DNA. [n this section, we review the current understandmg of how
repressors and activators control chromatin structure and pre-
• Activation and repression domains in transcription factors
initiation complex assembly. In the next section of the chapter,
exhibit a variety of amino acid sequences and three-dimensional
we discuss how the concentrations and activities of activators
structures. In general, these functional domains interact with
and repressors themselves are controlled, so that gene expres-
co-activators or co-repressors, which are critical to the ability
ston is precisely attuned to the needs of the cell and organism.
of transcription factors to modulate gene expression.
• The transcription-control regions of most genes contain
Formation of Heterochromatin Silences Gene
binding sites for multiple transcription factors. Transcription
of such genes varies depending on the particular repertoire of Expression at Telomeres, Near Centromeres,
transcription factors that are expressed and activated in a par- and in Other Regions
ticular cell at a particular time. For many years it has been clear that inactive genes 111 eu
·.· • Combinatorial complexity in transcription control results karyotic cells are often associated with heterochromatin, re-
from alternative combinations of monomers that form het- gions of chromatin that are more highly condensed and stain
crodimeric transcription factors (sec Figure 7-31) and from more darkly with DNA dyes than euchromatin, where most
cooperative binding of transcription factors to composite transcribed genes are located (see Figure 6-33a). Regions of
control sites (see Figure 7-32). chromosomes near the centromeres and telomeres and addi-
tional specific regions that vary in different cell types are or-
• Binding of multiple activators to nearby sites in an en-
ganized into heterochromatin. The DNA in heterochromatin
hancer forms a multiprotein complex called an enhancco-
is less accessible to externally added proteins than DNA in
some (sec Figure 7-33).
euchromatin and consequently is often referred to as "closed"
chromatin . For instance, in an experiment described in
Chapter 6, the DNA of inactive genes was found to be far
more resistant to digestion by DNasc I than the DNA of tran-
7.5 Molecular Mechanisms of Transcription scribed genes (see Figure 6-32).
Study of DNA regions inS. cereuisrae that behave like the
Repression and Activation
heterochromatin of higher eukaryotcs provided early insight
.· The repressors and activators that bind to specific sires in into the chromatin-mediated repressron of transcription. This
DNA and regulate expression of the associated protein-coding yeast can grow either as haploid or diploid cells. Haploid
genes do so by three general mechanisms. First, these regula- cells exhibit one of two possible mating types, called a and o:.
tory proteins act in concer..t with other proteins to modulate Cells of different mating type can "mate," or fuse, to generate
chromatin structure, inhibiting or stimulating the ability of a diploid cell. When a haploid cell divides by budding, the
general transcription factors to bind to promoters. Recall larger "mother" cell switches its mating type. Genetic and
from Chapter 6 that the DNA in eukaryotic cells is not free molecular analyses have revealed that three genetic loci on
but is associated with a roughly equal mass of protein in the yeast chromosome mcontrol the mating type of yeast cells
form of chromatin. The basic structural unit of chromatin is (Figure 7-34) . Only the central mating-type locus, termed
the nucleosome, which is composed of = 147 base pairs of MAT, is actively transcribed and expresses transcription factors
DNA wrapped tightly around a disk-shaped core of histone (a l, or o:1 and o:2) that regulate genes controlling the mating
proteins. Residues within the N-terminal region of each his- type. In any one cell, eithf'r an a oro: DNA sequence is lo-
tone, and the C-termmal regions of histones H2A and H2B, cated at the MAT. The two additional loci, termed HML and
called histone tails, extend from the surface of the nucleo- HMR, near the left and right telomere, respectivel y, contain
some and can be reversibly modified (see Figure 6-31 b). "silent" (nontranscribed) copies of the a or ex genes. These
Such modifications influence the relative condensation of sequences are transferred alternately from HMLo: or HMRa
chromatin and thus its accessibility to proteins required for into the MAT locus by a type of nonreciprocal recombination
transcription initiation. In addition to their role in such between sister chromatids during cell division. When the

7.5 Molecular Mechanisms of Transcription Repression and Activation 315

 Yeast chromosome Ill
Silencer Centromere S1lencer
Telomere 1 _ _ _____ Telomere
1 1
~~~:F'~
'--------~
~ MATaora ~

/
a sequences at MAT locus
~
a sequences at MAT locus

----
a2
FIGURE 7-34 Arrangement of mating-type loci on chromosome
a1 --a1
transcribed into mRNAs whose encoded proteins specify the mating-
Ill in the yeastS. cerevisiae. Silent (unexpressed) mating-type genes type phenotype of the cell. The silencer sequences near HML and HMR
(either a or a, depending on the strain) are located at the HML locus. bind proteins that are critical for repression of these silent loci. Haploid
The opposite mating-type gene is present at the silent HMR locus. cells can switch mating types in a process that transfers the DNA
When the a or a sequences are present at the MAT locus, they can be sequence from HML or HMR to the transcriptionally active MAT locus.

MAT locus contains the DNA sequence from HMJ.o:, the polymerase. Similar experiments conducted with various
cells behave as a. cells. When the MAT locus contains the yeast histone mutants indicated that specific interactions in-
DNA sequence from H MRa, the cells behave like a cells. volving the histone ta ils of H 3 and H4 are 'required for for-
Our interest here is how transcription of the silent mating- mation of a fully repressed chromatin structure. Other studies
type loci at HML and HMR is repressed. If the genes at these have shown that the telomeres of every yeast chromosome
loci are expressed, as they are in yeast mutants with defects in also behave like silencer sequences. For instance, when a gene
the repressing mechanism, both a and o: proteins are expressed, is placed within a few ki lobases of any yeast telomere, its
causing the cells to behave like diploid cells, which cannot expression is repressed. In addition, this repression is relieved
mate. The promoters and UASs controlling transcription of the by the same mutations in the H3 and H4 histone tails that
a and o: genes lie near the center of the DNA sequence that is interfere with repression at the silent mating-type loci.
transferred and are identical whether the sequences are at the Genetic studies led to identification of several proteins,
MAT locus or at one of the silent loci. This indicates that the RAPl and three SIR proteins, that are required for repression of
function of the rranscription factors that interact with these the silent mating-type loci and the telomeres in yeast. RAP1 was
sequences must somehow be blocked at HML and HMR but found to bind within the DNA silencer sequences associated
not at the MAT locus. This repression of the silent loci depends with HML and HMR and to a sequence that is repeated multi-
on silencer sequences located next to the region of transferred ple times at each yeast chromosome telomere. Further biochem-
DNA at HML and HMR (Figure 7-34). If the silencer is de- ical studies showed that the SIR2 protein is a histone deacetylase;
leted, the adjacent locus is transcribed. Remarkably, any gene it removes acetyl groups on lysines of the histone tails. Also, the
placed near the yeast mating-type silencer sequence by recom- RAPl, and SIR2, 3, and 4 proteins bind to one another, and
binant DNA techniques is repressed, or "silenced," even a SIR3 and SIR4 bind to theN-terminal tails of histones H3 and
tRNA gene transcribed by RNA polymerase III, which uses a H4 that are maintained in a la rgely unacetylated state by the
different set of general transcription factors than RNA poly- deacetylase activity of SIR2. Several experiments using fluores-
merase II uses, as discussed later. cence confocal microscopy of yeast cells either stained with
Several lines of evidence indicate that repression of the fluorescent-labeled antibody to any one of the SIR proteins o r
HML and HMR loci results from a condensed chromatin RAPl or hybridized to a labeled telomere-specific DNA probe
structure that sterically blocks transcription factors from in- revealed that these proteins fo rm large, condensed telomeric nu-
teracting with the DNA. In one telling experiment, the gene cleoprotein structures resembling the heterochromatin found in
encoding an F.. coli enzyme that methylates adenine residues higher eukaryotes (Figure 7-35a, b, c).
in GATC sequences was introduced into yeast cells under the Figure 7-35d depicts a model for the chromatin-mediated
control of a yeast promoter so that the enzyme was expressed. si lencing at yeast telomeres based o n these and o ther studies.
Researchers found that GATC sequences w ithin the MAT Formation of heterochromatin at telomeres is nucleated by
locus and most other regions of the genome in these cells multiple RAPl proteins bound to repeated sequences in a
were methylated, but not those within the HML and H MR nucleosome-free region at the extreme end of a telomere. A
loci. These results indicate that the DNA of the silent loci is network of protein-protein interactions involving telomere-
inaccessible to the E. coli methylase and presumably to pro- bound RAPl, three SIR proteins (2, 3, and 4 ), and hypoacety-
teins in general, including transcription facto rs and RNA lated histones H3 and H4 creates a higher-order nucleoprotein

316 CHAPTER 7 • Transcriptiona l Control of Gene Expression

 (a} Nuclei and telomeres (b}Telomeres (c) SIR3 protein

(d) Hypoacetylated histone

S r2
S r4 N-terminal tails
Sr3 ~.....,~....l-4
/1
Rap l
~~~

Sir2, Sir3, Sir4 proteins

Hypoacetyrated histone
N-terminal tails

/1

Nucleosomes condense

1 and n;ultiple telomeres

assoc1ate

EXPERIMENTAL FIGURE 7 35 Antibody and DNA probes sequence at each telomere region that lacks nucleosomes. SIR3 and
co localize SIR3 protein with telomeric heterochromatin in yeast SIR4 bind to RAPl, and SIR2 binds to SIR4. SIR2 is a histone deacetylase
nuclei. (a} Confocal micrograph 0.3 mm thick through three diploid that deacetylates the tails on the hi stones neighboring the repeated
yeast cells, each containing 68 telomeres. Telomeres were labeled by RAPl-binding site. (Middle) The hypoacetylated histone tails are also
hybridization to a fluorescent telomere-specific probe (yellow}. DNA binding sites for SIR3 and SIR4, which in turn bind additional SIR2,
was stained red to reveal the nuclei. The 68 telomeres coalesce into a deacetylating neighboring h1stones. Repetition of this process results
much smaller number of regions near the nuclear periphery. (b, c) in spreading of the region of hypoacetylated hi stones with associated
Confoca l micrographs of yeast cells labeled with a telomere-specific SIR2, SIR3, and SIR4. (Bottom) Interactions between complexes of SIR2,
hybridization probe (b) and a fluorescent-labeled antibody specific for SIR3, and SIR4 cause the chromatin to condense and several telomeres
SIR3 (c). Note that SIR3 is localized in the repressed telomeric hetero- to associate, as shown in a-c. The higher-order chromatin structure
chromatin. Similar experiments with RAPl, SIR2, and SIR4 have shown generated sterically blocks other proteins from interacting with the
that these proteins also colocalize with the repressed telomeric underlying DNA. [Parts (a)-(c) from M. Gotta et al., 1996, J. Cell Bioi. 1 34:1349;
heterochromatin. (d) Schematic model of silencing mechanism at yeast courtesy of M. Gatta, T. Laroche, and S.M. Gasser. Part (d) adapted from
telomeres. (Top left) Multiple copies of RAPl bind to a simple repeated M. Grunstein, 1997, Curr. Opm. Cell Bioi. 9:383.]
complex that includes several telomeres and in which the clearer when the eDNA encoding a human histone deacety-
DNA is largely inaccessible to external proteins. One addi- lase was found to have high homology to the yeast RPD3
tional protein, SIRl, is also required for silencing of the mating- gene, known to be required for the normal repression of a
type loci. It binds to the silencer regions associated with HML number of yeast genes. Further work showed that RPD3
and HMR together with RAPl and other proteins to initiate protein has histone deacetylase activity. The ability of RPD3 .·
assembly of a similar multiprotein silencing complex that to deacetylate histones at a number of promoters depends on
encompasses HML and HMR. two other proteins: UME6, a repressor that binds to a spe-
An important feature of this model is the dependence of cific upstream regulatory sequence (URSl), and SIN3, which
repression on hypoacetylation of the histone tails. This was is part of a large, mulnprotetn complex that also contains
shown in experiments with yeast mutants expressing histones RPD3. SIN3 also binds to the repression domain of UME6,
in which lysines in histone N-termini were substituted with thus positioning the RPD3 histone deacetylase in the com-
either arginines or glutamines or glycines. Arginine is posi- plex so it can interact with nearby promoter-associated nu-
tively charged like lysine but cannot be acetylated. Glutamine, cleosomes and remove acetyl groups from histone taillysines.
on the other hand, is neutral and simulates the neutral charge Additional experiments, using the chromatin immunopre-
of acetylared lysine, and glycine, with no side chain, also mim- cipitation technique outlined in Figure 7-16a and antibodies
ICS the absence of a positively charged lysine. Repression at to specific histone acetylated lysines demonstrated that in
tclomeres and at the silent mating-type loci was defective in wild-type yeast, one or two nucleosomes in the immediate
the mutants with glutamine and glycine substitutions but not vicinity of UME6-binding sites are hypoacetylated. These
111 mutants with arginine substitutions. Further, acetylation of DNA regions include the promoters of genes repressed by
H3 and H4 lysines interferes with binding by SIR3 and SIR4 UME6. In sin3 and rpd3 deletion mutants, not only were
and consequently prevents repression at the silent loci and these promoters derepressed, the nucleosomes near the
telomeres. Finally, chromatin immunoprecipitation experi- UME6-binding sites were hyperacetylated.
ments (Figure 7-16a) using antibodies specific for acetylated All these findings provide considerable support for the
lysines at particular positions in the histone N-terminal tails model of repressor-directed deacetylation "shown in figure
(Figure 6-3la) confirmed that histones in repressed regions 7-36a. The SIN3-RPD3 complex functions as a co-repressor.
ncar telomeres and at the silent mating loci are hypoacety- Co-repressor complexes containing histone deacetylases also
lated, but become hyperacetylated in sir mutants when genes have been found associated with many repressors from mam-
in these regions are derepressed. malian cells. Some of these complexes contain the mammalian
homolog of SIN3 (mSin3 ), which interacts with the repression
domain of repressors, as in yeast. Other histone deacetylase
complexes identified in mammalian cells appear to contain ad-
Repressors Can Direct Histone Deacetylation
ditional or different repression domain-binding proteins. These
at Specific Genes various repressor and co-repressor combinations mediate his-
The importance of histone deacetylation in chromatin-mediated tone deacetylation at specific promoters by a mechanism simi-
gene repression was further supported by studies of eukary- lar to the yeast mechanism (see Figure 7-36a). In addition to
otic repressors that regulate genes at internal chromosomal repression through the formation of "closed" chromatin struc-
positions. These proteins are now known to act in part by tures, some repression domains also inhibit the assembly of
causing deacetylation of histone tails in nucleosomes that preinitiation complexes in in vitro experiments with purified
bind to the TATA box and promoter-proximal region of the general transcription factors in the absence of histones. This
genes they repress. In vitro studies have shown that when activity probably contributes to the repression of transcription
promoter DNA is assembled onto a nucleosome with un- by these repression domains in \'ivo as well.
acetylated histones, the general transcription factors cannot
bind to the TATA box and initiation region. In unacetylated
histones, the N-terminal lysines are positively charged and
Activators Can Direct Histone
may interact with DNA phosphates. The unacerylated his-
tone rails also interact with neighboring histone octamers Acetylation at Specific Genes
and other chromatin-associated proteins, favoring the fold- Just as repressors function through co-repressors that bind to
ing of chromatin into condensed, higher-order structures their repression domains, the activation domains of DNA-
whose precise conformation is not well understood. The net binding activators function by binding multisubunit co-activator
effect is that general transcription factors cannot assemble complexes. One of the first co-activator complexes to be char-
into a preinitiation complex on a promoter associated with acterized was the yeast Si\GA complex, which functions with
hypoacerylated histones. In contrast, binding of general tran- the GCN4 activator protein described in Section 7.4. Early
scription factors is repressed much less by histones with hy- genetic studies indicated that full activity of the GCN4 activa-
peracetylated tails in which the positively charged lysines are tor required a protein called GCN5. The clue to GCNS's func-
neutralized and electrostatic interactions are eliminated. tion came from biochemical studies of a histone acetylase
The connection between histone deacetylation and re- purified from the protozoan Tetrahymena, the first histone
pression of transcription at specific yeast promoters became acetylase to be purified. Sequence analysis revealed homology

3 18 CHAPTER 7 • Transcriptional Control of Gene Expression

(a) Repressor-directed histone deacetylation FIGURE 7-36 Proposed mechanism of
histone deacetylation and hyperacetylation in

I• - Acetyl group j Deacetylation of histone

yeast transcription control. (a) Repressor·
directed deacetylation of histone N-terminal tails.
~mlooltoll< The DNA-binding domain (DBD) of the repressor
UME6 interacts with a specific upstream control
element (URS 1) of the genes it regulates. The
UME6 repression domain (RD) binds SIN3, a
subunit of a multiprotein complex that includes
RPD3, a histone deacetylase. Deacetylation of
\ histone N-terminal tails on nucleosomes in the
' '' region of the UME6·binding site inhibits binding
of general transcription factors at the TATA box,
thereby repressing gene expression. (b) Activator·
directed hyperacetylation of histone N-terminal
tails. The DNA-binding domain of the activator
GCN4 interacts with specific upstream activating
(b) Activator-directed histone hyperacetylation sequences (UAS) of the genes it regulates. The
GCN4 activation domain (AD) then interacts with a
Hyperacetylation of histone multiprotein histone acetylase complex that
GCNS ~erminal tails includes the GCNS catalytic subunit. Subsequent
hyperacetylation of histone N·terminal tails on
nucleosomes in the vicinity of the GCN4-binding
site facilitates access of the general transcription
factors required for initiation. Repression and
activation of many genes in higher eukaryotes
\ occurs by similar mechanisms..
'

between the Tetrahymena protein and yeast GCN5, which initiation complex (see Figure 7-17) . Nuclcosomes at pro-
was soon shown to have histone acetylase activity as well. moter regions of virtually all active genes are hyperacetylated.
Further genetic and biochemical studies revealed that GCN5 A similar activation mechanism operates in higher eukary-
is one subunit of'a multiprotein co-activator complex, named otes. Mammalian cells contain multisubunit histone acetylase
the SAGA complex after genes encoding some of the subunits. co-activator complexes homologous to the yeast SAGA com-
Another subunit of this histone acetylase complex binds to plex. They also express two related = 300-kDa, multidomain
activation domains in multiple yeast activator proteins, in- proteins called CBP and PJOO, which function similarly. As
cluding GCN4. The model shown in Figure 7-36b is consis- noted earlier, one domain of CBP binds the phosphorylated
tent with the observation that nucleosomcs near the promoter acidic activation domain in the CREB transcription factor.
region of a gene regulated ,by the GCN4 activator arc specifi- Other domains of CBP interact with different activation do-
cally hyperacerylated compared to most histones in the cell. mains in other activators. Yet another domain of CBP has
This activator-directed hyperacetylation of nucleosomes near histone acetylase activity, and another CBP domain associates
a promoter region opens the chromatin structure so as to fa- with additional multisubunit histone acetylase complexes.
cilitate the binding of other proteins required for transcription CREB and many other mammalian activators function in part
initiation. The chromatin structure is less condensed com- by directing CBP and the associated histone acetylase complex
pared to most chromatin, as indicated by its sensitivity to di- to specific nuclcosomes, where they acerylate histone tails, fa-
gestion with nucleases in isolated nuclei. cilitating the interaction of general transcription factors with
In addition to leading to the decondensation of chroma- promoter DNA.
tin, the aceLylation of specific histone lysmes generates bind-
ing sites for proteins with bromodomains that bind them.
Chromatin-Remodeling Factors Help
For example, a subunit of the general transcription factor
TFIID contains two bromodomains that bind to acctylated Activate or Repress Transcription
nucleosomes with high affinity. Recall that TFIID binding to In addition to histone acetylase complexes, multiprotein
a promoter initiates assembly of an RNA polymerase II pre- chromatin-remodeling complexes also are required for acnvation

7.5 Molecular Mechanisms of Transcription Repression and Activation 319

(a) (b) DNA repair. Several types of chromatin-remodeling com-
plexes are found in eukaryotic cells, all with homo logous
DNA helicase domains. SWIISNF complexes and related
chromatin-remodeli ng complexes in mu lticellular organisms
contain subunits with bromodomains that bind to acetylated
histone tails. Consequently, SWIISNF complexes remain as-
sociated with activated, acetylated regions of chromatin, pre-
sumably maintaining them in a decondensed conformation.
Chromatm-remodeling complexes can also participate in
transcriptional repression. These chromatin-remodeling com-
plexes bind to transcription repression domains of repressors
and contribute to repression, presumably by folding chroma-
tin into condensed structures. Much remains to be learned
about how this important class of proteins alters chromatin
structure to influence gene expression and other processes.
FIGURE 7 -37 Expression offusion proteins demonstrates
chromatin decondensation in response to an activation domain. The Mediator Complex Forms a Molecular Bridge
A cultured hamster cell line was engineered to contain multiple copies
of a tandem array of E. coli lac operator sequences integrated into a
Between Activation Domains and Pol II
chromosome in a region of heterochromatin. (a) When an expression Once the interaction of activation domains with histone acety-
vector for the lac repressor was transfected into these cells, lac lase complexes and chromatin remodeling complexes converts
repressors bound to the lac operator sites could be visualized in a the chromatin of a promoter region to an "open" chromatin
region of condensed chromatin using an antibody against the lac structure that allows the binding of general transcription fac-
repressor (red). DNA was visualized by staining with DAPI (blue). tors, activation domains interact with another multisubunit
revealing the nucleus. (b) When an expression vector for the lac co-activator complex, the mediator (figure 7-38). Activation
repressor fused to an activation domain was transfected into these
domain-mediator interactions stimulate assembly of the pre-
cells, staining as in (a) revealed that the activation domain causes this
initiation complex on the promoter. The head and middle do-
region of chromatin to decondense into a thinner chromatin fiber that
mains of the mediator complex are proposed ro interact
fills a much larger volume of the nucleus. [Courtesy of AndrewS. Belmont,
1999.). Cell Bioi. 145:1341.]
directly with subunits RBP3, 4, 7, and 11 of Pol II. Several
mediator subunits bind to activation domains in various acti-
vator proteins. Thus mediator can form a molecular bridge
between an activator bound to irs cognate site in DNA and
at many promoters. The first of these characterized was the Pol II at a promoter.
yeast SWI/SNF !:hromatin-remodeling complex. One of the Experiments with temperature-sensitive yeast mutants in-
SWI/SNF subunits has homology to DNA helicases, enzymes dicate that some mediator subunits are required for transcrip-
that use energy from ATP hydrolysis to disrupt interactions tion of virtually all yeast genes. These subunits most likely
between base-paired nucleic acids or between nucleic acids help maintain the overall structure of the mediator complex
and proteins. In vitro, the SWI/SNF complex is thought to or bind to Pol II and therefore are requi red for activation by
pump or push DNA into the nuclcosome so that DNA bound all activators. In contrast, other mediator subunits are re-
to the surface of the histone octamer transiently dissociates quired for normal activation or repression of specific subsets
from the surface and translocates, causing the nucleosomes of genes. D A microarray analysis of yeast gene expression in
to "slide" along the DNA. The net result of such chromatin mutants with defects in these mediator subunits indicates that
remodeling is to facilitate the binding of transcription fac- each such subunit influences transcription of =3-1 0 percent
tors to specific DNA sequences in chromatin. Many activa- of all genes to the extent that its deletion either increases or
tion domains bind to chromatin-remodeling complexes, and decreases mRNA expression by a factor of twofold or more
this binding stimulates in vitro transcription from chromatin (see Figure 5-29 for DNA microarray technique). These me-
templates (DNA bound to nucleosomes). Thus the SWI/SNF diator subunits are thought to interact with specific activation
complex represents another type of co-activator complex. domains; thus when one subunit is defective, transcription of
The experiment shown in Figure 7-37 demonstrates dramat- genes regu lated by activators that bind to that subunit is se-
ically how an activation domain can cause decondensation verely depressed, but transcription of other genes is unaf-
of a region of chromatin. This results from association of the fected. Recent studies suggest that most activation domains
activation domain with chromatin-remodeling and histone may interact with more than one mediator subunit.
acetylase complexes. The various experimental results indicating that individual
Chromatin-remodeling complexes are required for many mediator subunits bind to specific activation domains suggest
procec;ses involving DNA in eukaryotic cells, including tran- that multiple activators influence transcription from a single pro-
scription control, DNA replication, recombination, and moter by interacting with a mediator complex simultaneously

320 CHAPTER 7 • Transcriptional Control of Gene Expression

(a) Yeast mediator-Pol II complex FIGURE 7-38 Structure of yeast and human mediator
complexes. (a) Reconstructed image of mediator from
5. cerevisiae bound to Poll I. Multiple electron microscopy
images were aligned and computer-processed to produce this
average image in which the three-dimensional Poll I structure
(light orange) is shown associated with the yeast mediator
complex (dark blue). (b) Diagrammatic representation of
mediator subunits from 5. cerevisiae. Subunits shown in the
same color are thought to form a module. Mutations 1n one
subunit of a module may inhibit association of other subunits in
the same module with the rest of the complex. (c) Diagram-
matic representation of human mediator subunits. [Part (a). from
S. Hahn, 2004, Nat. Struct. Mol. Bioi. 1 1:394, based on J. Davis et al., 2002,
Mol. Cell 10:409. Part (b), from B. Guglielmi et al., 2004, Nucl. Acids Res.
32:5379. Part (c), adapted from S. Malik and R. G. Roeder, 2010, Nat. Rev.
(b) S. cerevisiae mediator Genet. 1 1:761. See H. M. Bourbon, 2008, Nucl. Acids Res. 36:3993.]

(c) Human mediator

(;) Head
Q Middle
0 Tail
Q CDK module MED30

(figure 7-39). Activators bound at enhancers or promoter- The Yeast Two-Hybrid System
proximal elements can interact with mediator associated with A powerful molecular genetic method called the yeast two-
a promoter because chromatin, like DNA, is flexible and can hybrid system exploits the flexibility in activator structures
form a loop bringing the regulatory regions and the promoter to identify genes whose products bind to a <>pecific protein of
close together, as ob5\:rved for the E. coli NtrL activator and interest. Because of the importance of protein-protein inter-
a' 4 -RNA polymerase (see Figure 7-4). The multiprotein nu- actions in virtually every biological process, the yeast two-
cleoprotein complexes that form on eukaryotic promoters may hybrid system is used widely in biological research.
comprise as many as 100 polypeptides with a total mass of This method employs a yeast vector for expressing a
=3 megadaltons (MDa), as large as a ribosome. DNA-binding domain and flexible linker region without the

7.5 Molecular Mechanisms of Transcription Repression and Activation 321

 DNA-binding yields cDNAs encoding protein domains that interact with
domain the bait domain.
Activation
·.
domain

KEY CONCfPTS of Section 7.5

Molecular Mechanisms of Transcription
Repression and Activation
• Eukaryotic transcription activators and repressors exert
their effects largely by binJing to multisubunit co-activators
or co-repressors that influence assembly of Pol II transcrip-
Ill
tion preinitiation complexes either by modulating chromatin
structure (indirect effect) or by interacting with Pol ll and
general transcription factors (direct effect).

FIGURE 7-39 Model of several DNA-bound activators interacting The DNA in condensed regions of chromatin (heterochro-
with a single mediator complex. The ability of different mediator matin) is relatively inaccessible to, transcription factors anJ
subunits to interact with specific activation domains may contribute to other proteins, so that gene expression is repressed.
the integration of signals from several activators at a single promoter. The interactions of several proteins with each other and
See the text for discussion. with the hypoacetylated N-terminal tails of histones H3 and
H4 are responsible for the chromatin-mediated repression of
transcription that occurs in the telomeres and the silent mat-
associated activation domain, such as the deleted GAL4- ing-type loci inS. cerevisiae (see Figure 7-3~).
containing amino acids l-692 (see Figure 7-26b). A eDNA
Some repression domains function by interacting with co-
sequence encoding a protein or protein domain of interest,
repressors that are histone deacetylase complexes. The subse-
called the bait domam, is fused in frame to the flexible linker
quent deacetylation of histone N-terminal tails in nucleosomes
region so that the vector will express a hybrid protein com-
near the repressor-binding site inhibits interaction between
posed of the DNA-binding domain, linker region, and bait
the promoter DNA and general transcription factors, thereby
domain (Figure 7-40a, left). A eDNA library is cloned into
repressing transcription initiation (see Figure 7-36a).
multiple copies of a second yeast vector that encodes a strong
activation domain and flexible linker to produce a vector li- Some activation domains function by binding multiprotein
brary expressing multiple hybrid proteins, each containing a co-activator complexes such as histone acetylase complexes.
different fish domain (Figure 7-40a, right). The subsequent hyperacetylation of histone N-terminal tails
The bait veqor and library of fish vectors are then trans- in nucleosomes near the activator-binding site facilitates inter-
fecteJ into engineered yeast cells in which the only copy of a actions between the promoter DNA and general transcription
gene required for histidine synth esis (HIS) is under control factors, thereby stimulating transcription initiation (see Figure
of a UAS with binding sites for the DNA-binding domain of 7-36b).
the hybrid bait protein. Transcription of the HIS gene re- SWl/SNf chromatin-remodeling factors constitute another
quires activation by proteins bound to the UAS. Trans- type of co-activator. These multisubunit complexes can tran-
formeJ cells that express the bait hybrid and an interacting siently dissociate DNA from histone cores in an ATP-dependent
fish hybrid will he able to activate transcription of the HIS reaction and may also decondense regions of chromatin, thereby
gene (Figure 7-40b). This system works because of the flex- promoting the binding of DNA-binding proteins needed for ini-
ibility in the spacing between the DNA-binding and activa- tiation to occur at some promoters.
tion domains of eukaryotic activators.
Mediator, another type of co-activator, is an =30-subu nit
A two-step selection process is used (Figure 7-40c). The
complex that forms a molecular bridge between activation
bait vector also expresses a wild-type TRP gene, and the hy-
domains and RNA polymerase II by binding directly to the
briJ vector expresses a wild-type LEU gene. Transfected cells
polymerase and activation domains. By binding to several dif-
are first grown in a medium that lacks tryptophan and leu-
ferent activators simultaneously, mediator probably helps in-
cine hut contains histidine. Only cells that have taken up the
tegrate the effects of multiple activators on a single promoter
bait vector and one of the fish plasmids will survive in this
(see Figure 7-39).
medium. The cell<, rh::n <;nrvive then are plated on a medium
that lacks histidine. Those cells expressing a fish hybrid that Activaton bound to a distant enhancer can interact with
does not binJ to the bait hybrid cannot transcribe the HIS transcription factors bound to a promoter because DNA is
gene and consequently will not form a colony on medium flexible and the intervening DNA can form a large loop.
lacking histidine. The few cells that express a bait- binding • The highly cooperative assembly of preinitiation complexes
fish hybrid will grow and form colonies in the absence of in vivo generally requires several activators. A cell must produce
histidine. Recovery of the fish vectors from these colonies

322 CHAPTER 7 • Transcriptional Control of Gene Expression

(;} TECHNIQUE ANIMATION: YeastTwo-Hybrid System
(a) Hybrid proteins EXPERIMENTAL FIGURE 7"40 The yeast two-hybrid system
DNA-binding Bait Activation provides a way of screening a eDNA library for clones encoding
domain domain fi'h~o;o protei ns that interact with a specific protein of interest. (a) Two

0<3 Bait hybrid Fish hybrid

vectors are constructed containing genes that encode hybrid (chime-
ric) proteins. In one vector (left), the coding sequence for the DNA·
binding domain of a transcription factor is fused to the sequences for a
known protein, referred to as the "bait" domain (light blue). The second
vector (right) expresses an activation domain fused to a "fish" domain
(b) Transcriptional activation by hybrid proteins in yeast
(green) that interacts with the bait domain. (b) If yeast cells are
transformed with vectors expressing both hybrids, the bait and fish
portions of the chimeric protein s interact to produce a functional
transcriptional activator. In this example, the activator promotes
Transfect yeast cells transcription of a HIS gene. One end of this protein complex binds to

1
with genes encoding
bait and fish hybrids

/
Co-activators and
transcription pre-
the upstream activating sequence (UAS) of the HI53 gene; the other
end, consisting of t he activation domain, stimulates assembly of the
transcription preinitiation complex at the promoter (yellow). (c) To
initiation complex screen a eDNA library for clones encoding proteins that interact w ith a
particu lar bait protein of interest, t he library is cloned into the vector
encoding the activation domain so that hybrid proteins are expressed.
The bait vector and fish vectors contain wild-type selectable genes
+ (e.g., a TRP or LEU gene). The only transformed cells that survive the
HISmRNA indicated selection scheme are those that express the bait hybrid and a
fish hybrid that interacts with it. See the text for discussion. [SeeS. Fields
(c) Fishing for proteins that interact with bait domain and 0. Song, 1989, Nature 340:245.]

O Q
Baitgene Fish eDNA from
library

TRP LEU
Bait vector Fish vector
7.6 Regulation of Transcription-
1. Transfect into trp, leu, his
mutant yeast cells Factor Activity
2. Select for cells that grow in
absence of tryptophan We have seen in the preceding discussion how comhinations
and leucine
3. Plate-selected cells on medium
of activators and repressors that bind to specific DNA regu-
lacking histidine latory sequences control transcription of eukaryotic genes.
Whether or not a specific gene in a mu lticel lular organism is

Si
~ B~~rid
~ Bait e"pressed in a particular cell at a particular time is largely a

~Q~
consequence of the nuclear concentrations and activities of

Q~
the tra nscription factors that interact with the regulatory se-
quences of tha t gene. (Exceptio ns are due to "transcriptional

~~
memory" of the functions of activators and repressors ex-
'~ HI
Bait-interacting Noninteracting
pressed in embryonic cells from which the cell has descended
as the result of efJigenetic mechanisms discussed in the next
hybrid hybrid
section.) Which t ranscription factors are expressed in a par-
ticular cell type, and the amounts produced, are determined
hy mu ltiple regu latory interactions between transcription-
Colony No colony fac tor genes t hat occur during t he development and differen-
formation formation
tiation of a particular cell type.
Jn addition to controlling the expression of thousands of
specific transcription factors, cells also regulate the activities
of ma ny of the transcription factors expressed in a particular
t he specific set of activators requireJ fur transcription of a cell type. For example, transcription factors are often regu-
particular gene in order to express tha t gene. lated in response to extracellu lar signals. Interactions be-
• The yeast two-hybrid system is widely used to detect cDNAs tween the extracell ular domains of transmembrane receptor
encod ing protein d omains that hind to a specific protein of proteins on the surface of the cell and specific protein ligands
interest (sec figure 7-40). for these receptors activate protein domains associated with
the intracellu lar domains of these tra nsmembrane proteins,

7.6 Regulation of Transcnpt1on- Factor Act1v1ty 323

FIGURE 7-41 Examples of hormones that bind to
nuclear receptors. These and related lipid-soluble
hormones diffuse through the plasma and nuclear
membranes and bind to receptors located in the cytosol
or nucleus. The ligand-receptor complex functions as a
Retinoic acid
transcription activator.

0
Co rtisol

Thyroxine

transducing the signal received on the outside of the cell to a All the nuclear receptors have a unique N-terminal region of
signal on the inside of the cel l that event ua ll y reaches tran- variable length (100-500 amino acids). Portions of this vari-
scription factors in the nucleus. Jn Chapter 16, we describe able region fu nction as activation doma ins in most nuclear
the major types of cell-surface receptors and intracellular sig- receptors. The DNA-binding domain maps ncar the center of
naling pathways that regulate transcription-factor activity. the primary sequence and has a repeat of the c4 zinc-finger
In this section, we discuss t he second major group of extra- motif (Figure 7-29 b). T he hormone-binding doma in, located
cellular signals, the small, lipid-soluble hormones-including nea r the C-terminal end, contains a hormone-dependent ac-
many different steroid hormones, retinoids, and thyroid hor- tivation domain (see Figure 7-30b, c). In some nuclear recep-
mones-that can diffuse through plasma and nuclear mem- tors, the hormone-binding domain functions as a repression
branes and interact d irectly with t he transcription factors they domain in the absence of ligand. ·
control (rigurc 7-41). As noted earlier, the intracellular recep-
tors for most of these lipid-soluble hormones, which constitute
the nuclear-receptor superfamily, function as transcription ac-
Nuclear-Receptor Response Elements
tivators when bound to their ligands. Contain Inverted or Direct Repeats
The characteristic n ucleotide sequences of the DNA sites,
called response elements, that bind several nuclear receptors
All Nuclear Receptors Share a Common
have been determined . The sequences of the consensus re-
Domain Structure sponse elements fo r the glucocorticoid and estrogen receptors
~equencing of cDNAs derived from mRNAs encoding various are 6-bp inverted repeats separated by any three base pai rs
nuclear recepto~s revealed a remarkable conservation in their (Figure 7-43a, b). This finding suggested that the cognate ste-

.. ,,
amino acid sequences and three functional regions (Figure 7-42). roid hormone receptors would bind to DNA as symmetrica l .·

,, 1553 Estrogen receptor (ER )

,,
,, .,
1 ~ 408
1946
1777
Progesterone recepto r (PR)

Glucocorticoid receptor (GR)

Thyroxine receptor (TR)

1 ~ 432 Retinoic acid receptor (RAR)

N -JL- - - r - - - - --__,J- L...L..-----.:---L.....--'f- C General primary structure

7 ~ \
Variable region DNA-bi nding Ligand-bi nding
(100-500 aa) domain (68 aa) domain (2 25-285 aa)
Amino acid identity: 0 42-94% 15-57%
FIGURE 7-42 General design of transcription factors in the Figure 7-29b). The (-terminal hormone-binding domain exhibits
nuclear-receptor superfamily. The centrally located DNA-binding somewhat less homology. The N-termina l regions in various receptors
domain exhibits considerable sequence homology among different vary in length, have unique sequences, and may contain one or more
receptors and contains two copies of the C4 zinc-finger motif (see activation domains. [SeeR. M. Evans, 1988, Science 240:889.]

324 CHAPTER 7 • Transcriptional Control of Gene Expression

 5' AGAACA(N) 3 TGTTCT 3'
receptors. Heterodimeric nuclear receptors (e.g., RXR-VDR,
(a) GRE RXR-TR, and RXR-RAR) are located exclusively in the nu-
3' TCTTGT(N) 3 ACAAGA 5'
cleus. In the absence of their hormone ligand, they repress
transcription when bound tO their cognate sites in DNA. They
5' AGGTCA(N) 3 TGACCT 3' do so by directing histone deacetylation at nearby nucleo-
(b) ERE somes by the mechanism described earlier (see Figure 7-36a).
3' TCCAGT(N) 3 ACTGGA 5'
~
In the ligand-bound conformation, heterodimeric nuclear re-
ceptors containing RXR can direct hyperacetylation of his-
5' AGGTCA(N)~GGTCA 3' tones in nearby nucleu:.omes, thereby reversing the reprcssmg
(c) VDRE
3' TCCAGT(N) 3 TCCAGT 5' effects of the free ligand-binding domain. ln the presence of
ligand, ligand-binding domains of nuclear receptors also bind

5' AGGTCA(N)4 AGGTCA 3'

. mediator, stimulating preinitiation complex assembly .
In contrast to heterodimeric nuclear receptors, homodi-
(d) T RE
3' TCCAGT(N) 4 TCCAGT 5' meric receptors are found in the cytoplasm in the absence of
their ligands. Hormone binding to these receptors leads to

5' AGGTCA(N) 5AGGTCA 3'

. their translocation to the nucleus. The hormone-dependent
translocation of the homodimeric glucocorticoid receptor
(e) RARE
3' TCCAGT(N) 5 TCCAGT 5' (GR) was demonstrated in the transfection experiments shown
in Figure 7-44. The GR hormone-binding domain alone medi-
FIGURE 7 - 43 Consensus sequences of DNA response elements
that bind three nuclear receptors. The response elements for the
ates this transport. Subsequent studies showed that, in the ab-
glucocorticoid receptor (GRE) and estrogen receptor (ERE) contain
sence of hormone, GR is anchored in the cytoplasm m a large
inverted repeats that bind these homodimeric proteins. The response protein complex with inhibitor proteins, including Hsp90, a
elements for heterodimeric receptors contain a common direct repeat protein related to Hsp70, the major heat-shock chaperone in
separated by three to five base pairs for the vitamin 0 3 receptor (VDRE). eukaryotic cells. As long as the receptor is confined to the cy-
thyroid hormone receptor (TRE), and retinoic acid receptor (RARE). The toplasm, it cannot interact with target genes <:~nd hence cannot
repeat sequences are indicated by red arrows. [SeeK. Umesono et al., activate transcription. Hormone binding to a homodimeric
1991, Ce//65:1255, and A.M. Naar et al., 1991 , Ce//65:1267.) nuclear receptor releases the inhibitor proteins, allowing the
receptor to enter the nucleus, where it can bind to response
elements associated with target genes (Figure 7-44d). Once the
receptor with bound hormone binds to a response clement, it
dimers, as was later shown from the x-ray crystallographic
activates transcription by interacting with chromatin-remodel-
analysis of the homodimeric glucocorticoid receptor's C 4
ing and histone acetylase complexes and mediator.
zinc-finger DNA-binding domain (see Figure 7-29b).
Some nuclear-receptor response elements, such as those
for the receptors that bind vitamin D 3, thyroid hormone, and Metazoans Regulate the Pol II Transition
retinoic acid, are direct repeats of the same sequence recog- from Initiation to Elongation
nized by the estrogen receptor, separated by three to five base
A recent unexpected discovery that resulted from application
pairs (Figure 7-43c-e). The specificity for responding to these
of the chromatin immunoprecipitation technique is that a
different hormones by binding distinct receptors is determined
large fraction of genes in metazoans have a paused elongating
by the spacing between the repeats. The receptors that bind to
Pol II within =200 base pairs of the transcription start site
such direct-repeat response elements do so as heterodimers
(Figure 7-16). T h us expression of the encoded protein is con-
with a common nuclear-receptor monomer called RXR. The
trolled not only by transcription initiation, but also by tran-
vitamin D 3 response element, for example, is bound by the
scription elongation early in the transcription unit. The first
RXR- VDR heterodimer, ;nd the retinoic acid response ele-
genes discovered to be regu lated by controlling transcription
ment is bound by RXR-RAR. The monomers composing
elongation were heat-shock genes (e.g., hsp70) encoding
these heterodimers interact with each other in such a way that
protein chaperonins that help to refold denatured proteins
the two DNA-binding domains lie in the same rather than in-
and other proteins that help the cell to deal with denatured
verted orientation, allowing the RXR heterodimers to bind to
proteins. When heat shock occurs, the heat-shock transcrip-
direct repeats of the binding site for each monomer. In con-
tion factor (HSTF) is activated. Binding of activated HSTF
trast, the monomers in homodimeric nuclear receptors (e.g.,
to specific sites in the promoter-proximal region of heat-
GRE and ERE) have an inverted orientation.
shock gene~ stimulates the paused polymerase to continue
chain elongation and promotes rapid reinitiation by addi-
tional RNA polymerase II molecules, leading to many tran-
Hormone Binding to a Nuclear Receptor
. ' Regulates Its Activity as a Transcription Factor
scription initiations per minute. This mechamsm of
transcriptional control permits a rapid response: these genes
The mechanism whereby hormone binding controls the activity are always paused in a state of suspended transcription and
of nuclear receptors differs for heterodimeric and homodimeric therefore, when an emergency arises, require no time to

7.6 Regulation of Transcription -Factor Activity 325

0 VIDEO: Hormone-Regulated Nuclear Translocation ofthe Glucocorticoid Receptor

(a) (b) (c)

-Dex

+ Dex

Proteins
expressed:
N~C
~
N~C
'--v---------'
N---tJ-c
[3·Galactosidase Glucocorticoid GR ligand-binding
receptor domain
EXPERIMENTAL FIGURE 7··44 Fusion proteins from expres-
sion vectors demonstrate that the hormone-binding domain of the (d)
glucocorticoid receptor (GR) mediates translocation to the nucleus Exterior
in the presence of hormone. Cultured animal cells were transfected
with expression vectors encoding the proteins diagrammed at the
bottom. Immunofluorescence with a labeled antibody specific for
13-galactosidase was used to detect the expressed proteins in trans-
fected cells. (a) In cells that expressed 13-galactosidase alone, the
enzyme was localized to the cytoplasm in the presence and absence of
the glucocorticoid hormone dexamethasone (Dex). (b) In cells that
expressed a fusion protein consisting of 13-galactosidase and the entire
glucocorticoid receptor (GR), the fusion protein was present in the
cytoplasm in the absence of hormone but was transported to the
nucleus in the presence of hormone. (c) Cells that expressed a fusion
protein composed of 13-galactosidase and just the GR ligand-binding
domain (light purple) also exhibited hormone-dependent transport of
the fusion protein to the nucleus. (d) Model of hormone-dependent
gene activation by a homodimeric nuclear receptor. In the absence of
hormone, the receptor is kept in the cytoplasm by interaction between
its ligand-binding domain (LBD) and inhibitor proteins. When hormone
is present, it diffuses through the plasma membrane and binds to the
ligand-binding domain, causing a conformational change that releases capable of differentiation into any cell type. The abi lity to
the receptor from the inhibitor proteins. The receptor with bound induce differentiated cells to convert to pluripotent stem cells
ligand is then translocated into the nucleus, where its DNA-binding has elicited enormous research interest because of its poten-
domain (DBD) binds to response elements, allowing the ligand-binding
tial for the development of therapeutic treatments for trau-
domain and an additional activation domain (AD) at theN-terminus to
matic injuries to the nervous system and degenerative diseases
stimulate transcription of target genes. [Parts (a)-( c) from D. Picard and K. R.
(Chapter 21).
Yamamoto, 1987, EMBO J. 6:3333; courtesy of the authors.]

Pol II Termination Is Also Regulated

Once Pol II has transcribed =200 nucleotides from the tran-
remodel and acetylate chromatin over the promoter and as- scription start site, elongation through most genes is highly
semble a transcription preinitiation complex. processive, although the L.brumatin immunoprecipitation
Another transcription factor shown to regulate transcrip- with antibody to Pol II indicates that the amount of Pol II at
tion by controlling elongation of Pol II paused near the tran- various positions in a transcription unit in a population of
scription start site is MYC, which functions in the regulation cells varies greatly (Figure 7-16b, right). This indicates that
of cell growth and division. MYC is often expressed at high the enzyme can elongate through some regions much more
level in cancer cells and is a key transcription factor in the re- rapidly than others. In most cases, Pol II does not terminate
programming of somatic cells into pluripotent stem cells until after a sequence is transcribed that directs cleavage and

326 CHAPTER 7 • Transcriptional Control of Gene Expression

 polyadenylation of the RNA at the sequence that forms the blood cells, or the cells that generate platelets involved in
3' end of the encoded mRNA. RNA polymerase II then can blood clotting. Lymphoid and myeloid stem cells both have
terminate at multiple sites located over a distance of 0.5-2 kb the identical DNA sequence as the zygote generated by fer-
beyond this poly(A) addition site. Experiments with mutant tilization of the egg cell by a sperm cell from which all cells
genes show that termination is coupled to the process that develop, but they have restricted developmental potential
cleaves and polyadenylates the 3' end of a transcript, which because of epigenetic differences between them. Such epigen-
is discussed in the next chapter. etic changes are initially the consequence of the expression of
specific transcription factors that are master regulators of cel-
lular differentiation, cunLrolling the expression of other genes
encoding transcription factors and proteins involved in cell-
KEY CONCEPTS of Section 7 6 cell communication in complex network!. of gene control that
Regulation of Transcription-Factor Activity are currently the subject of intense investigation. Changes in
gene expression initiated by transcription factors are often
The activities of many transcription factors are indirectly reinforced and maintained over multiple cell divisions by
regulated by binding of extracellular proteins and peptides post-translational modifications of histones and methylation
to cell-surface receptors. These receptors activate intracellu- of DNA at position 5 of the cytosine pyrimidine ring (fig-
lar signal-transduction pathways that regulate specific tran- ure 2-17) that are maintained and propagated to daughter
scription factors through a variety of mechanisms discussed cells when cells divide. Consequently, the term epigenetic is
in Chapter 16. used to refer to such post-translational modi fica nons of his-
• Nuclear receptors constitute a superfamily of dimeric C 4 tones and 5-methyl C modification of DNA.
zinc-finger transcription factors that bind lipid-soluble hor-
mones and interact with specific response elements in DNA
(see Figures 7-41-43). Epigenetic Repression by DNA Methylation
• Hormone binding to nuclear receptors induces conforma- As mentioned earlier, most promoters in mammals fall into the
tional changes that modify their interactions with other pro- CpG island class. Active CpG island promoters have Cs in CG
teins (Figure 7-30b, c). sequences that are unmethylated. Unmethybted CpG island
promoters are generally depleted of histone octamers, but nu-
.· • Heterodimeric nuclear receptors (e.g., those for retinoids, cleosomes immediately neighboring the unmethylated CpG is-
vitamin D, and thyroid hormone) are found only in the nu- land promoters are modified by histone H3 lysine 4 di- or
cleus. In the absence of hormone, they repress transcription trimethylation and have associated Pol II molecules that are
of target genes with the corresponding response element. paused during transcription of both the sense and non-sense
When bound to their ligands, they activate transcription. template DNA strands, as discussed earlier (Figures 7-16
• Steroid hormone receptors are homodimeric nuclear re- and 17). Recent research indicates that methylation of histone
ceptors. ln the absence of hormone, they are trapped in the H3lysine 4 occurs in mouse cells because a protein named Cfpl
cytoplasm by inhibitor proteins. When bound to their ligands, (CXXC finger protein 1) binds unmethylated CpG-rich DNA
they can translocate to the nucleus and activate transcription through a zinc-finger domain (CXXC) and associates with a
of target genes (.see Figure 7-44). histone methylase specific for histone H3 lysine 4 (Setd1).
Chromatin-remodeling complexes and the general transcrip-
tion factor TFIID, which initiates Pol II preinitiation complex
assembly (Figure 7-17), associate with nucleosomes bearing
the H3 lysine 4 trimethyl mark, promoting Pol II transcription
initiation.
7.7 Epigenetic Re_gulation of Transcription
However, in differentiated cells, a few percent of specific
The term epigenetic refers to inherited changes in the pheno- CpG island promoters, depending on the cell type, have
type of a cell that do not result from changes in DNA se- CpGs marked by 5-methyl C. This modification of CpG is-
quence. For example, during the differentiation of bone marrow land DNA triggers chromatin condensation. A family of pro-
stem cells into the several different types of blood cells, a hema- teins that bind to DNA rich in 5 methyi-C modified CpGs
topoietic stem cell (HSC) divides into two daughter cells, one (methyl CpG-binding proteins, MBDs) associate v.ith his-
of which continues to have the properties of an HSC with the tone deacetylases and repress chromatin-remodeling com-
potential to differentiate into all of the different types of blood plexes that condense chromatin, resulting in transcriptional
cells. But the other daughter cell becomes either a lymphoid repression. These methyl groups are added by de novo DNA
stem cell or a myeloid stem cell (see Figure 21-18). Lymphoid methyl transferases named DNMT3a and DNMT3b. Much
stem cells generate daughter cells that differentiate into lym- remains to be learned about how these enzymes are directed
phocytes, which perform many of the functions involved in to specific CpG islands, but once they have methylated a
immune responses to pathogens (Chapter 23). Myeloid stem DNA sequence, methylation is passed on through DNA rep-
cells divide into daughter cells that are committed to differen- lication through the action of the ubiquitous mamtenance
tiating into red blood cells, different kinds of phagocytic white methyl transferase DNMT1:

7.7 Epigenetic Regulation of Transcnption 327

 5' eMeG 3' 5' eMeG 3'
eMeG DNA 3 G-e 5 DNMTI 3' G-eM• 5'
5' 3'
3' G-eM• 5' Replication 5' e-G 3 5' eN'G 3'
3' G-eM• 5' 3' G - eM• 5'

(Red indicates daughter strands.) DNMT 1 also maintains side chain (see figure 2-14). Lysines can be modified by the
methylation of the Cs in CpG sequences that are statisticall) addition of one, two, or three methyl groups to this terminal
underrepre<;ented throughout most of the genome. As dis- nitrogen atom, generating mono-, di-, and trimcthylated ly-
cussed above, the CG sequence is underrepresented in most sine, all of wh ich carry a single positi\'e charge. Pulse-chase
of the ~equence of mammalian genomes, probably because radiolabeling experiments have shown that acetyl groups on
spontaneous deamination of 5-methyl C generates thymi - histone lysines turn over rapid ly, whereas methyl groups arc
dine, leading to the substitution of CpGs with TpGs over the much more stable. The acetylation state at a specific histone
period of mammalian evolution, unless there is selection lysine on a particular nucleosome results from a dynamic
against the resulting mutation, as probably occurs when equilibrium between acetylation and deacetylation by histone
CpG island promoters are mutated. This mechanism of epi- acetylases and histone deacetylases, respectively. Acetylation
genetic repression is intensely investigated because tumor- of histones in a localized region of chromatin predominates
suppressor genes encoding proteins that function to suppress when local DNA-bound activators transiently bind histone
the development of cancer are often inactivated in cancer acetyla<;e complexes. De-acetylation predominates when re-
cells by abnormal CpG methylation of their promoter re- pressors transiently bind histone deacetylase complexes.
gions, as discussed further in Chapter 24. In contrast to acetyl groups, methyl groups on histone
lysines are much more stable and turn over much less rapid ly
Histone Methylation at Other Specific Lysines than acetyl groups. Histone lysine methyl groups can be re-
moved by histone lysine demethylases. But the resulting
Are Linked to Epigenetic Mechanisms
turnover of histone lysine methyl groups is much slower
of Gene Repression than the turno\er of h1stone lys ine acetyl groups, making
Figure 6-31 b summarizes the different types of post- them appropriate post-translational modifications for prop-
translational modifications that arc found on histoncs, in- agating epigenetic information. Several other post-transla-
cluding acetylation of lysines and methylation of lysines on tional modifications have been characterized on histones
the nitrogen atom of the terminal £-amino group of the lysine (Figure 6-31 b). These all have the potential to positively or

'LH'*I'
Modif ication
Histone Post-Translational Modifications Associated with Active and Repressed Genes

Sites of Modificat ion Effect on Transcription

Acerylated lysine H3 (K9, K14, Kl8, K27, K56 ) Activation

H4 (K5, KS, Kl3, K16 )
H2A (K5, K9, Kl3l
H2B (K5, Kl2, Kl5, K20)

Hypoacetylated lysine Repression

Phosphorylated H3 (T3, S 10, S28) Activation

serine/threonine H2A (S I, Tl20 )
H2B (S 14)

Methylated argmme H3 (Rl7, R23) Activation

H4 (R3)

Methylated lysine H3 (K4 ) Me3 in promoter region Activation

HJ (K4) ~lei m enhancers
HJ (K36, K""9 ) in transcribed reg1on Elongation
H3 (K9, K27) Repress ion
H4 (K20 )

Ubiquitmated lysine H2B (K120 in mammals, K123 inS. ceret,istae) Activation

H2A (Kl19 in mammals) Repression

328 CHAPTER 7 • Transcriptional Co ntrol of Gene Expression

 negatively regulate the binding of proteins that interact with methyl transferase subunits that methylate histone H3 at lysine
the chromatin fiber to regulate transcription and other pro- 9, generating di- and trimethyllysines. These methylated lysmes
cesses such as chromosome folding into the highly condensed are binding sites for isoforms of HPl protein that function in
structures that form during mitosis (Figures 6-39 and 6-40). the condensation of heterochromatin, as discussed in Chapter
A picture of chromatin has emerged in which histone tails 6 (see Figure 6-34). For example, the KAP1 co-repressor com-
extending as random coils from the chromatin fiber are post- plex functions with a class of more than 200 zinc-finger tran-
translationally modified to generate one of many possible scription factors encoded in the human genome. This
combinations of modifications that regulate transcription co-repressor complex includes an H3 lysine 9 methyl transfer-
and other processes by regulating the binding of a large a:.e that methylates nucleosomes over the promoter region of
number of different protein complexes. This control of pro- repressed genes, leading to HP1 binding and repression of
tein interactions with specific regions of chromatin resulting transcription. An integrated transgene in cultured mouse fi-
from the combined influences of various post-translational broblasts that was repressed through the action of the KAP1
modifications of histones has been called a histone code. co-repressor associated with heterochromatin in most cells,
Some of these modifications, such as histone lysine acetyla- whereas the active form of the same transgcne associated
tion, are rapidly reversible, whereas others, such as histone with euchromatin (Figure 7-45). Chromatin immunoprecipi-
lysine methylation, can be templated through chromatin rep- tation assays (see Figure 7-16) showed that the repressed gene
lication, generating epigenetic inheritance in addition to in- was associated with histone H3 methylated at lysine 9 and
heritance of DNA sequence. Table 7-3 summarizes the HPl, whereas the active gene was not.
influence that post-translational modifications of specific Importantly, H3 lysine 9 methylation is maintained follow
histone amino acid residue~ usually have in transcription. ing chromosome replication in S phase by the mechanism dia-
grammed in Figure 7-46. When chromosomes replicate in S
Histone H3 Lysi ne 9 Methylation in Hetero chromat in In phase, the nucleosomes associated with the parental DNA are
most eukaryotes, some co-repressor complexes contain histone randomly distributed to the daughter DNA molecules. New

Active Repressed

..·

Transgene Heterochromatin Transgene

FIGURE 7 -45 Association of a repressed transgene with fluorescently labeled complementary probe (green). When the
heterochromatin. Mouse fibroblasts were stably transformed with a recombinant repressor was retained in the cytoplasm, the transgene
transgene with binding sites for an engineered repressor. The repressor was transcribed (left) and was associated with euchromatin in most
was a fusion between a DNA-binding domain, a repression domain that cells. When hormone was added so that the recombinant repressor
interacts with the KAPl co-repressor complex, and the ligand-binding entered the nucleus, the transgene was repressed (right) and associ·
domain of a nuclear receptor that allows the nuclear import of the ated with heterochromatin. Chromatin immunoprecipitation assays
fusion protein to be controlled experimentally (see Figure 7-44). DNA (see Figure 7-16) showed that the repressed gene was associated with
was stained blue with the dye DAPI. Brighter-staining regions are histone H3 methylated at lysine 9 and HPl, whereas the active gene
regions of heterochromatin, where the DNA concentration is higher was not. [Courtesy of Frank Rauscher, from Ayyanathan et al., 2003, Genes
than in euchromatin. The transgene was detected by hybridization of a Dev. 17:1855.]

7.7 Epigenetic Regulation of Transcription 329

 ~
combinations of H ox transcription facto rs help to direct the
development of specific tissues and organs in a developing
embryo. Early in embryogenesis, expression of Hox genes is
Replication controlled by typical activator and repressor protei ns. How-

~
ever, the expression of these activators and repressors stops
at an early point in embryogenesis. Correct expression of the
Hox genes in the descendants of the early embryonic cells is
then maintained throughout the remainder of embryogenesis

~
3K9 and on into adult life by the Polycomb proteins w hich main-
HMT tain the represswn of specific Hox genes. Trithorax proteins

~
perform the opposing function to Polycomb proteins, main-
taining the expression of t he Hox genes that were expressed
in a specific cell early in embryogenesis and in all the subse-
Methylation quent descendants of that cell. Polycomb and Trithorax pro-

~
3K9 teins control thousands of genes, including genes that
HMT regulate cell growth and division (i.e., the cell cycle, as dis-

~
cussed in Chapter 19 ). Polycomb and Trithorax genes arc
often mutated in cancer cells, contributing importantly to
the abnormal properties of these cells (Chapter 24).
FIGURE 7-46 Maintenance of histone H3 lysine 9 methylation Remarkably, virtua ll y a ll cells in the developing embryo
during chromosome replication. When chromosomal DNA is and adu lt express a simi lar set of Polycomb and T rithorax
replicated, the parental h1stones randomly associate with the two proteins, and all cells contain the same set of Hox genes. Yet
daughter DNA molecules while unmethylated histones synthesized only the Hox genes in cells where they we re .initially repressed
during S phase comprise other nucleosomes on the sister chromo-
in early embryogenesis remain repressed, even though the
somes. Association of histone H3 lysine 9 methyl transferases (H3K9
same Hox genes in other cells remain active in the presence of
HMn with parental nucleosomes bearing the histone 3 lysine 9 di- or
the same Polycomb proteins. Consequently, as in the case of
trimethylation mark methylate the newly added unmodified nucleo-
the yeast silent mating-type loci, the expression of Hox genes
somes. Consequently, histone H3 lysine 9 methylations are maintained
during repeated cell divisions unless they are specifically removed by
is regulated by a process that involves more than simply spe-
a histone demethylase. cific DNA sequences interacting with proteins that diffuse
through the nucleoplasm .
A current model for repression by Polycomb proteins is
histone octamers that arc not methylated on lysine 9 associate depicted in Figure 7-47 . .Most Polycomb proteins are sub-
with the new daughter chromosomes, but since the parental units of one of two classes of multi protein complexes, PRC 1-
nucleosomes are associated with hoth daughter chromosomes, typc complexes and PRC2 complexes. PRC2 complexes arc
<lpproximately half of them are methylated on lysine 9. Asso- thought to act initially by associating with specific repressors
ciation of histone H3 lysine methyl transferases (directly or bound to their cognate DNA sequences early in embryogen-
indirectly) with the parental methylated nudeosomes leads esis, or ribonucleoprotein complexes containing long non-
to methylanon of the newly assembled histone octamers. coding RNAs as discussed in a later section. The PRC2
Repetition of this process with each cell division results in complexes contain histone deacetylases that inhibit tran-
maintenance of H3 lysine 9 methylation of this region of the scription as discussed above. They also contain a subun it
chromosome. [E(z) in DrosofJhila, EZH2 in mammals] with a SET do-
main, the enzymatically active domain of several histone
methyl transferases. This SET domain in PRC2 complexes
Epigenetic Control by Polycomb
mcthylates histone H3 on lysine 27, generating di- and tr i-
and Trithorax Complexes methyl lysines. The PRCl complex then binds the methyl-
Another kmd of epigenetic mark that is essential for repres- ated nuclcosomes through dimeric Pc subunits (CBXs in
sion of genes in specific cell types in multicellular animals mammals), each containing a methyl-lysine binding domain
and plants involves a set of proteins known collectively as (called a chromodomam) specific for methylated H3 lysine
Polycomb proteins and a counteracting set of proteins 27. Binding of the dimeric Pc to neighboring nucleosomes is
known as T rithorax proteins, after phenotypes of mutations proposed to condense the chromatin into a structure that
in the genes encoding them in Drosophila, where they were inhibits transcription. This is ~uppurted by electron m icros-
first discO\·ered. The Polycomb repression mechanism is es- copy studies showing that PRC l complexes cause nucleo-
sential for maintaining the repression of genes in specific somes to associate in Yitro (Figure 7-47d, e).
types of cells and in all of the subsequent cells that develop PRCl complexes also contain a ubiquitin ligase that
from them throughout the life of an organism. Important monoubiquit inates histone H2A at lysine 119 in the H2A
genes regulated by Polycomb proteins include the Hox genes, C-termina l tail. This modification of H2A inh ibits Pol II
encoding master regulatory transcription factors. Different elongation through chromatin by inhibiting the association

330 CHAPTER 7 • Transcriptional Control of Gene Expression

 (c)
PRC1 complex

(d) Nucleosomes o n D NA (e) Nucleosomes + PRC1 complex on DNA

50 nm 1

FIGURE 7-47 Model for repression by Polycomb complexes. maintain H3 lysine 27 methylation of neighboring histones. As a
(a) During early embryogenesis, repressors associate with the PRC2 consequence, PRC1 and PRC2 association with the region is maintained
complex. (b) This results in methylation (Me) of neighboring nucleo- when expression of the repressor proteins in (a) ceases. (d, e) Electron
somes on histone H3 lysine 27 (K27) by the SET-domain-containing micrograph of an ~ 1-kb fragment of DNA bound by four nucleosomes
subunit E(z). (c) The PRC1 complexes bind nucleosomes methylated at in the absence (d) and presence (e) of one PRC1 complex per five
H3 lysine 27 through a dimeric, chromodomain-containing subunit Pc. nucleosomes. [Parts (a)-(c), adapted from A. H. Lund and M. van Lohuizen,
The PRC1 complex condenses the chromatin into a repressed chroma- 2004. Curr. Opm. Cell Bioi. 16:239. Parts (d, e), from N.J. Francis, R. E. Kingston, and
tin structure. PRC2 complexes associate with PRC1 complexes to C. L. Woodcock, 2004, Science 306: 1574.]

of a histone chaperone required to remove histone octamers ylates histone H3 lysine 4, a histone methylation associated
from DNA as Pol II transcribes through a nucleosome and with the promoters of actively transcribed genes. This his-
then replaces th~m as the polymerase passes. PRC2 complexes tone modification creates a binding site for histone acetylase
arc postulated to associate with nuclcosomes bearing the his- and chromatin-remodeling complexes that promote tran-
tone H3 lysine 27 trimethylation mark, maintaining methyla- scription, as well as TFIID, the general transcription factor
tion of H3 lysine 27 in nucleosomes in the region. This results that initiates preinitiation-complex assembly (Figure 7-17).
in association of rhe chromatin w ith PRC1 and PRC2 com- Nucleosomes with the H3 lysine 4 methyl modifications arc
plexes even after expression of the initial repressor proteins in also oinding sites for specific histone demethylases that pre-
Figure 7-47a, b has ceased. This would also maintain histone vent methylation of histone H3 at lysine 9, preventing the
H3 lysine 27 merhylatiorl'and histone H2A monoubiquitina- binding of HPl, and at lysine 27, preventing the binding of
tion following DNA replication, by a mechanism analogous the PRC-repressing complexes. Likewise, a hi~tone demeth
to that diagrammed in Figure 7-46. This is a key feature of ylasc specific for histone H3 lysine 4 associates with PRC2
Polycomb repression, which is maintained through successive complexes. Nucleosomes marked with histone H3 lysine 4
cell divisions for the life of an organism(= 100 years for some methylation also are thought to be distributed to both
vertebrates, 2000 years for a sugar cone pine!). daughter DNA molecules during DNA replication, resulting
Trithorax proteins counteract the repressive mechanism in maintenance of this epigenetic mark oy a strategy similar
of Polycomb proteins, as shown in studies of expression of to that diagrammed in Figure 7-46.
the Hox transcription factor Abd-B in the DrosofJhila em-
bryo (Figure 7-48). When the Polycomb system is defective,
Noncoding RNAs Direct Epigenetic
Abd-B is derepressed in all cells of the embryo. When the
Trithorax system is defective and cannot counteract repression Repression in Metazoans
b} the Polycomb system, Abd-B is repressed in most cells, Repressing complexes also have been discovered that are com-
except those in the very posterior of the embryo. Trithorax plexes of proteins bound to RNA molecules. ln some cases, this
complexes include a histone methyl transferase that trimeth- results in repression of genes on the same chromosome from

7.7 Epigenetic Regulation of Transcription 331

 Anterior Posterior (a)
Xist RepA

wt

(b)

Scm- Abd-8
(PcG)

FIGURE 7-49 The Xist noncoding RNA encoded in the

X-inactivation center coats the inactive X chromosome in cells of
human females. (a) The region of the hu'man X-inactivation center
encoding the noncoding RNAs Xist, RepA, and Tsix. (b) A cultured
fibroblast from a fema le was analyzed by in situ hybridization with a
trx-
p robe complementary to Xist RNA labeled with a red fluorescent dye
(trxG)
(left), a chromosome paint set of probes for the X chromosome labeled
with a green fluorescent dye (center), and an overlay of the two
fluorescent micrographs. The condensed inactive X chromosome is
FIGURE 7-48 Opposing influence of Polycomb and Trithorax associated w ith Xist RNA. [Part {a) adapted from J. T. Le~. 2010, Cold Spring
complexes on expression of the Hox transcription factor Abd-8 HarborPerspect. Bioi. 2:a003749. Part {b) from C. M. Clemson et al., 1996, J. Cell
in Drosoph ila embryos. At the stage of Drosophila embryogenesis BIOI. 132:259.]
shown, Abd-B is normally expressed only in posterior segments of the
developing embryo, as shown at the top by immunostaining with a
specific anti- Abd-B antibody. In embryos with homozygous mutations In differentiated female cells, the inactive X chromosome
of Scm, a Polycomb gene (PeG) encoding a protein associated with
is associated with Xist R A-protein complexes along its en-
the PRCl complex, Abd-B expression is de-repressed in all embryo
tire length. Targeted deletion of the Xist gene (Figure .5-42)
segments. In contrast, in homozygous mutants of trx, a Trithorax gene
in cultured embryonic stem cel ls showed that it is required
(trxG), Abd-B repression is increased so that it is only expressed to high
level in the most posterior segment. [Courtesy of Juerg Mueller, European
for X inactivation. As opposed to most protein-coding genes
Molecular Biology Laboratory.] on the inactive X chromosome, Xist is transcribed from the
X-inactivation center of the o therwise mostly inactive X
chromosome. The Xist RNA-protein complexes do not diffuse
which the RNA is transcribed, as in the case of X-chromosome to interact with the acuve X chromosome, but remain asso-
inactivation in female mammals. In other cases, these repressing ciated with the inactive X chromosome. Since the full length
RNA-protein complexes can be targeted to genes transcribed of the inactive X becomes coated by Xist RNA-protei n com-
from other chromosomes by base-pairing with nascent RNAs p lexes (Figu re 7-49b), t hese complexes must spread along
as they are being transcribed. the chromosome from the X-inactivation center where Xist
is transcribed. The inactive X chromosome is also associated
X-Chromosome Inactivation in Mammals The phenomenon with Polycomb PRC2 complexes that catalyze the trimethyl-
of X-chromosome inactivation in fema le mammals is one of ation of histone H3 lysine 27. This results in association of
the most intensely studied examples of epigenetic repression the PRC l complex and transc riptional repression as dis-
mediated by a long, non-protein-coding RNA. X inactivation cussed above.
is controlled by an = 100-kb domain o n the X chromosome In the earl y female embryo comprised of embryonic stem
called the X-inactivation center. Remarkably, the X-inactivation cells capable of differentiating into all cell t ypes (see Chapter
center does nor express proteins, but rather several noncod- 21 ), genes on both X chromosomes are transcribed and the
ing RNAs (ncRNAs) that participate in the random inactiva- 40-kb Tsix ncRNA is transcribed from the X-inactivation
tion of one entire X chromosome early in the development of center of both copies of the X chromosome. [xperimenl~
female mammals. The ncRNAs whose functions are partially employing engineered deletions in the X-inactivation center
understood are transcribed from the complementary D A have shown that Tsix transcription prevents significant tran-
strands near the middle of the X-inactivation center: the scription of the 17-kb Xist RNA from the complementary
40-kb Tsix RNA, the Xist RNA which is spliced into an RNA DNA strand. Later in development of the early emb ryo, as
of = 17 kb, and the shorter 1.6-kb RepA RNA from the 5' re- cells begin to differentiate, Tsix becomes transcribed only
giOn of the Xist RNA (Figure 7-49a). from the active X chromosome. The mechanism(s) controlling

332 CHAPTER 7 • Transcriptional Control of Gene Expression

this asymmetric transcription of Tsix are not yet under- conserved between mammals. This conservation of sequence
stood. However, the process occurs randomly on the two X strongly suggests that these noncoding RNAs have important
chromosomes. functions. The examples of Xist, HOTAIR, and two other
In a current model of X inactivation, inhibition of Tsix recently discovered ncRNAs that target Polycomb repression
transcription allows transcription of RcpA RNA from the mechanisms to specific genes raise the possibility that many
complementary DNA strand (figure 7-49a). RepA RNA has of these may abo target Polycomb repression. Consequently,
a repeating sequence that forms stem-loop secondary struc- the study of these conserved noncoding RNAs is another area
tures that are bound directly by subunits of the Polycomh of intense current investigation.
PRC2 complex. This interaction occur~ on nascent RepA
transcripts that are tethered to the X chromosome during
transcription and leads to methylation of histone H3 at ly-
Plants and Fission Yeast Use Short RNA-Directed
sine 27 in the surrounding chromatin. By mechanisms that
are not yet understood, this activates transcription from the Methylation of Hi stones and DNA
nearby Xist promoter. The transcribed Xist RNA contains Centromercs (hgure 6-45c) of the fission yeast Schizosac
RNA sequences that by unknown mechanisms cause it to charomyces pombe are composed of multiple sequence re-
spread along the X chromosome. The RepA repeated se- peats as they arc in multicellular organisms. Proper functioning
quence near the 5' end of the Xist RNA binds the PRC2 of these centromeres during chromosome segregation in mi-
polycomb complex leading to H3K27 di- and trimethylation tosis and meiosis (Figures 18-36, 5- l Oa, 19-38) requires cen-
along the entire length of the X chromosome. This in turn tromeres to form heterochromatin. Heterochromatin
results in binding of the PRCl polycomb complex and tran- formation at S.fwmhe cenrromcres is directed by short inter-
scriptional repression as discussed earlier. At the same time, fering RNAs (siRNAs), initially discovered in C. e/egans for
continued transcription of Tsix from the other, active X their function in the cytoplasm, where they direct degrada-
chromosome continues, represses Xist transcription from that tion of mRNAs to which they hybridize (Figure 5-45 and
X chromosome, and consequently prevents Xist-mediated re- discussed further in Chapter 8). RNA polymerase 11 tran-
pression of the active X. A short time later in development, the scribes low levels of noncoding transcripts from the centro-
DNA of the inactive X also becomes methylated at most of its meric repeats (cenRNA, Figure 7-50). This ·is converted into
associated CpG island promoters, probably contributing to its double-stranded RNA by an RNA-dependent RNA poly-
stable inactivation through the multiple cell divisions that merase found in plants and many fungi (but not in the bud-
occur later during embryogenesis and throughout adult life. ding yeastS. cerevisiae, where the siRNA system docs not
occur, and not in mammals, where this mechanism of tran-
Trans Repression by Long Noncoding RNAs Another exam- scriptional repression may not occur). The resulting long
ple of transcriptional repression by a long noncoding RNA double-stranded RNAs are cleaved by a double-strand RNA
was discovered recently by researchers studying the function specific ribonuclease called Dicer into 22-nuclcotide fragments
of noncoding RNAs transcribed from a region encoding a with two-nucleotide 3' overhangs. One strand of these Dicer
cluster of HOX genes, the HOXC locus, in cultured human fragments is bound by a member of a protein family called
fibroblasts. Depletion of a 2.2-kb noncoding RNA expressed Argonaut proteins that associate with siRNAs in both trans-
from the HOXC locus by siRNA (Figure 5-45) unexpectedly lational and transcriptional repression mechanisms. The S.
led to de-repression of the HOXD locus in these cells, an pombe Argonaut protein, Agol, associates with two other
= 40-kb region on another chromosome encoding several proteins to form the RITS complex (for .RNA-induced .tran-
HOX proteins and multiple other noncoding RNAs. Assays scriptional ~ilencing).
similar to chromatin immunoprecipitation showed that this The RITS complex associates with centromeric regions
noncoding RNA, named HOTAIR for HOX Antisense Inter- by base-pairing between the si RNA associated with its Ago 1
genic .R~A, associates wi.ch the HOXD loci and with Poly- subunit and nascent transcripts from the region and interac-
comb PRC2 complexes. This results in histone H3 lysine 27 tions of its Chpl (chromodomain protein I) subunit which
di- and trimethylation, PRCl association, histone H3 lysine contains a methyl lysine-binding chromodomain specific for
4 demethylation, and transcriptional repression. This is sim- binding histone H3 di- and trimethyllysine 9 associated with
ilar to the recruitment of Polycomb complexes by Xist RNA heterochromatin. The RITS complex also associates with an
except that Xist RNA functions in cis, remaining in associa- RNA-dependent RNA polymerase-containing complex,
tion with the chromosome from which it is transcribed, RDRC. Since multiple siRNAs are generated from the double-
whereas HOTAIR leads to Polycomb repression h1 trans on stranded RNA, this results in a positive feedback loop that
both copies of another chromosome. increases the association of RITS complexes with centro
Recently, characterization of DNA associated with the meric heterochromatin. The RITS complex also associates
histone H3 lysine 4 trimethylation mark associated with pro- with a histone H3 lysine 9 methyl transferase. The resulting
moter regions and H3 lysine 36 methylation associated with histone H3 lysine 9 methyl marks on the centromeric chro-
Pol li transcriptional elongation led to the discovery of matin arc binding sites for S. pombe HPl proteins and a
= 1600 long noncoding RNA~ transcribed from inrergenic histone deacetylase (HDAC), leading to the condensation of
regions berween protein-coding genes that are evolutionarily the centromere region into heterochromatin.

7.7 Epigenetic Regulation of Transcription 333

 II

H3K9 methylation

FIGURE 7· 50 Model for the generation of heterochromatin at nucleotides with two base overhangs at the 3 ' end ,of each strand.
S. pombe centromeres by noncoding RNAs. Step (0 ): Poll I tran- Step (lit ): One of the two = 22 nucleotide strands generated is bound
scripts of the repeated nonprotein-coding sequences of the centro- by the Agol subunit of a RITS complex. Since multiple siRNAs
mere are transcribed at a low level. Step (8 ): The nascent RNA is bound associated with RITS complexes are generated from each Pol II
by the RITS complex by base-pairing of the complementary short transcript, this results in a positive-feedback loop that concentrates
interfering RNA (siRNA) associated with the Agol subunit of the RITS RITS complexes at the centromere region. Step (f'a ): The RITS complex
complex and the interaction of the Chpl subunit with histone H3 also associates with a histone H3 lysine 9 methyl transferase (H3K9
methylated on lysine 9. Step (II): The RITS complex associates with the HMn, which methylates histone H3 in the centromeric region. This
RDRC complex, which includes an RNA-dependent RNA polymerase generates a binding site for S. pombe HPl proteins, as well as the Chpl
that converts the nascent Pol II transcript into double-stranded RNA. subunit of the RITS complex. Binding of HPl condenses the region into
Step (rJ): The double-stranded RNA is cleaved by the Dicer double- heterochromatin as diagrammed in Figure 6-35a. [Adapted from
strand-specific ribonuclease into double-stranded fragments of = 22 D. Moazed, 2009, Nature 457:413.]

rw..J 5-Methyl C Induction by ncRNAs in Plants The model The f WA gene encodes a homeodomain transcription
- plant Arabidopsis thaliana uses DNA methylation ex- factor involved in regulation of the flowering time in response
tensively to repress transcription of transposons and ret- to temperature, so that plants do not flower until the warm
rotransposons (discussed in Chapter 6) and certain specific days of spring. In wild-type A. thaliana, FWA is repressed by
genes. In addition to methylating C at the 5 position in the CHH methylation of its promoter region. Failure to methyl-
sequence CG, plants also methylate genes at CHG (where H ate the FWA promoter resu lts in an easily recognized late-
is any of the other nucleotides) and CHH. There is a degree of flowering phenotype, allowing the isolation of A. thaliana
redundancy, but the DNA methyl transferase METl largely mutants in multiple genes that fail to methylate CHH se-
carries out CpG methylation and is functionally similar to quences. These genes have been cloned by methods described
DNMTl in multicellular animals. CMT3 (chromomethylase in Chapter 5, revealing a complex mechanism of RNA-di -
3) methylates CHG, and DRM2 is the primary methyl trans- rected DNA methylation that involves the plant-specific RNA
ferase of CHH. Methylation of CpG and CHG sequences are polymerases IV and V mentioned earlier (Figure 7-51) and
maintained following DNA replication by METl and CMT3, plant-specific nuclear siRNAs that are 24 nucleotides long.
respectively, by recognition of the methyl C in the parental The FWA gene has a direct duplication in its promoter
strand of newly replicated DNA and methylation of the region, and multiple copies of transposons are present in
daughter strand C, as discus<;ed ::~bove for human DNMTl. plant genomes. By a mechanism yet to be elucidated, Pol IV
However, one of the daughter chromosomes of a CHH meth- is directed to transcribe repeated DNA no matter what its
ylation site has an unmodified Gat the position complemen- sequence. An RNA-dependent RNA polymerase (RDR2)
tary to the methylated C, and hence carries no DNA converts the single-stranded Pol IV transcript into double-
modification that can be recognized by the DRM2 methyl stranded RNA, which is cleaved by Dicer ribonucleases, es-
transferase. Consequently, CHH methylation sites must be pecially DCL3, into 24 nucleotide double-stranded fragments
maintained through cell division by an alternative mechanism. with two base overhangs. One strand of these RNA fragments

334 CHAPTER 7 • Transcriptional Control of Gene Expression

.'·
Histone H3 lysine 9 di- and trimethylation creates binding
sites for the heterochromatin-associated protein HPl, which
results in the condensation of chromatin and transcriptional
repression. These post-translational modifications are perpet-
uated following chromosome rephcation because the methyl-
ated histones are randomly associated with the daughter DNA
molecules and associate with histone H3 lysine 9 methyl
transferases that methylate histone 3 ly<:ine 9 on newl y synthe-
sized histone H3 assembled on the daughter DNA.

• Polycomb complexes maintain repression of genes initially

repressed by sequence-specific binding transcription factor
FIGURE 7-51 Model of the mechanism of DNA methylation at Cs repressors expressed early during embryogenesis. One class
in CHH sequences in A. thaliana. The plant-specific RNA polymerase
of Polycomb repression complexes, PRC2 complexes, is
IV transcribes repeated sequences such as transposons and the
.· thought to associate with these repressors in early embryonic
promoter-proximal region of the FWA gene (blue DNA, with the
cells, resulting in methylation of histone H3 lysine 27. This
duplicated region indicated by blue arrows). The RNA-dependent
RNA polymerase RDR2 converts this to double-stranded RNA, which is
creates binding Sites for subunits m the PRC2 complex and
cleaved by the Dicer enzyme DCL3 into 24-nucleotide double-stranded
PRCl-type complexes that inhibit the assembly of Pol II ini-
RNA fragments with two base overhangs. One strand is bound by the tiation complexes or inhibit transcription elongation. Since
Argonaut protein AG04 or AG06 and base-pairs with transcripts of parental histone octamers with H3 methylated at lysine 27
repeated DNA transcribed by the plant-specific RNA polymerase V. This are distributed to both daughter DNA molecules following
leads to methylation of Cs (M) by the DNA methyl transferase DRM2. DNA replication, PRC2 complexes that associate with these
Several other proteins that participate in this elaborate process are nucleosomes maintain histone H3 lysine 2 7 methylation
represented by colored circles. They were identified because mutations through cell division.
in them produce a late-flowering phenotype and they fail to methylate
Cs in the FWA promoter region. [Adapted from M. V. C. Greenberg et al., • Trithorax complexes oppose repression by Polycomb com-
2011 , Epigenetics 6:344.] plexes by methylating H3 at lysine 4 and maintaining this
activating mark through chromosome replication.

• X-chromosome inactivation in female mammals requires

is bound by an Argonaut protein (AG04 or 6 ) in dense bod- a long noncoding RNA (ncRNA) called Xist that is tran-
ies in the nucleus called Cajal bodies, after the Spanish biolo- scribed from the X-inactivation center and then spreads by a
gist who first described them early in the twentieth century. poorly understood mechanism along the length of the same
The 24-nucleotide single-stranded RNA in these Argonaut chromosome. Xist is bound by PRC2 complexes at an early
complexes then base-pairs with a nascent transcript of re- stage of embryogenesis, initiating X inactivation that is
petitive DNA synthesized by Pol V. This directs the DRM2 maintained throughout the remainder of embryogenesis and
DNA-methyl transferase to methylate Cs in the repeated adult life.
DNA. As in metazoans, a histone deacetylase interacts with
the methylated Cs, leading to hypoacetylation of nucleo- • Long ncRNAs also have been discovered that lead to re-
somes associated with repeated DNA and repression of tran- pression of genes in trans, as opposed to the cis inactivation
scription by Pol II. • imposed by Xist. Repression is initiated by their interaction
with PRC2 complexes. Much remains to be learned about
how they are targeted to specific chromosomal regions, but
the discovery of =1600 long ncRNAs conserved between
mammals raises the possibility that this is a widely utilized
KEY CONCEPTS of Section 7.7
mechanism of repression.
Epigenetic Regulation of Transcription
• In many fungi and plants, RNA-dependent RNA polymer-
• The term epigenetic control of transcription refers to re- ases generate double-stranded RNAs from nascent tran-
pression or activation that is maintained after cells replicate scripts of repeated sequences. These double-stranded RNAs
as the result of DNA methylation and/or post-translation are processed by Dicer ribonucleases into 22- or 24-nucleotide
modification of histones, especially histone methylation. siRNAs bound by Argonaut proteins. The siRNAs base-pair
• Methylation of CpG sequences in CpG island promoters with nascent transcripts from the repeated DNA sequences,
in mammals generates binding sites for a family of methyl- inducing histone H3 lysine 9 methylation at centromeric re-
binding proteins (MBTs) that associate with histone deacety- peats in the fission yeastS. pombe, and DNA methylation
lases, inducing hypoacetylation of the promoter regions and in plants, resulting in the formation of transcriptionally re-
transcriptional repression. pressed heterochromatin.

7.7 Epigenetic Regulation of Transcription 335

7.8 Other Eukaryotic Transcription Systems transcription 10-fold. In humans, assembly of the Pol I pre-
initiation complex (Figure 7-52) is initiated by the coopera-
We conclude this chapter with a brief discussion of transcrip- tive binding of UBF (upstream binding factor) and SL l
tion initiation by the other two eu karyotic nuclear R NA poly- (selectivity factor), a multisubunit factor containing TBP and
merases, Pol I and Pol Ill, and by the distinct polymerases that four Poll-specific TBP-associated factors (TAF1s) to the Pol I
transcribe mitochondrial and chloroplast DNA. Al though promoter region. The TAF1 subunits interact directly with
these systems, particularly their regulation, are less thoroughly Pol !-specific subunits, directing this specific nuclear RNA
understood than transcription by RNA polymerase II, they polymerase to the transcription start sire. TIF-LA, the mam-
;1re equally as fundamental to the life of eubryotic cells. malian homolog of 5. cereuisiae RRNJ, is another required
factor, as well as the abundant nuclear protein kinase CK2
Transcription Initiation by Poll and Pol Ill (casein kinase 2), nuclear actin, nuclear myosin, the protein
dcacetylasc SIR T7, and topoisomerase I, which prevents
Is Analogous to That by Pol II DNA supercoils (Figure 4-8) from forming during rapid Pol
The formation of transcription-initiation complexes involving I transcription of the = 14-kb transcription unit.
Pol I and Pol !II is similar in some respects to assembly of Polll Transcription of the = 14-kb precursor of 185, 5.85, and
initiation complexes (see Figure 7- 17). However, each of the 285 rRNAs (sec Chapter 8) is highly regulated to coordinate
three eukaryotic nuclear RNA polymcrases requires its own ribosome synthesis with cell growth and division. This is
polymerase-specific general transcr iption factors and recognizes achieved through regulation of the activities of the Pol I ini-
different DNA control clements. Moreover, neither Pol I nor tiation factors by post-translational modifications including
Pol III requires ATP hydrolysis by a DNA hclicase to help melt phosphorylation and acetylation at specific sites, control of
the DNA template strands to initiate transcription, whereas Pol the rate of Poll elongation, and control of the number of the
II docs. Transcription initiation by Poll, which synthesizes pre- =300 human rRNA genes that are transcriptionally active
rRNA, and by Pollli, which synthesizes tRNAs, 55 rRNA, and by epigenetic mechanisms that assemble inactive copies into
other short, stable RNAs (see Table 7-2), is tightly coupled to heterochromatin. Switching between the active and hetero-
the rate of cell growth and proliferation. chromatic silent copies of rRNA genes is accomplished by a
multisubunit chromatin-remodeling complex ca lled NoRC
Initiation by Poll The regulatory elements directing Poll ini- ("No" for nucleolus, the site of rRNA transcription within
tiation arc similarly located relative to the transcription start nuclei ). NoRC localizes a nucleosome over the Pol I tran-
site in both yeast and mammals. A core element spanning the scription start site, blocking preinitiation complex assembly.
transcription start site from -40 to +5 is essential for Poll It also interacts with a DNA methyl transferase that methyl-
transcription. An additional upstream control element ex- ares a critical CpG in the upstream control element, inhibit-
tending from roughly -155 to -60 stimulates in vitro Pol I ing binding by UBF, as well as histone methyltransferases

I
I
I
I
I
I
I
I
.·

FIGURE 7-52 Transcription of the rRNA precursor RNA by RNA transcription of the noncoding pRNA required for transcriptional
polymerase 1. Top: electron micrograph of RNA protein complexes silencing. Regions of DNA shown in blue are contained in the primary
transcribed from repeated rRNA genes. One Poll transcription unit is transcript, but are removed and degraded during rRNA processing. The
diagrammed in the middle. Enhancers that stimulate Poll transcription core promoter element and upstream control element are diagrammed
from a single transcription start site are represented by blue boxes. Poll below with the location of Poll and its general transcription factors UBF,
transcription termination sites (To. T 1- T10) bound by the Pol l-specific SL 1, and TIF- 1A represented, as well as other proteins required for Poll
termination factor TIF-1 are shown as red rectangles. pRNA indicates elongation and control. [Adapted from 1. Grummt, 201 o, FEBS J. 277:4626.]

336 CHAPTER 7 • Transcriptional Control of Gene Expression

that di- and trimethylate histone H3 Irsine 9, creating bind- The N-terminal half of one TFIIIB subunit, called BRF
ing sites for heterochromatic HPl, and histone deacetylases. (for THIB-related factor), is similar in sequence to TfiiB (a
Moreover, an =250-nucleotide noncoding RNA ~ai l ed Pol II factor). This similarity suggests that BRF and TFIIB
pRNA (for promoter associated) transcribed by Pol l from perform a similar functio n in initiation, namely, to assist in
""2 kb upstream of the rRNA transcription unit (red arrow in separating the template DNA strands at the transcription
Figure 7-52) is bound by a subunit of NoRC and is required start site (figure 7-19 ). Once TFIIIB has bound to either a
for transcriptional silencing. pRNA is believed to target tRNA or 55-rRl'\A gene, Pol lii can bind and initiate tran-
NoRC to Pol I promoter regions by forming an RNA:DNA scription in the presence of ribonucleoside triphosphates.
triplex with the T 0 terminator ~rquence. This creates a bind- The BRF subunit of TFIIIB interacts specifically with one of
ing site for the DNA methyl transferase DNMT3b that meth- the polymerase subunits unique to Pol Ill, accounting for
ylates the critical CpG in the upstream promoter element. initiation by this specific nuclear RNA polymerase.
Another of the three subunits composing TFIIIB is TBP,
Initiation by Pol Ill Unlike protein-coding genes and pre- which we can now see is a component of a general transcrip-
rRNA genes, the promoter regions of tRNA and 55-rRNA tion fnctor for all three eukaryotic nuclear RNA polymer-
genes lie entirely within the transcribed sequence (Figure ases. The finding that TBP participates in transcription
7-53a, b). Two such internal promoter elements, termed the initiation by Pol I and Pol III was surprising, since the pro-
A box and the B box, are present in all tRNA genes. These moters recognized by these enzymes often do not contain
highly conserved sequences not only function as promoters TATA boxes. Nonetheless, in the case of Pol III transcrip-
but also encode two invariant portions of eukaryotic tRNAs tion, the TBP subunit of TFIIIB interacts with DNA similarly
that are required for protein synthesis. In 55-rRNA genes, a to the way it interacts with TATA boxes.
single internal control region, the C box, acts as a promoter. Pol III also transcribes genes for small, stable RNAs with
Th ree general transcription factors are required for Pol upstream promoters containing a T ATA box. One example
III to initiate transcription of tRNA and 55-rRNA genes in is the U6 snRNA involved in pre-mRNA splicing, as dis-
vitro. Two multimeric factors, TFIIIC and TFIII B, partici- cussed in Chapter 8. In mammals, this gene contains an up-
pate in initiation at both tRKA and 55-rRNA promoters; a stream promoter element called the PSE in addition to the
third factor, TFIUA, is required for initiation at 55-rRNA TATA box (Figure 7-53c), which is bound by a multisubunit
promoters. As with assembly of Pol I and Pol II initiation complex called SNAP,, while the TATA box is bound by the
complexes, the Pol Ill general transcription factors bind to TBP subunit of a specialized form of TFIIIB containing an
promoter DNA in a defined sequence. alternative BRF subunit.
MAF1 is a specific inhibitor of Pol III transcription that
functions by interacting with the BRF subunit of TFHIB nnd
(a) Pol Ill. Its function is regulated by controlling its import from
the cytoplasm into nuclei by phosphorylations at specific sites
in response to signal transduction protein kinase cascades that
respond to cell stress and nutrient deprivation (see Chapters
tRNA gene ·
16 and 24). In mammals, Pol III transcription is also repressed
by the critical tumor suppressors p53 and the retinblasroma
(RB) family. In humans there are two genes encoding subunit
Pol Ill RPC32. One of these is expressed specifically in replicating
cells, and its forced expression can contribute to oncogenic
transformation of cultured human fibroblasts.

FIGURE 7-53 Transcription-control elements in genes tran-

scribed by RNA polymerase Ill. Both tRNA {a) and 55-rRNA {b) genes
contain internal promoter elements {yellow) located downstream from
the start site and named A, B, and C boxes, as indicated. Assembly of
Po l Ill transcription initiation complexes on these genes begins with the
binding of Pol Ill-specific general transcription factors TFIIIA, TFIIIB, and
TFIIIC to these control elements. Green arrows indicate strong, sequence-
specific protein-DNA interactions. Blue arrows indicate interactions
between general transcription factors. Purple arrows indicate interac-
tions between general transcription factors and Pol Ill. {c) Transcription of
the U6 snRNA gene in mammals is controlled by an upstream promoter
with a TATA box bound by the TBP subunit of a specialized form ofTFIIIB
with an alternative BRF subunit and an upstream regulatory element
called the PSE bound by a multisubunit factor called SNAP,. [From
L. Schramm and N. Hernandez, 2002, Genes Dev. 16:2593.]

7.8 Other Eukaryot1c Transcription Systems 337

Mitochondrial and Chloroplast DNAs Are regulatory transcription initiation factors haYe been trans-
Transcribed by Organelle-Specific ferred to the nucleus, where th e control of thei r t ranscription
by nuclear RNA polymerase II likely indirectly controls the
RNA Polymerases
expression of sets of chloroplast genes. The bacterial-like
As discussed in Cha pter 6, mitochondria and chloroplasts chloroplast RNA polymerase is ca lled th e plastid polymerase
probably evolved from eubactcria that were endocytoscd because its catalytic co re is encoded b y th e chl oro plast
into ancestral cells containing a eukaryotic nucleus. In mod- genome. Most chloroplast genes are transcribed by these en-
ern-cia} cukaryotes, both organelles contain distinct D:t\As zymes and have - 3S and - 10 control regions similar to pro-
that encode some of the proteins essentia l to their specific moters in cyanobacteria, from which they evolved. The
functions. Interestingly, the R NA polymerases th at tran- chloroplast T7-like RNA polymerase is a lso encoded in the
scribe mitochondrial (mt) DNA and chl oroplast DNA are nuclear genome of higher plants. It transcribes a different set
similar to polymcrases from eubactcria and bacteriophages, of chlo roplast ge nes . Cu ri ously, this includes genes encoding
reflecting their evolutio nary origins. s u bunits of the bacterial-like multisubunit plastid poly-
merase. Recent results indicate that transcription by the mul-
Mitochondrial Transcription The RNA polymerase that tran- tisubunit polymerase is regulated by sigma fa ctors whose
scribes mtDNA is encoded in nuclear DNA. After synthesis of activities arc regulated by light and metabolic stress.
the enzyme in the cytosol, it is imported into the mitochon-
drial matrix by mechanisms described in Chapter 13. The mi-
tochondrial RNA polymcrases from S. cereuisiae and the frog KEY CONCEPTS of Section 7.8
Xenopus laeuis both consist of a large subunit with ribonucle-
otide-polymerizing activity and a small B subunit (TFBM). In Other Eukaryotic Transcription Systems
mammals, another matrix protein, mitochondrial transcrip- • The process of transcription initiation by Pol I and Pol Ill
tion facto r A (TFAM), binds to mtDNA promoters and is es- is similar to that by Pol II but requires different general tran-
sential for initiating transcription at the start sites used in the scription factors, is directed by different pr~moter clements,
cell. The large subu nit of yeast mitochondrial RN A poly- and does not require hydrolysis of ATP 13--y phosphodiesrer
merase clearly is related to the monomeric RNA polymerases bonds to separate the DNA strands at the starr sire.
of bacteriophage T7 and similar bacteriophages. However,
• Mitochondrial DNA is transcribed by a nuclear-encoded
the mitochondrial enzyme is functionally distinct from the
RNA polymerase composed of two subu nits. One ~ubunit is
bacteriophage enzyme in irs dependence on two other poly-
homologous to the monomeric RNA polymerase from bac-
peptide~ for transcription from the proper starr si tes.
teriopha ge T?; the other resembles bacteria l a factors.
The promoter sequences recognized by mitochondrial
RNA polymerases incl ude the transcription start site. These • Ch loroplast DNA is transcribed by a chloroplast-encoded
promoter sequences, w h1ch arc rich in A residues, ha ve been RNA polymerase homologous ro bacterial RNA polymer-
characterized in the mtDNA from yeast, plants, and an imals. ases, with several alternative nuclear encoded a-factors, and
The circular hun~an mitochondrial genome contains two related a single subunit bacteriophage T7-like RNA polymerase.
15-bp promoter sequences, one for the transcription of each
strand. Each strand is transcribed in its enti rety; the long
primary transcripts are then processed by cleavage at tRNA
Perspectives for the Future
genes that separate each of the mitochondrial mRNAs and
rRNAs. A second promoter appears to be responsible for A great deal has been learned in recent years about transcrip-
transcnbing additional cop ies of the rRNAs. Currently, tion control in eukaryores. Genes encoding =2000 activa-
there is rclati\ely little understanding of how transcription tors and repressors can be recognized in the human genome.
of the mitochondrial genome is regulated to coordinate the We now ha ve a glim pse of how t he astronomical number of
production of the few mitochondrial proteins it encodes possible combinations of these transcription factors can gen-
with synthesis and import of the thousands of nuclear DNA- erate the complexity of gene control required to produce or-
.·.
encoded proteins that comprise the mitochondria. ganisms as remarkable as those we see around us. But very
much remains ro be understood. Altho ugh we now have
Chloroplast Transcription Chloroplast DNA is transcribed some understand ing of what processes turn a gene on and
by two types of RNA polymerases, one multisubunit protein off, we have very little understanding of how the frequency
similar to bacterial RNA polymerases and one similar to the of transcription is controlled in order to provide a cell with
single subunit enzymes of bacteriophages and mitochondria. the appropriate amounts of its various proteins. In a red
The core subunits of the bacterial-type enzyme, a,~' W, and blood cell precursor, for example, the globin genes are tran-
w subunits, are encoded in the ch loroplast DNAs of higher scribed at a far greater rate than the genes encoding the en-
plants, whereas six a -0 -like a factors arc encoded in the n u- zymes of intermediary metabolism (the so-called housekeeping
clear DNA of higher plants. This is another example of the genes). How are the vast differences in the frequency of tran-
transfer of genes from organellar gcnomes to nuclear ge- scription initi ation at various genes ach ieved? What happens
nomes during e\·olution. In this case, genes encoding the to the mu ltiple interactions between activation domains,

338 CHAPTER 7 • Transcriptional Control of Gene Expression

 '
co-activator complexes, general transcription factors, and geted to specific genes? Do the = 1600 long noncoding RNAs
RNA polymerase II when the polymerase initiates transcrip- that are conserved between mammals all function to regulate
tion and transcribes away from the promoter region? Do transcription of specific target genes, adding to the complexity
these completely dissociate at promoters that are transcribed of transcription control by sequence-specific DNA-binding
infrequently, so that the combination of multiple factors re- proteins? Research to address these questions will he an
quired for transcription must be reassembled anew for each exciting area of investigation in the coming years.
round of transcription? Do complexes of activators with A thorough understanding of normal development and of
their multiple interacting co-activators remain assembled at abnormal processes associated with disease will require an-
promoters from which reinitiation takes place at a high rate, swers to these and many related questions. As further under-
so that the entire assembly does not have to be reconstructed standing of the principles of transcription control arc
each time a polymerase initiates? discovered, applications of the knowledge will likely be made.
,· Much remains to be learned about the structure of chro- This understanding may allow fine control of the expression
matin and how that structure influences transcription. What of therapeutic genes introduced by gene therapy vectors as
additional components besides HPl and methylated hisrone they arc developed. Detailed understanding of the molecular
H3 lysine 9 are required to direct certain regions of chroma- Interactions that regulate transcription may provide new tar-
tin to form heterochromatin, where transcription is re- gets for the development of therapeutic drugs that inhibit or
pressed? Precisely how is the structure of chromatin changed stimulate the expression of specific genes. A more complete
by activators and repressors, and how docs this promote or understanding of the mechanisms of transcriptional control
inhibit transcription? Once chromatin-remodeling com- may allow improved engineering of crops with desirable char-
plexes and histone acetylase complexes become associated acteristics. Certainly, further advances in the area of transcrip-
with a promoter region, how do they remain associated? tion control will help to satisfy our desire to understand how
Current models suggest that certain subunits of these com- complex organisms such as ourselves develop and function.
plexes associate with modified histone tails so that the com-
bination of binding to a specific histone tail modification
plus modification of neighboring histone tails in the same Key Terms
way results in retention of the modifying complex at an acti-
vated promoter region. In some cases, this type of assembly activation domain 307 MAT locus (in yeast) 315
mechanism causes the complexes to spread along the length activators 281 mediator 315
of a chromatin fiber. What controls when such complexes anritermination factor 301 nuclear receptors 309
spread and how far they will spread? bromodomain 319 promoter 281
Single activation domains have been discovered to inter-
carboxyl-terminal domain promoter-proximal
act with several co-activator complexes. Are these interac-
(CTD) 293 elements 302
tions transient, so that the same activation domain can
interact with several co-activators sequentially? Is a specific chromatin-mediated repression domain 308
order of co-activator interaction required? How does the in- repression 315 repressors 281
teraction of ac.t ivation domains with mediator stimulate chromodomain 330 RNA polymerase II 290
transcription? Do these interactions simply stimulate the as- co-activator 311 silencer sequences 316
sembly of a preinitiation complex, or do they also influence co-repressor 312 specific transcription
the rate at which RNA polymerase II initiates transcription factors 305
DNase I footprinting 305
from an assembled preinitiation complex?
enhanceosome 314 TATA box 295
Transcriptional activation is a highly cooperative process
so that genes expressed in a specific type of cell are expressed enhancers 285 TATA box-binding protein
only when the complete'set of activators that control that general transcription (TBP) 298
gene are expressed and activated. As mentioned earlier, some factors 297 upstream activating
of the transcription factors that control expression of the heat-shock genes 325 sequences (UASs) 305
TTR gene in the liver are also expressed in intestinal and histone deacetylation 318 yeast two-hybrid system 321
kidney cells. Yet the ITR gene is not expressed in these other zinc finger 309
leucine zipper 311
tissues, since its transcription requires two additional tran-
scription factors expressed only in the liver. What mecha-
nisms account for this highly cooperative action of Review the Concepts
transcription factors that is critical to cell-type-<;pecific gene
expression? 1. Describe the molecular events that occur at the lac op-
The discovery that long noncoding RNAs can repress tran- eron when E. coli cells are shifted from a glucose-containing
scription of specific target genes has stimulated tremendous medium to a lactose-containing medium.
interest. Do these always repress transcription by targeting 2. The concentration of free glutamine affects transcription
Polycomb complexes? Can long noncoding RNAs also acti- of the enz} me glutamine synthetase in E. coli. D escribe the
vate transcription of specific target genes? How are they tar- mechanism for this.

Review the Concepts 339

3. What types of genes are transcribed by RNA polymcrases sequences. What arc the comparable sequences found in
I, II, and III? Design an experiment to determine whether a higher eukaryotic species?
specific gene is transcribed by RNA polymerase II. 17. Recall that the Trp repressor binds to a site in the opera-
4. The CTD of the largest subunit of RNA polymerase II tor region of tryptophan-producing genes when tryptophan
can be phosphorylated at multiple serine residues. What are is abundant, thereby preventing transcription. What would
the conditions that lead to the phosphorylated versus un- happen to the expression of the tryptophan biosynthetic en-
phosphorylated RNA polymerase II CTD? zyme genes in the following scenarios? Fill in the blanks with
5. What do TATA boxes, initiators, and CpG islands have in one of the following phrases:
common? Which was the first of these to be identified? Why?
never be expressed/always (constitutively) be expressed
6. Describe the methods used to identify the location of DNA-
control elements in promoter-proximal regions of genes. a. The cell produces a mutant Trp repressor that cannot
7. What i~ the difference between a promoter-proximal ele- bind to the operator. The enzyme genes will _ _ _ _ _ __
ment and a distal enhancer? What are the similarities? b. The cell produces a mutant Trp repressor that binds
8. Describe the methods used to identify the location of to its operator site even if no Tryptophan in present. The
DNA-binding proteins in the regulatory regions of genes. enzyme genes will _ _ _ _ _ __
9. Describe the structural features of transcriptional activa- c. The cell produces a mutant,sigma factor that cannot
tor and repressor proteins. bind the promoter region. The enzyme genes will _ _ _ __
10. Give two examples of how gene expression may be re- d. Elongation of the leader sequence is always stalled after
transcription of region 1. The enzyme genes will _ _ _ _ __
pressed without altering the gene-coding sequence.
18. Compare/contrast bacterial and eukaryotic gene expres-
11. Using CREB and nuclear receptors as examples, com-
sion mechanisms.
pare and contrast the structural changes that take place
when these transcription factors bind to their co-activators. 19. You are curious to identify the region of.gene X sequence
12. What general transcription factors associate with an that serves as an enhancer for gene expression. Design an
experiment to investigate this issue.
RNA polymerase II promoter in addition to the polymerase?
In what order do they bind in vitro? What structural change 20. Some organisms have mechanisms in place that will
occurs in the DNA when an "open" transcription-initiation override transcription termination. One such mechanism
complex is formed? using the Tat protein is employed by the HIV retrovirus. Ex-
13. Expression of recombinant proteins in yeast is an impor- plain why Tat is therefore a good target for HIV vaccination.
tant tool for biotechnology companies that produce new drugs 21. Upon identification of the DNA regulatory sequence re-
for human use. In an attempt to get a new gene X expressed in sponsible for translating a given gene, you note that it is en-
yeast, a researcher has integrated gene X into the yeast genome riched with CG sequences. Is the corresponding gene likely to
near a telomere. Will this strategy result in good expression of be a highly expressed transcript?
gene X? Why or 'why not? Would the outcome of this experi- 22. Name four major classes of DNA-binding proteins that
ment differ if the experiment had been performed in a yeast are responsible for controlling transcription, and describe
line containing mutations in the H3 or H4 histone tails? their structural features.
14. You have isolated a new protein called STICKY. You
can predict from comparisons with other known proteins
that STICKY contains a bHLH domain and a Sin3-interact- Analyze the Data
ing domain. Predict the function of STICKY and rationale In eukaryotes, the three RNA polymerases, Pol I, II, and III,
for the importance of these domains in STICKY function. each transcribes unique genes required for the synthesis of
15. The yeast two-hybrid method is a powerful molecular ribosomes: 25S and 18S rRNAs (Pol 1), 55 rRNA (Pol III),
genetic method to identify a protcin(s) that interacts with a and mRNAs for ribosomal proteins (Pol II). Researchers
known protein or protein domain. You have isolated the have long speculated that the activities of the three RNA
glucocorticoid receptor (GR) and have evidence that it is a polymerases are coordinately regulated according to the de-
modular protein containing an activation domain, a DNA- mand for ribosome synthesis: high in replicating cells in rich
binding domain, and a second ligand-binding activation do- nutrient conditions and low when nutrients are scarce. To
main. Further analysis reveals that in pituitary cells, the determine whether the activities of the three polymerases are
protein is anchored in the cytoplasm in the absence of its coordinated, Laferte and colleagues engineered a strain of
hormone ligand, a result leading you to speculate that it yeast to be partially reSIStant to the inhibition of cell growth
binds to other inhibitory proteins. Describe how a two-hy- by the drug rapamycin (2006, Genes Dev. 20:2030-2040).
brid analysis could be useJ to identify the protein(s) with As discussed in Chapter 8, rapamycin inhibits a protein kinase
which GR interacts. How would you specifically identify the (called TOR, for target of rapamycin) that regulates the
domain in the GR that binds the inhibitor(s)? overall rate of protein synthesis and ribosome synthesis.
16. Prokaryotcs and lower eukaryotes such as yeast have When TOR is inhibited by rapamycin, the transcription of
DNA-regulatory elements called upstream activating rRNAs by Pol I and Pol III and ribosomal protein mRNAs

340 CHAPTER 7 • Transcriptional Control of Gene Expression

by RNA polymerase II are all rapidly repressed. Part of the and CARA cells with 1 H uracil (for 20 minutes ) at various
inhibition of Pol I rRNA synthesi'> results from the dissocia- rimes after addition of rapamycin to the media. Total cellu-
tion of the Pol I transcription factor Rrn3 from Pol I. In the lar RNA was isolated and subjected to gel electrophoresis
strain constructed by Lafertc and colleagues, the wild-type and autoradiography. The lower autoradiogram shows the
Rnr3 gene and the wild-type A43 gene, encoding the Pol I region of the gel containing 5S rRNA. Based on these data,
subunit to which Rrn3 binds, were replaced with a gene en- what can be concluded about the influence of Pol I transcrip-
coding a fusion protein of the A43 Poll subunit with Rrn3. tion on the transcription of ribosomal protein genes by Pol II
The idea was that the covalent fusion of the two proteins and 5S rRNA by Pol Ill?
would prevent the Rrn3 dissociarion from Pol I otherwise c. To determine whether the difference in behavior of
caused by rapamycin treatment. The resulting CARA (con- wild-type and CARA cells can be observed under normal
stitutive association of Rrn3 and A43) strain was found to physiological conditions (i.e., without drug treatment), cells
be partially resistant to rapamycin. In the absence of rapa- were subjected to a shift in their food source, from nutrient-
mycin, the CARA strain grew at the same rate and had equal rich media to nutnent-poor media. Under these condition<>,
numbers of ribosomes as wild-type cells. in wild-type cells, the TOR protein kinase becomes inactive.
a. To analyze rRNA transcription by Pol I, total RNA Consequently, shifting cells from nutrient-rich media to nu-
was isolated from rapidly growing wild-type (WT) and CARA trient-poor media should result in a normal physiological
cells at various times following the addition of rapamycin. response that is equivalent to treating cells with rapamycm,
The concentration of the 355 rRNA precursor transcribed by which inhibits TOR. To determine how the CARA fusion
Pol I (sec Figure 8-38) was assayed by the primer-extension protein affected the response ro this media shift, RNA was
method. Since the 5' end of the 35S rRNA precursor is de- extracted from wild-type and CARA cells and used to probe
graded during the processing of 255 and 18S rRNA, this microarrays containing all yeast open reading frames. The
method measures the relatively short-lived pre-rRNA precur- extent of RNA hybridization with the arrays was quantified
sor. This is an indirect measure of the rate of rRNA transcrip- and is expressed in the graphs as log2 of the ratio of CARA-
tion by Pol I. The results of this primer extension assay are cell RNA concentration to wild-type-cell RNA concentra-
shown below. How does the CARA Pol I-Rrn3 fusion affect tion for each open reading frame. A value of zero indicates
the response of Poll transcription to rapamycin? that the two strains of yeast exhibit the same level of expre~­
sion for those specific RNAs. A value of 1 indicates that the

,__.
CARA cells contain twice as much of that particular RNA as
Minutes after WT CARA
do wild-type cells. The graphs below show the number of
rapamycin 0 20 40 60 80 100 0 20 40 60 80 100
open reading frames (y axis) that have values for log2 of this
35S rRNA ratio, indicated by the x axis. The results of hybndization to
open reading frames encoding mRNAs for ribosomal pro-
reins are shown by black bars, those for all other mRNAs by
b. The concentrations of four mRNAs encoding ribo- white bars. The graph on the left gives results for cells grown
somal proteins, RPL30, RPS6a, RPL7a, and RPL5, and the in nutrient-rich medium, the graph on the right for cells
mRNA for actin (ACTl ), a protein present in the cytoskele- shifted to nutrient-poor medium for 90 minutes. What do
ton, were assessed in wild-type and CARA cells by Northern these data suggest about the regulation of ribosomal protein
blotting at various times after addition of rapamycin to rap- gene transcription by Pol II?
idly growing cells (upper autoradiograms). 5S rRNA tran-
scription was assayed by pulse labeling rapidly growing WT Cells grown in rich media Cells grown in poor media

Minutes after WT_ CARA

rapamycin o 20 40 60 80 100 0 20 40 60 80 100

RPL30-
"'c::
Q)

Q)
RPS6a Cl

0
RPL7a Qi
.0
E
RPL5 ::J
z
ACT1

Minutes after WT CARA

rapamycin 0 20 40 60 80 100
5S , _ _ Expression ratio Expression ratio
(log 2 CARA/WT) (log 2 CARA /WT)

Analyze the Data 341

 Jun, S. H., et al. 2011. Archaeal RNA polymerase and transcnp-
References rion regulation. Crit. Reu. Biochem. Mol. Bioi. 46:27-40.
Kornberg, R. D. 2007. The molecular basis of eukaryotic
Control of Gene Expression in Bacteria
transcription. Proc. Nat! Acad. Sc1. USA. 104:12955-12961.
Campbell, E. A., I . f. Westblade, and S. A. Darst. 2008. ~hiller, F., A. Zaucker, and L. Tora. 2010. Developmental
Regulation of bactenal RNA polymerase ~igma factor activity: a regulation of transcription initiation: more than just changing the
structural perspective. Curr. Opin. Micro/no/. 11:121-127. actors. Curr. Opm. Genet. Dev. 20:533-540.
Ca~mo, P., V. Rubio, and A. .\1arina. 2010. The mechamsm of Papai, G., P. A. Wei!, and P. Schultz. 2011. New m~1ghts into
signal transduction by two-component systems. Curr. Opm. Struct. rhc function of transcription factor TFJTD from recent structural
Rio/. 20:763-771. studie~. Curr. Opin. Genet. Dev. 21:219-224.
Halford, S. E., and J. F. Marko. 2004. How do Site-specific Pnce, D. H. 2008. Poi~ed polymera~es: on your mark ... get set ...
D;\A-bindmg proteins find the1r targets? Nucl. ACids Res. go! Mol. Cel/30:7-!0.
32:3040-3052. Roeder, R. G. 1996. The role of general iniwHion factors in
Hs1eh, Y. J., and B. L. Wanner. 20 I 0. Global regulation by the transcription b> RNA polymerase II. Trends Biochem. Sci. 21:327-335.
seven-component Pi s1gnaling system. Curr. Opm. Mrcrobwl. Sandelin, A., er al. 200-. Mammalian Rl'\A polymerase II core
13:198-203. promoters: insights from genome-wide studies. Nat. Rev. Genet.
Lawson, C. L., et al. 2004. Catabolite activator protein: DNA 8:424-436.
bmding and transcnption activation. Curr. Oprn. Struct. Bioi. Seila, A. C., er al. 2009. Divergent transcription: a new feature
14:10-20. of awve promoters. Cell Cycle 8:2557!..2564.
Muller-Hill, B. 1998. Some repressors of bacterial transcription. Seith, L.A., S. Sigurdsson, and J. Q. Svejstrup. 2010. Transcript
Curr. Opm. Microbial. 1:145-151. elongation b) RNA polymerase II. Ann. Reu. Brochem. 79:271-293.
Murakami, K. S., and S. A. Darst. 2003. Bacterial Rl\A Sikorski, T. W., and S. Burarowsk1. 2009. The basal initiation
polymerases: the whole story. Curr. Opm. Struct. Bioi. 13:31-39. machinery: beyond the general transcription factors. Curr. Opm.
Sharma, U.K., and D. Chatterji. 2010. Transcriptional Cell Brol. 21:344-351.
switchmg in Eschenchia coli during stress and ~tarvation by Thomas, M. C., and C. M. Chiang. 2006. The general transcrip-
modulanon of sigma activ1ty. FEMS Mrcrobiol. Ret'. 34:646-657. tion machinery and general cofacrors. Crit. Rev. Brochem. Mol. Bioi.
Wigneshweraraj, S. R., et al. 2008. Modus operandi of the 41:105-178.
bacterial RNA polymerase containing the sigma54 promoter- Wade, J. T., and K. Srruhl. 2008. The transition from transcrip-
specificity factor. Mol. Mrcrobrol. 68:538-546. tional initiation to elongatton. Curr. Opm. Genet. Deu. 18:130-136.
Yamada, T., er al. 2006. P-TEFb-mediated phosphorylation of
Overview of Eukaryotic Gene Control and RNA Polymerases hSpt5 C-terminal repeats is critical for processive transcription
Brenner, S., et al. 2002. Conserved regulanon of the lympho- elongation. Mol. Cel/21:227-237.
cyte-specific expression of lck in the f-ugu and mammals. Proc. Nat/
Acad. Set. USA 99:2936-2941. Regulatory Sequences in Protein-Coding Genes and the Proteins
Cramer P., et al. 2008. Structure of eukaryotic Rl'\A polymer- Through Which They Function
ases. A1111. Rev. Brophys. 37:337-352. Fuxreiter, .VI., et al. 2008. Malleable machines rake shape in
Maston, G. A., S. K. Evans, and M. R. Green. 2006. Transcrip- eukaryoric transcriptional regulation. Nat. Chem. Bioi. 4:728-737.
tional regulatory elements in the human genome. Ann. Rev. Garvie, C. W., and C. Wolberger. 2001. Recognition of specific
Genomrcs Hum. Genet. 7:29-59. DNA sequences. Mol. Cell. 8:937-946.
Ptashne, M., and A. Gann. 2001. Transcription mitianon: Kadonaga, J. T. 2004. Regulation of R!';A polymerase II
1m posing specificity by localization. Essays Biochem. 3 7: 1-15. transcription by sequence-specific DNA binding factors. Cell
Strub!, K. 1999. Fundamentally different logic of gene regulation 116:247-25-.
m eukaryotes and prokaryotes. Cel/98:1-4. Kaufmann, K., A. Pajoro, and G. C. Angenent. 2010. Regula-
V1sel, A., E. M. Rubin, and L. A. Pennacchio. 2009. Genomic tion of transcription in plants: mechanisms controlling developmen-
views of distant-acting enhancers. Nature 461:199-205. tal switches. Nat. Rev. Genet. 11:830-842.
Wallace,]. A., and G. Felsenfeld. 2007. We gather together: Riechmann, J. L., et al. 2000. Arabrdopsrs transcription factors:
msularors and genome organization. Curr. Opm. Genet. Deu. genome-wide comparative analysis among eukaryotes. Science
17:400-407. 290:2105-2110.
Tupler, R., G. Perini, and M. R. Green. 200 L. Expressing rhe
RNA Polymerase II and the General Transcription Factors human genome. Nature 409:832-833.
Baumann, M., J. Pontiller, and W. Ernst. 2010. Structure and
basal transcription complex of RNA polymerase II core promoters m Molecular Mechanisms ofTranscription Repression and Activation
rhe mammahan genome: an overview. Mol. Bioteclmol. 45:241-247. Bannister, A. J., and T. Kouzaride~. 2011. Regulation of
Brueckner, F., J. Ortiz, and P. Cramer. 2009. A movie of the chromatin by histone modifications. Cell Res. 21:381-395.
RNA polymerase nucleotide addition cycle. Curr. Opin. Struct. Bioi. Bulger, M., and M. Groudine. 2011. Functional and mechanistic
19:294-299. diver•my of disml rr~nscripnon enhancers. Cell 144:327-339.
Fuda, N.J., ~1. B. Ardehah, and J. T. Lis. 2009. Defining Cairns, B. R. 2009. The logic of chromatin architecture and
mcchani~ms that regulate RNA polymerase II transcription in vivo. remodelling at promoters. Nature 461:193-198.
Nature 461:186-192. Conaway, R. C., and J. W. Conaway. 2011. Function and
Hahn, S. 2004. Structure and mechanism of the RNA polymerase regula non of the Ylediator complex. Curr. Oprn. Genet. Deu.
II rranswption machmery. Nat. Struct. Mol. Bzol. 11:394-403. 21:225-230.
Illingworth, R. S., and A. P. Bird. 2009. CpG islands-'a rough Courey, A. J., and S. Jia. 2001. Transcriptional repression: the
guide'. FFBS Lett. 583:1713-1720. long and the short of it. Genes Det'. 15:2-86-2796.

342 CHAPTER 7 • Transcriptional Control of Gene Expression

 Deaton, A. :..1., and A. Bird. 20 II. CpG I'> lands and the York, B., and B. W . O'Malley. 2010. Steroid receptor coactiva-
regulation of transcription. Genes Dev. 25:1010-1022. ror (SRC) family: masters of systems biology.}. Bwl. Chem.
Hargreaves, D. C., and G. R. Crahtree. 2011. ATP-dependenr 285:38743-3850.
chromatin remodeling: genetics, genomics and mechamsms. Cell
Res. 21:396-420. Epigenetic Regulation ofTranscription
Kornberg, R. D. 2005. Mediator and the mechanism of Beisel, C., and R. Paro. 2011. Silencing chromatin: comparing
rranscripnonal activation. Trends Biochem. Sci. 30:235-239. mode~ and mechanisms. Nat .Reu. Genet. 12:123-135.
L1, B., ;\1. Carey, and J. L. Workman. 2007. The role of Black, J. C., and J. R. Whetstine. 2011. Chromatin landscape:
chromatin during transcription. Cell 128: 7 07-719. methylation beyond tr<mscripnon. l:..prgenet1cs 6:9-1 S.
Malik, S., and R. G. Roeder. 2010. The metazoan Mediator Clouaire, T., and I. Stanchcva. 2008. Meth} 1-CpG bmding
co-acti\·aror complex as an integrative huh for tran'icriptional proteins: specialized transcnpt1onal repre~sors or structural
regulation. Nat. Rev. Genet. 11:761-772. components of chromatin? Cell Mol. Life Sci. 65:1509-1522.
:..Iohrmann, L., and C. P. YernJ7er. 200). Composmon and Grewal, S. !. 20 I 0. RNA1-dependem formation of heterochromatin
functional ~pcc1ficity of SWI2/SNF2 clas~ chromatin remodeling and its diverse funcnon~. Curr. Opm. Genet. Del'. 20:1 34-141.
complexe~. B1oclmn. B1uphys. Acta 1681:59-73. Lee, J. T. 20 I 0 . The X as model for RNA\ niche in epigenom1c
Perissi, V., et al. 2010. Deconstructing repression: evolving regulanon. Cold Sprmg Harbor Perspect. Bioi. 2 (9):a003749 ( 1-12).
model'> of co-repressor action. Nat. Reu. Genet. 11:109-123. Mmks, J., and C. J. Brown. 2009. Getting to the center of
Smith, C. L., and C. L. Peterson. 2005. ATP-dependent X-chromosome macnvarion: the role of transgenes. B10chem. Cell.
chromatin remodeling. Ctm. Top. Dev. Bioi. 65: I 15-148. BIOI. 87:759-766.
Taaqes, D. J. 2010. The human Mediator complex: a versanle, ;\1oazed, D. 2009. Small RNAs 111 transcnptional gene silencing
genome-wide regulator of tran~cription. Trends Bl()chem. Sci. and genome defence. Nature 457:413-420.
35:31 S-322. Simon, J. A., and R. E. Kmgsron. 2009. Mechanisms of
Venters, B. J., and B. F. Pugh. 2009. HO\\ cukar}·onc genes arc polycomb gene silencmg: known~ and unknowm. Nat. Reu. Alo/.
transcribed. Cnt. Rev. Bwchem. Mol. 810/. 44:117-141. Cell B10l. 10:697-708
Yun, ;\I., ct al. 2011. Readers of histone modifications. Cell Vermaak, D., and H. S. ~1ahk. 2009. ~1ulnple roles for
Res. 21:S64-578. heterochromatin protem l genes in Drosophila. Ann. Reu. Genet.
43:467-492.
Wurz, A., and J. Gribnau. 2007. X inacti,·anon Xplained. Curr.
Opm. Genet. Deu. 17:387-393. ·
Regulation of Transcription-Factor Activity
AltareJOS, .J. Y., and M. Montmmy. 20 I\. CRI:.B and the CRTC Other Eukaryotic Transcription Systems
co-activators: sensors for hormonal and metabolic <;ignals. Nat. Rev. Bona\\ 1tz, :N. D., D. A. Clayton, and G. S. Shadel. 2006.
Mol. Cell 810/. 12:141-151. Initiation and beyond: multiple functions of the human mitochondrial
Brivanlou, A. H., and J. F. Darnell, Jr. 2002. Signal transduc- transcnpnon machinery. Mol. Cell24:8l3-825.
tion and the control of gene expression. Science 295:813-818. Dumay-Odelor, H., et al. 2010. Cell growth- and ddfcrenna-
Chen, H., M. Tmi, and R. ,\1. Evans. 2001. HATs on and tion-dcpendent regulation of RNA polymerase Ill transcnption. Cell
beyond chromatin. Curr. Opin. Cell Bioi. 13:218-224. Cycle 9:3687-3699.
Echeverria, P. C., and D. Picard. 2010. ;\\olecular chaperones, Grummt, I. 2010. Wisely chosen paths-regulation of rRNA
essential partners of steroid hormone receptors for activity and symhes1s. FEBS]. 277:4626-4639.
mobdity. Bioclnm. B10phys. Acta 1803:641-649. Leigh-Brown, S., J. A. Enriquez, and D. T. Odom. 20 I 0.
Lefstin, J. A., and K. R. Yamamoto. 1998. Allosteric effects of :Nuclear transcnption factOrs in mammalian mltochondna. Genome
DNA on transcriptional regulators. Nature 392:885-888. BIOI. 11:215:1-9.
Perissi, V., and YL G. Rosenfeld. 2005. Controlling nuclear Schweer, J., cr al. 2010. Role and regulation of plastid sigma
receptors: the circular logic of cofactor cycles. Nat. Rev. Mol. Cell factors and their functional imeractors durmg chloroplast
Bwl. 6:542-554. transcnption-recem lessons from Arabidops1s thal!ana. Eur. .f. Cell
Wu, S. C., andY. Zhang. 2009. Minireview: role of protein Bioi. 89:940-946.
methylation and demethylation in nuclear hormone signaling. Mol. Willis, I. M., and R. D. Moir. 2007. Integration of nutritional and
Endocrinol. 23:1323-1334. , stress Signalmg pathways by ;\1af1. Trends Biochem. Sci. 32:51-53.

References 343
.·

.·
 CHAPTER

Post-transcriptional
Gene Control
Portion of a "lampbrush chromosome" from an oocyte of the newt
Nophtha/mus viridescens; hnRNP protein associated with nascent RNA
transcripts fluoresces red after staining with a monoclonal antibody.
[Courtesy of M. Roth and J. Gall.]

n the previous chapter, we saw that most genes are regulated that -95 percent of human genes give rise to alternatively

I at the first step in gene expression, transcription, by regu-

lating the assembly of the transcription pre-initiation com-
plex on a promoter DNA sequence and regulating transcription
spliced mRNAs. These alternatively spliced mRNAs encode
related proteins with differences in sequences limited to spe-
cific functional domains. In many cases, alternative RNA
elongation in the promoter proximal region . Once transcrip- splicing is regulated to meet the need for a specific protein
tion has been initiated, synthesis of the encoded RNA re- isoform in a specific cell type. Given the complexity of pre-
quires that RNA polymerase transcribe the entire gene and mRNA splicing, it is not surprising that mistakes are occasion-
not terminate prematurely. Moreover, the initial primary ally made, giving rise to mRNA precursors with improperly
transcripts produced from eukaryotic genes must undergo spliced exons. However, eukaryotic cells have evolved RNA
various processing reactions to yield the corresponding func- surveillance mechanisms that prevent the transport of mcor-
tional RNAs. For mRNAs, the 5' cap structure necessary for rectly processed RNAs to the cytoplasm or lead to their deg-
translation must be added (see Figure 4-14), introns must be radation if they are transported.
spliced out of pre-mRNAs (Table 8-1), and the 3' end must Additional control of gene expression can occur in the
be polyadenylated (see Figure 4-1 5 ). Once formed in the nu- cytoplasm. In the case of protein-coding genes, for instance,
cleus, mature, functional RNAs are exported to the cyto- the amount of protein produced depends on the stability of
plasm as components of ribonucleoproteins. Both processing the corresponding mRNAs in the cytoplasm and the rate of
of RNAs and their export from the nucleus offer opportuni- their translation. For example, during an immune response,
ties for further regulating gene expression after the initiation lymphocytes communicate by secreting polypeptide hor-
of transcription. mones called cytokines that signal neighboring lymphocytes
Recently, the vast ampunt of sequence data on human through cytokine receptors that span their plasma mem-
mRNAs expressed in different tissues and at various times branes (Chapter 23). lt is important for lymphocytes to syn-
during embryogenesis and cellular differentiation has revealed thesize and secrete cytokines in short bursts. This is possible

OUTLINE

8.1 Processing of Eukaryotic Pre-mRNA 348 8.4 Cytoplasmic Mechanisms of Post -transcriptional
Control 370
8.2 Regulation of Pre-mRNA Processing 360
8.5 Processing of rRNA and tRNA 384
8.3 Transport of mRNA Across the Nuclear Envelope 365
FiJ:i!J:ii RNAs Discussed in Chapter 8

mRNA Fully processed messenger RNA with 5' cap, introns removed b} RNA splicing, and a poly(A) tail

pre-mRNA An mRNA precursor conraming introns and not cleaved at the poly(A) site

hnRNA Heterogeneous nuclear RNAs. These include pre-mRNAs and RNA processing intermediates containing one or
more inrrons.

snR~A Five small nuclear RNAs that function in the removal of introns from pre-mRNAs by RNA splicing, plus rwo
small nuclear RNAs that substitute for the first two at rare inrrons

pre-tRNA A tRNA precursor containing additional transcribed bases at the 5' and 3' ends compared to the mature tRNA.
Some pre-tRNAs also contain an intron in the anti-codon loop.

pre rR~A The precursor to mature J8S, 5.8S, and 28S ribosomal RNAs. The mature rRNAs are processed from this long
precursor RNA molecule by cleavage, removal of bases from the ends of the cleaved products, and modification
of specific bases. '

snoRNA Small nucleolar RNAs. These base-pair with complementary regions of the pre-RNA molecule, directing
cleavage of the RNA chain and modification of bases during maturation of the rRNAs.

siRNA Short interfering RNAs, -22 bases long, that arc each perfectly complementary to a sequence in an mRNA.
Together with associated proteins, siRNAs cause cleavage of the "target" RNA, leading to its ra~id degradation.

m1RNA M1cro RNAs, -22 bases long, that base-pa1r extensively, but not completely, with mRNAs, especially over the
SIX base pairs at the 5' end of the miRNA. Th1s inhibits translation of the "target" mRNA.

because cytokine mRNAs are extremely unstable. Conse- expressed in the multiple types of human cells. Although some
quently, the concentration of the mRNA in the cytoplasm have recently been discovered to function through inhibition
falls rapidly once its synthesis is stopped. In contrast, mRNAs of target gene expression in the appropriate tissue and at the
encoding proteins required in large amounts that function appropriate time in development, the functions of the vast
over long period.s, such as ribosomal proteins, are extremely majority of human miRNAs are unknown and are the subject
stable so that multiple polypeptides are transcribed from each of a growing new area of research. If most miRNAs do indeed
mRNA. have significant functions, miRNA genes constitute an impor-
In addition to regulation of pre-mRNA processing, nu - tant subset of the -25,000 human gene~- A closely related
clear export, and translation, the cellular locations of many, process called RNA interference (RNAi) leads to the degrada-
if not most, mRNAs are regulated so that newly synthesized tion of viral RNAs in infected cells and the degradation of
protein is concentrated where it is needed. Particularly strik- transposon-encoded RNAs in many eukaryotes. This is of tre-
ing examples of this occur in the nervous systems of multi- mendous significance to biological researchers because it is
cellular animals. Some neurons in the human brain generate possible to design short interfering RNAs (siRNA) to inhibit
more than 1000 separate synapses with other neurons. Dur- the translation of specific mRNAs experimentally by a process
ing the process of learning, synapses that fire more frequently called RNA knockdown. This makes it possible to inhibit the
than others increase in size many times, while other synapses function of any desired gene, even in organisms that arc not
made by the same neuron do not. This can occur because amenable to classic genetic methods for isolating mutants.
mRNAs encoding proteins crittcal for synapse enlargement We refer to all the mechanisms that regulate gene expres-
are stored at all synapses, but translation of these localized, sion following transcription as post-transcriptional gene con-
stored mRNAs is regulated at each synapse independently by trol (Figure 8-1 ). Since the stability and translation rate of an
the frequency at which it initiates firing. In this way, synthe- mRNA contribute to the amount of protein expressed from a
sis of synapse-associated protein<, c;1n hP regulated independ- gene, these post-transcriptional processes are important com-
ently at each of the many synapses made by the same neuron. ponents of gene control. Indeed, the protein output of a gene
Another type of gene regulation that has recently come to is regulated at every step in the life of an m RNA from the ini-
light involves micro RNAs (miRNAs), which regulate the sta- tiation of its synthesis to its degradation. Thus genetic regula -
bility and translation of specific target mRNAs in multicellu- tory processes act on RNA as well as DNA. In this chapter,
lar animals and plants. Analyses of these short miRNAs in we consider the events that occur in the processing of mRNA
vanous human tissues indicate that there are -500 miRNAs following transcription initiation and promoter proximal

346 CHAPTER 8 • Post-transcriptional Gene Co ntrol

 Nucleolus

Pre-rR NA
Pre-mRNA transcription
transcription 5S rRNA

~
\
~~p..
Pre-tRNA
transcription
Ribosomal
subunit ,Sf)
Exc1sed
synthesis in
nucleolus pre-rRNA
a q ~ =~~i~Tn~NA mD
.......-----------~-
1'U:111 polyadenylation
Cleavage'
~

.-------------~-
...-
Improperly
processed Exosome
mRNA ~

~0 m ~d
pre-tRNA
mANA
export Nucleus

~AAAAA

liJ iI
Cytoplasmic
Poly(A) pol
Cytoplasmic
exosome

~
Cytoplasmic
De~~j CJ
enzyme Deadenylase

~ liJ
polyadenylation

~AAAAA
miRNA \ IJ /
i
~~AAAAA
miRNA
t ranslation inhibition Translation Cytoplasmic
initiation deadenylation

FIGURE 8-1 Overview of RNA processing and post- by degradation by cytoplasmic exosomes. The degradation rate of each
transcriptional gene control. Nearly all cytoplasmic RNAs are mRNA is controlled, thereby regu lating the mRNA concentration and,
processed from primary transcripts in the nucleus before t hey are consequently, the amount of protein translated. Some mRNAs are
exported to the cytoplasm. For protein-coding genes transcribed by synthesized without long poly(A) tails. Their translation is regulated by
RNA polymerase II, gene cont rol can be exerted through 0 the choice [iJ controlling the synthesis of a long poly(A) tail by a cytoplasmic
of alternative exons during pre-mRNA splicing and f) the choice of poly(A) polymerase. fJ Translation is also regulated by other mecha-
alternative poly(A) sites. Improperly processed mRNAs are blocked nisms including miRNAs. When expressed, t hese - 22-nucleotide RNAs
from export to the cytoplasm and deqraded 11 by a large complex inhibit translation of mRNAs to whic.h they hybridize, usually in the
called the exosome that contains multiple ribonucleases. Once 3'-untranslated region. tRNAs and rRNAs are also synthesized as
exported to the cytop l asm, ~ translation initiation factors bind to t he precursor RNAs that must be (;) processed before they are functional.
mRNA 5' -cap cooperatively with poly(A)-binding protein I bound to Regions of precursors cleaved from the mature RNAs are degraded by
m
the poly(A) tail and initiate translation (see Figure 4-28). mRNA is nuclear exosomes ~ . [Adapted from Houseley, et. al., 2006, Nat. Rev. Mol. Cell
degraded in the cytoplasm by de-adenylation and decapping followed Bioi. 7:529.)

CH A PTER 8 • Post-transcriptional Gene Control 347

elongation and the various mechanisms t hat are known to precursor is being transcribed. Thus pre-mRNA processing is
regulate these events. In the last section, we briefly discuss the co-transcriptional. As the RNA emerges from the surface of
processing of primary transcripts produced from genes encod- RNA polymerase II, its 5' end is immediately modified by the
mg rRNAs and tRNAs. addition of the 5' cap structure found on all mRNAs (see Fig-
ure 4-14). As the nascent pre-mRNA continues to emerge from
the surface of the polymerase, it is immediately bound by
members of a complex group of RNA-binding proteins that
8.1 Processing of Eukaryotic Pre-mRNA
assist in RNA splicing and export of the fully processed mRNA
In this section, we take a closer luuk at huw t:ukaryoric cells through nuclear pore complexes into the cytoplasm. ~orne of
convert the initial primary transcript synthesized by RNA these proteins remain associated with the mRNA in the cyto-
polymerase II into a functional mRNA. T h ree major events plasm, but most either remain in the nucleus or shuttle back
occur during the process: 5' capping, 3' cleavagelpolyadenyla- into the nucleus shortly after the mRNA is exported to the
tion, and RNA splicing (Figure 8-2). Adding these specific cytoplasm. Cytoplasmic RNA-bindi ng proteins are exchanged
modifications to the 5 ' and 3' ends of the pre-mRNA is impor- for the nuclear ones. Consequently, mRNAs never occur as
tant to protect it from enzymes that quickly digest uncapped free RNA molecules in the cell but are always associated with
RNAs generated by RNA processing, such as spliced-out in- protein as 1ibonucleoprotein (RNP) complexes, first as nascent
trons and RNA transcribed downstream from a polyadenyla- pre-mRNPs that are capped and- spliced as they are tran-
tion site. The 5' cap and 3' poly(A) tail distinguish pre-mRNA scribed. Then, following cleavage and polyadenylation, they
molecules from the many other kinds of RNAs in the nucleus. are referred to as nuclear mRNPs. Following the exchange of
Pre-mRNA molecules (see Table 8-1 ) are bound by nuclear proteins that accompany export to the cytoplasm, they are
proteins that function in mRNA export to the cytoplasm. After called cytofJ!asmic mRNPs. Although we frequently refer to
mRNAs are exported to the cytoplasm, they are bound by a set pre-mRNAs and mRNAs, it is important to remember that
of cytoplasmic proteins that stimulate translation and are criti- they are always associated with proteins as RNP complexes.
cal for mRNA stabi lity in the cytoplasm. Prior to nuclear ex-
port, introns must be removed to generate the correct coding
The 5 ' Cap Is Added to Nascent RNAs
region of the mRNA. In higher eukaryotes, including humans,
alternative splicing is intricately regulated in order to substitute Shortly After Transcription Initiation
different functional domains into proteins, producing a consid- As the nascent RNA transcript emerges from the RNA chan -
erable expansion of the proteome of these organisms. nel of RNA polymerase II and reaches a length of - 25 nu-
The pre-mRNA processing events of capping, splicing, and cleotides, a protective cap composed of 7 -methylguanosine
polyadenylation occur in the nucleus as the nascent mRNA and methylated riboses is added to the 5' end of eukaryotic

0 ANIMATION: Life Cycle of an mRNA

Poly(A) Termination
site sites

DNA

Primary RNA
Cap
tII Transcription, 5' capping
•••• 3'
transcript
5'
Endo~ucle
tfJ Cleavage at poly(A) site
"'J
ro
3:xJ
5'
A• polymetase
tIJ Polyadenylation
3'
z
l>
'0
0
5'
tIIJ RNA splicing
A1oo-2so3' C)
(!)
(/)

!:'!.
::J
co
mRNA 5' A100-2so 3 ' J
FIGURE 8 - 2 Overview of mRNA processing in eukaryotes. Shortly adenosine (A) residues is added (step iJ). The poly(A) tail contains - 250 A
after RNA polymerase II initiate~ trdmc.ription at the first nucleotide of the residues in mammals, - 1SO in insects, and - 100 in yeasts. For short
first exon of a gene, the 5' end of the nascent RNA is capped with primary transcripts with few introns, splicing (step 9 ) usually follows cleav-
7-methylguanylate (step 0 ). Transcription by RNA polymerase II age and polyadenylation, as shown. For large genes with multiple introns,
terminates at any one of multiple termination sites downstream from the introns often are spliced out of the nascent RNA during its transcription,
poly(A) site, which is located at the 3' end of the final exon. After the i.e., before transcription of the gene is complete. Note that the 5 ' cap and
primary transcript is cleaved at the poly(A) site (step f)), a string of sequence adjacent to the poly(A) tail are retained in mature mRNAs.

348 CHAPTER B • Post -transcriptional Gene Control

 5' end of RNA activation of the capping enzyme by the phosphorylated
CTD result in specific capping of R0!As transcribed by RNA
N -J@ijj,J;J~t;.i polymerase II.
One subunit of the capping enzyme removes the-y-
phosphate from the 5' end of the nascent RNA (Figure 8-3).
Another domain of this subunit transfers the GMP moiety
from GTP to the 5 '-diphosphate of the nascent transcript,
a ~ y ~ a
creating the unusual guanosine 5'-5' -triphosphate structure.
G 0
P P + ~ fl N-i@ij .. @r!J
'----v----' In the final steps, separate eu.qmes transfer methyl groups
GTP from S-adenosylmethionine to the N7 position of the guanine
and one or two 2' oxygens of riboses at the 5' end of the
nascent RNA.
Considerable evidence indicates that capping of the nas-
N -i@ij,J;J~f!l
cent transcript is coupled to elongation by RNA polymerase

l
Gua~ine 7-mot~ +CH 3 from
II so that all of its transcripts are capped during the earliest
t ansfe S-Ado-Met phase of elongation. As discussed in Chapter 7, in metazoans,
during the initial phase of transcription the polymerase elon-
gates the nascent transcript very slowly due to association of
N-i@ij,I;J@f!l
NELF (negative elongation factor) with RNA polymerase II

l+CH 3 from in the promoter proximal region (see Figure 7-20). Once the
S-Ado-Met 5' end of the nascent RNA is capped, phosphorylation of the
RNA polymerase CTD at position 2 in the heptapeptide re-
peat and of NELF and DSIF by the CDK9-cyclin T protein
kinase causes the release of NELF. This allows RNA polymer-
FIGURE 8-3 Synthesis of 5' -cap on eukaryotic mRNAs. The 5' end ase Il to enter into a faster mode of elongat_ion that rapidly
of a nascent RNA contains a 5 '-triphosphate from the initiating NTP. transcribes away from the promoter. The net effect of this
The-y-phosphate is removed in the first step of capping, while the mechanism is that the polymerase waits for the nascent RNA
remaining a- and (3-phosphates (orange) remain associated with the to be capped before elongating at a rapid rate.
cap. The third phosphate of the 5',5'-triphosphate bond is derived
from the ex-phosphate ofthe GTP that donates the guanine. The
methyl donor for methylation of the cap guanine and the first one
or two riboses of the mRNA is 5-adenosylmethionine (5-Ado-Met). A Diverse Set of Proteins with Conserved
[From S. Venkatesan and B. Moss, 1982, Proc. Nor/. Acad. Sci. USA 79:304.] RNA-Binding Domains Associate
with Pre-mRNAs
As noted earlier, neither nascent RNA transcripts from protein-
mRNAs (see Figure 4-14 ). The 5' cap marks RNA molecules coding genes nor the intermediates of mRJ:\A processing, col-
as mRNA precursors and protects them from RNA-digesting lectively referred to as pre-mRNA, exist as free RNA molecules
enzymes (5' -exoribonucleascs) in the nucleus and cytoplasm. in the nuclei of eukaryotic cells. From the time nascent tran-
This initial step in RNA processing is catalyzed by a dimeric scripts first emerge from RNA polymerase II until mature
capping enzyme, which associates with the phosphorylated mRNAs are transported into the cytoplasm, the RNA mole-
carboxyl-terminal domain (CTD) of RNA polymerase II. Re- cules are associated with an abundant set of nuclear proteins.
call that the CTD becomes phosphorylated by the TFIIH These are the major protein components of heterogeneous ribo-
general transcription facto.r at multiple serines at the 5 posi- nucleoprotein particles (lmRNPs), which contain heterogene-
tion in the CTD heptapeptide repeat during transcription ous nuclear RNA (hnRNA), a collective term referring to
initiation (see Figure 7-17). Binding to the phosphorylated pre-mRNA and other nuclear RNAs of various sizes. These
CTD stimulates the activity of the capping enzymes so that hnRNP proteins contribute to further steps in RNA processing,
they are focused on R NAs containing a 5'-triphosphate that including splicing, polyadenylation, and export through nu-
emerge from RNA polymerase II and nor on RNAs tran- clear pore complexes to the cytoplasm.
scribed by RNA polymerases I or III, which do not have a Researchers identified hnRNP proteins by first exposing
CTD. This is important because pre-mRNA synthesis ac- cultured cells to high-dose UV irradiation, which causes cova-
counts for only -80 percent of the total RNA synthesized in lent cross-links to form between RNA bases and closely associ-
replicating cells. About 20 percent is preribosomal RNA, ated proteins. Chromatography of nuclear extracts from treated
which is transcribed by RNA polymerase I, and 55 rRNA, cells on an oligo-dT cellulose column, which binds RNAs with
tRNAs, and other stable small RNAs, which are transcribed a poly( A) tail, was used to recover the proteins that had become
by RNA polymerase IlL The two mechanisms of (1) binding cross-linked to nuclear polyadenylared RNA. Subsequent treat-
of the capping enzyme to initiated RNA polymerase II spe- ment of cell extracts from un-irradiated cells with monoclonal
cifically through its unique, phosphorylated CTD and (2) antibodies specific for the major proteins identified by this

8.1 Processing of Eukaryotic Pre-mRNA 349

~ VIDEO: hnRNP A 1 Nucleocytoplasmic Shuttling

(a) (b) (c)

.
......
.···~!

FIGURE 8-4 Human hnRNP A1 protein can cycle in and out of the right of the oval-shaped Xenopus nucleus. (b, c) When the same
cytoplasm, but human hnRNP C protein cannot. Cultured Hela cells preparation was viewed by fluorescence microscopy, the stained
and Xenopus cells were fused by treatment with polyethylene glycol, hnRNP C protein appeared green and thi stained hnRNP A 1 protein
producing heterokaryons containing nuclei from each cell type. The appeared red. Note that the unfused Xenopus cell on the left is
hybrid cells were treated with cycloheximide immediately after fusion unstained, confirming that the antibodies are specific for the human
to prevent protein synthesis. After 2 hours, the cells were fixed and proteins. In the heterokaryon, hnRNP C protein appears only in the
stained with fluorescent-labeled antibodies specific for human hnRNP Hela-cell nucleus (b), whereas the A 1 protein appears in both the
C and A 1 proteins. These antibodies do not bind to the homologous Hela-cell nucleus and the Xenopus nucleus (c). Since protein synthesis
Xenopus proteins. (a) A fixed preparation viewed by phase-contrast was blocked after cell fusion, some of the human hnRNP A 1 protein
microscopy includes unfused Hela cells (a rrowhead) and Xenopus cells must have left the Hela-cell nucleus, moved through the cytoplasm,
(dotted arrow), as well as fused heterokaryons (solid arrow). ln the and entered the Xenopus nucleus in the heterokaryon. [SeeS. Pinoi-Roma
heterokaryon in this micrograph, the round Hela-cell nucleus is to the and G. Dreyfuss, 1992, Nature 355:730; courtesy of G. Dreyfuss.]

cross-linking technique revealed a complex set of abundant out of the cytoplasm, suggesting that they function in the
hnRJ\.rp proteins ranging in size from -30 to - 120 kDa. transport of mRNA (Figure 8-4).
Like transcription factors, most hnR, P proteins have a
modular structure. They contain one or more RNA-binding Conserved RNA-Binding Motifs The RNA recognztion motif
domains and at least one other domain that interacts with (RRM), also called the RNP motif and the RNA-binding do-
other proteins. Several different RNA-binding motifs have main (RBD), is the most common RNA-binding domain in
been identified by creating hnRNP proteins with missing hnRNP proteins. This -SO-residue domain, which occurs in
amino acid seq uences and testing their ability to bind RNA. many other RNA-binding proteins, contains two highly con-
served sequences (RNP1 and RNP2) that are found across
Functions of hnRNP Proteins The association of pre-mRNAs organisms ranging from yeast to human-indicating that
with hnRNP proteins prevents the pre-mRNAs from forming like many D A-binding domains, they evolved early in
short secondary structures dependent on base pairing of com- eukaryotic evolution.
plementary regions, thereby making the pre-mRNAs accessi- Structural analyses have shown that the RRM domain con-
ble for interaction with other RNA molecules or proteins. sists of a four-stranded ~sheet flanked on one side by two a
Pre-mRNAs associated with hnRNP proteins present a more helices. To interact with the negatively charged RNA phos-
uniform substrate for subsequent processing steps than would phates, the~ sheet forms a positively charged surface. The con-
free, unbound pre-mRNAs, in which each mRNA forms a served RNPl and RNP2 sequences lie side by side on the two
unique secondary structure due to its specific sequence. central ~ strands, and their side chains make multiple contacts
Binding studies with purified hnRNP proteins indicate that with a single-stranded region of RNA that lies across the sur-
different hnRNP proteins associate with different regions of a face of the 13 sheet (Figure 8-5).
newly made pre-mRNA molecule. For example, the hnRNP The 45-residue KH motif is found in the hnRNP K pro-
proteins A1, C, and D bind preferentially to the pyrimidine- tein and several other RNA-binding proteins. The three-
rich sequences at the 3' ends of introns (see Figure 8-7). Some dimensional structure of representative KH domains is similar
hnRNP proteins interact with the RNA sequences that specify to that of the RRM domain but smaller, consisting of a
RNA splicing or cleavage/polyadenylation and contribute to three-stranded 13 sheet supported from one side by two a
the structure recognized by RNA-processing factors. Finally, helices. Nonetheless, the KH domain interacts with RNA
cell-fusion experiments have shown that some hnRNP proteins much differently than does the RRM domain. RNA binds to
remain localized in the nucleus, whereas others cycle in and the KH domain by interacting with a hydrophobic surface

350 CHAPTER 8 • Post-transcriptional Gene Control

 (a) RNA recognition motif (RRM) (b) Sex-lethal (Sxl) RRM domains (c) Polypyrimidine tract binding protein {PTB)

FIGURE 8-5 Structure of the RRM domain and its interaction with regions, in shades of red. The pre-mRNA is bound to the surfaces of the
RNA. {a) Diagram of the RRM domain showin g the two n helices positively charged [3 sheets, making most of its contacts with the RNP 1
{green) and four [3 strands {red) that characterize this motif. The and RNP2 regions of each RRM. {c) Strikingly different orientation of
conserved RNPl and RNP2 regions are located in the two central RRM domains in a different hnRNP, the polypyrimidine tract binding
[3 stra nds. (b) Surface representation of the two RRM domains in (PTB) protein, illustrating that RRMs are oriented in different relative
Drosophila Sex-lethal (Sxl) protein, which bind a nine-base sequence in positions in different hnRNPs; colors are as in (b). Polypyrimidine (p{Y))
transformer pre-mRNA (yellow). The two RRMs are oriented like the two single-stranded RNA is bound to the upward (RRM3) and downward
parts of an open pair of castanets, with the [3 sheet of RRMl facing (RRM4) facing [3-sheets. RNA is shown in yellow. [Part (a) adapted from
upward and the [3 sheet of RRM2 facing downward. Positively charged K. Nagai et al., 1995, Trends Biochem. Sci. 20:235. Part (b) after N. Harada et al.,
regions in Sxl protein are shown in shad es of blue; negatively charged 1999, Nature 398:579. Part (c) after F. C. Oberstrass et al., 2006, Science 309:2054.)

formed by the o: helices and one 13 strand. The RGG box, the polyadenylated end of mRNA processing intermediates
another RNA-binding motif found in hnRNP proteins, con- in hnRNA are retained in the mature mRNA in the cyto-
tains five Arg-Gly-Gly (RGG} repeats with several interspersed plasm. The solution to this apparent conundrum came from
aromatic amino acids. Although the structure of this motif the discovery of introns by electron microscopy of RNA-
has not yet been determined, its arginine-rich nature is simi- DNA hybrids of adenovirus DNA and the mRNA encoding
lar to the RNA-oinding domains of the HIV Tat protein. KH hexon, a major virion capsid protein (Figure 8-6). Other
domains and RGG repeats are often interspersed in two or studies revealed nuclear viral RNAs that were colinear with
more sets in a single RNA-binding protein. the viral DNA (primary transcripts), and RNAs with one or
two of the introns removed (processing intermediates). These
Splicing Occurs at Short, Conserved results, together with the earlier findings that the 5' cap and
3' poly(A) tai l at each end of long mRNA precursors are re-
Sequences in Pre-mRNAs via Two tained in shorter mature cytoplasmic mRNAs, led to the re-
Transesterification Reactions alization that introns are removed from primary transcripts
During formation of a mature, functional mRNA, the introns as exons are spliced together.
are removed and exons are spliced together. For short tran- The location of splzce sites-that is, exon-inrron junc-
scription units, RNA splicing usua lly follows cleavage and tions-in a pre-mRNA can be determined by comparing the
polyadenylation of the 3' end of the primary transcript, as sequence of genomic DNA with that of the eDNA prepared
depicted in Figure 8-2. However, for lo ng transcription units from the corresponding mRNA (see Figure 5-15). Sequences
containing multiple exons, splicing of exons in the nascent that are present in the genomic DNA but absent from the
RNA begins before transcription of the gene is complete. eDNA represent introns ;:~nd indicate the positions of splice
Early pioneenng research on the nuclear processing of sites. Such analysis of a large number of different mRNAs
mRNAs revealed that mRNAs are initially transcribed as revealed moderately conserved, short consensus sequences at
much longer RNA molecules than the mature mRNAs in the the splice sites flanking inrrons in eukaryotic pre-mRNAs; a
cytoplasm. It was also shown that RNA sequences near the pyrimidine-rich region just upstream of the 3' splice site also
5' cap added shortly after transcription initiation are re- is common (Figure 8-7). Studies of mutant genes with dele-
tained in the mature mRNA, and that RNA sequences near tions introduced into introns have shown that much of the

8.1 Processing of Eukaryotic Pre-mRNA 351

G) PODCAST: Discovery of lntrons

EXPERIMENTAL F GURE 8-6 Electron (a) Adenovirus hexon gene

microscopy of mRNA-template DNA hybrids shows
that introns are spliced out during pre-mRNA
processing. (a) Diagram of the EcoRI A fragment of 5' ====1
1~A IIT
s JI
E=~c~==:J-=---#3'
adenovirus DNA, which extends from the left end of ~------------------EcoRIA------------------~
f--i
the genome to just before the end of the final exon
• Exons 0 lntrons lkb
of the hexon gene. The gene consists of three short
(b)
exons and one long (- 3.5 kb) exon separated by
three introns of 1, 2.5, and 9 kb. (b) Electron
micrograph (left) and schematic drawing (right) of a
hybrid between an EcoRI A DNA fragment and a
hexon mRNA. The loops marked A, B, and C corre-
spond to the introns indicated in (a). Since these
intron sequences in the viral genomic DNA are not
present in the mature hexon mRNA, they loop out
between the exon sequences that hybridize to their
complementary sequences in the mRNA. [Micrograph
from S.M. Berget et al., 1977, Proc. Not'/. Acad. Sci. USA
74:3171; courtesy of P. A. Sharp.]

\
mRNA

center portion of introns can be removed without affecting net result of these two reactions is that two exons are ligated
splicing; generally only 30-40 nucleotides at each end of an and the intervening intron is released as a branched lariat
1ntron are necessary for splicing to occur at normal rates. structure.
Analysis of the intermediates formed during splicing of
pre-mRNAs in vitro led to the discovery that splicing of exons
proceeds via two sequential transesterification reactions (Fig-
During Splicing, snRNAs Base-Pair
ure 8-8). lntrons are removed as a lariat structure in which the
5' G of the intron is joined in an unusual2' ,5 '-phosphodiester with Pre-mRNA
bond to an adenosine ncar the 3' end of the intron. This A Splicing requires the presence of small n uclear RNAs
residue is called the branch point A because it forms an RNA (snRNAs), important for base pairing with the pre-mRNA,
branch in the lariat structure. In each transestcrification re- and - 170 associated proteins. Five U-rich snRNAs, desig-
action, one phosphoester bond is exchanged for another. nated U1, U2, U4, US, and U6, participate in pre-mRNA
Since the number of phosphoester bonds in the molecule is splicing. Ranging in length from 107-210 nucleotides, these
not changed in either reaction, no energy is consumed. The snRNAs are associated with 6- 10 proteins each in the many

5' splice site Branch point 3' splice site

Pre-mRNA

Frequency of 70 60 80 100 100 95 70 80 45 80 90 80 100 80 80 100 100 60

occurrence (%) 1+---------- 20- 50 b ----------~

FIGURE 8 -7 Consensus se quences around splice sites in adenosine, also invariant, usually is 20-50 bases from the 3' splice site.
vertebrate pre-mRNAs. The only nearly invariant bases are the 5' GU The central region of the intron, which may range from 40 bases to
and the 3 ' AG of the intron (blue), although the flanking bases 50 kilo bases in length, generally is unnecessary for splicing to occur.
indicated are found at frequencies higher than expected based on a [SeeR. A. Padgett et al., 1986, Ann. Rev. Biochem. 55:1119, and E. B. Keller and
random distribution. A pyrimidine-rich region (hatch marked) near the W. A. Noon, 1984, Proc. Not'/. Acad. Sci. USA 81 :7417.]
3' end of the intron is found in most cases. The branch-point

352 CHAPTER 8 • Post-transcriptional Gene Control

 lntron (Figure 8-9b). Involvement of U2 snRNA in splicing initially
was suspected when it was found to have an internal se-
( 2' A
quence that is largely complementary to the consensus se-
I quence flanking the branch point in pre-mRNAs (see Figure
5'0 H- 3' 8-7). Compensating mutation experiments, similar to those
1/
O=P-o- -o-P=O
I conducted with Ul snRNA and 5' splice sites, demonstrated
1 I that base pairing between U2 snRNA and the branch-point
5'- - - -0 0 - Exon 2 - 3 ' sequence in pre-mRNA also is critical to splicing.
- - - - 3' 5'
Figure 8-9a illu~trates the general structures of the U1 and
0 = 3' oxygen of
exon 1
l First transesterification
U2 snRNAs and how they base-pair with pre-mRNA during
splicing. Significantly, the branch-point A itself, which is not
0 = 2' oxygen of base-paired to U2 snRNA, "bulges out" (Figure 8-lOa), al-
branch point A lowing its 2' hydroxyl to participate in the first transesterifi-
I. · =3' oxygen of cation reaction of RNA splicing (see Figure 8-8).
intron Similar studies with other snRNAs demonstrated that
5'0 ba se pairing between them also occurs during splicing.
I 2> A Moreover, rearrangements in these RNA-RNA interactions
O=P-0 are critical in the splicing pathway, as we describe next.
I
0 3'
f
-o- P=O Spliceosomes, Assembled from snRNPs
·. ~I
5'- - - -0 -H 0- - 3' and a Pre-mRNA, Carry Out Splicing
3' 5'
The five splicing snRNPs and other proteins involved in splicing
l Second transesterification
assemble on a pre-mRNA, forming a large ribonucleoprotein
complex called a spliceosome (Figure 8-11 ). '.fhe spliceosome
has a mass similar to that of a ribosome. Assembly of a spliceo-
some begins with the base pairing of the Ul snRNA to the
0 5' splice site and the cooperative binding of protein SFl (splic-
I
+ 5'- - - -0 - P - 0 - io--3' ing factor l ) to the branch point A and the heterodimeric pro-
5'0 11
tein U2AF (U2 associated factor) to the pyrimidine tract and
I 2> A 0
O=P- 0 Spliced exons the 3' AG of the intron via its large and small subunits, respec-
I 3' tively. The U2 snRNP then base-pairs with the branch point
o- OH
region (Figure 8-9a) as Sfl is released. Extensive base pairing
Excised lariat intron
between the snRNAs in the U4 and U6 snRNPs forms a
FIGURE 8-8 Two transesterification reactions that result in complex that associates with U5 snRNP. The U4/U6/U5
splicing of exons in pre-mRNA. In the first reaction, the ester bond "tri-snRNP" then associates with the previously formed U1/
between the 5' pho51Jhorus of the intron and the 3' oxygen (dark red)
U2/pre-mRNA complex to generate a spliceosome.
of exon 1 is exchanged for an ester bond with the 2' oxygen (blue) of the
After formation of the spliceosome, extensive rearrange-
branch-site A residue. In the second reaction, the ester bond between
ments in the pairing of snRNAs and the pre-mRNA lead to
the 5' phosphorus of exon 2 and the 3' oxygen (orange) of the intron is
exchanged for an ester bond with the 3' oxygen of exon 1, releasing the
release of the U1 snRNP. Figure 8-10b shows a cryoelectron-
intron as a lariat structure and joining the two exons. Arrows show where
microscopy structure of this intermediate in the splicing pro-
activated hydroxyl oxygens react with phosphorus atoms. cess. A further rearrangement of spliceosomal components
occurs with the loss of U4 snRNP. This generates a complex
that catalyzes the first transesterification reaction that forms
small nuclear ribonucleoprotein particles (s nRNPs) in the the 2' ,5 '-phosphodiester bond between the 2' hydroxyl on the
nuclei of eukaryotic cells. branch point A and the phosphate at the 5' end of the intron
Definitive evidence for the role of Ul snRNA in splicing (Figure 8-9). Following another rearrangement of the snRNPs,
came from experiments indicating that base pairing between the second transesterification reaction ligates the two exons in
the 5' splice site of a pre-mRNA and the 5' region ofUl snRNA a standard 3' ,5' -phosphodiester bond, releasing the intron as
is required for RNA splicing (Figure 8-9a). In vitro experiments a lariat structure associated with the snRNPs. This final intron-
showed that a synthetic oligonucleotide that hybridizes with snRNP complex rapidly dissociates, and the individual snRNPs
the 5' -end region of U 1 snRNA blocks RNA splicing. In vivo released can participate in a new cycle of splicing. The excised
experiments showed that base-pairing-disrupting mutations intron is then rapidly degraded by a debranching enzyme and
in the pre-mRNA 5' splice site also block RNA splicing; in other nuclear RNases discussed later.
this case, however, splicing can be restored by expression of As mentioned above, a spliceosome is roughly the sile of a
a mutant Ul snRNA with a compensating mutation that re- ribosome and is composed of - 170 proteins, including -100
stores base pairing to the mutant pre-mRNA 5' splice site "splicing factors" in addition ro the proteins associated with the

8.1 Processing of Eukaryotic Pre-mRNA 353

~ FOCUS ANIMATION: mRNA Splicing

(a)

~Ji~1snRNA
l
l
U2 snRNA
Sm
3' --

cap 5 '
Sm
3' - Gl - • · ·· · _ ca 5'
-----l.JJ. IIIIII P I I II II Py
5 ' ~UAAGU-- ----------UACUAC CAG G Exon 2 <-- 3'
Pre·mRNA A
'-...._B ranch point

(b)
W.·t. U1 snRNA 3 ' 1 - . cap 5' Mutant U1 snRNA 3 ' ( _ _ . . . U." cap 5'

Mutant pre·mRNA 5'~~liA~--- - 3' Mutant pre·mRNA

-----l.JJ.ll I I I I
5' ~ U AAAU - 3'

Mutation in pre-m RNA 5' splice site Compensatory mutat ion in U 1

blocks splicing restores splicing
FIGURE 8 -9 Base pairing between pre-mRNA, U1 snRNA, and U2 characterizing components of the splicing reaction."(b) Only t he 5' ends
snRNA early in the splicing process. (a) In this diagram, secondary of U 1 snRNAs and 5' splice sites in pre-mRNAs are shown. (Left) A
structures in the snRNAs that are not altered during splicing are mutation (A) in a pre-mRNA splice site that interferes with base pairing
depicted schematically. The yeast branch-point sequence is shown to the 5' end of Ul snRNA blocks splicing. (Right) Expression of a Ul
here. Note that U2 snRNA base-pa irs w ith a sequence that includes the snRNA w ith a compensating mutation (U) that restores base pairing
branch-point A, although this residue is not base-paired. The purple also restores splicing of the mutant pre-mRNA. [Part (a) adapted from
rectangles represent sequences that bind snRNP proteins recognized M. J. Moore et al., 1993, in R. Gesteland and J. Atkins, eds., The RNA World, Cold
by anti-Sm antibodies. For unknown reasons, antisera from patients Spring Harbor Press, pp. 303-357. Part (b) see Y. Zhuang and A.M. Weiner, 1986,
with the autoimmune disease systemic lup us eryt hematosus (SLE) Ce// 46:827.]
contain antibodies to snRNP p roteins, w h ich have been useful in

t,;) VIDEO: Dynamic Nature of Pre-mRNA Splicing Factor Movement in Living Cells

FIGURE 8-10 Structures of a bulged A in an RNA· RNA helix and (a) Self-complementary (c) Spliceosome structure
an intermediate in the splicing process. (a) The structure of an RNA sequence w ith bulging A
duplex w ith the sequence shown, cont aining bulged A residues (red) at
5'UACUA ACGU AGUA
position 5 in the RNA helix, was determined by x-ray crystallography.
AUGA UGCA UCAU 5'
(b) The bulged A residues extend from the side of an A-form RNA-RNA A
helix. The phosphate backbone of one strand is shown in green; the
other strand in blue. The structure on th e right is turned 90 degrees for (b) X-ray crystallog ra phy
a view down the axis o f the helix. (c) 40 A reso lution structure of a structure
spliceosomal splicing intermediate containing U2, U4, US, and U6 37oA
snRNPs, determined by cryoelectron microscopy and image recon-
struction. The U4/U6/US tri-snRNP complex has a similar structure to
th~ triangu lar body of this complex, suggesting that t hese snRNPs are
at the bottom o f the structure shown here and t hat the head is
composed largely of U2 snRNP. [Parts (a) and (b) from J. A. Berglund et al., As As
2001, RNA 7:682. Part (c) from D. Boehringer et al., 2004, Nor. Srruct. Mol. Bioi. (top) (bottom)
11 :463. See also H. Stark and R. Luhrmann, 2006, Annu. Rev. Biophys. Biomol.
Srruct. 35:435.]

354 CHAPTER 8 • Post-transcriptional Gene Control

 0 FOCUS ANIMATION: mRNA Splicing
FIGURE 8-11 Model of spliceosome-mediated splicing of SF1 U2AF
pre-mRNA. Step 0 : After U1 base pairs with the consensus S'splice U1
site, SF1 (splicing factor 1) binds the branch point A; U2AF (U2 snRNP 5- pG - "A Yn.rJ.G 3

associated factor) associates with the polypyrimidine tract and 3'splice

site; and the U2 snRNP associates with the branch point A via base-
pairing interactions shown in Figure 8-9, displacing SF1 . Step IFJ: A D~ SFl
trim eric snRNP complex of U4, US, and U6 joins the initial complex
to form the spliccosome. Step ~ : Rearrang~ments of base-pa1rmg U1
interactions between snRNAs converts the spliceosome into a 5- pG ___. - A p 3
catalytically active conformation and destabilizes the U1 and U4 U2
snRNPs, which are released. Step~: The catalytic core, thought to
be formed by U6 and U2, then catalyzes the first transesterification fJ ~ U4/ U6/ U5
reaction, forming the intermediate containing a 2'.5'-phosphodiester
bond as shown in Figure 8-8. Step I,'J: Following further rearrangements
between the snRNPs, the second transesterification reaction joins the
two exons by a standard 3 ',S'-phosphodiester bond and releases the U1 \ A
I U2
intron as a lariat structure and the remaining snRNPs. Step r;J: The U6
"-..:.:p 3'
excised lariat intron is converted into a linear RNA by a debranching
enzyme. [Adapted from T. Villa et al .. 2002, Ce// 109:149.] U5 Spliceosome

five snRNPs. This makes RNA splicing comparable in complex-

~~~ Ul, U4
ity to initiation of transcription and protein synthesis. Some of
the splicing factors are associated with snRNPs, but others are ·~
not. For instance, the 65-kD subunit of the U2-associated factor U6 A./
OH
(U2AF) binds to the pyrimidine-rich region near the 3' end of G/ ) .' U2
s P;.__/
introns and to the U2 snRNP. The 35-kD subunit of U2AF U5 P 3'
binds to the AG dinucleotide at the 3' end of the intron and also -
interacts with the larger U2AF subunit bound nearby. These two
U2AF subunits act together with SF1 to help specify the 3' splice
site by promoting interaction of U2 snRNP with the branch
IJ 1 First transesterification

G
point (see Figure 8-11, step 0 ). Some splicing factors also ex-
hibit sequence homologies to known RNA helicases; these are
probably necessary for the base-pairing rearrangements that
U6 (
occur in snRNAs during the spliceosomal splicing cycle.
.· Following RNA splicing, a specific set of hnRNP proteins --o~P 3
remain bound to the spliced RNA approximately 20 nucle- s U5
otides 5' to each exon-exon junction, thus forming an exon-
junction complex. One of the hnRNP proteins associated with
the exon-junction complex is the RNA export factor (REF),
II 1 Second transesterification

which functions in the export of fully processed mRNPs from

the nucleus to the cytoplaslll, as discussed in Section 8.3. Other + U2, U5, U6
proteins associated with the exon-junction complex function Spliced exons
in a quality-control mechanism that leads to the degradation of
improperly spliced mRNAs, known as nonsense-mediated
decay (Section 8.4 ).
A small fraction of pre-mRNAs (< 1% in humans) con-
tain introns whose splice sites do not conform to the standard 5 pG- - - -A - OH3
consensus sequence. This class of introns begins with AU and Linear intron RNA
ends with AC rather than following the usual "GU-AG rule"
(see Figure 8-7). Splicmg ot this special class of introns occurs
via a splicing cycle analogous to that shown in Figure 8-11, sponding pre-mRNA by removal of internal intron s and
except that four novel, low-abundance snRNPs, together spl icing of exons. However, in two types of protozoans-
with the standard US snRNP, are involved. trypanosomes and euglenoids-mRNAs are constructed by
Nea rly all functional mRNAs in vertebrate, insect, and splicing together separate RNA molecules. This process, re-
plant cells are derived from a single molecule of the corre- ferred to as trans-splicing, is also used in the synthesis of

8.1 Processing of Eukaryotic Pre -mRNA 355

 RNA polymerase II

~
~. Linker CTD
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
I< 28 nm 65 nm --------------+1
FIGURE 8·12 Schematic diagram of RNA polymerase II with the polymerase. The CTD of mammalian RNA polymerase II is twice as long.
CTD ext ended. The length of the fully extended yeast RNA polymerase In its extended form, the CTD can associate with mu ltiple RNA-processing
II carboxyl-terminal domain (CTD) and the linker region that connects it factors simultaneously. [From P. Cramer, D. A. Bushnell, and R. D. Kornberg,
to the polymerase is shown relative to the globular domain of the 2001, Science 292:1863.)

10-15 percent of the mRNAs in the nematode (roundworm) (-3500 bases). The longest introns contain upward of 500 kb!
Caenorhabditis elegans, an important model organism for Because the sequences of 5' and 3' splice sites and branch
studying embryonic development. Trans-splicing is carried po.ints are so degenerate, multiple copies are likely ro occur
out by snRNPs by a process similar to the splicing of exons random ly in long introns. Consequently, additional sequence
m a single pre-mRNA. information is required to define the exons that should be
spliced together in higher organisms with long introns.
Chain Elongation by RNA Polymerase II The information for defining the splice sites that demar-
cate exons is encoded within the sequences of the exons. A
Is Coupled to the Presence of RNA- family of RNA-binding proteins, the SR proteins, interact
Processing Factors with sequences within exons called exonic spJicing enhancers.
I fow is RNA processing efficiently coupled with the transcrip- SR proteins are a subset of the hnRNP proteins discussed
tion of a pre-mRNA? The key lies in the long carboxyl-terminal earlier and so contain one or more RRM RNA-binding do-
domain (CTD) of RNA polymerase II, which, as discussed in mains. They also contain several protein-protein interaction
Chapter 7, is composed of multiple repeats of a seven-residue domains rich in arginine (R ) and serine (S) residues called RS
(heptapeptide) sequence. When fully extended, the CTD do- domains. When bound to exonic splicing enhancers, SR pro-
main in the yeast enzyme is about 65 nm long (Figure 8-12); teins mediate the cooperative binding of Ul snRNP to a true
the CTD in human RNA polymerase II is about twice as long. 5' splice site and U2 snRNP to a branch point through a net-
The remarkable length of the CTD apparently allows multiple work of protein-protein interactions that span an exon (Fig- _I
proteins to associate simultaneously with a single RNA ure 8-13). The complex of SR proteins, snRNPs, and other
polymerase II molecule. For instance, as mentioned earlier, the splicing factors (e.g., U2AF) that assemble across an exon,
enzymes that add the 5' cap to nascent transcripts associate which has been called a cross-exon recognition complex, per-
with the CTD phosphorylated on multiple serines at the fifth mits precise specification of exons in long pre-mRNAs.
position in the heptapeptide repeat (Ser-5) during or shortly
after transcription initiation by a subunit of TFIIH. In addi- In the transcription units of higher organisms with
tion, RNA splicing and polyadenylation factors have been long introns, exons not only encode the amino acid
found to associate with the phosphorylated CTD. As a conse- sequences of different portions of a protein but also contain
quence, these processing factors are present at high local con- binding sites for SR proteins. Mutations that interfere with
centrations when splice sites and poly( A) signals are transcribed the binding of an SR protein to an exonic splicing enhancer,
hy the polymerase, enhancing the rate and specificity of RNA even if they do not change the encoded amino acid sequence,
processing. In a reciprocal fashion, the association of hnRNP prevent formation of the cross-exon recognition complex. As
proteins with the nascent RNA enhances the interaction of a result, the affected exon is "skipped" during splicing and is
RNA polymerase II with elongation factors such as DSIF and not included in the final processed mRNA. The truncated
CDK9-cyclin T (P-TEFb) (Figure 7-20), increasing the rate of mRNA produced in this case is either degraded or translated
transcription. As a consequence, the rate of transcription is into a mutant, abnormally functioning protein. This type of
coordinated with the rate of nascent RNA association with mutation occurs in some human genetic diseases. For exam-
hnRNPs and RNA-processing factors. This mechanism may ple, spinal muscular atrophy is one of the most common ge-
ensure that a pre-mRNA is not synthesized unless the machin- netic causes of childhood mortality. This disease results from
ery for processing it is properly positioned. mutations in a region of [he genumt: cumaining two closely
related genes, SMNJ and SMN2, that arose by gene duplica-
tion. SMN2 encodes a protein identical with SMNJ. SMN2
SR Proteins Contribute to Exon Definition
is expressed at a much lower level because a silent mutation
in Long Pre-mRNAs in one exon interferes with the binding of an SR protein.
The average length of an exon in the human genome is - 150 This leads to exon skipping in most of the SMN2 mRNAs.
bases, whereas the average length of an intron is much longer The homologous SMN gene in the mouse, where there is

356 CHAPTER 8 • Post-transcriptional Gene Control

 Spliceosome

U2 U1

~---11"-
-
1 \
- -.l...
Jr
U
2
U2AF65
A ..-yyyy. · AG
<.....:f
3~· Iu GU
1

__;L.._
7r-
3'

sranch point 3' splice Site ESE s· splice site

Cross-exon recognition complex Cross-exon recognition complex

FIGURE 8-13 Exon recognition through cooperative binding of spans an exon and activates the correct splice sites for RNA splicing.
SR proteins and splicing factors to pre-mRNA. The correct 5' GU and Note that the Ul and U2 snRNPs in this unit do not become part of the
3' AG splice sites are recognized by splicing factors on the basis oftheir same spliceosome. The U2 snRNP on the right forms a spliceosome
proximity to exons. The exons contain exonic splicing enhancers (ESEs) with the Ul snRNP bound to the 5' end of the same intron. The Ul
that are binding sites for SR proteins. When bound to ESEs, the SR snRNP shown on the right forms a spliceosome with the U2 snRNP
proteins interact with one another and promote the cooperative bound to the branch point of the downstream intron (not shown), and
binding of the Ul snRNP to the 5' splice site of the downstream intron, the U2 snRNP on the left forms a spliceosome with a Ul snRNP bound
SFl and then the U2 snRNP to the branch point of the upstream intron, to the 5' splice site of the upstream intron (not shown). Double-headed
the 65- and 35-kD subunits of U2AF to the pyrimidine-rich region and arrows indicate protein-protein interactions. [Adapted from T. Maniatis,
AG 3' splice site ofthe upstream intron, and other splicing factors (not 2002, Nature 41 8:236; see also S.M. Berget, 1995, J. Bioi. Chern. 270:2411.)
shown). The resulting RNA-protein cross-exon recognition complex

only a single copy, is essential for cell viability. Spinal mus- fungi. Discovery of the catalytic activity of self-splicing in-
cular atrophy in huma ns resu lts from homozygous muta- trans revolutionized concepts about the functions of RNA.
tions that inactivate SMNl. The low level of protein translated As discussed in Chapter 4, RNA is now kn.own to catalyze
from the small fraction of SMN2 mRNAs that are correctly peptide-bond formation during protein synthesis in ribo-
spliced is sufficient to maintain cell viability during embryo- somes. Here we discuss the probable role of group II introns,
genesis and fetal development, but it is not sufficient to now found only in mitochondrial and chloroplast DNA, in
maintain viability of spi nal cord motor neurons in childhood, the evolution of snRNAs; the functioning of group I introns
resulting in their death and the associated disease. is considered in the later section on rRNA processing.
Approximately 15 percent of the single-base-pair muta- Even though their precise sequences are not highly con-
tions that cause human genetic diseases interfere with proper served, all group II introns fo ld into a conserved, complex
exon definition. Some of these mutations occur in 5' or 3' secondary structure containing numerous stem loops (Fig-
splice sites, often resu lting in the use of nearby alternative ure 8-14a). Self-splicing by a group II intron occurs via two
"cryptic" splice sites present in the normal gene sequence. ln transesterification reactions, involving intermediates and
the absence of the normal splice site, the cross-exon recogni- products analogous to those found in nuclear pre-mRNA
tion complex recognizes these alternative sites . Other muta- splicing. The mechanistic similarities between group II intron
tions that cause abnormal splicing result in a new consensus self-splicing and spliceosomal splicing led to the hypothesis
splice-site sequence that becomes recognized in place of the that snRNAs function analogously to the stem loops in the
normal splice site. Finally, some mutations can interfere with secondary structure of group II imrons. According to this hy-
the binding of specific SR proteins to pre-mRNAs. These pothesis, snRNAs interact with 5' and 3' splice sites of pre-
mutations inhibit splicing at normal splice sites, as in the case mRNAs and with each other to produce a three-dimensional
of the SMN2 gene, and thus lead to exon skipping. • RNA structure functionally analogous to that of group II self-
splicing introns (Figure 8-14b).
An extension of this hypothesis is that introns in ancient
pre-mRNAs evolved from group II self-splicing introns
Self-Splicing Group II lntrons Provide
through the progressive loss of internal RNA structures,
Clues to the Evolution of snRNAs which concurrently evolved into trans-acting snRNAs that
Under certain nonphysio logical in vitro conditions, pure perform the same functions. Support for this type of evolu-
preparations of some RNA transcripts slowly splice out in- tionary model comes from experiments with group I! intron
trans in the absence of any protein. This observation led to mutants in which domain V and part of domain I are deleted.
the recognition that some introns are self-splicing. Two types Rl'\A transcripts containing such mutant introns are defective
of self-splicing introns have been discovered: group I introns, in self-splicing, but when RNA molecules equivalent to the
present in nuclear rRNA genes of protozoans, and group IT deleted regions are added to the in vitro reaction, self-splicing
introns, present in protein-coding genes and some rRNA and occurs. This finding demonstrates that these domains in group
tRNA genes in mitochondria and chloroplasts of plants and II introns can be trans-acting, like snRNAs.

8.1 Processing of Eukaryotic Pre-mRNA 357

(a) Group II intron (b) U sn RNAs in spliceosome processing that involves cleavage but not polyadenylation.
Specialized RNA-binding proteins that help to regulate his-
tone mRNA translation bind to the 3' end generated by this
specialized system.) Early studies of pulse-labeled adenovirus
and SV40 RNA demonstrated that the viral primary tran-
scripts extend beyond the site from which the poly(A) tail
extends. These results suggested that A residues are added to
a 3' hydroxyl generated by endonucleolytic cleavage of a
longer transcript, but the predicted downstream RNA frag-
5' 3'
- Pre-mRNA
ments never were detected in vivo, presumably because of
intron their rapid degradation. However, detection of both pre-
FIGURE 8-14 Comparison of group II self-splicing introns and
dicted cleavage products was observed in in vitro processing
the spliceosome. The schematic diagrams compare the secondary reactions performed with nuclear extracts of cultured human
structures of (a) group II self-splicing introns and (b) U snRNAs present cells. The cleavage/polyadenylation process and degradation
in the spliceosome. The first transesterification reaction is indicated by of the RNA downstream of the cleavage site occurs much
light green arrows; the second reaction, by blue arrows. The branch- more slowly in these in vitro reactions, simplifying detection
point A is boldfaced. The similarity in these structures suggests that the of the downstream cleavage prodl1ct.
spliceosomal snRNAs evolved from group II introns, with the trans-acting Early sequencing of eDNA clones from animal cells showed
snRNAs being functionally analogous to the corresponding domains in that nearly all mRNAs contain the sequence AAUAAA 10-35
group II introns. The colored bars flanking the introns in (a) and nucleotides upstream from the poly(A) tail (Figure 8-15). Poly-
(b) represent exons. [Adapted from P. A. Sharp, 1991, Science 254:663.] adenylation of RNA transcripts is virtually eliminated when
the corresponding sequence in the template DNA is mutated to
any other sequence except one encoding a closely related se-
The similarity in the mechanisms of group II intron self- quence (AUUAAA). The unprocessed RNA transcripts pro-
splteing and spliceosomal splicing of pre-mRNAs also sug- duced from such mutant templates do not accumulate in nuclei
gests that the splicing reaction is catalyzed by the snRNA, not but are rapidly degraded. Further mutagenesis studies revealed
the protein, components of spliceosomes. Although group II that a second signal downstream from the cleavage site is re-
introns can self-splice in vitro at elevated temperatures and quired for efficient cleavage and polyadenylation of most pre-
Mg2 concentrations, under in vivo conditions proteins called mRNAs in animal cells. This downstream signal is not a
maturases, which bind to group II intron RNA, are required specific sequence but rather a GU-rich or simply a U-rich re-
for rapid splicing. Maturases are thought to stabilize the pre- gion within - 50 nucleotides of the cleavage site.
cise three-dimensional interactions of the intron RNA re- Identification and purification of the proteins required for
quired to catalyze the two splicing transesterification reactions. cleavage and polyadenylation of pre-mRNA have led to the
By analogy, snRNP proteins in spliceosomes are thought to model shown in Figure 8-15. A 360-kDa cleavage and poly-
stabilize the precise geometry of snRNAs and intron nucle- adenylation specificity factor (CPSF), composed of five differ-
otides required to catalyze pre-mRNA splicing. ent polypeptides, first forms an unstable complex with the
The evolution of snRNAs may have been an important upstream AAUAAA poly(A) signal. Then at least three addi-
step in the rapid evolution of higher eukaryotes. As internal tional proteins bind to the CPSF-RNA complex: a 200-kDa
intron sequences were lost and their functions in RNA splicing heterotrimer called cleavage stimulatory factor (CStF), which
supplanted by trans-acting snRNAs, the remaining intron se- interacts with the G/U-rich sequence; a 150-kDa hetero-
quences would be free to diverge. This in turn likely facilitated tetramer called cleavage factor I (CFI); and a second het-
the evolution of new genes through exon shuffling since there erodimeric cleavage factor (CFII). An -150-kDa protein
are few constraints on the sequence of new introns generated called symplekin is thought to form a scaffold on which these
in the process (see Figures 6-18 and 6-19). It also permitted the cleavage/polyadenylation factors assemble. Finally, poly(A)
increase in protein diversity that results from alternative RNA polymerase (PAP) binds to the complex before cleavage can
splicing and an additional level of gene control resulting from occur. This requirement for PAP binding links cleavage and
regulated RNA splicing. polyadenylation, so that the free 3' end generated is rapidly
polyadenylated and no essential information is lost to exonu-
clease degradation of an unprotected 3' end.
Assembly of the large, multiprotein cleavage/polyadenyla-
3' Cleavage and Polyadenylation
tion complex around the AU-rid1 poly(A) signal in a pre-mRNA
of Pre-mRNAs Are Tightly Coupled IS analogous in many ways to formation of the transcription-
In eukaryotic cells, all mRNAs, except histone mRNAs, have preinitiation complex at the AT-rich TATA box of a template
a 3' poly(A) tail. (The major histone mRNAs are transcribed DNA molecule (see Figure 7-16). In both cases, multiprotein
from repeated genes at prodigious levels in replicating cells complexes assemble cooperatively through a network of specific
during the S phase. They undergo a special form of 3'-end protein-nucleic acid and protein-protein interactions.

358 CHAPTER 8 • Post-transcriptional Gene Control

 Poly(A) Poly(A) FIGURE 8-15 Model for cleavage and polyadenylat ion o f
signal Poly(A) site signal pre-mRNAs in mammalian cells. Cleavage and polyadenylation

5' e>o-----!I~~~U~AAA~I--*L--~ 3' specificity factor (CPSF) binds to the upstream AAUAAA poly(A) signal.
CStF interacts with a downstream GU- or U-rich sequence and with
CPSF, CStF, CFI, CFII i p,~mRNA bound CPSF, forming a loop in the RNA; binding of CFI and CFII helps
stabilize the complex. Binding of poly(A) polymerase (PAP) then
stimulates cleavage at a poly(A) site, which usually is 10-35 nucleotides
3' of the upstream poly(A) signal. The cleavage factors are released, as
CPSF is the downstream RNA cleavage product, which is rapidly degraded.
Bound PAP then adds - 12 A residues at a slow rate to the 3'-hydroxyl
group generated b y the cleavage reaction. Binding of poly(A)-binding
protein II (PABPII) to the initial short poly(A) tail accelerates the rate of
addition by PAP. After 200-250 A residues have been added, PABPII
signals PAP to stop polym erization.

Following cleavage at the poly(A) site, polyadenylation pro-

ceeds in two phases. Add ition of the first 12 or so A residues
5'
occurs slowly, followed by rapid addition of up to 200-250
more A residues. The rapid phase requires the binding of mul-
tiple copies of a poly(A)-binding protein containing the RRM
motif. This protein is designated PABPII to distinguish it from

1
3'
the poly(A)-bind ing protein present in the cytoplasm. PABPII
c....... binds to the short A tail initially added by PAP, stimulating the
rate of polymerization of additional A residues by PAP, result-
ing in the fast phase of polyadenylation (Figur.e 8-15). PABPII is
also responsible for signaling poly(A) polymerase to terminate
polymerization when the poly(A) tail reaches a length of 200-
250 residues, although the mechanism for controlling the length
of the ta il is not yet understood. Bi nding of PABPII to the
poly(A) tail is essential for mRNA export into the cytoplasm.

p~
J' + ATP pp.
Slow
polyadenylation Nuclear Exonucleases Degrade RNA
+ CStF, CFI, CFII I That Is Processed Out of Pre-mRNAs
Because the h uman genome contains long introns, only - 5
percent of the nucleotides that are polymerized by RNA
polymerase II during transcription are retained in mature,
5' A
processed mRNAs. Although this process appears ineffi-
cient, it probably evolved in mu lticellular organisms because
PABPIIi the process of exon shuffling facilitated the evolution of new
genes in organisms with long introns (Chapter 6). The in-
trons that are spliced out and the region downstream from
the cleavage and polyadenylation site are degraded by nu-
5' clear exoribonucleases that hydrolyze one base at a time
from either the 5 ' or 3' end of an RNA strand .
As mentioned earlier, the 2',5'-phosphodiester bond in
PABPIIF ATP Rapid excised introns is hydrolyzed by a debranching enzyme,
polyadenylation
PP; yielding a linear molecule with unprotected ends that can be
attacked by exonucleases (see Figure 8-11 ). The predomi-
nant nuclear decay pathway is 3'-4S' hydrolysis by 11 exo-
nucleases that associate with one another in a large protein
5' H 3' complex called the exosome. O ther proteins in the complex
include RNA helicases that disr upt base pairing and RNA-
protein interactions that wou ld o therwise impede the exonu-
cleases. Exosomes also function in the cytoplasm, as discussed

8.1 Processi ng of Eukaryotic Pre-mRNA 359

later. In addition to introns, the exosome appears to degrade
pre-mRNAs that have not been properly spliced or poly- come associated with the CTD, where they are poised to inter-
adenylated. It is not yet clear how the exosome recognizes act with the nascent prc-mRNA as it emerges from the surface
improperly processed pre-mRNAs. But in yeast cells with of the polymerase.
temperature-sensitive mutant poly(A) polymerase (Figure • Five different snRNPs interact via base pairing with one
8-15), pre-mRNAs are retained at their sites of transcription another and with pre-mRNA to form the spliceosome (see
in the nucleus at the nonpermissive temperature. These ab- Figure 8-11 ). This very large ribonucleoprotein complex
normally processed pre-mRNAs are released in cells with a catalyzes two transesterification reactions that join two ex-
second mutation in a subunit uf the exosome found only in ons and remove the intron as a lariat structure, wh ich is sub-
nuclear and not in cytoplasmic exosomes (PM-Scl; 100 kD sequently degraded (see Figure 8-8).
in humans). Also, exosomes are found concentrated at sites
• SR proteins that bind to exonic splicing enhancer sequences in
of transcription in Drosophila polytene chromosomes,
exons are critical in defining exons in the large pre-mRNAs of
where they are associated with RNA polymerase II elonga-
higher organisms. A network of interactions between SR pro- , '
tion factors. These results suggest that the exosome partici-
teins, snRNPs, and splicing factors forms a cross-exon recogni-
pates in an as yet poorly understood quality-control mechanism
tion complex that specifies correct splice sites (see Figure 8- 13 ).
that recognizes aberrantly processed pre-mRNAs, prevent-
ing their export to the cytoplasm and ultimately leading to • The snRNAs in the spliccosorn.e are thought to have an
their degradation. overall tertiary structure similar to that of group II self-splicing
To avoid being degraded by nuclear exonucleases, nascent introns.
transcripts, pre-mRNA processing intermediates, and mature • For long transcription units in higher organisms, splicing of
mRNAs in the nucleus must have their ends protected. As dis- exons usually begins as the pre-mRNA is still being formed.
cussed above, the 5' end of a nascent transcript is protected by Cleavage and polyadenylation to form the 3' end of the mRNA
addition of the 5' cap structure as soon as the 5' end emerges occur after the poly(A) site is transcribed.
from the polymerase. The 5' cap is protected because it is
• In most protein-coding genes, a conserved AAUAAA poly(A)
bound by a nuclear cap-binding complex, which protects it
signal lies slightly upstream from a poly(A) site where cleavage
from 5' exonucleases and also functions in export of mRNA
and polyadenylation occur. A GU- or U-rich sequence down-
to the cytoplasm. The 3' end of a nascent transcript lies within
stream from the poly(A) site contributes to the efficiency of
the RNA polymerase and thus is inaccessible to exonucleases
cleavage and polyadenylation.
(see figure 4-12). As discussed previously, the free 3' end gen-
erated by cleavage of a pre-mRNA downstream from the A multiprotein complex that includes poly(A) polymerase
poly(A) signal is rapidly polyadenylated by the poly(A) (PAP) carries out the cleavage and polyadenylation of a pre-
polymerase associated with the other 3' processing factors, mRNA. A nuclear poly(A)-binding protein, PABPII, stimu-
and the resulting poly(A) rail is bound by PABPII (Figure lates addition of A residues by PAP and stops addition once
8-15). This tigh~ coupling of cleavage and polyadenylation the poly(A) tail reaches 200-250 residues (see Figure 8-15).
protects the 3' end from exonuclease attack. • Excised introns and RNA downstream from the cleavage/
poly(A) site arc degraded primarily by exosomes, multipro-
tcin complexes that contain eleven 3 '-75' exonucleases as
well as RNA helicases. Exosomes also degrade improperly
KEY CONCEPTS of Section 8.1 processed pre-mRNAs.
Processing of Eukaryotic Pre-mRNA
In the nucleus of eukaryotic cells, pre-mRNAs are associated
with hnRNP proteins and processed by 5' capping, 3' cleavage
and polyadcnylation, and splicing before being transported to 8.2 Regulation of Pre-mRNA Processing
the cytoplasm (see Figure 8-2). Now that we've seen how pre-mRNAs are processed into ma-
Shortly after transcription initiation, capping enzymes asso- ture, functional mRNAs, we consider how regulation of this
ciated with the carboxyl-terminal domain (CTD) of RNA process can contribute to gene control. Recall from Chapter 6
polymerase II, which is phosphorylated multiple rimes at ser- that higher eukaryotes contain both simple and complex tran-
ine 5 of the heptapeptide repeat by TFIIH during transcription scription units encoded in their DNA. The primary transcripts
initiation, add the 5' cap to the nascent transcript. Other RNA- produced from the former contain one poly(A) site and exhibit
processing factors involved in RNA splicing, 3' cleavage, ami only one pattern of RNA splicing, even if multiple introm are
polyadenylation associate with the CTD when it is phosphory- present; thus simple transcription units encode a single mRNA.
lated at serine 2 of the hcptapeptidc repeat, increasing the rate In contrast, the primary transcripts produced from complex
of transcription elongation. Consequently, transcription does transcription units (-92-94 % of all human transcription
not proceed at a high rate until RNA-processing factors be- units) can be processed in alternative ways to yield different
mRNAs that encode distinct proteins (see Figure 6-3).

360 CHAPTER 8 • Post-transcriptional Gene Control

 Alternative Splicing Generates Transcripts A Cascade of Regulated RNA Splicing Controls
with Different Combinations of Exons Drosophila Sexual Differentiation
The discovery that a large fraction of transcription units in One of the earliest examples of regulated alternative splicing
higher organisms encode alternatively spliced mRNAs and of pre-mRNA came from studies of sexual differentiation in
that differently spliced mRNAs are expressed in different cell Drosophila. Genes required for normal Drosophila sexual dif-
types revealed that regulation of RNA splicing is an impor- ferentiation were first characterized by isolating Drosophila
tant gene-control mechanism in higher eukaryotes. Although mutants defective in the process. When the proteins encoded
many examples of cleavage at alternative poly(A) sites in pre- by the wild-type genes were characterized biochemically, two
mRNA!. an: known, alternative splicing of different exons is of them were found to regulate a cascade of alternative RNA
the more common mechanism for expressing different pro- splicing in Drosophila embryos. More recent research has
teins from one complex transcnption unit. In Chapter 4, for provided insight into how these proteins regulate RNA
example, we mentioned that fibroblasts produce one type of processing and ultimately lead to the creation of two different
the extracellular protein fibronectin, whereas hepatocytes sex-specific transcriptional repressors that suppress the devel-
produce another type. Both fibronectin isoforms are encoded opment of characteristics of the opposite sex.
by the same transcription unit, which is spliced differently in The Sxl protein, encoded by the sex-lethal gene, is the
... the two cell types to yield two different mRNAs (see Figure first protein to act in the cascade (Figure 8-16). The Sxl pro-
4-16). In other cases, alternative processing may occur simul- tein is present only in female embryos. Early in development,
taneously in the same cell type in response to different devel- the gene is transcribed from a promoter that functions only in
opmental or environmental signals. First we discuss one of female embryos. Later in development, this female-specific
the best-understood examples of regulated RNA processing promoter is shut off and another promoter for sex-lethal be-
and then we briefly consider the consequences of RNA splic- comes active in both male and female embryos. However, in
ing in the development of the nervous system. the absence of early Sxl protein, the sex-lethal pre-mRNA in

Pre-mRNAs mRNAs

<j> / -i 2 I 4 ~ -+ • Sxl protein

'I§!/
(a) sxl

' v / 'v /

------------------------
(b) tra 5~ ~
' v /

Rbp1 + Tca2'~' ---

Dsx protein
(c) dsx 4
· Dsx protein

FIGURE 8 - 16 Cascade of regulated splicing that controls sex Rbpl and Tra2, activates splicing between exons 3 and 4 and cleavage/
determination in Drosophila embryos. For clarity, only the exons polyadenylation(An)at the 3' end of exon 4 in dsx pre-mRNA in female
(boxes) and introns (black lines) where regulated splicing occurs are embryos. In male embryos, which lack functional Tra, the SR proteins
shown. Splicing is indicated by red dashed lines above (female) and do not bind to exon 4, and consequently exon 3 is spliced to exon S.
blue dashed lines below (male) the pre-mRNAs. Vertical red lines in The distinct Dsx proteins produced in female and male embryos as the
exons indicate in-frame stop codons, which prevent synthesis of result of this cascade of regulated splicing repress transcription of
functional protein. Only female embryos produce functional Sxl genes required for sexual differentiation of the opposite sex. [Adapted
protein, which represses splicing between exons 2 and 3 in sxl pre- from M. J. Moore et al., 1993, in R. Gesteland and J. Atkins, eds., The RNA World,
mRNA (a) and between exons 1 and 2 in tra pre-mRNA (b). (c) In Cold Spring Harbor Press, pp. 303- 357.]
contrast, the cooperative binding ofTra protein and two SR proteins,

8.2 Regulation of Pre -mRNA Processing 361

male embryos is spliced to produce an mRNA that contains
a stop codon early in the sequence. The net result is that
male embryos produce no functional Sxl protein either earl}
or later in development.
ln contrast, the Sxl protein expressed in early female em-
bryos directs splicing of the sex-lethal pre-mRNA so that a
functional sex-lethal mRNA is produced (Figure 8-16a). Sxl
accomplishes this by binding to a sequence in the pre-mRNA
near the 3' end of the intron between exon 2 and exon 3,
thereby blocking the proper association of U2AF and U2 5' 3
snRNP (Figure 8-1 I). As a consequence, the Ul snRNP bound
to the 3' end of exon 2 assembles into a spliceosome with U2
snRNP bound to the branch point at the 3' end of the intron FIGURE 8-17 Model of splicing activation by Tra protein and the
between exons 3 and 4, leading to splicing of exon 2 to exon SR proteins Rbpl and Tra2. In female Drosophila embryos, splicing of
4 and skipping of exon 3. The binding site for Sxl in the Sxl exons 3 and 4 in dsx pre-mRNA is activated by binding ofTrafTra2/
pre-mRNA is called an intronic SfJlicing silencer because of its Rbpl complexes to six sites in exon 4. Because Rbpl and Tra2 cannot
location in an intron and its function in blocking, or "silenc- bind to the pre-mRNA in the absence onra, exon 4 is skip ped in male
ing," use of a splice site. The resulting female-specific sex-lethal embryos. See the text for discussion. An= polyadenylation. [Adapted
mRNA is translated into functional Sxl protein, which rein- from T. Maniatis and B. Tasic, 2002. Nature 418:236.]
forces its own expression in female embryos by continuing to
cause skipping of exon 3. The absence of Sxl protein in male for female development. Conversely, the female Dsx protein
embryos allows the inclusion of exon 3 and, consequently, of represses transcription of genes required for male development.
the stop codon that prevents translation of functional Sxl Figure 8- I 7 illustrates how the Tra/Tra2/Rbpl complex is
protein. thought to interact with double-sex (dsx) pre-mRNA. Rbpl
Sxl protein also regulates alternative RNA splicing of the and T ra2 are SR proteins, but they do not interact with exon 4
transformer gene pre-mRNA (Figure 8-16b). In male embryos, in the absence of the Tra protein. Tra protein interacts with
where no Sxl is expressed, exon l is spliced to exon 2, which Rbpl and Tra2, resulting in the cooperative binding of all
contains a stop codon that prevents synthesis of a functional three proteins to six exonic splicing enhancers in exon 4. The
transformer protein. ln female embryos, however, binding of bound Tra2 and Rbpl proteins then promote the binding of
Sxl protein to an intronic splicing silencer at the 3' end of the U2AF and U2 snRNP to the 3' end of the intron between
inrron between exons 1 and 2 blocks binding of U2AF ar this exons 3 and 4, just as other SR proteins do for constitutively
site. The interaction of Sxl with transformer pre-mRNA is me- spliced exons (see Figure 8-13). The Tra/Tra2/Rbp1 complexes
diated by two RRM domains in the protein (see Figure 8-5). may also enhance binding of the cleavage/polyadenylation
When Sxl is bound, U2AF binds to a lower-affinity site farther complex to the 3' end of cxon 4.
3' in the pre-mRNA; as a result exon 1 is spliced to this alter-
native 3' splice c;ite, eliminating exon 2 with irs stop codon.
Splicing Repressors and Activators Control
The resulting female-specific transformer mRNA, which con-
tains additional constitutively spliced exons, is translated into Splicing at Alternative Sites
functional Transformer (Tra) protein. As is evident from Figure 8-16, the Drosophila Sxl protein
Finally, Tra protein regulates the alternative processing and Tra protein have opposite effects: Sxl prevents splicing,
of pre-mRNA transcribed from the double-sex gene (Figure causing exons to be skipped, whereas Tra promotes splicing.
8 16c). In female embryos, a complex ofT ra and two consti- The action of similar proteins may explain the cell-type-specific
tutively expressed proteins, Rbpl and Tra2, directs spl icing expression of fibronectin isoforms in humans. For instance,
of exon 3 to exon 4 and also promotes cleavage/polyaden- an Sxl-like splicing repressor expressed in hepatocytes might
ylation at the alternative poly(A) site at the 3' end of exon bind to splice sites for the EIIIA and EIIIB exons in the fi -
4-leading to a short, female-specific version of the Dsx pro- bronectin pre-mRNA, causing them to be skipped during
tein. In male embryos, which produce no Tra protein, exon RNA splicing (see Figure 4-16). Alternatively, a Tra-like
4 is skipped, so that exon 3 is spliced to exon 5. Exon 5 is splicing activator expressed in fibroblasts might activate the
constitutively spliced to exon 6, which is polyadenylated at splice sites associated with the fibronectin EIIIA and ElliB
its 3' end-leading to a longer, male-specific version of the Dsx exons, leading to inclusion of these exons in the mature
protein. The RNA sequence ro which Tra binds in exon 4 is mRNA. Experimental examination in some systems has re-
called an exomc splicing enhancer since it enhances splicing at vealed that inclusion of an exon in some cell types versus
a nearby splice site. skipping of the same exon in other cell types results from the
As a result of the cascade of regulated RNA processing de- combined influence of several splicing repressors and en-
picted in Figure 8-16, different Dsx proteins are expressed in hancers. RNA binding sites for repressors, usually hnRNP
male and female embryos. The male Dsx protein is a transcrip- proteins, can also occur in exons, where they are called exonic
tional repressor that inhibits the expression of genes required splicing silencers. And binding sites for splicing activators,

362 CHAPTERs • Post-transcriptional Gene Control

 usually SR proteins, can also occur in introns, where they arc (a)
called intronic splicing enhancers.
Alternative splicing of cxons is especially common in the
nervous system, generating multiple isoforms of many proteins
required for neuronal development and function in both verte-
brates and invertebrates. The primary transcripts from these ./"' ~~'7'1Wl':mY
genes often show complex splicing patterns that can generate Apical Auditory
hair cell nerve Basal
several different mRNAs, with different spliced forms expressed (50 Hz) cell body Auditory hair cell
. ' in different anatomical locations within the cenrr::~l nervous sys- nP.rve (5000 Hz)

tem. We consider two remarkable examples that illustrate the

(b)
critical role of this process in neural function. N
Exterior
Expression of K+ -Channel Proteins in Vertebrate Hair Cells In
the inner car of verteb rates, individual "hair cells," which
are ciliated neurons, respond most strongly to a specific fre-
quency of sound. Cells tuned to low frequency (-50 Hz) arc
Cytosol
found at one end of the tubular cochlea that makes up the inner
car; cells responding to high frequency (-5000Hz) are found
at the other end (Figure 8-18a). Cells in between the ends
respond to a gradient of frequencies between these extremes.
One component in the tuning of hair cells in reptiles and
birds is the opening of K + ion channels in response to in-
creased intracellular Ca 2 concentrations. The Cal+ concen-
tration at which the channel opens determines the frequency
with which the membrane potential oscillates and hence the
frequency to which the cell is tuned. FIGURE 8-18 Role of alternative splicing in the perception of
The gene encoding this Ca2 + -activated K channel is ex- sounds of different frequency. (a) The chicken cochlea, a 5-mm-long
pressed as multiple, alternatively spliced mRNAs. The vari- tube, contains an epithelium of auditory hair cells that are tuned to a
ous proteins encoded by these alternative mRNAs open at gradient of vibrational frequencies from SO Hz at the apical end (left)
different Ca 2 + concentrations. Hair cells with differcnr re- 2
to 5000 Hz at the basal end (right). (b) The Ca -activated K channel
sponse frequencies express different isoforms of the channel contains seven transmembrane a helices (S0-56), which associate to
protein depending on their position along the length of the form the channel. The cytosolic domain, which includes four hydropho-
cochlea. The sequence variation in the protein is very com- bic regions (57-510), regulates opening of the channel in response
plex: there are at least eight regions in the mRNA where alter- to Ca 2 +. lsoforms of the channel, encoded by alternatively spliced
native exons are utilized, permitting the expression of 576 mRNAs produced from the same primary transcript, open at different
possible isoforms (Figure 8-18b). PCR analysis of mRNAs Ca 2 concentrations and thus respond to different frequencies.
from individ ual "hair cells has shown that each hair cell ex- Red numbers refer to regions where alternative splicing produces
different amino acid sequences in the various isoforms. [Adapted from
presses a mixture of different alternative Ca 2 +-activated K+-
K. P. Rosenblatt et al., 1997, Neuron 19:1061.]
channel mRNAs, with different forms predominating in
different cells according to their position along the cochlea.
This remarkable arrangement suggests that splicing of the
Ca 2 -activated K+-channel pre-mRNA is regulated in re- Many examples of genes similar to those that encode
sponse to extracellular signals that inform the cell of its posi- the cochlear K+-channel have been observed in verte-
tion along the cochlea. brate neurons; alternatively spliced mRNAs co-expressed
Other studies demonstrated that splicing at one of the al- from a specific gene in one type of neuron are expressed at
ternative splice sites in the Ca 2 -activated K+ -channel pre- different relative concentrations in different regions of the
mRNA in the rat is suppressed when a specific protein kinase central nervous system. Expansions in the number of micro-
is activated by neuron depolarization in response to synaptic satellite repeats within the transcribed regions of genes ex-
activity from interacting neurons. This observation raises the pressed in neurons can cause an alteration in the relative
possibility that a splicing repressor specific for this site may be concentrations of alternatively spliced mRNAs transcribed
activated when it is phosphorylated by this protein kinase, from multiple genes. In Chapter l'l, we discussed how back-
whose activity in turn is regulated by synaptic activity. Since ward slippage during DNA replication can lead to expansion
hnRNP and SR proteins are extensively modified by phos- of a microsatellite repeat (see Figure 6-5). At least 14 different
phorylation and other post-translational modifications, it types of neurological disease result from expansion of micro-
seems likely that complex regulation of alternative RNA splic- satellite regions within transcription units expressed in neu-
ing through post-translational modifications of splicing fac- rons. The resulting long regions of repeated simple sequences
tors plays a significant role in modulating neuron function. in nuclear RNAs of these neurons result in abnormalities in

8.2 Regulation of Pre-mRNA Processing 363

 CAA TAA
~
~:::::t:t:
. ~ F:'f=
l ::::;:
1 ::t-:!~~
.UJ~---+~-~
apoB
3'
gene !
i =

apoB
mRNA 5'
...
CAA
...
UAA
An 5'
CAA-+UAA UAA

~ ~
1 4536 1 2152

Proteins NH21 d cooH NH 2 ll--~-~....j~ COOH

ApoB-100 ApoB-48
FIGURE 8·19 RNA editing of apo-8 pre-mRNA. The apoB mRNA receptors on cell membranes. In the apo-8 mRNA produced in the
produced in the liver has the same sequence as the exons in the intestine, the CAA codon in exon 26 is edited to a UAA stop codon.
primary transcript. Th is mRNA is translated into apoB-1 00, which has As a result, intestinal cells produce apoB-48, which corresponds to the
two functional domains: anN-terminal domain (green) that associates N-terminal domain of apoB-100. [Adapted from P. Hodges and J. Scott,
,
with lipids and a (-terminal domain (orange) that binds to LDL 1992, Trends Biochem. Sci. 17:77.]

the relative concentrations of alternatively spliced mRNAs. RNA Editing Alters the Sequences
For example, the most common of these types of diseases, of Some Pre-mRNAs
myotonic dystrophy, is characterized by paralysis, cognitive
impairment, and personality and behavior disorders. Myo- In the mid-1980s, sequencing of numerous eDNA clones and
tonic dystrophy results from increased copies of either CUG corresponding genomic DNAs from multiple organisms led
repeats in one transcript, in some patients, or CCUG repeats to the unexpected discovery of another type of pre-mRNA
in another transcript, in other patients. When the number of processing. In this type of processing, called RNA editing, the
these repeats increases to 10 or more times the normal num- sequence of a pre-mRNA is altered; as a result, the sequence
ber of repeats in these genes, abnormalities are observed in of the corresponding mature mRNA differs from the exons
the level of two hnRNP proteins that hind to these repeated encoding it in genomic DNA.
sequences. This probably results because these hnRNPs are RNA editing is widespread in the mitochondria of proto-
bound by the abnormally high concentration of this RNA zoans and plants and also in chloroplasts. In the mitochon-
sequence in the.nuclei of neurons in such patients. The ab- dria of certain pathogenic trypanosomes, more than half the
normal concentrations of these hnRNP proteins are thought sequence of some mRNAs is altered from the sequence of the
to lead to alterations in the rate of splicing of different alter- corresponding primary transcripts. Additions and deletions
native splice sites in multiple pre-mRNAs normally regulated of specific numbers of U's fo llows temp lates provided by
by these hnRNP proteins. • base-paired short "guide" RNAs. These RNAs are encoded
by thousands of mini-mitochondrial DNA circles catenated
to many fewer large mitochondrial DNA molecules. The rea-
Expression of Dscam lsoforms in Drosophila Retinal Neurons son for this baroque mechanism for encoding mitochondrial
The most extreme example of regulated alternative RNA proteins in such protOzoans is not clear. But this system does
processing yet uncovered occurs in expression of the Dscam represent a potential target for drugs to inhibit the complex
gene in Drosophila. Mutations in this gene interfere with the processing enzymes essential to the microbe that do not exist
normal synaptic connections made between axons and den- in the cells of their human or other vertebrate hosts.
drites during fly development. Analysis of the Dscam gene In higher eukaryotes, RNA editing is much rarer, and
showed that it contains 95 alternatively spliced exons that thus far, only single-base changes have been observed. Such
could be spliced to generate 38,016 possible isoforms! Re- minor editing, however, turns out to have significant func-
cent results have shown that Drosophila mutants with aver- tional consequences in some cases. An important example of
sion of the gene that can be spliced in only about 22,000 RNA editing in mamma ls involves the apoB gene. This gene
different ways have specific defects in connectivity between encodes two alternative forms of a c;erum protein central to
neurons. These results indicate that expression of most of the uptake and transport of cholesterol. Consequently, it is
the possible Dscam isoforms through regulated RNA splic- important in the pathogen ic processes that lead to athero-
ing helps to specify the tens of millions of different specific sclerosis, the arterial disease that is the major cause of death
synapti<.. connections between neurons in the Drosophila in the developed world. The apoB gene expresses both the
brain. In other words, the correct w iring of neurons in the serum protein apolipoprotein B-100 (apoB-100) in hepato-
brain requires regulated RNA splicing. cytes, the major cell type in the liver, and apoB-48, expressed

364 CHAPTERs • Post-transcriptional Gene Control

 in intestinal epithelial cells. The ~240-kDa apoB-48 corre- The nuclear envelope is a double membrane that separates
sponds to theN-terminal region of the ~500-kDa apoB-100. the nucleus from the cytoplasm (see Figure 9-32). Like the
As we detail in Chapter 10, both apoB proteins are compo- plasma membrane surrounding cells, each nuclear mem-
nents of large lipoprotein complexes that transport lipids in brane consists of a water-impermeable phospholipid bilayer
the serum. However, only low-density lipoprotein (LDL) and multiple associated proteins. mRNPs and other macro-
complexes, which contain apoB-100 on their surface, deliver molecules including tRNAs and ribosomal subunits traverse
cholesterol to body tissues by binding to the LDL receptor the nuclear envelope through nuclear pores. This section will
present on all cells. focus on the export of mRNPs through the nuclear pore and
The cell-type-specific expression of tht> two forms of apoB the mechanisms that allow some lt:vcl of regulation of this
results from editing of apoB pre-mRNA so as to change the step. Transport of other cargoes across the nuclear pore is
nucleotide at position 6666 in the sequence from a C to a U. discussed in Chapter 13.
This alteration, which occurs only in intestinal cells, converts
a CAA codon for glutamine to a UAA stop codon, leading to
Macromolecules Exit and Enter the Nucleus
synthesis of the shorter apoB-48 (Figure 8-19). Studies with
the partially purified enzyme that performs the post-tran- Through Nuclear Pore Complexes
scriptional deamination of C 6666 to U shows that it can recog- Nuclear pore complexes (NPCs) arc large, symmetrical struc-
nize and edit an RNA as short as 26 nucleotides with the tures composed of multiple copies of approximately 30 dif-
sequence surrounding C 6666 in the a{JOB primary transcript. ferent proteins called nucleoporins. Embedded in the nuclear
envelope, NPCs are cylindrical in shape with a diameter of
- 30 nm (Figure 8-20a). A special class of nucleoporins called
FG-nucleoporins line the central channel through the NPC.
KEY CONCEPTS of Section 8.2 FG-nucleoporins form a semi-permeable barrier that allows
small molecules to diffuse freely, but restricts the passage of
Regulation of Pre-mRNA Processing larger molecules. The FG-nucleoporin globular domain an-
Because of alternative splicing of primary transcripts, the chors the protein in the NPC scaffold. From these anchors,
use of alternative promoters, and cleavage at different poly(A) long random-coil stretches extend into the channel. These ex-
sites, different mRNAs may be expressed from the same gene tensions are composed of hydrophilic amino acid sequences
in different cell types or at different developmental stages (sec punctuated by hydrophobic PG-repeats, short sequences rich
Figure 6-3 and Figure 8-16). in hydrophobic phenylalanine (F) and glycine (G). These FG-
Alternative splicing can be regulated by RNA-binding pro- rcpeat domains form a cloud of polypeptide chains in the
teins that bind to specific sequences near regulated splice central channel and extending into the nucleoplasm and cyto-
plasm, effectively limiting the free diffusion of macromole-
·. sites. Splicing repressors may sterically block the binding of
cules across the channel. Water, ions, metabolites, and small
splicing factors to specific sites in pre-mRNAs or inhibit their
function. Splicing activators enhance splicing by interacting globular proteins up to - 40-60 kDa can diffuse through the
with splicing factors, thus promoting their association with a cloud of FG-repeat domains. However, the FG-domains in
regulated splice.site. The RNA sequences bound by splicing the central channel form a barrier restricting the diffusion of
repressors are called intronic or exonic splicing silencers, de- larger macromolecules between the cytoplasm and nucleus.
pending on their location in an intron or exon. RNA se- Proteins and RNPs larger than ~40-60 kDa must be se-
quences bound by splicing activators arc called intronic or lectively transported across the nuclear envelope with the as-
exonic splicing enhancers. sistance of soluble transporter proteins that bind them and
also interact reversibly with the FG-repeats of FG-nucleoporins.
In RNA editing the nucleotide sequence of a pre-mRNA is As a consequence of these reversible interactions with FG-
a ltered in the nucleus. In vertebrates, this process is fairly domains, the transporter and its stably bound cargo can be
rare and entails deamina~ion of a single base in the mRNA passed from FG-domain to FG-domain, allowing them both
sequence, resulting in a change in the amino acid specified by to diffuse down a concentration gradient from the nucleus to
the corresponding codon and production of a functionally the cytoplasm.
different protein (see Figure 8-19). mRNPs are transported through the NPC by the mRNP
exporter, a heterodimer consisting of a large subunit, called
nuclear export factor 1 (NXF1 ), and a small subunit, nuclear
export transporter 1 (NXT1) (Figure 8-20b) . NXFl binds
8.3 Transport of mRNA Across nuclear mRNPs through associations with both RNA and
other proteins in the mRNP complex. One of the most im-
the Nuclear Envelope
portant of these is REF (RNA export factor), a component
Fully processed mRNAs in the nucleus remain bound by of the exon-junction complexes discussed earlier, which is
hnRNP proteins in complexes referred to as nuclear mRNPs. bound approximately 20 nucleotides 5' to each cxon-exon
Before an mRNA can be translated into its encoded protein, junction (see Figure 8-21 ). The NXF L/NXT1 mRNP exporter
it must be exported out of the nucleus into the cytoplasm. also associates with SR proteins bound to cxonic splicing

8.3 Transport of mRNA Across the Nuclear Envelope 365

 (a) Cytosol (b) Cytosol
FG Nucleoporins ~mRNP
Direction of t
Central
channel
Cytoplasmic
fi laments
tra nsport I
Outer nuclear
Outer nuclea r
Core scaffold membrane membrane ""'
/
Outer ri ng

REF, ot her
'\Inner nuclear
adapter membrane
Lumen Luminal Nuclear
proteins
ring basket

Nucleus Nucleus

FIGURE 8 -20 Model of transporter passage through an NPC. in pu rple. The FG-domains of FG-nucleop'orins (tan worms) have an
(a) Diagram of NPC structure. The nuclear envelope lipid bilayers are extended, random-coil conformation that forms a molecular cloud of
represented in green. Transmembrane proteins represented in red continuously moving random-coil polypeptide. (b) Nuclear transport-
form a luminal ring around which the nuclear envelope membrane folds ers (NXF 1/NXT1) have hydrophobic regions on their surface that bind
to form the inner and outer nuclear membranes. Transmembrane reversibly to the FG-domains in the FG-nucleoporins. As a conse-
domains of these proteins are connected to globular domains on the quence, they can penetrate the molecular cloud in the NPC central
inside of the pore to which the other nucleoporins bind, forming the channel and diffuse in and out of the nucleus. [Adapted from D. Grunwald,
core scaffold. The globular domains of the FG-nudeoporins are shown R. H. Singer, and M. Rout, 2011 , Nature 475:333.)

enhancers. Thus SR proteins associated with exons function Protein filaments extend from the core scaffold into the
to direct both the splicing of pre-mRNAs and the export of nucleoplasm forming a "nuclear basket " (see Figure 8-20a).
fully processed mRNAs through NPCs to the cytoplasm. Protein filaments also extend into the cytoplasm. T hese fila-
mRNPs are probably bound along their length by multiple ments assist in mRNP export. Gle2, an adapter protein that
NXFl/NXTl mRNP exporters, which interact with the FG- reversibly binds both NXFl and a nucleoporin in the nuclear
domains of FG-nucleoporins to faci litate export of mRNPs basket, brings nuclear mRNPs to the pore in preparation for
through the NPC central channel (Figure 8-20b). export. A nucleoporin in the cytoplasmic filaments of the NPC

CBC

Nucleus

Cytoplasm

FIGURE 8-21 Remodeling of mRNPs during

nuclear export. Some mRNP proteins (rectan-
gles) dissociate from nuclear mRNP complexes
before expurllhrough an NPC. Some (ovals) are ciF4E
exported through the NPC associated with the
mRNP but dissociate in the cytoplasm and are
shuttled back into the nucleus through an NPC.
In the cytoplasm, translation initiation factor
eiF4E replaces CBC bound to the 5' cap and
PABPI replaces PABPII.

366 CHAPTER 8 • Post-transcriptional Gene Control

binds an RNA helicase (Dbp5) that functions in the dissocia- plasm, where they dissociate. As a result, the direction of
tion of NXFl/NXTl and other hnRNP proteins from the mRNP export may be driven by simple diffusion down a
mRNP as it reaches the cytoplasm. concentration gradient of the transport-competent mRNP
In a process called mRNP remodeling, the proteins associ- exporter-mRNP complex across the NPC, from high in the
ated with an mRNA in the nuclear mRNP complex are ex- nucleus to low in the cytoplasm.
changed for a different set of proteins as the mRNP is
transported through the NPC (Figure 8-21). Some nuclear Nuclear Export of Balbiani Ring mRNPs The larval salivary
mRNP proteins dissociate early in transport, remaining in the glands of the insect Chironomous tentans provide a good
nucleus to bind to newly synthesized nascent pre-mRNA. model system for electron mil:roscopy studies of the forma-
Other nuclear mRNP proteins remain with the mRNP com- tion of hnRNPs and their export through NPCs. In these
plex as it traverses the pore and do not dissociate from the larvae, genes in large chromosomal puffs called Balbiani
mRNP until the complex reaches the cytoplasm. Proteins in rings are abundantly transcribed into nascent pre-mRNAs
this category include the NXFl/NXTl mRNP exporter, cap- that associate with hnRNP proteins and are processed into
binding complex (CBC) bound to the 5 ' cap, and PABPII coiled mRNPs with a final mRNA length of -75 kb (Figure
bound to the poly(A) tail. They dissociate from the mRNP on 8-23a, b). These giant mRNAs encode large glue proteins
the cytoplasmic side of the NPC through the action of the that adhere the developing larvae to a leaf. After processing
Dbp5 RNA helicase that associates with cytoplasmic NPC fila- of the pre-mRNA in Balbiani ring hnRNPs, the resulting
ments, as discussed above. These proteins are then imported mRNPs move through nuclear pores to the cytoplasm. Elec-
back into the nucleus as discussed for other nuclear proteins in tron micrographs of sections of these cells show mRNPs that
Chapter 13, where they can function in the export of another appear to uncoil during their passage through nuclear pores
mRNP. In the cytoplasm, the cap-binding translation initiation and then bind to ribosomes as they enter the cytoplasm. This
factor eiF4E replaces CBC bound to the 5' cap of nuclear uncoiling is probably a consequence of the remodeling of
mRNPs (Figure 4-24). ln vertebrates, the nuclear poly(A)- mRNPs as the result of phosphorylation of mRNP proteins
binding protein PABPII is replaced with the cytoplasmic by cytoplasmic kinases and the action of the RNA helicase
poly(A)-binding protein PABPI (so named because it was dis- associated with NPC cytoplasmic filament~, as discussed in
covered before PABPII). Only a single PABP is found in bud- the previous section. The observation that mRNPs become
ding yeast, in both the nucleus and the cytoplasm. associated with ribosomes during transport indicates that
the 5' end leads the way through the nuclear pore complex.
Yeast SR Protein Studies of S. cerevisiae indicate that the di- Detailed electron microscopic studies of the transport of Bal-
rection of mRNP export from the nucleus into the cytoplasm biani ring mRNPs through nuclear pore complexes led to the
is controlled by phosphorylation and dephosphorylation of model depicted in Figure 8-23c.
mRNP adapter proteins such as REF that assist in the binding
of the NXFl/NXTl exporter to mRNPs. In one case, a yeast
SR protein (Npl3) functions as an adapter protein that pro-
Pre-mRNAs in Spliceosomes Are Not
motes the binding of the yeast mRNP exporter (Figure 8-22).
The SR protein initially binds to nascent pre-mRNA in its Exported from the Nucleus
phosphorylated form. When 3' cleavage and polyadenylation It is critical that only fully processed mature mRNAs be ex-
are completed, the adapter protein is dephosphorylated by a ported from the nucleus because translation of incompletely
specific nuclear protein phosphatase essential for mRNP ex- processed pre-mRNAs containing introns would produce de-
port. Only the dephosphorylated adapter protein can bind the fective proteins that might interfere with the functioning of
mRNP exporter, thereby coupling mRNP export to correct the cell. To prevent this, pre-mRNAs associated with snRNPs
polyadenylation. This is one form of mRNA "quality con- in spliceosomes usually are prevented from being transported
trol." If the nascent mRNf is not correctly processed, it is not to the cytoplasm.
recognized by the phosphatase that dephosphorylates Npl3. In one type of experiment demonstrating this restriction,
Consequently, it is not bound by the mRNA exporter and not a gene encoding a pre-mRNA with a single intron that nor-
exported from the nucleus. Instead it is degraded by exo- mally is spliced out was mutated to introduce deviations
somes, the multiprotein complexes that degrade unprotected from the consensus splice-site sequences. Mutation of either
RNAs in the nucleus and cytoplasm (see Figure 8-1). the 5' or the 3' invariant splice-site bases at the ends of the
Following export to the cytoplasm, the Npl3 SR protein inrron resulted in pre-mRNAs that were bound by snRNPs
is phosphorylated by a specific cytoplasmic protein kinase. to form spliceosomes; however, RNA splicing was blocked,
This causes it to dissociate from the mRNP, along with the and the pre-mRNA was retained in the nucleus. In contrast,
mRNP exporLtr. In this way, dephosphorylation of adapter mutation of both the 5' and 3' splice sites in the same pre-
mRNP proteins in the nucleus once RNA processing is com- mRNA resulted in export of the unspliced pre-mRNA, al-
plete and their phosphorylation and resulting dissociation in though less efficiently than for the spliced mRNA. When
the cytoplasm results in a higher concentration of mRNP both splice sites were mutated, the pre-mRNAs were not ef-
exporter-mRNP complexes in the nucleus, where they form, ficiently bound by snRNPs, and, consequently, their export
and a lower concentration of these complexes in the cyto- was not blocked.

8.3 Transport of mRNA Across the Nuclear Envelope 367

 RNA pol II

RNA pol II RNA pol II

II
Nucleoplsam

·.
NPC

Cytoplasm

lmportin
AAAAAAA
II AAAAAAA
~NXF1/NXT1
NXF1/N~
II

e a NXF1/NXT1

FIGURE 8 -22 Reversible phosphorylation and direction of mRNP Skyl phosphorylates Npl3 in the cytoplasm, causing lit dissociation of
nuclear export. Step 0 :The yeast SR protein Npl3 binds nascent the mRNP exporter and phosphorylated Npl3, probably through the
pre-mRNAs in its phosphorylated form. Step fJ: When polyadenylation action of an RNA helicase associated with NPC cytoplasmic filaments.
has occurred successfully, the Glc? nuclear phosphatase essential for mThe mRNA transporter and phosphorylated Npl3 are transported
mRNP export dephosphorylates Npl3, promoting the binding of the back into the nucleus through NPCs. 6 Transported mRNA is available
yeast mRNP exporter, NXFl /NXTl . Step 11:The mRNP exporter allows for translation in the cytoplasm. [From E. lzaurralde, 2004, Nat. Struct. Mol.
diffusion of the mRNP complex through the central channel of the Bioi. 11 :210-212. See W. Gilbert and C. Guthrie, 2004, Mol. Ce/113:201-212.]
nuclear pore complex (NPC). Step m: The cytoplasmic protein kinase

Recent studies in yeast have shown that a nuclear protein HIV Rev Protein Regulates the Transport
that associates with a nucleoporin in the NPC nuclear basket of Unspliced Viral mRNAs
is required to retain pre-mRNAs associated with snRNPs in
the nucleus. If either this protein or the nucleoporin to which As discussed earlier, transport of mRNPs containing mature,
it binds is deleted, unspliced prc-mRNAs are exported. functional mRNAs from the nucleus to the cytoplasm entails
a complex mechanism that is crucial to gene expression (sec
Many cases of thalassemia, an inherited disease that re- Figures 8-21, 8-22, and 8-23). Regulation of this transport
sults in abnormally low levels of globin proteins, are due theoretically could provide another means of gene control,
to mutations in globin-gene splice sites that decrease the effi- although it appears to be relatively rare. Indeed, the only
ciency of splicing but do not prevent association of the pre- known examples of regulated mRNA export occur during
mRNA with snRNPs. The resulting unspliced globin pre-mRNAs the cellular response to conditions (e.g., heat shock) that
are retained in reticulocyte nuclei and are rapidly degraded. • cause protein denaturation or during viral infection when

368 CHAPTER 8 • Post-transcriptional Gene Control

 (b) FIGURE 8 -23 Formation of heterogeneous ribonucleoprotein
particles (hnRNPs) and export of mRNPs from t he nucleus.
(a) Model of a single chromatin transcription loop and assembly of
Balbiani ring (BR) mRNP in Chironomous ten tans. Nascent RNA transcripts
produced from the template DNA rapidly associate with proteins,
forming hnRNPs. The gradual increase in size of the hnRNPs reflects

\ the increasing length of RNA transcripts at greater distances from the

transcription start site. The model was reconstructed from electron
micrographs of serial thin sections of salivary gland cells. (b) Schematic
diagram of the biogenesis of hnRNPs. Following processing of the
pre-mRNA, the resulting ribonucleoprotein particle is referred to as
an mRNP. (c) Model for the transport of BR mRNPs through the nuclear
pore complex (NPC) based on electron microscopic studies. Note that
the curved mRNPs appear to uncoil as they pass through nuclear pores.
As the mRNA enters the cytoplasm, it rapidly associates with ribosomes,
indicating that the 5' end passes through the NPC first. [Part (a) from
C. Erricson et al., 1989, Cell 56:631 ; courtesy of B. Daneholt. Parts (b) and
(c) adapted from B. Daneholt, 1997, Ce//88:585. See also B. Daneholt, 2001,
Template mRNP Proc. Nat'/. Acad. Sci. USA 98:7012.)
DNA

m RNA

virus-induced alterations in nuclear transport maximize viral have some mechanism for overcoming this block, permitting
replication. Here we describe the regulation of mRNP ex- export of the longer viral mRNAs. Some retroviruses have
port mediated by a protein encoded by human immunodefi- evolved a sequence called the collstitutive transport element
ciency virus (HIV). (CTE), which binds to the NXFl/NXTl mRNP exporter with
A retrovirus, HJV integrates a DNA copy of its RNA ge- high affinity, thereby permitting export of unspliced retroviral
nome into the host-cell DNA (see Figure 4 -49). The integrated Rl"'A into the cytoplasm. HJV solved the problem differently.
viral DNA, or provirus, contains a single transcription unit, Studies with H IV mutants showed that transport of un-
which is transcribed into a. single primary transcript by cell ular spliced 9-k b and singly spliced 4 -kb viral mRNAs from the
RNA polymerase II. The HIV transcript can be spliced in alter- nucleus to the cytoplasm requires the virus-encoded Rev pro-
native ways to yield three classes of mRNAs: a 9-kb unspliced tein. Subsequent biochemical experiments demonstrated that
mRNA; ~4-kb mRNAs formed by removal of one intron; and Rev binds to a specific Rev-response element (RRE) present in
~2-kb mRNAs formed by removal of two or more introns HJV RNA. In cells infected with HJV mutants lacking the RRE,
(Figure 8-24). After their synthesis in the host-cell nucleus, all unspliced and singly spliced viral mRNAs remain in the nu -
three classes of HJV mRNAs are transported to the cytoplasm cleus, demonstrating that the RRE is required for Rev-mediated
and translated into viral proteins; some of the 9-kb unspliced stimulation of nuclear export. Early in an infection, before
RNA is used as the viral genome in progeny virions that bud any Rev protein is synthesized, only the multiply spliced 2-kb
from the cell surface. mRNAs can be exported. One of these 2-kb mRNAs encodes
Since the 9-kb and 4-kb HIV mRNAs contain splice sites, Rev, which contains a leucine-rich nuclear export signal that
they can be viewed as incompletely spliced mRNAs. As dis- interacts with transporter Exportin 1. As discussed in Chapter
cussed earlier, association of such incompletely spliced mRNAs 13, translation and nuclear import of Rev results in export of
with snRNPs in spliceosomes normally b locks their export the larger unspliced and singly spliced HIV mRNAs through
from the nucleus. Thus HJV, as well as other retroviruses, must the nuclear pore complex.

8.3 Transport of mRNA Across the Nuclear Envelope 369

 HIV provirus

~~----------------~·~--~c==
/

NUCLEAR mRNAs
l RRE

Transcription, splicing
CYTOPLASMIC mRNAs
9-kb +Rev
('
Unspliced 9 kb

4-kb ~'
Singly spliced 4 kb

2-kb ~
-Rev •0:>.::;;;;;;;;;;;;:;~ _____., 0 Rev protein
Multiply spliced ' 2 kb Translation

Nucleoplasm Cytopl asm

FIGURE 8-24 Transport of HIV mRNAs from the nucleus to the RNA species are translated into different viral proteins. Rev protein,
cytoplasm. The HIV genome, which contains several coding regions, is encoded by a 2-kb mRNA, interact s with the Rev-response element
transcribed into a single 9-kb primary transcript. Several -4-kb mRNAs (RRE) in the unspliced and singly spliced mRNAs, stimulating their
result from alternative splicing out of any one of several introns transport to the cytoplasm. [Adapted from B. R. Cullen and M. H. Malim, 1991,
(dashed lines) and several -2-kb mRNAs from splicing out oftwo or Trends Biochem. Sci. 16:346.]
more alternative introns. After transport to the cytoplasm, the various

scription elongation in the promoter proximal region are the

KEY CONCEPTS of Section 8.3 initial mechanisms for controlling the expression of genes in
the gene expression pathway of DNA~RNA~protein. In
Transport of mRNA Across the Nuclear Envelope
preceding sections of this chapter, we learned that the expres-
• Most mRNPs are exported from the nucleus by a het- sion of protein isoforms is controlled by regulating alternative
erodimeric mRNP exporter that interacts with FG-rcpcats of RNA splicing and cleavage and polyadenylation at alternative
FG-nucleoporins (see Figure 8-20) . The direction of trans- poly(A) sites. Although nuclear export of fully and correctly
port (nucleus~cytoplasm) may result from dissociation of processed mRNPs to the cytoplasm is rarely regulated, the
the exporter-mRNP complex in the cytoplasm by phosphor- export of improperly processed or aberrantly remodeled pre-
ylation of mRNP proteins by cytoplasmic kinases and the mRNPs is prevented, and such abnormal transcripts are de-
action of an RNA helicase associated with cytoplasmic fila- graded by the exosome. However, retroviruses, including
ments of the nuclear pore complexes. HIV, have evolved mechanisms that permit pre-mRNAs that
• The mRNP exporter binds to most mRNAs cooperatively retain splice sites to be exported and t ranslated.
v.·ith SR proteins bound to exons and with REF associated with In this section we consider other mechanisms of post-
the exon-junction complexes that bind to mRNAs following transcriptional control that contribute to regulating the ex-
RNA splicing, and also binds to additional mRNP proteins. pression of some genes. Most of these mechanisms operate
in the cytoplasm, controlling the stability or loca lization of
Pre-mRNAs bound by a spliceosome normally arc not ex-
mRNA or its translation into protein. We begin by discuss-
ported from the nucleus, ensuring that only fully processed,
ing two recently discovered and related mechanisms of gene
functional mRNAs reach the cytoplasm for translation.
control that provide powerful new techniques for manipulat-
ing the expression of specific genes for experimental and
therapeutic purposes. These mechanisms are controlled by
short, - 22-nuclcotide, single-stranded RNAs called micro
8.4 Cytoplasmic Mechanisms RNAs (miRNAs) and short interfering (siRNAs). Both base-
pair with specific target mRNAs, either inhibiting their
of Post-transcriptional Control translation (miRNAs) or causing their degradation (siRNAs).
Before proceeding, let's quickly review the steps in gene ex- Humans express -500 miRNAs. Most of these are expressed
pression at which control is exerted. We saw in the previous in specific cell types at particular times during embryogenesis
chapter that regulation of transcription initiation and tran- and after birth. Many miRNAs can target more than one

370 CHAPTER 8 • Post-transcriptional Gene Control

.··
 (a) miRNA ~translation inhibition (b) siRNA ~RNA cleavage

,o ~ OH I 1 I I ;1'-~""'l ~..,.~,..,1
:"T"IT"""j ;~~
II"'TI"T"Inii'"T'I'I"1111"'TI'T'l:-;
• I I I L.LI I I I I I I I I I I I I LLr.
Target RNA Target RNA t
uc
C A
C A
5' -UCCCUGAGA GUGUGA 3' ~-UAGGUAGUUUCAUGUUGUUGGG-3'
11111 1 11 I I II I 1111111111111 111111111
3' -UCCAGGGACUCAACCAACACUCAA- 5' 3' - CUUAUCCGUCAAAGUACAACAACCUUCU- 5'

lin-4 miRNA and lin-14 mRNA (C.elegans) miR-196a and HOXBB mRNA (H. sapiens)

5' -UGUUAGCUGGAUGAAAACTT- 3' 5' -UCGGACCAGGCUUCAUUCCCC_ 3'

11111111 111111111 1111111111111 1 11111
3' -GCCACAAUCGAAACACUUUUGAAGGC- 5' 3' - UUAGGCCUGGUCCGAAGUAGGGUUAGU- 5'

CXCR4 m iRNA and target mRNA (H. sapiens) miR-166 and PHAVOLUTA mRNA (A. thaliana)

FIGURE 8-25 Base pairing with target RNAs distinguishes miRNA mRNA. (b) siRNA hybridizes perfectly with its target mRNA, causing
and siRNA. (a) miRNAs hybridize imperfectly with their target mRNAs, cleavage of the mANA at the position indicated by the red arrow,
repressing translation of the mRNA. Nucleotides 2 to 7 of an miRNA triggering its rapid degradation. [Adapted from P. D. Zamore and B. Haley,
(highlighted blue) are the most critical for targeting it to a specific 2005, Science 3 09:151 9.]

mRNA. Consequently, these newly discovered mechanisms that they regulate need not be perfect (Figure 8-25). In fact,
contribute significantly to the regulation of gene expression. considerable experimentation with synthetic miRNAs has
siRNAs, involved in the process called RNA interference, are shown that complementarity between the six or seven 5 ' nucle-
also an important cellular defense against viral infection and otides of an miRNA and its target mRNA 3' untranslated
excessive transposition by transposons. region are most critical for target mRNA selection.
Most miRNAs are processed from RNA polymerase II tran-
scripts of several hundred to thousands of nucleotides in length
Micro RNAs Repress Translation
called pri (for primary transcript)-miRNAs (figure 8-26). Pri-
of Specific mRNAs miRNAs can contain the sequence of one or more miRNAs.
Micro RNAs (miRNAs) were fim discovered during analysis miRNAs are also processed out of some excised introns and
of mutations in the lin-4 and let-7 genes of the nematode C. from 3' untranslated regions of some pre-mRNAs. Within
elegans, which influence development of the organism, Clon- these long transcripts are sequences that fold into hairpin struc-
ing and analysis 'o f wild-type lin-4 and let-7 revealed that they tures of -70 nucleotides in length with imperfect base pairing
encode not protein products but rather RNAs only 21 and 22 in the stem. A nuclear RNase specific for double-stranded
nucleotides long, respectively. The RNAs hybridize to the 3' RNA called Drosha acts with a nuclear double-stranded RNA-
untranslated regions of specific target mRNAs. For example, binding protein called DGCR8 in humans (Pasha in Dro-
the lin-4 miRNA, which is expressed early in embryogenesis, sophila) and cleaves the hairpin region out of the long precursor
hybridizes to the 3' untranslatcd regions of both the lin-14 RNA, generating a pre-miRNA. Pre-miRNAs are recognized
and lin-28 mRNAs in rhe.cytoplasm, thereby repressing their and bound by a specific nuclear transporter, Exportin5, which
translation by a mechanism discussed below, Expression of interacts with the FG-domains of nucleoporins, allowing the
lin-4 miRNA ceases later in development, allowing translation complex to diffuse through the inner channel of the nuclear
of newly synthesized lin-14 and lin-28 mRNAs at that time. pore complex, as discussed above (see Figure 8-20, and Chap-
Expression of let-7 miRNA occurs at comparable times dur- ter 13). Once in the cytoplasm, a cytoplasmic double-stranded
ing embryogenesis of all bilaterally symmetric animals. RNA-specific RNase called Dicer acts with a cytoplasmic
miRNA regulation of translation appears to be widespread double-stranded RNA-binding protein called TRBP in
in a ll multicellular plants and animals. In the past few years, humans (for Tar binding protein; called Loquacious in Dro-
small RNAs of 20- 26 nucleotides have been isolated, cloned, sophila) to further process the pre-miRNA into a double-
and sequenced from various tissues ot multiple model organ- stranded miRNA. The double-stranded miRNA is approximately
isms. Recent estimates suggest the expression of one-third of two turns of an A-form RNA helix in length, with strands
all human genes is regulated by -500 human miRNAs iso- 21-23 nucleotides long and two unpaired 3'-nucleotides at
lated from various tissues. The potential for regulation of mul- each end. Finally, one of the two strands is selected for assem-
tiple mRNAs by one miRNA is great because base pairing bly into a mature RNA-induced silencing complex (RISC)
between the miRNA and the sequence in the 3' ends of mRNAs containing a single-stranded macure miRNA bound by a

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 37 1

 multidomain Argonaute protein, a member of a protein
family with a recognizable conserved sequence. Several Ar-
miR-1-1 gene gonaute proteins are expressed in some organisms, espe-
..F'~J.'"'ff~' cially plants, and are found in distinct RISC complexes with
different functions.
The miRNA-RISC complexes associate with target
mRNPs by base pairing between the Argonaute-bound ma-
ture miRNA and complementary regions in the 3' untrans-
lated regions (3' UTRs) of target mRNAs (see Figure 8-25).
Inhibition of target mRNA translation requires the binding
of two or more RISC complexes to distinct complementary
regions in the target mRNA 3' UTR. It has been suggested
that this may allow combinatorial regulation of mRNA

~
translation by separately regulating the transcription of two or
Drosha more different pri-miRNAs, which are processed to miRNAs
Pasha that are required in combination tO suppress the translation
of a specific target mRNA.
pre-miR-1-1
The binding of several RISC complexes to an mRNA in-
LO Nucleus hibits translation initiation by a mechanism currently being
c
·e analyzed. Binding of RISC complexes causes the bound
0
c. mRNPs to associate with dense cytOplasmic domains many
X
TRBP w Cytoplasm times the size of a ribosome called cytoplasmic RNA-processing
Dicer bodies, or simply P bodies. P bodies, which will be described
in greater detail below, are sites of RNA degradation that
TRBP Dicer '\
contain no ribosomes or translation factors, potentially ex-
"'JGAAcc
. s· -p CAUACUUCUUUAUAUGCCCAUA U plaining the inhibition of translation. The association with P
pre-m1R-1-1 1111111111111 11 IIII I G
3' -AUGUAUGAAGAAAUGUA G GGUAUC C bodies may also explain why expression of an miRNA often
----~ GAAU decreases the stability of a targeted mRNA.

5
l
pCAUACUUCUUUAUAUGCCCAUA- 3'
As mentioned earlier, approximately 500 different human
miRNAs have been observed, many of them expressed only
miR-1-1 111111111111111 Ill in specific cell types. Determining the function of these miRNAs
3'-AUGUAUGAAGAAAUGUA G GGUp- s·
is currently a highly active area of research. In one example,

l a specific miRNA called miR-133 is induced when myoblasts

differentiate into muscle cells. miR-133 suppresses the trans-
Mature miR-1-1 lation of PTB, a regulatory splicing factor that functions sim-
bound to an ilarly to Sxl in Drosophila (see Figure 8-16). PTB binds to the
Argonaute RISC
protein 3' splice-site region in the pre-mRNAs of many genes, leading
to exon skipping or use of alternative 3 ' splice sites. When
miR-133 is expressed in differentiating myoblasts, the PTB
concentration falls without a significant decrease in the con-
FIGURE 8·26 miRNA processing. This diagram shows transcription centration of PTB mRNA. As a result, alternative isoforms of
and processing of the miR-1-1 miRNA. The primary miRNA transcript multiple proteins important for muscle-cell function are ex-
(pri-miRNA) is transcribed by RNA polymerase II. The nuclear double- pressed in the differentiated cells.
stranded RNA-specific endoribonuclease Drosha, with its partner
Other examples of miRNA regulation in various organ-
double-stranded RNA-binding protein DGCRS (Pasha in Drosophila),
isms are being discovered at a rapid pace. Knocking out the
make the initial cleavages in the pri-miRNA, generating a - 70-nucle-
otide pre-miRNA that is exported to the cytoplasm by nuclear dicer gene eliminates the generation of miRNA in mammals.
transporter ExportinS. The pre-miRNA is further processed in the This causes embryonic death early in mouse development.
cytoplasm to a double-stranded miRNA with a two-base single- However, when dicer is knocked out only in limb primordia,
stranded 3' end by Dicer in conjunction with the double-stranded the influence of miRNA on the development of the nones-
RNA-binding protein TRBP (loquatious in Drosophila). Finally, one of sential limbs can be observed (Figure 8-27). Although all
the two strands i~ incorporated into an RISC complex, where it is bound major cell types differentiate and the fundamental aspects of
by an Argonaute protein. [Adapted from P. D. Zamore and B. Haley, 2005, limb patterning are maintained, development is abnormal-
Science 309:1 519.] demonstrating the importance of miRNAs in regulating
the proper level of translation of multiple mRNAs. Of the
-500 human miRNAs, 53 appear to be unique to primates.
It seems likely that new miRNAs arose readily during evolution
by the duplication of a pri-miRNA gene followed by mutation

372 CHAPTER 8 • Post-transcriptional Gene Control

 Wild type Dicer mut ant cases, the mature short single-stranded RNA, either mature
siRNA or mature miRNA, is assembled into RISC complexes
in which the short RNAs are bound by an Argonaute pro-
tein. What distinguishes a RISC complex containing an
siRNA from one containing an miRNA is that the siRNA
base-pairs extensively with its target RNA and induces its
cleavage, whereas a RISC complex associated with an
miRNA recognizes its target through imperfect base-pairing
and results in inhibition of translation.
The Argonaute protein is responsible for cleavage of tar-
EXPERIMENTAL FIGURE 8 · 27 miRNA function in limb get RNA; one domain of the Argonaute protein is homolo-
development. Micrographs comparing normal (left) and Dicer
gous to RNase H enzymes that degrade the RNA of an
knockout (right) limbs of embryonic development day-13 mouse
RNA-DNA hybrid (see Figure 6-14). When the 5' end of the
embryos immunostained for the GdS protein, a marker of joint
short RNA of a RISC complex base-pairs precisely with a
formation. Dicer is knocked out in developing mouse embryos by
conditional expression of Cre to induce deletion of the Dicer gene only
target mRNA over a distance of one turn of an RNA helix
in these cells (see Figure 5-42). [From B. D. Harfe et al., 2005, Proc. Nor'/. Acad.
(10-12 base pairs), th is domain of Argonaute cleaves the
Sci. USA 102:10896.] phosphodiesrer bond of the target RNA across from nucle-
otides 10 and 11 of the siRNA (see Figure 8-25 ). The cleaved
RNAs arc released and subsequently degraded by cytoplas-
mic exosomes and 5' exoribonucleases. If base pairing is not
of bases encoding the mature miRNA. miRNAs are particu- perfect, the Argonaute domain does not cleave or release the
larly abundant in plants-more than 1.5 million distinct target mRNA. Instead, if several miRNA-RISC complexes
miRNAs have been characterized in Arabidopsis thaliana! associate with a target mRNA, its translation is tnhibired
and the mRNA becomes associated with P bodies, where, as
mentioned earlier, it is probably degraded by a different and
RNA Interference Induces Degradation
slower mechanism than the degradation pathway initiated
of Precisely Complementary mRNAs by RISC cleavage of a perfectly complementary target RNA.
RNA interference (RNAi) was discovered unexpectedly dur- When double-stranded RNA is introduced into the cyto-
ing attempts to experimentally manipulate the expression of plasm of eukaryotic cells, it enters the pathway for assembly
specific genes. Researchers tried to inhibit the expression of a of siRNAs into a RISC complex because it is recognized by
gene in C. elegans by microinjecting a single-stranded, com- the cytoplasmic Dicer enzyme and TRBP double-stranded
plementary RNA that would hybridize to the encoded mRNA RNA-binding protein that process pre-miRNAs (see Figure
and prevent its translation, a method called antisense inhibi- 8-26). This process of RNA interference is believed to be an
tion. But in control experiments, perfectly base-paired double- ancient cellular defense against certain viruses and mobile
stranded RNA a few hundred base pairs long was much more genetic elements in both plants and animals. Plants with mu-
effective at inhibiting expression of the gene than the anti- tations in the genes encoding Dicer and RISC proteins ex-
sense strand a lone (see Figure 5-45). Similar inhibition of hibit increased sensitivity to infection by RNA viruses and
gene expression by an introduced double-stranded RNA soon increased movement of transposons within their genomes.
was observed in plants. In each case, the double-stranded The double-stranded RNA intermediates generated during
RNA induced degradation of all cellular RNAs containing a replication of RNA viruses are thought to be recognized by
sequence that was exactly the same as one strand of the double- the Dicer ribonuclease, inducing an RNAi response that ulti-
stranded RNA. Because of the specificity of RNA interference mately degrades viral mRNAs. During transposition, trans-
in targeting mRNAs for destruction, it has become a power- posons are inserted into cellular genes in a random orientation,
ful experimental tool for studying gene function. and their transcription from different promoters produces
Subsequent biochemical studies with extracts of Dro- complementary RKAs that can hybrid ize with each other, initi-
sophila embryos showed that a long double-stranded RNA ating the RNAi system that then interferes with the expression
that mediates interference is initially processed into a double- of transposon proteins required for additional transpositions.
stranded short interfering RNA (siRNA). The strands in In plants and C. elegans the RNAi response can be in-
siRNA contain 21-23 nucleotides hybridized to each other duced in all cells of the organism by introduction of double-
so that the two bases at the 3' end of each strand are single- stranded RNA into just a few cell s. Such organism-wide
stranded. Further studies revealed that the cytoplasmic double- induction requires production of a protein that is homolo-
stranded RNA-specific ribonuclease that cleaves long gous to the RNA replicases of RNA viruses. It has been
double-stranded RNA into siRNAs is the same Dicer en- revealed that double-stranded siRNAs arc replicated and then
zyme involved in processing pre-miRNAs after their nuclear transferred to other cells in these organisms. In plants, transfer
export to the cytoplasm (see Figure 8-26). This discovery led of siRNAs might occur through plasmodesmata, the cytoplas-
to the realization that RNA interference and miRNA-medi- mic connections between plant cells that traverse the cell walls
ated translational repression are related processes. In both between them (see Figure 20-38 ). Organism-wide induction

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 373

of RNA interference does not occur in Drosophila or mam- mRNAs stored in oocytes are not translated efficiently. At
mals, presumably because their genomes do not encode the appropriate time during oocyte maturation or after ferti-
RNA replicase homologs. lization of an egg cell, usually in response to an external sig-
In mammalian cells, the introduction of long RNA-RNA nal, approximately 150 A residues are added to the short
duplex molecules into the cytoplasm results in the generalized poly(A) tails on these mRNAs in the cytoplasm, stimulating
inhibition of protein synthesis via the PKR pathway, discussed their translation.
further below. This greatly limits the use of long double- Studies with mRNAs stored in Xenopus oocytes have
stranded RNAs to experimentally induce an RNAi response helped elucidate the mechanism of this type of translational
against a specific targeted mRNA. furtunacdy, researchers control. Experiments in which short-tailed mRNAs are in-
discovered that one strand of double-stranded siRNAs 21-23 jected into oocytes have shown that two sequences in their 3'
nucleotides in length with two-base 3' single-stranded regions UTR are required for their polyadenylation in the cytoplasm:
leads to the generation of mature siRNA RISC complexes the AAUAAA poly(A) signal that is also required for the nu- .
without inducing the generalized inhibition of protein synthe- clear polyadenylation of pre-mRNAs, and one or more cop-
sis. This has allowed researchers to use synthetic double- ies of an upstream U-rich cytoplasmic polyadenylation
stranded siRNAs to "knock down" the expression of specific element (CPE). This regulatory element is bound by a highly
genes in human cells as well as in other mammals. This conserved CPE-binding protein (CPEB) that contains an
method of siRNA knockdown is now widely used in studies RRM domain and a zinc-finger domain.
of diverse processes, including the RNAi pathway itself. According to the current model, in the absence of a stimu-
latory signal, CPEB bound to the U-rich CPE interacts with the
protein Maskin, which in turn binds to eiF4E associated with
Cytoplasmic Polyadenylation Promotes
the mRNA 5' cap (Figure 8-28, left). As a result, eiF4E cannot
Translation of Some mRNAs interact with other initiation factors and the 40S ribosomal
In addition to repression of translation by miRNAs, other subunit, so translation initiation is blocked. During oocyte
protein-mediated translational controls help regulate expres- maturation, a specific CPEB serine is phosphorylated, causing
sion of some genes. Regulatory sequences, or elements, in Maskin to dissociate from the complex. This allows cytoplas-
mRNAs that interact with specific proteins to control transla- mic forms of the cleavage and polyadenylation specificity fac-
tion generally are present in the untranslated region (UTR) at tor (CPSF) and poly(A) polymerase to bind to the mRNA
the 3' or 5' end of an mRNA. Here we discuss a type of pro- cooperatively with CPEB. Once the poly(A) polymerase cata-
tein-mediated translational control involving 3' regulatory ele- lyzes the addition of A residues, PABPI can bind to the length-
ments. A different mechanism involving RNA-binding proteins ened poly(A) tail, leading to the stabilized interaction of all the
that interact with 5' regulatory elements is discussed later. factors needed to initiate translation (Figure 8-28, right; see
Translation of many eukaryotic mRNAs is regulated by also Figure 4-24). In the case of Xenopus oocyte maturation,
sequence-specific RNA-binding proteins that bind coopera- the protein kinase that phosphorylates CPEB is activated in
tively to neighb<:>ring sites in 3' UTRs . This allows them to response ro the hormone progesterone. Thus timing of the
function in a combinatorial manner, similar to the cooperative translation of stored mRNAs encoding proteins needed for
binding of transcription factors to regulatory sites in an en- oocyte maturation is regulated by this external signal.
hancer or promoter region. In most cases studied, translation Considerable evidence indicates that a similar mechanism
is repressed by protein binding to 3' regulatory elements and of translational control plays a role in learning and memory.
regulation results from derepression at the appropriate time or In the central nervous system, the axons from a thousand or
place in a cell or developing embryo. The mechanism of such so neurons can make connections (synapses) with the den-
repression is best understood for mRNAs that must undergo drites of a single postsynaptic neuron (Figure 22-23) . When
cytoplasmic polyadenylat1on before they can be translated. one of these axons is stimulated, the postsynaptic neuron "re-
Cytoplasmic polyadenylation is a critical aspect of gene members" which one of these thousands of synapses was
expression in the early embryo of animals. The egg cells stimulated. The next time that synapse is stimulated, the
(oocytes) of multicellular animals contain many mRNAs, en- strength of the response triggered in the postsynaptic cell dif- ·.
coding numerous different proteins that are not translated fers from the first time. This change in response has been
until after the egg is fertilized by a sperm cell. Some of these shown to result largely from the translational activation of
"stored" mRNAs have a short poly(A) tail, consisting of mRNAs stored in the region of the synapse, leading to the
only -20-40 A residues, to which just a few molecules of local synthesis of new proteins that increase the size and alter
cytoplasmic poly(A)-binding protein (PABPI) can bind. As the neurophysiological characteristics of the synapse. The
discussed in Chapter 4, multi pit> PA RPI molecules hound to finding that CPEB is present in neuronal dendrites has led to
the long poly(A) tail of an mRNA interact with the eiF4G the proposal that cytoplasmic polyadenylation stimulates
initiation factor, thereby stabilizing the interaction of the translation of specific mRNAs in dendrites, much as it does
mRNA 5' cap with eiF4E, which is required for translation in oocytes. In this case, presumably, synaptic activity (rather
initiation (see Figure 4-24). Because this stabilization cannot than a hormone) is the signal that induces phosphorylation of
occur with mRNAs that have short poly(A) tails, such CPEB and subsequent activation of translation.

374 CHAPTERs • Post-transcriptional Gene Control

 Translationally dormant Translationally active

1 - - - - - :UUUUAU • AAUAAA-A
I I I I

mANA I I
CPE Poly(A)
signal
FIGURE 8-28 Model for control of cytoplasmic polyadenylation specificity factor (CPSF) then binds to the poly(A) site, interacting with
and translation initiation. (Left) In immature oocytes, mRNAs both bound CPEB and the cytoplasmic form of poly(A) polymerase
containing the U-rich cytoplasmic polyadenylation element (CPE) have (PAP). After the poly(A) tail is lengthened, multiple copies of the
short poly(A) tails. CPE-binding protein (CPEB) mediates repression of cytoplasmic poly(A)-binding protein I (PABPI) can bind to it and
interact with eiF4G, which functions with other initiation factors to
... translation through the interactions depicted, which prevent assembly
of an initiation complex at the 5' end ofthe mRNA. (Right) Hormone bind the 405 ribosome subunit and initiate translation. [Adapted from
stimulation of oocytes activates a protein kinase that phosphorylates R. Mendez and J. D. Richter, 2001, Nature Rev. Mol. Cell Bioi. 2:521 .]
CPEB, causing it to release Maskin. The cleavage and polyadenylation

Cytoplasmic mRNAs are degraded by one of the three

Degradation of mRNAs in the Cytoplasm
pathways shown in Figure 8-29. For mQst mRNAs, the
Occurs by Several Mechanisms deadenylation-dependent pathway is followed: the length of
The concentration of an mRNA is a function of both its rate the poly(A) tail gradually decreases with time through the
of synthesis and its rate of degradation. For this reason, if action of a deadenylating nuclease. When it is shortened suf-
two genes are transcribed at the same rate, the steady-state ficiently, PABPI molecules can no longer bind and stabilize
concentration of the corresponding mRNA that is more sta- the interaction of the 5' cap and translation initiation factors
ble will be higher than the concentration of the other. The (see Figure 4-24). The exposed cap then is removed by a de-
stability of an mRNA also determines how rapidly synthesis capping enzyme (Dcp1/Dcp2 in 5. cerevisiae), and the un -
.· of the encoded protein can be shut down . For a stable protected mRNA is degraded by a 5'~3' exonuclease (Xrnl
mRNA, synthesis of the encoded protein persists long after in 5. cerevisiae). Removal of the poly(A) tail also makes
transcription of the gene is repressed. Most bacterial mRNAs mRNAs susceptible to degradation by cytoplasmic exosomes
are unstable, decaying exponentially with a typical half-life containing 3'~5' exonucleases. The 5'~3' exonucleases
of a few minute~. For this reason, a bacterial cell can rapidly predominate in yeast, and the 3' ~5 ' exosome predominates
adjust the synthesis of proteins to accommodate changes in in mammalian cells. The decapping enzymes and 5'~3' exo-
the cellular environment. Most cells in multicellular organ- nuclease are concentrated in the P bodies, regions of the cy-
isms, on the other hand, exist in a fairly constant environ- toplasm of unusually high density.
ment and carry out a specific set of functions over periods of Some mRNAs are degraded primarily by a deadenylation-
days to months or even the lifetime of the organism (nerve independent decapping pathway (see Figure 8-29). This is be-
cells, for example). Accoroingly, most mRNAs of higher eu- cause certain sequences at the 5' end of an mRNA seem to
karyotes have half-lives of many hours. make the cap sensitive to the decapping enzyme. For these
However, some proteins in eukaryotic cells are required mRNAs, the rate at which they are decapped controls the rate
only for short periods and must be expressed in bursts. For at which they are degraded because once the 5' cap is removed,
example, as discussed in the chapter introduction, certain sig- the RNA is rapidly hydrolyzed by the 5' ~3' exonuclease.
naling molecules called cytokines, which are involved in the The rate of mRNA deadenylation varies inversely with the
immune response of mammals, are synthesized and secreted frequency of translation initiation for an mRNA: the higher the
in short bursts (see Chapter 23). Similarly, many of the tran- frequency of initiation, the slower the rate of deadenylation.
scription factors that regulate the onset of the S phase of the This relation probably is due to the reciprocal interactions be-
cell cycle, such as c-Fos and c-Jun, are synthesized for brief tween translation initiation factors bound at the 5' cap and
periods only (Chapter 19 ). Expression of such proteins oc- PABPI bound to the poly(A) tail. For an mRNA that is trans-
curs in short bursts because transcription of their genes can lated at a high rate, initiation factors are bound to the cap much
be rapidly turned on and off, and their mRNAs have unusu- of the time, stabilizing the binding of PABPI and thereby pro-
ally short half-lives, on the order of 30 minutes or less. tecting the poly(A) tail from the deadenylation exonuclease.

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 375

 Decapping pathway Deadenylation-dependent Endonucleolytic
(deadenylation-independent) pathways pathway

I~·w••o
- - - - - - AAAAAA - - - - - - AAAAAA - - - - - - AAAAAA

1
1a
- -----A
Poly{A) shortenmg

-----~- AAAAAA
Endonucleolytic
cleavage

Decapping / ""- . Exo~~c:e~lyt1c /

/ ~ decay / _

• - - - - - - AAAAAA • ------ A
5->3 / Exosome
Exonucleolytic

c decay

A
FIGURE 8 -29 Pathways for degradation of eukaryotic mRNAs. may either (1) be decapped and degraded by a 5' -t3' exonuclease or
In the deadenylation-dependent (middle) pathways, the poly(A) tail is (2) be degraded by a 3 '-t5' exonuclease in cytoplasmic exosomes.
progressively shortened by a deadenylase (orange) until it reaches a Some mRNAs (right) are cleaved internally by an endonuclease and the
length of 20 or fewer A residues, at which point the interaction with fragment s degraded by an exosome. Oth;r mRNAs (left) are decapped
PABPI is destabilized, leading to weakened interactions between the before they are deadenylated and then degraded by a 5' -t3' exonucle·
5' cap and translation-initiation factors. The deadenylated mRNA then ase. [Adapted from M. Tucker and R. Parker, 2000, Ann. Rev. Biochem. 69:571.]

Many short-lived mRNAs in mammalian cells contain mul- translational modifications such as phosphorylation. Such
tiple, sometimes overlapping copies of the sequence AUUUA in mechanisms affect the translation rate of mpst mRNAs and
their 3' untranslated region. Specific RNA-binding proteins hence the overall rate of cellular protein synthesis.
have been found that both bind to these 3' AU-rich sequences
and also interact with a deadenylating enzyme and with the TOR Pathway The TOR pathway was discovered through
exosome. This causes rapid deadenylation and subsequent research into the mechanism of action of rapamycin, an an-
3'-t5' degradation of these mRNAs. In this mechanism, the tibiotic produced by a strain of Streptomyces bacteria, which
rate of mRNA degradation is uncoupled from the frequency of is useful fo r suppressing t he immune response in organ
translation. Thus mRNAs containing the AUUUA sequence can transplant patients . The target of rapamycin (TOR) was
be translated at high frequency yet also be degraded rapidly, identified by isolating yeast m utants resistant to rapa mycin
allowing the encoded proteins to be expressed in short bursts. inhibition of cell growth. TOR is a large (-2400 amino acid
As shown in Figure 8-29, some mRNAs are degraded by residue) protein kinase that regulates several cellular pro-
an endonucleolytic pathway that docs not involve decapping cesses in yeast cells in response to nutritional status. In mul-
or significant deadenylation. One example of this type of path- ticellular eukaryotes, metazoan TOR (mTOR) also responds
way is the RNAi pathway discussed above (see Figure 8-25). to multiple signals from cell-surface-signaling proteins to co-
Each siRNA-RISC complex can degrade thousands of targeted ordinate cell growth with developmental programs as well as
RNA molecules. The fragments generated by internal cleavage nutritional status.
then are degraded by exonucleases. Current understanding of the mTOR pathway is sum-
marized in Figure 8-30. Active mTOR stimulates the overall
P Bodies As mentioned above, P bodies are sires of transla-
rate of protein synthesis by phosphorylating two critical pro-
tional repression of mRNAs bound by miRNA-RISC com-
teins that regulate translation directly. mTOR also activates
plexes. They arc also the major sites of mRNA degradation in
t ranscription factors that contr o l exp ression of ribosomal
the cytoplasm. These dense regions of cytoplasm contain the
components, tRNAs, and tra nslation fac tors, furthe r activat-
dccapping enzyme (Dcp 1/Dcp2 in yeast), activators of decap-
ing protein synthesis and cell growth.
ping (Dhh, Patl, Lsml-7 in yeast), the major 5'-t3' exonucle-
Recall that the first step in translation of a eukaryotic
ase (Xrnl), as well as densely associated mRNAs. P bodies arc
mRNA is binding of the ei F4 initiation complex to t he 5 ' cap
dynamic structures that grow and shrink in size depending on
via its eiF4E cap-bind ing subunit (see Figure 4-24). The con-
the rate at which mR.t"iPs associate with them, the rate at which
centration of active elF4E is regulated by a small family of
mRNAs are degraded, and the rate at which mRNPs exit P
homologous elf4E-binding proteins (4E-BPs) that inhibit
bodies and reenter the pool of translated mRNPs. mRNAs
Lhe iutera<.:tion of e!F4E with mRNA 5' caps. 4£-BPs are di-
whose translation is inhibited by imperfect base-pa iring of
rect targets of mTOR. When phosphorylated by mTOR, 4E-
miRNAs (Figure 8-25) are major components of P-bodies.
BPs release elF4E, stimulating translation initiation. mTOR
also phosphorylates and activates another protein kinase
Protein Synthesis Can Be Globally Regulated (56K) that phosphorylates the small ribosomal subun it pro-
Like proteins involved in other processes, translation initia- tein S6 and probably additional substrates, leading to a fur-
tion factors and ribosomal proteins can be regulated by post- ther increase in the rate of protein synthesis.

376 CHAPTER 8 • Post-transcriptional Gene Control

 Growth factor
receptor
•••
Nutrients

Cytoplasm

n utrients

~~/\
Protein
synthesis
Rib osom e Pol Ill M acroautophagy
bio g enesis transcript ion

FIGURE 8-30 mTOR pathway. mTOR is an active protein mTOR protein kinase activity. Low nutrient concentration also
kinase when bound by a complex of Rheb and an associated GTP regulates Rheb GTPase activity, by a mechanism that does not require
(lower /eft).ln contrast, mTOR is inactive when bound by a complex TSClfTSC2. Active mTOR phosphorylates 4E-BP, causing it to release
of Rheb associated with GDP (lower right). When active, the TSCl fTSC2 eiF4E, stimulating translation initiation. It also phosphorylates and
Rheb-GTPase activating protein (Rheb-GAP) causes hydrolysis of Rheb- activates 56 kinase (S6K), which in turn phosphorylates ribosomal
bound GTP to GDP, thereby inactivating mTOR. The TSClfTSC2 proteins, stimulating translation. Activated mTOR also activates
Rheb-GAP is activated (arrows) by phosphorylation by AMP kinase transcription factors for RNA polymerases I, II, and Ill, leading to
(AMPK) when cellular energy charge is low and by other cellular stress synthesis and assembly of ribosomes, tRNAs, and translation factors.
responses. Signal-transduction pathways activated by cell-surface In the absence of mTOR activity, all of these processes are inhibited. In
growth factor receptors lead to phosphorylation of inactivating sites contrast, activated mTOR inhibits macroautophagy, which is stimulated
on TSClfTSC2, inhibiting its GAP activity. Consequently, they leave a in cells with inactive mTOR. [Adapted from S. Wullschleger et al., 2006, Cell
higher fraction of cellular Rheb in the GTP conformation that activates 124:471.]

Translation of a specific subset of mRNAs that have a string translation factor genes. Finally, mTOR stimulates process-
of pyrimidines in their 5' untranslated regions (called TOP ing of the rRNA precursor (Section 8.5). As a consequence of
mRNAs for tract of oligopyrimidine) is stimulated particularly phosphorylation of these several mTOR substrates, the syn-
strongly by mTOR. The T0P mRNAs encode ribosomal pro- thesis and assembly of ribosomes as well as the synthesis of
teins and translation elongation factors. mTOR also activates translation factors and tRNAs are greatly increased. Alterna-
the RNA polymerase I transcription factor TIF-lA, stimulating tively, when mTOR kinase activity is inhibited, these sub-
transcription of the large rRNA precursor (see Figure 7-52). strates become dephosphorylated, greatly decreasing the rate
mTOR activates transcription by RNA polymerase III as of protein synthesis and the production of ribosomes, transla-
well, by phosphorylating and thereby activating protein ki- tion factors, and tRNAs, thus halting cell growth.
nases that phosphorylate MAFl, a protein inhibitor of RNA mTOR activity is regulated by a monomeric small G pro-
polymerase III transcription. MAFl phosphorylation causes it tein in the Ras protein family called Rheb. Like other small G
to be exported from the nucleus, relieving repression of RNA proteins, Rheb is in its :=~ctive conformation when it is bound
polymerase III transcription. When mTOR activity falls , to GTP. Rheb·GTP binds the mTOR complex, stimulating
MAFl in the cytoplasm is rapidly dephosphorylated and im- mTOR kinase activity, probably by inducing a conformation
ported into the nucleus where it represses transcription by change in its kinase domain. Rheb is in turn regulated by a
RNA polymerase III. heterodimer composed of subunits TSCl and TSC2, named
In addition, mTOR activates two RNA polymerase II ac- for their involvement in the medical syndrome tuberous scle-
tivators that stimulate transcription of ribosomal protein and rosis complex, as discussed below. In the active conformation,

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 377

the TSClfTSC2 heterodimer functions as a GTPase activat- High mTOR protein kinase activity in tumors correlates
ing protein for Rheb (Rheb-GAP), causing hydrolysis of the with a poor clinical prognosis. Consequently, mTOR inhibi-
Rheb-bound GTP to GDP. This converts Rheb to its GOP- tors are currently in clinical trials to test their effectiveness for
bound conformation, which binds to the mTOR complex treating cancers in conjunction with other modes of therapy.
and inhibits its kinase activity. Finally, the activity of the Rapamycin and other structurally related mTOR inhibitors
TSClfTSC2 Rheb-GAP is regulated by several inputs, allow- are potent suppressors of the immune response because they
ing the cell to integrate different cellular signaling pathways to inhibit activation and replication ofT lymphocytes in response
control the overall rate of protein synthesis. Signaling from cell- to foreign antigens (Chapter 23). Several viruses encode pro-
surface growth factor receptors leads to phosphorylaLiun of teins that activate mTOR early after viral infection. The result-
TSClfTSC2 at inhibitory sites, causing an increase in Rheb·GTP ing stimulation of translation has an obvious selective advantage
and activation of mTOR kinase activity. This type of regula- for these cellular parasites. •
tion through cell-surface receptors links the control of cell
growth to developmental processes controlled by cell-cell eiF2 Kinases eiF2 kinases also regulate the global rate of cel-
interactions. lular protein synthesis. Figure 4-24 summarizes the steps in
mTOR activity also is regulated in response to nutritional translation initiation. Translation initiation factor e!F2
status. When energy from nutrients is not sufficient for cell brings the charged initiator tRNA to the small ribosome sub-
growth, the resulting fall in the ratio of ATP to AMP concen- unit P site. e!F2 is a trimeric G protein and consequently
trations is detected by AMP kinase (AMPK). The activated exists in either a GTP-bound or a GDP-bound conforma-
AMP kinase phosphorylates TSC1fTSC2 at activating sites, tion. Only the GTP-bound form of elf2 is able to bind the
stimulating its Rheb-GAP activity and consequently inhibit- charged initiator tRNA and associate with the small ribo-
ing mTOR kinase activity and the global rate of translation. somal subunit. The small subunit with bound initiation fac-
Hypox1a and other cellular stresses also activate the TSCll tors and charged initiator tRNA then interacts with the eiF4
TSC2 Rheb-GAP. Finally, the concentration of nutrients in complex bound to the 5' cap of an mRNA via its eiF4E sub-
the extracellular space also regulates Rheb, by an unknown unit. The small ribosomal subunit then "scans down the
mechanism that does not require the TSClfTSC2 complex. mRNA in the 3' direction until it reaches an AUG initiation
In addition to regulating the global rate of cellular pro- codon that can base-pair with the initiator tRNA in its P site.
tein synthesis and the production of ribosomes, tRNAs, and When this occurs, the GTP bound by eiF2 is hydrolyzed to
' '
translation factors, mTOR regulates at least one other pro- GDP and the resulting elF2 ·GDP complex is released. GTP
cess involved in the response to low levels of nutrients: mac- hydrolysis results in an irreversible "proofreading" step that
roautophagy (or simply autophagy). Starved cells degrade prepares the small ribosomal subunit to associate with the
cytoplasmic constituents, including whole organelles, to sup- large subunit only when an initiator tRNA is properly bound
ply energy and precursors for essential cellular processes. in the P site and is properly base-paired with the AUG start
During this process a large, double-membrane structure en- codon. Before ciF2 can participate in another round of ini-
gulfs a reg10n ~f cytoplasm to form an autophagosome, tiation, its bound GDP must be replaced with a GTP. This
which then fuses with a lysosome where the entrapped pro- process is catalyzed by the translation initiation factor eiF2B,
teins, lipids, and other macromolecules are degraded, com- a guanine nucleotide exchange factor (GEF) specific for eiF2.
pleting the process of macroautophagy. Activated mTOR A mechanism for inhibiting general protein synthesis in
inhibits macroautophagy in growing cells when nutrients are stressed cells involves phosphorylation of the eiF2 o: subunit
plentiful. Macroautophagy is stimulated when mTOR activ- at a specific serine. Phosphorylation at this site does not in-
ity falls in nutrient-deprived cells. terfere with ciF2 function in protein synthesis directly.
Rather, phosphorylated eiF2 has very high affinity for the
Genes encoding components of the mTOR pathway elF2 guanine nucleotide exchange factor, eiF2B, which can-
are mutated in many human cancers, resulting in cell not release the phosphorylated eiF2 and consequently is
growth in the absence of normal growth signals. TSCl and blocked from catalyzing GTP exchange of additional elf2
TSC2 (see Figure 8-30) were initially identified because one factors. Since there is an excess of eiF2 over eiF2B, phos-
or the other of the proteins is mutant in a rare human genetic phorylation of a fraction of elF2 results in inhibition of all
syndrome: tuberous sclerosis complex. Patients with this dis- the cellular eiF2B. The remaining ciF2 accumulates in its
order develop benign tumors in multiple tissues. The disease GDP-bound form, which cannot participate in protein syn-
results because inactivation of either TSC1 or TSC2 elimi- thesis, thereby inhibiting nearly all cellular protein synthesis.
nates the Rheb-GAP activity of the TSC1!TSC2 heterodi- However, some mRNAs have 5' regions that allow transla-
mer, resulting in an abnormally high and unregulated level tion initiation at the low elF2-GTP concentration that rc
of Rheb·GTP and the resulting high, unregulated activity of suits from elf2 phosphorylation. These mRNAs include
mTOR. Mutations in components of cell-surface receptor those for chaperone proteins that function to refold cellular
signal-transduction pathways that lead to inhibition ofTSCl/ proteins denatured as the result of cellular stress, additional
TSC2 Rheb-GAP activity are also common in human tumors proteins that help the cell to cope with stress, and transcrip-
and contribute to cell growth and replication in the absence tion factors that activate transcription of the genes encoding
of normal signals for growth and proliferation. these stress-induced proteins.

378 CHAPTER 8 • Post-transcriptional Gene Control

 Human cells contain four eiF2 kinases that phosphor- Control of intracellular iron concentration by the iron
ylate the same inhibitory eiF2a serine. Each of these is regu- response element-binding protein (JRE-BP) is an elegant ex-
lated by a different type of cellular stress, inhibiting protein ample of a single protein that regulates the translation of one
synthesis and allowing cells to divert the large fraction of mRNA and the degradation of another. Precise regulation
cellular resources usually devoted to protein synthesis in of cellular iron ion concentration is critical to the cell. Mul-
growing cells for use in responding to the stress. tiple enzymes and proteins contain Fe2+ as a cofactor, such as
The GCN2 (general control non-dercpressible 2) elf2- enzymes of the Krebs cycle (see Figure 12-1 0) and electron-
kinase is activated by binding uncharged tRNAs. The con- carrying proteins involved in the genera non of ATP by mito-
centration of unch::~rged tRNAs increases when cells are chonJria and chloroplasts (Chapter 12). On the other hand,
starved for amino acids, activating GCN2 eiF2-kinase activ- excess Fe2 generates free radicals that react with and dam-
ity and greatly inhibiting protein synthesis. age cellular macromolecules. When intracellular iron stores
PEK (pancreatic eiF2 kinase) is activated when proteins are low, a dual-control system operates to increase the level
translocated into the endoplasmic reticulum (ER) do not fold of cellular iron; when iron is in excess, the system operates
properly because of abnormalities in the ER lumen environ- ro prevent accumulation of toxic levels of free ions.
ment. Inducers include abnormal carbohydrate concentration, One component in this system is the regulation of the
because this inhibits the glycosylation of many ER proteins, production of ferritin, an intracellular iron-binding protem
and inactivating mutations in an ER chaperone required for that binds and stores excess cellular iron. The 5' untranslated
proper folding of many ER proteins (Chapters 13 and 14). region of ferritin mRNA contains iron-response elements
Heme-regulated inhibitor (HRI) is activated in develop- (IREs) that have a stem-loop structure. The IRE-binding pro-
ing red blood cells when the supply of the heme prosthetic tein (IRE-BP) recognizes five specific bases in the IRE loop
group is too low to accommodate the rate of globin protein and the duplex nature of the stem. At low iron concentra-
synthesis. This negative feedback loop lowers the rate of tions, IRE-BP is in an active conformation that binds to the
globin protein synthesis until it matches the rate of heme syn- IREs (Figure 8-Jla). The bound IRE-BP blocks the small ri-
thesis. HRI is also activated in other types of cells in response bosomal subunit from scanning for the AUG start codon (see
._ to oxidative stress or heat shock . Figure 4-24 ), thereby inhibiting translation initiation. The
Finally, protein kinase RNA activated (PKR) is activated resulting decrease in ferritin means less iron is complexed
by double-stranded RNAs longer than -30 base pairs. Under with the ferritin and therefore more is available to iron-
normal circumstances in mammalian cells, such double- requiring enzymes. At high iron concentrations, IRE-BP is in
stranded RNAs are produced only during a viral infection. an inactive conformation that does not bind to the 5' IREs,
Long regions of double-stranded RNA are generated in repli- so translation initiation can proceed. The newly synthesized
cation intermediates of RNA viruses or from hybridization of ferritin then binds free iron ions, preventing their accumula-
complementary regions of RNA transcribed from both tion to harmful levels.
strands of DNA virus genomes. Inhibition of protein synthe- The other part of this regulatory system controls the im-
sis prevents the production of progeny virions, protecting port of iron into cells. In vertebrates, ingested iron is carried
neighboring cells from infection. Interestingly, adenoviruses through the circulatory system bound to a protein called
evolved a defense against PKR: they express prodigious transferrin. After binding to the transferrin receptor (TfR) tn
amounts of an -1 60-nucleotide virus-associated (VA) RNA the plasma membrane, the transferrin-iron complex is
with long double-stranded hairpin regions. VA RNA is tran- brought into cells by receptor-mediated endocytosis (Chap-
scribed by RNA polymerase !II and exported from the nu- ter 14 ). The 3' untranslated region of TfR mRNA contains
cleus by Exportin5, the exportin for pre-miRNAs (see Figure IREs whose stems have AU-rich destabilizing sequences (Fig-
8-27). VA RNA binds to PKR with high affinity, inhibiting ure 8-Jlb). At high iron concentrations, when the IRE-BP is
its protein kinase activity and preventing the inhibition of in the inactive, nonbinding conformation, these AU-rich se-
protein synthesis observed in cells infected with a mutant quences promote degradation of TfR mRNA by the same
adenovirus from which the VA gene was deleted. mechanism that leads to rapid degradation of other short-
lived mRNAs, as described previously. The resulting de-
crease in production of the transferrin receptor quickly
Sequence-Specific RNA-Binding Proteins
reduces iron import, thus protecting the cell from excess
Control Specific mRNA Translation iron. At low iron concentrations, however, IRE-BP can bind
In contrast to global mRNA regulation, mechanisms have to the 3' IREs in TfR mRNA. The bound IRE-BP blocks
also evolved for controlling the translation of certain specific recognition of the destabilizing AU-rich sequences by the
mRNAs. This is usually done by sequence-specific RNA- proteins that would otherwise rapidly degrade lhc mRNAs.
binding proteins that bind to a particular sequence or RNA As a result, production of the transferrin receptor increases
structure in the mRNA. When binding is in the 5' untrans- and more iron is brought into the cell.
lated region (5' UTR) of an mRNA, the ribosome's ability to Other regulated RNA-binding proteins may also func-
scan to the first initiation codon is blocked, inhibiting trans- tion to control mRNA translation or degradation, much like
lation initiation. Binding in other regions can either promote the dual-acting IRE-BP. For example, a heme-sensitive RNA-
or inhibit mRNA degradation. binding protein controls translation of the mRNA encoding

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 379

(a) Ferritin mRNA their degradation by the exosome, and the general restriction
IREs coding reg1on COOH against nuclear export of incompletely spliced pre-mRNAs

~
that remain associated with a spliceosome.
High iron H2 "! Another mechanism called nonsense-mediated decay
5' ~B!!IIIl!II!!!D- An ~ ~ (NMD) causes degradation of mRNAs in wh ich o ne or more
exons have been incorrectly spliced. Such incorrect splicing
Inactive IRE-BP ~ Translated
often will alter the open reading frame of the mRNA 3' to the

Active IRE-BP Q 1 ferritin improper exon junction, resulting in introduction of an out-of-

frame missense mutation and an incorrect stop codon. For
nearly all properly spliced mRNAs, the stop codon is in the last
exon. The process of nonsense-mediated decay resu lts in the
Low iron rapid degradation of mRNAs with stop codons that occur be-
fore the last splice junction in the mRNA since in most cases,
5' L --llir!'mlnl!mlml- An -If--+ No translation
such mRNAs arise from errors in RNA splicing. However,
initiation

v
NMD can also result from a mutation creating a stop codon
(b) TfR mRNA IREs within a gene or a frame-shifting deletion or insertion. NMD

..::.:~~;.. gJ1
AU-rich region was initially discovered during the'study of patients with 13°-
~ ,,,~~ thalasemia, who produce a low level of 13-globin protein asso-
ciated with a low level of 13-globin mRNA (Figure 8-32a, b).
5 A ---+ ',,.,.,
n 1"':,'- J 1 A search for possible molecular signals that might indi-
( .1( - ,

Inactive IRE-BP ~ I,.,,. --:._, cate the positions of splice junctions in a processed mRNA

01
Degraded led to the discovery of exon-junction complexes. As noted
mononucleotides
already, these complexes of several proteins. (including Y14,
Active IRE-BP
Magoh, eiF4IIIA, UPF2, UPF3, and REF), bind -20 nucle-
otides 5' to an exon-exon junction following RNA splicing

~ ~ ~
(Figure 8-32c), stimulate export of mRNPs from the nucleus
Low iron
by interacting with the mRNA exporter (see Figure 8-21 ).
5' .ii!ffiii.I.WQ An -If--+ Little Analysis of yeast mutants indicated that one of the proteins
degradation
in exon-junction complexes (UPF3) functions in nonsense-
FIGURE 8 -31 Iron-dependent regulation of mRNA translation mediated decay. In the cytoplasm, this component of exon-
and degradation. The iron response element-binding protein (IRE-BP) j unction complexes interacts with a protein (UPFl) and a
controls (a) translation of ferritin mRNA and (b) degradation of protein kinase (SMGl) that phosphorylates UPFl, causing
transferrin-receptor (TfR) mRNA. At low intracellular iron concentra-
the mRNA to associate with P bodies, repressing translation
tions IRE-BP binds to iron-response elements (IREs) in the 5' or 3'
of the mRNA. An additional protein (UPF2) associated with
untranslated regiori of these mRNAs. At high iron concentrations,
the mRNP complex binds a P-body associated deadenylase
IRE-BP undergoes a conformational change and cannot bind either
mRNA. The dual control by IRE-BP precisely regulates the level of free
that rapidly removes the poly(A) tail from an associated
iron ions within cells. See the text for discussion.
mRNA, leading to its rapid decapping and degrada tion by
the P-body associated 5' ~3' exonuclease (see Figure 8-29).
In the case of properly spliced mRNAs, the exon-junction
complexes associate with the nuclear cap-binding complex
aminolevulinate (ALA) synthase, a key enzyme in the synthe- (CBP80, CBP20) as the mRNP is transported through a nu-
sis of heme. Similarly, in vitro studies have shown that the clear pore complex, thereby protecting the mRNA from deg-
mRNA encoding the milk protein casein is stabilized by the radation. The exon-junction complexes arc thought to be
hormone prolactin and rapidly degraded in its absence. dislodged from the mRNA by passage of the first "pioneer"
ribosome to translate the mRNA. However, for mRNAs with
a stop codon before the final exon junction, one or more
Surveillance Mechanisms Prevent Translation
exon-junction complexes remain associated with the mRNA,
of Improperly Processed mRNAs resulting in nonsense-mediated decay (Figure 8-32).
Translation of an improperly processed mRNA could lead to
production of an abnormal protein that interferes with the Localization of mRNAs Permits Production
gene's normal function. This effect is equiva lent to that re
of Proteins at Specific Regions
suiting from dominant-negative mLJtations, discussed in
Chapter 5 (Figure 5-44). Several mechanisms collectively Within the Cytoplasm
termed mRNA surveillance help cells avoid the translation of Many cellular processes depend on localization of particular
improperly processed mRNA molecules. We have previously proteins to specific structures or regions of the cel l. In later
mentioned two such surveillance mechanisms: the recogni- chapters we examine how some proteins are transported after
tion of improperly processed pre-mRNAs in the nucleus and their synthesis to their proper cellular location. Alternatively,

380 CHAPTER 8 • Post-transcriptional Gene Control

 (a) (c) EJC
Exon-exon
CG deletion CBP80
Initiation of
. CBP20
p1oneer round m7 lip')
~-Globin of translation Gppp
genomic -20-24 nt Norm Ter
t t
DNA
AUG Poly(A)
site l
(b) CBP80 and
UPF1

.. ..
wt ~- ~0- transiently
globin thalasemia or weakly
interact PTC
- + - + Act D

+-
CBP80 and UPF1
promote SMG1-UPF1
binding to eRF1-eRF3
to promote SURF
complex formation

CBP80 and UPF1

promote SMG1-UPF1
binding to a
PTC-distal EJC

l
SMG1 CBP80
phosphorylates CBP20
m Gppp llip')-~--11!11-lllild ..I!II~AAAAA
l
7
UPF1
t t t
AUG PTC Norm Ter

Translational repression and mRNA decay

FIGURE 8-32 Discovery of nonsense-mediated mRNA decay (3°-thalasemia had much less (3-globin mRNA than the patient with
(NMD). (a) Patients with (3°-thalasemia express very low levels of a wild-type (3-g lobin gene (- Act D). The mutant (3-globin mRNA
(3-globin mRNA. A common cause of this syndrome is a single-base- decayed rapidly when transcription was inhibited ( + Act D), whereas
pair deletion in exon 1 or exon 2 of the (3-globin gene. Ribosomes the wild-type (3-globin mRNA remained stable. (c) Current model of
translating the mutant mRNA read out of frame following the deletion NMD. PTC. premature termination codon. Norm Term, the normal
and encounter a stop codon in the wrong reading frame before they termination codon. SURF, complex of protein kinase SMG1 , UPF1 , and
translate across the last exon junction in the mRNA. Consequently, translation termination factors eRF1 and eRF3. Formation of the SURF
they do not displace an exon-junction complex (EJC) from the mRNA. complex leads to phosphorylation of UPF1 and association of
Cytoplasmic proteins associate with the EJC and induce degradation phospho-UPF1 with a UPF2-UPF3 complex bound to any exon-exon
of the mRNA. (b) Bone marrow was obtained from a patient with a junction complexes that were not displaced from the mRNA by the
wild-type (3-globin gene and from a patient with (3°-thalasemia. RNA first, pioneer ribosome to translate the message. This leads to the
was isolated from the bone marrow cells shortly after collection, or association ofthe PTe-containing mRNA with P-bodies, removal of the
30 min after incubation in media with Actinomycin D, a drug that poly(A) tail, and degradation of the mRNA. [Part (b) from L. E. Maquat et al.,
inhibits transcription. The amount of (3-globin RNA was measured 1981, Ce// 27:543. Part (c) adapted from J. Hwang et al., 2010, Mol. Ce//39:396.]
using the S1-nuclease protection method (arrow). The patient with

protein localization can be achieved by localization of mRNAs of 3000 mRNAs analyzed were localized to specific subcellular
to specific regions of the cytoplasm in which their encoded pro- regions, raising the possibility that this is a much more general
teins function. In most cases examined thus far, such mRNA phenomenon than previously appreciated.
localization is specified by sequences in the 3' untranslated re-
gion of the mRNA. A recent genomic-level study of mRNA Localization of mRNAs to the b ud in S. cerevisiae The most
localization in Drosophila embryos revealed that -70 percent thoroughly understood example o f mRNA localization occurs

8.4 Cytop lasmic Mechanisms of Post-tran sc riptional Control 381

FIGURE 8-33 Switching of mating type in haploid yeast cells. (a)
(a) Division by budding forms a larger mother cell (M) and smaller
~Bud
daughter cell (D), both of which have the same mating type as the
original cell (a in this example). The mother cell can switch mating type
during G1 of the next cell cycle and then divide again, producing two / \Division
cells ofthe opposite type (a in this example). Switching depends on
transcription of the HO gene, which occurs only in the absence of Ashl
protein. The smaller daughter cells, which produce Ash 1 protein,
MG) G) D
cannot switch; after growing in size through interphase, they dividP to
form a mother cell and daughter cell. (b) Model for restriction of
HO transcripti~shl r j
Switching /
mating-type switching to mother cells in 5. cerevisiae. Ashl protein
prevents a cell from transcribing the HO gene whose encoded protein
initiates the DNA rearrangement that results in mating-type switching
from a to ex or ex to a. Switching occurs only in the mother cell, after it
M0 G)D
separates from a newly budded daughter cell, because the Ashl /1
protein is present only in the daughter cell. The molecular basis for this
differential localization of Ash 1 is the one-way transport of ASH I mRNA
into the bud. A linking protein, She2, binds to specific 3'untranslated
88 M D
a

M
sequences in the ASH I mRNA and also binds to She3 protein. This
protein in turn binds to a myosin motor, Myo4, which moves along (b)
actin filaments into the bud. [SeeS. Koon and B. J. Schnapp, 2001, Curr.
Biology 11 :R166.]

Ash1
mRNA

>---c~~
in the budding yeastS. cerevisiae. As discussed in Chapter 7,
whether a haploid yeast cell exhibits the a or a mating type
is determined by whether a or a genes are present at the ex-
pressed MAT locus on chromosome III (see Figure 7-33). ~ She2
.... J9 ~~
Myo4 ""
The process that transfers a or a genes from the silent mating- Actin
type locus to the expressed MAT locus is initiated by a se-
quence-specific endonuclease ca lled HO. Transcription of
the HO gene is dependent on the SWI/SNF chromatin-
remodeling complex (see Chapter 7, Section 7.5). Daughter
yeast cells arising by budding from mother cells contain a
transcriptional repressor called Ash 1 (for Asymmetric syn-
thesis of HO) that prevents recruitment of the SWIISNF tion signal to which She2 binds, usually in their 3' UTR. The
complex to the HO gene, thereby preventing its transcrip- process can be visualized in live cells by the experiment
tion. The absence of Ashl from mother cells allows them to shown in Figure 8-34. RNAs can be fluorescently labeled by
transcribe the HO gene. As a consequence mother cells including in their sequence high-affinity binding sites for
switch their mating type, while daughter cells generated by RNA-binding proteins, such as bacteriophage proteins MS2
budding do not (Figure 8-33a). coat protein and bacteriophage }..N protein, that bind to dif-
Ashl protein accumulates only in daughter cells because ferent stem loops of specific sequence (Figure 8-34a). When
the mRNA encoding it is localized to daughter cells. The lo- such engineered mRNAs are expressed in budding yeast cells
calization process requires three proteins: She2 (for SWI- along with the bacteriophage proteins fused to proteins that
dependent HO expression), an RNA-binding protein that fluoresce different colors, the fusion proteins bind to these
binds specifically to a localization signal with a specific RNA specific RNA sequences, thereby labeling the RNAs that
structure in the ASHl mRNA; Myo4, a myosin motor pro- contain them with different colors. In the experiment shown
tein that moves cargos on actin filaments (see Chapter 17); in Figure 8-34b, ASHl mRNA was labeled by the binding of
and She3, which links She2 and therefore ASHl mRNA to green fluorescent protein fused to ;\N. Another mRNA local-
Myo4 (Figure 8-33b). ASHl mRNA is transcribed in the nu- ized to the bud by this system, the IST2 mRNA encoding a
cleus of the mother cell before mitosis. Movement of Myo4 component of the growing bud membrane, was labeled by
with its bound ASHl mRNA along actin filaments that ex- the binding of red fluorescent protein fused to M$2 coat
tend from the mother cell into the bud carries the ASHl protein. Video of a budding cell showed that the differently
mRNA into the growing bud before cell division. labeled ASHl and IST2 mRNAs accumulated in the same
At least 23 other mRNAs were found to be transported large cytoplasmic RNP particle containing multiple mRNAs
by the She2, She3, Myo4 ~ystem. All have an RNA localiza- in the mother cell cytoplasm, as can be seen from the merge

382 CHAPTER 8 • Post-transcriptional Gene Control

 (a) Binding sites for GFP-f..N

A
(b)

, ' , " , ' , .." , " '

ASH1 m RNA I
~ ' I
.. - ;

...
~

. M ' '.. •
:

,, .,
Binding sites for RFP-MS2
N
(/)
::E " " ' 't
" 'l " "
~ N

, ,.
1-
!a

... " ' , ".

, "
• '.. • ... • '• '
AAAA
" "
I

•'
t
1""1 ,.
I j I

') .~
I

• • ,.

EXPERIMENTAL FIGURE 8·34 Transport of mRNP particles the right, GFP-A.N and RFP-MS2 were independently visualized by using
from a yeast mother cell into the bud. (a) Yeast cells were engineered millisecond alternating laser excitation of GFP and RFP. (b) Frames from
to express an ASH1 mRNA with binding sites for the bacteriophage A.N a video of fluorescing cells are shown. The nucleus next to the large
protein in its 5' untranslated region and an IST2 mRNA with binding vacuole in the mother cell near the center of the micrographs, as well
sites for bacteriophage MS2 coat protein in its 3' untranslated region. A as nuclei in neighboring cells, was observed by green and red
fusion of green fluorescent protein to A.N protein (GFP-A.N) and a fusion fluorescence as shown in the top and middle rows. A merge of the two
of red fluorescent protein to MS2 coat protein (RFP-MS2) also were images is shown in the bottom row, which also indicates the time
expressed in the same cells. In other experiments, these fluorescently elapsed between images. An RNP particle containing both the ASH 1
tagged sequence-specific RNA-binding proteins were shown to bind mRNA with A.N-binding sites and the IST2 mRNA with MS2-binding
to their own specific binding sites engineered into the ASH1 and IST2 sites was observed in the mother cell cytoplasm in the left column of
mRNAs, and not to each others' binding sites. Both fluorescently images (arrow). The particle increased in intensity between 0.00 and
tagged proteins also contained a nuclear localization signal so that the 46.80 seconds, indicating that more of these mRNAs joined the RNP
fluorescent proteins that were not bound to their high-affinity binding particle. The RNP particle was transported into the bud between 46.80
sites in these mRNAs were transported into nuclei through nuclear and 85.17 seconds and then became localized to the bud tip.
pore complexes (see Chapter 13). This was necessary to prevent high [From 5. Lange et al., 2008, Traffic 9:1256. See this paper to v1ew the v1deo.]
fluorescence from excess GFP-A.N and RFP-MS2 in the cytoplasm. At

of the green and red fluorescent signals. The RNP particle short poly(A) tails that do not allow translation initiation.
was then transported into the bud within about one minute. Once again, large RNP particles containing multiple mRNAs
bearing localization signals form in the cytoplasm near the
Localization of mRNAs to synapses in the mammalian ner- cell nucleus. In this case, the RNP particles arc transported
vous system As mentioned earlier, in neurons, localization down the axon to synapses by kinesin motor proteins that
of specific mRNAs at synapses far from the nucleus in the cell travel down microtubules extending the length of the axon
body plays an essential function in learning and memory (Fig- (see Chapter 18). Electrical activity at a given synapse may
ure 8-35). Like the localized mRNAs in yeast, these mRNAs
contain RNA localization signals in their 3' untranslated re-
gion. Some of these mRNAs are initially synthesized with EXPERIME TAL FIGURII:: 8·35 A specific neuronal mRNA
localizes to synapses. Sensory neurons from the sea slug Ap/ysia
californica were cultured with target motor neurons so that processes
from the sensory neurons formed synapses with processes from the
motor neurons. The micrograph at the left shows motor neuron
processes visualized with a blue fluorescent dye. GFP-VAMP (green)
was expressed in sensory neurons and marks the location of synapses
formed between sensory and motor neuron processes (arrows). The
micrograph on the right shows red fluorescence from in situ hybridiza-
tion of an antlsensonn mHNA probe. Sensorin is a neurotransmitter
expressed by the sensory neuron only; sensory neuron processes are
not otherwise visualized in this preparation, but they lie adjacent to the
motor neuron processes. The in situ hybridization results indicate that
sensorin mRNA is localized to synapses. [From V. Lyles, Y. Zhao, and K. C.
Martin, 2006, Neuron 49:323.]

8.4 Cytoplasmic Mechanisms of Post-transcriptional Control 383

then stimulate the mRNAs' polyadcnylation in the region
of the synapse, activating the translation of encoded proteins • Many mRNAs are transported to specific subcellular loca-
that increase the size and alter the neurophysiological proper- tions by sequence-specific RNA-binding proteins that bind
tics of the one synapse whi le leaving unaffected the hundreds to localization sequences usually found in the 3' UTR. These
thousands of other synapses made by the neuron. RNA-binding proteins then associate directly or via interme-
diary proteins with motor proteins that carry large RNP
complexes w ith many mRNAs bearing the same localization
signal on actin or microtubule fibers to specific locations in
KEY CONCEPTS of Section 8.4 the cytoplasm.
Cytoplasmic Mechanisms of
Post-transcriptional Control
Translation can be repressed by micro RNAs (miRNAs),
which form imperfect hybrids with sequences in the 3' un-
translated region (UTR) of specific target mRNAs. 8.5 Processing of rRNA and tRNA
~ The related phenomenon of RNA interference, which Approximately 80 percent of the total RNA in rapidly grow-
probably evolved as an early defense system against viruses ing mammalian cells (e.g., cultured,HeLa cells) is rRNA, and
and transposons, leads to degradation of mRNAs that form 15 percent is tRNA; protein-coding mRNA thus constitutes
perfect hybrids with short interfer ing RNAs (siRNAs). only a small portion of the total RNA. The primary tran-
Both miRNAs and siRNAs contain 21-23 nucleotides, are scripts produced from most rRNA genes and from tRNA
generated from longer precursor molecules, and are bound genes, like pre-mRNAs, are extensively processed to yield
by an Argonaute protein and assembled into a multiprotein the mature, functional forms of these RNAs.
RNA-induced silencing complex (RISC) that either represses The ribosome is a highly evolved, complex structure (see
translation of target mRNAs or cleaves them (see figures Figure 4-23 ), optimized for its function in protein synthesis.
8-25 and 8-26). Ribosome synthesis requires t he function and coordination of
all three nuclear RNA polymerases. The 28S and 5.8S rRNAs
Cytoplasmic polyadenylation is required for translation of associated with the large ribosomal subunit and the single
mRNAs with a short poly( A) tail. Binding of a specific protein 185 rRNA of the small subunit are transcribed by RNA
to regu latory elements in their 3' UTRs represses translation
polymerase I. The 5S rRNA of the large subunit is transcribed
of these mRNAs. Phosphorylation of this RNA-binding pro- by RNA polymerase Ill, and the mRNAs encoding the ribo-
tein, induced by an external signal, leads to lengthening of somal proteins are transcribed by RNA polymerase II. In ad-
the 3' poly(A) tail and thus translation (see Figure 8-28). dition to the four rRNAs and -70 ribosomal proteins, at
Most mRNAs are degraded as the result of the gradual least 150 other RNAs and proteins interact transiently with
shortening of their poly(A) tail (deadenylation) followed by the two ribosomal subunits during their assembly through a
exosome-mediated 3' ~5· digestion, or removal of the 5' cap series of coordinated steps. Furthermore, multiple specific
and digestion by a 5'~3' exonuclease (see Figure 8-29). bases and riboses of the mature rRNAs are modified to opti-
Eukaryotic mRNAs encoding proteins that are expressed mize their function in protein synthesis. Although most of the
in short bursts generally have repeated copies of an AU-rich steps in ribosomal subunit synthesis and assembly occur in
sequence in their 3' UTR. Specific proteins that bind to these the nucleolus (a subcompartment of the nucleus not bounded
elements also interact with the deadenylating enzyme and by a membrane), some occur in the nucleoplasm during pas-
cytoplasmic exosomes, promoting rapid RNA degradation. sage from the nucleolus to nuclear pore complexes. A quality-
control step occurs before nuclear export so that only fully
Binding of various proteins to regulatory elements in the functional subunits are exported to the cytoplasm, where the
3' or 5' UTRs of mRNAs regulates the translation or degra- final steps of ribosome subunit maturation occur. tRNAs also
dation of many mRNAs in the cytoplasm. are processed from precursor primary transcripts in the nu-
Translation of ferritin mRNA and degradation of trans- cleus and modified extensively before they are exported to
ferrin receptor (TfR) mRNA are both regulated by the same the cytoplasm and used in protein synthesis. First we'll dis-
iron-sensitive RNA-binding protein. At low iron concentra- cuss the processing and modification of rRNA and the as-
tions, this protein is in a conformation that binds to specific sembly and nuclear export of ribosomes. Then we'll consider
elements in the mRNAs, inhibiting ferritin mRNA transla- the processing and modification of tRNAs.
tion or degradation of TfR mRN A (see Figure 8-31 ). This
dual control precisely regulates the iron level within cells.
Pre-rRNA Genes Function as Nucleolar
Nonsense-mediated decay and other mRNA surveillance Organizers and Are Similar in All Eukaryotes
mechanisms prevent the translation of improperly processed
mRNAs encoding abnormal proteins that might interfere The 28S and 5.8S rRNAs associated with the large (60S)
with functioning of the corresponding normal proteins. ribosomal subunit and the 18S rRNA associated with the
small (405) ribosomal subunit in higher eukaryores (a nd the

384 CHAPTER 8 • Post-transcriptional Gene Control

..
I '
 The synthesis and most of the processing of pre-rRNA
occurs in the nucleolus. When prc-rRNA genes initially were
identified in the nucleolus by in situ hybridization, it was not
known whether any other DNA was required to form the
nuc leolus. Subsequent experiments with transgenic Dro-
sophila strains demonstrated that a single complete pre-
Transcription rRNA transcription unit induces formation of a sma ll
unit
nucleolus. Thus a single pre-rRNA gene is sufficient to be a
nucleolar organizer, and all the OLher components of the ri-
bosome d iffuse to the newly formed prc-rRNA. The structure
of the nucleolus observed by light and electron microscopy
results from the processing of pre-RNA and the assembly of
ribosomal subunits.

Nont ranscribed
} spacer Small Nucleolar RNAs Assist
in Processing Pre-rRNAs
Ribosomal subunit assembly, maturation, and export to the
cytoplasm are best understood in the yeast S. cereviswe.
However, nearly al l the proteins and RNAs involved are
Transcription highly conserved in multicellular eukaryotes, where the fun-
unit damental aspects of ribosome biosynthesis are likely to be
the same. As for pre-mRNAs, nascent pre-rRNA transcripts
arc immediately bound by proteins, forming preribosomal
ribonucleoprotein particles (pre-rRNPs). For reasons not yet
known, cleavage of the pre-rRNA does not begin until tran-
EXPERIMENTAL FIGURE 8·36 Electron micrograph of scription of the pre-rRNA is nearly complete. In yeast, it
pre-rRNA transcription units from the nucleolus of a frog oocyte. takes approximately six minutes for a pre-rRNA to be tran-
Each "feather• represents multiple pre-rRNA molecules associated with scribed. Once transcription is complete, the rRNA is cleaved,
protein in a pre-ribonucleoprotein complex (pre-rRNP) emerging from and bases and riboses are modified in about 10 seconds. In a
a transcription unit. Note the dense " knob" at the 5' end of each rapidly growing yeast cell, -40 pairs of ribosomal subumts
nascent pre-RNP thought to be a processome. Pre-rRNA transcription are synthesized, processed, and transported to the cytoplasm
units are arran ged in tandem, separated by nontranscribed spacer every second. This extremely high rate of ribosome synthesis
regions of nucleolar chromatin. [Courtesy of Y. Osheim and 0. J. Miller, Jr.]
despite the seemingly long period required to transc ribe a
pre-rRNA is possible because pre-rRNA genes are packed

functionally equivalent rRNAs in all other eukaryotes) are

185 5.85 285
encoded by a single type of p re-rRNA transcription unit. In
Human, - 13.7 kb
human cells, transcription by RNA polymerase I yields a 455 3'
5'
(-13.7 kb) primary transcript (pre-rRNA), which is processed
into the mature 285, L8S, and 5.85 rRNAs found in cytoplas- X. laevis (frog), -7.9 kb
mic ribosomes. The fourt11 rRNA, 55, is encoded separately
and transcribed outside the nucleolus. Sequencing of the D. melanogaster (fruit fly), -7.7 kb
DNA encoding the 455 pre-rRNA from many species showed
that this DNA shares several properties in all eukaryotes. First, S. cerevisiae (yeast), - 6.6 kb
the pre-rRNA genes are arranged in long tandem arrays sepa-
rated by nontranscribed spacer regions ranging in length from - Transcribed spacer
- 2 kb in frogs to -30 kb in humans (Figure 8-36). Second, the
- Region preserved in rRNA
genomic regions corresponding to the three mature rRNAs
are always arranged in the same 5'~3' order: 185, 5.85, and FIGURE 8-37 General structu re of eukaryotic pre-rRNA t ran-
285. T hird, in all eukaryotic cells (and even in bacteria), the scription units. The three coding regions (red) encode the 185, 5.85,
pre-rRNA gene codes for regions that are removed during and 285 rRNAs found in ribosomes of higher eukaryotes or their
processing and rapid ly degraded. These regions probably equivalents in other species. The order of these coding regions in the
contribute to proper folding of the rRNAs but are notre- genome is always 5 '~3'. Variations in the lengths of the transcribed
quired once the folding has occurred. The general structure of spacer regions (blue) account for the major difference in the lengths
pre-rRNAs is diagrammed in Figure 8-37. of pre-rRNA transcription units in different organisms.

8.5 Processing of rAN A and tRNA 385

with RNA polymerase I molecules transcribing the same involved in protein synthesis. The positions of the specific
gene simultaneously (see Figure 8-36) and because there are sites of 2' -0-methylation and pseudouridine formation are
I 00-200 such genes on chromosome XU, the yeast nucleolar determined by approximately 150 different small nucleolus-
organizer. restricted RNA species, called small nucleolar RNAs
The primary transcript of~ 7 kb is cut in a series of cleav- (snoRNAs), which hybridize transiently to pre-rRNA mole-
age and cxonuclcolytic steps that ultimately yield the mature cules. Like the snRNAs that function in prc-mRNA processing,
rRNAs found in nbosomes (Figure 8-38). During processing, snoRNAs associate with proteins, forming ribonucleopro-
pre-rRNA also is extensively modified, mostly by methyla- tein particles called snoRNPs. One class of more than 40
tion of the 2'-hyJroxyl group of specific riboses and conver- snoRNPs (contaming box C+D snoRNAs) positions a meth-
sion of specific uridine residues to pseudouridine. These yltransferase enzyme near methylation sites in the pre-
post-transcriptional modifications of rRNA are probably mRNA. The multiple different box C+ D snoRNAs direct
important for protein synthesis, because they are highly con- methylation at multiple sites through a similar mechanism.
served. Virtually all of these modifications occur in the most They share common sequence and structural features and
conserved core structure of the ribosome, which is directly are bound by a common set of proteins. One or two regions

Primary
transcript
,
5- - - -- - - -- - - -
~ 3'----+ e ,
.. Rat1
Co transcr~pt1onal l
..,,.'11~-rucleolvt.c cleavage

355 ------------~---------------

l
•• ·~vie
,_ C+D s~r- · ·-
L.a.o..ArA

V CH3 CH 3

~ ~~
t ----~-----------------
1 ~ I
355
Exosome • -t') ~
1 Cleav

335 J£~----------------------­
Xrn1 r >••
Rat1 \.....>
~ Cleavag.e l

'\.IR., qJl
325 ------~--------

- - - - - - -- 275A2
Xrn1
Rat1
G·· e

Xrn1 ~

Cleavage on
cytoplasm
Rat1 ~ -1--------- 275A3
Exxuclease prccesson:. 1
27585 --:l 275BL ~:+------
•

Q ·· :essong
vage
ress.ng
svage
Xrn1
75 5 - ··D 75L - ··t)
l Exosom~ l Exosom)

- + - +
185 5.855 255 or 5.85L 255
FIGURE 8-38 rRNA processing. Endoribonuc!eases that make rRNAs occurs following the initial cleavage at the 3 ' end, before the
internal cleavages are represented as scissors. Exoribonuc!eases that initial cleavage at the 5' end. Proteins and snoRNPs known to partici-
digest from one end, either 5' or 3', are shown as Pac-Men. Most pate in these steps are indicated. [From J. Venema and D. Tollervey, 1999,
2' -0-ribose methylation (CH 3) and generation of pseudouridines in the Ann. Rev. Genetics 33:261.)

386 CHAPTER 8 • Post-transcriptional Gene Control

 OH OH
Uridine

Box H Box ACA

OH OH
Pseudouridine

FIGURE 8 - 39 snoRNP-directed modification of pre-rRNA. (a) A in the stems. Pre-rRNA hybridizes to the single-stranded bulges,
snoRNA called box C +D snoRNA is involved in ribose 2'-0-methylation. demarcating a site of pseudouridylation. (c) Conversion from uridine
Sequences in this snoRNA hybridize to two different regions in the to pseudouridine directed by the box H + ACA snoRNAs of part (b ).
pre-rRNA, directing methylation at the indicated sites. (b) Box H + ACA [Part (a) from T. Kiss, 2001 , EMBOJ. 2 0:3617. Part (b) from U. T. Meier, 2005,
snoRNAs fold into two stem loops with internal single-stranded bulges Chromosoma 114:1 .]

of each of these snoRNAs are precisely complementary to Some snoRNAs are expressed from their own promoters
sites on the pre-rRNA and direct the methyltransferase to by RNA polymerase II or III. Remarkabl}, however, the
specific riboses in the hybrid region (Figure 8-39a). A second large majority of snoRNAs are processed from spliced-out
major class of snoRNPs (containing box H + ACA snoRNAs) intron s of genes encoding functional mRNAs for proteins
positions the enzyme that converts uridine to pseudouridine involved in ribosome synthesis or translation. Some snoRNAs
(Figure 8-39b). This conversion involves rotation of the py- are processed from imrons spliced from apparently nonfunc-
rimidine ring (Figure 8-39c). Bases on either side of the mod- tional mRNAs. The genes encoding these mRNAs seem to
ified uridine in the pre-rRNA base-pair with bases in the exist only to express snoRNAs from excised introns.
bulge of a stem in the H + ACA snoRNAs, leaving the modified Unlike 185, 5 .85, and 285 genes, 55 rRNA genes are
uridine bulged out of the helical double-manded region, like the transcribed by RNA polymerase Ill in the nucleoplasm out-
branch point A bulges out in pre-mRNA spliceosomal splicing side the nucleolus. With only minor additional processing to
(see Figure 8-1 0). Other modifications of pre-rRNA nucle- remove nucleotides at the 3' end, 55 rRNA diffuses to the
otides, such as adenine dimethylation, are carried out by spe- nucleolus, where it assembles with the pre-rRNA precursor
cific proteins without the assistance of guiding snoRNAs. and remains associated with the region that is cleaved into
The U3 snoRNA is assembled into a large snoRNP contain- the precursor of the large ribosomal subunit.
ing -72 proteins called the small subunit (SSU) processomc, Most of the ribosomal proteins of the small 40S ribosomal
which specifies cleavage at site A 0 , the initial cut near the 5' subunit associate with the nascent pre-rRNA during transcrip-
end of the pre-rRNA (see Figure 8-38). U3 snoRNA base-pairs tion (figure 8-40). Cleavage of the full-length pre-rRNA in the
with an upstream region of the pre-rl~.NA to specify the loca- 90S RNP precursor releases a pre-405 particle that requires
tion of the cleavage. The processome is thought to form the only a few more remodeling steps before it is transported to the
"5' knob" visible in electron micrographs of pre-rRNPs (see cytoplasm. Once the pre-405 particle leaves the nucleolus, it
Figure 8-36). Base pairing of other snoRNPs specify additional traverses the nucleoplasm quickly and is exported through nu-
cleavage reactions that remove transcribed spacer regions. The clear pore complexes (NPCs), as discussed below. final matu-
first cleavage to initiate processing of t he 5.85 and 255 rRNAs ration of t he small ribosomal subunit occurs in the cytoplasm:
of the large subunit is performed by RNase M RP, a complex exonucleol) tic processing of the 205 rRNA into mature small
of nine proteins with an RNA. Once cleaved from pre-rRNAs, subunit 185 rRNA by the cytoplasmic 5' ~3 ' exoribonuclease
these sequences arc degraded by the same exosome-associated Xrn 1 and the dimethylation of two adjacent adenines near the
3'~5' nuclear exonucleases that degrade introns spliced from 3' end of 185 rRNA by the cytoplasmic enzyme Diml.
pre-mRNAs. Nuclear 5'~3' exoribonucleases (Ratl; Xrnl) In contrast to the pre-405 particle, the precursor of the
also remove some regions of 5' spacer. large subunit requires considerably more remodeling through

8.5 Processing of rRNA and tRNA 387

 Nucleolus Nucleoplasm Cytoplasm

Early Intermediate Late Mature

rONA

SSU processome

rRNA C) Helicases
RNA polymerase I 0 Intranuclear transport (Noc proteins)

0 () U3-associated factors 0 GTPases

¥ U3-snoRNP o- AAA-type ATPase

0 rRNA processing/modification factors Export factors (Nmd3, Nxt1, Ran-GTP)

FIGURE 8-40 Ril?osomal subunit assembly. Ribosomal proteins subunits in the cytoplasm. Other factors that associate transiently with
and RNAs in the maturing small and large ribosomal subunits are the maturing subunits are depicted in different colors, as shown in the
depicted in blue, with a shape similar to the icons for the mature key. [From H. Tschochner and E. Hurt, 2003, Trends Cell Bio/.13:255.]

many more transient interactions with nonribosomal pro- occur in the nucleoplasm, during passage from the nucleolus
teins before it is sufficiently mature for export to the cyto- to nuclear pore complexes (see Figure 8-40). Much remains
plasm. Consequently, it takes a considerably longer period to be learned about the complex, fascinating, and essential
for the maturing 60S subunit to exit the nucleus (30 minutes remodeling processes that occur during formation of the ri-
compared to 5 minutes for export of the 40S subunit in cul- bosomal subunits.
tured human cells). Multiple presumptive RNA helicases and The large ribosomal subunit is one of the largest struc-
small G proteins are associated w ith the maturing pre-60S tures to pass through nuclear pore complexes. Maturation of
subunits. Some RNA helicases arc necessary to dislodge the the large subunit in the nucleoplasm leads to the generation
snoRNPs that base-pair perfectly with pre-rRNA over up to of binding sites for a nuclear export adapter called Nmd3 .
30 base pairs. Other RNA helicases may function in the dis- Nmd3 is bound by the nuclear transporter Exportinl (also
ruption of protein-RNA interactions. The requirement for so called Crml ). This is another quality-control step, because
many GTPases suggests that there are many quality-control only correctl y assembled subunits can bind Nmd3 and be ex-
checkpoints in the assembly and remodeling of the large sub- ported. The small ~ubunit of the mRNP exporter (NXT 1) also
unit RNP, where one step must be completed before a GTPase becomes associated with the nearly mature large ribosomal
is activated to allow the next step to proceed. Members of the subunit. These nuclear transporters interact with FG-domains
AAA ATPase family are also bound transiently. This class of of FG-nucleoporins. This mechanism allows penetration of
proteins is often involved in large molecular movements and the molecular "cloud" that fills most of the central channel of
may be required to fold the complex, large rRNA into the the NPC (sec Figure 8-20). Several specific nucleoporins
proper conformation. Some steps in 60S subunit maturation without FG-domains are also required for ribosomal subunit

388 CHAPTER 8 • Post-transcriptional Gene Control

 export and may have additional functions specific for this A whole raft of self-splicing sequences subsequently were
task. The dimensions of ribosomal subunits (- 25-30 nm in found in pre-rRNAs from other single-celled organisms, m
diameter) and the central channel of the NPC are compara- mitochondrial and chloroplast pre-rRNAs, in several pre-
ble, so passage may not require distortion of either the ribo- mRNAs from certain E. coli bacteriophages, and in some
somal subunit or the channel. Final maturation of the large bacterial tRNA primary transcripts. The self-splicing se-
subunit in the cytoplasm includes removal of these export quences in all these precursors, referred to as group I m-
factors. As for the export of most macromolecules from the trons, use guanosine as a cofactor and can fold by internal
nucleus, including tRNAs and pre-miRNAs (but not most base pairi ng to juxtapose closely the two exons that must be
mRNPs), ribosome ~uhunit export requires the function of a joined. As discussed earl ier, certain mitochondrial and chlo-
small G protein called Ran, as discussed in Chapter 13. roplast pre-mRNAs and tRNAs contain a second type of
self-splicing intron, designated group 11.
The splicing mechanisms used by group I introns, group
Self-Splicing Group llntrons Were the First
II introns, and spliceosomes are generally similar, involving
Examples of Catalytic RNA two transesterification reactions, which require no input of
During the 1970s, the pre-rRNA genes of the protozoan Tet- energy (Figure 8-4 1). Structural studies of the group [ intron
rahymena thermophda were discovered to contain an intron. from Tetrahymena pre-rRNA combined with mutational and
Careful searches failed to uncover even one pre-rRNA gene hiochemical experiments have revea led that the RNA folds
without the extra sequence, indicating that splicing is required into a precise three-dimensional structure that, like protein
to produce mature rRNA in t hese organisms. In 1982, in vitro enzymes, contains deep grooves for binding substrates and
studies showing that the pre-rRNA was spliced at the correct solvent-inaccessible regions that function in catalysis. The
sites in the absence of any protein provided the first indication group I intron functions like a metalloenzyme to precisely
that RNA can function as a catalyst, like enz}mes. orient the atoms that participate in the two transestcrification

Self-splicing introns Spliceosome-catalyzed splicing

of pre-mRNA
Group I Group II

a
Spliceosome

.•
5'
P~p
G

3' 5'
r1 HO
P+------'p/
A

3'
~

3'

1 1 1
__Q_G_3'0H
~p
D-
--• o~P ..;

1 1 1
p~ C\ _Q
- P HO --- p HO 3'

FIGURE 8-41 Splicing mechanisms in group I and group II involving th~ 2'-hydroxyl groups of branch-s1te As in group II introns
self-splicing introns and spliceosome-catalyzed splicing of and pre-mRNA introns spliced in spliceosomes {see Figure 8-8). The
pre-mRNA. The intron is shown in gray, the exons to be joined in red. subsequent transesterification that links the 5' and 3' exons is similar in
In group I introns, a guanosine cofactor {G) that is not part of the RNA all three splicing mechanisms. Note that spliced-out group I introns are
chain associates with the active site. The 3'-hydroxyl group of this linear structures, unlike the branched intron products in the other two
guanosine participates in a transesterification reaction with the cases. (Adapted from P. A. Sharp, 1987, Science 235:769.]
phosphate at the 5' end of the intron; this reaction is analogous to that

8.5 Processing of rRNA and tRNA 389

reactions adjacent to catalytic Mg 2 ions. Considerable evi- RNase P polypeptide increases the rate of cleavage by the RNA,
dence now indicates that splicing by group II introns and by allowing it to proceed at physiological Mg2 + concentrations. A
snRNAs in the spliceosome also involves bound catalytic comparable RNase P functions in eukaryotes.
Mg2 -'- ions. In both the group I and II self-splicing introns and About 10 percent of the bases in pre-tRNAs are modified
probably in the spliceosome, RNA functions as a ribozyme, enzymatically during processing. Three classes of base modifi-
an RNA sequence with catalytic ability. cations occur (Figure 8-42): (1) U residues at the 3' end of
pre-tRNA are replaced with a CCA sequence. The CCA se-
quence is found at the 3' end of all tRNAs and is required for
Pre-tRNAs Undergo Extensive Modification
their charging by aminoacyl-tRNA synthetases during protein
in the Nucleus synthesis. This step in tRNA synthesis likely functions as a
Mature cytosolic tRNAs, which average 75-80 nucleotides in quality-control point, since only properly folded tRNAs arc
length, arc produced from larger precursors (pre-tRNAs) syn- recognized by the CCA addition enzyme. (2) Methyl and iso-
thesized by RNA polymerase I1I in the nucleoplasm. Mature pentenyl groups are added to the heterocyclic ring of purine
tRNAs also contain numerous modified bases that are not bases, and the 2' -OH groups in the ribose of specific residues
present in tRNA primary transcripts. Cleavage and base are methylated. (3) Specific uridincs are converted to dihy-
modification occur during processing of all pre-tRNAs; some drouridine, pseudouridine, or ribothymidine residues. The
pre-tRNAs also are spliced during processing. All of these functions of these base and ribose modifications are not well
processing and modification events occur in the nucleus. understood, but since they are highly conserved, they proba-
A 5' sequence of variable length that is absent from mature bly have a positive influence on protein synthesis.
tRNAs is present in all prc-tRNAs (Figure 8-42). This occurs As shown in Figure 8-42, the pre-tRNA expressed from the
because the 5' end of mature tRNAs is generated by an endo- yeast tyrosine tRNA (tRNAT>r) gene contains a 14-base intron
nucleolrtic cleavage specified by the tRNA three-dimensional that is not present in mature tRNATrr. Some other eukaryotic
structure rather than the start site of transcription. These extra tRNA genes and some archaeal tRNA gene!! also contain in-
5' nuclcotides are removed by ribonuclease P (RNase P), a ri- trans. The introns in nuclear pre-tRNAs arc shorter than those
bonucleoprotein endonuclease. Studies with E. coli RNase P in pre-mRNAs and lack the consensus splice-site sequences
indicate that at high Mg2 concentrations, the RNA compo- found in pre-mRNAs (see Figure 8-7). Pre-tRNA introns also
nent alone can recognize and cleave E. coli pre-tRNAs. The are clearly distinct from the much longer self-splicing group I

3' 3'
OH OH

Processing

Mature tRNATyr

FIGURE 8-42 Changes that occur during the processing of the stem loops are converted to characteristic modified bases (yellow).
tyrosine pre-tRNA. A 14-nucleotide intron (blue) in the anticodon Not all pre-tRNAs contain introns that are spliced out during process-
loop is removed by splicing. A 16-nucleotide sequence (green) at the ing, but they all undergo the other types of changes shown here.
5' end is cleaved by RNase P. U residues at the 3' end are replaced by D = dihydrouridine; ljJ = pseudouridine.
the CCA sequence (red) found in all mature tRNAs. Numerous bases in

390 CHAPTERs • Post-transcriptional Gene Control

and group II intro ns found in chloroplast and mitochondrial 3 kDa 10 kDa 40 kDa 40 kDa 70 kDa 500 kDa 2000 kDa

pre-rRNAs. The mechanism of pre-tRNA splicing differs in

three fundamental ways from the mechanisms utilized by self-
splicing introns and spliceosomes (see Figure 8-41). First, splic-
ing of pre-tRNAs is catalyzed by proteins, not by RNAs.
Second, a pre-tRNA intron is excised in one step that entails
simultaneous cleavage at both ends of the intron. Finally, hy-
drolysis of GTP and ATP is required to join the two rRNA FIGURE 8·43 Nuclear bodies are different ially permeable t o
halves generated by cleavage on either side of the intron. molecules in the bulk nucleoplasm. Each pair of panels shows a single
After pre-tRNAs are processed in the nucleoplasm, the area through a living Xenopus oocyte nucleus that was previously
mature tRNAs are transported to the cytoplasm through nu- injected with fluorescent dextran of the indicated molecular mass
clear pore complexes by Exportin-t, as discussed previously. (3-2000 kDa). Each section ofthe upper panel is a confocal image in
In the cytoplasm, tRNAs are passed between aminoacyl- which the intensity of fluorescence is a measure of dextran concentra-
rRNA synthetases, elongation factors, and ribosomes during tion (i.e., darker areas show regions where dextran has been excluded).
protein synthesis (Chapter 4). Thus tRNAs generally are as- Each section of the lower panel is a differential interference contrast
sociated with proteins and spend tittle time free in the cell, as image of the same field. Open arrowheads indicate nucleoli, closed
is also the case for mRNAs and rRNAs. arrowheads Cajal bodies (CBs) with attached nuclear speckles which are
much larger in Xenopus oocytes than in most somatic cells. Dextrans of
low molecular mass (e.g., 3 kDa) almost completely penetrated CBs but
Nuclear Bodies Are Functionally Specialized were excluded more from nuclear speckles and nucleoli. Exclusion of
Nuclear Domains dextran increased with molecular mass. Bar= 10 f.l.m. [From
K. E. Handwerger et al., 2005, Mol. Bioi. Cell 16:202.]
High- resolution visualization of plant- and animal-cell nu-
clei by electron microscopy and subsequent staining with
fluorescently labeled antibod ies has revealed domains in nu-
clei in addition to chromosome territories and nucleoli. methyl groups to the 2'-hydroxyl groups of specific riboses.
These specialized nuclear domains, called nuclear bod ies, are T hese post-transcriptional modifications a·rc important for
not surrounded by membranes but are nonetheless regions the proper assembly and function of snRNPs in pre-mRNA
of high concentrations of specific proteins and RNAs that splicing. These modifications occur in Cajal bodies, where
form distinct, rough ly spherical structures within the nu- they are directed by a class of snoRNA-like guide RNA mol-
cleus. T he most prominent nuclear bodies are nucleoli, the ecules called scaRNAs (small Cajal body-associated RNAs ).
sites of ribosomal subunit synthesis and assembly discussed There is also evidence that the Cajal body is the site of reas-
earlier. Several other types of nuclear bodies also have been sembly of the U4/U6/U5 tri-snRNP complexes required for
described in structural studies. pre-mRNA splicing from the free U4, US, and U6 snRNPs
Experiments with fluorescenrly labeled nuclear proteins released during the removal of each intron (see Figure 8-11 ).
have shown that the nucleus is a highly dynamic environ- Since Cajal bodies also contain a high concentration of the
ment, with rapid diffusion of proteins through the nucleo- U7 snRNP involved in the specialized 3'-end processing of
plasm. Proteins associated with nuclear bodies are often also the major histone mRNAs, it is likely that this process also
observed at lower concentrations in the nucleoplasm outside occurs in Cajal bodies, as may the assembly of the telom-
the nuclear bodies, and fluorescence studies indicate that erase RNP.
they diffuse in and out of the nuclear bodies. Based on these
measurements of molecular mobility in living cells, nuclear Nuclear Speckles Nuclear speckles were observed, using
bodies can be mathematically modeled as the expected fluorescently labeled antibodies to snRNP proteins and other
steady state for diffusin~ proteins that interact with suffi- proteins involved in pre-mRNA splicing, as approximately
cient affinity to form self-organized regions of high concen- 25-50 irregular, amorphous structures 0.5-2 1-Lm in diame-
trations of specific proteins but with low enough affinity for ter that are distributed through the nucleoplasm of verte-
each other to be able to diffuse in and out of the structure. In brate cells. Since speckles are not located at sites of
electron micrographs these structures appear to be a hetero- co-transcriptional pre-mRNA splicing, which are associated
geneous, sponge! ike network of interacting components. We closely with chromatin, they are thought to be storage re-
discuss a few examples of nuclear bodies here. gions for snRNPs and proteins involved in pre-mRNA splic-
ing that are released into the nucleoplasm when required.
Cajal Bodies Cajal bodies are -0.2-11-Lm spherical structures
that have been observed in large nuclei for more than a cen- Promyelocytic leukemia (PML) Nuclear Bodies The PML
tury (Figure 8-43 ). Current research indicates that like nucle- gene was originally discovered when chromosomal translo-
o li, Cajal bodies are centers of RNP-complex assembly for cations within the gene were observed in the leukemic cells
spliceosomal snRNPs and other RNPs. Like rRNAs, snRNAs of patients with the rare disease promyelocytic leukemia
undergo specific modifications, such as the conversion of (PML). When antibodies specific for the PML protein were
specific uridine residues to pseudouridine and addition of used in immunofluorescence microscopy studies, the protein

8.5 Processing of rRNA and tRNA 391

was found to localize to -10-30 roughly spherical regions
0.3-1 1-1-m in diameter in the nuclei of mammalian cells. • Synthesis and processing of pre-rRNA occur in the nucleo-
Multiple functions have been proposed for these PML nu- lus. The SS rRNA component of the large ribosomal subunit
clear bodies, but a consensus is emerging that they function is synthesized in the nucleoplasm by RNA polymerase III.
as sites for the assembly and modification of protein com- • Approximately 150 snoRNAs, associated with proteins in
plexes involved in DNA repair and the induction of apopto- snoRNPs, base-pair with specific sites in pre-rRNA where
sis. For example, the important p53 tumor suppressor they direct ribose methylation, modification of uridine to
protein appears to be post-translationally modified by phos- pseudouridine, and cleavage at specific sites during rRNA
phorylation and acetylation in PMI nuclear bodies in re- processing in the nucleulu~.
sponse to DNA damage, increasing its ability to activate the
• Group I and group II self-splicing introns and probably
expression of DNA-damage response genes. PML nuclear
snRNAs in spliceosomes all function as ribozymes, or cata-
bodies are also required for cellular defenses against DNA
lytically active RNA sequences, that carry out splicing by
viruses that are induced by interferons, proteins secreted b)
analogous transesterificarion reactions requiring bound Mg2
virus-infected cells and T-lymphocytes involved in the im-
ions (see Figure 8-41).
mune response (sec Chapter 23).
PML nuclear bodies arc also sites of protein post- • Pre-tRNAs synthesized by RNA polymerase Jll in the nu-
translational modificatton through the addition of a small, cleoplasm are processed by remo":al of the 5' -end sequence,
ubiquitin-like protein called SUMOl (small ubiquitin-like addition of CCA to the 3' end, and modification of multiple
moiety- I), which can control the activity and subcellular lo- internal bases (see Figure 8-42).
calization of the modified protein. Many transcriptional acti- • Some pre-tRNAs contain a short intron that is removed by
vators are mhibited when they are sumoylated, and mutation a protein-catalyzed mechanism distinct from the splicing of
of their sire of sumoylation increases their activity in stimu- pre-mRNA and self-splicing introns.
lating transcription. These observations indicate that PML
• All species of RNA molecules are associ a red with proteins
nuclear bodies arc involved in a mechanism of transcrip-
in various types of ribonucleoprotein particles, both in the
tional repression that remains to be studied and thoroughly
nucleus and after export to the cytoplasm.
understood.
Nuclear bodies are functionally specialized regions in the nu-
Nucleolar Functions in Addition to Ribosomal Subunit Syn- cleus where interacting proteins form self-organized structures.
Many of these, like the nucleolus, are regions of assembly of
thesis The first nuclear bodies to be observed, the nucleoli,
RNP complexes.
.. ·.
may have specialized regions of substructure that arc dedi-
cated to functions other than ribosome biogenesis. There is
evidence that immature SRP ribonucleoprotein complexes
involved in protein secretion and ER membrane insertion
(Chapter 13) are assembled in nucleoli and then exported to
Perspectives for the Future
the cytoplasm, where their final maturation takes place. The
Cdc14 protein phosphatase that regulates processes in the In this and the previous chapter, we have seen that in eu-
final stages of mitosis is sequestered in nucleoli in yeast cells karyotic cells, mRNAs are synthesized and processed in the
until chromosomes have been properly segregated into the nucleus, transported through nuclear pore complexes to the
bud (Chapter 19). Also, a tumor suppressor protein called cytoplasm, and then, in some cases, transported to specific
ARF, which is involved in the regulation of the protein en- areas of the cytoplasm before being translated by ribosomes.
coded by the most frequently mutated gene in human can- Each of these fundamental processes is carried out by com-
cers, p5 3, is sequestered in nucleoli and released in response plex macromolecular machines composed of scores of pro-
to DNA damage (Chapter 24). In addition, heterochromatin teins and in many cases RNAs as well. The complexity of
often forms on the surface of nucleoli (Figure 6-33), suggest- these macromolecular machines ensures accuracy in finding
ing that proteins associated with nucleoli also participate in promoters and splice sites in the long length of DNA and
the formation of this repressing chromatin structure. RNA sequences and provides various avenues for regulating
synthesis of a polypeptide chain. Much remains to be learned
about the structure, operation, and regulation of such
KEY CONCEPTS of Section 8.5 complex machines as spliceosomes and rhe cleavage/poly-
adenylation apparatus.
Processing of rRNA and tRNA Recent examples of the regulation of pre-mRNA splicing
A large precursor pre-rRNA ( 13.7 kb in humans) synthe- raise the question of how extracellular signals might control
sized by RNA polymerase I undergoes cleavage, exonucleo- such events, especially in the nervous system of vertebrates.
lytic digestion, and base modifications to yield mature 28S, A case in point is the remarkable situation in the chick inner
185, and 5.8S rRNAs, which associate with ribosomal pro- car, where multiple isoforms of the Ca 2 + -activated K chan-
teins into ribosomal subunits. nel called S/o are produced by alternative RNA splicing.
Cell-cell interactions appear to inform cells of their position

392 CHAPTER 8 • Post- transcriptional Gene Control

 in the cochlea, leading to alternative splicing of Slo pre-mRNA.
The challenging task facing researchers is to discover how
Key Terms
such cell-cell interactions regulate the activity of RNA- 5' cap 349 poly(A) tail 358
processing facrors. alternative splicing 361 pre-mRNA 349
The mechanism of mRNP transport through nuclear pore pre-rRNA 385
cleavage/polyadenylation
complexes poses many intriguing questions. Future research
complex 358 ribozyme 390
will likely reveal additional activities of hnRNP and nuclear
mRNP proteins and clarify their mechanisms of action. For cross-exon recognition RNA editing 364
instance, there i<> a small gene family encoding proteins ho- complex 356 RNA-induced silencing
mologous to the large subunit of the mRNA exporter. What Dicer 371 complex (RISC) 3 71
are the functions of these related proteins? Do they partici- Drosha 371 RNA interference (RNAi) 373
pate in the transport of overlapping sets of mRNPs? Some exosome 359 RNA splicing 348
hnRNP proteins contain nuclear-retention signals that pre- FG-nucleoporins 365 short interfering RNAs
vent nuclear export when fused to hnRNP proteins with (siRNA) 346
group I introns 357
nuclear-export signals (NESs). How are these hnRNP proteins
selectively removed from processed mRNAs in the nucleus, group II introns 357 siRNA knockdown 374
allowing the mRNAs to be transported to the cytoplasm? iron-response element- small nuclear RNAs
The localization of certain mRNAs to specific subcellular binding protein (snRNAs) 352
locations is fundamental to the development of multicellular (IRE-BP) 379 small nucleolar RNAs
organisms. As we will discuss in Chapter 21, during develop- micro RNAs (miRNAs) 370 (snoRNAs) 386
ment an individual cell frequently divides into daughter cells mRNA surveillance 380 spliceosome 353
that function differently from each other. In the language of mRNP exporter 365 SR proteins 356
developmental biology, the two daughter cells are said to
nuclear pore complex
have different developmental fates. In many cases, this dif-
(NPC) 365
ference in developmental fate results from the localization of
an mRNA to one region of the cell before mitosis so that
after cell division, it is present in one daughter cell and not
the other. Much exciting work remains to be done to fully
Review the Concepts
understand the molecular mechanisms controlling mRNA
localization that are critical for the normal development of 1. Describe three types of post-transcriptional regulation of
multicellular organisms. protein-coding genes.
Some of the most exciting and unanticipated discoveries 2. True or False? The CTD is responsible for mR.t'\IA-processing
·. in molecular cell biology in recent years have concerned the steps that are specific for mRNA, and not other forms of RNA.
existence and function of miRNAs and the process of RNA Explain why you chose true or false.
interference. RNA interference (RNAi) provides molecular
3. There are a number of conserved sequences found in an
cell biologists with a powerful method for studying gene
mRNA which dictate where splicing occurs. Where are these
function. The discovery of -500 miRNAs in humans and
sequences found relative to the exon/intron junctions? What
other organisms suggests that multiple significant examples
is the significance of these sequences in the splicing process?
of translational control by this mechanism await characteri -
One of these important regions is the branch point A found
zation. Recent studies in S. pombe and plants link similar
in the intron. What is the role of the branch pmnt A in the
short nuclear R.NAs to the control of DNA methylation and
splicing process, and can this be accomplished with the OH
the formation of heterochromatin. Will similar processes
group on either the 2' or the 3' carbon?
control gene expression through the assembly of heterochro-
matin in humans and other animals? What other regulatory 4. What is the difference between hnRNAs, snRNAs,
processes might be directed by other kinds of small RNAs? miRNAs, siRNAs, and snoRNAs?
Since control by these mechanisms depends on base pairing 5. What are the mechanistic similarities between group II
between miRNAs and target mRNAs or genes, genomic and intron self-splicing and spliceosomal splicing? What is the
bioinformatic methods will probably suggest genes that may evidence that there may be an evolutionary relationship be-
be controlled by these mechanisms. What other processes in tween the rwo?
addition to translation control, mRNA degradation, and 6. You obtain the sequence of a gene containing 10 exons, 9
heterochromatin assembly might be controlled by miRNAs? introns, ;Jnd a 3' UTR containing a polyadenylation consen-
These are just a few of the fascinating questions concern- sus sequence. The fifth intron also contains a polyadenyla-
ing RNA processing, post-transcriptional control, and nu- tion site. To test whether both polyadenylation sites are
clear transport that will challenge molecular cell biologists in used, you isolate mRNA and find a longer transcript from
the coming decades. The astounding discoveries of entirely muscle tissue and a shorter transcript from all other tissues.
unanticipated mechanisms of gene control by miRNAs re- Speculate about the mechanism involved in the production
mind us that many more surprises are likely in the future. of these different transcripts.

Review the Concepts 393

7. RNA editing is a common process occurring in the mito- expresses a Pst-Miu fragment of the LAT gene (see diagram in
chondria of trypanosomes and plants, in chloroplasts, and in part b). The percentage of these transfected cells that then un-
rare cases in higher eukaryotes. What is RNA editing, and derwent drug-induced cell death was compared to that of con-
what benefit does it demonstrate in the documented example trol cells. The experiment was repeated in cells in which Dicer
of apoB in humans? expression was knocked down using Dicer siRNA. The data
8. A1> DNA is found in the nucleus, transcription is a nuclear- obtained are shown in the graph below. What conclusions can
localized process. Ribosomes responsible for protein synthe- be drawn from these data? Why did the scientists who con-
sis are found in the cytoplasm. Why is hnRNP trafficking to ducted this study examine the effects of silencing Dicer?
the cytoplasm restncted to the nuclear pore complexes? How
do the rG-repeats of the nuclear pore complexes act as a
specificity barrier in nuclear transport? • Cells transfected with
(/)
60 control expression vector
9. A protein complex in the nucleus is responsible for trans- (/)·en
=o
Q)- • Cells transfected with LA T
porting mRNA molecules into the cytoplasm. Describe the (.)0.
expression vector
-0 o
a.
proteins that form this exporter. What two protein groups Q) 0)
are likely behind the mechanism involved in the directional ClCl
0) c
40 0 Cells transfected with LA T
expression vector and
movement of the mRNP and exporter into the cytosol. c:Q)C)
·o expressing control siRNA
(.) ~
~ Q)
10. RNA knockdown has become a powerful tool in the arse- C>-o
a..c 0 Cells transfected with LA T
nal of methods ro deregulate gene expression. Briefly describe ::J
20 expression vector and
expressing Dicer siRNA
how gene expression can be knocked down. What effect would
introducing siRNAs to TSC I have on human cells?
11. Speculate about why plants deficient in Dicer activity b. Cells were transfected with an expression vector ex-
show increased sensitivity to infection by RNA viruses. pressing the Pst-Miu fragment of the LA T gene from which
12. mRNA sta bility is a key regulator of protein levels in a the region between the two Sty restriction sites was deleted
cell. Briefly describe the three mRNA degradation pathways. (DSty; diagram below). When these cells were induced to
A yea~t cell has a mutation in the DCPl gene, resulting in undergo apoptosis, they died at the same rate as did non-
decreased uncapping acti\ity. Would you expect to see a transfected cells. In additional studies, cells were transfected
change in the P bodies found in this mutant cell? with an expression vector expressing the Sty-Sty region of
13. mRNA localization now appears to be a common phe- the LA T gene. These cells exhibited the same resistance to
nomenon. What benefit docs mRNA localization have for a apoptosis as did cells transfected with the Pst-Mlu fragment.
cell? What is the evidence that some mRNAs are directed to What can be deduced from these findings about the region of
accumulate in specific subcellular location s? the LAT gene required to protect cells from apoptosis?

Pst/Miu fragment
I<
Analyze the Data
~Region deleted in t1Sty
~1ost humans are infected with herpes simplex virus-1 (HSV-1),
Pst Mlu
the causative agent of cold sores. The HSV-1 genome com-
prises about I 00 genes, most of which arc expressed in in-
fected host cells at the site of oral sores. The infectious
process involves replication of viral DNA, transcription and
translanon of viral genes, assembly of new viral particles,
and death of the host cell as the viral progeny are released. 5' - GTGGCGGCCCGGCCCGGGGCCCCGG• G :CAAGGGGCCCCGGCCCGGGGCCCCAc- 3'
Unlike most other types of viruses, herpesvirus also has a Stem (5' arm) LOOP Stem (3' arm) I
latent phase, 10 which the virus remains hidden in neurons.
These latently infected neurons are the source of active infec- c. RNA encoded within the Sty-Sty region is predicted to
tions, causing cold sores when latency is overcome. form a stem loop (see diagram in part b). Northern blot anal-
Interestingly, only a single viral transcript is expressed ysis was performed on total-cell RNA isolated from control
during latency. This transcript, LA T (latency-associated cells (mock ), cells infected with wild-type HSV-1, cells in-
transcript), does not encode a protein, and neurons mfected fected with an HSV-1 deletion mutant from which these-
with mutant HSV - I lacking the LA T gene undergo cell death quence between the two Sty site:, in the LA T gene was deleted
by apoprosis at a rate twice that of cells infected with wild- (DSty), and cells infected with a rescued DSry virus into which
type HSV-1. To determine if LAT functions to block apop- the deleted region was re-inserted into the viral genome
tosis by encoding a miRKA, the following studies were done (StyR). The probe used for the Northern blot was the labeled
(see Gupta et al., 2006, Nature 442:82-85). 3' stem region of the LAT RNA in the Sty-Sty region, as dia-
a. A cell line was transfected (a process in which foreign grammed in part (b). The RNAs recognized by this probe
DNA is inserted into a cell) with an expression vector that were either - S 5 nucleotidcs or 20 nucleotides, as shown in

394 CHAPTER 8 • Post-transcriptional Gene Control

.·
the Northern blot below. Why were two different-sized RNAs ~loore, ~1. J., and N.J. Proudfoot. 2009. Pre-mR~ .-\ proccssmg
detected? When a second probe was used that was the labeled reache> back to transcription and ahead to translation. Ce// 136:
5' stem region of the RNA sequence shown in part (b), only 688-700.
Pawlicki, .J. ~!.,and J. A. SteltZ. 20 I 0. Nuclear networking
the -55-nucleotide RNA was detected. What can you deduce
fashions pre-messenger RNA and pnmary m1croRN \transcripts for
about the processing of RNA expressed from the LAT gene? function. Trends Cell Bioi. 20:52-61.
What enzyme likely produced the -55-nucleotide RNA? In Perales, R ., and D. Bentley. 2009. "Corranscriptionality": the
what part of t he cell ? What enzyme likely produced the 20- transcnption elongation complex as a nexu> for nuclear transac-
nucleotide RNA? In what part of the cell? tions. Mol. Cell. 36:178-191.
Pyl", A. ~1. 20 I 0. The rcrnar} ,truLture oi group 11 introns:
unplications for biological function and evolution. Crit. Ret•.
Control wt HSV-1 ll Sty Biochem. Mol. Brol. 45:21 S-232.
Sharp, P. A. 2005. The discovery of split genes a nd RNA
u;
Q) splicing. Trends Bwchem. Sci. 30:2"9-281.
-o 60
·g Valadkhan, S. 20 I 0. Role of the snR~As in sphceosomal aLnve
Q) 50 s1te. RNA Bioi. 7:345-353.
u:;,
c::
Wahl, :vi. C...., C. l.. Will, and R. Lt1hrmann. 2009. The spliceo-
'!t ~ome: design principles of a dynamic RNP machine. Cell136:701-718.
Q)
N
·u; Regulation of Pre-mRNA Processing
<(
z Black, D. I. 2003. ~l ec hamsms of alternative pre-mR]';"A
a: 20
splicing. Amz. Rev. Bwchem. 72:291-336.
Chen, M., and J. l.. Manley. 2009. ).lechanisms of alternanve
>plicing regulation: insights from molecular and genomics approaches.
d. T GF-13 mR NA encodes a protein, t ransforming N,1t. Ret•. Mol. Cell Bioi. 10:741-754.
growth factor 13, that inhibits cell growth and induces apop- l.icaralos1, D. D., and R. B. Darnell. 2010. RNA processing and
tosis. The 3' unrranslated region (3' UTR) of TGF-13 mRNA irs regulation: global insights IntO biologicalnerworks. Nat. Rez•.
can form an imperfect duplex with miRNA encoded by the Genet. 11:7 5-87.
~laniatis, T., and B. Tasic_ 2002. Alternative prc-mR!\A splicing
5' stem region of the LA T Sty-Sty domain (miR-LAT), as
and proteome expansion 111 metazoans. Nature 418:236-243.
shown below. In what way might the expression levels of Raponi, ~!.,and D. Baralle. 20 I 0. Alternative splicing: good
TGF-13 differ in cell s infected with wild-type HSV-1 com- and bad effec t~ of translatlonall} silent subsmutions. FEBS .f.
pared to uninfected cells? What can you infer about latent 2 77:836-840.
HSV- 1 infections from these stud ies? \X1ang, E. T., et al. 2008. Alternative 1soform regulation in
human tissue transcriptomes. Nature 456:470-476.
Zhong, X. Y., et al. 2009. SR protems 111 \'Crtlcal inregration of
c gene n.pression from transcription to RNA processmg to transla-
miR-LAT 1\
C C : G ~ G ~ C ~ ~ ~ C _ C ' ~ 5' tion. Curr. Opm. Genet. Dev. 19:424-436.
I I I I I I I I I I I I I I I I I I I
5' C- G-C- C- C-C- G-G G-C-C-C-G-G-C-C-C-C-A Transport of mRNA Across the Nuclear Envelope
I I TGF-[3-3' UTR
Cole, C. N., and J. .J. Scarcel11. 2006. Tra nsport of messenger
CAG
Rl\'A from the nucleus ro the cytoplasm. Curr. Opm. Cell Bioi.
18:299-306.
Grunwald, D., R. H. Singer, and M. Rout. 2011. Nuclear
References export dynamiCS of R~A -protem complexes. Nature 4 7~ :333-341.
Iglesias, N., and F. Stutz. 2008. Regulation of mRNP dynamics
Processing of Eukaryotic Pre-,mRNA along the export pathwa}. FEBS /.ett . 582:198.,-1 996.
Bergkessel, M., G. M. Wilmes, and C. Guthrie. 2009. SnapShot: Kata hira, J., andY. Yoneda. 2009. Roles of the TREX complex
formation of mRNPs. Ce/1136:"94. m nuclear export of mR~A. RNA Bwl. 6:149-1 52.
Blanc, V., and N. 0. Davidson. 2010. APOBFC-1-mediated Rodriguez-Navarro, S., and E. Hurr. 20 I I. Lmkmg gene
RNA editing. Wiley InterdisCifJ. Ret•. Syst. 810/. 'vfed. 2: 'i94-602. regulation to mRNA production and export. Curr. Opin. Cell Bwl.
de Almeida, S. F., and M. Carm o-rc)Tlseca. 2008 . The CTD role 23:302-309.
in cocra nscriptional Ri','A processing and sun·eillance. H: BS Lett. Stewart, 1\1. 20 I 0. Nuclear export of mRNA. Trends Biochem.
582: 1971-19.,6. So. 35:609-617.
Gu, M., and C. D. Lima. 2005. Processmg the mes>age: Wente, S. R., and M. P. Rout. 20 I 0. The nuclear pore complex
structural msights mto capping and decappmg mR~A. Curr. Ot1ilr. and nuclear transport. Cold Sprmg Har/1. Perspect. Bioi.
Struct. Bioi. 15:99- 106. 2( 10):a000562.
Hocine, S., R. I L Singer, and D. C..runwald. 2010. R~A process-
ing and export. Cold Spring Harb. Perspect. 810/. 2 ( 12):a000752. Cytoplasmic Mechanisms of Po st-transcriptional Control
Housele),]., and D. Tolleney. 2009. The many pathways of Ambros, V. 2004. The functions of animal microR::-..As. N,lfure
RNA degradation. Ce/1136:763-776. 431:350-355.
La mbowitz, A. \1., and S. Zimmer!}. 2004. ~1obile group II Buchan, J. R., and R. Parke r. 2009. Fukaryonc stres~ ~ranulcs:
mtrons. Amw. Reu. Genet. 38:1-35. the im and outs of translation. Mol. Ce/136:932-94 L.

References 395
 Carthew, R. W., and F. J. Sontheimcr. 2009. Origins and Stmpson, L., et al. 2004. Mitochondrial proteins and complexes
mechamsms of miRNA~ and stRNAs. Cell 136:642-655. 111Letshmania and Trypanosoma involved 111 U-insertlon/deletion
Doma, \1. K., and R. Parker. 2007. RNA quality control tn RNA editing. RNA 10:159-170.
eukaryote~. Ce/1131:660-668. Siomi, M.. C., et al. 2011. PTWI-interacting small RNAs: the
Eulalio, A., I. Behm-Ansmant, and E. Izaurralde. 2007. P bodies: vanguard of genome defence. Nat. Rev. Mol. Cell Bzol. 12:246-258.
at the crossroads of post-transcriptional pathways. Nat. Rev. Mol. Willis, I. M., and R. D. Moir. 2007. Integration of nutritional and
Cell Bzol. 8:9-22. stress signaling pathways b) Mafl. Trends Biochem. SCI. 32:51-53.
Fabian, M. R., N. Sonenberg, and W. Filipowicz. 2010. Wullschleger, S., R. Loewith, and M. N. Hall. 2006. TOR
Regulation of mRNA translation and stability by microRNAs. signaling in growth and metabolism. Cell 124:471-484.
Annu. Rev. B10chem. 79:351-3 79. Zhang, H., J. M. Mamar, and A. Z. Fire. 2011. 'lnc-miRs':
Ghildiyal, .\1., and P. D. Zamore. 2009. Small silencing RNAs: funcnonal intron-interrupted miRNA genes. Genes Dev. 25:1589-
an expanding universe. Nat. Rev. Genet. 10:94-108. 1594. A. Z. Fire's Nobel Pnze lecture can be viewed at http://
Groppo, R., and j. D. Richter. 2009. Translational control from nohclprize.org/nobel_prizes/medicine/laureatcs/2006/announcement.
head to tail. Curr. Opi11. Cell Bioi. 21:444-451. html
Hirokawa, !':. 2006. mRNA transport in dendmes: RNA Zoncu, R., A. Efeyan, and D. M. Sahatmi. 201 I. mTOR: from
granules, motors, and tracks.]. Neuroscz. 26:7139-7142. growth signal integration to cancer, diabetes and ageing. Nat. Rev.
Huntzinger, E., and E. Izaurralde. 2011. Gene silencing by Mol. Cell Bioi. 12:21-35.
microRNAs: contributions of translational repression and mRNA
decay. Nat. Rev. Genet. 12:99-110. Processing of rRNA and tRNA
Jobson, R. W., andY. L. Qiu. 2008. Did RNA edtring in plant Evans, D., S.M. \1arquez, and ::-.l. R. Pace. 2006. RNase P: interface
organellar genomes originate under natural selecnon or through of the RNA and protein worlds. Trends Biochem. Sci. 31:333-341.
genetic drift? Bioi. Direct. 3:43. Farica, A., and D. Tollervey. 2002. Making ribosomes. Curr.
Kidner, C. A., and R. A. Marrienssen. 2005. The developmental Opm. Cell Bzol. 14:313-318.
role of microRNA in plants. Curr. Opm. Plant Bioi. 8:38-44. Hage. A. E., and D. Tollervey. 2004. A surfeit of factors: why ts
Leung, A. K., and P. A. Sharp. 2010. MicroRNA functions 111 nbosome assembly so much more complicated in eukaryores than
stress responses. Mol. Cell. 40:205-2 15. bacteria? RNA Bioi. 1:10-15.
Lodish, H. F., et al. 2008. Micromanagement of the immune Hamma, T., and A. R. Ferrc-D'Amarc. 2010. The box H/ACA
system by mtcroR;'\/As. Nat. Rev. lmmzmol. 8:120-130. ribonucleoprotein complex: mterplay of RNA and protem structures 111
Maquat, L. E., W. Y. Tarn, and 0. lsken. 2010. The pioneer post-transcripnonal RNA modification.]. Bzol. Chem. 285:805-809.
round of translation: features and functions. Cell 142:368-374. Handwcrger, K. E., and J. G. Gall. 2006. Subnuclear organelles:
Marrin, K. C., and A. Ephrussi. 2009. mRNA localization: gene new insights into form and function. Trends Cell Bzol. 16:19-26.
expression in the spatial dimension. Ce/1136:719-730. Kressler, D., E. Hurt, and J. Bassler. 2010. Drinng rihosomc
Mello, C. C., and D. Conte Jr. 2004. Revealing the world of assembly. Bioclnm. Biophys. Acta. 1803:673-683.
RNA Interference. Nature 431:338-342. C. C. :V1ello\ Nobel Pme Liang, B., and H. Li. 201 1. Structures of ribonucleoprotein
lecture can he viewed at http://nobelprize.org/nobel_prizcs/medicme/ particle modification enzymes. Q. Rev. Biophys. 44:95-122.
laureates/2006/announcement.html Marvin, M. C., and D. R. Engelke. 2009. RNase P: increased
Michels, A. A. 2011. MAF1: a new target of mTORCl. versatility through protein complexity? RNA Bioi. 6:40-42.
Biochem. Soc. TraJIS. 39:487-491. Ntzami, Z., S. Deryusheva, and J. G. Gall. 2010. The Cajal
Mihaylova, M. M., and R. j. Shaw. 2011. The AMPK signalling body and histone locus body. Cold Spnng Harb. Perspect. Bioi.
pathway coordinates cell growth, autophagy and metabolism. 2(7):a000653.
Nature Cell Bioi. 13:1016-1023. Phizicky, E. M., and A. K. Hopper. 2010. tRK'A biology charges
Parker, R., and H. Song. 2004. The enzymes and control of to the front. Genes Dev. 24:1832-1860.
eukaryotic mRNA turnover. Nat. Struct. Mol. Bzol. 11:121-127. Schmid, M., and T. H. Jemen. 2008. The exosome: a multipur-
Richter, J. D., and N. Sonenberg. 2005. Regulation of cap- pose RNA-decay machine. Trends Biochem. Sci. 33:501-510.
dependent translation by eiF4E inhibitory proteins. Nature Smhley, \1. R., and S. A. Strobel. 2006. RNA splicing: group I
433:4..,--480. mtron crystal structures reveal the basts of splice site ~election and
Ruvkun, G. B. 2004. The tiny RNA world. Haruey Lect. 99:1-21. metal ton catalysis. Curr. Opm. Struct. Rio/. 16:319-326.
Shaw, R. J. 2008. mTOR signaling: RAG GTPases transmit the Tschochner, H., and E. Hurr. 2003. Pre-ribosomes on the road
amino acid signal. Trends Riochem. Sci. 33:565-568. from the nucleolus to the cytoplasm. Trends Cell Bioi. 13:255-26 ."'!.

396 CHAPTER 8 • Post-transcriptional Gene Control

 CHAPTER

Culturing, Visualizing,
and Perturbing Cells

Fluorescence microscopy shows the location of DNA and multiple

proteins within the same cell. Here fluorescent tagging and staining
techniques using different fluorescent molecules reveal the cytoskel-
etal proteins a-tubulin (green) and actin (red), DNA (blue), the Golgi
complex (yellow), and mitochondria (purple). The images along the top
are false-colored images of each structure stained individually. The
larger image merges these sepa rate images to depict the full cell.
[From B. N. G. Giepmans, et al., 2006, Science 312:217 .]

t is difficult to believe that 400 years ago, it was not yet organs, even isolated ones, is sufficiently complex to pose nu-

I known that all living things are made of cells. In L655,

Robert Hooke used a primitive microscope to examine a
piece of cork and saw an orderly arrangement of rectangles-
merous problems for research. Thus molecular cell biologists
often conduct experimental studies on cells isolated from an
organism. In Section 9. 1, we learn how to maintain and grow
the walls of the dead plant cells-that reminded him of monk diverse cell types and how to isolate specific types of cells
cells in a monastery, so he coined the term cells. Shortly after from complex mixtures. However, cells in culture are not in
this, Antonie van Leeuwenhook described the microorgan- their native setting, so we discuss how researchers are now
isms that he saw in his simple microscope, the first description growing and examining cells in three-dimensional em iron-
of living cells. Two hundred years later, Matthias Schleiden ment~ to more closely mimic their situation in an animal.
and Theodore Schwann observed that individual cells consti- In many cases, isolated cells can be maintained in the
tute the fundamental unit of life in a variety of plants, ani- laboratory under conditions that permit their survival and
mals, and single-celled organisms. Collectively, these were growth, a procedure known as wlturing. Cultured cells have
some of the greatest discoveries in biology and posed the ques- several advantages over intact organisms for cell biology re-
tion of how cells are organized and function. However, many search. Cells of a single specific type can be grown in culture,
technical constraints harrfper studies of cells in intact animals experimenta l conditions can be better controlled, and in
and plants. One alternative is the use of intact organs that are many cases a single cell can be readily grown into a colony
removed from animals and treated to maintain their physio- of many identical cells. T he resulting strain of cells, which is
logic integrity and function. However, the organization of genetically homogeneous, is called a clone.

OUTLINE

9.1 Growing Cells in Culture 398 9 .4 Isolation and Characterization of Cell Organelles 424

9.2 Light Microscopy: Exploring Cell Structure 9.5 Perturbing Specific Cell Functions 430
and Visualizing Proteins Within Cells 404

9.3 Electron Microscopy: High-Resolution Imaging 419

 Discoveries about cellular organization have been inti- Microscopy and organelle characterization are inherently
mately tied with developments in both light and electron mi- descriptive technologies. How can one investigate the mo-
croscopy. This is as true today as it was 400 years ago. Light lecular mechanisms underlying cell biological processes? If
microscopy initially revealed the beautiful internal organiza- we want to understand how a car works, we can explore the
tiOn of cells, and today highly sophisticated microscopes are effects of removing or interfering with individual components
continually being improved to probe deeper and deeper to re- to see what happens. This, of course, is in principle the con-
veal the molecular mechanism by which cells function. In Sec- cept of genetic analysis we described in Chapter 5: how inter-
tion 9.2, we discuss light microscopy and the different fering with a specific component can be used to explore its
technologies available, long standing but still valuable meth- funcLiun. In Set:tion 9.5, we describe how small molecules
ods, and then trace through several clever methods that have that interfere with the function of specific proteins can also
been developed since, culminating with the newest, cutting- be used to dissect cellular processes. Finally, we describe how
edge technologies. A major advance came in the 1960s and the discovery of small interfering RNAs that target specific
1970s with the development of immunofluorescence micros- mRNAs for destruction has been exploited to suppress ex-
copy to allow the localization of specific proteins within fixed pression of specific proteins in both cultured cells and whole
cells, thus providing a static image of their location, as illus- animals to expand and complement the use of classical genet-
trated in the opening figure. Such studies led to the important ics in the analysis of biological processes.
concept that the membranes and interior spaces of each type
of organelle contain a distinctive group of proteins that are
essential for the organelle to carry out its unique functions. A
major advance came in the mid-1990s with the simple idea of 9.1 Growing Cells in Culture
expressing chimeric proteins-<:onsisting of a protein of inter-
est covalently linked to a naturally fluorescent protein-to en- The study of cells is greatly facilitated by growing them in
able biologists to visualize the movements of individual culture, where they can be examined by microscopy and sub-
proteins in live cells. Suddenly, the dynamic nature of cells jected to specific treatments under controlled conditions. It
could be appreciated, which changed the view of cells from is generally quite easy to grow unicellular bacterial, fungal,
the previously available static images. In addition, it presented or protist cells; for example, by placing them in a rich me-
a technological challenge-the more sensitive a microscope dium that supports their growth. However, animal cells
could be made to detect the fluorescent protein, the more in- come from multicellular organisms, making it more difficult
formation the investigator could glean from the data. It also to culture single or small groups of cells. In this section, we
opened up the development of fluorescent techniques to mon- discuss how animal cells are grown in culture and how dif-
itor protein-protein interactions in living cells, as well as a ferent cell types can be purified for study.
myriad of other sophisticated molecular technologies, some of
which we also discuss in this section.
Culture of Animal Cells Requires Nutrient-Rich
Despite the amazing developments in light microscopy,
visible light provides too low a resolution to examine cells in Media and Special Solid Surfaces
ultrastructural detail. The electron microscope gives a much To permit the survival and normal function of cultured tis-
higher resolution, but the technology generally requires that sues or cells, the temperature, pH, ionic strength, and access
the cell be fixed and sectioned and so all cell movements are to essential nutrients must simulate as closely as possible the
frozen in time. Electron microscopy also allows investigators conditions within an intact organism. Isolated animal cells
to examine the structure of macromolecular complexes or are typically placed in a nutrient-rich liquid, called the cul-
single macromolecules. In Section 9.3, we outline the various ture medium, within specially coated plastic dishes or flasks.
approaches for preparing specimens for observation in the The cultures are kept in incubators in which the temperature,
electron microscope and describe the type of information atmosphere, and humidity can be controlled. To reduce the
that can be derived from them. chances of bacterial or fungal contamination, antibiotics are
Light and electron microscopy revealed that all eukaryotic often added to the culture medium. To further guard against
cells-whether of fungal, plant, or animal origin-<:ontain a contamination, investigators usually transfer cells between
similar repertoire of membrane-limited compartments dishes, add reagents to the culture medium, and otherwise
termed organelles. In Section 9.4, we provide a simple intro- manipulate the specimens within special sterile cabinets con-
duction to the basic structure and function of the major or- taining circulating air that is filtered to remove microorgan-
ganelles in animal and plant cells, as a prelude to their isms and other airborne contaminants.
detailed description in subsequent chapters. In parallel with Media for culturing animal cells must supply the nine
the developments in microscopy, subcellular fractionation amino acids (phenylalanine, valine, threonine, tryptophan,
methods were developed that have enabled cell biologists to isoleucine, methionine, leucine, lysine, and histidine) that
isolate individual organelles to a high degree of purity. These cannot be synthesized by adult vertebrate animal cells. In
techniques, also detailed in Section 9.4, continue to provide addition, most cultured cells require three other amino acids
important information about the protein composition and (cysteine, tyrosine, and arginine) that are synthesized only
biochemical function of organelles. by specialized cells in intact animals, as well as glutamine,

398 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

.·.
 which serves as a nitrogen source. The other necessary com- When cells removed from an embryo or an adult animal
·. ponents of a medium for culturing animal cells are vitamins, are cultured, most of the adherent cells will divide a finite num-
various salts, fatty acids, glucose, and serum-the fluid re- ber of times and then cease growing (cell senescence). For in-
maining after the noncellular part of blood (plasma) has stance, human fetal fibroblasts divide about 50 times before
been allowed to clot. Serum contains various protein factors they cease growth (Figure 9-1 a). Starting with 10 6 cells, 50
that are needed for the proliferation of mammalian cells in doublings has the potential to produce 106 X 2 50, or more than
culture, including the polypeptide hormone insulin; transfer- 1020 cells, which is equivalent to the weight of about 1000
rin, which supplies iron in a bioaccessible form; and numer- people. Normally, only a very small fraction of these cells are
ous growth factors. In addition, certain cell types require used in :my one experiment. Thus, even though its lifetime is
specialized protein growth factors not present in serum. For limited, a single culture, if carefully maintained, can be studied
instance, progenitors of red blood cells require erythropoietin, through many cell generations. Such a lineage of cells originat-
and T lymphocytes require interleukin 2 (see Chapter 16). A ing from one initial primary culture is called a cell strain.
few mammalian cell types can be grown in a chemically de- One important exception to the finite life of normal cells
fined, serum-free medium containing amino acids, glucose, is the embryonic stem cell, which, as its name implies, is
vitamins, and salts plus certain trace minerals, specific pro-
tein growth factors, and other components. {a) Human cells
Unlike bacterial and yeast cells, which can be grown in Phase Phase
suspension, most animal cell types will grow only attached I Phase II Ill
to a solid surface. This requirement highlights the impor-
Cell strain
tance of cell-surface proteins, called cell-adhesion molecules
(CAMs), that cells use to bind to adjacent cells and to com- Growth
ponents of the extracellular matrix (ECM) such as collagen rate of
culture
or fibronectin (see Chapter 20). These ECM proteins coat
the solid surface (usually glass or plastic) and either come Cell
from the serum or are secreted by the cells in culture. A sin- senescence
gle cell cultured on a glass or a plastic dish proliferates to I
form a visible mass, or colony, containing thousands of ge- 25 50
Cell generations
netically identical cells in 4 to 14 days, depending on the
growth rate. Some specialized blood cells and tumor cells
can be maintained or grown in suspension as single cells. (b) Mouse cells

Initial loss Emergence of

/ ofgrowth immortal
Primary Cell Cultures and Cell Strains potential variant
Have a Finite Life Span {cell line)
Growth
Normal animal tissues (e.g., skin, kidney, liver) or whole em- rate of
culture
bryos are commonly used to establish primary cell cultures.
To prepare indi'vidual tissue cells for a primary culture, the
cell-cell and cell-matrix interactions must be broken. To do
so, tissue fragments are treated with a combination of a pro-
tease (e.g., trypsin, the collagen-hydrolyzing enzyme collage- 30 60
nase, or both) and a divalent cation chelator (e.g., EDTA) that Days after initiation of culture
depletes the medium of free Ca2+ . Many cell-adhesion mole- FIGURE 9-1 Stages in the establishment of a cell culture.
cules require calcium and are thus inactivated when calcium is (a) When cells isolated from human tissue are initially cultured, some
removed; other cell-adhesion molecules that are not calcium cells die and others (mainly fibroblasts) start to grow; overall, the
dependent need to be proteolyzed for the cells to separate. The growth rate increases (phase 1). If the remaining cells are harvested,
released cells are then placed in dishes in a nutrient-rich, se- diluted, and replated into dishes again and again, the cell strain
rum-supplemented medium, where they can adhere to the sur- continues to divide at a constant rate for about 50 cell generations
face and to one another. The same protease-chelator solution (phase II), after which the growth rate falls rapidly. In the ensuing
period (phase Ill), all the cells in the culture stop growing {senescence).
is used to remove adherent cells from a culture dish for bio-
(b) In a culture prepared from mouse or other rodent cells, initial cell
chemical studies or subculturing (transfer to another dish).
death (not shown) is coupled with the emergence of healthy growing
Fibroblasts are the predominant cells in connective tissue cells. As these dividing cells arc diluted and allowed to continue
and normally produce ECM components such as collagen that growth, they soon begin to lose growth potential, and most stop
bind to cell-adhesion molecules, thereby anchoring cells to a growing (i.e., the culture goes into senescence). Very rare cells undergo
surface. In culture, fibroblasts usually divide more rapidly than oncogenic mutations that allow them to survive and continue dividing
other cells from a tissue, eventually becoming the predominant until their progeny overgrow the culture. These cells constitute a cell
cell type in a primary culture unless special precautions are line, which will grow indefinitely if it is appropriately diluted and fed
taken to remove them when isolating other types of cells. with nutrients. Such cells are said to be immortal.

9.1 Growing Cells in Culture 399

derived from an embryo and will divide and give rise to all riod, most of the cells stop growing, but often a rapidly divid-
tissues during development. As we discuss in Chapter 21, ing transformed cell arises spontaneously and takes over, or
embryonic stems cells can be cultured indefinitely under the overgrows, the culture. A cell line derived from such a trans-
appropriate conditions. formed variant will grow indefinitely if provided w ith the
Research with cell strains is simplified by t he ability to necessary nutrients. ln contrast to rodent cells, normal human
freeze and successfully thaw them at a later time for experimen- cells rarely undergo spontaneous transformation into a cell
tal analysis. Cell strains can be frozen in a state of suspended line. The Hela cell line, the first human cell line established,
animation and stored for extended periods at liquid nitrogen was originally obtained in 1952 from a malignant tumor
temperature, provided that a preservative that prevem~ the for- (carcinoma) of the utenne cervix. Other human cell lines are
mation of damaging ice crystals is used. Although not all cells often derived from cancers, and others have been rendered
survive thawing, many do survive and resume growth. immortal by transforming them to express oncogenes.
Regard less of the source, cells in immortalized lines
often have chromosomes with abnormal DNA sequences. In
Transformed Cells Can Grow
addition, the number of chromosomes in such cel ls is usu-
Indefinitely in Culture ally greater than that in the normal cel l from which they
To be able to clone individual cells, modify cell behavior, or arose, and the chromosome number changes as the cells
select mutants, biologists often want to maintain cell cul- continue to divide in culture. A noteworthy exception is the
tures for many more than 50 doublings. Such prolonged Chinese hamster ovary (CHO) line and its derivatives,
growth is exhibited by cells derived from some tumors. In which have fewe r chromosomes than their hamster progen-
addition, rare cells in a population of primary cells may un - itors. Cells with an abnormal number of chromosomes are
dergo spontaneous oncogenic mutations, leading to onco- said to be aneuploid.
genic transformation (see Chapter 24 ). Such cells, said to
be oncogenically transformed or simply transformed, arc
able ro gro\\ indefinitely. A culture of cells with an indefinite Flow Cytometry Separates Different Cell Types
life span is considered immortal and is called a cell li ne. Some cell types differ sufficiently in density that t hey can be
Primary cell cultures of normal rodent cells commonly separated on the basis of this physical property. White blood
undergo spontaneous transformation into a cell line. After cells (leukocytes) and red blood cells (erythrocytes), for in-
rodent cell5 are grown in culture for several generations, the stance, have very different densities because erythrocytes have
culture goes into senescence (Figure 9-1 b). During this pe- no nucleus; thus these cells can be separated by equi librium

Cell suspension

Sheath fluid

FIGURE 9 -2 Fluorescence-activated cell sorter

(FACS) separates cells that are labeled differen-
t ially with a fluorescent reagent. Step 0 : A Condenser
Scattered
concentrated suspension of labeled cells is mixed E1 light
with a buffer (the sheath fluid) so that the cells detector
pass single-file through a laser light beam. E1
Step fl: Both the fluorescent light emitted and the
light scattered by each cell are measured; from
measurements of the scattered light, the size and
shape of the cell can be determined. Step il: The
laser beam.............._
suspension is then forced through a nozzle, which
forms tiny droplets containing at most a single cell.
At the time of formation at the nozzle tip, each
droplet containing a cell is given a negative electric
charge proportional to the fluorescence of that cell
determined from the earlier measurement.
Step 9 : Droplets now pass through an electric
field, so that those with no charge are di!>l.drded,
whereas those with different electric charges are
• } Fluorescent cells
separated and collected. Because it takes only
milliseconds to sort each droplet, as many as • Nonfluorescent cell
10 million cells per hour can pass through the : } Fluorescent cell droplets Sorted charged
machine. [Adapted from D. R. Parks and L.A. Herzenberg, droplets containing
• Nonfluorescent cell droplet fluorescent cells
1982, Meth. Cell Bioi. 2 6:283.]

400 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

density centrifugation (described in Section 9.4). Because most 10 4 ~--------.---------------------,
cell types cannot be differentiated so easily, other techniques
such as flow cytometry must be used to separate them.
To identify one type of cell from a complex mixture, it IS
necessary to have some way to mark and then sort out the
desired cells. Different cell types often express different mol-
i
ecules on their cell surface. ff a particular surface molecule is
only expressed on the desired cell type, this can be used to
mark those cells. The cell mixture can be incubated with a
fluorescent dye linked to an antibody to the specific cell-
surface molecule, thus rendering just the desired cells fluo-
rescent. The cells can be analyzed in a flow cytometer. This
machine flows cells past a laser beam that measures the light
that they scatter and the fluorescence that they emit; thus it
can quantify the numbers of cells of the desired type from a
mixture. A fluorescence-activated cell sorter (FACS), which 102
is based on flow cytometry, can both analyze the cells and CD3
select one or a few cells from thousands of others and sort Red fluorescence~
them into a separate culture d1sh (Figure 9-2). To sort the EXPERIMENTA FIGURE 9 · " T cells bound t o fluorescence-
cells, their concentration has to be adjusted so that the tiny tagged antibodies to two cell-surface proteins are separated fro m
droplets that the FACS machine makes and analyzes contain other w hite blood cells by FACS. Spleen cells from a mouse were
only one cell each. A stream of droplets is ana lyzed for fluo- treated with a red fluorescent monoclonal antibody specific for the
rescence, and those that have the desired signal are sorted CD3 cell-surface protein and with a green fluorescent monoclonal
away from those that do not. Having been sorted from other antibody specific for a second cell-surface protein, Thy1.2. As the cells
cells, the selected cells can be grown in culture. were passed through a FACS machine, the intensity of the green and
The FACS procedure is commonly used to purify the dif- red fluorescence emitted by each cell was recorded. Each dot repre-
ferent types of white blood cells, each of which bears on its sents a single cell. This plot of the green fluorescence (vertical axis)
surface one or more d istinctive proteins and so will bind versus red fluorescence (horizontal axis) for thousands of spleen cells
monoclonal antibodies specific for that protein. Only the T shows that about half of them-the T cells-express both CD3 and
Thy1.2 proteins on their surfaces (upper-right quadrant). The remaining
cells of the immune system, for instance, have both CD3 and
cells, which exhibit low fluorescence (lower-left quadrant), express only
Thy1.2 proteins on their surfaces. The presence of these sur-
background levels of these proteins and are other types of white blood
face proteins allows T cells to be separated easily from other
cells. Note the logarithmic scale on both axes. (Courtesy of Chengcheng
types of blood cells or spleen cells (Figure 9-3). Zhang, Whitehead Institute.)
O ther uses of flow cytometry include the measurement of
a cell's DNA and RNA content and the determination of its
general shape and size. The FACS can make simultaneous
measurements 6f the size of a cell (from the amount of scat- cell types function only when closely linked to other cells.
tered light) and the amount of DNA that it contains (from Key examples are the sheet-like layers of epithelial tissue,
the amount of fluorescence emitted from a DNA-binding called epithelia (singular, epithelium), which cover the exter-
dye). Measurements of the DNA content of individual cells nal and internal surfaces of organs. Typically, the distinct
are used to follow replication of DNA as the cells progress surfaces of a polarized epithelial cell are called the apical
through the cell cycle (see Chapter 19). (top), basal (base or bottom ), and lateral (side) surfaces (see
An alternative method for separating specific types of cells Figure 20-10). The basal surface usually contacts an underly-
uses small magnetic beads coupled to antibodies for the spe- ing extracellular matrix called the basal lamina, whose com-
cific surface molecule. For example, to isolate T cells, the beads position and function are discussed in Section 20.3. Epithelial
are coated w ith a monoclonal antibody specific for a surface cells often function to transport specific classes of molecules
protein such as CD3 or Thy1.2. Only cells with these proteins across the epithelial sheet; for example, the epithelial lining of
will stick to the beads and can be recovered from the prepara- the intestine transports nutrients into the cell through the api-
tion by adhesion to a small magnet on the side of the test tube. cal surface and out toward the bloodstream across the baso-
lateral surface. When grown on plastic or glass, epithelial
Growth of Cells in Two-Dimensional cells cannot easi ly perform this function. Therefore, special
containers have been designed with a porous surface that acts
and Three-Dimensional Culture Mimics
as the basal lamina to which epithelial cells attach and form
the In Vivo Environment a uniform two-dimensional sheet (Figure 9-4). A commonly
While much has been learned using cells grown on a plastic used cultured cell line derived from dog kidney epithelium is
or glass surface, these surfaces are far removed from cells' called Mad in-Darby cmzine kidney (MDCK) cells and is often
normal tissue environment. As detailed in Chapter 20, many used to study the formation and function of epithelial sheets.

9.1 Growing Cells in Culture 401

FIGURE 9-4 Madin-Oarby canine kidney (MOCK) cells grown in
Culture dish
specialized containers provide a useful experimental system for
studying epithe lial cells. MOCK cells form a polarized epithelium
Lateral
l
when grown on a porous membrane fifter coated on one side with surface
collagen and other components of the basal lamina. With the use of
the special culture dish shown here, the medium on each side of the Basal
filter (apical and basal sides of the monolayer) can be experimentally surface
manipulated and the movement of molecules across the layer
monitored. Several cell junctions that interconnect the cells form only
if the growth medium contains sufficient Ca 2 +. Basal medium

Basal Porous Monolayer

However, even a two-dimensional sheer often does nor lamina filter of MOCK cells
allow cells to fully mimic behavior in their normal environ-
ment. Methods have now been developed ro grow cells in three
dimensions by providing a support infiltrated with compo- introducing genes encoding insulin, growth factors, and other
nents of the extracellular matrix. If MDCK cells are cultured therapeutically useful proteins into bacterial or eukaryotic cells
under appropriate conditions, they will form a tubular sheet can be used to express and recover .{hese proteins (see Figures
mimicking a tubular organ or the duct of the secretory gland. 5-31 and 5-32). Here we consider th e use of special cultured
In these three-dimensional structures, the apical aspect of the cells to generate monoclonal antibodies, which are experimen-
epithelial sheet lines the lumen, whereas the basal side of each tal tools widely used in many aspects of cell b iological research.
cell ism contact with the extracellular matrix (Figure 9-5). Increasingly, they are being used for diagnostic and therapeutic
purposes in medicine, as we discuss in later chapters.
To understand the challenge of generating monoclonal an-
Hybrid Cells Called Hybridomas Produce
tibodies, we need to briefly review how mammals produce an-
Abundant Monoclonal Antibodies tibodies; more detail is provided in Chapt e r 23. R eca ll that
In addition to serving as research models for studies on cell antibodies are proteins secreted by white blood cells that bind
function, cultured cells can be conYcrted into "factories" for w1th high affinity to their antigen (see Figure 3-19). Each nor-
producing specific proteins. In Chapter 5, we described how mal antibody-producing B lymphocyte in a mammal is capable

(a) (b)

·.

EXPERIMEN' AL FIGURE 9-5 MOCK cells can form cysts in membranes (green). these cells can be seen to be fully polarized with
culture. (a) MOCK cells grown in a supported extracellular matrix will the apical side facing the lumen, which recapitulates their organization
form groups of cells that polarize to form a spherical single layer of in the kidney tubules from which they are derived. The nuclear DNA is
cells with a lumen in the middle, called a cyst. (b) By examining the stained blue. [Parts (a) and (b) from D. M. Bryant et al., 2010, Nat. Cell Bioi.
localization of proteins found in the apical (red) and basolateral 12:1035.]

402 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 of producing a single type of antibody that can bind to a par- Inject mouse -~
ticular detenninant or epitope on an antigen molecule. An epi- with antigen X/

~
tope is generally a small region on the antigen, for example,
consisting of just a few amino acids. If an animal is injected
with an antigen, the B lymphocytes that make antibodies rec-
ognizing that antigen are stimulated to grow and secrete the
antibodies. Each antigen-activated B lymphocyte forms a clone !
of cells in the spleen or lymph nodes, with each cell of the clone
producing the identical antibody-that ic;, a nzrmoclonal anti-
body. Because most natural antigens contain multiple epitopes, •••••
••••
exposure of an animal to an antigen usually stimulates the for-
mation of multiple different B-lymphocyte clones, each pro-
ducing a different specific antibody. The resu lti ng mixture of
antibodies from rhe many B-lymphocyte clones that recognize
Mutant mouse
myeloma cells
••
Mouse spleen cells;
some cells (red) make
different epitopes on the same antigen is said to be polyclonal. unable to grow
in selective medium
D antibody to antigen X
Such polyclonal antibodies circulate in the blood and can be
isolated as a group.
Although polyclonal antibodies are very useful, monoclo-
nal antibodies are suitable for many types of experiments and
medical applications when you need a reagent that binds to
Mixand
fuse cells
• } 0
Q!
:}o
vo

just one site on a protein; for example, one that competes with
a ligand on a cell-surface receptor. Unfortunately, the bio-
chemical purification of any one type of monoclonal antibody
fJ 1 Transfer to
selective medium
from blood is not feasible for two main reasons: the concen-
tration of any given antibody is quite low, and all antibodies
have the same basic molecular architecture (sec Figure 3-19). Unfused cells
To produce a nd then purify monoclonal antibodies, one ( O e e l die
first needs to be able to grow the appropriate B-lymphocytc Fused cells
clone. I !owever, primary cultures of normal B lymphocytes ( 0 0) grow
are of limited usefulness for the production of monoclonal
antibodies because they have a limited life span. Thus the first
step in producing a monoclonal antibody is to generate im-
111 Culture single cells
in separate wells
mortal, antibody-produci ng cells (Figure 9-6). This immor-
.· tality is achieved by fusing normal B lymphocytes from an
immunized animal with transformed, immortal lymphocytes
ca lled myeloma cells that themselves synthesize neither the
heavy nor the ligh t polypeptides that constitute all antibodies
(see Figure 3-19). Treatment with certain viral glycoproteins
or the ch emical polyethylene glycol promotes the plasma
membranes of two cells to fuse, allowing their cytosols and
organelles to intermingle. Some of the fused cells undergo
Test each well for antibody to antigen X
division, and their nuclei eventually coalesce, producing via-
ble hybrid cells with a si~gle nucleus that contains chromo- FIGURE 9-6 Use of cell fusion and selection to obtain
somes from both "parents." The fusion of two cells that arc hybridomas producing monoclonal antibody to a specific protein.
genetically different can yield a hybrid cell with novel charac- Step 0 : Immortal myeloma cells that cannot synthesize purines
teristics. For instance, the fusion of a myeloma cell with a under special conditions because they lack thymidine kinase are
normal antibody-producing cell from a rat or mouse spleen fused with normal antibody-producing spleen cells from an animal
that was immunized with antigen X. StepfJ: When cultured in/on a
yields a hybrid that proliferates into a clone called a hybrid-
special selective medium, unfused and self-fused cells do not grow:
oma. Like myeloma cells, hybridoma cells grow rapidly and
the mutant myeloma cells do not grow because the selective medium
arc immortal. Each hybridoma produce~ the monoclonal an-
does not contain purines and the spleen cells because they have
tibody encoded by its B-lymphocyte parent. a limited life span in culture. Thus only fused cells formed from a
The second step in this procedure for producing monoclo- myeloma cell and a spleen cell survive in the special medium,
nal antibody is to separate, or select, the hybridoma cells from proliferating into clones called hybridomas. Each hybridoma
the unfused parental cells and the self-fused cells generated by produces a single antibody. Step 1): Testing of individual clones
the fusion reaction. This selection is usually performed by in- identifies those that recognize antigen X. After a hybridoma that
cubating the mixture of cells in a special culture medium, produces a desired antibody has been identified, the clone can be
called selection medium, that permits the growth of only the cultured to yield large amounts of that antibody.

9.1 Growing Cells in Culture 403

hybridoma cells because of their novel characteristics. The
myeloma cells used for the fusion carry a mutation that blocks • Using fluorescent antibodies to cell-surface molecules, a
a metabolic pathway, so a selection medium can be used that machine called a fluorescent-activated cell sorter can sort
is lethal to them and not their lymphocyte fusion partners that out cells with different surface markers.
do not have the mutation. In the immortal hybrid cells, the • To mimic growth in tissues, epithelial cells are often
functional gene from the lymphocyte can supply the missing grown is special containers to mimic their functional polar-
gene product, and thus the hybridoma cells will be able to ity. Cells can also be grown in three-dimensional matrices to
grow in the selection medium. Because the lymphocytes used more accurately reflect their normal environment.
in the fusion are not immortali£eJ, only the hybridoma cells
• Monoclonal antibodies, reagents that bind one epitope on
will proliferate rapidly in the selection medium and so can be
an antigen, can be secreted by cultured cells called hybridomas.
readily isolated from the initial mixture of cells. Finally, each
These hybrid cells are made by fusing an antibody-producing
selected hybridoma clone is then tested for the production of
B-cell with an immortalized myeloma cell and then identifying
the desired antibody; any clone producing that antibody is
those clones that produce the antibody. Monoclonal antibod-
then grown in large cultures, from which a substantial quan-
ies are important for basic research and as therapeutic agents.
tity of pure monoclonal antibody can be obtained.
Monoclonal antibodies have become very valuable re-
agents as specific research tools. They arc commonly em-
ployed in affinity chromatography to isolate and purify
proteins from complex mixtures (see Figure 3-38c). As we
discuss later in this chapter, they can also be employed in
9.2 Light Microscopy: Exploring Cell
immunofluorescence microscopy to bind and so locate a par- Structure and Visualizing Proteins
ticular protein within cells. They can also be used to identify With in Cells
specific proteins in cell fractions with the use of immuno-
blotting (see Figure 3-39). Monoclonal antibodies have become The existence of the cellular basis of life ~as first appreci-
important diagnostic and therapeutic tools in medicine; for ated using primitive light microscopes. Since then, progress
example, monoclonal antibodies that bind to and inactivate in cell biology has paralleled and often been driven by tech-
toxins secreted by bacterial pathogens are used to treat dis- nological advances in light microscopy (Figure 9-7). Here we
eases. Other monoclonal antibodies are specific for cell-surface discuss each of these major developments and how they ad-
proteins expressed by certain types of tumor cells. Several of vanced the study of cellular processes. First we describe basic
these anti-tumor antibodies are widely used in cancer ther- uses of a light microscope to observe unstained cells and
apy, including monoclonal antibody against a mutant form structures. Next we describe the development of fluores-
of the Her2 receptor that is overexpressed in some breast cence microscopy and its use to localize specific proteins in
cancers (see Figure 16-7) . fixed cells. By using molecular genetic approaches to express
a fusion between a protein of interest and a naturally fluo-
rescent protein, it is possible to follow the localization of
specific proteins in living cells-an ability that revealed how
KEY CONCEPTS of Section 9.1 dynamic the organization of living cells is. In parallel with
these advances in specimen preparation, optical advances
Growing Cells in Culture were being made to enhance and sharpen the images pro-
• Animal cells have to be grown in culture under conditions vided by fluorescence microscopy to reveal cellular structure
that mimic their natural environment, which generally requires in unprecedented clarity. Man y specialized technologies
them to be supplied with necessary amino acids and growth have emerged from these advances, and we describe some of
factor supplements. the more important ones.
Most animal cells need to adhere to a solid surface to grow.
Primary cells-those isolated directly from tissue-have a The Resolution of the Light Microscope
finite life span. Is About 0.2 J.lm
Transformed cells, like cells derived from tumors, can All microscopes produce a magnified image of a small ob-
grow indefinitely in culture. ject, but the nature of the image depends on the type of mi-
croscope employed and on rhe way the specimen is prepared.
Cells that can be grown indefinitely are called a cell line.
The compound microscope, used in conventional bright-
Many cells lines are aneuploid, having a different number field light microscopy, contains several lenses that magnify
of chromosomes than the parent animal from which they the image of a specimen under study (Figure 9-8a). The total
were derived. magnification is a product of the magnification of the indi-
• Different cells express different marker proteins on their vidual lenses: if the objective lens, the lens closest to the
cell surface, which can be used to distinguish them. specimen, magnifies 100-fold (a 1 OO X lens, the maximum
usually employed) and the projection lens, sometimes called

404 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 (a) (b) (c)

FIGURE 9 -7 Development of the light microscope. (a) Early sophisticated microscopes limited only by the resolution of light were
microscopes, like ones used by Robert Hooke in the 1660s, used lenses common. (c) In the second half of the twentieth century, fluorescence
or a mirror to illuminate the specimen. (b) Optics in general and light microscopy and digital imaging together with confocal techniques
microscopes in particular developed enormously during the nine- were developed to yield the versatile microscopes of today. [Part (a) SSPL
teenth century, and by the middle of the twentieth century highly via Getty Images; part (b) courtesy of Carl Zeiss Archive; part (c) Zeiss.com.]

the ocular or eyepiece, magnifies 10-fold, the final magnifi- Owing to limitations in the values of a, A., and N based on
cation recorded by the human eye or on a camera will be rhe physical properties of light, the limit of resolution of a
1000-fold. ltght microscope using visible light is about 0.2 IJ.m (200 nm).
However, the most important property of any microscope No matter how many times the image is magnified, a conven-
is not its magnification but its resolving power, or resolution- tional light microscope can ne\er resolve objects that are le~s
the ability to distinguish between two very closely positioned than - 0 .2 IJ.m apart or reveal details smaller than - 0.2 IJ.m
objects. Merely enlarging the image of a specimen accom- in size. However, some ne\\ sophisticated technologies ha\'e
plishes nothing if the image is blurred. The resolution of a mi- been devised to 'beat' this resolution barrier and can resolve
croscope lens is numerically equivalent to D, the minimum objects just a few nanometers apart; we discuss such a super-
distance between two distinguishable objects. The smaller the resolution microscope in a later section.
value of D, the better the resolution . The value of Dis given by Despite this lack of resolution, a conventional mtcroscope
the equation can track a single object to within a few nanometers. If we
know the precise size and shape of an object-say, a 5 nm
D = 0.61>.. sphere of gold that is attached to an antibod) in turn bound
(9-1)
N ~ina to a cell-surface protein on a living cell-and if we use a cam-
era to rapid!) take multiple digital images, then a computer
where a is the angular aperture, or half-angle, of the cone of can calculate the average position to reveal the center of the
light entering the objective lens from the specimen (see Figure object to within a few nanometers. In this wa:--, computer al-
9-8a), N is the refractive index of the medium between the gorithms can be used to locate smgle objects at a more precise
specimen and the objective lens (i.e., the relative velocity of level-in this case the location and movement with time of a
light in the medium compared with the velocity in air), and X. cell-surface protein labeled with the gold-tagged antibody-
is the wavelength of the incident light. Resolution is improved than would be possible ba5ed on the light microscope's reso-
by using shorter wavelengths of light (decreasing the value of lution alone. This technique has been used to measure
X.) or gathering more light (increasing either Nor a). Lenses for nanometer-size steps as molecules and vesicles move along
high-resolution microscopy are designed to work with oil be- cytoskeletal filaments (see Figures 17-29 and 17-30).
tween the lens and the specimen since oil has a higher refrac-
tive index (1.56, compared with 1.0 for air and 1.3 for water). Phase-Contrast and Differential-Interference-
To maximize the angle a, and hence sina, the lenses are also
Contrast Microscopy Visualize Unstained
designed to focus very close to the thin coverslip covering
the specimen. The term N sma is known as the numerical afJ- Living Cells
erture (NA) and is usually marked on the objective lens. A Cells are about 70 percent water, 15 percent protein, 6 percent
good high-magnification lens has anNA of about 1.4 and the RNA, and smaller amounts of lipids, DNA, and small mole-
very best lenses a value approaching 1.7, and costing as much cules. Since none of these major classes of molecules are col-
as a medium-size car! Notice that the magnification is not ored, other methods have to be used to see cells in a microscope.
part of this equation. The simplest microscope views cells under bright-field optics

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 405
(a) Optical microscope

Detector~
~ -- w 1
11

Projection--~ Excitation
lens filter

o;~~:~~:-1-7 LJ)
Objective~
rr= ,•·

Specimen
stage

Condenser--~ Collector
lens

M;no' ~~r__jj'
~~ ~ ~

(b) (c) (d )

Brightfield Phase-contrast Epifluorescence

Light
source

L- - -- Unobstructed
----.,....-~ .......- -· light - -- - Objective
lens

- Annular diaphragm
!.
•.• I

406 CHAPTER 9 • Cultu rin g, Visualizing, and Pe rturbing Cells

FIGURE 9 -8 Optical microscopes are commonly configured for light is refracted (bent) once as it passes from the medium mto
bright-field (transmitted), phase-contrast, and epifluorescence a transparent object and again when it departs. In a phase-
microscopy. (a) In a typical light microscope, the specimen is usually contrast microscope, a co ne of light generated by an annular
mounted on a transparent glass slide and positioned on the movable diaphragm in the condenser illuminates the specimen (see Fig-
specimen stage. (b) In bright-field light microscopy, light from a ure 9-8c). The light passes through the specimen into the objec-
tungsten lamp is focused on the specimen by a condenser lens below
tive, and the unobstructed direct light passes through a region
the stage; the light travels the pathway shown in yellow. (c) In
of the phase plate that both transmits only a small percentage
phase-contrast microscopy, incident light passes through an annular
diaphragm, which focuses a circular annulus (ring) of light on the
of the light and changes its phase slightly. The part of a light
sample. Light that passes unobstructed through the specimen is
wave that passes through a specimen will be refracted and will
focused by the objective lens onto the thicker gray ring of the phase be out of phase (out of synchrony) with the part of the wave
plate, which absorbs some of the direct light and alters its phase by that does not pass through the specimen. How much their
one-quarter of a wavelength. If a specimen refracts (bends) or diffracts phases differ depends on the difference in refractive index
the light, the phase of some light waves is altered (green lines) and the along the two paths and on the thickness of the specimen. The
light waves pass through the clear region of the phase plate. The refracted and unrefracted light are recombined at the image
refracted and unrefracted light are recombined at the image plane to plane to form the image. If the two parts of the light wave arc
form the image. (d) In epifluorescence microscopy, a beam of light recombined, the resu ltant light will be brighter if they are in
from a mercury lamp (gray lines) is directed to the excitation filter that phase and less bright if they are out of phase. Phase-contrast
allows just the correct wavelength light to pass (green lines). The light microscopy is suitable for observing single cells or thin cell lay-
is then reflected off a dichroic filter and through the objective that ers but not thick tissues. It is particularly useful for examining
focuses it on the sample. The fluorescent light emitted by the sample
the location and movement of larger organelles in live cells.
(red lines) passes up through the objective, then through the dichroic
DIC microscopy is based on interference between polar-
mirror and is focused and recorded on the detector at the image plane.
ized light and is the method of choice for visualizing ex-
tremely small details and thick objects. Contrast is generated
(Figure 9-8b), and little detail can be seen (Figure 9-9). Two by differences in the refractive index of the object and its
common methods for imaging live cells and unstained tissues to surrounding medium. In DIC images, objects appear to cast
generate contrast takes advantage of differences in the refractive a shadow to one side. The "shadow" primarily represents a
index and thickness of cellular materials. These methods, called difference in the refractive index of a specimen rather than
phase-contrast microscopy and differential-interference-con- its topography. DIC microscopy easily defines the outlines of
trast (DJC) microscopy (or Nomarski interference microscopy), large organelles, such as the nucleus and vacuole. In addition
produce images that differ in appearance and reveal different to having a "relief"-like appearance, a DIC image is a thin
features of cell architecture. Figure 9-9 compares images of live, OfJtical section, or slice, through the object (Figure 9-9,
cultured cells obtained with these two methods and standard right) . Thus details of the n ucleus in thick specimens (e.g., an
bright-field microscopy. Since optical microscopes are expen- intact Caenorhabditis elegans roundworm; see Figure 21-31)
sive, they are often set up to perform many different types of can be observed in a series of such optical sections, and the
microscopy on the same microscope stand (see Figure 9-8a-d). three-dimensional struct ure of the object can be recon-
Phase-cont rast microscopy generates an image in which the structed by combining the individual DIC images.
degree of darkness or brightness of a region of the sample de- Both phase-contrast and DIC microscopy can be used in
pends on the refractive index of that region. Light moves more time-lapse microscopy, in which the same cell is photo-
slowly in a medium of higher refractive index. Thus a beam of graphed at regu lar intervals over time to generate a movie.

FIGURE 9-9 Live cells can be visualized by microscopy image, cells are surrounded by alternating dark and light bands;
techniques that generate contrast by interference. These micro- in-focus and out-of-focus details are simultaneously imaged in a
graphs show live, cu ltured macrophage cells viewed by bright-field phase-contrast microscope. In a DIC image, cells appear in pseudo-
microscopy (left), phase-contrast microscopy (middle), and differential- relief. Because only a narrow in-focus region is imaged, a DIC image is
interference-contrast (DIC) microscopy (right). In a phase-contrast an optical slice through the object. [Courtesy of N. Watson and J. Evans.)

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 407
This procedure allows the observer to study cell movement, Hematoxylin binds to basic amino acids (lysine and argi-
provided the microscope's stage can contro l the temperature nine) on many different kinds of proteins, w hereas eosin
of the specimen and the appropriate environment. binds to acidic molecules (such as DNA and side c hains of
aspartate and glutamate). Because of their different binding
properties, these dyes stain var ious cell types sufficiently dif-
Imaging Subcellular Details Often Requires That ferently that they are distinguishable visually (Figure 9-lOb).
the Samples Be Fixed, Sectioned, and Stained If an enzyme catalyzes a reaction that produces a colored or
otherwise visible precipitate from a colorless precursor, the
Live cell~ and tis~ues gener:::JIIy lack compounds that absorb
enzyme can be detected in cell sections by their colored reac-
light and so a re nearly invisible in a light microscope. Although
tion products. Such staining techniques, although once quite
such specimens can be visualized by the special techniques we
common, have been largely replaced by other techniques for
just discussed, these methods do not reveal the fine details of
visua lizing particular p roteins, as we discms next.
structure.
Specimens for light and electron microscopy are commonly
Fluorescence Microscopy Can Localize and
fixed wtth a solution containing chemicals that cross-link most
proteins and nucleic acids. Formaldehyde, a common fixative, Quantify Specific Molecules in Live Cells
cross-links amino groups on adjacent molecules; these covalent Perhaps the most versatile and pO\yerful technique for local-
bonds stabilize protein-protein and protein-nucleic acid inter- izing molecules within a cell by light microscopy is fluorescent
actions and render the molecules insoluble and stable for sub- staining of cells and observation by fluorescence microscopy.
sequent procedures. After fixation, a tissue sample for A chemical is said to be fluorescent if it absorbs light at one
examination by light microscopy is usually embedded in paraf- wavelength (the excitation wavelength) and emits light (fluo-
fin and cut into sections about 50 ~J-m thick (f igure 9-lOa). resces) at a specific and longer wavelength. Modern mi cro-
Cultured cells growing on glass coverslips, as described above, scopes for observing fluorescent samples are configured to
arc thin enough so they can be fixed in situ and visualized by pass the excitation light through the objective into the sam-
light microscopy without the need for sectioning. ple and then selectively observe the emitted fluorescent light
A final step in prepari ng a specimen for light microscopy coming back through the objective from the sample. This is
IS to stam 1t so as to visualize the main structural features of achieved by reflecting the excitation light on a special type of
the cell or tissue. Many chemica l stains bind to molecules fi lter called a dichroic mirror into the sample and allowing
that have specific features. For example, histological samples the light emitted at the longer wavelength to pass through to
are often stained with hematoxylin and eosm ("H&E stain'' ). the observer (see Figure 9-8d ).

(a) (b)

Specimen Specimen
holder block Block Specimen

section

Microtome

I
I
Microscope slide
Copper
mesh
grid

FIGURE 9-10 Tissues for light microscopy are commonly fixed, specimen has hardened, it is mounted on the arm of a microtome and
embedded in a solid medium, and cut into thin sections. (a) A fixed slices are cut with a knife. Typical sections cut for light microscopy are
tissue is dehydrated by soaking in a series of alcohol-water solutions, 0.5 to 50 IJ..m thick. The sections are collected on microscope slides
ending with an organic solvent compatible with the embedding and stained with an appropriate agent. (b) A section of mouse intestine
medium. To embed the tissue for sectioning, the tissue is placed in stained with H&E. [Part (b) it> Dr. Gladden WillisNisuals Unlimited/Corbis.]
liquid paraffin for light microscopy. After the block containing the

408 CHAPTER 9 • Culturing, Visualizmg, and Perturbing Cells

 fraction of fura -2 that has a bound Ca2 ion and thus in the
concentration of cytosolic Ca2 + (Figure 9- 11 ).
Fluorescent dyes (e.g., SNARF-1) that are sensitive to the
H + concentration can similarly be used to monitor the cytosolic
pH of living cells. Other useful probes consist of a fluorochrome
linked to a weak base that is only partially protonated at neu-
tral pH and that can freely permeate cell membranes. In acidic
organelles, however, these probes become protonated; because
the protonatcd probes cannoL n:l:ru~s the organelle membrane,
they accumulate in the lumen in concentrations manyfold
greater than in the cytosol. Thus this type of fluorescent dye can
be used to specifically stain mitochondria and lysosomes in liv-
ing cells (Figure 9-12).

EXPERIMENTAL FIGURE 9-11 Fura-2, a Ca2+ -sensitive Immunofluorescence Microscopy Can Detect
fluorochrome, can be used to monitor the relative concentrations
of cytosolic Ca2+ in different regions of live cells. (Left) In a moving
Specific Proteins in Fixed Cells
leukocyte, a Ca2+ gradient is established. The highest levels (green) are The common chemical dyes mentioned above stain nucleic
at the rear of the cell, where cortical contractions take place, and the acids or broad classes of proteins, but it is much more infor-
lowest levels (blue) are at the cell front, where actin undergoes mative to detect the presence and location of specific proteins.
polymerization. (Right) When a pipette filled with chemotactic Immunofluorescence microscopy is the most widely used
molecules placed to the side of the cell induces the cell to turn, the method to detect specific proteins with an anttbody to which
Ca 2 concentration momentarily increases throughout the cytoplasm a fluorescent dye has been covalently attached. To do this,
and a new gradient is established. The gradient is oriented such that
you first need to generate antibodies to your specific protein.
the region of lowest Ca 2 ~ (blue) lies in the direction that the cell will
As discussed briefly in Section 9.1 and in detail in Chapter 23,
turn, whereas a region of high Ca 2 .. (yellow) always forms at the site
as part of the response to infection the vertebrate immune
that will become the rear of the cell. [From R. A. Brundage et al., 1991,
Science 254:703; courtesy of F. Fay.]

Determination of Intracellular Ca2+ and H+

Levels with Jon-Sensitive Fluorescent Dyes
The concentration of Ca 2 + or H ~ within live cells can be
measured with the aid of fluorescent dyes, or fluorochromes,
whose fluorescence depends on the concentration of these
ions. As discussed in later chapters, intracellu lar Ca2+ and
H concentrations have pronounced effects on many cellular
processes. For instance, many hormones and other stimuli
cause a rise in cytosolic Ca 2 + from the resting level of about
10 7 M to 10-6 M, which induces vario us cellular responses
such as the contraction of muscle.
The fluorescent dye fura-2, which is sensitive to Ca 2 ~, con-
tains five carboxylate groups that form ester linkages with etha-
nol. The resulting fura-2 esfer is lipophilic and can diffuse from
the medium across the plasma membrane into cells. Within the
cytosol, esterases hydrolyze fu ra-2 ester, yielding fura-2, whose
free carboxylate groups render the molecule nonlipophilic and
thus unable to cross cellular membranes, so it remains in the
cytosol. Inside cells, each fura-2 molecule can bind a single
Cal+ ion but no other cellular cation. This binding, which is
proportional to the cytosolic Ca2 .. concentration over a certain
r ange, increases the fluorescence of fura-2 at one particular EXPER Ill. TAL FIGURE 9-12 Location oflysosomes and
wavelength. At a second wavelength, the fluorescence of fura-2 mitochondria in a cultured living bovine pulmonary artery endothe-
is the same whether or not Ca 2 + is bound and provides a mea- lial cell. The cell was stained with a green-fluorescing dye that is specifi-
sure of the total amount of fura-2 in a region of the cell. By cally bound to mitochondria and a red-fluorescing dye that is specifically
examining cells continuously in the fluorescence microscope incorporated into lysosomes. The image was sharpened using a deconvo-
and measuring rapid changes in the ratio of fura-2 fluorescence lution computer program discussed later in the chapter. N, nucleus.
at these two wavelengths, one can quantify rapid changes in the [Courtesy Invitrogen/Molecular Probes Inc.]

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 409
system generates proteins called antibodies that bind specifi- which emit red light; Cy3, which emits orange light; and fluo-
cally to the infectious agent. Cell biologists have made use of rescein, which emirs green light. When a fluorochrome-
this immunological response to generate antibodies to specific antibody complex is added to a permeabilized cell or tissue
proteins. Consider you have purified protein X and then inject section, the complex will bind to the corresponding antigen,
it into an experimental animal so that it responds to the pro- then light up when illuminated by the exciting wavelength.
tein as a foreign molecule. Over a period of weeks, the animal Staining a specimen with different dyes that fluoresce at differ-
will mount an immune response and make antibodies to pro- ent wavelengths allows multiple proteins as well as DNA to be
tein X (the "antigen"). If you collect the blood from the ani- localized within the same cell (see chapter opening figure).
mal, it will have antibodies to protein X m1xed 111 w1th The most commonly used variation of this technique is
antibodies to many other different antigens, together with all called indzrect immunofluorescence microscopy since the spe-
the other blood proteins. You can now covalently bind pro- cific antibody is detected indirectly. In this technique, an unla-
tein X to a resin and, using affinity chromatography, bind and beled monoclonal or polyclonal antibody is applied to the cells
selectively retain just those antibodies specific to protein X. or fixed tissue section, followed by a second fluorochrome-
The antibodies can be eluted from the resin, and now you tagged antibody that binds to the constant (Fe) segment of the
have a reagent that binds specifically to protein X. This ap- first antibody. For example, a "second" antibody can be gener-
proach generates polyclonal antibodies since many different ated by immunizing a goat with the Fe segment that is common
cells in the animal have contributed the antibodies. Alterna- to all rabbit IgG antibodies; when ~oupled to a fluorochrome,
tively, as we described earlier in this chapter, it is possible to this second antibody preparation (called "goat anti-rabbit")
generate a clonal cell line that secretes antibodies to a specific will detect any rabbit antibody used to stain a tissue or cell
epitope on protein X; these arc called monoclonal antibodies. (Figure 9-13). Because several goat anti-rabbit antibody mole-
To use either type of antibody to localize the protein, the cules can bind to a single rabbit antibody molecule in a section,
cells or tissue must first be fixed to ensure that all components the fluorescence is generally much brighter than if a directly
remain in place and the cell permeabilized to allow entry of coupled single fluorochrome antibody is used. This approach is
the antibody, commonly done by incubating the cells with a often extended to do double-label fluorescence microscopy, in
non-ionic detergent or extracting the lipids with an organic which two proteins can be visualized simultaneously. For ex-
solvent. In one version of immunofluorescence microscopy, ample, both proteins can be visualized by indirect immunofluo-
the antibody is covalently linked to a fluorochrome. Com- rescence microscopy using first antibodies made in different
monly used fluorochromcs include rhodamine and Texas red, animals (e.g., rabbit and chicken) and the second antibodies

0 Prepare sample and place

on microscope slide

c~ ~
!fJ Incubate with primary antibody;
wash away unbound antibody

/_
Incubate with fluorochrome-
conjugated secondary antibody;
wash away unbound antibody

I II
Mount specimen and observe
in fluorescence microscope , 20 IJ..m

FIGURE 9-13 A specific protein can be localized in fixed tissue medium and examined in a fluorescence microscope (step 19). In this
sections by indirect immunofluorescence microscopy. To localize a example, a section of the rat intestinal wall was stained with Evans
protein by immunofluorescence microscopy, a tissue section, or sample blue, which generates a nonspecific red fluorescence, and GLUT2, a
of cells, has to be chemically fixed and made permeable to antibodies glucose transport protein, was localized by indirect immunofluores-
(step 0). The sample is then incubated with a primary antibody that cence microscopy. GLUT2 is seen to be present in the basal and lateral
binds specifically to the antigen of interest and then unbound antibody sides of the intestinal cells but is absent from the brush border,
removed by washing (step f)). The sample is next incubated with a composed of closely packed microvilli on the apical surface facing the
fluorochrome-labeled secondary antibody that specifically binds to the intestinal lumen. Capillaries run through the lamina propria, a loose
primary antibody, and again excess secondary antibody is removed by connective tissue beneath the epithelial layer. [B. Thorens et al., 1990,
washing (step 0). The sample is then mounted in specialized mounting Am. J. Physio/. 259:(279; courtesy of B. Thorens.]

410 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 Rhodamine-labeled phalloidin
(fluorochrome-conjugated
drug that binds actin filaments, red)

Primary antibodies (rabbit, black)

that recognize microtubules and
fluorochrome-conjugated secondary
antibodies (goat-anti-rabbit, green)
EXPERIMENTAL FIGURE 9-14 Double-label fluorescence
microscopy can visualize the relative distributions of two proteins.
In double-label fluorescence microscopy, each protein has to be
labeled specifically with different fluorochromes. The diagram at left
shows how this can be done: a cultured cell was fixed and permeabi-
lized and then incubated with Rhodamine-labeled phalloidin, a reagent
that specifically binds to filamentous actin. It was also incubated with
rabbit antibodies to tubulin, the major component of microtubules,
followed by a fluorescein-labeled second goat-anti-rabbit antibody.
The upper panels on the right show the fluorescein-stained tubulin
(left) and Rhodamine-stained actin (right) and the lower panel the
electronically merged images. [Part (right) Courtesy of A. Bretscher.]

(e.g., goat-anti-rabbit and sheep-anti-chicken) labeled with dif- chromophore when illuminated with blue light. Using re-
ferent fluorochromes. In another variation, one protein can be combinant DNA technologies, it is possible to make a DNA
visualized by indirect immunofluorescence microscopy and the construct in which the coding sequence of GFP is fused to
second protein by a dye that specifically binds to it. Once the the coding sequence of a protein of interest. When intro-
individual images are taken on the fluorescence microscope, duced and expressed in cells, a GFP "tagged" protein is
their images can be merged electronically (Figure 9-14 ). made in which the protein of interest is covalently ltnked to
ln another widely used version of this technology, molecu- GFP as part of the same polypeptide. Although GFP is a
lar biology techl)iques are used to make a eDNA encoding a moderate-size protein, the function of the protein of interest
recombinant protein to which is fused a short sequence of is often not changed by fusing it to GFP. This now allows
amino acids called an epitope tag. When expressed in cells, one to visualize GFP-and hence the protein of interest. Not
this eDNA will generate the protein linked to the specific rag. only can one immediately sec the localization of the GFP-
Two commonly used epitope tags are called FLAG, encoding tagged protein, but one can view its distribution in a living
the amino acid sequence DYKDDDDK (single-letter code), cell over time and thereby assess its dynamics or track its
and myc, encoding the sequence EQKUSEEDL. Commercial localization following various cell treatments. The simple
fluorochrome-coupled mo'noclonal antibodies to the FLAG or idea of tagging specific proteins with GFP has revolutionized
myc epitopes can then he used to detect the recombinant pro- cell biology and led to the development of many different
tein in the cell. In an extension of rhis technology to allow the fluorescent proteins (Figure 9-15). One use of this colorful
simultaneous visualization of two proteins, one protein can be ,·ariety of fluorescent proteins allows one to visualize two or
tagged with FLAG and a different protein with myc. Each more proteins simultaneously if they are each tagged with a
tagged protein is then visualized with a different color, for different-colored fluorescent protein. We describe additional
example, with a rhodamine-labeled antibody to the myc epi- techniques that exploit fluorescent proteins in later sections.
tope and a fluorescein-labeled antibod} to the FLAG epitope.
Deconvolution and Confocal Microscopy
Tagging with Fluorescent Proteins Allows the Enhance Visualization of Three-Dimensional
Visualization of Specific Proteins in Living Cells Fluorescent Objects
The jellyfish Aequorea victoria expresses a naturally fluores- Conventional fluorescence microscopy has two major limita-
cent protein, called green fluorescent protein (GFP, -27 kD). tions. First, the fluorescent light emitted by a sample comes
GFP contains a serine, tyrosine, and glycine sequence whose not only from the plane of focus but also from molecules
side chains spontaneously cyclize to form a green-fluorescing above and below it; thus the observer sees a blurred image

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 411
(a) (b)

m m m (")
3 3 3 3 3 3 3 3 3 3
.,.,Ill .,., Cl.,., s:
(")
I
0
Ill
Q)
0
a.
ol a;! ~ (")
~
Cl :JJ
Q)
Cl
.., ::!!
c:
::> ::> iil ::> iil iil Q)
(1) ::>
"'a.
Q) ::> 3Q) co :E
~ "0 "'
"0 "0 3
'< ::>
Q)
co
(1)
0 ~- C1" < (1)
..... C1"
..,
(1)
(1)
N
(1) ::> ~ .....
:E <
(1)
'<

FIGURE 9-15 Many different colors of fluorescent proteins are to show growing bacteria expressing several different-colored
now available. (a) Tubes show the emission colors and names of fluorescent proteins. [R. Tsien.)
many different fluorescent proteins, and (b) an agar dish is illuminated

caused by the superposition of fluorescent images from mol- methods require the image to be collected electronically so
ecules at many depths in the cell. The blurring effect makes it that it can then be computationally manipulated as necessary.
difficult to determme the actual molecular arrangements. Sec- The first approach is called deconvolution microscopy,
ond, to visualize thick specimens, consecutive (serial) images w hich uses computational methods to remove fluorescence
at various depths through the sample must be collected and contributed from out-of-focus parts of the sample. Consider a
then aligned to reconstruct structures in the original thick tis- three-dimensional sample in which images from three differ-
sue. Two general approaches have been developed to obtain ent focal planes a re recorded. Since the w hole sample is illumi-
high-resolution three-dimensional informatio n. Both these nated, the image from plane 2 will contain out-of-focus

PE ENTA fiGURE: - 6 Deconvolution fluorescence this three-dimensional reconstruction of the raw images, the DNA,
microscopy yields high-resolution optical sections that can be microtubules, and actin appear as diffuse zones in the cell. (b) After
reconstructed into one three-dimensional image. A macrophage application of the deconvolution algorithm to the images, the fibrillar
cell was stained with fluorochrome-labeled reagents specific for DNA organization of microtubules and the localization of actin to adhesions
(blue), microtubules (green), and actin microfilaments (red). The series become readily visible in the reconstruction. [Courtesy of J. Evans,
of fluorescent images obtained at consecutive focal planes (optical Whitehead Institute.)
sections) through the cell were recombined in three dimensions. (a) In

412 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells ..

 fluorescen<.:c from planes I and 3. If you knew exactly how sample, an accurate three-dimensional representation can he
·. out-of-focus fluorescence from planes 1 and 3 contributed to computationally generated. Two types of confocal micro-
light collected in plane 2, you could computationally remove scopes are in common use today, a point-scanning confocal
it. To gain this information for a particular microscope, a se- microscope (also known as a laser-scanning confocal micro-
ries of images of focal planes are made from a test slide con- scope, or LSCM) and a spinning disk confocal microscope.
taining tiny fluorescent heads. Each bead represents a pinpoint The idea behind each microscope is to both illuminate and
of light that becomes a blurred object outside its focal plane; collect emitted fluorescent light in just one small area of a
from these images a point stJread function is determined that focal plane at a time in such a way that out-of-focus light is
enables the investigator to calculate the di~rribution of fluo- excluded. This can be achieved by cullt:cting the emitted
rescent point sources that contributed to the "blur" w hen out light through a pinhole before reaching t he detector-light
of focus. Having calibrated the microscope in this manner, the from the focal plane passes through, whereas light from
experimental series of images can be computationally decon- other focal planes is largely excluded. The illuminated area
volved. Images restored by deconvolution display impressive is then moved across the whole focal plane to build up the
detail without any blurring, as illustrated in Figure 9-16. image electronically. The two types of microscopes differ in
The second approach to obtaining better three-dimen- how they cover the image. The point-scanning microscope
sional information is called confocal microscopy because it uses a point laser light source at the excitation wavelength to
uses optical methods to obtain images from a specific focal rapidly scan the foca l plane in a raster pattern, with the emit-
plane and exclude light from other planes. By collecting a ted fluorescence collected by a photomultiplier tube, and
series of images focused through the vertical depth of the thereby build up an image (Figure 9-17a). It can then take a

(a) Laser-Scanning Confocal Microscope (b) Spinning Disk Confocal Microscope

Lasers

Spinning DiskHead Unit

Pinhole Microlens Spread
array /array laser
Image - -= .;:;::::. Image ~ •- =:-
plane plane beam

Projection - _ - ._
lens

mirror mirror

Objective
lens ~­
Scanning Head Unit

Specimen ,_ __ ___

Computer
() i1 Computer

FIGURE 9-17 Light paths for two types of confocal microscopy. tube comes almost exclusively from the illuminated spot in the focal
Both types of microscopy are assembled around a conventional plane. A computer then takes these signals and reconstructs the image.
fluorescence microscope (yellow shading). The diagram in (a) depicts the The diagram in (b) depicts the light path in a spinning disk confocal
light path in a laser scanning confocal microscope. A single-wavelength microscope. Here instead of using two scanning mirrors, the beam
point of light from an appropriate laser is reflected off a dichroic mirror from the laser is spread to focus on pinholes in a spinning disk. The
and bounces off two scanning mirrors and from there through the excitation light passes through the objective to provide point
objective to illuminate a spot in thP ~pecime n _ The scanning mirrors illumination of a number of ~pots in the sample. The fluorescence
rock back and forth in such a way that the light scans the sample in a emitted passes back through the objective, through the holes in the
raster fashion (see green lines in the sample). The fluorescence emitted spinning disk, and is then bounced off a dichroic mirror into a sensitive
by the sample passes back through the objective and is bounced off digital camera. The pinholes in the disk are arranged so that as it spins,
the scanning mirrors onto the dichroic mirror. This allows the light to it rapidly illuminates all parts of the sample several times. As the disk
pass through toward the pinhole. This pinhole excludes light from spins fast, for example at 3,000 rpm, very dynamic events in living cells
out-of-focus foca l planes, so the light reaching the photomultiplier can be recorded.

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 413
 (a) Conventional fluo rescence m icroscopy (b) Confocal f lu orescence microscopy

1
40 ~-tm

Focal plane Focal plane---+~-

~volume
Imaged

FIGURE 9-18 Confocal microscopy produces an in-focus optical occurs because background fluorescence is detected from tubulin
section through thick cells. A mitotic fertilized egg from a sea urchin above and below t he focal plane as depicted in the sketch. (b) The
(Psammechinus) was lysed with a detergent, exposed to an anti-tubulin confocal microscopic image is sharp, particularly in the center of the
antibody, and then exposed to a fluorescein-tagged antibody t hat mitotic spindle. In this case, fluorescence is detected only from
binds to the anti-tubulin antibody. (a) When viewed by conventional molecules in the focal plane, generating a very thin optical section.
fluorescence microscopy, the mitot ic spindle is blurred. This blurring [Micrographs from J. G. White et al., 1987, J. Cell Bioi. 104:41.]

series of images at different depths in the sample to generate focuses the laser light on 20,000 pinholes of the second disk.
a three-dimensional reconstructio n. A point-scanning confo- T he pinholes are arr anged in such a way that they completely
cal microscope can provide exceptionally high-resolu tion scan the foca l plane of the sample several times with each
images in both two and three dimensions (Figure 9- 18), al- turn of the disk. T he emitted fluorescent light returns through
though it has two minor li m itations. First, it can take sig- the pinholes of the second disk and is reflected by a dichroic
nificant time to scan each focal plane, so if a very dyna mic mirror and focused onto a highly sensitive digital camera. In
process is being_ imaged, the microscope may not be able to this way, the sample is scanned in less than a millisecond, and
collect images fast enough to follow the dynamics. Second, it so the real-time location of a fluorescent reporter can be cap-
illuminates each spot with intense laser light that can bleach tured even if it is highly dynamic (Figure 9-19). A current
the fluorochrome being imaged and therefore limit the num- lim itation of a spinning disk microscope is that the pinhole
ber of images that can be collected. size is fixed and has to be matched to the magn ification of the
The spinning disk microscope circumvents these two objective, so it is generally configured for use with a 63X or
problems (see Figure 9-l?b). The excitation light from a laser lOOX objective and is less useful for the lower-magnification
is spread out and illuminates a small part of the disk spinning imagi ng that might be required in tissue sections. Thus the
at high speed, for example at 3000 rpm. T he disk in fact con- point-scanning and spinning disk confocal microscopes have
sists of two linked disks: one with 20,000 lenses that precisely overlapping and complementary strengths.

(i) VIDEO: Microtubule Dynamics in Fission Yeast

00 s 30 s 58 s 85 s 113 s 140 s

EXPERIMENTAL FIGURE 9 19 The dynamics of microtubules Six frames from a movie of GFP-tubulin in two of the rod-shaped cells
can be imaged on the spinning disk confocal microscope. of fission yeast are shown. [Courtesy of Fred Chang.]

414 CHAPTER 9 • Culturing, Visua l izing, and Perturbing Cells

(a) Emitted light (b)

A
Excitation beam - -- 1 -- 1

--Objective
Internal reflection at
glass-water interface
generates evanescent wave
~Immersion oil
Coverslip

Specimen in water ---=~~~-~Jiillii~~~- Evanescent wave, depth of

illumination 50-100 nm

Microscope slide >

EX PEn MENTAL FIGURE 9 20 Fluorescent samples in a

restricted focal plane can be Imaged by total internal reflection
(TIRF) microscopy. (a) In TIRF microscopy only about 50 to 100 nm
adjacent to the coverslip is illuminated so that fluorescent molecules in
the rest of the sample are not excited. This limited illumination is
ach ieved by directing the illuminating light at an angle where it is
reflected from the glass-water interface of the coverslip rather than
passing through it. Whereas most of the light is reflected, it also
generates a very small region of illumination called the evanescent
wave (depicted in light blue). (b) Immunofluorescence microscopy with
tubulin antibody has been used to visualize microtubules viewed by
conventional fluorescence microscopy (top panel), TIRF (middle panel),
and a merged image. The two images were collected and false colored
red and green so that the merge could highlight those microtubules
that are close to the coverslip (green). [Part (b) from J. B. Man neville et al.,
2010,J.Cel/8io/. 191:585.]

TIRF Microscopy Provides Exceptional

Imaging in One Focal Plane
The confocal microscopes we have just described provide
amazing and in'for mative images. However, they are not TIRF has been exceptionall y useful in identifying structures
perfect, and some experimental situations call for fluores- on the bottom of cells and therefore close to the coverslip
cence imaging in a thin foca l plane adjacent to a surface. For (Figure 9-20b) and for measuring the kinetics of assembly
example, confocal imaging is not ideal for exploring the de- and disassembly of structures such as microtubules and actin
tails of proteins at adhesion sites between a cell and a cover- filaments (see Chapters 17 and 18).
slip or to follow t he kinetics of assembly of microtubules
attached to a coverslip. Both these situations can be imaged
at high sensitivity using total internal reflection fluorescence
FRAP Reveals the Dynamics
(TIRF) microscopy. In the most common configuration of
T IRF microscopy, th e excitation beam of light comes of Cellular Components
through the objective lens (Figure 9-20a). However, the Live cell fluorescence imaging reveals the location and bulk
angle at which the light arrives at the coverslip is adj usted to dynamics of populations of fluorescent molecu les, but it
the critical angle so that the light is reflected off the coverslip doesn't tell you how dynamic individual molecules are. For
and returns up through the objective. This generates a nar- example, if we see that a GFP-labeled protein forms a patch at
row band, called an evanescent wave, that illuminates only the surface of a cell, does this represent a stable collection of
about 50 to 100 nm of the sample adjacent to the coverslip, fluorescent protein molecules or a dynamic equilibrium with
with no illumination to the rest of the sample. Thus if you fluorescent proteins coming in and out of the patch? We can
have a complex mixture of fluorescent structures in a speci- investigate th is question by observing the dynamics of the
men, in the TIRF microscope you w ill see only those that are molecules in the patch (Figure 9-21 ). If we use a high-intensity
within 50 to100 nm (2 to 4 times the thickness of a microtu- light to permanently bleach the fluorochrome (e.g., GFP) just
bule) of the coverslip. When cells are grown on a coverslip, in the patch, there will initially be no fluorescence coming

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 415
 (a)

Bl each a small region Follow recovery of fluorescence over time in ROI

Pre-bleach of interest (ROI) -------------------~ nme

(b) Pre-bleach Bleach Post-bleach Po~t-h l each

t =-7.04 s t- 0 s t = 5.56 s t=30.16s

(c)

Ql
g 0.8
Ql
(,)

"'
Ql
.... >
g.~ 0.6
;:;::c

1l~
.~ ·-
(ij 0.4
E
0
z
0.2

-5 0 5 10 15 20 25 30

t
Bleach
Normalized time (s)

XPER II' EN 'A FIGURE 9·21 Fluorescence recovery after represents unbleached molecules that have moved into the ROI.
photobleaching (FRAP) reveals the dynamics of molecules. In a (b) The mobility of GFP-serotonin receptors on the cell surface observed
living cell, following the distribution of a GFP-Iabeled protein provides using FRAP. The fluorescence in two regions is followed-one that is
a view of the overall distribution of the protein, but it doesn't tell you bleached (region 1) and a control region that is not (reg ion 2). (c) By
how dynamic populations of individual molecules might be. This can quantitating the recovery, the dynamic properties ofthe serotonin
be determined by FRAP. (a) In this technique, the GFP signal is receptor can be established. [Part (b) from S. Kalipatnapu, 2007, Membrane
bleached by a short burst of strong laser light focused on the region of Organization and Dynamics of the Serotonin 1A Receptor Monitored using
interest (ROI). This rapidly bleaches the molecules irreversibly, so they Fluorescence Microscopic Approaches, in Serotonin Receptors in Neurobiology,
are not detected again. Restoration of fluorescence into the region A. Chattopadhyay, ed. CRC press.]

from it and it will look dark in the fluorescence microscope. (see Figure 10-10), the dynamics of specific components of
However, if the components in the patch are in dynamic equi- the secretory pathway.
librium with unbleached molecules elsewhere in the cell, the
bleached molecules will be replaced by unbleached ones, and
the fluorescence will begin to come back. The rate of fluores- FRET Measures Distance Between Chromophores
cence recovery is a measure of the dynamics of the molecules. Fluorescence microscopy can also be used to determine if two
This technique, known as fluorescence recovery after photo- proteins interact in vivo using a technique called Forster reso-
bleaching (FRAP), has revealed how very dynamic many nance energy transfer (FRET). This technique utilizes two fluo-
components in cells are. For example, it has been used to rescent proteins in which the emission wavelength of the first is
determine the diffusion coefficient of membrane proteins the same as the excitation wavelength of the second (Figure

416 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 (a) FIGURE 9-22 Protein-protein interactions can be visualized
by FRET. The idea behind FRET is to use two different fluorescent
480 n m
proteins so that when one is excited, its emission will excite the second
/ fluorescent protein provided they are sufficiently close (upper panel).
YFP In this example, cyan fluorescent protein (CFP) is fused to protein X,
yellow fluorescent protein (YFP) is fused to protein Y, and both proteins
@> are expressed in a living cell. lfthe cell is now illuminated with 440-nm
light, the CFP will emit a fluorescent signal at 480 nm. If YFP is not close
by, it will not absorb the 480-nm light and no 535-nm light will be emit-
(b)
ted. However, if protein X interacts with protein Y (as shown), it will
bring CFP close t o YFP, t h e emitted 480-nm light will be captured by
YFP, and it will emit light at 535 nm. (b) In this mouse fibroblast, FRET
has been used to reveal that the interaction between an active
regulatory protein (Rae) and its binding partner is localized to the
front of a migrating cell. [Part (b) from R. B. Sekar and Periasamy, 2003, J. Cell
Bioi. 160:629.]

sensitive to small changes in distance and in practice is not

detectable at >10 nm. Thus by illuminating an appropriate
sample with 440-nm light and observing at 535 nm, one can
tell if proteins separately tagged with CFP and YFP are in very
close proximity. For example, FRET sensors have been devel-
oped to determine where in a cell signaling between a small
GTP-binding protein and its effector occurs (Figure 9-22b).
9-22). For example, when cyan fluorescent protein (CFP) is A modified version of this technique can be used to measure
excited with 440-nm wavelength light, it fluoresces and emits conformational changes in a protein. For example, a sensor
light at 480 nm. If yellow fluorescent protein (YFP) is close by, called cameleon has been designed to measure intracellular lev-
it wi ll absorb the 480-nm light and emit light at 535 nm. The els of Ca2 • Cameleon consists of a single polypeptide contain-
efficiency of energy transfer is proportional to R6, where R is ing both CFP and YFP joined by a piece of polypeptide capable
the distance between the fluorop hores-it is therefore very of binding Ca2+ (Figure 9-23a). ln the absence of Ca2 , the CFP

(a) EXPERIMENTAL FIGURE 9-23 FRET biosensors can detect

480 nm
local biochemical environments. (a) A FRET biosensor is a fusion
protein containing two fluorescent proteins linked by a region
.· sensitive to the environment under study. In this example, a protein
YFP construct called came leon consists of CFP linked to YFP by a sequence
based on the protein calmodulin that undergoes a large conforma-
tional change when it binds Ca 2• . In the absence of Ca 2- , the two
fluorescent proteins are too fa r apart to undergo FRET, whereas when
the local Ca 2 • rises, they are brought sufficiently close to undergo FRET.
440 nm
(b) An example of the use of came leon reveals the oscillation in local
Ca 2 levels in an Arabidopsis growing root tip.
[Part (b) from G. B. Monshausen et al., 2008, PlantPhysiol. 147:1690-1698.]
YFP
-~535nm

(b)

9.2 Light Microscopy: Exploring Cell Structure and Visualizing Proteins Within Cells 417
and YFP are not close enough for FRET to occur. However, in overlap so much that they look like one structure. New
the presence of an appropriate local concentration of Ca:!.+ , methods have been developed to get around this problem,
cameleon binds Ca 2 and undergoes a conformational change one of which is called photo-activated localization micros-
that brings CFP and YFP in close proximity, and they can now copy (PALM). It relies on the abi lity of a variant of GFP to
undergo FRET. Sensors such as cameleon can be used to mea- be photoactivated; that is, it only becomes fluorescent when
sure the level of Ca 2 in living cells, for example in a growing activated by a specific wavelength of light, different from its
root tip (Figure 9-23b). Creative researchers are developing excitation wavelength. Consider what would happen if you
FRET sensors to illuminate many different types of local envi- could activate just one GFP molecule. When you excite the
ronment~; for example, it is possible to make a probe that un- sample, the one activated GFP would emit many hundreds of
dergoes FRET only when it becomes phosphorylated by a photons, giving rise to a Gaussian distribution (Figure
specific kinase and thereby reveal where in the cell the active 9-24a). Although analysis of each photon does not tell you
kinase is localized. precisely where the GFP is, the center of the peak can tell
you where the GFP is located with nanometer accuracy. If
you now activate another GFP, you can localize it individu-
ally with the same precision. In PALM, a small percentage of
Super-Resolution Microscopy Can Localize
GFPs are activated and each localized witb high precision, and
Proteins to Nanometer Accuracy then another set is activated and localized, and as additional
As we discussed earlier, the theoretical resolution limit of the cycles of activation and localization are recorded, a high-reso-
fluorescence microscope is about 0.2 !J.m (200 nm). To un- lution image emerges. For example, the three-dimensional dis-
derstand why this is, consider two fluorescent structures tribution of microtubules can be seen with much greater
separated by 100 nm. When you try to image them, they clarity than with any other light-microscopic method (Figure
each generate a Gaussian distribution of fluorescence, which 9-24b) and a clathrin-coated pit-about 100 nm in diameter-

(a) EXPERIMENTAL FIGURE 9-24 Super-resolution microscopy can generate

light-microscope images with nanometer resolution. The theoretical resolution of
the light microscope can be circumvented by super-resolution microscopy, where
single molecules are imaged individually/ separately to generate a composite image.
"'c0 400 ~
One version of this technology images proteins fused to photo-activatable GFP in
0
.s= fixed samples. (a) When a GFP is activated and then excited, it will emit thousands of
a.
0 photons that can be collected. This generates a Gaussian curve centered around the
Qi location of the emitting GFP; the center provides the location of the GFP to nanome-
.0 200 f- ter accuracy. This process is reiterated hundreds of times to excite other GFP
E
::J
z molecules, and a high-resolution image emerges. (b) A confocal image of microtu-
bules (left) is compared with a corresponding super-resolution image (right) in which

0 I l ~ I
the three-dimensional arrangement of the microtubules is color coded. (c) The circular
nature of a clathrin-coated pit (discussed in Chapter 14) is shown-a confocal image
-100 - 50 0 50 100 of this structure would look like two bright spots without any detail visible. [Parts (b)
x(nm) and (c) from B. Huang et al.. 2008, Science 319:81 0- 813.)

(b) (c)

418 CHAPTER 9 • Culturing, Visualizing , and Perturbing Cells

can be seen in remarkable detail (Figure 9-24c). These types of 9.3 Electron Microscopy:
images can take up to an hour to generate, so they are re-
stricted to fixed images and currently cannot be used for im- High-Resolution Imaging
aging proteins in live cells. Electron microscopy of biological samples, such as single pro-
reins, organelles, cells, and tissues, offers a much higher resolu-
tion of ultrastructure than can be obtained by light microscopy.
KEY CONCEPTS of Section 9.2 The short wavelength of electrons means that the limit of
resolution for a transmission electron microscope is theo-
Light Microscopy: Exploring Cell Structure retical!) 0.005 nm (less than the diameter of a smgle atom),
and Visualizing Proteins Within Cells or 40,000 times better than the resolution of a light micro-
• The resolution of the light microscope, about 0.2 11-m, is scope and 2 million times better than that of the unaided
limited by the wavelength of light. human eye. However, the effective resolution of the transmis-
sion electron microscope in the study of biological systems is
Because most cellular components are not colored, differ-
considerably less than this ideal. Under optimal conditions, a
ences in refractive index can be used to observe parts of
resolution of 0.10 nm can be obtained with transmission elec-
single cells employing phase-contrast and interference-contrast
tron microscopes, about 2000 times better than with conven-
microscopy.
tional high-resolution light microscopes.
• Tissues generally have to be fixed, sectioned, and stained The fundamental principles of electron microscopy are
for cells and subcellular structures to be observed. similar to those of light microscopy; the major difference is
that electromagnetic lenses focus a high-velocity electron
• Fluorescence microscopy makes use of compounds that ab- beam instead of the visible light used by optical lenses. In the
sorb light at one wavelength and emit it at a longer wavelength. transmission electron microscope (TEM), electrons are emit-
• Ion-sensitive fluorescent dyes can measure intracellular ted from a filament and accelerated in an electric field . A con-
concentrations of ions, such as Ca2+. denser lens focuses the electron beam onto the sample;
objective and projector lenses focus the el\!ctrons that pass
• Immunofluorescence microscopy makes use of antibodies to through the specimen and project them onto a viewing screen
localize specific components in fixed and permeablized cells. or other detector (Figure 9-25, left). Because atoms in air ab-
sorb electrons, the entire tube between the electron source and
• Indirect immunofluorescence microscopy uses an unla-
the detector is maintained under an ultrahigh vacuum. Thus
beled primary antibody, followed by a fluorescently labeled
living material cannot be imaged by electron microscopy.
secondary antibody, that recognizes the primary one and al-
In this section, we describe various different approaches
lows it to be localized.
to viewing biological material by electron microscopy. The
• Short sequences encoding epitope tags can be appended to most widely used instrument is the transmission electron mi-
protein-encoding sequences to allow localization of the ex- croscope, but also in common usage is the scanning electron
pressed protein using an antibody to the epitope tag. microscope (SEM), which provides complementary informa-
tion as we discuss at the end of this section.
• Green fluore~cence protein (GFP) and its derivatives are
naturally occurring fluorescent proteins.
• Fusing GFP to a protein of interest allows its localization Single Molecules or Structures Can Be
and dynamics to be explored in a living cell. Imaged After a Negative Stain or Metal
• Deconvolution and confocal microscopy provide greatly Shadowing
improved clarity in fluor~scent images by removing out-of- It is common in biology to explore the detailed shape of sin-
focus fluorescent light. gle macromolecules, such as proteins or nucleic acids, or of
• Total internal reflection (TIRF) microscopy allows fluores- structures, such as viruses and the filaments that make up the
cent samples adjacent to a coverslip to be seen with great clarity. cytoskeleton. It is relatively easy to view these in the transmis-
sion electron microscope provided they are stained with a
• Fluorescence recovery after photobleaching (FRAP) of a
heavy metal that scatters the incident electrons. To prepare a
GFP fusion protein allows the dynamics of the population of
sample, it is first absorbed to a 3-mm electron microscope
molecules to be analyzed.
grid (Figure 9-26a) coated with a thin film of plastic and car-
• Forster resonance energy transfer (FRET) is a technique in bon. The sample is then bathed in a solution of a heavy metal,
which light energy i~ transferred from one fluorescent pro- such as uranyl acetate, and excess solution removed (Figure
tein to another when the proteins are very close, thereby re- 9-26b). As a result of this procedure, the uranyl acetate coats
vealing when two molecules are close in the cell. the grid but is excluded from the regions where the sample
• Super-resolution microscopy allows for detailed fluores- has adhered. When viewed in the TEM, you observe where
cent images at nanometer resolution. the stain has been excluded, and so the sample is said to be
negatively stained. Because the stain can precisely reveal the

9.3 Electron Microscopy: High-Resolution Imaging 419

 TEM SEM (a} (b)
----Tungsten filament - - - -
(cathode) ~

fn.lt..cl - - - - - - Anode
·-\+;
I
~

- Add
sample
Stain sample
with heavy
metal

Scanning _ _ _ _ __./'..............,..,.... (c)

coils
.·

Electromagnetic
objective lens

Projector l ens

Specimen .,
FIGURE 9-25 In electron microscopy, images are formed from
electrons that pass through a specimen or are scattered from a
metal-coated specimen. In a transmission electron microscope
(TEM), electrons are extracted from a heated filament, accelerated by
an electric field, and focused on the specimen by a magnetic
condenser lens. Electrons that pass th rough the specimen are
focused by a series of magnetic objective and projector lenses to
form a magnified image of the specimen on a detector, which may
be a fluorescent viewing screen, a photographic film, or a charged- 100 nm
couple-device (CCD) camera. In a scanning electron microscope
(SEM), electrons are focused by condenser and objective lenses on a FIGURE 9 -26 Transmission electron microscopy of negatively
metal-coated specimen. Scanning coils move the beam across the stained samples reveals fine features. (a) Samples for transmission
specimen, and electrons scattered from the metal are collected by a electron microscopy (TEM) are usually mounted on a small copper or
photomultiplier tube detector. In both types of microscopes, because gold grid. The grid is usually covered with a very thin film of plastic
electrons are easily scattered by air molecules, the entire column is and carbon to which a sample can adhere. (b) The specimen is then
maintained at a very high vacuum. incubated in a heavy metal, such as uranyl acetate, and excess stain
removed. (c) When observed in the TEM, the sample excludes the stain,
so it is seen in negative outline. The example in (c) is a negative stain
of rotaviruses. [Part (c) ISM/Phototake.)

topology of the sample, a high-resolution image can be ob- niques provide information about the three-dimensional to-
tained (figure 9-26c). pology of the sample (Figure 9-27).
Samples can also be prepared by metal shadowing. In this
technique, the sample is absorbed to a small piece of mica,
Cells and Tissues Are Cut into Thin Sections
then coated with a thin film of platinum by evaporation of
the metal, followed by dissolving the sample with acid or for Viewing by Electron Microscopy
bleach. The platinum coating can be generated from a fixed Single cells and pieces of tissue are too thick to be viewed
angle or at a low angle as the sample is rotated, in which case directly in the standard transmission electron microscope. To
it is called low-angle rotary shadowing. When the sample is overcome this, methods were developed to prepare and cut
transferred to a grid and examined in the TEM, these tech- thin sections of cells and tissues. When these were examined

420 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 (a) (b)
Sample......._ Mica surface
D 2..~ ~ 1

fJ ~/
~ ~ ..
Evaporated platinum

.
Metal replica

Evaporated carbon

~ ~ ~ ~
II ~"boofHm
1p.m

Metal replica ready

for visualization

FIGURE 9-27 Metal shadowing makes surface details on very electron micrographs of such preparations, the carbon-coated areas
small objects visible by transmission electron microscopy. (a) The appear light-the reverse of micrographs of simple metal-stained
sample is spread on a mica surface and then dried in a vacuum preparations, in which the areas of heaviest metal staining appear the
evaporator (step 0 ). The sample grid is coated w ith a thin film of a darkest. (b) A platinum-shadowed replica of the substructural fibers of
heavy metal, such as p latinum or gold, evaporated from an electrically calfskin collagen, the major structural protein of tendons, bone, and
heated metal filament (step f)). To stabilize the replica, the specimen is similar tissues. The fibers are about 200 nm thick; a characteristic
then coated with a carbon film evaporated from an overhead electrode 64-nm repeated pattern (white parallel lines) is visible along the length
(step 0 ). The biological material is then dissolved by acid and bleach of each fiber. [Courtesy R. Kessel and R. Kardon.]
(step B l. and the remaining metal replica is viewed in a TEM. In

in the electron microscope, the organization, beauty, and developed to use antibodies to localize proteins in thin sec-
complexiry of the cell interior was revealed and led to a revo- tions at the electron microscope level. However, the harsh
lution in cell biology-for the first time, new organelles and procedures used to prepare traditional thin sections-chemical
the first glimpses of the cytoskeleton were seen. fixation and embedding in plastic-can denature or modify
To prepare thin sections, it is necessary to chemically fix the the antigens so that they are no longer recognized by the spe-
sample, dehydrate it, impregnate it with a liquid plastic that cific antibodies. Gentler methods, such as a light fixation,
hardens (similar to Plexiglas), and then cut sections of about 5 sectioning material after freezing at the temperature of liquid
to 100 run in thickness. For structures to be seen, the sample nitrogen, followed by antibody incubations at room tempera-
has to be stained with heavy metals such as uranium and lead ture, have been developed. To make the antibody visible in
salts, which can be done either before embedding in the plastic the electron microscope, it has to be attached to an electron-
or after sections are cut. Examples of cells and tissues viewed by dense marker. One way to do this is to use electron-dense
thin-section electron microscopy appear throughout this book gold particles coated with protein A, a bacterial protein that
(see, for example, Figure 9-33). It is important to realize that binds the Fe segment of all antibody molecules (Figure 9-29).
the images obtained represent just a thin slice through a cell, so Because the gold particles diffract incident electrons, they ap-
to get a three-dimensional view, it is necessary to cut serial sec- pear as dark spots.
tions through the sample and reconstruct the sample from a
series of sequential images (Figure 9-28).
Cryoelectron Microscopy Allows Visualization
lmmunoelectron Microscopy Localizes Proteins of Specimens Without Fixation or Staining
at the Ultrastructural Level Standard transmission electron microscopy cannot be used to
Just like immunofluorescence microscopy for localizing pro- study live cells, and the absence of water causes macromole-
teins at the light-microscope level, methods have been cules to become denatured and nonfunctional. However,

9.3 Electron Microscopy: High-Resolution Imaging 421

0 VIDEO: Three-Dimensional Model of a Golgi Complex

FIGURE 9 - 28 Model of the Golgi complex based on three-

dimensional reconstruction of electron microscopy images.
Transport vesicles (white spheres) that have budded off the rough ER
fuse with the cis membranes (light blue) of the Golgi complex. By
mechanisms described in Chapter 14, proteins move from the cis
region to the medial region and finally to the trans region of the Golgi
complex. Eventually, vesicles bud off the trans-Golgi membranes
(orange and red); some move to the cell surface and others move to
lysosomes. The Golgi complex, like the rough endoplasm ic reticulum, is
especially prominent in secretory cells. [Brad J. Marsh & Katheryn E. Howell,
Nature Reviews Molecular Cell Biology 3, 789-785 (2002).]

hydrated, unfixed, and unstained biological specimens can be of multiple images can make use of the symmetry of the par-
viewed directly in a transmission electron microscope if the ticle to calculate the three-dimensional srru,cture of the cap-
sample is frozen. In this technique of cryoelectron microscopy, sid to about 5-nm resolution. Examples of such images arc
an aqueous suspension of a sample is applied to a grid in an shown in Figure 4-44.
extremely thin film, frozen in liquid nitrogen, and maintained An extension of this technique, cryoelectron tomogra-
in this state by means of a special mount. The frozen sample phy, allows researchers to determine the three-dimensional
then is placed in the electron microscope. The very low tem-
perature ( - 196 oq keeps water from evaporating, even in a
vacuum. Thus the sample can be observed in detail in its native,
hydrated state without fixing or heavy metal staining. By com-
puter-based averaging of hundreds of images, a three-dimen-
sional model can be generated almost to atomic resolution. For
example, this method has been used to generate models of ribo-
somes, the muscle calcium pump discussed in Chapter 11, and
other large proteins that are difficult to crystallize.
Many viruses have coats, or capsids, that contain multi-
ple copies of one or a few proteins arranged in a symmetric
array. In a cryoelectron microscope, images of these particles
can be viewed from a number of angles. A computer analysis (b) Peroxisomes

FIGURE 9-29 Gold particles coated with protein A are used

to detect an antibody-bound protein by transmission electron
microscopy. (a) First, antibodies are allowed to interact with their
specific antigen (e.g., catalase) in a section of fixed tissue. Then the
section is treated with electron-dense gold particles coated with
protein A from the bacterium 5. aureus. Binding of the bound protein A
to the Fe domains of the antibody molecules makes the location of the
target protein, catalase in this case, visible in the electron microscope.
(b) A slice of liver tissue was fixed w ith glutaraldehyde, sectioned, and
then treated as described in part (a) to localize catalase. The gold
particles (black dots) indicating the presence of catalase are located
exclusively in peroxisomes. [From H. J. Geuze et al., 1981,1. Cell Bioi. 89:653.
Reproduced from the Journal of Cell Biology by copyright permission
ofThe Rockefeller Unrversity Press.]

422 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 ~
(a)
\ I
....... .itt"

•
, ,~I ' ,
'' 61\"W
~

(d) ' ~
t
~
......

FIGURE 9 - 30 Structure of the nuclear pore complex (NPC) by nitrogen and maintained in this state as the sample was observed in
cryoelectron tomography. (a) In electron tomography, a semicircular the electron microscope. The panel shows three sequential tilted
series of two-dimensional projection images is recorded from the images. Different orientations of NPCs (arrows) are shown in top view
three-dimensional specimen that is located at the center; the specimen (left and center) and side view (rig ht). Ribosomes connected to the
is tilted while the electron optics and detector remain stationary. The outer nuclear membrane are visible, as is a patch of rough ER (arrow-
three-dimensional structure is computed from the individual two- heads). (c) Computer-generated surface-rendered representation of a
dimensional images that are obtained when the object is imaged by segment of the nuclear envelope membrane (yellow) studded with
electrons coming from different directions (arrows in left panel). These NPCs (blue). (d) By averaging the images of multiple nuclear pores,
individual images are used to generate a three-dimensional image of much more detail can be discerned. [Part (a) after s. Nickell et al., 2006,
the object (arrows, right panel). (b) Isolated nuclei from the cellular Narure Rev. Mol. Cell. Bioi. 7:225. Parts (b), (c), and (d) from M. Becket al., 2004,
slime mold Dictyoste/ium discoideum were quick-frozen in liquid Science 306:1387.]

architecture of organelles or even whole cells embedded in Scanning Electron Microscopy

ice, that is, in a state close to life. In this technique, the spec-
of Metal-Coated Specimens Reveals
imen holder is tilted in small increments around the axis per-
pendicular to the electron beam; thus images of the object Surface Features
viewed from different directions are obtained (Figure 9-30a, b). Scanning electron microscopy (SEM) allows investigator~ to
The images are then merged computationally into a three- view the surfaces of unsectioned metal-coated specimens. An
dimensional reconstruction termed a tomogram (Figure intense electron beam inside the microscope scans rapidly
9-30c, d). A disadvantage of cryoelectron tomography is over the sample. Molecules in the coating are excited and
that the samples must be relatively thin, about 200 nm; this release secondary electrons that are focused onto a scintilla-
is much thinner than the samples (200 f.Lm thick ) that can be tion detector; the resulting signal is displayed on a cathode-
studied by confocal light microscopy. ray tube much like a conventional television (see Figure 9-25,

9.3 Electron Microscopy: High-Resolution Imaging 423

FIGURE 9-31 Scanning electron microscopy (SEMI produces
a three-dimensional image of the surface of an unsectioned
specimen. Shown here is an SEM image of the epithelium lining the
lumen of the intestine. Abundant fingerlike microvilli extend from
the lumen-facing surface of each cell. The basal lamina beneath the Absorptive
epithelial
epithelium helps support and anchor it to the underlying connective cells } Microvilli
tissue. Compare this image of intestinal cells with that in Figure 9-13,
a fluorescence micrograph. [From R. Kessel and R. Kardon, 1979, Tissues and
Organs: A Text-Atlas ofScanning Electron Microscopy, W. H. Freeman and
Company, p. 176.]
Basal ·.
lamina

right ). The resulting scanning electron micrograph has a

three-dimensional appearance because the number of sec-
ondary electrons produced by any one point on the sample
depends on the angle of the electron beam in relation to the
surface (Figure 9-31 ). The resolving power of scanning elec-
specific set of proteins to perform the function. Cell bio logists
tron microscopes, which is limited by the thickness of the
use this fact to identify specific organelles. For example, as
metal coating, is only about 10 nm, much less than that of
discussed in Chapter 12, most of the ATP in a cell is made by
transmission instruments.
ATP synthase, which converts ADP to ATP and is localized in
mitochondria, and ATP synthase is a good marker for mito-
chondria. As we will discuss below, the availability of specific
markers for organelles has helped in the development of organ-
KEY CONCEPTS of Section 9.3 elle purification.
Electron Microscopy: High-Resolution Imaging In this section, we first give a brief overview of the organ-
elles of eukaryotic cells as a prelude to their discussion in
Electron microscopy provides very high-resolution images much more detail in later chapters. We then discuss methods
because of the short wavelength of the high-energy electrons
that are used to open up cells for the purification of organ-
used to image the sample. elles. We end with recent advances in proteomics aimed at
Simple specimens, such as proteins or viruses, can be nega- defining the complete protein inventory of organelles.
tively stained or shadowed with heavy metals for examina-
tion in a transmission electron microscope (TEM).
Thicker sectiqns generally must be fixed, dehydrated, em- Organelles of the Eukaryotic Cell
bedded in plastic, sectioned, and then stained with electron- The major organelles in animal and plant cells are depicted in
dense heavy metals before viewing by TIM. Figure 9-32, and some are shown in more detail in Figure 9-33.
Specific proteins can be localized by TEM by employing The plasma membrane encloses the cell and is a vitally im-
specific antibodies associated with a heavy metal marker, portant barrier since it is the interface a cell has with its envi-
ronment, separating the outside world from the internal
such as small gold particles.
cytoplasm . It is made up of about an equal mass of lipids
Cryoelectron microscopy allows examination of hydrated, and proteins. Although the physical properties of the plasma
unfixed, and unstained biological specimens in the TEM by membrane are largely determined by its lipid composition, a
maintaining them at a few degrees above absolute zero. membrane's complement of proteins is primarily responsible
Scanning electron microscopy (SEM) of metal-shadowed for the membrane's functional properties. As we d iscuss in
material reveals the surface features of specimens. Chapter 13, the plasma membrane is a permeability barrier
that contains specific membrane tr ansport proteins neces-
sary to bring ions and metabolites into the cell. It is also the
site of reception of chemical signals from other cells, so it
contains receptors that sense these signals and transmits the
9.4 Isolation and Characterization information across the plasma membrane into the cytosol, a
topic we discuss in Chapter 15. The plasma membrane de-
of Cell Organelles
fines the shape of a cell, and so it is intimately associated
The examination of cells by light and electron microscopy led with the cytoskeleton and with other cells and the cellular
to the appreciation that eukaryotic cells contain a common set matrix, interactions we cover in Chapters 17 through 19.
of organelles. Most organelles are enclosed in a lipid bilayer Larger molecules than ions and metabolites can be taken
and perform a specific function. To perform this function, each up by endocytosis, a process involving the invagination of
type of organelle has a recognizable structure and contains a the plasma membrane, which then pinches off to form an

424 CHAPTER 9 • Culturing, Visualizing , and Perturbing Cells

Animal cell D Plasma membrane controls movement of molecules in and
out of the cell and functions in cell·cell signaling and cell
adhesion .
fJ Mitochondria, which are surrounded by a double membrane,
generate ATP by oxidation of glucose and fatty acids.
D Lysosomes, which have an acidic lumen, degrade material
internalized by the cell and worn-out cellular membranes and
organelles.
a Nuclear envelope, a double membrane, encloses the contents
of the nucleus; the outer nuclear membrane is continuous
with the rough ER.
II!J Nucleolus is a nuclear subcompartment where most of the
cell's rRNA is synthesized.
I!J Nucleus is filled with chromatin composed of DNA and
proteins; site of mRNA and tRNA synthesis.
IJ Smooth endoplasmic reticulum (ER) synthesizes lipids and
detoxifies certain hydrophobic compounds.

a Rough endoplasmic reticulum (ER) functions in the synthesis,

processing, and sorting of secreted proteins, lysosomal
proteins, and certain membrane proteins.
EJ Golgi complex processes and sorts secreted protei ns,
lysosomal proteins, and membrane proteins synthesized on
the rough ER.
ill] Secretory vesicles store secreted proteins and fuse with the
plasma membrane to release their contents.
II] Peroxisomes detoxify various molecules and also break down
fatty acids to produce acetyl groups for biosynthesis.

mCytoskeletal fibers form networks and bundles that support

cellular membranes, help organize organelles, and participate
in cell movement.
I!] Microvilli increase surface area for absorption of nutrients
from surrounding medium.
iiJ Cell wall, composed largely of cellulose, helps maintain the
cell's shape and provides protection against mechanical
stress.
iE Vacuole stores water, ions, and nutrients, degrades
macromolecules, and functions in cell elongation during
growth.
1m Chloroplasts, which carry out photosynthesis, are surrounded
by a double membrane and contain a network of internal
membrane-bounded sacs.

FIGURE 9 · 32 Schematic overview of a "typical" animal cell (top) here, and other substructures can be present in some. Cells also differ
and plant cell (bottom) and their major substructures. Not every cell considerably in shape and in the prominence of various organelles and
will contain all the organelles, granules, and fibrous structures shown substructures.

endosome in the cytopla~. During the best-studied form of interconnected network of flattened membrane-bound sacs and
endocytosis, special regions of the plasma membrane called tubules. The ER can be divided into the smooth endoplasmic
coated pits are formed in which receptors collect and bring reticulum, so called because the membrane has a smooth sur-
specific molecules or particles into the cell (Figure 9-33a). This face, and the rough endoplasmic reticulum, which is studded
process is known as receptor-mediated endocytosis. Once ma- with ribosomes (Figure 9-33b). The smooth endoplasmic reticu-
terials are internalized, they are sorted and can either be re- lum is the site of synthesis of fatty acids and phospholipids. In
turned to the plasma membrane or delivered to lysosomes for contrast, the rough endoplasmic reticulum with its associated
degradation. Lysosomes contain a battery of degradative en- ribosomes is the site of synthesis of membrane proteins and pro-
zymes that can break down essentially any biological molecule teins that will be secreted out of a cell, accounting for about
into smaller components. The lumen of lysosomes has an acidic one-third of all the different types of proteins synthesized by a
pH of 4.5; this helps to denature proteins and the degradative cell. After synthesis on the ER, proteins destined for the plasma
enzymes-collectively known as acid hydrolases-can with- membrane or for secretion are first transported to the Golgi
stand this environment and in fact work optimally at this pH. complex, a stack of flattened membranes called cisternae (Figure
The largest internal membrane system is an organelle known 9-33b), in which the proteins are modified and sorted before
as the endoplasmic reticulum (ER), consisting of an extensive being transported to their destination at the plasma membrane

9.4 Isolation and Characterization of Cell Organelles 425

@ VIDEO: Three-Dimensional Model of a Mitochrondrion

Gold-labeled transferrin Clathrin-coated pit (b)

Rough
endoplasmic
reticulum

200 nm
100 nm I I
I I

Outer (d)
(c) Inner membrane Cristae membrane

Grana~~----------~•
Thylakoid
membrane
Stroma _ ___,.,;.._;_

Chloroplast----
membranes
(outer and inner)

Starch--:.,..;+:--
lntermembrane Matrix Matrix
granule
space granules

1 I'-m
I

FIGURE 9-33 Examples of organelles viewed by transmission (d) Plants contain chloroplasts, another double-membrane organelle.
electron microscopy of thin sections. (a) The plasma membrane The thylakoid membranes contain the enzymes of the photosynthetic
contains a clathrin-coated pit. Before fixation, the cells were incubated pathway that involves the conversion of light energy into ATP. [Part (a)
with colloidal gold-labeled transferrin, a protein involved in iron uptake from C. Lamaze et al., 1997, Journal of Biological Chemistry 272:20332; part (b)
through a receptor localized in coated pits (see Chapter 14). (b) A from G. Palade collection; part (c) from D. W. Fawcett, 1981, The Cell, 2d ed.,
section through a secretory cell shows the ribosome-studded endoplas- Saunders, p. 415; part (d) courtesy of Biophoto Associates/M. C. Ledbetter/
mic reticulum and the Golgi complex. (c) The two membranes of a Brookhaven National Laboratory.
mitochondrion and the membrane infoldings called cristae are shown.

or, in some cases, delivered to endosomes. Because proteins des- membrane proteins that will remain in the endoplasmic reticu-
tined for secretion are made on the endoplasmic reticulum, lum, the Golgi complex, and the plasma membrane and are
transported through the Golgi complex, and released from the therefore not secreted. These so-called membrane-trafficking
cell, this whole process is collectively known as the secretory pathways, encompassing both the endocytic and the secretory
pathway, although it also includes the synthesis and transport of pathways, are discussed in detail in Chapter 14.

. . '
426 CHAPTER 9 • Culturing, Visualizing , and Perturbing Cells
 Another common organelle is the peroxisome, a class of Disruption of Cells Releases Their Organelles
roughly spherical organelles that contain oxidases--enzymes and Other Contents
that use molecular oxygen to oxidize toxins to produce harm-
less products and for the oxidation of fatty acids for the pro- The initial step in purifying subcellular structures is to release
duction of acetyl groups, a topic we come to in Chapter 12. the cell's contents by rupturing the plasma membrane and
All the organelles discussed so far are enclosed by a single the cell wall, if present. First, the cells are suspended in a
lipid bilayer membrane. Some organelles, namely the nucleus, solution of appropriate pH and salt content, usually isotonic
mitochondria, and, in plant cells, chloroplasts, have an ad - sucrose (0.25 M) or a combination of salts similar in compo-
ditional membrane that serves various functions, as we sition to those in the cell's interior. Many cells can then be
describe below. broken by stirring the cell suspension in a high-speed blender
The nucleus contains the DNA of the genome and is the or by exposing it to ultrahigh-frequency sound (sonication ).
site of transcription of the DNA into messenger RNA. The Alternatively, plasma membranes can be sheared by special
nucleus has an inner membrane that defines the nucleus itself. pressurized tissue homogenizers in which the cells are forced
It also contains an outer membrane that is continuous with through a very narrow space between a plunger and the ves-
the membrane of the endoplasmic reticulum in such a way sel wall; the pressure of being forced between the wall of the
that the space between the inner and outer nuclear mem- vessel and the plunger ruptures the cell.
branes is continuous with the endoplasmic reticulum (see Fig- Recall that water flows into cells when they are placed in a
ure 9-32). Access in and out of the nucleus is provided by hypotonic solution, that is, one with a lower concentration of
tubular connections between the inner and outer membrane ions and small molecules than found inside the cell. This osmotic
stabilized by structures called nuclear pores. Nuclear pores tlow can be used to cause cells to swell, weakening the plasma
not only define the site of transport across the nuclear mem- membrane and facilitating its rupture. Generally, the cell solu-
branes but act as gatekeepers, only allowing transport of spe- tion is kept at 0 °C to best preserve enzymes and other constitu-
cific macromolecules in and out of the nucleus-an important ents after their release from the stabilizing forces of the cell.
and fascinating topic we mentioned in Chapter 8 and will Disrupting the cell produces a mix of suspended cellular
discuss in more detail in Chapter 13. components, the homogenate, from which the desired organ-
Mitochondria and chloroplasts are believed to have elles can be retrieved. Because rat liver contains an abun-
evolved from an event a long time ago when a eukaryotic cell dance of a single cell type, this tissue has been used in many
engulfed one type of bacterium that gave rise to mitochondria classic studies of cell organelles. However, the same isolation
and a different kind that gave rise to chloroplasts. That mito- principles apply to virtually all cells and tissues, and modifi-
chondria and chloroplasts have two membranes is evidence cations of these cell-fractionation techniques can be used to
supporting this hypothesis. The inner membrane is most separate and purify any desired components.
likely derived from the original bacterial membrane, whereas
the outer membrane is a vestige of the plasma membrane
Centrifugation Can Separate Many Types
from the engulfment event. There is a lot of evidence for the
bacterial original of these organelles, including the fact that of Organelles
mitochondria and chloroplasts both have their own DNA ge- In Chapter 3, we considered the principles of centrifugation
nome and the biosynthesis of proteins in these organelles is and the uses of centrifugation techniques for separating pro-
more similar to bacterial protein synthesis than eukaryotic teins and nucleic acids. Similar approaches are used for sepa-
protein synthesis. rating and purifying the various organelles, which differ in
Mitochondria can occupy as much as 25 percent of the both size and density and thus undergo sedimentation at
volume of the cytoplasm. They are thread-like organelles in different rates.
which the outer membrane contains porin proteins that ren - Most cell-fractionation procedures begin with differen-
der the membrane permeable to molecules up to a molecular tral centrifugation of a filtered cell homogenate at increas-
weight of 10,000. The inner mitochondrial membrane is ingly higher speeds (Figure 9-34). After centrifugation at
highly convoluted with infoldings called cristae that pro- each speed for an appropriate time, the liquid that remains
trude into the central space, called the matrix (Figure 9-33c). at the top of the vessel, called the supernatant, is poured off
A major function of mitochondria is to complete the termi- and centrifuged at higher speed. The pelleted fractions ob-
nal stages of degradation of glucose by oxidation to generate tained by differential centrifugation generally contain a mix-
most of the ATP supply of the cell. Thus mitochondria can ture of organelles, although nuclei and viral particles can
be regarded as the "power plants" of the cell. sometimes be purified completely by this procedure.
All plants and green algae are characterized by the pres- An impure organelle fraction obtained by differential cen-
t:m.:e of chloroplasts (Figure 9-33d), organelles that use pho- trifugation can be further purified by equilibrium density-
tosynthesis to capture the energy of light with colored gradient centrifugation, which separates cellular components
pigments, including the green pigment chlorophyll, and ulti- according to their density. After the fraction is resuspended, it
mately store the captured energy in the form of ATP. The is layered on top of a solution that contains a gradient of a
processes by which ATP is made in mitochondria and chlo- dense non-ionic substance (e.g., sucrose or glycerol). The
roplasts is described in Chapter 12. tube is centrifuged at a high speed (about 40,000 rpm) for

9.4 Isolation and Characterization of Cell Organelles 427

 C Filter
homogenate
to remove
clumps of
unbroken
cells,
connective
tissue, etc.

Centrifuge

600g x
10min
Pour. out:
15,QQQg X
5 m1n
A }Po u,r o u tg: x
100 0 0 0
60 min
_
l Pour out:
J3QQ,QQQg X
2h _

I
}

~Nuclei [Mit~c~dria, ~~
I
Filtered Soluble
homogenate chloroplasts, membrane, subunits, part of
lysosomes, microsomal small cytoplasm
and fraction polyribo- (cytosol)
peroxisomes (fragments of somes
endoplasmic
reticulum),
and large
polyribosomes
FIGURE 9-34 Differential centrifugation is Subsequent centrifugation in the ultracentrifuge at 1OO,OOOg for 60
a common first step in fractionating a cell homogenate. The minutes results in deposition of the plasma membrane, fragments of
homogenate resulting from disrupting cells is usually filtered to the endoplasmic reticulum, and large polyribosomes. The recovery of
remove unbroken cells and then centrifuged at a fairly low speed to ribosomal subunits, small polyribosomes, and particles such as
selectively pellet the nucleus-the largest organelle. The undeposited complexes of enzymes requires additional centrifugation at still higher
material (the supernatant) is next centrifuged at a higher speed to speeds. Only the cytosol-the soluble aqueous part of the cytoplasm- . I
sediment the mitochondria, chloroplasts, lysosomes, and peroxisomes. remains in the supernatant after centrifugation at 300,000g for 2 hours.

several hours, allowing each particle to migrate to an equi- elle-specific marker molecules can be quantified. For example,
librium position where the density of the surrounding liquid the protein cytochrome c is present only in mitochondria, so
is equal to the density of the particle (Figure 9-35). The dif- the presence of this protein in a fraction of lysosomes would
ferent layers of liquid are then recovered by pumping out the indicate its contamination by mitochondria. Similarly, cata-
contents of the centrifuge tube through a narrow piece of lase is present only in peroxisomes; acid phosphatase, only in
tubing and collecting fractions.
Because each organelle has unique morphological fea-
tures, the purity of organelle preparations can be assessed by
examination in an electron microscope. Alternatively, organ-

FIGURE 9-35 A mixed-organelle fraction can be further

separated by equilibrium density-gradient centrifugation. In this ....0
example, utilizing rat liver, material in the pellet from centrifugation at i!'
'iii-
1S,OOOg (see Figure 9-34) is resuspended and layered on a gradient of c"'
E
Ill
increasingly dense sucrose solutions in a centrifuge tube. During 'OU

centrifugation for several hours, each organelle migrates to its OlOl Mitochondria_,
c- (1.18 g/cm 3 )
·- Q)
appropriate equilibrium density and remains there. To obtain a good "'"'
<~~o
Ill ...
separation of lysosomes from mitochondria, the liver is perfused with ... u
u :J
a solution containing a small amount of detergent before the tissue is .E"' Peroxisomes {
(1.23 g/cm 3 )
disrupted. During this perfusion period, detergent is taken into the
cells by endocytosis and transferred to the lysosomes, making them
less dense than they would normally be and permitting a "clean" Before After
separation of lysosomes from mitochondria. centrifugation centrifugation

428 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

lysosomes; and ribosomes, only in the rough endoplasmic re- nm in diameter ) and density, which makes them difficult to
ticulum or the cytosol. separate from one another by centrifugation techniques. Im-
munological techniques are particularly useful for purifying
specific classes of such vesicles. Fat and muscle cells, for in-
Organelle-Specific Antibodies Are Useful
stance, contain a particular glucose transporter (GLUT4)
in Preparing Highly Purified Organelles that is localized to the membrane of one of these types of
Cell fractions remaining after differential and equilibrium vesicle. When insulin is added to the cells, these vesicles fuse
density-gradient centrifugation usually contain more than one with the plasma membrane and increase the number of glu-
type of organelle. Monoclonal antibodies for various organelle- cose transporters able ro take up glucose from the blood. As
specific membrane proteins are a powerful tool for further we will see in Chapter 15, this process is critical to maintain-
purifying such fractions. One example is the purification of ing the appropriate concentration of sugar in the blood. The
vesicles whose outer surface is covered with the protein clath- GLUT4-containing vesicles can be purified by using an anti-
rin; these coated vesicles are derived from coated pits at the body that binds to a segment of the GLUT4 protein that
plasma membrane during receptor-mediated endocytosis (see faces the cytosol. Likewise, the various transport vesicles dis-
Figure 9-33a), a topic discussed in detail in Chapter 14. An cussed in Chapter 14 are characterized by unique surface
antibody to clathrin, bound ro a bacterial carrier, can selec- proteins that permit their separation with the aid of specific
tively bind these vesicles in a crude preparation of membranes, antibodies.
and the whole antibody complex can then be isolated by low- A variation of this technique is employed when no anti-
speed centrifugation (Figure 9-36). A related technique uses body specific for the organelle under study is available. A gene
tiny metallic beads coated with specific antibodies. Organelles encoding an organelle-specific membrane protein is modified
that bind to the antibodies, and are thus linked to the metallic by the addition of a segment encoding an epitope tag; the tag
beads, are recovered from rhe preparation by adhesion to a is placed on a segment of the protein that faces the cytosol.
small magnet on the side of the test tube. Following stable expression of the recombinant protein in the
All cells contain a dozen or more different types of small cell under study, an anti-epitope monoclonal antibody (de-
membrane-limited vesicles of about the same size (50 to 100 scribed above) can be used to purify the organelle.

(a) (b)
Coated
vesicles
Clathrin Bacterial cell
Antibody to clathrin

Protein A

Coated vesicle

FIGURE 9·36 Small coated vesicles can be purified by binding region of antibodies. (a) Interaction of protein A with dntibodies bound
of antibody specific for a vesicle surface protein and linkage to to clathrin-coated vesicles links the vesicles to the bacterial cells. The
bacterial cells. In this example, a suspension of membranes from rat vesicle-bacteria complexes can then be recovered by low-speed
liver is incubated with an antibody specific for clathrin, a protein that centrifugation. (b) A thin-section electron micrograph reveals
coats the outer surface of certain cytosolic vesicles. To this mixture is clathrin-coated vesicles bound to an 5. au reus cell. [See E. Merisko et al ..
added a suspension of killed Staphylococcus au reus bacteria, whose 1982, J. Cell Bioi. 93:846. Micrograph courtesy of G. Palade.)
surface membrane contains protein A, which binds to the constant (Fe)

9.4 Isolation and Characterization of Cell Organelles 429

Proteomics Reveals the Protein complicated biochemical processes, such as DNA replication
Composition of Organelles or protein synthesis. These biochemical approaches have
been complemented by genetic approaches, and as we have
To identify all the proteins in an organelle requires three steps. seen in Chapter 5, mutations can be used to identify genes
First, one has to be able to obtain the organelle in high purity. whose products play specific functions. As we will see in
Second, one has to have a way to identify all the sequences of Chapters 14 and 19, classic genetic screens in yeast were used
the proteins in the organelle. This identification is generally to identify proteins that participate in the secretory pathway
done by digesting all the proteins with a protease such as tryp- and the cell cycle, respectively. Genetic approaches in other
sin, which cleaves all polypeptides at lysine and arginine resi- orgauisrn~, such as the nematode worm, the fruit fly, and the
dues, and then determining the mass and sequence of all these mouse, have contributed immensely to uncovering basic as-
peptides by mass spectrometry. Third, one has to have a ge- pects of cell biology and development (see Chapter 1).
nomic sequence to identify the proteins from which all the Over the last few years, additional new and very powerful
peptides came. In this way, the "proteome" of many organ- approaches have been developed to perturb specific compo-
elles has been determined. As one example, a recent proteomic nents in living cells and thereby shed light on their functions.
study on mitochondria purified from mouse brain, heart, kid- In this section, we discuss two of these approaches: the use of
ney, and liver revealed 591 mitochondrial proteins, including specific chemicals to perturb cell function and the use of in-
163 proteins not previously known to be associated with this terfering RNA to suppress the expPession of specific genes.
organelle. Several proteins were found in mitochondria only
in specific cell types. Determining the functions associated
with these newly identified mitochondrial proteins is a major Drugs Are Commonly Used in Cell Biology
objective of current research on this organelle. Naturally occurring drugs have been used for centuries, but
how they worked was often not known. For example, ex-
tracts of the meadow saffron were used to treat gout, a pain-
ful disease resulting from inflammation of j"oints. Today we
KEY CONCEPTS of Section 9.4 know that the extract contains colchicine, a drug that de-
polymerizes microtubules and interferes with the ability of
Isolation and Characterization of Cell Organelles white blood cells to move to the sites of inflammation (see
Microscopy has revealed a common set of organelles pres- Chapter 18). Alexander Fleming discovered that certain
ent in eukaryotic cells (see Figure 9-32). fungi secrete compounds that kill bacteria (antibiotics), re-
• Disruption of cells by vigorous homogenization, sonica- sulting in the discovery of penicillin. Only later was it was
tion, or other techniques releases their organelles. Swelling of discovered that penicillin inhibits bacterial cell division by
blocking the assembly of the cell walls of certain bacteria.
cells in a hypotonic solution weakens the plasma membrane,
making it easier to rupture. Many examples like these have resulted in the discovery
of a very wide range of drugs available to inhibit specific and
• Sequential differential centrifugation of a cell homogenate essential processes of cells. In most cases, researchers have
yields fractions of partly purified organelles that differ in mass eventually been able to identify the molecular target of the
and density. drug. For example, there are many other antibiotic drugs that
• Equilibrium density-gradient centrifugation, which separates affect aspects of prokaryotic protein synthesis. A selection of
cellular components according to their densities, can further some of the more commonly used drugs that affect a broad
purify cell fractions obtained by differential centrifugation. variety of cell biological processes are listed in Table 9-1,
grouped according to the process they inhibit.
• Immunological techniques using antibodies against organ-
elle-specific membrane proteins are particularly useful in pu-
rifying organelles and vesicles of similar sizes and densities. Chemical Screens Can Identify
• Proteomic analysis can identify all the protein components New Specific Drugs
in a preparation of a purified organelle. How does one discover a new drug? One widely used ap-
proach makes use of chemica/libraries consisting of 1O,OOOs
to 100,000s of different compounds to search for chemicals
that inhibit a specific process. The screening of chemical li-
braries in conjunction with high-throughput microscopic
9.5 Perturbing Specific Cell Functions techniques has now become one of the major routes for new
What general approaches have scientists used to understand leads in drug discovery. Here we give just one case to illus-
the function of specific proteins in cell biological processes? trate how this type of approach works.
We have already discussed in Chapter 3 how proteins can be In our example (Figu re 9-37a ), researchers wanted to
purified and their properties characterized in detail. In many identify compounds that inhibit mitosis, the process where
cases, this has led to the in vitro biochemical reconstitution of duplicated chromosomes are accurately segregated by a

430 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 TABLE 9-1 Selected Set of Small Molecules Used in Cell Biological Research

Some of the following molecules have broad specificity, whereas others are highly specific. More information about many of these
compounds can be found in the relevant chapters in this text.

DNA replication inhibitors Aphidicolin (eukaryonc DNA polymerase inhibitor); camptothecin, etoposide (eukaryotic
topoisomerase inhibitors)

Transcription inhi bit ors a-Amanitin (eukaryotic RNA polymerase II inhibitor); actinomycin D (eukaryoric tran~crip­
tion elongation inhibitor); rifampicin (bacterial RNA polymerase inhibitor); thiolutin (bac-
terial and yeast RNA polymerase inhibitor)

Protein synthesis inhibitors- Cycloheximide (translational inhibitor in eukaryotes); genericin/G418, hygromycin, puromycin
block general protein production; (translation inhibitors in bacteria and eukaryotes); chloramphenicol (translation inhibitor m
toxic after extended exposure bacteria and mitochondria); tetracycline (translation inhibitor in bacteria)

Protease inhibitors- block MG-132, lactacystin (proteasome inhibitors); E-64, leupeptin (serine and/or cysteine protease
prot ein degradation inhibitOrs); phenylmethanesulfonylfluoride (PMSF) (serine proteases inhibitor); tosyi-L-lysme
chloromcthyl ketone (TLCK) (trypsin-like serine protease inhibitor)
·.· --
Compounds affecting Phalloidin, jasplakinolide (F-actin stabilizer); latrunculin, cytochalasin (F-actin polymerization
the cytoskeleton inhibitors); taxol (microtubule stabilizer); colchicine, nocodazole, vinblastine, podophyllo-
toxin (microtubule polymerization inhibitors); monastrol (kinesin-5 mhibitor)

Compounds affect ing membrane Brefeldin A (secretion inhibitor); lepromycin B (nuclear protein export inhibitOr);
traffic, intracellular movement dynasore (dynamin inhibitor); tunicamycin (N-Iinked glycosylation inhibitOr)
and t he secretory pathway,
protein glycosylation

Kinase inhibitors Genistein, rapamycin, gleevec (tyrosine kinase inhibirors with various specificities);
worrmannin, LY294002 (PB kinase inhibitors); staurosporine (protein kinase inhibitor);
roscovmne (cell cycle CDKl and CDK2 inhibitors)

Phosphatase inhibitors Cyclosporine A, FK506, calyculin (protein phosphatase inhibitors with various specificities);
okadaic acid (general inhibitor of serine/threonine phosphatases); phenylarsine oxide, sodium
orthovanadate (tyrosine phosphatase inhibitors)

Compounds affecting Forskolin (adenylate cyclase activator)

intracellular cAMP levels

Compounds affecting ions A23187 (Ca2 · ionophore); valinomycin (K+ ionophore); BAPTA (divalent cation (e.g., Cah)
(e.g., K+ Ca2+) binding/sequestering agent); thapsigargin (endoplasmic reticulum Ca2 + ATPase inhibitor);
ouabain (Na IK ... ATPase inhibitor)
+

Some drugs used in medicine Propranolol (13-adrenergic receptor antagonist), statins (HMG-CoA reductase mhibitors,
block cholesterol synthesis)

microtubule-based machine called the mitOtic spindle (dis- to tubulin, the major protein of microtubules. Over 16,000
cussed in Chapter 18). It was known that if spindle assembly compounds were screened, and a compound was identified
is compromised, cells arrest in mitosis. Therefore, the screen that resulted in cells with abnormal spindles-instead of having
first used an automated robotic method ro look for com- two asters, they had a single aster, what is called a mono-
pounds that arrest cells in mitosis. The basis for the inhibi- astral array (Figure 9-37b). This drug, now called monastrol,
tion of the candidate compounds wa~ then explored tO see if was found to interfere with the assembly of the spindle by
they affected assembly of the microtubules. Since inhibition inhibiting a microtubule-based motor called kinesin-5 (see
of microtubule assembly was not of interest, the effect of the Chapter 18 for more details about the mitotic spindle). De-
remaining candidates on the structure of the spindle was de- rivatives of monastrol are now being tested as anti-tumor
termined by immunofluorescence microscopy with antibodies agents for the treatment of certain cancers.

9.5 Perturbing Specific Cell Functions 431

FIGURE 9 -37 Screening for drugs that affect (a) (b)
specific biological processes. In this example, a 16,320
chemical library of 16,320 different chemicals was Chemical Compounds
subject to a series of screens for inhibitors of mitosis.
Since such an inhibitor is expected to arrest cells at
Screen for those
the mitotic stage of the cell cycle, the first screen 0 D th~t a~rest cells in
was to see if any of the chemicals enhanced the level
of a marker for mitotic cells, and this yielded 139
candidates. Microtubules make up the structure of
the mitotic spindle, and the researchers were not
1
139
mitOSIS

interested in new drugs that target microtubules, so Screen for those

in the second screen fJ they tested the 139 n that do not affect
U microtubule

1
compounds for their ability to affect microtubule
assembly in vitro
assembly, and this eliminated 53 candidates.
Immunofluorescence microscopy with antibodies to
86
tubulin (the major subunit of microtubules) together
with a stain for DNA was then used in the third
Screen for those
screen lJ to identify compounds that disrupt the that specifically
structure of the spindle. (b) Localization of tubulin affect spindle
(green) and DNA (blue) are shown for an untreated morphology
mitotic spindle and one treated with one of the
recovered compounds, now called monastrol. 5
Monastrol inhibits a microtubule-based motor called
kinesin-5, discussed in Chapter 18, necessary to
separate the poles of the mitotic spindle. When
kinesin-5 is inhibited, the two poles remain
associated to give a monopolar spind le. [Part (b)
T. U. Mayer et al., Science 286:971 - 974.]

degraded by the associated argonaute protein. If the single-

Small Interfering RNAs (siRNAs) Can Knock
stranded siRNA sequence can base pair exactly with a target
Down Expression of Specific Proteins mRNA sequence, the argonaute protein-RNA complex
RNA interference (RNAi) is a mechanism that cells use to sup- cleaves the target mRNA, which is then degraded (Figure
press the expression of genes by either blocking translation of 9-38a). Although this system probably evolved as a defense
specific mRNAs 'through micro-RNAs (miRNAs) or degrada- mechanism against invading viruses, it has provided researchers
tion of specific mRNAs targeted by small interfering RNA with a very powerful tool to experimentally suppress the ex-
(siRNA). The extensive use of miRNAs to regulate gene ex- pression of particular genes and explore the resulting conse-
pression, especially during development, has been discussed in quences. It has been used very effectively in many different
Chapter 8. Here we focus on the experimental use of siRNA systems, as we summarize below.
technology to suppress the expression of genes in animal cells.
The discovery of the siRNA pathway arose from many siRNA Knockdown in Cultured Cells Since the discovery in
different observations. For example, it was discovered in 2001 that treatment of cultured cells with siRNAs suppresses
plants that recombinant expression of a gene could lead to gene expression by degrading the target mRNA, siRNAs have
the down-regulation of the target gene rather than the ex- been used in thousands of studies to suppress-or "knock
pected result of enhanced expression. A similar type of result down "-the levels of target proteins. To do this, researchers
was seen in the nematode C. elegans. Investigating this phe- use computer programs to identify a - 2 1-base sequence in the
nomenon, Andrew Fire and Craig .Mello reported in 1998 mRNA that is unique to the target gene and has the characteris-
that suppression could not be achieved by expressing either tics optimal for siRNA. Double-stranded RNA is then synthe-
sense or antisense mRNA, but that it required expression of sized and applied to cells in culrure (Figure 9-38a). If effective,
double-stranded RNA. Fire and Mello were awarded the this will result in degradation of the specific mRNA, and no new
Nobel Prize in Physiology or Medicine in 2006 for this dis- target protein will be synrhesi7ed. However, the target protein is
covery. Subsequent work in a number of systems showed that present at the beginning of the experiment, so the cells have to
the double-stranded RNA has to be cleaved by a protein be able to grow to allow the endogenous protein to undergo its
called Dicer to produce double-stranded fragments of 21 to normal turnover as well as get diluted by cell division- this usu-
23 base pairs with a two-nucleotide overhang at each of the ally takes 24 to 72 hours. The level of the target protein is gener-
3' ends. This double-stranded RNA is recognized by RISC ally determined by immunoblotting (see Figure 3-39), and if
(RNA-induced silencing complex), and one of the strands is significantly reduced, the phenotype of the cells is examined.

432 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

(a} Synthetic siRNA
3' llilliillllllllllllllli 5'
5'J.W11LII'''"'"''"' 3'

3' 111111111 II IIIII IIIII 5'

5' 11111111111111111111111 3'

RISC/a siRNA ~ e
D1cer
m
~
5'
r D
r.;3'5.:::p• shRNA

\IJ
( 5'~ 3'

~ ,t~=-1
5'J \a ~ ~ ,,
Target mRNA /
---.
u
,
, ,,
, ,,

,,,
DNA with construct· to express shRNA

EBP50 (c) EPB50

(b) Control siRNA Control siRNA

EBP50

- Tubulin

FIGURE 9-38 siRNA and DNA expressing shRNA can target the nucleus, the RNA hairpin becomes a substrate of the nuclease Dicer to
degradation of specific mRNAs in cultured cells. In the first step (0 ), generate the appropriate siRNA. (b) As an example of this technology,
a double-stranded siRNA tha~has homology to the target mRNA is researchers wanted to examine the effects of knocking down a protein,
introduced into cells by transfection. This double-stranded RNA is called EBPSO, that is a component of cell-surface microvilli (see Figure
recognized by the RISC complex (f)), which degrades one strand of the 17-21d). siRNAs were designed and their ability to knock down EBPSO
RNA and targets the mRNA with the homologous sequence (D). The in cultured cells assessed by doing an immunoblot with EBPSO
target mRNA is cleaved (D) and degraded ~ ). In an alternative antibodies and tubulin antibodies as a control. (c) They then examined
strategy, a DNA construct containing a sequence that when transcribed the cells for microvilli by staining for the microvillar-specific protein
will form a hairpin is introduced into the cell (I':J). This DNA can either ezrin. In these confocal sections at the top of the cell, untreated cells
be introduced by transfection or carried in a virus particle. In either have abundant microvilli, whereas cells in which EBPSO has been
case, it is engineered to carry with it a drug-selectable marker (not knocked down only have a few microvilli around the cell periphery.
shown) so that cells in which this DNA is integrated into the genome [Parts (h) and (c) from Hanono et al., J. Cell Bioi. 175:803.]
can be selected. When transcribed (6) and transported out of the

9.5 Perturbing Specific Cell Fun ctions 433

 (a)

Grow nematode larval worms on Tra nsfer some adults to lay Allow embryos to grow into adults
special strain of E. coli embryos, then remove adults and observe their phenotype

a IJ
(b) Wild type tac-1 RNAi

FIGURE 9-39 RNAi screens can explore the function of all the corresponding to a nematode gene. (a) In this approach, a specific
genes in the nematode Caenorhabdit;s e/egans. The determination E. coli strain is fed to larval nematodes D and the expression of the .,
of the C. elegans genome sequence in 1998 revealed it contains about target gene is suppressed in the germ line. Because these nematodes
20,000 protein-encoding genes. This information opened the possibility are self-fertilizing hermaphrod ites (having the reproductive organs of
of using RNAi to knock down expression of each gene to explore what both sexes), it is not necessary to mate them but merely to let the adult
effect it would have. Today this is routinely done. The nematode can worms lay eggs f). When these grow into adults, the effect of RNAi on
live by eating the bacterium Escherichia coli, and it is possible to the target gene can be assessed. (b) In this example, the researchers
express in the bacterium a long stretch of double-stranded RNA were screening for genes necessary for nuclear movement. They
corresponding to a single nematode gene. Remarkably, when the identified a gene called TAC-1, whose product is located at centro-
nematode eats the bacteria, the double-stranded RNA enters the cells somes and is necessary for the norma l distribution of microtubules, as
of the gut and is reC!)gnized by Dicer and processed into siRNAs that revealed by immunofluorescence microscopy with tubulin antibodies.
spread to almost all the cells in the animal. Therefore, researchers have [Part (b) from N. Le Bot et al., Currenr Biology 13:1499 (2003).]
made a library of E. coli strains each expressing double-stranded RNA

Suppression of gene expression by siRNAs has become a stan- appropriate conditions that allow the cells to take it up or by
dard technique; an example is shown in Figure 9-38b and 9-38c, use of viral vectors that more efficiently introduce the DNA.
and many other examples can be found throughout this book. Massive efforts are cu rrently under way to explore the
An alternative strategy to knockdown protein expression is effects of knocking down expression of each gene in cultured
to introduce appropriate DNA constructs into cells that will cell lines and then examining the effects on specific path-
generate siRNAs when they are transcribed (see Figure 9-38a). ways. This effort is in its infancy, so future refinements and
To achieve this, the target sequence is present as an inverted analysis will provide a "systems biology" view of cell orga-
repeat in the DNA sequence. When transcribed, the mRNA will nization and function.
form a double-stranded short hairpin RNA (shRNA), which is
recognized and cleaved by Dicer to generate siRNAs. This ap-
proach has the advantage that once the shRNA is expressed,
Genomic Screens Using siRNA
the siRNAs are always made, resulting in permanent knock-
down of the target protein. This will not work if the protein is in the Nematode C. e/egans
essential, m which case treating cells with siRNAs is the method When the annotated genome sequence of the nematode worm
of choice. The DNA construct to express shRNAs can be intro- Caenorhabditis elegans was determined in 1998, this provided
duced into cells by simple addition of the DNA under the the first catalog of all the genes present in an animal. It also

434 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 (a)

- - -
14,425 RNAs 4,000,000 images Automated image analysis Researcher observation

II fJ EJ II
l
150 candidates
(b)

FIGURE 9-40 RNAi screens can explore the function of all the
genes in cultured cells. Cultured cells from the fruit fly Drosophila
melanogaster can take up RNAi to inhibit the expression of specific genes. In
this example, researchers wanted to identify genes that affected the mitotic
spindle. 14,425 different double-stranded RNAs are arrayed in 96 well plates
0 .Drosophila cells are added to each well, the target gene is suppressed,
and the cells are stained for microtubules and a component of the spindle
pole. The cells are then examined in an automated microscope fl.The
images are then analyzed by a computer, and a gallery of images assem-
bled 10. A researcher then examines the images for those that have a
consistent effect on spindle organization. (b) Various examples of aberrant
spindles were recovered in this screen and stained for microtubules (green),
a spindle pole component (red), and DNA (blue). [Parts (a) step 4 and part (b)
from G. Goshima et al., 2007, Science 316:417.]

·. provided the possibility to explore the function of each gene been developed to explore the consequence of using siRNA to
using RNAi to suppress expression of each individual gene. In suppress each of these in cultured fruit fly cells. With such a
fact, this nematode was the first animal in which a genomic large number to be tested, automatic high-throughput ap-
RNAi screen was anempted. C. elegans can live by eating the proaches were developed (Figure 9-40a). For example, about
bacterium E. coli. Remarkably, if the bacterium expresses a 150 96 well plates are made, with each well containing one
double-stranded RNA homologous to a nematode gene, when double-stranded RNA for a specific gene. Cells are added, and
the bacteria are eaten, they are broken open and the RNA is the double-stranded RNA is taken up and processed by Dicer
absorbed through the intesrine, then processed by Dicer to sup- into siRNAs, which then suppress expression of the target gene.
press expression of the target gene. Since there are about The cells can then be examined for a specific phenotype. In the
20,000 different genes in the nematode, this many different E. example shown in Figure 9-40b, the investigators explored the
coli strains were generated, each expressing double-stranded effect of gene suppression on cells arrested in mitosis. Since this
RNA targeted to a specific nematode gene. In a typical experi- is a morphological screen, they stained cells with appropriate
ment to suppress expression of a single gene (Figure 9-39a), markers and used a robotic microscope to take pictures and a
nematodes are grown on the bacteria expressing the specific computer program to analyze them. In this way they identified
double-stranded RNA, and this suppresses expression of the about 150 new genes whose products contribute to mitosis and
target gene in their embryos. After the adult nematodes have are therefore excellent subjects for further in-depth studies.
laid eggs, they are removed and the effect on the growing em- Unlike in the nematode described above, it is not possible
bryos is examined (Figure 9-39b). to suppress expression of genes by feeding fly larvae double-
stranded RNA. However, it is possible to use RNAi for tissue-
Genomic Screens Using siRNA in Fruit Flies The fruit fly has specific suppression . This is achieved by making a fly in which
about 14,000 protein-encoding genes, and techniques have a specific hairpin RNA is expressed behind an upstream-

9.5 Perturbing Specific Cell Funct1ons 435

activating sequence (UAS) for the DNA-binding protein Gal4. effect of suppression of each gene in the genome can be tested
In the absence of Gal4, the hairpin is not expressed, so no in each tissue. This type of approach is currently ongoing, and
suppression occurs. However, if Gal4 is expressed behind a exciting results are expected in the near future.
tissue-specific promoter, it will be expressed in that tissue, bind
to the UAS, and drive expression of the hairpin RNA, which
will be processed by Dicer into siRNAs and suppress the spe-
cific gene (Figure 9-41a). The system has been set up in this
way so that it can be done on a genomic scale: 14,000 different KEY CONCEPTS of Section 9.5
fly lines have been made, each with the UAS to regulate expres-
Perturbing Specific Cell Functions
sion of a hairpin specific for a target gene. When each of these
flies is mated to a fly carrying the tissue-specific Gal4 gene, the • Genetic techniques have been critical in the analysis of
complex cell biological pathways. Genetic approaches are
now being extended and complemented by chemical screens
and the use of RNAi technology.
Large chemical libraries can be screened for compounds
(a) Parent A Parent B
that target specific processes to study those processes and to
Tissue-specific identify new components.
promoter
• Treatment of cultured cells with appropriate siRNAs leads
to the destruction of target mRNAs and hence "knock-
down" of the encoded protein.
With the availability of annotated genomes, RNAi screens
Progeny can be used to explore the effect of suppressing expression of
each individual gene in an organism. This has been achieved
in the nematode worm and in cultured cells from the fruit fly.
• RNAi can be used to suppress expression of specific genes
l RNAi pathway
in a tissue-specific manner; this technique is currently being
applied on a genomic scale in the fruit fly.
Tissue-specific
gene knockdown

Perspectives for the Future

This chapter has introduced many aspects of technology cur-
rently used by cell biologists. Science is driven by the tech-
nology available, and with each development we can peer
more deeply into the mysteries of life.
The ability to grow cells in culture was a tremendous ad-
vance in technology-it allowed researchers to examine and
explore the inner workings of cells. Techniques in cell culture
are still developing; for example, they are currently contribut-
ing to the exciting developments in stem-cell research (see
Chapter 21). Although most studies have used flat dishes to
FIGURE 9-41 RNAi can be used to suppress genes in a tissue- grow these cells, in the body they form a three-dimensional
specific manner in the fruit fly. Genetic approaches have been used structure. Major areas of research are now examining the
to develop methods to suppress target gene expression in specific functions of cells in three-dimensional environments and gen-
tissues. This involves making use of two large sets of flies. In the first set
erating three-dimensional cell organizations, such as epithe-
of about 14,000 different flies each designed to target one gene (set A),
lial tubes, in supported culture systems.
flies are generated in which a specific hairpin RNA for the target gene is
The discovery and use of GFP and other fluorescent pro-
under the control of an upstream activating sequence (UAS) for the
teinc; h:1c; revolutionized cell biology. By tagging proteins
transcriptional activator Gal4. In each of the second set of flies (set B),
expression of GAL4 is controlled by a tissue-specific promoter. When a
with GFP and following their localization in live cells, it has
fly from set A is mated with a fly from set B, Gal4 will be expressed in become apparent that the cytoplasm of cells is far more dy-
one tissue and therefore the RNAi hairpin will be expressed and the namic than previously envisaged. Every year brings new tech-
gene silenced in this one tissue. (b) A wild-type fly eye (left), and one in nologies associated with fluorescent proteins; approaches
which the white gene has been specifically suppressed in the eye (right). such as FRAP, FRET, and TIRF have become widespread
[Part (b) from N. Perrimon eta!., 2010, Cold Spring Harbor Perspect. Bioi. 2:a003640.] tools to explore the dynamics and molecular mechanisms of

436 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 proteins, either in vivo or in vitro. At the time of this writing, deconvolution microscopy indirect immunofluores-
.· we are seeing a revolution in super-resolution microscopy, 412 cence microscopy 410
opening up the ability to localize molecules by light micros- differential centrifugation lysosome 425
copy several times more accurately than was believed to be 427 membrane transport
possible. Super-resolution microscopy can currently be done differential-interference- protein 424
only on fixed samples; optimists believe it will soon be pos- contrast (DIC) micros- metal shadowing 420
sible to achieve this level of resolution in living cells and copy 427
thereby open up the possibility of watching dynamic pro- mitochondria 427
endoplasmic reticulum monoclonal antibody 403
cesses at the molecular level. As these techniques develop,
(ER) 425
fewer people need to use electron microscopy, and so the organelles 398
expertise in this important area is waning. endosome 425
peroxisome 427
RNAi has provided an awesome and unanticipated new equilibrium density-gradient
phase-contrast
technology to the arsenal of techniques available to cell and centrifugation 427
microscopy 407
developmental biologists. The ability to perform genome-wide fluorescence-activated cell
photo-activated localization
screens in both the nematode worm and fruit fly has made sorter (FACS) 401
microscopy (PALM) 418
traditionally excellent genetic systems even more powerful. fluorescence recovery after
Coupling these technologies with visual screens opens up yet polyclonal antibody 403
photobleaching
another dimension. Consider the following problem: which (FRAP) 416 resolution 405
genes in the nematode affect the organization of a small subset fluorescent staining 408 scanning electron
of neurons? A few years ago, this would have been a techni- microscope (SEM) 419
Forster resonance energy
cally challenging problem. Now it is possible to make a nema- total internal reflection
transfer (FRET) 416
tode in which just those neurons are marked with GFP and fluorescence (TIRF)
then subject them to a visual genomics RNAi screen to see Golgi complex 425
microscopy 415
which gene products are necessary for the normal morphology hybridoma 403
transmission electron
of those neurons. More and more imaginative approaches are immunofluorescence microscope (TEM) 419
being developed combining RNAi with visual and functional microscopy 398
screens in both nematode worms and the fruit fly, permitting
an ever-deepening understanding of life processes. In addition,
efforts are currently under way to explore the effects of knock-
ing down expression of each gene in cultured cell lines and Review the Concepts
then examining the effects on specific pathways. This effort is 1. Both light and electron microscopy are commonly used to
in its infancy, so future refinements and analysis will provide a visualize cells, cell structures, and the location of specific
systems biology view of cell organization and function. molecules. Explain why a scientist may choose one or the
Can RNAi technology be used in medicine? Could it be other microscopy technique for use in research.
used to suppress expression of oncogenes in the treatment of
2. The magnification possible with any type of microscope
cancer? The technological delivery problems are significant
is an important property, but its resolution, the ability to
since the siRNA needs to be delivered to the right cells, re-
distinguish between two very closely apposed objects, is even
main stable in the patient, and be effective at knocking down
more critical. Describe why the resolving power of a micro-
the appropriate protein. Currently, at least a dozen clinical
scope is more important for seeing finer details than its mag-
trials are testing the feasibility of this approach. If the techni-
nification. What is the formula to describe the resolution of
cal hurdles can be overcome, RNAi might become a major
a microscope lens and what are the limitations placed on the
class of therapeutic agent.
values in the formula?
What new technologies will the next decade bring? In the
last decade, RNAi and GFP have revolutionized cell biology. 3. Why are chemical stains required for visualizing cells and
No doubt the next decade will bring about unexpected new tissues with the basic light microscope? What advantage do
developments, so we should be excited about what is to come. fluorescent dyes and fluorescence microscopy provide in
comparison to the chemical dyes used to stain specimens for
light microscopy? What advantages do confocal scanning
microscopy and deconvolution microscopy provide in com-
Key Terms parison to conventional fluorescence microscopy?
4. In certain electron microscopy methods, the specimen is
bright-field light clone 398
not d1rectly imaged. How do these methods provide infor-
microscopy 404 confoca I microscopy 413 mation about cellular structure, and what types of structures
cell line 400 cryoelectron microscopy do they visualize? What limitation applies to most forms of
cell strain 399 422 electron microscopy?
chimeric proteins 398 culturing 397 5. What is the difference between a cell strain, a cell line,
chloroplast 427 cytoplasm 424 and a clone?

Review the Concepts 437

6. Explain why the process of cell fusion is necessary to pro· radation of glucose or fatty acids; ribosomal RNA forms part
duce monoclonal antibodies used for research. of the protein-synthesizing ribosomes; catalase catalyzes de-
7. Much of what we know about cellular function depends composition of hydrogen peroxide; acid phosphatase hydro-
on experiments utilizing specific cells and specific parts (e.g., lyzes monophosphoric esters at acid pH; cytidylyl transferase
organelles) of cells. What techniques do scientists commonly is involved in phospholipid biosynthesis; and amino acid per-
use to isolate cells and organelles from complex mixtures, mease aids in transport of amino acids across membranes.
and how do these techniques work? a. Name the marker molecule and give the number of
8. Hoechst 33258 is a chemical dye that binds specifically to the fraction that is most enriched for each of the following
DNA in live cells, and when excited by UV light, it fluoresces cell components: lysosomes; peroxisomes; mitochondria;
·.
m the visible spectrum. Name and describe one specific plasma membrane; rough endoplasmic reticulum; smooth
method, employing Hoechst 33258, an investigator would endoplasmic reticulum.
usc to isolate fibroblasts in the G 2 phase of the cell cycle b. Is the rough endoplasmic reticulum more or less dense
from those fibroblasts in interphase. than the smooth endoplasmic reticulum? Why?
9. shR~As and siRNAs can be used to successfully knock c. Describe an alternative approach by which you could
down the expression of any specific protein in a given cell identify which fraction was enriched for which organelle.
line or organism. The utility of one over the other is debar- d. How would addition of a de~ergent to the homogenate,
able, but there are merits to using one for therapeutic appli- which disrupts membranes by solubilizing their lipid and pro-
cations in a living organism. Which of the two methods is tein components, affect the equilibrium density·gradient
likely to be more advantageous in the long-term and what is results?
one of its limitations? 2. The nematode worm C. elegans is amenable as a model for
siRNA studies. In this experiment, adult nematodes are fed
Analyze the Data bacteria expressing a double-stranded RNA to suppress the
expression of the unc18 gene, whose mammalian homolog en-
1. Mouse liver cells were homogenized and the homogenate codes a protein that participates in the integration of GLUT4
subjected to equilibrium density-gradient centrifugation with storage vesicles into the plasma membrane. Before evaluating
sucrose gradients. Fractions obtained from these gradients the effects on the organism itself, it is necessary to use RT-PCR
were assayed for marker molecules (i.e., molecules that are to analyze if the siRNA experiment worked. The samples are
limited to specific organelles). The results of these assays are from embryos from five adults fed the bacteria, and the results
shown in the figure. The marker molecules have the following are those following the RT-PCR using primers to amplify
functions: cytochrome oxidase is an enzyme involved in the mRNA from the two different genes, unc18 and GLUT4.
process by which ATP is formed in the complete aerobic deg-
Worms
A B C DE F B c D E
100

80

--
~ 60
E
·x<tJ

-E
0
#.
40 -- - - -
20

10
Fraction number
20
__j
-- - - - - -
50%..,...__ _ _ _ _ _ _ _ _ _ _ _ Sucrose 0%
a. Of the embryos examined, which samples show that
Curve A = cytochrome oxidase CurveD= acid phosphatase the adults successfully took up bacteria containing the
Curve B = ribosomal RNA Curve E = cytidylyl transferase double-stranded unc18 RNA? What made you come to
Curve C =catalase Curve F =amino acid permease those conclusions?

438 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 b. Label which set of bands is the result of the amplifica- Egner, A., and S. Hell. 2005. Fluorescence microscopy with
tion by the unc18 primers and which by the GLUT4 primers. super-resolved optical sections. Trends Cell Bioi. 15:20"'-215.
Why was RT-PCR with GLUT4 primers used as the positive Gaictta, G., et al. 2002. ~lulticolor and electron microscopic
imaging of connexin trafficking. SCience 296:503-507.
control? What does this tell you about the relationship be-
Giepmans, B. N. G., et al. 2006. The fluorescent toolbox for
tween the knock down of the unc18 and GLUT4 mRNAs?
assessing protein location and function. Science 312:217-224.
c. With antibodies against unc18, draw a representative Gilroy, S. 1997. Fluorescence m1croscopy of hvmg plant cells.
Western blot showing the expected results from protein sam- Ann. Rev. Plant Phys10l. Plant Mol. Bioi. 48:165-190.
ples from each of these five samples. Huang, B., H. Babcock, and X. Zhuang. 2010. Breakmg the
diffraction barner: supcr-n:~uluriun imaging of cells. Ce//143:
d. To further invt:~rigate the relationship between the
1047-1058.
unc18 and GLUT4 proteins, the siRNA experiment is re-
Inoue, S., and K. Spnng. 1997. V1deo MICroscopy, 2d ed.
peated, but in embryonic cells expressing a GFP-tagged GLUT4 Plenum Press.
protein. Draw a cell including its mitochondria, nucleus, and Lippincott-Schwartz, J. 20 l 0. Imaging: visualizmg the possibili-
rough endoplasmic reticulum and show where, if using a ties.}. Cell SCience 123:3619-3620.
laser scanning confocal microscope, the GFP fluorescence Lippmcott-Schwartz, J. 2011. Emerging m vivo analyses of cell
would localize in the cells expressing the unc 18 siRNAs. Do function using fluorescence imaging. Ann. Rev. Biochem. 80:327-3 32.
not forget to draw the cell representing the control. Matsumoto, B., ed. 2002. Methods in Cell Bwlogy. Vol. 70: Cell
Biological Applications of Confocal Microscopy. Academic Press.
e. The ability of a protein to colocalize with another pro-
~1ayor, S., and S. Bllgram1. 2007. Frettmg about FRET in cell
tein suggests but does not prove that the two physically inter- and structural biology. In Evaluating Tcchmques in Biochemical
act with each other. Having each protein labeled, in this case Research, D. Zuk, ed. Cell Press.
unc18 to green fluorescent protein and GLUT4 to red fluores- Misteli, T., and D. L. Spector. 1997. Applications of the green
cent protein, provides the experimenter with reagents to tease fluorescent protein in cell biology and biotechnology. Nature
apart the question of colocalizatton versus interaction. Using B10tech. 15:961-964.
these reagents, describe one technique that would simply dem- Pepperkok, R., and Ellenberg, J. 2006. High-throughput fluores-
cence microscopy for systems bwlogy. Nature Rev. Mol. Cell B1ol.
onstrate uncl8 colocalizes with GLUT4 and another tech-
AOP, published online July 19, 2006 (doi: I 0.103?fnrm1979).
nique that proves the two proteins interact physically.
Roukos, V., T. Misteli, and C. K. Schmidt. 20 I 0. Descriptive no
more: the dawn of high-throughput microscopy. Trends m Cell
Biology 20:503-506.
References Saku, Y., S. Minoguchi, and T. Yanag1da. 2000. Single-molecule
Growing Cells in Culture Imaging of EGFR signalling on the surface of living cells. Nature
Bissell, M. J., A. Rizki, and I. S. Mian. 2003. Tissue architec- Cell Bioi. 2:168-172.
ture: the ultimate regulator of breast epithelial function. Curr. Opm. Simon, S., and J. ja1swal. 2004. Potentials and pitfalls of fluorescent
Cell Bioi. 15:753-762. quantum dots for biological imaging. Trends Cell Bioi. 14:497-504.
Barrye, F. L., and K. Shorrman. 1991. Flow cytometry and Sluder, G., and D. Wolf, eds. 1998. Methods m Cell B10/ogy.
cell-separation procedures. Curr. Opm. bnmunol. 3:238-241. Vol. 56: Video Microscopy. Academic Press.
Davis, J. M., ed . 1994. Basic Cell Culture: A Practical So, P. T. C., et al. 2000. Two-photon excitation fluorescence
Approach. IRL Pn;ss. microscopy. Ann. Rev. Biomed. Eng. 2:399-429.
Edwards, B., et al. 2004. Flow cyrometry for high-throughput, Tsien, R. Y. 2009. Indicators based on fluorescence resonance
high-content screening. Curr. Opm. Chem. Bioi. 8:392-398. energy transfer (FRET). Cold Sprmg Harbor Protoc., dod0.1101/
Goding, J. W. 1996. Monoclonal Antibod1es: Principles and pdb.top57.
Practice. Production and Application of Monoclonal Antibodies in Willig, K. !., et al. 2006. STED microscopy reveals that
Cell Biology, B10chenustry, and Immunology, 3d ed. Academic Press. synaptotagmin remains clustered after synaptic vesicle exocytosis.
Griffith, L. G., and M. A. Swartz. 2006. Capturing complex 3D Nature 440:935-939.
tissue phys10logy in vitro. Nature Rev. Mol. Cell. B1ol. 7:211-224.
Krutzik, P., et al. 2004. Analysis of protein phosphorylation and Electron Microscopy: High-Resolution Imaging
cellular signaling events by flow cytometry: techniques and clinical Beck, ~L, et al. 2004. Nuclear pore complex structure and dynamics
applications. Clin. lmmunol. 110:206- 221. revealed by cryoelectron tomography. Science 306:1387-1390.
Paszek, M. ]., and V. M. Weaver. 2004. The tension mounts: Frey, T. G., G. A. Perkins, and M. H. Ellisman. 2006. Electron
mechanics meets morphogeneSIS and malignancy.). Mammary
tomography of membrane-bound cellular organelles. Ann. Rev.
Gland B10l. Neoplasia 9:325-342.
Biophy. Biomol. Stnlc. 35:199-224.
Shaw, A. ]., ed. 1996. Epithelial Cell Culture. IRL Press. Hyatt, M.. A. Pnnciples and Techniques of Electron MICroscopy,
Tyson, C. A., and J. A. Frazier, eds. 1993. Methods in Tox~eol­
4th ed. 2000. Cambridge University Press.
ogy. Vol. 1 (Part A): In Vitro Biological Systems. Academic Press. Koster, A., and J. Klumperman. 2003. Electron microscopy 111
Describes methods for growing many types of primary cells 111 culture.
cell biology: integrating structure and function. Nature Rev. Mol.
Cell Bioi. 4:SS6-SS10.
Light Microscopy: Exploring Cell Structure and Visualizing Lucie, V., et al. 2005. Structural stud1es by electron tomogra-
Proteins within Cells phy: from cells to molecules. Ann. Rev. Biochem. 74:833-865.
Chen, X., M. Velliste, and R. F. Murphy. 2006. Automated Medalia, 0., et al. 2002. Macromolecular architecture m
interpretation of subcellular patterns in fluorescence microscope eukaryotic cells visualized by cryoelectron tomography. SCience
images for location proteomics. Cytometry (Part A) 69A:631-640. 298:1209-1213.

References 439
 Nickell, S., et a!. 2006. A visual approach to proteomics. Nature Ormerod, ,\1. G., ed. 1990. Flow Cytometry: A Practrcal
Reu. Mol. Cell Brol. 7:225-230. Approach. IRL Press.
Rickwood, D. 1992. Preparative Centrifugation: A Practical
Isolation and Characterization of Cell Organelles Approach. IRL Press.
Bamwn, D. 1981. The discovery of lysosomes.). Cell Brol. Wanders, R., and H. R. Waterham. 2006. Biochemistry of
91:66s-76s. mammalian peroxisomes revisited. Ann. Rev. Biochem. 75:295-332.
Cuervo, A. ~1., and J. F. Dice. 1998. Lysosomes: a meermg
point of protems, chaperones, and proteases. }. Mol. Med. 76:6-12. Perturbing Specific Cell Functions
de Duve, C. 1996. The peroxisome in retrospect. Ann. NY Eggert, U.S., and T.]. Mitchison. 2001'>. '\mall molecule
Acad. Scr. 804:1-10. screening by imaging. Curr. Opin. Chern. Bioi. 10:232-237.
de Duve, C. 1975. Explonng cells With a cenmfuge. Science Elbashir, S. .M., et al. 2001. Duplexes of 21-nucleotide RNAs
189:186-194. The Nobel Prize lecture of a pioneer in rhe study of mediate RNA interference in cultured mammalian cells. Nature
cellular organelles. 411:494-498.
de Duve, C., and H. Beaufay. 1981. A short hiswry of tissue Fire, A., et al. 1998. Potent and specific generic interference by
fractionation.}. Cell Brol. 91:293s-299s. double-stranded RNA in Caenorhabditrs elegans. Nature 391:806-81 I.
Foster, L J., et al. 2006. A mammalian organelle map by Goshima, G., er al. 2007. Genes required for mitotic spindle
protein correlation profiling. Cel/125:187-199. assembly in Drosophila S2 cells. Science 316:417-421.
Holtzman, E. 1989. Lysosomes. Plenum Press. Kamath, R. S., et a!. 2003. Systematic functional analysis of the
Howell, K. E., E. Devaney, and J. Gruenberg. 1989. Subcellular Caenorhabdrtrs elegans genome using RNAi. Nature 421:231-237.
fractionation of tissue culture cells. Trends Biochem. Sci. 14:44-48. Mayer, T. U., eta!. 1999. Small molecule inhibitor of mitotic
Lamond, A., and W. Earnshaw. 1998. Structure and function in spindle hipolamy Identified m a phenotype-based screen. Scrence
the nucleus. Science 280:547-553. 286:971-974.
Mootha , V. K., er al. 2003. Integrated analysis of protein Meister, G., and T. Tuschl2004. Mechanisms of gene silencing
composition, tissue diversity, and gene regulation in mouse by douhle-stranded RNA. Nature 431:343-349.
mirochondna. Cel/115:629-640. Mohr, S., et al. 2010. Genomic screening wah RNAi: result and
Palade, G. 1975. Intracellular aspects of the process of protein challenges. Ann. Rev. Biochem. 79:37-64.
synthesis. Science 189:347-358. The Nobel Prize lecrure of a Perrimon, N., et al. 2010. In vivo RNAi: today and tomorrow.
pioneer in the study of cellular organelles. Cold Spring Harbor Perspect. Bioi. 2:a003640.

' '
440 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells
 CLASSIC EXPERIMENT 9.1

Separating Organelles
·. H. Beaufay et al., 1964, Biochemical Jouma/92:191

n the 1950s and 1960s, scientists used a subcellular population will contain subsequent enzyme analysis. Once the
I two techniques to study cell organ-
elles: microscopy and fractionation.
tl11:: ~arne enzymes and that each en-
zyme is located at a discrete site within
fractionation was complete, de Duve
performed enzyme assays to determine
Christian de Duve was at the forefront the cell. Armed with these hypotheses the subcellular distribution of each en-
of cell fractionation. In the early 1950s, and the powerful tool of centrifuga- zyme. He then graphically plotted the
he used centrifugation to distinguish a tion, de Duve further subdivided the distribution of the enzyme throughout
new organelle, the lysosome, from pre- mitochondrial-rich fraction. First, he the cell. As had been shown previously,
viously characterized fractions: the nu- identified the light mitochondrial frac- the activity of cytochrome oxidase, an
cleus, the mitochondrial-rich fraction, tion, which is made up of hydrolytic important enzyme in the electron-
and the microsomes. Soon thereafter, enzymes that are now known to com- transfer system, was found primarily in
he used equilibrium-density centrifuga- pose the lysosome. Then, in a series of the heavy mitochondrial fractions. The
tion to uncover yet another organelle. experiments described here, he identi- microsomal fraction was shown to
fied another discrete subcellular frac- contain another previously character-
tion, which he called the peroxisome, ized enzyme, glucose-6-phosphatase.
Background
within the mitochondrial-rich fraction. The light mitochondrial fraction, which
Eukaryotic cells are highly organized is made up of the lysosome, showed the
and composed of cell structures known characteristic acid phosphatase activ-
The Experiment
as organelles that perform specific ity. Unexpectedly, de Duve observed a
functions. Although microscopy has De Duve studied the distribution of en- fourth pattern when he assayed uricase
allowed biologists to describe the loca- zymes in rat liver cells. Highly active in activity. Rather than following the pat-
tion and appearance of various organ- energy metabolism, the liver contains a tern of the reference enzymes, uricase
elles, it is of limited use in uncovering number of useful enzymes to study. To activity was sharply concentrated
an organelle's function. To do this, cell look for the presence of various en- within the light mitochondrial fraction.
biologists have relied on a technique zymes during the fractionation, de Duve This sharp concentration, in contrast to
known as cell fractionation. Here cells relied on known tests, called enzyme as- the broad distribution, suggested to de
are broken open and the cellular com- says, for enzyme activity. To retain Duve that the uricase might be secluded
ponents are separated on the basis of maximum enzyme activity, he had to in another subcellular population sepa-
size, mass, and density using a variety take precautions, which included per- rate from the lysosomal enzymes.
of centrifugation techniques. Scientists forming all fractionation steps at 0 oc To test this theory, de Duve em-
could then isolate and analyze cell to reduce protease activity. ployed a technique known as equilib-
components of different densities, De Duve used rate-zonal centrifu- rium density-gradient centrifugation,
called fractions. Using this method, bi- gation to separate cellular components which separates macromolecules on the
ologists had divided the cell into four by successive centrifugation steps. He basis of density. Equilibrium density-
fractions: nuclei, mitochondrial-rich removed the rat's liver and broke it gradient centrifugation can be per-
fraction, microsomes, and cell sap. apart by homogenization. The crude formed using a number of different
De Duve was a biocHemist inter- preparation of homogenized cells was gradients, including sucrose and glyco-
ested in the subcellular locations of then subjected to relatively low-speed gen. In addition, the gradient can be
metabolic enzymes. He had already centrifugation. This initial step sepa- made up in either water or "heavy
completed a large body of work on the rated the cell nucleus, which collects as water," which contains the hydrogen
fractionation of liver cells, in which he sediment at the bottom of the tube, isotope deuterium in place of hydro-
had determined the subcellular loca- from the cytoplasmic extract, which re- gen. In his experiment, de Duve sepa-
tion of numerous enzymes. By locating mains in the supernatant. Next, de rated the mitochondrial-rich fraction
these enzymes in specific cell fractions, Duve further subdivided the cytoplas- prepared by rate-zonal centrifugation
he could begin to elucidate the func- mic extract into heavy mitochondrial in each of the~c: different gradients (see
tion of the organelle. He noted that his fraction, light mitochondrial fraction, Figure 9-35). If uricase were part of a
work was guided by two hypotheses: and microsomal fraction. He accom- separate subcellular compartment, it
the "postulate of biochemical homoge- plished separating the cytoplasm by would separate from the lysosomal en-
neity" and "the postulate of single lo- employing successive centrifugation zymes in each gradient tested. De Duve
cation." In short, these hypotheses steps of increasing force. At each step performed the fractionations in this
propose that the entire composition of he collected and stored the fractions for series of gradients, then performed

Separating Organelles 441

FIGURE 1 Graphical representation of the enzyme analysis of 5
products from a sucrose gradient. The mitochondrial-rich fraction
was separated as depicted in Figure 9-35, and then enzyme assays
were performed. The relative concentration of active enzyme is plotted
3
on they axis; the height in the tube is plotted on the x axis. The peak Cytochrome oxidase
activities of cytochrome oxidase (top) and acid phosphatase (bottom) 2
are observed near the top of the tube. The peak activity of uricase
(middle) migrates to the bottom of the tube.

80
c 5
0
·;:;
~
4
cQl
() 3
c Uricase
0
()

Ql
2
.:::
~
a;
a:
20 40 60 80
5

4
3
Acid phosphatase
2

Percent height in tube

enzyme assays as before. In each case, hydrogen peroxide, de Duve proposed the lysosome and the peroxisome. His
he found uricase in a separate popula- that this fraction represented an organ- work also provided important clues to
tion than the lysosomal enzyme acid elle responsible for the peroxide me- the organelles' function. The lysosome,
phosphatase and the mitochondrial en- tabolism and dubbed it the peroxisome. where de Duve found so many poten-
zyme cytochrome oxidase (Figure 1). tially destructive enzymes, is now
By repeatedly observing uricase activ- known to be an important site for deg-
ity in a distinct fraction from the activ-
Discussion radation of biomolecules. The peroxi-
ity of the lysosomal and mitochondrial De Duve's work on cellular fraction- some has been shown to be the site of
enzymes, de Duve concluded that uri- ation provided an insight into the func- fatty acid and amino acid oxidation,
case was part of a separate organelle. tion of cell structures as he sought to reactions that produce a large amount
The experiment also showed that two map the location of known enzymes. of hydrogen peroxide. In 1974, de
other enzymes, catalase and o-amino Examining the inventory of enzymes in Duve received the Nobel Prize for
acid oxidase, segregated into the same a given cell fraction gave him clues to Physiology or Medicine in recognition
fractions as uricase. Because each of its function. His careful work resulted of his pioneering work.
these enzymes either produced or used in the uncovering of two organelles:

442 CHAPTER 9 • Culturing, Visualizing, and Perturbing Cells

 CHAPTER

Biomembrane Structure

Molecular model of a lipid bilayer with embedded membrane proteins.

Integral membrane proteins have distinct exoplasmic, cytosolic, and
membrane-spanning domains. Shown here are portions of the insulin
receptor, which regulates cell metabolism. [Ramon Andrade 3Dciencia/
Science Photo Library.]

embranes participate in many aspects of cell structure single plasma membrane contains hundreds of different types

M and function. The plasma membrane defines the cell

and separates the inside from the outside. In eukary-
otes, membranes also define the intracellular organelles such as
of proteins that are integral to the function of the cell. Some
of these proteins catalyze ATP synthesis and initiation of DNA
replication, for instance. Others include the many types of
the nucleus, mitochondrion, and lysosome. These biomem- membrane transport proteins that enable specific ions, sugars,
branes all have the same basic architecture-a phospholipid amino acids, and vitamins to cross the otherwise impermeable
bilayer in which proteins are embedded (Figure 10-1). By pre- phospholipid bilayer to enter the cell and that allow specific
venting the unassisted movement of most water-soluble sub- metabolic products to exit. Receptors in the plasma membrane
stances from one side of the membrane to the other, the are proteins that allow the cell to recognize chemical signals
phospholipid bilayer serves as a permeability barrier, helping present in its environment and adjust its metabolism or pattern
to maintain the 'characteristic differences between the inside of gene expression in response.
and outside of the cell or organelle; in turn, the embedded pro- Eukaryotes also have a plasma membrane studded with a
teins endow the membrane with specific functions, such as multitude of proteins that perform a variety of functions, in-
regulated transport of substances from one side to the other. cluding membrane transport, cell signaling, and connecting
Each cellular membrane has irs own set of proteins that allow cells into tissues. In addition, eukaryotic cells-which are
it to carry out a multitude of different functions. generally much larger than prokaryotes-also have a variety
Prokaryotes, the simplest and smallest cells, are about of internal membrane-bound organelles (see Figure 9-32).
1-2 j.J..m in length and are surrounded by a single plasma Each organelle membrane has a unique complement of pro-
membrane; in most cases they contain no internal membrane- teins that enable it to carry out its characteristic cellular func-
limited subcompartments (see Figure 1-11 ). However, this tions, such as ATP generation (in mitochondria) and DNA

OUTLINE

10.1 The lipid Bilayer: Composit ion 10.3 Phospholipids, Sphingolipids, and Cholesterol:
and Structural Organization 445 Synthesis and Int racellular Movement 464

10.2 Membrane Proteins: Struct ure

and Basic Functions 455
 Hydrophilic
phospholipid
head group

Phospholipid
bilayer

Hydrophobic
fatty acyl
side chains

FIGURE 10-1 Fluid mosaic model of biomembranes. A bilayer of substances from one side to the other.lntegral (transmembrane)
phospholipids -3 nm thick provides the basic architecture of all cellular proteins span the bilayer and often form dimers and higher-order
membranes; membrane proteins give each cellular membrane its oligomers. Lipid-anchored proteins are tethered to one leaflet by a
unique set of functions. Individual phospholipids can move laterally covalently attached hydrocarbon chain. Peripheral proteins associate
and spin within the plane of the membrane, giving the membrane a with the membrane primarily by specific noncovaler:lt interactions
fluidlike consistency similar to that of olive oil. Noncovalent interactions wit h integral p roteins or membrane lipids. Proteins in t he plasma
between phospholipids, and between phospholipids and proteins, membrane also make extensive contact with the cytoskeleton.
lend strength and resilience to the membrane, while the hydrophobic [After D. Engelman, 2005, Nature 438:578-580.]
core of the bilayer prevents the unassisted movement of water-soluble

(a)
synthesis (in the nucleus). Many plasma membrane proteins
also bind components of the cytoskeleton, a dense networ k
of protein filaments that crisscrosses the cytosol to provide
mechanical support for cellular membranes, interactions that
are essential for the cell to assume its specific shape and for
many types of cell movements.
Despite playing a structural role in cells, membranes are
nor rigid structures. They can bend and flex in three dimen-
sions while still maintaining their integrity, due in part to
abundant noncovalent interactions that hold lipids and pro-
teins together. Moreover, within the plane of the membrane,
there is considerable mobility of individual lipids and proteins.
According to the fluid mosaic model of biomembranes, first
proposed by researchers in the 1970s, the lipid bilayer behaves (b)
in some respects like a two-dimensional fluid, with individual
lipid molecules able to move past one another as well as spin
in place. Such fluidity and flexibility not only allows organelles
to assume their typical shapes, but also enables the dynamic

FIGURE 10-2 Eukaryotic cell membranes are dynamic structures.

(a) An electron micrograph of the plasma membrane of an HIV-infected
cell, show1ng HIV particles budding into the culture medium. As the
virus core buds from the cell, it becomes enveloped by a membrane
derived from the cell's plasma membrane that contains specific viral
proteins. (b) Stacked membranes of the Golgi complex with budding
vesicles. Note the irregular shape and curvature of these membranes.
[Part (a) from W. Sundquist and U. von Schwedler, University of Utah;
part (b) from Biology Pies/Photo Researchers, Inc.]

444 CHAPTER 10 • Biomembrane Structure

property of membrane budding and fusion, such as occurs (a) Membrane bilayer
when viruses are released from an infected cell (Figure 10-2a)
and when the internal cellular membranes of the Golgi com-
plex bud into vesicles in the cytosol (Figure 10-2b) and then
fuse with other membranes to transport their contents from
one organelle to another (Chapter 14).
We begin our examination of biomembranes by consid-
ering their lipid components. These not only affect mem-
brane shape and function but also help anchor proteins to
the membrane, modify membrane protein activities, and
transduce signals to the cytoplasm. We then consider the
structure of membrane proteins . Many of these proteins
have large segments that are embedded in the hydrocarbon
core of the phospholipid bilayer, and we will focus on the
principal classes of such membrane proteins. Finally, we
consider how lipids such as phospholipids and cholesterol
are synthesized in cells and distributed to the many mem- (b) Polar head
branes and organelles. Cholesterol is an essential component
of the plasma membrane of all animal cells but is toxic to the
organism if present in excess.

1 0.1 The Lipid Bilayer: Composition and

Structural Organization
(c)
In Chapter 2 we learned that phospholipids are the principal
building blocks of biomembranes. The most common phos-
pholipids in membranes are the phosphoglycerides (see Fig-
ure 2-20), but as we will see in this chapter, there are multiple
types of phospholipid. All phospholipids are amphipathic
molecules-they consist of two segments with very different
chemical properties: a fatty acid-based (fatty acyl) hydrocar-
bon "tail" that is hydrophobic and partitions away from Micelle
water, and a polar "head group" that is strongly hydrophilic,
Liposome
or water loving, and tends to interact with water molecules.
The interactions ·o f phospholipids with each other and with FIGURE 1 0- 3 The bilayer st ructure of biomembranes. (a) Electron
water largely determine the structure of biomembranes. micrograph of a thin section through an erythrocyte membrane
Besides phospholipids, biomembranes contain smaller stained with osmium tetroxide. The characteristic "railroad track"
amounts of other amphipathic lipids, such as glycolipids and appearance of the membrane indicates the presence of two polar
cholesterol, which contribute to membrane function in im- layers, consistent with the bilayer structure of phospholipid mem-
branes. (b) Schematic interpretation ofthe phospholipid bilayer in
portant ways. We first consider the structure and properties
which polar groups face outward to shield the hydrophobic fatty acyl
of pure phospholipid bilar,ers and then discuss the composi-
tails from water. The hydrophobic effect and van der Waals interactions
tion and behavior of natural cell membranes. We will see
between the fatty acyl tails drive the assembly of the bilayer (Chapter 2).
how the precise lipid composition of a given membrane in- (c) Cross-sectional views of two other structures formed by dispersal of
fluences its physical properties. phospholipids in water. A spherical micelle has a hydrophobic interior
composed entirely of fatty acyl chains; a sphericalliposome consists of
a phospholipid bilayer surrounding an aqueous center.
Phospholipids Spontaneously [Part (a) courtesy of J. D. Robertson.]
Form Bilayers
The amphipathic nature of phospholipids, which governs
their interactions, ts cnrical to the structure of biomembranes.
When a suspension of phospholipids is mechanically dispersed a mixture of phospholipids depends on several factors, includ-
in aqueous solution, the phospholipids aggregate into one of ing the length of the fatty acyl chains in the hydrophobic tail,
three forms: spherical micelles and liposomes, or sheetlike their degree of saturation (i.e., the number of C-C and C=C
phospholipid bilayers, which arc two molecules thick (Figure bonds), and temperature. In all three structures, the hydro-
10-3). The type of structure formed by pure phospholipids or phobic effect causes the fatty acyl chains to aggregate and

10.1 The Lipid Bilayer: CompositiOn and St ru ctural Organization 445

exclude water molecules from the "core.'' Micelles are rarely
formed from natural phospholipids, whose fatty acyl chains /Oligosaccharide
generally arc too bulky to fit into the interior of a micelle.
However, micelles arc formed if one of the two fatty acyl
chains that make up the tail of a phospholipid is removed by
hydrolysis, forming a lysophospholipid, as occurs upon treat-
ment with the enzyme phospholipase. In aqueous solution,
common detergents and soaps form micelles that behave like
the balls in tiny ball bearings, thus giving soap solutions thc1r
slippery feel and lubricating properties.
Phospholipids of the composition present in cells sponta-
neously form symmetric phospholipid bilayers. Each phos-
pholipid layer in this lamellar structure is called a leaflet. The
Treat withl
hydrophobic fatty acyl cha ins in each leaflet minimize their organic 1\ Proteins and oligosaccharides
contact with water by aligning themselves tightly together in solvent I~ form insoluble residue
the center of the bilayer, forming a hydrophobic core that is that is removed
about 3-4 nm thick (Figure 10-3b). The close packing of cJ~:\
f'..v '
these nonpolar tails is stabilized by van der Waals interac-
tions between the hydrocarbon chains. Ionic and hydrogen Phospholipids in solution

1
bonds stabilize the interactions of the phospholipid polar Evaporate
head groups with one another and with water. Electron mi- solvent
croscopy of thin membrane sections of cells stained with os- ...==~A"'-==.....,
mium tetroxide, which binds strongly to the polar head Disperse [ Dissolve phospholipids
phospholipids in solvent and apply
groups of phospholipids, reveals the bilayer structure (Figure m water to small hole
10-3a). A cross section of a single membrane stained with in partition
osmium tetroxide looks like a railroad track: two thin dark Planar / Plastic
lines (the stained head group complexes) with a uniform light bilayer partition

\
space of about 2 nm between them (the hydrophobic tails).
A phospholipid bilayer can be of almost un limited size-
from micrometers (f.Lm) to millimeters (mm) in length or
width-and can contain tens of millions of phospholipid
molecules. The phospholipid bilayer is the basic structural
unit of nearly all biological membranes. Its hydrophobic core Water Water
prevents most water-soluble substances from crossing from
one side of the membrane to the other. Although biomem-
branes contain other molecules (e.g., cholesterol, glycolipids, EXPERIMEN . AL FIGUR -4 Formation and study of pure
proteins), it is the phospholipid bilayer that separates two phospholipid bilayers. (Top) A preparation of biological membranes
aqueous solutions and acts as a permeability barrier. The is treated with an organic solvent, such as a mixture of chloroform and
lipid bilayer thus defines cellular compartments and allows a methanol (3:1 ), which selectively solubilizes the phospholipids and
separation of the cell's interior from the outside world. cholesterol. Proteins and carbohydrates remain in an insoluble residue.
The solvent is removed by evaporation. (Bottom left) If the lipids are
mechanically dispersed in water, they spontaneously form a liposome,
shown in cross section, with an internal aqueous compartment.
Phospholipid Bilayers Form a Sealed
(Bottom right) A planar bilayer, also shown in cross section, can form
Compartment Surrounding an Internal over a small hole in a partition separating two aqueous phases; such a
Aqueous Space system can be used to study the physical properties of bilayers, such as
their permeability to solutes.
Phospholipid bilayers can be generated in the laboratory by
simple means, using either chemically pure phospholipids or
lipid mixtures of the composition found in cell membranes
(Figure 10-4 ). Such synthetic bilayers possess three impor- Even though the exterior aqueous environment can vary
tant properties. First, they are virtually impermeable to water widely in ionic strength and pH, Lht: bilayt:r has the strength
soluble (hydrophilic) solutes, which do not readily diffuse to retain its characteristic architecture. Third, all phospho-
across the bilayer. This includes salts, sugars, and most other lipid bilayers can spontaneously form sealed closed compart-
small hydrophilic molecules-including water itself. The sec- ments where the aqueous space on the inside is separated
ond property of the bilayer is its stability. Hydrophobic and from that on the outside. An "edge" of a phospholipid bi-
van der Waals interactions between the fatty acyl chains layer, as depicted in Figure 10-3b, with the hydrocarbon core
maintain the integrity of the interior of the bilayer structure. of the bilayer exposed to an aqueous solution, is unstable; the

446 CHAPTER 10 • Biomembrane Structure

 Mitochondrion FIGURE 10-5 The faces of cellular membranes. The
plasma membrane, a single bilayer membrane, encloses
the cell. In this highly schematic representation, internal
Outer] Mitochondrial cytosol (tan) and external environment (white) define the
Inner membranes cytosolic (red) and exoplasmic (gray) faces of the bilayer.
Vesicles and some organelles have a single membrane
and their internal aqueous space (white) is topologically
~:.-Cf--1--1--- Matrix
lntermembrane space equivalent to the outside of the cell. Three organelles-
the nucleus, mitochondrion, and chloroplast (which is not
Exoplasmic shown)-are enclosed by two membranes separated by a
face
small intermembrane space. The exoplasmic faces of the
Lysosome
I inner and outer membranes around these organelles
border the intermembrane space between them. For
simplicity, the hydrophobic membrane interior rs not
indicated in this diagram.

Cytosolic
face

Plasma membrane

lntermembrane space

exposed fatty acyl side chains would be in an energetically partments, similar in basic architecture to liposomes. Because
much more stable state if they were not adjacent to water all cellular membranes enclose an entire cell or an internal
molecules but surrounded by other fatty acyl chains (hydro- compartment, they have an internal face (the surface ori-
phobic effect; Chapter 2). Thus in aqueous solution, sheets ented toward the interior of the compartment) and an exter-
of phospholipid bilayers spontaneously sea l their edges, nal face (the surface presented to the environment). More
forming a spherical bilayer that encloses an aqueous central commonly, we designate the two surfaces of a cellular mem-
compartment. The liposome depicted in Figure 10-3c is an brane as the cytosolic face and the exoplasmic face. This
example of such a structure viewed in cross section. nomenclature is useful in highlighting the topological equiv-
This physical chemical property of a phospholipid bi- alence of the faces in different membranes, as diagrammed in
layer has important implications for cellular membranes: no figures 10-5 and I 0-6. For example, the exoplasmic face of
membrane in a cell can have an "edge" with exposed hydro- the plasma membrane is directed away from the cytosol, to-
carbon fatty acyl chains. All membranes form closed com- ward the extracellular space or external environment, and
defines the outer limit of the cell. The cytosolic face of the
plasma membrane faces the cytosol. Similarly for organelles
Exoplasmic face Membrane protein and vesicles surrounded by a single membrane, the cytosolic

FIGURE 10-6 Faces of cellular membranes are conserved during

membrane budding and fusion. Red membrane surfaces are
cytosolic faces; gray are exoplasmic faces. During endocytosis a
segment of the plasma membrane buds inward toward the cytosol and
eventually pinches off a separate vesicle. During this process the
cytosolic face of the plasma membrane remains facing the cytosol and
the exoplasmic face of the new vesicle membrane faces the vesicle
Cytosolic lumen. During exocytosis an intracellular vesicle fuses with the plasma
face
membrane, and the lumen of the vesicle (exoplasmic face) connects
with the extracellular medium. Proteins that span the membrane retain
their asymmetric orientation during vesicle budding and fusion; in
particular the same segment always faces the cytosol.

10.1 The Lipid Bilayer: Composition and Structural Organization 447

face faces the cytosol. The exoplasmic face is always directed (a)
away from the cytosol and in this case is on the inside of the
organelle in contact with the internal aq ueous space, or
lumen. The lumen of these vesicle~ is topologically equiva-
lent to the extracellular space, a concept most easily under-
stood for vesicles that arise by invagination (endocytosis) of
the plasma membrane. The external face of the plasma mem-
brane becomes the internal face of the vesicle membrane,
while in the vesicle the cytosolic face of the plasma mem-
brane still faces the cytosol (figure 10-6).
Three organelles-the nucleus, mitochondrion, and
chloroplast-are surrounded not by a single membrane, but
by two. The exoplasmic surface of each membrane faces the
space between the two membranes. This can perhaps best be
understood by reference to the endosymbiont hypothesis, dis-
cussed in Chapter 6, which posits that mitochondria and
chloroplasts arose early in the evolution of eukaryotic cells by
the engulfment of bacteria capable of oxidative phosphoryla-
tion or photosynthesis, respectively (see Figure 6-20).
Natural membranes from different cell types exhibit a
variety of shapes, which complement a cell's function. The
smooth, flexible surface of the erythrocyte plasma membrane
allows the cell to squeeze through narrow blood capillaries
(Figure 10-7a). Some cells have a long, slender extension of the (b)
plasma membrane, called a cilium or flagellum, which beats in
.,
a whiplike manner (Hgure 10-7b). This motion causes fluid to
flow across the surface of a sheet of cells, or a sperm cell to
swim toward an egg. The differing shapes and properties of
biomembranes raise a key question in cell biology, namely
how the composition of biological membranes is regulated to
establish and maintain the identity of the different membrane
structures and membrane-delimited compartments. We return
to this question in Section 10.3 and in Chapter 14.

Biomembranes Contain Three Principal 10 1-lm

Classes of Lipids
FIGURE 10-7 Variation in biomembranes in different cell types.
The term phospholipid is a somewhat generic term, encompass- (a) A smooth, flexible membrane covers the surface of the discoid
ing multiple distinct molecules from multiple classes. It refers to erythrocyte cell as seen in this scanning electron micrograph.
any amphipathic lipid with a phosphate-based head group and (b) Tufts of cilia (Ci) project from the ependymal cells that line the brain
a two-pronged hydrophobic tail. A typical biomembrane actu- ventricles. [Part (a) Copyright 0 Omi Kron/Photo Researchers, Inc. Part (b) from
ally contains three classes of amphipathic lipids: phosphoglyc- R. G. Kessel and R. H. Kardon, 1979, Tissues and Organs: A Text-Atlas of Scanning
erides, sphingolipids, and sterols, which differ in their chemical Electron Microscopy, W. H. Freeman and Company.]
structures, abundance, and functions in the membrane (Figure
10-8). While all phosphoglycerides are phospholipids, only cer-
tain sphingolipids are, and no sterols are. lipids in the plasma membrane, the head group consists of
Phosphoglycerides, the most abundant class of phospholip- choline, a positively charged alcohol, esterified to the nega-
ids in most membranes, are derivatives of glycerol 3-phosphate tively charged phosphate. In other phosphoglycerides, an
(sec Figure 10-Sa). A typical phosphoglyceride molecule con- OH-containing molecule su\:h as ethanolamine, serine, or the
sists of a hydrophobic tail composed of two fatty acid-based sugar derivative inositol is linked to the phosphate group.
(acyl) chains esterified to the two hydroxyl groups in glycerol The negatively charged phosphate group and the positively
phosphate and a polar head group attached to the phosphate charged groups or hydroxyl groups on the head group inter-
group. The two fatty acyl chains may differ in the number of act strongly with water. At neutral pH, some phosphoglycer-
carbons that they contain (commonly 16 or 18) and their ides (e.g., phosphatidylcholine and phosphatidylethanolamine)
degree of saturation (0, 1, or 2 double bonds) . A phospho- carry no net electric charge, whereas others (e.g., phosphati-
glyceride is classified according to the nature of its head dylinositol and phosphatidylserine) carry a single net nega-
group. In phosphatidylcholines, the most abundant phospho- tive charge. Nonetheless, the polar head groups in all these

448 CHAPTER 10 • Biomembrane Structure

 Variable p ortion of
(a) Phosphoglycerides head group
H
I .....-H
N+ PE
o/'-....../ ---...,.H
Glycerol
0 0
' 3 II
o·
"""- 2 /'-....
"'Y' ·o" p1 PC
0

PS

OH OH
HO~OH
O~OH
1 3 PI

0
Plasmalogen , 3 II
~
,... P. X
0 Q I
0 Variable head
0 group

(b) Sphingolipids

0
o~o~oH
HO~OH
OH GlcCer

(c) Sterols

Cholesterol Ergosterol Stigm asterol

(animal) (f ungal) (plant )

FIGURE 10-8 Three classes of membrane lipids. (a) Most to sphingosine by an amide bond . The sphingomyelins (SM), which
phosphoglycerides are derivatives of glycerol 3-phosphate (red), which contain a phosphocholine head group, are phospholipids. Other
contains two esterified fatty acyl chains that constitute the hydropho· sphingolipids are glycolipids in which a single sugar residue or
bic "tail" and a polar •head group" esterified to the phosphate. The branched oligosaccharide is attached to the sphingosine backbone.
fatty acids can vary in length and be saturated (no double bonds) or For instance, the simple glycolipid glucosylcerebroside (GicCer) has a
unsaturated (one, two, or three double bonds). In phosphatidylcholinP glucose head group. (c) The major sterols in animdls (cholesterol), fung1
(PC), the head group is choline. Also shown are the molecules attached (ergosterol), and plants (stigmasterol) differ slightly in structure, but all
to the phosphate group in three other common phosphoglycerides: serve as key components of cellular membranes. The basic structure of
phosphatidylethanolamine (PE), phosphatidylserine (PS), and phospha- sterols is a four-ring hydrocarbon (yellow). Like other membrane lipids,
tidylinositol (PI). Plasma logens contain one fatty acyl chain attached sterols are amphipathic. The single hydroxyl group is equivalent to the
to glycerol by an ester linkage and one attached by an ether linkage; polar head group in other lipids; the conjugated ring and short
these contain similar head groups as other phosphoglycerides. hydrocarbon chain form the hydrophobic tail. [See H. Sprong et al., 200 1,
(b) Sphingolipids are derivatives of sphingosine (red), an amino alcohol Nature Rev. Mol. Cell Bioi. 2:504.]
with a long hydrocarbon chain. Various fatty acyl chains are connected

10.1 The Lipid Bilayer: Compos ition and Structural Organization 449
phospholipids can pack together into the characteristic bi- the cholesterol in most mammalian cells is present in the
layer structure. When phospholipases act on phosphoglycer- plasma membrane and associated vesicles. Cholesterol and
ides, they produce lysophospholipids, which lack one of the other sterols are roo hydrophobic to form a bilayer structure
two acyl chains. Lysophospholipids are not only important on their own. Instead, at concentrations found in natural
signaling molecules, released from cells and recognized by membranes, these sterols must intercalate between phospho-
specific receptors; their presence can also affect the physical lipid molecules to be incorporated into biomembranes. When
properties of the membranes in which they reside. so intercalated, sterols provide structural support to mem-
The plasmalogens are a group of phosphoglycerides that branes, preventing too close a packing of the phospholipids'
wntain one fatty acyl chain attached to carbon 2 of glycerol acyl chains to maintain a s•gmficant measure of membrane
by an ester linkage and one long hydrocarbon chain attached fluidity, and at the same time conferring the necessary rigidity
to carbon 1 of glycerol by an ether (C-0-C) rather than required for mechanical support. Some of these effects can be
an ester linkage. Plasmalogens are particularly abundant in highly local, as in the case of lipid rafts, discussed below.
human brain and heart tissue. The greater chemical stability In addition to its structural role in membranes, choles-
of the ether linkage in plasmalogens, compared to the ester terol is the precursor for several important bioactive mole-
linkage, or the subtle differences in their three-dimensional cules. They include bile acids, which are made in the liver and
structure compared with that of other phosphoglycerides help emulsify dietary fats for digestion and absorption in the
may have as yet unrecognized physiologic significance. intestines; steroid hormones produced by cndocri ne cells
A second class of membrane lipid is the sphingolipids. (e.g., adrenal gland, ovary, testes); and vitamin D produced
All of these compounds are derived from sphingosine, an in the skin and kidneys. Another critical function of choles-
amino alcohol with a long hydrocarbon chain, and contain a terol is its covalent addition to Hedgehog protein, a key sig-
long-chain fatty acid attached in amide linkage to the sphin- naling molecule in embryonic development (Chapter 16).
gosine amino group (see Figure 1 0-8b) . Like phosphoglycer-
ides, some sphingolipids have a phosphate-based polar head
Most Lipids and Many Proteins Are.laterally
group. In sphingomyelin, the most abundant sphingolipid,
phosphocholine is attached to the terminal hydroxyl group Mobile in Biomembranes
of sphingosine (see Figure 10-8b, SM). Thus sphingomyelin In the two-dimensional plane of a bilayer, thermal motion
is a phospholipid, and its overall structure is quite similar to permits lipid molecules to rotate freely around their long
that of phosphatidylcholine. Sphingomyelins are similar in axes and to diffuse laterally within each leaflet. Because such
shape to phosphoglycerides and can form mixed bilayers movements are lateral or rotational, the fatty acyl chains
with them. Other sphingolipids are amphipathic glycolipids
whose polar head groups are sugars that are not linked via a
phosphate group (and so technically are not phospholipids).
Glucosylcerebroside, the simplest glycosphingolipid, con-
tains a single glucose unit attached to sphingosine. In the Heat
complex glycosphingolipids called gangliosides, one or two
branched sugar chains (oligosaccharides) containing sialic
acid groups arc attached to sphingosine. Glycolipids consti-
tute 2-10 percent of the total lipid in plasma membranes;
they are most abundant in nervous tissue. Gel-like consistency Fluid like consistency
Cholesterol and its analogs constitute the third important
class of membrane lipids, the sterols. The basic structure of
sterols is a four-ring isoprenoid-based hydrocarbon. The
structures of the principal yeast sterol (ergosterol) and plant
phytosterols (e.g., stigmasterol) differ slightly from that of
cholesterol, the major animal sterol (see Figure 1 0-Sc). The
small differences in the biosynthetic pathways and structures
of fungal and animal sterols arc the basis of most antifungal
drugs currently in use. Cholesterol, like the two other sterols,
FIGURE 10-9 Gel and fluid forms of the phospholipid bilayer.
has a hydroxyl substituent on one ring. Although cholesterol
(Top) Depiction of gel-to-fluid transition. Phospholipids with long
IS almost entirely hydrocarbon in composition, it is amphipa-
saturated fatty acyl chains tend to assemble into a highly ordered,
thic hecause its hydroxyl group can interact with water. Be- gel-like bilayer in which there is little overlap of the nonpolar tails in
cause it lacks a phosphate-based head group, it is not a the two leaflets. Heat disorders the nonpolar tails and induces a
phospholipid. Cholesterol is especially abundant in the transition from a gel to a fluid within a temperature range of only a few
plasma membranes of mammalian cells but is absent from degrees. As the chains become disordered, the bilayer also decreases in
most prokaryotic and all plant cells. As much as 30-50 per- thickness. (Bottom) Molecular models of phospholipid monolayers in
cent of the lipids in plant plasma membranes consists of cer- gel and fluid states, as determined by molecular dynamics calculations.
tam steroids unique to plants. Berween 50 and 90 percent of [Bottom based on H. Heller et al., 1993, J. Phys. Chem. 97:8343.]

450 CHAPTER 10 • Biomembrane Structure

remain in the hydrophobic interior of the bilayer. In both In pure membrane bilayers (i.e., in the absence of pro-
natural and artificial membranes, a typical lipid molecule ex- tein), phospholipids and sphingolipids rotate and move lat-
changes places with its neighbors in a leaflet about 10- times erally, but they do not spontaneously migrate, or flip-flop,
per second and diffuses several micrometers per second at 37 oc. from one leaflet to the other. The energetic barrier is roo
These diffusion rates indicate that the bilayer is 100 times high; migration would require moving the polar head group
more viscous than water-about the same as the viscosity of from its aqueous environment through the hydrocarbon core
olive oil. Even though lipids diffuse more slowly in the bi- of t he bilayer to the aqueous solution on the other side. Spe-
layer than in an aqueous solvent, a membrane lipid could cial membrane proteins discussed in Chapter 11 are required
diffuse the length of a typical bacterial cell ( 1 p m) in only to flip membrane lipids and other polar molecules from one
1 second and the length of an animal cell in about 20 seconds. leaflet to the other.
When artificial p u re phospholipid membranes are cooled The lateral movements of specific plasma-membrane
below 37 °C, the lipids can undergo a phase transition from proteins and lipids can be quantified by a technique called
a liquidlike (fluid) state to a gel-like (semisolid) state, analo- fluorescence recovery after photobleaching (FRAP). Phos-
gous to the liquid-solid transition when liquid water freezes pholipids containing a fluorescent substituent are used to
(Figure 10-9). Below the p hase-transition temperature, the monitor lipid movement. For proteins, a fragment of a
rate of d iffusion of the lipids drops precipitously. At usual monoclonal antibody that is specific for the exoplasmic do-
p hysiologic temperatures, the hydrophobic interior of natu- main of the desired protein and that has only a single antigen-
ral membranes generally has a low viscosity and a fluid like binding site is tagged with a fluorescent dye. With this
consistency, in contrast to the gel-like consistency observed method, described in Figure 10-10, the rate at which mem-
at lower temperatures. brane molecules move-the diffusion coefficient-can be

(a)

Membrane protein Fluorescent reagent

I I

Bleach with Fluorescence

Label laser recovery
---+
D fJ IJ

(b)

~
c
;;;s Fluorescence before bleaching
.em
c
"iii
3000
./
... · .. - - - - - - - - - - - - - - - - - - - - - - - }-5~~
c
...
::L~~\~~·~;~·~ ~~:~:.~:~~·~ ~~::·~·~· ~~} ~!~:ile
Q) 2000
.::
Q)
(.)
c
1000
Q)
(.)
(JJ
Q)
Bleach
t
0;;;s
u:::
Time (s)

EXPERIMENTAL FIGURE 10-10 Fluorescence recovery after proportional to the fraction of labeled molecules that are mobile in the
photobleaching (FRAP) experiments can quantify the lateral membrane. (b) Results of a FRAP experiment with human hepatoma
movement of proteins and lipids within the plasma membrane. cells treated with a fluorescent antibody specific for the asialoglycopro-
(a) Experimental protocol. Step D Cells are first labeled with a fluores- tein receptor protein. The finding that SO percent of the fluorescence
cent reagent that binds uniformly to a spPcific membrane lipid or returned to the bleacht:!d area indicates that :,o percent of the receptor
protein. Step f) A laser light is then focused on a small area of the molecules in the illuminated membrane patch were mobile and
surface, irreversibly bleaching the bound reagent and thus reducing SO percent were immobile. Because the rate of fluorescence recovery
t he fluorescence in the illuminated area. Step tl ln t ime, the fluores- is proportional to the rate at which labeled molecules move into the
cence of th e bleached patch increases as unbleached fluorescent bleached region, the diffusion coefficient of a protein or lipid in the
surface molecules diffuse into it and bleached ones diffuse outward. membrane can be calculated from such data. [See V.I. Henis et al., 1990,
The extent of recovery of fluorescence in the bleached patch is J. Cell Bioi. 111:1409.]

10.1 The Lipid Bilayer: Composition and Structural Organization 451

determined, as well as the proportion of the molecules that reticulum (ER), where phospholipids are synthesized, and the
arc laterally mobile. Golgi, where sphingolipids arc synthesized. The proportion
The results of FRAP studies with fluorescence-labeled of sphingomyelin as a percentage of total membrane lipid
phospholipids have shown that in fibroblast plasma mem- phosphorus is about six times as high in Golgi membranes as
branes, all the phospholipids are freely mobile over distances it is in ER membranes. In other cases, the movement of mem-
of about 0.5 IJ-m, but most cannot diffuse over much longer branes from one cellular compartment to another can selec-
distances. These findings suggest that protein-rich regions of tively enrich certain membranes in lipids such as cholesterol.
the plasma membrane about 1 IJ-m in diameter separate In responding to differing environments throughout an or-
lipid-rich regions containing the bulk of the membrane phos- ganism, different types of cells generate membranes with
pholipid. Phospholipids are free to diffuse within such re- differing lipid compositions. In the cells that line the intesti-
gions but not from one lipid-rich region to an adjacent one. nal tract, for example, the membranes that face the harsh
Furthermore, the rate of lateral diffusion of lipids in the environment in which dietary nutrients are digested have a
plasma membrane is nearly an order of magnitude slower sphingolipid-to-phosphoglyceride-to-cholesterol ratio of
than in pure phospholipid bilayers: diffusion constants of 1:1:1 rather than the 0.5:1.5:1 ratio found in cells subject to
10 H cm 2/s and 10 ~ cm 2/s are characteristic of the plasma less stress. The relatively high concentration of sphingolipid
membrane and a lipid bilayer, respectively. This difference in this intestinal membrane may increase its stability because
suggests that lipids may be tightly but not irreversibly bound of extensive hydrogen bonding by the free -OH group in the
to certain integral proteins in some membranes, as indeed sphingosine moiety (see Figure 10-8).
has recently been demonstrated (see discussion of annular The degree of bilayer fluidity depends on the lipid compo-
phospholipids, below). sition, the structure of the phospholipid hydrophobic tails,
and temperature. As already noted, van der Waals interac-
tions and the hydrophobic effect cause the nonpolar tails of
phospholipids to aggregate. Long, saturated fatty acyl chains
Lipid Composition Influences the Physical
have the greatest tendency to aggregate, pa~king tightly to-
Properties of Membranes gether into a gel-like state. Phospholipids with short fatty
A typical cell contains many different types of membranes, acyl chains, which have less surface area and therefore fewer
each with unique properties derived from its particular mix van der Waals interactions, form more fluid bilayers . Like-
of lipids and proteins. The data in Table 10-1 illustrate the wise, the kinks in cis-unsaturated fatty acyl chains (Chapter 2)
variation in lipid composition in different biomembranes. result in their forming less stable van der Waals interactions
Several phenomena contribute to these differences. For in- with other lipids, and hence more fluid bilayers, than
stance, the relative abundances of phosphoglycerides and do straight saturated chains, which can pack more tightly
sphingolipids differ between membranes in the endoplasmic together.

TABLE 10-1 Major Lipid Components of Selected Biomembranes

Composition (mol %)
- --- ---
Source/Location PC PE + PS SM Cholesterol
- ---

Plasma membrane (human erythrocytes) 21 29 21 26

.Viyelin membrane (human neurons) 16 37 13 34

Plasma membrane (£. coli) 0 85 0 0

Endoplasmic reticulum membrane (rat) 54 26 5 7

Golgi membrane (rat) 45 20 13 13

Inner mitochondrial membrane (rat) 45 45 2 7

Outer mitochondrial membrane (rat) 34 46 2 11

Primary leaflet location Exoplasmic Cyrosolic Exoplasmic Both

PC phosphatidykholme; PE = phospharidylerhanolamme; PS = phosphandylserine; SM sphingomyelin.

Dow han and .\1. Bogdanov, 2002, in D. f. Vance and J. F. Vance, eds., Biochemistry of T.~pids, Ltpoprotems, and i'•1embranes, Elsevier.
~OL:R<:r: \XI.

452 CHAPTER 10 • Biomembrane Structure

 other membrane components, such as proteins, in a particu
lar membrane. It has been argued that relatively short trans-
membrane segments of certain Golgi-resident enzymes
(glycosyltransferases) are an adaptation to the lipid composi-
tion of the Golgi membrane and contribute to the retention
of these enzymes in the Golgi apparatus. The results of bio-
physical studie~ on artificial membranes demonstrate that
sphingomyelin associates into a more gel-like and thicker bi-
layer than phosphoglycerides uo (Figure 10-1 la ). Cholesterol
and other molecules that decrease membrane fluidity also in-
crease membrane thickness. Because sphingomyelin tails are
already optimally stabilized, the addition of cholesterol has
PC and
cholesterol
SM SM and no effect on the thickness of a sphingomyelin bilayer.
cholesterol Another property dependent on the lipid composition of
a bilayer is its curvature, which depends on the relative sizes
(b)
of the polar head groups and nonpolar tails of its constituent
phospholipids. Lipids with long tails and large head groups
arc cylindrical in shape; those with small head groups are
cone shaped (Figure 10-11 b). As a result, bilayers composed
of cylindrical lipids are relatively flat, whereas those contain-
ing large numbers of cone-shaped lipids form curved bilayers
(figure 10-llc). This effect of lipid composition on bilayer
curvature may play a role in the formation of highl y curved
membranes, such as sites of viral budding (see Figure 10-2)
and formation of internal vesicles from the plasma mem-
brane (see Figure 10-6), and in specialized ~table membrane
FIGURE 10-11 Effect of lipid composition on bilayer thickness structures such as microvilli. Several proteins bind to the sur-
and curvature. (a) A pure sphingomyelin (SM) bilayer is thicker than face of phospholipid bilayers and cause the membrane to
one formed from a phosphoglyceride such as phosphatidylcholine curve; such proteins are important in formation of transport
(PC). Cholesterol has a lipid-ordering effect on phosphoglyceride vesicles that bud from a donor membrane (Chapter 14 ).
bilayers that increases their thickness, but it does not affect the
thickness of the more-ordered SM bilayer. (b) Phospholipids such as PC
have a cylindrical shape and form essentially flat mono layers, whereas Lipid Composition Is Different in the Exoplasmic
those with smaller head groups such as phosphatidylethanolamine and Cytosolic leaflets
(PE) have a conical shape. (c) A bilayer enriched with PC in the
A characteristic of all biomembranes is an asymmetry in
exoplasmic leaflet and with PE in the cytosolic face, as in many plasma
lipid composition across the bilayer. Although most phos-
membranes, would· have a natural curvature. [Adapted from H. Sprong
pholipids are present in both membrane leaflets, some are
et al., 2001, Nature Rev. Mol. Cell Bioi. 2:504.]
commonly more abundant in one or the other leaflet. For
instance, in plasma membranes from human erythrocytes
and Maclin Darby canine kidney (MDCK) cells grown in cul-
Cholesterol is important in maintaining the appropriate ture, almost all the sphingomyelin and phosphatidylcholine,
fluidity of natural membranes, a property that appears to be both of which form less fluid bilayers, are found in the exo-
essential for normal cell gr0wth and reproduction. Cholesterol plasmic leaflet. In contrast, phosphatidylethanolamine,
restricts the random movement of phospholipid head groups phosphatidylserine, and phosphatidylinositol, which form
at the outer surfaces of the leaflets, but its effect on the move- more fluid bilayers, are preferentially located in the cyrosolic
ment of long phospholipid tails depends on concentration. At leaflet. Because phosphatidylserine and phosphatidylinositol
cholesterol concentrations present in the plasma membrane, carry a net negative charge, the stretch of amino acids on the
the interaction of the steroid ring with the long hydrophobic cytoplasmic face of a single-pass membrane protein, 111 close
tails of phospholipids tends to immobilize these lipids and thus proximity to the transmembrane segment, is often enriched
decrease biomembrane fluidity. It is this property that can help in positively charged (Lys, Arg) residues, the "inside posi-
organize the plasma membrane into discrete subdomains of tive" rule. This segrt-gation of lipids across the bilayer may
. ' unique hptd and protein composition. At lower cholesterol influence membrane curvature (see Figure J 0-llc). Unlike
concentrations, however, the steroid ring separates and dis- particular phospholipids, cholesterol is relatively evenly dis-
perses phospho! ipid tails, causing the inner regions of the tributed in both leaflets of cellular membranes. The relative
membrane to become slightly more fluid. abundance of a particular phospholipid in the two leaflets of
The lipid composition of a bilayer a lso influences its a plasma membrane can be determined experimentally on
,.
thickness, which in turn may influence the distribution of the basis of the susceptibility of phospholipids to hydrolysis

10.1 The Lipid Bilayer: Composition and Structural Organization 453

 Polar head group ~ r.::l results in activation of the cytosolic enzyme phospholipase C,

?~~
which can then hydrolyze the bond connecting the phos-
phoinositols to the diacylglycerol. As we will see in Chapter
-o-P=O 15, both water-soluble phosphoinositols and membrane-

f'----.0
CH2 0
embedded diacylglycerol participate in intracellular signaling
pathways that affect many aspects of cell ular metabolism.
Phosphatidylserine also is normally most abundant in the cy-
I II
?H-O -
~C-(CH2lnCH3 tosolic leaflet of the plasma membrane. In the initial stages of
CH 2 platelet stimulation by serum, phosphatidylsenne 1s briefly
I translocated ro the exoplasmic face, presumably by a flippase
enzyme, where it activates enzymes participating in blood
(;] b=o5] clotting. When cells d ie, lipid asymmetry is no longer main-
I tained, and phosphatidylserine, normally enriched in the cy-
(CH 2 )
I n rosolic leaflet, is increasingly found in the exoplasmic one.
CH3 This increased exposure is detected by use of a labeled ver-
FIGURE 10-12 Specificity of phospholipases. Each type of sion of Annexin V, a protein that specifically binds to phos-
phospholipase cleaves one of the susceptible bonds shown in red. The phatidylserine, to measure the ooset of programmed cell
glycerol carbon atoms are indicated by small numbers. In intact cells, death (apoptosis).
only phospholipids in the exoplasmic leaflet of the plasma membrane
are cleaved by phospholipases in the surrounding medium. Phospholi-
pase C, a cytosolic enzyme, cleaves certain phospholipids in the
cytosolic leaflet of the plasma membrane. Cholesterol and Sphingolipids Cluster
with Specific Proteins in Membrane
Microdomains
by phospholipases, enzymes that cleave the ester bonds via Membrane lipids are not randomly distributed (evenly mixed)
which acyl chains and head groups are connected to the lipid in each leaflet of a bilayer. One hint that lipids may be orga-
molecule (Figure 10-12). When added to the external me- nized within the leaflets was the discovery that the lipids re-
dium, phospholipases cannot cross the membrane, and thus maining after the extraction (solubilization) of plasma
they cleave off the head groups of only those lipids present in membranes with nonionic detergents such as Triton-XlOO
the exoplasmic face; phospholipids in the cytosolic leaflet predominantly contain two species: cholesterol and sphingo-
are resistant to hydrolysis because the enzymes cannot pen- myelin. Because these two lipids are found in more ordered,
etrate to the cytosolic face of the plasma membrane. less fluid bilayers, researchers hypothesized that they form
How the asymmetric distribution of phospholipids in microdomains, termed lipid rafts, surrounded by other, more
membrane leafl.ets arises is still unclear. As noted, in pure fluid phospholipids that are more readily extracted by non-
bilayers phospholipids do not spontaneously migrate, or ionic detergents. (We discuss more fully the role of ionic and
flip-flop, from one leaflet to the other. In part, the asym- non ionic detergents in extracting membrane proteins in
metry in phospholipid distribution may reflect where these Section 10.2.)
lipids are synthesized in the endoplasmic reticulum and Some biochemical and microscopic evidence supports the
Golgi. Sphingomyelin is synthesized on the luminal (exo- existence of lipid rafts, which in natural membranes are typi-
plasmic) face of the Golgi, which becomes the exoplasmic cally 50 nm in diameter. Rafts can be disrupted by methyl-~­
face of the plasma membrane. In contrast, phosphoglycer- cyclodextrin, wh ich specifically extracts cholesterol out of
Jdes are synthesized on the cyrosolic face of the ER mem- membranes, or by antibiotics, such as filipin, that sequester
brane, which is topologically equivalent to the cytosolic cholesterol into aggregates within the membrane. Such find-
face of the plasma membrane (see Figure 1 0-5). Clearly, ings indicate the importance of cholesterol in maintaining the
however, this explanation does not account for the prefer- integrity of these rafts. These raft fractions, defined by their
ential location of phospharidylcholine (a phosphoglyceride) insolubility in non ionic detergents, contain a subset of plasma
in the exoplasmic leaflet. Movement of this phosphoglycer- membrane proteins, many of which arc implicated in sensing
ide, and perhaps others, from one leaflet to the other in extracellular signals and transmitting them into the cytosol.
some natural membranes is most likely catalyzed by ATP- Because raft fractions arc enriched in glycolipids, an impor-
powered transport proteins called flippases, which are dis- tant tool for microscopic visualization of raft-type structures
cussed in Chapter 11. in intact cells is the use of fluoresccntly labeled cholera toxin,
The preferential location of lipids on one face of the bi- a protein that specifically binds to certain gangliosides. By
layer is necessary for a variety of membrane-based functions. bringing many key proteins into dose proximity and stabiliz-
For example, the head groups of all phosphorylated forms of ing their interactions, lipid rafts may facilitate signaling by
phosphatidylinositol (see Figure 10-8; PI), an important cell-surface receptors and the subsequent activation of cyro-
source of second messengers, face the cytosol. Stimulation of solic events. However, much remains to be learned about the
many cell-surface receptors by their corresponding ligand structure and biologica l function of lipid rafts.

.··
454 CHAPTER 10 • Biomembrane Structure
 Cells Store Excess Lipids in Lipid Droplets
KEY CONCEPTS of Section 10.1
Lipid droplets are vesicular structures, composed of triglycer-
·. ides and cholesterol esters, that originate from the ER and The Lipid Bilayer: Composition
serve a lipid-storage function. When a cell's supply of lipids and Structural Organization
exceeds the immediate need for membrane construction, ex- • Membranes are crucial to cell structure and function. The
cess lipids are relegated to these lipid droplets, readily visual- eukaryotic cell is demarcated from the external environment
ized in living cells by staining with a lipophilic dye such as by the plasma membrane and organized into membrane-
Congo red. Feeding cells with oleic acid, a type of fatty acid, limited internal compartments (organelles and vesicles).
enhances lipid droplet formation. Lipid droplets are not only
storage compartments for triglycerides and cholesterol esters, • The phospholipid bilayer, the basic structural unit of all
but may also serve as platforms for storage of proteins targeted biomembranes, is a two-dimensional lipid sheet with hydro-
for degradation. The biogenesis of lipid droplets starts with philic faces and a hydrophobic core, which is impermeable
delamination of the lipid bilayer of the ER, through insertion to water-soluble molecules and ions; proteins embedded in
of triglycerides and cholesterol esters (Figure 10-lJ). The lipid the bilayer endow the membrane with specific functions (see
"lens" continues to grow by insertion of more lipid, until fi- Figure 10-1).
nally a lipid droplet is hatched by scission from the ER. The • The primary lipid components of biomembranes are phos-
resulting cytoplasmic droplet is thereby enwrapped by a phos- phoglyceridcs, sphingolipids, and sterols such as cholesterol
pholipid monolayer. The details of lipid droplet biogenesis as (see Figure 10-8). The term "phospholipid" applies to any
well as their functions remain to be defined more clearly. amphipathic lipid molecule with a fatty acyl hydrocarbon
tail and a phosphate-based polar head group.
• Phospholipids spontaneously form bilayers and scaled com-
partments surrounding an aqueous space (see Figure 10-3 ).
• As bilayers, all membranes have an internal (cytosolic) face
and an external (exoplasmic) face (see Figure 10-5). Some
ER membrane
organelles are surrounded by two, rather than one, membrane
bilayer.
• Most lipids and many proteins are laterally mobile in bio-
membranes (see Figure 10-1 0). Membranes can undergo phase
transitions from fluid- to gel-like states depending on the tem-
fiffr&m~~·~ perature and composition of the membrane (see Figure 10-9).

~~Cholesterol and • Different cellular membranes vary in lipid composition (see

Table 10-1). Phospholipids and sphingolipids are asymmetri-
triglycerides
cally distributed in the two leaflets of the bilayer, whereas
cholesterol is fairly evenly distributed in both leaflets.
• Natural biomembranes generally have a viscous consis-
tency with fluidlike properties. In general, membrane fluidity
is decreased by sphingolipids and cholesterol and increased
hy phosphoglycerides. The lipid composition of a membrane
"Lens"
also influences its thickness and curvature (see Figure 10-11 ).
• Lipid rafts are microdomains containing cholesterol,
sphingolipids, and certain membrane proteins that form in
the plane of the bilayer. These lipid-protein aggregates might
facilitate signaling by certain plasma membrane receptors.
Lipid droplet
• Lipid droplets are storage vesicles for lipids, originating in
formed from
cytoplasmic the ER (see Figure 10-13).
leaflet

FIGURE 10- 13 lipid droplets form by budding and scission from

the ER membrane. Lipid droplet formntion begins with the accumula-
tion of cholesterol esters and triglycerides (yellow) within the hydro- 10.2 Membrane Proteins: Structure and
phobic core of the lipid bilayer. The resulting delamination of the two Basic Functions
lipid monolayers causes a "lens" to form, the further growth of which
creates a spherical droplet that is then released by scission at the neck. Membrane proteins are defined by their location within or at
The newly formed droplet is surrounded by a lipid monolayer, derived the surface of a phospholipid bilayer. Although every bio-
from the cytosolicleaflet of the ER membrane. logical membrane has the same basic bilayer structure, the

10.2 Membrane Proteins: Structure and Basic Functions 455

proteins associated with a particular membrane are respon- soluble proteins in their amino acid composition and struc-
sible for its distinctive activities. The kinds and amounts of ture. In contrast, the membrane-spanning segments usually
proteins associated with biomembranes vary depending on contain many hydrophobic amino acids whose side chains
cell type and subcellular location. For example, the inner mi- protrude outward and interact with the hydrophobic hydro-
tochondrial membrane is 76 percent protein; the myelin carbon core of the phospholipid bilayer. In all transmem-
membrane that surrounds nerve axons, only 18 percent. The brane proteins examined to date, the membrane-spanning
high phospholipid content of myelin allows it to electrically domains consist of one or more a helices or of multiple f3
insulate the nerve from its environment, as we discuss in strands. We discussed the ribosomal synthesis and post-
Chapter 22. The importance of membrane proteins is evJdent translational processing of soluble cytosolic proteins in
from the finding that approximately a third of all yeast genes Chapters 4 and 8; the process by which integral membrane
encode a membrane protein. The relative abundance of genes proteins arc inserted into membranes as part of their synthe-
for membrane proteins is greater in multicellular organisms, sis is discussed in Chapter 13.
in which membrane proteins have additional functions in cell Lipid-anchored membrane proteins are bound covalently
adhesion. to one or more lipid molecules. The hydrophobic segment of
The lip1d bilayer presents a distinctive two-dimensional hy- the attached lipid is embedded in one leaflet of the mem-
drophobic environment for membrane proteins. Some proteins brane and anchors the protein to the membrane. The poly-
contain segments that are embedded within the hydrophobic peptide chain itself docs not enter'the phospholipid bilayer.
core of the phospholipid bilayer; other proteins are associated Peripheral membrane proteins do not directly contact the
with the exoplasmic or cytosolic leaflet of the bilayer. Protein hydrophobic core of the phospholipid bilayer. Instead they
domains on the extracellular surface of the plasma membrane are bound to the membrane either indirectly by interactions
generally bind to extracellular molecules, including external with integral or lipid-anchored membrane proteins or di-
signaling proteins, ions, and small metabolites (e.g., glucose, rectly by interactions with lipid head groups. Peripheral pro-
fatty acids), as well as proteins on other cells or in the external teins can be bound to either the cytosolic orthe exoplasmic
environment. Segments of proteins within the plasma mem- face of the plasma membrane. In addition to these proteins,
brane perform multiple functions, such as forming the chan- which are closely associated with the bilayer, cytoskeletal
nels and pores through which molecules and ions move into filaments can be more loosely associated with the cytosolic
and out of cells. Intramembrane segments also serve to orga- face, usually through one or more peripheral (adapter) pro-
nize multiple membrane proteins into larger assemblies within teins. Such associations with the cytoskeleton provide sup-
the plane of the membrane. Domains lying along the cytosolic port for various cellular membranes, helping to determine
face of the plasma membrane have a wide range of functions, cell shape and mechanical properties, and play a role in the
from anchoring cytoskeletal proteins to the membrane to trig- two-way communication between the cell interior and the
gering intracellular signaling pathways. exterior, as we learn in Chapter 17. Finally, peripheral pro-
In many cases, the function of a membrane protein and the teins on the outer surface of the plasma membrane and the
topology of its polypeptide chain in the membrane can be pre- exoplasmic domains of integral membrane proteins are often
dicted on the basis of its similarity with other well-characterized attached to components of the extracellular matrix or to the
proteins. In this section, we examine the characteristic struc- cell wall surrounding bacterial and plant cells, providing a
tural features of membrane proteins and some of their basic crucial interface between the cell and its environment.
functions. We will describe the structures of several proteins
to help you get a feel for the way membrane proteins interact
Most Transmembrane Proteins Have
with membranes. More complete characterization of the
properties of various types of membrane proteins is presented Membrane-Spanning a Helices
in later chapters that focus on their structures and activities Soluble proteins exhibit hundreds of distinct localized folded
in the context of their cellular functions. structures, or motifs (see Figure 3-9). In comparison, the reper-
toire of folded structures in the transmembrane domains of
integral membrane proteins is quite limited, with the hydro-
Proteins Interact with Membranes
phobic a helix predominating. Proteins containing membrane-
in Three Different Ways spanning a-helical domains are stably embedded in membranes
Membrane proteins can be classified into three categories- because of energetically favorable hydrophobic and van der
integral, lipid-anchored, and peripheral-{)n the basis of their Waals interactions of the hydrophobic side chains in the do-
position with respect to the membrane (see Figure 10-1). main with specific lipids and probably also by ionic interac-
Integral membrane proteins, also called transmembrane pro- tions witlt the polar head groups of the phospholipids.
tems, span a phospholipid bilayer and comprise three A single a-helical domain is sufficient to incorporate an
segments. The cytosolic and exoplasmic domains have hydro- integral membrane protein into a membrane. However,
philic exterior surfaces that interact with the aqueous envi- many proteins have more than one transmembrane a helix.
ronment on the cytosolic and exoplasmic faces of the Typically, a membrane-embedded a helix is composed of a
membrane. These domains resemble segments of other water- continuous segment of 20-25 hydrophobic (uncharged )

456 CHAPTER 10 • Biomembrane Structure

 (a) Glycophorin A dimer (b) Transmembrane coiled-coil domain

t:w.:;->~---Coiled-coil dimer
stabilized by van
der Waals
interactions
between adjacent
side chains

.·
Cytosolic
domain

FIGURE 10- 14 Structure of glycophorin A, a typical single-pass domain is heavily glycosylated, with the carbohydrate side chains
transmembrane protein. (a) Diagram of dime ric glycophorin showing (green diamonds) attached to specific serine, threo'nine, and aspara-
major sequence features and its relation to the membrane. The single gine residues. (b) Molecular model of the transmembrane domain of
23-residue membrane-spanning a helix in each monomer is composed dime ric glycophorin corresponding to residues 73-96. The hydrophobic
of amino acids with hydrophobic (uncharged) side chains (red and side chains of the a helix in one monomer are shown in pink; those
green spheres). By binding negatively charged phospholipid head in the other monomer, in green. Residues depicted as space-filling
groups, the positively charged arginine and lysine residues (blue structures participate in van der Waals interactions that stabilize the
spheres) near the cytosolic side of the helix help anchor glycophorin coiled-coil dimer. Note how the hydrophobic side chains project
in the membrane. Both the extracellular and the cytosolic domains are outward from the helix, toward what would be the surrounding fatty
rich in charged residues and polar uncharged residues; the extracellular acyl chains. [Part (b) adapted from K. R. MacKenzie et al., 1997, Science 276:131.]

amino acids (see Figure 2-14). The predicted length of such brane protein, which contains only one membrane-spanning
an a helix (3.75 nm) is just sufficient to span the hydrocar- a helix (Figure 10-14 ). The 23-residue membrane-spanning a
bon core of a phospholipid bilayer. In many membrane pro- helix is composed of amino acids with hydrophobic (un-
teins, these helices are perpendicular to the plane of the charged) side chains, which interact with fatty acyl chains in
membrane, whereas in others, the helices traverse the mem- the surrounding bilayer. In cells, glycophorin A typically
brane at an oblique angle. The hydrophobic side chains pro- forms dimers: the transmembrane helix of one glycophorin A
trude outward from the helix and form van der Waals polypeptide associates with the corresponding transmem-
interactions with the farcy acyl chains in the bilayer. In con- brane helix in a second glycophorin A to form a coiled-coil
trast, the hydrophilic amide peptide bonds are in the interior structure (Figure 10-14b). Such interactions of membrane-
of the a helix (see Figure 3-4); each carbonyl (C=O) group spanning a helices are a common mechanism for creating
forms a hydrogen bond with the amide hydrogen atom of dtmeric membrane proteins, and many membrane proteins
the amino acid four residues toward the C-terminus of the form oligomers (two or more polypeptides bound together
helix. These polar groups are shielded from the hydrophobic noncovalenrly) by interactions between their membrane-
interior of the membrane. spanning helices.
To help you get a better sense of the structures of proteins A large and important group of integral proteins is de-
with a-helical domains, we will briefly discuss four different fined by the presence of seven membrane-spanning a helices;
kinds of such protems: glycophorin A, G protein-coupled this includes the large family of G protein-<:oupled cell-surface
receptors, aquaporins (water/glycerol channels), and T-cell receptors discussed in Chapter 15, several of which have
receptor for antigen. been crystallized. One such multipass transmembrane
Glycophorin A, the major protein in the erythrocyte protein of known structure is bacteriorhodopsin, a protein
plasma membrane, is a representative single-pass transmem- found in the membrane of certain photosynthetic bacteria; it

10.2 Membran e Proteins: Structure and Bas1c Functions 457

(a) Bacteriorhodopsin (b) Glycerol channel
Back Front Half elices

Exterior

Membrane

Cytosol

FIGURE 10-1 5 Structural models of two multipass membrane several membrane-spanni ng a helices that are at oblique angles, the
proteins. (a) Bacteriorhodopsin, a photoreceptor in certain bacteria. two helices that penetrate only halfway through th e membran e
The seven hydrophobic a helices in bact eriorhodopsin traverse the (purple with yellow arrows), and one long membrane-spanning helix
lipid bilayer roughly perpendicu lar to the plane of the membrane. A with a " break" or distortion in the middle (purple with yellow line). The
retinal molecule (black) covalently attached to one helix absorbs light. glycerol molecule in the hydrophilic "core" is colored red. The st ructure
The large class of G protein-coupled receptors in eukaryotic cells also was approximately positioned in the hydrocarbon core ofthe
has seven membrane-spanning a helices; their t hree-dimensional membrane by finding the most hydrophobic 3-~m slab of the protein
structure is thought to be similar to that of bacteriorhodopsin. (b) Two perpendicular to the membrane plane. [Part (a) after H. Luecke et at., 1999,
views of the glycerol channel Glpf, rotated 180° with respect to each J. Mol. Bioi. 291:899. Part (b) after J. Bowie, 2005, Nature 438:581-589, and D. Fu
other along an axis perpendicular to the plane of the membrane. Note et at., 2000, Science 290:481-486.]

illustrates the general structure of all these proteins (Figure aquaporin has one long transmembrane helix w ith a bend in
10-15a). Absorption of light by the retinal group covalently the middle, and more strikingly, there are two a helices that
attached to this protein causes a conformational change in penetrate only halfway through the membrane. TheN-termini
the protein that results in the pumping of protons from the of these he lices face eac h other (yellow N's in the figure),
cytosol across the bacterial membrane to the extracellular and together they span the membrane at an oblique angle.
space. The proton concentration gradient thus generated Thus some membrane-embedded helices- and other, nonhe-
across the membrane is used to synthesize ATP during pho- lical, structures we will encounter later-do not traverse the
tosynthesis (Chapter 12). In the high-resolution structure of entire bilayer. As we will see in Chapter 11, these short heli -
bacteriorhodopsin the positions of all the individual amino ces in aquapor ins form part of the glycerol/water-selective
acids, retinal, and the surrounding lipids are clearly defined. pore in the middle of each subunit. This highlights the con-
As might be expected, virtually all of the amino acids on the siderable diversity in the ways membrane-spanning a helices
exterior of the membrane-spanning segments of bacteriorho- interact with the lipid bilayer and with other segments of the
·.
dopsin are hydrophobic, permitting energetically favorable protein.
interactions with the hydrocarbon core of the surrounding The specificity of phospholipid-protein interactions is evi-
lipid bilayer. dent from the structure of a different aquaporin, aquaporin 0
The aquaporins are a large family of high ly conserved (Figu re 1 0- 16). Aquaporin 0 is the most abundant protein in
proteins that transport water, glycerol, and other hydro- the plasma membrane of the fiber cells that make up t he bulk
philic molecules across biomembranes. They illustrate sev- of the lens of the mammalian eye. Like other aquaporins, it is
eral aspects of the structure of multipass transmembrane a tetramer of identical subunits. The protein's surface is not
proteins. Aquaporins arc tc tramers of four identical sub- t-uvered by a set of uniform bind ing sites for phospholipid
units. Each of the four subunits has six membrane-spanning molecules. Instead, fatty acyl side chains pack tightly against
a helixes, some of which traverse the membrane at oblique the irregular hydrophobic outer surface of the protein; these
angles rather than perpendicularly. Because the aquaporins lipids are referred to as annular phospholipids, because they
have similar structures, we w ill focus on one, the glycerol from a tight ring (annulus) of lipids that exchange less easily
channel Glpf, that has an especially well-defined structure with bulk phospholipids in the bilayer. Some of t he fatty acyl
determined by x-ray diffraction studies (Figure 10-15b). This chains are straight, in the all-trans conformation (Chapter 2),

458 CHAP TER 10 • Biomembrane Structure

 @ PODCAST: Annular Phospholipids

FIGURE 10-16 Annular phospholipids. Side view of the three-

Exterior dimen sional structure of one subunit of the lens-specific aquaporin
0 homotetramer, crystal lized in the presence of the phospholipid
dimyristoylphosphatidylcholine, a phospholipid with 14 carbon-
saturated fatty acyl chains. Note the lipid molecules forming a bilayer
shell around the protein. The protein is shown as a surface plot (the
liqhter background molecule). The lipid molecules are shown in
Membrane
space-fill format; the polar lipid head groups (grey and red) and the
lipid fatty acyl chains (black and grey) form a bilayer with almost
uniform thickness around the protein. Presumably, in the membrane,
lipid fatty acyl chains will cover the whole of the hydrophobic surface
of the protein; only the most ordered of the lipid molecules will be
Interior resolved in the crystallographic structure. [After A. Lee, 2005,
Nature 438:569-570, and T. Gonen et al., 2005, Nature 438:633-688.)

'·

whereas others are kinked in order ro interact with bulky affected by the specific types of phospholipid present in the
hydroph ilic side chains on the surface of the protein. Some bilayer.
of the li pid head groups are parallel to the su rface of the In addition to the predominantly hydrophobic (un-
membrane, as is the case in purified phospholipid bilayers. charged) residues that serve to embed int~gral membrane
Others, however, are oriented almost at right angles to the proteins in the bi layer, many such a-helical transmembrane
plane of the membrane. Thus there can be specific interac- segments do contain polar and/or charged residues. Their
tions between phospholipids and membrane-spanni ng pro- amino acid side chains can be used to guide the assembly and
teins, and the function of many membrane proteins can be stabilization of multimeric membrane proteins. The T-ccll re-
ceptor for antigen is a case in point: it is composed of four
separate dimers, the interactions of which are driven by
charge-charge interactions between a helices at the appropri-
ER lumen
ate "depth" in the hydrocarbon core of the lipid bi layer (Fig-
ure 10- 17). The electrostatic attraction of positive and
C030E negative charges on each dimer helps the d imers to "find each
other." T h us charged residues in otherwise hydrophobic
transmembrane segments can help guide assembly of multi-
meric membra ne proteins.

FIGURE 10- 17 Charged residues can orchestrate assembly of

multimeric membrane proteins. The T-cell receptor (TCR) for antigen
is composed of four separate dimers: an al3 pair directly responsible for
Cytosol antigen recognition, and accessory subunits collectively referred to as
o,
the CD3 complex. These accessories include the-y, E:, and~ subunits.
The~ subunits form a disulfide-linked homodimer. The "Y and 8
subunits occur in complex with an E subunit, to generate a "YE and a 1iE
pair. The transmembrane segment s of the TCR a and 13 chains each
ER lumen
contain positively charged residues (blue). These allow recruitment of
corresponding 8E: and 'fE: heterodimers, which carry negative charges
(red) at the appropriate depth in the hydrophobic core of the bilayer.
The~ homodimer docks onto the charges in the TCR a chain (dark
green), w hile the 'fE: and 8E: subunit pairs find their corresponding
~ 1:
partners deeper down in the hydrophobic core on both the TCR a and
Membrane TCR 13 chain (light green). Charged residues in otherwise nonpolar
I
'-r
transmembrane segments ca n thu s guide assembly of higher order
Cytosol structures. [After K.W. Wucherpfennig, et al., 2010, Cold Spring Horb Perspect
Bio/, 2.)

10.2 Membrane Proteins: Structure and Basic Functions 459

Multiple 13 Strands in Porins Form protein, a porin has a hydrophilic interior and a hydrophobic
Membrane-Spanning "Barrels" exterior; in this sense, porins are inside out. In a porin mono-
mer, the ourward-facing side groups on each of the 13 strands
The porins are a class of transmembrane proteins whose struc- are hydrophobic and form a nonpolar ribbonlike band that
ture differs radically from that of other integral proteins based encircles the outside of the barrel. This hydrophobic band in-
on a-helical transmembrane domains. Several types of porins teracts with the fatty acyl groups of the membrane lipids or
are found in the outer membrane of gram-negative bacteria with other porin monomers. The side groups facing the inside
such as E. coli and in the outer membranes of mitochondria of a porin monomer are predominantly hydroph ilic; they line
and chloroplasts. The outer membrane protects an intestinal the pore through which small water-soluble molecules cross
bacterium from harmful agents (e.g., antibiotics, bile salts, and the membrane. (Note that the aquaporins discussed above, de-
proteases) but permits the uptake and disposal of small hydro- spite their name, are not porins and contain mu ltiple trans-
philic molecules, including nutrients and waste products. Dif- membrane a helices.)
ferent types of porins in the outer membrane of an E. coli cell
provide channels for the passage of specific types of disaccha-
rides or other small molecules as well as of ions such as phos- Covalently Attached Lipids Anchor
phate. The amino acid sequences of porins contain none of the Some Proteins to Membranes
long, continuous hydrophobic segments typical of integral pro- In eukaryotic cells, covalently attached lipids can anchor some
teins with a-helical membrane-spanning domains. Rather, it is otherwise typically water-soluble proteins to one or the other
the entire outer surface of the fully folded porin that displays leaflet of the membrane. In such lipid-anchored proteins, the
its hydrophobic character to the hydrocarbon core of the lipid lipid hydrocarbon chains are embedded in the bilayer, but the
bilayer. X-ray crystallography shows that porins are trimers of protein itself does not enter the bilayer. The lipid anchors used
identical subunits. In each subunit, 16 13 strands form a sheet to anchor proteins to the cytosolic face are not used for the
that twists into a barrel-shaped structure with a pore in the exoplasmic face and vice versa. .
center (Figure 10-18). Unlike a typical water-soluble globular One group of cytosolic proteins are anchored to the cyto-
solic face of a membrane by a fatty acyl group (e.g., myristate
or palmitate) covalently attached to an N-terminal glycine
residue, a process called acylation (Figure 10-19a ). Retention
of such proteins at the membrane by the N-terminal acyl an-
chor may play an important role in a membrane-associated
function. For example, v-Src, a mutant form of a cellular tyro-
sine kinase, induces abnormal cellular growth that can lead to
cancer but does so only when it has a myristylated N-terminus.
A second group of cytosolic proteins are anchored to
membranes by a hydrocarbon chain attached to a cysteine
residue at or near the C-terminus, a process called prenyl-
ation (Figure 10-19b). Prcnyl anchors are built from 5-carbon
isoprene units, which, as detailed in the following section,
are also used in the synthesis of cholesterol. In prenylation, a
15-carbon farnesyl or 20-carbon geranylgeranyl g roup is
bound through a thioether bond to the -SH group of a C-
terminal cysteine residue of the protein. In some cases, a sec-
ond geranylgeranyl group or a fatty acyl palmitate group is
linked to a nearby cysteine residue. The additional hydrocar-
bon anchor is thought to reinforce the attachment of the
protein to the membrane. For example, Ras, a GTPase su-
perfamily protein that functions in intracellular signaling
(Chapter 15), is recruited to the cytosolic face of the plasma
membrane by such a double anchor. Rab proteins, which
also belong to the GTPase superfamily, are similarly bound
Periplasm to the cytosolic surface of intracellular vesicles by prenyl an-
FIGURE 10-1 8 Structural model of one subunit of OmpX, a porin
chors; these proteins arc requireu fur the fusion of vesicles
found in t he outer membrane of E. coli. All porins are trimeric with their target membranes (Chapter 14).
transmembrane proteins. Each subunit is barrel shaped, with 13 strands Some cell-surface proteins and specialized proteins with
forming the wall and a transmembrane pore in the center. A band of distinctive covalently attached polysaccharides called pro-
aliphatic (hydrophobic and noncyclic) side chains (yellow) and a border teoglycans (Chapter 20 ) are bound to the exoplasmic face
of aromatic (ring-containing) side chains (red) position the protein in of the plasma membrane by a third type of anchor group,
the bilayer. [After G. E. Schulz, 2000, Curr. Opm. Struc. Bioi. 1 0:443.] glycosylphosphatidylinositol (GPI). The exact structures of

460 CHAPTER 10 • Biomembrane Structure

 and exoplasmic segments always face the opposite side of the
membrane. This asymmetry in protein orientation confers
different properties on the two membrane faces. The orienta-
tion of different types of transmembrane proteins is estab-
lished during their synthesis, as we describe in Chapter 13.
Membrane proteins have never been observed to flip-flop
(c) GPI anchor
across a membrane; such movement, requiring a transient
Exterior movement of hydrophilic amino acid residues through the
hydrophobic interior of the membrane, would be energeti-
cally unfavorable. Accordingly, the asymmetric topology of a
transmembrane protein, which is established during its bin-
synthetic insertion into a membrane, is maintained through-
. ' out the protein's lifetime. As Figure 10-6 shows, membrane
proteins retain their asymmetric orientation in the membrane
Cytosol during membrane budding and fusion events; the same seg-
ment always faces the cytosol and the same segment is always
exposed to the exoplasmic face. In proteins with multiple
transmembrane segments (multipass or polytopic membrane
proteins), orientation of individual transmembrane segments
(a) Acylation (b) Prenylation can be affected by changes in phospholipid composition.
FIGURE 10-19 Anchoring of plasma-membrane proteins to the Many transmembrane proteins contain carbohydrate
bilayer by covalently linked hydrocarbon groups. (a) Cytosolic chains covalently linked to serine, threonine, or asparagine side
proteins such as v-Src are associated with the plasma membrane chains of the polypeptide. Such transmembrane glycoproteins
through a single fatty acyl chain attached to the N-terminal glycine (Giy) are always oriented so that all carbohydrate chains are in the
residue of the polypeptide. Myristate ((14) and palmitate (C 16) are exoplasmic domain (see Figure 10-14 for the example of gly-
common acyl anchors. (b) Other cytosolic proteins (e.g., Ras and Rab cophorin A). Likewise, glycolipids, in which a carbohydrate
proteins) are anchored to the membrane by prenylation of one or two chain is attached to the glycerol or sphingosine backbone of a
cysteine (Cys) residues, at or near the (-terminus. The anchors are membrane lipid, are always located in the exoplasmic leaflet
farnesyl ((15) and geranylgeranyl (C20) groups, both of which are with the carbohydrate chain protruding from the membrane
unsaturated. (c) The lipid anchor on the exoplasmic surface of the plasma surface. The biosynthetic basis for the asymmetric glycosyl-
membrane is glycosylphosphatidylinositol (GPI). The phosphatidylinosi-
ation of proteins is described in Chapter 14. Both glycopro-
tol part (red) of this anchor contains two fatty acyl chains that extend
teins and glycolipids are especially abundant in the plasma
into the bilayer. The phosphoethanolamine unit (purple) in the anchor
membranes of eukaryotic cells and in the membranes of the
links it to the protein. The two green hexagons represent sugar units,
which vary in number, natu re, and arrangement in different GPI anchors.
intracellular compartments that establish the secretory and en-
The complete structure of a yeast GPI anchor is shown in Figure 13-15. docytic pathways; they are absent from the inner mitochon-
[Adapted from H. Sprong et al., 2001 , Nature Rev. Mol. Cell Bioi. 2:504.] drial membrane, chloroplast lamellae, and several other
intracellular membranes. Because the carbohydrate chains of
glycoproteins and glycolipids in the plasma membrane extend
G Pi anchors vary greatly in different cell types, but they always into the extracellular space, they are available to interact with
contain phosphatidylinositol (PI), whose two fatty acyl chains components of the extracellular matrix as well as lectins (pro-
extend into the lipid bilayer just like those of typical membrane teins that bind specific sugars), growth factors, and antibodies.
phospholipids; phosphoethanolamine, which covalently links
the anchor to the C-terminus of a protein; and several sugar
residues (Figure 10-19c). Therefore GPI anchors are glycolip- One important consequence of such interactions is il-
ids. The GPI anchor is both necessary and sufficient for binding lustrated by the A, B, and 0 blood group antigens.
proteins co the membrane. for instance, treatment of cells with These three structurally related oligosaccharide components
phospholipase C, which cleaves the phosphate-glycerol bond in of certain glycoproteins and glycolipids are expressed on the
phospholipids and in GPI anchors (see Figure 10-12), releases surfaces of human red blood cells and many other cell types
GPI-anchored proteins such as Thy-1 and placental alkaline (Figure 10-20). All humans have the enzymes for synthesizing
phosphatase (PLAP) from the cell surface. 0 antigen. Persons with type A blood also have a glycosyl-
transferase enzyme that adds an extra modified monosaccha-
ride called N-acetylgalactosamine to 0 antigen to form A
All Transmembrane Proteins and Glycolipids
antigen. Those with type B blood have a different transferase
Are Asymmetrically Oriented in the Bilayer that adds an extra galactose to 0 antigen to form B antigen.
Every type of transmembrane protein has a specific orienta- People with both transferases produce both A and B antigen
tion, known as its topology, with respect to the membrane (AB blood type); those who lack these transferases produce 0
faces. Its cytosolic segments are alway!> facing the cytoplasm, antigen only (0 blood type).

10.2 Membrane Proteins: Structure and Basic Functions 461

Lipid or lipases, for example, hydrolyze various bonds in the head
protein A antigen groups of phospholipids (see Figure 10-12), and thereby play
a variety of roles in cells-helping to degrade damaged or

GaiNAc
transferase
1 aged cell membranes, generating precursors to signaling
molecules, and even serving as the active components in
many snake venoms. Many such enzymes, including phos-
pholipases, initially bind to the polar head groups of mem-
Lipid or
protein 0 antigen brane phospholipids to carry out their catalytic functions.
The mechanism of action uf phospholipase A 2 illustrates
how such water-soluble enzymes can reversibly interact with
membranes and catalyze reactions at the interface of an
aqueous solution and lipid surface. When this enzyme is in
aqueous solution, its Ca 2 -containing active site is buried in
Lipid or
8 antigen a channel lined with hydrophobic amino acids. The enzyme
protein
binds with greatest affinity to bilayers composed of nega-
tively charged phospholipids (e.g., phosphotidylserine). This
Glc = Glucose finding suggests that the rim of positively charged lysine and
Gal = Galactose
GlcNAc = N-Acetylglucosamine
arginine residues around the entrance to the catalytic chan-
GaiNAc = N-Acetylgalactosamine nel is particularly important in binding (Figure 10-21 a).
Fuc =Fucose Binding induces a conformational change in phospholipase
A2 that strengthens its binding to the phospholipid heads
FIGURE 10-20 Human ABO blood group antigens. These antigens
and opens the hydrophobic channel. As a phospholipid mol-
are oligosaccharide chains covalently attached to glycolipids or
ecule moves from the bilayer into the channel, the enzyme-
glycoproteins in the plasma membrane. The terminal oligosaccharide
bound Ca 2 + binds to the phosphate in the head group,
sugars distinguish the three antigens. The presence or absence of the
glycosyltransferases that add galactose (Gal) or N-acetylgalactosamine
thereby positioning the ester bond to be cleaved in the cata-
(GaiNAc) to 0 antigen determine a person's blood type.
lytic site (Figure 10-21b) and so releasing the acyl chain.

People whose erythrocytes lack the A antigen, the B anti-

Proteins Can Be Removed from Membranes
gen, or both on their surface normally have antibodies against by Detergents or High-Salt Solutions
the missing antigen(s) in their serum. Thus if a type A or 0 Membrane proteins are often difficult to purify and study,
person receives a transfusion of type B blood, antibodies mostly because of their tight association with membrane lip-
against the B antigen will bind to the introduced red cells and ids and other membrane proteins. Detergents are amphipa-
trigger their destruction. To prevent such harmful reactions, thic molecules that disrupt membranes by intercalating into
blood group typi~g and appropriate matching of blood donors phospholipid bilayers and can thus be used to solubilize lip-
and recipients are required in all transfusions (Ta ble 10-2). • ids and many membrane proteins. The hydrophobic part of
a detergent molecule is attracted to the phospholipid hydro-
carbons and mingles with them readily; the hydrophilic part
Lipid-Binding Motifs Help Target Peripheral
is strongly attracted to water. Some detergents such as the
Proteins to the Membrane bile salts are natural products, but most are synthetic mole-
Many water-soluble enzymes use phospholipids as their sub- cules developed for cleaning and for dispersing mixtures of
strates and thus must bind to membrane surfaces. Phospho- oil and water in the food industry (e.g., creamy peanut butter)

TABLE 10-2 ABO Blood Groups

Blood Group Antigens on RBCs* Seru m Antibodies Can Receive Blood Types
- --
A A Anti-B A and 0

B B Ami-A Band 0

AB A and B None All

0 0 Anti-A and anti-B 0

•sec Figure 10-20 for anrigen structures.

462 CHAPTER 10 • Biomembrane Structure

 (Figure 10-22). Ionic detergents, such as sodium deoxycholate
and sodium dodecylsulfate (S DS), contain a charged group;
nonionic detergents, such as Triton X-100 and octylglucoside,
lack a charged group. At very low concentrations, detergents
dissolve in pure water as isolated molecules. As the concentra-
tion increases, the molecules begin to form micelles-small,
spherical aggregates in w hich the hydrophilic parts of the
molecules face outward and the hydrophobic parts cluster in
the center (see Figure 10-3c). The critical micelle concmtration
(CMC) at which micelles form is characteristic of each deter-
gent and is a function of the structures of its hydrophobic
and hydrophilic parts.
Ionic and nonionic detergents interact differently with pro-
teins and have different uses in the lab. Ionic detergents bind to
the exposed hydrophobic regions of membrane proteins as well
as to the hydrophobic cores of water-soluble proteins. Because
of their charge, these detergents can also disrupt ionic and hy-
drogen bonds. At high concentrations, for example, sodium do-
decylsulfate completely denatures proteins by binding to every
side chain, a property that is exploited in SDS gel electrophoresis
(see Figure 3-36). Nonionic detergents generally do not denature
proteins and are thus useful in extracting proteins in their folded
FIGURE 10-21 Lipid-binding surface and mechanism of action
and active form from membranes before the proteins are puri-
of phospholipase A2 • (a) A structural model of the enzyme showing
fied. Protein-protein interactions, especially the weaker ones,
the surface that interacts w ith a membrane. This lipid-binding surface
can be sensitive to both ionic and nonionic detergents.
contains a rim of positively cha rged arginine and lysine residues,
At high concentrations (a bove the CMC), nonionic de-
shown in blue surrounding the cavity of the catalytic active site, in
which a substrate lipid (red stick structure) is bound. (b) Diagram of
tergents solubilize biological membranes by forming mixed
catalysis by phospholipase A2• When docked on a model lipid mem- micelles of detergent, phospholipid, and integral membrane
brane, positively charged residues of the binding site bind to nega- proteins, bulky hydrophobic structures that do not dissolve
tively charged polar groups at the membrane surface. This binding in aqueous solution (Figure 10-23, top). At low concentra-
triggers a small conformational change, opening a channel lined with tions (below the CMC), these detergents bind to the hydro-
hydrophobic amino acids that leads from the bilayer to the catalytic phobic regions of most integral membrane proteins, but
site. As a phospholipid moves into the channel, an enzyme-bound Ca2 without forming micelles, allowing them to remain soluble
ion (green) binds to the head group, positioning the ester bond to be in aqueous solution (Figure 10-23, bottom ). Creating such
cleaved (red) next to the catalytic site. [Part (a) adapted from M. H. Gelb et al., an aqueous solution of integral membrane proteins IS a nec-
1999, Curr. Opm. Struc. Bioi. 9:428. Part (b), see D. Blow, 1991, Nature 351 :444.] essary first step in protein purification.

IONIC DETERGENTS
H3 C L....__ _ __
0
I
HC - <;H 2 - CH 2 - COO-Na+ H3C (CH 2 l 111- 0 - S - 0-Na+
....__-~- II
0

Sodium deoxycholate Sodium dodecylsulfate (SDS)

FIGURE 1 0-22 Structures of four
common detergents. The hydro-
phobic part of each molecule is
NONIONIC DETERGENTS shown in yellow; the hydrophilic
HOCH 2
part, in blue. The bile salt sodium

0
deoxycholate is a naturdl product;
H3 C CH3 the others are synthetic. Although
H3C C CH 2 C- ;, r O (CH - 2 CH 2 - 0l9 .5 - H HO
H
ionic detergents commonly cause
denaturation of proteins, non ionic
(Average)
OH detergents do not and are thu s
Triton X-100 Octylglucoside useful in solubilizing integral
(polyoxyethylene(9.5)p-t-octylphenol) (octyl·~-o-glucopyranoside) membrane proteins.

10.2 Membrane Proteins: Structure and Basic Functions 463

FIGURE 10-23 Solubilization of integral
membrane proteins by nonionic detergents.
At a concentration higher than its critical micelle
concentration (CMC), a detergent solubilizes Concentration
above CMC
lipids and integral membrane proteins, forming

mmru ~ 1
mixed micelles containing detergent, protein,
and lipid molecules. At concentrations below the
CMC. non ionic detergents (e.g., octylglucoside,
Triton X-100) can dissolve membrane proteins

mm~~~L.._______.~
Detergent
without forming micelles by coating the
membrane-spanning regions.

Dissolved
Concentration but not
below CMC form ing
micelles

Treatment of cultured cells with a buffered salt solution

containmg a nonionic detergent such as Triton X-1 00 extracts surrounding the cytosolic and exoplasmic faces of the mem-
water-soluble proteins as well as integral membrane proteins. brane (see Figures 10-14, 10-15, and 10-17).
As noted earlier, the exoplasmic and cytosolic domains of inte- • Fatty acyl side chains as well as the polar head groups of mem-
gral membrane proteins are generally hydrophilic and soluble brane lipids pack tightly and irregularly around the hydrophobic
in water. The membrane-spanning domains, however, are rich segments of integral membrane proteins (see Figure 10-16).
in hydrophobic and uncharged residues (sec Figure 10-14 ).
• The porins, unlike other integral proteins, contain membrane-
When separated from membranes, these exposed hydrophobic
spanning 13 sheets that form a barrel-like channel through the
segments tend to interact with one another, causing the protein
bilayer (see Figure 10- 18).
molecules to aggregate and precipitate from aqueous solutions.
The hydrophobic parts of nonionic detergent molecules prefer- • Long-chain lipids attached to certain amino acids anchor some
entiaUy bind to the hydrophobic segments of transmembrane proteins to one or the other membrane leaflet (see Figure 10-19).
proteins, preventing protein aggregation and allowing the pro- All transmembrane proteins and glycolipids are asymmetri-
teins to remain in the aqueous solution. Detergent-solubilized cally oriented in the bilayer. Invariably, carbohydrate chains
transmembrane proteins can then be purified by affinity chro- are present only on the exoplasmic surface of a glycoprotein
matography an~ other techniques used in purifying water- or glycolipid.
soluble proteins (see Chapter 3 ).
Many water-soluble enzymes (e.g., phospholipases) use
As discussed previously, most peripheral membrane pro-
phospholipids as their substrates and must bind to the mem-
teins are bound to specific transmembrane proteins or mem-
brane surface to carry out their function. Such bindi ng is often
brane phospholipids by ionic or other weak noncovalent
due to the attraction between positive charges on basic resi-
interactions. Generally, peripheral proteins can be removed
dues in the protein and negative charges on phospholipid
from the membrane by solutions of high ionic strength (high
head groups in the bilayer.
salt concentrations), which disrupt ionic bonds, or b) chem-
icals that bind divalent cations such as Mgh . Unlike integral Transmembrane proteins are selectively extracted (solubi-
proteins, most peripheral proteins arc soluble in aqueous so- lized) and purified with the use of nonionic detergents.
lution and need not be solubilized by nonionic detergents.

1 0.3 Phospholipids, Sphingolipids,

KEY CONCEPTS of s~("tion 1 0 .,
and Cholesterol: Synthesis
Membrane Proteins: Structure and Basic Functions
and Intracellular Movement
• Biological membranes usuaUy contain both integral (trans-
membrane) protein" and peripheral membrane proteins, In this section, we consider some of the special challenges
which do not enter the hydrophobic core of the bilayer (see that a cell faces in synthesizing and transporting lipids, which
Figure 10-1). are poorly soluble in the aqueous interior of cells. The focus
of our discussion will be the biosynthesis and movement of
• Most integral membrane proteins contain one or more
membrane-spanning hydrophobic a helices bracketed by hy- the major lipids found in cellular membranes-phospholipids,
drophilic domains that extend into the aqueous environment sphingolipids, and cholesterol-and their precursors. In lipid
biosynthesis, water-soluble precursors are assembled into

464 CHAPTER 10 • Biomembrane Structure

membrane-associated intermediates that are then converted Acetyl CoA is an important intermediate in the metabolism
into membrane lipid products. The movement of lipids, espe- of glucose, fatty acids, and many amino acids, as detailed in
cially membrane components, between different organelles is Chapter 12. lt also contributes acetyl groups in many bio-
critical for maintaining the proper composition and proper- synthetic pathways. Saturated fatty acids (no carbon-carbon
ties of membranes and overall cell structure. double bonds) containing 14 or 16 carbon atoms are made
A fundamental principle of membrane biosynthesis is that from acetyl CoA by two enzymes, acetyl-CoA carboxylase and
cells synthesize new membranes only by the expansion of exist- fatty acid synthase. In animal cells, these enzymes are found
ing membranes. (The one exception may be autophagy, where in the cytosol; in plants, they are found in chloroplasts. Pal-
new membrane is formed first through the formation of an mitoyl CoA (16 carbon fatty acyl group linked to CoA) can
autophagic crescent, the construction of which involves modi- be elongated to 18-24 carbons by the sequential addition of
fication of phosphatidylethanolamine with the ubiquitin-like two-carbon units in the endoplasmic reticulum (ER) or some-
modifier Atg8 Isee Figure 14-351.) Although some early steps times in the mitochondrion. Desaturase enzymes, also located
in the synthesis of membrane lipids take place in the cyto- in the ER, introduce double bonds at specific positions in
plasm, the final steps are catalyzed by enzymes bound to pre- some fatty acids, yielding unsaturated fatty acids. Oleyl CoA
existing cellular membranes, and the products are incorporated (oleate linked to CoA, see Table 2-4), for example, is formed
into the membranes as they are generated. Evidence for this by removal of two H atoms from stearyl CoA. In contrast to
phenomenon is seen when cells are briefly exposed to radioac- free fatty acids, fatty acyl CoA derivatives are soluble in
tive precursors (e.g., phosphate or fatty acids): all the phospho- aqueous solutions because of the hydrophilicity of the CoA
lipids and sphingolipids incorporating these precursor segment.
substances are associated with intracellular membranes; as ex-
pected from the hydrophobicity of the fatty acyl chains, none
Small Cytosolic Proteins Facilitate
are found free in the cytosol. After they are formed, membrane
lipids must be distributed appropriately both in leaflets of a Movement of Fatty Acids
given membrane and among the independent membranes of In order to be transported through the cell cytoplasm, free, or
different organelles in eukaryotic cells, as well as in the plasma unesterified, fatty acids (those unlinked to a CoA), commonly
membrane. Here, we consider how this precise lipid distribu- are bound by fatty-acid-binding proteins (FABPs), which be-
tion is accomplished; in Chapters 13 and 14 we discuss how long to a group of small cytosolic proteins that facilitate the
membrane proteins arc inserted into cell membranes and traf- intracellular movement of many lipids. These proteins con-
ficked to their appropriate location within the cell. tain a hydrophobic pocket lined by [3 sheets (Figure 10-24 ). A
long-chain fatty acid can fit into this pocket and interact non-
covalently with the surrounding protein.
Fatty Acids Are Assembled from Two-Carbon
The expression of cellular FABPs is regulated coordi-
Building Blocks by Several Important Enzymes nately with cellular requirements for the uptake and release
Fatty acids (Chapter 2) play a number of important roles in of fatty acids. Thus FABP levels are high in active muscles
cells. In addition ro being a cellular fuel source (see discussion that are using fatty acids for generation of ATP, and in adi-
of aerobic oxidation in Chapter 12 ), fatty acids are key com- pocytes (fat-storing cells) when they are either taking up
ponents of both the phospholipids and sphingolipids making fatty acids to be stored as triglycerides or releasing fatty
up cell membranes; they also anchor some proteins to cellular acids for use by other cells. The importance of FABPs in fatry
membranes (see Figure 10-19). Thus the regulation of fatty acid metabolism is highlighted by the observations that they
acid synthesis plays a key role in the regulation of membrane can compose as much as 5 percent of all cytosolic proteins in
synthesis as a whole. The major fatty acids in phospholipids the liver and that genetic inactivation of cardiac muscle
contain 14, 16, 18, or 20 carbon atoms and include both satu- FABP converts the heart from a muscle that primanly burns
rated and unsaturated chains. The fatty acyl chains found on fatty acids for energy into one that primarily burns glucose.
sphingolipids can be longer than those in the phosphoglycer-
ides, containing up to 26 carbon atoms, and may bear other
Fatty Acids Are Incorporated into Phospholipids
chemical modifications (e.g., hydroxylation) as well.
Fatty acids are synthesized from the two-carbon building Primarily on the ER Membrane
block acetate, CH3 COO . In cells, both acetate and the in- Fatty acids are not directly incorporated into phospholipids;
termediates in fatty acid biosynthesis are esterified to the rather, in eukaryotic cells they are first converted into CoA
large water-soluble molecule coenzyme A (CoA), as exempli- esters. The subsequent synthesis of phospholipids such as the
fied by the structure of acetyl CoA: phosphoglycerides is carried out hy emymes associated with

H CH 3 0 0
I II
C -C-CH 2 - 0 - P - 0 - P - O - Ribose - Adenine
I I I I I
OH CH3 0 0 Phosphate
Coenzyme A (CoA)

10.3 Phospholipids, Sphingolipids, and Cholesterol: Synthesis and Intracellular Movement 465
FIGURE 10-24 Binding of a fatty acid to the hydrophobic pocket
of a fatty-acid-bindi ng protein (FABP). The crystal structure of
adipocyte FABP (ribbon diagram) reveals that the hydrophobic binding
pocket is generated from two 13 sheets that are nearly at right angles to
each other, forming a clam-shell-like structure. A fatty acid (carbons
yellow; oxygens red) interacts noncovalently with hydrophobic amino
acid residues within this pocket. [See A. Reese-Wagoner et al .. 1999, Biochim.
Biophys. Acta 23:1441(2-3):106--1 16.]

·.
the cytosolic face of the ER membrane, usually the smooth
ER, in animal cells; through a series of steps, fatty acyl CoAs,
glycerol 3-phosphatc, and polar head-group precursors are
linked together and then inserted into the ER membrane (Fig-
ure 10-25). The fact that these enzymes arc located on the
cytosolic side of the membrane means that there is an inherent
asymmetry in membrane biogenesis: new membranes are ini-
tially synthesized only on one leaflet-a fact with important (ceramide) also takes place in the ER. Later, in the Golgi, a
consequences for the asymmetric distribution of lipids in polar head group is added to ceramide yield ing sphingomy-
membrane leaflets. Once synthesized on the ER, phospholip- elin, whose head group is phosphorylcholine, and various
ids are transported to other organelles and to the plasma glycosphingolipids, in which the head group may be a mono-
membrane. Mitochondria synthesize some of their own mem- saccharide or a more complex oligosaccharide (see Figure
brane lipids and import others. 10-8b). Some sphingolipid synthesis can also take place in mi-
Sphingolipids are also synthesized indirectly from multiple tochondria. In addition to serving as the bacl<bone for sphin-
precursors. Sphingosine, the building block of these lipids, is golipids, ceramide and its metabolic products are important
made in the ER, beginning with the coupling of a palmitoyl signaling molecules that can influence cell growth, prolifera-
group from palmitoyl CoA to serine; the subsequent addi- tion, endocytosis, resistance to stress, and programmed cell
tion of a second fatty acyl group to form N-acyl sphingosine death (apoptosis).

Acetyl CoA

CoA~
E osolic e~zymes

.
~c - OH
II

?!
Fatty acid

~c - S - CoA Fatty acyl CoA

CMP
+
c-@

Cytosol .:2
o_
J--
(/) Q)
0-
>-<1l
ER u~

'] ·~
membrane

ro_
- a._
Q)
Lumen o-
GPAT LPAAT Flippase x ro
(acyl transferases) UJ~

Choline

FIGURE 10 -25 Phosp holipid synt hesis in ER membrane. Because bon chains anchor the molecule to the membrane. Step fJ:A phospha-
phospholipids are amphipathic molecules, the last stages of their tase converts phosphatidic acid into diacylglycerol. Step 10:A polar head
multistep synthesis take place at the interface between a membrane and group (e.g., phosphorylcholine) is transferred from cytosine diphospho-
the cytosol and are catalyzed by membrane-associated enzymes. Step 0 : choline (COP-choline) to the exposed hydroxyl group. Step 0 :Flippase
Two fatty acids from fatty acyl CoA are esterified to the phosphorylated proteins catalyze the movement of phospholipids from the cytosolic
glycerol backbone, forming phosphatidic acid, whose two long hydrocar- leaflet in which they are initially formed to the exoplasmic leaflet.

466 CHAPTER 10 • Biomembrane Structure

 After their synthesis is completed in the Golgi, sphingo- phospholipid asymmetry are not well understood, it is clear
lipids are transported to other cellular compartments that flippases play a key role. As described in Chapter 11,
through vesicle-mediated mechanisms similar to those for these integral membrane proteins use the energy of ATP
the transport of proteins, discussed in Chapter 14. Any type hydrolysis to facilitate the movement of phospholipid mole-
of vesicular transport resu lts in movement not only of the cules from one leaflet to the other (see Figure 11-15).
protein payload, but also of the lipids that compose the ve-
sicular membrane. In addition, phospholipids such as phos-
Cholesterol Is Synthesized by Enzymes
phoglycerides, as well as cholesterol, can move between
organelles by different mechanismc;, described below. in the Cytosol and ER Membrane
Next we focus on cholesterol, the principal sterol in animal
cells. Cholesterol is synthesized mainly in the liver. The first
Flippases Move Phospholipids from One
steps of cholesterol synthesis (Figure 10-26)-convcrsion of
Membrane Leaflet to the Opposite Leaflet three acetyl groups linked to CoA (acetyl CoA ) forming the
Even though phospholipids are initial ly incorporated into 6-carbon molecule 13-hydroxy-13-methylglutaryllinked to CoA
the cytosolic leaflet of the ER membrane, various phospho- (HMG-CoA)-take place in the cytosol. The conversion of
lipids are asymmetrically distributed in the two leaflets of HMG-CoA into mevalonate, the key rate-controlling step in
the ER membrane and of other cellular membranes. As cholesterol biosynthesis, is catalyzed by HMG-CoA reductase,
noted above, phospholipids do not readily flip-flop from one an ER integral membrane protein, even though both its sub-
leaflet to the other. For the ER membrane to expand b) strate and its product are water soluble. The water-soluble
growth of both leaflets and have asymmetrically distributed catalytic domain of HMG-CoA reductase extends into the cy-
phospholipids, its phospholipid components must be able to tosol, but its eight transmembrane a helices firmly embed the
move from one membrane leaflet to the other. Although the enzyme in the ER membrane. Five of the transmembrane a he-
mechanisms employed to generate and maintain membrane lices compose the so-called sterol-sensing domain and regulate

0 0 0
+
~S-CoA
Acetyl CoA acetoacetyl CoA
A s-CoA +

0 OH 0 1
O~S-CoA
HMG-CoA
HMG-CoA

0 OH l
reductase

O~OH
Mevalonate
!
!
~OPP /
lsopentenyl pyrophosphate ,
(IPP)
lsopentenyl adenosine
Many other isoprenoids

Dolichol

OPP
1 / ~~~~inone
Farnesyl pyrophosphate ;;;;: Vitamins (A, E, K)
Chlorophyll
Lipid-anchored proteins (Ras)

( 1
Squalene

Vitamin 0
Bile acids
Cholesterol E Steroid hormones
Cholesterol esters
HO Modified proteins (Hedgehog)
FIGURE 10-26 Cholesterol biosynthetic pathway. The regulated five-carbon isoprenoid structure. IPP can be converted into cholesterol
rate-cont rolling step in cholesterol biosynthesis is the conversion of and into many other lipids, often through the polyisoprenoid interme-
~-hydroxy-13-methylglutaryl CoA (HMG-CoA) into mevalonic acid by diates shown here. Some of the numerous compounds derived from
HMG-CoA reductase, an ER-membrane protein. Mevalonate is then isoprenoid intermediates and cholesterol itself are indicated.
converted into isopentenyl pyrophosphate (IPP), which has the basic

10.3 Phospholipids, Sphingolipids, and Cholesterol: Synthesis and Intracellular Movement 467
enzyme stability. When levels of cholesterol in the ER mem- this pathway, farnesyl pyrophosphate, is the precursor of the
brane are high, binding of cholesterol to this domain causes the prenyl lipid that anchors Ras and related proteins to the cy-
protein to bind to two other integral ER membrane proteins, tosolic surface of the plasma membrane (see Figure 1 0-19) as
Insig-1 and lnsig-2. This in turn induces ubiquitination (sec Fig- well as other important biomolecules (sec Figure 10-26 ).
ure 3-29) of HMG-CoA reductase and its degradation by the
proteasome pathway, reducing the production of mevalonate,
Cholesterol and Phospholipids Are Transported
the key intermediate in cholesterol biosynthesis.
Between Organelles by Several Mechanisms
Atherosclerosis, frequently called cholesterol-dependent As already noted, the final steps in the synthesis of cholesterol
clogging of the arteries, is characterized by the progres- and phospholipids take place primarily in the ER. Thus the
sive deposition of cholesterol and other lipids, cells, and extra- plasma membrane and the membranes bounding other organ-
cellular matrix material in the inner layer of the wall of an elles must obtain these lipids by means of one or more intra-
artery. The resulting distOrtion of the artery's wall can lead, cellular transport processes. Membrane lipids can and do
either alone or in combination with a blood clot, to major accompany both soluble and membrane proteins during the
blockage of blood flow. Atherosclerosis accounts for 75 percent secretory pathway described in Chapter 14; membrane vesi-
of deaths due to cardiovascular disease in the United States. cles bud from the ER and fuse with membranes in the Golgi
Perhaps the most successful anti-atherosclerosis medica- complex, and other membrane v~sicles bud from the Golgi
tions are the statins. These drugs bind to HMG-CoA reductase complex and fuse with the plasma membrane (Figure 10-2 7a).
and directly inhibit its activity, thereby lowering cholesterol However, several lines of evidence suggest that there is sub-
biosynthesis. As a consequence, the amount of low-density stantial inter-organelle movement of cholesterol and phospho-
lipoproteins (see Figure 14-27)-the small, membrane-en- lipids through other mechanisms. For example, chemical
veloped particles containing cholesterol esterified to fatty inhibitors of the classic secretory pathway and mutations that
acids that often and rightly are called "bad cholesterol"- impede vesicular traffic in this pathway do not prevent choles-
drops in the blood, reducing the formation of atheroscle- terol or phospholipid transport between membranes.
rotic plaques. • A second mechanism entails direct protein-mediated contact
of ER or ER-derived membranes with membranes of other or-
Mevalonate, the six-carbon product formed by HMG- ganelles (Figure l0-27b). In the third mechanism, small lipid-
CoA reductase, is converted in several steps into the five-carbon transfer proteins facilitate the exchange of phospholipids or
isoprenoid compound isopentenyl pyrophosphate (IPP) and cholesterol between different membranes (Figure 1 0-27c). Al-
its stereoisomer, dimethylallyl pyrophosphate (DMPP) (see though such transfer proteins have been identified in assays in
Figure 10-26). These reactions are catalyzed by cytosolic vitro, their role in intracellular movements of most phospholip-
enzymes, as are the subsequent reactions in the cholesterol ids is not well defined. For instance, mice with a knockout muta-
synthesis pathway, in which six IPP units condense to yield tion in the gene encoding the phosphatidylcholine-transfer
squalene, a branched-chain 30-carbon intermediate. Enzymes protein appear to be normal in most respects, indicating that this
bound to the ER membrane catalyze the multiple reactions protein is not essential for cellular phospholipid metabolism.
that convert squalene into cholesterol in mammals or into As noted earlier, the lipid compositions of different or-
related sterols in other species. One of the intermediates in ganelle membranes vary considerably (see Table 10-1 ). Some

(a) (b) (c)

Binding
Phospholipid protein
\ Cholesterol OH

L__. ~ HO
HO.
/'0~

rr
HO OH J: HO
~
;>-<

Vesicle

Hypothetical Binding
proteins protein
Cytosol Cytosol Cytosol

FIGURE 10-27 Proposed mechanisms of transport of cholesterol by membrane-embedded proteins. In mechanism (c), transfer is
and phospholipids between membranes. In mechanism (a), vesicles mediated by small, soluble lipid-transfer proteins. [Adapted from
transfer lipids between membranes. In mechanism (b). lipid transfer is F. R. Maxfield and D. Wustner, 2002, J. Clin. Invest. 11 0:891.]
a consequence of direct contact between membranes that is mediated

' '
468 CHAPTER 10 • Biomembrane Structure
of these differences are due to different sites of synthesis. For
Perspectives for the Future
example, a phospholipid called cardiolipin, which is localized
to the mitochondrial membrane, is made only in mitochon- One fundamental question in lipid biology concerns the gen-
dria and little is transferred to other organelles. Differential eration, maintenance, and function of the asymmetric distri-
transport of lipids also plays a role in determining the lipid bution of lipids within the leaflets of one membrane and the
compositions of different cellular membranes. For instance, variation in lipid composition among the membranes of dif-
even though cholesterol is made in the ER, the cholesterol ferent organelles. What arc the mechanisms underlying this
concentration (cholcsrerol-ro-phospholipid molar ratio) is complexity, and why is such complexity needed? We already
~1.5-13 - fold higher in the plasma membrane than in other know that certain lipids can specifically interact with and
organelles (ER, Golgi, mitochondrion, lysosome). Although influence the activity of some proteins. For example, the
the mechanisms responsible for establishing and maintaining large multimeric proteins that participate in oxidative phos-
these differences are not well understood, we have seen that phorylation in the inner mitochondrial membrane appear to
the distinctive lipid composition of each membrane has a assemble into supcrcomplexcs whose stability may depend
major influence on its physical and biological properties. on the physical properties and binding of specialized phos-
pholipids such as cardiolipin (see Chapter 12).
The existence of lipid rafts in biological membranes and
their function in cell signaling remains a topic of heated de-
bate. Many biochemical studies using model membranes
KEY CONCEPTS of Section 1 0.3 show that stable lateral assemblies of sphingolipids and
Phospholipids, Sphingolipids, and Cholesterol: cholesterol-lipid rafts--<:an facilitate selective protein-protein
Synthesis and Intracellular Movement interactions by excluding or including specific proteins. But
whether or not lipid rafts exist in natural biological mem-
• Saturated and unsaturated fatty acids of various chain
branes, as well as their dimensions and dynamics, is under
lengths are components of phospholipids and sphingolipids.
intense investigation. New biophysical and microscopic
• Fatty acids are synthesized from acetyl CoA by water-soluble tools are beginning to provide a more solid basis for their
enzymes and modified by elongation and desaturarion in the existence, size, and behavior. ·
endoplasmic reticulum (ER). Despite considerable progress in our understanding of
• Free fatty acids are transported within cells by fatty-acid- the cellular metabolism and movement of lipids, the mecha-
binding proteins (FABPs). nisms for transporting cholesterol and phospholipids be-
tween organelle membranes remain poorly characterized. In
• fatty acids are incorporated into phospholipids through a
particular, we lack a detailed understanding of how various
multi-step process. The final steps in the synthesis of phos-
transport proteins move lipids from one membrane leaflet to
phoglycerides and sphingolipids are catalyzed by membrane-
another (flippase activity) and into and out of cells. Such un-
associated enzymes primarily on the cyrosolic face of the ER
derstanding will require a determination of high-resolution
(sec Figure 10-25).
structures of these molecules, their capture in various stages
• Each type of newly synthesized lipid is incorporated into of the transport process, and careful kinetic and other bio-
the preexisting membranes on which it is made; rhus, mem- physical analyses of their function, similar to the approaches
branes are themselves the platform for the synthesis of new discussed in Chapter 11 for elucidating the operation of ion
membrane material. channels and ATP-powered pumps.
• Most membrane phospholipids are preferentially distrib- Recent advances in solubilizing and crystallizing integral
uted in either the exoplasmic or the cytosolic leaflet. This membrane proteins have led to the delineation of the mo-
asymmetry results in part from the action of phospholipid lecular structures of many important types of proteins, such
flippases, which flip lipids from one leaflet to the other. as ion channels, G protein-coupled receptors, ATP-powered
ion pumps, and aquaporins, as we will see in Chapter 1 l.
• The initial steps in cholesterol bios} nthesis take place in
However, many important classes of membrane proteins
the cytosol, whereas the last steps are catalyzed by enzymes
have proven recalcitrant to even these new approaches. For
associated with the ER membrane.
example, we lack the structure of any protein that transports
• The rate-controlling step in cholesterol biosynthesis is cata- glucose into a eukaryotic cell. As we will learn in Chapters
lyzed by HMG-CoA reductase, whose transmembrane segments 15 and 16, many classes of receptors span the plasma mem-
are embedded in the ER membrane and contain a sterol-sensing brane with one or more a helixes. Perhaps surprisingly, we
domain. lack the molecular structure of the transmembrane segment
• Considerable evidence indicates that Golgi-independent of any single-pass eukaryotic cell-surface receptor, and so
vesicular transport, direct protein-mediated contacts be- many aspects of the function of these proteins arc still mys-
tween different membranes, soluble protein carriers, or all terious. The transmittal of information across the mem-
three may account for some inter-organelle transport of cho- brane, as it occurs when a single-pass receptor binds an
lesterol and phospholipids (see Figure 10-27) . appropriate ligand, remains to be described at adequate mo-
lecular resolution. Elucidating the molecular structures of

Perspectives for the Future 469

these and many other types of membrane proteins will clarify 7. ldentify the following membrane-associated proteins based
many aspects of molecular cell biology. on their structure: (a) tetramers of identical subunits, each with
six membrane-spanning o: helices; (b) trimers of identical sub-
units, each with 16 13 sheets forming a barrel-like structure.
Key Terms 8. Proteins may be bound to the exoplasmic or cytosolic
face of the plasma membrane by way of covalently attached
amphipathic 445 lipid raft 454 lipids. What arc the three types of lipid anchors responsible
aquaporin 458 liposome 445 for tethering proteins to the plasma membrane bilayer, and
atherosclerosis 468 lumen 448 which type is used by cell-surface proteins that face the ex-
cholesterol 450 membrane transport ternal medium and by glycosylated proteoglycans?
cilium 448 protein 443 9. Although both faces of a biomembranc are composed of
the same general types of macromolecules, principally lipids
cytoskeleton 443 micelle 445
and proteins, the two faces of the bilayer are not identical.
cytosolic face 447 peripheral membrane
What accounts for the asymmetry between the two faces?
exoplasmic face 44 7 protein 456
10. What are detergents? How do ionic and nonionic deter-
flagellum 448 phosphoglyceride 448
gents differ in their ability to disrupf cell membrane structure?
flippase 454 phospholipase 454
11. What is the likely identity of these membrane-associated
glycolipid 450 phospholipid bilayer 445 proteins: (a) released from the membrane with a high-salt
glycoprotein 461 plasma membrane 443 solution causing disruption of ionic linkages; (b) not released
porin 458 from the membrane upon exposure to a high-salt solution
hydrophilic 445
receptor protein 443 alone, but released when incubated with an enzyme that
hydrophobic 445
saturated 465 cleaves phosphate-glycerol bonds and covalent linkages are
integral membrane disrupted; (c) not released from the membrane upon expo-
protein 456 sphingolipid 450
sure to a high-salt solution, but released after addition of the
lectin 461 statin 468 detergent sodium dodecyl sulfate (SDS). Will the activity of
lipid-anchored membrane sterol 450 the protein released in part (c) be preserved following release?
protein 456 unsaturated 465 12. Following the production of membrane extracts using
lipid droplet 455 the nonionic detergent Triton X-100, you analyze the mem-
brane lysates via mass spectrometry and note a high content
of cholesterol and sphingolipids. furthermore, biochemical
analysis of the lysates reveals potential kinase activity. What
Review the Concepts have you likely isolated?
1. When viewed by electron microscopy, the lipid bilayer is 13. Phospholipid biosynthesis at the interface between the
often described as looking like a railroad track. Explain how endoplasmic reticulum (ER) and the cytosol presents a number
the structure of the bilayer creates this image. of challenges that must be solved by the cell. Explain how each
2. Explain the following statement: The structure of all of the following is handled.
biomembranes depends on the chemical properties of phos- a. The substrates for phospholipid biosynthesis are all
pholipids, whereas the function of each specific biomem- water soluble, yet the end products are not.
brane depends on the specific proteins associated with that b. The immediate site of incorporation of all newly syn-
membrane. thesized phospholipids is the cytosolic leaflet of the ER
3. Biomembranes contain many different types of lipid mol- membrane, yet phospholipids must be incorporated into
ecules. What arc the three main types of lipid molecules both leaflets.
found in biomembranes? How are the three types similar, c. Many membrane systems in the cell, for example the
and how are they different? plasma membrane, are unable to synthesize their own phos-
4. Lipid bilayers are considered to be two-dimensional flu- pholipids, yet these membranes must also expand if the cell
ids. What does this mean? What drives the movement of is to grow and divide.
lipid molecules and proteins within the bilayer? How can 14. What are the common fatty acid chains in phosphoglyc-
such movement be measured? What factors affect the degree crides, and why do these fatty acid chains differ in their
of membrane fluidity? number of carbon atoms by multiples of 2?
5. Why are water-soluble substances unable to freely cross 15. Fatty acids must associate with lipid chaperones in order
the lipid bilayer of the cell membrane? How does the cell to move within the cell. Why are these chaperones needed,
overcome this permeability barrier? and what is the name given to a group of proteins that arc
6. Name the three groups into which membrane-associated responsible for this intracellular trafficking of fatty acids?
proteins may be classified. Explain the mechanism by which What is the key distinguishing feature of these proteins that
each group associates with a biomembrane. allows fatty acids to move within the cell?

470 CHAPTER 10 • Biomembrane Structure

 16. The biosynthesis of cholesterol is a highly regulated pro- The tracks generated during a 5-second observational period
cess. What is the key regulated enzyme in cholesterol biosyn- by a gold particle attached to XR present in a cell (left) or in
thesis? This enzyme is subject to feedback inhibition. What a liposome (middle) or to XR adhered to a microscope slide
is feedback inhibition? How does this enzyme sense choles- (right) are shown below.
terol levels in a cell?
17. Phospholipids and cholesterol must be transported from
their site of synthesis to various membrane systems within
cells. One way of doing this is through vesicular transport,
as is the case for many proteins in the secretory pathwa}
XR present XR present XR present on a
(Chapter 14). However, most phospholipid and cholesterol in a cell microscope slide
in a liposome
membrane-to-membrane transport in cells is nor by vesicular
transport. What is the evidence for this statement? What
appear to be the major mechanisms for phospholipid and What additional information do these data provide beyond
cholesterol transport?
what can be determined from the fRAP data?
18. Explain the mechanism by which statins lower "bad" c. Fluorescence resonance energy transfer (fRET) is a
cholesterol. technique by which a fluorescent molecule, following its ex-
citation with the appropriate wavelength of light, can trans-
.·
fer its emission energy to and excite a nearby different
fluorescent molecule (see Figure 15-18). Cyan fluorescent
Analyze the Data protein (CFP) and yellow fluorescent protein (YfP) are re-
1. The behavior of receptor X (XR), a transmembrane pro- lated to GFP but fluoresce at cyan and yellow wavelengths
tein present in the plasma membrane of mammalian cells, is rather than at green. If CFP is excited with the appropriate
being investigated. The protein has been engineered as a fu- wavelength of light and a YFP molecule is very near, then
sion protein containing the green fluorescent protein (GFP) energy can be transferred from CFP emission and used to
at its N-rerminus. GFP-XR is a functional protein and can excite YFP, as indicated by a loss of emission of cyan fluores-
replace XR in cells. cence and an increase in emission of yellow fluorescence.
a. Cells expressing GfP-XR or artificial lipid vesicles (lipo- CFP-XR and YFP-XR are expressed together in a cell line or
somes) containing GFP-XR are subjected to fluorescence are both incorporated into liposomes. The number of mole
recovery after photobleaching (FRAP). The intensity of the cu les of YFP-XR and CFP-XR per cm 2 of membrane is
fluorescence of a small spot on the surface of the cells (solid equivalent in the cells and the liposomes. The cells and lipo-
line) or on the su rface of the liposomes (dashed line) is mea- somes are then irradiated with a wavelength of light that
sured prior to and following laser bleaching (arrow). The data causes CFP but not YFP to fluoresce. The amount of cyan
are shown below. (CFP) and yellow (YFP) fluorescence emitted by the cells
(solid line) or liposomes (dashed line) is then monitored, as
~hown below.

5000

~ .. .......... ·······~;~rface of liposomes

'iii
c:
...,Q)
.. .. 20

.. ...
~
.!: 'iii
Q) c:
c.J
c:
Q) .. . c
Q)

... .
u
f/) Q)
Q) c.J
0 c:
:J
u::: .. Surface of cells
~
f/)
10

...
Q)

1000
0
:J
u:::
25 50 75 100
Time {s)
475 500 525
Wavelength {nm) of emitted fluorescent light
What explanation could at:cuunr for the differing behav10r of
GFP-XR in liposomes versus in the plasma membrane of a cell?
b. Tiny gold particles can be attached to individual mol- What can be deduced about XR from these data?
ecules and their movement then followed in a light micro- 2. After performing differential scanning calorimetry (a pro-
scope by single-particle tracking. This method allows one to cedure used to determine the transition temperature of a given
observe the behavior of individual proteins in a membrane. membrane by recording the amount of heat absorbed prior to

Analyze the Data 471

phase transition [solid to fluid statel) on membranes from Vance, D. l:.., and J. E. Vance. 2002. Brochemistry of Lrpids,
three different organisms, you obtain the following results: Upoproteins, and Membranes, 4rh ed. Elsevier.
Van Meer, G. 2006. Cellular lipidomics.l:MBO]. 24:3159-3165.
Yeager, P. L 2001. L1pids. Encyclopedra of Ufe Sciences.
~arure Publishmg Group.
Zimmerberg, J., and M. M. Kozlov. 2006. How proteins produce
cellular membrane cun·ature. Nature Ret'. Mol. Cell Bwl. 7:9-19.

Membrane Proteins: Structure and Basic Functions

Bowie,]. 2005. Solving rhe membrane protein folding problem.
1
c
Nature 438:581-589.
0 Cullen, P. J., (,_ E. Co£Jer, G. Banrmg, and H. Mellor. 2001.
·g Modular phosphoinosiride-binding domains: rheir role in ~ignalling
0 and membrane trafficking. Curr. Bioi. ll:R882-R893.
"'
.I:l
Engelman, D . .\lembranes are more mosa1c than flmd. 2005.
...."' Nature 438:578-580 .
"'
Q)
I Lany1, J. K., and H. Luecke. 2001. Bacreriorhodopsin. Curr.
Oprn. Struc. Bioi. 11:415-519.
Lee, A. G. 2005. A greasy grip. Nahtre 438:569-570.
MacKenzie, K. R., J. H. Presregard, and D. M. Engelman. 1997.
A rransmembrane helix dimer: structure and implications. Science
276:131-133.
Mclmosh, T. J., and S. A. Simon. 2006. Roles of hi layer
0 25 50 75 100 125 150 material properties in function and distribution of membrane
·c prorems Ann. Ret•. Biophys. Biomolec. Struct. 35:177-198.
Wucherpfennig, K.W., E. Gagnon, M .]. Call,- E. S. Huseby, and
M. E. Call. 2010. Cold Spring Harb Perspect Bioi, 2.:a005140.
Which of the following statements is likely true about the Schulz, G. E. 2000. (3-Barrcl membrane proteins. Curr. Opin.
lipid composition of membrane C? Struc. Bra/. 10:443-447.
i. It has high levels of saturated hydrocarbons and long
hydrocarbon tails compared to A and B. Phospholipids, Sphingolipids, and Cholesterol: Synthesis
ii. It has high levels of saturated hydrocarbons and short and Intracellular Movement
hydrocarbon tails compared to A and B. Bloch, K. 1965. The biological symhesis of cholesrerol. Science
iii. It has high levels of unsaturated hydrocarbons and 150:19-28.
Daleke, D. L., and j. V. Lyles. 2000. Idenrificauon and
long hydrocarbon tails compared to A and B.
purificauon of aminophospholipid fltppases. Brochim. Brophys.
iv. It has high levels of unsaturated hydrocarbons and Acta 1486:108-127.
short hydrocarb.on tails compared to A and B. Furerman, A., and H. Riezman. 2005. The ins and ours of
sphingohp1d synrhcsis. Trends Cell Rrol. 15:312-318.
HaJri, T., and N. A. Abumrad. 2002. Farry actd transport across
membranes: relevance to nurririon and metabolic pathology. A1111.
References Rei'. Nutr. 22:383-415.
Henneberry, A. 1.., M ..'v1. Wnghr, and C. R . .\lcMasrcr. 2002.
The Lipid Bilayer: Composition and Structural Organization The major sires of cellular phospholipid synthesis and molecular
McMahon, H., ,md J. L. Gallop. 2005. Membrane curvature dererminanrs of faery acid and lipid head group specificiry. Mol.
and mechamsms of dynamic cell membrane remodeling. Nature Bwl. Cell13:3148-3161.
438:590-596. Holthuis,j. C. M., and T. P. Levine. 2005. Lipid traffic: flopp}' ·.
Mukherjee, S., and F. R. Maxfield. 2004. Membrane domains. drives and a superhighway Nature Reu. Molec. Cell Bioi. 6:209-220.
Annu. Rei'. Cell Del'. Bwl. 20:839-866. Ioannou, Y. A. 2001. Mulndrug permeascs and subcellular
Ploegh, H. 2007. A lipid-based model for rhe crearion of an cholcsrerol transport. Nature Ret'. Mol. Cell Bwl. 2:657-668.
escape harch from rhc endoplasmic rericulum. N,zture 448:435-438. Kenr, C. 199). Eukaryoric pho~pholipid biosynthesis. Ann. Rev.
Simon~, K., and D. Toomre. 2000. Lipid rafrs and signal Biochem. 64:3 I S-343 .
rransducnon. Natllre Reu. Mol. Cell BICJI. 1:31-41. .\1axfield, F. R., and L Tabas. 2005. Role of cholesrerol and
Simons, K., and W. L C. Vaz. 2004. Model \ysrems, bp1d rafrs, lipid organizanon 111 disease. Nature 438:612-621.
and cell membranes. Annu. Reu. Biophys. Biomulec. Struct. Srahl, A., R. E. Gimcno, LA. Tarraglia, and H. F. Lodi~h.
33:269-2 9 5. 200 l. Farry ac1d rransporr proteins: a current v1ew of a growing
Tamm, I . K., V. K. Kiessling, ami M. L Wagner. 200 I. fatTIJI}. Trends l:ndocrmol. '>1etab. 12(6):266-273.
Membrane dynam1cs. F.ncyclopedw of Life Scrences. Narure van Meer, G., and H. Sprong. 2004. Membrane lipids and
Publishing Group. vesicular traffic. Curr. Opin. Cell Bioi. 16:373-378.

472 CHAPTER 10 • Biomembrane Structure

 CHAPTER

Transmembrane
Transport of Ions
and Small Molecules
Outside-in view of a bacterial aquaporin protein, which transports
water and glycerol into and out of the cell, embedded in a phospho-
lipid membrane. The four identical monomers are colored individually;
each has a channel in its center. [After M. 0. Jensen et al., 2002, Proc. Nat'/
..· Acad. Sci. USA 99:6731-6736.]

n al l cells, the p lasma membrane forms the permeability trations in the interi or of other organelles, Sl}Ch as the endo-

I barrier that separates the cytoplasm from the exterior en-

vironment, thus defining a cell's physical and chemical
boundaries. By preventing the unimpeded movement of mol-
plasmic reticulum or the Golgi complex, than in the cytosol.
All cellular membranes, both plasma and organelle, con-
sist of a bilayer of phospholipids into which other lipids and
ecules and ions into and out of cells, the plasma membrane specific types of proteins are embedded. It is this combina-
maintains essential differences between the composition of tion of lipids and proteins that gives cellular membranes
the extracellular fluid and that of the cytosol; for example, their distinctive permeability qualities. If cellular membranes
the concentration of NaCI in the blood and extracellular flu- we re pure phospholipid bila yers (see Figure 10-4), they
ids of animals is genera lly above 150 mM, similar to that of would be excellent chemical barriers, impermeable to virtu-
the seawater in which a ll cells are thought to have evolved, ally all ions, amino acids, sugars, and other water-soluble
whereas the Na concentration in the cytosol is tenfold molecules. In fact, only a few gases and uncharged, small,
lower. In contrast, the potassium ion (K ) concentration is water-soluble molecules can readily diffuse across a pure
higher in the cytosol than outside. phospholipid bilayer (Figure 11-1 ). But cellular membranes
Organelle membranes, which separate the cytosol from must serve not only as barriers; they must a lso act as con-
the interior of the organelle, also fo rm permeability barriers. duits, selectively transporting molecules and ions from one
For example, the proton concentration in the lysosome inte- side of the membrane to the other. Energy-rich glucose, for
rio r, pH 5, is about 100-fold greater than that of the cytosol, example, must be imported into the cell, and wastes must be
and many specific metabolites accumulate at higher concen- shipped out.

OUTLINE

11.1 Overview of Transmembrane Transport 474 11.4 Nongated lon Channels and the Resting
Membrane Potential 495
11.2 Facilitated Transport of Glucose and Water 477
11.5 Cotransport by Symporters and Anti porters 502
11.3 AlP-Powered Pumps and the Intracellular Ionic
Environment 483 11.6 Transcellular Transport 508
Gases We begin our discussion of membrane transport proteins
C0 2, N2, 0 2 )I
Permeable
by rev iewing some of the general pr inciples of transport
across membranes and distinguishi ng between three major
Small
Ethanol )I classes of such proteins. In subsequent sectio ns, we describe
uncharged Permeable
polar the structure and operation of specific examples of each class
molecules and show how members of families of homo logous trans-
Water, Urea
port proteins have d ifferent pr operties that ena ble d iffe rent
cell types to function appropriately. We also explain how
both the pla~ma membrane and organellar mem branes con-
Large tain specific com binations of transport proteins that enable
uncharged cells to carry out essential physiological processes, including
polar
molecules Glucose, fructose the ma intenance of cytosolic pH , the accumulation of su-
crose and sa lts in plant cell vacuoles, and the di rected flow
of water in both plants and animals. The cell's resting mem-
Ions
K -. Mg2•, Ca2+, Cl , brane potential is an important consequence of selective ion
HC03 , HP04 2- transport across membranes, and we consider how this po-
tential arises. Epithelial cells, such as those li ning the small
Charged
intestine, use a combination of membrane proteins to trans-
polar Am ino acids, ATP, port ions, sugars, and other small molecules and water from
molecules glucose 6-phosphate, one side of the cell to the other. We will see how this under-
proteins, nucleic acids
standing has led to the development of sports drinks as well
as therapies fo r cholera.
FIGURE 11 - 1 Relative permeability of a pure phospholipid Note that in this chapter we cover only transpo rt of small
bilayer to various molecules and ions. A bilayer is permeable to
molecules and ions; transport of larger mo lecules, such as pro-
many gases and to small, uncharged, water- soluble (polar) molecules.
teins and oligosaccharides, is covered in Chapters 13 and 14.
It is slightly permeable to water, and essentially impermeable to ions
and to large polar molecules.

11.1 Overview of Transmembrane

Transport
Movement of virtually all small molecules and ions across
cell membranes is mediated by membrane transport proteins- In this section, we first describe the factors that influence the
integral membrane proteins embedded by multiple transmem- permeabil ity of lipid membranes, and then briefly describe
brane domains in cellular membranes. These membrane- the three major classes of membrane transport proteins that
spanning proteins act variously as shuttles, channels, or allow molecules and ions to cross them. Different kinds of
pumps for transporting molecules and ions through a mem- membrane-embedded proteins accomplish the task of mov-
brane's hydrophobic interior. In some cases, molecules or ions ing molecules and ions in different ways.
are transported from a higher to a lower concentration, a
thermodynamically favored process powered by an increase in
Only Gases and Small Uncharged Molecules
entropy. Examples include the transport of water or glucose
from the blood into most body cells. In other cases, molecules Cross Membranes by Simple Diffusion
or ions must be pumped from a lower to a higher concentra- With its dense hydrophobic core, a phospholipid bilayer is
tion, a thermodynamically unfavorable process that can only largely impermeable to water-soluble mo lecules and ions.
occur when an external source of energy is available to push Only gases, such as 0 2 and C02 , and small uncharged polar
the molecules "uphill" against a concentration gradient. An molecules, such as urea and ethanol, can readily move by sim-
example is the cell's ability to concentrate protons within ly- ple diffusion across an artificial membrane composed of pure
sosomes to generate a low pH in the lumen. Often the re- phospholipid or of phospholipid and cholesterol (see Figure
quired energy is provided by mechanistically coupling the 11-1 ). Such molecules also can diffuse across cell ular mem-
energy-releasing hydrolysis of the terminal phosphoanhydride branes without the aid of transport proteins. No metabolic
bond in ATP with the movement of a molecule or ion across energy is expended because movement is from a high to a low
the membrane. Other proteins couple the movement of one concentration of the molecule, down its chemical concentra-
molecule or ion against its concentration gradient with the tion gradient. As noted in Chapter 2, such movements are
movement of another down its gradient, using the energy re- spontaneous because they have a positive 6.5 value (increase in
leased by the downhill movement of one molecule or ion to entropy) and th us a negative ~G (decrease in free energy).
thermodynamically drive the uphill movement of another. The relative diffusion rate of any substance across a pure
Proper functioning of any cell relies on a precise balance be- phospholipid bilayer is proportional to its concentration gra-
tween such import and export of various molecules and ions. dient across the bilayer and to irs hydrophobicity and size;

474 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

the movement of charged molecules is also affected by any If a substance carries a net charge, its movement across a
electric potentia I across the membrane. When a pure phos- membrane is influenced by both its concentration gradient
pholipid bilayer separates two aqueous spaces, or "compart- and the membrane potential, the electric potential (voltage)
ments," membrane permeability can be easily determined by across the membrane. The combination of these two forces,
adding a small amount of radioactive material to one com- called the electrochemical gradient, determines the energeti-
partment and measuring its rate of appearance in the other cally favorable direction of movement of a charged molecule
compartment. The greater the concentration gradient of the across a membrane. The electric potential that exists across
substance, the faster its rate of movement across a bilayer. most cellular membranes results from a small imbalance 10
The hydrophobicity of a substance is determined by mea- the concentration of positively and negatively charged ion:.
suring its partition coefficient K, the equilibrium constant for on the two sides of the membrane. We discuss how this ionic
its partition between oil and water. The higher a substance's imbalance, and resulting potential, arise and are maintained
partition coefficient (the greater the fraction found in oil rela- in Sections 11.4 and 11.5.
tive to water), the more lipid soluble it is and, therefore, the
faster its rate of movement across a bilayer. The first and Three Main Classes of Membrane
rate-limiting step in transport by simple diffusion is move-
ment of a molecule from the aqueous solution into the hydro- Proteins Transport Molecules and
phobic interior of the phospholipid bilayer, which resembles Ions Across Biomembranes
olive oil in its chemical properties. This is the reason that the As is evident from Figure 11-1, very few molecules and no
more hydrophobic a molecule is, the faster it diffuses across ions can cross a pure phospholipid bilayer at appreciable
a pure phospholipid bilayer. For example, diethylurea, with rates by simple diffusion. Thus transport of most molecules
an ethyl group attached to each nitrogen atom: into and out of cells requires the assistance of specialized
0 membrane proteins. Even in the case of molecules with rela-
II tively large partition coefficients (e.g., urea, fatty acids), and
CH 3 -CH 2 -NH-C - NH - CH 2 - CH 3
certain gases such as CO1 (carbon dioxide) and NH1 (am-
has a K of 0.01, whereas urea monia), transport is frequently accelerated hy specific pro-
0
teins because simple diffusion usually docs not occur rapidly
II enough to meet cellular needs.
NH 2 -C - NH 2
All transport proteins are transmembrane proteins con-
has a K of 0.0002. Diethylurea, which is 50 times (0.0 110.0002) taining multiple membrane-spanning segments that gener-
more hydrophobic than urea, will therefore diffuse through ally are a helices. By forming a protein-lined pathway across
phospholipid bilayer membranes about 50 times faster than the membrane, transport proteins are thought to allow
urea. Similarly, fatty acids with longer hydrocarbon chains movement of hydrophilic substances without their coming
are more hydrophobic than those with shorter chains and at into contact with the hydrophobic interior of the membrane.
all concentrations will diffuse more rapidly across a pure Here we introduce the three main types of membrane trans-
phospholipid bilayer. port proteins covered in this chapter (Figure 11-2 ).

D IJ
ATP-p owered pumps lon channels Tran sporters
( 1oo-1 03 ions/s) (107-108 ions/s) (10 2- 104 molecules/s)

Exterior
-..,
• • • •
I
Cytosol

ATP ADP+P; Uniporter Symporter

• • •
Anti porter

II
FIGURE 11 - 2 Overview of membrane transport protein s. transport a single type of molecule down its concentration gradient
Gradients are indicated by triangles with the tip pointing toward lower El,;,'!. Cotransport proteins (symporters, D}], and anti porters, ml catalyze
concentration, electric potential, or both. 0 Pumps utilize the energy the movement of one molecule against its concentration gradient
released by ATP hydrolysis to power movement of specific ions or small (black circles), driven by movement of one or more ions down an
molecules (red circles) against their electrochemical gradient. electrochemical gradient (red circles). Differences in the mechanisms of
fJ Channels permit movement of specific ions (or water) down their transport by these three major classes of proteins account for their
electrochemical gradient. ID Transporters, which fall into three groups, varying rates of solute movement.
facilitate movement of specific small molecules or ions. Uniporters

11 .1 Overview of Transmembrane Transport 475

 ATP-powered pumps (or simply pumps) are ATPases that
use the energy of ATP hydrolysis to move ions or small mol- Exterior
ecules across a membrane against a chemical concentration
gradient, an electric potential, or both. This process, referred Plasma
to as active transport, is an example of a coupled chemical membrane
reaction (Chapter 2 ). In this case, transport of ions or !>mall Cytosol
molecules "uphill" against an electrochemical gradient, which
requires energy, is coupled to the hydrolysis of ATP, which ADP + P;
releases energy. The overall reaction ATP hydrolysis and the
"uphill" movement of tons or small molecules-is energeti-
cally favorable. K+ Na+/lysine
Channels transport water, specific ions, or hydrophilic channel symporter
small molecules across membranes down their concentration FIGURE 11 -3 Multiple membrane transport proteins function
or electric potential gradients. Because this process requires together in the plasma membrane of metazoan cells. Gradients are
transport proteins but not energy, it is sometimes referred to indicated by triangles with the tip pointing toward lower concentra-
as "passive transport" or "facilitated diffusion," but it is tion. The Na /K ATPase in the plasma membrane uses energy
more properly called facilitated transport. Channels form a released by ATP hydrolysis to pump Na QUt of the cel l and K+ inward;
hydrophilic "tube'' or passageway across the membrane this creates a concentration gradient of Na that is greater outside the
through which multiple water molecules or ions move simul- cell than inside and one of K that is greater inside than outside.
taneously, single file, at a very rapid rate. Some channels are Movement of positively charged K + ions out of the cell through
.I
membrane K+ channel proteins creates an electric potential across the
open much of the time; these are referred to as nongated
plasma membrane-the cytosolic face is negative with respect to the
channels. Most ion channels, however, open only in response
extracellular face. A Na +/lysine transporter, a typical sodium/amino
to specific chemical or electric signals. These are referred to
acid cotransporter, moves 2 Na ions together with one lysine from the
as gated channels because a protein "gate" alternatively extracellular medium into the cell. "Uphill" movement ofthe amino ~ .'
blocks the channel or moves out of the way to open the chan- acid is powered by "downhill" movement of Na + ions, powered both
nel (see Figure 11-2). Channels, like all transport proteins, by the outside-greater-than-inside Na · concentration gradient and by
are very selective for the type of molecule they transport. the negative potential on the inside of the cell membrane, which
Transporters (also called carriers) move a wide variety of attracts the positively charged Na ions. The ultimate source of the
ions and molecules across cell membranes, but at a much energy to power amino acid uptake comes from the ATP hydrolyzed by
slower rate than channels. Three types of transporters have the Na /K ATPase, since this pump creates both the Na ion
been identified. Uniporters transport a single type of molecule concentration gradient and, via the K channels, the membrane
down its concentration gradient. Glucose and amino acids potential, which together power influx of Na ions.
cross the plasma membrane into most mammalian cells with
the aid of uniporters. Collectively, channels and uniporters are
sometimes called facilitated transporters, indicating move- to the other side in a second conformation. Because each such
ment down a concentration or electrochemical gradient. cycle results in movement of only one (or a few) substrate mol-
In contrast, antiporters and symporters couple the move- ecules, these proteins are characterized by relatively slow rates
ment of one type of ion or molecule against its concentration of transport ranging from 10° to 10 4 ions or molecules per
gradient with the movement of one or more different ions second (see Figure 11-2). Most ion channels shuttle between a
down its concentration gradient, in the same (symporter) or closed state and an open state, but many ions can pass through
different (anti porter) directions. These proteins often are an open channel without any further conformational change.
called cotranstwrters, referring to their ability to transport For this reason, channels are characterized by very fast rates of
two or more different solutes simultaneously. transport, up to 108 ions per second.
Like ATP pumps, cotransporters mediate coupled reac- frequently, several different types of transport proteins
tions in \\·hich an energetically unfavorable reaction (i.e., uphill work in concert to achieve a physiological function. An ex-
movement of one type of molecule or ion ) is coupled to an ample is seen in Figure 11-3, where an ATPase pumps Na
energetically favorable reaction (i.e., the downhill movement out of the cell and K ions inward; this pump, which is
of another). Note, however, that the nature of the energy- found in virtually all metazoan cells, establishes the oppositely
sup plying reaction driving active transport by these two directed concentration gradients of Na and K ions across
classes of proteins differs. ATP pumps use energy from hy- the plasma membrane (relatively high concentrations of K
drolysis of ATP, whereas cotransporters use the energy inside and Na outside of cells) that are used to power the
stored in an electrochemical gradient. This latter process import of amino acids. The human genome encodes hun-
sometimes i~ referred to as secondary active transport. dreds of different types of transport proteins that use the
Conformational changes are essential to the function of all energy stored across the plasma membrane in the Na + con-
transport proteins. ATP-powered pumps and transporters un- centration gradient and its associated electric potential to
dergo a cycle of conformational change exposing a binding site transport a wide variety of molecules into cells against their
(or sites) to one side of the membrane in one conformation and concentration gradients.

476 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 TABLE11-1 Mechanisms for Transporting Ions and Small Molecules Across Cell Membranes

Property Simple Diffusion Facilitated Transport Active Transport Cotransport*

- - - - - - -
Requires specific
protein + + +

Solute transported
against its gradient + +
Coupled to ATP
hydrolysis +
Driven by movement
of a cotransported ion
down its gradient +
Examples of molecules 02, C02, steroid Glucose and Ions, small hydrophilic Glucose and amino acids
transported hormones, many amino acids (uniporters); molecules, lipids (symporters); vanous tons
drugs ions and water (channels) (ATP-powered pumps) and sucrose (antiporters)

*Abo called secoudary actwe transport.

Table l l-1 summarizes the four mechanisms by w hich

• Transporters fall into three groups: Uniporters transport a
small molecules a nd ions a re transported across cellular
molecule down its concentration gradient (facilitated trans-
membranes. In th e next section we consider some of the sim-
port); symporters and antiporters couple movement of a
plest membrane transport proteins, those responsible for the
substrate against its concentration gradient to the movement
transport of glucose and water.
of a second substrate down its concentration gradient, a pro-
cess known as seconda ry active t ransport o r cotransport (see
Table 11-1).
KEY CONCEPTS of Section 11.1
Conformational changes are essential to the function of all
Overview of Transmembrane Transport membrane transport proteins; speed of transport depends on the
• Cellular membranes regulate the traffic of molecules and number of substrates tha t can pass through a protein at once.
ions into and out of cells a nd their organelles. The rate of
·. simple diffusion of a substance across a membrane is pro-
portional to its concentration gradient and hyd rophobicity.
• With the exception of gases (e.g., 0 2 a nd C02 ) and small,
uncharged, water-soluble molecules, most mo lecules cannot 11.2 Facilitated Transport of Glucose
diffuse across a pure phospholipid bilayer at rates sufficient
to meet cellular needs.
and Water
• Membrane transport proteins provide a hydrophilic pas- Most anima l cells utilize glucose as a substrate for ATP pro-
sageway for molecules a nd ions to travel through the hydro- duction; they usually employ a glucose uniporter to take up
phobic interi or of a membrane. glucose from the blood or other extracellular fluid. Many
cells utilize c hannel-like membrane transport proteins called
Three classes of transmembrane proteins mediate transport aquaporins to increase the rate of wa ter movement across
of ions, suga rs, am ino acids, and other metabolites across cell their surface membranes. Here, we discuss the structure and
membranes: ATP-powered pumps, channels, and transporters function of these and other faci litated transporters.
(see Figure 11-2).
• ATP-powered pumps couple t he movement of a substrate
against its concentration gradient to ATP hydrolysts, a pro-
cess known as active transport.
Uniport Transport Is Faster and More
Specific than Simple Diffusion
Channels form a hydrophilic "tube" thro ugh which water
o r ions move down a concentration gradient, a process The protein-mediated transport of a single type of molecu le,
known as facilitated transport or facilitated diffusion. such as glucose or other small hydrophilic molecules, down
a concentration gradient across a cellula r membrane is

11.2 Facilitated Transport of Glucose and Water 477

known as uniport. Several features distinguish uniport from
simple diffusion:

1. The rate of substrate movement by uniportcrs is far

higher than simple diffusion through a pure phospholipid GLUT1 (erythrocytes)

bilayer.
2. Because the transported molecule never enters the
hydrophobic core of the phospholipid bilayer, its partition
coefficient K IS Irrelevant.
3. Transport occurs via a limited number of uniporter
molecules. Consequently, there is a maximum transport
rate Vmax that depends on the number of uniporters in the
membrane. v max is achieved when the concentration 0 3 4 5 6 7 8 9 10 11 12 13 14
gradient across the membrane is very large and each External concentration of glucose (mM)
uniporter is working at its maximal rate.
EXPERIMENTAL FIGURE 11-4 Cellular uptake of glucose
4. Transport is reversible, and the direction of transport will mediated by GLUT proteins exhibits ;imple enzyme kinetics. The
change if the direction of the concentration gradient changes. initial rate of glucose uptake, u (measured as micromoles per milliliter
of cells per hour), in the first few seconds is plotted as a percentage of
5. Transport is specific. Each uniporter transports only a
the maximum rate, Vmax• against increasing glucose concentration in
single type of molecule or a single group of closely related
the extracellular medium. In this experiment, the initial concentration
molecules. A measure of the affinity of a transporter for ItS of glucose in the cells is always zero. Both GLUTl, expressed by erythro-
substrate IS Km, which is the concentration of substrate at cytes, and GLUT2, expressed by liver cells, catalyze glucose uptake
which transport is half maximal. (burgundy and tan curves). Like enzyme-catalyzed 'reactions, GLUT-
facilitated uptake of glucose exhibits a maximum rate Wmaxl· The Km is
These properties also apply to transport mediated by the
the concentration at which the rate of glucose uptake is half maximal.
other classes of proteins depicted in Figure 11-2. GLUT2, with a Km of about 20 mM (not shown), has a much lower
One of the best-understood uniporters is the glucose affinity for glucose than GLUTl, with a Km of about 1.5 mM.
transporter called GLUT1 found in the plasma membrane of
most mammalian cells. GLUTl is especially abundant in the
erythrocyte plasma membrane. Because these cells have a
single membrane and no nucleus or other internal organelles into the cell. Figure 11-5 depicts the sequence of events oc-
(see Figure 10-7a), the properties of GLUT I and many other curring during the unidirectiona l transport of glucose from
transport proteins from mature erythrocytes have been ex- the cell exterior inward to the cytosol. GLUT I also can cata-
tensively studied. The simplified structure of these cells lyze the net export of glucose from the cytosol to the extra-
makes isolating and purifying a transport protein a fairly cellular medium, when the glucose concentration is higher
straightforward procedure. inside the cell than outside.
Figure 11-4 shows that glucose uptake by erythrocytes The kinetics of the unidirectiona l transport of g lucose
and liver cells exhibits kinetics similar to that of a simple from the outside of a cell inward via GLUT1 can be de-
enzyme-catalyzed reaction involving a single substrate. The scribed by the same type of equation used to describe a sim-
kinetics of transport reactions mediated by other types of ple enzyme-catalyzed chemical reaction. For simplicity, let's
proteins are more complicated than those for uniporters. assume that the substrate glucose, S, is present initially only
Nonetheless, all protein-assisted (facilitated) transport reac- on the outside of the cell; this can be achieved by first incu-
tions occur faster than simple diffusion across the bi layer, bating the cells in a medium lacking glucose so their internal
are substrate-specific, and exhibit a maximal rate ( V maxl· stores are depleted. In this case, we can write

Km Vmax

The Low Km of the GLUT1 Uniporter Enables It Sout + GLUTl ~ Sout - GLUTl ~ S'" + GLUTl
to Transport Glucose into Most
where Sam - GLUTl represents GLUTl in the outward-facing
Mammalian Cells conformation with a bound glucose. This equation is similar
Like other uniporters, GLUTl alternates between two con- to the one describing the path of a simple enzyme-catalyL.eJ
formational states: in one, a glucose-binding site faces the reaction where the protein binds a single substrate and then
outside of the cell; in the other, a glucose-binding site faces transforms it into a different molecule. Here, however, no
the cytosol. Since the glucose concentration usually is higher chemical modification occurs to the GLUTl-bound sugar;
in the extracellular medium (blood, in the case of erythro- rather, it is moved across a cellular membrane. Nonetheless,
cytes) than in the cell, the GLUTl uniporter generally cata- the kinetics of this transport reaction are similar to those of
lyzes the net import of glucose from the extracellular medium simple enzyme-catalyzed reactions, and we can use the same

478 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 L

Exterior GLUT1 e Glucose

.· Glucose

Cytosol
Outward-facing
conformation
D
~
fJ

"
Inward-facing
conformation
~
II

•
a
________,._

H
Outward-facing
conformation

FIGURE 11-5 Model of uniport t ransport by GLUT1.1n one transporter undergoes the reverse conformational change, regenerat-
conformation, the glucose-binding site faces outward; in the other, ing the outward-facing binding site (step EIJ. If the concentration of
the binding site faces inward. Binding of glucose to the outward-facing glucose is higher inside the cell than outside, the cycle will work in
site (step O Jtriggers a conformational change in the transporter such reverse (step EJ ~ step OJ, resulting in net movement of glucose out
that the binding site now faces inward toward the cytosol (step fJ J. of the cell. The actual conformational changes are probably smaller
Glucose then is released to the inside of the cell (step IJJ. Finally, the than those depicted here.

derivation as that of the Michaelis-Menten equation in Chap- GLUT1 accounts for 2 percent of the protein in the plasma
ter 3 to derive the following expression for v 0 , the initial membrane of erythrocytes. After glucose is transported into the
transport rate for S into the cell catalyzed by GLUTl: erythrocyte, it is rapidly phosphorylated, forming glucose
6-phosphate, which cannot leave the cell. Because this reaction,
the first step in the metabolism of glucose (see Figure 12-3), is
Vo = (11-1) rapid and occurs at a constant rate, the intracellular concentra-
Km
1 +- tion of glucose is kept low even when glucose is imported from
c the extracellular environment. Consequently ·the concentration
where Cis the concentration of Sour (initially, the concentra- gradient of glucose (outside greater than inside the cell) is main-
tion of SID = 0). vmm the rate of transport when all mole- tained sufficiently high to support continuous, rapid import of
cules of GLUT1 contain a bound S, occurs at an infinitely additional glucose molecules and provide sufficient glucose for
high Sour concentration. The lower the value of K01, the more cellular metabolism.
tightly the substrate binds to the transporter. Equation 11-1
describes the curve for glucose uptake by erythrocytes shown
The Human Genome Encodes a Family
in Figure 11-4 as well as similar curves for other uniporters.
For GLUTl in the human erythrocyte membrane, the K01 of Sugar-Transporting GLUT Proteins
for glucose transport is 1.5 mM. Thus when the extracellular The human genome encodes at least 14 highly homologous
glucose concent,ration is 1.5 mM, roughly half the GLUTl GLUT proteins, GLUT1-GLUT14, that are all thought to
transporters with outward-facing binding sites will have a contain 12 membrane-spanning a helices, suggesting that they
bound glucose and transport will occur at 50 percent of the evolved from a single ancestral transport protein. Although
maximal rate. Since blood glucose is normally 5 mM, the no three-dimensional structure of any GLUT protein is avail-
erythrocyte glucose transporter usually is functioning at 77 able, detailed biochemical studies on GLUTJ have shown that
percent of its maximal rate, as can be seen from Equation the amino acid residues in the transmembrane a helices are
11-1. The GLUTl transporter (or the very similar GLUT3 predominantly hydrophobic; several helices, however, bear
glucose transporter) is expressed by all body cells that need amino acid residues (e.g., serine, threonine, asparagine, and
to take up glucose from the blood continuously at high rates; glutamine) whose side chains can form hydrogen bonds with
the rate of glucose uptake by such cells will remain high re- the hydroxyl groups on glucose. These residues are thought to
gardless of small changes in the concentration of blood glu- form the inward-facing and outward-facing glucose-binding
cose, because the blood concentration remains much higher sites in the interior of the protein (see Figure 11-5).
than the Km and the intracellular glucose concentration is The structures of all GLUT isoforms are thought to be
kept low by metabolism. quite similar, and all transport sugars. Nonetheless, their
In addition to glucose, the isomeric sugars o-mannose differential expression in various cell types, the regulation
and o-galactose, which differ from o-glucose in tht>ir con- of the number of GLUT transporters un cell surfaces, and
.·
figuration at only one carbon atom, are transported by isoform-specific functional properties enable different body
GLUTl at measurable rates. However, the Km for glucose cells to regulate glucose metabolism differently and at the
(1.5 mM ) is much lower than it is for o-mannose (20 mM) same time allow a constant concentration of glucose in the
or o-galactose (30 mM). Thus GLUTl is quite specific, hav- blood to be maintained. For instance, GLUT3 is found in
ing a much higher affinity (indicated by a lower Kml for the neuronal cells of the brain. Neurons depend on a constant
normal substrate o-glucose than for other substrates. influx of glucose for metabolism, and the low Km of GLUT3

11.2 Facilitated Transport of Glucose and Water 479

for glucose (K 111 = 1.5 mM), like that of GLUT I, ensures that movement of substrates across membranes) can be studied
these cells incorporate glucose from brain extracellular fluids only when they are associated with a membrane. Thus, the
at a high and constant rate. purified protein is usually reincorporated into pure phospho-
GLUT2, expressed in liver cells and the insulin-secreting lipid bilayer membranes, such as liposomes (see Figure 10-3),
islet 13 cells of the pancreas, has a Km of -20 mM, a bout 13 across which substrate transport can be readily measured.
times higher than the K 111 of GLUT I. As a result, when blood One good source of GLUT! is erythrocyte membranes. An-
glucose rises after a meal from its basal level of 5 mM to 10 other is recombinant cultured mammalian cells expressing a
mM or so, the rate of glucose influx will almost double in GLUTl transgene, often one expressing a modified GLUTl
GLUT2-expressing cells, whereas it will increase only slightly that contains an epitope tag (a portion of a molecule to
in GLUT1-expressing cells (see Figure 11-4). ln liver, the which a monoclonal antibody [see Chapter 9] can bind)
"excess" glucose brought into the cell is stored as the poly- fused to its N- or C-terminus. All of the integral proteins in
mer glycogen. In islet 13 cells, the rise in glucose triggers se- either of these two types of cells can be solubilized by a non-
cretion of the hormone insulin (see Figure 16-38), which in ionic detergent, such as octylglucoside. The glucose uni-
turn lowers blood glucose by increasing glucose uptake and porter GLUTl can be purified from the solubilized mixture
metabolism in muscle and by inhibiting glucose production by antibody affinity chromatography (Chapter 3) on a col-
in liver (see Figure 15-38). Indeed, cell-specific inactivation umn containing either a GLUTl-specific monoclonal anti-
of GLUT2 in pancreatic 13-cells prevents glucose-stimulated body or an antibody specific for the epitope tag, and then
insulin secretion and in liver cells (hepatocytes) disrupts the incorporated into liposomes made of pure phospholipids.
regulated expression of glucose-sensitive genes. Alternatively, the gene encoding a specific transport pro-
Another GLUT isoform, GLUT4, is expressed only in fat tein can be expressed at high levels in a cell type that nor-
and muscle cells, the cells that respond to insulin by increas- mally does not express it. The difference in transport of a
ing their uptake of glucose, thereby removing glucose from substance by the transfected and by control nontransfected
the blood. In the absence of insulin, GLUT4 resides in intra- cells will be due to the expressed transport protein. In these
cellular membranes, not the plasma membrane, and is unable systems, the functional properties of the various membrane
to facilitate glucose uptake from the extracellular fluid. By a proteins can be examined without ambiguity caused, for in-
process detailed in Figure 16-39, insulin causes these GLUT4- stance, by partial protein denaturation during isolation and
rich internal membranes to fuse with the plasma membrane, purification procedures. As an example, overexpressing
increasing the number of GLUT4 molecules present on the GLUTl in lines of cultured fibroblasts increases severalfold
cell surface and thus the rate of glucose uptake. This is one their rate of uptake of glucose, and expression of mutant
principal mechanism by which insulin lowers blood glucose; GLUTl proteins with specific amino acid alterations can
defects m the movement of GLUT4 to the plasma membrane identify residues important for substrate binding. .._
is one of the causes of adult onset, or type II, diabetes, a dis-
ease marked by continuously high blood glucose.
Osmotic Pressure Causes Water to Move
GLUTS is the only GLUT protein with a high specificity
(preference) for fructose; its principal site of expression is the Across Membranes
apical membrane of intestinal epithelial cells, where it trans- Movement of water into and our of cells is an important feature
ports dietary fructose from the intestinal lumen to inside of the life of microorganisms, plants, and animals. Aquaporins
the cells. are a family of membrane proteins that allow water and a few
other small uncharged molecules, such as glycerol, to cross bio-
membranes efficiently. But before discussing these transport
Transport Proteins Can Be Studied Using
proteins, we need to review osmosis, the force that powers the
Artificial Membranes and Recombinant Cells movement of water across membranes.
There are a variety of approaches for studying the intrinsic Water spontaneously moves "downhill" across a semi-
properties of transport proteins, such as defining the Vmax permeable membrane from a solution of low solute concen-
and K111 parameters and identifying key residues responsible tration (relatively higher water concentration) to one of high
for binding. Most cellular membranes contain many differ- solute concentration (relatively lower water concentration),
ent types of transport proteins but a relatively low concen- a process termed osmosis, or osmotic flow. In effect, osmosis
tration of any particular one, making functional studies of a is equivalent to "diffusion" of water across the semiperme-
single protein difficult. To facilitate such studies, researchers able membrane. Osmotic pressure is defined as the hydro-
use two approaches for enriching a transport protein of in- static pressure required to stop the net flow of water across
n: rt:M so that it predominates in the membrane: purification a membrane separating solutions of d1fferent water concen-
and reconstitution into artificial membranes, and overex- trations (Figure 11-6). In other words, the osmotic pressure
pression in recombinant cells. balances the entropy-driven thermodynamic force of the
In the first approach, a specific transport protein is ex- water concentration gradient. In this context, the "mem-
tracted from its membrane with detergent and purified. Al- brane" may be a layer of cells or a plasma membrane that is
though transport proteins can be isolated from membranes permeable to water but not to the solutes. The osmotic pressure
and purified, their functional properties (i.e., their role in the is directly proportional ro the difference in the concentration

480 CHAPTER 11 • Transmembrane Transport of Ion s and Small Molecules

 sugars and salts) usually is higher in the vacuole (see Figure
9-32) than in the cytosol, which in turn has a higher solute
concentration than the extracellular space. The osmotic pres-
sure, called turgor pressure, generated from the entry of water
into the cytosol and then into the vacuole pushes the cytosol
Solution A Solution B and the plasma membrane against the resistant cell wall. Plant
CA Ca cells can harness this pressure to help them stand upright, and
also grow. Cell elongation during growth occurs by a hor-
mone-induced localized loosening of a defined region of the
cell wall, followed by influx of water into the vacuole, increas-
ing its size and thus the size of the cell. •
Water flow if C8> CA
Although most protozoans (like animal cells) do not have
FIGURE 11-6 Osmotic pressure. Solutions A and Bare separated by a rigid cell wall, many contain a contractile vacuole that per-
a membrane that is permeable to water but impermeable to all solutes. mits them to avoid osmotic lysis. A contractile vacuole takes
If C8 (the total concentration of solutes in solution B) is greater than CA. up water from the cytosol and, unlike a plant vacuole, peri-
water will tend to flow across the membrane from solution A to odically discharges its contents through fusion with the
solution B. The osmotic pressure" between the solutions is the
plasma membrane. Thus even though water continuously
hydrostatic pressure that would have to be applied to solution B to
enters the protozoan cell by osmotic flow, the contractile
prevent this water flow. From the van't Hoff equation, osmotic pressure
vacuole prevents too much water from accumulating in the
is given by" = RT(C8 - CA), where R is the gas constant and Tis the
cell and swelling it to the bursting point.
absolute temperature.

Aquaporins Increase the Water Permeability

of the total number of solute molecules on each side of the
membrane. For example, a 0.5 M NaCI solution is actually of Cell Membranes
0.5 M Na + ions and 0.5 M Cl ions and has the same os- The natural tendency of water to flow across cell membranes
motic pressure as a 1 M solution of glucose or sucrose. as a result of osmotic pressure raises an obvious question:
The movement of water across the plasma membrane de- why don't the cells of fresh-water animals burst in water?
termines the volume of individual cells, which must be regu- For example, frogs lay their eggs in pond water (a hypotomc
lated to avoid damage to the cell. Small changes in extracellular solution), but frog oocytes and eggs do not swell with water
osmotic conditions cause most animal cells to swell or shrink even though their internal salt (mainl y KCI) concentration is
rapidly. When placed in a hypotonic solution (i.e., one in comparable to that of other cells (- 150 mM KCI). These
which the concentration of solutes is lower than in the cyto- observations were what first led investigators to suspect that
sol), animal cells swell owing to the osmotic flow of water the plasma membranes of most cell types, but not frog
inward. Conversely, when placed in a hypertonic solution oocytes, contain water-channel proteins that accelerate the
(i.e., one in which the concentration of solutes is higher than osmotic flow of water. The experimental results shown in
in the cytosol ), animal cells shrink as cytosolic water leaves Figure 11-7 demonstrate that an aquaporin from the eryth-
the cell by osmotic flow. Consequently, cultured animal cells rocyte plasma membrane functions as a water channel.
must be maintained in an isotonic medium, which has a sol- In its functional form, aquaporin is a tetra mer of identical
ute concentration and thus osmotic strength similar to that 28-kDa subunits (Figure 11 -8a). Each subunit contains six
of the cell cytosol. membrane-spanning a helices that form a central pore
through which water can move in either direction, depending
riiJ In vascular plants, water and minerals are absorbed on the osmotic gradient (Figure 1 1-8b, c). At the center of
- from the soil by the roots and move up the plant through each monomer, the -2-nm-long water-selective channel, or
conducting tubes (the xylem); water loss from the plant, mainly pore, is only 0.28 nm in diameter-only slightly larger than
by evaporation from the leaves, drives this movement of water. the diameter of a water molecule. The molecular sieving
Unlike animal cells, plant, algal, fungal, and bacterial cells are properties of the constriction are determined by several con-
surrounded by a rigid cell wall, which resists the expansion of served hydrophilic amino acid residues whose side-chain and
the volume of the cell when the intracellular osmotic pressure carbonyl groups extend into the middle of the channel and by
increases. Without such a wall, animal cells expand when in- a relatively hydrophobic wall that lines one side of the chan-
ternal osmotic pressure increases-if that pressure rises too nel. Several water molecules move simultaneously through
much, the cells will burst like overinflated balloons. Because of the channel, each molecule sequentially forming specific hy-
the cell wall in plants, the osmotic influx of water that occurs drogen bonds with channel-lining amino acids and displacing
when such cells are placed in a hypotonic solution (even pure another water molecule downstream. Since aquaporins do
water) leads to an increase in intracellular pressure but not in not undergo conformational changes during water transport,
cell volume. In plant cells, the concentration of solutes (e.g., they transport water orders of magnitude faster than GLUT!

11.2 Facilitated Transport of Glucose and Water 481

(;) VIDEO: Frog Oocyte Expressing Aquaporin Bursts in Hypotonic Solution

0.5 min 1.5 min 2.5 min 3.5 min

EXPERIMENTAL FIGURE 11-7 Expression of aquaporin by frog solution (0.035 M). The volume of the control oocytes remained
oocytes increases their permeability to water. Frog oocytes, which unchanged because they are poorly permeable to water. In contrast,
normally are impermeable to water and do not express an aquaporin the microinjected oocytes expressing aquaporin swelled and then
protein, were microinjected w ith mRNA encoding aquaporin. These burst because of an osmotic influx of water, indicating that aquaporin
photographs show control oocytes (bottom cell in each panel) and is a water-channel protein. [Courtesy of Gregory M. Preston and Peter Agre,
microinjected oocytes (top cell in each panel) at the indicated t imes Johns Hopkins University School of Medicine. See L. S. King, D. Kozono, and P.
after transfer from an isotonic salt solution (0. 1 M) to a hypotonic salt Agre, 2004, Nat. Rev. Mol. Cell Bioi. 5:687-698.]

Pore

(b) .·
Exterior

Cytosol
Cytosolic
NH 3 + coo vestibule

FIGURE 11 -8 Structure of the water-channel protein aquaporin. gate. (c) Side view of the pore in a single aquaporin subunit in which
(a) Structural model of the tetra meric protein comprising four identical several water molecules (red oxygens and white hydrogens) are seen
subunits. Each subunit forms a water channel, as seen in this view within the 2-nm-lonq water-selective gate that separates the water-
looking down on the protein from the exoplasmic side. One of the filled cytosolic and extracellu lar vestibules. The gate contains highly
monomers is shown with a molecular surface in which the pore conserved arginine and histidine residues, as well as the two aspara-
entrance can be seen. (b) Schematic diagram of the topology of a gine residues whose side chains form hydrogen bonds w ith transport-
single aquaporin subunit in relation to the membrane. Three pairs of ed water molecules. (Key gate residues, including the two asparagines,
homologous transmembrane a helices (A and A', Band 8', and C and are highlighted in blue.) Transported water molecules also form
C') are oriented in the opposite direction with respect to the mem- hydrogen bonds to the main-chain carbonyl group of a cysteine
brane and are connected by two hydrophilic loops containing short residue. The arrangement of these hydrogen bonds and the narrow
non-membrane-spanning helices and conserved asparagine (N) pore diameter of 0.28 nm prevent passage of protons (i.e., H 30 ) or
residues. The loops bend into the cavity formed by the six transmem- other ions. [After H. Sui et al., 2001, Nature 41 4:872. See also T. Zeuthen, 2001,
brane helices, meeting in the middle to form part of the water-selective Trends Biochem. Sci. 26:77, and K. Murata et al., 2000, Nature 407:599.]
transports glucose. The formation of hydrogen bonds be-
tween the oxygen atom of water and the amino groups of expression in different cell types, and substrate specificities
two amino acid side chains ensures that only uncharged are important for proper sugar metabolism in the body.
water (i.e., H 20, but not H 3 0+) passes through the channel; • Two common experimental systems for studying the func-
the orientations of the water molecules in the channel prevent tions of transport proteins are liposomes containing a puri-
protons from jumping from one to the next and thus prevent fied transport protein and cells transfected with the gene
the net movement of protons through the channel. As a con- encoding a particular transport protein.
sequence ionic gradients are maintained across membranes
• Most biological membranes are semipermeable, more per-
even when water is flowing :1cross them through aquaporins.
meable to water than to ions or most other solutes. Water
moves by osmosis across membranes from a solution of lower
IF!j Mammals express a family of aquaporins; 11 such genes
solute concentration to one of higher solute concentration.
U are known in humans . Aquaporin 1 is expressed in
abundance in erythrocytes, and the homologous aquaporin 2 • The rigid cell wall surrounding plant cells prevents their
is found in the kidney epithelial cells that resorb water from swelling and leads to generation of turgor pressure in re-
the urine, thus controlling the amount of water in the body. sponse to the osmotic influx of water.
The activity of aquaporin 2 is regulated by vasopressin, also • Aquaporins are water-channel proteins that specifically
called antidiuretic hormone. The regulation of the activity of increase the permeability of biomembranes to water (see
aquaporin 2 in resting kidney cells resembles that of GLUT4 in Figure 11-8).
fat and muscle in that when its activity is not required, when
• Aquaporin 2 in the plasma membrane of certain kidney
the cells are in their resting state and water is excreted to form
cells is essential for resorption of water from urine being
urine, aquaporin 2 is sequestered in intracellular vesicle mem-
formed; the absence of aquaporin 2leads to the medical con-
branes and so is unable to mediate water import into the cell.
dition diabetes insipidus.
When the polypeptide hormone vasopressin binds to the cell-
surface vasopressin receptor, it activates a signaling pathway
using cAMP as the intracellular signal (detailed in Chapter 15)
that causes these aquaporin 2-containing vesicles to fuse with 11.3 ATP-Powered Pumps and the
the plasma membrane, increasing the rate of water uptake and
its return into the circulation instead of the urine. Inactivating Intracellular Ionic Environment
mutations in either the vasopressin receptor or the aquaporin In previous sections, we focused on transport proteins that
2 gene cause diabetes insipidus, a disease marked by excretion move molecules down their concentration gradients (facili-
of large volumes of dilute urine. This finding demonstrates that tated transport). Here we focus our attention on a major
the level of aquaporin 2 is rate limiting for water resorption class of proteins-the ATP-powered pumps-that use the
from urine being formed by the kidney. • energy released by hydrolysis of the terminal phosphoanhy-
dride bond of ATP to transport ions and various small mol-
Other members of the aquaporin family transport hydroxyl- ecules across membrane s against their concentration
containing molecules such as glycerol rather than water. gradients. All ATP-powered pumps are transmembrane pro-
Human aquapo;in 3, for instance, transports glycerol and is teins with one or more binding sites for ATP located on sub-
similar in amino acid sequence and structure to the Esche- units or segments of the protein that always face the cytosol.
richia coli glycerol transport protein GlpF. These proteins commonly arc called A TPases and they nor-
mally do not hydrolyze ATP into ADP and P, unless ions or
other molecules arc simultaneously transported. Because of
this tight coupling between ATP hydrolysis and transport,
KEY CONCEPTS of ~ection 11.2
the energy stored in the phosphoanhydride bond is not dis-
Facilitated Transport of Glucose and Water sipated as heat hut rather is used to move ions or other mol-
• Protein-catalyzed transport of biological solutes across a ecules uphill against an electrochemical gradient.
membrane occurs much faster than simple diffusion, exhib-
its a Vmax when the limited number of transporter molecules There are Four Main Classes
are saturated with substrate, and is highly specific for sub- of ATP-Powered Pumps
strate (see Figure 11-4).
The general structures of the four classes of ATP-powcred
• Uniport proteins, such as the glucose transporters (GLUTs), pumps are depicted in figure 11-9, with specific examples in
are thought to shuttle between two conformational states, each class listed below the figure. Note that the members of
one in which the substrate-binding site faces outward and three of the classes (P, F, and V) only transport ions, as do
one in which the binding site faces inward (see Figure 11-5). some members of the fourth class, the ABC superfamily. Most
• All members of the GLUT protein family transport sugars members of the ABC superfamily transport small molecules
and have similar structures. Differences in their Km values, such as amino acids, sugars, peptides, lipids, and other small
molecules including many types of drugs.

11.3 ATP-Powered Pumps and the Intracellular Ionic Environment 483

Exoplasmic 2W 4W
face

Cytosolic
face

ATP ATP
ADP ADP + P,

ADP+P,
P-class pumps V-class proton pumps F-class proton pumps ABC superfamily
Plasma membrane of plants and Vacuolar membranes in Bacterial plasma Bacterial plasma
fungi (W pump) plants, yeast, other fungi membrane membranes (amino acid,
sugar, and peptide
Plasma membrane of higher Endosomal and lysosmal Inner mitochondrial
eukaryotes (Na+/K+ pump) transporters)
membranes in animal membrane
cells Mammalian plasma
Apica l plasma membrane of T hylakoid membrane
membranes (transporters
mammalian stomach (W/K pump) Plasma mem brane of of chloroplast
of phospholipids, small
osteoclasts and some
Plasma membrane of all eukaryotic lipophilic drugs, cholesterol,
cells (Cah pump) kidney tubule cells
other small mo lecules)
Sarcoplasmic reticu lum membrane
in muscle cells (Ca 2+ pump)

FIGURE 11 -9 The four classes of ATP-powered transport gradient, whereas F-class pumps normally operate in the reverse
proteins. The locations of specific pumps are indicated below each direction to utilize energy in a proton concentrat ion or voltage gradient
class. P-class pumps are composed of two catalytic a subunits, which to synthesize ATP. All members of the large ABC superfamily of proteins
become phosphorylated as part of the transport cycle. Two f3 subunits, contain two tran smembrane m domains and two cytosolic ATP-binding
present in some of these pumps, may regulate transport. Only one a (A) domains, which couple ATP hydrolysis to solute movement. These
and p subunit are depicted. V-class and F-class pumps do not form core domains are present as separate subunits in some ABC proteins
phosphoprotein intermediates and almost all t ransport only protons. (depicted here) but are fused into a single polypeptide in ot her ABC
Their structures are similar and contain similar proteins, but none of proteins. [See T. Nishi and M. Forgac, 2002, Nature Rev. Mol. Cell Bioi. 3:94; C.
their subunits are related to those of P-el ass pumps. V-class pumps Toyoshtma et al., 2000, Nature 405:647; D. Mcintosh, 2000, Nature Struc. Bioi.
couple ATP hydroly5is to t ransport of protons against a concentration 7:532; and T. Elston, H. Wang, and G. Oster, 1998, Nature 391 :51 0.]

All P-class ion pumps possess rwo idenrical cara lyric a transmembrane and cytosolic subunirs. Virtually all known V
subunits, each of which contains an ATP-binding site. Most and F pumps transport only protons and do so in a process
also have two smaller 13 subunits that usually have regula- that does not involve a phosphoprotein intermed iate. V-class
tory functions. During rra nsport, at least one of the a sub- pum ps generall y fu nction to generate the low pH of plant
units becomes phosphorylated (hence the name "P" class), vacuoles and of lysosomes and other acidic vesicles in animal
and the transported ions move through the phosphorylated cells by pumping protons from the cytosolic to the exoplasmic
subunit. The amino acid seq uences aro und th e phosphory- face of the membrane aga inst a proton electrochemical gradi-
lated residues are homologous in different pumps. This class ent. In contrast, the H + pumps that generate and maintain the
includes the Na + !K ATPase in the plasma membrane, plasma mem brane electric potential in p lant, fungal, and
which generates the low cytosolic Na and high cytosolic K+ many bacterial cells belong to the P-class of proton pumps.
concentrations typical of animal cells (see Figure 11 -3 ). Cer- F-class pu mps are fo und in bacterial plasma membranes
tain Ca 2 + ATPases pump Ca2 + ions out of the cytosol into and in mitochondria and chloroplasts. In contrast to V-class
the external medium; others pump Ca 2 + from the cytosol pumps, they generally function as reverse proton pumps, in
into the endoplasmic reticulum or into the specialized ER which the energy released by the energetically favored move-
called the sarcoplasmic reticulum tha t i~ found in muscle ment of protons fro m the exoplasm ic to the cyrosolic face ot
cells. Another member of the P class, found in acid-secreti ng the membrane down the proton electrochemical gradient is
cells of the mammalian stomach, transports protons (H used to power the energeticall y unfavorable synthesis of
ions) out of and K+ ions mto the cell. ATP from ADP and P,. Because of their importance in ATP
The structures of V-class and F-class ion pumps are simi la r synthesis in chloroplasts and mitochondria, F-class proton
to one another but unrelated to, and more complicated than, pumps, commonly called ATP synthases, are treated sepa-
P-class pumps. V- and F-class pumps contain several different rately in Chapter 12 (Cellular Energetics).

484 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 The final class of ATP-powered pumps is a large family
TABLE 11-2 Typical Intracellular and Extracellular
of multiple members that are more diverse in function than
those of the other classes. Referred to as the ABC (ATP- lon Concentrations
binding cassette) superfamily, this class includes several hun- Cell(mM) Blood(mM)
Ion
dred different transport proteins found in organisms ranging - -
from bacteria to humans. As detailed below, some of these Squid Axon
transport proteins were first identified as multidrug-resistance (invertebrate)*
proteins that, when overexpressed in cancer cells, export an- - - -
ticancer drugs and render tumors rcsisranr ro their auion. 400 20
Each ABC protein is specific for a single substrate or group
of related substrates, which may be ions, sugars, amino Na 50 440
acids, phospholipids, cholesterol, peptides, polysaccharides,
or even proteins. All ABC transport proteins share a struc- 40-150 560
tural organization consisting of four "core" domains: two
transmembrane (T) domains, forming the passageway 0.0003 10
through which transported molecules cross the membrane,
and two cytosolic ATP-binding (A) domains. ln some ABC 300-400 5-10
proteins, mostly those in bacteria, the core domains are pres-
Mammalian Cell
ent in four separate polypeptides; in others, the core domains
{vertebrate)
are fused into one or two multidomain polypeptides.
139 4
AlP-Powered Ion Pumps Generate and Maintain
Ionic Gradients Across Cellular Membranes
Na~ 12 145
The specific ionic composition of the cytosol usually differs 4 116
greatly from that of the surrounding extracellular fluid. In
virtually all cells-including microbial, plant, and animal 12 29
cells-the cytosolic pH is kept near 7.2 regardless of the ex-
tracellular pH. In the most extreme case, there is a million- X 138 9
fold difference in H + concentration between the pH of the
cytosol of the epithelial cells lining the stomach and the pH 0.8 1.5
of the stomach contents after a meal. Also, the cytosolic con-
centration of K is much higher than that of Na . In both < 0.0002 1.8
invertebrates and vertebrates the concentration of K is
The large nerve axon of rhe squid has been widely used in ~rudies of the
20-40 times higher in the cytosol than in the blood, while mechamsm of conduction of electric impulses.
the concentratiop of Na + is 8-12 times lower in the cytosol ·X represents protetns, wh1ch have a ner neganve charge ar rhe neutral
than in the blood (Table 11-2 ). Some Ca2 in the cytosol is pi I of blood and cells.
hound to the negatively charged groups in ATP and in pro-
teins and other molecules, but it is the concentration of un-
bound (or "free") Cal+ that is critical to its functions in
signaling pathways and muscle contraction. The concentra- ability to conduct electric signals rapidly and efficiently, as
tion of free Ca 2 · in the cytosol is generally less than 0.2 mi- we detail in Chapter 22. Certain enzymes required for pro-
cromolar (2 X 10 - M), 'a thousand or more times lower tein synthesis in all cells require a high K + concentration and
than that in the blood. Plant cells and many microorganisms are inhibited by high concentrations of Na ; these would
maintain similarly high cytosolic concentrations of K and cease to function without the operation of the Na /K+
low concentrations of Ca 2 + and Na even if the cells are pump. In cells treated with poisons that inhibit the produc-
cultured in very dilute salt solutions. tion of ATP (e.g., 2,4-dinitrophenol in aerobic cells), the
The ion pumps discussed in this section are largely re- pumping stops and the ion concentrations inside the cell
sponsible for establishing and maintaining the usual ionic gradually approach those of the exterior environment as
gradients across the plasma and intracellular membranes. In ions spontaneously move through channels in the plasma
carrying out this task, cells expend considerable energy. For membrane down their electrochemical gradients. Eventually
example, up to 25 percent of the ATP produced by nerve and poison-treated cells die: partly because protein synthesis re-
kidney cells is used for ion transport, and human erythro- quires a high concentration of K + ions and partly because in
cytes consume up to 50 percent of their available ATP for the absence of a Na + gradient across the cell membrane, a
this purpose; in both cases, most of this ATP is used to cell cannot import certain nutrients such as amino acids (see
power the Na /K• pump (see Figure 11-3). The resultant Figure 11-3). Studies on the effects of such poisons provided
Na ... and K + gradients in nerve cells are essential for their early evidence for the existence and significance of ion pumps.

11.3 ATP·Powered Pumps and the Intracellular Ionic Environment 485

Muscle Relaxation Depends on Ca2+ ATPases in the plasma membrane to ensure that the cytosolic concen-
That Pump Ca2+ from the Cytosol into the tration of free Ca 2 + in resting muscle remains below 0.1 1-1M.
Sarcoplasmic Reticulum
The Mechanism of Action of the Ca2+ Pump
In skeletal muscle cells, Ca 2 + ions are concentrated and
stored in the sarcoplasmic reticulum (SR), a specialized type Is Known in Detail
of endoplasmic reticulum (ER). The release via ion channels Because the calcium pump constitutes more than 80 percent
of stored Ca 2 + ions from the SR lumen into the cytosol causes of the integral protein in muscle SR membranes, it is easily
muscle contraction, as d1scussed in Chapter 17. A P-class purified from other membrane proteins and has been studied
Ca2 · ATPase located in the SR membrane of skeletal muscle extensively. Determination of the three-dimensional structure
pumps Ca 2 - from the cytosol back into the lumen of the SR, of this protein in several conformational states representing
thereby inducing muscle relaxation. different steps in the pumping process has revealed much
In the cytosol of muscle cells, the free Ca 1 concentration about its mechanism of action, and serves as a paradigm for
ranges from 10 M (resting cells) to more than 10 6 M (con-
7
understanding many P-class ATPase pumps.
tracting cells), whereas the total Ca 2 • concentration in the SR The current model for the mechanism of action of the Ca1 t
lumen can be as high as 10 1 M. The lumen of the SR contains ATPase in the SR membrane involves multiple conformational
two abundant proteins, calsequestrin and the so-called high- states. For simplicity, we group these into £1 states, in which the
affinity Ca 2 + binding protein, each of which binds multiple two binding sites for Ca2 ... , located in the center of the mem-
Ca 2 ions at high affinity. By binding much of the Ca 2 in the
+ + brane-spanning domain, face the cytosol, and £2 states, in which
SR lumen these proteins reduce the concentration of "free" these binding sites face the exoplasmic face of the membrane,
Ca2 - ions in the SR vesicles. This reduces the Ca2 ... concentra- pointing into the lumen of the SR. Coupling of ATP hydrolysis
tion gradient between the cytosol and the SR lumen and con- with ion pumping involves several conformational changes in
sequently reduces the energy needed to pump Cah ions into the protein that must occur in a defined order, as shown in Fig-
the SR from the cytosol. The activity of the muscle Ca 2 c ure 11-10. When the protein is in the £1 conformation, two
ATPase increases as the free Ca 2 + concentration in the cytosol Ca2 + ions bind to two high-affinity binding sites accessible from
rises. In skeletal muscle cells the calcium pump in the SR the cytosolic side; even though the cytosolic Ca2+ concentration
membrane works in concert with a similar Ca 2 + pump located is low (see Table 11-2), calcium ions still fill these sites.

Ca 2 ~-binding
SR lumen sites
E1 E1 E1
ca 2 Calcium and Phosphorylation
ATP binding of aspartate

---11--+ --II-+

Cytosol

ATP
Conformational Conformational
ATP- change change ·.
binding
site Dephosphorylation Calcium
release

+-a- +-II-

E2 E2 E2

FIGURE 11 - 10 Operational model of the Cal+ ATPase i n the SR membrane. In the figure, - P indicates a high-energy aspartyl phosphate
membrane of skeletal muscle cells. Only one of the two catalytic bond; -P indicates a low-energy bond. Because the affinity of CaF for
a subunits of this P-class pump is depicted. E1 and E2 are alternative the cytosolic-facing binding sites in E1 is a thousandfold greater than
conformations of the protein in which the Ca 2... -binding sites are the affinity of CaH for the exoplasmic-facing sites in E2, this pump
accessible to the cytosolic and exoplasmic (SR lumen) faces, respec- transports Ca2+ unidirectionally from the cytosol to the SR lumen. See
tively. An ordered sequence of steps, as diagrammed here, is essential the text and Figure 11-11 for more details. [See C. Toyoshima and G. lnesi,
for coupling ATP hydrolysis and the transport of Ca 2... ions across the 2004, Ann. Rev. Biochem. 73:269-292.]

486 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 Next, an ATP binds to a site on the cytosolic surface Binding of Ca2 • ions to the Cal+ pump illustrates a gen-
(step 0 ). The bound ATP is hydrolyzed to ADP m a reaction eral principle of ion binding to channel and transport pro-
that requires Mg 2 + , and the liberated phosphate is trans- teins that we will encounter repeatedly in this chapter: as
ferred to a specific aspartate residue in the protein, forming ions bind they lose most of their waters of hydration but in-
the high-energy acyl phosphate bond denoted by El - P teract with oxygen atoms in the protein that are in a similar
(step f)). The protein then undergoes a conformational geometry to the water oxygens that are bound to the ion in
change that generates E2, in which the affinity of the two aqueous solution. This reduces the thermodynamic barrier
Ca 1 -binding sites is reduced (shown in detail in the next for ion binding to the protein and allows tight binding of the
figure) and in which these sites are now accessible to the SR ion even from solminns of relatively low concentrations.
lumen (step ID). The free energy of hydrolysis of the aspartyl- The cytoplasmic region of the Cah pump consists of three
phosphate bond in E l -Pis greater than that in E2-P, and domains that are well separated from each other in the £1
this reduction in free energy of the aspartyl-phosphate bond state (Figure 11-llb}. Each of these domains is connected to
can be said to power the El ~ E2 conformational change. the membrane-spanning helices by short segments of amino
The Ca 2 + ions spontaneously dissociate from the low- acids . Movements of these cytosolic domains during the
affinity sites to enter the SR lumen, because even though the pumping cycle cause movements of the connecting segments
Ca 1 concentration there is higher than in the cytosol, it is that are transmitted into movements of the attached membrane-
lower than the Kd for Ca 2 + binding in the low-affinity state spanning ex helices. For example, the phosphorylated residue,
(step 19). Finally, the aspartyl-phosphate bond is hydrolyzed Asp 351, is located in the P domain. The adenosine moiety of
(step H ). This dephosphorylation, coupled with subsequent ATP binds to the N domain, but the -y-phosphate of ATP
binding of cytosolic Ca 2 + to the high-affinity El Ca2 + bind- binds to specific residues on the P domain, requiring move-
.. • ing sites, stabilizes the E1 conformational state relative to £2,
and can be said to power the E2 ~ E1 conformational change
ments of both theN and P domains. Thus, following ATP and
Ca 1 binding, the-y phosphate of the bound ATP sits adjacent
(step m). Now El is ready to transport two more Ca 1 ions. to the aspartate on the P domain that is to receive the phos-
Thus the cycle is complete and hydrolysis of a phosphoanhy- phate. Although the precise details of these and other protein
dride bond in ATP has powered the pumping of 2 Ca2 • ions conformational changes are not yet clear, the movements of
against its concentration gradient into the SR lumen. theN and P domains are transmitted by lever-like motions of
Much structural and biophysical evidence supports the the connecting segments into rearrangements of several mem-
model depicted in Figure 11-10. For instance, the muscle cal- brane-spanning ex helices. These changes are especially appar-
cium pump has been isolated with phosphate linked to the ent in the four helices that contain the two Ca 2 -binding sites:
key aspartate residue, and spectroscopic studies have de- the changes prevent the bound Ca 2 + ions from moving back
tected slight alterations in protein conformation during the into the cytosol when released but enable them to dissociate
£1 ~ E2 conversion. The two phosphorylated states can into the exoplasmic space (lumen).
also be distinguished biochemically; addition of ADP to All P-class ion pumps, regardless of which ion they trans-
phosphorylated El results in synthesis of ATP, the reverse of port, are phosphorylated on a highly conserved aspartate
step f) in Figure 11-10, whereas addition of ADP to phos- residue during the transport process. As deduced from eDNA
phorylated £2 does not. Each principal conformational state sequences, the catalytic ex subunits of all the P pumps examined
of the reaction cycle can also be characterized by a different to date have similar amino acid sequences and thus are pre-
susceptibility to various proteolytic enzymes such as trypsin. sumed to have similar arrangements of transmembrane ex heli-
Figure 11-11 shows the three-dimensional structure of the ces and cytosol-facing A, P, and N domains (see Figure 11-1 0 ).
Ca 2 · pump in the £1 state. As can be seen in the right two pan- These findings strongly suggest that all such proteins evolved
els of part c in the bottom half of the figure, the 10 membrane- from a common precursor although they now transport differ-
spanning ex helices in the catalytic subunit form the passageway ent ions. This suggestion is borne out by the similarities of the
through which Ca 1 + ions )Tiove. Amino acids in four of these three-dimensional structures of the membrane-spanning
helices form the two high-affinity E1 Cal+ -binding sites (Figure segments of the Na + /K'" ATPase with that of the Ca 1 + pump
11-11a, left). One site is formed out of negatively charged oxy- (Figure 11-12 ); the molecular structures of the three cyto-
gen atoms from the carboxyl groups (COO-) of glutamate and plasmic domains are also very similar. Thus, the operational
aspartate side chains, as well as from water molecules. The model in Figure 11-11 is generally applicable to all of the
other site is formed from side- and main-chain oxygen atoms. P-class ATP-powered ion pumps.
Thus, as Cal+ -ions bind to the Ca 2 "' pump they lose the
water molecules that normally surround a Cal+ ion in aque-
ous solution (see Figure 2-7), but these waters are replaced Calmodulin Regulates the Plasma Membrane
by oxygen a rom~ with a similar geometry that are part of the
Pumps That Control Cytosolic
transport protein. In contrast, in the £2 state (Figure 11-11a,
right), several of these binding side chains have moved frac-
Ca 2 Concentrations
-j

tions of a nanometer and are unable to interact with bound As we explain in Chapter 15, in many types of cells in addition
Cal+ ions, accounting for the low affinity of the £2 state for to muscle cells, small increases in the concentration of free
Ca 2 "'" ions. Cah ions in the cytosol trigger a variety of cellular responses.

11.3 ATP-Powered Pumps and the Intracellular Ionic Environment 487

 (a) E1 state E2 state
High affinity for Ca2• Low affinity for Ca 2•
Two bou nd Ca 2 • No bound Ca 2•

~- 4 f./ Ca 2•

1.
\) )~
'\
-~j
·._:-v
r 6

(b) (c)
'
-
SR lumen

Cytosol

Actuator
domain (A) Phosphorylation
site
Nucleotide·
binding domain (N)
FIGURE 11 -11 Structure of the catalytic a subunit of the muscle three domains: the nucleotide-binding domain N (blue), the phosphor-
Cal+ ATPase. (a) Ca 2 +·binding sites in the El state (left), with two ylation domain P (green), and the actuator domain A (beige) that
bound calcium ions, and the low-affinity E2 state (right), without bound connects two of the membrane-spanning helices. (c) Models of the
ions. Side chains of key amino acids are white, and the oxygen atoms pump in the El state (left) and E2 state (right). Note the differences
on the glutamate and aspartate side chains are red. In t he high-affinity between the El and E2 states in the conformations of the nuc leotide·
El conformation, Ca 2 ions bind at two sites between helices 4, 5, 6, binding and actuator domains; these movements power the conform a·
and 8 inside the membrane. One site is formed out of negatively tiona! changes of the membrane-spanning Ct. helices (purple) that
charged oxygen atoms from glutamate and aspartate side chains and constitute the Ca 2 ' -binding sites, converting them from one in which
of water molecules (not shown), an d the other is formed out of side- the Ca 2 -binding sites are accessible to the cytosolic face (E 1 state) to
and main-chain oxygen atoms. Seven oxygen atoms su rround the Ca2 one in which the now loosely bound Ca 2 ions are accessible to the
ion in both sites. (b) Three-dimensional model of the protein in the El exoplasmic face (E2 state). [Adapted from C. Toyoshima and H. Nomura,
state based on the structure determined by x-ray crystallography. 2002, Nature 418:605- 611; C. Toyoshima and G.lnesi, 2004, Ann. Rev. Biochem.
There are 10 transmembrane a helices, f our of which (purple) contain 73:269-292; and E. Gouaux and R. MacKinnon, 2005, SCience 31 0 :1461.]
residues that participate in Ca 2 binding. The cytosolic segment forms

In order for Ca 2 to funt.tiuu in intracellular signaling, the con- The activity of plasma membrane Ca 2 ATPases is regu-
centration of Ca 2 ions free in the cytosol usually must be kept lated by calmod ulin, a cytosolic Ca 2 - -binding protein (see
below 0.1-0.2 f..lM. Animal, yeast, and probably plant cells Figure 3-31 ). A rise in cytosolic Ca 2 - induces t he binding of
express plasma membrane Ca 1 + ATPases that transport Ca 2 Ca 1 ions to calmodulin, which triggers activation of the
out of the cell against its electrochemical gradient. The cata- Cal<- ATPase. As a result, the export of Ca 1 + ions from the cell
lytic a subunit of these P-class pumps is similar in structure accelerates, qu ickly restoring the low concentration of free
and sequence to the a subunit of the muscle SR Ca1 - pump. cytosolic Ca 1 characteristic of the resting cell.

488 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 FIGURE 11 - 12 Structural comparison of Na+/K+ ATPase and
muscle Ca2+ ATPase. Three-dimensional structure of the Na /K+
ATPase (gold) compared to that ofthe muscle Ca1 · ATPase (purple), as
seen from the cytoplasmic surface. aM1-1 0 denote the 10 membrane-
spanning a-helices of the Na • /K ATPase. [After J.P. North et. al., 2007,
Narure450:1043 and H. Ogawa et. al., 2009, Proc. Nat'/ Acad. Sci. USA 106:13742]

involved directly in ion pumping. During the catalytiC cycle of

the Na ~ /1("' ATPase it moves three Na ions out of and two K+
ions into the cell per ATP molecule hydrolyzed. The mechanism
of action of the Na + !K"' ATPase, outlined in Figure 11-13, is
similar to that of the muscle SR calcium pump, except that ions
are pumped in both directions across the membrane, with each
ion moving against its concentration gradient. In its El confor-
mation, the Na /K+ ATPase has three high-affinity Na -binding
sites and two low-affinity K -binding sites accessible to the cy-
Na + /K +- ATPase Maintains the Intracellular Na + rosolic surface of the protein. The Km for binding of Na to
these cyrosolic sites is 0.6 m.'v1, a value considerably lower than
and K+ Concentrations in Animal Cells the intracellular Na + concentration of -12 mM; as a result,
An important P-class ion pump present in the plasma mem- Na ~ ions normally will fully occupy these sites. Conversely, the
brane of all animal cells is the Na +IK ATPase. This ion pump affinity of the cytosolic K +-binding sires is low enough that K+
is a tetramer of subunit composition a 213 2• and shares structural ions, transported inward through the protein, dissoctate from
homology with the Ca2+ pump (see Figure 11-12). The small, El into the cytosol despite the high intracellular K+ concentra-
glycosylated 13 transmembrane polypeptide apparently is not tion. During the E 1 ~ E2 transition, the three bound Na ions

G) OVERVIEW ANIMATION: Biological Energy lnterconversions

Exterior E1
Na+ and
E1
Phosphorylation
E1
Conformational
E2
•• 2 K+

Na• ATP binding of aspartate change

;jj
,\
--.--. ~ ~

K+
Cytosol

3 Na•

f TP ADP

K+ release
Dephosphorylation
and conformational
•••
change

+-0- +-II-

E1 E1 E2
FIGURE 11-13 Operational model of the plasma membrane (see Figure 11-11 ). In this case, hydrolysis of the E2-P intermediate
Na+/K+ ATPase. On ly one of the two catalytic a subunits of this powers the E2 ~ E1 conformational change and concomitant
P-class pump is depicted. It is not known whether just one or both transport of two K ions inward. Na + ions are indicated by red circles;
subunits in a single ATPase molecule transport ions. lon pumping by K ions, by purple squares; high-energy acyl phosphate bond, by - P;
the Na · /K+ ATPase involves phosphorylation, dephosphorylation, and low-energy phosphoester bond, by -P.
conformational changes similar to those in the muscle Ca1 ATPase

11.3 ATP-Powered Pumps and the Intracellular Ionic Environment 489

become accessible to the exoplasmic face, and simultaneously By themselves, ATP-powered proton pumps cannot acid-
the affinity of the three Na--binding ~ires drops. The three Na + ify the lumen of an organelle (or the extracellular space) be-
ions now bound to low-affinity Na + sites dissociate one at a cause these pumps are electrogenic; that is, a net movement
time into the extracellular medium despite the high extracellu- of electric charge occurs during transport. Pumping of just a
lar Na +concentration. Transition to the E2 conformation also few protons causes a buildup of positively charged H ions
generates two high-affinity K- sites accessible to the exoplasmic on the exoplasmic (inside ) face of the organelle membrane.
face. Because the Km for K binding to these sites (0.2 mM) is For each H .,. pumped across, a negative ion (e.g., OH or
lower than the extracellular K+ concentration (4 mM), these Cl ) will be "left behind" on the cytosolic face, causing a
sites will fill with K ions as the Na ions dissociate. Similarly, buildup of negatively charged ions there. These oppoSitely
during the subsequent E2 ~ El transition, the two bound K- charged ions attract each other on opposite faces of the
ions are transported inward and then released into the cytosol. membrane, generating a charge separation, or electric poten-
Certain drugs (e.g., ouabain and digoxin) bind to the exo- tial, across the membrane. The lysosome membrane thus
plasmic domain of the plasma membrane Na+/K+ ATPase functions as a capacitor in an electric circuit, storing oppos-
and specifically inhibit its ATPase activity. The resulting dis- ing charges (a nions and cations) on opposite sides of a bar-
ruption in the Na +/K balance of cells is strong evidence for rier impermeable to movement of charged particles.
the critical role of this ion pump in maintaining the normal As more and more protons are pumped and build up excess
K+ and Na + ion concentration gradients. Classic Experiment positive charge on the exoplasmic ~ce, the energy required to
11.1, which directly follows this chapter, describes the dis- move additional protons against this rising electric potential gra-
covery of this important enzyme, which is required for life. dient increases dramatically and prevents pumping of additional
protons long before a significant transmembrane H ... concentra-
tion gradient is established (Figure 11 -14a). In fact, this is the
way that P-class H . pumps generate a cytosol-negative potential
V-Ciass H '" ATPases Maintain the Acidity
across plant and yeast plasma membranes.
of lysosomes and Vacuoles In order for an organelle lumen or an extracellular space
~ .'

All V-class A TPases transport only H- ions. These proton (e.g., the lumen of the stomach) to become acidic, movement
pumps, present in the membranes of lysosomes, endosomes,
and plant vacuoles, function to acidify the lumen of these or-
ganelles. The pH of the lysosomal lumen can be measured (a) ATP ADP + P,
precisely in living cells by use of particles labeled with a pH-
sensitive fluorescent dye. When these particles are added to Cl
the extracellular fluid, the cells engulf and internalize them
Cytosol
(phagocytOsis; see Chapter 17), ultimately transporting them
into lysosomes. The lysosomal pH can be calculated from the
spectrum of the. fluorescence emitted. The DNA encoding a Electric
Neutral pH potential
naturally fluorescent protein whose fluorescence depends on
the pH can be modified (by adding DNA segments encoding
"signal sequences," detailed in Chapters 13 and 14) such that
the protein is targeted to the lysosome lumen; fluorescence
measurements can then be used to determine the pH in the
(b) ATP ADP + P;
organelle lumen. Maintenance of the hundredfold or more

H~
proton gradient between the lysosomal lumen (pH -4.5-5.0)
and the cytosol (pH -7.0) depends on a V-class ATPase and
thus ATP production by the cell. The low lysosomal pH is
necessary for optimal function of the many proteases, nucle-
ases, and other hydrolytic enzymes in the lumen; on the other Cl
hand, a cytosolic pH of 5 would disrupt the functions of many No electric
Acidic pH potential
proteins optimized to act at pH 7 and lead to death of the cell.
Pumping of relatively few protons is required to acidify an
intracellular vesicle. To understand why, recall that a solution
of pH 4 has an H+ ion concentration of 10 4 moles per liter,
FIGURE 11-14 Effect of V-class H+ pumps on H+ concentration
or 10 7 moles of H t- ions per milliliter. Since there are 6.02 X
gradients and electric potential gradients across cellular
1023 atoms of H per mole (Avogad ro's number), then a milli- membranes. (a) If an intracellular organelle contains only V-class
liter of a pH 4 solution contains 6.02 X 10 16 H ions. Thus at pumps, proton pumping generates an electric potential across the
pH 4, a primary spherical lysosome with a volume of 4.18 X membrane (cytosol-facing side negative and luminal-side positive)
10 15 ml (diameter of 0.2 J..Lm) will contain just 252 protons. but no significant change in the intraluminal pH. (b) If the organelle
At pH 7, the same organelle would have an average of only membrane also contains Cl channels, anions passively follow the
0.2 protons in its lumen, and thus pumping of only approxi- pumped protons, resulting in an accumulation of H and o- ions in the
mately 250 protons is necessary for lysosome acidification. lumen (low luminal pH) but no electric potential across the membrane.

490 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

of protons must be accompanied either by ( I ) movement of an ABC Proteins Export a Wide Variety
equal number of anions (e.g., Cl ) in the same direction or b) of Drugs and Toxins from the Cell
(2) movement of equal numbers of a different cation in the op-
posite direction. The first process occurs in lysosomes and As noted earlier, all members of the very large and diverse
plant vacuoles, w hose membranes contain V-class H - ATPases ABC superfamily of transport proteins contain two trans-
and anion channels through which accompanying Cl ions membrane (T) domains and two cytosolic ATP-binding (A)
move (Figure 11-14b). The second process occurs in the lining domains (see Figure 11-9). The T domains, each built of 10
of the stomach, which comains a P-class H +IK- ATPase that is membrane-spanning a helices, form the pathway through
not electrogenic and pumps one H+ outward and one K in- which the transported substance (substrate) crosses the mem-
ward. Operation of this pump is discussed later in the chapter. brane (Figure 11 -15a) and determine the substrate specificity
The ATP-powered proton pumps in lysosomal and vacu- of each ABC protein. The sequences of the A domains are
olar membranes have been solubilized, purified, and incor- approximately 30-40 percent homologous in all members of
porated into liposomes. As shown in Figure ll-9, these this superfamily, indicating a common evolutionary origin.
V-class proton pumps contain two discrete domains: a cyto- Discovery of the first eukaryotic ABC protein to be recog-
solic hydrophilic domain (Vd and a transmembrane domain nized came from studies on tumor cells and cultured cells that
(Vo) w ith multiple subunits forming each domain. Binding exhibited resistance to several drugs with unrelated chemical
and hydrolysis of ATP by the B subunits in V 1 provide the structures. Such cells eventually were shown to express ele-
energy for pumping of H + ions through t he proton-conduct- vated levels of a multidrug-resistance (MDR) transport pro-
ing channel formed by the 'c' and 'a' subunits in V0 • Unlike tein originally called MDRJ and now known as ABCBl. This
P-class ion pumps, V-class proton pumps are not p hosphory- protein uses the energy derived from ATP hydrolysis to export
lated and dephosphorylated during proton transport. The a large variety of drugs from the cytosol to the extracellular
structurally similar F-class proton pumps, which we describe medium. The Mdrl gene is frequently amplified in multidrug-
in Chapter 12, norma lly operate in the "reverse" direction to resistant cells, resulting in a large overproduction of the
generate ATP rather than pump protons; their structure and MDRl protein. In contrast to bacterial ABC proteins, which
mechanism of action is understood in great detail. are built of four discrete subunits, all four domains of mam-
malian ABCB1 are fused into a single 170,000-MW protein.
The substrates of mammalian ABCBl are primarily
planar, lipid-soluble molecules with one or more positive
charges; they all compete with one another for transport,
(a) suggesting that they bind to the same or overlapping sites on
Transmembrane the protein. Many drugs transported by ABCBl diffuse from
domains 290 the extracellular medium across the plasma membrane,
Outside

Inside FIGURE 11-15 The multidrug transporter ABCB1 (MDR1):

structure and model of ligand export. (a) Cross-sectional view
through the center of an ABCBl protein bound to two molecules of a
drug analog qz59-sss (black) reveals the central location ofthe
ATP-binding { ligand-binding site in relation to the phospholipid bilayer: the central
domains ligand-binding cavity is close to the leaflet-leaflet interface of the
membrane. During transport, this binding cavity is alternately exposed
to the exoplasmic and the cytosolic surface of the membrane. Serines
289 and 290 affect the ligand specificity of the t ransporter, and are
(b)
shown as red spheres to highlight their juxtaposition to the bound
ligand. Surface residues are colored yellow to denote hydrophobic and
Outside blue to denote hydrophilic amino acids. (b) Three-dimensional
structure of ABCBl with its ligand-binding site facing inward toward
the cytosol. In this conformation a hydrophilic ligand can bind directly
·------- ---
@ ~~~
from the cytosol. A more hydrophobic ligand can partition into the
inner leaflet of the plasma membrane and then enter the ligand-
Inside !t binding site through a gap in the protein that Is accessible directly to
the hydrophobic core of the inner leaflet. (c) Model for the structure of
0 ABCBl with its ligand-binding site facing outward, based on the
structures of homologous bacterial ABC proteins. When the transporter
assumes this conformation t he ligand can either diffuse into the
exoplasmic leaflet or directly into the aqueous extracellular medium.
[After D. Gutman et. al., 2009, Trends Biochem. Sci. 35:36-42. Structures from
5. G. Aller et al., 2009, Sc1ence 323:1718-1722.]

11.3 ATP-Powered Pumps and the Intracellular Ionic Environment 491

unaided by transport proteins, into the cell cytosol, where the hydrophobic core of the inner leaflet of the membrane
they block various cellular functions. Two such drugs are bilayer; this allows hydrophobic molecules to enter the binding
colchicine and vinblastine, which block assembly of microtu- site directly from the inner leaflet of the phospholipid bilayer
bules (Chapter 18). ATP-powered export of such drugs by (Figure ll-15b). After the ATP-powcred change to the out-
MDRl reduces their concentration in the cytosol. As a re- ward-facing conformation, molecules can exit the binding
sult, a much higher extracellular drug concentration is re- site into the exoplasmic membrane leaflet or directly into the
quired to kill cells that express ABCB 1 than those that do extracellular medium (Figure 11-15c).
not. That ABCBl is an ATP-powered small-molecule pump About 50 different mammalian ABC transport proteins
has been demonstrated wiLh lipu:.umes containing the puri- are now recognized (Table 11-3). Several are expressed in
fied protein. Different drugs enhance the ATPase activity of abundance in the liver, intestines, and kidney-sites where
these liposomes in a dose-dependent manner corresponding natural toxic and waste products are removed from the body.
to their ability to be transported by ABCBl. Substrates for these ABC proteins include sugars, amino acids,
The three-dimensional structures of ABCBl, together cholesterol, bile acids, phospholipids, peptides, proteins, tox-
with those of homologous bacterial ABC transporters, re- ins, and foreign substances. The normal function of ABCBl
vealed its mechanism of transport as well as its ability to bind most likely is to transport various natural and metabolic tox-
and transport a wide array of hydrophilic and hydrophobic ins into the bile or intestinal lumen for excretion or into the
substrates (Figure 11-15). The two T domains form a binding urine being formed in the kidney•. During the course of its
site in the center of the membrane that alternates between an evolution, ABCBl appears to have acquired the ability to
inward (Figure 11-lSb) and an outward (Figure ll-15c) fac- transport drugs whose structures are similar to those of these
ing orientation. The alternation between these two confor- endogenous toxins. Tumors derived from MDR-expressing
mational states of the protein is powered by ATP binding to cell types, such as hepatomas (liver cancers), frequently are
the two A subunits and subsequent hydrolysis to ADP and P, resistant to virtually all chemotherapeutic agents and are thus
but precisely how this happens is not known. difficult to treat, presumably because the tumors exhibit in-
The substrate-binding cavity formed by ABCBl is large. creased expression of ABCBl or a related ABC protein.
Some of the amino acids that line the cavity have aromatic
side chains, mainly tyrosine and phenylalanine, allowing Certain ABC Proteins "Flip" Phospholipids
ABCBl to bind multiple types of hydrophobic ligands. Other
and Other Lipid-Soluble Substrates from
segments of the cavity are lined with hydrophilic residues,
allowing hydrophilic or amphipathic molecules to bind. In One Membrane Leaflet to the Other
the inward-facing conformation the binding site is open As shown in Figure 11-15b and c, ABCBl can move, or "flip,"
directly to the surrounding aqueous solutions, allowing hy- a hydrophobic or amphipathic substrate molecule from the
drophilic molecules to enter the binding site directly from inner leaflet of the membrane to the outer leaflet. This is an
the cytosol. A gap in the protein is accessible directly from otherwise energetically unfavorable reaction powered by the

TABLE 1 1-3 Selected Human ABC Proteins

Protein Tissue Expression Function Disease Caused by Defective Protein

ABCBl ( ~lDR 1) Adrenal, kidney, brain Exports lipophilic drugs

ABCB4 (MDR2) Liver Exports phosphatidylcholine

into bile

ABCBll Liver Exports bile salts into bile

CFTR Exocrine tissue Transports Cl ions Cystic fibrosis

ABCDI Ub1quirous in peroxisomal Influences activity of peroxisomal Adrenoleukodystrophy (ADL)

membrane enzyme that oxidizes very long chain
fatty acids

ABCG5/8 Liver, intestme Exports cholesterol and other sterols 13-Sirosterolemia

ABCAl Ubiquitous Exports cholesterol and phospholipid Tangier's disease

for uptake into high-density
lipoprotein (HDL)

492 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

coupled ATPase activity of the protein. Support for this so- in the cellular export of various lipids, presumably by mecha-
called flippase model of transport by ABCBl comes from ex- nisms similar to that of ABCBl (see Table 11-3).
periments on ABCB4 (originally called MDR2), a protein ABCB4 was first suspected of having phospholipid tlippase
homologous to ABCBl that is present in the region of the activity because mice with homozygous loss-of-function muta-
liver-cell plasma membrane that faces the bile canaliculi. tions in the ABCB4 gene exhibited defects in the secretion of
ABCB4 moves phosphatidylcholine from the cyrosolic to the phosphatidylcholine into bile. To determine directly if ABCB4
exoplasmic leaflet of the plasma membrane for subsequent was in fact a flippase, researchers performed experiments on a
release into the bile in combination with cholesterol and bile homogeneous population of purified vesicles isolated from
acids, which themselves are transported by other ABC family special mutant yeast cells with ABCB4 in the membrane and
members. ~everal other ABC superfamily members participate with the cyrosolic face directed outward (Figure 11-16). After

Internal labeled lipids:

/ protected, fluorescent

II II
+ Ouencher +Detergent
+ Ouencher
+Light
Micelle:
unprotected,
all labeled lipids
quenched

ADP
Add
Add
ATP
quencher
l l
~
·c;;
c
Q)
E
Q) Add
u
cQ)
u
(J)
Q)

0
::l
u:::
Time (min)
,
EXPERIMENTAL FIGURE 11-16 In vitro fluorescence - vesicles. Dithionite reacts with the fluorescent head group, destroying
quenching assay can detect phospholipid flippase activity of its ability to fluoresce (gray). In the presence of the quencher, only
ABCB4. A homogeneous population of secretory vesicles containing labeled phospholipid in the protected environment on the inner leaflet
ABCB4 protein was obtained by introducing the eDNA encoding will fluoresce. Subsequent to the addition of the quenching agent, the
mammalian ABCB4 into a temperature-sensitive yeast sec mutant such total fluorescence decreases with time until it plateaus at the point at
that ABCB4 was localized to intracellular endoplasmic reticulum which all external fluorescence is quenched and only the internal
vesicles in its normal orientation and with the cytosolic face of the phospholipid fluorescence can be detected. The observation of greater
vesicles facing outward (see Figure 14-4). Step 0 : Synthetic phospho- fluorescence (less quenching) in the presence of ATP than in its
lipids containing a fluorescently modified head qroup (blue) were abse>nce indicates that ABCB-1 has flipped some of the labeled
incorporated primarily into the outer, cytosolic leaflets of the purified phospholipid to the inside. Not shown here are "control" vesicles
vesicles. Step 6 : If ABCB4 acted as a flippase, then on addition of ATP isolated from cells that did not express ABCB4 and that exhibited no
to the outside of the vesicles, a small fraction of the outward-facing flippase activity. Step 19: Addition of detergent to the vesicles
labeled phospholipids would be flipped to the inside leaflet. Step 10: generates micelles and makes all fluorescent lipids accessible to the
Flipping was detected by adding a membrane-impermeable quench- quenching agent, lowering the fluorescence to baseline values.
ing compound called dithionite to the medium surrounding the [Adapted from S. Ruetz and P. Gros, 1994, Cel/77:1 071.]

11.3 ATP-Powered Pumps and the Intracellular Ionic Environment 493

purifying these vesicles, investigators labeled them in vitro with mutant protein fails to fold properly and to move to the cell
a fluorescent phosphatidylcholine derivative. The fluorescence- surface where it normally functions. Interestingly, if cells ex-
quenching assay outlined in Figure 11-16 was used to dem- pressing the mutant protein are incubated at room temperature,
onstrate that the vesicles containing ABCB4 exhibited an the protein accumulates normally on the plasma membrane,
ATP-dependent flippase activity. where it functions nearly as well as the wild-type CFTR chan-
nel. Much effort is now being directed toward identifying
sma ll molecules that might allow the mutant protein in CF
The ABC Cystic Fibrosis Transmembrane
patients to traffic normally to the cell surface and thus re-
Regulator (CFTR) Is a Chloride Channel, verse the effe~.:t~ uf the disease. •
Nota Pump
Several human genetic diseases are associated with defecti ve
ABC proteins (see Table 11-3). The best studied and most KEY CONCEPTS of Section 11.3
widespread is cystic fibrosis (CF), caused by a mutation in
ATP-Powered Pumps and the Intracellular
the gene encoding the cystic fibrosis transmembrane regula-
Ionic Environment
tor (CFTR, also called ABCC7). Like other ABC proteins,
CFTR has two transmembrane T domains and two cytosolic • Four classes of transmembrane proteins couple the energy-
A, or ATP-binding, domains. CFTR contains an additional releasing hydrolysis of ATP with the energy-requiring trans-
R (regulatory) domain on the cytosolic face; R links the two port of substances against their concentration gradients: P-, V-,
homologous halves of the protein, creating an overall do- and F-class pumps and ABC proteins (sec Figure 11-9).
main organization of Tl-A 1-R-T2-A2. But CFTR is a Cl • The combined action of P-class Na+/K + ATPases in the
channel protein, not an ion pump. lt is expressed in the api- plasma membrane and homologous Cal+ ATPases in the
cal plasma membranes of epithelial cells in the lung, sweat plasma membrane or sarcoplasmic reticulum creates the
glands, pancreas, and other tissues. For instance, CFTR pro- usual ionic milieu of animal cells: high K +, low Cal+, and
tein is important for reuptake, into the cells of sweat glands, low Na + in the cytosol; low K +, high Ca 2 ... , and high Na in
of Cl- lost by sweating; babies with cystic fibrosis, if licked, the extracellular fluid.
often taste "salty" because this reuptake is inhibited.
• Jn P-class pumps, phosphorylation of the a: (catalytic ) sub-
The Cl channel of CFTR is normally shut. Channel
unit and changes in conformational states are essential for
opening is activated by phosphorylation of the R domain by
coupling ATP hydrolysis tO transport of H +, Na ... , K +, or
a protein kinase (PKA, discussed in Chapter 15) that in turn
Cal· ions (see Figures 11-10 through 11-13 ).
is activated by an increase in cyclic AMP (cAMP), a small
intracellular signaling molecule. Opening of the channel also • V- and F-class ATPases, which transport protons exclusively,
requires sequential binding of two ATP molecul es to the two are large, multisubunit complexes with a proton-conducting
A domains (Figure 11-17). channel in the transmembrane domain and ATP-binding sites
in the cytosolic domain.
About two-thirds of all CF disease cases can be attrib- • V-class H pumps in animal lysosomal and endosomal
uted to a single mutation in CFTR: deletion of Phe 508 membranes and plant vacuolar membranes are responsible
in the ATP-binding A 1 domain. At body temperature, the

FIGURE 11-17 Structure and function of

the cystic fibrosis transmembrane regulator
(CFTR). Structural interpretation of the AlP-
External CFTR
dependent gating cycle of phosphorylated CFTR face Closed channel

..
channels. The regulatory (R) domain (not
depicted) must be phosphorylated before ATP e ATP O ATP
is able to support channel opening. One ATP
(yellow circle) becomes tightly bound to the A1 \... \..
Cytosolic
domain (green). Binding of the second ATP to the face
A2 domain (blue) is followed by formation of a
tight intramolecular A1-A2 heterodimer and Al
slow channel opening. The relatively stable open
AlP-binding
state becomes destabilized by hydrolysis of the domains
ATP bound at A2 to ADP (red crescent) and P.
The ensuing disruption of the tight A1-A2 dimer
interface leads to channel closure. T=transmem-
brane domain; A=cytosolic AlP-binding domain.
(After D. C. Gadsby et. at., 2006, Nature 440:477 and
S. G. Aller, 2009, Science 323:1 718.]

.·
494 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules
 processes. As noted previously, a rise in the cytosolic Ca2t
for maintaining a lower pH inside the organelles than in the concentration is an important regulatory signal, initiatmg
surrounding cytosol (see Figure 11-14 ). contraction in muscle cells and triggering in many cells secre-
• All members of the large and diverse ABC superfamily of tion of proteins, such as digestive enzymes from pancreatic
transport proteins contain four core domains: two trans- cells. In many animal cells, the combined force of the Na
membrane domains, which form a pathway for solute move- concentration gradient and membrane electric potential
ment and determine substrate specificity, and two cytosolic drives the uptake of amino acids and other molecules against
ATP-binding domains (see Figure 11-15). their concentration gradients by symport and antiport pro-
teins (see Figure 11-3 and (\ecrinn 11. 5). Furthermore, the
• The ARC superfamily includes bacterial amino acid and
electrical signaling by nerve cells depends on the opening
sugar permeases and about 50 mammalian proteins (e.g.,
and closing of ion channels in response to changes in the
ABCBl, ABCA 1) that transport a wide array of substrates,
membrane electric potential (Chapter 22).
including toxins, drugs, phospholipids, peptides, and pro-
Here, we discuss the origin of the membrane electric po-
teins, into or out of the cell.
tential in resting non-neuronal cells, often called the cell's
• The two T domains of the multidrug transporter ABCBl "resting potential"; how ion channels mediate the selective
form a ligand-binding site in the middle of the plane of the movement of ions across a membrane; and useful experi-
membrane; ligands can bind directly from the cytosol or from mental techniques for characterizing the functional proper-
the inner membrane leaflet through a gap in the protein. ties of channel proteins.
• Biochemical experiments directly demonstrate that ABCB4
(MDR2) possesses phospholipid flippase activity (see Selective Movement of Ions Creates
Figure 11-16 ).
a Transmembrane Electric Gradient
• CFTR, an ABC protein, is a Cl- channel protein, not an
To help explain how an electric potential across the plasma
ion pump. Channel opening is triggered by protein phos-
membrane can arise, we first consider a set of simplified ex-
phorylation and by binding of ATP to the two A domains
perimental systems in which a membrane separates a 150 mM
(Figure 11-17).
NaCI/15 mM KCI solution (similar to the ~xtracellular me-
dium surrounding metazoan cells) on the right from a 15 mM
NaCl/1 50 mM KCI solution (similar to that of the cytosol) on
the left (Figure 11- 18a ). A potentiometer (voltmeter) is con-
11.4 Nongated Jon Channels and the nected to both solutions to measure any difference in electric
potential across the membrane. If the membrane is imperme-
Resting Membrane Potential
able to all ions, no ions will flow across it. Initially both so-
In addition to ATP-powered ion pumps, which transport ions lutions contain an equal number of positive and negative
against their concentration gradients, the plasma membrane ions. Furthermore, there will be no difference in voltage, or
contains channel proteins that allow the principal cellular ions electric potential gradient, across the membrane, as shown in
(Na +, K+, Ca 2 , and Cl -) to move through them at different Figure 11-18a.
rates down thei~ concentration gradients. Ion concentration Now suppose that the membrane contains Na -channel
gradients generated by pumps and selective movements of proteins that accommodate Na tons but exclude K and Cl
ions through channels constitute the principal mechanism by ions (Figure 11-18b). Na ions then tend to move down their
which a difference in voltage, or electric potential, is generated concentration gradient from the right side to the left, leanng
across the plasma membrane. In other words, ATP-powered an excess of negative Cl- ions compared with Na ions on the
ion pumps generate differences in ion concentrations across right side and generating an excess of positive Na ions com-
the plasma membrane, a11d ion channels utilize these concen- pared with Cl- ions on the left side. The excess Na on the left
tration gradients tO generate a tightly controlled electric po- and Cl on the right remain near the respective surfaces of the
tential across the membrane (see Figure 11-3). membrane because the excess positive charges on one side of
In all cells the magnitude of this electric potential gener- the membrane are attracted to the excess negative charges on
ally is -70 millivolts (m V), with the inside cytosolic face of the other side. The resulting separation of charge across the
the cell membrane always negative with respect to the exo- membrane constitutes an electric potential, or voltage, with the
plasmic face. This value does not seem like much until we left (cytosolic) side of the membrane having excess positive
consider that the thickness of the plasma membrane is only charge with respect to the right.
-3 .5 nm. Thus the voltage gradient across the plasma mem- As more and more Na + ions move through channels
brane is 0.07 V per 3.5 X 10 - em, or 200,000 volts per across the membrane, the magnitude of this charge difference
centimeter! (To appreciate what this means, consider that (i.e., voltage) increases. However, continued nght-to-lcft
high-voltage transmission lines for electricity utilize gradi- movement of the Na ions eventually is inhibited by the mu-
ents of about 200,000 volts per kilometer, 10 5-fold less!) tual repulsion between the excess positive (Na +)charges ac-
The ionic gradients and electric potential across the cumulated on the left side of the membrane and by the
plasma membrane play crucial roles in many biological attraction of Na + ions to the excess negative charges built up

11.4 Nongated lon Channels and the Resting Membrane Potential 495
(a) Membrane impermeable to Na , K'", and Cl EXPERIMENTAL FIGURE 11-18 Generation of a transmem-
brane electric potential (voltage) depends on the selective
0 movemen t of ions across a semipermeable membrane. In this

Potentiometer _ _ , 6~ 60 potential
Memb~ane electric 0
experimental system, a membrane separates a 15 mM NaCI/150 mM
KCI solution (left) from a 150 mM NaCI/1 5 mM KCI solution (right); these
ion concentrations are similar to those in cytosol and blood, respec-
tively.lf the membrane separating the two solutions is impermeable to

Cell cytosol
medium
!
Extracellular~;
all ions (a), no ions can move across the membrane and no difference
in electric potential is registered on the potentiometer connecting the
two soluttons.lfthe membrane is selectively permeable only to Na+
(b) or to K (c), then diffusion of ions through their respective channels
leads to a separation of charge across the membrane. At equ ilibrium,
15mM 150mM the membrane potential caused by the charge separation becomes
Na... CI Na+CI
equal to the Nernst potential ENa or EK registered on the potentiometer.
150mM 15 mM See the text for further explanation.
K'"CI
Cytosolic Exoplasmic
face face
'-- I
___-#'
on the right side. The system soon reaches an equilibrium
point at which the two opposing factors that determine the
movement of Na ... ions-the membrane electric potential
(b) Membrane permeable only to Na and the ion concentration gradient-balance each other out.
At equilibrium, no net movement of Na + ions occurs across
Membrane electric potential -
+59 mV, cytosolic face of the the membrane. Thus this membrane, like all biological mem-
membrane posit ive with branes, acts as a capacitor-a device consisting of a thin
respect to the exoplasmic face. sheet of nonconducting material (the hydrophobic interior)
surrounded on both sides by electrically conducting material
(the polar phospholipid head groups and the ions in the sur-
rounding aqueous solution) that can store positive charges
on one side and negative charges on the other.
If a membrane is permeable only to Na + ions, then at
equilibrium the measured electric potential across the mem-
Na-channel
brane equals the sodium equilibrium potential in volts, £"•·
The magnitude of ENa is given by the Nernst equation, which
is derived from basic principles of physical chemistry:

(11-2)
Charge separation across membrane

where R (the gas constant) = 1.987 cal/(degree ·mol ), or

(c) Membrane permeable only to K 8.28 joules/(degree ·mol); T (the absolute temperature in
z
degrees Kelvin ) = 293 °K at 20 °C; (the charge, also called
0 the valency) here equal to + 1; F (the Faraday constant) =
Membrane electric pot ential =

,~,
59 mV, cytosolic face of the 23,062 cal/(mol · Y), or 96,000 coulombs/(mol · V); and
membrane negative with (Na,ef,] and [Na,;ghrl are the Na+ concentrations on the left
respect to the exoplasmic face.
and right sides, respectively, at equilibrium. By convention
the potential is expressed as the cytosolic face of the mem-
- + brane relative to the exoplasmic face, and the equation is
+
+ written with the concentration of ion in the extracellular so-
lution (here the right side of the membrane) placed in the
numerator and that of the cytosol in the denominator.
At 20 oc, Equation 11-2 reduces to
K K+ channel

(11-3)

If INanghrllfNaletrl = 10, a tenfold ratio of concentrations as

Charge separation across membrane in Figure ll-18b, then E:-.~. = +0.05 9 V (or + 59 mY), with

496 CHAPTER 11 • Transmembrane Transport of Ion s and Small Molecules

 the left, cytosolic side positive with respect to the exoplasmic 0
right side. Potentiometer
If the membrane is permeable only to K+ ions and not to
Na or Cl ions, then a similar equation describes the potas-
sium equilibrium potential EK:
·.
_ [Knght] Microelectrode
E~-: = 0.0.)9loglo -- ( 11-4) Reference electrode
filled with in contact with
Kieft l conductin~ bathing medium
The magrutude of the membrane electric potential ts the salt solution
same (59 mY, for a tenfold difference in ion concentrations), ~
except that the left, cytosolic side is now negative with respect Bathing medium
to the right (Figure 11-18c), opposite to the polarity obtained
t + + + + t + +
across a membrane selectively permeable to Na _,. ions.

Cytosol
The Resting Membrane Potential in Animal
Cells Depends Largely on the Outward Flow EXPERIMENTAL FIGURE 1 1·19 The electric potential across
of K+ Ions Through Open K+ Channels the plasma membrane of living cells can be measured. A microelec-
trode, constructed by filling a glass tube of extremely small diameter
The plasma membranes of animal cells contain many open K_,.
with a conducting fluid such as a KCI solution, is inserted into a cell in
channels but few open Na , Cl , or Ca2 ; channels. As a result, such a way that the surface membrane seals itself around the tip of the
the major ionic movement across the plasma membrane is that electrode. A reference electrode is placed in the bathing medium. A
of K+ from the inside outward, powered by the K concentra- potentiometer connecting the two electrodes registers the potential,
tion gradient, leaving an excess of negative charge on the cyto- in this case - 60 mV with the cytosolic face negative with respect to the
solic face of the plasma membrane and creating an excess of exoplasmic face of the membrane. A potential difference is registered
positive charge on the exoplasmic face, similar to the experi- only when the microelectrode is inserted into the c'ell; no potential is
mental system shown in Figure 11-18c. This outward flow of registered if the microelectrode is in the bathmg fluid.
K+ ions through these channels, called resting K+ channels, is
the major determinant of the inside-negative membrane poten- ton pumps, a process similar to what occurs in lysosomal
tial. Like all channels, these alternate between an open and a membranes lacking Cl channels (see Figure ll-14a ): each
closed state (Figure 11-2), but since their opening and closing H pumped out of the cell leaves behind a Cl ion, generat-
are not affected by the membrane potential or by small signal- ing an electric potential gradient (cytosolic face negative )
ing molecules, these channels are called nongated. The various across the membrane. In aerobic bacterial cells the inside
gated channels in nerve cells (Chapter 22) open only in re- negative potential is generated by outward pumping of pro-
sponse to specific ligands or to changes in membrane potential. tons during electron transport, a process similar to proton
Quantitatively, the usual resting membrane potential of pumping in mitochondrial inner membranes that will be dis-
-70 m Vis close to the potassium equilibrium potential, cal- cussed in detail in Chapter 12 (sec Figure 12-16 ).
culated from the Nernst equation and the K_,. concentrations The potential across the plasma membrane of large cells can
in cells and surrounding media depicted in Table 11 -2. Usu- be measured with a microelectrode inserted inside the cell and
ally the potential is lower (less negative) than that calculated a reference electrode placed in the extracellular fluid. The two
from the Nernst equation because of the presence of a few are connected to a potentiometer capable of measuring small
open Na channels. These open Na ... channels allow the net potential differences (figure 11-19). The potential across the
inward flow of Na ion$, making the cytosolic face of the surface membrane of most animal cells generally does not vary
plasma membrane more positive, that is, less negative, than with time. In contrast, neurons and muscle cells-the principal
predicted by the Nernst equation forK .... The K concentra- types of electrically active cells-undergo controlled changes in
tion gradient that drives the flow of ions through resting K their membrane potential, as we discuss in Chapter 22.
channels is generated by the Na + /K ATPase described pre-
viously (see Figures 11-3 and 11-13 ). fn the absence of this
pump, or when it is inhibited, the K+ concentration gradient lon Channels Are Selective for Certain Ions
cannot be maintained, the magnitude of the membrane po- by Virtue of a Molecular //Selectivity Filterll
tential falls to zero, and the cell eventually dies. All ion channels exhibit specificity for particular ionc;: K
Although resting K+ channels play the dominant role in channels allow K+ but nor closely related Na • ions to enter,
generating the electric potential across the plasma membrane whereas Na + channels admit Na + but not K . Determination
of animal cells, this is not the case in bacterial, plant, and of the three-dimensional structure of a bacterial K+ channel
fungal cells. The inside-negative membrane potential in plant first revealed how this exquisite ion selectivity is achieved.
and fungal cells is generated by transport of positively Comparisons of the sequences and structures of other K+
charged protons (H.,.) out of the cell by ATP-powered pro- channels from organisms as diverse as bacteria, fungi, and

11.4 Nongated ion Channels and the Resting Membrane Potential 497
(a) Single subunit (b) Tetrameric channel

Selectivity
filter Exterior

P helix

Outer
helix
(55)

Cytosol

coo-
FIGURE 11-20 Structure of a resting K+ channel from the consist of a nonhelical "turret," which lines the upper part of the pore; a
bacterium Streptomyces lividans. All K... channel proteins are short a helix; and an extended loop that protrudes into the narrowest
tetramers comprising four identical subunits, each containing two part of the pore and forms the ion-selectivity filter. This filter allows K
conserved membrane-spanning a helices, called by convention 55 and (purple spheres) but not other ions to pass. Below the filter is the
56, and a shorter P, or pore segment. (a) One of the subunits, viewed central cavity or vestibule lined by the inner, or 56, a helixes. The
from the side, with key structural features indicated. (b) The complete subunits in gated K channels, which open and close in response to
tetrameric channel viewed from the side {left) and the top, or extracel- specific stimuli, contain additional transmembrane helices not shown
lular end (right). The P segments (pink) are located near the exoplasmic here; these are discussed in Chapter 22. [See Y. Zhou et al., 2001, Nature
surface and connect the 55 and 56 a helices (yellow and silver); they 414:43.]

humans established that all share a common structure and to select KT over Na + is due mainly to backbone carbonyl oxy-
probably evolved from a single type of channel protein. gens on residues located in a Gly-Tyr-G iy sequence that is
Like all other K+ channels, bacterial K ... channels are built found in an ana logous position in the P segment in every known
of four 1dentical transmembrane subunits symmetrically ar- K... channel. As a K ion enters the narrow selectivity filter-the
ranged around a central pore (Figure 11-20). Each subunit space between the P segment filter sequences contributed by the
contains two me!nbrane-spanning ex helices (55 and S6) and a four adjacent subunits-it loses its eight waters of hydration
short P (pore) segment that partly penetrates the membrane but becomes bound in the same geometry to eight backbone
bilayer from the exoplasmic surface. In the tetrameric K carbonyl oxygens, two from the extended loop in each of the
channel, the eight transmembrane ex helices (two from each four P segments lining the channel (Figure 1 L-2la, bottom left).
subunit) form an inverted cone, generating a water-filled cav- Thus little energy is required to strip off the eight waters of
Ity called the vestibule in the central portion of the channel hydration of a K- ion, and as a result, a relatively low activa-
that extends halfway through the membrane toward the cyto- tion energy is required for passage of K+ ions into the channel
solic side. Four extended loops that are part of the four P seg- from an aqueous solution. A dehydrated Na- ion is too small
ments form the actual ion-selectivity filter in the narrow part to bind to all eight carbonyl oxygens that line the selectivity
of the pore near the exoplasmic surface, above the vestibule. filter with the same geometry as a Na- ion surrounded by its
Several related pieces of evidence support the role of P norma l eight water molecules in aqueous solution. As a result,
segments in ion selection. First, the amino acid sequence of Na + ions would ''prefer" to remain in water rather than enter
the P segment is highly homologous in all known K chan- the selectivity filter, and thus the change in free energy for entry
nels and is different from that in other ion channels. Second, of Na + ions into the channel is relatively high (Figure 11-2la,
mutation of certain amino acids in this segment alters the right). This difference in free energies favors passage of K ... ions
ability of a K- channel to distinguish Na from K . Finally, through the channel over Na- by a factor of 1000. Like Na ,
replacing the P segment of a bacterial KT channel with the the dehydrated Ca 2+ ion is smaller than the dehydrated K+ ion
homologous segment from a marnm<1lian K ~ channel yields and cannot interact properly with the o:>.:y gen atoms in the se-
a chimenc protein that exhibits normal selectivity forK - lectivity filter. Also, because a Ca 2 ion has two positive charges
over other ions. Thus all K+ channels are thought to use the and binds water oxygens more tightly than does a single posi-
same mechanism to distinguish K from other ions. tive Na or K ion, more energy is required to strip the waters
Na 1ons are smaller thanK ... ions. How, then, can a chan- of hydration from Cah than from K +-or Na .
nel protein exclude smaller Na + ions, yet allow passage of Recent x-ray crystallographic studies reveal that both
larger K ? The ability of the ion-selectivity filter inK ... channels when open and when closed, the channel contains K+ ions

498 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

(a) K+ and Na+ ions in the pore of a K+ channel (top view) FIGURE 11 -21 Mechanism of ion selectivity and transport in
rest ing K+ channels. (a) Schematic diagrams of K and Na ions
K· in water Na in water
hydrated in solution and in the pore of a K+ channel. AsK... ions pass
-Q ;; ?--H through the selectivity filter, they lose their bound water molecules and
0 0 ..) 0 0 ~H become coordinated instead to eight backbone carbonyl oxygens, four
'-"
~ Q of which are shown, that are part of the conserved amino acids in the
channel-lining selectivity filter loop of each P segment. The smaller Na
0 0 j '-1 0 \ 0 ,J ions with their tighter shell of water molecules cannot perfectly
"'' :5 J "-0 coordinate with the channel oxygen atoms and therefore pass through
the channel only rarely. (b) High-resolution electron density map
K+in K pore Na• inK pore obtained from x-ray crystallography showing K ions (purple spheres)
---.., passing through the selectivity filter. Only two of the diagonally opposed
0

0 0
0

-
~
0

0
0

0
channel subunits are shown. Within the selectivity filter each unhydrated
K ion interacts with eight carbonyl oxygen atoms (red sticks) lining the
channel, two from each of the four subunits, as if to mimic the eight
waters of hydration. (c) Interpretation of the electron density map
showing the two alternating states by which K ions move through the
channel. In state 1, numbered top-to-bottom from the exoplasmic side
(b) K... ions in the pore of a K· channel (side view) of the channel inward, one sees a hydrated K ion with its eight bound
water molecules, K+ ions at positions 1 and 3 within the selectivity filter,

Exoplasmic
face
• • and a fully hydrated K ion within the vestibule. During K movement
each ion in state 1 moves one step inward, forming state 2. Thus in state
2 the K ion on the exoplasmic side of the channel has lost four of its
eight waters, the ion at position 1 in state 1 has moved to position 2, and
the ion at position 3 in state 1 has moved to position 4. In going from
state 2 to state 1, the K at position 4 moves into the vestibule and picks
up eight water molecules, while another hydrated K. ion moves into the
channel opening and the other K... ions move down one step. Note that
K... ions are shown here moving from the exoplasmic side of the channel
to the cytosolic side because that is the normal direction of movement in
bacteria. In animal cells the direction of K. movement is typically the
reverse-from inside to outside. [Part (a) adapted from C. Armstrong, 1998,
Science 280:56. Parts (b) and (c) adapted from Y. Zhou et al., 2001 , Nature 414:43.]

Vestibule ~- - - Water
e-K within the selectivity filter; without these ions the channel

•• probably would collapse. The K . ions are thought to be

present either at positions 1 and 3 or at 2 and 4 , each sur-
rounded by eight carbonyl oxygen atoms (Figure 11-21b and
(c) lon movement through selectivity filter
c). Several K ions move simultaneously through the channel
such that when the ion on the exoplasmic face that has been
•••• partially stripped of its water of hydration moves into posi-

•••• • • •••
tion 1, the ion at position 2 jumps to position 3 and the one
at position 4 exits the channel (Figure ll-2lc) .

0
• Although the amino acid sequences of the P segments in
Na • and K + channels differ somewhat, they are similar
enough to suggest that the general structure of the ion-selec-
tivity filters are comparable in both types of channels. Pre-
sumably the diameter of the filter in Na • channels is small
0 enough that it permits dehydrated Na + ions to bind to the
backbone carbonyl oxygens but excludes the larger K+ ions
from entering, but the first three-dimensional structure of a
••• sodium channel was determined only in late 20 11 and the

•• • mechanism of ion selectivity is only now being detennim:d.

State 1 State 2 Patch Clamps Permit Measurement of lon

Movements Through Single Channels
Once it was realized that in most cells there are only one or a
few ion channels per square micrometer of plasma membrane,

11.4 Nongated lon Channels and the Resting Membrane Potential 499
 (a) ,--------~ EXPERIMENTAL FIGURE 11-22 Current fl ow through
.-------""1 Device to maintain constant II individual ion channels can be measured by patch-clamping •
v oltage across m embrane and (a) Basic experimental arrangement for measuring current flow
to measure current flow across
through individual ion channels in the plasma membrane of a living
membra ne at ti p of patch
electrode cell. The patch electrode, filled with a current-conducting saline
solution, is applied, with a slight suction, to the plasma membrane. The
0.5-fJ..m·diameter tip covers a region that conta ins on ly one or a few ion
channels. The second electrode is inserted through the membrane into
Patch electrode filled the cytosol. A recording device measures current flow only through the
w1th conducting salt channels in the patch of plasma membrane. (b) Photomicrograph of
solution the cell body of a cultured neuron and the tip of a patch pipette
Intracellular
/ ' / ion channels
electrode touching the cell membrane. (c) Different patch-clamping configura-
tions. Isolated, detached patches are the best configurations for
L ,t-11-lt-IHI-It-IH I-II-I I /,..lt-lt-lt-IHI-It-lt-lt-11-1MI-11-IMI-It-IHI-II-II-II-i

I Cytosol Intact cell studying the effects on channels of different ion concentrations and
solutes such as extracellular hormones and intracellular second
messengers (e.g., cAMP). An inside-out patch, in which an on-cell
(b) patch is formed and then the cell pulled away, is used in the experiment
in Figure 11-23. [Part (b) from B. Sakmann, 1!192, Neuron 8:613 (Nobel lecture);
also published in E. Neher and B. Sakmann, 1992, Sci. Am. 266(3):44.
Part (c) adapted from B. Hille, 1992, lon Channels ofExcitable Membranes, 2d ed.,
Sinauer Associates, p. B9.]

The inward or outward movement of ions across a patch

of membrane is quantified from the amount of electric current
needed to maintain the memb rane potential at a particular
"clamped" value (Figure 11-22a and b). To preserve electro-
neutrality and to keep the membrane potential constant even
(c) Tip of micropipette
though ions are moving through channels in the membrane

(-r--~---- ion channel

patch, the entry of each positive ion (e.g., a Na .... ion) into the
cell through a channel in the patch of membrane is balanced

"-,~
by the addition of an electron into the cytosol through a mi-
croelectrode inserted into the cytosol; an electronic device
measures the numbers of electrons (cu rrent) required to coun-
On-cell pat ch measures indirect effect of extracellular solutes
on channels within membrane patch on intact cell terbalance the inflow of ions through the membrane channels.
Conversely, the exit of each positive ion from the cell (e.g., a
Cytosolic . ~ Exoplasmic ~ K+ ion) is balanced by the withdrawal of an electron from the
face--......__~ face '-........['(,1 cytosol. The patch-clamping technique can be employed on
whole cells or isolated membrane parches to measure the ef-
tr

'
fects of different substances and ion concentrations on ion
~ flow (Figure 11-22c).
Inside-out det ached pat ch Outside-out detached patch The patch-clamp tracings in Figure 11-23 illustrate the
measures effects of intra- measures effects of extra- use of this technique to study the properties of voltage-gated
cellular solutes on channels cellular solutes on channels
within isolated patch within isolated patch Na • channels in the plasma membrane of muscle cells. As we
discuss in Chapter 22, these channels normally are closed in
resting muscle cells and open following nervous stimulation.
it became possible to record ion movements through single Patches of muscle membrane, each containing on average one
ion channels, and to measure the rates at which these channels Na + channel, were clamped at a predetermined voltage that,
open and close and conduct specific ions, using a technique in this study, was slightly less than the resting membrane po-
known as patch clamping or voltage clamping. As illustrated tential. Under these circumstances, transient pulses of posi-
in Figure 11-22, a tiny pipette is tightly applied to the surface tive charges (Na + ions ) cross the membrane from the
of a cell; the segment of the plasma membrane within rhe rip exoplasmic to the cytosolic face as individua l Na + channels
will contain only one or a few ion channels. An electrical re- open and then close. Each channel is either fully open or
cording device detects ion Aow, measured as electric current, completely closed. From such tracings, it is possible to deter-
through the channels; this usually occurs in small bursts when mine the time that a channel is open and the ion flux through
the channel is open. An electrical device "clamps" or locks the it. For the channels measured in Figure 11-23, the flux is
electric potential across the membrane at a predetermined about 10 million Na + ions per channel per second, a typical
value (hence the term patch clamping). value for ion channels. Replacement of the NaCl within the

500 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 Microinject mRNA
D encoding channel
protein of interest

membrane
EXPERIMENTAL FIGURE 11·23 lon flux through individual
Na + channels can be calculated from patch clamp t racings. Two
1
inside-out patches of muscle plasma membrane were clamped at a
potential of slightly less than that of the resting membrane potential. -\ ~~ Newly
The patch el~ttrode contained NaCI. I he transient pulses of electric
current in picoamperes (pA), recorded as large downward deviations
(blue arrows), indicate the opening of a Na channel and movement of
fJ
lncui.Ji:lte 24-48 h for
synthesis and
movement of channel
protein to plasma
fr0
"\:;_.,~'
)
synthesized
channel
pcoto;o
positive charges (Na • ions) inward across the membrane. The smaller membrane
deviations in current represent background noise. The average current
through an open channel is 1.6 pA, or 1.6 X 1o- 12 amperes. Since

~otcod•
1 ampere = 1 cou lomb (C) of charge per second, this current is
equivalent to the movement of about 9900 Na + ions per channel per
millisecond: (1.6 X 10 12 C/s)(1 0 3 s/ms)(6 x 1023 molecules/mol) 7 Measure channel-

\:/
96,500 C/mol. [See F. J. Si9worth and E. Neher, 1980, Nature 287:447.) a protein activity by
U patch-clamping
technique

patch pipette (cor responding to the outside of the cell) with EXPERIMENTAL FIGURE 11-24 Oocyte expression assay is
KCl or choline chloride abolishes current through the chan- useful in comparing the function of normal and mutant forms of
nels, confirming that they conduct only Na- ions, not K or a channel protein. A follicular frog oocyte is first treated with
other ions. collagenase to remove the surrounding follicle cells_. leaving a denuded
oocyte, which is microinjected with mRNA encoding the channel
protein under study. [Adapted from T. P. Smith, 1988, Trends Neurosci. 11 :250.)
Novellon Channels Can Be Characterized
by a Combination of Oocyte Expression
and Patch Clamping
Cloning of human-disease-causing genes and sequencing of which establishes Na +and K +concentration gradients across
the human genome have identified many genes encoding puta- the membrane, and resting K channels which permit selec-
tive channel proteins, including 67 putative K · channel pro- tive movement only of K ' ions back down their concentra-
teins. One way of characterizing the function of these proteins tion gradient to the external medium (sec Figure 11-3).
is to transcribe a cloned eDNA in a cell-free system to prod uce • Unlike the more common gated ion channels, which open
the corresponding mRNA. Injecting this mRNA into frog only in response to various signals, these nongated K+ chan-
oocytes and taking patch clamp measurements of the newly nels are usually open.
synthesized channel protein can often reveal its function (Fig-
• The electric potential generated by the selective flow of
ure 11-24). This experimental approach is especially useful
ions across a membrane can be calculated using the Nernst
because frog oocytes normally do not express any channel
equation (see Equa tion 11-2).
proteins on their surface membrane, so only the channel under
study is present in the membrane. Tn addition, because of the • In plants and fungi, the membrane potential is maintained
large size of frog oocytes,'patch-clamping studies are techni- by the ATP-driven pumpi ng of protons from the cytosol to
cally easier to perform o n them than on smaller cells. the exterior of the cel l.
• K+ channels are assembled from four identical subunits,
each of which has at least two conserved membrane-spanning
a hel ices and a nonhelical P segment that lines the ion pore
and forms the selectivity filter (see Figure 11-20).
KEY CONCEPTS of Section 11.4
• The ion specificity of K • channel proteins is due mainly to
Nongated lon Channels and the Resting coordination of t he selected ion with eight carbonyl oxygen
Membrane Potential atoms of specific amino acids in the P segments, which low-
• An inside-negative electric potential (voltage) of about ers the activation energy for passage of the selected K com-
-70 m V exists across the plasma membrane of all cells. pared with Na or other ions (see Figure 11-21).
• The resting membrane potential in an ima l cells is the resu lt • Patch-clamping techn iq ues, which permit measurement of
of t he combined action of t he ATP-powered Na +/K • pump, ion movements through single channels, are used to determine

11.4 Nongated lon Channels and the Resting Membrane Potential 501
 to calculate the change in free energy (.lG) that occurs dur-
the ion conductivity of a channel and the effect of various ing Na entry. As mentioned earlier, two forces govern the
signals on its activity (see Figure 11-22). movement of ions across selectively permeable membranes:
• Recombinant DNA techniques and patch clamping allow the voltage and the ion concentration gradient across the
the expression and functional characterization of channel membrane. The sum of these forces constitutes the electro-
proteins in frog oocytes (see Figure 11-24 ). chemical gradient. To calculate the free-energy change, .lG,
corresponding to the transport of any ion across a mem-
brane, we need to consider the independent contributions
from each of the force~ Lu tht: decrrochemical gradient.
For example, when Na moves from outside to inside the
11.5 Cotransport by Symporters cell, the free-energy change generated from the Na concen-
tration gradient is given by
and Antiporters
In prev1ous sections we saw how ATP-powereu pumps gener- Na,"
ate ion concentration gradients across cell membranes and .lG, = RTln [ - - (11-5 )
Naour 1
how K+ channel proteins use the K+ concentration gradient to
establish an electric potential across the plasma membrane. In
At the concentrations of Na," and Na0 ur shown in Figure 11-25,
this section we see how cotransporters use the energy stored in
which are typical for many mammalian cells, .lG., the change
the electric potential and concentration gradients of Na or
in free energy due to the concentration gradient, is -1.45 kcal
H + ions to power the uphill movement of another substance,
for transport of 1 mol of Na ... ions from outside to inside a cell,
which may be a small organic molecule such as glucose or an
assuming there is no membrane electric potential. Note the free
amino acid or a different ion. An important feature of such
energy is negative, indicating spontaneous movement of Na
cotransport is that neither substance can move alone; move-
into the cell down its concentration gradient.
ment of both substances together is obligatory, or coupled.
The free-energy change generated from the membrane
Cotransporters share common features with uniporters
electric potential is given by
such as the GLUT proteins. The two types of transporters
exhibit certain structural similarities, operate at equivalent
(11-6)
rates, and undergo cyclical conformational changes during
transport of their substrates. They differ in that uniporters
where F is the Faraday constant [= 23,062 cal/(mol· V)]; and
can only accelerate thermodynamically favorable transport
E is the membrane electric potential. If E = - 70 m V, then
down a concentration gradient, whereas cotransporters can
~Gn, the free-energy change due to the membrane potential,
harness the energy released when one substance moves down
is - 1.61 kcal for transport of 1 mol of Na + ions from outside
its concentration gradient to drive the movement of another
to inside a cell, assuming there is noNa ~ concentration gradi-
substance against its concentration gradient.
ent. Since both forces do in fact act on Na ~ ions, the total ~G
When the transported molecule and cotransported ion
is the sum of the two partial values:
move in the same direction, the process is called symport;
when they move in opposite directions, the process is called
.lG = .l Gc + ~Gm = (-1.45) + (-1.61) = -3.06 kcal/mole
antiport (see Figure 11-2). Some cotransporters transport only
positive ions (cations), while others transport only negative
In this example, the Na _,. concentration gradient and the mem-
ions (anions). Yet other cotransporters mediate movement of
brane electric potential contribute almost equally to the total
both cations and anions together. Cotransporters are present
In all orgamsms, including bacteria, plants, and animals, and
.lG for transport of Na + ions. Since .l G is < 0, the inward
movement of Na + ions is thermodynamically favored. As dis-
in this section we describe the operation and function of sev-
eral physiologically important symporters and antiporters. cussed in the next section, the inward movement of Na1 is
used to power the uphill movement of other ions and several
types of small molecules into or out of animal cells. The rapid,
Na + Entry into Mammalian Cells Is energetically favorable movement of Na + ions through gated
Thermodynamically Favored Na channels also is critical in generating action potentials in
nerve and muscle cells, as we discuss in Chapter 22.
Mammalian cells express many types of Na -coupled sym-
porters. The human genome encodes literally hundreds of
different types of transport proteins that use the energy Na +-Linked Symporters Enable Animal Cells
stored across the plasma membrane in the Na + concentra-
to Import Glucose and Amino Acids Against
tion gradient and in the inside-negative electric potential
across the membrane to transport a wide variety of mole- High Concentration Gradients
cules into cells against their concentration gradients. To see Most body cells import glucose from the blood down a con-
why such transporters allow cells to accumulate substrates centration gradient of glucose, utilizing GLUT proteins to
against a considerable concentration gradient we first need facilitate this transport. However, certain cells, such as those

502 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

9 OVERVIEW ANIMATION: Biological Energy lnterconversions

FIGURE 1 1 ·25 Transmembrane forces acting on Na + ions. ion concentration Membrane electric
gradient potential
As with all ions, the movement of Na ions across the plasma
membrane is governed by the sum of two separate forces-the ion
Inside Outside Inside Outside
concentration gradient and the membrane electric potential. At the
+
internal and externa I Na concentrations typical of mammalian cells, 145 mM Na ~
+
these forces usually act in the same direction, making the inward
movement of Na ions energetically favorable. -< -70 mV

t.Gc = -1.45 kcal/mol t.Gm = - 1.61 kcal/mol

lining the small intestine and the kidney tubules, need to im- Free-energy change during transport
of Na+ from outside to inside
port glucose from extracellular fluids (digestive products or
urine) against a very large concentration gradient (glucose
concentration higher inside the cell ). Such cells utilize a two- Inside Outside
Na .,. /one-glucose symporter, a protein that couples import of
one glucose molecule to the import of two Na ... ions:

2 Na\ur + glucoseour ~ 2 Na '" + glucose'"

Quantitatively, the free-energy change for the symport trans-

port of two Na ~ ions and one glucose molecule can be written

glucose'" 'Na m]
~G = RTln + 2RTln . + ] + 2FE (11-7)
fglucoseout INa out
one-glucose symporter permits cells to accumulate a very high
Thus the ~G for the overall reaction is the sum of the free- concentration of glucose relative to the external concentration.
energy changes generated by the glucose concentration gradi- This means that glucose present even at very low concentra-
ent (1 molecule transported), the Na + concentration gradient tions in the lumen of the intestine or in the kidney tubules can
(2 Na + ions transported), and the membrane potential (2 Na + be efficiently transported into the lining cells and not lost from
ions transported). As illustrated in Figure 11-25, the free energy the body.
released by movement of 1 mole of Na + ions into mammalian The two-Na +/one-glucose symporter is thought to contain
cells down its electrochemical gradient has a free-energy change, 14 transmembrane a: helices with both its N-and C-termini
~G, of about -3 kcal per mole of Na + transported. Thus the
extending into the cytosol. A truncated recombinant protein
~G for transport· of two moles of ).fa+ inward would be twice
consisting of only the five C-terminal transmembrane a heli-
this amount, or about -6 kcal. This negative free-energy ces can transport glucose independently of Na + across the
change of sodium import is coupled to the uphill transport of plasma membrane, down its concentration gradient. This por-
glucose, a process with a positive ~ G. We can calculate the tion of the molecule thus functions as a glucose uniporter. The
glucose concentration gradient, inside greater than outside, N-terminal portion of the protein, including helices 1-9, is
that can be established by the action of this Na ·-powered sym- required to couple Na binding and influx to the transport of
porter by realizing that at equilibrium for sodium-coupled glu- glucose against a concentration gradient.
cose import, /1G = 0. By substituting the values for sodium Figure 11-26 depicts the current model of transport by
import into Equation 11-7 and setting .lG = 0, we see that Na +/glucose symporters. This model entails conformational
changes in the protein analogous to those that occur in uni-
port transporters, such as GLUTl, which do not require a
0 = RTln lglucosem l - 6 kca l cotransported ion (compare to Figure 11-5). Binding of all
glucoseuur substrates to their sites on the extracellular domain is re-
and we can calculate that at equilibrium, the ratio of glucose'"/ quired before the protein undergoes the conformational
glucoseour = - 30,000. Thus the inward flow of two moles change that converts the substrate-binding sites from out-
of Na + can generate an intracellular glucose concentration ward to inward facing; this ensure~ chat inward transport of
that is - 30,000 times greater than the exterior concentra- glucose and Na ions are coupled.
tion. If only one Na" ion were imported (AG of approximately Note that cells use comparable Na ... -powered symporters
- 3 kcaUmol) per glucose molecule, then the available energy to transport substances other than glucose into the cell against
could generate a glucose concentration gradient (inside/outside) high concentration gradients. For example, several types of
of only about 170-fold. Thus by coupling the transport of Na ... /amino acid symporters a llow cells to import many amino
two Na ions to the transport of one glucose, the two-Na •t acids into the cell.

11.5 Cotransport by Symporters and Antiporters 503

@ OVERVIEW ANIMATION: Biological Energy lnterconversions

2 Na• • e Glucose
Exterior •
Na

If --. u H --&+ --11-+ ...... --11-+

Glucose Outward-facing
Cytosol conformation
" " "
FIGURE 11-26 Operational model for the two-Na +/one-glucose
Occluded
conformation

symporter. Simultaneous binding of Na + and glucose to the confor-

Inward-facing
conformation • • Outward-facing
conformation

sites. Dissociation of the bound Na · and glucose into the cytosol

(step B J allows the protein to revert to its original outward-facing
mation with outward-facing binding sites (step OJ causes a conforma- conformation (step 1.'.1), ready to transport additional substrate.
tional change in the protein such that the bound substrates are [See H. Krishnamurthy et al., 2009, Nature 459:347-355 for details on the
transiently occluded, unable to dissociate into either medium (step f) ). structure and function of this and related Na -coupled transporters.]
In step iJ the protein assumes a third conformat ion with inward-facing '

A Bacterial Na +IAmino Acid Symporter surrounding extracellular or cytoplasmic media. This struc-
..
Reveals How Symport Works ture represents an intermediate in the transport process (see
Figure 11-26) in which the protein appears to be changing from
No three-dimensional structure has yet been determined for a conformation with an exoplasmic- to o ne with a cytosolic-
any mammalian sodium symporter, but the structures of sev- facing binding site.
eral homologous bacterial sodium/substrate symporters have
provided considerable information abou t symport function.
The bacterial two-Na /one-leucine symporter shown in Fig- A Na+-Linked Ca2+ Antiporter Regulates
ure ll-27a consists of 12 membrane-spanning a helices.
Two of the helices (numbers l and 6) have nonhelical seg- the Strength of Cardiac Muscle Contraction
ments in the middle of the membrane that form part of the In all muscle cells, a rise in the cytosolic Ca1 " concentration
leucine-binding site. triggers contraction. In cardiac muscle cells a three-Na+l011e-
Amino acid residues involved in binding the leucine and Ca2+ antiporter, rather than the plasma membrane Ca 2
the two Na ions are located in the middle of the membrane- ATPase discussed earlier, plays the principal role in maintain-
spanning segment (as depicted for the two-Na .. /one-glucose ing a low concentration of Ca1 - in the cytosol. The transport
symporter in Figure ll-26) and are close together in three- reaction mediated by this cation antiporter can be written
dimensional space. This demonstrates that the coupling of
amino acid and ion transport in these transporters is the con-
sequence of direct or nearly direct physical interactions of the
substrates. Indeed, one of the Na ions is bound to the car- Note that the inward movement of three Na _._ ions is re-
boxyl group of the transported leucine (Figure ll -27b). Thus quired to power t he export of one Ca 2 ion from the cytosol,
neither substance can bind to the transporter without the with a !Ca 1 ' ]of -2 X 10 ~ M, to the extracellular medium,
other, indicating how transport of sodium and leucine are with a [Ca 1+1 of -2 X 10 ·1 M, a gradient of some 10,000-
coupled. Each of the two Na + ions is bound to six oxygen fold (higher on the outside). By lowering cytosolic Ca 2 , op-
atoms. Sodium I, for example, is bound to carbonyl oxygens eration of the Na /Ca 1 anti porter reduces the strength of
c

of several transporter amino acids as well as to carbonyl oxy- heart muscle contraction.
gens and the hydroxyl O:>..')'gen of one threonine. Equally im-
portantly, there are no water molecules surroundi ng either of .. The Na /K ATPase in the plasma membrane of car-
the bound Na ... atoms, as is the case forK ' ions in potassium diac muscle cells, as in other body cells, creates the Na ~
channels (see Figure 11-21 ). Thus as the Na · io ns lose their concentratio n gradient necessary for export of Ca 1 ~ by the
water of hydration in binding to rht: transporter, they bind ro Na -linked Ca 1 • antiporrer. As mentioned earlier, inhibition
SIX oxygen atoms with a similar geometry. This reduces the of the Na +/K ATPase by the drugs ouabain and digoxin
energy change required for binding of Na' ions and prevents lowers the cytosolic K" concentration and, more relevant
other ions, such as K.. , from binding in place of Na-+. here, simultaneously increases cytosolic Na-+. The resulting
One striking feature of the structure depicted in Figure reduced Na + electrochemical gradient across the membrane
11 2 7 is that the bound Na ions and leucine are occluded- causes the Na -linked Ca 2 antiporter to function less effi-
that is, they cannot diffuse out of the protein to either the ciently. As a result, fewer Ca 2 io ns arc expo rted and the

504 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

9 PODCAST: The Two-Na +/One-Leucine Symporter

(a)
Heli

Exoplasmic
face

Membrane

Cytosolic
face

Helix 8

FIGURE 11 -27 Three-dimensional structure of the two-Na +/one- oxygen atoms or carboxyl side-chain oxygens {red) that are part of
leucine symporter from the bacterium Aquifex aeo/icus. {a) The helices 1 {brown), 6 {blue). or 8 {orange). It is important that one of the
bound L-leucine, two Na ions, and a Cl ion are shown in yellow, sodium ions is also bound to the carboxyl group of the transported
purple and green, respectively. The three membrane-spanning a leucine (part b). [From A. Yamashita et al., 2005, Narure 437:215; see also
helices that bind the Na ... or leucine are colored brown, blue, and H. Krishnamurthy et al., 2009, Nature 459:347-355 for details on the structure
orange. {b,c) Binding of the two Na ... ions to carbonyl main-chain and function of this and related Na- -coupled transporters.)

cytosolic Ca 2 • concentration increases, causing the muscle tion of the imported HC0 3 - ions into C0 2 and an OH
to contract more strongly. Because of their ability to increase (hydroxyl) ion:
the force of heart muscle contractions, drugs such as oua-
bain and digoxin that inhibit the Na + fK ... ATPase are widely Carbomc
used in the treatment of congestive heart failu re. • anhydrase

The OH ions combine with intracellular protons, forming

Several Cotransporters Regulate Cytosolic pH water, and the C02 diffuses out of the cell. Thus the overall
The anaerobic metabolism of glucose yields lactic acid, and action of this transporter is to consume cytosolic H - ions,
aerobic metabolism yields C0 2, which combines with water thereby raising the cytosolic pH. Also important in raising
to form carbonic acid (H2 C03 ). These weak acids dissociate, cyrosolic pH is aNa+IH+ antiporter, which couples entry of
yielding H+ ions (protons); if these excess protons were not one Na + ion into the cell down its concentration gradient to
removed from cells, the cytosolic pH would drop precipi- the export of one H + ion.
tously, endangering cellular functions. Two types of cotrans- Under certain circumstances, the cytosolic pH can rise
port proteins help remove some of the "excess" protons beyond the normal range of 7.2-7.5. To cope with the excess
generated during metabolism in animal cells. One is a OH- ions associated with elevated pH, many animal cells
Na · HC01- !Cr antiporter, which imports one Na ... ion to- utilize an anion antlporter that catalyzes the one-for-one ex-
gether with one HC03 , in exchange for export of one Cl- ion. change of HC0 3 - and Cl- across the plasma membrane. At
The cytosolic enzyme carbonic anhydrase catalyzes dissocia- high pH, this cr
/HC03- antiporter exports one molecule of

11.5 Cotransport by Symporters and Antiporters 505

HC0 3 - (which can be viewed as a "complex" of OH and conformation that enables a histidine side chain of a globin
C0 2 ) in exchange for import of one molecule of Cl -, thus polypeptide to bind a proton. Thus when red cells are in
lowering the cytosolic pH. The import of Cl down its con- systemic capillaries water is split into a proton that binds
centration gradient (CI medium > Cl cvro>o1; see Table 11-2) hemoglobin and an OH that reacts with C0 2 to form an
powers the export of HC0 3 -. . HC0 3 anion.
The activity of all three of these anti port proteins is regu- In a reaction catalyzed by the red cell anti porter AE 1,
lated by the cytosolic pH, providing cells with a finely tuned cytosolic HC03 is transported out of the erythrocyte in ex-
mechanism for controlling cytosolic pH. The two antiporters change fo r an entering Cl anion:
that operate to mcrease cytosolic pH are activated when the
pH of the cytosol falls. Similarly, a rise in pH above 7.2 HC03 m + Cl out ~ HC01 out+ Cl on
stimulates the Cl-/HC0 3 - anti porter, leading to a more
rapid export of HC0 1 and a drop in the cytosolic pH. In (sec Figure 11-28a). The entire anion-exchange process is
this manner, the cytosolic pH of growing cells is maintained completed within 50 milliseconds (ms), during which time
very close to pH 7.4. 5 X 10 9 HC01 ions are exported from each cell down its
concentration gradient. If anion exchange did not occur,
during periods such as exercise, when much C0 2 is gener-
An Anion Anti porter Is Essential for Transport
ated, HC0 3 would accumulate 'inside the erythrocyte to
of C02 by Red Blood Cells toxic levels as the cytosol would become alkaline. The ex-
Transmembrane anion exchange is essential for an impor- change of HC0 3 (equal to OH- + C0 2 ) for Cl causes the
tant function of erythrocytes-the transport of waste C0 2 cytosolic pH to remain nearly neutral. Normally about 80
from peripheral tissues to the lungs for exhalation. Waste percent of the C0 2 in blood is transported as HC0 3- gener-
col released from cells into the capillary blood freely dif- ated inside erythrocytes; anion exchange allows about two-
fuses across the erythrocyte membrane (Figure 11-28a). In thirds of this HC0 3 - to be transported by blood plasma
its gaseous form, C0 2 dissolves poorly in aqueous solutions, external to the cells, increasing the amount of col that can
such as the cytosol or blood plasma, as is apparent to anyone be transported from tissues to the lungs. In the lungs, where
who has opened a bottle of a carbonated beverage. How- carbon dioxide leaves the body, the overall direction of this
ever, the large amount of the potent enzyme carbonic anhy- anion-exchange process is reversed (Figure ll-28b).
drase inside the erythrocyte combines col with hydroxyl AEl cata lyzes the precise one-for-one sequential ex-
ions (OH ) to form water-soluble bicarbonate (HC0 1-) an- change of anions on opposite sides of the membrane required
ions. This process occurs while red cells are in systemic (tis- to preserve electroneutra lity in the cell; only once every
sue) capillaries and releasing oxygen into the blood plasma. 10,000 or so transport cycles does an anion move unidirec-
Release of oxygen from hemoglobin induces a change in its tionally from one side of the membrane to the other. AEl is

(a) In systemic capillaries (b) In pulm onary capillaries

High C0 2 pressure Low C02 pressure
Low 0 2 pressure High 0 2 pressure
Hemoglobin

+----- HC03-
C.,boni<
•nhyd"" A
.~

Erythrocyte AE1 protein

plasma membrane HC03 Cl
FIGURE 11 -28 Carbon dioxide transport in blood requires a HC03 ions across the membrane. The overall reaction causes HC03 to
CI-/HC03 - anti porter. (a) In systemic capillaries, carbon dioxide gas be released from the cell, which is essential for maximal C02 transport
diffuses across the erythrocyte plasma membrane and is converted from the tissues to the lungs, and for maintaining pH neutrality in the
into soluble HC03 - by the enzyme carbonic anhydrase; at the same blood ceiL (b) In the lungs, where carbon dioxide is excreted, the
time, oxygen leaves the cell and hemoglobin binds a proton. The anion overall reaction is reversed. See text for additional discussion.
antiporter AEl (purple) catalyzes the reversible exchange of Cl and

506 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 composed of a membrane-embedded domain, folded into at The proton electrochemical gradient across the plant
least 12 transmembrane a helices, that catalyzes anion trans- vacuole membrane is used in much the same way as the Na
port, and a cytosolic-facing domain that anchors certain cy- electrochemical gradient across the animal-cell plasma mem-
toskelctal proteins to the membrane (see Figure 17-21). brane: to power the selective uptake or extrusion of ions and
small molecules by various antiporters. In the leaf, for ex-
ample, excess sucrose generated during photosynthesis in the
.· day is stored in the vacuole; during the night, the stored su-
Numerous Transport Proteins Enable Plant crose moves into the cytoplasm and is metabolized to col
Vacuoles to Accumulate Metabolites and Ions and H 2 0 with concomitant generation of ATP from ADP
The lumen of plant vacuoles is much more acidic (pH 3-6) and P,. A proton/sucrose antiporter in the vacuolar mem-
than is the cytosol (pH 7.5). The acidity of vacuoles is main- brane operates to accumulate sucrose in plant vacuoles. The
tained by a V-class ATP-powered proton pump (sec Figure inward movement of sucrose is powered by the outward
11-9) and by a pyrophosphate-powered proton pump that is movement of H , which is favored by its concentration
unique to plants. Both of these pumps, located in the vacu- gradient (lumen > cytosol) and by the cytosolic-negative
olar membrane, import H+ ions into the vacuolar lumen potential across the vacuolar membrane (see Figure 11-29).
against a concentration gradient. The vacuolar membrane Uptake of Ca 2 + and Na + into the vacuole from the cytosol
also contains Cl and N0 3 - channels that transport these against their concentration gradients is similarly mediated by
anions from the cytosol into the vacuole. Entry of these an- proton antiporters.
ions against their concentration gradients is driven by the
inside-positive potential generated by the H pumps. The &:;1 Understanding of the transport proteins in plant vacu-
combined operation of these proton pumps and anion chan- ~ olar membranes has the potential for increasing agri-
nels produces an inside-positive electric potential of about cultural production in high-salt (NaCI) soils, which are
20 mV across the vacuolar membrane and also a substantial found throughout the world. Because most agriculturally
pH gradient (Figure 11-29). useful crops cannot grow in such saline soils, agricultural
scientists have long sought to develop salt-t~lerant plants by
traditional breeding methods. With the availability of the
cloned gene encoding the vacuolar Na /H antiporter, re-
searchers can now produce transgenic plants that overex-
W -pumping proteins press this transport protein, leading to increased sequestration
of Na + in the vacuole. For instance, transgenic tomato plants

,-----A~o~----~
that overexpress the vacuolar Na +/H + antiporter can grow,
flower, and produce fruit in the presence of soil NaCI con-
centrations that kill wild-type plants. Interestingly, although

II 2W w
the leaves of these transgenic tomato plants accumulate large
amounts of salt, the fruit has a very low salt content. •

Jon-channel B el Plant vacuole lumen

(pH = 3-6)
KEY CONCEPTS of Section 11 .5
Cotransport by Symporters and Anti porters
• The electrochemical gradient across a semipermeable mem-
brane determines the direction of ion movement through trans-
Cytosol membrane proteins. The two forces constituting the electro-
H H (pH = 7.5) chemical gradient-the membrane electric potential and the ion
Proton antiport proteins concentration gradient-may act in the same or opposite direc
FIGURE 11 - 29 Concentration of ions and sucrose by the plant tions (see Figure 11-25).
vacuole. The vacuolar membrane contains two types of proton pumps Cotransporters use the energy released by movement of an
(orange): a V-class H ATPase (left) and a pyrophosphate-hydrolyzing ion (usually H+ or Na "' ) down its electrochemical gradient
proton pump (right) that differs from all other ion transport proteins
to power the import or export of a small molecule or differ-
and probably is unique to plants. These pumps generate a low luminal
ent ion against its concentration gradient.
pH as well as an inside-positive electric potential across the vacuolar
membrane owing to the inward pumping of H+ ions. The inside-positive • The cells lining the small intestine and kidney tubules con-
potential powers the movement of Cl and N0 3 from the cytosol tain symport proteins that couple the energetically favorable
through separate channel proteins (purple). Proton antiporters (green), entry of Na to the import of glucose against its concentra-
powered by theW gradient, accumulate Na · , Cah , and sucrose inside tion gradient (see Figure 11-26). Amino acids also enter cells
.· the vacuole. [After B. J. Barkla and 0. Pantoja, 1996, Rev. Plant Physiol. Plant Mol. by Na -coupled symporters.
Btol. 47:159- 184 and P. A. Rea et al., 1992, Trends Biochem. Sci. 1 7:348.]

11.5 Cotransport by Symporters and Ant1porters 507

 The molecular structure of a bacterial Na /amino acid
2 Na+/ glucose
symporter reveals how binding of Na +and leucine are cou-
pled and provides a snapshot of an occluded transport inter-
Glucose
mediate in which the bound substrates cannot diffuse out of
•.
the protein (see Figure 11-2 7).
In cardiac muscle cells, the export of Ca 2 + is coupled to and
powered by the import of Na + by a cation anti porter, which
transports three Na ions inward for each Cal ion exported. Apical
Two cotransporters that arc activated at low pH help membrane
maintain the cytosolic pH in animal cells very close to 7.4
Tight junction
despite metabolic production of carbonic and lactic acids.
Cytosol Intestinal lumen
One, a Na •IH anti porter, exports excess protons. The other,
Low Na Dietary glucose
aNa HCOl /Cl cotransporter, imports HC0 3 , which dis- High K~ High dietary Na'CI" 1
sociates in the cytosol to yield pH-raising OH ions.
FIGURE 11-30 Transcellula r transpo,rt of glucose from the
• A Cl /HC0 3 anti porter that is activa ted at high pH func-
intestinal lumen into the blood. The Na+ /K ATPase in the basolateral
tions to export HC0 3 - when the cytosolic pH rises above
surface membrane generates Na and K+ concentration gradients
normal and causes a decrease in pH. (step 0 ). The outward movement of K- ions through nongated K+
AEl, a Cl IHCO~ anti porter in the erythrocyte mem- channels generates an inside-negative membrane potential across the
brane, increases the ability of blood to transport C0 2 from entire plasma membrane. Both the Na+ concentration gradient and
tissues to the lungs (see Figure 11-28). the membrane potential are used to drive the uptake of gl ucose from
the intestinal lumen by the two-Na /one~glucose symporter located in
• Uptake of sucrose, Na +, Ca 1 , and other substances into
the apical surface membrane (step f)). Glucose leaves the cell via
plant vacuoles is carried out by proton antiporters in the vacu- facilitated diffusion catalyzed by GLUT2, a glucose uniporter located in
olar membrane. Ion channels and proton pumps in the mem- the basolateral membrane (step D l.
brane are critical in generating a large enough proton concen-
tration gradient to power accumulation of ions and metabolites
in vacuoles by these proton antiporters (see Figure 11-29).

son, absorption of many nutrients from the intestinal lumen

across the epithelial cell layer and eventually into the blood
occurs by the two-stage process cal led transcellular trans-
port: import of molecules through the plasma membrane on
1 1 .6 Transcellular Transport the apical side of intestinal epithelial cells and their export
through the plasma membrane on the basolateral (blood-
Previous sections illustrated how several types of transport- facing) side (Figure 11-30). The apical portion of the plasma
ers function together to carry out important cell functions. membrane, which faces the intestinal lumen, is specialized
Here, we extend this concept by focusing on the transport of for absorption of sugars, amino acids, and other molecules
several types of molecules and ions across polarized cells, that are produced from food by multiple digestive enzymes.
which are cells that are asymmetric (have different "sides")
and thus have biochemically distinct regions of the plasma
membrane. A particularly well-studied class of polarized
Multiple Transport Proteins Are Needed to Move
cells are the epithelial cells that form sheetlike layers (epithe-
lia) covering most external and internal surfaces of body or- Glucose and Amino Acids Across Epithelia
gans. Epithelial cells are discussed in greater detail in Chapter Figure 11-30 depicts the proteins that mediate absorption of
20. Like many epithelial cells, an intestinal epithelial cell in- glucose from the intestinal lumen into the blood and illus-
volved in absorbing nutrients from the gastrointestinal tract trates the important concept that different types of proteins
has a plasma membrane organized into two major discrete are localized to the apica l and basolateral membranes of
regions: the surface that faces the outside of the organism, epithelial cells. In the first stage of this process, a two-Na I
called the apical, or top, surface, and the surface that faces one-glucose symporter located in the apical membrane im-
rhe inside (or bloodstream-facing side) of the organism, port~ glm:ose, against irs concentration gradient, from the
called the basolateral surface (see Figure 20-1 0). intestinal lumen across the apical surface of the epithelial
Specialized regions of the epithelial-cell plasma mem- cells. As noted above, this symporter couples the energeti-
brane, called tight junctions, separate the apical and basolat- cally unfavorable inward movement of one glucose molecule
eral membranes and prevent many, but not all, water-soluble to the energetically favorable inward transport of two Na +
substances on one side from moving across to the other side ions (see Figure 11-26 ). In the steady state, all the Na+ ions
through the extracellular space between cells. For this rea- transported from the intestinal lumen into the cell during

508 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

Na /glucose symport, or the similar process of ::--Ja /amino layer and ultimately into the blood. Thus giving affected
acid symport that also takes place on the apical membrane, children a solution of sugar and salt to drink (but not sugar or
are pumped out across the basolateral membrane, which salt alone) causes increased sodium and sugar transepithelial
faces the blood. Thus the low intracellular Na + concentra- transport and consequently increased osmotic flow of water
tion is maintained. The Na /K' ATPase that accomplishes into the blood from the intestinal lumen, leading to rehydra-
this is found exclusively in the basolateral membrane of in- tion. Similar sugar-salt solutions are the basis of popular
testinal epithelial cells. The coordinated operation of these drinks used by athletes to get sugar as well as water into the
two transport proteins allows uphill movement of glucose body quickly and efficiently. •
and amino acids from the intestine into the cell. This first
stage in transcellular transport ultimately is powered by ATP
hydrolysis by the Na "'"/K + ATPase.
Parietal Cells Acidify the Stomach Contents
In the second stage, glucose and amino acids concen-
trated inside intestinal cells by apical symporters are ex- While Maintaining a Neutral Cytosolic pH
ported down their concentration gradients into the blood via The mammalian stomach contains a 0.1 M solution of h}-
uniport proteins in the basolateral membrane. In rhe case of drochloric acid (HCI ). This strongly acidic medium kills
glucose, this movement is mediated by GLUT2 (see Figure many ingested pathogens and denatures many ingested pro-
11-30). As noted earlier, this GLUT isoform has a relatively teins so that they can be degraded by proteolytic enzymes
low affinity for glucose but increases its rare of transport (e.g., pepsin ) that function at acidic pH. Hydrochlonc actd is
substantially when the glucose gradient across the mem- secreted into the stomach by specialized epithelial cells called
brane rises (see Figure 11-4 ). parietal cells (also known as oxyntic cells) in the stomach
The net result of this two-stage process is movement of lining. These cells contain an H /K ATPase in their apical
Na + ions, glucose, and amino acids from the intestinal lumen membrane, which faces the stomach lumen and generates a
across the intestinal epithelium into the extracellular medium millionfold H~ concentration gradient: pH -1.0 in the stom-
that surrounds the basolateral surface of intestinal epithelial ach lumen versus pH -7.2 in the cell cytosol. This transport
cells, and eventually into the blood. Tight junctions between protein is a P-class ATP-powered ion pump ~tmilar in struc-
the epithelial cells prevent these molecules from diffusing back ture and function to the plasma membrane Na '"/K ATPase
into the intestinal lumen. The increased osmotic pressure cre- discussed earlier. The numerous mitochondria in panetal
ated by transcellular transport of salt, glucose, and amino cells produce abundant ATP for use by the H /K~ ATPase.
acids across the intestinal epithelium draws water from the If parietal cells simply exported H + ions in exchange for
intestinal lumen, mainly through the tight junctions, into the K + ions, the loss of protons would lead to a rise in the con-
extracellular medium that surrounds the basolateral surface; centration of OH- ions in the cytosol and thus a marked
aquaporins do not appear to play a major role. In a sense, salts, increase in cytosolic pH. (Recall that [H ] X [OH ] always
glucose, and amino acids "carry" the water along with them. is a constant, I 0 14 M 2 .) Parietal cells avoid this rise in cvto-
solic pH in conjunction with acidification of the stomach
lumen by using Cl-/HC0 3 antiporrers in the basolateral
Simple Rehydration Therapy Depends membrane to export the "excess" OH- ions from the cyto-
sol to the blood. As noted earlier, this a nion antiporter is
on the Osmotic Gradient Created
activated at high cytosolic pH.
by Absorption of Glucose and Na + The overall process by which parietal cells acidify the
8 An understanding of osmosis and the intestinal absorp- stomach lumen is illustrated in Figure 1 1-31. In a reaction
H tion of salt and glucose forms the basis for a simple catalyzed by carbonic anhydrase, the "excess" cytosolic
therapy that saves millions of lives each year, particularly in OH- combines with C0 2 that diffuses in from the blood,
less-developed countries. ,ln these countries, cholera and forming HC0 1 . Catalyzed by the basolateral anion Cl I
other intestinal pathogens are major causes of death of young HC0 1 antiporter, this bicarbonate ion is exported across
children. A toxin released by the bacteria activates chloride the basolateral membrane (and ultimately into the blood) in
secretion from the apical surface of the intestinal epithelial exchange for a Cl - ion. The Cl - ions then exit through Cl
cells into the lumen; water follows osmotically, and the resul - channels in the apical membrane, entering th e stomach
tant massive loss of water causes diarrhea, dehydration, and lumen. To preserve electroneutrality, each Cl ion that
ultimately death. A cure demands not only killing the bac- moves into the stomach lumen across the apical membrane
teria with antibiotics but also rehydration-replacement of is accompanied by a K.- ion that moves outward through a
the water that is lost from the blood and other tissues. separatt' K channel. In this way, the excess K ions pumpeJ
Simply drinking water does not help, because it is ex- inward by the H /K ATPase are returned to the stomach
creted from the gastrointestinal tract almost as soon as it lumen, thus maintaining the normal intracellular K concen-
enters. However, as we have just learned, the coordinated tration. The net result is secretion of equal amounts of H +
transport of glucose and Na + across the intestinal epithelium and Cl ions (i .e., HCI ) into the stomach lumen, while the
creates a transepithelial osmotic gradient, forcing movement pH of the cytosol remains neutral and the excess OH 10ns,
of water from the intestinal lumen across the epithelial cell as HC03-, are transported into the blood.

11.6 Transcellular Transport 509

 channel

rl

Apical
- membrane Bone
Tight junction

Cytosol Stomach lumen

pH 7.2 pH 1.0
V-el ass W pump
FIGURE 11-31 Acidification of the stomach lumen by parietal FIGURE 11-32 Dissolution of bone by polarized osteoclast cells
cells in the gastric lining. The apical membrane of parietal cells requires a V-class proton pump and the CIC-7 chloride channel
contains an H /K ATPase (a P-class pump) as well aso- and K• protein. The osteoclast plasma membrane is divided into two
channel proteins. Note the cyclic K transport across the apical domains separated by the tight seal between a ring of membrane
membrane: K~ ions are pumped inward by the H /K ATPase and exit and the bone surface. The membrane domain facing the bone
via a K channel. The basolateral membrane contains an anion contains V-class proton pumps and CIC-7 o-
channels. The opposing
anti porter that exchanges HC0 3 and Cl ions. The combined membrane domain contains anion antiporters that exchange HC03
operation of these four different transport proteins and carbonic and Cl ions. The combined operation of these three transport proteins
anhydrase acidifies the stomach lumen while maintaining the neutral and carbonic anhydrase acidifies the enclosed space and allows bone
pH of the cytosol. resorption while maintaining the neutral pH of the cytosol.
[SeeR. Planells-Cases and T. Jentsch, 2009, Biochim. Biophys. Aera 1792:1 73 for
discussion of CIC-7.]
Bone Resorption Requires Coordinated Function
of a V-Ciass Proton Pump and a Specific resorption. Many patients have an inactivating mutation in
the gene encoding ClC-7, the chloride channel protein local-
Chloride Channel Protein
ized to the domain of the osteoclast plasma membrane that
Net bone growth in mammals subsides just after puberty, faces the space near the bone. As with lysosomes (see Figure
but a finely balanced, highly dynamic process of disassembly 11-14 ), in the absence of a chloride channel the proton pump
(resorption) and reassembly (bone formation) goes on cannot acid ify the enclosed extracellular space and thus bone
throughout adulthood. Such continual bone remodeling per- resorption is defective. •
mits the repair of damaged bones and can release calcium,
phosphate, and other ions from mineralized bone into the
blood for use elsewhere in the body. KEY CONCEPTS of Section 11.6
Osteoclasts, the bone-dissolving cells, are a type of mac-
rophage best known for their role in protecting the body Transcellular Transport
from infections. Osteoclasts are polarized cells that form • The apical and basolateral plasma membrane domains of
specialized, very tight seals between themselves and bone, epithelial cells contain different transport proteins and carry
creating an enclosed extracellular space (Figure 11-32). An out quite different transport processes.
adher ed osteoclast then secretes into this space a corrosive
In the intestinal epithelial cell, the coordinated operation
mixture of HCI and proteases that dissolves the inorganic
of Na.,. -linked symporters in the apical membrane and Na I
components of the bone into Ca 2 · and phosphate and di-
K"' ATPases and uniporters in the basolateral membrane me-
gests its protein components. The mechanism of HCI secre-
diates transcellular transport of amino acids and glucose
tton is similar to that used by the stomach to generate
from the intestinal lumen to the blood (see Figure 11-30) .
digestive JUice (see Figure l 1-31 ). As in gastric HCI secre-
tion, carbonic anhydrase and an anion antiport protein are • The increased osmotic pressure created by transcell ular
tmportant for osteoclast function. Osteoclasrs employ a V transport of salt, glucose, and amino acids across the intesti-
type proton pump to export H ions into the bone-facing nal epithelium draws water from the intestinal lumen into
space rather than the P-el ass ATP-powered H • /K + pump the body, a phenomenon that serves as the basis for rehydra-
used by gastric epithelial cells tion therapy using sugar-salt solutions. . ., '
• The combined action of carbonic anhydrase and four dif- ' .
The rare hereditary disease osteopetrosis, marked by ferent transport proteins permits parietal cells in the stomach
increased bone density, is due to abnormally low bone

510 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 pressin to certain kidney cells similarly causes an aquaporin
lining to secrete HCl into the lumen while maintaining their to traffic to the plasma membrane, increasmg the rate of
cyrosolic pH near neutrality (see Figure 11-31). water transport. But despite much research, the underlying
• Bone resorption requires coordinated function in osteo- cellular mechanisms by which hormones stimulate the move-
clasts of a V-class proton pump and the CIC-7 chloride ment of transport proteins to and from the plasma mem-
channel protein (Figure 11-32). brane, and the regulation of these processes, remain obscure.

Key Terms
Perspectives for the Future ABC superfamily 485 membrane potential475
In this chapter, we have explained the action of specific mem- active transport 4 76 Na /K+ ATPase 489
brane transport proteins and their impact on certain aspects antiport 502 patch clamping 500
of human physiology; such a molecular physiology approach aquaporins 480 P-class pump 484
has many medical applications. Even today, specific inhibitors ATP-powered pump 4 76 resting K channel 497
or activators of channels, pumps, and transporters constitute cotransport 502 resting potential 495
the largest single class of drugs. For instance, an inhibitor of
electrochemical gradient 475 sarcoplasmic reticulum 486
the gastric H .._/K+ ATPase that acidifies the stomach is the
most widely used drug for treating stomach ulcers and gastric facilitated transport 4 76 simple diffusion 474
reflux syndrome. Inhibitors of channel proteins in the kidney F-class pump 484 symport 502
are widely used to control hypertension (high blood pressure); flippase 493 tight junction 508
by blocking resorption of water into the blood from urine gated channel 476 transporter 4 76
forming in the kidneys, these drugs reduce blood volume and GLUT proteins 479 rranscellular transport 508
thus blood pressure. Calcium-channel blockers are widely em-
hypertonic 481 uniport 478
ployed to control the intensity of contraction of the heart.
Drugs that inhibit a particular potassium channel in (3 islet hypotonic 481 V-class pump 484
cells enhance secretion of insulin (see Figure 16-38), and are isotonic 481
widely used to treat adult-onset (type II ) diabetes.
With the completion of the human genome project, we
have in hand the sequences of all human membrane-transport Review the Concepts
proteins. Already we know that mutations in many of them
1. Nitric oxide (NO) is a gaseous molecule with lipid solu-
cause disease-cystic fibrosis, due to mutations in CFTR, is
bility similar to that of 0 2 and C02 • Endothelial cells lining
one example, and osteopetrosis, caused by mutations in the
arteries use NO to signal surrounding smooth muscle cells to
ClC-7 chloride channel, is another. More recently it was shown
relax, thereby increasing blood flow. What mechanism or
that loss-of-function mutations in either subunit of a different
mechanisms would transport NO from where it is produced
chloride channel 1 ClC-K, cause both salt loss by the kidney and
in the cytoplasm of an endothelial cell into the cytoplasm of
deafness. This explosion of basic knowledge, associating spe-
a smooth muscle cell where it acts?
cific genetic diseases with specific transport proteins, will en-
able researchers to identify new types of compounds that 2. Acetic acid (a weak acid with a pK. of 4.75) and ethanol
inhibit or activate just one of these membrane transport pro- (an alcohol) are each composed of two carbons, hydrogen,
teins and not its homologs. An important challenge, however, and oxygen, and both enter cells by passive diffusion. At pH 7,
is to understand the role of an individual transport protein in one is much more membrane permeable than the other.
each of the several tissues in which it is expressed. Which is more permeable and why? Predict how the perme-
Another major challenge is to understand how each ability of each is altered when the pH is reduced to 1.0, a
channel, transporter, and pump is regulated to meet the value typical of the stomach.
needs of the cell. Like other cellular proteins, many of these 3. Uniporrers and ion channels support facilitated diffusion
proteins undergo reversible phosphorylation, ubiquitination, across biomembranes. Although both are examples of facili-
and other covalent modifications that affect their activity, tated diffusion, the rates of ion movement via an ion channel
but in the vast majority of cases, we do not understand how are roughly 104 - to 105 -fold faster than that of molecule move-
' ' this regulation affects cellular function. Many channels, ment via a uniporter. What key mechanistic difference results
transporters, and pumps normally reside on intracellular in this large difference in transport rate? What contribution to
membranes, not on the plasma membrane, and move to the free energy (~G) determines the direction of transport?
plasma membrane only when a particular hormone is pres- 4. Name the three classes of transporters. Explain which
ent. The addition of insulin to muscle, for instance, causes one or more of these classes is able to move glucose and
•' the GLUT4 glucose transporter to move from intracellular which move bicarbonate (HC0 3 ) against an electrochemical
membranes to the plasma membrane, increasing the rate of gradient. In the case of bicarbonate, but not glucose, the ~G
glucose uptake. We noted earlier that the addition of vaso- of the transport process has two terms. What are these two

Review the Concepts 511

terms, and why does the second not apply to glucose? Why coding for a putative K channel actually codes for a K',. or
are cotransporters often referred to as examples of second- Na channel.
ary active transport? 13. Plants use the proton electrochemical gradient across
5. An H + ion is smaller than an H 2 0 molecule, and a glyc- the vacuole membrane to power the accumulation of salts
erol molecule, a three-carbon alcohol, is much larger. Both and sugars in the organelle. This creates a hypertonic situa-
readily dissolve in H 2 0. Why do aquaporins fail to transport tion. Why does this not result in the plant cell swelling and
H whereas some can transport glycerol? bursting? Even under isotonic conditions, there is a slow
6. GLUT1, found in the plasma membrane of erythrocytes, leakage of ions into animal cells. How does the plasma mem-
is a classic example of a uniporrer. brane Na-/K ATPase enable animal cells to avoid osmotic
a. Design a set of experiments to prove that GLUT1 is lysis under isotonic conditions?
indeed a glucose-specific uniporter rather than a galactose- 14. In the case of the bacterial sodium/leucine transporter,
or mannose-specific uniporrer. what is the key distinguishing feature about the bound sodium
b. Glucose is a 6-carbon sugar while ribose is a 5-carbon ions that ensures that other ions, particularly K , do not bind?
sugar. Despite this smaller size, ribose is not efficiently trans- 15. Describe the symport process by which cells lining the
ported by GLUTl. How can this be explained? small intestine import glucose. What ion is responsible for the
c. A drop in blood sugar from 5 mM to 2.8 mM or below transport, and what two particular, features facilitate the en-
can cause confusion and fainting. Calculate the effect of this ergetically favored movement of this ion across the plasma
drop on glucose transport into cells expressing GLUT1. membrane?
d. How do liver and muscle cells maximize glucose uptake 16. Movement of glucose from one side to the other side of
without changing vm"'? the intestinal epithelium is a major example of transcellular
e. Tumor cells expressing GLUT1 often have a higher transport. How does the Na+/K• ATPase power the process?
Vmax for glucose transport than do normal cells of the same
Why are tight junctions essential for the process? Why is lo-
type. How could these cells increase the Ymax? calization of the transporters specifically in the apical and
f. Fat and muscle cells modulate the V""' for glucose basolateral membrane crucial for transcellular transport?
uptake in response to insulin signaling. How? Rehydration supplements such as sport drinks include a
7. Name the four classes of ATP-powered pumps that pro- sugar and a salt. Why are both important to rehydration?
duce active transport of ions and molecules. Indicate which
of these classes transport ions only and which transport pri-
marily small organic molecules. The initial discovery of one
Analyze the Data
class of these ATP-powered pumps came from studying the Imagine that you are investigating the transepithelial trans-
transport of not a natural substrate but rather artificial sub- port of radioactive glucose. Intestinal epithelia l cells are
strates used as cancer chemotherapy drugs. What do investi- grown in culture to form a complete sheet so that the fluid
gators nov, think are common examples of the natural bathing the apical domain of the cells (the apical medium) is
substrates of this particular class of ATP-powered pumps? completely separated from the fluid bathing the basolateral
8. Explain why the coupled reaction ATP ~ ADP + P, in domain of the cells (the basolateral medium). Radioactive
the P-class ion pump mechanism does not involve direct hy- e 4
C-Iabeled) glucose is added to the apical medium, and the
drolysis of the phosphoanhydride bond. appearance of radioactivity in the basolateral medium is
9. Describe a negative feedback mechanism for controlling monitored in terms of counts per minute per milliliter (cpm/ml),
rising cytoplasmic Ca 2 concentration in cells that require a measure of radioactivity per unit volume.
rapid changes in Ca2 concentration for normal functioning. Treatment 1: The apical and basolateral media each con-
How would a drug that inhibits calmodulin activity affect tain 150 mM Na+ (curve 1).
cytoplasmic Ca 2 concentration regulated by this mecha- Treatment 2: The apical medium contains 1 mM Na ,
nism? What would be the effect on the function of, for ex- and the basolatcral medium contains 150 mM Na - (curve 2).
ample, a skeletal muscle cell? Treatment 3: The apical medium contains 150 mM Na ,
10. Certain proton pump inhibitors inhibit secretion of and the basolateral medium contains I mM Na (curve 3).
stomach acid and are among the most widely sold drugs in
the world today. What pump does this type of drug inhibit, Radioactivity in Basolateral Medium
and where is this pump located?
=
.§ 200
11. The membrane potential in animal cells, but not in plants,
depends largely on resting K"'" channels. How do these channels 8.. 150
~
contribute to the resting potential? Why are these channels con-
.~ 100
sidered to be nongated channels? How do these channels >
·;::;
achieve selectivity forK + versus Na +,which is smaller than K ? ~ 50 Treatment 2
.Q
12. Patch clamping can be used to measure the conductance
properties of individual ion channels. Describe how patch 2 4 6 8 10
Time (min) in 14 C-glucose
clamping can be used to determine whether or not the gene

512 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

a. What is a likely explanation for the different results ob- Gottesman, M. M., and V. Ling. 2006. The molecular basi\ of
tained in treatments 1 and 3 versus treatment 2? multidrug resistance in cancer: the early vears of P-glycoprotem
re~earch. Fl:.BS Lett. 580:998-1009.
In additional studies, the drug ouabain, which inhibits Guerini, D., L. Colcrto, and E. Carafoli. 2005. Exporting
Na + fK ATPases, is included as noted. calclllm from cell\. Cell Cafctum 38:281-289.
Guttmann, D., et al. 2009. Undersranding polyspeC1fic1ty of
Treatment 4: The apical and basolateral media contain multidrug ABC transporters: closing in on rhe gaps 111 ABCB 1.
150 mM Na +and the apical media contains ouabain (curve 4). Trends Biochem. Set. 35:36-42.
Treatment 5: The apical and basolateral media contain Hall, M., et al. 2009. Is resistance useless? Multidrug resistance
150 mM Na + and the basolatcral media contains ouabain and collateral sensiti\'ity. Trends J>harmacol. Set. 30:546-556.
(curve 5). Jencks, W. P. 1995. The mechanism of coupling chem1cal and
physical reacnons b} the calcium ATPase of sarcoplasmK reticulum
and other coupled vectorial systems. Biosci. Rept. 15:283-287.
Radioactivity in Basolateral Medium Locher, K. P., A. Lee, and D. C. Rees. 2002. The F. co/1 BtuCD
structure: a framework for ABC transporter architecture and
=
.§ 200
mcchamsm. Science 296:1091.
Treatment 4
g_ 150 Ogawa, H., et. al. 2009. C rystal structure of the sodium-
~ potassium pump (Na ,K -ATPase) with hound potassium and
.? 100 ouabain. Proc. Nat'/ Acad. Sci. USA 106:13742-13747.
.::: Raggers, R. j., et al. 2000. Lip1d traffic: the ABC of rransbilayer
u
"'
0 50 movement. Traffic 1:226-234.
-o R1ordan, J. 2005. A'lsemhly of functional CFTR chloride
"'
a: channels. Ann. Rev. Physiol. 67:701-718.
Time (min) in 14C-glucose Shmoda, T., et. al. 2009. Crystal structure of the sodium-
potassium pump at 2.4 A resolution. Nature 459:446-450.
Toel, ~1., R. Saum, and M. Forgac. 2010. Regulation and
b. What is a likely explanation for the different results ob- isoform function of the V -ATPases. Biochemistry 49:4 715-4 72.3.
tained in treatment 4 versus treatment 5? Toyosh1ma, C. 2009. How Ca2 -ATPase pumps 1ons across the
c. Certa in natural compounds and drugs being tested as sarcoplasmic reticulum membrane. Biochim. Biophys. Acta 1793:
treatments for diabetes lower glucose transport in intestinal 941-946.
or kidney epithelial cell s, and thereby lower blood sugar lev- Verkman, A. S., G. L. Lukacs, and L.]. Galiena. 2006. CFTR
els. Addition of one such drug to the apical medium yields chloride channel drug discovery-111hibaors as amid1arrheals
transport similar to that in Treatment 5, whereas its addition and activators for therapy of cysnc fibrosis. Curr. Pharm. Des.
12:2235-2247.
to the basolateral medium yields transport similar to that in
Treatment 4. What is the most likel y target of this drug, and
what is the effect on this target? Nongated Ion Channels and the Resting Membrane Potential
Dutzler, R., et al. 2002. X-ra} structure of a CIC chlonde
channel at 3.0 A re\'eals the molecular basis of anion selecti\·ity.
References Nature 415:287-294.
Hibino, H., et al. 2010. Inwardly recnfying porasslllm channels:
Facilitated Transport of Glucose and Water their structure, function, and physiOlogical roles. Physiol. Rev.
Engel, A., Y. Fujiyoshi, and P. Agre. 2000. The importance of 90:291-366.
aquaporin water channel protem structures. EMBO ]. 19:800-806. Hille, B. 200 I . !on Cha111zels of Excitable Membranes, 3d ed.
Hedfalk, K., et al. 2006. Aquaporin gating. Curr. Opimon Smauer Associates.
Structural Btofogy 16:1-10. Jentsch, T.J. 2008. ClC chloride channels and transporters: from
Hruz, P. W., and M . .'vi. Mueckler. 2001. Srructural ana l y~is of genes to protem structure, pathology and physiology. Crrt. Ret•.
the GLUT1 facilitative glucos~ transporter (rev1ew). Mol. Memb. 810chem. Mol. 43:3-36.
Bioi. 18:1 83-193. Jouhaux, E., and R ..Mackinnon. 2005. Pnnciples of .,elecnve
King, L. S., D. Kozono, and P. Agre. 2004. From structure to ion transport 111 channels and pumps. Science 310:1461-1465.
disease: the evolving tale of aquaporin biology. Nat. Reu. Mol. Cell MacKinnon, R. 2004. Potassium channels and the atomic basis
Btof. 5:687-698. of selective ion conduction. Nobel Lecture reprinted in Btosci. Rep.
Thorens, B., and M . .Viueckler, Glucose transporters 111 the 2 1st 24:75-100.
century. Am.]. Physiof.-Endoc. M. 298:E 141-E145, 2010. Montello, C., L. B1rnbaumer, and V. Flickers. 2002. The TRP
Verkman, A. S. 2009. Knock-out models reveal new aquaponn channels, a remarkably functional family. Cel/ 108:595- 598.
functions. Handb. I:.xp. Pharmacal. 190:359-38 1. Neher, E. 1992. Ion channels for communication between and
Wang, Y., K. Schulten, and E. Tajkhorshid. 2005. What makes w1thin cells. Nobel Lectun: reprinred in Neuron 8:605-612. and
an aquaporin a glycerol channel? A comparative study of AqpZ and Sctence 256:498-502.
GlpF structure. Structure 13:11 07- 111 8. Neher, E., and B. Sakmann. 1992. The parch clamp rechmque.
Sci. Am. 266(3):28-35.
ATP-Powered Pumps and the Intracellular Ionic Environment Planells-Cases, R., and T. J. Jentsch. 2009. Chlonde channelo-
Aller, S., et al. 2009. Structure of P-glycoprotein reveals a pathies. Biochim. Biophys. Acta 1792:173-189.
molecular basis for poly-specific drug bmdmg. Science 323: Roux, B. 2005. lon conduction and selectivity 111 K+ channels.
1718- 1722. 2005. Ann. Rev. Biophys. Biomol. Struct. 34:153-171.

References 513
 Zhou, Y., et al. 200 I. Chemistry of 10n coordmation and \X'akabayashi, S., :Vt. Shigekawa, and]. Pouyssegur. 1997.
hydration revealed by a K channei-Fab complex at 2 A resolution. Molecular physiology of vcrrebrate Na + IH • exchangers. Physiol.
Nature 414:43-48. Ret-'. 77: .5 1-74.
\XInght, E. ~1. 2004. The sodium/glucose cotransport family
Cotransport by Symporters and Anti porters
Sl.C5. P(/ugers Arch. 447:510-518.
Alper, S. L. 2009. Molecular physiology and genetics of Na"- Wright, E. M., and D. D. Loo. 2000. Coupling between Na + ,
independent SLC4 anion exchangers.]. f-xp. B10/. 212:1672-1683. sugar, and water transport across the mtestine. Ann. NY Acad. Set.
Barkla, B. J., R. Vera-Estrella, and 0. PantOJa. 1999. Towards 915:54-66.
the production of salt-tolerant crops. Adl'. Exp. Med. Bioi.
464:77-89.
Diallmas. G. 2008. An almost- complete movie: structural Transcellular Transport
~napshots of transporrer protems re,·eal how they transport species Anderson, J. M., and C. M. Van ltallie. 2009. Physwlogy and
across membranes. Science 322: 1644-1645. function of the nght junction. Cold Stning Harbor Perspect. Bioi.
Gao, X., et al. 2009. Structure and mechanism of an amino acid 1:a002584.
ami porter. Soence 324:1565-1568. l:lkouby-Naor, L., and T. Ben-Yoscf. 2010. Functions of claudm
Gouau", E. 2009. Re,·iew: The molecular logic of sodium- right junction proteins and their complex interactions in various
coupled neurotransmitter transporters. Phil. Trans. R. Soc. Land. physiological systems. Int. Reu. Cell Mol. Bioi. 279:1-32.
B B10/. Set. 364:149-154. Hubner, C. A., and T. J. Jentsch. 2008. Channelopathtes of
Knshnamurrhy, H., C. L. Piscitelli, and E. Gouaux. 2009. transeptthelial transport and \'esicular klnction. Adv. Genet.
Unlocking the molecular secrets of sodium-coupled transporters. 63:113-152.
Nature 459:34 7-.15 5. Rao, :VL 2004. Oral rehydration therapy: new explanations for
Orlowski, J., and S. Grinstein. 2007. Emergmg roles of alkali an old remedy. Ann. Reu. Phys10/. 66:385-41 ~.
cation/proton exchangers in organellar homeostasis. Curr. Opm. Schafer, J. A. 2004. Renal water reabsorption: a physiologic
Cell Bioi. 19:483-492. retnhpecrive in a molecular era. Kidney Int. Suppl. 91:520-27.
Shabala, S., and T. A. Cum. 2008. Potassium transport and Schultz, S. (,. 2001. Fptthelial water absorption: osmosis or
plant salt tolerance. Physiol. Plant 133:651-669. cotransporr? Proc. Nat"/. Acad. Set. USA 98:3628,-3630.

514 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 CLASSIC EXPERIMENT 11.1

Stumbling upon Active Transport

J. Skou, 1957, Biochem. Biophys. Acta 23:394

n the mid-1950s jens Skou was a

I young physician researching the
effects of local anesthetics on isolated
that this would be an ideal enzyme for
his purposes, Skou set our to isolate such
the concentrations of salts a nd other
cofactors, which bring cations into the
an ATPase from a more readily available reaction. He could easily determine a
lipid bilayers. He needed an easily as- source, crab leg neurons. It was during pH optimum as well as an optimal
sayed membrane-associated enzyme to his characterization of this enzyme that concentration of Mg2 ' , but optimizing
use as a marker in his studies. What he he discovered the protein's function. Na and K + proved to be more diffi-
discovered was an enzyme critical to cult. Regardless of the amount of K+
the maintenance of membrane poten- added to the reaction, the enzyme was
tial, the Na _._ /K ATPase, a molecular The Experiment
inactive without Na '. Similarly, with-
pump that catalyzes active transport. Since the original goal of his study was out K , Skou observed only a low-level
to characterize the ATPase for usc in ATPase activity that did not increase
subsequent studies, Skou wanted to with increasing amounts of Na .
Background
know under what experimental condi- These results suggested that the en-
During the 1950s many researchers tion its activity was both robust and zyme required both Na and K ~for op-
around the world were actively investi- reproducible. As often is the case with timal activity. To demonstrate that this
gating the physiology of the cell mem- the characterization of a new enzyme, was the case, Skou performed a series of
brane, which plays a role in a number this requires careful titration of the experiments in which he measured the
..• of biological processes. lt was well various components of the reaction . enzyme activity as he varied both the
known that the concentration of many Before this can be done, one must be Na and K concentrations in the reac-
ions differs inside and outside the cell. sure the system is free from outside tion (Figure 1 ). Although both cations
For example, the cell maintains a lower sources of contamination. clearly were required for significant ac-
intracellular sodium (Na ') concentra- In order to study the influence of tivity, something interesting occurred at
tion and higher intracellular potassium various cations, including three that are high concentrations of each cation. At
(K ... ) concentration than is found out- critical for the rcaction-Na , K ~, and the optimal concentration of :--.ra + and
side the cell. Somehow the membrane Mg2 -Skou had to make sure that no K+, the ATPase activity reached a peak.
can regulate intracellular salt concen- contaminating ions were brought into Once at that peak, further increasing
trations. Additionally, movement of the reaction from another source. There- the concentration did not affect the
ions across cell membranes had been fore all buffers used in the purification ATPase activity. Na- thus behaved like
observed, sugge:;ting that some sort of of the enzyme were prepared from salts a classic enzyme substrate, with increas-
transport system is present. To main- that did not contain these cations. An ing input leading to increased activity
tain normal intracellular Na+ and K additional source of contaminating cat- until a saturating concentration was
concentrations, the transport system ions was the ATP substrate, which con- achieved, at which the activity pla-
could not rely on passive diffusion be- tains three phosphate groups, giving it teaued. K •, on the other hand, behaved
cause both ions must move across the an overall negative charge. Because differently. When the K ... concentration
membrane against their concentration stock solutions of ATP often included a was increased beyond the optimum,
gradients. This energy-r~quiring pro- cation to balance the charge, Skou con- ATPase activity declined. Thus while
cess was termed active transport. verted the ATP used in his reactions to K was required for optimal activity, at
At the time of Skou's experiments, the acid form so that balancing cations high concentrations it inhibited the en-
the mechanism of active transport was would not affect the experiments. Once zyme. Skou reasoned that the enzyme
still unclear. Surprisingly, Skou had no he had a well-controlled environment, must have separate binding sites for
intention of helping to clarify the field. he could characterize the enzyme activ- Na + and K . For optimal ATPase activ-
He found the Na +/K+ ATPase com- ity. These precautions were fundamen- ity, both must be filled. However, at
pletely by accident in his search for an tal to his discovery. high concentrations K+ could compete
abundant, easily measured enzyme ac- Skou first showed that his enzyme for the Na -binding site, leading to en-
tivity associated with lipid membranes. could indeed catalyze the cleavage of zyme inhibition. He hypothesized that
A recent study had shown that mem- ATP into ADP and inorganic phos- this enzyme was involved 1n active
branes derived from squid axons con- phate. He then moved on to look for transport, that is, the pumping of Na •
tained a membrane-associated enzyme the optimal conditions for this activity out of the cell, coupled to the import of
that could hydrolyze ATP. Thinking by varying the pH of the reaction, and K into the cell. Later studies would

Stumbling upon Active Transport 515

(a) 40 (b) 40
Mg 6mM/ I
K 120 m M / 1

30 30

a..
~ 20

Mg 6mM/ I

10 10

0 ~--------L--------L------~~------~
0 50 100 150 200
KCI mM/1 NaCI mM/1
FIGURE 1 Demonstration of the dependence of Na+/K+ ATPase increases up to a peak and then levels out. This graph also demonstrates
activity on the concentration of each ion. The graph on the left shows the dependence of the activity on low levels of K+. [~dapted from J. Skou,
that increasing K+ leads to an inhibition of the ATPase activity. The 1957, Biochem. Biophys. Acta 23:394.)
graph on the right shows that with increasing Na , the enzyme activity

prove that this enzyme was indeed the active transport on a molecular level. nation by Na and K+ at all stages, he
pump that catalyzed active transport. How did Skou know to test both Na obtained clear-cut, reproducible results.
This finding was so exciting that Skou and K ~ ? In his Nobel lecture in 1997, The discovery of the Na +!K+ ATPase
devoted his subsequent research to he explained that in his first attempts at had an enormous impact on membrane
studying the enzyme, never using it as a characterizing the ATPase, he took no biology, leading to a better understand-
marker, as he initially intended. precautions to avoid the use of buffers ing of the membrane potential. The gen-
and ATP stock solutions that contained eration and disruption of membrane
Na and K .,. . Pondering the puzzling potential forms the basis of many bio-
and unreproducible results that he ob- logical processes, including neurotrans-
Discussion tained led to the realization that con- mission and the coupling of chemical
Skou's finding that a membrane ATPase taminating salts might be influencing and electrical energy. For this funda-
used both Na .,. and K as substrates the reaction. When he repeated the ex- mental discovery, Skou was awarded
was the first step in understanding periments, this time avoiding contami- the Nobel Prize for Chemistry in 1997.

516 CHAPTER 11 • Transmembrane Transport of Ions and Small Molecules

 CHAPTER

Cellular Energetics

Immunofluorescence micrograph showing the intert wined network of

mitochondria (red) in cultured human Hela cells. The nuclei of the cells
are stained purple. [Dr. Gopal Murti/Photo Researchers.]

F
rom the growth and division of a cell to the beating of power many otherwise energetically unfavorable processes.
a heart to the electrical activity of a neuron that under- Examples include the synthesis of proteins from amino acids
lies thinking, life requires energy. Energy is defined as and of nucleic acids from nucleotides (Chapter 4 ), transport of
the capacity to do work, and on a cellular level that work molecules against a concentration gradient by ATP-powered
includes conducting and regulating a mu ltitude of chemical pumps (Chapter 11 ), contraction of muscle (Chapter 17),
reactions and transport processes, growing and dividing, and beating of cilia (Chapter 18 ). A key theme of cellular ener-
generating and mainta ining a highly o rganized structu re, getics is the use of proteins to use, or "couple," energy released
and interacting with other cells. This chapter describes the from one process (e.g., ATP hydrolysis) to drive another pro-
molecular mecha nisms by which cells use sunlight or chemi- cess (e.g., movement of molecules across membranes) that
cal nutrients as sources of energy, with a special focus on otherwise would be thermodynamically unfavorable.
how cells convert these external sources of energy into a bio- The energy to drive ATP synthesis from ADP (.lG 0 ' =
logically universal, intracellular, chemical energy carrier, 7.3 kcallmol) derives primarily from two sources: the energy
adenosine triphosphate, or ATP (Figu re 12- 1). ATP, found in the chemical bonds of nutrients and the energy in sunlight
in all types of orga nisms and presumably present in the ear- (figure 12-1 ). The two processes primarily responsib le for
liest life-forms, is genera~ed from the chemical addition of converting these energy sources into ATP are aerobic oxida-
inorganic phosphate (H PO/ , often abbreviated as P,) to tion (also known as aerobic respiratio n), which occurs in mi-
adenosine diphosphate, or ADP, a process called phosphory- tochondria in nearly all eukaryotic cells (Figure 12-1, top),
lation. Cells use the energy released during hydrolysis of the and pho tosynthesis, which occurs in chloroplasts only in leaf
terminal phosphoanhydride bond in ATP (see Figure 2-3 1) to cells of plants (Figure 12-1, bottom) and certain single-cell

OUTLINE

12.1 First Step of Harvesting Energy 12.5 Photosynthesis and Light-Absorbing Pigments 552
from Glucose: Glycolysis 519
12.6 Molecular Analysis of Photosystems 559
12.2 Mitochondria and the Citric Acid Cycle 524
12.7 C02 Metabolism During Photosynthesis 567
12.3 The Electron Transport Chain and Generation
of the Proton-Motive Force 532

12.4 Harnessing the Proton-Motive Force

to Synthesize ATP 544
 Energy Cytosol
source Mitochondrion
Stage I Stage Ill
Chemical ---f- - + Lipid or --~---+ Substrate ---+ NADH---+ Electron ---+Proton- ---+ ATP
bond sugar oxidation FADH 2 transport motive
(in glycolysis (citric acid cycle) (electron ~ force
glucose - r ;yruvate) ""\ ""\ carriers) o2 H2o (W gradient)
ATP NADH

Chloroplast
Stage 2
Photons
(sunlight)

FIGURE 12-1 Overview of aerobic oxidation and photosynthesis. to 0 2 and establish conditions (stage 2 ) necessary for the generation of
Eukaryotic cells use two fundamental mechanisms to convert external ATP (stage 3) and carbohydrates from C02 (carbon fixation, stage 4).
sources of energy into ATP. (Top) In aerobic oxidation, "fuel" molecules Both mechanisms involve the production of reduced high-energy
(primarily sugars and fatty acids) undergo preliminary processing in electron carriers (NADH, NADPH, FADH 2 ) and movement of electrons
the cytosol, e.g., breakdown of glucose to pyruvate (stage 1}, and down an electric potential gradient in an electron transport chain
are then transferred into mitochondria, where they are converted by through specialized membranes. Energy from these electrons is
oxidation with 0 2 to carbon dioxide and water (stages II and Ill) and released and captured as a proton electrochemical gradient (proton-
ATP is generated (stage IV). (Bottom) In photosynthesis, which occurs motive force) that is then used to drive ATP synthesis. Bacteria utilize
in chloroplasts, the radiant energy of light is absorbed by specialized comparable processes.
pigments (stage 1); the absorbed energy is used to both oxidize water

organisms, such as algae and cyanobacteria. Two additional carbohydrates-primarily sucrose and starch. Unlike aerobic
processes, glycolysis and the citric acid cycle (Figure 12-1, tofJ), oxidation, which uses carbohydrates and 0 2 to generate
are also important direct or indirect sources of ATP in both C0 2, photosynthesis uses C0 2 as a substrate and generates
animal and plant cells. 0 2 and carbohydrates as products.
In aerobic oxidation, breakdown products of sugars (ca r- This reciprocal relationship between aerobic oxidation
bohydrates) and fatty acids (hydrocarbons)-both derived in occurring in mitochondria and photosynthesis in chloro-
animals from the digestion of food-are converted by oxida- plasts underlies a profound symbiotic relationship between
tion with 0 2 to carbon dioxide and water. The energy re- photosynthetic and nonphotosynthetic organisms. The oxy-
leased from this overall reaction is transformed into the gen generated during photosynthesis is the source of virtu-
chemical energy of phosphoanhydride bonds in ATP. This is ally all the oxygen in the air, and the carbohydrates produced
analogous to burning wood (carbohydrates) or oil (hydro- are the ultimate source of energy for virtually all nonphoto-
carbons) to generate heat in furnaces or motion in automo- synthetic organisms on earth. (An exception is bacteria liv-
bile engines: both consume 0 2 and generate carbon dioxide ing in deep ocean vents-and the organisms that feed on
and water. The key difference is that cells break the overall them-that obtain energy for converting col
into carbohy-
reaction down into many intermed iate steps, with the drates by oxidation of geologically generated reduced inor-
amount of energy released in any given step closely matched ganic compounds released by the vents.)
to the amount of energy that can be stored-for example as At first glance, it might seem that the molecular mecha-
ATP-or that is required for the next intermediate step. If nisms of photosynthesis and aerobic oxidation have little in
there were not such a close match, excess released energy common, besides the fact that they buLh produce ATP. How-
would be lost as heat (which would be very inefficient ) or ever, a revolutionarr d iscovery in cell biology established
not enough energy would be released to generate energy that bacteria, mitochondria, and chloroplasts all use the
storage molecules such as ATP or to drive the next step in same mechanism, known as chemiosmosis, to generate ATP
the process (w hich wou ld be ineffective). from ADP and P,. In chemiosmosis (also known as chemios-
In photosynthesis, the radiant energy of light is absorbed motic coupling), a proton electrochemical gradient is first
by pigments such as chlorophyll and used to make ATP and generated across a membrane, driven by energy released as

518 CHAPTER 12 • Cellular Energetics

 Low pH and water. The process is relatively inefficient in that substan-
Radiant H H H tial amounts of the chemical energy stored in the fuel are
(loltv Positive
wasted as they are converted to unused heat, and substantial
ene,. electric amounts of fuel are only partially oxidized and are released as
9
'YJ + potential carbonaceous, sometimes toxic, exhaust. ln the competition
High pH + to survive, organisms cannot afford to squander their some-
- +
Negative - + times limited energy sources on an equivalently inefficient
electric
potential
process, and have therefore evolved a more efficient mecha-
nism for converting fuel into work. That mechanism, known
as aerobic oxidation, provides the following advantages:
• By dividing the energy conversion process into multiple
steps that generate several energy-carrying intermediates,
chemical bond energy is efficiently channeled into the syn-
thesis of ATP, with less energy lost as heat.
Chemical
bonds in Exoplasmic face • Different fuels (sugars and fatty acids) are reduced to com-
carbohydrates H+
mon intermediates that can then share subsequent pathways
and lipids Synthesis
ofATP
for combustion and ATP synthesis.
FIGURE 12-2 The proton-motive force powers ATP synthesis. • Because the total energy stored in the bonds of the initial
Transmembra ne proton concentration and electrical (voltage) fuel molecu les is substantially greater than that required to
gradients, collectively called the proton-motive force, are generated drive the synthesis of a single ATP molecule ( -7.3 kcal/mol),
during aerobic oxida tion and photosynthesis in eukaryotes and many A TP molecules are produced.
prokaryotes (bacteria). High-energy electrons generated by light
absorption by pigments (e.g., chlorophyll), or held in the reduced form An important feature of ATP production from the break-
of electron carriers (e.g., NADH, FADH 2) made during the catabolism of down of nutrient fuels into C02 and water (see Figure 12-1,
sugars and lipids, pass down an electron transport chain (blue arrows). tOfJ) is a set of reactions, called respiration, involving a series
releasing energy throughout the process. The released energy is used of oxidation and reduction reactions called an electron
to pump protons across the membrane (red arrows), generating the transport chain. The combination of these reactions with
proton-motive force. In chemiosmotic coupling, the energy released phosphorylation of ADP to form ATP is called oxidative
when protons flow down the gradient t hrough ATP synthase drives phosphorylation and occurs in mitochondria in nearly all
the synthesis of ATP. The proton-motive force can also power other e ukaryotic cells. When oxygen is available and used as the
processes, such as the transport of metabolites across the membrane fina l recipient of the electrons transported via the electron
against their concentration gradient and rotation of bacterial flagella.
transport chain, the respiratory process that converts nutri-
ent energy into ATP is called aerobic respiration or aerobic
oxidation. Aerobic respiration is an especially effic1ent way
to maximize the conversion of nutrient energy into ATP be-
electrons travel down their electric potential gradient
cause oxygen is a relatively strong oxidant. If some mole-
through an elect ron t ransport ch ain. The energy stored in
cule other than oxygen, for example the weaker oxidant
this proton electrochemical gradient, called the proton-motive
sulfate (SO / ) or nitrate (N0 3 -) , is the final recipient of the
force, is then used to power the synthesis of A TP (figure
electrons in the electron transport chain, the process is
12-2) or other energy-requiring processes. As prorons move
called anaerobic respiration. Anaerobic respiration is typi-
down their electrochemical gradient through the ATP syn-
cal of some prokaryotic microorganisms. Although there
·. thesis enzyme, ATP is synthesized from ADP and P., a pro-
cess that is the reverse 'of the ATP-powered ion pumps
are exceptions, most known multicellular (metazoan) eu-
karyotic organisms usc aerobic respiration to generate most
discussed in the previous chapter. In this chapter, we explore
of their ATP.
the molecular mechanisms of the two processes that share
In our discussion of aerobic oxidation, we will be tracing
this central mechanism, focusing first on aerobic oxidation
the fate of the two main cellular fuels: sugars (principall y
and then on photosynthesis.
glucose) and fatty acids. Under certain conditions amino
acids also feed into these metabolic pathways. We first con-
sider glucose oxidation, and then turn to fatty acids.
The complete aerobic oxidation of one molecule of glu-
12.1 First Step of Harvesting Energy cose yields 6 molecules of C0 2 and the energy released is
coupled to the synthesis of as many as 30 molecules of ATP.
from Glucose: Glycolysis
The overall reaction is
In an automobile engine, hydrocarbon fuel is oxidatively and
explosively converted in an essentially one-step process to me- C6H 120 6 + 6 02 + 30 P, 2 + 30 ADP 3 + 30 H ~
chanical work (i.e., driving a piston) plus the products C0 2 6 C02 + 30 ATP4 - + 36 H 2 0

12.1 First Step of Harvesting Energy from Glucose: Glycolysis 519

Glucose oxidation in eukaryotes takes place in four stages from two ATPs. These can be thought of as "pump priming"
(see Figure 12-1, top): reactions, which introduce a little energy up front in order to
effectively recover more energy downstream. Thus glycolysis
Stage 1: Glycolysis In the cytosol, one 6-carbon glucose mol-
yields a net of only two ATP molecules per glucose molecule.
ecule is converted by a series of reactions to two 3-carbon
The balanced chemical equation for the conversion of
pyruvate molecules; a net of 2 ATPs are produced for each
glucose to pyruvate shows that four h ydrogen atoms (four
glucose molecule.
protons and four electrons) are also released:
Stage II: Citric Acid Cycle In the mitochondrion, pyruvate
oxidation to C0 2 is coupled to the generation of the high- 0 0
energy electron carriers NADH and FADH2, which store the
energy for later use.
Glucose Pyruvate

Stage III: Electron Transport Chain High-energy electrons

flow down their electric potential gradient from NAD H and
(For convenience, we show pyruvate here in its un-ionized
FADH 2 to 0 1 via membrane proteins that convert the energy
form, pyruvic acid, a lthough at physiological pH it would be
released into a proton-motive force (H gradient).
largely dissociated.) All four electrons and two of the four
Stage IV: ATP Synthesis The proton-motive force powers protons are transferred (Figure 12, 3, reaction 6) to two mol-
the synthesis of ATP as protons flow down their concentra- ecules of the oxidized form of nicotinamide adenine dinucle-
tion and voltage gradients through the ATP synthesis en- otide (NAD+) to produce the reduced form of the coenzyme,
zyme. For each original glucose molecule, an estimated 28 NADH (see Figure 2-33):
additional ATPs are produced by this mechanism of oxida-
tive phosphorylation.
2H + + 4 e +2 NAD~ ---7 2 NADH
In this section, we discuss stage 1: the biochemical path-
ways that break down glucose into pyruvate in the cytosol.
Later we will see that the energy carried by the electrons in
We also discuss how these pathways are regulated, and con-
NADH and the analogous electron carrier FADH 2, the re-
trast the metabolism of glucose under anaerobic and aerobic
duced form of the coenzyme flavin adenine dinucleotide
conditions. The ultimate fate of pyruvate, once it enters mi-
(FAD), can be used to make additional ATPs via the electron
tochondria, is discussed in Section 12.2.
transport chain. The overall chemical equation for this first
stage of glucose metabolism is

During Glycolysis (Stage I), Cytosolic Enzymes C6H1206 + 2 NAD + + 2 ADP 3 + 2 P, 2 ---7
Convert Glucose to Pyruvate 2 C3H 4 0 3 + 2 NADH + 2 ATP4
Glycolysis, the first stage of glucose oxidation, occurs in the
cytosol in both eukaryotes and prokaryotes; it does not re- After glycolysis, only a fraction of the energy available in
quire molecular oxygen (0 2 ), and is thus an anaerobic pro- glucose has been extracted and co n verted to ATP and
cess. Glycolysis is an example of catabolism, the biological NADH. The rest remains trapped in the covalent bonds of
breakdown of complex substances into simpler ones. A set the two pyruvate molecules. The ability to efficiently convert
of 10 water-soluble cytosolic enzymes catalyze the reactions the energy remain ing in pyruvate to ATP depends on the
constituting the glycolytic pathway (glyco, "sweet"; lysis, presence of molecular oxygen. As we will see, energy conver-
'"split"), in which one molecule of glucose is converted to sion is substantially more efficient under aerobic conditions
two molecules of pyruvate (Figu re 12-3). All the reaction than under anaerobic conditions.
intermediates produced by these enqmes are water-soluble,
phosphorylated compounds called metabolic intermediates.
In addition to chemically converting one glucose molecule
The Rate of Glycolysis Is Adjusted
into two pyruvates, the glycolytic pathway generates four
ATP molecules by phosphorylation of four ADPs (reactions to Meet the Cell's Need for ATP
7 and 10). ATP is formed directly through the enzyme- To maintain appropriate levels of ATP, cells must control the
catalyzed joining of ADP and P, that is derived from phos- rate of glucose catabolism. The operation of the glycolytic
phorylated metabolic intermediates; this process is ca lled p;:rrhway (stage I), as well as the citric acid cycle (stage II), is
substrate-level phosphorylation (to distingu ish it from the continuously regu lated, primarily by allosteric mechanisms
oxidative phosphorylation that generates ATP in stages III (see Chapter 3 for gene ral principles of allosteric control).
and IV). Substrate-level phosphorylation in glycolysis, which Three allosteric enzymes involved in glycolysis play a key role
does not involve the use of a proton-motive force, requires in regulating the entire glycolytic pathway. Hexokmase (Fig-
the prior addition (in reactions 1 and 3) of two phosphates ure 12-3, step 0 ) is inhibited by its reaction product, glucose

520 CHAPTER 12 • Cellular Energetics

 vt~~~
FIGURE 12-3 The glycolytic pathway. A series of ten reactions
degrades glucose to pyruvate. Two reactions consume ATP, forming
ADP and phosphorylated sugars (red), two generate ATP from ADP
Glucose
by substrate-level phosphorylation (green), and one yields NADH by
H~H
reduction of NAD (yellow). Note that all the intermediates between
D ~ ATP 2
glucose and pyruvate are phosphorylated compounds. Reactions 1, 3, Hexokir
f ADP
CH 2 - 0PO - H OH

vr-0~
and 10, with single arrows, are essentially irreversible (large negative
~G values) under ordinary conditions in cells.
Glucose 6-phosphate
.· HO~bH
P~osphogl~cos•1l
:omerese H OH CH 2 OPO •-

Fructose 6-phosphatef- ~O~:,OH

..~. . ···~·r ATP 'H'
Fructo:.-~~;,p:::phote ~~::;H, O
;:, H
~
II ~~6H
((:::========= OH H
H 0 OH IV '9" 1VS8
Ill phosphate
:omerase
Glyceraldehyde 0 H
II I
H
I
- HO3 PO-C-C-C-H Dihydroxyacetone
3-phosphate HC - C-C-H
I I phosphate
H H (2 molecules ) I I
HO OPO/ -
GI 3-ohosphate
d t J It
1F 2 NAD•.,..+•. "l
2 NADH + 2 H7
2m •

6-phosphate. Pyruvate kinase (step mll is inhibited by ATP, 1,3-Bisphosphoglycerate

(2 molecules)
so glycolysis slows down if too much ATP is present. The
third enzyme, phosphofructokinase-] (step II), is the principal IJ Phosr _. " ·-1~
2 ADP
k1nase ~ 2 ATP
HO

rate-limiting enzyme of the glycolytic pathway. Emblematic of

0 H H
its critical role in regulating the rate of glycolysis, this enzyme 3-Phosphoglycerate
is allosterically controlled by several molecules (Figure 12-4). -o - C- C- C- H
(2 molecules)
I
For example, phosphofructokinase-] is allostericall y in- HO
hibited by ATP and allosterically activated by adenosine II eoc -"...;·1l
monophosphate (AMP). As a result, the rate of glycolysis is 0 H H
2-Phosphoglycerate II I I
very sensitive to the cell's energy charge, a measure of the frac- - o - C- C- C- H
(2 molecules) I
tion of total adenosine phosphates that have "high-energy"

- - ·12 H,O
2
0 3 PO OH
phosphoanhydride bonds, which is equal to ([ATP] + 0.5
[ADP])/([ATP] + [ADP] + [AMP]). The allosteric inhibition 0 H
of phosphofructokinase-] by ATP may seem unusual, be- Phosphoenolpyruvate
cause ATP is also a substrate of this enzyme. But the affinity (2 molecules)
of the substrate-binding site for ATP is much higher (has a
.,."··••g ~ 2ADP
lower Kml than that of the allosteric site. Thus at low con-
centrations, ATP binds tQ the catalytic but not to the inhibi-
k;n.•At 2ATP
0 0 H
Pyruvate II II I
tory allosteric site, and enzymatic catalysis proceeds at near
(2 molecules)
-o-c- c- c- H
maximal rates. At high concentrations, ATP also binds to I
H
the allosteric site, inducing a conformational change that re-
duces the affinity of the enzyme for the other substrate, fruc-
tose 6-phosphate, and thus reduces the rate of this reaction
and the overall rate of glycolysis. of a metabolite (here, fructose 6-phosphate) accelerates its
Another important allosteric activator of phosphofructo- subsequent metabolism. Fructose 2,6-bisphosphate allosteri-
kinase-] is fructose 2,6-bisphosphate. This metabolite is call y activates phosphofructokinase-1 in livrr cells by de-
formc::J from fructose 6-phosphate by an enzyme called creasing the inhibitory effect of high ATP and by increasing
phosphofructokinase-2. Fructose 6-phosphate acce lerates the affinity of phosphofructokinase-1 for one of its sub-
the formation of fructose 2,6-bisphosphate, which in turn strates, fructose 6-phosphate.
activates phosphofructokinase-]. This type of control is The three glycolytic enzymes that are regulated by allo-
known as feed-forward activation, in which the abundance stery catalyze reactions with large negative .lG 0 ' values-

12.1 First Step of Harvesting Energy from Glucose: Glycolysis 521

 H;gh [AMP]~ t
High [ATPI

/ gh [dt<•tol
·.
Fructose Fructose

ho•ph•to .,?{'· .,: 1,6·bl•pho•ph•te-- ->To "'""'"

Phosp lofructo-

Gl"'"" ::;:p\
-7

Insulin -o- fructokrnas T

A Fructose
2,6-bisphosphate
FIGURE 12-4 Allosteric regulation of glucose metabolism. kinase activity forms fructose 2,6-bisphosphate from fructose
The key regulatory enzyme in glycolysis, phosphofructokinase-1, is 6-phosphate, and its phosphatase activity catalyzes the reverse
allosterically activated by AMP and fructose 2,6-bisphosphate, which reaction. Insulin, which is released by the pancreas when blood
are elevated w hen the cell's energy stores are low. The enzyme is glucose levels are high, promotes PFK2 kinase activity and thus
inhibited by ATP and citrate, both of w hich are elevated when the cell stimulates glycolysis. At low blood glucose, glucagon is released by
is actively oxidizing glucose to C02 (i.e., when energy stores are high). the pancreas and promotes PFK2 phosphatase activity in the liver,
Later we will see that citrate is generated during stage II of glucose indirectly slowing down glycolysis.
oxidat ion. Phosphofructokinase-2 (PFK2) is a bifunctional enzyme: its

reactions that arc essen ti ally irreversible und er o rdinary with the concomitant production of a large amount of ATP.
conditions. These enzymes thus are particularly suitable fo r Most eukaryotes, however, can generate some ATP by an-
regulating the entire glyco lytic pathway. Additional control aerobic metabolism. A few eukaryotes are facultative anaer-
is exerted by glyceraldeh yde 3-phosp hate dehydrogenase, obes: th ey g row in either the presence or the absence of
which catalyzes the reduction of NAD to NADH (see Fig- oxygen. For example, annelids, mollusks, and some yeasts
ure 12-3, step (:'!). As we shall see, NADH is a high-energy can survive without oxygen, relying on the ATP produced by
electron carrier used subsequently during oxidative phos- fermentation.
phorylation in mitochondria. If cytosolic NADH builds up In the absence of oxygen, yeasts convert the pyruvate
owing to a slowdown in mitochondrial oxidation, step (:'! produced by glycolysis to one molecule each of ethano l and
becomes thermodynamically less favorable. C0 2; in these reactions two NADH molecules are oxidized
Glucose metabolism is controlled differently in various to NAD for each two pyruvates converted to ethanol,
mammalian tissues to meet the metabolic needs of the organ- thereby regenerating the supply of NAD •, which is necessary
ism as a whole. I)uring periods of carbohydrate starvation, for for glycolysis to continue (Figu re 12-Sa, left). This anaerobic
instance, it is necessary for the liver to release glucose into the catabolism of glucose, ca lled fermentation, is the basis of
bloodstream. To do this, the liver converts the polymer glyco- beer and wine production.
gen, a storage form of glucose (Ch apter 2), directly to glucose Fermentation also occurs in animal cells, although lactic
6-phosphate (wit hout involvement of hexokinase, step 0 ). acid rather than alcohol is the product. During prolonged
Under these conditions, there is a reduction in fructose 2, contraction of mammalian skeletal muscle cells-for exam-
6-bisphosphate levels and decreased phosphofructokinase- I ple, during exercise-oxygen w ithin t he muscle tissue can
activity (Figure 12-4 ). As a result, g lucose 6-phosphate derived become scarce. As a consequence, glucose catabolism is lim-
from glycogen is not metabolized to pyruvate; rather, it is con- ited to glycolysis and muscle cells convert pyruvate to two
vened to glucose by a phosphatase and released into the blood molecules of lactic acid by a red uction reaction that also
to nourish the brain and red blood cells, which depend primar- oxidizes two NADHs to two NAD+s (figure 12-Sa, right).
ily on glucose for their energy. In all cases, the activity of these Although the lactic acid is released from the muscle into the
regulated enzymes is controlled by the level of small -molecule blood, if the contracti o ns ar e sufficiently rapid and strong,
metabolites, generally by allosteric interactions, or by hormone- the lactic acid can t ransiently accumu late in the tissue and
mediated phosphorylation and dephosphorylation reactions. contribute to muscle and joint pain during exercise. Once it
(Chapter 15 gives a more detailed discussion of hormonal con- is secreted into the blood, some of t he lactic acid passes into
trol of glucose metabolism in liver and muscle.) the li ver, where it is reoxid ized to pyruvate and ei ther further
metabolized to CO1 aerobically or converted back to glu-
cose. Much lactate is metaboli zed to C0 2 by the heart, which
is highly perfused by blood and ca n continue aerobic me-
Glucose Is Fermented When Oxygen Is Scarce tabolism at times when exercising, oxygen-poor skeletal
\1any eukaryotes, including humans, are obligate aerobes: muscles secrete lactate. Lactic acid bacteria (the organisms
they grow only in the presence of molecular oxygen and can that spoil milk) and other prokaryotes also generate ATP by
metabolize glucose (or related sugars) completely to C0 2 , the fermentation of glucose to lactic acid.

522 CHAPTER 12 • Cellular Energetics

 ......

(a) (b)
ANAEROBIC METABOLISM (FERMENTATION) AEROBIC METABOLISM

Yeast Muscle
CYTOSOL CYTOSOL CYTOSOL

'J ADP + 2 NAO+ + 2 P1 2 ADP + 2 NAD+ + 2 P, 2 ADP + 2 NAD + 2 P,

Glycolysis

0
K 0
2 ATP + 2 NADH + 2 P1
+2 H20
Glycolysis

K
0
2 ATP + 2 NADH + 2 P;
+ 2 H20
Glycolysis

0
K
0
2 ATP + 2 NADH
+ 2 H20
~ 2 P;

,I

h
C OH C OH -C C-OH
Pyruvic acid Pyruvic acid

Pyruvate
decarboxylase >2
Lactate
'lya·og~:'l:se
~ NADH + W Transfer into
mitochondrion
1
C02 NAD~

0 OH 0 MITOCHONDRION
i II
CH 3 ·CH CH 3 CH C-OH 0 0

F
X2 II II
Acetaldehyde Lactic acid CH3- C- C- OH
Pyruvic acid
Alcohol NADH + W
dehydrogenal
NAD+
C02
CoA-SH
'>( 2

NADH

Overall reactions of anaerobic metabolism:

Glucose+ 2 ADP + 2 P; ~ 2 ethanol+ 2 C0 2 + 2 ATP + 2 H20
Glucose+ 2 ADP + 2 P1 ~ 2 lactate-+ 2 ATP + 2 H2 0
NADH

NAD•

-28 ADP + -28 P,

Overall reaction of aerobic metabolism:

Glucose+ 6 0 2 + -30 ADP + -30 P, ~

6 C02 + 36 H2 0 + -30 ATP

FIGURE 12-5 Anaerobic versus aerobic metabolism of glucose. NADH reduces pyruvate to form lactic acid, regenerating NAD , a
The ultimat e fate of pyruvate formed during glycolysis depends on the process ca lled lactic acid fermentation. (b) In the presence of oxygen,
presence or absence of oxygen. (a) In the ab~Pnce of oxygen, pyruvate pyruvate is transported into mit ochondria, where first It IS converted
is o nly pa rtially degraded and no further ATP is made. However, by pyruvate dehydrogenase into one molecule of C02 and one of
two electrons are transferred from each NADH molecule produced acetic acid, the latter linked to coenzyme A (CoA-SH) to form acetyl
during glycolysis to an acceptor molecule to regenerate NAD• , which CoA, concomitant with reduction of one molecule of NAD to NADH.
is required for continued glycolysis. In yeast (left), acetaldehyde is the Further metabolism of acetyl CoA and NADH generates approximately
electron acceptor and ethanol is the product. This process is called an additional 28 molecules of ATP per glucose molecule oxidized.
alcoholic fermentation. When oxygen is scarce in muscle cells (right),

12.1 First Step of Harvesting Energy from Glucose: Glycolysis 523

 Fermentation is a much less efficient way to generate 12.2 Mitochondria and the Citric
ATP than aerobic oxidation, and therefore in animal cells
only occurs when oxygen is scarce. In the presence of oxy-
Acid Cycle
gen, pyruvate formed by glycolysis is transported into mito- Oxygen-producing photosynthetic cyanobacteria appeared
chondria, where it is oxidized by 0 2 to C0 2 and H 2 0 via the about 2.7 bill ion years ago. The subsequent buildup in the
series of reactions outlined in Figure 12-5b. This aerobic earth's atmosphere of sufficient oxygen (0 2 ) during the next
metabolism of glucose, which occurs in stages II-TV of the approximately billion years opened the way for organisms to
process outlined in Figure 12-1, generates an estimated 28 ad- evolve the very efficient aerobic oxidation pathway, which
dition;~ I ATP molecules per original glucose molecule, far in turn pt:nnim:d the evolution, especially during what JS
outstripping the ATP yield from anaerobic glucose metabo- called the Cambrian explosion, of large and complex body
lism (fermentation). forms and associated metabolic activities. In eukaryotic cells,
To understand how ATP is generated so efficiently by aerobic oxidation is carried out by mitochondria (stages II-IV).
aerobic oxidation, we must consider first the structure and In effect, mitochondria are ATP-generating factories, taking
function of the organelle responsible, the mitochondrion. full advantage of this plentiful oxygen. We first describe
Mitochondria, and the reactions that take place within them, their structure and then the reactions they employ to degrade
are the subjects of the next section. pyruvate and make ATP.

Mitochondria Are Dynamic Organelles

with Two Structurally and Functionally
KEY CONCEPTS of Section 12.1
Distinct Membranes
First Step of Harvesting Energy
Mitochondria (Figure 12-6) are among the larger organelles
from Glucose: Glycolysis
in the cell. A mitochondrion is about the size of an E. coli
• In a process known as aerobic oxidation, cells convert the bacterium, which is not surprising, because bacteria are
energy released by the oxidation ("burning") of glucose or thought to be the evolutionary precursors of mitochondria
fatty acids into the terminal phosphoanhydride bond of ATP. (see discussion of the endosymbiont hypothesis, below). Most
• The complete aerobic oxidation of each molecule of glu- eukaryotic cells contain many mitochondria, which may col-
cose produces six molecules of C02 and approximately 30 ATP lectively occupy as much as 25 percent of the volume of the
molecules. The entire process, which starts in the cytosol and cytoplasm. The numbers of mitochondria in a cell, hundreds
is completed in the mitochondrion, can be divided into four to thousands in mammalian cells, are regulated to match the
stages: (T) degradation of glucose to pyruvate in the cytosol cell's requirements for ATP (e.g., stomach cells, which use a
(glycolysis); (II) pyruvate oxidation to C0 2 in the mitochon- lot of ATP for acid secretion, have many mitochondria).
drion coupled t9 generation of the high-energy electron car- The details of mitochondrial structure can be observed
riers NADH and FADH2 (via the citric acid cycle); (III) elec- with electron microscopy (see Figure 9-33). Each mitochon-
tron transport to generate a proton-motive force together drion has two distinct, concentric membranes: the inner and
with conversion of molecular oxygen to water; and (IV) ATP outer membrane. The outer membrane defines the smooth
synthesis (see Figure 12-1). From each glucose molecule two outer perimeter of the mitochondrion . The inner membrane
ATPs are generated by glycolysis (stage I) and approximately lies immediately underneath the outer membrane and has
28 from stages II-IV. numerous invaginations, called cristae, which extend from
the perimeter of the inner membrane into the center of the
In glycolysis (stage 1), cytosolic enzymes convert glucose mitochondrion. The high membrane curvature at the tips of
to two molecules of pyruvate and generate two molecules the cristae may be due to the presence of a high concentra-
each ofNADH and ATP (see Figure 12-3). tion of dimers of an integral membrane protein that synthe-
The rate of glucose oxidation via glycolysis is regulated by sizes ATP (the F0 F 1 complex, discussed in Section 12.3). The
the inhibition or stimulation of several enzymes, depending outer and inner membranes topologically define two submi-
on the cell's need for ATP. Glucose is stored (as glycogen or tochondrial compartments: the intermembrane space, be-
fat) when ATP is abundant (see Figure 12-4). tween the outer and inner membranes, and the matrix, or
• In the absence of oxygen (anaerobic conditions), cells can central compartment, which forms the lumen within the
metabolize pyruvate to lactic acid or (in the case of yeast) to inner membrane. The invaginating cristae greatly expand the
ethanol and col, in the proCt>'iS converting NADH back to surface area of the inner mitochonclri;~l membrane, thus in-
, ··
NAD ' , which is necessary for continued glycolysis. ln the pres- creasing the ability to synthesize ATP. In typical liver mito-
ence of oxygen (aerobic conditions), pyruvate is transported chondria, for example, the area of the inner membrane,
into the mitochondrion, where it is metabolized to C02 , in the including cristae, is about five times that of the outer mem-
process generating abundant ATP (see Figure 12-5). brane. In fact, the total area of all inner mitochondrial mem-
branes in liver cells is about 17 times that of the plasma

524 CHAPTER 12 • Cellular Energetics

 0 VIDEO: Mitochondrion Reconstructed by Electron Tomography

(a) (b)
F0 F1 complexes

lntermembrane Cristae
space
Outer Cristae junctions
membrane
Inner
·. membrane

FIGURE 12-6 Internal structure of a mitochondrion. (a) Schematic contains the mitochondrial DNA (blue strand), ribosomes (small blue
diagram showing the principal membranes and compartments. The spheres), and granules (large yellow spheres). (b) Computer-generated
smooth outer membrane forms the outside boundary of the mito- model of a section of a mitochondrion from chicken brain. This model
chondrion. The inner membrane is distinct from the outer membrane is based on a three-dimensional electron microscopic image calculated
and is highly invaginated to form sheets and tubes called cristae. The from a series of two-dimensional electron micrographs recorded at
relatively small uniform tubular structures that connect the cristae to regular intervals. This technique is analogous to a three-dimensional
the portions of the inner membrane that are juxtaposed to the outer x-ray tomogram or CAT scan used in medical imaging. Note the tightly
membrane are called crista junctions. The intermembrane space is con- packed cristae (yellow-green), the inner membrane (light blue), and the
tinuous with the lumen of each crista. The F0F1 complexes (small red outer membrane (dark blue). [Part {b) courtesy ofT. Frey, from T. Frey and
spheres). which synthesize ATP, are intramembrane particles that pro- C. Mannella, 2000, Trends Biochem. Sci. 25:319.]
trude from the cristae and inner membrane into the matrix. The matrix

membrane. The mitochondria in heart and skeletal muscles and cause human disease. An example is the inherited neuro-
contain three times as many cristae as are found in typical muscular disease Charcot-!\.larie-Tooth subtype 2A, in which
liver mitochondria-presumably reflecting the greater de- defects in peripheral nerve function lead to progressive mus-
mand for ATP by muscle cells. cle weakness, mainly in the feet and hands. The ongoing fu-
Analysis of fluoresceptly labeled mitochondria in living sion and fission process appears to protect mitochondrial
cells has shown that mitochondria are highly dynamic. They DNA from accumulating mutations, and may permit the iso-
undergo frequent fusions and fissions that generate tubular, lation of dysfunctional or damaged segments of mitochon-
sometimes branched networks (figure 12-7), which may ac- dria that can be specifically targeted for destruction in the cell
count for the wide variety of mitochondrial morphologies by a process called autophagy (see Chapter 14).
seen in different types of cells. When individual mitochondria Fractionation and purification of mitochondrial mem-
fuse, each of the two membranes fuse (inner with inner, and branes and compartments have made it possible to determine
outer with outer) and each of their distinct compartments their protein, DNA, and phospholipid compositions and to lo-
intermix (matrix with matrix, intermembrane space with in- calize each enzyme-catalyzed reaction to a specific membrane
rcrmembrane spat.:t:). Fusions and fissions apparently play a or compartment. Over 1000 different types of polypeptides are
functional role as well, because genetic disruptions in several required to make and maintain mitochondria and permit them
GTPase superfamily genes that are required for these dy- to function. Detailed biochemical analysis has established that
namic processes can disrupt mitochondrial function, such as there are at least 1098 proteins in mammalian mitochondria
maintenance of proper inner membrane electric potential, and perhaps as many at l500. Only a small number of these-

12.2 Mitochondria and the Citric Acid Cycle 525

0 VIDEO: Mitochondrial Fusion and Fission

EXPERIMENTAL FIGURE 12-7 M itochondria undergo rapid

f usion and fission in living cells. Mitochondria labeled with a
fluorescent protein in a living normal murine embryonic fibroblast
Fusio n
were observed using time-lapse fluorescence microscopy. Several
mitochondria undergoing fusion (top) or fission (bottom) are artificially
highlighted in blue and with arrows. [Modified from D. C. Chan, 2006, Cell
125(7):1241-1 252.)

Fission

13 in humans-are encoded by mitochondrial DNA genes and Note that plants have mitochoudria and perform aerobic
synthesized inside the mitochondrial matrix space. The re- oxidation as well. In plants, stored carbohydrates, mostly in
maining proteins are encoded by nuclear genes (Chapter 6), the form of starch, are hydrolyzed to glucose. Glycolysis then
synthesized in the cytosol, and then imported into mitochon- produces pyruvate that is transported into mitochondria, as
dria (Chapter 13). Defective functioning of the mitochondrial- in animal cells. Mitochondrial oxidation of pyruvate and
associated proteins, due for example to inherited genetic concomitant formation of ATP occur in photosynthetic cells
mutations, leads to over 150 human diseases. The most com- during dark periods when photosynthesis is not possible, and
mon of these are electron transport chain diseases, which result in roots and other nonphotosynthetic tissues at all times.
from mutations in any one of 92 protein-encoding genes and The mitochondrial inner membrane, cristae, and matrix arc
exhibit a very wide variety of clinical abnormalities affecting the sites of most reactions involving the oxidation of pyruvate
muscles, the heart, the nervous system, and the liver, among and fatty acids to C02 and H 20 and the coupled synthesis of
other physiologic systems. Other mitochondrial-associated dis- ATP from ADP and P, with each reaction occurring in a dis-
eases include Miller syndrome, which results in multiple ana- crete membrane or space in the mitochondrion (Figure 12-8).
tomical malformations, and connective tissue defects. We now continue our detailed discussion of glucose oxi-
The most abundant protein in the outer membrane is mi- dation and ATP generation, exploring what happens to the
tochondrial porin, a transmembrane channel protein similar pyruvate generated during glycolysis (stage I) after it is trans-
in structure to bacterial porins (see Figure l 0-18). Ions and ported into the mitochondrial matrix. The last three of the
most small molecules (up to about 5000 Da) can readily pass four stages of glucose oxidation are:
through these channel proteins when they are open. Al-
though there may be metabolic regulation of the opening of • Stage II. Stage II can be subdivided into two distinct partS:
mitochondrial porins and thus the flow of metabolites across (1) the conversion of pyruvate to acetyl CoA, followed by
the outer membrane, the inner membrane and its cristae are (2) oxidation of acetyl CoA to C02 in the citric acid cycle. These
the major permeability barriers between the cytosol and the oxidations are coupled to reduction of NAD ... to NADH and of
mitochondrial matrix, limiting the rate of mitochondrial FAD to FADH 2 • (Fatty acid oxidation follows a similar route,
oxidation and ATP generation. with conversion of fatty acyl CoA to acetyl CoA.) Most of the
Protem constitutes 76 percent of the total mass of the reactions occur in or on the membrane facing the matrix.
inner mitochondrial membrane-a higher fraction than in • Stage ill. Electron transfer from NADH and FADH 2 to 0 2
any other cellular membrane. Many of these proteins are ke) via an electron transport chain within the inner membrane,
participants in oxidative phosphorylation. They include ATP which generates a proton-motive force across that membrane.
synthase, proteins responsible for electron transport, and a
wide variety of transport proteins that permit the movement • Stage IV. Harnessing the energy of the proton-motive force
of metabolites between the cytosol and the mitochondrial for ATP synthesis in the mitochondrial inner membrane. Stages
matrix. The human genome encodes 48 members of a family III and IV are together called oxidative phosphorylation.
of mitochondrial transport proteins. One of these is called
the ADP/ATP carrier, an antiporter that moves newly syn-
In the First Part of Stage II, Pyruvate Is Converted
thesized ATP out of the matrix and into the inner membrane
space (and subsequently the cytosol) in exchange for ADP to Acetyl CoA and High-Energy Electrons
originating from the cytosol. Without this essential anti- Within the mitochond rial matrix, pyruvate reacts with co-
porter, the energy trapped in the chemical bonds in mito- enzyme A, forming C0 2 , acetyl CoA, and NADH (Figure
chondrial ATP would not be available to the rest of the cell. 12-8, Stage 11/eft). This reaction, catalyzed by pyruvate

526 CHAPTER 12 • Cellular Energetics

 Outer mitochondrial membrane (penneable to metabolites)

lntennembrane space

Stage I

22 ~::~ +
Glucose

2 Pyruvate
2ATP

Fatty acid
-~--> Pyruvate -E--+---+-...:::....-,.......,..-~ Acetyl CoA _ ___.....
NADH

H~~~~ ~ NAD·
FADH 2
AMP +
PP; FAD
Mitochond1a1 matrix
Fatty acyl ---.,.-----~ 3 NADH
Co A

Stage Ill Stage IV ATP

NADH~NAD•
Succinate
+ Traneporters ADP
2e +2W+ I0 2--> H 0
~ T)/' ;~P,
2

+ - - - - - NAD NADH 02
.~ Fumarate~
FA
H20
IVj 3W
~ OH

" . ~
W 3H
Electron transport chain F0F1 complex

FIGURE 12-8 Summary of aerobic oxidation of glucose and fatty and C02• Stage Ill: Electron transport reduces oxygen to water and
acids. Stage 1: In the cytosol. glucose is converted to pyruvate (glycoly- generates a proton-motive force. Electrons (blue} from reduced coen-
sis} and fatty acid to fatty acyl CoA. Pyruvate and fatty acyl CoA then zymes are transferred via electron-transport complexes (blue boxes} to
move into the mitochondrion. Mitochondrial porins make the outer 0 2 concomitant with transport of H ions (red) from the matrix to the
membrane permeable to these metabolites, but specific transport intermembrane space, generating the proton-motive force. Electrons
.· proteins (colored ovals} in the inner membrane are required to import from NADH flow directly from complex I to complex Ill, bypassing com-
pyruvate (yellow} and fatty acids (blue} into the matrix. Fatty acyl plex II. Electrons from FADH2 flow directly from complex II to complex
groups are transferred from fatty acyl CoA to an intermediate carrier, Ill, bypassing complex I. Stage IV: ATP synthase, the F0 F1 complex
transported across the inner membrane (blue oval}, and then reat- (orange}, harnesses the proton-motive force to synthesize ATP in the
tached to CoA on the matrix side. Stage II: In the mitochondrial matrix, matrix. Antiporter proteins (purple and green ovals} transport ADP and
pyruvate and fatty acyl CoA are converted to acetyl CoA and then P, into the matrix and export hydroxyl groups and ATP. NADH gener-
oxidized, releasing C02 • Pyruvate is converted to acetyl CoA with the ated in the cytosol is not transported directly to the matrix because the
formation of NADH and C02; two carbons from fatty acyl CoA are con- inner membrane is impermeable to NAD and NADH; instead, a shuttle
verted to acetyl CoA with the formation of FADH 2 and NADH. Oxidation system (red) transports electrons from cytosolic NADH to NAD in the
of acetyl CoA in the citric acid cycle generates NADH and FADH 2, GTP, matrix. 0 2 diffuses into the matrix, and C02 diffuses out.

dehydrogenase, is hi ghly excrgonic (~Go' = -8.0 kcal!mol) oxidized to C0 2 via the citric acid cycle. Note that the two
and essentially irreversible. carbons in the acetyl group come from pyruvate; the third
Acetyl CoA is a molecule consisting of a two-carbon ace- carbon of pyruvate is released as carbon dioxide.
tyl group covalently linked to a longer molecule known as
coenzyme A (CoA) (Figure 12-9). It plays a central role in In the Second Part of Stage II, the Citric Acid
the oxidation of pyruvate, fatty acids, and amino acids. In
Cycle Oxidizes the Acetyl Group in Acetyl CoA
addition, it is an intermediate in numerous biosynthctic reac-
tions, including transfer of an acetyl group to histone pro- to C02 and Generates High-Energy Electrons
teins and many mammalian proteins, and synthesis of lipids Nine sequential reactions operate in a cycle to oxidize the ace-
such as cholesterol. In respiring mitochondria, however, the tyl group of acetyl CoA to C02 (Figure 12-8, Stage II right).
two-carbon acetyl gro up of acetyl CoA is almost always The cycle is referred to by several names: the citric acid cycle,

12.2 Mitochondria and the Citric Acid Cycle 527

 0 H H H CH 3 0 0
II I
H3C- C - S (CH2b-N C -(CH 2)2- N - C-C- C- CH 2- 0 - P- 0-P - 0- Ribose - Adenine
Acetyl I I I I I
0 0 OH CH 3 0 0 Phosphate

Coenzyme A (CoAl

FIGURE 12-9 The structure of acetyl CoA. This compound, pyruvate, fatty acids, and many amino acids. It also contributes acetyl
consisting of an acetyl group covalently linked to a coenzyme A (CoAl groups in many biosynthetic pathways.
molecule, is an important intermediate in the aerobic oxidation of

the tricarboxylic acid (TCA) cycle, and the Krebs cycle. The reaction 9; thus three NADHs are generated per turn of the
net result is that for each acetyl group entering the cycle as cycle. In reaction 7, two electrons and two protons are trans-
acetyl CoA, two molecules of C02 , three of NADH, and one ferred to FAD, yielding the reduced form of this coenzyme,
each of FADH2 and GTP arc produced. NADH and FADH2 FADH2 • Reaction 7 is distinctive because not only is it an
are high-energy electron carriers that will play a major role in intrinsic part of the citric acid cycle (stage II), but it is also
stage III of mitochondrial oxidation: electron transport. catalyzed by a membrane-attached enzyme that, as we shall
As shown in Figure 12-10, the cycle begins with conden- see, also plays an important role 'i n stage III. In reaction 6,
sation of the two-carbon acetyl group from acetyl CoA with hydrolysis of the high-energy thioester bond in succinyl CoA
the four-carbon molecule oxaloacetate to yield the six-carbon is coupled to synthesis of one GTP by substrate-level phos-
citric acid (hence the name citric acid cycle) . In both reactions phorylation. Because GTP and ATP are interconvertible,
4 and 5, a C02 molecule is released and NAD+ is reduced to
NADH. Reduction of NAD to NADH also occurs during GTP + ADP ~ GDP + ATP

NADH + H " coo~

HO

Malate cis-Aconitate

II
coo
tH
coo
HC
too
F u m a r a t e f r CH2
CH 2
coo coo
I
CH2
I

CH2
CH2
t=O
T-
' H- e- coo
HO-T-H
COO
lsocitrate
~~----- CH 2 ~----~~-
D
FAD COO
1 coo
C SCoA a-Keto-
0 + HSCoA glutarate C02 + NADH .. H •
Succinyl
GT 0
.._ HSCoA CoA CO NADH · t-' T

FIGURE 12-1 0 The citric acid cycle. Acetyl CoA is metabolized to FADH 2. and one molecule of GTP. The two carbon atoms that enter the
C02 and the high-energy electron carriers NADH and FADH 2• In reac- c..ycle with acetyl CoA are highlighted in blue through succinyl CoA. In
tion 1, a two-carbon acetyl residue from acetyl CoA condenses with succinate and fumarate, which are symmetric molecules, they can no
the four-carbon molecule oxaloacetate to form the six-carbon citrate. longer be specifically denoted. Isotope-labeling studies have shown
In the remaining reactions (2~9) each molecule of citrate is eventu- that these carbon atoms are not lost in the turn of the cycle in which
ally converted back to oxaloacetate, losing two C02 molecules in the they enter; on average, one will be lost as C02 during the next turn of
process. In each turn of the cycle, four pairs of electrons are removed the cycle and the other in subsequent turns.
from carbon atoms, forming three molecules of NADH, one molecule of

528 CHAPTER 12 • Cellular Energetics

 TABLE 12-1 Net Result ofthe Glycolytic Pathway and the Citric Acid Cycle

C02 Molecules NAD+ Molecules FAD Molecules

Reaction Produced Reduced to NADH Reduced to FADH 2 ATP(ORGTP)

I glucose molecule to 2 pyruvate molecules 0 2 0 2

2 pyruvates to 2 acetyl CoA molecules 2 2 0 0

2 acetyl CoA to 4 C02 molecules 4 6 2 2

Total 6 10 2 4

this can be considered an ATP-generating step. Reaction 9 FAD-dependent reactions are to continue.) As we will see in
regenerates oxaloacetate, so the cycle can begin again. Note the next section, the electron transport chain within the mito-
that molecular 0 2 does not participate in the citric acid cycle. chondrial inner membrane converts NADH to NAD and
Most enzymes and small molecules involved in the citric FADH2 to FAD as it reduces 0 2 to water and converts the en-
acid cycle are soluble in the aqueous mitochondrial matrix. ergy stored in the high-energy electrons in the reduced forms of
These include CoA, acetyl CoA, succinyl CoA, NAD+, and these molecules into a proton-motive force (stage III ). Even
NADH, as well as most of the eight cycle enzymes. Succinate though 0 2 is not involved in any reaction of the citric acid
dehydrogenase (reaction 7), however, is a component of an cycle, in the absence of 0 2 this cycle soon stops operating as
integral membrane protein in the inner membrane, with its the intramitochondrial supplies of NAD+ and FAD dwindle
active site facing the matrix. When mitochondria are dis- due to the inability of the electron transport chain to oxidize
rupted by gentle ultrasonic vibration or by osmotic lysis, non- NADH and FADH2 • These observations raise the question of
membrane-bound enzymes in the citric acid cycle are released how a supply of NAD~ in the cytosol is regenerated.
•.
as very large multiprotei n complexes. It is believed that, If the NADH from the cytosol could move into the mito-
within such complexes, the reaction product of one enzyme chondrial matrix and be oxidized by the electron transport
passes directly to the next enzyme without diffusing through chain and if the NAD product could be transported back
the solution. Much work is needed to determine the struc- into the cytosol, regeneration of cytosolic NAD would be
tures of these large enzyme complexes as they exist in the cell. simple. However, the inner mitochondrial membrane is im-
Because glycolysis of one glucose molecule generates two permeable to NADH. To hypass this problem and permit the
acetyl CoA molecules, the reactions in the glycolytic pathway electrons from cytosolic NADH to be transferred indirectly
and citric acid cycle produce six C0 2 molecules, 10 NADH to 0 2 via the mitochondrial electron transport chain, cells
molecules, and two FADH 2 molecules per glucose molecule use several electron shuttles to transfer electrons from cyto-
(Table 12-1 ). Although these reactions also generate four solic NADH to NAD+ in the matrix. Operation of the most
high-energy phosphoanhydride bonds in the form of two ATP widespread shuttle-the malate-aspartate shuttle-is de-
and two GTP molecules, this represents only a small fraction picted in Figure 12-11.
of the available energy released in the complete aerobic oxida- For every complete "turn" of the cycle, there is no over-
tion of glucose. The remaining energy is stored as high-energy all change in the numbers of NADH and NAD molecules
electrons in the reduced coenzymes NADH and FADH2 , or the intermediates aspartate or malate used by the shuttle.
which can be thought of as "electron carriers." The goal of However, in the cytosol, NADH is oxidized to NAD+, which
stages III and IV is to rec~ver this energy in the form of ATP. can be used for glycolysis, and in the matrix, NAD is re-
duced to NADH, which can be used for electron transport:
Transporters in the Inner Mitochondrial
NADH,"o'ol + NAD matm - t NAD q'tosol + NADHm ..mx
Membrane Help Maintain Appropriate Cytosolic
and Matrix Concentrations of NAD+ and NADH
Mitochondrial Oxidation of Fatty Acids
In the cytosol NAD is required for step mof glycolysis (see
Figure 12-3), and in the mitochondrial matrix !'\AD+ is re- Generates ATP
quired for conversion of pyruvate to acetyl CoA and for three Up to now, we have focused mainly on the oxidation of car-
steps in the citric acid cycle (19, ~.and m in Figure 12-10 ). ln bohydrates, namely glucose, for ATP generation. Fatty acids
each case, NADH is a product of the reaction. If glycolysis are another important source of cellular energy. Cells can take
and oxidation of pyruvate are to continue, NAD ~ must be up either glucose or fatty acids from the extracellular space
regenerated by oxidation of NADH. (Similarly, the FADH 2 with the help of specific transporter proteins (Chapter 11).
generated in stage II reactions must be reoxidized to FAD if Should a cell not need to immediately burn these molecules, it

12.2 Mitochondria and the Citric Acid Cycle 529

 Cytosol NADHcytosol NAD +cytosol

Aspartate

-
Tran~~~::-onase
(inf\
------::>.......,--~>

o Ketoglutarate Glutamate
G luta~ate t_ J
~II/
Oxaloacetate _ ___,_"'-------~> Malate
Maille
J ehydrc.genase

-----«- a-Ke~glutarate ~
coo-
+H3 N - C-H

CH 2
I
coo-
Aspartate
coo-
C= O

CH 2
I
coo-
Oxaloacetate
coo-
I
HO - C- H

CH 2
I
coo-
Malate
Mitochondrial
inner membrane coo- coo-
Glutamate o Ketoglutarate
+H3 N - C-H c-o
I
r 't- CH2
I
CH 2
,. Ketog l"'"'"''"' Glutamate CH2 CH 2
I
Aspartate~<----..::....."""---
\.II/ Oxaloacetate ...;<'---(i---:>E"""""'i\,------
J Malate
coo- coo-
3se Glutamate a-Ketoglutarate

Matrix NADH matrix NAD+ matrix

FIGURE 12-11 The malate-aspartate shuttle. This cyclical series of cannot directly cross the inner membrane, is converted to aspartate
reactions transfers electrons from NADH in the cytosol (intermembrane by addition of an amino group from glutamate. In this transaminase-
space) across the inner mitochondrial membrane, which is imperme- catalyzed reaction in the matrix, glutamate is converted to
able to NADH itself, to NAD in the matrix. The net result is the replace- a-ketoglutarate. Step 1.=1: A second antiporter (red oval) exports
ment of cytosolic NADH with NAD and matrix NAD with NADH. Step aspartate to the cytosol in exchange for glutamate. Step Iii: A cytosolic
0 : Cytosolic malate dehydrogenase transfers electrons from cytosolic transaminase converts aspartate to oxaloacetate and a-ketoglutarate
NADH to oxaloacetate, forming malate. Step f) : An antiporter (blue to glutamate, completing the cycle. The blue arrows reflect the
oval) in the inner mitochondrial membrane transports malate into the movement of the a-ketoglutarate, the red arrows the movement of
matrix in exchange for a-ketoglutarate. Step IJ: Mitochondrial malate glutamate, and the black arrows that of aspartate/malate. It is note-
dehydrogenase converts malate back to oxaloacetate, reducing NAD worthy that, as aspartate and malate cycle clockwise, glutamate and
in the matrix to NADH in the process. Step D: Oxaloacetate, which a-ketoglutarate cycle in the opposite direction.

can store them as a polymer of glucose called glycogen (espe- of glucose. The oxidation of 1 g of triacylglyceride to C0 2
cially in muscle or liver) or as a trimer of fatty acids covalently generates about six times as much ATP as does the oxidation
linked to glycerol, called a triacylglycerol or triglyceride. In of 1 g of hydrated glycogen. Thus triglycerides are more ef-
some cells, exce~s glucose is converted into fatty acids and ficient than carbohydrates for storage of energy, in part be-
then triacylglycerols for storage. However, unlike microor- cause they arc stored in an hydrous form and can yield more
ganisms, animals are unable to convert fatty acids to glucose. energy when oxidized and in part because they are intrinsi-
When the cells need to burn these energy stores to make ATP cally more reduced (have more hydrogens) than carbohydrates.
(e.g., when a resting muscle begins to do work and needs to In mammals, the primary site of storage of triacylglycerides is
burn glucose or fatty acids as fuel), enzymes break down gly- fat (adipose) tissue, whereas the primary sites for glycogen
cogen to glucose or hydrolyze triacylglycerols to fatty acids, storage are muscle and the liver.
which are then oxidized to generate ATP: Just as there are four stages in the oxidation of glucose,
there arc four stages in the oxidation of fatty acids. To opti-
0 mize the efficiency of ATP generation, part of stage II (citric
acid cycle oxidation of acetyl CoA) and all of stages III and
CH 3 - (CH 2 ln - C
IV of fatty acid oxidation are identical to those of glucose
0
I oxidation. The differences lie in the cytosolic stage I and the
CH 3 (CH 2 ln - C first part of the mitochondrial stage II. In stage I, fatty acids
0
are converted to a fatty acyl CoA in the cytosol in a reaction
CH 3 - (CH 2 ln - C coupled to the hydrolysis of ATP to AMP and PP, (i norganic
HO
Triacylglycerol pyrophosphate) (see Figure 12-8):
0 HO CH
0
(CH2 ln C- OH + HO CH 2
Fatty acid Glycerol R C o- + HSCoA + ATP -->
Fatty acid

Fatty acids are the major energy source for some tissues, 0
particularly adult heart muscle. In humam, in fact, more R-C SCoA + AMP+ PP,
ATP is generated by the oxidation of fats than the oxidation Fatty acyl CoA

530 CHAPTER 12 • Cellular Energetics

 Subsequent hydrolysis of PP, to two molecules of P, releases enter the citric acid cycle and are oxidized to C02 • As will be
energy that drives this reaction to completi on. T o transfer the described in detail in the next section, the reduced NADH
fatty acyl group into the mitochondrial matrix, it is covalently and FADH2 with their high-energy electrons will be used in
transferred to a molecule called carnitine and moved across the stage Ill to generate a proton-motive force that in turn is
inner mitochondrial membrane by an acylcarnitine transporter used in stage IV to power ATP synthesis.
protein (see Figure 12-8, blue oval); then, on the matrix side,
the fatty acyl group is released from carnitine and reattached to
Peroxisomal Oxidation of Fatty Acids
another CoA molecule. The activity of the acylcarnitine t rans-
porter is regulated to prevcnr oxidation of fatty acids when Generates No ATP
cells have adequate energy (ATP) supplies. Mitochondrial oxidation of fatty acids is the major source of
In the first part of stage II, each molecule of a fatty acyl ATP in mammalian liver cells, and biochemists at one time
CoA in the mitochondrion is oxidized in a cyclical sequence of believed this was true in a ll cell types. However, rats treated
four reactions in which all the carbon atoms are converted two with clofibrate, a drug t hat affects many features of lipid
at a ti me to acetyl CoA with generation of FADH2 and NADH metabolism, were found to exhibit an increased rate of fatty
(Figure 12-12a). For example, mitochondrial oxidation of each acid oxidation and a large increase in the number of peroxi-
molecule of the 18-carbon stearic acid, CH 3 (CH 2 ) 16COOH, somes in their liver cells. This finding suggested that peroxi-
yields nine molecules of acetyl CoA and eight molecules each somes, as well as mitochondria, can oxidize fatty acids.
of NADH and FADH2 • In the second part of stage IT, as w ith These small organelles, -0.2- 1 f-lm in diameter, are lined by
acetyl CoA gener ated from pyruvate, the~e acetyl groups a single membrane (see Figure 9-32 ). They are present in all

(a) MITOCHONDRIAL OXIDATION (b) PEROXISOMAL OXIDATION

0
R- CH 2 -CH 2 - ' •i C SCoA
Fatty acyl CoA

Electron FAD ~ FAD H2 0 2

transport Dehydr genase 0 >91fase Cat4lase
) chain FADH 2 FADH 2 02
Hp ~
0 H2 0 + 112 0 2
ADP ATP 'I'
+ P; R- CH 2 CH= CH C SCoA

FIGURE 12-12 Oxidation of fatty acids in

mitochondria and peroxisomes.ln both mito-
chondrial oxidation (a) and peroxisomal oxidation
R- CH - CH- CH
2 I 2
c SCoA
(b), fatty acids are converted to acetyl CoA by a
OH series of four enzyme-catalyzed reactions (shown
02
Electron _, NAD · NAD+ NADH down the center of the figure). A fatty acyl CoA
l tranSJ?Ort (.__
~ NADH
Dehydfogenase exported for
NADH - - - - - - - - - r reoxidation
molecule is converted to acetyl CoA and a fatty acyl
CoA shortened by two carbon atoms. Concomi-
0
ATP tantly, one FAD molecule is reduced to FADH2 and
R -CH - C- CH C SCoA one NAD+ molecule is reduced to NADH. The cycle
2 II 2
is repeated on the shortened acyl CoA until fatty
0
acids with an even number of carbon atoms are
CoASH Thi1tase CoASH
completely converted to acetyl CoA. In mitochon-
dria, electrons from FADH2 and NADH enter the
R-CH -C-SCoA electron transport chain and ultimately are used to
2 II generate ATP; the acetyl CoA generated is oxidized
0 in the citric acid cycle, resulting in release of C02
Acyl CoA shortened
by two carbon atoms and ultimately the synthesis of additional ATP.
+ Because peroxisomes lack the electron transport
0 complexes composing the electron transport chain
Citric acid - H C C SCoA - - - - - - -7---+ Acetyl CoA and the enzymes of the citric acid cycle, oxidation
cycle ·' exported
Acetyl CoA of fatty acids in these organelles yields no ATP.
[Adapted from D. L. Nelson and M. M. Cox, Lehninger
Principles of Biochemistry, 3d ed., 2000, Worth Publishers.]

12 .2 Mitochondria and t he Ci tr ic Acid Cycle 531

mammalian cells except erythrocytes and are also found in
plant cells, yeasts, and probably most other eukaryotic cells. • Like glucose oxidation, the oxidation of fatty acids takes
Mitochondria preferentially oxidize short-chain (fewer place in four stages. In stage I, fatty acids are converted to
than 8 carbons in the fatty acyl chain, or <C 8 ), medium- fatty acyl CoA in the cytosol. In stage II, the fatty acyl CoA
chain (C 8-C 12 ), and long-chain (C 1cC 20 ) fatty acids, whereas is first converted into m ultiple acetyl CoA molecules with
peroxisomes preferentially oxidize very long-chain fatty generation of NADH and FADH2 . Then, as in glucose oxi-
acids (VLCFAs, > C 20 ), which cannot be oxidized by mito- dation, the acetyl CoA enters the citric acid cycle. Stages lll
chondria. Most dietary fatty acids have long chains, which and JV are identical for fatty acid and glucose oxidation (see
means they are oxidized mostly in mitochondria. In contrast Figure 12-8).
to mitochondrial oxidation of fatty acids, which is coupled • In most eukaryotic cells, oxidation of short- to long-chain
to generation of ATP, peroxisomal oxidation of fatty acids is fatty acids occurs in mitochondria with production of ATP,
not linked to ATP formation, and energy is released as heat. whereas oxidation of very long-chain fatty acids occurs pri-
The reaction pathway by which fatty acids are degraded to marily in peroxisomes and is not linked to ATP production
acetyl CoA in peroxisomes is similar to that used in mitochon- (see Figure 12-12); the energy released during peroxisomal
dria (Figure 12-12b). However, peroxisomes lack an electron oxidation of fatty acids is converted to heat.
transport chain, and electrons from the FADH2 produced dur-
ing the oxidation of fatty acids are immediately transferred to
0 2 by oxidases, regenerating FAD and forming hydrogen per-
oxide (H 20 2 ). In addition to oxidases, peroxisomes contain ~ ...
abundant catalase, which quickly decomposes the H 20 2, a 12.3 The Electron Transport Chain and
highly cytotoxic metabolite. NADH produced during oxidation
of fatty acids is exported and reoxidized in the cytosol; there is Generation of the Proton-Motive Force
no need for a malate-aspartate shuttle here. Peroxisomes also Most of the energy released during the oxidation of glucose
lack the citric acid cycle, so acetyl CoA generated during per- and fatty acids to C0 2 (stages I and II) is co~verted into high-
oxisomal degradation of fatty acids cannot be oxidized further; energy electrons in the reduced coenzymes NADH and
instead it is transported into the cytosol for use in the synthesis FADH2 • We now turn to stage Ill, in which the energy tran-
of cholesterol (Chapter 10) and other metabolites. siently stored in these reduced coenzymes is converted by an
electron transport chain, also known as the respiratory chain,
into the proton-motive force. We first describe the logic and
KEY CONCEPTS of Section 12.2 components of the electron transport chain. Next we follow
the path of electrons as they flow through the chain, and the
Mitochondria and the Citric Acid Cycle mechanism of proton pumping across the mitochondrial inner
• The mitochondrion has two distinct membranes (outer membrane. We conclude this section with a discussion of the
and inner) and. two distinct subcompartments (intermem- magnitude of the proton-motive force produced hy electron
brane space between the two membranes, and the matrix transport and proton pumping. In Section 12.4, we will see
surrounded by the inner membrane). Aerobic oxidation oc- how the proton-motive force is used to synthesize ATP.
curs in the mitochondrial matrix and on the inner mitochon-
drial membrane (see Figure 12-6). Oxidation of NADH and FADH 2 Releases
• In stage II of glucose oxidation, the three-carbon pyruvate a Significant Amount of Energy
molecule is first oxidized to generate one molecule each of C02,
During electron transport, electrons are released from NADH
NADH, and acetyl CoA. The acetyl group of acetyl CoA is then
and FADH 2 and eventually transferred to 0 2 , forming H 2 0
oxidized to C0 2 hy the citric acid cycle (see Figure 12-8).
according to the following overall reactions:
• Each turn of the citric acid cycle releases two molecules of
C02 and generates three ~ADH molecules, one FADH2 mol- NADH + H"' + \12 0 1 ---* NAD + + H20,
ecule, and one GTP (see Figure 12-10). tiC = -52.6 kcal/mol
• Most of the energy released in stages I and II of glucose FADH2 + V2 0 2 ~FAD+ H 2 0,
oxidation is temporarily stored in the reduced coenzymes .lG = -43.4 kcal/mol
NADH or l-'ADH 2 , which carry high-energy electrons that
subsequently drive the electron transport chain (stage Ill). Recall that the conversion of 1 glucose molecule to C0 2 via
Neither glycolys1s nor the citric acid cycle directly use mo- the glycolytic pathway and citric acid cycle yields 10 NADH
lecular oxygen (0 2). and 2 FADH2 molecules (see Table 12-1 ). Oxidation of these
• The malate-aspartate shuttle regenerates the supply of reduced coenzymes has a total t1G 0 ' of -613 kcal/mol
cytosolic NAD necessary for continued glycolysis (see [10(-52.6) + 2(-43.4)]. Thus of the potential free energy
Figure 12-11). present in the chemical bonds of glucose ( -686 kcal/mol),
about 90 percent is conserved in the reduced coenzymes.

532 CHAPTER 12 • Cellular Energetics

Why should there be two different coenzymes, NADH and to the intermembrane space. In other words, the free energy
FADH2? Although many of the reactions involved in glucose released during the oxidation of NADH or FADH, is stored
and fatty acid oxidation are sufficiently energetic to reduce both as a proton concentration gradient and an electrical gra-
NAD+, several are not, so those reactions are coupled tore- dient across the membrane--collectively, the proton-motive
duction of FAD, which requires less energy. force (see Figure 12-2 ). As we will see in Section 12.4, the
The energy carried in the reduced coenzymes can be re- movement of protons back across the inner membrane,
leased by oxidizing them. The biochemical challenge faced driven by this force, is coupled to the synthesis of ATP from
by the mitochondrion is to transfer, as efficiently as possible, ADP and Pi by ATP synthase (stage IV).
the energy released by this oxidation into the energy in the The synthesis of ATP from ADP ;~nd Pi, driYcn by the
terminal phosphoanhydride bond in ATP. energy released by transfer of electrons from NADH or
FADH2. to 0 2 , is the major source of ATP in aerobic nonpho-
P, 2- + H + ADP 3 ~ ATP4 + H 20, tosynthetic cells. Much evidence shows that in mitochondria
~G = +7.3 kcaUmol and bacteria this process of oxidatiue phosfJhorylation de-
pends on generation of a proton-motive force across the inner
A relatively simple one-to-one reaction involving reduction membrane (mitochondria) or bacterial plasma membrane,
of one coenzyme molecule and synthesis of one ATP would with electron transport, proton pumping, and ATP formation
be terribly inefficient, because the .lG0 ' for ATP generation occurring simultaneously. In the laboratory, for instance, ad-
from ADP and P, is substantially less than that for the coen- dition of 0 1 and an oxidizable substrate such as pyruvate or
zyme oxidation and much energy would be lost as heat. To succinate to isolated intact mitochondria results in a net syn-
efficiently recover the energy, the mitochondrion converts thesis of ATP if the inner mitochondrial membrane is intact.
the energy of coenzyme oxidation into a proton-motive force In the presence of minute amounts of detergents that make
using a series of electron carriers, all but one of which are the membrane leaky, electron transport and the oxidation of
integral components of the inner membrane (see Figure 12-8 ). these metabolites by 0 2 still occurs. However, under these
The proton-motive force can then be used to very efficiently conditions no ATP is made, because the proton leak prevents
generate ATP. the maintenance of the proton-motive force.
The coupling between electron transport from NADH
(or FADH2 ) to 0 2 and proton transport across the mner mi-
Electron Transport in Mitochondria
tochondrial membrane can be demonstrated experimentally
Is Coupled to Proton Pumping with isolated, intact mitochondria (Figure 12-13). As soon
During electron transport from NADH and FADH 2 to 0 2 , as 0 2 is added to a suspension of mitochondria in an other-
protons from the mitochondrial matrix are pumped across wise Orfree solution that contains NADH, the medium out-
the inner membrane. This pumping raises the pH of the mito- side the mitochondria transiently becomes more acidic
chondrial matrix relative to the intermembrane space and (i ncreased proton concentration), because the mitochondrial
cytosol and also makes the matrix more negative with respect outer membrane is freely permeable to protons. (Remember

>2 added

0 2 solution c
.g
60 t
~
c
~=
c 0
40
8E
"'
Io
·= .s
Q)
20
Ol
c
u"'
..c
0

Mitochondrion Elapsed time (s)

EXPERIMENTAL FIGURE 12-13 Electron transfer from NADH tons in the surrounding medium outside the mitochondria (decrease in
to 0 2 is coupled to proton transport across the mitochondrial pH). Thus the oxidation of NADH by 0 2 is coupled to the movement of
membrane. If NADH is added to a suspension of mitochondria de- protons out of the matrix. Once the 0 2 is depleted, the excess protons
pleted of 0 2, no NADH is oxidized. When a small amount of 0 2 is added slowly move back into the mitochondria (powering the synthesis of
to the system (arrow), there is a sharp rise in the concentration of pro- ATP) and the pH ofthe extracellular medium returns to its initial value.

12.3 The Electron Transport Chain and Generation of the Proton-Motive Force 533
 ·,

that malate-aspartate and other shuttles can convert the

NADH in the solution into NADH in the matrix.) Once the 0 2
TABLE 12-2 Electron-Carrying Prosthetic Groups
is depleted by its reduction, the excess protons in the medium in the Respiratory Chain
'

slowly leak back into the matrix. From analysis of the mea- Protein Component Prosthetic Groups*
sured pH change in such experiments, one can calculate that
about 10 protons are transported out of the matrix for every NADII-CoQ reductase FMN
electron pair transferred from NADH to 0 2 • (complex I) Fe-S
To obtain numbers for FADH 2 • the above experiment
can he repeated using succinate instead of NADH as the sub- Succinarc-CoQ reductase FAD
strate. (Recall that oxidation of succinate to fumarate in the (complex II ) Fe-S
citric acid cycle generates FADH 2 ; see Figure 12-10). The
amount of succinate added can be adjusted so that the CoQHr-cyrochrome c reductase Heme bL
amount of FADH2 generated is equivalent to the amount of (complex III) Heme bH
NADH in the first experiment. As in the first experiment, Fe-S
H eme c 1
addition of oxygen causes the medium o utside the mitochon-
dria to become acidic, but less so than with NADH. This is
Cytochrome c Hemec
not surprising because electrons in FADH1 have less poten-
tial energy (43.4 kcallmol) than electrons in NADH (52.6
Cytochrome c oxidase (complex IV) Cu.2
kcal/mol), and thus it drives the translocation of fewer pro- Heme a
tons from the matrix and a smaller change in pH. Cu~ 2
Heme a 3

Electrons Flow "Downhill" Through a Series •Not included is coenzyme Q, an electron carrier thar i; nor
permanently bound ro a prorein complex.
of Electron Carriers
~OURCE: J. W. De P1crre and l. Ernster, 1977, Ann. Rev. Biochem.
We now examine more closely the energetically favored 46:201.
movement of electrons from NADH and FADH 2 to the final
electron acceptor, 0 2• For simplicity, we will focus our dis-
cussion on NADH. In respiring mitochondria, each NADH
molecule rel eases two electrons to the electron transport Four large multiprotein complexes (complexes I-IV)
chain; these electrons ultimately reduce one oxygen atom compose an electron transport chain in the inner mitochon-
(half of an 0 2 molecule), forming one molecule of water: drial membrane that is responsible for the generation of the
proton-motive force (see Figure 12-8, stage III). Each com-
NADH--+ NAD + H .. +2e plex contains several prosthetic groups that participate in
the process of moving electrons from donor molecules to ac-
1 e + 2 H + lh 0 2 --+ H zO ceptor molecules in coupled oxidation-reduction reactions
(see Chapter 2). These small nonpeptide organic molecules
As electrons move from NADH to 0 1 , their electric potential or metal ions arc tightly and specifically associated with the
declines by 1.14 V, which corresponds to 26.2 kcallmol of multiprotein complexes (Table 12-2).
electrons transferred, or -53 kcallmol for a pair of elec-
trons. As noted earlier, much of this energy is conserved in He me and the Cytochromes Several types of heme, an iron-
the proton-motive force generated across the inner mito- containing pro~thetic group similar to that found in hemo-
chondrial membrane. globin and myog lobin (Figure 12-14a), are tightly bound

FIGURE 12 - 14 Heme and iron-sulfur pros- (a) (b)

t hetic g roups in the electron transport chain. H 2 C= CH
(a) Heme portion of cytochromes bLand bH, which
are components of CoQH 2-cytochrome c reductase
(complex Ill). The same porphyrin ring (yellow) is CH = CH 2
present in all hemes. The chemical substituents Protein
attached to the porphyrin ring differ in the other
cytochromes in the electron transport chain. All
hemes accept and release one electron at a time.
(b) Dime ric iron-sulfur cluster (Fe-5). Each Fe atom
is bonded to four 5 atoms: two are inorganic sulfur
and two are in cysteine side chains of the associ-
ated protein. All Fe-5 clusters accept and release
one electron at a time.

534 CHAPTER 12 • Cellular Energetics

 (covalently or noncovalently) to a set of mitochondrial proteins 0
called cytochromes. Each cytochrome is designated by a letter,
Ubiquinone (CoO}
H3CO CH 3 _ CH
I 3
such as a, b, c, or c 1• Electron flow through the cytochromes (oxidized form} H3CO !CH2 - CH C-CH 2 ) 10- H
occurs by oxidation and reduction of the Fe atom in the center
of the heme molecule: 0

.· Because the heme ring in cyrochromes consiM~ of alternating

o-
double- and single-bonded atoms, a large number of reso- H3CO¢ · CH3 CH3
Semiquinone (Coo· I 1
nance hybrid forms e>. ist. These allow the extra electron de- !free radical} H3CO ~ (CH 2 - CH = C- CH 2 ) 10- H
livered to the cytochrome to be delocalized throughout the
heme carbon and nitrogen atoms as well as the Fe ion. 0·
The various cytochromes have slightly different heme
groups and surrounding atoms (called axial ligands), which
generate different environments for the Fe ion. Therefore,
OH
each cytochrome has a different reduction potential, or ten-
dency to accept an electron-an important property dictat- H3C0 0 CH 3 CH3
Dihydroquinone
ing the unidirecti onal "downhill" electron flow a long the (CoOH2 } H3co V <cH 2 -CH = C- CH 2 l, 0-H
chain. Just as water spontaneously flows downhill from a (fully reduced form )
OH
higher to lower potential energy state-but not uphill-so
too do electrons flow in only one direction from one heme FIGURE 12-15 Oxidized and reduced forms of coenzyme Q (CoQ),
(or other prosthetic group) to another due to their differing which can carry two protons and two electrons. Because of its long
hydrocarbon "tail" of isoprene units, CoO, also called ubiquinone, is
reduction potentials. (For more o n the concept of reduction
soluble in the hydrophobic core of phospholipid bilayers and is very
potential,£, see Chapter 2.) All the cytochromes, except cy-
mobile. Reduction of CoO to the fully reduced form; OH2 (dihydro-
tochrome c, are components of integral membrane multipro-
quinone), occurs in two steps with a half-reduced free-radical interme-
tein complexes in the inner mitochondrial membrane.
diate, called semiquinone.

Iron-Sulfur Clusters Iron-sulfur clusters are non heme, iron-

containing prosthetic groups consisting of Fe atoms bonded Four Large Multiprotein Complexes Couple
both to inorganic S atoms and to S atoms on cysteine resi- Electron Transport to Proton Pumping Across
dues in a protein (Figure 12-14b). Some Fe atoms in the clus- the Mitochondrial Inner Membrane
ter bear a + 2 charge; others have a + 3 charge. However, the
As electrons flow downhill from one electron carrier to the
net charge of each Fe atom is actually between +2 and + 3,
next in the electron transport chain, the energy released is
because electrons in their outermost orbitals together with
used to power the pumping of protons against their electro-
the extra electr6n delivered via the transport chain are dis-
chemical gradient across the inner mitochondrial membrane.
persed among the Fe atoms and move rapidly from one atom
Fou.r large multiprotein complexes couple the movement of
to another. Iron-sulfur clusters accept and release electrons
electrons to proton pumping: NADH-CoQ reductase (com-
one at a time.
plex I, > 40 subunits), succinate-CoQ reductase (complex II,
4 subunits), CoQHrcytochrome c reductase (complex Ill,
Coenzyme Q (CoQ) Coenzyme Q (CoQ), also called ubiqui- 11 subunits), and cytochrome c oxidase (complex IV, 13
none, is the only small-molecule electron carrier in the chain subunits) (Figure 12-16). Electrons from NADH flow from
that is not an essentially irreversibly protein-bound prosthetic complex I via CoQ/CoQH 1 to complex III and then via the
group (Figure 12-15). It is a carrier of both protons and elec- soluble protein cytochrome c (cyt c) to complex IV to reduce
trons. The oxidized quinone form of CoQ can accept a single molecular oxygen (complex 11 is bypassed) (see Figure 12-l6a);
electron to form a semiquinone, a charged free radical denoted alternatively, electrons from FADH2 flow from complex II
by CoQ" . Addition of a second electron and t\vo protons (thus via CoQ/CoQH2 to complex lii and then via cytochrome c to
a total of two hydrogen atoms) to CoQ·- forms dihydroubiqui- complex IV to reduce molecular oxygen (complex I is bypassed)
none (CoQH 2 ), the fully reduced form. Both CoQ and CoQH2 (see Figure 12-16b).
are soluble in phospholipids and diffuse freely in the hydropho- As shown in Figure 12-16, CoQ accepts electron~ released
bic center of the inner mitochondrial membrane. This is how it from NADH-CoQ reductase (complex 1) or succinate-CoQ re-
participates in the electron transport chain-carrying electrons ductase (complex 11) and donates them to CoQHr-cytochrome
and protons between the protein complexes of the chain. c reductase (complex III). Protons are simultaneously trans-
We now consider in detail the multiprotein complexes ported from the matrix (also called the cytosolic) side of the
which use these prosthetic groups, and the paths taken by the membrane to the intermembrane space (also called the exo-
electrons and protons as they pass through the complexes. plasmic side). Whenever CoQ accepts electrons, it does so at a

12.3 The Electron Transport Chain and Generation of the Proton-Motive Force 535
(;} FOCUS ANIMATION: Electron Transport
·.
(a) From NADH (b) From succinate

lntermembrane space
(exoplasmic) 2H
4H

+++I~Fe-SlJ r::cCoQ ~oQrl ,. t bl

Q .~ C tbH
- FMN C ° ''z
4H R'l'
Matrix
Cytosolic
2H
(cytosolic) W
2
Succinate Fumar,ate + 2 H•
NADH NAD'+ H
CoOH 2-cytochrome c Cytochrome c oxidase Complex Ill
NADH-CoO reductase Succinate-CoO reductase
reductase {complex Ill) {complex IV)
{complex I) {complex Ill

FIGURE 12·16 The electron transport chain. Electrons (blue A total of 10 protons are translocated per pair of electrons that flow
arrows) flow through four major multiprotein complexes (I-IV). Electron from NADH to 0 2• The protons released into the matrix space during
movement between complexes is mediated either by the lipid-soluble oxidation of NADH by complex I are consumed in t he formation of
molecule coenzyme Q (CoQ. oxidized form; CoQH2, reduced form) or water from 0 2 by complex IV, resulting in no net proton translocation
the water-soluble protein cytochrome c (Cyt c). The multiple multi- from these reaction s. (b) Pathway from succinate. Electrons flow from
protein complexes use the energy released from passing electrons succinate to complex II via FAD/FADH2 and iron-sulfur clusters (Fe-5).
to pump protons from the matrix to the intermembrane space (red from complex II to complex Ill via CoQ/CoQHb and then to complex IV
arrows). (a) Pathway from NADH. Electrons from NADH flow through via Cyt c. Electrons released during oxidation of succinate to fumarate
complex I, in itially via a flavin mononucleotide (FMN) and then via in complex II are used to reduce CoQ to CoQH 2 without translocating
seven iron-sulfur clusters (Fe-5), to CoQ, to w hich two protons bind, additional protons. The remainder of electron transport from CoQH2
forming CoQH 2• Conformational changes in complex I that accom- proceeds by the same pathway as for the NADH pathway in (a). Thus
pany the electron flow drive proton pumping from the matrix to the for every pair of electrons transported from succinate to 0 2, six protons
intramembrane space (red arrows). Electrons then flow via the released are translocated by complexes Ill and IV.
(and recycled) CoQH2 to complex Ill, and then via Cyt c to complex IV.

binding Site on the matrix side of a protein complex, always NAD is exclusively a two-electron ca r rier: it accepts or
picking up protons from the med ium there. Whenever CoQH2 releases a pair of electrons simultaneously. In NADH-CoQ
releases its electrons, it does so at a site on the intermembrane reductase (complex 1), the NADH-binding site is at the tip of
space side of a protein complex, releasing protons into the the peripheral arm (see Figure 12-17a); electrons released
fluid of the intermembrane space. T hus transport of each pair from NADH first flow to FMN (flavin monon ucleotide ), a
of electrons by CoQ is obligatorily coupled to movement of cofactor rela ted to FAD, then are sh uttled - 95 A down that
two protons from the matrix to the intermembrane space. arm through seven iron-sulfur clusters and finally to CoQ,
which is bound at a site at least partiall y in the plane of the
NADH-CoQ Reductase (Complex I) Electrons are transferred membrane. FMN, like FAD, can accept two electro ns but
from NADH to CoQ by NADH-CoQ reductase (see Figure does so one electron at a time.
12-16a). Electron microscopy and x-ray crysta llography of Each transported electron undergoes a drop in potential of
complex I from both bacteria (mass - 500 kDa, with 14 sub- - 360 mY, equivalent to a .lG 0 ' of -16.6 kcaVmol for the two
units) and eukaryotes (- 1 MD a, with 14 high ly conserved electrons transported. Much of this released energy is used to
core subunits shared with bacteria plus about 26-32 acces- transport four pro tons across the inner membrane per mole-
sory subunits) established that it is L-shaped (Figure cule of NADH oxidized by complex I. Those fo ur protons are
12 17a). The membrane-embedded arm of Lhe L is slighrly distinct from the two protons that are transferred to the CoQ
curved, - 180 A long, and comprises proteins with more as illustrated in Figures 12-15, 12-16a, and l2-17a. The struc-
than 60 transmembrane alpha helices. Thi s arm has four ture of complex I suggests that t he energy released by the elec-
subdomains, three of which have proteins that are members tron transport in the peripheral arm is used to change the
of a family of cation antiporrers. The hydrophilic peripheral conformation of subunits in the membrane arm and thus medi-
arm extends over 130 A away from the membrane in to the ate movement of four protons across the membrane. Three
cytosolic space. protons are likely to pas!> through the three cation antiporter

536 CHAPTER 12 • Cellu lar Energetics

 (a) Complex I (b) Complex II

lntermembrane space
(exoplasmic)
H H· W
++-r

2H
Matrix '"-
H'--------'H
...:...yr----'-'----.:...-
(cytosolic) H
4

CoO=l FMN or= ~

FAD

Fe-S =. NADH
Succinate Fumarate + 2H •

Succinate-CoO reductase
NADH-CoQ reductase
(complex I) (complex II)

FIGURE 12-17 Electron and proton transport through pumping through the transmembrane subunits from the matrix to the
complexes I and II. (a) Model of complex I based on its three- intramembrane space (red arrows). (b) Model of complex II based on its
dimensional structure. The outline of the shape of complex I, as three-dimensional structure. Electrons flow from succinate to complex
determined by x-ray crystallography, is shown in light blue, and distinct II via FAD/FADH 2 and iron-sulfur clusters (Fe-5, blue diamonds), and
structural subunits are indicated by thin dashed black lines. From from complex II to complex Ill via CoQ/CoQH 2• Electrons released dur-
NADH electrons flow first to a flavin mononucleotide (FMN) and then, ing oxidation of succinate to fumarate in complex ll_are used to reduce
via seven of the nine iron-sulfur clusters (Fe-5, blue diamonds), to CoQ, CoQ to CoQH 2 without translocating additional protons. (Structure of
to which two protons from the matrix bind (red arrow) to form CoQH 2• complex I adapted from C. Hunte, V. Zickermann, and U. Brandt, 2010, Science
Conformational changes due to the electron flow, which probably 329:448- 451 and R. G. Efremov, R. Baradaran, and L A. Sazanov, 2010, Nature
include a pistonlike horizontal movement of the t-helix, drive proton 465:441-445.]

domains while the route of the forth is through a different type The pathway is somewhat reminiscent of that in complex I
of domain. An - 110 A long, kinked, transverse alpha helix (Figure 12-17a).
(t-helix) in the membrane arm runs parallel to the plane of The overall reaction catalyzed by this complex is
the membrane, potentially mechanically linking the anti-
porter domains to the peripheral arm (Figure 12-17a) and Succinate + CoQ ~ fumarate + CoQH 1
thereby transmirting electron-transport-induced conforma-
(Reduced ) (Oxidized) (Oxidized) (Reduced)
tional changes in the peripheral arm to the distant anti porter
domains to drive proton transport.
The overall reaction catalyzed by this complex is Although the ~G ' for this reaction is negative, the released
0

energy is insufficient for proton pumping in addition to re-

NADH + CoQ + 6 H+'" ~ duction of CoQ to form CoQH1 . Thus no protons are trans-
(Reduced) (Oxid1zed )
located directly across the membrane by the succinate-CoQ
reductase complex, and no proton-motive force is generated
in this part of the respiratory chain. Shortly we will see how
(Ox1dized) (Reduced) the protons and electrons in the CoQH 2 molecules generated
by complexes I and II contribute to the generation of the
Succinate-CoQ Reduct ase (Complex II} Succinate dehydroge- proton-motive force.
nase, the enzyme that oxidizes a molecule of succinate to Complex H generates CoQH 2 from succinate via FAD/
fumarate in the citric acid cycle (and in the process generates FADH2-mediated redox reactions. Another set of proteins in
the reduced coenzyme FADH 2), is one of the four subunits of the matrix and inner mitochondrial membrane performs a
complex II. Thus the citric acid cycle is physically as well as comparable set of FAD/FADHr mediated redox reactions to
functionally linked to the electron transport chain. The two generate CoQH2 from fatty acyl CoA. Fatty acyl-CoA dehy-
electrons released in conversion of succinate tO fumarate are drogenase, which is a water-soluble enzyme, catalyzes the
transferred first to FAD in succinate dehydrogenase, then to first step of the oxidation of fatty acyl CoA in the mitochon-
iron-sulfur clusters-regenerating FAD-and finally to CoQ, drial matrix (see Figure 12-12). There are several fatty acyi-
which binds to a cleft on the matrix side of the transmem- CoA dehydrogenase enzymes with specificities for fatty acyl
brane portions of complex II (Figures l2-16b and 12-17b). chains of different lengths. These enzymes mediate the initial

12.3 The Electron Transport Chain and Generation of the Proton-Motive Force 537
step in a four-step process that removes two carbons from the two-for-one transport of protons and electrons by com-
the fatty acyl group by oxidizing the carbon in the 13 position plex III (Figure 12-18).
of the fatty acyl chain (thus the entire process is often re- The substrate for complex Ill, CoQH2, is generated by sev-
ferred to as 13-oxidation). These reactions generate acetyl eral enzymes, including NADH-CoQ reductase (complex I),
CoA, which in turn enters the citric acid cycle. They also succinate-CoQ reductase (complex II ), electron transfer
generate an FADH 2 intermediate and NADH. The FADH2
generated remains bound to the enzyme dur ing the redox
reaction, as is the case for complex II. A water-soluble pro-
tein called electron transfer flavoprotein (ETF) transfers the
high-energy electrons from the FADH 2 in the acyi-CoA de-
hydrogenase to electron transfer flavoprotein:ubiquinone lntermembrane
oxidoreductase (ETF:QO), a membrane protein that reduces sp ace
CoQ to CoQH2 in the inner membrane. This CoQH.! inter-
mixes in the membrane with the other CoQH 2 molecules
generated by complexes I and II.

CoQH 2- Cytochro me c Reductase (Complex Ill ) A CoQH 2

generated either by complex I, complex ll, or ETF:QO donates
two electrons to CoQHr-cytochrome c reductase (complex M atrix
lll), regenerating oxidized CoQ. Concomitantly it releases into
the intermembrane space two protons previously picked up by
CoQ on the matrix face, generating part of the proton-motive
force (see Figure 12-16). Within complex lll, the released elec- CoQH 2- cytochrome c
trons first are transferred to an iron-sulfur cluster within the reductase (complex Ill)
complex and then to cytochrome c 1 or to two b-type cyto-
chromes (bLand bH, see Q cycle below). Finally, the two elec- At 0 0 sit e: 2 CoOH 2 + 2 Cyt 2•---+
trons are transferred sequentially to two molecules of the (4 H • ) 2 CoO+ 2 Cyt CZ• + ! e· + 4 H !outside!
(2 e )
oxidized form of cytochrome c, a water-soluble peripheral pro-
tein that diffuses in the intermembrane space. For each pair of At Oi sit e: CoO+ e + 2 H•!matrix side!---+ CoOH2
electrons transferred, the overall reaction catalyzed by the (2 H", L

CoQHrcytochrome c reductase complex is Net 0 cycle (sum of react ions at 0 0 and O i):
CoOH2 + 2 Cyt 2+ + 2 H (matrix s1del---+
CoQHz + 2 Cyt c3 + + 2 H\,-+ CoQ + 4 H +our+ 2 Cyt c2 + (2 H , 2 e ) CoO+ 2 Cyt c-
-'• + 4 H (outside)

(Ox1dized) (Reduced)
( )

Per e transferred through complex Ill to cytochrome c, 4 W

The ~G ' for this reaction is sufficiently negative that two pro- released to the intermembrane space .·
tons in addition to those from CoQH 2 are translocated from FIGURE 12-18 The Q cycle. The a cycle of complex Ill uses the net
the mitochondrial matrix across the inner membrane for each oxidation of one CoaH 2 molecule to transfer four protons into the
pair of electrons transferred; this involves the proton-motive Q intermembrane space and two electrons to two cytochrome c molecules.
cycle, discussed below. The heme protein cytochrome c and The cycle begins when a molecule from the combined pool of reduced
the small lipid-soluble molecule CoQ play similar roles in the CoaH 2 in the membrane binds to the ao site on the intermembrane space
electron transport chain in that they both serve as mobile elec- (outer) side of the transmembrane portion of complex Ill (step 0 ). There,
tron shuttles, transferring electrons (and thus energy) between CoaH 2 releases two protons into the intermembrane space (step fll),
the complexes of the electron transport chain. and two electrons and the resulting Coa dissociate (step D l. One of the
electrons is transported, via an iron-sulfur protein and cytochrome c~<
The Q Cycle Experiments have shown that four protons arc directly to cytochrome c (step fm ). (Recall that each cytochrome c
translocated across the membrane per electron pair trans- shuttles one electron from complex Ill to complex IV.) The other electron
moves through cytochromes bLand bH and partially reduces an oxidized
ported from CoQH 2 through CoQH 2-cytochrome c reduc -
Coa molecule bound to the second, a. site on the matrix (inner) side of
tase (complex III). These four protons arc those carried on
the complex, forming a Coa semiquinone anion, a ·- (step !1). The process
two CoQH 2 molecules, which are converted to two CoQ
is repeated with the binding of a second coaH 2 at the Go site (step D l.
molecules during the cycle. However, another CoQ molecult> proton release (step m3), reduction of another cytochrome c (step rim),
receives two other protons from the matrix space and is con- and addition of the other electron to the a· bound at the a , site (step fJ).
verted to one CoQH2 molecule. Thus the net overall reaction There, the addition of two protons from the matrix yields a fully reduced
involves the conversion of on ly one CoQH 2 molecule to CoaH2 molecule at the a , site, which then dissociates (steps 13 and I!]),
CoQ as two electrons are transferred one at a time to two freeing the a; to bind a new molecule of Coa (step llil) and begin the
molecules of the acceptor cytochrome c. An evolutionarily a cycle over again. [Adapted from B. Trumpower, 1990,1. Bioi. Chern. 265:11409,
conserved mechanism, called the Q cycle, is responsible for and E. Darrouzet et al., 2001, Trends Biochem. Sci. 26:445.)

538 CHAPTER 12 • Ce llular Energetics

flavoprotein:uhiquinone oxidoreductase (ETF:QO, during simplicity, Figure 12-16 shows only two electrons moving
13-oxidation), and, as we shall see, by complex III itself. and 112 0 2 being reduced.) Proposed intermediates 111 oxygen
As shown in Figure 12-18, in one turn of the Q cycle, reduction include the peroxide anion (0 2 2 ) and probably the
two molecules of CoQH 2 are oxidized to CoQ at the Q 0 site hydroxyl radical (OH·), as well as unusual complexes of iron
and release a total of four protons into the intermembrane and oxygen atoms. These intermediates would be harmful to
space, but at the Q, site one molecule of CoQH 2 is regener- the cell if they escaped from complex IV, but they do so only
ated from CoQ and two additional proteins from the matrix rarely (see the discussion of reactive oxygen species below).
space. The translocated protons are all derived from CoQH 2• During transport of four electrons through the cytochrome c
which obtained its protons from the matrix as described oxidase complex, four protons from the matrix space are
above. Although seemingly cumbersome, the Q cycle opti- translocated across the membrane. Thus, complex IV trans-
mizes the number of protons pumped per pair of electrons ports only one proton per electron transferred, whereas com-
moving through complex III. The Q cycle is found in all plex II, using the Q cycle, transports two protons per electron
plants and animals, as well as in bacteria. Its formation at a transferred. However, the mechanism by which complex IV
very early stage of cellular evolution was likely essential for translocates these protons is not known.
the success of all life-forms, as a way of converting the po- For each four electrons transferred, the overall reaction
tential energy in reduced coenzyme Q into the maximum catalyzed by cytochrome c oxidase is
proton-motive force across a membrane. In turn this maxi-
mizes the number of ATP molecules synthesized from each
electron that moves down the electron transport chain from (Reduced) (Oxidized)
NADH or FADH2 to oxygen.
How are the two electrons released from CoQH2 at the The poison cyanide, which has been used as a chemical
Q 0 site directed to different acceptors, either to Fe-S, cyto- warfare agent, by spies to commit suicide when cap-
chrome ch and then cytochrome c (upward pathway in Fig- tured, in gas chambers to execute prisoners, and by the Nazis
ure 12-18), or alternatively to cytochrome b 1 , cytochrome (Zyklon B gas) for the mass murder of Jews and others, IS
b 1" and then CoQ at the Q, site (downward pathway in Fig-
ure 12-18)? The mechanism involves a flexible hinge in the
toxic because it binds to the heme a,
in mitochondrial cyto-
chrome c oxidase (complex IV), inhibiting electron transport
Fe-S-containing protein subunit of complex III. Initially the and thus oxidative phosphorylation and production of ATP.
Fe-S cluster is close enough to the Q" site to pick up an elec- Cyanide is one of many toxic small molecules that interfere
tron from CoQH 2 bound there. Once this happens, a seg- with energy production in mitochondria. •
ment of the protein containing this Fe-S cluster swings the
cluster away from the Q o site to a position near enough to
the heme on cytochrome c 1 for electron transfer to occur. Reduction Potentials of Electron Carriers
With the Fe-S subunit in this alternate conformation, the sec- in the Electron Transport Chain Favor
ond electron released from CoQH2 bound to the Q 0 site can-
Electron Flow from NADH to 0 2
not move to the Fe-S cluster-it is too far away, so it takes
an alternative path open to it via a somewhat less thermody- As we saw in Chapter 2, the reduction potential E for a par-
namically favored route to cytochrome bLand through cyto- tial reduction reaction
chrome bH to the CoQ at the Q , site.
Oxidized molecule + e ~ reduced molecule
Cytochrome c Oxidase (Complex IV) Cytochrome c, after
being reduced by one electron from CoQH2-cytochrome c is a measure of the equilibrium constant of that partial reac-
reductase (complex III ), is reoxidized as it transports its elec- tion. With the exception of the b cyrochromes in the CoQH 2-
tron to cytochrome c oxidase (complex IV) (see Figure 12-16). cytochrome c reductase complex, the standard reduction
Mitochondrial cytochrome c oxidases contain 13 different potential e· of the electron carriers in the mitochondrial
subunits, but the catalytic core of the enzyme consists of respiratory chain increases steadily from NADH to 0 2 • For
only three subunits. The functions of the remaining subunits instance, for the partial reaction
are not well understood. Bacterial cytochrome c oxidases
contain only the three catalytic subunits. In both mitochon- NAD + H + 2e ~ NADH
dria and bacteria, four molecules of reduced cytochrome c
bind, one at a time, to the oxidase. An electron is transferred the value of the standard reduction potential is -320 mY,
from the heme of each cytochrome c, first to the pair of cop- which is equivalent to a .1G0 ' of + 14.8 kcallmol for transfer of
per 1ons called Cu/ , then to the heme in cytochrome a, and two electrons. Thus this partial reaction tends to proceed to-
next to the Cubl+ and the heme in cytochrome a, that to- ward the left, that is, toward the oxidation of NADH to NAD .
gether make up the oxygen reduction center. The four elec- By contrast, the standard reduction potential for the
trons are finall y passed to 0 2, the ultimate electron acceptor, partial reaction
yielding four H 20, which together with C02 is one of the end
products of the overall oxidation pathway. (Note that, for

12.3 The Electron Transport Chain and Generation of the Proton-Motive Force 539
FIGURE 12-19 Changes in redox potential and Redox potential Free energy
free energy during stepwise flow of electrons lmVI (kcal/mol)
through the respiratory chain. Blue arrows indicate 60
electron flow; red arrows, translocation of protons NADH-CoO reductase
- 400 (complex II
across the inner mitochondrial membrane. Electrons
pass through the multi protein complexes from those NADH NAD +W
at a lower reduction potential to those with a higher
(more positive) reduction potential (left scale), with a
~ Fumarate + 2 H'
2e
50
corresponding reduction in free energy (right scale).
~
The energy released as electrons flow through three of - 200 FMN
the complexes is sufficient to power the pumping of
H ions across the membrane, establishing a proton-
motive force.
w
")
!
Fe·S
Succinate-CoO
reductase (complex II)
H'oul
40
0-
CoO ~

H';n ) Fe-S
H +out
30
200 CoOH 2 -cytochrome c
reductase (complex Ill)

20
400

600 10

Cytochrome c oxidase
(complex lVI 2e

800 0

0
is +220 mY (.lG = 5.1 kcal/mol) for transfer of one elec-
' The Multi protein Complexes of the Electron
tron. Thus this partial reaction tends to proceed toward the Transport Chain Assemble into Supercomplexes
right, that is, toward the reduction of cytochrome c (Fe3 +) to
cytochrome c (Fe 2- ). Over 50 years ago Britton Chance proposed that electron
The fina l reaction in the respiratory chain, the reduction transport complexes might assemble in to large supercom-
of 0 2 to H20 plexes. Doing so wou ld bring the complexes into close and
highly organized proximity, w hich might improve the speed
and efficiency of the overall process. Indeed, genetic, bio-
chemical, and biophysical studies have provided very strong
has a standard reduction potential of + 816 mY (.l G 0 ' evidence for the existence of electron transport chain super-
- 37.8 kcal/mol for transfer of two electrons), the most pos- complexes. These studies involved gel electrophoretic meth-
itive in the whole series; thus this reaction also tends to pro- ods ca ll ed blue native (BN)-PAGE and color less native
ceed towa rd the right. (CN)-PAGE, wh ich permit separation of very large macro-
As illustrated in F1gure 12-19, the steady increase in 1:. ·• molecul ar protein complexes, and electron microscopic anal-
values, and rh e corresponding decrease in ~C ' values, of the
0
ysis of t he ir three-dimemiunal st ru ctures. O ne such
carriers in the electron transport chain favors the flow of elec- supercomplex contains one copy of complex I, a dimer of
trons from NADH and FADH 2 (generated from succinate) to complex III (III 2), and one or more copies of complex IV
oxygen. The energy released as electrons flow "downhill " en- (Figure 12-20). One supercomplex that contains all of the com-
ergeticall y through the electron transport chain complexes ponents tho ught to play a role in respiration--complexes I-IV,
drives the pumping of protons aga inst their concentration ubiquinone (CoQ), and cytochrome c-was isolated from
gradient across the mitochondrial inner membrane. BN-PAGE gels and shown to transfer electrons from NADH

540 CHAPTER 12 • Cellular Energetics

 (a) (b)
Complex Ill dimer Complex IV

f Supercomplex 1/111 2 /IV

=- Supercomplex 1/111 2
-Complex I
- - ATP synthase
-Complex Il l dimer (111 2 )

-Complex IV

-Complex II

EXPERIMENTAL FIGURE 12-20 Electrophoresis and electron of complex or supercomplex present. (b) Supercomplex 1/ 111/ IV was
microscopic imaging identifies an electron transport chain extracted from the gel, and the particles were negatively stained with
supercomplex containing complexes I, Ill, and IV. (a) Membrane 1% uranyl acetate and visualized by transmission electron microscopy.
proteins in isolated bovine heart mitochondria were solubilized with Images of 228 particles were combined at a resolution of - 3.4 nm to
a detergent, and the complexes and supercomplexes were separated generate an averaged image of the complex viewed from the side in
by gel electrophoresis using the blue native (BN)-PAGE method. Each the plane of the membrane. Approximate locations of the complex
blue-stained band within the gel represents the indicated protein Ill dimer and complex IV are indicated by dashed ovals; the outline of
complex or supercomplex, with 111 2 representing a dimer of complex Ill. complex I is also indicated by a dashed line (white). Scale bar is 10 nm.
Intensity of the blue stain is approximately proportional to the amount [Adapted from E. Schafer et al., 2006, J. Bioi. Chern. 28 1 (22):15370- 15375.]

to 0 2 ; in other words, this supercomplex can respire-it is Reactive Oxygen Species (ROS) Are Toxic
a resp1rasome. By-products of Electron Transport
The unique phospholipid cardiolipin (diphosphatidyl
glycerol) appears to play an important role in the assembly
That Can Damage Cells
and function of these supercomplexes. Generally not observed B About 1-2 percent of the oxygen metabolized by acro-
H hie organisms, rather than being converted to water, is
partially reduced to the superoxide anion radical (O.i ,
Cardiolipin
where the "dot" represents an unpaired electron). Radicals
0
are atoms that have one or more unpaired electrons in an
outer (valence) shell, or molecules that contain such an atom.
Many, though not all, radicals are generally highly chemi-
cally reactive, altering the structures and properties of those
molecules with wh1ch the) react. The products of such reac-
tions often arc themselves radicals and thus can propagate a
chain reaction that alters many additional molecules. Super-
oxide and other highly reactive oxygen-containing mole-
cules, both radicals (e.g., Oi.-) and non-radicals (hydrogen
peroxide, H 20 2 ) arc called reactive oxygen species (ROS ).
0 ROS arc of great interest because they can react v.ith and
thus damage many key biological molecules, including lipids
(particularly unsaturated fatty acids and their derivatives ),
in other membranes of eukaryotic cells, cardiolipin has been protein5, and DNA, and thus severely interfere with their
observed to bind to integral membrane proteins of the inner normal functions. At moderate to high levels, ROS contrib-
membrane (e.g., complex II). Genetic and biochemical studies ute to what is often called cellular oxidative stress and can be
in yeast mutants in which cardiolipin synthesis is blocked highly toxic. Indeed, ROS are purposefully generated by
have established that cardiolipin contributes to the formation hody defense cells (e.g., macrophages, neutrophils) to kill
and activity of mitochondrial supcrcomplexes, and thus it has pathogens. In humans, excessive or inappropriate generation
been called the glue that holds together the electron transport of ROS has been implicated in many diverse diseases, includ-
chain, though the precise mechanism remains to be defined. ing heart failure, neurodegcnerative diseases, alcohol-induced
In addition, there is evidence that cardiolipin may influence liver disease, diabetes, and aging.
the inner membrane's binding and permeability to protons Although there are several mechanisms for generating
and consequently the proton-motive force. ROS in cells, the major source in eukaryotic cells is electron

12.3 The Electron Transport Chain and Generation of the Proton-Motive Force 541
transport in the mitochondria (or in chloroplasts as de- hydroxyl radical (OH"). Thus cells depend on the inactiva-
scribed below). Electrons passing through the mitochondrial tion of H 2 0 2 by catalase and other enzymes, such as perox-
electron transport chain can have sufficient energy to reduce iredoxin and glutathione peroxidase, which also detoxify the
molecular oxygen (0 2 ) to form superoxide anions (Figure lipid hydroperoxide products formed when ROS react with
12-21, top). This can only occur, however, when molecular unsaturated fatty acyl groups. Small molecule antioxidant
oxygen comes in close contact with the reduced electron car- radical scavengers such as vitamin E and a-lipoic acid also
riers (iron, FMN, CoQH 2 ) in the chain. Usually such contact protect against oxidative stress. Altho ugh in many cells cata-
is prevented by sequestration of the carriers within the pro- lase is located only in peroxisomes, in heart muscle cells it is
teins involved. However, there are some sites (particularly m found in mitochondria. This is not surprising, because the
complex I and CoQ·- , see Figure 12-15) and some conditions heart is the most oxygen-consuming organ per gram weight
(e.g., high NADH/NAD+ ratio in the matrix, high proton- in mammals.
motive force when ATP is not generated) when electrons can As the rate of ROS production by mitochondria and
more readily "leak" out of the chain and reduce 0 2 too;- . chloroplasts reflects the metabolic state of the::.e organelles
Superoxidc anion is an especially unstable and reactive (e.g., size of proton-motive force, NADH/NAD ratio), cells
ROS. Mitochondria have evolved several defense mecha- have developed ROS-sensing systems, such as ROS/redox-
nisms that help protect against Oi toxicity, including the sensitive transcription factors, to monitor the metabolic state
use of enzymes that inactivate superoxide, first by converting of these organelles and respond accordingly, for example by
it to H 20 2 (Mn-containing superoxide dismutase, SOD) and changing the rate of transcription of nuclear genes that en-
then to H 2 0 (catalase) (Figure 12-21 ). Because Oi is so code organelle-specific proteins. •
highly reactive and toxic, SOD and catalase are some of the
fastest enzymes known. SOD is found within mitochondria
and other cellular compartments. Hydrogen peroxide itself is Experiments Using Purified Electron Transport
a ROS that can diffuse readily across membranes and react Chain Complexes Established the Sloichiometry
with molecules throughout the cell. It can also be converted of Proton Pumping
by certain metals such as Fe2 + into the even more dangerous
The multiprotein complexes of the electron transport chain
that are responsible for proton pumping have been identified
by selectively extracting mitochondrial membranes with de-
tergents, isolating each of the complexes in nearly pure form,
Complex I }
CoOi-
etc.
e i and then preparing artificial phospholipid vesicles (lipo-
somes) containing each complex. When an appropriate elec-
tron donor and electron acceptor are added to such
liposomes, a change in pH of the medium will occur if the
embedded complex transports protons (Figure 12-22). Stud-
ies of this type indicate that NADH-CoQ reductase (com- ·.
plex I) translocates four protons per pair of electrons
transported, whereas cytochrome c oxidase (complex IV)
translocates two protons per electron pair transported.
Current evidence suggests that a total of 10 protons are
transported from the matrix space across the inner mito-
chondrial membrane for every electron pair that is trans-
ferred from NADH ro 0 2 (see Figure 12-16). Because
succinate-CoQ reductase (complex II) does not transport
protons and complex I is bypassed when the electrons come
@ : Superoxide dismutase from succinate-derived FADHl> only six protons arc trans-
FIGURE 12-2 1 Generation and inactivation of toxic reactive ported across the membrane for every electron pair that is
oxygen species. Electrons from the electron transport chains of transferred from this FADH2 to 0 2 •
mitochondria and chloroplasts as well as some generated through
other enzymatic reactions reduce molecular oxygen (02 ), forming the
highly reactive radical anion superoxide (02 ). Superoxide is rapidly The Proton-Motive Force in Mitochondria
converted by superoxide dismutase (SOD) to hydrogen peroxide Is Due Largely to a Voltage Gradient
(H 20 2), which in turn can be converted by metal ions such as Fel to Across the Inner Membrane
hydroxyl radicals (OH") or inactivated to water by enzymes such as
catalase. Because of their high chemical reactivityo;-, H20 2, OH", and The main result of the electron transport chain is the genera-
similar molecules are called reactive oxygen species (ROS). They cause tion of the proton-motive force (pmf), which is the sum of a
oxidative and free radical damage to many biomolecules, including transmembrane proton concentration (pH) gradient and
lipids, proteins, and DNA. This damage leads to cellular oxidative stress electric potential, or voltage gradient, across the mitochon-
that can cause disease and, if sufficiently severe, can kill cells. drial inner membrane. The relative contribution of the two

542 CHAPTER 12 • Cellular Energetics

 (a) to the matrix (exchange of like charge) would a lso short-
Cytochrome c circuit voltage gradient formation. Indeed, the inner mito-
oxidase complex chondrial membrane is poorly permeable to other ions. Thus
Phospho! i pid proton pumping generates a voltage gradient that makes it
~membrane
energetically difficult for additional protons to move across
2 because of charge repulsion. As a consequence, proton
pumping by the electron transport chain establishes a robust
voltage gradient in the context of a rather small pH gradient.
Because mitochondria arc much too ~mall to be impaled
with electrodes, the electric potential and pH gradient across
the inner mitochondrial membrane cannot be directly mea-
sured. However, the electric potential can be measured indi-
rectly by adding radioactive 42K + ions and a trace amount of
valinomycin to a suspension of respiring mitochondria and
measuring the amount of radioactivity that accumulates in
Valinomycin-bound K+ the matrix. Although the inner membrane is normally imper-
meable to K +, valinomycin is an ionophore, a small lipid-
(b)
soluble molecule that selectively binds a specific ion (in this
case, K .. ) and carries it across otherwise impermeable mem-
branes. In the presence of valinomycin, 42 K + equilibrates
E across the inner membrane of isolated mitochondria in ac-
:J
:.0
Q)
cordance with the electric potential: the more negative the
E matrix side of the membrane, the more 42 K + will be attracted
0 to and accumulate in the matrix.
J:
c. At equilibrium, the measured concentration of radioactive
K+ ions in the matrix, [Kinl, is about 500 times greater than that
in the surrounding medium, [Kourl· Substitution of this value
·· ' into the Nernst equation (Chapter 11 ) shows that the electric
Elapsed time (min) potential E (in mV) across the inner membrane in respiring
EXPERIMENTAL FIGURE 12·22 Electron transfer from mitochondria is - 160 mV, with the matrix (inside) negative:
reduced cytochrome c to 0 2 via cytochrome c oxidase (complex
IV) is coupled to proton transport. The oxidase complex is incorpo-
Km
E = -59 log [ - = - 59log500 = -160 mV
rated into liposomes with the binding site for cytochrome c positioned Kout
on the outer surface. (a) When 0 2 and reduced cytochrome care
added, electrons are transferred to 0 2 to form H20 and protons are
Researchers can measure the matrix (instde) pH by trap-
transported from the inside to the medium outside of the vesicles. ping pH-sensitive fluorescent dyes inside vesicles formed
A drug called valinomycin was added to the medium to dissipate the from the inner mitochondrial membrane, with the matrix
voltage gradient generated by the translocation of H , which would side of the membrane facing inward. They also can measure
otherwise reduce the number of protons moved across the membrane. the pH outside of the vesicles (equivalent to the intermem-
(b) Monitoring of the medium's pH reveals a sharp drop in pH brane space) and thus determine the pH gradient (.lpH),
following addition of 0 2• As the reduced cytochrome c becomes fully which turns out to be - 1 pH unit. Because a difference of
oxidized, protons leak back into the vesicles, and the pH of the medium one pH unit represents a tenfold difference in H concentra-
returns to it s initial value. Measurements show that two protons are tion, according to the Nernst equation a pH gradient of one
transported per 0 atom reduced. Two electrons are needed to reduce unit across a membrane is equivalent to an electric potential
one 0 atom, but cytochrome c transfers only one electron; thus two of 59 mV at 20 oc. Thus, knowing the voltage and pH gra-
molecules of Cyt c 2 ~ are oxidized for each 0 reduced. [Adapted from
dients, we can calculate the proton-motive force, pmf, as
B. Reynafarje et al., 1986, J. Bioi. Chern. 261:8254.]

pmf = 'JI - (RFT X .lpH) = 'I' - 59 .lpH

components to the total pmf has been determined experi- where R is the gas constant of 1.987 caU(degree · mol ), Tis
mentally, and depends on the permeability of the membrane the temperature (in degrees Kelvin), F is the Faraday constant
to ions other than H +. A significant voltage gradient can 123,062 caU(V · mol)], and 'I' is the transmembrane electric
develop only if the membrane is poorly permeable to other potential; 'I' and pmf are measured in millivolts. The electric
cations and to anions. Otherwise, anions would leak across potential 'I' across the inner membrane is -160 mV (negative
from the matrix to the intermembrane space along with the inside matrix) and .lpH is equivalent to - 60 mV. Thus the
protons and prevent a voltage gradient from forming. Simi- total pmf is - 220mV, with the transmembrane electric poten-
larly, cations leaking across from the intermembrane space tial responsible for about 73 percent of the total.

12.3 The Electron Tran sport Chain and Generation of the Proton -Motive Force 543
KEY CONCEPTS of Section 1 2.3 • Reactive oxygen species (ROS) are toxic by-products of the
electron transport chain that can modify and damage proteins,
The Electron Transport Chain and Generation DNA, and lipids. Specific enzymes (e.g., glurathinone permci-
of the Proton-Motive Force dase, catalase) and small molecule antioxidants (e.g., vitamin E)
By the end of the citric acid cycle (stage II), much of the help protect against ROS-induced damage (see Figure 12-21 ).
energy origmally present in the covalent bonds of glucose ROS can also be used as intracellular signaling molecules.
and fatty acids is converted into high-energy electrons in the • A total of 10 H + ions are translocated from the matrix
reduced coemymes NADH and FADH2 • The energy from across the inner membrane per electron pair flowi ng from
these electrons is used to generate the proton-motive force. NADH to 0 2 (see Figure 12-16), whereas 6 H+ ions are trans-
In the mitochondrion, the proton-motive force is generated located per electron pair flowing from FADH2 to 0 2•
by coupling electron flow (from N ADH and FADH2 to 0 2) to • The proton-motive force is due largely to a voltage gradi-
the uphill transport of protons from the matrix across the in- ent across the inner membrane produced by proton pumping;
ncr membrane to the intermembrane space. This process to- the pH gradient plays a quantitatively less important role.
gether with the synthesis of ATP from ADP and P, driven by
the proton-motive force is called oxidative phosphorylation.
As electrons flow from FADH2 and NADH to 0 2 they pass
through multi protein complexes. The four major complexes are 12.4 Harnessing the Proton-Motive
NADH-CoQ reductase (complex 1), succinate-CoQ reductase Force to Synthesize ATP
(complex II), CoQHr-cytochrome c reductase (complex III ),
and cytochrome c oxidase (complex IV) (see Figure 12-16). The hypothesis that a proton-motive force across the inner mi-
tochondrial membrane is the immediate source of energy for
Each complex contains one or more electron-carrying ATP synthesis was proposed in 196 1 by Petef' Mitchell. Virtu-
prosthetic groups: iron-sulfur clusters, flavins, heme groups, ally all researchers studying oxidative p hosphorylation and
and copper ions (see Table 12-2 ). Cytochrome c, which con- photosynthesis initially rejected his chemiosmotic hypothesis.
tains heme, and coenzyme Q (CoQ), a lipid-soluble small They favored a mechanism similar to the then well-elucidated
molecule, are mobi le carriers that shuttle electrons between substrate-level phosphorylation in glycolysis, in which chemi-
the complexes. cal transformation of a substrate molecule (in the case of gly-
Complexes I, Ili, and IV pump protons from the matrix colysis, phosphoenolpyruvate) is directly coupled to ATP
into the intermembrane space. Complexes I and II reduce synthesis. Despite intense efforts by a large number of investi-
CoQ to CoQH2 , which carries protons and high-energy elec- gators, however, compelling evidence for such a substrate-level
trons to complex Ill. The heme protein cytochrome c carries phosphorylation-mediated mechanism was never observed.
electrons from complex liT to complex IV, which uses them Definitive evidence supporting Mitchell's hypothesis de-
to pump protons· and reduce molecular oxygen to water. pended on developing techniques to purify and reconstitute
organelle membranes and membrane proteins. The experi-
The high-energy electrons from NADH enter the electron
ment with vesicles made from chloroplast thylakoid mem-
transport chain through complex I, whereas the high-energy
branes (equivalent to the inner membrane of mitochondria)
electrons from FADH2 (derived from succinate in the citric
containing ATP synth ase, outlined in Figure 12-23, was one
acid cycle) enter the electron transport chain through com-
of several demonstrating that this protein is an ATP-generat-
plex IT. Additional electrons derived from FADH2 by the ini-
ing enzyme and that ATP generation is dependent on proton
tial step of fatty acyl-CoA 13-oxida tion increase the supply of
movement down an electrochemical gradient. It turns our that
CoQH2 available for electron transport.
the protons actually move through the ATP synthase as they
The Q cycle allows four protons to be translocated per pair traverse the membrane.
of electrons moving through complex III (see Figure 12-18 ). As we shall see, rhe ATP synthase is a multiprotein com-
plex that can be subdivided into two subcomplexes called F0
Each electron carrier in the chain accepts an electron or
(contai ning the transmembrane portions of the complex)
electron pair from a carrier with a less positive reduction
and r, (containing the globular portions of the complex that
potential and transfers the electron to a carrier with a more
sit above the membrane and point toward the matrix space
positive reduction potential. Thus the reduction potentials of
in mitochondria). Thus the ATP synthase is often also called
electron carriers favor unidirectional, "downhill," electron
the F0 F 1 complex; we will use the terms interchangeably.
flo\.\: from NADH and FADH2 to 0 2 (see Figure 12-19).
Within the inner membrane, electron transport complexes The Mechanism of ATP Synthesis Is Shared
assemble into supercomplexes held together by cardiolipin, a
Among Bacteria, Mitochondria, and Chloroplasts
specialized phospholipid. Supercomplex formation may en-
hance the speed and efficiency of generation of the proton- Although bacteria lack internal membranes, aerobic bacte-
motive force. ria nonetheless carry our oxidative phosphorylation by the
same processes that occur in cukaryoric mitochondria and

544 CHAPTER 12 • Cellular Energetics

 Bacterium
H
Fo ----=-f~:---::-;.
pH 7.5 F , - -- , - - t

Plasma--~
Thylakoid membrane

l Soak for several minutes

at pH 4.0
membrane

Mitochondrion
Outer lntermembrane space
pH 4.0
membrane\

F0 ---7"'-'---:rr Matrix
F, --r--~0-..t:

l Add a solution of pH 8.0

that contains ADP and P,

Inner
membrane

pH 8.0
Chloroplast
Outer Light lntermembrane
EXPERIMENTAL FIGURE 12·23 Synthesis of ATP by ATP membrane
synthase depends on a pH gradient across the membrane. Isolated
chloroplast t hylakoid vesicles containing ATP synthase (F0 F1 particles)
were equilibrated in the dark with a buffered solution at pH 4.0. When
the pH in the thylakoid lumen became 4.0, the vesicles were rapidly
mixed with a solution at pH 8.0 containing ADP and P,. A burst of ATP
synthesis accompanied the transmembrane movement of protons
driven by the 10,000-fold H+ concentration gradient (1o- 4 M versus
1o-s M). In similar experiments using "inside-out" preparations of
mitochondrial membrane vesicles, an artificially generated membrane Inner
electric potential also resulted in ATP synthesis. membrane
Thylakoid membrane

chloroplasts (Figure 12-24). Enzymes that catalyze the reac- FIGURE 12·24 ATP synthesis by chemiosmosis is similar in
tions of both the glycolytic pathway and the citric acid cycle bacteria, mitochondria, and chloroplasts. In chemiosmosis, a
are present in the cytoso1 of bacteria; enzymes that oxidize proton-motive force generated by proton pumping across a
NADH to NAD+ and transfer the electrons to the ultimate membrane is used to power ATP synthesis. The mechanism and
acceptor 0 2 reside in the bacterial plasma membrane. The membrane orientation of the process are similar in bacteria, mito-
movement of electrons through these membrane carriers is chondria, and chloroplasts. In each illustration, the membrane surface
facing a shaded area is a cytosolic face; the surface facing an unshaded,
coupled to the pumping of protons out of the cell. The move-
white area is an exoplasmic face. Note that the cytosolic face of the
ment of protons back into the cell, down their concentration
bacterial plasma membrane, the matrix face of the inner mitochondrial
gradient through ATP synthase, drives the synthesis of ATP.
membrane, and the stromal face of the thylakoid membrane are all
The bacterial ATP synthase (F0 F 1 complex) is essentially
equivalent. During electron transport, protons are always pumped
identical in structure and function to the mitochondrial and from the cytosolic face to the exoplasmic face, creating a proton con-
chloroplast ATP synthases but is simpler to purify and study. centration gradient (exoplasmic face > cytosolic face) and an electric
Why is the mechanism of ATP synthesis shared among potential (negative cytosolic face and positive exoplasmic face) across
both prokaryotic organisms and eukaryotic organelles? the membrane. During the synthesis of ATP, protons flow in the reverse
Primitive aerobic bacteria were probably the progenitors of direction (down their electrochemical gradient) through ATP synthase
both mitochondria and chloroplasts in eukaryotic cells (Fig- (F0 F1 complex), which protrudes in a knob at the cytosolic face in
ure 12-25). According to this endosymbiont hypothesis, the all cases.

12.4 Harnessing the Proton-Motive Force to Synthesize ATP 545

 Eukaryotic
plasma membrane
I

Bacterial plasma
membrane becomes
inner membrane inner membrane
of mitochondrion of chloroplast

Inner membrane buds

off thylakoid vesicles
Mitochondrial matrix T hylakoid
membrane
FIGURE 12-25 Endosymbiont hypothesis for the evolutionary mitochondrion (left) or chloroplast (right). Budding of vesicles from
origin of mitochondria and chloroplasts. Endocytosis of a bacterium the inner chloroplast membrane, such as occurs during development
by an ancestral eukaryotic cell (step 0 ) would generate an organelle of chloroplasts in contemporary plants, would generate the thylakoid
with two membranes, the outer membrane derived from the eukary- membranes with the F1 subunit remaining on the cytosolic face,
otic plasma membrane and the inner one from the bacterial membrane facing the chloroplast stroma (step i) ). Membrane surfaces facing a
(step fJ). The F1 subunit of ATP synthase, localized to the cytosolic face shaded area are cytosolic faces; surfaces facing an unshaded area are
of the bacterial membrane, would then face the matrix of the evolving exoplasmic faces.

inner mitochondrial membrane would be derived from the (Figure 12-26a). The Fa component contains three types of in-
bacterial plasma membrane with its cyrosolic face pointing tegral membrane proteins, designated a, b, and c. In bacteria
toward what became the matrix space of the mitochondrion. and in yeast mitochondria the most common subunit stoichi-
Similarly, in plants the progenitor's plasma membrane be- ometry is a 1b 1 c 10, but F0 complexes in animal mitochondria
came the chloroplast's thylakoid membrane and its cyrosolic have 12 c subunits and those in chloroplasts have 14. In all
face pointed toward what became the stromal space of the chlo- cases the c subunits form a doughnut-shaped ring ("c ring") in
roplast. In all cases, ATP synthase is positioned with the the plane of the membrane. The a and two b subunits are rig-
globular F 1 domain, which catalyzes ATP synthesis, on the idly linked to one another but not to the c ring, a critical fea-
cytosolic face of the membrane, so ATP is always formed on ture of the protein to which we will return shortly.
the cytosolic face 'of the membrane (sec Figure 12-24 ). Protons The F 1 portion is a water-soluble complex of five distinct
always flow through ATP synthase from the exoplasmic to polypeptides with the composition a 3 f3:!'Y8£ that is normally
the cytosolic face of the membrane. This flow is driven by firmly bound to the F0 subcomplex at the surface of the mem-
the proton motive force. Invariably, the cytosolic face has a brane. The lower end of the rodlike -y subunit of the F 1 sub-
negative electric potential relative to the exoplasmic face. complex is a coiled coil that fits into the center of the c-subunit
In addition to ATP synthesis, the proton-motive force across ring of Fa and appears rigidly attached to it. Thus when the c-
the bacterial plasma membrane is used to power other pro- subunit ring rotates, the rodlike -y subunit moves with it. The
cesses, including the uptake of nutrients such as sugars (using F 1 E subunit is rigidly attached to-y and also forms tight con-
proton/sugar symporters) and the rotation of bacterial flagella. tacts with several of the c subunits of F0 . The a and [3 subunits
Chemiosmotic coupling thus illustrates an important principle are responsible for the overall globu lar shape of the F 1 sub-
introduced in our discussion of active transport in Chapter 11: complex and associate in alternating order to form a hexamer,
the membrane potential, the concentration gradients ofprotons a[3a[3a[3, or (a[3)}, which rests atop the single long -y subunit.
(and other ions) across a membrane, a11d the phosphoanhydride The f 1 8 subunit is permanently linked to one of the F 1 a sub-
bonds in A TP are equiualent and interc01wertible forms of units and also binds to the b subunit of Fa. Thus the Fa a and b
chemical potential energy. Indeed, ATP synthesis through ATP subunits and the 8 subunit and (af3)} hexamer of the F 1 com-
synthase can be thought of as active transport in reverse. plex form a rigid structure anchored in the membrane. The
rodlike b !.ubunit<> form a "stator" that prevents the (aJ3)3 hex
amer from moving while it rests on the-y subunit, whose rota-
ATP Synthase Comprises F0 and F1
tion together with the c subunits of F0 plays an essential role in
Multiprotein Complexes the ATP synthesis mechanism described below.
With general acceptance of Mitchell's chemiosmotic mecha- When ATP synthase is embedded in a membrane, the F 1
nism, researchers turned their attention to the structure and component forms a knob that protrudes from the cytosolic
operation of ATP synthase. The complex has two principal (in the mitochondrion this is the matrix ) face. Because F 1
components, Fa and F" both of which are multimeric proteins separated from membranes is capable of catalyzing ATP

546 CHAPTER 12 • Cellular Energetics

(;} FOCUS ANIMATION: Proton Translocating, Rotarv F-ATPase
(a) 100 nm (b)
w II Adjacent proton exits
IJ c ring
half-channel II rotates

Cytosolic
medium

Exoplasmic
medium
Proton bound
Proton to negative charge
half-channel on Asp-61
Static Rotates

FIGURE 12-26 Structure of ATP synthase (the F0 F 1 complex) in subunits and the F1 I) subunit and (al)h hexamer form a rigid structure
the bacterial plasma membrane and mechanism of proton translo- anchored in the membrane (orange). During proton flow, the c ring and
cation across the membrane. (a) The F0 membrane-embedded portion the attached F1 E and-y subunits rotate as a unit (green), causing
of ATP synthase is built of three integral membrane proteins: one copy of conformation changes in the F1 13 subunits, leading to ATP synthesis.
a, two copies of b, and on average 10 copies of c arranged in a ring in the (b) Potential mechanism of proton translocation. Step 0 : A proton from
plane of the membrane. Two proton half-channels in subunit a mediate the exoplasmic space enters half-channel I and moves toward the
proton movement across the membrane (proton path indicated by red "empty" (unprotonated) Asp-61 proton-binding site. The negative charge
arrows). Half-channel I allows protons to move one at a time from the (blue"-") on the unprotonated side chain Asp-61 is balanced, in part, by
exoplasmic medium to the negatively charged side chain of Asp-61 in a positive charge on the side chain of Arg-210 (red" + "). Step H :The
the center of a c subunit near the middle of the membrane. The proton fills the empty proton-binding site and simultaneously displaces
proton-binding site in each c subunit is represented as a white circle with the Arg-21 0 side chain, which swings over to the filled proton-binding
a blue" -" representing the negative charge on the side chain of Asp-61. site on the adjacent c subunit. As a consequence the proton bound at
Half-channel II permits protons to move from the Asp-61 of an adjacent that adjacent site is displaced. Step 10: The displaced adjacent proton
c subunit into the cytosolic medium. The F1 portion of ATP synthase moves through half-channel II and is released into the cytosolic space,
contains three copies each of subunits a and I) that form a hexamer leaving an empty proton-binding site on Asp-61. Step B: Counterclock-
resting atop the single rod-shaped -y subunit, which is inserted into the wise rotation of the entire c ring moves the "empty" c subunit over half-
c ring of F0• TheE subunit is rigidly attached to the-y subunit and also to channel I. Step 111: the process is repeated. [Adapted from M. J. Schnitzer,
several of the c subunits. The 8 subunit permanently links one of the a 2001, Nature 41 0:878; P. D. Boyer, 1999, Nature 402:247; and C. von Ballmoos,
subunits in the F1 complex to the b subunit of F0. Thus the F0 a and b A. Wiedenmann, and P. Dimroth, 2009, Ann. Rev. Biochem. 78:649-672.]

hydrolysis (ATP conversion to ADP plus P,) in the absence of protons from the exoplasmic medium (intermembrane space in
the F0 component, it has been called the F 1 ATPase; however, the mitochondrion ) ro the cytosolic (matrix) medium. How-
its function in cells is the reverse, to synthesize ATP. ATP ever, the coupling berween proton flow and ATP synthesis must
hydrolysis is a spontaneous process (.l G < 0); thus energy is not occur in the same portions of the protein, because the
required to drive the ATPase "in reverse" and generate ATP. nucleotide-binding si tes on the f3 subunits of Ft. where ATP
synthesis occurs, are 9-10 nm from the surface of the mito-
chondrial membrane. The most widely accepted model for ATP
Rotation of the F1 "' Subunit, Driven by Proton
synthesis by the F0 F 1 complex-the binding-change mecha-
Movement Through F0, Powers ATP Synthesis llism-posits just such an indirect coupling (Figure 12-27).
Each of t he three f3 subunits in the globular F 1 portion of the According to this mechan ism, energy released by the
complete F0F 1 complex can bind ADP and P, and catalyze the "downhill" movement of protons through F0 directl} powers
endergonic synthesis of ATP when coupled to the flow of rotation of the c-subunit ring together with its attached"' and

12.4 Harnessing the Proton-Motive Force to Synthesize ATP 547

(;} FOCUS ANIMATION: ATP Synthesis
PODCAST: ATP Synthesis

Reaction
Rotation (no rotat ion)

D fJ

ATP

Rot.,;oo 1· ADP
~
P,

Reaction
(no rotation)

a
.
p
ATP

F IGURE 12-27 The binding-change mechanism of ATP synthesis and a decrease in the binding affinity of the fl 2 subunit for a previously
from ADP and P;. This view is looking up at F1 from the membrane sur- bound ATP (from T--+ 0 ), causing release of the bound ATP. Step f) :
face (see Figure 12-26). As the-y subunit rotates by 120° in the center, Without additional rotation the ADP and P, in the T site (here the [l3
each of the otherwise identical F1 fl subunits alternates between three subunit) form ATP, a reaction t hat does not requ ire an input of addi-
conformational states (0, open with oval representation of the binding tional energy due to the special environment in the active site of the T
site; L, loose with a rectangular binding site; T, tight with a triangular state. At the same time a new ADP and P, bind loosely to the unoccu-
site) that differ in their binding affinities for ATP, ADP, and P,. The cycle pied 0 site on ~2 • Step 10:Proton flux powers anot her 120° rotation of
begins (upper left) when ADP and P, bind loosely to one of the three the-y subunit, consequent conformational changes in the binding sites
fl subunits (here, arbitrarily designated [l 1) whose nucleotide-binding (L--+ T, 0--+ L, T--+ 0), and release of ATP from [l 3• Step m:
Without
site is in the 0 (open) conformation. Proton flux through the F0 portion additional rotation the ADP and P in the T site of (3 1 form ATP, and
of the protein powers a 120° rotation of the-y subunit (relative to the addit iona l ADP and P; bind to the unoccupied 0 site on [l 3• The process
fixed fl subunits) (step 0 ). This causes the rotating -y subunit, which is continues with rotation (step 1,'1) and ATP formation (step l'al until the
asymmetric, to push differentially against the fl subunits, resulting in a cycle is complete, with three ATPs having been produced for every
conformational change and an increase in the binding affinity of the~ ~ 360° rotation of -y. [Adapted from P. Boyer, 1989, FASEBJ. 3:2164; Y. Zhou
subunit for ADP and P, (from 0--+ L), an increase in the binding affinity et al., 1997, Proc. N at'l. Acad. Sci. USA 94:1 0583; and M. Yoshida, E. Muneyuki, and
of the ~3 subunit for ADP and P, that were previously bound (from L--+ n. T. Hisabori, 2001, Nat. Rev. Mol. Cell Bioi. 2:669-677.]

e subunits (see Figure 12-26a). The -y subunit acts as a cam, 3. AT (tight} state that binds ADP and P, so tightly that
or nonsymmetrical rotating shaft, whose rotation within the they spontaneously react and form ATP
center of the static (a(3 h hexamer of F1 causes it to push se-
In the T state the ATP produced is bound so tightly that it
quentially against each of the (3 subunits and thus cause cycli-
cannot readily dissociate from the site-it is trapped until an-
cal changes in their conformations between three different
other rotation of the-y subunit returns that~ subunit to the 0
states. As schematically depicted in a view of the bottom of
state, thereby re leasing ATP and begi nning the cycle again.
the (a(3)J hexamer's globular structure in Figure 12-27, rota-
ATP or ADP also binds to regu latory or a llosteric sites on the
tion of the-y subunit relative to the fixed (a(3 h hexamer
three a subunits; this binding modifies the rate of ATP synthe-
causes the nucleotide-binding site of each (3 subunit to cycle
sis accordiug to the level of ATP and ADP m the matrix, but
through three conformatiOnal states in the following order:
is not directly involved in synthesis of ATP from ADP and P,.
1. An 0 (open) state that binds ATP very poorly and ADP Several types of evidence support the binding-change mech-
and P; weakly anism. First, biochemical studies showed that one of the three (3
subunits on isolated F 1 particles can tightly bind ADP and P,
2. An L (loose ) state that binds ADP and P, more strongly and then form ATP, which remains tightly bound. The mea-
but cannot bind ATP sured tiC for this reaction is near zero, indicating that once

548 CHAPTER 12 • Cellular Energetics

 0 , .

~ VIDEO: Rotation of Actin Filament Bound to ATP Synthase

-----------------------------------------------------
EXPERIMENTAL FIGURE 12-28 They subunit of the F1 com-
plex rotates relative to the (otllh hexamer. F1 complexes were engi-
neered that contained J3 subunits with an additional His-6 sequence,
which causes them to adhere to a glass plate coated with a metal re-
agent that binds polyhistidine. They subunit in the engineered F1 com-
plexes was linked covalently to a fluorescently labeled actin filament.
When viewed in a fluorescence microscope, the actin filaments were
seen to rotate counterclockwise in discrete 120° steps in the presence
of ATP, powered by ATP hydrolysis by the J3 subunits. (Adapted from H.
Noji et al., 1997, Nature 386:299, and R. Yasuda et al., 1998, Ce/193:1117.]

ADP and P, are bound to the T state of a ~subunit, they spon- These observations established that the-y subunit, along with
taneously form ATP. Importantly, dissociation of the bound the attached c ring and E subunit, does indeed rotate, thereby
ATP from the~ s ubunit on isolated F 1 particles occurs ex- driving the conformational changes in the 13 subunits that are
tremely slowly. This finding suggested that dissociation of ATP required for binding of ADP and P., followed by synthesis and
would have to be powered by a conformational change in the ~ subsequent release of ATP. .
subunit, which in turn would be caused by proton movement.
X-ray crystal log raphic analysis of the (a~), hexamer Multiple Protons Must Pass Through ATP
yielded a striking conclusion: although the three ~ subunits
Synthase to Synthesize One ATP
arc identical in sequence and overall structure, the ADP/
ATP-binding sites have different conformations in eac h sub- A simple calculation indicates that the passage of more than
unit. The most reasonable conclusion was that the three ~ one proton is required to synthesize one molecule of ATP
subunits cycle in an ene rgy-dependent reaction between from ADP and P,. Although the .l G for this reaction under
three conformational states (0, L, T), in which the nucleotide- standa rd conditions is + 7.3 kcal/mol, at the concentrations
binding site has substantiall y different structures. of reactants in the mitochondrion, .l G is probably higher
In other stu~ies, intact F1l 1 complexes were treated with (+ 10 to + 12 kcal/mol). We can calculate the amount of free
chemical cross-linking agents that covalently linked the-y energy released by the passage of 1 mol of protons down an
and E subunits and the c-subunit ring. The observation that electroc hemical gradient of 220 mY (0.22 V) from the
such treated complexes could synthesize ATP or use ATP to Nernst equation, !letting n = 1 and measuring .lE in volts:
power proton pumping indicates that the cross-linked pro-
teins normally rotate together. .lG(cal/mol) = -nF!:.E = -(23,062 cal · V 1
• mol 1
).lE
1 1
Finally, rotation of the-y subunit relative to the fixed (a~lJ = (23,062 cal · V • mol )(0.22 V)
hexamer, as proposed in rhe binding-change mechanism, was = -5074 cal/mol, or -5.1 kcal/mol
observed directly in the clever experiment depicted in Figure
12-28. In one modification of this experiment in which tiny Because the downhill movement of 1 mol of protons releases
gold particles, rather than an actin filament, were attached to just over 5 kcal of free energy, the passage of at least two
the-y subunit, rotation rates of 134 revolutions per second protons is required for synthesis of each molecule of ATP
were observed. H ydrolysis of three ATPs, which you recall is from ADP and P,.
the reverse reaction catalyzed by the same enzyme, is thought
to power one revolution; this result is close to the experimen-
F0 c Ring Rotation Is Driven by Protons Flowing
tally determined rate of ATP hydrolysis by F0 f 1 complexes:
about 400 ATPs per second. In a related experiment, a -y sub- Through Transmembrane Channels
unit linked to an E subunit and a ring of c subunits was seen Each copy of subu nit c contains two membrane-spanning a
to rotate relative to the fixed (a~b hexamer. Rotation of the helices that form a hairpin-like structure. An aspartate resi-
'Y subunit in these experiments was powered by ATP hydroly- due, Asp-61 (E. coli ATPase numbering), in the center of one
sis, the reverse of the normal process in which proton move- of these helices in each subunit is thought to play a key role
ment through the F0 complex drives rotation of the-y subunit. in proton movement by binding and releasing protons as

12.4 Harnessing the Proton-Motive Force to Synthesize ATP 549

they traverse the membrane. Chemical modification of this ATPs. Why these otherwise similar F0 F 1 complexes have
aspartate by the poison dicyclohexylcarbodiimide or its mu- evolved to have different H- :ATP ratios is not clear.
tation to alanine specifically blocks proton movement
through F0 • According to one current model, the protons tra- ATP-ADP Exchange Across the Inner
verse the membrane via two staggered, proton ha lf-cha nnels,
Mitochondrial Membrane Is Powered
I and II (see Figure 12-26a and b). They are ca ll ed half-
channels because each only extends halfway across the mem- by the Proton-Motive Force
brane; the intramembrane termini of the channels are at the The proton-motive force is used to power multiple energy-
levd of Asp-61 in the middle of the membrane. Half-channel requiring processes in cells. In addition to powering ATP syn-
I is open only to the exoplasmic surface and II is open only thesis, the proton-motive force across the inner mitochondria l
to the cytosolic face. Prior to rotation, each of the Asp-61 membrane powers the exchange of ATP formed by oxidative
carboxylate side chains in the c subunits are bound to a pro- phosphorylation inside the mitochondrion for ADP and P, in
ton, except that on the c subunit in contact with half-channel the cytosol. This exchange, which is required to supply ADP
I. The negative charge on that unprotona ted carboxylate (the and P, substrate for oxidative phosphorylation to continue, is
"empty" proton-binding site; see Figure 12-26b, bottom) is mediated by two proteins in the inner membra ne: a phos-
neutralized by interaction with the positively charged side phate transporter (HPO/ /OH- antiporter) that mediates
chain of Arg-21 0 from the a subunit. Proton translocation the import of one HPO/- coupled lo the export of one OH ,
across the membrane begins w hen a proton from the cxo- and an ATPIADP antiporter (figure 12-29).
plasmic medium moves upwards through half-channel I (Fig- The ATP/ADP antiporter allows one molecule of ADP to
ure 12-26b, step 0 ). As that proton moves into the empty enter only if one molecule of ATP exits simultaneously. The
proton-binding site, it displaces the Arg-210 side chain, ATP/ADP antiporter, a dimer of two 30,000-Da subunits,
which swings toward the filled proton-binding site of the makes up 10-15 percent of the protein in the inner membrane,
adjacent c subunit in contact wit h half-channel II (step f)).
As a consequence, the positive side chain of Arg-210 dis-
places the proton bound to Asp-61 of the adjacent c subunit.
Inner mitochondrial
This displaced proton is now free to travel up half-channel II H + co ncentration membrane
and out into the cytosolic medium (step ID). Thus, when one gradient
proton entering from half-channel I binds to the c ring, a dif-
ferent proton is released to the opposite side of the membrane Membrane { +
electric + Matrix
via half-channel ll. Rotation of the entire c ring due to ther-
potential +
maVBrownian motion (step 19) then allows the newly unpro-
tonated c subunit to move into alignment above half-channel
I as an adjacent, protonated c subunit rotates in to take its Trans location of H
-H
place under half-channel II. The entire cycle is then repeated during electron transport
(step ~), as additional protons move down their electrochem-
ical gradient from the exoplasmic medium to the cytosolic
: OH } Phosphate transporter
medium. During each partial rotation (360° divided by the HP0 4 2
number of c subunits in the ring), the c ring rotation is ratch-
eted in that net movement of the ring only occurs in one di-
rection. The energy driving the protons across the membrane, 3 3
and thus rotation of the c ring, comes from the electric poten- ADP (:==?) ., ADP } ATP/AOP antiporter
ATP4 .,._ _.
_-.:
~ -----,"7'-:...-- ATP4
tial and pH gradient across the membrane. If the direction of
proton flow is reversed, which can be done by experimentally
reversing the direction of the proton gradient and proton-
motive force, the direction of c ring rotation is reversed. lntermembrane
space
Because the-y subunit ofF 1 is tightly attached to the c ring
of F0 , rotation of the c ring associated with proton movement
causes rotation of the -y subunit. According to the binding- ATP4 + OH
change mechanism, a 120° rotation of -y powers synthesis of
one ATP (see Figure 12-2 7). Thus complete rotation of the c
ring by 360° would generate three ATPs. In E. colt, where the
FIGURE 12-29 The phosphate and ATP/ADP transport system in
I-- 0 composition is a 1b 2 c 10, movement of 10 protons drives one the inner mitochondrial membrane. The coordinated action of two
complete rotation and thus synthesis of three ATPs. This value anti porters (purple and green) results in the uptake of one ADP 3 and
is consistent with experimental data on proton flux during ATP 2 4
one HP04 in exchange for one ATP and one hydroxyl, powered by
synthesis, providing indirect support for the model coupling the outward translocation of one proton (mediated by the proteins of
proton movement to c-ring rotation depicted in Figure 12-26. the electron transport chain, blue) during electron transport. The outer
The F0 from chloroplasts contains 14 c subunits per ring, and membrane is not shown here because it is permeable to molecules
movement of 14 protons would be needed for synthesis of three smaller than 5000 Da.

550 CHAPTER 12 • Cellular Energetics

so it is one of the more abundant mitochondrial proteins. mitochondrial membrane. If the resulting proton-motive force
Functioning of the two ami porters together produces an influx is not dissipated during the synthesis of ATP from ADP and P,
of one ADP 3- and one P,2 and efflux of one ATP4 together (or during other energy-requiring processes), both the trans-
with one OH . Each OH- transported ounvard combines with membrane proton concentration gradient and the membrane
a proton, translocated during electron transport to the inter- electric potential will increase to very high levels. At this point,
membrane space, to form H 20. This drives the overall reaction pumping of additional protons across the inner membrane re-
in the direction of ATP export and ADP and P, import. quires so much energy that it eventually ceases, blocking the
Because some of the protons translocated out of the mi- coupled oxidation of NADH and other substrates.
tochondrion during electron transport provide the power (by
combining with the exported OH ) for the ATP-ADP ex-
Brown-Fat Mitochondria Use the Proton-Motive
change, fewer protons are available for ATP synthesis. It is
estimated that for every four protons translocated out, three Force to Generate Heat
are used to synthesize one ATP molecule and one is used to Brown-fat tissue, whose color is due to the presence of abun-
power the export of ATP from the mitochondrion in ex- dant mitochondria, is specialized for the generation of heat.
change for ADP and P,. This expenditure of energy from the In contrast, white-fat tissue is specialized for the storage of
proton concentration gradient to export ATP from the mito- fat and contains relatively few mitochondria.
chondrion in exchange for ADP and P, ensures a high ratio The inner membrane of brown-fat mitochondria contains
of ATP to ADP in the cytosol, where hydrolysis of the high- thermogenin, a protein that functions as a natural uncoupler of
energy phosphoanhydride bond of ATP is utilized to power oxidative phosphorylation and generation of a proton-motive
many energy-requiring reactions. force. Thermogenin, or UCP1, is one of several uncoupling pro-
teins (UCPs) found in most eukaryotes (but not in fermentative
B Studies of what turned out to be ATP/ADP antiporter yeasts). Thermogenin dissipates the proton-motive force by ren-
H activity were first recorded abou t 2000 years ago, dering the inner mitochondrial membrane permeable to pro-
when Dioscorides (-AD 40-90) described a poisonous herb tons. As a consequence the energy released by NADH oxidation
from the thistle Atractylis gummifera, found commonly in in the electron transport chain and used to create a proton gra-
the Mediterranean region. The same agent is found in the dient is not then used to synthesize ATP via ATP synthase.
traditional Zulu multipurpose herbal remedy impila (Calli- Instead, when protons move back into the matrix down their
lepis /aureola). In Zulu impila means "health," although it concentration gradient via thermogenin, the energy is released
has been associated with numerous poisonings. In 1962 the as heat. Thermogenin is a proton transporter, not a proton
active agent in the herb, atractyloside, which inhibits the channel, and shuttles protons across the membrane at a rate
ATP/ADP antiporrer, was shown to inhibit oxidative phos- that is a millionfold slower than that of typical ion channels (see
phorylation of extramitochondrial ADP but not intramito- Figure 11-2). Thermogenin is similar in sequence to the mito-
chondrial ADP. This demonstrated the importance of the chondrial ATP/ADP transporter, as are many other mitochon-
ATP/ADP antiporter and has provided a powerful tool to drial transporter proteins that compose the ATP/ ADP
study the mechanism by which this transporter functions. transporter family. Certa in small-molecule poisons also func-
Dioscorides lived near Tarsus, at the time a province of tion as uncouplers by rendering the inner mitochondrial mem-
Rome in southeastern Asia Minor in what is now Turkey. brane permeable to protons. One example is the lipid-soluble
His five-volume De Materia Medica (The Materials of Medi- chemical 2,4-dinitrophenol (DNP), which can reversibly bind
cine) "on the preparation, properties, and testing of drugs" to and release protons and shuttle them across the inner mem-
described the medicinal properties of about 1000 natural brane from the intermembrane space into the matrix.
products and 4 740 medicinal usages of them. For approxi- Environmental conditions regulate the amount of ther-
mately 1600 years it was the basic reference in medicine from mogenin in brown-fat mitochondria. For instance, during
northern Europe to the lpdian Ocean, comparable to today's the adaptation of rats to cold, the ability of their tissues to
Physicians' Desk Reference as a guide for using drugs. • generate heat is increased by the induction of thermogenin
synthesis. In cold-adapted animals, thermogenin may con-
stitute up to 15 percent of the total protein in the inner
Rate of Mitochondrial Oxidation Normally
mitochondrial membrane.
Depends on ADP Levels For many years it was known that small animals and
If intact isolated mitochondria are provided with NADH (or human infants expressed significant amounts of brown fat,
a source of FADH2 such as succinate) plus 0 2 and P, but not but there was scant evidence for it playing a significant role
ADP, the oxidation of NADH and the reduction of 0 2 rapidly in adult humans. In the newborn human, thermogenesis b}
cease, because the amount of endogenous ADP IS depleted by brown-fat mitochondria is vital to survival, as it is in hiber-
ATP formation. If ADP is then added, the oxidation of NADH nating mammals. In fur seals and other animals naturally
is rapid ly restored. Thus mitochondria can oxidize FADH2 acclimated to the cold, muscle-cell mitochondria contain
and NADH only as long as there is a source of ADP and P, to thermogenin; as a result, much of the proton-motive force is
generate ATP. This phenomenon, termed respiratory control, used for generating heat, thereby maintaining body tempera-
occurs because oxidation of NADH and succinate (FADH2 ) ture. Recently investigators have used sophisticated func-
is obligatorily coupled to proton transport across the inner tional imaging methods (for example, positron-emission

12.4 Harnessing the Proton -Motive Force to Synthesize ATP 551

tomography) to definitively establish the presence of brown 12.5 Photosynthesis and
fat m adult humans in the neck, clavicle, and other sites, the
levels of which are significantly increased on cold exposure. Light-Absorbing Pigments
8 We now shift our attention to photosynthesis, the second
H main process for synthesizing ATP. In plants, photo-
KEY CONCEPTS of Section 12.4 synthesis occurs in chloroplasts, large orga nelles found
mainly in leaf cells. During photosynthesis, chloroplasts cap-
Harnessing the Proton-Motive Force ture the energy of sunlight, convert it into chemical energy in
to Synthesize A TP the form of A1 P and NADPH, and then use this energy to
Peter Mitchell proposed the chemiosmotic hypothesis that make complex carboydrates our of carbon dioxide and water.
a proton-motive force across the inner mitochondrial mem- The principal carbohydrates produced are polymers of hex-
brane is the immediate source of energy for ATP synthesis. ose (six-carbon) sugars: sucrose, a glucose-fructose disaccha-
ride (see Figure 2-19), and leaf starch, a mixture of two types
Bacteria, mitochondria, and chloroplasts all use the same of a large insoluble glucose polymer called amylose and amy-
chemiosmotic mechanism and a similar ATP synthase to lopectin. Starch is the primary storage carbohydrate in plants
generate ATP (see Figure 12-24 ). (Figure 12-30). Leaf starch is synthesized and stored in the
ATP synthase (the F0 F 1 complex) catalyzes ATP synthesis chloroplast. Sucrose is synthesized in the leaf cytosol from
as protons flow through the inner mitochondrial membrane three-carbon precursors generated in the ch loroplast; it is
(plasma membrane in bacteria) down their electrochemical transported to nonphotosynthetic (nongrecn) plant tissues
proton gradient. (e.g., roots and seeds), which metabolize sucrose for energy
- F0 contains a ring of 10-14 c subunits that is rigidly linked by the pathways described in the previous sections. Photosyn-
to the rod-shaped 'Y subunit and thee subunit of F1. Together thesis in plants, as well as in eukaryotic single-celled algae
they rotate during ATP synthesis. Resting atop the 'Y subunit and in several photosynthetic bacteria (e.g:, the cyanobac-
is the hexameric knob of F 1 f(o:l3hl, which protrudes into the teria and prochlorophytes), also generates oxygen. The over-
mitochondrial matrix (cytosol in bacteria). The three 13 sub- a II reaction of oxygen-genera ring photosynthesis,
units are the sites of ATP synthesis (see Figure 12-26).
Movement of protons across the membrane via two half-
channels at the interface of the F0 a subunit and the c ring pow- is the reverse of the overall reaction by which carbohydrates
ers rotation of the c ring with its attached F 1 e and 'Y subunits. are oxidized to C0 2 and H 20. In effect, photosynthesis in
chloroplasts produces energy-rich sugars that are broken
Rotation of the F 1 'Y subunit, which is inserted in the cen-
down a nd harvested for energy by mitoc hondria using
ter of the nonrotating (o:l3h hexamer and operates like a cam-
oxidative phosphorylation.
shaft, leads to changes in the conformation of the nucleotide-
Although green and purple bacteria also carry our photo-
binding sites in t~e three F 1 13 subunits (see Figure 12-27). By
synthesis, they use a process that does not generate oxygen.
means of this binding-change mechanism, the 13 subunits
As discussed in Section 12.6, detailed analysis of the photo-
bind ADP and P, condense them to form ATP, and then re-
synthetic system in these bacteria has he! ped elucidate the
lease the ATP. Three ATPs are made for each revolution
first stages in the more common process of oxygen-generating
made by the assembly of c, -y, and e subunits.
photosynthesis. In this section, we provide an overview of
The proton-motive force also powers the uptake of P, and the stages in oxygen-generating photosynthesis and introduce
ADP from the cytosol in exchange for mitochondrial ATP the main molecular components, including the chlorophylls, the
and OH , thus reducing some of the energy available for principal light-absorbing pigments. •
ATP synthesis. The ATP/ADP antiporter that participates in
this exchange is one of the most abundant proteins in the
Glucose
inner mitochondrial membrane (see Figure 12-29).
Continued mitochondrial oxidation of NADH and reduc-

~~~
tion of 0 2 are dependent on sufficient ADP being present in
the matrix. This phenomenon, termed respiratory control, is
, ·,
an important mechanism for coordinating oxidation and

~
- 0 -0 0
ATP synthesis in mitochondria.
• In brown fat, the inner mitochondrial membrane contains H OH
the uncoupler protein thermogenin, a proton transporter
that dissipates the proton-motive force into heat. Certain Starch
[poly(a 1->4 glucose))
chemicals also function as uncouplers (e.g., DNP) and have
the same effect, uncoupling oxidative phosphorylation from FIGURE 12-30 Structure of starch. This large glucose polymer and
electron transport. the disaccharide sucrose (see Figure 2-19) are the principal end products
of photosynthesis. Both are built of six-carbon sugars (hexoses).

552 CHAPTER 12 • Cellular Energetics

 Thylakoid Membranes in Chloroplasts Are
the Sites of Photosynthesis in Plants Cuticle~ Leaf

Chloroplasts are lens shaped with a diameter of approximately

Upper ~
5 IJ.m and a width of approximately 2.5 IJ.m. They contain epidermis
about 3000 different proteins, 95 percent of which are encoded
in the nucleus, made in the cytosol, imported into the organelle, Chloroplasts
and then transported to their appropriate membrane or space
(Chapter 13). They are bounded by two membranes, which do Mesophyll
not contain chlorophyll and do not participate di rectly in the
generation of ATP and NADPH driven by light (Figure 12-31 ).
As in mitochondria, the outer membrane of chloroplasts con-
tains porins and thus is permeable to metabolites of small mo-
·. lecular weight. The inner membrane forms a permeability
barrier that contains transport proteins for regulating the move-
ment of metabolites into and out of the organelle.
Unlike mitochondria, chloroplasts contain a third mem-
brane-the thylakoid membrane-on which the light-driven
Chloroplast
generation of ATP and NADPH occurs. The chloroplast thyla-
koid membra ne is believed to constitute a single sheet that Stroma: enzymes that
forms numerous small, interconnected flattened structures, the catalyze C02 fixation
and starch synthesis
thylakoids, wh ich commonly are arranged in stacks termed
grana (Figure 12-3 1). T he spaces within all the thylakoids con-
stitute a single continuous compartment, the thylakoid lumen.
The thylakoid membrane contains a number of integral mem-
brane proteins to w hich are bound several important prosthetic
groups and light-absorbing pigments, most notably chloro-
Thylakoid membrane:
phyll. Starch synthesis and storage occurs in the stroma, the absorption of light by Outer
soluble phase between the thylakoid membrane and the inner chlorophyll, synthesis membrane:
membrane. In photosynthetic bacteria extensive invaginations of ATf>4 , NADPH, and permeable to
electron transport small molecules
of the plasma membrane form a set of internal membranes, also
termed thylakoid membranes, where photosynthesis occurs.

Three of the Four Stages in Photosynthesis

Occur Only During Illumination
The photosynth.etic process in plants can be divided into four
stages (Figure 12-32), each localized to a defined area of the
chloroplast: (1) absorption of light, generation of a high-en-
ergy electrons, and formation of 0 2 from H 2 0; (2) electron
transport leading to red uction of NADP + to NADPH, and Thylakoid
membrane
generation of a proton-motive force; (3) synthesis of ATP;
and (4) conversion of col into carbohydrates, commonly re-
ferred to as carbon fixati6 n. All four stages of photosynthesis
are tightly coupled and con tro lled so as to produce the 0.1 !Jm
amount of carbohydrate required by the plant. All the reac- L__j

tions in stages 1-3 are catalyzed by multiprotein complexes

in the thylakoid membrane. The generation of a pmf and the FIGURE 12-31 Cellular structure of a leaf and chloroplast.
use of the pmf to synthesize ATP resemble stages III and IV of Like mitochondria, plant chloroplasts are bounded by a double
mitochondrial oxidative phosphorylation. The enzymes that membrane separated by an intermembrane space. Photosynthesis
incorporate C0 2 into chemical intermediates and then con- occurs on a third membrane, the thylakoid membrane, which is
vert them to starch are soluble constituents of the ch loroplast surrounded by the inner membrilne and forms a series of flattened
stroma; the enzymes that form sucrose from three-carbon vesicles (thylakoids) that enclose a single interconnected luminal space.
intermediates are in the cytosol. The green color of plants is due to the green color of chlorophyll, all
of which is localized to the thylakoid membrane. A granum is a stack
Stage 1: Absorption of Light Energy, Generation of High- of adjacent thylakoids. The stroma is the space enclosed by the inner
Energy Electrons, and 0 2 Formation The initial step in pho- membrane and surrounding the thylakoids. [Photomicrograph courtesy
tosynthesis is the absorption of light by chlorophylls attached of Katherine Esau, University of California, Davis.]

12.5 Photosynthesis and Light-Absorbing Pigments 553

 Stage 4

Carbon fixation ,
carbohydrate synthesis
Sucrose
t Cytosol

Inner rn
Stage 1 Stage 2 Stage 3 embra
6 C0 2 ~ 2 Glyceraldehyde ne
3-phosphate
Light absorption, Electron transport, formation ATP synthesis
(carbon f ixation)
generation of high- of proton-motive force NADP t H
energy electron, ~
0 2 formation ' ADP + P1 Starch synthesis
. W Light NADPH in the stroma

~AR~~ \)~~
~ l
ATP

~~
Light
LHC [ Oh>J. _

~t
~SIIJ I ,,•
~rPC
PSI ~ 14W
_j~_
Thylakoid ; _ - -
membrane

FIGURE 12-32 Overview of the four stages of photosynthesis. In energy is introduced by absorption of light in photosystem I (PSI), to
stage 1, light is absorbed by light-harvesting complexes (LHC) and the synthesize the high-energy electron carrier NADPH. In stage 3, flow
reaction center of photosystem II (PSII). The LHCs transfer the absorbed of protons down their concentration and voltage gradient through
energy to the reaction centers, which use it, or the energy absorbed the F0F1 ATP synthase drives ATP synthesis. Stagesl-3 in plants take
directly from a photon, to oxidize water to molecular oxygen and gen- place in the thylakoid membrane of the chloroplast. In stage 4, in
erate high-energy electrons (electron paths shown by blue arrows). In the chloroplast stroma, the energy stored in NADPH and ATP is used
stage 2, these electrons move down an electron transport chain, which to convert C02 initially into three-carbon molecules (glyceraldehyde
uses either lipid-sol~ble {Q/QH 2) or water-soluble (plastocyanin, PC) 3-phosphate), a process known as ca rbon fixation. These molecules
electron carriers to shuttle electrons between multiple protein com- are then transported to the cytosol of t he cell for conversion to hexose
plexes. As electrons move down the chain, t hey release energy that the sugars in the form of sucrose. Glyceraldehyde 3-phosphate is also used
complexes use to generate a proton-motive force and, after additional to make starch within the chloroplast.

to proteins in the thylakoid membranes. Like the heme com- Chlorophyll a

ponent of cytochromes, chlorophylls consist of a porphyrin CH 2
ring attached to a long hydrocarbon side chain (Figure 12-33 ). I
CH H
In contrast to the hemes (see Figure 12-14), chlorophylls con- i
tain a central Mg2 + ion (rather than Fe2 +) and have an ad- c
ditional five-membered ring. The energy of the absorbed

FIGURE 12-33 Str ucture of chlorophyll a, the principal pigment

t hat t raps light energy. Electrons are delocalized among three of
chlorophyll a's four central rings (yellow) and the atoms that inter-
connect them. In chlorophyll, a Mg 2 ion, rather than the Fe2+ ion
found in heme, sits at the center of the porphyrin ring and an addi-
tional five-membered ring (blue) is present; otherwise, the structure
of chlorophyll is similar to that of heme, found in molecules such as
hemoglobin and cytochromes (see Figure 12-1 4a). The hydro-carbon ·t

phytol "tail" facilitates binding of chlorophyll to hydrophobic regions

of chlorophyll-binding proteins. The CH 3 group (green) is replaced by a • 0

formaldehyde (CHO) group in chlorophyll b.

554 CHAPTER 12 • Cellular Energetics

light ultimately is used ro remove electrons from a donor Each Photon of Light Has a Defined
(water in the case of green plants), forming oxygen: Amount of Energy
Quantum mechanics established that light, a form of elec-
tromagnetic radiation, has properties of both waves and
particles. When light interacts with matter, it behaves as dis-
The electrons are transferred to a primary electron acceptor,
crete packets of energy (quanta) called photons. The energy
a quinone designated Q, which is similar to CoQ in mito-
of a photon, £, is proportional to the frequency of the light
chondria. In plants the oxidation of water takes place in a
wave:£= h-y, where his Planck's constant (1.58 X 10 34
multiprotein complex called photosystem II (PSII).
cal · s, or 6.63 X 10 34 ] • s) and-y is the frequency of the
light wave. It is customary in biology to refer to the wave-
Stage 2: Electron Transport and Generation of a Proton-Motive length of the light wave, A, rather than to its frequency, -y.
Force Electrons move from the quinone primary electron The two are related by the simple equation -y = c ..;.- X., where
acceptOr through a series of electron carriers until they reach cis the velocity of light (3 X 10 10 cm/s in a vacuum). Note
the ultimate electron acceptor, usually the oxidized form of that photons of shorter wavelength have higher energies.
nicotinamide adenine dinucleotide phosphate (NADP+), reduc- Also, the energy in 1 mol of photons can be denoted by E =
ing it to NADPH. The structure of NADP is identical to that Ne, where N is Avogadro's number (6.02 X 10 23 molecules
of NAD+ except for the presence of an additional phosphate or photons/mol). Thus
group. Both molecules gain and lose electrons in the same way
(see Figure 2-33). In plants the reduction of NADP takes place Nhc
in a complex called photosystem I (PSl). The transport of elec- E~ = Nh-y = X.
trons in the thylakoid membrane is coupled to the movement of
protons from the stroma to the thylakoid lumen, forming a pH The energy of light is considerable, as we can calculate for
gradient across the membrane (pHiumcn < pHmom.l· This pro- light with a wavelength of 550 nm (550 X 10 -em), typical
cess is analogous to generation of a proton-motive force across of sunlight:
the inner mitochondrial membrane and in bacterial membranes
during electron transport (see Figure 12-23 ). (6.02 X 10 23 photons/mol)(1.58 X 10- 34caJ.s)(3 X 10 10cm/s.
Thus the overall reaction of stages 1 and 2 can be
E= ., -
550 X 10 em
summarized as = 51,881 caUmol

2 H 20 + 2 NADP li~hr) 2 H + + 2 NADPH + 0 2 or about 52 kcal/mol. This is enough energy to synthesize

several moles of ATP from ADP and P; if all the energy were
Stage 3: Synthesis of ATP Protons move down their concen- used for this purpose.
tration gradient from the thy lakoid lumen to the stroma
through the F0 F1 complex (ATP synthase), which couples pro-
ton movement to the synthesis of ATP from ADP and P,. The Photosystems Comprise a Reaction Center
chloroplast ATP synthase works similarly to the synthases of and Associated Light-Harvesting Complexes
mitochondria and bacteria (see Figures 12-26 and 12-27). The absorption of light energy and its conversion into chem-
ical energy occurs in multiprotein complexes called photo-
Stage 4: Carbon Fixation The NADPH and ATP generated systems. Found in all photosynthetic organisms, both
by the second and third stages of photosynthesis provide the eukaryoric and prokaryotic, photosystems consist of two
energy and the electrons to drive the synthesis of polymers of closely linked components: a reactwn center, where rhe pn-
six-carbon sugars from C0 2 and H 2 0. The overall chemical mary events of photosynthesis-generation of high-energy
equation is written as electrons-occur; and an antenna complex consisting of nu-
merous protein complexes, mcluding internal antenna pro-
6 C0 2 + 18 ATP 4 + 12 NADPH + 12 H 20 ~ teins within the photosystem proper and external antenna
C 6 H 12 0 6 + 18 ADP3 + 18 P,2 + 12 NADP + 6 H- complexes termed light-harvesting complexes (LHCs), made
up of specialized proteins which capture light energy and
The react ions that generate the ATP and NADPH used in transmit it to the reaction center (see Figure 12-32).
carbon fixation are directly dependent on light energy; thus Both reaction centers and antennas contain tightly bound
stages 1- 3 are called the light reactions of photosynthesis. light-absorbing pigment molecu les. Chlorophyll a i~ the
The reactions in stage 4 are indirectly dependent on light principal pigment involved in photosynthesis, being present
energy; they are sometimes called the dark reactiOns of pho- in both reaction centers and antennas. In addition to chloro-
tosynthesis because they can occur in the dark, utilizing the phyll a, antennas contain other light-absorbing pigments:
supplies of ATP and NADPH generated by light energy. chlorophyll b in vascular plants and carotenoids in both
However, the reactions in stage 4 are not confined to the plants and photosynthetic bacteria. Carotenoids consist of
dark; in fact, they occur primarily during illumination. long branched hydrocarbon chains with alternating single

12.5 Photosynthesis and Light-Absorbing Pigments 555

 Action spectrum phyll a is bound in the unique protein environment of the
of photosynthesis reaction center, dissipation of excited-state energy occurs by a

100 Chlorophyll a

Chlorophyll b
\I :0
(I)
quite different process that is the key to photosynthesis.

?i
;::·
80 (I)
Photoelectron Transport from Energized
c
w
.... Reaction-Center Chlorophyll a Produces
(I)
0
·~ 60 Q. a Charge Separation
"0
0 ~ The absorption of a photon of light of wavelength -680 nm
"'
.0
<{
0
0 by one of the two "special-pair" chlorophyll a molecules in
40
"'
<:J
,.. the reaction center increases the molecules' energy by 42 kcaU
~
(I) mol (the first excited state). Such an energized chlorophyll a
20 "'iii" molecule in a plant reaction center rapidly donates an electron
to an intermediate acceptor, and the electron is rapidly passed
0
on to the primary electron acceptor, quinone Q, near the stro-
400 500 600 mal surface of the thylakoid membrane (Figure 12-35). This
Wavelength {nm) light-driven electron transfer, called photoelectron transport,
depends on the unique environment of both the chlorophylls
EXPERIMENTAL FIGURE 12-34 The rate of photosynt hesis is
and the acceptor within the reaction center. Photoelectron
greatest at wavelengths of light abso rbed by t hree pigments. The
transport, which occurs nearly every time a photon is absorbed,
action spectrum of photosynthesis in plants {the ability of light of dif-
ferent wavelengths to support photosynthesis) is shown in black. The
leaves a positive charge on the chlorophyll a close to the lumi-
energy from light can be converted into ATP only if it can be absorbed
nal surface of the thylakoid membrane (opposite side from the
by pigments in the chloroplast. Absorption spectra {showing how well stroma) and generates a reduced, negatively charged acceptor
light of different wavelengths is absorbed) for three photosynthetic (Q- ) near the stromal surface.
pigments present in the antennas of plant photosystems are shown in The Q produced by photoelectron transport is a power-
color. Comparison of the action spectrum with the individual absorp- ful reducing agent with a strong tendency to transfer an elec-
tion spectra suggests that photosynthesis at 680 nm is primarily due to tron to another molecule, ultimately to NADP+. The positively
light absorbed by chlorophyll a; at 650 nm, to light absorbed by chloro- charged chlorophyll a , a strong oxidizing agent, attracts an
phyll b; and at shorter wavelengths, to light absorbed by chlorophylls a electron from an electron donor on the luminal surface to
and band by carotenoid pigments, including [3-carotene. regenerate the original chlorophyll a. In plants, the oxidizing
power of four chlorophyll a+ molecules is used, by way of
intermediates, to remove four electrons from 2 H 2 0 molecules
bound to a site on the luminal surface to form 0 2 :
and double bonds; they are similar in structure to the visual
pigment retinal, which absorbs light in the eye. The presence 2 H 20 + 4 chlorophyll a+~ 4 H + + 0 2 + 4 chlorophyll a
of various antenna pigments, which absorb light at different
wavelengths, greatly extends the range of light that can be These potent biological reductants and oxidants provide all
absorbed and used for photosynthesis. the energy needed to drive all subsequent reactions of photo-
One of the strongest pieces of evidence for the involve- synthesis: electron transport (stage 2), ATP synthesis (stage 3 ),
ment of chlorophylls and carotenoids in photosynthesis is and C0 2 fixation (stage 4).
that the absorption spectrum of these pigments is similar to Chlorophyll a also absorbs light at discrete wavelengths
the action spectrum of photosynthesis (Figure 12-34 ). The shorter than 680 nm (see Figure 12-34). Such absorption raises
latter is a measure of the relative ability of light of different the molecule into one of several excited states, whose ener-
wavelengths to support photosynthesis. gies are higher than that of the first excited state described
When chlorophyll a (or any other molecule) absorbs visi- above and which decay by releasing energy within 2 X 10- tl
ble light, the absorbed light energy raises electrons in the chlo- seconds (2 picoseconds, ps) to the lower-energy first excited
rophyll a to a higher-energy (excited) state. This state differs state with loss of the extra energy as heat. Because photoelec-
from the ground (unexcited) state largely in the distribution of tron transport and the resulting charge separation occur only
the electrons around the C and N atoms of the porphyrin ring. from the first excited state of the reaction-center chlorophyll a,
Excited states are unstable, and the electrons return ro the the quantum yield-the amount of photosynthesis pt:r ab-
ground state by one of several competing processes. For chlo- sorbed photon-is the same for all wavelengths of visible light
rophyll a molecules dissolved in organic solvents such as etha- shorter (a nd therefore of higher energy) than 680 nm. How
nol, the principal reactions that dissipate the excited-state closely the wavelength of light matches the absorption spectra
energy are the emission of light (fluorescence and phosphores- of the pigments determines how likely it is that the photon will
cence) and thermal emission (heat). When the same chloro- be absorbed. Once absorbed, the photon's exact wavelength is

556 CHAPTER 12 • Cellular Energetics

 @ FOCUS ANIMATION: Photosynthesis
FIGURE 12- 35 Photoelectron t ransport, the primary event in Primary electron Strong reducing
photosynthesis. After absorption of a photon of light, one of the Ligh Reaction acceptor agent (electron donor)
excited special pair of chlorophyll a molecules in the reaction center Stroma
(left) donates via several intermediates (not shown) an electron to
a loosely bound acceptor molecule, the quinone Q, on the stromal
surface of the thyla koid membrane, creating an essentially irreversible
charge separation across the membrane (right). Subsequent transfers
of this electron rel~d~~ energy that Is used to generate ATP and NADPH
(Figures 12-38 and 12-39). The positively charged chlorophyll a gener-
ated when the light-excited electron moves to Q is eventually neutral-
ized by the transfer to the chlorophyll a... of another electron. In plants lumen
the oxidation of water to molecular oxygen provides this neutralizing Chlorophyll a Strong Oxidizing
electron and takes place in a multiprotein complex called photosystem agent (electron acceptor)
II (Figure 12-39). The complex photosystem I uses a similar photoelec-
tron transport pathway, but instead of oxidizing water, it receives an
electron from a protein carrier called plastocyanin to neutralize the
positive charge on chlorophyll a (Figure 12-39).

not critical, provided it is at least energetic enough to push the then rapidly transferred (in <10 9 seconds) to one of the two
chlorophyll into the first excited state. "special-pair" chlorophyll a molecules in the associated re-
action center, where it promotes the primary photosynthetic
Internal Antenna and Light-Harvesting charge separation (Figure 12-35). Photosystem core proteins
and LHC proteins maintain the pigment molecules in the
Complexes Increase the Efficiency
precise orientation and position optimal for light absorption
of Photosynthesis and energy transfer, thereby max1mizing the very rapid and
Although chlorophyll a molecules within a reaction center efficient resonance transfer of energy from antenna pigments
'·
that are involved directly with charge separation and elec- to reaction-center chlorophylls. Resonance energy transfer
tron transfer are capable of directly absorbing light and ini- does not involve the transfer of an electron. Studies on one
tiating photosynthesis, they most commonly are energized of the two photosystems in cyanobacteria, which are similar
indirectly by energy transferred to them from other light- to those in multicellular, seed-bearing plants, suggest that
absorbing and energy-transferring pigments. These other energy from absorbed light is funneled first to a "bridging"
pigments, which include many other chlorophyll molecules, chlorophyll in each LHC and then to the special pair of
are involved with absorption of photons and passing the en- reaction-center chlorophylls (Figure 12-36a ). Surprisingly,
ergy to the chlorophyll a molecules in the reaction center. however, the molecular structures of LHCs from plants and
Some are bound to protein subunits that are considered to be cyanobacteria are completely different from those in green
intrinsic components of the photosystem and thus are called and purple bacteria, even though both types contain carot-
internal antennas; others are bound to protein complexes enoids and chlorophylls in a clustered arrangement within
that bind tO, but arc dist~nct from, the photosystem core pro- the membrane. Figure 12-36b shows the distribution of the
teins and are called light-harvesting complexes (LHCs). Even chlorophyll pigments in photosystem I from Pisum sativum
at the maximum light intensity encountered by photosyn- (garden pea) together with those from peripheral LHC an-
thetic organisms (tropical noontime sunlight), each reaction- tennas. The large number of internal and LHC antenna chlo-
·' center chlorophyl l a molecule absorbs only about one rophylls surround the core reaction center to permit efficient
photon per second, which is not enough to support photo- transfer of absorbed light energy to the special chlorophylls
synthesis sufficient for the needs of the plant. The involve- in the reaction center.
ment of internal antennas and LHCs greatly increases the Although LHC antenna chlorophylls can transfer light
efficiency of photosynthesis, especially at more typical light energy absorbed from a photon, they cannot rt>le;1se an elec-
intensities, by increasing absorption of 680-nm light and by tron. As we've seen already, this function resides in the two
extending the range of wavelengths of light that can be ab- reaction-center chlorophylls. To understand their electron-
sorbed by other antenna pigments. releasing abi lity, we examine the structure and function of
Photons can be absorbed by any of the pigment mole- the reaction center in bacterial and plant photosystems in the
cules in internal antennas or LHCs. The absorbed energy is next section.

12.5 Photosynthesis and Light-Absorbing Pigments 557

 (a) Reaction
Light

Stroma

Lumen\'------v------/
Thylakoid LHC Special-pair LHC
membrane chlorophylls

(b)

center
LHC LHC
FIGURE 12-36 Light-harvesting complexes and photosystems chlorophylls (squares, dark green) and thence to chlorophylls in the
in cyanobacteria and plants. (a) Diagram of the membrane of a reaction center. (b) Three-dimensional organization of the photosystem I
cyanobacterium, in which the multiprotein light-harvesting complex (PSI) and associated LHCs of Pisum sativum (garden pea), as determined
(LHC) contains 90 chlorophyll molecules (green) and 31 other small by x-ray crystallography and seen from the plane of the membrane.
molecules, all held in a specific geometric arrangement for optimal Only the chlorophylls together with the reaction-center electron
light absorption and energy transfer. Of the six chlorophyll molecules carriers are shown. (c) Expanded view of the reaction center from (b),
in the reaction center, two constitute the special-pair chlorophylls rotated 90° about a vertical axis. [Part (a) adapted from W. Ki.ihlbrandt, 2001,
(ovals, dark green) that can initiate photoelectron transport when Nature 411:896, and P. Jordan et al., 2001, Nature 411:909. Parts (b) and (c) based
excited (blue arrow). Resonance transfer of energy (red arrows) on the structural determination by A. Ben-Sham et al., 2003, Nature 426:630.]
rapidly funnels energy from absorbed light to one of two "bridging"

KEY CONCEPTS of Section 12.5 reaction-center proteins in the thylakoid membrane. The en-
ergized chlorophylls donate, via intermediates, an electron
Photosynthesis and Light-Absorbing Pigments to a quinone on the opposite side of the membrane, creating
• The principal end products of photosynthesis in plants are a charge separation (see Figure 12-35). In green plants, the
molecular oxygen and polymers of six-carbon sugars (starch positively charged chlorophylls then remove electrons from
and sucrose). water, forming molecular oxygen (0 2).
• The light-capturing and ATP-generating reactions of pho- • In stage 2, electrons are transported from the reduced qui-
tosynthesis occur in the thylakoid membrane located within none via carriers in the thylakoid membrane until they reach
chloroplasts. The permeable outer membrane and inner the ultimate electron acceptor, usually NADP-, reducing it
membrane surrounding chloroplasts do not participate di- to NADPH. Electron transport is coupled to movement of
rectly in photosynthesis (see Figure 12-31 ). protons across the membrane from the stroma ro the thyla-
• There are four stages in photosynthesis: (1) absorption of koid lumen, forming a pH gradient (proton-motive force)
light, generation of a high-energy electrons, and formation across the thylakoid membrane.
of 0 2 from H 2 0; (2) electron transport leading to reduction
ofNADP' to NADPH, and to generation of a proton-motive • In stage 3, movement of protons down their electrochemi-
force; (3) synthesis of ATP; and (4 ) conversion of C02 into cal gradient through F0 F 1 complexes (ATP synthase) powers
carbohydrates (carbon fixation). the synthesis of ATP from ADP and P,.

• In stage 1 of photosynthesis, light energy is absorbed by • [n stage 4, the NADPH and ATP generated in stages 2 and
one of two "special-pair" chlorophyll a molecules bound to 3 provide the energy and the electrons to drive the fixation

558 CHAPTER 12 • Cellular Energetics

of col, w hich results in the synthesis of carbohydrates.
These reactions occur in the thylakoid stroma and cytosol.
• Associated with each reaction center are multiple internal
antenna and light-harvesting complexes (LHCs), which con-
tain chlorophylls a and b, carotenoids, and other pigments
that absorb light at multiple wavelengths. Energy, but not an
electron, is transferred from the internal antenna and LHC
chlorophyll molecules to reaction-center chlorophyll s hy
resonance energy transfer (see Figure 12-36).

12.6 Molecular Analysis of Photosystems

As noted in the previous section, photosynthesis in the green
and purple bacteria does not generate oxygen, whereas pho-
tosynthesis in cyanobacteria, algae, and plants does.* This
difference is attributable to the presence of two types of pho-
tosystem (PS) in the latter organisms: PST reduces NADP+ to
NADPH, and PSII fo rms 0 2 from H 20. In contrast, the
green and purple bacteria have only one type of photosys-
tem, which cannot form 0 2 • We first discuss the simpler
photosystem of purple bacteria and then consider the more
complicated photosynthetic machinery in chloroplasts.
Special-pair
chlorophyll
The Single Photosystem of Purple Bacteria FIGURE 12-37 Three-dimensional structure of t he photosyn-
Generates a Proton-Motive Force but No 0 2 thetic reaction center from the purple bacterium Rhodobacter
spheroides. (Top) The l subunit (yellow) and M subunit (gray) each
The three-dimensional structures of photosynthetic reaction
form five transmembrane a helices and have a very similar structure
centers have been determined, permitting scientists to trace in
overall; the H subunit (light blue) is anchored to the membrane by
detail the paths of electrons during and after the absorption
a single transmembrane Ct. helix. A fourth subunit (not shown) is a
of light. The reaction center of purple bacteria contains three peripheral protein that binds to the exoplasmic segments of the
protein subunits (L, M, and H) located in the plasma mem- other subunits. (Bottom) Within each reaction center, but not easily
brane (Figure 12-37). Bound to these proteins are the pros- distinguished in the top image, is a specia l pair of bacteriochlorophyll
thetic groups that absorb light and transport electrons during a molecules (green), capable of initiating photoelectron transport;
photosynthesis. The prosthetic groups include a "special two accessory chlorophylls (purple); two pheophytins (dark blue), and
pair" of bacteriochlorophyll a molecules equivalent to the two qui nones, QA and Q 8 (orange). Q 8 is the primary electron acceptor
reaction-center chlorophyll a molecules in plants, as well as during photosynthesis. [After M. H. Stowell et at., 1997, Science 276:812.]
several other pigments and two quinones, termed QA and Q 8,
that are structurally similar to mitochondrial ubiquinone.
chlorophyll thereby acquires a positive charge, and Qll ac-
Initial Charge Separatio~ The mechanism of charge separa- quires a negative charge. To determine the pathway tra-
tion in the photosystem of purple bacteria is identical to that versed by electrons through the bacterial reaction center,
in plants outlined earlier; that is, energy from absorbed light researchers exploited the fact that each pigment absorbs
is used to strip an electron from a reaction-center bacterio- light of only certain wavelengths, and its absorption spec-
chlorophyll a mo lecule and transfer it, via several different trum changes when it possesses an extra electron. Because
pigments, to the primary electron acceptor Q 6 , which is these electron movements are completed in less than 1 mil-
loosely bound to a site on the cytosolic membrane face. The lisecond (ms), a special technique called picosecond absorp-
tion spectroscopy is required to mon itor the changes in the
absorption spectra of the various pigments as a function of
• A verr different type of mechanism used to harvest the energ} of light, time shortly after the absorption of a light photon.
which occurs only in certain archaebacteria, IS not discussed here because
it is very different from reaction center mechanisms described here. In this
When a preparation of bacterial membrane vesicles is ex-
other mechanism, the plasma-membrane protein that absorbs a photon of posed to an intense pulse of laser light lasting less than 1 ps,
light, called bacteriorhodopsin, also pumps one proton from the cytosol each reaction center absorbs one photon (Figure 12-38). Light
to the extracellular space for every photon of light absorbed. absorbed by the chlorophyll a molecules in each reactton center

12.6 Molecular Analysis of Photosystems 559

 .·
P,+
Q cycle: additional ADP ATP

2W proton transport H

Cytosol

Plasma
membrane
p

++++
Periplasmic
space

Bacterial reaction F0 F1 complex

center

FIGURE 12-38 Cyclic electron flow in the single photosystem cytosol to form QH 2• (Center) After diffusing through the membrane
of purple bacteria. Cyclic electron flow generates a proton-motive and binding to the Oo site on the exoplasmic face of the cytochrome
force but no 0 2• Blue arrows indicate flow of electrons; red arrows be 1 complex, QH 2 donates two electrons and simultaneously gives
indicate proton movement. (Left) Energy absorbed directly from light up two protons to the external medium in the periplasmic space,
or funneled from an associated LHC (not illustrated here) energizes one generating a proton electrochemical gradient (proton-motive force).
of the special-pair chlorophylls in the reaction center. Photoelectron Electrons are transported back to the reaction-center chlorophyll via a
transport from the energized chlorophyll, via an accessory chlorophyll, soluble cytochrome, which diffuses in the periplasmic space. Note the
pheophytin (Ph), and quinone A (QA), to quinone B (Q8) forms the semi- cyclic path (blue) of electrons. Operation of a Q cycle in the cytochrome
quinone o·- and leaves a positive charge on the chlorophyll. Following be 1 complex pumps additional protons across the membrane to the
absorption of a second photon and transfer of a second electron to external medium, as in mitochondria. [Adapted from J. Deisenhofer and
the semiquinone, the quinone rapidly picks up two protons from the H. Michael, 1991, Ann. Rev. Cell Bioi. 7:1.]

converts them to the excited state, and the subsequent electron two protons into the periplasmic space (the space between
transfer processes are synchronized in all reaction centers in the the plasma membrane and the bacterial cell wall). This pro-
experimental sam'ple. Within 4 X I 0 11 seconds (4 ps), an elec- cess moves protons from the cytosol to the outside of the
tron moves, via the accessory bacterial chlorophyll (sec Figure cell, generating a proton-motive force across the plasma
12-37, bottom) as an intermediate, to the pheophytin molecules membrane. Simultaneously, QH2 releases its two electrons,
(Ph), leaving a positive charge on the chlorophyll a. This state which move through the cytochrome be 1 complex exactly
exists for about 200 ps before the electron moves to QA, and as depicted for the mitochondrial comp lex III (CoQHz-
then, in the slowest step, it takes -200 f.LS for it to move to QB. cytochrome e reductase) in Figure 12-18. The Q cycle in the
This pathwa} of electron flow is traced in the left parr of Figure bacterial reaction center, like the Q cycle in mitochondria,
12-3 8. The later steps are slower than inherently rapid electron pumps additional protons from the cytosol to the in termem-
movements because they involve relatively slow protein confor- brane space, thereby increasing the proton-motive force.
mational changes. The acceptor for electrons transferred through the cyto-
chrome be 1 complex is a soluble cytochrome, a one-electron
Subsequent Electron Flow and Coupled Proton Move- carrier, in the periplasmic space, which is reduced from the
ment After the primary electron acceptor, Ql\, in the bac- FeH to the Fe2 + state. The reduced cytochrome (analogous
terial reaction center accepts one electron, forming Qs' , it to cytochrome c in mitochondria) then diffuses to a reaction
accepts a second electron from the same reaction-center center, where it releases irs electron to a positively charged
chlorophyll following irs re-cxcitation (e.g., by absorption of chlorophyll a~, returning that chlorophyll to the uncharged
a second photon or transfer of energy from antenna mole- ground state and thf' cytochrome to the Fe H state. This cyclic
cules) . l he quinone then binds two protons from the cyto- electron flow generates no oxygen and no reduced coenzymes,
sol, forming the reduced quinone (QH2 ), which is released but it has generated a proton-motive force.
from the reaction center (Figure 12-38). QH2 diffuses within As in other systems, this proton-motive force is used by
the bacterial membrane to the Q 0 site on the exoplasmic face the F0 F 1 complex located in the bacterial plasma membrane
of a cytochrome be 1 electron transport complex similar in to synthesize ATP and also to transport molecules across the
structure to complex Ill in mitochondria. There it releases its membrane against a concentration gradient.

560 CHAPTER 12 • Cellular Energetics

Chloroplasts Contain Two Functionally PSII, especially LCHII. Evidence for this distribution came
and Spatially Distinct Photosystems from studies in which thylakoid membranes were gently
fragmented into vesicles by ultrasound. Stacked and un-
In the 1940s, biophysicist R. Emerson discovered that the stacked thylakoid vesicles were then fractionated by density-
rate of plant photosynthesis generated by light of wavelength gradient centrifugation. The stacked fractions contained
700 om can be greatly enhanced by adding light of shorter primarily PSII protein and the unstacked fractions PSI.
wavelength (higher energy). He found that a combination of Finally, and most importantly, the two chloroplast photo-
light at, say, 600 and 700 nm supports a greater rate of pho- systems differ significantly in their functions (Figure 12-39):
tosynthesis than the sum of the rates for the two separate only PSII oxidi7es water to form molecular oxygen, whereas
wavelengths. This so-called Emerson eflect led researchers to only PSI transfers electrons to the final electron acceptor,
conclude that photosynthesis in plants involves the interac- NADP . Photosynthesis in chloroplasts can follow a linear
tion of two separate photosystems, referred to as PSI and or cyclic pathway. The linear pathway, which we discuss
PSlf. PSI is driven by light of wavelength 700 om or less; first, can support carbon fixation as well as ATP synthesis.
PSII, only by shorter-wavelength light (<680 om). In contrast, the cyclic pathway supports only ATP synthesis
In chloroplasts, the special-pair reaction-center chloro- and generates no reduced NADPH for use in carbon fixa-
phylls that initiate photoelectron transport in PSI and PSII tion. Photosynthetic algae and cyanobacteria contain two
differ in their light-absorption maxima because of differences photosystems analogous to those in chloroplasts. Similar pro-
in their protein environments. For this reason, these chloro- teins and pigments compose photosystems I and II of plants
phylls are often denoted P680 (PSU) and P 700 (PSI). Like a bac- and photosynthetic bacteria.
terial reaction center, each chloroplast reaction center is
associated with multiple internal antenna and light-harvesting
complexes (LHCs); the LHCs associated with PSII (e.g., LHCU) Linear Electron Flow Through Both Plant
and PSI (e.g., LHCI) contain different proteins.
Photosystems, PSII and PSI, Generates
The two photosystems also are distributed differently in
thylakoid membranes: PSil primarily in stacked regions a Proton-Motive Force, 0 2, and NADPH
(grana, see Figure 12-31 ) and PSI primarily in unstacked re- Linear electron flow in chloroplasts involves PSII and PSI in
gions. The stacking of the thylakoid membranes may be due an obligate series in which electrons are transferred from
to the binding properties of the proteins associated with H 20 to NADP+. The process begins with absorption of a

NADP" + W

P,+
0 cycle: additional NADPH ADP ATP
2W proton transport H

Stroma

Thylakoid
membrane

++++
Lumen
Psao p700
chlorophyll chlorophyll
H20 F0 F1 complex
PSII reaction center center

FIGURE 12- 39 Linear electron flow in plants, which requires additional protons across the membrane to the thylakoid lumen,
both chloroplast photosystems PSI and PSII. Blue arrows indicate increasing the proton-motive force. (Right) In the PSI reaction center,
flow of electrons; red arrows indicate proton movement. LHCs are not each electron released from light-excited P700 chlorophylls moves via
shown. (Left) In the PSII reaction center, two sequential light-induced a series of carriers in the reaction center to the stromal surface, where
excitations of the same P680 chlorophylls result in reduction of the primary soluble ferredoxin (an Fe-S protein) transfers the electron to ferredoxin-
electron acceptor Q8 to QH 2• On the luminal side of PSII, electrons NADP+ reductase (FNR). This enzyme uses the prosthetic group flavin
removed from H20 in the thylakoid lumen are transferred to P680 , adenine dinucleotide (FAD) and a proton to reduce NADP~, forming
restoring the reaction-center chlorophylls to the ground state and gen- NADPH. P700 + is restored to its ground state by addition of an electron
erating 0 2• (Center) The cytochrome bfcomplex then accepts electrons carried from PSI I via the cytochrome bf complex and plastocyanin, a
from QH 2, coupled to the release of two protons into the lumen. soluble electron carrier.
Operation of a Q cycle in the cytochrome bf complex translocates

12.6 Molecular Analysis of Photosystems 561

photon by PSII, causing an electron to move from a P680 PSII reaction
center
chlorophyll a to an acceptor plastoquinone (Q 8 ) on the stro-
mal surface (Figure 12-39 ). The resulting oxidized P6 Ro +
strips one electron from the relatively unwilling donor H 20, Stroma
forming an intermediate in 0 2 formation and a proton,
which remains in the thylakoid lumen and contributes to the Thylakoid
proton-motive force. After P680 absorbs a second photon, the membrane
semiquinone Q' accepts a second electron and picks up two
protom from the stromal space, generating QH 2• After dif-
4 photons
fusing in the membrane, QH 2 binds to the Q 0 site on a cyto-
chrome bf complex that is analogous to the bacterial
Thylakoid
cytochrome bc1 complex and to the mitochondrial complex lumen
III. As in these systems, a Q cycle operates, thereby increas- 0 2 ·evolving
complex
ing the proton-motive force generated by electron transport. }
After the cytochrome bf complex accepts electrons from
QH 2, it transfers them, one at a time, to the Cu2 form of the
soluble electron carrier plastocyanin (analogous to cyto-
chrome c), reducing it to the Cu 1 f form. Reduced plastocya- FIGURE 12-40 Electron flow and 0 2 evolution in chloroplast PSII.
nin then diffuses in the thylakoid I umen, carrying the electron The PSII reaction center, comprising the two integral proteins 01 and
to PSI. 02, special-pair chlorophylls (P680), and other electron carriers, is associ-
Absorption of a photon by PSI leads to removal of an ated with an oxygen-evolving complex on the luminal surface. Bound
electron from the reaction-center chlorophyll a, P-00 (Figure to the three extrinsic proteins (33, 23, and 17 kOa) of the oxygen-
12-39 ). The resu lting oxidized P-,00 + is reduced by an elec- evolving complex are four manganese ions (Mn, red)_. a Ca 1 - ion (blue),
tron passed from the PSII reaction center via the cytochrome and a Cl ion (yellow). These bound ions function in the splitting of H20
bf complex and plastocyanin. Again, this is ana logous to the and maintain the environment essential for high rat es of 0 1 evolution.
situation in mitochondria, where cytochrome c acts as a Tyrosine-161 (Y161) ofthe 01 polypeptide conducts electrons from the
single-electron shuttle from complex III to complex IV (see Mn ions to the oxidized reaction-center chlorophyll (P680 -). reducing it
to the ground state P680• [Adapted from C. Hoganson and G. Babcock, 1997,
Figure 12-16). The electron taken up at the luminal surface
Science 277:1953.)
by the P-00 energized by photon absorption moves within PSI
via several carriers to the stromal surface of the thylakoid
membrane, where it is accepted by ferredoxin, an iron-sulfur
(Fe-S) protein. In linear electron flow electrons excited in PSI
are transferred from ferredoxin via the enzyme ferredoxin- two proteins in PSII, called 01 and D2, whose sequences are
NADP reductase (FN R). This enzyme uses the prosthetic remarkably similar to the sequences of the Land M subunits
group FAD as ari. electron carrier to reduce NADP+, form- of the bacterial reaction center (Figure 12-37), attesting to
ing, together with one proton picked up from the stroma, the their common evolutionary origins . When PSII absorbs a
reduced molecule NADPH. photon with a wavelength of <680 nm, it triggers the loss of
F0 F1 complexes in the thylakoid membrane use the proton- an electron from a P6 80 molecule, generating P 680 "'".As in
motive force generated during linear electron flow to synthe- photosynthetic purple bacteria, the electron is transported rap-
size ATP on the stromal side of the membrane. Thus this idly, probably via an accessory chlorophyll, to a pheophytin,
pathway exploits the energy from multiple photons absorbed then to a quinone (QA), and then to the primary electron ac-
by both PSII and PSI and their antennas to generate both ceptor, Q 8 , on the outer (stromal) surface of the thylakoid
NADPH and ATP in the stroma of the chloroplast, where membrane (Figures 12-39 and 12-40 ).
they are utili zed for C0 2 fixation. The photochemically oxidized reaction-center chloro-
phyll of PSII, P680 , is the strongest biological oxidant known.
The reduction potential of P680 is more positive than that of
water, and thus it can oxidize water to generate 0 2 and H
An Oxygen-Evolving Complex Is Located on the
ions. Photosynthetic bacteria cannot oxidize water because
Luminal Surface of the PSII Reaction Center the excited chlorophyll a+ in the bacterial reaction center is
Somewhat surprisingly, the structure of the PSII reaction not a sufficiently strong oxidant. Thus they use other sources
center, which removes electrons from H 2 0 to form 0 2 , rt>- of electrons, such :J.S H 2 S and H 2 •
sembles that of the reaction center of photosynthetic purple The oxidation of H 20, which provides the electrons for
bacteria, which does not form 0 2 . Like the bacterial reaction reduction of P680 in PSII, is catalyzed by a three-protein
center, the PSII reaction center contains two molecules of complex, the oxygen-euolving complex, located on the luminal
chlorophyll a (P680 ), as well as two other accessory chloro- surface of PSII in the thylakoid membrane. The oxygen-
phylls, two pheophytins, two quinones (QA and Q 8 ), and evolving complex contains four manganese (M n) ions con-
one nonheme iron atom. These small molecules are bound to nected by bridging oxygen atoms, as well as bound Cl and

562 CHAPTER 12 • Cellular Energetics

'·

t;';fl Herbicides that inhibit photosynthesis not only are

~ ,·ery important in agriculture but also have proved use-
~
ful in dissecting the pathway of photoelectron transport in
"'
"'
;;::: plants. One such class of herbicides, the s-triazincs (e.g., at-
Qi razine ), binds specifically to the Dl subunit in the PSII reac-
c.
"'C
Q) tion center, thus inhibiting binding of oxidized Q 8 to its site
~
0 on the stromal surface of the thylakoid membrane. When
>
Q) added to illuminated chloroplasts, s-triazines cause all down-
0 c;rream electron carriers to accumulate in the oxidized form,
because no electrons can be released from PSII. In atrazine-
resistant mutants, a single amino acid change in Dl renders
2 3 4 5 6 7 8 9 10 11 12 it unable to bind the herbicide, so photosynthesis proceeds at
Flash number
normal rates. Such resistant weeds are prevalent and present
.' EXPER MENTA FIGURE 2--< A single PSII absorbs a photon a major agricultural problem. •
and transfers an electron four times to generate one 0 2 • Dark-
adapted chloroplasts were exposed to a series of closely spaced, short
(5 fJ.S) pulses of light that activated virtually all the PSIIs in the prepa-
Multiple Mechanisms Protect Cells Against
ration. The peaks in 0 2 evolution occurred after every fourth pulse,
indicating that absorption of four photons by one PSII is required to Damage from Reactive Oxygen Species
generate each 0 2 molecule. Because the dark-adapted chloroplasts During Photoelectron Transport
were initially in a partially reduced state, the peaks in 0 2 evolution
As we saw earlier in the case of mitochondria, ROS generated
occurred after flashes 3, 7, and 11. [From J. Berget al., 2002, Biochemistry,
during electron transport through the electron transport chain
5th ed., W. H. Freeman and Company.]
(see Figure 12-21) can both serve as signals to regulate organ-
elle function and cause damage to a variety of biomolecules.
The same is true for chloroplasts. Even though the PSI and
PSII photosystems with their associated light-han·estmg com-
Ca 2 + ions (Figure 12-40); this is one of the very few cases in plexes are remarkably efficient at converting radiant energy to
which manganese plays a role in a biological system. These useful chemical energy in the form of ATP and NADPH, they
manganese ions together with the three extrinsic proteins are not perfect. Depending on the intensity of the light and the
can be removed from the reaction center by treatment with physiologic conditions of the cells, a relatively small-but
solutions of concentrated salts; this abolishes 0 2 formation significant-amount of energy absorbed by chlorophylls in
but does not affect light absorption or the initial stages of the light-harvesting antennas and reaction centers results in
electron transport. the chlorophyll being converted to an activated state called
The oxidation of two molecules of H 20 to form 0 2 re- "triplet" chlorophyll. In this state, the chlorophyll can trans-
quires the removal of four electrons, but absorption of each fer some of its energy to molecular oxygen (0 2 ), converting it
photon by PSI! results in the transfer of just one electron. A from its normal, relatively unreactive ground state, called trip-
simple experiment, described in Figure 12-41, resolved whether let oxygen eo 2 ), to a very highly reactive (ROS ) singlet state
the formation of 0 2 depends on a single PSII or multiple ones form, 10 2 • Some of this 1 0 2 can be used for signalmg to the
acting in concert. The results indicated that a single PSII must nucleus to communicate the metabolic state of the chloroplast
lose an electron and then oxidize the oxygen-evolving complex to the rest of the cell. However, if the majority of the 10 2 is
four times in a row for an 0 2 molecule to be formed. not quickly quenched by reacting with specialized 10 2 "scav-
Manganese is known to exist in multiple oxidation states enger molecules," it will react with and usually damage nearby
with from two to five pasitive charges. Indeed, spectroscopic molecules. This damage can suppress the efficiency of thyla-
studies showed that the bound Mn ions in the oxygen-evolving koid activity and is called photoinhibition. Carotenoids (poly-
complex cycle through five different oxidation states, S0-S4 • In mers of unsaturated isoprene groups, including beta-carotene,
this S cycle, a total of two H 20 molecules are split to generate which gives carrots their orange color) and a.-tocopherol (a
four protons, four electrons, and one 0 2 molecule. Channels in form of vitamin E) are hydrophobic small molecules that play
the structure of the oxygen-evolving complex have been pro- important roles as 10 2 quenchers to protect plants. For exam-
posed to serve as conduits for the delivery of H 20 to and the ple, inhibition of tocopherol synthesis in the unicellular green
removal of 0 2 from the active site through the surrounding alga Chlamydomonas reinhardtii by the herbicide pyrazo-
protein of the oxygen-evolving complex. The electrons released lynate can result in greater light-induced photoinhibition. The
from H 20 are transferred, one at a time, via the Mn ions and carotenoids, which very efficiently siphon off energy from the
a nearby tyrosine side chain on the Dl subunit to the reaction- dangerous triplet chlorophyll when they are in close proxim-
center P680 +, where they regenerate the reduced chlorophyll, ity, are the quantitatively most important molecules for pre-
P680 , ground state by replacing the electron that was removed venting 10 2 formation. There are about 11 carotenoid
by light absorption. The protons released from H 20 remain in molecules and 35 chlorophylls in the PSII monomer from the
the thylakoid lumen. cyanobacterium Thermosynechococcus elongatus.

12.6 Molecular Analysis of Photosystems 563

 Photoinhibition Recovery a protease and replaced by newly synthesized D 1 protem 111
(2400 11E m 1 s 1 ) (20 11E m-1 s 1)
what is called the Dl protein damage-repair cycle. The rapid
0.8 replacement of damaged Dl, which requires a high rate of
01 synthesis, helps the PSII recover from photoinactivation
and maintain sufficient activity. The experiment in Figure
12-42 shows that an important component in the damage-
0.7
repair cycle is the chaperone protein HSP70B (see Chapter 3),
which binds to the damaged PSII and helps prevent loss of
the odu::r components of the complex as the Dl subunit is
0.6 replaced. The extent of photoinhibition can depend on the
amount of HSP70B available to the chloroplasts.

0.5 Cyclic Electron Flow Through PSI Generates

~
a Proton-Motive Force but No NADPH or 0 2
·:;:
·~ As we've seen, electrons from reduced ferredoxin in PSI are
0.4
="'
(/)
transferred to NADP during linea-r electron flow, resulting
in production of NADPH (see Figure 12-39). In some cir-
a..
cumstances, such as drought, high light intensity, or low carbon
0.3 dioxide levels, cells must generate relatively greater amounts
of ATP relative to NADPH than that produced by linear
electron flow_ To do this, they photosynthetically produce
0.2
ATP from PSI without concomitant NADPH production.
This is accomplished using a PS!I-independent process called
cyclic photophosphorylation, or cyclic electron flow_ In this
process electrons cycle between PSI, ferredoxin, plastoqui-
0.1 none (Q), and the cytochrome bf complex (Figure 12-43 );
thus no net NADPH is generated, and there is no need to
oxidize water and produce 0 2 _ There are two distinct cyclic
0.0~--~-----L----~----~--~-----L----~ electron flow pathways: the NAD(P)H dehydrogenase (Ndh)-
-60 -30 0 30 60 90 120 150 dependent (shown in Figure 12-43) and Ndh-independent
Time (min) pathways. Ndh is an enzyme complex, very similar to the
EXPERIMENTAL FIGURE 12-42 The chaperone HSP708 helps mitochondrial complex I (see Figure 12-16), that oxidizes
PSII recover from p ~oto i nh ibition after exposure to intense light. NADPH or NADH while reducing Q to QH 2 and thus con-
The unicellular green alga Chlamydomonas reinhardtii was geneti- tributes to the proton motive force by transporting protons.
cally manipulated so that it had abnormally high or low levels of the During cyclic electron flow, the substrate for the Ndh is the
chaperone protein HSP70B. The high, low, and normal strains were NADPH generated by light absorption by PSI, ferredoxin,
2 1
then exposed to high-intensity light (2400 1.1-E m s ) for 60 minutes and ferredoxin-NADP reductase (FNR)_ The QH2 formed by
to induce photoinhibition followed by exposure to low light (20 1.1-E Ndh then diffuses through the thylakoid membrane to the
m 2 s 1) for up to 150 minutes. The effects of photoinhibition by the Qo binding site on the luminal surface of the cytochrome bf
high-intensity light and the ability of PSII to recover from the photoin- complex. There it releases two electrons to the cytochrome
hibition were measured using fluorescence spectroscopy to determine
bf complex and two protons to the thylakoid lumen, gener-
PSII activity. The ability of the cells to recover PSII activity depends on
ating a proton-motive force. As in linear electron flow, these
the levels of HSP70B-the more HSP70B available, the more rapid the
electrons return to PSI via plastocyanin. This cyclic electron
recovery-due to HSP70B protection of the PSII reaction centers that
had withstood 10 2 -induced 01 subunit damage. [From Schroda et al.,
flow is similar to the cyclic process that occurs in the single
1999, Plant Ce// 11 :1165.]
photosystem of purple bacteria (sec Figure 12-38). A Q cycle
operates in the cytochrome bf complex during cyclic electron
flow, leading to transport of rwo additional protons into the
lumen for each pair of electrons transported and a greater
Under intense illumination, photosystem PSII is espe- proton-motive force.
cially prone to generating 10 2 , whereas PSI will produce Tn Ndh-independenr cyclic electron flow, the mechanism
other ROS, including superoxide, hydrogen peroxide, and of which has not yet been completely defined, electrons
hydroxyl radicals. The D1 subunit in the PSII reaction center from the ferredoxin are used to reduce Q, either via a hypo-
(see Figure 12-40 ) is, even under low light conditions, sub- thetical membrane-associated ferredoxin:plastoquinone
jected to almost constant 10rmediated damage. A damaged oxidoreductase (FQR) or via the Q, site, which is part of the
reaction center moves from the grana to the unstacked re- Q cycle in the cytochrome bf complex. Genetic analysis of
gions of the thylakoid, where the D 1 subunit is degraded by Arabidopsis thaliana has identified several genes involved in

564 CHAPTER 12 • Cellular Energetics

 NADP+H'

( Q cycle: additional~ NADPH

P, +
ADP ATP
NADPH NADP + W proton transport H'

Stroma

Thylakoid
membrane

++++
Lumen
2W W p700
chlorophyll
NAD(P)H dehydrogenase Cytochrome bf PSI reaction F0 F1 complex
(Ndh) complex center

FIGURE 12-43 Cyclic electron flow in plants, which generates a fix carbon-is oxidized by Ndh. The released electrons are transferred
proton-motive force and ATP but no oxygen or net NADPH.In the to plastoquinone (Q) within the membrane to generate QH2, which
NAD(P)H-dehydrogenase (Ndh)-dependent pathway for cyclic electron then transfers the electrons to the cytochrome bf complex, then to
flow, light energy is used by PSI to transport electrons in a cycle to plastocyanin, and finally back to PSI, as is the case for the linear elec-
generate a proton-motive force and ATP without oxidizing water. The tron flow pathway (see Figure 12-39).
NADPH formed via the PSI/ferredoxin/FNR-instead of being used to

Ndh-independent cyclic electron flow, including the integral associated kinase and an apparently constitutively active
membrane protein PGRL1. phosphatase. LHCIJ's unphosphorylated form is preferen-
tially associated with PSII, and the phosphorylated form dif-
fuses in the thylakoid membrane from the grana to the
Relative Activities of Photosystems I
unstacked region and associates with PSI more than the un-
and II Are Regulated phosphorylated form. Light conditions in which there is
In order for PSII, which is preferentially located in the preferential absorption of light by PSII result in the produc-
stacked grana, and PSI, which is preferentially located in the tion of high levels of QH 2 that bind to the cytochrome bf
unstacked thylakoid membranes, to act in sequence during complex (see Figure 12-39). Consequent conformational
linear electron flow, the amount of light energy delivered to changes in this complex are apparently responsible for acti-
the two reaction centers must be controlled so that each cen- vation of the LHCII kinase, increased LHCII phosphoryla-
ter activates the same number of electrons. This balanced tion, compensatory increased activation of PSI relative to
condition is called state 1 (Figure 12-44a ). If the two photo- PSII, and thus an increase in cyclic electron flow in state 2
systems are nor equally excited, then cyclic electron flow (Figure 12-44a). When the green alga Chlamydomonas rem-
occurs in PSI and PSIT becomes less active (state 2). Varia- hardtii was forced into state 2, it was possible to isolate a
tions in the wavelengths and intensities of ambient light (as "super-supercomplex" containing PSI, LHCI, LHCII, Cyt
a consequence of the tim~ of day, cloudiness, etc.) can change bf, ferredoxin (Fd), NADPH oxidoreductase (FNR), and the
the relative activation of the two phorosystems, potentially integral membrane protein PGRL 1 that participates in Ndh-
upsetting the appropriate relative amounts of linear and cyclic independent cyclic electron flow (Figure 12-44b). Thus it ap-
electron flow necessary for production of optimal ratios of pears that the efficient operation of electron transport chains
ATP and NADPH. has involved the evolution of functional complexes of in-
One mechanism for regulating the relative contributions creasing size and complexity, from individual proteins to
of PSI and PSII, in response to varying lighting conditions complexes to supercomplexes to super-supercomplcxes.
and thus the relative amounts of linear and cyclic electron Regu lating the supramolecular organization of the pho-
flow, entails redistributing the light-harvesting complex rosystems in plants has the effect of directing them toward
LHCII between the two photosystems. The more LHCII as- ATP production (state 2) or toward generation of reducing
sociated with a particular phorosystem, the more efficiently equivalents (NADPH) and ATP (state 1 ), depending on am-
that system will be activated by light and the greater its con- bient light conditions and the metabolic needs of the plant.
tribution to electron flow. The distribution of LHCII be- Both NADPH and ATP are required to convert C02 to su-
tween PSI and PSII is mediated by reversible phosphorylation crose or starch, the fourth stage in photosynthesis, which we
and dephosphorylation of LHCII by a regulated, membrane- cover in the last section of this chapter.

12.6 Molecular Analysis of Photosystems 565

 (a)
State 1, linear electron flow

Carbohydrate
PSII membrane domains synthesis
(stacked)

Stroma

Lumen
synthase

State 2, cyclic electron flow

No carbohydrate
ATP synthesis
Thylakoid membrane

Stroma

Lumen ATP
synthase
Super-supercomplex

FIGURE 12-44 Phosphorylation of LHCII and the regulation of PSII, and diffuses into the unstacked membranes, where it associates
linear versus cyclic electron flow. (a, top) In normal sunlight, PSI with PSI and its permanently associated LHCI. In this alternative
and PSII are equally activated, and the photosystems are organized supramolecular organization (state 2), most of the absorbed light
in state 1. In this arrangement, light-harvesting complex II (LHCII) energy is transferred to PSI, supporting cyclic electron flow and ATP
is not phosphorylated and six copies of LHCII trimers together with production but no formation of NADPH and thus no C0 2 fixation.
several other light-harvesting proteins encircle a dimeric PSI I reaction (b) Model of a PSI "super-supercomplex" involved with Ndh-
center in a tightly associated supercomplex in the grana (for clarity, independent cyclic electron flow that was isolated from green algae
molecular details of the supercomplexes not shown). As a result, PSII in stage 2. The super-supercomplex contains multiple complexes,
and PSI can function in parallel in linear electron flow. (a, bottom) including the integral membrane protein PGRL 1 that was identified
When light excitation of the two photosystems is unbalanced (e.g., by genetic analysis. [Adapted from F. A. Wollman, 2001, EMBO J. 20:3623;
too much via PSII), ~HCII becomes phosphorylated, dissociates from and M.lwai, et al., 2010, Nature 464:1210-1213.]

KI=Y CONCEPTS of Section 12.6 system, photochemically oxidized P680 + in PSII is regenerated to
P680 by electrons derived from the evolution of 0 2 from H 2 0
Molecular Analysis of Photosystems (see Figure 12-39, left).
In the single photosystem of purple bacteria, cyclic elec- In linear electron flow, photochemically oxidized P 700 in
tron flow from light-excited, special-pair chlorophyll a mol- PSI is reduced, regenerating P- 00 , by electrons transferred
ecules in the reaction center generates a proton-motive force, from PSII via the cytochrome bf complex and soluble plasto-
which is used mainly to power ATP synthesis by the F0F 1 cyanin. Electrons released from P-00 following excitation of
complex in the plasma membrane (see Figure 12-38). PSI are transported via several carriers ultimately to NADP~,
• Plants contain two photosystems, PSI and PSII, which generating NADPH (see Figure 12-39, right).
have different functions and are physically separated in the The absorption of light by pigments in the chloroplast can
thylakoid membrane. PSII converts H 20 into 0 2, and PSI generate reactive oxygen species (ROS ), including singlet
reduces NADP to NADPH. Cyanobacteria have two analo- oxygen, 10 2, and hydrogen peroxide, H 20 2 • In small
gous phorosystems. amounts they are not toxic and are used as intracellular sig-
naling molecules to control cellul ar metabolism. In larger
• In chloroplasts, light t::nergy absorbed by light-harvesting
amounts they can be toxic. Small molecule scavengers and
complexes (LHCs) is transferred to chlorophyll a molecules
antioxidant enzymes help to protect against ROS-induced
in the reaction centers (P680 in PSII and P- 00 in PSI).
damage; however, singlet oxygen damage to the Dl subunit
Electrons flow through PSII via the same carriers that are of PSII still occurs, causing photoinhibition. An HSP70
present in the bacterial photosystem. In contrast tO the bacterial chaperone helps PSII recover from the damage.

566 CHAPTER 12 • Cellular Energetics

 identical small subunits. One subunit is encoded in chloroplast
• In contrast to linear electron flow, which requires both DNA; the other, in nuclear DNA. Because the catalytic rate of
PSII and PSI, cyclic electron flow in plants involves only PSI. rubisco is quite low, many copies of the enzyme are needed
In this pathway, neither net NADPH nor 0 2 is formed al- to fix sufficient C02 . Indeed, this enzyme makes up almost
though a proton-motive force is generated. Very large super- 50 percent of the chloroplast soluble protein and is believed to
supercomplexes can be involved in cyclic electron flow. be the most abundant protein on earth. It is estimated that
• Reversible phosphorylation and dephosphorylation of the rubisco fixes more than 10 I I tOnS of atmOSpheriC C02each year.
light-harvesting complex II (LHCII) control the functional When photosynthetic algae are exposed to a brief pulse
organit.aLion of the photosynthetic apparatus m thylakoid of 14 C-Iabeled C0 2 and the cells are then quickly disrupted,
membranes. State 1 favors linear electron flow, whereas 3-phosphoglycerate is radiolabeled most rapidly, and all the
state 2 favors cyclic electron flow (see Figure 12-44 ). radioactivity is found in the carboxyl group. Because C0 2 is
initially incorporated into a three-carbon compound, the
Calvin cycle is also called the C 1 pathway of carbon fixation
(Figure 12-46).
12.7 C0 2 Metabolism The fate of 3-phosphoglycerate formed by rubisco is
·. During Photosynthesis
complex: some is converted to hexoses incorporated into
starch or sucrose, but some is used to regenerate ribulose 1,5-
Chloroplasts perform many metabolic reactions in green bisphosphate. At least nine enzymes are required to regener-
leaves. In addition to C01 fixation-incorporation of gas- ate ribulose 1,5-bisphosphate from 3-phosphoglycerate.
eous col into small organic molecules and then sugars-the Quantitatively, for every 12 molecules of 3-phosphoglycer-
synthesis of almost all amino acids, all fatty acids and caro- ate generated by rubisco (a total of 36 C atoms), 2 of them
tenes, all pyrimidines, and probably all purines occurs in (6 C atoms) are converted to 2 molecules of glyceraldehyde
chloroplasts. However, the synthesis of sugars from C0 2 is 3-phosphate (and later to 1 hexose), whereas 10 molecules
the most extensively studied biosynthetic pathway in plant (30 C atoms) are converted to 6 molecules of ribulose 1,5-
cells. We first consider the unique pathway, known as the bisphosphate (Figure 12-46, top). The fixation of 6 C02 mole-
Calvin cycle (after discoverer Melvin Calvin), that fixes C01 cules and the net formation of 2 glyceraldehyde 3-phosphate
into three-carbon compounds, powered by energy released molecules require the consumption of 18 ATPs and 12 NADPHs,
during ATP hydrolysis and oxidation of NADPH. generated by the light-requiring processes of photosynthesis.

Rubisco Fixes C0 2 in the Chloroplast Stroma Synthesis of Sucrose Using Fixed C02
The enzyme ribulose 1,5-bisphosphate carboxylase, or rubisco, Is Completed in the Cytosol
fixes C02 into precursor molecules that are subsequently con- After its formation in the chloroplast stroma, glyceraldehyde
verted into carbohydrates. Rubisco is located in the stromal 3-phosphate is transported to the cytosol in exchange for
space of the chloroplast. This enzyme adds C0 2 to the five- phosphate. The final steps of sucrose synthesis (Figure l2-46,
carbon sugar ribulose 1,5-bisphosphate to form two molecules bottom) occur in the cytosol of leaf cells.
of the three-carbon-containing 3-phosphoglycerate (Figure An antiporter transport protein in the chloroplast mem-
12-45). Rubisco is a large enzyme (- 500 kDa ), with the most brane brings fixed C0 2 (as glyceraldehyde 3-phosphate ) into
common form composed of eight identical large and eight the cytosol when the cell is exporting sucrose vigorously. No

CH - 0 - PO H
I 2 3

CH - 0 - PO H 0 CH 2 - 0 P03 H H2 0 H- C- OH
I 2 3

O= C=O + C= O 0- C C OH c 0
I
H-C-OH C= O 0
I I +
H- C- OH H- C- OH o-
1
CH 2 - 0 - P0 3 W C= O

H- C- OH

CH 2 - 0-P0 3 H
Ribulose Enzyme-bound intermediate 3-Phosphoglycerate
1,5·bisphosphate (two molecules)

FIGURE 12-45 The initial reaction of rubisco that fixes C02 five-carbon sugar ribulose 1,5-bisphosphate. The products are two
into organic compounds. In this reaction, catalyzed by ribulose molecules of 3-phosphoglycerate.
1,5-bisphosphate carboxylase (rubisco). C02 condenses with the

12.7 C0 2 Metabolism During Photosynthesis 567

 6 C02 -= 1C
C02 O=C = O
-}

~
0 CH 2 0H

= SC
c-o- C= O
6 Ribulose 12 3-Phosphoglycerate = 3C l
1,5-bisphosphate I
H C - OH
6AOP < y C02 FIXATION
Y - - - - - 1 2 ATP
CH 2 - 0P0/-
H- C- OH

H-C - OH
6ATP - - - - 1 (CALVIN CYCLE)
~ 12ADP 3-Phosphoglycerate I 2-
CH2- 0P03 I
Ribulose = 5C
6 12 1,3-Bisphosphoglycerate = 3C 0 Ribulose
5-phosphate
5-phosphate
12 NADPH C- OP032 -

4P,~ }7e"'yme< .
12 NADP ~ H
I
C-OH CH 2 OP032 -
CH 2 -0P0 3 2-
Glyceraldehyde _ lC Glyceraldehyde = lC C= O
10 12 1,3-Bisphosphoglycerate I
3-phosphate - 3-phosphate
H- C- OH

~
H I
I H-C - OH
C= O I 2
I CH 2 OP0 3 -
Glyceraldehyde = lC
2 H C- OH Ribulose
3-phosphate
1,5-bisphosphate
CH 2 - 0P0 32 -
Stroma Giyceraldehyde 3-phosphate
Phosphate-
triosephosphate Inner chloroplast membrane
antiport protein
Cytosol
2 P,

2
Glyceraldehyde = lC
3-phosphate

~
Fructose
1 1,6-bisphosphate &C
CH 2 O-P032 -
~ x2 C= O
Fructose I
2 1,6-bisphosphate &C HO - C - H

P,~
H- C - OH
P, H- C OH
Fructose CH 2 O- P032 -
6C
1-phosphate Fructose 1,6-bisphosphate

j ciUCROSE SYNTHESIF

Glucose Fructose = &C

1 1-phosphate &C 6-phosphate

UTP - - - - - \

PP;~
1 UOP-glucose = 6C

CH20H CH OH

~
2

0~
0
UDP

12c
Sucrose = 0
6-phosphate , 2 HO
HO L- O CH 2 0P03 2 -

~P, 0H
Sucrose 6-phosphate
OH
Sucrose 12C

568 CHAPTER 12 • Cellular Energetics

FIGURE 12-46 The pathway of carbon during photosynthesis. Reduced thioredoxin then activates severa l Calvin cycle
(Top) Six molecules of C02 are converted into two molecules of glyc- enzymes by reducing their disulfide bonds. In the dark, when
eraldehyde 3-phosphate. These reactions, which constitute the Calvin thioredoxin becomes reoxidized, these enzymes are rcoxtdized
cycle, occur in the stroma of the chloroplast. Via the phosphate/ and so inactivated. Thus these enzymes are sensitive to the
triosephosphate antiporter, some glyceraldehyde 3-phosphate is redox state of the stroma, which in turn is light sensitive-an
transported to the cytosol in exchange for phosphate. (Bottom) In
elegant mechanism for regulating enzymatic activity by light.
the cytosol, an exergonic series of reactions converts glyceraldehyde
Rubisco is one such light/redox-sensitive enzyme, al-
3-phosphate to fructose 1,6-bisphosphate. Two molecules of fructose
though its regulation is very complex and not yet fully un-
1,6-bisphosphate are used to synthesize one of the disaccharide
sucrose. Some g lyceraldehyde 3-phosphate (not shown here) is
derstood. Rubtsco is spontaneously activated in the presence
also converted to amino acids and fats, compounds essential for
of high C0 2 and Mg2+ concentrations. The activating reac-
plant growth. tion entails covalent addition of C0 2 to the side-chain amino
group of a lysine in the active site, forming a carbamate
group that then binds a Mg 2 - ion required for enzymatic
fixed C0 2 leaves the chloroplast unless phosphate is fed into it activity. Under normal conditions, however, with ambient
to replace the phosphate carried out of the stroma in the form levels of col, the reaction is slow and usually requires ca-
of glyceraldehyde 3-phosphate. During the synthesis of sucrose talysis by rubisco activase, a member of the AAA + family of
from glyceraldehyde 3-phosphate, inorganic phosphate groups ATPases. Rubisco activase hydrolyzes ATP and uses the en-
are released (Figure 12-46, bottom left) . Thus the synthesis of ergy released to clear the active site of rubisco so that C0 2
sucrose facilitates the transport of additional glyceraldehyde can be added to its active site lysine. Rubisco activase also
3-phosphate from the chloroplast to the cytosol by providing accelerates an activating conformational change in rubisco
phosphate for the anti porter. It is worth noting that glyceralde- (inactive-closed to active-opened state). The regulation of
hyde 3-phosphate is a glycolytic intermedia te and that the rubisco activase by thioredoxin is, at least in part in some
mechanism of the conversion of glyceraldehyde 3-phosphate to species, responsible for ru bisco's light/redox sensitivity. Fur-
hexoses is almost the reverse of that in glycolysis. thermore, rubisco activase's activity is sensitive to the ratio
The synthests of starch is more complex. The key monomer of ATP:ADP. If that ratio is low (relatively. high ADP), then
substrate used to build large starch polymers is ADP-glucose. the activase will not activate rubisco (and so the cell will
This polymerization takes place in the stroma and starch poly- expend less of its scarce ATP to fix carbon). Photosynthesis
mers are stored in densely packed crystalline aggregates called is sensitive to a variety of typical plant stresses-moderate
granules. The enzymes that generate ADP-glucose from heat, cool temperatures, drought (limited water), high salt,
glucose-1 phosphate and ATP are found in both the stroma high light intensity, and UV radiation. At least some of these
and the cytosol, indicating that hexoses of various structures in influence C0 2 fixation by reducing the activity of rubisco
the cytosol are imported into to stroma for starch synthesis. activase and thus rubisco. Inhibition of C02 fixation reduces
consumption of NADPH. Under strong light conditions the
excess NADPH/NADP ratio can reduce electron flow to
Light and Rubisco Activase Stimulate
NADP and increase leakage to 0 2, resulting in increased
C02 Fixation ROS formation, which can interfere with a variety of cellular
The Calvin cycle enzymes that catalyze C02 fixation are rap- processes. Given the key ro le of rubisco in controlling energy
idly inactivated in the dark, thereby conserving ATP that is utilization and carbon flux-both in an individual chloro-
generated in the dark (for example by the breakdown of plast and, in a sense, th roughout the entire biosphere-it is
starch) for other synthetic reactions, such as lipid and amino not surprising that its activity is tightly regulated.
acid biosynthesis. One mechanism that contributes to this
control is the pH dependence of several Calvin cycle enzymes.
Because protons are traflsported from the stroma into the
Photo respiration Competes with Carbon
thylakoid lumen during photoelectron transport (see Figure
12-39), the pH of the stroma increases from -7 in the dark to Fixation and Is Reduced in ( 4 Plants
-8 in the light. The increased activity of several Calvin cycle As noted above, rubisco catalyzes the incorporation of C0 2
enzymes at the higher pH promotes C02 fixation in the light. into ribu lose 1,5-bisphosphate as part of photosynthesis. It
A stromal protein called thioredoxin (Tx) also plays a can catalyze a second, distinct, and competing reaction with
role in controlling some Calvin cycle enzymes. In the dark, the same substrate-ribulose 1,5-bisphosphate-but with 0 2
thioredoxin contains a disulfide bond; in the light, electrons are in place of C02 as a second substrate, in a process known as
transferred from PSI, via ferredoxin, to thioredoxin, reducing photorcspiration (figure 12-47). The products of the second
its d isulfide bond: reaction are one molecule of 3-phosphoglycerate and one
molecule of the two-carbon compound phosphoglycolate.
PS I
2H+l 2 The carbon-fixing reaction is favored when the ambient C0 2

e:-----''------ e :: concentration is relatively high, whereas photorespiration is

favored when C02 is low and 0 2 relatively high. Photorespi-
ration takes place in light, consumes 0 2, and converts ribulose

12.7 C02 Metabolism During Photosynthesis 569

 0
I
c 0
~~~S ugars
H-C OH

CH 2 -0P0- 2

0
ADP ATP
C02+ H20 c-o co,
CH 2 OPO,' \. -~ CHOH
~

H-C-OH
c 0
~ C0 2 fixation
3-Phosphoglycoroto
CH 20H
Glycerate
\
H-C OH I
Photorespirotlon /
0 0
CH 2 OPO/ -..., ....__fJ II
Ribulose
/>-. c 0 0 c-o Two glycolate
1,5-bisphosphote
H-C OH c 0 CH2 0H
o. H20 Glycolote
CH 2 · OPO/ CH 2 OPO/ P,
3-Phosphoglycoroto Phosphoglycolato

FIGURE 12-47 C02 fixation and photorespiration. These compet- complex set of reactions that take place in peroxisomes and mitochon-
ing pathways are both initiated by ribulose 1,5-bisphosphate carboxyl- dria, as well as chloroplasts. The net result: for every two molecules
ase (rubisco), and both utilize ribulose 1,5-bisphosphate. C02 fixation, of phosphoglycolate formed by photorespiration (four C atoms), one
pathway 1, is favored by high C02 and low 0 2 pressures; photorespira- molecule of ]-phosphoglycerate is ultimately formed and recycled and
tion, pathway 2, occurs at low C02 and high 0 2 pressures (that is, under one molecule of C0 2 is lost.
normal atmospheric conditions). Phosphoglycolate is recycled via a

1,5-bisphosphate in part to C0 2 . As Figure 12-47 shows, The C 4 pathway involves two types of cells: mesophyll
photorcspiration is wasteful to the energy economy of the cells, which are adjacent to the air spaces in the leaf interior,
plant: it consumes ATP and 0 2, and it generates C0 2 with-
out fixing carbon. Indeed, when C0 2 is low and 0 2 is high,
and bundle sheath cells, which surround the vascular tissue ·.
and are sequestered away from the high oxygen levels to
much of the C0 2 fixed by the Calvin cycle is lost as the result which mesophyll cells arc exposed (figure 12-48a ). In the
of photorespiration. Recent studies have suggested that this mesophyll cells of c4 plants, phosphoenolpyruvate, a three-
surprising, wasteful alternative reaction catalyzed by rubisco carbon molecule derived from pyruvate, reacts with C0 2 to
may be a conseq!Jencc of the inherent difficulty the enzyme generate oxaloacetate, a four-carbon compound (Figure
has in specifically binding the relatively featureless col mol- 12-48b). The enzyme that cata lyzes this reaction, phospho-
ecule and of the ability of both co~ and 02 to react and enolpyruvate carboxylase, is found almost exclusively in c4
form distinct products with the same initial enzyme/ribulose plants and unlike rubisco is insensitive to 0 2 . The overall
1,5-bisphosphate intermediate. reaction from pyruvate to oxaloacetate involves the hydroly-
Excessive photorespiration could become a problem for sis of one ATP and has a negative ~ G. Therefore, C0 2 fixa-
plants 111 a hot, dry environment, because they must keep the tion will proceed even when the C02 concentration is low.
gas-exchange pores (stomata) in their leaves closed much of The oxaloacetatc formed in mesophyll cells is reduced to
the time to prevent excessive loss of moisture. As a conse- malate, which is transferred by a special transporter to the
quence, the C0 2 level inside the leaf can fall below the Km of bundle sheath cells, where the C0 2 released by decarboxyl-
rubisco for C0 1 . Under these conditions, the rate of photo- ation enters the Calvin cycle (Figure 12-48b).
synthesis is slowed, photorespiration is greatly favored, and Because of the transport of C0 2 from mesophyll cells,
the plant might be in danger of fixing inadequate amounts of the C02 concentration in the bundle sheath cells of C 4 plants
C0 1 • Corn, sugarcane, crabgrass, and other plants that can is much higher than it is in the normal atmosphere. Bundle
grow in hot, dry environments have evolved a way to avoid sheath cells are also unusual in that they lack PSIJ and carry
this problem by utilizing a two-step pathway of C02 fixation our only cyclic electron flow catalyzed by PSI, so no 0 2 is
in which a C01 -hoarding step precedes the Calvin cycle. The evolved. The high C0 2 and reduced 0 1 concentrations in the
pathway has been n:1med the C4 pathway because f14 C]C0 2 bundle sheath cells favor the fixation of col by rubisco to
labeling showed that the first radioactive molecules formed form 3-phosphoglycerate and inhibit the utilization of ribu -
during photosynthesis in this pathway are four-carbon com- lose 1,5-bisphosphate in photorespiration.
pounds, such as oxaloacetatc and ma late, rather than the In contrast, the high 0 2 concentration in the atmosphere
three-carbon molecules that initiate the Calvin cycle favors photorespiration in the mesophyll cells of cl plants
(C 3 pathway). (pathway 2 in figure 12-47); as a result, as much as 50 percent

570 CHAPTER 12 • Cellular Energetics

 Vascular bundle
(a) (xylem, phloem)

Mesophyll
cells
Bundle
sheath
cells

Epidermis

Stoma Chloroplast

(b)
Mesophyll cell 0 NADPH 0 0 Bundle sheath cell

c 0 +H NADP
c 0 c 0
CH 2 \...! CH 2 CH 2
I I
c 0 H C- OH H C- OH
CO __. Calvin
2 cycle
C- 0 C- 0 C- 0
II
0 0 0
Oxaloacetate Malate Malate NADP

CH 2 CH 3 CH 3
Pyruvate·
•. ' c- o PO/ phosphate
dikinase C= O C= O
NADPH "' H+
I ( I I
C- 0
7"\ATP
AMP
c- o
I
c- o
II
0 0 0
Phosphoenolpyruvate +
PP, P Pyruvate Pyruvate

FIGURE 12-48 Leaf anatomy of C4 plants and the C4 pathway. synthesis. Sucrose is carried to the rest of the plant via the phloem. In
(a) In ( 4 plants, bundle sheath cells line the vascular bundles C3 plants, which lack bundle sheath cells, the Calvin cycle operates in
containing the xylem and phloem. Mesophyll cells, which are the mesophyll cells to fix C0 2• (b) The key enzyme in the C4 pathway
adjacent to the substomal air spaces, can assimilate C01 into is phosphoenolpyruvate carboxylase, which assimilates C02 to form
four-carbon molecules at low ambient col and deliver it to the oxaloacetate in mesophyll cells. Decarboxylation of malate or other
interior bundle sheath cel ls. Bundle sheath cells contain abundant C4 intermediates in bundle sheath cells releases C02, which enters the
chloroplasts and are the sites of photosynthesis and sucrose standard Calvin cycle (see Figure 12·46, top).

of the carbon fixed by rubisco may be reoxidized to C02 in lower than it is in C3 plants, which use only the Calvin cycle
c3 plants. c4 plants are superior to cl plants in utilizing th e for C0 2 fixation. Nonetheless, the net rates of photosynthe-
available C02 , because the C4 enzyme phosphoenolpyruvate sis for C 4 grasses, such as corn or sugarcam:, can be two to
carboxylase has a higher affinity for col than does rubisco three times the rates for otherwise similar cl grasses, such
in the Calvin cycle. However, one ATP is converted to one as wheat, rice, or oats, owing to the elimination of losses
AMP in the cyclic C 4 process (to generate phosphoenolpyru- from photorcspiration.
vate from pyruvate}; thus the overall efficiency of the photo· Of the two carbohydrate products of photosynthesis, starch
synthetic production of sugars from NAOPH and ATP is remains in the mesophyll cells of C 3 plants and the bundle

12.7 C02 Metabolism During Photosynthesis 571

sheaf cells in C 4 plants. In these cells, starch is subjected to and P;, formation of ATP, and then release of ATP. Nor has
glycolysis, mainly in the dark, forming ATP, NADH, and small the detailed pathway of proton movement though the c
molecules that are used as building blocks for the synthesis of ring been defined. In addition, many questions remain
amino acids, lipids, and other cellular constituents. Sucrose, in about the precise mechanism of action of transport pro-
contrast, is exported from the photosynthetic cells and trans- teins in the inner mitochondrial and chloroplast mem-
ported throughout the plant. branes that play key roles in oxidative phosphorylation
and photosynthesis.
We now know that release of cytochrome c and other
proteins from the intermembrane space of mitochondria into
the cytosol plays a major role in triggering apoptosis (Chap-
KEY CONCEPTS of Section 12.7 ter 21). Certain members of the Bcl-2 family of apoptotic
C02 Metabolism During Photosynthesis proteins and ion channels localized in part to the outer mito-
chondrial membrane participate in this process. The connec-
In the Calvin cycle, C0 2 is fixed into organic molecules in tions between energy metabolism and mechanisms underlying
a series of reactions that occur in the chloroplast stroma. The apoptosis remain to be clearly defined.
initial reaction, catalyzed by rubisco, forms a three-carbon The recognition over the past decade of the importance
intermediate. Some of the glyceraldehyde 3-phosphate gen- of mitochondrial dynamics (e.g., fnsion and fission) to mito-
erated in the cycle is transported to the cytosol and converted chondrial function has set the stage for detailed genetic mo-
to sucrose (see Figure 12-46). lecular analysis of these processes. Several of the key players
• The light-dependent activation of several Calvin cycle en- in fusion and fission have been identified, but many addi-
zymes and other mechanisms increases fixation of C02 in tional components have yet to be discovered, and the mecha-
the light. The redox state of the stroma plays a key role in nisms of these complex processes, such as the coordinated
this regulation as does the regulation of the activity of fusion of inner membranes with each other and outer mem-
rubisco by rubisco activase. branes with each other, are waiting to be el'ucidated.
• In C 3 plants, a substantial fraction of the C0 2 fixed by the The role of reactive oxygen species (ROS ) in cell biology
Calvin cycle can be lost as the result of photorespiration, a is an active area of research. ROS-mediated cellular stress is
wasteful reaction catalyzed by rubisco that is favored at low now thought to play a role in many diseases and will likely
C02 and high 0 2 levels (see Figure 12-47). continue to be a major area of research in the coming years.
In addition to their role in cellular oxidative stress, ROS can
In C 4 plants, C02 is fixed initially in the outer mesophyll also serve as signaling molecules that alter nuclear gene ex-
cells by reaction with phosphoenolpyruvate. The four-carbon pression, sometimes called retrograde signaling. It appears
molecules so generated are shuttled to the interior bundle that ROS and other small molecules released from the mito-
sheath cells, where the C0 2 is released and then used in the chondrion and chloroplast can be used to inform the nucleus
Calvin cycle. The rate of photorespiration in C 4 plants is about the metabolic status of each organelle and thus permit
much lower than it is in C 3 plants. appropriate regulation of gene expression in response. In
some cases this involves compensatory activation of protec-
tive genes. In others it may involve increasing or decreasing
the production of nuclear-encoded proteins to insure proper
organelle functioning. The mechanisms of these signaling
Perspectives for the Future pathways, which in some cases involve redox reactions with
Although the overall processes of photosynthesis and mito- thiols on signaling molecules, remain to be determined.
chondrial oxidation are well understood, many important As we better understand the mechanisms underlying
details remain to be uncovered. For example, while increas- photosynthesis, particularly the action of rubisco-both its
ingly high-resolution structures of complexes and supercom- regulation and its influence on photosynthesis and overall
plexes are being determined, many of the mechanistic details chloroplast metabolism-it is possible that we will be able to
underlying the function and regulation of electron transport exploit these insights to improve crop yields to provide
chains and their associated reactions (proton translocation, abundant and inexpensive food to all who need it.
oxygen generation, etc. ) remain to be established. Moving
beyond this static picture of these remarkably complex struc-
tures requires additional biophysical analysis of the dynam-
ics underlying their acLivitie~. For example, we do not know
with certainty the pathway taken by protons during proton Key Terms
pumping in some of the electron transport complexes.
aerobic oxidation 517 Calvin cycle 567
Although the binding-change mechanism for ATP syn-
thesis by the F0 F 1 complex is now generally accepted, we do ATP synthase 544 carbon fixation 553
not understand precisely how conformational changes in binding-change mechanism catabolism 520
each 13 subunit are coupled to the cyclical binding of ADP 547 c4pathway 570

572 CHAPTER 12 • Cellular Energetics

L
chemiosmosis 518 oxidative phosphorylation organelle, besides the mitochondrion, can oxidize fatty acids?
chlorophylls 552 519 What is the fundamental difference between oxidation occur-
chloroplast 552 peroxisomal oxidation 531 ring in this organelle and mitochondrial oxidatton?
citric acid cycle 520 photoelectron transport 556 7. Each of the cyrochromes in the mitochondrion contains
phororespiration 569 prosthetic groups. What is a prosthetic group? Which type
coenzyme Q 535
of prosthetic group is associated with the cyrochromcs?
cytochrome 535 photosynthesis 517
What property of the various cytochromes ensures unidirec-
electron carrier 529 photosystem 555 tional electron flow along the electron transport chain?
electron transport chain 519 prosthetic group SJ4 8. The electron transport chain consists of a number of mul-
endosymbiont hypothesis proton-motive force 519 tiprotein complexes, which work in conjunction to pass elec-
545 Q cycle 538 trons from an electron carrier, like NADH, to 0 2 . What is
fermentation 522 reactive oxygen species 541 the role of these complexes in ATP synthesis? It has been
flavin adenine dinucleotide reduction potential 539 demonstrated that respiration supercomplexes contain all
(FAD) 520 the protein components necessary for respiration. Why is
respiration 519
this beneficial for ATP synthesis, and what is one way that
f' 0 F 1 complex 544 respiratory control 551 the existence of supercomplexes has been demonstrated ex-
glycolysis 520 rubisco 567 perimentally? Coenzyme Q (CoQ) is not a protein, but a
mitochondrion 524 substrate-level small, hydrophobic molecule. Why is it important for the
mitochondrial inner phosphorylation 520 functioning of the electron transport chain that CoQ is a
membrane 525 thylakoids 553 hydrophobic molecule?
nicotinamide adenine uncoupler 551 9. It is estimated that each electron pair donated by NADH
dinucleotide (NAD+) 520 leads to the synthesis of approximately three ATP molecules,
whereas each electron pair donated by FADH2 leads to the
synthesis of approximately two ATP molecules. What is the
underlying reason for the difference in yield for electrons do-
Review the Concepts nated by FADH2 versus NADH?
10. Describe the main functions of the different components
1. The proton-motive force (pmf) is essential for both mito- of ATP synthase enzyme in the mitochondrion. A structur-
chondrial and chloroplast function. What produces the pmf, ally similar enzyme is responsible for the acidification of
and what is its relationship to ATP? The compound 2,4- lysosomes and endosomes. Given what you know about the
dinitrophenol (DNP), which was used in diet pills in the mechanism of ATP synthesis, explain how this acidification
1 930s but later shown to have dangerous side effects, allows might occur.
protons to diffuse across membranes. Why is it dangerous to 11. Much of our understanding of ATP synthase is derived
consume DNP? from research on aerobic bacteria. What makes these organ-
2. The mitochondrial inner membrane exhibits all of the isms useful for this research? Where do the reactions of gly-
fundamental characteristics of a typical cell membrane, but colysis, the citric acid cycle, and the electron transport chain
it also has several unique characteristics that are closely as- occur in these organisms? Where is the pmf generated in
sociated with its role in oxidative phosphorylation. What are aerobic bacteria? What other cellular processes depend on
these unique characteristics? How docs each contribute to the pmf in these organisms?
the function of the inner membrane? 12. An important function of the mitochondrial inner mem-
3. Maximal production of ATP from glucose involves the brane is to provide a selectively permeable barrier to the
reactions of glycolysis, the> citric acid cycle, and the electron movement of water-soluble molecules and thus generate dif-
transport chain. Which of these reactions requires 0 2 , and ferent chemical environments on either side of the mem-
why? Which, in certain organisms or physiological condi- brane. However, many of the substrates and products of
tions, can proceed in the absence of 0 2 ? oxidative phosphorylation are water soluble and must cross
4. Fermentation permits the continued extraction of energy the inner membrane. How does this transport occur?
from glucose in the absence of oxygen. If glucose catabolism 13. The Q cycle plays a major role in the electron transport
is anaerobic, why is fermentation necessary for glycolysis to chain of mitochondria, chloroplasts, and bacteria. What is the
continue? function of the Q cycle, and how does it carry out this func-
5. Describe the step-by-step process by which electrons from tion? WhaL ekctron transport components parttctpate in the Q
glucose catabolism in the cytoplasm are transferred to the cycle in mitochondria, in purple bacteria, and in chloroplasts?
electron transport chain in the mitochondrial inner mem- 14. True or False: Since ATP is generated in chloroplasts,
brane. In your answer, note whether the electron transfer at cells capable of undergoing photosynthesis do not require
each step is direct or indirect. mitochondria. Explain. Name and describe the idea that ex-
6. Mitochondrial oxidation of fatty acids is a major source plains how mitochondria and chloroplasts are thought to
of ATP, yet fatty acids can be oxidized elsewhere. What have originated in eukaryotic cells.

Review the Concepts 573

15. Write the overall reaction of oxygen-generating photo-
synthesis. Explain the following statement: the 0 2 generated
by photosynthesis is simply a by-product of the pathway's
generation of carbohydrates and ATP.
16. Photosynthesis can be divided into multiple stages.
What are the stages of photosynthesis, and where does each
occur within the chloroplast? Where is the sucrose produced
.£
(J)
cQ)
by photosynthesis generated? c
17. The photosystems responsible for absorption of light c
0
energy are each composed of two linked components, the ·a;
-~
reaction center and an antenna complex. What is the pig- E
LU
ment composition and role of each component in the process
of light absorption? What evidence exists that the pigments
found in these components are involved in photosynthesis?
18. Photosynthesis in green and purple bacteria does not
produce 0). Why? How can these organisms still use photo-
synthesis to produce ATP? What molecules serve as electron
donors in these organisms? 550
Wavelength (nm)
19. Chloroplasts contain two photosystems. What is the func-
tion of each? ror linear electron flow, diagram the flow of elec-
trons from photon absorption to NADPH formation. What c. After the vesicles were incubated in buffer containing
does the energy stored in the form of NADPH synthesize? ADP, P, and 0 2 for a period of time, additipn of dinitrophe-
20. The Calvin cycle reactions that fix C02 do not function nol caused an increase in BCECF fluorescence. ln contrast,
in the dark. What are the likely reasons for this? How are addition of valinomycin produced only a small transient ef-
these reactions regulated by light? fect. Explain these findings.
21. Rubisco, which may be the most abundant protein on d. What result would you expect to see if the source of
earth, plays a key role in the synthesis of carbohydrates in the mitochondrial membrane were brown-fat mitochondria?
organisms that use photosynthesis. What is rubisco, where is Explain.
It located, and what function does it serve? c. Chloroplasts could also be used as a source of mem-
branes in a similar experiment (as in part a) involving
BCECF. ln this case, the BCECF would be surrounded by
what membrane? How would the fluorescence change upon
addition of light, ADP, and P,?

Analyze the Data

References
A proton gradient can be analyzed with fluorescent dyes whose
emission-intensity profiles depend on pH. One of the most use- First Step of Harvesting Energy from Glucose: Glycolysis
ful dyes for measuring the pH gradient across mitochondrial Berg, J., J. Tymoczko, and L. Stryer. 2002. Biochen11stry, 5th cd.
membranes is the membrane-impermeant, water-soluble fluo- W. H. Freeman and Company, chaps. 16 and 17.
rophore 2', 7'-bis-(2-carboxyethyl )-5( 6)-carboxyfluorescein Depre, C., M. Rider, and L. Hue. 1998. Mechanasms of control
of heart glycolysis. Eur.]. Biochem. 258:277-290.
(BCECF). The effect of pH on the emission intensity ofBCECF,
Fersht, A. 1999. Structure and Mechanism in Protein Science: A
excited at 505 nm, is shown in the accompanying figure. In G111de to Enzyme Catalyszs and Protem 1:-olding. W. H. Freeman
one study, sealed vesicles containing this compound were pre- and Company.
pared by mixing unsealed, isolated inner mitochondrial mem- Fothergill-Gilmore, L. A., and P. A. Michels. 1993. Evolution
branes with BCECF; after resealing of the membranes, the of glycolysis. Prog. Biophys. Mol. Bioi. 59:105-135.
vesicles were collected by centrifugation and then resuspended Nelson, D. L., and M. M. Cox. 2000. Lelminger Prmciples of
in nonfluorescent medium. Bzochemistry. Worrh, chaps. 14-17, 19.
a. When these vesicles were incubated in a physiological Pilkis, S. J., T. H. Claus, I. J. Kurland, and A. J. Lange. 1995.
6-Phosphofructo-2-kinase/fructose-2,6-bisphosph~ta~C': ~ metabolic
buffer containing NADH, ADP, P, and 0 2 , the fluorescence
signaling enzyme. Ann. Rev. Biochem. 64:799-835.
of BCECF trapped inside gradually decreased in intensity.
What does this decrease in fluorescent intensity suggest Mitochondria and the Citric Acid Cycle
about this vesicular preparation? Canfield, D. E. 2005. The early history of atmospheric oxygen:
b. How would you expect the concentrations of ADP, P, homage to Robert M. Garrels. Annu. Rev. l::.arth Planet. Sci. 33:1-36.
and 0 2 to change during the course of the experiment de- Chan, D. C. 2006. Mitochondria: dynamic organelles in disease,
scribed m part a? Why? aging, and development. Ce/1125(7):1241-1252.

574 CHAPTER 12 • Cellular Energetics

 Eaton, S., K. Bartlett, and M. Pourfarzam. 1996. Mammalian Ramirez, B. E., B. Malmstrom, J. R. Winkler, and H. B. Gray.
mitochondrial beta-oxidanon. Biochem. f. 320 (Part 2):345-557. 1995. The currents of life: the terminal electron-transfer complex of
Guest, J. R., and G. C. Russell. 1992. Complexes and complexi- re~piration. Proc. Nat'/. Acad. Sci. USA 92:11949-11951.
ties of the citric acid cycle in l::.schenclna coil. Curr. Top. Cell Reg. Rllltenberg, M., ct a!. 2002. Reduction of cytochrome c oxidase by
33:231-247. a second electron lead\ to proton translocation. Nat11re 417:99-102.
.· Krebs, H. A. 1970. The h1srory of the tricarboxylic acid cycle . Saraste, M. 1999. Oxidative phosphorylation at the fin de siecle.
Perspect. Bioi. Med. 14:154-170. 5czence 283:1488-1492.
Rasmussen, B., and R. Wolfe. 1999. Regulation of fatty acid Schafer, E., et al. 2006. Architecture of active mammalian respira-
oxidanon in skeletal muscle. Ann. Rev. Nutntion 19:463-484. tory cham supercomplexes.]. Bzol. Chem. 281(22): 15370-15375.
VPinr, C., M. Mixon, M. Teige, and l'. Srcre. 1997. Model of a Schultz, B., and S. Chan. 2001. Structures and proton-pumping
quinary structure between Krebs TCA cycle enzymes: J model for srraregtes of mitochondrial resptratoq enzymes. Amz. Reu. Bzophys.
the metabolon. Biochemistry 36:14271-14276. Bwmol. Struc. 30:23-65.
Wanders, R. j., and H. R. Waterham. 2006. Bwchemistry of Sheeran, F. L., and S. Pepe. 2006. Energy deficiency in the
mammalian peroxisomes revisited . Anmt. Reu. Biochem. 75:295-332. fa11ing heart: linking mcreased reactJve oxygen spectes and disrup-
tion of oxidative pho~phorylation rate. Biochzm. Bzophys. Acta
The Electron Transport Chain and Generation
1757(5-6):543-552.
of the Proton-Motive Force
Tsukihara, T., et al. 1996. The whole structure of the 13-subunit
Acin-Perez, R., P. Fernandez-Stlva, M. L. Peleato, A. Perez- ox1dizcd cytochrome c oxidase at 2.8 A. Science 272:1136-1144.
Marros, and J. A. Enriquez. 2008. Respiratory active mitochondrial Walker, J. I:.. 199S. Determmation of the structures of respira-
supercomplexes. Mol. Cell32:519-539. tory enzyme complexes from mammalian mitochondria. Bzochim.
Babcock, G. 1999. How oxygen is activated and reduced in Bzophys. Acta 1271:221-227.
resptration . Proc. Nat'/. Acad. Scz. USA 96:12971-12973. Wallace, D. C. 2005. A mitochondrial paradigm of metabolic
Beinerr, H., R. Holm, and E. ~1iinck. 1997. Iron-sulfur clusters: and degenerattve diseases, aging, and cancer: a dawn for evolution-
nature's modular, multipurpose structures. Sczence 277:653-659. ary medicine. Anm1. Rev. Genet. 39:359-407.
Brandt, U. 2006. Energy Converting NAOH:quinone O'l.ldore- Xta, D., et al. 1997. Crystal structure of the cytochrome hc 1
ductase (complex 1). Amzu. Rev. Bwchem. 75:165-187. complex from bovine heart mitochondria. Science 277:60-66.
Brandt, U., and B. Trumpower. 1994. The protonmotive Q cycle Zaslavsky, D., and R. Genms. 2000. Proton pumping by
in mitochondria and bacteria. Crit. Ret'. Biochem. Mol. Bioi. cytochrome oxidase: progress and postulates. Bwchzm. Biophys.
29:165-197. Acta 1458:164-179.
Daiber, A. 2010. Redox signaling (cross-talk) from and to Zhang, M., 1:.. Mtleykovskaya, and W. Dowhan. 2005.
mitochondria involves mitochondrial pores and reactive oxygen Cardiolipin is essential for organizarion of complexes UI and IV into
species. Bzochzm. Biophys. Acta 6-7:897-906. a supercomplex in intact yeast mitochondria. f. Bioi. Chem.
Oarrouzet, E., C. Moser, P. L. Dutton, and F. Daldal. 2001. 280(33 ):29403-29408.
Large scale domain movement in cytochrome be I: a new device for Zhang, Z., et al. 1998. Electron transfer by domain movement
electron transfer in proteins. Trends Biochem. Sci. 26:445-451. in cytochrome bc 1• Nature 392:6-7-684.
Dickinson, B. C., D. Srikun, and C. J. Chang. 2010. Mitochondrial-
targeted fluorescent probes for reactive oxygen species. Cztrr. Opin. Harnessing the Proton-Motive Force to Synthesize ATP
Chem. Bioi. 14:50-56. Aksimcntiev, A., I. A. Balabin, R. H. Fillingame, and K.
Efremov, R. G., R. Baradaran, and L.A. Sazanov. 2010. The Schulten. 2004. Insights into the molecular mechamsm of rotation
architecture of respiratory complex I. Nature 465:441-445. in the F0 sector of ATP synthase. Biophys.]. 86(3 ): 1332-1344.
Finkel, T. 201 l. Signal transduction by reactive oxygen species. Stancher, M.A., J. Hullihen, P. Pedersen, and M. Amzel. 1998.
journal Cell BioiOg)' 194:7-15. The 2.8 A structure of rat liver f1-ATPase: configuration of a
Grigoneff, N. 1999. Structure of the resptratory critical intermediate 111 ATP symhests/hydrolysts. Proc. Nat'/. Acad.
NADH:ubiquinone oxidoreductase (complex 1). Curr. Opm. Struc. Scz. USA 95:11065-11070.
Bioi. 9:476-483. Boyer, P. D. 1997. The ATP synthase-a splendid molecular
Hosler, J. P., S. Ferguson-Miller, and D. A. ~1ills. 2006. Energy machme. Atm. Rev. Bwchem. 66:717-749.
transduction: proton transfer through the respiratory complexes. Capaldi, R., and R. Aggeler. 2002. ~lechanism of the Fni- 1-type
Amm. Rev. Bzochem. 75:165-..187. ATP synthase-a biological rotary motor. Trends Bzochem. Scr.
Hunte, C., V. Zickermann, and U. Brandt. 20 I 0. Functional 27:154-160.
modules and \tructural basis of conformational coupling in Elston, T., H. Wang, and G. Oster. 1998. Energy transduction
mitochondrial complex I. Science 329:448-451. in ATP synrhase. Nature 391:510-512.
Hyde, B. B., G. Twig, and 0. S. Shtrihai.2010. Organellar vs Hinkle, P. C. 1005. P/0 ratios of mitochondrial oxidative
cellular control of mitochondrial dynam1cs. Semm. Cell Det'. Bioi. phosphorylation. Bzochim. Biophys. Acta 1706(1-2):1-11.
21:575-581. Junge, W., S. Hendrik, and S. Engelbrecht. 2009. Torque
Koopman, W. J., et al. 2010. Mammalian mitochondrial generation and elastic power transmission in the rotary F0 1' 1-
·. complex 1: biogenesis, regulation, and reactive oxygen species ATPase. Nature 459:364-370.
generation. Antioxzd. Redox Szgnal. 12:1431-1470. Kinosir~, K., ct a!. 1998. F 1-ATPase: a rotary motor maJc of a
Michel, H.,]. Behr, A. Harrenga, and A. Kannt. 1998. Cyto- single molecule. Ce/193:21-24.
chrome c oxidase. Ann. Rev. Bzophys. Biomol. Struc. 27:329-356. Klingenberg, M., and S. Huang. 1999. Structure and function of
Mitchell, P. 1979. Keilin's resptratory cham concept and its the uncouplmg protem from brown adipose tissue. Bzochzm.
chemiosmotic consequences. Science 206:1148-1159. (Nobel Prize Biophys. Acta 1415:271-296.
Lecture.) Nury, H., et al. 2006. Relations between structure and function
Murphy, M.P. 2009. How mitochondria produce reactive of the mitochondrial ADP/ATP carrier. Amzu. Rev. Bzochem.
oxygen species. Biochem.]. 417:1-13. 75:713-741.

References 575
 Tsunoda, S., R. Aggeler, ,\1. Yoshida, and R. Capaldi. 2001. Hankamer, B., J. Barber, and E. Boekema. 1997. Structure and
Rotation of the c subunit oligomer m fully functional F.lF ATP membrane organizatton of phorosystem II from green plants. Ann.
synthase. Proc. Nat'[. Acad. Sci. USA 98:898-902. Reu. Plant Phys10l. Plant Mol. Bioi. 48:641-672.
Vercest, A. E., et al. 2006. Plant uncoupling mitochondrial Heathcote, P., P. Fyfe, and M. Jones. 2002. Reaction centres:
proteins. Annu. Rev. Plant 810l. 57:383-404. the \tructure and enJiution of hiological solar power. Trends
von Ballmoos, C., A. Wiedenmann, and P. Dimroth. 2009. Bwchem. Set. 27:79-87.
Essentials for ATP synrhesis hy F1F0 ATP synthases. A1mu. Ret•. Horron, P., A. Ruban, and R. Walters. 1996. Regulation of ltght
B10chem. 78:649-672. h.uvcsting in green plants. Ann. Rev. Plant Physiol. Plant Mol. Bioi.
Yasuda, R., et al. 200 I. Resolution of distinct rotational 47:655-684.
subs reps by suhmdlisecond kinetic analysis of F1-ATPase. Nature lwai, M., et al. 2010. Isolation of the elusive supercomplex that
410:898-904. dnves cyclic electron flow 111 photosynrhesis. Nature 464:1210-1213.
joltot, P., and A. Joliot. 2005. Quantification of cyclic and linear
Photosynthesis and Light-Absorbing Pigments flows in plants. Proc. Nat/. Acad. Sci. USA 102(13):4913-4918.
Ben-Shem, A., F. Frolow, and N. Nelson. 2003. Crystal Jordan, P., et al. 2001. Three-dimensional structure of cyano-
structure of plant photosystem I. Nature 426(6967):630-635. bactenal phorosystcm I at 2.5 A resolunon. Nature 411:909-917.
Blankenship, R. E. 2002. Molecular Mechamsms of Photosyn- Kiihlbrandt, \X . 2001. Chlorophylls galore. Nature 411:896-898.
thesis. Blackwell. Martin, J. L., and M. H. Vos. 1992. Femtosecond biology. Ann.
Detsenhofer, J., and J. R. Norris, cds. 1993. The Photosynthetic Ret•. 8iophys. Btamol. Struc. 21:199-222.
ReactiOn Center, vols. l and 2. Acadcmtc Press. Penner-Hahn, J. 1998. Structural characterization of the Mn sire
,\lcDermorr, G., et al. 1995. Crystal structure of an mtegral in the photosynthetic oxygen-evolving complex. Stmc. Bondmg
membrane light-harvesting complex from photosynthetic bacteria. 90:1-36.
Nature 364:517. Tommos, C., and G. Babcock. 1998. Oxygen production in
1\:elson, N., and C. F. Yocum. 2006. Structure and function of nature: a light-driven metalloradical emyme process. Ace. Chem.
phorosysrems 1 and II. Anmt. Rev. Plant Btol. 57:521-565. Res. 31:18-25.
Prince, R. 1996. Photosynthesis: the Z-scheme revisited. Trends
Birn·hem. Set. 21:121-122. C02 Metabolism During Photosynthesis
Wollman, F. A. 2001. State transitions reveal the dynamics and Buchanan, B. B. 1991. Regulation of C01 assimilanon in
flexihdity of the photosynthetic apparatus. f.MBO]. 20:3623-3630. oxygenic photosynthesis: the ferredoxin/thtOredoxm system.
Perspective on tts discovery, present status, and future development.
Molecular Analysis of Photosystems Arch. Biochem. Biophys. 288:1-9.
Allen, J. F. 2002. PhotoS)nthesis of ATP--electrons, proton Gutteridge, S., and j. Pierce. 2006. A unified theory for the basts of
pumps, rotors, and pmse. Cell 1 10:2""'3-276. the ltmttations of the primary reactton of photosynthetic col fixation:
Amunts, A., H. Toponk, A. BorovlkOV<I, and N. Nelson. 2010. was Dr. Pangloss right? Proc. "Jatl. Acad. Sci. USA 103:7203-7204.
Structure determmation and improved model of plant phorosystem Portis, A. 1992. Regulation of ribulose 1,5-bisphosphate
I.]. Bioi. Chem. 285:3478-3486. carboxylase/oxygenase activity. Ami. Reu. Plant Physiol. Plant Mol. '
Aro, E. M., l. Virgin, and B. Andersson. 199.1. Photoinhihtnon Bioi. -B:415-437.
of phorosystem II: Inacnvanon, protein damage, and turnover. Rawsthorne, S. 1992. Towards an understanding of C 1-C4
Broclmn. Bwphys:Acta 1143:113-134. photosynthesis. l:.ssays Bwchem. 27:135-146.
Deisenhofcr, j., and H. Michel. 1989. The photosynthetic Rokka, A., I. Zhang, and E.-M. Aro. 200 I. Rubisco activase: an
reaction center from the purple bacterium Rhodopseudomonas emyme with a temperature-dependent dual function? Plant].
l'trtdis. Science 245:1463-1473. (;-;obel Prize Lecture. ) 25:463-4 ~2.
Deisenhofer,]., and H. Michel. 1991. Structures of bacterial Sage, R., and J. Colemana. 2001. Effects of low atmospheric
phorosynthenc reaction centers. Ami. Rev. Cell Bioi. 7:1-23. C01 on plants: more than a thing of the past. Trends Plant Set.
Dekker, J.P., and E. .J. Boekema. 2005. Supramolecular 6:18-24.
organization of thylakoid membrane proteins in green plants. Schneider, G., Y. Lindqnst, and C. I. Branden. 1992. Rubisco:
B10clnm. Bwphys. Acta 1706( 1-2): 12-39. structure and mechanism. Ann. Rev. Biophys. Btomol. Struc.
Finazzi, G. 2005. The cenrral role of the green alga Chlamydo- 21:119-153.
monas remhardttr in reve•1ling the mechanism of state transitions J. Tcherkez, G. G., G. D. Farquhar, and T. J. Andrews. 2006.
Exp. Bot. 56(411 ):383-388. Despite slow catalysis and confused substrate specifictty, all ribulose
Guskov, A., et al. 2010. Recent progress in the crystallographtc btsphosphate carboxylases may he nearly perfectly optimized. Proc.
studtes of photosystem II. Chemphyschem. 11(6): 1160-1171. Natl. Acad. Set. USA 103(19):7246-7251.
Haldrup, A., P. Jensen, C. Lunde, and H. Scheller. 200 I. Wolosiuk, R. A., M.A. Ballicora, and K. Hagelin. 1993. The
Balance of power: a view of the mechanism of photosynthetic state reducrive penrose phosphate cycle for photosynthetic col assimtla-
transitions. Trends Plant Set. 6:301-305. tion: enzyme modulation. FASFB ]. 7:622-637.

576 CHAPTER 12 • Cellular Energetics

... ~
CHAPTER

Moving Proteins
into Membranes
and Organelles

Fluorescence micrograph of a cultured mammalian (COS-7) cell showing

the distribution of endoplasmic reticulum (green), Golgi apparatus (red),
and nucleus (blue). Newly synthesized secretory proteins are first
targeted to the ER. where they are folded and modified before being
exported to the Golgi for sorting to downstream destinations. [Courtesy of
Jennifer Lippincott-Schwartz and Prasanna Satpute]

typical mammalian cell contains up to 10,000 different or protein sorting, encompasses two very different kinds of

A kinds of proteins; a yeast cell, about 5000. The vast

majority of these proteins are synthesized by cytosolic
ribosomes, and many remain within the cytosol (Chapter 4 ).
processes: signal-based targeting and vesicle-based trafficking.
The first general process involves targeting of a newly synthe-
sized protein from the cytoplasm to an intracellular organelle.
However, as many as half of the different kinds of proteins Targeting can occur during translation or soon after synthesis
produced in a typical cell are delivered to one or another of the of the protein is complete. For membrane proteins, targeting
various membrane-bounded organelles within the cell or to the leads to insertion of the protein into the lipid bilayer of the
cell surface. For example, many receptor proteins and trans- membrane, whereas for water-soluble proteins, targeting leads
porter proteins must be delivered to the plasma membrane, to translocation of the entire protein across the membrane into
some water-soluble enzymes such as RNA and D A polymer- the aqueous interior of the organelle. Proteins are sorted to the
ases must be targeted to the nucleus, and components of the endoplasmic reticulum (ER), mitochondria, chloroplasts, per-
extracellular matrix as well as digestive enzymes and polypep- oxisomes, and nucleus by this general process (Figure 13-1 ).
tide signaling molecules must be directed to the cell surface for The second general sorting process is known as the secre-
secretion from the cell. These and all the other proteins pro- tory pathway, and involves transport of proteins from the
duced by a cell must reach their correct locations for the cell to ER to their final destination within membrane-enclosed ves-
function properly. icles. For many proteins, including those that make up the
The delivery of newly synthesized proteins to their proper extracellular matrix, the final destination is the outside of
cellular destinations, usually referred to as protein targeting the cell (hence the name); integral membrane protems are also

OUTLINE

13.1 Targeting Proteins to and Across 13.4 Targeting of Prot eins to Mitochondria
the ER Membrane 579 and Chloroplasts 601

13.2 Insertion of Membrane Proteins into the ER 587 13.5 Targeting of Peroxisomal Proteins 612

13.3 Protein Modifications, Folding, 13.6 Transport into and out of the Nucleus 615
and Quality Control in the ER 594
0 OVERVIEW ANIMATION: Protein Sorting

Outer nuclea r
membrane -
Inner nuclear
6
memb ra ne - . .
"'f"
/~
Nuclear
pore
mRNA ""-.

(-~--,.:."''
Cytosol

sequence
fJ

Rough endoplasmic seq uence Peroxisome

reticulum

\f\N Matrix

11\ Golgi
~ complex
Inner
mem brane
Mitochondrion

ek-
~ Vl -cr- Chloroplast

Plasma
membrane
m/ \.m Lysosome

SECRETORY PATHWAY

FIGURE 13-1 Overview of major protein-sorting pathways in subcompartments of these organelles by additional sorting steps.
eukaryotes. All nuclear-encoded mRNAs are translated on cytosolic Nuclear proteins enter and exit through visible po res in the nuclear
ribosomes. Right (nonsecretory pathways): Synthesis of proteins lacking envelope. Left (secretory pathway): Ribosomes synthesizing nascent
an ER signal sequence is completed on free ribosomes (step OJ. Those proteins in the secretory pathway are directed to the rough endoplas-
proteins that contain no targeting sequence are released into the mic reticulum (ER) by an ER signa l sequence (pi nk; steps D . f) ). After
cytosol and remain there (step fl). Proteins with an organelle-specific translation is completed on the ER, these proteins can move via
targeting sequence (pink) first are released into the cytosol (step fll transport vesicles to the Golgi complex (step Ill. Further sorting
but then are imported into mitochondria, chloroplasts, peroxisomes, or delivers proteins either to the plasma membrane or to lysosomes
the nucleus (steps 11-l'iJ). Mitochondrial and chloroplast proteins (steps 9'1. m1). The vesicle-based processes underlying the secretory
typically pass through the outer and inner membranes to enter the pathway (steps 11. 9 , shaded box) are discussed in Chapter 14.
matrix or stromal space, respectively. Other proteins are sorted to other

transported to the Golgi, lysosome, and plasma membrane ER membrane. Once translocated across the ER membrane,
by this process. The secretory pathway begins in the ER; proteins are assembled into their native conformation by
thus all proteins slated ro enter the secretory pathway arc protein-folding catalysts present in the lumen of t he ER. In-
initially targeted to th is organelle. deed, the ER is the location where about one-third of the pro-
Targeting to the ER generally involves nascent proteins still teins in a typical cell fold into their native conformations,
m the process of being synthesized on a ribosome. Newly made and most of the resident ER proteins either directly or indi-
proteins are thus extruded from the ribosome directly into the rectly contribute to the folding process. As part of the folding

578 CHAPTER 13 • Moving Proteins into Membranes and Organelles

process, proteins also undergo specific post-translational mod- For each of the protein-targeting events discussed in this
ifications in the ER. These processes are monitored carefully, chapter, we will seek to answer four fundamental questions:
and only after their folding and assembly is complete are
1. What is the nature of the signal sequence, and what
proteins permitted to be transported out of the ER to other
distinguishes it from other types of signal sequences?
destinations. Proteins whose final destination is the Golgi,
lysosome, plasma membrane, or cell exterior are transported 2. What is the receptor for the signal sequence?
along the secretory pathway by the action of small vesicles 3. What is the structure of the translocation channel that
that bud from the membrane of one organelle and then fuse allows transfer of proteins across the membrane bilayer? In
with the membrane of another (see Figure 13-1, shaded box). particular, is the channel so narrow that proteins can pass
We discuss vesicle-based protein trafficking in the next chap- through only in an unfolded state, or will it accommodate
ter because mechanistically it differs significantly from non- folded protein domains?
vesicle-based protein targeting to intracellular organelles.
[n this chapter, we examine how proteins are targeted to 4. What is the source of energy that drives unidirectional
five intracellular organelles: ER, mitochondria, chloroplast, transfer across the membrane?
peroxisome, and nucleus. Two features of this protein-targeting In the first part of the chapter, we cover targeting of pro-
process initially were quite baffling: how a given protein could teins to the ER, including the post-translational modificattons
be directed to only one specific membrane, and how relatively that occur to proteins as they enter the secretory pathway.
large hydrophilic protein molecules could be translocated Targeting of proteins to the ER is the best-understood exam-
across a hydrophobic membrane without disrupting the bilayer ple of protein targeting, and will serve as an exemplar of the
as a barrier to ions and small molecules. Using a combination process in general. We then describe targeting of proteins to
of biochemical purification methods and genetic screens for mitochondria, chloroplasts, and peroxisomes. Finally, we
identifying mutants unable to execute particular translocation cover the transport of proteins into and out of the nucleus
steps, cell biologists have identified many of the cellular com- through nuclear pores.
ponents required for translocation across each of the different
intracell ular membranes. [n addition, many of the major
translocation processes in the cell have been reconstituted 13.1 Targeting Proteins to and Across
using the purified protein components incorporated into arti-
ficial lipid bilayers. Such in vitro systems can be freely the ER Membrane
manipulated experimentally. All eukaryotic cells have an endoplasmic reticulum (ER). The
These studies have shown that, despite some variations, ER is a large, convoluted organelle made up of tubules and flat
the same basic mechanisms govern protein sorting to all the tened sacs, whose membrane is continuous with the mem-
various intracellular organelles. We now know, for instance, brane of the nucleus. The ER membrane is where cellular lipids
that the information to target a protein to a particular or- are synthesized (Chapter 10), and the ER is where most mem-
ganelle destination is encoded within the amino acid se- brane proteins are assembled, including those of the plasma
quence of the protein itself, usually within sequences of membrane and the membrane of the lysosomes, ER, and Golgi.
about 20 am ino acids, known generically as signal sequences ln addition, all soluble proteins that will eventually be secreted
(see Figure 13-1 i; these are also called uptake-targeting se- from the cell-as well as those destined for the lumen of the ER,
quences or signal peptides. Such targeting sequences usually Golgi, or lysosomes-are initially delivered to the ER lumen
occur at the N-terminus of a protein and arc thus the first (see Figure 13-1 ). Since the ER plays such an important role in
part of a protein to be synthesized. More rarely, targeting protein secretion, we refer to the pathway of protein trafficking
sequences can occur at either the C-terminus or within the that flows through the ERas the "secretory pathway." For sim-
interior of a protein sequence. Each organelle carries a set of plicity, we will refer to all proteins initially targeted to the ER
receptor proteins that birrd on ly to specific kinds of signal as "secretory proteins," but keep in mind that not all proteins
sequences, thus ensuring that the information encoded in a that are targeted to the ER are actually secreted from the cell.
signal sequence governs the specificity of targeting. Once a In this first section, we discuss how proteins are initially
protein containing a signal sequence has interacted with the identified as secretory proteins, and how such proteim are
corresponding receptor, the protein chain is transferred to translocated across the ER membrane. We deal first with solu-
some kind of translocation channel that allows the protein ble proteins-those that pass all the way through the ER mem-
to pass into or through the membrane bilayer. The unidirec- brane, into the lumen. In the next section, we discuss integral
tional transfer of a protein into an organelle, without sliding membrane proteins, wh ich are inserted into the ER membrane.
back out into the cytoplasm, is usually achieved by coupling
translocation to an energetically favorable process such as Pulse-Labeling Experiments with Purified ER
hydrolysis of GTP or ATP. Some proteins are subsequently
Membranes Demonstrated That Secreted
sorted further to reach a subcompartment within the target
organelle; such sorting depends on yet other signal sequences Proteins Cross the ER Membrane
and other receptor proteins. Finally, signal sequences often Although all cells secrete a variety of proteins (e.g., extracel-
are removed from the mature protein by specific proteases lular matrix proteins), certain types of cells are specialized for
once translocation across the membrane is completed. secretion of large amounts of specific proteins. Pancreatic

13.1 Targeting Proteins to and Across the ER Membrane 579

 Cytosol ER lumen ER membrane that receives proteins entering the secretory pathway is known
(a) as the rough ER because tt is so densely studded with ribo-
somes that its surface appears morphologically distinct from
other ER membranes (Figure 13-2). From these experiments,
it became clear that du ring or immediately after their synthe-
sis on the ribosome, secretory proteins translocate across the
ER membrane into the lumen of the ER.
To delineate the steps in the translocation process, it was
necessary to isolate the ER from the rest of the cell. Isolation
of intact ER with its delicate lacelike structure and intercon-
nectedness with other organelles is not feasible. However,
Free ribosomes Attached scientists discovered that after cells are homogenized, the
ribosomes
rough ER breaks up into small closed vesicles with ribo-
somes on the outside, termed rough microsomes, which re-
(b)
Free -~ tain most of the biochemical properties of the ER, including
ribosome~
the capability of protein translocation. T he experiments de-

\ ., picted in Figure 13-3, in w hich microsomes isolated from

pulse-labeled cells are treated with a protease, demonstrate
that although secretory proteins are synthesized on ribosomes
bound to the cytosolic face of the ER membrane, the poly-
peptides produced by these ribosomes end up within the
Ribosomal lumen of ER vesicles. Experiments such as this raised the
subunits
question of how polypeptides are recogni.zed as secretory
proteins shortly after their synthesis begins and how a nascent
/ secretory protein is threaded across the ER membrane.

Cytosol
A Hydrophobic N-Terminal Signal Sequence
Targets Nascent Secretory Proteins to the ER
ER membrane
After synthesis of a secretory protein begins on free ribo- .·
ER lumen somes in the cytosol, a 16- to 30-residue ER signa l sequence
in the nascent protein directs the ribosome to the ER mem-
brane and initiates translocation of the growing polypeptide
across the ER membrane (see Figure 13-1, left). An ER signal
FIGURE 13-2 Structure of the roughER. (a) Electron micrograph sequence typica lly is located at the N-terminus of the pro-
of ribosomes attached to the rough ER in a pancreatic acinar cell. tein, the first parr of the protein to be synthesized. The signal
Most of the proteins synthesized by this type of cell are to be secreted sequences of different secretory proteins all contain one or
and are formed on membrane-attached ribosomes. A few membrane- more positively charged amino acids adjacent to a continu-
unattached (free) ribosomes are evident; presumably, these are ous stretch of 6- 12 hydrophobic resid ues (k nown as the hy-
synthesizing cytosolic or other non secretory proteins. (b) Schematic drophobic core), but otherwise they have little in common.
representation of protein synthesis on the ER. Note that membrane-
For most secretory proteins, the signal sequence is cleaved
bound and free cytosolic ribosomes are identical. Membrane-bound
from the protein while it is still elongating on the ribosome;
ribosomes get recruited to the endoplasmic reticulum during protein
thus signal sequences are usually not present in the "mature"
synthesis of a polypeptide containing an ER signal sequence.
[Part (a) courtesy of G. Palade.]
proteins found in cells.
The hydrophobic core of ER signal sequences is essential
for their function. For instance, the specific deletion of sev-
acinar cells, for instance, synthesize large quantities of several eral of the hydrophobic amino acids from a signal sequence
digestive enzymes that are secreted into ducru les that lead to or the introduction of charged amino acids into the hydro-
the intestine. Because such secretory cells contain the organ- phobic core by mutation can abolish the ability of the N-
elles of the secretory pathway (e.g., ER and Golgi) in great terminus of a protein to function as a signal sequence. As a
abundance, they have been widely used in studying this path- consequence, the modifieJ prutt:in remains in the cytosol,
way, including the initial steps that occur at the ER membrane. unable to cross the ER membrane into the lumen. Con-
The sequence of events that occurs immediately after the versely, signal sequences can be added to normally cytosolic
synthesis of a secretory protein were first elucidated by pulse- proteins using recombinant DNA techniques. Provided the
labeling experiments with pancreatic acinar cells. ln such added sequence is sufficiently long and hydrophobic, such a
cells, radioactively labeled amino acids are incorporated into modified cytosolic protein acquires the ability to be translo-
secretory proteins as they are synthesized on ribosomes that cated to the ER lumen. Thus the hydrophobic residues in the
are bound to the surface of the ER. The portion of the ER core of ER signal sequences form a binding site that is critical

580 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 (a) Cell-free protein synthesis; no microsomes present

Add microsome
membranes

N-terminal Completed proteins

signal sequence with signal sequences

No incorporation
into microsomes;
no removal of
signal sequence

(b) Cell-free protein synthesis; microsomes present

Treat with
~ detergent
Mature protein
Cotranslational transport
?: chain without

l:MQ; of protein into microsome

and removal of signal
sequence
· signal sequence

!kc c Add
ccc protease
c • Add
c c protease
c
EXPERIMENTAL IGURE 1 3 Translation and t ranslocation
occur simultaneously. Cell-free experiments demonstrate that
translocation of secretory proteins into microsomes is coupled to
... ~,
c
c
translation. Treatment of microsomes with EDTA, which chelates Mg 2..
ions, strips them of associated ribosomes, allowing isolation of
c ( c ' c \ c'
c ribosome-free microsomes, which are equivalent to ER membranes
• c
.~,.,~ (see Figure 13-3). Protein synthesis is carried out in a cell-free system
/ I 'liiiiU' c
c containing functional ribosomes, tRNAs, ATP, GTP, and cytosolic
Digestion of No digestion of
enzymes to which mRNA encoding a secretory protein is added. The
secretory protein secretory protein
secretory protein is synthesized in the absence of microsomes (a) but is
EXPERIMENTAl FIGURE 13-3 Secretory proteins enter the ER. translocated across the vesicle membrane and loses its signal sequence
Labeling experiments demonstrate that secretory proteins are localized (resulting in a decrease in molecular weight) only if microsomes are
to the ER lumen shortly aher synthesis. Cells are incubated for a brief present during protein synthesis (b).
time with radiolabeled amino acids so that only newly synthesized
proteins become labeled. The cells then are homogenized, fracturing
the plasma membrane and shearing the roughER into small vesicles
called microsomes. Beca use they have bound ribosomes, microsomes function and fate of ER signal sequences. initial experiments
have a much greater buoyant density than other membranous with this system demonstrated that a typical secretory pro-
organelles and can be separated from them by a combination of tein is incorporated into microsomes and has its signal se-
differential and sucrose density-gradient centrifugation (Chapter 9). The quence removed only if the microsomes are present during
purified microsomes are treated with a protease in the presence or
protein synthesis. If microsomes are added to the system
absence of a detergent. The labeled secretory proteins associated with
after protein synthesis is completed, no protein transport
the microsomes are digested by the protease only if the permeability
into the microsomes occurs (Figure 13-4 ). Subsequent ex-
barrier of the microsomal membrane is first destroyed by treatment
with detergent. This finding indicates that the newly made proteins are
periments were designed to determine the precise stage of
inside the microsomes, equivalent to the lumen of the roughER.
protein symhesis at which microsomes must be present in
order for translocation to occur. In these experiments, mi-
crosomes were added to the reaction mixtures at different
for the interaction of signal sequences with the machinery umes after protein synthesis had begun. These experiments
responsible for targeting the protein to the ER membrane. showed that microsomes must be added before the first 70
Biochemical studies utilizing a cell-free protein-synthesizing or so amino acids are translated in order for the completed
system, mRNA encoding a secretory protein, and microsomes secretory protein to be localized in the microsomal lumen.
stripped of their own bound ribosomes have elucidated the At this point, the first 40 amino acids or so protrude from

13.1 Targeting Proteins to and Across the ER Membrane 581

0 PODCAST: Structure and Function of the Signal-Recognition Particle in Protein Translocation
FIGURE 13- 5 Structure of the signal-recognition particle (SRP). (a) Ffh signal-sequence-binding domain
(a) Signal·sequence binding domain: The bacterial Ffh protein is (related to P54 subunit of SRP)
homologous to the portion of P54 that binds ER signal sequences. This
surface model shows the binding domain in Ffh, which contains a large
cleft lined with hydrophobic amino acids (purple) whose side chains
interact with signal sequences. (b) GTP- and receptor-binding domain:
The structure of GTP bound to FtsY (the bdcterial homolog of the ex
subunit of SRP receptor) and Ffh proteins illustrates how the interac-
tion between these proteins is controlled by GTP binding and
hydrolysis. Ffh and FtsY each can bind to one molecule of GTP, and
when Ffh and FtsY bind to each other, the two bound molecules of
GTP fit in the interface between the protein subunits and stabilize the
dimer. Assembly of the semisymmetrical dimer allows formation of two
active sites for the hydrolysis of both bound GTP molecules. Hydrolysis
to GDP destabilizes the interface, causing disassembly of the dimer.
[Part (a) adapted from R. J. Keenan et al., 1998, Cell 94:181. Part (b) adapted from
P. J. Focia et al., 2004, Science 303:373.]
Hydrophobic
binding groove

(b)
FtsY Ffh
(SRP receptor ex subunit) GTP (SRP P54 subunit)

the ribosome, including the signal sequence that later will be

cleaved off, and the next 30 o r so ami no acids are still bu ried
within a channel in the ribosome (see Figure 4-26). Thus the
transport of most secretory proteins into the ER lumen begins
while the incompletely synthesized (nascent) protein is still
bound to the ribosome, a process referred to as cotranslational
translocation.

Cotranslational Translocation Is Initiated

by Two GTP-Hydrolyzing Proteins
Since secretory proteins are synthesized in association with GTP
the ER membrane but not with any other cellular membrane,
a signal-sequence recognition mechanism must target them
there. The two key components in t his targeting are the signal-
recognition particle (SRP) and its receptor, located in the ER
membrane. The SRP is a cytosolic ribonucleoprotein particle
that transiently binds to both the ER signa l sequence in a The SRP brings the nascent chain-ribosome complex to the
nascent protein as well as the large ribosomal subunit, form- ER membrane by docking with the SRP receptor, an integral
ing a large complex; SRP then targets t he nascent protein- protein of t he ER membrane made up of two subunits: an a
ribosome complex to the ER membrane by binding to the subunit and a smaller 13 subunit. Interaction of the SRP/nascem
SRP receptor on the membrane. chain/ribosome complex with the SRP receptor is strengthened
The SRP is made up of six proteins bound to a 300- when both the P54 subunit of SRP and the a subunit of the
nucleotide RNA, which acts as a scaffold for the hexamer. SRP receptor are bound to GTP. The structure of the P54 sub-
One of the SRP proteins (P54) can be chemically cross-linked unit of SRP and the SRP receptor a subunit (FtsY), from the
to ER signal sequences, showing that this is the subunit that archae bacteria Thermus aquaticus, provides insight into how a
binds to the signal sequence in a nascent secretory protein. A cycle of GTP binding and hydrolysis can drive the binding and
region uf P54 known as theM doma in, containing many me- dtssoctatlon ot these proteins. Figure 13-5b shows that the P54
thionine and other amino acid residues with hydrophobic side and Fts Y each bound to a single molecule of GTP come to-
chains, contains a cleft whose inner surface is lined by hydro- gether to fo rm a pseudo-symmetrical heterodimer. Neither
phobic side chains (Figure 13-Sa). The hydrophobic core of the subunit alone contains a complete active site for the hydrolysis
signal peptide binds to this cleft via hydrophobic interactions. of GTP, but when rhe two proteins come together, they form
Other polypeptides in the SRP interact with the ribosome or two complete active sites that are capable of hydrolyzi ng both
are required for protein translocation into the ER lumen. bound GTP molecules.

582 CHAPTER 13 • Moving Proteins into Membranes and Organelles

0 FOCUS ANIMATION: Synthesis of Secreted and Membrane-Bound Proteins

m RNA

5'

Translocon Translocon
(closed) (open) Signal
peptidase -" '
Cleaved
signal
sequence Folded
protein

FIGURE 13-6 Cotranslational translocation. Steps D . fJ : Once the GTP and then are ready to initiate the insertion of another polypeptide
ER signal sequence emerges from the ribosome, it is bound by a chain. Step ~ : As the polypeptide chain elongates, it passes through
signal-recognition particle (SRP). Step D : The SRP delivers the the translocon channel into the ER lumen, where the signal sequence is
ribosome/nascent polypeptide complex to the SRP receptor in the ER cleaved by signal peptidase and is rapidly degraded. Step m: The
membrane. This interaction is strengthened by binding of GTP to both peptide chain continues to elongate as the mRNA is translated toward
the SRP and its receptor. Step EJ:Transfer of the ribosome/nascent the 3' end. Because the ribosome is attached to the translocon, the
polypeptide to the translocon leads to opening of this translocation growing chain is extruded through the translocon into the ER lumen.
channel and insertion of the signal sequence and adjacent segment of Steps fl. rlJ: Once translation is complete, the ribosome is released, the
the growing polypeptide into the central pore. Both the SRP and SRP remainder of the protein is drawn into the ER lumen, the translocon
receptor, once diss.ociated from the translocon, hydrolyze their bound closes, and the protein assumes its native folded conformation.

Figure 13-6 summarizes our current understanding of se- complex of proteins that forms a channel embedded within
cretory protein synthesis and the role of the SRP and its re- the ER membrane. As translation continues, the elongating
ceptor in this process. Hydrolysis of the bound GTP chain passes directly from the large ribosomal subunit into
accompanies disassembly, of the SRP and SRP receptor and, the central pore of the translocon. The 60S ribosomal subunit
in a manner that is not understood, initiates transfer of the is aligned with the pore of the translocon in such a way that
nascent chain and ribosome to a site on the ER membrane, the growing chain is never exposed to the cytoplasm and is
where translocation can take place. After dissociating from prevented from folding until it reaches the ER lumen (see
each other, SRP and its receptor each release their bound Figure 13-6).
GDP, SRP recycles back to the cytosol, and both are ready to The translocon was first identified by mutations in the
initiate another round of interaction between ribosomes syn- yeast gene encoding Sec61ex, which caused a block in the
thesizing nascent secretory proteins and the ER membrane. translocation of secretory proteins into the lumen of the ER.
Subsequently, three proteins called the Sec61 complr>x were
found to form the mammalian translocon: Sec61ex, an inte-
Passage of Growing Polypeptides Through
gral membrane protein with 10 membrane-spanning ex heli-
the Translocon Is Driven by Translation ces, and two smaller proteins, termed Sec61 j3 and Sec61 -y.
Once the SRP and its receptor have targeted a ribosome synthe- Chemical cross-linking experiments-in which amino acid
sizing a secretory protein to the ER membrane, the ribosome side chains from a nascent secretory protein can become co-
and nascent chain are rapidly transferred to the translocon, a valently attached to the Sec61 ex subunit-demonstrated that

13.1 Targeting Proteins to and Across the ER Membrane 583

 elongation at the ribosome appears to be sufficient to push the
polypeptide chain across the membrane in one direction.
The translocon must be able to allow passage of a wide
-40S}
variety of polypeptide sequences whi le remaining sealed to
Ribosome
small molecules such as ATP, in order to maintain the perme-
--60S ability barrier of the ER membrane. Furthermore, there must
Cytosol be some way to regulate the translocon so that it is closed in
its default state, opening only when a ribosome-nascent chain
complex is bound. A high-resolution structure of the Sec6 1
Sec61u complex from archaebacteria shows how the translocon pre-
serves the integrity of the membrane (Figure 13-8). The 10
membrane Cross- transmembrane helices of Sec61 a form a centra l channel
linking through which the translocating peptide chain passes. A con-
agent
striction in the middle of the central pore is lined with hydro-
phobic isoleucine residues that in effect form a gasket around
the translocating peptide. The structura l model of the Sec61
Microsomal
lumen complex, which was isolated with~ut a translocating peptide
and therefore is presumed to be in a closed confo rmation,
reveals a short helical peptide plugging the central channel.
x'PER'Ii\11E NTAL F'I GURE 3 -7 Sec61a is a translocon compo- Biochemical studies of the Sec61 complex have shown that,
nent. Cross-linking experiments show that Sec61 a is a translocon in the absence of a translocating polypeptide, the peptide that
component that contacts nascent secretory proteins as they pass into forms the plug effectively seals the translocon to prevent pas-
the ER lumen. An mRNA encoding theN-terminal 70 amino acids of sage of ions and small molecules. Once a translocating pep-
the secreted protein prolactin was translated in a cell-free system
tide enters the channel, the plug peptide swings away to allow
containing microsomes (see Figure 13-4b). The mRNA lacked a
translocation to proceed.
chain-termination codon and contained one lysine codon, near the
As the growing polypeptide chain enters the lumen of the
middle of the sequence. The reactions contained a chemically modified
lysyl-tRNA in which a light-activated cross-linking reagent was attached
ER, the signal sequence is cleaved by signal peptidase, which
to the lysine side chain. Although the entire mRNA was translated, the is a transmembrane ER protein associated w ith the translo- ·.
completed polypeptide cou ld not be released from the ribosome con (see Figure 13-6, step ~). Signal peptidase recognizes a
without a chain-termination codon and thus became "stuck" crossing sequence on the C-terminal side of the hydrophobic core of
the ER membrane. The reaction mixtures then were exposed to an the signal peptide and cleaves th e chain specifica ll y at this
intense light, causing the nascent chain to become covalently bound sequence once it has emerged into the luminal space of the
to whatever proteins were near it in the translocon. When the ER. After the signal sequence has been cleaved, the growing
experiment was performed using micro somes from mammalian cells, po lypeptide moves through the translocon into the ER
the nascent chain became covalently linked to Sec61 a. Different lumen. The translocon remains open until trans lation is
versions of the prolactin mRNA were created so that the modified completed and the entire polypeptide chain has moved into
lysine residue would be placed at different distances from the ribosome; the ER lumen. After translocation is complete, t he plug helix
cross-linking to Sec61a was observed only when the modified lysine returns to the pore to reseal the tra nslocon channel.
was positioned within the translocation channel. (Adapted from
T. A. Rapoport, 1992, Science 258:931, and D. Gorlich and T. A. Rapoport, 1993,
Cell 75:61 5.]
ATP Hydrolysis Powers Post-translational
Translocation of Some Secretory Proteins
in Yeast
In most eukaryotes, secretory proteins enter the ER by co-
the translocating polypeptide chain comes into contact with translational translocation. In yeast, however, some secre-
the Sec6la protein, confirming its identity as the translocon tory proteins enter the ER lumen after translation has been
pore (Figure 13-7) . completed. In such post-translational translocation, the
When microsomes in the cell-free translocation system translocating protein passes through the same Sec61 translo-
were replaced with reconstituted phospholipid vesicles con- con that is used in cotranslational t ranslocation . However,
taining only the SRP receptor and Sec61 complex, na~cenr '>e- the SRP and SRP receptor are not involved in post-translational
cretory protein was translocated from its SRP/ribosome translocation, and in such cases a direct interaction between
complex into the vesicles. This finding indicates that the SRP the translocon and the signal seq uence of the completed pro-
receptor and the Sec61 complex are the only ER-membrane tein appears to be sufficient fo r targeti ng to the ER mem-
proteins that are absolutely required for translocation. Because brane. In addition, the driving force for unidirectional
neither of these can hydrolyze ATP or otherwise provide en- translocation across the ER membrane is provided by an
ergy to drive the translocation, the energy derived from chain additional protein complex known as the Sec63 complex

584 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 0 VIDEO: Three-Dimensional Model of a Protein Translocation Channel

EXPERIMENTAL FIGURE 13-8 Structure of a bacterial Sec61 (a) Side view

complex. The structure of the detergent-solubilized Sec61 complex
from the archaebacterium M.jannaschii (also known as the SecY
complex) was determined by x-ray crystallography. (a) A side view
shows the hourglass-shaped channel through the center of the pore.
A ring of isoleucine residues at the constricted waist of the pore forms
a gasket that keeps the channel sealed to small molecules even as a
translocating polypeptide passes through the channel. When no trans-
locating peptide is present, the channel is closed by a short helical
plug (red). This plug moves out of the channel during translocation. In
this view the front half of protein has been removed to better show
the pore. (b) A view looking through the center of the channel shows a
region (on the left side) where helices may separate, allowing lateral
passage of a hydrophobic transmembrane domain into the lipid
bilayer. [Adapted from A. R. Osborne et al., 2005, Ann. Rev. Cell Dev. Bioi.
21:529.)

(b) Top view Pore ring

and a member of the Hsc70 family of molecular chaperones

known as BiP (see Chapter 3 for further discussion of mo-
lecular chaperones). The tetrameric Sec63 complex is em-
bedded in the ER membrane in the vicinity of the translocon,
whereas BiP is within the ER lumen. Like other members of
the Hsc70 family, BiP has a peptide-binding domain and an
ATPase domain. These chaperones bind and stabilize un-
folded or partially folded proteins (see Figure 3-16).
The current model for post-translational translocation of
a protein into the ER is outlined in Figure 13-9. Once theN-
terminal segment of the protein enters the ER lumen, signal
'· peptidase cleaves the signal sequence just as in cotranslational
translocation (step 0 ). Interaction of BiP· ATP with the lumi-
nal portion of the Sec63 complex causes hydrolysis of the
bound ATP, producing a conformational change in BiP that their role in this process are not well understood, but they are
promotes its binding to an exposed polypeptide chain (step f)). thought to act at an early stage of the process, such as thread-
Since the Sec63 complex is located near the translocon, BiP is ing the signal peptide into the pore of the translocon.
thus activated at sites where nascent polypeptides can enter The overall reaction carried out by BiP is an important
the ER. Certain experiments suggest that, in the absence of example of how the chemical energy released by the h ydro-
binding to BiP, an unfolded polypeptide slides back and forth lysis of ATP can power the mechanical movement of a pro-
within the translocon chMlnel. Such random sliding motions tein across a membrane. Bacterial cells also use an ATP-driven
rarely result in the entire polypeptide's crossing the ER mem- process for rranslocating completed proteins across the
brane. Binding of a molecule of BiP·ADP to the luminal por- plasma membrane-in this case to be released from the cell.
tion of the polypeptide prevents backsliding of the polypeptide In bacteria the driving force for translocation comes from a
out of the ER. As further inward random sliding exposes cytosolic ATPase known as the SecA protein. SecA binds to
more of the polypeptide on the luminal side of the ER mem- the cytoplasmic side of the translocon and hydrolyzes cyro-
brane, successive binding of BiP·ADP molecules to the poly- solic ATP. By a mechanism that resembles the needle on a
peptide chain acts as a ratchet, ultimately drawing the entire sewing machine, the SecA protein pushes segments of the
polypeptide into the ER within a few seconds (steps IJ and 19). polypeptide through the membrane in a mechanical cycle
On a slower time scale, the BiP molecules spontaneously ex- coupled to the hydrolysis of ATP.
change their bound ADP for ATP, leading to release of the As we will see, translocation of proteins across other eu-
polypeptide, which can then fold into its native conformation karyotic organelle membranes, such as those of mitochondria
(steps ~ and ~). The recycled BiP·ATP then is ready for an- and chloroplasts, also typically occurs by post-translational
other interaction with Sec63. BiP and the Sec63 complex are translocation. This explains why ribosomes are not found
also required for cotranslational translocation. The details of bound to these other organelles, as they are to the rough ER.

13.1 Targeting Proteins to and Across the ER Membrane 585

 Translocating I
polypeptide
chain

Cleaved
signal \
sequence

FIGURE 13-9 Post-translational translocation. This mechanism

is fairly common in yeast and probably occurs occasionally in higher
eukaryotes. Small arrows inside the translocon represent random
sliding ofthe translocating polypeptide inward and outward. Succes-
sive binding of BiP·ADP to entering segments of the polypeptide
prevents the chain from sliding out toward the cytosol. [SeeK. E. Matlack
et al., 1997, SCience 277:938.]

KEY CONCEPTS of Section 13.1 translation, the unfolded protein chain is extruded into the ER
lumen. No additional energy is required for translocation.
Targeting Proteins to and Across the ER Membrane
• Synthesis of secreted proteins, integral plasma-membrane The translocon contains a central channel lined with hy-
proteins, and proteins destined for the ER, Golgi complex, drophobic residues that allows transit of an unfolded protein
or lysosome begins on cyrosolic ribosomes, which become chain while remaining sealed to ions and small hydrophilic
attached to the membrane of the ER, forming the rough ER molecules. In addition, the channel is gated so that it is open
(see Figure I 3-l, left). only when a polypeptide is being translocated.

The ER signal sequence on a nascent secretory protein • In post-translational translocation, a completed secretory
consists of a segment of hydrophobic amino acids located at protein is targeted to the ER membrane by interaction of the
the N-terminus. signal sequence with the translocon. The polypeptide chain
In cotranslational translocation, the signal-recognition is then pulled into the ER by a ratcheting mechanism that
particle (SRP) first recognizes and binds the ER signal se- requires ATP hydrolysis by the chaperone BiP, which stabi-
quence on a nascent secretory protein and in turn is bound lizes the entering polypeptide (see Figure 13-9). In bacteria,
by an SRP receptor on the ER membrane, thereby targeting the driving force for post-translational translocation comes
the ribosome/nascent chain complex to the ER. from SecA, a cytosolic ATPase that pushes polypeptides
through the translocon channel.
• The SRP and SRP receptor then mediate insertion of the
nascent secretory protein into the translocon (Sec61 com- In both cotranslational and post-translational transloca-
plex). Hydrolysis of two molecules of GTP by the SRP and its tion, a signal peptidase in the ER membrane cleaves the ER
receptor cause the dissociation of SRP (see Figures 13-5 and signal sequence from a secretory protein soon after the N-
13-6). As the ribosome attached to the translocon continues terminus enters the lumen.

586 CHAPTER 13 • Moving Protems into Membranes and Organelles

13.2 Insertion of Membrane Proteins Several Topological Classes of Integral
into the ER Membrane Proteins Are Synthesized on the ER
The topology of a membrane protein refers to the number of
In previous chapters we have encountered many of the vast
times that its polypeptide chain spa ns the membrane and the
array of integral (transmembrane) proteins that are present
orientation of these membrane-spanning segments within the
throughout the cell. Each such protein ha~ a unique orien-
membrane. The key elements of a protein that determine its
tation with re~pect to the membrane's phospholipid bi-
topology are membrane-spanning segments themselves, which
layer. Integral membrane proteins located in the ER, Golgi,
usually arc a helices containing 20-2) hydrophobic amino
and lysosomes, and ::~lso proteins in the plasma membrane,
acids that contribute to energetically favorable interactions
which are all synthesized on the rough ER, remain embed-
within the hydrophobic interior of the phospholipid bilayer.
ded in the membrane in their unique orientation as they
Most integral membrane protems fall into one of the five
move to their final destinations along the same pathway
topological classes illustrated in Figure 13-l 0. Topological
followed by soluble secretory proteins (see figure 13-l,
classes I, II, III, and tail-anchored proteins comprise single-pass
left). During this transport, the orientation of a membrane
proteins, which have only one membrane-spanning a-helical
protein is preserved; that is, the same segments of the pro-
segment. Type I proteins have a cleaved N-terminal ER signal
tein always face the cytosol, whereas other segments always
sequence and are anchored in the membrane with their hydro-
face in the opposite direction. Thus the final orientation of
philic N-terminal region on the luminal face (also known as
these membrane proteins is established during their biosyn -
the exoplasmic face) and their hydrophilic C-terminal region
thesis on the ER membrane. In this section, we first see
on the cytosolic face. Type II proteins do not contain a cleav-
how integral proteins can interact with membranes and
able ER signal sequence and are oriented with their hydro-
then examine how several types of ~equences, known col-
philic N-terminal region on the cyrosolic face and their
lectively as topogenic sequences, direct the membrane in-
hydrophilic C-tcrminal region on the exoplasmic face (i.e., op-
sertion and orientation of various classes of integral
posite to type I proteins). Type III proteins have a hydrophobic
proteins. These processes occur via modifications of the
membrane-spanning segment at their N-tcrminus and thus
basic mechanism used to translocatc soluble secretory pro-
have the same orientation as type I proteins but do not contain
teins across the ER membrane.

coo

Cytosol

Exoplasmic
space
(ER or Golgi signal
lumen; sequence
cell exterior) I
Type I Type II Type Ill Tail-anchored protein Type IV GPI-Iinked protein

LDL receptor Asialoglycoprotein Cytochrome v-SNARE and G protein-coupled receptors Plasminogen

Influenza receptor P450 t-SNARE Glucose transporters activator receptor
HA protein Transferrin receptor Voltage-gated Ca2+ channels Fasciclin II
Insulin receptor Golgi galactosyl- ABC small molecule pumps
transferase
Growth hormone CFTR (CI-) channel
receptor Golgi sialyltrans-
Sec61
ferase
FIGURE 13- 10 ER membrane proteins. Five topological classes of type IV proteins have multiple transmembrane o: helices. The type IV
integral membrane proteim are synthesized on the rough ER <:~swell topology depicted here corresponds to that of G protein-coupled
as a sixth type tethered to the membrane by a phospholipid anchor. receptors: seven o: helices, theN-terminus on the exoplasmic side
Membrane proteins are classified by their orientation in the membrane of the membrane, and the (-terminus on the cytosolic side. Other
and the types of signals they contain to direct them there. For the type IV proteins may have a different number of helices and various
integral membrane proteins, hydrophobic segments of the protein chain orientations of theN-terminus and (-terminus. [See E. Hartmann et al.,
form o: helices embedded in the membrane bilayer; the regions outside 1989, Proc. Nat'/Acad.Sci. USA 86:5786, and C. A. Brown and S.D. Black, 1989,
the membrane are hydrophilic and fold into various conformations. All J. Bioi. Chern. 264 :4442.]

13.2 Insertion of Membrane Proteins into the ER 587

 .·

a cleavable signal sequence. Finally, tail-anchored proteins sequences that are used to direct proteins to the ER mem-
have a hydrophobic segment at their C-terminus that spans the brane and to orient them within it. We have already intro-
membrane. These different topologies reflect distinct mecha- duced one, th e N-term inal signal sequence. The other two,
nisms used by the cell to establish the membrane orientation of introduced here, are internal sequences known as stop-transfer
transmembrane segments, as discussed in the next section. anchor sequences and signal-anchor sequences. Unlike signal
The proteins forming topological class IV contain two or sequences, the two types of internal topogenic sequences end
more membrane-spanning segments and are sometimes called up in the mature protein as membrane-spanning segments.
multipass proteins. For example, many of the membrane However, the two types of internal topogenic sequences differ
transport proteins discussed in Chapter 11 and the numerou~ in their final o rientation in the membrane.
G protein-coupled receptors covered in Chapter 15 belong to
this class. A final type of membrane protein lacks a hydro- Type I Proteins All type I transmembrane proteins possess
phobic membrane-spanning segment altogether; instead, an N-terminal signal sequence that ta rgets them to the ERas
these proteins are linked to an amphipathic phospholipid an- well as an internal hydrophobic sequence that becomes the
chor that is embedded in the membrane (Figure 13-10, right). membrane-spanning a helix. The N-terminal signal sequence
on a nascent type I protein, like that of a soluble secretory
Internal Stop-Transfer and Signal-Anchor protein, initiates cotranslational translocation of the protein
through the combined action of tl'le SRP and SRP receptor.
Sequences Determine Topology
Once the N-terminus of the growing polypeptide enters the
of Single-Pass Proteins lumen of the ER, the signal sequence is cleaved, and the grow-
We begm our discussion of how membrane protein topology ing chain continues to be extruded across the ER membrane.
is determined with the membrane insertion of integral pro- However, unlike the case with soluble secretory proteins, when
teins that contain a single, hydrophobic membrane-spanning the sequence of approximately 22 hydrophobic ami no acids
segment. Two sequences are involved in targeting and orient- that will become a transmembrane domaip of the nascent
ing type I proteins in the ER membrane, whereas type II and chain enters the translocon, it stops transfer of the protein
type III proteins contain a single, internal topogenic se- through the channel (Figure 13- 11). The Sec61 complex is
quence. As we will see, there are three main types of topogcnic then able to open like a clamshell, allowing the hydrophobic

Cytosol
Q
3'

Open
translocon

ER lumen
sequence

FIGURE 13-11 Positioning type I single-pass proteins. Step 0 : transfer anchor sequence moves laterally between the translocon
After the ribosome/nascent cham complex becomes assotidted with subunits and becomes anchored in the phospholipid bilayer. At this
a translocon in the ER membrane, the N-terminal signal sequence is time, the translocon probably closes. Step Ill : As synthesis continues,
cleaved. This process occurs by the same mechanism as the one for the elongating chain may loop out into the cytoso l through the small
soluble secretory proteins (see Figure 13-6). Steps fJ. IJ:The chain is space between the ribosome and translocon. Step r;'l : When synthesis
elongated until the hydrophobic stop-transfer anchor sequence is is complete, the ribosomal subunits are released into the cytosol,
synthesized and enters the translocon, where it prevents the nascent leaving the protein free to diffuse in the membrane. [See H. Do et al.,
chain from extruding farther into the ER lumen. Step B :The stop- 1996, Ce// 85:369, and W. Mothes et al., 1997, Ce// 89:523.)

588 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 transmembrane segment of the translocating peptide to move single a-helical membrane-spanning segment or that is missing
laterally benveen the protein domains constituting the translo- most of this segment is translocated entirely into the ER lumen
con wall (see Figure 13-8). When the peptide exits the translo- and is eventually secreted from the cell as a soluble protein.
con in this manner, it becomes anchored in the phospholipid These kinds of experiments establish that the hydrophobic
bilayer of the membrane. Because of the dual function of such membrane-spanning a helix of the HGH receptor and of other
a sequence to both stop passage of the polypeptide chain type I proteins functions both as a stop-transfer sequence and
t hrough the translocon and to become a hydrophobic trans- as a membrane anchor that prevents the C-terminus of the pro-
membrane segment in the membrane bilayer, it is called a stop- tein from crossing the ER membrane.
transfer anchor sequence.
Once translocation is interrupted, translation continues Type II and Type Ill Proteins Unlike type I proteins, type II
at the ribosome, which is still anchored to the now unoccu- and type lll proteins lack a cleavable N-terminal ER signal
pied and closed rranslocon. As the C-rerminus of the protein sequence. Instead, both possess a single internal hydrophobic
chain is synthesized, it loops out on the cyrosolic side of the signal-anchor sequence that functions as both an ER signal
membrane. When translation is completed, the ribosome is sequence and a membrane-anchor sequence. Recall that type
re leased from the translocon and the C-terminus of the II and type II1 proteins have opposite orientations in the mem-
newly synthesized type I protein remains in the cytosol. brane (see Figure 13-1 0); this difference depends on the orien-
Support for this mechanism has come from studies in tation that their respective signal-anchor sequences assume
which cDNAs encoding various mutant receptors for human within the translocon. The internal signal-anchor sequence in
growth hormone (HGH) are expressed in cultured mammalian type II proteins directs insertion of the nascent chain into the
cells. The wild-type HGH receptor, a typical type I protein, is ER membrane so that the N-terminus of the chain faces the
transported normally to the plasma membrane. However, a cytosol, using the same SRP-dependent mechanism described
mutant receptor that has charged residues inserted into the for signal sequences (Figure 13-12a). However, the internal

(a) (b)
11
Nascent
polypeptide
chain

Cytosol II D 11 coo

Translocon

ER lumen
sequence

FIGURE 13- 12 Positioning type II and type Ill single-pass and anchors the chain in the phospholipid bilayer. Step ID: Once
proteins. (a) Type II protein s. Step 0 : After the internal signal-anchor protein synthesis is completed, the (-terminus of the polypeptide is
sequence is synthesized on a cytosolic ribosome, it is bound by an SRP released into the lumen, and the ribosomal subunits are released into
(not shown), which directs the ribosome/nascent chain complex to the the cytosol. (b) Type Ill proteins. Step 0 : Assembly is by a similar
·. ER membrane. This is similar to targeting of soluble secretory proteins pathway to that of type II proteins except that positively charged
except that the hydrophobic signal sequence is not locilted at the residues on the (-terminal side of the signdl-anchor sequence cause
N-terminus and is not subsequently cleaved. The nascent chain the transmembrane segment to be oriented within the translocon with
becomes oriented in t he translocon with its N-terminal portion toward its (-terminal portion oriented to the cytosol and the N-terminal side of
the cytosol. This orien tation is believed to be mediated by the the protein in the ER lumen. Steps fJ, ID:Chain elongation of the
positively charged residues shown N-terminal to the signal-anchor (-terminal portion ofthe protein is completed in the cytosol, and
sequence. Step fJ: As the chain is elongated and extruded into the ribosomal subunits are released. [SeeM. Spiess and H. F. Lodish, 1986, Cell
lumen, the internal signal-anchor moves laterally out of the translocon 44:177, and H. Do et al., 1996, Ce// 85:369.]

13.2 Insertion of Membrane Proteins into the ER 589

signal-anchor sequence is not cleaved and moves laterally be- coat of influenza virus. Three arginine residues are located
tween the protein domains of the translocon wall into the just N-terminal to the internal signal-anchor sequence in
phospholipid bilayer, where it functions as a membrane an- neuraminidase. M utation of these three positively charged
chor. As elongation continues, the C-term inal region of the residues to negatively charged glutamate residues causes
growing chain is extruded through the translocon into the ER neuraminidase to acquire the reverse orientation. Similar ex-
lumen by cotranslational translocation. periments have shown that other proteins, with either type II
In the case of type III proteins, the signa l-anchor se- or type III orientation, can be made to "flip" their orienta-
quence, which is located near theN-terminus, inserts the na- tion in the ER membrane by mutating charged residues that
scent chain into the ER membrane with its N-tcrminus facing flank the inLt:rnal signal-anchor segment.
the lumen, in the opposite orientation of the signal anchor in
type II proteins. The signal-anchor sequence of type III pro- Tail-Anchored Proteins For all topological classes of proteins
teins also functions like a stop-transfer sequence and prevents we have considered so far, membrane insertion begins when
further extrusion of the nascent chain into the ER lumen (Fig- SRP recognizes a hydrophobic topogenic peptide as it emerges
ure 13-12b). Continued elongation of the chain C-terminal to from the ribosome. Recognition of tail-anchored proteins,
the signal-anchor/stop-transfer sequence proceeds as it does which have a single hydrophobic topogenic sequence at the
for type I proteins, with the hydrophobic sequence movi ng C-terminus, present a unique challenge since the hydrophobic
laterally between the translocon subunits to anchor the poly- C-terminus only becomes available lor recognition after com-
peptide m the ER membrane (see Figure 13-11). pletion of translation and the protein has been released from
One of the features of signal-anchor sequences that ap- the ribosome. Insertion of tail-anchored proteins into the ER
pears to determine their insertion orientation is a high density membrane does not employ SRP, SRP receptor, or the trans-
of positively charged amino acids adjacent to one end of the locon, but instead depends on a pathway dedicated for this
hydrophobic segment. These positively charged residues tend purpose as depicted in Figure 13-13. Targeting of tail-anchored
to remain on the cytosolic side of the membrane, not travers- proteins involves an ATPase known as Get3, which binds to
ing the membrane into the ER lumen. Thus the position of the C-terminal hydrophobic segment of tail-anchored pro-
the charged residues dictates the orientation of the signal- teins. The complex of Get3 bound to a tai l-anchored protein
anchor sequence within the translocon as well as whether the is recruited to the ER by a dimeric integral membrane receptor
rest of the polypeptide chain continues to pass into the ER known as Getl/Get2, and the tail-anchored protein is released
lumen: type II proteins tend to have positively charged resi- from Get3 for insertion into the membrane. Insertion of tail-
dues on the N-terminal side of their signal-anchor sequence, anchored protein into the ER membranes by Get proteins
orienting the N-terminus in the cytosol and allowing passage shares fundamental mechanistic similarities to the targeting of
of the C-terminal side into the ER (Figure 13-12a), whereas signal sequence-bearing proteins to the ER by SRP and SRP
type ill proteins tend to have positively charged residues on the receptor. Two major differences between the two targeting
C-terminal side of their signal-anchor sequence, inserting the processes are that, after release from Get3, tail-anchored pro-
N-terminus into the translocon and restricting the C-terminus teins may be inserted directly into the membrane bilayer,
to the cytosol (Fi.gure 13-12b). whereas SRP transfers a signal sequence to the translocon,
A striking experimental demonstration of the importance and that Get3 couples targeting and transfer of tail-anchored
of the flanking charge in determining membrane orientation proteins to ATP hydrolysis, whereas SRP couples secretory
is provided by neuraminidase, a type II protein in the surface protein targeting to GTP hydrolysis.

D
FIGURE 13- 13 Insertion of tail-anchored proteins. For
(-terminal tail-anchored proteins the hydrophobic (-terminus is not
~~-n-.J
Hydrophobic
Get3

available for membrane insertion until protein synthesis is complete C-terminal tail
and the protein has been released from the ribosome. Step 0 : Get3

D~ATP
in an ATP-bound state binds to the hydrophobic (-terminal tail. This
binding reaction is facilitated by a complex of three proteins, Sgt2,
Get4, and GetS, which sequester the hydrophobic (-terminal tail
before transferring it to Get3·ATP (not shown). Step FJ: The ternary
complex Get3·ATP bound to the (-terminal tail docks onto the Getl
and Get2 proteins, which are embedded in the ER membrane.
Step iD: In succession, ATP is hydrolyzed and ADP is released from
Cytosol IJ
____. ~
~
. . i~ADP
-
DP P,
~ '· -

ER membrane ' I
Get3. At the same time, the hydrophobic (-terminal tail is released
from Get3 and becomes embedded in the ER membrane. ~•.., -"•-... r't~. t' • "' ·•"1!•l
Step 19:Get3 binds to ATP and Get3·ATP is released from the ER lumen Get1 Get2
complex of Getl and Get2 in a soluble form, ready for another round
of binding to a hydrophobic ( -terminal tail.

590 CHAPTER 13 • Moving Proteins into Membranes and Organelles

.'· Multipass Proteins Have Multiple initiates insertion of the nascent chain into the ER membrane
Internal Topogenic Sequences with the N-terminus oriented toward the cytosol; thus this
a-helical segment functions like the internal signal-anchor
figure 13-14 summarizes the arrangements of topogenic sequence of a type II protein (see Figure 13-12a). As the na-
sequences in single-pass and multipass transmembrane pro- scent chain following the first a helix elongates, it moves
teins. In multipass (type IV) proteins, each of the membrane- through the translocon until the second hydrophobic a helix
spanning a helices acts as a topogenic sequence in the ways is formed. This helix prevents further extrusion of the na-
that we have already discussed: they can act to direct the scent chain through the translocon; thus its function is simi-
protein to the ER, to anchor the protein in the ER membrane, lar to that of the stop-transfer anchor sequence in a type I
or to stop transfer of tbe protein through the membrane. protein (see Figure 13-11).
Multipass proteins fall into one of two types, depending on After synthesis of the first two transmembrane a helices,
whether the N-terminus extends into the cytosol or the both ends of the nascent chain face the cytosol and the loop
exoplasmic space (e.g., the ER lumen, cell exterior). This N- between them extends into the ER lumen. The C-terminus of
terminal topology usually is determined by the hydrophobic the nascent chain then continues to grow into the cytosol, as
segment closest to the N-terminus and the charge of the se- it does in synthesis of type I and type III proteins. According
quences flanking it. If a type IV protein has an even number of to this mechanism, the third a helix acts as another type II
.· transmembrane a helices, both its N-terminus and C-terminus signal-anchor sequence and the fourth as another stop-transfer
will be oriented toward the same side of the membrane (Fig- anchor sequence (Figure 13-14d). Apparently, once the first
ure 13-14d). Conversely, if a type IV protein has an odd num- topogenic sequence of a multipass polypeptide initiates asso-
ber of a helices, its two ends will have opposite orientations ciation with the translocon, the ribosome remains attached to
(Figure 13-14e). the translocon, and topogenic sequences that subsequently
emerge from the ribosome are threaded into the translocon
Type IV Proteins with N-Te rm inus in Cytosol Among the without the need for the SRP and the SRP receptor.
multipass proteins whose N-terminus extends into the cyto- Experiments that use recombinant DNA techmques to
sol are the various glucose transporters (GLUTs) and most exchange hydrophobic a helices have provided insight into
ion-channel proteins, discussed in Chapter 11. In these pro- the functioning of the topogenic sequences in type IV-A mul-
teins, the hydrophobic segment closest to the N-terminus tipass proteins. These experiments indicate that the order of

FIGURE 13-14 Topogenic sequences determine

orientation of ER membrane proteins. Topogenic
sequences are shown in red; soluble, hydrophilic portions
in blue. The internal topogenic sequences form transmem-
STA = Internal stop-transfer anchor sequence brane a helices that anchor the proteins or segments of
SA =Internal signal-anchor sequence
proteins in the membrane. (a) Type I proteins contain a
cleaved signal sequence and a single internal stop-transfer
anchor (STA). (b, c) Type II and type Ill proteins contain a
Lumen Cytosol
---·~ - coc
(a) Type I sing le internal signal-anchor (SA) sequence. The difference
Signal in the orientation of these proteins depends largely on
STA
sequence whether there is a high density of positively charged amino
acids (+++ )on theN-terminal side of the SA sequence
(type II) or on the C-terminal side of the SA sequence
Cytosol Lumen
(b) Type II NH 3+ - -JtNtltl\__ - - coo (type Ill). (d, e) Nearly all multipass proteins lack a cleavable
+++ SA signal sequence, as depicted in the examples shown here.
Type IV-A proteins, whose N-terminus faces the cytosol,
contain alternating type II S sequences and STA sequences.
Lumen Cytosol Type IV-B proteins, whose N-terminus faces the lumen,
(c) Type Ill NH ~- coo begin with a type Ill SA sequence followed by alternating
+++
SA type II SA and STA sequences. Proteins of each type with
different numbers of a helices (odd or even) are known.

Cytosol Lumen Lumen Cytosol

(d) Type IV-A .N/tN\__ .Jtlt/tltl\- - coo
+++
SA STA SA STA

Lumen Cytosol
(e) Type IV-B I\IH 3 ' ~---J ~~~--~~~---- coo-
SA ++++++ SA STA SA STA SA STA

13.2 Insertion of Membrane Proteins into the ER 591

the hydrophobic a helices relative to each other in the grow- (a) 0 =Inositol
ing chain largely determines whether a given helix functions 0 = Glucosamine
as a signal-anchor sequence or a stop-transfer anchor se- e = Mannose
quence. Other than its hydrophobicity, the specific amino PO•-(>- NH 2 = Phosphoethanolamine
acid sequence of a particular helix has little bearing on its
function. Thus the first N-tenninal a helix and the subse-
quent odd-numbered ones function as signal-anchor se-
quences, whereas the interveni ng even-numbered he lices
function as stop-transfer anchor sequences. This odd-eYen
relationship among signal-anchor and stop-transfer anchor
(.....)~_ _ _ ~r NH 3•
sequences is dictated by the fact that the transmembrane a Hy::...d_ro..;.p_ho
_ b_ic_ _-'-
) ____Po_la_r_ __
helices assume alternating orientations as a multipass protein
is woven back and fo rth aero% the membrane; signal anchor
sequences are oriented with their N-tcrmini toward the cyto- (b)
plasmic side of the bilayer, whereas stop-transfer sequences GPI
have their N-termini oriented toward the expolasmic side of Cytoso l
the bilayer.

Type IV Protei ns with N -Terminus in the Exoplasm i c

Space The large family of G protein-coupled receptors, all of
which contain seven transmembrane a helices, constitute the
most numerous type IV-B proteins, whose N-terminus ex-
tends into the exoplasmic space. In these proteins, the hydro-
phobic a helix closest to the N-terminus often is followed by
a cluster of positively charged amino acids, similar to a type 1\11

III signal-anchor sequence (see Figure 13-12b). As a result, the Mature GPI-Iinked
first a helix inserts the nascent chain into the translocon with ER lumen
protein
the N-rerminus extending into the lumen (see Figure 13-1 4e).
FIGURE 13- 1 S GPI-anchored proteins. (a) Structure of a glyco-
As the chain is elongated, it is inserted into the ER mem-
sylphosphatidylinositol (GPI) from yeast. The hydrophobic portion of
brane by alternating type II signal-anchor sequences and stop- the molecule is composed of fatty acyl chains, whereas the polar
transfer sequences, as just described for type IV-A proteins. (hydrophilic) portion of the molecule is com posed of carbo hydrate resi-
dues and phosphate groups. In other organisms, both the length of the
A Phospholipid Anchor Tethers Some acyl chains and the carbohydrate moieties may vary somewhat from
the structure shown. (b) Formation of GPI-anchored proteins in the ER
Cell-Surface Proteins to the Membrane membrane. The protein is synthesized and initially inserted into the ER
Some cell-surface proteins are anchored to the phospholipid membrane as shown in Figure 13-11. A specific transamidase simulta-
bilayer nor by a sequence of hydrophobic amino acids bur by neously cleaves the precursor protein with in the exoplasmic-facing
a covalently attached amphipathic molecule, glycosylphos- domain, near the stop-transfer anchor sequence (red), and transfers the
phatidylinositol (GPI) (Figure t3-15a and Chapter 10). These carboxyl group of the new (-terminus to the terminal amino group of a
proteins are synthesized and initially anchored to the ER preformed GPI anchor. [See C. Abeijon and C. B. Hirschberg, 1992, Trends
Biochem. Sci. 17:32, and K. Kodukula et al., 1992, Proc. Nat'/ Acad. Sci. USA
membrane exactly like type I transmembrane proteins, with
89:4982.]
a cleaved N-terminal signal sequence and an internal stop-
transfer anchor sequence directing the process (see Figure
13-Ll). However, a short sequence of amino acids in the lu-
minal domain, adjacent to the membrane-spanning domain, from moving laterally in the membrane because their cytosol-
is recognized by a transamidase located within the ER mem- facing segments interact with the cytoskeleton. In addition,
brane. This enzyme simultaneously cleaves off the original the GPI anchor targets the attached protein to the apical do-
stop-transfer anchor sequence and transfers the luminal por- main of the plasma membrane (instead of the basolateral
tion of the protein to a preformed CPT anchor in the mem- domain) in certain polarized epithelia l cells, as we discuss in
brane (figure 13-15b). Chapter 14.
Why change one type of membrane anchor for another?
Attachment ot the GPI anchor, which results in removal of
The Topology of a Membrane Protein Often
the cytosol-facing hydrophilic domain from the protein, can
have several consequences. Proteins with GPI anchors, for Can Be Deduced from Its Sequence
example, can diffuse relatively rapidly in the plane of the As we have seen, various topogenic sequences in integral
phospholipid bilayer membrane. Tn contrast, many proteins membrane proteins synthesized on the ER govern interaction
anchored by membrane-spanning a helices are impeded of the nascent chain with the translocon. When scientists

592 CHAPTER 13 • Moving Proteins into Membranes and Organelles

begin to study a protein of unknown function, the identifica- step is to identify longer segments of sufficient overall hydro-
tion of potential topogenic sequences within the correspond- phobicity to be N-terminal signal sequences or internal stop-
ing gene sequence can provide important clues about the transfer sequences and signal-anchor sequences. To accomplish
protein's topological class and function. Suppose, for exam- this, the total hydropathic index for each successive segment of
ple, that the gene for a protein known to be required for a 20 consecutive amino acids is calculated along the entire length
cell-to-cell signaling pa thway contains nucleotide sequences of the protein. Plots of these calculated va lues against position
that encode an apparent N-terminal signal sequence and an in the amino acid sequence yield a hydropathy profile.
internal hydrophobic sequence. These findings suggest that Figure 13-16 shows the hydropathy profiles for three dif-
the protein is a type I integral membrane protein and there- ferent membrane p roteins. The prominent praks in such
fore may be a cell-surface receptor for an extracellu lar li- plots identify probable topogenic sequences as well as their
gand. Furthermore, the implied type I topology suggests that position and approximate length. For example, the hydropa-
theN-terminal segment that lies between the signal sequence thy profile of the human growth hormone receptor reveals
and the internal hydrophobic sequence constitutes the extra- the presence of both a hydrophobic signal sequence at the
cellu lar domain which would likely have a part in ligand extreme N-terminus of the protem and an internal hydro-
binding, whereas the C-terminal segment that lies after the phobic stop-transfer sequence (Figure 13-16a). On the basis
internal hydrophobic sequence would likely be cytosolic and of this profile, we can deduce, correctly, that the human
may have a part in intracellular signalling. growth hormone receptor is a type I integral membrane pro-
Identification of topogenic sequences requires a way to tein. The hydropathy profile of the asialoglycoprotein recep-
scan sequence databases for segments that are sufficiently hy- tOr, a cell-surface protein that mediates removal of abnormal
drophobic to be either a signal sequence or a transmembrane extracellular glycoproteins, reveals a prominent internal hy-
anchor sequence. Topogenic sequences can often be identified drophobic signal-anchor sequence but gives no indication of
with the aid of computer programs that generate a hydropathy a hydrophobic N-terminal signal sequence (Figure 13-16b).
profile for the protein of interest. The first step is to assign a Thus we can predict that the asia loglycoprotein receptor is a
value known as the hydropathic index to each amino acid in type II or type III membrane protein. The distribution of
the protein. By convention, hydrophobic amino acids are as- charged residues on either side of the signal-anchor sequence
signed positive va lues and hydrophilic amino acids negative often can differentiate between these possibilities, since posi-
values. Although different scales for the hydropathic index tively charged amino acids flanking a membrane-spanning
exist, all assign the most positive va lues to amino acids with segment usually are oriented toward the cyrosolic face of the
side chains made up of mostly hydrocarbon residues (e.g., phe- membrane. For instance, in the case of the asialoglycoprotein
nylalanine and methionine) and the most negative values to receptor, examination of the residues flanking the signal-
charged amino acids (e.g., arginine and aspartate). The second anchor sequence reveals that the residues on the N-terminal

(a) Human growth hormone receptor (type I)

4
.3 Signal sequence
2
1
0 ~~----n--n~~~~-++-~~~~~~---,n-~~--~~~~~ftr~~~
-1
-2
-3
N-terminus 100 200 300 400 500 C-terminus
,
(b) Asialoglycop rotein receptor (type II) (c) GLUT1 (type IV)

4 4
Signal-anchor sequence
3 3
2 2
1 1
or-~--~~--~~~,p~---.r-~ O hr~~~~~~~L-~~-;~TW~~~~~~~~~~~~
-1 -1
-2 -2
-3 -3
100 200 100 200 300 400
EXPERIMENTAL FIGURE 13·16 Hydropathy profiles. portions; negative values, relatively polar portions of the protein.
Hydropathy profiles can identify likely topogenic sequences in integral Probable topogenic sequences are marked. The complex profiles for
membrane proteins. They are generated by plotting the total hydro- multipass (type IV) proteins, such as GLUTl in part (c), often must be
phobicity of each segment of 20 contiguous amino acids along the supplemented with other analyses to determine the topology of these
length of a protein. Positive values indicate relatively hydrophobic proteins.

13.2 Insertion of Membrane Proteins into the ER 593

side carry a net positive charge, rhus correctly predicting that 13.3 Protein Modifications, Folding,
this is a type II protein.
The hydropathy profile of the GLUTI glucose trans-
and Quality Control in the ER
porter, a multipass membrane protein, shows the presence Membrane and soluble secretory proteins synthesized on the
of many segments that arc sufficiently hydrophobic to be rough ER undergo four principal modifications before they
membrane-spanning helices (Figure 13-16c). The complexity reach their final destinations: ( 1) covalent addition and pro-
of this profile illustrates the difficulty both in unambiguously cessing of carbohydrates (glycosylation) in the ER and Golgi,
identifying all the membrane-spanning segments in a multi- (2) formation of disulfide bonds in the ER, (3) proper fold-
pass protein and in predicting the topology of individual ing of polypeptide chains and assembly of multtsubunit pro-
signal-anchor and stop-transfer sequences. More sophisti- teins in the ER, and (4) specific proteolytic cleavages in the
cated computer algorithms have been developed that take ER, Golgi, and secretory vesicles. Generally speaking, these
into account the presence of positively charged amino acids modifications promote folding of secretory proteins into
adjacent to hydrophobic segments as well as the length of their native structures and add structural stability to proteins
and spacing between segments. Using all this information, the exposed to the extracellular environment. Modifications
best algorithms can predict the complex topology of multi- such as glycosylation also allow the cell to produce a vast
pass proteins with an accuracy of greater than 75 percent. array of chemically distinct molecules at the cell surface that
Finally, sequence homology to a known protein may per- are the basis of specific moleculaP interactions used in cell-
mit accurate prediction of the topology of a newly discov- to-cell adhesion and communication.
ered multipass protein. For example, the genomes of One or more carbohydrate chains are added ro the ma-
multicellular organisms encode a very large number of mul- jority of proteins that are synthesized on the rough ER; in-
tipass proteins with seven transmembrane o: helices. The deed, glycosylation is the principal chemical modification to
similarities between the sequences of these proteins strongly most of these proteins. Proteins with attached carbohydrates
suggest that all have the same topology as the well-studied G arc known as glycoproteins. Carbohydrate chains in glyco-
protein-coupled receptors, which have the N-terminus ori- proteins attached to the hydroxyl (-OH)' group in serine
ented to the exoplasmic side and the C-terminus oriented to and threonine residues are referred to as 0-linked oligosac-
the cytoso!Ic side of the membrane. charides, and carbohydrate chains attached to the amide
nitrogen of asparagine are referred to as N-linked oligosac-
charides. The various types of 0-linked oligosaccharides in-
clude the mucin-type 0-linked chains (named after abundant .·
glycoproteins found in mucus ) and the carbohydrate modifi-
KEY CONCEPTS of Section 13.2 cations on proteoglycans described in Chapter 20. 0-linked
chains typically contain only one to four sugar residues,
Insertion of Membrane Proteins into the ER
which are added to proteins by enzymes known as glycosly-
• Integral membrane proteins synthesized on the rough ER transferases, located in the lumen of the Golgi complex. The
fall into five topological classes as well as a lipid-linked type more common N-linked oligosaccharides are larger and
(see Figure 13-10). more complex, containing several branches. In this section
• Topogenic sequences-N-terminal signal sequences, inter- we focus on N-linked oligosaccharides, whose initial synthe-
nal stop-transfer anchor sequences, and internal signal-anchor sis occurs in the ER. After the initial N-glycosylation of a
sequences-direct the insertion and orientation of nascent protein in the ER, the oligosaccharide chain is modified in
proteins within the ER membrane. This orientation is re- the ER and commonly in the Golgi as well.
tained during transport of the completed membrane protem Disulfide bond formation, protein folding, and assembly
to its final destination-e.g., the plasma membrane. of multimeric proteins, which take place exclusively in the
roughER, also are discussed in this section. Only properly
Single-pass membrane proteins contain one or two tope- folded and assembled proteins are transported from the . ~
.
.,
genic sequences. In multipass membrane proteins, each ex- rough ER to the Golgi complex and ultimately to the cell
helical segment can function as an internal topogenic se- surface or other final destination through the secretory path-
quence, depending on its location in the polypeptide chain way. Unfolded, misfolded, or partly folded and assembled
and the presence of adjacent positively charged residues (see proteins are selectively retained in the rough ER and then
Figure 13-14). can be degraded. We consider several features of such "qual-
Some cell-surface proteins are initially synthesized as type ity control'' in the latter parr of this section.
I proteins on the ER and then are cleaved with their luminal As discussed previously, N-terminal ER signal sequences
domain transferred to a GPI anchor (see Figure 13-15). arc cleaved from soluble secretory proteins and type I mem-
• The topology of membrane proteins can often be correctly brane proteins in the ER. Some proteins also undergo other
predicted by computer programs that identify hydrophobic specific proteolytic cleavages in the Golgi complex or secre-
ropogenic segments within the amino acid sequence and gen- tory vesicles. We cover these cleavages, as well as carbohy-
erate hydropathy profiles (see Figure 13-16). drate modifications that occur primarily or exclusively in the
Golgi complex, in the next chapter.

594 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 A Preformed N-Linked Oligosaccharide Is Added oligosaccharide is oriented so that the oligosaccharide por-
to Many Proteins in the Rough ER tion faces the ER lumen.
The entire 14-residue precursor is transferred from the dol-
Biosynthesis of all N-linked o ligosaccharides begins in the ichol carrier to an asparagine residue on a na!.cent polypeptide
rough ER with addition of a preformed oligosaccharide pre- as it emerges into the ER lumen (figure 13-18, step 0 ). Only
cursor containing 14 residues (Figure 13-17). The structure of asparagine residues in the tripeptide sequences Asn-X-Ser and
this precursor is the same in plants, animals, and single-celled Asn-X-Thr (where X is any amino acid except proline) are sub-
eukaryotes: a branched oligosaccharide, containing three glu- strates for oligosaccharyl transferase, the enzyme that catalyzes
cose (Glc), nine mannose (Man), and two N-acetylglu- this reaction. Two of the three subunits of this enzyme are ER
cosamme (GicNAc) molecules, wh ich can be written as membrane proteins whose cytosol-facing domains bind to the
Glc3 Man 9 (GicNAch. Once added to a protein, this branched ribosome, localizing a third subunit of the transferase, the cata-
carbohydrate structure is modified by addition or removal of lytic subunit, near the growing polypeptide chain in the ER
monosaccharides in the ER and Golgi compartments. The lumen. Not all Asn-X-SerfThr sequences become glycosylated,
modifications toN-linked chains differ from one glycoprotein and it is not possible to predict from the amino acid !.equence
to another and differ among different organisms, but a core of alone which potential N-linked glycosylation sites will be mod-
5 of the 14 residues is conserved in the structures of all N- ified; for instance, rapid folding of a segment of a protein con-
linked oligosaccharides on secretory and membrane proteins. taining an Asn-X-SerfThr sequence may prevent transfer of the
Prior to transfer to a nascent chain in the lumen of the ER, oligosaccharide precursor to it.
·. the precursor oligosaccharide is assembled on a membrane- Immediately after the entire precursor, Glc~Man 9 (GlcNAch,
attached anchor called dolichol phosphate, a long-chain is transferred to a nascent polypeptide, three different enzymes,
polyisoprenoid lipid (Chapter 10). After the first sugar, called glycosidases, remove all three glucose residues and one
GlcNAc, is attached to t he dolichol phosphate by a pyro- particular mannose residue (Figure 13-18, steps PJ-m). The
phosphate bond, t he other sugars are added by glycosidic three glucose re!>idues, which are the last residues added during
bonds in a complex set of reactions catalyzed by enzymes at- synthesis of the precursor on the dolichol carrier, appear to act
tached to the cytosolic or luminal faces of the r ough ER as a signal that the oligosaccharide is complete and ready to be
membrane (Figure I 3-17). The final dolichol pyrophosphoryl transferred to a protein. ·

Blocked
by tunicamycin
Cytosol

• • •
UOP UDP 5 GOP

4 GOP 3 UOP

~ aj flip
4 GOP

>-< •

A ><
3 UOP !

fl : IB
{V ~ ~ {V
.P'· rp (p :
[v
(P l
'J ~
:t~~ff\~il ~~· .f't!\'t;tl~~~
· ~ ,t~~·%
} ,~~
Dolichol
phosphate .\

• = N-Acetylglucosamine Completed
e = Mannose precursor
A = Glucose
ER lumen

FIGURE 13-17 Biosynthesis of the oligosaccha ride precursor. seven-residue dolichol pyrophosphoryl intermediate is flipped to
Oolichol phosphate is a strongly hydrophobic lipid, containing the luminal face (step !1), the remain ing four man nose and all three
75-95 carbon atoms, that is embedded in the ER membrane. Two glucose residues are added one at a time (steps "'· m ). In the later
N-acetylglucosamine (GicNAc) <Jnd five man nose residues are added redctions, the sugar to be added is first transferred from a nucleotide-
one at a time to a dolichol phosphate on the cytosolic face of the ER sugar to a carrier dolichol phosphate on the cytosolic face of the ER;
membrane (steps 0-IJ). The nucleotide-sugar donors in these and the carrier is then flipped to the luminal face, where the sugar is
later reactions are synthesized in the cytosol. Note that the first sugar transferred to the growing oligosaccharide, after which the "empty"
residue is attached to dolichol by a high-energy pyrophosphate carrier is flipped back to the cytosolic face. [After C. Abeijon and C. B.
linkage. Tunicamycin, which blocks the first enzyme in this pathway, Hirschberg, 1992, Trends Biochem. Sci. 1 7:32.)
inhibits the synthesis of all N-linked oligosaccharides in cells. After the

13.3 Protein Modifications, Folding, and Quality Control in the ER 595

 ER lumen

Dol = Dolichol • = Mannose

• = N-Acetylglucosamine & = Glucose

FIGURE 13-18 Addition and initial processing of N-llnked removed. Re-addition of one glucose residue (step mil plays a role in
oligosaccharides.ln the rough ER of vertebrate cells, the the correct folding of many proteins in t~e ER, as discussed later. The
Glc3Man9(GicNAch precursor is transferred from the dolichol carrier to process of N-linked glycosylation of a soluble secretory protein is
a susceptible asparagine residue on a nascent protein as soon as the shown here, but the luminal portions of an integral m embrane protein
asparagine crosses to the luminal side of the ER (step 0 ). In three can be modified on asparagine residues by the same mechanism. [See
separate reactions, first one glucose residue (step f)), then two glucose R. Kornfeld and S. Kornfeld, 1985, Ann. Rev. Biochem. 45:631, and M. Sousa and
residues (step 11), and finally one man nose residue (step !]) are A. J. Parodi, 1995, EMBO J. 14:4196.]

Oligosaccharide Side Chains May Promote domain in certain CAMs found on endothelial cells lining
Folding and Stability of Glycoproteins blood vessels. This interaction tethers the leukocytes to the
endothelium and assists in their movement into tissues during
The oligosaccharides attached to glycoproteins serve various an inflammatory response to infection (see Figure 20-39).
functions. For example, some proteins require N-linked oli- Other cell-surface glycoproteins possess oligosaccharide side
gosaccharides in order to fold properly in the ER. This func- chains that can induce an immune response. A common ex-
tion has been demonstrated in studies with the antibiotic ample is the A, B, 0 blood-group antigens, which are O-lin ked
tunicamycin, which blocks the first step in the formation of o ligosaccharides attached to glycoproteins and glycolipids on
the dolichol-linked oligosaccharide precursor and therefore the surface of erythrocytes and other cell types (Figure I 0-20).
inhibits synthesis of all N-linked oligosaccharides in cells (see In both cases, oligosaccharides are added to the luminal face
Figure 13-17, top left). For example, in the presence of tu- of these membrane proteins, in a manner similar to what is
nicamycin, the flu virus hemagglutinin precursor polypeptide shown in Figure 13-18 for solu ble proteins. The luminal face
(HA 0 ) is synthesized, but it cannot fold properly and form a of these membrane proteins is topologically equivalent to the
normal trimer; in this case, the protein remains, misfolded, in exterior face of the plasma membrane, where these proteins
the rough ER. Moreover, mutation of a particular asparagine eventually end up.
in the HA sequence to a glutamine residue prevents addition
of an N-linked oligosaccharide to that site and causes the
protein to accumulate in the ER in an unfolded state.
Disulfide Bonds Are Formed and Rearranged
In addition tO promoting proper folding, N-linked oligo-
saccharides also confer stability on many secreted glycopro- by Proteins in the ER Lumen
teins. Many secretory proteins fold properly and are In Chapter 3 we learned that both intramolecular and intermo-
transported to their final destination even if the addition of lecular disulfide bonds (-S-S-) help stabilize the tertiary
all N-linked oligosaccharides is blocked, for example, by tu- and quaternary structure of many proteins. These covalent
nicamycin. However, such nonglycosylated proteins have bonds form by the oxidative linkage of sulfhydryl groups
been shown to be less stable than their glycosylated forms. (-SH), also known as thiol groups, on two cysteine residues in
For instance, glycosylated fibronectin, a normal component the same or different polypeptide chains. This reaction can pro-
of the extracellular matrix, is degraded much more slowly by ceed spont~neously only when a suitable oxidant is present. In
tissue proteases than is nonglycosylated fibronectin. eukaryotic cells, disulfide bonds are formed only in the lumen
Oligosaccharides on certain cell-surface glycoproteins also of the roughER. Thus disulfide bonds are found only in soluble
play a role in cell-cell adhesion. For example, the plasma mem- secretory proteins and in the exoplasmic domains of membrane
brane of white blood cells (leukocytes) contains cell-adhesion proteins. Cyrosolic proteins and organelle proteins synthesized
molecules (CAMs) that are extensively glycosylated. T he oligo- on free ribosomes (i.e., those destined for mitochondria, chloro-
saccharides in these molecules interact with a sugar-binding plasts, peroxisomes, etc.) usually lack disulfide bonds.

596 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 The efficient formation of disulfide bonds in the lumen of sequence while a polypeptide is still growing on the ribo-
the ER depends on the enzyme protein disulfide isomerase some. Such sequential formation, however, sometimes yields
(POI), which is present in all eukaryotic cells. This enzyme is disulfide bonds between the wrong cysteines. For example,
especially abundant in the ER of secretory cells in such organs proinsulin, a precursor to the peptide hormone insulin, has
as the liver and pancreas, where large quantities of proteins three disulfide bonds that link cysteines 1 and 4, 2 and 6, and
that contain disulfide bonds are produced. As shown in figure 3 and 5. In this case, a disulfide bond that initially formed
13-19a, the disulfide bond in the active site of PDI can be read- sequentially (e.g., between cysteines I and 2) would have to
ily transferred to a protein by two sequential thiol-disulfide be rearranged for the protein to achieve its proper folded con-
transfer reactions. The reduced PDI generated by this reaction formation. In cells, the rearrangement of disulfide bonds also
is returned to an oxidized form by the action of an ER-residcnt is accelerated by POI, which acts on a broad range of protein
protein, called Erol, which carries a disulfide bond that can be substrates, allowing them to reach their thermodynamicall y
transferred to PDl. Ero 1 itself becomes oxidized by reaction most stable conformations (Figure 13-19b). Disulfide bonds
with molecular oxygen that has diffused into the ER. generally form in a specific order, first stabilizing small do
In proteins that contain more than one disulfide bond, mains of a polypeptide, then stabilizing the interactions of
the proper pairing of cysteine residues is essential for normal more distant segments; this phenomenon is illustrated by the
structure and activity. Disulfide bonds commonly arc formed folding of influenza hemagglutinin (HA) protein, discussed in
between cystcines that occur sequentially in the amino acid the next section.

s s

(a) Formation of a disulfide bond

. vS
,..-SH
Reduced
Ox•dized
POl 'S ...-SH POl ' SH
s~ 1
2
SH SH
!.
s s s sJ
Reduced
substrate Oxidized
protein substrate
protein

(b) Rearrangement of disulfide bonds

,..- SH

"-S · 3 Reduced vSH

POl "-SH
s·
_)
2
~ s si
\-s-
s_j
~s-,J
_L J-s s
.............-
Protein with
incorrect disulfide bonds Protein with
correct disulfide bonds
FIGURE 13- 19 Act ion of protein disulfide isomerase (POl). POl the substrate then reacts with this intermediate, forming a disulfide
forms and rearranges disulfide bonds via an active site with two closely bond within the substrate protein and releasing reduced POl. POl, in
spaced cysteine residues that are easily interconverted between the> turn, transfers electrons to a disulfide bond in the luminal protein Ero1,
reduced dithiol form and the oxidized disulfide form. Numbered red thereby regenerating the oxidized form of POl. (b) Reduced POl can
arrows indicate the sequence of electron transfers. Yellow bars catalyze rearrangement of improperly formed disulfide bonds by
represent disulfide bonds. (a) In the formation of disulfide bonds, the similar thiol-disulfide transfer reactions. In this case, reduced POl both
ionized (-5-) form of a cysteine thiol in the substrate protein reacts initiates and is regenerated in the reaction pathway. These reactions
with the disulfide (5-5) bond in oxidized POl to form a disulfide- are repeated until the most stable conformation of the protein is
bonded POl-substrate protein intermediate. A second ionized thiol in achieved. [SeeM. M. Lyles and H. F. Gilbert, 1991 . Biochemistry 30:619 ]

13.3 Protein Modifications, Folding, and Quality Control in the ER 597

Chaperones and Other ER Proteins Facilitate nascent chains as they enter the ER during cotranslational
Folding and Assembly of Proteins translocation. Bound BiP is thought to prevent segments of a
nascent chain from misfolding or forming aggregates,
Although many denatured proteins can spontaneously refold thereby promoting folding of the entire polypeptide into the
into their native state in vitro, such refolding usually requires proper conformation. Protein d isulfide isomerase (PDI) also
hours to reach completion. Yet new soluble and membrane contributes to proper folding, because correct 3-D confor-
proteins produced in the ER generally fold into their proper mation is stabilized by disulfide bonds in many proteins.
conformation within minutes after their synthesis. The rapid As illustrated in Figure 13-20, two other ER proteins, the
folding of these newly synthesized proteins in cells depends homologous lectim (carbohydrate-binding proteins) cal-
on the sequential action of several proteins present within nexin and calreticulin, bind selectively to certain N-linked
the ER lumen. We have already seen how the molecular oligosaccharides on growing nascent chains. The ligand for
chaperone BiP can drive post-translational translocation in these two lectins, which resembles theN-linked oligosaccha-
yeast by binding fully synthesized polypeptides as they enter ride precursor but that has only a single glucose residue
the ER (see Figure 13-9) . BiP can also bind transiently to

(a) Oligosaccharyl
transferase
Dolichol
oligosaccharide Membrane-spanning Luminal
a hel ix a helix
Cytosol

ER lumen

SiP

HA 0 trimer

Cal reticulin Completed

HA0 monomer

(b)
FIGURE 13-20 Hemagglutinin folding and assembly.
(a) Mechanism of (HA0) trimer assembly. Transient binding of the
chaperone SiP (step Ul'll to the nascent chain and of two lectins,
calnexin and cal reticulin, to certain oligosaccharide chains (step
mll promotes proper folding of adjacent segments. A total of
seven N-linked oligosaccharide chains are added to the luminal
portion of the nascent chain during cotranslational translocation,
and PDI catalyzes the formation of six disulfide bonds per
monomer. Completed HAo monomers are anchored in the
membrane by a single membrane-spanning a helix with the
N-terminus in the lumen (step f)). Interaction of three HA0 chains
with one another, initially via their transmembrane a helices,
apparently triggers formation of a long stem containing one
"' helix from the luminal part of each HA0 polypeptide. Finally,
interactions occur among the three globular heads, generating a
stable HAo trimer (step ~ ). (b) Electron micrograph of a complete
influenza virion showing trimers of HA protein protruding as
spikes from the surface of the viral membrane. [Part (a}, see U. Tatu
et al., 1995, EMBO J. 14:1340, and D. Hebert et al., 1997, J. Cell Bioi. 139:613.
Part (b), Chris Bjorn berg/Photo Researchers, Inc.]

598 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 [Glc 1Man 9 (GicNAchJ, is generated by a specific glucosyl- and membrane-spanning portions of the HA subunits also help
transferase in the ER lumen (see Figure 13-18, step m!). This stabilize the trimeric protein. Studies have shown that it takes
enzyme acts only on polypeptide chains that arc unfolded or just 10 minutes for the HA 0 polypeptides to fold and assemble
misfolded, and in this respect the glucosyltransferase acts as into their proper trimeric conformation.
one of the primary surveillance mechanisms to ensure qual
ity control of protein folding in the ER, but the mechanism
Improperly Folded Proteins in the ER Induce
by which the glucosyltransferase distinguishes folded and
unfolded proteins is not yet understood. Binding of calnexin Expression of Protein -Folding Catalysts
and cal reticulin to unfolded na~ccnt chains marked with g1u- Wild-type proteins that are synthesized on the rough ER can-
cosylated N-linked oligosaccharides prevents aggregation of not exit this compartment until they achieve their completely
adjacent segments of a protein as it is being made on the ER. folded conformation. Likewise, almost any mutation that pre-
Thus calnexin and cal reticulin, like BiP, help prevent prema- vents proper folding of a protein in the ER also blocks move-
ture, incorrect folding of segments of a newly made protein. ment of the polypeptide from the ER lumen or membrane to
Other important protein-folding catalysts in the ER the Golgi complex. The mechanisms for retaining unfolded or
lumen are peptidyl-prolyl isomerases, a family of enzymes incompletely folded proteins within the ER probably increase
that accelerate the rotation about peptidyl-prolyl bonds at the overall efficiency of folding by keeping intermediate forms
proline residues in unfolded segments of a polypeptide: in proximity to folding catalysts, which are most abundant in
the ER. Improperly folded proteins retained within the ER
generally are seen bound to the ER chaperones BiP and cal-
nexin. Thus these luminal folding catalysts perform two re-
lated functions: assisting in the folding of normal proteins by
preventing their aggregation and binding to misfolded pro-
Rotation about teins to retain them in the ER.
peptide bond Both mammalian cells and yeasts respond to the presence
of unfolded proteins in the rough ER by increasing transcrip-
tion of several genes encoding ER chaperones and other fold-
ing catalysts. A key participant in this unfolded-protein
cis trans response is Ire 1, an ER membrane protein that exists as both
a monomer and a dimer. The dimeric form, but not the mono-
meric form, promotes formation of Hacl, a transcription fac-
Such isomerizations sometimes are the rate-limiting step in tor in yeast that activates expression of the genes induced in
the folding of protein domains. Many peptidyl-prolyl isom- the unfolded-protein response. As depicted in Figure 13-21,
erases can catalyze the rotation of exposed peptidyl-prolyl binding of BiP to the luminal domain of monomeric Ire1 pre-
bonds indiscriminately in numerous proteins, but some have vents formation of the Irel dimer. Thus the quantity of free
'·
very specific protein substrates. BiP in the ER lumen determines the relative proportion of mo-
Many important soluble secretory and membrane proteins nomeric and dimeric Trel. Accumulation of unfolded proteins
synthesized on the ER are built of two or more polypeptide within the ER lumen sequesters BiP molecules, making them
subunits. In all cases, the assembly of subunits constituting unavailable for binding to Irel. As a result, the level of di-
these multisubunit (multimeric) proteins occurs in the ER. An meric lre1 increases, leading to an increase in the level of Hacl
important class of multimeric secreted proteins is the immu- and production of proteins that assist in protein folding.
noglobulins, which contain two heavy (H) and two light (L) Mammalian cells contain an additional regulatory path-
chains, all linked by intrachain disulfide bonds. Hemaggluti- way that operates in response to unfolded proteins in the
nin (HA) is another multimeric protein that provides a good ER. In this pathway, accumulation of unfolded proteins in
illustration of folding and subunit assembly (see figure 13-20). the ER triggers proteolysis of ATF6, a transmembrane pro-
This trimeric protein forms the spikes that protrude from the tein in the ER membrane, at a site within the membrane-
surface of an influenza virus particle. The HA trimer is spanning segment. The cytosolic domain of ATF6 released
formed within the ER of an infected host cell from three cop- by proteolysis then moves to the nucleus, where it stimulates
ies of a precursor protein termed HA0 , which has a single transcription of the genes encoding ER chaperones. Activa-
membrane-spanning a helix. In the Golgi complex, each of tion of a transcription factor by such regulated intramem-
the three HA 0 proteins is cleaved to form two polypeptides, brane proteolysis also occurs in the Notch signaling pathway
HA 1 and HA2 ; thus eilch H A molecule that eventually resides and during activation of the cholesterol-responsive tran-
on the viral surface contains three copies of HA 1 and three of scription factor SREBP (see Figures 16-35 and 16-37) .
HA2 (see Figure 3-1 0). The trimer is stabilized by interactions
between the large exoplasmic domains of the constituent A hereditary form of emphysema illustrates the detri-
polypeptides, which extend into the ER lumen; after HA is mental effects that can result from misfolding of proteins
transported to the cell surface, these domains extend into the in the ER. This disease is caused by a point mutation in a 1-
extracellular space. Interactions between the smaller cytosolic antitrypsin, which normally is secreted by hepatocytes and

13.3 Protein Modifications, Folding, and Quality Control in the ER 599

FIGURE 13-21 The unfolded-protein response. lrel, a Hac1
transmembrane protein in the ER membrane, has a binding site for t ranscription
BiP on its luminal domain; the cytosolic domain contains a specific ICI ~ factor
RNA endonuclease. Step 0: Accumulating unfolded proteins in the Spliced /Translation
ER lumen bind BiP molecules, releasing them from monomeric lrel. Hac1 mRNA " - . _ , /

H•~'-../
Dimerization of lrel then activates its endonuclease activity. Steps
f), ID: The unspliced mRNA precursor encoding the transcription
factor Hacl is cleaved by dimeric lrel, and the two exons are joined
to form functional Hacl mRNA. Current evidence indicates that this Eodooooloo"-'"'
processing occurs in the cytosol, although pre-mRNA processing
generally occurs in the nucleus. Step a :Hacl is translated into Unspliced '-... '*"
Hacl protein, which then moves back into the nucleus and Hac1mRNA ~
activates transcription of genes encoding several protein-folding
lre1
catalysts. [See U. Ruegsegger et al., 2001, Ce/1 107:1 03; A. Bertolotti et al.,
2000, Nat. Cell Bioi. 2:326; and C. Sidrauski and P. Walter, 1997, Ce//90:1031.)
Cytosol

ER
macrophage!.. The wild-type protein binds to and inhibits lumen
trypsin and a lso the blood protease elastase. In the absence
of arantitrypsin, elastase degrades the fine tissue in the lung
that participates in the absorption of oxygen, eventually pro-
ducing the symptoms of emphysema. Although the mutant l
a 1-antitrypsin is synthesized in the roughER, it does not fold Unfolded proteins Unfolded proteins
properly, forming an almost crystalline aggregate that is not with BiP bound
exported from the ER. In hepatocytes, the secretion of other
proteins also becomes impaired as the roughER is filled with
aggregated a 1-antitrypsin. • proteins that have transient partially folded states as they
acqu ire their fully folded conformation. One possibility is
Unassembled or Misfolded Proteins in the ER that trimming of N-linked carbohydra te chains by the a-
mannosidase 1 may occur slowly, such t hat only those glyco-
Are Often Transported to the Cytosol
proteins that remain misfoldcd in the ER lumen for a
for Degradation sufficiently long time are trimmed and therefore targeted for
Misfolded soluble secretory and membrane proteins, as well degradation. Luminal proteins that lack carbohydrate chains
as the unassembled subunits of multimeric proteins, often are altogether ca n also be targeted for degradation, indicating
degraded within an hour or two after their synthesis in the that other processes for the recognition of unfolded proteins
rough ER. For many years, researchers thought that proteo- must also exist. Other mechanisms to recognize unfolded
lytic enzymes within the ER lumen catalyzed degradation of proteins that do not involve trimming of N-linked carbohy-
misfolded or unassem bled polypeptides, but such proteases drate chains must exist because misfolded membrane pro-
were never found. More recent stud ies have shown that mis- teins that lack N-linked carbohydrate chains altogether can
folded secretory proteins are recognized by specific ER mem- nevertheless be targeted for degradation.
brane proteins and are targeted for transport from the ER Once an unfolded protein has been identified, it is tar-
lumen into the cytosol, by a process known as dislocation. geted for dislocation across the ER membrane. Some kind of
The dislocation of misfolded proteins out of the ER de- channel must exist for dislocation of misfolded proteins
pends on a set of proteins located in the ER membrane and across the ER membrane, and a complex of at least four inte-
in the cytosol that perform three basic functions. The first gral membrane proteins, known as the ERAD (ER-associated
function is recognition of misfolded proteins that will be degradation) complex, appears to serve this functio n. The
substrates for the dislocation reaction. One mechanism for structure of the dislocation channel and the mechanism by
recognition involves trimming of N-linked carbohydrate which misfolded proteins traverse the ER membrane are not
chains by the enzyme a-mannosidase I (Figure 13-22). Trimmed yet known.
glycans of the str ucture ManM(GlcNAch are recognized by As segments of the dislocated polypeptide arc exposed to
the lectin! ike protein known as EDEM, and glycans that are the cytosol, they encounter cytusolic enzymes that drive dis-
further trimmed to Man-(GicNAch arc recognized by the location. One of these enzymes is an ATPasc called p97 , a
lectinlike protein OS-9. Both EDEM and OS-9 target the member of a protein family known as t he AAA ATPase fam-
trimmed glycoprotein to the dislocation complex for degra- ily, which couple the energy of ATP hydrolysis to disassem-
dation. It is not known precisely how a-mannosidase I dis- bly of protein complexes. In retrotranslocation, hydrolysis of
tinguishes proteins that cannot fold properly, and are thus ATP by p97 may provide the driving force to pull misfolded
legitimate substrates for the dislocation process, from normal proteins from the ER membrane into the cytosol. As rnisfolded

600 CHAPTER 13 • Moving Proteins into Membranes and Organelles

.·
Cytosol {Gic) 1 {Man)9 {GicNAc) 2 (Man) 8 (GicNAcl 2 (Man)7 _5 {GicNAc) 2

ER lumen

= Glucose

{Gicb{Man) 9 {GicNAc) 2

Fold ing/retention Degradation Degradation

FIGURE 1 3-22 Modifications of N-linked oligosaccharides are that cannot fold and are therefore retained in the ER for longer times
used to monitor folding and quality control. After removal of three undergo man nose trimming by mannosidase I to form Man8(GicNAch,
glucose residues from theN-linked oligosaccharides in the ER, a which is recognized by the lectin EDEM, or further trimming to
single glucose can be re-added by a glucosyl transferase to form Man 7.5(GicNAch, which is recognized by OS-9. Recognition by either
Glc 1Man 9(GicNAch (see Figure 13-18, step m ). This modified N-linked EDEM or OS-9 leads to dislocation of the misfolded protin out of the ER,
carbohydrate binds the lectins calnexin {CNX) and cal reticulin {CRn for ubiquitination, and degradation by the proteasome.
retention in the ER and engagement of folding chaperones. Proteins

proteins reenter the cytosol, specific ubiquitin ligase enzymes

in the ER membrane add ubiquitin residues to the dislocated Disulfide bonds are added to many soluble secretory pro-
peptide. Like the action of p97, the ubiquitination reaction teins and the exoplasmic domain of membrane proteins in
is coupled to ATP hydrolysis; this release of energy possibly the ER. Protein disulfide isomerase (PDI), present in the ER
also contributes to trapping proteins in the cytosol. The re- lumen, catalyzes both the formation and the rearrangement
sulting pol yubiquitinated polypeptides, now fully in the cy- of disulfide bonds (see Figure 13-19).
tosol, are removed from the cell altogether by degradation in • The chaperone BiP, the lectins calnexin and calreticulin,
the proteasome. The ro le of polyubiquitination in targeting and peptidyl-prolyl isomerases work together to ensure
proteins to the proteasome is discussed more fully in Chap- proper folding of newly made secretory and membrane pro-
ter 3 (see Figure 3-29 and Figure 3-34). teins in the ER. The subunits of multimeric proteins also as
semble in the ER (see Figure 13-20).
Only properly folded proteins and assembled subunits arc
transported from the rough ER to the Golgi complex in
KEY CONCEPTS of Section 13.3 vesicles.
Protein Modifications, Folding, The accumulation of abnormally folded proteins and un-
and Quality Control in the ER assembled subunits in the ER can induce increased expres-
sion of ER protein-folding catalysts via the unfolded-protein
• All N-linked oligosaccharides, which are bound to aspara-
gine residues, contain a core of two N-acetylglucosamine response (see Figure 13-21).
and at least three mannos,e residues and usually have several Unassembled or misfolded proteins in the ER often are
branches. 0-linked o ligosaccharides, which are bound to transported back to the cytosol, where they are degraded in
serine or threonine residues, are generally short, often con- the ubiquitinlproteasome pathway (see Figure 13-22).
taining only one to four sugar residues.
• Formation of N-linked o ligosaccharides begins with as-
sembly of a conserved 14-residue high-mannose precursor
on dol ichol, a lipid in the membrane of the rough ER (see 13.4 Targeting of Proteins
Figure 13-17). After this preformed oligosaccharide is trans-
ferred to specific asparagine residues of nascent polypeptide to Mitochondria and Chloroplasts
chains in the ER lumen, three glucose residues and one man- In the remainder of this chapter, we examine hov.· proteins
nose residue are removed (see Figure 13-18). synthesized on cytosolic ribosomes are sorted to mitochondria,
Oligosaccharide side chains may assist in the proper folding chloroplasts, peroxisomes, and the nucleus (see Figure 13-1 ).
of glycoproteins, help protect the mature proteins from prote- In both mitochondria and chloroplasts, an internal lumen
olysis, participate in cell-cell adhesion, and function as antigens. called the matrix is surrounded by a double membrane, and
internal subcompartments exist within the matrix. In contrast,

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 601

peroxisomes are bounded by a single membrane and have a Proteins encoded by mitochondrial DNA or chloroplast
single luminal matrix compartment. Because of these and other DNA are synthesized on ribosomes within the organelles
differences, we consider peroxisomes separately in the next and directed to the correct subcompartment immediately
section. The mechanism of protein transport into and out of after synthesis. The majority of proteins located in mito-
the nucleus differs in many respects from sorting to other or- chondria and chloroplasts, however, are encoded by genes in
ganelles; this is discussed in the last section. the nucleus and are imported into the organelles after their
In addition to being bounded by two membranes, mito- synthesis in the cytosol. Apparently, as eukaryotic cells
chondria and chloroplasts share similar types of electron- evolved over a billion years, much of the genetic information
transport proteins and usc an F-class ATPast w ~ynthesize from the ancestral bactenal DNA in these endosymbiotic or-
ATP (see Figure 12-24 ). Remarkably, these characteristics ganelles moved, by an unknown mechanism, to the nucleus.
are shared by gram-negative bacteria. Also like bacterial Precursor proteins synthesized in the cytosol that arc des-
cells, mitochondria and chloroplasts contain their own tined for the matrix of mitochondria or the equivalent space
DNA, which encodes organelle rRNAs, tRNAs, and some in chloroplasts, the stroma, usually contain specific N-terminal
proteins (Chapter 6). Moreover, growth and division of mi- uptake-targeting sequences that specify binding to receptor
tochondria and chloroplasts are not coupled to nuclear divi- proteins on the organelle surface. Generally, this sequence is
sion. Rather, these organelles grow by the incorporation of cleaved once it reaches the matrix or stroma. Clearly, these
cellular proteins and lipids, and new organelles form by divi- uptake-targeting sequences are similar in their location and
sion of preexisting organelles. The numerous similarities of general function to the signal sequences that direct nascent
free-living bacterial cells with mitochondria and chloroplasts proteins to the ER lumen. Although the three types of signals
have led to the understanding that these organelles arose by share some common sequence features, their specific se-
the incorporation of bacteria into ancestral eukaryotic cells, quences differ considerably, as summarized in Table 13-1.
forming endosymbiotic organelles (see Figure 6-20). The se- In both mitochondria and chloroplasts, protein import
quence similarity of many membrane translocation proteins requires energy and occurs at points where the outer and
shared by mitochondria, chloroplasts, and bacteria provides inner organelle membranes are in close contact. Because mi-
the most striking evidence for this ancient evolutionary rela- tochondria and chloroplasts contain multiple membranes
tionship. In this section we will examine these membrane and membrane-limited spaces, sorting of many proteins to
translocation proteins in detail. their correct location often requires the sequential action of

TABLE 13-1 Uptake-Targeting Sequences That Direct Proteins from the Cytosol to Organelles*
I

Location of Sequence
Target Organell~ Within Protein Removal of Sequence Nature of Sequence

Endoplasmtc renculum (lumen) N-terminus Yes Core of 6-12 hydrophobic ammo

acids, often preceded by one or
more basic ammo acids (Arg, Lys)

Mitochondrion (matrix) N-terminus Yes Amphipathic helix, 20-50

residues in length, with Arg and
Lys residues on one side and
hydrophobic residues on the other

Chloroplast (stroma) N-terminus Yes No common motifs; generally

nch in Ser, Thr, and small
hydrophobic residues and poor
in Glu and Asp

Peroxisome (matrix) C-terminus (most proteins); No PTSl signal (Scr-Lys-Lcu) at

N-rerminus (few proteins) extreme C-terminus; PTS2 signal
at N-terminus

Nucleus (nucleoplasm) Varies No Multiple different kinds; a

common motif mcludcs a short
segment rich in Lys and Arg
residues

*Different or additional sequences target proteins to organelle membranes and subcompartments.

602 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 Uptake- EXPERIMENTAL FIGURE 13-23 Import of
targeting - - - - i -L,.. mitochondrial precursor proteins is assayed

-~:
sequence in a cell-free system. Mitochondrial precursor
Trypsin
proteins with attached uptake-targeting signals

Mitochondrial
Add energized
yeast
\.
a·· can be synthesized on ribosomes in a cell-free
reaction. When respiring mitochondria are
added to the synthesized mitochondrial precursor
protein mitochondria
.tg: proteins (top), the proteins are taken up by
mitochondria. Inside mitochondria, proteins are
protected from the action of proteases such as
Yeast mitochondrial Protein taken up Proteins seques- trypsin. When no mitochondria are present
proteins made by into mitochondria; tered within
(bottom), mitochondrial proteins are degraded
cytoplasmic ribosomes uptake-targeting mitochondria
in a cell-free system sequence removed are resistant to by added protease. Protein uptake occurs only
and degraded trypsin with energized (respiring) mitochondria, which
have a proton electrochemical gradient
.c . c (proton-motive force) across the inner mem-
Trypsin
:,· ....
•. • c .

: ..· c •
brane. The imported protein must contain an
appropriate uptake-targeting sequence. Uptake
also requires ATP and a cytosolic extract
·c .. • . • · containing chaperone proteins that maintain the
. c
Uptake-targeting A precursor proteins in an unfolded conformation.
sequence and .•· : •. t: This assay has been used to study target-
mitochondrial . c . :• ·
protein degraded oC . • c. ing sequences and other features of the
translocation process.

two targeting sequences and two membrane-bound translo- The cell-free assay outlined in Figure ·13-23 has been
cation systems: one to direct the protein into the organelle widely used in studies to define the biochemical steps in the
and the other to direct it into the correct organellar compart- import of mitochondrial precursor proteins. In this system,
ment or membrane. As we will see, the mechanisms for sort- respiring (energized) mitochondria extracted from cells can
ing various proteins to mitochondria and chloroplasts arc incorporate mitochondrial precursor proteins carrying appro-
related to some of the mechanisms discussed previously. priate uptake-targeting sequences that have been synthesized
in the absence of mitochondria. Successful incorporation of
the precursor into the organelle can be assayed either by resis-
Amphipathic N-Terminal Signal Sequences
tance to digestion by an added protease such as trypsin. In
Direct Proteins to the Mitochondrial Matrix other assays, successful import of a precursor protein can be
All proteins that travel from the cytosol to the same mito- shown by the proper cleavage of the N-terminal targeting se-
chondrial destination have targeting signals that share com- quences by specific mitochondrial proteases. The uptake of
mon motifs, although the signal sequences are generally not completely presynthesized mitochondrial precursor protems
identical. Thus the receptors that recognize such signals are by the organelle in this system contrasts with the cell-free co-
able to bind to a number of different but related sequences. translational translocation of secretory proteins into the ER,
The most extensively studied sequences for localizing proteins which generally occurs only when microsomal (ER-derived)
to mitochondria are the matrix-targeting sequences. These se- membranes are present during synthesis (see Figure 13-4).
quences, located at theN-terminus, are usually 20-50 amino
acids in length. They are rich in hydrophobic amino acids, Mitochondrial Protein Import Requires
positively charged basic amino acids (arginine and lysine), and
Outer-Membrane Receptors and Translocons
hydroxylated ones (serine and threonine) but tend to lack
negatively charged acidic residues (aspartate and glutamate). in Both Membranes
Mitochondrial matrix-targeting sequences are thought to Figure 13-24 presents an overview of protein import from
assume an a-helical conformation in which positively charged the cytosol into the mitochondrial matrix, the route into the
amino acids predominate on one side of the helix and hydro- mitochondrion followed by most imported proteins. We will
phobic amino acids predominate on the other side. Sequences discuss in det:~il each step of protein transport into the ma-
·. such as these that contain both hydrophobic and hydrophilic trix and then consider how some proteins subsequently are
regions are said to be amphipathic. Mutations that disrupt targeted to other compartments of the mitochondrion.
this amphipathic character usually disrupt targeting to the After synthesis in the cytosol, the soluble precursors of
matrix, although many other amino acid substitutions do mitochondrial proteins (including hydrophobic integral mem-
not. These findings indicate that the amphipathicity of brane proteins) interact directly with the mitochondrial mem-
matrix-targeting sequences is critical to their function. brane. In general, only unfolded proteins can be imported into

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 603

FIGURE 13-24 Protein import into the coo
mitochondrial matrix. Precursor proteins Precursor
protein
synthesized on cytosolic ribosomes are
maintained in an unfolded or partially folded
state by bound chaperones, such as Hsc70
(step 0 ). After a precursor protein binds to an
import receptor near a site of contact with the
inner membrane (step 8 ), it is transferred Cytosolic
into the general import pore (step 0 ). The _../ Hsc70
translocating protein then moves through this
channel and an adjacent channel in the inner
membrane (steps !), L't). Note that transloca- Matrix-targeting
tion occurs at rare "contact sites" at which the sequence
inner and outer membranes appear to touch.
Binding of the translocating protein by the
matrix chaperone Hsc70 and subsequent ATP
hydrolysis by Hsc70 helps drive import into the
matrix. Once the uptake-targeting sequence is
removed by a matrix protease and Hsc70 is General
released from the newly imported protein import pore
(step [;)), it folds into the mature, active (Tom40)
conformation within the matrix (step lf..il).
Folding of some proteins depends on matrix
chaperonins. [See G. Schatz, 1996, J. Bioi. Chem.
271:31763, and N. Pfanner eta!., 1997, Ann. Rev. Cell
Devel. Bioi. 1 3:25.]

lntennembrane
space

Inner membrane

Mitochondrial matrix
\
·~
Cleaved
Active
protein <..l-J targeting
sequence

the mitochondrion. Chaperone proteins such as cytosolic involved in targeting and import are designated Tom proteins,
Hsc70 keep nascent and newly made proteins in an unfolded for translocon of the outer membrane.)
state so that they can be taken up by mitochondria. This pro- The import receptOrs subsequently transfer the precursor
cess requires ATP hydrolysis. Import of an unfolded mito- proteins to an import channel in the outer membrane. This
chondrial precursor is initiated by the binding of a mitochondrial channel, composed mainly of the Tom40 protein, is known
targeting sequence to an import receptor in the outer mito- as the general import pore because all known mitochondrial
chondrial membrane. These receptors were first identified by precursor proteins gain access to the interior compartments
experiments in which antibodies to specific proteins of the of the mitochondrion through t his channel. When Tom40 is
outer mitochondrial membrane were shown to inhibit protein purified and incorporated into liposomes, it forms a trans-
import into i<;o]ared mitochondria. Subsequent genetic experi- membrane channel with a pore wide enough to accommo-
ments, in which the genes for specific mitochondrial outer- date an unfolded polypeptide chain . The general import pore
membrane proteins were mutated, showed that specific forms a largely passive channel through the outer mitochon-
receptOr proteins were responsible for the tmport of different drial membrane, and the driving force for unidirectional
classes of mitochondrial proteins. For example, N-terminal transport into mi tochondria comes from within the mito-
matrix-targeting sequences are recognized by Tom20 and chondrion. In the case of precursors destined fo r the mito-
Tom22 . (Proteins in the outer mitochondrial membrane chondrial matrix, transfer through the outer membrane

604 CHAPTER 13 • Moving Proteins into Membranes and Organelles

·.
 occurs simultaneously with transfer through an inner- cleave their uptake-targeting sequence normally, but the im-
membrane channel composed of the Tim23 and Tim 17 pro- ported polypeptides fail to fold and assemble into the nanve
•. '
teins. (Tim stands for translocon of the inner membrane.) tertiary and quaternary structures.
Translocation into the matrix thus occurs at "contact sites"
where the outer and inner membranes arc in close proximity.
Studies with Chimeric Proteins Demonstrate
Soon after the N-terminal matrix-targeting sequence of a
protein enters the mitochondrial matrix, it is removed by a Important Features of Mitochondrial Import
protease that resides within the matrix. The emerging protein Dramatic evidence for the ability of mitochondrial matrix-
also is bound by matrix Hsc70, a chaperone that is localized targeting sequences to Jirect import was obtained with chime-
to the translocation channels in the inner mitochondrial ric proteins produced by recombinant DNA techniques. For
membrane by interacting with transmembrane protein example, the matrix-targeting sequence of alcohol dehydroge-
Tim44 . This interaction stimulates ATP hydrolysis by matrix nase can be fused to theN-terminus of dihydrofolate reductase
Hsc70, and together these two proteins are thought to power (DHFR), which normally resides in the cytosol. In the presence
translocation of proteins into the matrix. of chaperones, which prevent the C-terminal DHFR segment
Some imported proteins can fold into their final, active from folding in the cytosol, cell-free translocation assays show
conformation without further assistance. Final folding of that the chimeric protein is transported into the matrix (Figure
many matrix proteins, however, requires a chaperonin. As 13-25a ). The inhibitor methotrexate, which binds tightly to
.· discussed in Chapter 3, chaperonin proteins actively facili- the active site of DHFR and greatly stabilizes its folded confor-
tate protein folding in a process that depends on ATP. For mation, renders the chimeric protein resistant to unfolding by
instance, yeast mutants defective in Hsc60, a chaperonin in cytosolic chaperones. When translocation assays are per-
the mitochondrial matrix, can import matrix proteins and formed in the presence of methotrexate, the chimeric protein

(a) coo (b) Bound (c)

methotrexate
inhibitor
Cytosol
"~
Outer ''- "'
Folded
DHFR-
~~~:t/membrane

Cytosol t \.
Outer membrane ~
lntermembrane
space

lntermembrane Inner
space membrane

l
~
Folded
'ODHFR
Translocation
int ermediate
I M itochondrial
matrix

~~.fo)
Cleaved . / Cleaved
Mitochondrial targeting targeting
matrix sequence , sequence 0.2f.l.m
Spacer sequence/

EXPERIMENTAL FIGURE 13-25 Experiments with chimeric the spacer sequence is long enough to extend across both transport
proteins el ucidate mitochondrial protei n import. These experi- channels, a stable translocation intermediate, with the targeting
ments show that a matrix-targeting sequence alone directs proteins sequence cleaved off, is generated in the presence of methotrexate, as
to the mitochondrial matrix and that only unfolded proteins are shown here. (c) The (-terminus of the translocation intermediate in (b)
translocated across both membranes. The chimeric protein in these can be detected by incubating the mitochondria with antibodies that
experiments contained a matrix-targeting signal at its N-terminus bind to the DHFR segment, followed by gold particles coated with
(red), followed by a spacer sequence of no pilrticular function (black) bclcterial protein A, which binds nonspecifically to antibody molecules
and then by dihydrofolate reductase (DHFR), an enzyme normally (see Figure 9-29). An electron micrograph of a sectioned sample reveals
present only in the cytosol. (a) When the DHFR segment is unfolded, gold particles (red arrowhead) bound to the translocation intermediate
the chimeric protein moves across both membranes to the matrix of at a contact site between the inner and outer membranes. Other
energized mitochondria and the matrix-targeting signal then is contact sites (black arrows) also are evident. [Parts (a) and (b) adapted
removed. (b) When the ( -terminus of the chimeric protein is locked in from J. Rassow et al., 1990, FEBS Lett. 275:190. Part (c) from M. Schweiger et al.,
the folded state by binding of methotrexate, translocation is blocked. If 1987, J. Cell Bioi. 105:235, courtesy of W. Neupert.]

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 605

does not completely enter the matrix. This finding demon- The sequential binding and ATP-driven release of multiple
strates that a precursor must be unfolded in order to traverse matrix Hsc70 molecules to a translocating protein may simply
the import pores in the mitochondrial membranes. trap the unfolded protein in the matrix. Alternatively, the ma-
Additional studies revealed that if a sufficiently long trix Hsc70, anchored to the membrane by the Tim44 protein,
spacer sequence separates the N-terminal matrix-targeting may act as a molecular motor to pull the protein into the ma-
sequence and DHFR portion of the chimeric protein, then, in trix (see Figure 13-24 ). In this case, the functions of matrix
the presence of methotrexate, a translocation intermediate Hsc70 and Tim44 would be analogous to those of the chaper-
that spans both membranes can be trapped if enough of the one BiP and Scc63 complex, respectively, in post-translational
polypeptide protrudes into the matrix to prevent the poly- translocation into the ER lumen (see Figure 13-9).
peptide chain from sliding back into the cytosol, possibly by The third energy input required for mitochondrial pro-
stably as~ociating with matrix Hsc70 (Figure 13-25b). In tein import is a H+ electrochemical gradient, or fJroton-motive
order for such a stable translocation intermediate to form, force, across the inner membrane. Recall from Chapter 12
the spacer sequence must be long enough to span both mem- that protons are pumped from the matrix into the intermem-
branes; a spacer of 50 amino acids extended to its maximum brane space during electron transport, creating a transmem-
possible length is adequate to do so. If the chimera contains brane potential across the inner membrane. ln general, only
a shorter spacer-say, 35 amino acids-no stable transloca- mitochondria that are actively undergoing respiration, and
tion intermediate is obtained because the spacer cannot span therefore have generated a proton-motive force across the
both membranes. These observations provide further evi- inner membrane, arc able to translocate precursor proteins
dence that translocated proteins can span both inner and from the cytosol into the mitochondrial matrix. Treatment
outer mitochondrial membranes and traverse these mem- of mitOchondria with inhibitors or uncouplers of oxidative
branes in an unfolded state. phosphorylation, such as cyanide or dinitrophenol, dissi-
Microscopy studies of stable translocation intermediates pates this proton-motive force. Although precursor proteins
show that they accumulate at sites where the inner and outer still can bind tightly to receptors on such poisoned mito-
mitochondrial membranes are close together, evidence that chondria, the proteins cannot be imported, either in intact cells
precursor proteins enter only at such sites (Figure 13-25c). or in cell-free systems, even in the presence of ATP and chap-
The distance from the cytosolic face of the outer membrane erone proteins. Scientists do not fully understand how the
to the matrix face of the inner membrane at these contact protOn-motive force is used to facilitate entry of a precursor
sites is consistent with the length of an unfolded spacer se- protein into the matrix. Once a protein is partially inserted
quence required for formation of a stable translocation in- into the inner membrane, it is subjected to a transmembrane
termediate. Moreover, stable translocation intermediates potential of 200 mY (matrix space negative). This seemingly
can be chemically cross-linked to the protein subunits that small potential difference is established across the very narrow
comprise the translocation channels of both the outer and hydrophobic core of the lipid bilayer, which gives an enor-
inner membranes. This finding demonstrates that imported mous electric gradient, equivalent to about 400,000 V/cm.
proteins can simultaneously engage channels in both the One hypothesis is that the positive charges in the amphipa-
outer and inner mitochondrial membrane, as depicted in Fig- thic matrix-targeting sequence could simply be "electropho-
ure 13-24. Since roughly 1000 stuck chimeric proteins can resed," or pulled, into the matrix space by the inside-negative
be observed in a typical yeast mitochondrion, it is thought membrane electric potential.
that mitochondria have approximately 1000 general import
pores for the uptake of mitochondrial proteins.
Multiple Signals and Pathways Target Proteins
to Submitochondrial Compartments
Three Energy Inputs Are Needed to Import Unlike targeting to the matrix, targeting of proteins to the
intermembrane space, inner membrane, and outer membrane
Proteins into Mitochondria of mitochondria generally requires more than one targeting
As noted previous!} and indicated in Figure 13-24, ATP hy- sequence and occurs via one of several pathways. Figure
drolysis by Hsc70 chaperone proteins in both the cytosol and 13-26 summarizes the organization of targeting sequences in
the mitochondrial matrix is required for import of mitochon- proteins sorted to different mitochondrial locations.
drial proteins. Cytosolic Hsc70 expends energy to maintain
bound precursor proteins in an unfolded state that is compe- Inner-Membrane Proteins Three separate pathways arc
tent for translocation into the matrix. The importance of ATP known to target proteins to the inner mitochondrial mem-
to this function was demonstrated in studies in which a mito- brane. One pathway makes use of the same machmery that
chondrial precursor protein was purified and then denatured is used for targeting of matrix proteins (Figure 13-27, path A).
(unfolded) by urea. When tested in the cell-free mitochondrial A cytochrome oxidase subunit called CoxVa is a protein trans-
translocation system, the denatured protein was incorporated ported by this pathway. The precursor form of CoxVa, which
into the matrix in the absence of ATP. In contrast, import of contains an N-terminal matrix-targeting sequence recognized
the native, undenatured precursor required ATP for the nor- by the Tom20/22 import receptor, is transferred through the
mal unfolding function of cyrosolic chaperones. Tom40 general import pore of the outer membrane and the

606 CHAPTER 13 • Moving Proteins into Membranes and Organelles

Location Imported Locations of targeting sequences FIGURE 13-26 Targeting sequences in
of imported protein in preprotein imported mitochondrial proteins. Most
protein mitochondrial proteins have anN-terminal
Cleavage by matrix-targeting sequence (pink) that is similar
matrix protease but not identical in different proteins. Proteins
destined for the inner membrane, the intermem-
Matrix Alcohol
brane space, or the outer membrane have one
dehydrogenase Ill
or more additional targeting sequences that
function to direct the proteins to these locations
sequence Mature protein
by several different pathways. The lettered
pathways correspond to those illustrated in
Inner Cleavage by Hydrophobic Figures 13-26 and 13-27. [See W. Neupert, 1997,
membrane matrix protease stop-transfer sequence
Ann. Rev. Biochem. 66:863.]

~-._?.- ~
Cytochrome
Path A oxidase subunit
CoxVa

Cleavage by Internal sequences

Path 8
ATP
synthase
subunit 9
:::ct:~::Z''
Internal sequences recognized
by Tom70 receptor and 'Tim22 complex

Path c ADP/ATP
anti porter
. . . . A .A.L -~ ~~
lntermembrane First cleavage by Second cleavage by protease
space matrix protease in intermembrane space

Path A Cytochrome b 2
~~
I
lntermembrane-space-targeting
sequence

Targeting sequence for

the general import pore

Path 8 Cytochrome c
· heme lyase ~ - .__.-...
Outer Stop-transfer and outer-membrane
membrane localization sequence

Porin (P70)

inner-membrane Tim23/17 translocation complex. In addi- domains recognized by an inner-membrane protein termed
tion to the matrix-targeting sequence, which is cleaved dur- Oxal. This pathway is thought to involve translocation of at
ing import, CoxVa contains a hydrophobic stop-transfer least a portion of the precursor into the matrix via the Tom40
sequence. As the protein passes through the Tim23/17 chan- and Tim23117 channels. After cleavage of the matrix-targeting
nel, the stop-transfer sequence blocks translocation of the sequence, the protein is inserted into the inner membrane by a
C-terminus across the inner membrane. The membrane- process that requires interaction with Oxa 1 and perhaps other
anchored intermediate is then transferred laterally into the inner-membrane proteins (Figure 13-27, path B). Oxal is re-
bilayer of the inner membrane much as type I integral mem- lated to a bacterial protein involved in inserting some cytoplas-
brane proteins are incorporated into the ER membrane (see mic membrane proteins in bacteria. This relatedness suggests
Figure 13-11 ). that Oxal may have descended from the translocation machin-
A second pathway to the inner membrane is followed by ery in the endosymbiotic bacterium that eventually became the
proteins (e.g., ATP synthase subunit 9) whose precursors con- mitochondrion. However, the proteins forming the inner-
tain both a matrix-targeting sequence and internal hydrophobic membrane channels in mitochondria are not related to the

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 607

 Path A Path B Path C

Oxa1-targeting coc
Stop-transfer Matr ix-targeting

e
Internal targeting
sequence sequence Matrix-targeting
sequences

p,Co!~..~ \
sequence

NH • NH,) ~ NH,

Cytosol

Outer
membrane

D D
lntermembrane
space

Mitochondrial Assembled
matrix p rotein .•
Hsc70

Cleaved
matrix-targeting
sequences----

FIGURE 13-27 Three pathways to the inner mitochondrial redirected to the inner membrane in pathway B. Matrix Hsc70 plays
membrane from the cytosol. Proteins with different targeting a role similar to its role in the import of soluble matrix proteins (see
sequences are directed to the inner membrane via different pathways. Figure 13-23). Proteins delivered by pathway C contain internal
In all three pathways, proteins cross the outer membrane via the sequences that are recognized by the Tom70/Tom22 import receptor;
Tom40 general import pore. Proteins delivered by pathways A and B a different inner-membrane translocation channel (Tim22/54) is used in
contain an N-terminal matrix-targeting sequence that is recognized by this pathway. Two intermembrane proteins (Tim9 and Tim 10) facilitate
the Tom20/22 import receptor in the outer membrane. Although both transfer between the outer and inner channels. See the text for .·.
these pathways use the Tim23/17 inner-membrane channel, they differ discussion. [See R. E. Dalbey and A. Kuhn, 2000, Ann. Rev. Cell Dev. Bioi. 16:51,
in that the entire precursor protein enters the matrix and then is and N. Pfanner and A. Geissler, 2001, Nature. Rev. Mol. Cell Bioi. 2:339.]

proteins in bacterial translocons. Oxal also participates in the outer-membrane proteins Tom70 and Tom22, the imported
inner-membrane insertion of certain proteins (e.g., subunit II of protein passes through the outer membrane via the general
cytochrome oxidase) that are encoded by mitochondrial DNA import pore (Figure 13-27, path C) . The protein then is
and synthesized in the matrix by mitochondrial ribosomes. transferred to a second translocation complex in the inner
The final pathway for insertion in the inner mitochon- membrane composed of the Tim22, Tim18, and Tim54 pro-
drial membrane is followed by multipass proteins that con- teins. Transfer to the Tim22/18/54 complex depends on a
tain six membrane-spanning domains, such as the ADP/ATP multimcric complex of two small proteins, Tim9 and TimlO,
antiporter. These proteins, which lack the usua l N-tcrminal which reside in the intermembranc space. The sma ll Tim
matrix-targeting sequence, contain multiple internal mito- proteins are thought to act as chaperones, guiding imported
chondrial targeting sequences. After the internal sequences protein precursors from the general import pore to the
are recognized by a second import receptor composed of Tim22118/54 complex in the inner membrane by binding to

608 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 their hydrophobic regions, preventing them from forming proteolytic cleavage, this pathway is simi lar to that of inner-
insoluble aggregates in the aqueous environment of t he in- membrane proteins such as CoxVa (see Figure 13-27, path A).
termembrane space. Ultimately the Tim22/18/54 complex is The smal l Tim9 and TimlO proteins, which reside in the
responsible for incorporating the multiple hydrophobic seg- intermembrane space, illustrate a second pathway for targeting
ments of the imported protein into the inner membrane. to the intermembrane space. In this pathway, the imported
proteins do not contain an N-terminal matrix-targeting se-
quence and arc delivered directly to the intermembrane space
lntermembrane-Space Proteins Two pathways deliver cyto- via the general import pore without involvement of any inner-
solic proteins to the ~p;~ce between the inner and outer mito membrane translocation fauors (Figure 13-28, path B). Trans-
chondrial membranes. The major pathway is followed by location through the Tom40 general import pore does not
proteins, such as cytochrome b2 , whose precursors carry two seem to be coupled to any energetically favorable process;
different N-terminal targeting sequences, both of which ulti- however, once located in the intermembrane space, Tim9 and
mately are cleaved. T he most N-terminal of the two sequences TimlO proteins acquire two disulfide bonds each and acquire
is a matrix-targeting sequence, which is removed by the ma- compact stable folded structures. Apparently, the mechanism
trix protease. The second targeting sequence is a hydrophobic that drives unidirectional translocation through the outer
segment that blocks complete translocation of the protein membrane involves passive diffusion through the outer mem-
across the inner membrane (Figure 1 3-28, path A). After the brane, followed by folding and disulfide bond formation which
resulting membrane-embedded intermediate diffuses laterally irreversibly traps the protein in the intermembrane space. In
away from the Tim23/17 translocation channel, a protease in ma ny respects the process of disulfide bond formation in the
·. the membrane cleaves the protein near the hydrophobic trans- intermembrane space resembles that of the ER lumen and in-
membrane segment, releasing the mature protein in a soluble volves a disulfide bond-generating protein Erv 1 and a disulfide
form into the intermembrane space. Except for the second transfer protein Mia40.

Path A Path 8
lntermembra ne-space-
targeting sequence

Tim9 or Tim10 protei~

~NH 3+
coo

Cytosol

Outer membrane

lntermembrane space

Heme 1-s· sj
~ .!!- 0 Mia40

~
Erv1

'""" meonb""'
Protease
rfJ li

)
Tim23/17

Mitochondrial
matrix /
Cleaved
matrix-targeting
sequence

FIGURE 13-28 Two pathways to the mitochondrial intermem- space. Pathway B is a specialized pathway for delivery to the intermem·
brane space. Pathway A, the major one for delivery of proteins brane space ofthe proteins Tim9 and Tim10. These proteins readily
lfrom thP cytosol to the intermembrane space, is similar to pathway pass through the Tom40 general1mport pore and, once they are in
A for delivery to the inner membrane (see Figure 13-26). The major the intermembrane space, fold and form disulfide bonds that
difference is that the internal targeting sequence in proteins such as prevent reverse translocat ion through Tom40. The disulfide bonds are
cytochrome b 2 destined for the intermembrane space is recognized generated by Ervl and are transferred to Tim9 and Tim 10 by Mia40.
by an inner-membrane protease, which cleaves the protein on the [SeeR. E. Dalbey and A. Kuhn, 2000, Ann. Rev. Cell Dev. Bioi. 16:51 ; N. Pfanner and
intermembrane-space side of the membrane. The released protein A. Geissler, 2001 , Nat. Rev. Mol. Cell Bioi. 2:339, and K. Tokatlidis, 2005, A disulfide
then folds and binds to its heme cofactor within the intermembrane relay system in mitochondria. Ce/1 121 :965-967.]

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 609

Outer-Membrane Proteins Many of the proteins that reside matrix and BiP in the ER lumen. Unlike mitochondria,
in the mitochondrial outer membrane, including the Tom40 chloroplasts do not generate an electrochemical gradient
pore itself and mitochondrial porin, have a (3-barrel struc- (proton-motive force) across their inner membrane. Thus
ture in which antiparallel strands form hydrophobic trans- protein import into the chloroplast stroma appears to be
membrane segments surrounding a central channel. Such powered solely by ATP hydrolysis.
proteins are incorporated into the outer membrane by first
interacting with the general import pore, Tom40, and then
they are transferred to a complex known as the SAM (sort- Proteins Are Targeted to Thylakoids
ing and assembly machinery) complex, whi\.:h is composed
by Mechanisms Related to Translocation
of at least three outer-membrane proteins. Presumably it is
the very stable hydrophobic nature of 13-barrel proteins that Across the Bacterial Cytoplasmic Membrane
ultimately causes them to be stably incorporated into the In addition to the double membrane that surrounds them,
outer membrane, but precisely how the SAM complex facili- chloroplasts contain a series of internal interconnected mem-
tates this process is not known. branous sacs, the thylakoids (see Figure 12-31). Proteins lo-
calized to the thylakoid membrane or lumen carry out
photosynthesis. Many of these proteins are synthesized in
Targeting of Chloroplast Stromal Proteins the cytosol as precursors contairring multiple targeting se-
quences. For example, plasrocyanin and other proteins des-
Is Similar to Import of Mitochondrial
tined for the thylakoid lumen require the successive action of
Matrix Proteins two uptake-targeting sequences. The first is an N-termina l
Among the proteins found in the chloroplast stroma are the stromal-import sequence that directs the protein to the
enzymes of the Calvin cycle, which functions in fixing car- stroma by the same pathway that imports the rubisco S sub-
bon dioxide into carbohydrates during photosynthesis (Chap- unit. The second sequence targets the protein from the
ter 12). The large (L) subunit of ribulose 1,5-bisphosphate stroma to the thylakoid lumen. The role ~f these targeting
carboxylase (rubisco) is encoded by chloroplast DNA and sequences has been shown in experiments measuring the up-
synthesized on chloroplast ribosomes in the stromal space. take of mutant proteins generated by recombinant DNA
The small (S) subunit of rubisco and all the other Calvin-cycle techniques into isolated chloroplasts. For instance, mutant
enzymes are encoded by nuclear genes and transported to plastocyanin that lacks the thylakoid-targeting sequence but
chloroplasts after their synthesis in the cytosol. The precur- contains an intact stromal-import sequence accumulates in
sor forms of these stromal proteins contain anN-terminal the stroma and is not transported into the thylakoid lumen.
stromal-import sequence (see Table 13-1). Four separate pathways for transporting proteins from
Experiments with isolated chloroplasts, similar to those the stroma into the thylakoid have been identified. All four
with mitochondria illustrated in Figure 13-23, have shown pathways have been found to be closely related to analogous
that they can import the S-subunit precursor after its synthe- transport mechanisms in bacteria, illustra ring the close evo-
sis. After the unfolded precursor enters the stromal space, it lutionary relationship between the stromal membrane and
binds transiently to a stromal Hsc70 chaperone and theN- the bacterial cytoplasmic membrane. Transport of plastocy-
terminal sequence is cleaved. In reactions facilitated by anin and related proteins into the thylakoid lumen from the
Hsc60 chaperonins that reside within the stromal space, stroma occurs by a chloroplast SRP-dependent pathway that
eightS subunits combine with the eight L subunits to yield utilizes a translocon similar to SecY, the bacterial version of
the acti\·e rubisco enzyme. the Sec61 complex (Figure 13-29, left). A second pathway
The general process of stromal import appears to be very for transporting proteins into the thylakoid lumen involves a
similar to that for importing proteins into the mitochondrial protein related to bacterial protein SecA, which uses the en-
matrix (see Figure 13-24 ). At least three chloroplast outer- ergy from ATP hydrolysis to drive protein translocation
membrane proteins, including a receptor that binds the through the SecY translocon. A third pathway, which targets
stromal-import sequence and a translocation channel pro- proteins to the thylakoid membrane, depends on a protein
tein, and five inner-membrane proteins are known to be es- related to the mitochondrial Oxa 1 protein and the homolo-
sential for directing proteins to the stroma. Although these gous bacterial protein (see Figure 13-27, path B). Some pro-
proteins are functionally analogous to the receptor and chan- teins encoded by chloroplast DNA and synthesized in the
nel proteins in the mitochondrial membrane, they are not stroma or transported into the stroma from the cytosol are
structurally homologous. The lack of sequence similarity be- inserted into the thylakoid membrane via this pathway.
tween these chloroplast and mitochondrial proteins suggests Finally, thylakoid proteins that bind metal-\.:untaining
that they may have arisen independently during evolution. cofactors follow another pathway into the thylakoid lumen
The available evidence suggests that chloroplast stromal (Figure 13-29, right) . The unfolded precursors of these pro-
proteins, like mitochondrial matrix proteins, are imported in teins are first targeted to the stroma, where theN-terminal
the unfolded state. Import into the stroma depends on ATP stromal-import sequence is cleaved off and the protein then
hydrolysis catalyzed by a stromal Hsc70 chaperone whose folds and binds its cofactor. A set of thylakoid-membrane
function is similar to that of Hsc70 in the mitochondrial proteins assists in translocating the folded protein and bound

610 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 Metal-binding
Thylakoid-targeting ~:0
recursor )
Plastocyanin sequence
COO precursor
St romal-import -------
/ sequence ------ r
.__. """\ NH3 •
NH •
3
Toe
Cytosol

Outer membrane

lntermembrane
space

Inner membrane

complex complex
Stroma

Plastocy~ 1 Cleaved import

sequence CI"~I::rt-'
1 ~r~~:~i~bi ndi ng
sequence ~
M 6 pH pathway

RR ~ metal
Bound
0 ions
0

jB
Thylakoid membrane

.· Thylakoid
lumen

Mature
Mat ure metal-binding
plastocyanin protein
FIGURE 13-29 Transporting proteins to chloroplast thylakoids. the thylakoid lumen by a separate endoprotease, the prot ein folds into
Two of the four pathways for transporting proteins from t he cytosol to its mature conformation (step 0 ). In the pH-dependent pathway
the thylakoid lumen are shown here. In these pathways, unfolded (right) , metal-bindin g proteins fold in the stroma, and complex redox
precursors are delivered to the stroma via t he same outer-membrane cofactors are added (step f)). Two arginine residues (RR) at the
proteins t hat import st romal-localized proteins. Cleavage of the N-termi nus of the thylakoid-targeting sequence and a pH gradient
N-terminal st romal-import sequence by a stromal protease then across the inner membrane are required for transport of the folded
reveals the thylakoid-targeting sequence (step 0 ). At t his poi nt the protein into t he thylakoid lumen (step 0 ). The translocon in the
two pathways diverge. ln the SRP-dependent pathway (left), plastocya- thylakoid membrane is composed of at least four proteins related to
nin and similar proteins are kept unfolded in the stromal space by a set proteins in the bacterial cytoplasmic membrane. The thylakoid
of chaperones (not shown) and, directed by the thylakoid-targeting targeting sequence conta ining the two arginine residues is cleaved in
sequence, bind to proteins t hat are closely related to the bacterial SRP. the thylilkoid lumen (step (1 ). [SeeR. Dalbey and C. Robinson, 1999, Trend>
SRP receptor, and SecY translocon, which mediate movement into the Btochem. Sci. 24:17; R. E. Dalbey and A. Kuhn, 2000, Ann. Rev. Cell Dev. Bioi. 16:51 ;
lumen (step f)). After t he thylakoid-target ing sequence is removed in and C. Robinson and A. Bolhuis, 2001 , Nat. Rev. Mol. Cell Bioi. 2:350.]

13.4 Targeting of Proteins to Mitochondria and Chloroplasts 611

cofactor into the thylakoid lumen, a process powered by the
H+ electrochemical gradient normally maintained across the • The four known pathways for moving proteins from the
thylakoid membrane. The thylakoid-targeting sequence that chloroplast stroma to the thylakoid closely resemble translo-
directs a protein to this pathway includes two closely spaced cation across the bacterial cytoplasmic membrane (see Figure
arginine residues that are crucia I for recognition. Bacterial 13-29). One of these systems can translocate folded proteins.
cells also have a mechanism for translocating folded proteins
with a similar arginine-containing sequence across the cyto-
plasmic membrane, known as the Tat (twin-arginine translo-
cation) pathway. The molecular mechamsm whereby these
13.5 Targeting of Peroxisomal Proteins
large folded globular proteins can be translocated across the
thylakoid membrane is currently under intense study. Peroxisomes are small organeJles bounded by a single mem-
brane. Unlike mitochondria and chloroplasts, peroxisomes
lack DNA and ribosomes. Thus all peroxisomal proteins are
encoded by nuclear genes, synthesized on ribosomes free in
KEY CONCEPTS of Section 13.4 the cytosol, and then incorporated into preexisting or newly
generated peroxisomes. As peroxisomes are enlarged by ad-
Targeting of Proteins to Mitochondria and Chloroplasts
dition of protein (and lipid), they eventually divide, forming
• Most mitochondrial and chloroplast proteins are encoded new ones, as is the case with mitochondria and chloroplasts.
by nuclear genes, synthesized on cytosolic ribosomes, and The size and enzyme composition of peroxisomes vary
tmported post-translationally into the organelles. considerably in different kinds of cells. However, all peroxi-
AJI the information required to target a precursor protein somes contain enzymes that use molecular oxygen to oxidize
from the cytosol to the mitochondrial matrix or chloroplast various substrates such as amino acids and fatt y acids,
stroma IS contained within its N-terminal uptake-targeting breaking them down into smaller componenJs for use in bio-
sequence. After protein import, the uptake-targeting sequence synthetic pathways. The hydrogen peroxide (H 2 0 2 ) gener-
is removed by proteases within the matrix or stroma. ated by these oxidation reactions is extremely reactive and ·.
potentiaJly harmful to ceJlular components; however, the
• Cytosolic chaperones maintain the precursors of mito-
peroxisome also contains enzymes, such catalase, that effi-
chondrial and chloroplast proteins in an unfolded state. Only
ciently convert H 2 0 2 into H 2 0. In mammals, peroxisomes
unfolded proteins can be imported into the organelles. Trans-
are most abundant in liver cells, where they constitute about
location in mitochondria occurs at sites where the outer and
1-2 percent of the cell volume.
inner membranes of the organelles are close together.
Proteins destined for the mitochondrial matrix bind to re-
ceptors on the outer mitochondrial membrane and then are
Cytosolic Receptor Targets Proteins
transferred to th~ general import pore (Tom40) in the outer
membrane. Translocation occurs concurrently through the with an SKL Sequence at the (-Terminus
outer and inner membranes, driven by the proton-motive into the Peroxisomal Matrix
force across the inner membrane and ATP hydrolysis by the Peroxisomal targeting signals were first identified by testing
Hsc70 ATPase in the matrix (see Figure 13-24). deletions of peroxisomal proteins for a specific defect in per-
• Proteins sorted to mitochondrial destinations other than oxisomal targeting. In one early study, the gene for firefly lu-
the matrix usuaJly contain two or more targeting sequences, ciferase was expressed in cultured insect cells and the resulting
one of which may be an N-terminal matrix-targeting se- protein was shown to be properly targeted to the peroxisome.
quence (see Figure 13-26). However, expression of a truncated gene missing a small por-
tion of the C-tcrminus of the protein led to luciferase that
• Some mitochondrial proteins destined for the intermem-
failed to be targeted to the peroxisome and remained in the
brane space or inner membrane are first imported into the
cytoplasm. By testing various mutant luciferase proteins in this
matrix and then redirected; others never enter the matrix but
system, researchers discovered that the sequence Ser-Lys-Leu
go directly to their final location.
(SKL in one-letter code) or a related sequence at the C-terminus
• Protein import into the chloroplast stroma occurs through was necessary for peroxisomal targeting. Further, addition of
inner-membrane and outer-membrane translocation chan- the SKL sequence to the C-terminus of a normaJly cytosolic
nels that are analogous in function to mitochondrial chan- protein leads to uptake of the altered protein by peroxisomes
nels, but compmed of proteins unrelated in sequence to the in cultured cdb. All but a few of the many different peroxi-
corresponding mitochondrial proteins. somal matrix proteins bear a sequence of this type, known as
• Proteins destined for the thylakoid have secondary target- peroxisomal-targeting sequence 1, or simply PTSl.
ing sequences. After entry of these proteins into the stroma, The pathway for import of catalase and other PTSl-bear-
cleavage of the stromal-targeting sequences reveals the ing proteins into the peroxisomal matrix is depicted in Figure
thylakoid-targeting sequences. 13-30. The PTSl binds to a soluble carrier protein in the cyto-
sol (Pex5), which in turn binds to a receptor in the peroxisome

612 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 C_H3+ protein translocation is not well understood but probably m-
volves the formation of oligomers of Pex5 bound to PTS !-
Q '\ coo bearing cargo molecules and the Pex14 receptor. There is
PTS1
evidence that the size of the oligomer adjusts according to the
peroxisomal-targeting
sequence size of the PTSl-bearing cargo molecules and that the oligo-
mers dissociate once the complex of Pex5 bound to PTSl-
a\ bearing cargo molecule enters the peroxisomal matrix. The

~
dynamic formation of oligomers apparently is the key mecha-
nism by which PTSl-bearing cargo molecules can be accom-
modated without the formation of large stable pores that
Pox5 ""PIO< would disrupt the integrity of the peroxisomal membrane.
Once the complex of a PTSl-bearing cargo molecule
bound to Pex5 enters the matrix, Pex5 dissociates from the
peroxisomal matrix protein for recycling back to the cyto-
II plasm. The peroxisomal membrane proteins PexlO, Pex12,
and Pex2 form a complex that is cructal for recycling of Pex5.
Pex5 is modified by ubiquitination and then is deubiquitinated
Cytosol
as part of the recycling process. Since ubiquitin modification
of proteins ultimately requires ATP hydrolysis, the energy-
dependent recycling of Pex5 may be the step in the import
process that uses energy to power unidirectional translocation
Peroxiso~ Pex14
matrix / II~
. •
0 of cargo molecules across the peroxisomal membrane.
A few peroxisomal matrix proteins such as thiolase are
~ synthesized as precursors with an N-terminal uptake-targeting
C"' Peroxisomal
matrix protein sequence known as PTS2. These proteins b~nd to a different
cytosolic receptor protein, but otherwise import is thought to
occur by the same mechanism as for PTSl-containing proteins.
FIGURE 13-30 PTSl -dlrected import of peroxisomal matrix
proteins. Step 0 : Most peroxisomal matrix proteins contain a (-terminal
PTS 1 uptake-targeting sequence (red) that binds to the cytosolic receptor Peroxisomal Membrane and Matrix Proteins
PexS. Step fJ: PexS with the bound matrix protein forms a multimeric Are Incorporated by Different Pathways
complex with the Pex14 receptor located on the peroxisome membrane.
Autosomal recessive mutations that cause defective
Step 0 : By a process that is not well understood, the matrix protein-PexS
complex is then transferred into the peroxisomal matrix, where PexS
peroxisome assembly occur naturally in the human
dissociates from the matrix protein. Step D: PexS is then returned to the
population. Such defects can lead to severe developmental
cytosol by a process that involves the peroxisomal membrane proteins defects often associated with craniofacial abnormalities. In
Pex2, Pexl 0, and Pexl2, as well as additional membrane and cytosolic Zellweger syndrome and related disorders, for example, the
proteins not shown. Note that folded proteins can be imported into transport of many or all proteins into the peroxisomal matrix
peroxisomes and that the targeting sequence is not removed in the is impaired: newly synthesized peroxisomal enzymes remain
matrix. [SeeP. E. Purdue and P. B. Lazarow, 2001,Ann. Rev. Cei/Dev. Bioi. in the cytosol and are eventually degraded. Genetic analyses
17:701; S. Subramani et al., 2000, Ann. Rev. Biochem. 69:399; and V. Dammai and of cultured cells from different Zellweger patients and of
S. Subramani, 2001, Ce// 105:187.] yeast cells carrying similar mutations have identified more
than 20 genes that are required for peroxisome biogenesis. •

membrane (Pex14). The protein to be imported then moves Studies with peroxisome-assembly mutants have shown
across the peroxisomal membrane while still bound to Pex5. that different pathways are used for importing peroxisomal
The peroxisome import machinery, unlike most systems that matrix proteins versus inserting proteins into the peroxi
mediate protein import into the ER, mitochondria, and chloro- somal membrane. For example, analysis of cells from some
plasts, can translocate folded proteins across the membrane. Zellweger patients led to identification of genes encoding the
For example, catalase assumes a folded conformation and Pex5-recycling proteins PexlO, Pexl2, and Pex2. 'vlutant
binds to heme in the cytoplasm before traversing the peroxi- cells defective in any one of these proteins cannot incorpo-
somal membrane. Cell-free studies have shown that the peroxi- rate matrix proteins into peroxisomes; nonetheless, the cells
some import machinery can transport large macromolecular contain empty peroxisomes that have a normal complement
objects, including gold particles of about 9 nm in diameter, as of peroxisomal membrane proteins (Figure 13-31b). Muta-
long as they have a PTSl tag attached to them. However, per- tions in any one of three other genes were found to block
oxisomal membranes do not appear to contain large stable insertion of peroxisomal membrane proteins as well as import
pore structures, such as the nuclear pore described in the next of matrix proteins (Figure 13-31c). These findings demon-
section. The fundamental mechanism of peroxisomal matrix strate that one set of proteins translocates soluble proteins

13.5 Targeting of Peroxisomal Proteins 613

EXPERIMENTAL FIGURE 13··31 Studies reveal different Stained for Stained for
pathways for incorporation of peroxisomal membrane and matrix PMP70 catalase
proteins. Cells were stained with fluorescent antibodies to PMP70, a
peroxisomal membrane protein, or with fluorescent antibodies to
catalase, a peroxisomal matrix protein, then viewed in a fluorescent
microscope. (a) In wild-type cells, both peroxisomal membrane and
matrix proteins are visible as bright foci in numerous peroxisomal bodies.
(b) In cells from a Pex12-deficient patient, catalase is distributed uniformly
throughout the cytosol, whereas PMP70 is localized normally to
peroxisomal bodies. (c) In cells from a Pex3-deficient patient, peroxisomal
membranes cannot assemble, and as a consequence, peroxisomal bodies
(b) Pex12 mutants (deficient
do not form. Thus both catalase and PMP70 are mis-localized to the in matrix protein import)
cytosol. [Courtesy of Stephen Gould, Johns Hopkins University.]

. .· ..·.·.o· ·..
·a
into the peroxisomal matrix, but a different set is required
for insertion of proteins into the peroxisomal membrane.
This situation differs markedly from that of the ER, mito-
chondnon, and chloroplast, for which, as we have seen,
. ,o ·
. ..
.. . .
0 •

membrane proteins and soluble proteins share many of the

(c) Pex3 mutants (deficient
same components for their insertion into these organelles. in membrane biogenesis)
Although most peroxisomes are generated by division of
preexisting organelles, these organelles can arise de novo by 0
the three-stage process depicted in Figure 13-32. In this case,
peroxisome assembly begins in the ER. At least two peroxi- .,
0 0
somal membrane proteins, Pex3 and Pex16, are inserted into
the ER membrane by the mechanisms described in Section
13.2. Pex3 and Pexl6 then recruit Pex19 to form a special- 0
ized region of the ER membrane that can bud off of the ER
to form a peroxisomal precursor membrane. Current evi-
dence indicates that peroxisomal membrane protein assem-
bly into mature peroxisomes may also follow the same Division of mature peroxisomes, which largely deter-
Pex19-dcpendent pathway for the de novo formation of new mines the number of peroxisomes within a cell, depends on
peroxisomes frqm the ER. The insertion of peroxisomal still another protein, Pexll. Overexpression of the Pexll
membrane proteins generates membranes that have all the protein causes a large increase in the number of peroxisomes,
components necessary for import of matrix proteins, leading suggesting that this protein controls the extent of peroxisome
to the formation of mature, functional peroxisomes. division. The small peroxisomes generated by division can be

Precursor
membrane
Peroxisomal Peroxisomal
membrane ghost PT S 1-bearing M ature pe roxisome
proteins matrix protein

e ta. PTS2-bearing
matrix protein
PMP70 Catalase

FIGURE 13-32 Model of pe roxisoma l biogenesis and division. protein~ tdrgeted to the matrix. The pathways for importing PTS 1- and
The first stage in the de novo formation of peroxisomes is the PTS2-bearing matrix proteins differ only in the identity of the cytosolic
incorporation of peroxisomal membrane proteins into precursor receptor (PexS and Pex7, respectively) that binds the targeting
membranes derived from the ER. Pex19 acts as the receptor for sequence (see Figure 13-30). Complete incorporation of matrix
membrane-targeting sequences. A complex of Pex3 and Pex16 is proteins yields a mature peroxisome. Although peroxisomes can form
required for proper insertion of proteins (e.g., PMP70) into the forming de novo as just described, under most conditions the proliferation of
peroxisomal membrane. Insertion of all peroxisomal membrane peroxisomes involves division of mature peroxisomes, a process that
proteins produces a peroxisomal ghost, which is capable of importing depends on the Pex11 protein.

614 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 enlarged by incorporation of additional matrix and mem- Large and Small Molecules Enter and Leave
brane proteins via the same pathways described previously. the Nucleus via Nuclear Pore Complexes
Numerous pores perforate the nuclear envelope in all eu-
KEY CONCEPTS of Section 13.5 karyotic cells. Each nuclear pore is formed from an elaborate
structure termed the nuclear pore complex (NPC), which is
Targeting of Peroxisomal Proteins one of the largest protein assemblages in the cell. The total
• All peroxisomal proteins are synthesized on cytosolic ribo- mass of the pore structure is 60-80 million Da in vertebrates,
somes and incorporated into the organelle post-translarionally. which is about 16 times larger than a ribosome. An NPC is
made up of multiple copies of some 30 different proteins
• Most peroxisomal matrix proteins contain a C-terminal
called nucleoporins. Electron micrographs of nuclear pore
PTSl targeting sequence; a few have an N-terminal PTS2
complexes reveal an octagonal, membrane-embedded nng
targeting sequence. Neither targeting sequence is cleaved
structure that surrounds a largely aqueous pore (Figure 13-33 ).
after import.
Eight approximately 100-nm-long filaments extend into the
• All proteins destined for the peroxisomal matrix bind to a nucleoplasm with the distal ends of these filaments joined by
cytosolic carrier protein, which differs for PTS1- and PTS2- a terminal ring, forming a structure called the nuclear bas-
bearing proteins, and then are directed to common import ket. Cytoplasmic filaments extend from the cytoplasmic side
receptor and translocation machinery on the peroxisomal of the NPC into the cytosol.
membrane (see Figure 13-30). Ions, small metabolites, and globular proteins up to
• Translocation of matrix proteins across the peroxisomal about 40 kDa can diffuse passively through the central aque-
membrane depends o n ATP hydrolysis. Unlike protein im- ous region of the nuclear pore complex. However, large pro-
port to the ER, mitochondrion, and chloroplast, many per- teins and ribonucleoprotein complexes cannot diffuse in and
oxisomal matrix proteins fold in the cytosol and traverse the out of the nucleus. Rather, these macromolecules are actively
membrane in a folded conformation. transported through the NPC with the assistance of soluble
Proteins destined for the peroxisomal membrane contain transporter proteins that bind macromolecules and also in-
different targeting sequences than peroxisomal matrix pro- teract with nucleoporins. The capacity and efficiency of the
teins and are imported by a different pathway. NPC for such active transport is remarkable. In one minute,
each NPC is estimated to transport 60,000 protein molecules
Unlike mitochondria and chloroplasts, peroxisomes can
into the nucleus, 50-250 mRNA molecules, 10-20 ribo-
arise de novo from precursor membranes probably derived
somal subunits, and 1000 tRNAs out of the nucleus.
from the ER as well as by division of preexisting organelles
In general terms, the nuclearporins are of three types:
(see Figure 13-32).
structural nucleoporins, membrane nucleoporins, and FG-
nucleoporins. The structural nucleoporins form the scaffold
of the nuclear pore, which is a ring of eightfold symmetry
13.6 Transport into and out that traverses both membranes of the nuclear envelope, cre-
ating an annulus. The membranes of the inner and outer
of the Nucleus leaflets of the nuclear envelope connect at the NPC by a
·. The nucleus is separated from the cytoplasm by two mem- highly curved region of membrane that contains the embed-
branes, which form the nuclear envelope (see Figure 9-32). ded membrane nucleoporins (see Figure 13-33b). A set of
The nuclear envelope is continuous with the ER and forms a seven structural nucleoporins forms a Y-shaped structure
part of it. Transport of proteins from the cytoplasm into the about the size of the ribosome, known as the ¥-complex.
nucleus and movement of macromolecules, including mRNAs, Sixteen copies of theY-complex form the basic structural
tRNAs, and ribosomal subunits, our of the nucleus occur scaffold of the pore, which has bilateral symmetry across the
through nuclear pores, which span both membranes of the nuclear envelope and eightfold rotational symmetry in the
nuclear envelope. Import of proteins into the nucleus shares plane of the envelope (see Figure 13-33c). A structural motif
some fundamental features with protein import into other or- repeated several times within theY-complex is closely re-
ganelles. For example, imported nuclear proteins carry spe- lated to a structure found in the COPII proteins that drive
cific targeting sequences known as nuclear localization formation of coated vesicles within cells (see Chapter 14).
sequences, or NLSs. However, proteins are imported into the This primordial relationship between nuclear pore structural
nucleus in a folded state, and thus nuclear import differs fun- proteins and vesicle coat proteins suggests that the two types
damentally from protein translocation across the membranes of membrane coat complexes share a common origin. The
of the ER, mitochondrion, and chloroplast, where proteins basic function of thi~ c:kment may be to form a protein lat-
are unfolded during translocation. In this section we discuss tice that, in a complex with membrane nucleoporins, de-
the main mechanism by which proteins and some ribonuclear forms the membrane into a highly curved structure.
proteins such as ribosomes enter and exit the nucleus. We also The FG-nucleoporins, which line the channel of the nu-
discuss how mRNAs and other ribonuclear protein complexes clear pore complex and also are found associated with the
are exported from the nucleus by a process that differs mecha- nuclear basket and the cytoplasmic filaments, contain mul-
nistically from nuclear protein import. tiple repeats of short hydrophobic sequences that are rich in

13.6 Transport into and out of the Nucleus 615

(a) (b)
Cytoplasmic
fi laments
Cytoplasm
FG-nucleop orins

Structural nucleoporins
(Y-complex)

Membrane nucleoporins

Outer nuc lear

/,membrane

N uclear
envelop e

In ner nuclear
membrane

Nucleoplasm

~ · · " ' "" 4

membrane ;;;
o
rrr
~ fk I, ',
Inner nuclear ~
membrane

(d)

Hydrophilic FG-repeat
region (hydrophobic)
FG-nucleoporin Matrix of FG-repeats
in central channel of pore

FIGURE 13- 33 Nuclear pore complex at different levels of membrane of the nucleus (left) and the eightfold rotational symmetry
resolution. (a) Visualized by scanning electron microscopy. nuclear around the axis of the pore (right). (d) The FG-nucleoporins have
envelopes from the large nuclei of Xenopus oocytes. Top: View of the extended disordered structures that are composed of repeats of the
cytoplasmic face reveals octagona l shape of membrane-embedded sequence Phe- Giy interspersed with hydrop hilic regions (left). The
portion of nuclear pore complexes. Bottom: View of the nucleoplasmic FG-nucleoporins are most abundant in the central part of the pore,
filce shows the nuclear basket that extend~ from the membrane and the FG-repeat sequences are thought to fill the central channel
portion. (b) Cutaway model of the pore complex showing the major with a gel-like matrix (right). [Part (a) from V. Doye and E. Hurt, 1997, Curr.
structural features formed by membrane nucleoporins, structural Opin. Cell Bioi. 9:401, courtesy of M. W. Goldberg and T. D. Allen. Part (b) adapted
nucleoporins, and FG-nucleoporins. (c) Sixteen copies of theY-complex from M.P. Rout and J.D. Atchison, 2001 , J. Bioi. Chern. 276:16593. Part (c)
forms a major part of the structura l scaffold of the nuclear pore courtesy ofThomas Schwartz. Part (d) adapted from K. Ribbeck and D. Gorlich,
complex. The three-dimensional structure of theY-complex is modeled 2001, EMBOJ. 20:1320- 1330.]
into the pore structure. Note the twofold symmetry across the double

616 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 phenylalanine (F) and glycine (G) residues (FG-repeats). The amino acids near the C-terminus of the protein: Pro-Lys-Lys-
hydrophobic FG-repeats are thought to occur in regions of Lys-Arg-Lys-Val. Experiments with engineered hybrid proteins
extended, otherwise hydrophilic polypeptide chains that fill in which this sequence was fused to a cytosolic protem demon-
.•
the central transporter channel. The FG-nucleoporins are es- strated that it directs transport into the nucleus and conse-
sential for the function of the NPC; however, the NPC re- quently functions as an NLS. NLS sequences subsequently were
mains functional even if up to half of the FG-repeats have identified in numerous other proteins imported into the
been deleted. The FG-nucleoporins are thought to form a flex- nucleus. Many of these are similar to the basic NLS in SV40
ible gel-like matrix with bulk properties that allow the diffu- large T-antigen, whereas other NLSs are chemically quite
sion of small molecules while excluding unchaperoned different. For instance, an NLS in the RNA binding protein
hydrophilic proteins larger than 40 kDa (see Figure 13-33d). hnRNP Al is hydrophobic. Thus there must be multiple mech-
anisms for the recognition of these very different sequences.
Early work on the mechanism of nuclear import showed
Nuclear Transport Receptors Escort Proteins that proteins containing a basic NLS, similar to the one in
Containing Nuclear-Localization Signals SV40 large T-antigen, will be efficiently transported into iso-
into the Nucleus lated nuclei if they are provided with a cytosolic extract (Fig-
All proteins found in the nucleus-such as histones, transcrip- ure 13-35). Using this assay system, researchers purified two
tion factors, and DNA and RNA polymerases-are synthesized
in the cytoplasm and imported into the nucleus through nuclear
pore complexes. Such proteins contain a nuclear-localization (a) Effect of digitonin
signal (NLS) that directs their selective transport into the nu-
cleus. NLSs were first discovered through the analysis of mu-
tants of the gene for large T-antigen encoded by simian virus 40
(SV40). The wild-type form of large T-antigen is localized to
the nucleus in virus-infected cells, whereas some mutated forms
of large T-antigen accumulate in the cytoplasm (Figure 13-34).
The mutations responsible for this altered cellular localization
all occur within a specific seven-residue sequence rich in basic

(a) (b)
-Digitonin +Digitonin

(b) Nuclear import by permeabilized cells

-

-Lysate
.c. +Lysate

EXPERIMENTAL FIGURE 13-35 Cytosolic proteins are

required for nuclear transport. The failure of nuclear transport to
occur in permeabilized cultured cells in the absence of lysate demon-
EXPERIMENTAL FIGURE 13-34 Nuclear-localization signal strates the involvement of soluble cytosolic components in the
(NLS) directs proteins to the cell nucleus. Cytoplasmic proteins can process. (a) Phase-contrast micrographs of untreated and digitonin-
be localized to the nucleus when they are fused to a nuclear localiza- permeabilized Hela cells. Treatment of a monolayer of cultured cells
tion signal. (a) Normal pyruvate kinase, visualized by immunofluores- with the mild, nonionic detergent digitonin permeabilizes the plasma
cence after cultured cells were treated with a specific antibody (yellow), membrane so that cytosolic constituents leak out but leaves the
is localized to the cytoplasm. This very large cytosolic protein functions nuclear envelope and NPCs intact. (b) Fluorescence micrographs of
in carbohydrate metabolism. (b) When a chimeric pyruvate kinase digitonin-permeabilized Hela cells incubated with a fluorescent
protein containing the SV40 NLS at its N-terminus was expressed in protein chemically coupled to a synthetic SV40 T-antigen NLS peptide
cells, it was localized to the nucleus. The chimeric protein was in the presence and absence of cytosol (lysate). Accumulation of this
expressed from a transfected engineered gene produced by fusing transport substrate in the nucleus occurred only when cytosol was
a viral gene fragment encoding the SV40 NLS to the pyruvate kinase included in the incubation (right). (From S. Adam et al., 1990,J. Cell Bioi.
gene. [From D. Kalderon et al., 1984, Cell 39:499, courtesy of Dr. Alan Sm1th.] 1 1 1:807, courtesy of Dr. Larry Gerace.]

13.6 Transport into and out of the Nucleus 617

FIGURE 13-36 Nuclear import. Mechanism Ran·GDP Ran·GTP lmportin Cargo

~
for nuclear import of "cargo" proteins. In the
cytoplasm (top), a free nuclear transport receptor
D
~ T ~
p p
(importin) binds to the NLS of a cargo p rotein,

w
forming a bimolecular cargo complex. The cargo
complex diffuses through the NPC by transiently
interacting with FG-nucleoporins. In the
nucleoplasm, Ran ·GTP binds to the importin,
causing a conformational change that decreases ·.
~ ../
Ca rgo
its affinity for the NLS and releasing the cargo. To
support another cycle of import, the exportin-
complex f) ~
Ran ·GTP complex is transported back to the
cytoplasm. A GTPase-accelerating protein (GAP)

<~····- J~ ti;--------
associated with the cytoplasmic filaments of the
NPC stimulates Ran to hydrolyze the bound GTP.
This generates a conformational change that
causes dissociation from the nuclear transport
receptor, which can then initiate another round
I NPC j

of import. Ran·GDP is returned to the nucleo- Nucleoplasm

plasm, where a guanine nucleotide-exchange FG-nucleoporins
factor (GEF) causes release of GDP and rebinding
ofGTP.

s T
p

required cytosolic components: Ran and a nuclear transport reaches the cytoplasmic side of the NPC, Ran interacts with a
receptor. Ran is a small monomeric G protein that exists in ei- specific GTPase-activating protein (Ran-GAP) that is a com-
ther GTP- or GDP-bound conformations (see Figure 3-32). The ponent of the NPC cytoplasmic filaments. This stimulates Ran
nuclear transport receptor binds to both the NLS on a cargo to hydrolyze its bound GTP to GDP, causing it to convert to
protein to be transported into the nucleus and to FG-repeats on a conformation that has low affinity for the nuclear transport
nucleoporins. By a physical process that is not well understood, receptor, so that the free nuclear transport receptor is released
by binding transiently to FG-repeats, nuclear transport recep- into the cytoplasm, where it can participate in another cycle of
tors have the ability to rapidly traverse the FG-repeat-contain- import. Ran-GDP travels back through the pore to the nucleo-
ing matrix in the central channel of the nuclear pore, whereas plasm, where it encounters a specific guanine nucleotide-
proteins of simi lar size that lack this property are excluded exchange factor (Ran-GEF) that causes Ran to release its bound
from the central channel. Nuclear transport receptors can be GDP in favor of GTP. The net result of this series of reactions
monomeric, with a single polypeptide that can bind to both an is the coupling of the hydrolysis of GTP to the transfer of an
NLS and FG-repears, or they can be dimeric, with one subunit NLS-bearing protein from the cytoplasm to the nuclear interior,
binding to the NLS and the other binding to FG-repeats. thus providing a driving force for nuclea r transport.
The mechanism for 1mport of cytoplasmic cargo pro- Although the nuclear transport receptor-cargo complex
teins mediated by a nuclear import receptor is shown in Fig- travels through the pore by random diffusion, the overall pro-
ure 13-36. Free nuclear transport receptor in the cytoplasm cess of transport of cargo into the nucleus is unidirectional.
binds to its cognate NLS in a cargo protein, forming an im- Because of the rapid dissociation of the import complex when
portin-cargo complex. The cargo complex then translocates it reaches the nucleoplasm, there is a concentration gradient of
through the NPC channel as the nuclear transport receptor the nuclear transport receptor-cargo complex across the NPC:
interacts with FG-repeats. The cargo complex r::~pidly reaches high in the cytoplasm, where the complex assembles, and low
the nucleoplasm, and there the nuclear transport receptor in- in the nucleoplasm, where it dissociates. This concentration
teracts with Ran·GTP, causing a conformational change in gradient is responsible for the unidirectional nature of nuclear
the nuclear transport receptor that displaces the NLS, releas- import. A similar concentration gradient is responsible for
ing the cargo protein into the nucleoplasm. The nuclear trans- driving the nuclear transport receptor in the nucleus back into
port receptor-Ran·GTP complex then diffuses back through the the cytoplasm. The concentration of the nuclear transport re-
NPC. Once the nuclear transport receptor-Ran·GTP complex ceptor-Ran·GTP complex is higher in the nucleoplasm, where

618 CHAPTER 13 • Moving Proteins into Membranes and Organelles

it assembles, than on the cytoplasmic side of the NPC, where difference: RawGTP is part of the cargo complex during
it dissociates. Ultimately, the direction of the transport pro- export but not during import. Apart from this dtfference,
cesses depends on the asymmetric distribution of the Ran- the two transport processes are remarkably similar. In both
GEF and the Ran-GAP. Ran-GEF in the nucleoplasm processes, association of a nuclear transport receptor with
maintains Ran in the Ran·GTP state, where it promotes dis- Ran·GTP in the nucleoplasm causes a conformational
sociation of the cargo complex. Ran-GAP on the cytoplasmic change that affects its affinity for the transport signal. Dur-
side of the NPC converts Ran-GTP to Ran·GDP, dissociating ing import, the interaction causes release of the cargo,
the nuclear transport recepror-Ran-GTP complex and releas- whereas during export, the interaction promotes association
ing free nuclear transport receptor into the cytosol. with the cargo. In both export and import, stimulation of
Ran-GTP hydrolysis in the cytoplasm by Ran-GAP pro-
A Second Type of Nuclear Transport Receptors duces a conformational change in Ran that releases the
transport signal receptor. During nuclear export, the cargo
Escort Proteins Containing Nuclear-Export
is also released. Localization of the Ran-GAP and -GEF to
Signals out of the Nucleus the cytoplasm and nucleus, respectively, is the basts for the
A very similar mechanism is used to export proteins, tRNAs, unidirectional transport of cargo proteins across the NPC.
and ribosomal subunits from the nucleus to the cytoplasm. In keeping with their similarity in function, the two types
This mechanism initially was elucidated from studies of cer- of nuclear transport receptors are highly homologous in se-
tain ribonuclear protein complexes that "shuttle" between quence and structure. The family of nuclear transport recep-
the nucleus and the cytoplasm. Such "shuttling" proteins tors has 14 members in yeast and more than 20 in mammalian
contain a nuclear-export signal (NES) that stimulates their cells. The NESs or NLSs to which they bind have been deter-
export from the nucleus to the cytoplasm through nuclear mined for only a fraction of them. Some individual nuclear
pores, in addition to an NLS that results in their uptake into transport receptors function in both import and export.
the nucleus. Experiments with engineered hybrid genes en- A similar shuttling mechanism has been shown to export
coding a nucleus-restricted protein fused to various segments other cargoes from the nucleus. For example, exportin-t
of a protein that shuttles in and out of the nucleus have iden- functions to export tRNAs. Exporrin-t binds fully processed
tified at least three different classes of NESs: a leucine-rich tRNAs in a complex with Ran·GTP that diffuses through
sequence found in PKI (an inhibitor of protein kinase A) and NPCs and dissociates when it interacts with Ran-GAP in the
in the Rev protein of human immunodeficiency virus (HIV), NPC cytoplasmic filaments, releasing the tRNA into the cy-
as well as two sequences identified in two different heteroge- tosol. A Ran-dependent process is also required for the nu-
neous ribonucleoprotein particles (hnRNPs). The function- clear export of ribosomal subunits through NPCs once the
ally significant structural features that specify nuclear export protein and RNA components have been properly assembled
remain poorly understood. in the nucleolus. Likewise, certain specific mRNAs that as-
The mechanism whereby shuttling proteins are exported sociate with particular hnRNP proteins can be exported by a
from the nucleus is best understood for those containing a Ran-dependent mechanism.
leucine-rich NES. According to the current model, shown in
Figure 13-3 7a, a specific nuclear transport receptor, in the
Most mRNAs Are Exported from the Nucleus
nucleus, called" exportin 1, first forms a complex with
Ran·GTP and then binds the NES in a cargo protein. Bind- by a Ran-Independent Mechanism
ing of exportin 1 to Ran-GTP causes a conformational Once the processing of an mRNA is completed in the nucleus,
change in exportin 1 that increases its affinity for the NES so it remains associated with specific hnRNP proteins in a mes-
that a tmnolecular cargo complex is formed. Like other nuclear senger ribonuclear protein complex, or mRNP. The principal
transport receptors, exportin 1 interacts transiently with FG- transporter of mRNPs out of the nucleus is the mRNP ex-
repeats in FG-nucleoporins and diffuses through the NPC. porter, a heterodimeric protein composed of a large subunit
The cargo complex dissociates when it encounters the Ran- called nuclear export factor 1 (NXF1) and a small subunit,
GAP in the NPC cytoplasmic filaments, which stimulates nuclear export transporter 1 (NXTl). Multiple NXFl/NXTl
Ran to hydrolyze the bound GTP, shifting it into a confor- dimers bind to nuclear mRNPs through cooperative interac-
mation that has low affinity for exportin 1. The released ex- tions with the RNA and other mRNP adapter proteins that
perrin 1 changes conformation to a structure that has low associate with nascent pre-mRNAs during transcription elon-
affinity for the NES, releasing the cargo into the cytosol. The gation and pre-mRNA processing. In many respects, NXFl/
direction of the export process is driven by this dissociation NXTl acts like a nuclear transport receptor that binds to an
of the cargo from exportin 1 in the cytoplasm, which causes NLS or NES in the sense that both subunits interact with the
a concentration gradient of the cargo complex across the FG-domains of FG-nucleoporins, allowing them to diffuse
NPC that is high in the nucleoplasm and low in the cyto- through the central channel of the NPC.
plasm. Exportin 1 and the Ran·GDP are then transported The process of mRNP export docs not require Ran, and
back into the nucleus through an NPC. thus the unidirectional transport of mRNA out of the nu-
By comparing this model for nuclear export with that in cleus requires a source of energy other than GTP hydrolysis
Figure 13-36 for nuclear import, we can see one obvious by Ran. Once the mRNP-NXFl/NXTl complex reaches the

13.6 Transport into and out of the Nucleus 619

FIGURE 13-37 Ran-dependent and (a}
Ran·GDP Ran·GTP Exportin 1 Cargo
Ran-independent nuclear export. (a) Ran-
Q ~ ~ ~NES
dependent mechanism for nuclear export of
T
cargo proteins containing a leucine-rich p

nuclear-export signal (N ES}. In the nucleoplasm

(bottom}, the protein exportin 1 binds coopera-
tively to the NES of the cargo protein to be
transported and to Ran·GTP. After the resulting GAP ~
cargo complex diffuses through an NPC via
(
transient interactions with FG repeats in
FG-nucleoporins, the GAP associated with the
NPC cytoplasmic filaments stimulates GTP
hydrolysis, converting Ran·GTP to Ran-GDP. The
accompanying conformational change in Ran Cytoplasm
leads to dissociation of the complex. The
NES-containing cargo protein is released into the
cytosol, whereas exportin 1 and Ran·GDP are
transported back into the nucleus through NPCs.
Ran-GEF in the nucleoplasm then stimulates
conversion of Ran·GDP to Ran·GTP. (b) Ran-
independent nuclear export of mRNAs. The
heterodimeric NXF1/NXT1 complex binds to
mRNA-protein complexes (mRNPs} in the
nucleus. NXF1/NXT1 act as a nuclear export
.____,.. ~ \.)
factor and directs the associated mRNP to the
central channel of the NPC by transiently G
T
p
~
T
'GEF G) D
p p
interacting with FG-nucleoporins. An RNA
Cargo complex
helicase (DbpS) located on the cytoplasmic side
of the NPC removes NXF1 and NXT1 from the
mRNA in a reaction that is powered by ATP
(b)
hydrolysis. Free NXF1 and NXT1 proteins are
Ribosome Cellular mRNA
recycled back into the nucleus by the Ran-
dependent import process depicted in
Figure 13-36.

mRNP
remodeling
ATP ADP

Cytoplasm

Re-import

@
®
~----------------------------
Assembly on processed mRNA

cytoplasmic side of the NPC, NXFl and NXTl dissociate RNA chains and dissociating RNA-protein complexes (Chap-
from the mRNP with the help of the RNA helicase, Dbp5, ter 4). This leads to the simple idea that Dpb5, which associ-
which associates with cytoplasmic NPC filaments. Recall that ates with the cytoplasmic side of the nuclear pore complex,
RNA helicases use the energy derived from hydrolysis of ATP acts as an ATP-driven motor to remove NXFl/NXTl from
to move along RNA molecules, separating double-stranded the mRNP complexes as they emerge on the cytoplasmic side

620 CHAPTER 13 • Moving Proteins into Membranes and Organelles

of the NPC. The assembly of NXFl/NXTl onto mRNPs on
the nucleoplasmic side of the NPC and the subsequent ATP- GTPase-activating protein (GAP) in the cytoplasm creates a
dependent disassembly of NXFl!NXT I from mRNPs on the gradient with high Ran·GTP in the nucleoplasm and Ran·GDP
cytoplasmic side of the NPC creates a concentration gradient of in the cytoplasm. The interaction of import cargo complexes
mRNP-NXFl/NXTl which drive!> unidirectional export. After with the Ran-GTP in the nucleoplasm causes dissociation of
being removed from the mRNP, the free NXFl and NXTl pro- the complex, releasing the cargo into the nucleoplasm (see Fig-
teins that have been stripped from the mRNA by Dbp5 helicase ure 13-36), whereas the assembly of export cargo complexes is
are imported back into the nucleus by a process that depends stimulated by interaction with Ran-GTP in the nucleoplasm
on Ran and a nuclear transport receptor (Figure 13-37b). (see Figure 13-37).
In Ran-dependent nuclear export (discussed in the previ- • Most mRNPs are exported from the nucleus by binding to
ous subsection), hydrolysis of GTP by Ran on the cytoplas- a heterodimeric mRNP exporter in the nucleoplasm that in-
mic side of the NPC causes dissociation of the nuclear teracts with FG-rcpeats. The direction of transport (nucleus
transport receptor from its cargo. In basic outline, the Ran- to cytoplasm) results from the action of an RNA helicase as-
independent nuclear export discussed here operates by a sociated with the cytoplasmic filaments of the nuclear pore
similar mechanism except that Dbp5p on the cytosolic side complexes that removes the heterodimeric mRNP exporter
of the NPC uses hydrolysis of ATP to dissociate the mRNP once the transport complex has reached the cytoplasm.
exporter from mRNA.

KEY CONCEPTS of Section 13.6 Perspectives for the Future

Transport into and out of the Nucleus As we have seen in this chapter, we now understand many
aspects of the basic processes responsible for selectively
• The nuclear envelope contains numerous nuclear pore
transporting proteins into the endoplasmic reticulum (r R),
complexes (NPCs), which are large, complicated structures
mitochondrion, chloroplast, peroxisome, and nucleus. Bio-
composed of multiple copies of 30 proteins called nucleo-
chemical and genetic studies, for instance, have identified
porins (see Figure 13-33). FG-nucleoporins, which contain
signal sequences responsible for targeting proteins ro the
multiple repeats of a short hydrophobic sequence (FG-rcpeats),
correct organelle membrane and the membrane receptors
line the central transporter channel and play a role in transport
that recognize these signal sequences. We also have learned
of all macromolecules through nuclear pores.
much about the underlying mechanisms that translocate pro-
• Transport of macromolecules larger than 20-40 kDa teins a<.:ross organelle membranes and have determined
through nuclear pores requires the assistance of nuclear whether energy is used to push or pull proteins across the
transport receptors that interact with both the transported membrane in one Jirection, the type of channel through
molecule and FG-repeats of FG-nucleoporins. which proteins pass, and whether proteins are translocated
• Proteins imported to or exported from the nucleus contain in a folded or an unfolded state. Nonetheless, fundamental
a specific amino acid sequence that functions as a nuclear- questions remain unanswered; probably the most puzzling is
localization sigi1al (NLS) or a nuclear-export signal (NES). how fully folded proteins move across a membrane.
Nucleus-restricted proteins contain an NLS but not anNES, The peroxisomal import machinery provides one exam-
whereas proteins that shuttle between the nucleus and cyto- ple of the translocation of folded proteins. It not only is ca-
plasm contain both signals. pable of translocating fully folded proteins with bound
cofactors into the peroxisomal matrix, it can even direct the
• Several different types of NES and NLS have been identi-
import of a large golJ particle decorated with a (PTSl) per-
fied. Each type of nuclear-transport signal is thought to in-
oxisomal targeting peptide. Some researchers have specu-
teract with a specific nuc1ear transport protein belonging to
lated that the mechanism of peroxisomal import may be
a family of homologous proteins.
related to that of nuclear import, the best-understood ex-
• A cargo protein bearing anNES or NLS trans locates through ample of post-translational translocation of folded proteins.
nuclear pores bound to its cognate nuclear transport protein. Both the peroxisomal and nuclear import machinery can
The transient interactions between nuclear transport receptors transport folded molecules of very divergent sizes, and both
and FG-repeats allow very rapid diffusion of nuclear transport appear to involve a component that cycles between the cyto-
protein-cargo complex through the central channel of the sol and the organelle interior-the Pex5 PTSJ receptor in the
NPC, which is filled with a hydrophobic matrix of FG-repeats. case of peroxisomal import and the Ran-importin complex
• The unidirectional nature of protein export and import in the case of nuclear import. However, there also appear to
through nuclear pores results from participation of Ran, a be crucial differences benveen the two translocation processes.
monomeric G protein that exists in different conformations For example, nuclear pores represent large, stable macromo-
when bound to GTP or GDP. Localization of the Ran guanine lecular assemblies that are readily observed by electron mi-
nucleotide-exchange factor (GEF) in the nucleus and of Ran croscopy, whereas analogous porelike structures have not
been observed in the peroxisomal membrane. Moreover, small

Perspectives for the Future 621

molecules can readily pass through nuclear pores, whereas Ran protein 618 stop-transfer anchor
peroxisomal membranes maintain a permanent barrier to roughER 580 sequence 588
the diffusion of small hydrophilic molecules. Taken together, signal-anchor topogenic sequences 587
these observations suggest that peroxisomal import may re- sequence 588 topology (membrane
quire an entirely new type of translocation mechanism. protein) 587
signal-recognition particle
The evolutionarily conserved mechanisms for translocat-
(SRP) 582 translocon 583
ing folded proteins across the cytoplasmic membrane of bac-
terial cells and across the thylakoid membrane of chloroplasts signal (uptake-targeting) trimolecular cargo
also are poorly undero;rood. A better understanding of all of sequences 579 complex 619
these processes for translocating folded proteins across a single-pass membrane unfolded-protein
membrane will likely hinge on future development of in vitro proteins 587 response 599
translocation systems that allow investigators to define the
biochemical mechanisms driving translocation and to iden-
tify the structures of trapped translocation intermediates.
Review the Concepts
Compared with our understanding of how soluble proteins 1. The following results were obtained in early studies on the
are translocated into the ER lumen and mitochondrial matrix, translation of secretory proteins. Based on what we now
our understanding of how targeting sequences specify the to- know of this process, explain the reason why each result was
pology of multipass membrane proteins is quite elementary. observed. (a) An in vitro translation system consisting only of
For instance, we do not know how the translocon channel ac- mRNA and ribosomes resulted in secretory proteins that were
commodates polypeptides that are oriented differently with re- larger than the identical protein when translated in a cell. (b)
spect to the membrane, nor do we understand how local A similar system that also included microsomes produced se-
polypeptide sequences interact with the translocon channel cretory proteins that were identical in size to those found in a
both to set the orientation of transmembrane spans and to sig- cell. (c) When the microsomes were added af~er in vitro trans-
nal for lateral passage into the membrane bilayer. A better un- lation, the synthesized proteins were again larger than those
derstanding of how the amino acid sequences of membrane made in a cell.
proteins can specify membrane topology will be crucial for de-
2. Describe the source or sources of energy needed for uni-
coding the vast amount of structural information for membrane
directional translocation across the membrane in (a) cotrans-
proteins contained within databases of genomic sequences.
lational translocation into the endoplasmic reticulum (ER);
A more detailed understanding of all translocation pro-
(b) post-translational translocation into the ER; (c) translo-
cesses should continue to emerge from genetic and biochemical
cation into the mitochondrial matrix.
studies, both in yeasts and in mammals. These studies will un-
doubtedly reveal additional key proteins involved in the recog- 3. Translocation into most organelles usually requires the ac-
nition of targeting sequences and in the translocation of proteins tiviry of one or more cyrosolic proteins. Describe the basic func-
across lipid bilayers. Finally, the structural studies of translocon tion of three different cytosolic factors required for translocation
channels will likely be extended in the future to reveal, at reso- into the ER, mitochondria, and peroxisomes, respectively.
lutions on the atomic scale, the conformational states that are 4. Describe the typical principles used to identify topogenic
associated with each step of the translocation cycle. sequences within proteins and how these can be used to de-
velop computer algorithms. How does the identification of
topogenic sequences lead to prediction of the membrane ar-
rangement of a multipass protein? What is the importance of
the arrangement of positive charges relative to the mem-
brane orientation of a signal-anchor sequence?
Key Terms 5. An abundance of misfolded proteins in the ER can result
cotranslational N-linked in the activation of the unfolded protein response (UPR) and
translocation 582 oligosaccharides 594 ER-associated degradation (ERAD) pathways. UPR de-
creases the amount of unfolded proteins by altering gene ex-
dislocation 600 nuclear pore complex
pression of what type of genes? What is one manner in which
dolichol phosphate 595 (NPC) 615
ERAD may identify misfolded proteins? Why is dislocation
FG-nucleoporins 615 nuclear transport of these misfolded proteins to the cytoplasm necessary?
receptor 618
general import pore 604 6. Temperature-sensitive yeast mutants have been isolated
O-lin ked that block each of the enzymatic steps in the synthesis of the
hydropathy profile 593 oligosaccharides 594 dolichol-oligosaccharide precursor for N-linked glycosyl-
microsomes 580 post-translational ation. Propose an explanation for why mutations that block
molecular chaperones 585 translocation 584 synthesis of the intermediate with the structure dolichol-PP-
multipass membrane protein disulfide (GlcNAchMan5 completely prevent addition of N-linked
proteins 588 isomerase 597 oligosaccharide chains to secretory proteins, whereas mutations

622 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 - -----~·
.....

that block conversion of this intermediate into the completed polypeptide ending with the last codon included on the
precursor-dolichoi-PP-(GicNAchMan 9 Gic 3-allow the mRNA will remain attached to the ribosome, thus allowing
addition of N-linked oligosaccharide chains to secretory a polypeptide of defined length to extend from the ribosome.
glycoproteins. You have generated a set of mRNAs that encode segments of
7. Name four different proteins that facilitate the modifica- the N-terminus of prolactin of increasing length, and each
tion and/or folding of secretory proteins within the lumen of mRNA can be translated in vitro by a cytosolic translation
the ER. Indicate which of these proteins covalently modifies extract containing ribosomes, tRNAs, aminoacyl-tRNA syn-
substrate proteins and which brings about only conforma- thetases, GTP, and translation initiation and elongation fac-
tional changes in substrate proteins. tors. When radiolabeled amino acids an." included in the
8. Describe what would happen to the precursor of a mito- translation mixture, only the polypeptide encoded by the
chondrial matrix protein in the following types of mitochon- added mRNA will be labeled. After completion of translation,
drial mutants: (a) a mutation in the Tom22 signal receptor; each reaction mixture was resolved by SDS poly-acrylamide
(b) a mutation in the Tom70 signal receptor; (c) a mutation gel electrophoresis, and the labeled polypeptides were identi-
in the matrix Hsc70; and (d) a mutation in the matrix signal fied by autoradiography.
peptidase. a. The autoradiogram depicted below shows the results
9. Describe the similarities and differences between the of an experiment in which each translation reaction was car-
mechanism of import into the mitochondrial matrix and the ried out either in the presence (+) or the absence (-) of mi-
chloroplast stroma. crosomal membranes. Based on the gel mobility of pcptides
synthesized in the presence or absence of microsomes, deduce
10. Design a set of experiments using chimeric proteins,
how long the prolactin nascent chain must be in order for the
composed of a mitochondrial precursor protein fused to di-
prolactin signal peptide to enter the ER lumen and to be
hydrofolate reductase (DHFR), that could be used to deter-
cleaved by signal peptidase. (Note that microsomes carry sig-
mine how much of the precursor protein must protrude into
nificant quantities of SRP weakly bound to the membranes. )
the mitochondrial matrix in order for the matrix-targeting
sequence to be cleaved by the matrix-processing protease.
11. Peroxisomes contain enzymes that use molecular oxy-
gen to oxidize various substrates, but in the process hydro- _,j
-- -- --
- - -
--
<ll
gen peroxide-a compound that can damage DNA and ]l~
proteins-is formed. What is the name of the enzyme re-

--
~a.
-<ll
sponsible for the breakdown of hydrogen peroxide to water? Oa.
What is the mechanism of the import of this protein into the .~%
peroxisome, and what proteins are involved? ~ 0.~--------------------------------------~
+ + + + + +
12. Suppose that you have identified a new mutant cell line 50 70 90 110 130 150
that lacks functional peroxisomes. Describe how you could Size of mRNA (in codons)
determine experimentally whether the mutant is primarily
defective for insertion/assembly of peroxisomal membrane
b. Given this length, what can you conclude about the
proteins or matrix proteins.
conformational state(s) of the nascent prolactin polypeptide
13. The nuclear import of proteins larger than 40 kDa re- when it is cleaved by signal peptidase? The following lengths
quires the presence of what amino acid sequence? Describe will be useful for your calculation: the prolactin signal se-
the mechanism of nuclear import. How are nuclear transport quence is cleaved after amino acid 31; the channel within the
receptors able to get through the nuclear pore complex? ribosome occupied by a nascent polypeptide is about 150 A
14. Why is localization of Ran-GAP in the nucleus and long; a membrane bilayer is about 50 A thick; in polypep-
Ran-GEF in the cytoplasm necessary for unidirectional tides with an a-helical conformation, one residue extends
transport of cargo proteins containing an NES? 1.5 A, whereas in fully extended polypeptides, one residue
extends about 3.5 A.
c. The experiment described in part {a) is carried out in
an identical manner except that microsomal membranes are
Analyze the Data not present during translation but are added after translation
1. Imagine that you are evaluating the early steps in translo- is complete. In this case none of the samples shows a differ-
cation and processing of the secretory protein prolactin. By ence in mobility in the presence or absence of microsomes.
using an experimental approach similar to that shown in What can you l:ondude about whether prolactin can be
Figure 13-7, you can use truncated prolactin mRNAs to con- translocated into isolated microsomes post-translationally?
trol the length of nascent prolactin polypeptides that are syn- d. In another experiment, each translation reaction was
thesized. When prolactin mRNA that lacks a chain-termination carried out in the presence of microsomes, and then the mi-
(stop) codon is translated in vitro, the newly synthesized crosomal membranes and bound ribosomes were separated

Analyze the Data 623

from free ribosomes and soluble proteins by centrifugation. q,o iY~ q,o iY~
~><.: .~'l><; o/-
For each translation reaction, both the total reaction (T) and <.q, <.0 .~'l>'
~-.::. <::-' <::-'
the membrane fraction (M) were resolved in neighboring gel v )v v<::-"' )v

lanes. Based on the amounts of labeled polypeptide in the A B

membrane fractions in the autoradiogram depicted below,

--
deduce how long the prolactin nascent chain must be in
order for ribosomes engaged in translation to engage the
SRP and thereby become bound to microsomal membranes.
108

90
-
-
-- - -
75

~i
-
~:2
~a.
Oc..
-
a>>-
a>

.!::!o
<ll

<ll
- --------
--
45
22
5
--
-
Actin

~ c..L-------------------------------------~
T M T M T M T M T M T M
50 70 90 110 130 150 Label each blot with the antibody that was used for the ana l-
Size of mRNA (in codons) ysis. How do you explain the increase in the intensity of the
signal seen in the juniferdin-treated cells in blot B?
c. An immunocytochemistry and fluorescence microscopy
2. Recently, researchers discovered that treating mamma- analysis was undertaken with the antibody used in blot Band
lian cells with juniferdin, a plant-derived compound, affects a secondary antibody labeled with rhodamine (red). Since ju-
protein secretion, and have reported that the target of this drug niferdinspecifically affects PDI, a resident rough ER (RER)
is protein disulfide isomerase (POI). In the following experi- protein, how do you explain the localization of the signal in
ment, cultured pancreatic 13-cells were treated with juniferdin the nucleus?
and protein lysates were isolated and compared to lysates
from untreated cell s using immunoblot analysis. Probing
blots with antibodies against POI (57 kDa), actin (43 kOa)
and pro-insulin (9 .8 kDa), show the following:

.....
Pro-insulin .u
••

88
--
Nucleus
J
- - -
60

-- - - -- --
55

45
3. Antibody labeling of proteins like that used in immuno-
23
9
- fluorescence analysis can be applied to electron microscopy,
but instead of using fluorescent labels attached to antibod-
ies, investigators use gold particles that are electron dense
and appear as uniform dots in an electron micrograph. Fur-
a. Given that approxtmately the same amount of protein thermore, by varying the size of the gold particles (e.g., 5 nm
was loaded in each lane, as evidenced by the actin signals, vs 10 nm), one can identify the localization of more than one
how do you explain the fact that the PDI levels also appear protein in the cell.
about the same, while most of the pro-insulin remains accu- a. Using this approach, investigators have determined
mulated in the juniferdin-treated cells? the subcellular localization of Tim and Tom proteins used
b. To confirm your results, protein lysates of juniferdin- for protein import into mitochondria. In the drawing of the
treated and untreated cells were separated by SDS-PAGE results below, label the gold particles showing the localiza-
and blotted to membranes and then probed with antibodies tion of Tim44 and Tom40. What made you come to these
against lrel and Hac1. conclusions?

624 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 How do you explain the apparent shift in the molecular mass
of actin in the mitochondrial fraction of cyanide-treated cells?
Note that there is a similar shift in the mass of the succinate
dehydrogenase A control. What would you expect to see on
blots if cyanide-treated cells were supplemented with ATP?

Mitochondrion
References
b. Genetically engineering theN-terminus of alcohol dehy-
drogenase (ADH) onto a cytosolic protein alters that protein's Targeting Proteins to and Across the ER Membrane
localization. The following blot was seen when cells were Egca, P. F., R. M. Stroud, and P. Walter. 2005. Targermg
transfected with an ADH-actin chimeric construct, and pro- proteins ro membranes: structure of the signal recognition particle.
teins were isolated from the cytosol and mitochondria. Anti- Curr. Opzn. Struc. Bzol. 15:213-220.
Osborne, A. R., T. A. Rapoport, and B. van den Berg. 2005.
bodies against actin (43 kDa), the cytosolic protein GAPDH
Protein translocation by the Sec61/SecY channel. Ann. Rev. Cell
(37 kDa), and the mitochondrial inner membrane protein suc-
Dev. Brol. 21:529-550.
cinate dehydrogenase A (72 kDa), were used in the analysis. Wickner, W., and R. Schekman. 2005. Protem translocation
across b1ological membranes. Sctence 310:1452-1456.

Insertion of Membrane Proteins into the ER

Englund, P. T. 1993. The structure and biosynthesis of glyco-
sylphosphatidylinositol protein anchors. Ann. Rev. Biochem.

-- -
62:121-138.
··' Mothes, W., et al. 1997. Molecular mechanism of membrane
75 protein inregranon into the endoplasmic reticulum. Cell89:523-533.
Shao, S., and R. S. Hegde. 2011. Membrane protein insertion at
52 the endoplasmic reticulum. Ann. Rev. Cell Dev. Bioi. 27:25-56.

45

35
23
-- -- - Wang, F., et al. 2011. The mechanism of tail~anchored protem
insertion into the ER membrane. Mol. Cell43:738-750.

Protein Modifications, Folding, and Quality Control in the ER

Braakman, 1., and N.J. Bulleid. 2011. Protein folding and modi-
fication in the mammalian endoplasmic reticulum. Ann. Rev.
Biochem. 80:71-99.
How do you explain the presence of actin in two distinct sub- Hegde, R. S., and H. L. Ploegh. 2010. Quality and quantity
cellular pools of protein? How do you explain it being the same control at the endoplasmic reticulum. Curr. Opin. Cell Bioi.
molecular mass in both pools? If the blot were stripped and re- 22:437-446.
Helenius, A., and M. Aebi. 2004. Roles of N-linked glycans in
pro bed with an antibody against theN-terminus of alcohol
the endoplasmic reticulum. Ann. Rev. Brochem. 73:1019-1049.
dehydrogenase; where would you expect to find the signal?
Kornfeld, R., and S. Kornfeld. 1985. Assembly of asparagine-
c. Cyanide is toxic to cells because it inhibits a specific Hnked oligosaccharides. Ann. Rev. Biochem. 45:631-664.
mitochondrial protein complex that is responsible for pro- Pari!, C., and P. Walter. 2001. Intracellular signaling from the
ducing ATP. An experiment like that described above was endoplasmic reticulum to the nucleus: the unfolded protem response
repeated and the results compared to those for cells exposed in yeast and mammals. Curr. Opin. Cell Bioi. 13:349-355.
to hydrogen cyanide. The following are the results from the Meusser, B., et al. 2005. ERAD: the long road to destruction.
Nat. Cell Bioi. 7:766-772.
immunoblot analysis:
Sevier, C. S., and C. A. Kaiser. 2002. Formation and transfer of
~,'b ~,'b
disulph1de bonds in living cells. Nat. Rev. Mol. Cell Bioi. 3:836-847.
~~: ~~ Tsa1, B., Y. Ye, and T. A. Rapoport. 2002. Retro-translocation
.,o' O'o .,<5- O'o of proteins from the endoplasmic reticulum into the cytosol. Nat.
~0 ·~0 ~0 ·~0
G ~' G ~ Rev. Mol. Cell Bioi. 3:246-255.
Cyanide

- - -
Untreated
Targeting of Proteins to Mitochondria and Chloroplasts
92 Dalbey, R. E., and A. Kuhn. 2000. Evolutionanly related

-- -- - -
75 insertion pathways of bacterial, mitochondrial, and thylakoid
membrane proteins. Ann. Rev. Cell Dev. Bro/. 16:51-87.
52 Dolezal, P., et al. 2006. Evolution of the molecular machines for
protem import into m1tochondna. Science 313:314-318.

-- -
45
Koehler, C. M. 2004. New developments in mitochondrial
35 assembly. Ann. Rev. Cell Dev. Brol. 20:309-335.
23 Li, H.-M., and C. -C. Chiu. 2010. Protein transport into
chloroplasts. Ann. Rev. Plant Bioi. 61:157-180.

References 625
 Matouschek, A., N. Pfanner, and W. Voos. 2000. Protein Ma, C., G. Agrawal, and S. Subramani. 20 I I. Peroxisome
unfoldmg by mitochondria: the H~p70 import motor. l:.MBO Rep. assembly: matrix and membrane protein biogenesis. J. Cell Btol.
1:404-410. 193:7-16.
Neupert, W., and M. Brunner. 2002. The protein import motor Purdue, P. E., and P. B. Lazarow. 2001. Peroxisome biogenesis.
of mitochondna. Nat. Rev. Mol. Cell Bioi. 3:555-565. Ann. Rev. Cell Dev. Bioi. 17:701-752.
Rapaport, D. 2005. How doe~ the TOM complex mediate
insertion of precursor proteins into the mitochondrial outer Transport into and out of the Nucleus
membrane?]. Cell Bioi. 171:419-423. Chook, Y. M., and G. Blobel. 2001. Karyopherins and nuclear
Robinson, C., and A. Bolhuis. 2001. Protein targetmg by the import. Curr. Opm. Struc. Bioi. 11:'03-715.
rwin-argmine translocation pathway. Nat. Reu. Mol. Cell Bioi. Cole, C. N., and J. J. Scarcelh. 2006. Transport of messenger
2:350-356. RNA from the nucleus to the cytoplasm. Curr. Opin. Cell Bioi.
Truscott, K. N., K. Brandner, and N. Pfanner. 2003. i\1echamsms 18:299-306.
of protem 1mport into mnochondna. Curr. Brol. 13:R326-R337. Johnson, A. W., E. Lund, and J. Dahlberg. 2002. Nuclear
ell.porr of nbosomal subumts. Trends Biochem. Scr. 27:580-585.
Targeting of Peroxisomal Proteins Ribbeck, K., and D. Gorlich. 200 I. Kinetic analysis of translo-
Dammai, V., and S. Subramani. 200 I. The human peroxisomal cation through nuclear pore complexes. l:.MBO J. 20:
targeting signal receptor, Pex5p, IS tran~located inro the perox1somal 1320-1330.
matrix and recycled to the cytosol. Celll05:187-196. Rout, ;\11. P., and J.D. Aitchi~on. 2001. The nuclear pore
Gould, S. J., and C. S. Collins. 2002. Opinion: perOXISOmal- complex as a transport machine.}. Biof Chem. 276:16593-16596.
protem 1mport: is it really that complex? Nat. Reu. Mol. Cell Brol. ~chwarrz, T. U. 2005. Modularity within the architecture of the
3:382-389. nuclear pore complex. Curr. Opin. Struc. Bioi. 15:221-226.
Gould, S. J., and D. Valle. 2000. Perox1some biogenes1s Stewart, M. 20 I 0. Nuclear export of mRNA. Trends Biochem.
disorders: genetiCS and cell h1ology. Tre11ds Genet. 16: Sci. 35:609-617.
340-345. Terry, L. J., and S. R. Wente. 2009. flexible gates: dynamic
Hoepfner, D., et al. 2005. Contribution of the endoplasmic topologies and funcnons for FG nucleoporins in nucleocytoplasmic
reticulum to peroxisome formation. Cell122:85-95. transport.l:.ukaryot. Cel/8:1814-1827.

626 CHAPTER 13 • Moving Proteins into Membranes and Organelles

 CHAPTER

Vesicular Traffic,
Secretion, and
Endocytosis
Scanning electron micrograph showing the formation of clathrin-
coated vesicles on the cytosolic face of the plasma membrane. [John
Heuser, Washington University School of Medicine.]

n the previous chapter we explored how proteins are tar- an acidic interior that is generally used for degradation of

I geted to and translocated across the membranes of several

different intracellular o rganelles, including the endoplas-
mic reticulum, mitochondria and chloroplasts, peroxisomes,
unneeded proteins and the storage of small molecules such
as amino acids. Accordingly, the types of proteins delivered
to the lysosomal membrane include subunits of the V-class
and the nucleus. In this chapter we turn our attention to the proton pump that pumps H + from the cytosol into the acidic
secretory pathway and the mechanisms of vesicular traffic lumen of the lysosome, as well as transporters that release
that allow proteins to be secreted from the cell or delivered small molecules stored in the lysosome into the cytoplasm.
to the plasma membrane and the lysosome. We wi ll also dis- Soluble proteins delivered by this pathway include lysosomal
cuss the related processes of endocytosis and autophagy, digestive enzymes such as proteases, glycosidases, phospha-
which deliver proteins and small molecules from either out- tases, and lipases.
side the cell or from the cytoplasm to the interior of the lyso- In contrast to the secretory pathway, which allows pro-
some for degradation. teins to be targeted to the cell surface, the endocytic pathway
The secretory pathway carries both soluble and mem- is used to take up substances from the cell su rface and move
brane proteins from the ER to their final destination at the them into the interior of the cell. The endocytic pathway is
cel l surface or in the lysosome. Proteins delivered to the used to ingest certain nutrients that are too large to be trans-
plasma membrane include cell-surface receptors, transport- ported across the plasma membrane by one of the transport
ers for nutrient 'uptake, and ion channels that maintain the mechanisms discussed in Chapter 11. For example, the endo-
proper ionic and electrochemical balance across the plasma cytic pathway is utilized in the uptake of cholesterol carried
membrane. Such membrane proteins, once they reach the in LDL particles, and iron atoms carried by the iron-binding
plasma membrane, become embedded within it. Soluble se- protein transferrin. In add ition, the endocytic pathway can
creted proteins also follow the secretory pathway to the cell be used to remove receptor proteins from the cell surface as
surface, but instead of remaining embedded in the mem- a way to down-regulate their activity.
brane they are released into the aqueous extracellular envi- A single unifying principle governs all protein trafficking
ronment. Examples of sec reted proteins are digestive in the secretory and endocytic pathways: transport of mem-
enzymes, peptide hormones, serum proteins, and collagen. brane and soluble proteins from one membrane-bounded
As described in Chapter 9, the lysosome is an organelle with compartment to another is mediated by transport vesicles

OUTLINE

14.1 Techniques for Studying the Secretory Pathway 629 14.4 Later Stages of the Secretory Pathway 646

14.2 Molecular Mechanisms of Vesicle Budding 14.5 Receptor-Mediated Endocytosis 654

and Fusion 634
14.6 Directing Membrane Proteins and Cytosolic
14.3 Early Stages of the Secretory Pathway 640 Materials to the Lysosome 661
0 OVERVIEW ANIMATION: Protein Secretion
FIGURE 14-1 Overvi ew of the Exterior
Plasma
secretory and endocytic pathways
of protein sorting. Secretory pathway: Cytosol ~ membrane

Synthesis of proteins bearing an ER n Consti.tutive /

signal sequence is completed on the Regul~ted
lfJ secret ton
( U secrett on (
II En docytosis \
rough ER D. and the newly made
polypeptide chains are inserted into
the ER membran~ or cross it into the
lumen (Chapter 13). Some proteins (e.g.,
ER enzymes or structural proteins)
~ Sec.retory
~ vestcle

(
e ~ En~ocytic
~ vestcle

remain within the ER. The remainder n Sorti ng to )

g lysosomes
are packaged into transport vesicles fJ

-+ @ ~ q) Late
that bud from the ER and fuse to form
new cis-Golgi cisternae. Missorted ER-
resident proteins and vesicle mem-
brane proteins that need to be reused
.,g.,
Tra~sport endosome
vestcle '
are retrieved to the ER by vesicles
iJ that bud from the cis-Golgi and fuse
with the ER. Each cis-Golgi cisterna,
with its protein content, physically
moves from the cis to the trans face of
the Golgi complex !J by a nonvesicu-
lar process called cisternal maturation. ..
Retrograde transport vesicles IJ move
Golgi-resident proteins to the proper
Golgi compartment. In all cells, certain
soluble proteins move to the cell
surface in transport vesicles m and are
Lysosome
secreted continuously (constitutive
mediai-
secretion). In certain cell types, some Golgi
soluble proteins are stored in secretory Retrograde
vesicles 6 and are released only after
the cell receives an appropriate neural
Cisternal
maturation
t II •
transport from
later to earlier
Golgi cisternae
or hormonal signal (regulated secre-
tion). Lysosome-destined membrane
and soluble proteins, which are trans-
ported in vesicles that bud from the
trans-Golgi 1;1, first move to the late t
endosome and then to the lysosome.
Endocytic pathway: Membrane and cis-Golgi
soluble extracellular proteins taken up network
in vesicles that bud from the plasma
membrane m also can move to the
lysosome via the endosome. @fJ Budding and fusion of
ER-to-Golgi vesicles to Retrograde Golgi-to-ER
form cis-Golgi transport

'
RoughER Protein synthesis on bound ribosomes;
D co-translationa l transport of proteins
into or across ER membrane

628 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 that collect "cargo" proteins in buds arising from the mem- to the plasma membrane and secreted immediately, stored for
brane of one compartment and then deliver these cargo pro- later release, or shipped to the lysosome (steps [i!--(3 ). The
teins to the next compartment by fusing with the membrane process by which a vesicle moves to and fuses with the plasma
of that compartment. Importantly, as transport vesicles bud membrane and releases its contents is known as exocytosis.
from one membrane and fuse with the next, the same face of In all cell types, at least some proteins are secreted continu-
the membrane remains oriented toward the cytosol. There- ously, while others are stored inside the cell until a signal for
fore once a protein has been inserted into the membrane or exocytosis causes them to be released. Secretory proteins des-
the lumen of the ER, the protein can be carried along the tined for lysosomes are first transported by vesicles from the
secretory pathway, moving from one organelle to the next trans-Golgi nenvork to a compartment usually called the late
without being translocated across another membrane oral- endosome; proteins then are transferred to the lysosome by
tering its orientation within the membrane. Similarly, the direct fusion of the endosome with the lysosomal membrane.
endocytic pathway uses vesicle traffic to transport proteins Endocytosis is related mechanistically to the secretory
from the plasma membrane to the endosome and lysosome pathway. In the endocyric pathway, vesicles bud inward
and thus preserves their orientation in the membrane of from the plasma membrane, bringing membrane proteins
these organelles. Figure 14-1 outlines the main secretory and and their bound ligands into the cell (see Figure 14 1, right).
endocytic pathways in the cell. After being internalized by endocytosis, some proteins are
Reduced to its simplest elements, the secretory pathway transported to lysosomes via the late endosome, whereas
operates in two stages. The first stage takes place in the others are recycled back to the cell surface.
rough endoplasmic reticulum (ER) (Figure 14-1, step 0 ). As In this chapter we first discuss the experimental techniques
described in Chapter 13, newly synthesized soluble and that have contributed to our knowledge of the secretory path-
membrane proteins are translocated into the ER, where they way and endocytosis. Then we focus on the general mecha-
fold into their proper conformation and receive covalent nisms of membrane budding and fusion. We will see that
modifications such as N-linked and 0-linked carbohydrates although different kinds of transport vesicles utilize distinct
and disulfide bonds. Once newly synthesized proteins are sets of proteins for their formation and fusion, all vesicles use
properly folded and have received their correct modifica- the same general mechanism for budding, selection of particu-
tions in the ER lumen, they progress to the second stage of lar sets of cargo molecules, and fusion with the appropriate
the secretory pathway, transport to and through the Golgi. target membrane. In the remaining sections of the chapter, we
The second stage of the secretory pathway can be sum- discuss both the early and late stages of the secretory pathway,
marized as follows. In the ER, cargo proteins are packaged including how specificity of targeting to different destinations
into anterograde (forward-moving) transport vesicles (figure is achieved, and conclude with a discussion of how proteins are
14-1, step fJ). These vesicles fuse with each other to form a transported to the lysosome by the endocytic pathway.
flattened membrane-bounded compartment known as the
cis-Golgi network or cis-Golgi cisterna (a "cistern" is a con-
tainer for holding water or other liquid). Certain proteins,
mainly proteins that function in the ER, can he retrieved 14.1 Techniques for Studying
from the cis-Golgi cisterna to the ER via a different set of the Secretory Pathway
retrograde (backward-moving) transport vesicles (step II). In
a manner reminiscent of an assembly line, the new cis-Golgi The key to understanding how proteins are transported
·. cisterna with its cargo of proteins physically moves from the through the organelles of the secretory pathway has been to
develop a basic description of the function of transport vesi-
cis position (nearest the ER) to the trans position (farthest
from the ER), successively becoming first a medw/-Golgi cis- cles. Many components required for the formation and fu-
terna and then a trans-Golgi cisterna (step 19). This process, sion of transport vesicles have been identified in the past
known as cisternal maturqtion, primarily involves retrograde decade by a remarkable convergence of the genetic and bio-
transport vesicles (step 11.1), w hich retrieve enzyme~ and other chemical approaches described in this section. All studies of
Golgi-resident proteins from later to earlier Golgi cisternae, intracellular protein trafficking employ some method for as-
thereby "maturing" the cis-Golgi to the medial-Golgi, and saying the transport of a given protein from one compart-
the mediai-Golgi to the trans-Golgi. As secretory proteins ment to another. We begin by describing how intracellular
move through the Golgi, they can receive further modifica- protein transport can be followed in living cells and then
tions to linked carbohydrates by specific glycosyl transferascs consider genetic and in vitro systems that have proved useful
that are housed in the different Golgi compartments. in elucidating the secretory pathway.
Proteins in the secretory pathway are eventually delivered
to a complex network of membranes and vesicles termed the
Transport of a Protein Through the Secretory
traus-Golgi network (TGN). The TGN is a major branch
point in the secretory pathway. It is at this point that proteins Pathway Can Be Assayed in Living Cells
are loaded into different kinds of vesicles and thereby traf- The classic studies of G. Palade and his colleagues in the
ficked to different destinations. Depending on which kind of 1960s first established the order in which proteins move
vesicle the protein is loaded into, it will be either transported from one organelle to the next in the secretory pathway.

14.1 Techniques for Studying the Secretory Pathway 629

These early studies also showed that secretory proteins are virus (VSV) is introduced into cultured mammalian cells ei-
never released into the cytosol, the firs t indication that trans- ther by transfection or simply by infecting the cells with the
ported proteins are a lways associated with some type of virus. The treated cells, even those that are not speciali zed
membrane-bounded intermediate. In these experiments, for secretion, rapidly synthesize the VSV G protein on the
which combined pulse-chase labeling (see Figure 3-40) and ER like normal cellular secretory proteins. Use of a mutant
autoradiography, radioactively labeled amino acids were in- encoding a temperatu re-sensitive VSV G protein allows re-
jected into the pancreas of hamsters. At different times after searchers to turn subsequent transport of this protein on and
injection, the animals were sacrificed and the pancreatic cells off. At the restrictive temperature of 40 °C, newly made VSV
were immediately fixed with glutara ldehyde, sectioned, and G prutein is misfolded and therefore retained within the ER
subjected to autorad iography to visualize the location of the by quality-control mechanisms discussed in Chapter 13,
radiolabeled proteins. Because the radioactive amino acids whereas at the permissive temperature of 32 °C, the protein
were administered in a short pulse, only those proteins syn- is correctly folded and is transported through the secretory
thesized immediately after injection were labeled, forming a pathway to the cell surface. Im portantly, the m isfolding of
distinct cohort of labeled proteins whose transport could be the temperature-sensitive VSV G protein is reversible; thus
followed. In addition, because pancreatic acinar cells are when cell s synthesizing mu tant VSV G protein are grown at
dedicated secretory cells, almost all of the labeled amino 40 °C and then shifted to 32 °C, the misfolded mutant VSV
acids in these cells are incorporated into secretory proteins, G protein that had accumulated in the ER will refold and be
facilitating the observation of transported proteins. transported normally. This clever use of a temperature-sensitive
Although autoradiography is rarely used today to localize mutation in effect defines a protein cohort whose subsequent
proteins within cells, these early experiments illustrate the transport can be followed.
two basic requirements for any assay of intercompartmental In two variations of this basic procedure, transport of
transport. First, it is necessary to label a cohort of proteins in VSV G protein is monitored by different techniques. Studies
an early compartment so that their subsequent transfer to using both of these modern trafficking assays and Palade's
later compartments can be followed with time. Second, it is early experiments all came to the same conclusion: in mam-
necessary to have a way to identify the compartment in which malian cells vesicle-mediated transport of a pr otein molecule
a labeled protein resides. Here we describe two modern ex- from its site of synthesis on the rough ER to its arrival at the
perimental procedures for observing the intracellular traffick- plasma membrane takes from 30 to 60 minutes.
ing of a secretory protein in almost any type of cell.
In both procedures, a gene encoding an abundant mem- Microscopy of GFP-Labeled VSV G Protein One approach for
brane glycoprotein (G protein) from vesicular stomatitis observing transport of VSV G protein employs a hybrid gene

~ VIDEO: Transport ofVSVG-GFPThrough the Secretory Pathway

(a) (b)
0 min 40 min 180 min 20
Total

.g; 15
2S.
Q_
u.. 10
(.!:)
I
(.!:)
>
(f)
> 5

Plasma
membrane 00
Time (min)

PER ll .NTAL IGURE 14-2 Protein transport through the finally to the cell surface occurred within 180 minutes. The scale bar is
secretory pathway can be visualized by fluorescence microscopy 5 1-1-m. (b) Plot of the levels of VSVG-GFP in the endoplasmic reticulum
of cells producing a GFP-tagged membrane protein. Cultured cells (ER), Golgi, and p lasma membrane (PM) at different times after shift to
were transfected with a hybrid gene encoding the viral membrane lower temperature. The kinetics of transport from one organelle to
glycoprotein VSV G protein linked to the gene for green fluorescent another can be reconstructed from computer analysis of these data.
protein (GFP). A mutant version of the viral gene was used so that The decrease in total fluorescence that occurs at later times probably
newly made hybrid protein (VSVG-GFP) was retained in the ER at 40 oc resu lts from slow inactivation of GFP fluorescence. [From Jennifer
but was released for transport at 32oc. (a) Fluorescence micrographs of Lippincott-Schwartz and Koret Hirschberg, Metabolism Branch, National Institute
cells just before and two times after they were shifted to the lower of Child Health and Human Development.)
temperature. Movement ofVSVG-GFP from the ER to the Go lgi and

630 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 in which the viral gene is fused to the gene encoding green drate side chains that occur at different stages of the secre-
fluorescent protein (GFP), a naturally fluorescent protein tory pathway. To understand this approach, recall that
(Chapter 9). The hybrid gene is rransfected into cultured cells many secretory proteins leaving the ER contain one or more
by techniques described in Chapter 5. When cells expressing copies of theN-linked oligosaccharide Man 8 (GicNAch,
the temperature-sensitive form of the hybrid protein (VSVG- which are synthesized and attached to secretory proteins in
GFP) are grown at the restrictive temperature, VSVG-GFP ac- the ER (see Figure 13-18). As a protein moves through the
cumulates in the ER, which appears as a lacy network of Golgi complex, different enzymes localized to the cis-, medial-,
membranes when cells arc observed in a fluorescent micro- and trans-Golgi cisternae catalyze an ordered series of reac-
scope. When the cells are subsequently shifted to a permissive Liun~ to these core Man8 (GicNAch chams, as discussed in a
temperature, the VSVG-GFP can be seen to move first to the later section of this chapter. For instance, glycosidases that
membranes of the Golgi apparatus, which are densely concen- reside specifically in the cis-Golgi compartment sequentially
trated at the edge of the nucleus, and then to the cell surface trim mannosc residues off the core oligosaccharide to yield a
(Figure 14-2a). By observing the distribution of VSVG-GFP at "trimmed" form, Man,(GlcNAch. Scientists can use a spe-
different times after shifting cells to the permissive tempera- cialized carbohydrate-cleaving enzyme known as endoglyco-
ture, researchers have determined how long VSVG-GFP re- sidase D to distinguish glycosylated proteins that remain in
sides in each organelle of the secretory pathway (Figure 14-2b). the ER from those that have entered the cis-Golgi: trimmed
cis-Golgi-specific oligosaccharides are cleaved from proteins
Detection of Compartment-Specific Oligosaccharide Modifi- by endoglycosidase D, whereas the core (untrimmed) oligo-
cations A second way to follow the transport of secretory saccharide chains on secretory proteins within the ER are
proteins takes advantage of modifications to their carbohy- resistant to cleavage by this enzyme (Figure l4-3a). Because

(a) ER Cis-Golgi

Man nose
trimming
(bl Time at 32 oc (mini 0 5 10 15 20 30 45 60
(ERI Resistant'-

t Extract glycoprotein
(cis-Golgi) Sensitive r

_y_
Man 8 (GicNAc) 2
(cl

·-c
0
Ql
(/)
1.0

0.8

t
Ql '0
... "'
0·-
... (/)
)rear with endoglycosidase D Q.O

_y_
No cleavage,
..
Cleavage,
.....
t?
-Cl
~ 0
0'0
... c
oQ)
cB
.2
...
U·-
Ql
>
co .t:
~

... (/) 0.2

u..c
0.6

0.4

endoglycosidase D-resistant endoglycosidase D-sensitive Ql

(/)

0
• = N-Acetylglucosamine 10 20 30 40 50 60
e = Mannose Time (mini

EXP RIMEIN TAL FIGURE 14-3 Transport of a membrane ER. (b) SDS gel electrophoresis of the digestion mixtures resolves the
glycoprotein from the ER to the Golgi can be assayed based on resistant, uncleaved (slower-migrating) and sensitive, cleaved
sensit ivity to cleavage by endoglycosidase D. Cells expressing a (faster-migrating) forms of labeled VSVG. As this electrophoretogram
temperature-sensitive VSV G protein (VSVG) were labeled with a pulse shows, initially all of the VSVG was resistant to digestion, but with time
of radioactive amino acids at the non permissive temperature so that an increasing fraction was sensitive to digestion, reflecting protein
labeled protein was retained in the ER. At periodic tirnt:'s after a return transported from the ER to the Golgi and processed there. In control
to the permissive temperature of 32 C, VSVG was extracted from cells cells kept at 40 oe, only slow-moving, digestion-resistant VSVG was
and digested with endoglycosidase D. (a) As proteins move to the detected after 60 minutes (not shown). (c) Plot of the proportion of
cis-Golgi from the ER, the core oligosaccharide Man 8(GicNAch is VSVG that is sensitive to digestion, derived from electrophoretic data,
trimmed to Man 5(GicNAch by enzymes that reside in the cis-Golgi reveals the time course of ER ~ Golgi transport. [From c. J. Beckers et al.,
compartment. Endoglycosidase D cleaves the oligosaccharide chains 1987, Cell SO:S23.)
from proteins processed in the cis-Golgi but not from proteins in the

14.1 Techniques for Studying the Secretory Pathway 631

.·
a deglycosylated protein produced by endoglycosidase D di- and the cell wall. The best studied of these, invertase, hydro-
gestion moves faster on an SDS gel than the corresponding lyzes the disaccharide sucrose to glucose and fructose.
glycosylated protein, these proteins can be readily distin- A large number of yeast mutants initially were identified
guished (figure 14-3b). based on their ability to secrete proteins at one temperature
This type of assay can be used to track movement of and inability tO do so at a higher, nonpermissive temperature.
VSV G protein in virus-infected cells pulse-labeled with ra- When these temperature-sensitive secretion (sec) mutants are ·.
dioactive amino acids. Immediately after labeling, all the transferred from the lower to the higher temperature, they
labeled VSV G protein is still in the ER and, upon extraction, accumulate secretory proteins at the point in the pathway
IS resistant to digestion by endoglycosidase D, but with time blocked by the mutation. Analysis of such mutants identified
the fraction of the extracted glycoprotein that is sensitive to five classes (A-E) characterized by protein accumulation in
digestion increases. This conversion of VSV G protein from the cytosol, rough ER, small vesicles taking proteins from the
an endoglycosidase D-resistant form to an endoglycosidase ER to the Golgi complex, Golgi cisternae, or constitutive se-
D-sensitive form corresponds to vesicular transport of the cretory vesicles (Figure 14-4 ). Subsequent characterization of
protein from the ER to the cis-Golgi. Note that transport of sec mutants in the various classes has helped elucidate the
VSV G protein from the ER to the Golgi takes about 30 min- fundamental components and molecular mechanisms of ves-
utes as measured by either the assay based on oligosaccha- icle trafficking that we discuss in later sections.
ride processing or fluorescence microscopy of VSVG-GFP To determine the order of the' steps in the pathway, re-
(Figure 14-3c). A variety of assays based on specific carbo- searchers analyzed double sec mutants. For instance, when
hydrate modifications that occur in later Golgi compart- yeast cells contain mutations in both class Band class D func-
ments have been developed to measure progression of VSV tions, proteins accumulate in the rough ER, not in the Golgi
G protein through each stage of the Golgi apparatus. cisternae. Since proteins accumulate at the earliest blocked
step, this finding shows that class B mutations must act at an
earlier point in the secretory pathway than class D mutations
do. These studies confirmed that as a secreted protein is synthe-
Yeast Mutants Define Major Stages and Many
sized and processed, it moves sequentially from the cytosol ~
Components in Vesicular Transport rough ER ~ ER-to-Golgi transport vesicles ~ Golgi cis-
The general organization of the secretory pathway and many ternae~ secretory vesicles and finally is exocyrosed.
of the molecular components required for vesicle trafficking The three methods outlined in this section have delineated
are similar mall eukaryotic cells. Because of this conservation, the major steps of the secretory pathway and have contrib-
genetic studies with yeast have been useful in confirming the uted to the identification of many of the proteins responsible
sequence of steps in the secretory pathway and in identifying for vesicle budding and fusion. Currently each of the indi-
man)' of the proteins that participate in vesicular traffic. Al- vidual steps in the secretory pathway is being studied in
though yeasts secrete few proteins into the growth medium, mechanistic detail, and increasingly, biochemical assays and
they continuous~y secrete a number of enzymes that remain molecular genetic studies are used to study each of these steps
localized in the narrow space between the plasma membrane in terms of the function of individual protein molecules.

Class A Class B Class C Class D Class E

ER

~
Golgi
~
••• .o ••
Fate of Normal Accumulation Accumulation Accumulation Accumulation Accumulation
secretory secretion in the cytosol in roughER in ER-to-Golgi in Golgi in secretory
proteins transport vesicles vesicles

Defective Transport Budding of Fusion of Transport f rom Transport from

function into the ER vesicles from transport vesicles Golgi to secretory secretory vesicles
the rough ER with Golgi vesicles to cell surface
EXPERIMENTAL FIGURE 14-4 Phenot ypes of yeast sec when cells are shifted from the permissive temperature to the higher,
m utants identified f ive stages in the secret ory pathway. These nonpermissive one. Analysis of double mutants permitted the
temperature-sensitive mutants can be grouped into five classes based sequential order of the steps to be determined. [See P. Novicket al., 1981 ,
on the site where newly made secretory proteins (red dots) accumulate Ce// 25:461, and C. A. Kaiser and R. Schekman, 1990, Ce//61 :723.]

632 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 (a) (b)
Golgi isolated from G protein in Golgi from
cis-Golgi mediai-Go lgi trans-Go lgi uninfected wild-type cells infected mutant cells

~y -....... G protein
VSV-infected w ild-type cells
N-A"Wiglooo":
transferase I reaction
~. -§1;)
..
Addition of
N-acetyl-
VSV-infected mutant cells glucosamine
(no N-acetylglucosamine
• = N-Acetylglucosamine e ~ Galactose toG protein
transferase I) e Mannose + = N-Acetylneuraminic acid
EXPERIMENTAL FIGURE 14-5 A cell-free assay demonstrates simpler high-mannose oligosaccharide containing only two
protein transport from one Golgi cisterna to another. (a) A mutant N-acetylglucosamine and five man nose residues. (b) When Golgi
line of cultured fibroblasts is essential in this type of assay.ln this cisternae isolated from infected mutant cells are incubated with Golgi
example, the cells lack the enzyme N-acetylglucosamine transferase I cisternae from normal. uninfected cells, the VSV G protein produced in
(step fJ in Figure 14-14). In wi ld-type cells, this enzyme is localized to vitro contains the additional N-acetylglucosamine. This modification is
the media/-Golgi and modifies N-linked oligosaccharides by the carried out by transferase enzyme that is moved by transport vesicles
addition of one N-acetylglucosamine.ln VSV-infected wild-type cells, from the wild-type media/-Golgi cisternae to the mutant cis-Golgi
the oligosaccharide on the viral G protein is modified to a typical cisternae in the reaction mixture. [See W. E. Balch et al.;1984, Ce//39:405
complex oligosaccharide, as shown in the trans-Golgi panel. In infected and 525; W. A. Braell et al.. 1984, Cell 39:511; and J. E. Rothman and T. Sollner,
mutant cells, however, the G protein reaches the cell surface with a 1997, Science 276:1212.]

Cell-Free Transport Assays Allow Dissection transferase I from the medial- to cis-Golgi can be purified
of Individual Steps in Vesicular Transport away from the donor wild-type Golgi membranes by cen-
trifugation. By examining the proteins that are enriched in
In vitro assays for intercompartmental transport are power- these vesicles, scientists have been able to identify many of
ful complementary approaches to studies with yeast sec mu- the integral membrane proteins and peripheral vesicle coat
tants for identifying and analyzing the cellular components proteins that are the structural components of this type of
responsible for vesicular trafficking. In one application of this vesicle. Moreover, fractionation of the cytosolic extract re-
approach, cultur~d mutant cells lacking one of the enzymes quired for transport in cell-free reaction mixtures has per-
that modify N-linked oligosaccharide chains in the Golgi are mitted isolation of the various proteins required for formation
infected with vesicular stomatitis virus (VSV), and the fate of of transport vesicles and of proteins required for the target-
the VSV G protein is followed. For example, if infected cells lack ing and fusion of vesicles with appropriate acceptor mem-
N-acetylglucosamine transferase I, they produce abundant branes. In vitro assays similar in general design to the one
amounts of VSV G protein but cannot add N-acerylglucosamine shown in Figure 14-5 have been used to study various trans-
residues to the oligosaccha-ride chains in the medial-Golgi as port steps in the secretory pathway.
wild-type cells do (Figure 14-Sa). When Golgi membranes
isolated from such mutant cells are mixed with Golgi mem-
branes from wild-type, uninfected cells, the addition of N-
acetylglucosamine to VSV G protein is restored (Figure 14-Sb).
This modification is the consequence of vesicular transport of KEY CONCEPTS of Section 14.1
N-acetylglucosamine transferase I from the wild-rype medial-
Golgi to the cis-Golgi isolated from virally infected mutant Techniques for Studying the Secretory Pathway
cells. Successful intercompartmental transport in this cell-free • All a~~:1ys for following the trafficking of proteins through
system depends on requirements that are typical of a normal the secretory pathway in living cells require a way to label a
physiological process, including a cytosolic extract, a source cohort of secretory proteins and a way to identify the com-
of chemical energy in the form of ATP and GTP, and incuba- partments where labeled proteins subsequently are located.
tion at physiological temperatures. • Pulse labeling with radioactive amino acids can specifically
In addition, under appropriate conditions a uniform popu- label a cohort of newly made proteins in the ER. Alternatively,
lation of the transport vesicles that move N-acerylglucosamine

14.1 Techniques for Studying the Secretory Pathway 633

 (a) Coated vesicle budding
a temperature-sensitive mutant protein that is retained in the

4(& ~~:;;,;~:"
ER due to misfolding at the nonpermissive temperature will
be released as a cohort for transport when cells are shifted to
the permissive temperature.
• Transport of a fluorescently labeled protein along the se-
cretory pathway can be observed by microscopy (see Fig- Soluble ~ Membrane
ure 14-2). Transport of a radiolabeled protein commonly is cargo 0 cargo.
protein........____ protetn

~ ~~ Membrane
trachJ by following compartment-specific covalent modifi-
cations to the protein.
, .}.: ~~ cargo-receptor
• Many of the components required for intracellular pro-
tein trafficking have been identified in yeast by analysis of ~ "' protein

temperature-sensitive sec mutants defective for the secretion Coat proteins

Donor
of proteins at the nonpermissive temperature (see Figure 14-4). membrane Cytosol
• Cell-free assays for intercompartmental protein transport
have allowed the biochemical dissection of individual steps (b) Uncoated vesicle fusion
of the secretory pathway. Such in vitro reactions can be used
to produce pure transport vesicles and to test the biochemi- Cytosol Target
membrane
cal function of individual transport proteins.

/. )L
14.2 Molecular Mechanisms of Vesicle
Budding and Fusion
Small membrane-bounded vesicles that transport proteins
i •).. . ~ ')t-SNARE
pmto;c.

from one organelle to another are common elements in the

secretory and endocytic pathways (see Figure 14-1). These
vesicles bud from the membrane of a particular "parent··
(donor) organelle and fuse with the membrane of a particu-
•-SNARE
protein
Al SNARE
complex

lar "target" (destination) organelle. Although each step in FIGURE 14-6 Overvi ew of vesicle budding and fusion w ith a
the secretory and endocytic pathways employs a different target membrane. (a) Budding is initiated by recruitment of a small
type of vesicle, studies employing genetic and biochemical GTP-binding protein to a patch of donor membrane. Complexes of coat
techniques have revealed that each of the different vesicular proteins in the cytosol then bind to the cytosolic domain of membrane
transport steps is simply a variation on a common theme. In cargo proteins, some of which also act as receptors that bind soluble
proteins in the lumen, thereby recruiting luminal cargo proteins into
this section we explore the basic mechanisms underlying
the budding vesicle. (b) After being released and shedding its coat, a
vesicle budding and fusion that all vesicle types have in com-
vesicle fuses with its target membrane in a process that involves
mon, before discussing the details unique to each pathway.
interaction of cognate SNARE proteins.

Assembly of a Protein Coat Drives Vesicle .·

Formation and Selection of Cargo Molecules fuse (Figure 14-6b). Regardless of target organelle, all trans-
port vesicles use v-SNARES and t-SNARES to bud and fuse.
The budding of vesicles from their parent membrane is driven
Three major types of coated vesicles have been character-
by the polymerization of soluble protein complexes onto the
ized, each with a different type of protein coat and each
membrane to form a proteinaceous vesicle coat (Figure 14-6a).
formed by reversible polymerization of a distinct set of pro-
Interactions between the cytosolic portions of integral mem-
tein subunits (Table 14-1 ). Each type of vesicle, named for its
brane proteins and the vesicle coat gather the appropriate
primary coat proteins, transports cargo proteins from par-
cargo proteins into the forming vesicle. Thus the coat gives
ticular parent organelles to particular destination organelles:
curvature to the membrane to form a vesicle and acts as the
filter to determine which proteins are admitted into the vesicle. • COPII vesicles transport proteins from the ER to the Golgi.
The coat is also respomihle for including in the vesicle
• COPI vesicles mainly transport proteins in the retrograde
fusion proteins known as v-SNAREs. After formation of a
direction between Golgi cisternae and from the cis-Golgi back
vesicle is completed, the coat is shed, exposing the vesicle's
to the ER.
v-SNARE proteins. The specific joining of v-SNAREs in the
vesicle membrane with cognate t-SNAREs in the target • Clathrin vesicles transport proteins from the plasma mem-
membrane to which the vesicle is docked brings the mem- brane (cell surface) and the trans-Golgi network to late
branes into close apposition, allowing the two bilayers to endosomes.

634 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 TABLE 14-1 Coated Vesicles Involved in Protein Trafficking

Vesicle Type Transport Step Mediated Coat Proteins Associated GTPase

COPII ER to os-Golg1 Sec23/Sec24 and Sec 13/Sec3 l Sarl

complexes, Sec 16

COPI c1s-Golgi to ER Coatomers containing seven ARF

Later to earlier Golgi cisternae different COP subumts

Clathrin and adapter proteins'· trans-Golgi to endosome Clathrin + APl complexes ARF

trans-Golgi to endosome Clathrin + GGA ARr

Plasma membrane to endosome Clathrin ;- AP2 complexes ARF

Golgi to lysosome, melanosome, AP3 complexes ARF

or platelet vesicles

• Each trpe of AP complex consiSts of four different subumts. lr 1s nor known whether the coat of AP3 ves1cles contains clathnn.

Every vesicle-mediated trafficking step is thought to utilize membrane. The polymerized coat proteins are thought to
some kind of vesicle coat; however, a specific coat protein form a curved lattice that drives the formation of a vesicle
complex has not been identified for every type of vesicle. For bud by adhering to the cytosolic face of the membrane.
example, vesicles that move proteins from the trans-Golgi to
the plasma membrane during either constitutive or regulated
secretion exhibit a uniform size and morphology suggesting
A Conserved Set of GTPase Switch Proteins
that their formation is driven by assembly of a regular coat Controls Assembly of Different Vesicle Coats
structure, yet researchers have not identified specific coat Based on in vitro vesicle-budding reactions with isolated
proteins surrounding these vesicles. membranes and purified coat proteins, scientists have deter-
The general scheme of vesicle budding shown in Figure mined the minimum set of coat components required to form
14-6a applies to all three known types of coated vesicles. each of the three major types of vesicles. Although most of
Experiments with isolated or artificial membranes and puri- the coat proteins differ considerably from one type of vesicle
fied coat proteins have shown that polymerization of the to another, the coats of all three vesicles contain a small GTP-
coat proteins onto the cytosolic face of the parent membrane binding protein that acts as a regulatory subunit to control
is necessary to produce the high curvature of the membrane coat assembly (see figure 14-6a). For both COPI and clathrin
that is typical of'a transport vesicle about 50 nm in diameter. vesicles, this GTP-binding protein is known as ARF protem.
Electron micrographs of in vitro budding reactions often re- A different but related GTP-binding protein known as Sarl
veal structures that exhibit discrete regions of the parent protem is present in the coat of COPII vesicles. Both ARF and
membrane bearing a dense coat accompanied by the curva- Sarl are monomeric proteins with an overall structure similar
ture characteristic of a completed vesicle (Figure 14-7). Such to that of Ras, a key intracellular signal-transducing protein
structures, usually called vesicle buds, appear to be interme- (see Figure 16-19). ARF and Sarl proteins, like Ras, belong
diates that are visible after the coat has begun to polymerize to the GTPase superfamily of switch proteins that cycle be-
but before the completed vesicle pinches off from the parent tween GDP-bound and GTP-bound forms (see Figure 3-32 to
review the mechanism of GTPase "switch" proteins).
The cycle of GTP binding and hydrolysis by ARF and
Sarl are thought to control the initiation of coat assembly,

>:.PERIMENTAL FIGURE 14·7 Vesicle buds can be visualized

during in vitr o budding reactions. When punfled COPII coat
components are incubated with isolated ER vesicles or artificial
phospholipid vesicles (liposomes), polymerization of the coat proteins
on the vesicle surface induces emergence of highly curved buds. In this
electron micrograph of an in vitro budding reaction, note the distinct
membrane coat, visible as a dark protein layer, present on the vesicle
buds. [From K. Matsuoka et al., 1988, Ce// 93(2):263.]

14.2 Molecular Mechanisms of Vesicle Budding and Fusion 635

D Sar1 membrane binding, as schematically depicted for the assembly of COPII vesicles
GTP exchange in Figure 14-8. First, an ER membrane pro tein known as
Hydrophobic
Secl2 catalyzes release of GOP from cytosolic Sarl ·GDP and
N-terminus
binding of GTP. This guanine nucleotide-exchange factor
Cytosol apparently receives and integrates multiple as-yet-unknown
signals, probably including the presence in the ER membrane
of cargo proteins that are ready to be transported. Binding of
ER lumen GTP causes a conformational change in Sarl that exposes its
hydrophobic N-terminus, which then becomes embedded in
the phospholipid bilayer and tethers Sarl ·GTP to the ER
membrane (Figure 14-8, step 0 ). The membrane-attached
Sarl ·GTP drives polymerization of cytosolic complexes of
COPII subunits on the membrane, eventually leading to for-
mation of vesicle buds (step f)). Once COPII vesicles are re-
leased from the donor membrane, the Sarl GTPase activity
COPII coat hydrolyzes Sarl·GTP in the vesicle membrane to Sarl ·GOP
El assembly
with the assistance of one of the coat subunits (step IJ). This
hydrolysis triggers disassembly of the COPll coat (step 19).
Thus Sarl couples a cycle of GTP binding and hydrolysis to
the formation and then dissociation of the COPII coat.
ARF protein undergoes a similar cycle of nucleotide ex-
change and hydrolysis coupled to the assembly of vesicle coats
composed either of COPI or of clathrin and other coat pro-
teins (AP complexes), discussed later. A covaient protein mod-
ification known as a myristate anchor on theN-terminus of
IJ GTP hydrolysis ARF protein weakly tethers ARF·GDP to the Golgi mem-
brane. When GTP is exchanged for the bound GOP by a nu-
cleotide-exchange factor attached to the Golgi membrane, the
resulting conformational change in ARF allows hydrophobic
residues in its N-terminal segment to insert into the membrane
bilayer. The resulting tight association of ARF·GTP with the
membrane serves as the foundation for further coat assembly.
Drawing on the structural similarities of Sarl and ARF
to other small GTPase switch proteins, researchers have con-
structed genes encoding mutant versions of the two proteins
II Coat disassembly that have predictable effects on vesicular traffic when trans-
fected into cultured cel ls. For example, in cells expressing
mutant versions of Sarl or ARF that cannot hydrolyze GTP,
vesicle coats form and vesicle buds pinch off. Howeve r, be-
Uncoated vesicle
cause the mutant proteins cannot trigger d isassembly of the
FIGURE 14-8 Model for the role of Sar1 in the assembly and coat, all available coat subunits eventually become perma-
disassembly of COPII coats. Step 0 : Interaction of soluble GOP- nently assembled into coated vesicles that arc unable to fuse
bound Sarl with the exchange factor Sec12, an ER integral membrane with target membranes. Addition of a nonhydrolyzable GTP
protein, catalyzes exchange of GTP for GOP on Sarl.ln the GTP-bound
analog to in vitro vesicle-budding reactions causes a similar
form of Sarl, its hydrophobic N-terminus extends outward from the
blocking of coat disassembly. The vesicles that form in such
protein's surface and anchors Sarl to the ER membrane. Step D: Sarl
reactions have coats that never dissociate, allowing their
attached to the membrane serves as a binding site for the Sec23/Sec24
coat protein complex. Membrane cargo proteins are recruited to the
composition and structure to be more readily ana lyzed. The
forming vesicle bud by binding of specific short sequences (sorting purified COPI vesicles shown in Figure 14-9 were produced
signals) in their cytosolic regions to sites on the Sec23/Sec24 complex. in such a budding reaction.
Some membrane cargo proteins also act as receptors that bind soluble
proteins in the lumen. The coat is completed by assembly of a second
Targeting Sequences on Cargo Proteins Make
type of coat complex composed of Sec13 and Sec31 (not shown).
Step ID: After the vesicle coat is complete, the Sec23 coat subunit
Specific Molecular Contacts with Coat Proteins
promotes GTP hydrolysis by Sarl. Step B : Release of Sarl ·GOP from In order for transport vesicles to move specific proteins from
the vesicle membrane causes disassembly of the coat. [Sees. Springer one compartment to the next, vesicle buds must be able to
etal., 1999,Ce//97:145.] discriminate among potential membrane and soluble cargo
proteins, accepting only those cargo proteins that should

636 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 E•. PERl J!fiENTAL FIGUR 14.. 9 Coated vesicles accumulate
during in vitro budding reacti ons in the presence of a nonhydro-
lyzable analog of GTP. When isolated Golgi membranes are incubated
with a cytosolic extract containing COP I coat proteins, vesicles form
and bud off from the membranes. Inclusion of a nonhydrolyzable
analog of GTP in the budding reaction prevents disassembly of the
coat after vesicle release. This micrograph shows COP I vesicles
generated in such a reaction and separated from membranes by
centrifugation. Coated vesicles prepared in this way can be analyzed
to determine their components and properties. [Courtesy of L Orci.]

proteins (see Figure 14-6a ). The polymerized coat thus acts as

an affinity matrix to cluster selected membrane cargo proteins
into forming vesicle buds. Since soluble proteins within the
60 nm
lumen of parent organelles cannot contact the coat directly,
they require a different kind of sorting signal. Soluble luminal
advance to the next compartment and excluding those that proteins often contain what can be thought of as luminal sort-
should remain as residents in the donor compartment. In ad- ing signals, which bind to the luminal domains of certain
dition to sculpting the curvature of a donor membrane, the membrane cargo proteins that act as receptors for luminal
vesicle coat functions in selecting specific proteins as cargo. cargo proteins. The properties of several known sorting sig-
The primary mechanism by which the vesicle coat selects nals in membrane and soluble proteins are summarized in
cargo molecules is by directly binding to specific sequences, or Table 14-2. We describe the role of these signals in more de-
sort ing sign als, in the cytosolic portion of membrane cargo tail in later sections.

TABLE 14-2 Known Sorting Signals That Direct Proteins to Specific Transport Vesicles

Vesicles That Incorporate

Signal Sequence* Proteins with Signal Signal Receptor Signal- Bearing Protein

LUMINAL SORTING SIGNALS

-- -- ------------ ---------
Lys-Asp-Glu-Leu (KDEL) ER-residem soluble proteins KDEL receptor in c1s-Golgi COPI
membrane

Mannose 6-phosphate (M6P) Soluble lysosomal enzymes M6P receptor in trans-Golgi Clathrin/APl
after processing in cis-Golgi membrane

Secreted lysosomal enzymes M6P receptor in plasma membrane Clathrin/AP2

CYTOPLASMIC SORTING SIGNALS

- - - - - - -- - - - - -- -------
Lys-Lys-X-X (KKXX) ER-residem membrane proteins COPI a and f3 subunits COPI

Di-arginine (X-Arg-Arg-X) ER-residem membrane proteins COPI a and f3 subunits COPI

Di-acidic (e.g., Asp-X-Glu) Cargo membrane proteins in ER COPII Sec24 subunit COPII

Asn-Pro-X-Tyr (NPXY) LDL receptor in plasma AP2 complex Clathrin/AP2

membrane

Tyr-X-X-<1> (YXX<l>) Membrane proteins in APl (f.Ll subunit) Clathrin/APl

trans-Golgi

Plasma membrane proteins AP2 (f.L2 subunit) Clathrin/AP2

Leu-Leu (LL) Plasma membrane proteins AP2 complexes Clathrin/AP2

•x = any ammo acid; <P hydrophobic ammo acid. Singlc-letrer amino acid abbreviarions are m parenrheses.

14.2 Molecular Mechanisms of Vesicle Budding and Fusion 637

(a) Transport (b) SNARE complex
vesicle

Ves icle docki ng lD

:~:..~
FIGURE 14-10 Model for docking and fusion of transport
vesicles with their target membranes. (a} The proteins shown in this
example participate in fusion of secretory vesicles with the plasma
membrane, but similar proteins mediate all vesicle-fusion events.
Step 0 : ARab protein tethered via a lipid anchor to a secretory vesicle
/ Syntaxin binds to an effector protein complex on the plasma membrane,
__.. SNAP-25 thereby docking the transport vesicle on the appropriate target
membrane. Step f) : A v-SNARE p rotein (in this case, VAMP} interacts
with the cytosolic domains of the cognate t-SNAREs (in this case,
Target~ · syntax in and SNAP-25}. The very stable (oiled-coil SNARE complexes
membrane
A sse mbly o f
SNARE c omplexes
1fJ Rab effector that are formed hold the vesicle close to t he target membrane.
Step ID: Fusion of the two membranes immediately follows formation
of SNARE complexes, but precisely how this occurs is not known.
Step r::l: Following membrane fusion, NSF in conjunction with o:-SNAP
protein binds to the SNARE complexes. The NSF-catalyzed hydrolysis of
ATP then drives dissociation of the SNARE complexes, freeing the
SNARE proteins for another round of vesicle fusion.'Aiso at this time,
Rab·GTP is hydrolyzed to Rab·GDP and dissociates from the Rab
effector (not shown}. (b) The SNARE complex. Numerous noncovalent
interactions between four long o: helices, two from SNAP-25 and one
each from syntaxin and VAMP, stabilize the coiled-coil structure. [See
Mem bra n e fus i o n 111 J. E. Rothman and T. Sbllner, 1997, 5cience 276:1212, and W. Weis and R Scheller,
1998, Nature 395:328. Part (b) from Y. A. Chen and R H. Scheller, 2001 , Nat. Rev.
Mol. Cell Bioi. 2(2):98.]
NSF

change in Rab that enables it to interact with a surface protein

on a particular transport vesicle and insert its isoprenoid
cis-SNARE/
complex
anchor into the vesicle membrane. Once Rab·GTP is tethered
ATP to the vesicle surface, it is thought to interact with one of a
ADP + p. number of d ifferent large proteins, known as Rab effectors, ·.

f.
Di sasse mbly of ' attached to the target membrane. Binding of Rab·GTP to a
SNARE co mpl exes IJ Rab effector docks the vesicle on an appropriate target mem-
brane (Figure 14-10, step 0 ). After vesicle fusion occurs, the
GTP bound to the Rab protein is hydrolyzed to GDP, trigger-
ing the release of Rab ·GDP, which then can undergo another
cycle of GDP-GTP exchange, binding, and hydrolysis.
Several lines of evidence support the involvement of spe-
cific Rab proteins in vesicle-fusion events. For instance, the
yeast SEC4 gene encodes a Rab protein, and yeast cells ex-
Rab GTPases Control Docking of Vesicles
pressing mutant Sec4 proteins accumulate secretory vesicles
on Target Membranes that are unable to fuse with the plasma membrane (class E
A second set of small GTP-binding proteins, known as Rab mutants in Figure 14-4 ). In mammalian cells, Rab5 protein
proteins, participate in the targeting of vesicles to the appro- i~ lut:alized to endocytic vesicles, also known as early endo-
priate target membrane. Like Sarl and ARF, Rab proteins somes. These uncoated vesicles form from clathrin-coated
belong to the GTPase superfamily of switch proteins. Rab vesicles just after they bud from the plasma membrane dur-
proteins also contain an isoprenoid anchor that allows them ing endocytosis (see Figure 14-1, Step m). The fusion of early
to become tethered to the vesicle membrane. Conversion of endosomes with each other in cell-free systems requires the
cytosolic Rab·GDP to Rab·GTP, catalyzed by a specific gua- presence of Rab5, and addition of Rab5 and GTP to cell-free
nine nucleotide-exchange factor, induces a conformational extracts accelerates the rate at which these vesicles fuse with

638 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 each other. A long coiled protein known as EEA 1 (early proteins. The likely explanation for this difference is that in
endosome antigen 1 ), which resides on the membrane of the the cell, other proteins such as Rab protems and their effectors
early endosome, functions as the effector for Rab5. In this are involved in targeting vesicles to the correct membrane.
case, Rab5·GTP on one endocytic vesicle is thought to spe- Ycast cells, like all eukaryotic cells, express more than 20
cifically bind to EEAl on the membrane of another endo- different related v-SNARE and t-SNARE proteins. Analyses
cytic vesicle, setting the stage for fusion of the two vesicles. of yeast mutants defective in each of the SNARE genes have
A different type of Rab effector appears to function for identified specific membrane-fusion events in which each
each vesicle type and at each step of the secretory pathway. SNARE protein participates. For all fusion events that have
Many questions remain about how R::1h proteins are tar- been examined, the SNARE!> form four-helix bundled com-
geted to the correct membrane and how specific complexes plexes, similar to the VAMP/syntaxin/SNAP-25 complexes
form between the different Rab proteins and their corre- that mediate fusion of secretory vesicles with the plasma
sponding effector proteins. membrane. However, in other fusion events (e.g., fusion of
COPII vesicle~ with the cis-Golgi network), each participat-
ing SNARE protein contributes only one a helix to the bun-
dle (unlike SNAP-25, which contributes two helices); in
Paired Sets of SNARE Proteins Mediate Fusion
these cases the SNARE complexes comprise one v-SNARE
of Vesicles with Target Membranes and three t-SNARE molecules.
As noted previously, shortly after a vesicle buds off from the Using the in vitro liposome fusion assay, researchers
donor membrane, the vesicle coat disassembles to uncover a have tested the ability of various combinations of individual
vesicle-specific membrane protein, a v-SNARE (see Figure v-SNARE and t-SNARE proteins to mediate fusion of donor
14-6b). Likewise, each type of target membrane in a cell con- and target membranes. Of the very large number of different
tains t-SNARE membrane proteins, which specifically inter- combinations tested, only a small number could efficiently
act with v-SNAREs. After Rab-mediated docking of a vesicle mediate membrane fusion. To a remarkable degree, the func-
on its target (destination) membrane, the interaction of cog- tional combinations of v-SNAREs and t-SNAREs revealed m
nate SNAREs brings the two membranes close enough to- these in vitro experiments correspond to t~e actual SNARE
gether that they can fuse. protein interactions that mediate known membrane-fusion
One of the best-understood examples of SNARE-mediated events in the yeast cell. Thus, together with the specificity of
fusion occurs during cxocytosis of secreted proteins (see Fig- interaction between Rab and Rab effector proteins, the spec-
ure 14-10, steps fJ and 10). In this case, the v-SNARE, known ificity of the interaction between SNARE proteins can ac-
as VAMP (vesicle-associated membrane protein), is incorpo- count for most, if not all, of the specificity of fusion between
rated into secretory vesicles as they bud from the trans-Golgi a particular vesicle type and its target membrane.
network. The t-SNAREs are syntaxin, an integral membrane
protein in the plasma membrane, and SNAP-25, which is
attached to the plasma membrane by a hydrophobic lipid Dissociation of SNARE Complexes
anchor in the middle of the protein. The cytosolic region in
After Membrane Fusion Is Driven
each of these three SNARE proteins contains a repeating
heptad sequence'that allows four a helices-one from VAMP, by ATP Hydrolysis
one from syntaxin, and two from SNAP-25-to coil around After a vesicle and its target membrane have fused, the
one another to form a four-helix bundle (Figure 14-l Ob). The SNARE complexes must dissociate to make the individual
.' unusual stability of this bundled SNARE complex is con- SNARE proteins available for additional fusion events. Be-
ferred by the arrangement of hydrophobic and charged amino cause of the stability of SNARE complexes, which arc held
residues in the heptad repeats. The hydrophobic amino acids together by numerous noncovalent intermolecular interac-
are buried in the central core of the bundle, and amino acids tions, their dissociation depends on additional proteins and
of opposite charge are aligned to form favorable electrostatic the input of energy.
interactions between helices. As the four-helix bundles form, The first clue that dissociation of SNARE complexes re-
the vesicle and target membranes are drawn into close ap- quired the assistance of other proteins came from in vitro
position by the embedded transmembrane domains of VAMP transport reactions depleted of certain cytosolic proteins.
and syntaxin. The observed accumulation of vesicles in these reactions in-
In vitro experiments have shown that when liposomes dicated that vesicles could form but were unable to fuse with
containing purified VAMP are incubated with other lipo- a target membrane. Eventually two proteins, designated NSF
somes containing syntaxin and SNAP-25, the two classes of and a-SNAP, were found to be required for ongoing vesicle
membranes fuse, albeit slowly. This finding is strong evidence fusion in the in vitro transport reaction. The function of NSF
that the close apposition of membranes resulting from forma- in vivo can be blocked selectively by N-ethylmaleimide
tion of SNARE complexes is sufficient to bring about mem- (NEM), a chemical that reacts with an essential -SH group
brane fusion. Fusion of a vesicle and target membrane occurs on NSF (hence the name, NEM-sensitive factor).
more rapidly and efficiently in the cell than it does in liposome Yeast mutants have also contributed to our understand-
experiments in which fusion is catalyzed only by SNARE ing of SNARE function. Among the class C yeast sec mutants

14.2 Molecular Mechanisms of Vesicle Budding and Fusion 639

are strains that lack functional Sec18 or Sec17, the yeast
counterparts of mammalian NSF and o:-SNAP, respectively. membrane, inducing fusion of the two membranes. After fu-
When these class C mutants are placed at the nonpermissive sion is completed, t he SNARE complex is disassembled in an
temperature, they accumulate ER-to-Golgi transport vesicles; ATP-dependent reaction med iated by other cytosolic pro-
when the cells arc shifted to the lower, permissive temperature, teins (see Figure 14-10).
the accumulated vesicles are able to fuse with the cis-Golgi.
Subsequent to the initial biochemical and genetic studies
identifying NSF and o:-SNAP, more sophisticated in vitro
transport assays were developed. Using these newer a:.:.ay~, 14.3 Early Stages of the Secretory
researchers have shown that NSF and o:-SNAP proteins are
not necessary for actual membrane fusion but rather are re- Pathway
quired for regeneration of free SNARE proteins. NSF, a hex- In this section we take a closer look at vesicular traffic be-
amer of identical subunits, associates with a SNARE complex tween the ER and the Golgi and some of the evidence support-
with the aid of o:-SNAP (soluble NSf attachment protein). ing the general mechanisms discussed in the previous section.
The bound NSF then hydrolyzes ATP, releasing sufficient Recall that anterograde transport from the ER to Golgi, the
energy to dissociate the SNARE complex (Figure 14-10, first vesicle trafficking step in the secretory pathway, is medi-
step a ). Evidently, the defects in vesicle fusion observed in ated by COPII vesicles. These vesides contain newly synthe-
the earlier in vitro fusion assays and in the yeast mutants sized proteins destined for the Golgi, cell surface, or lysosomes
after a loss of Sec17 or Sec18 were a consequence of free as well as vesicle components such as v-SNAREs that are re-
SNARE proteins rapidly becoming sequestered in undissoci- quired to target vesicles to the cis-Golgi membrane. Proper
ated SNARE complexes and thus being unavailable to medi- sorting of proteins between the ER and Golgi also requires
ate membrane fusion . retrograde (reverse) transport from the cis-Golgi to the ER,
which is mediated by COPI vesicles (Figure ,14-11 ). This ret-
rograde vesicle transport serves to retrieve v-SNARE proteins
and the membrane itself back to the ER to provide the neces-
KEY CONCEPTS o~ Section , 4.2 sary material for additional rounds of vesicle budding from
the ER. COPf-mediated retrograde transport also retrieves
Molecular Mechanisms of Vesicle Budding and Fusion missorted ER-resident proteins from the cis-Golgi to correct
The three well-characterized transport vesicles-COPI, sorting mistakes.
COPII, and clathrin vesicles-are distinguished by the pro- We also discuss in this section the process by which pro-
teins that form their coats and the transport routes they me- teins that have been correctly delivered to the Golgi advance
diate (see Table 14-1). through successive compartments of the Golgi, from cis- to
trans-network. This process, known as cisternal maturation,
All types of coated vesicles are formed by polymerization involves budding and fusion of retrograde rather than an-
of cytosolic coat proteins onto a donor (parent) membrane terograde transport vesicles.
to form vesicle buds that eventually pinch off from the mem-
brane to release a complete vesicle. Shortly after vesicle re-
lease, the coat is shed, exposing proteins required for fusion
COPII Vesicles Mediate Transport
with the target membrane (see Figure 14-6).
from the ER to the Golgi
Small GTP-binding proteins (ARF or SarI) belonging to
the GTPase superfamily control polymerization of coat pro- COPII vesicles were first recognized when cell-free extracts
teins, the initial step in vesicle budding (sec Figure 14-8). of yeast rough ER membranes were incubated with cytosol
After vesicles are released from the donor membrane, hydro- and a nonbydrolyzable analog of GTP. The vesicles that
lysis of GTP bound to ARF or Sarl triggers disassembly of formed from the ER membranes had a distinct coat similar
the vesicle coats. to that on COPI vesicles but composed of different proteins,
designated COPII proteins. Yeast cells with mutations in the
• Specific sorting signals in membrane and luminal proteins
genes for COPII proteins are class B sec mutants and accu-
of donor organelles interact with coat proteins during vesicle
mulate proteins in the roughER (see Figure 14-4). Ana lysis
budding, thereby recruiting cargo proteins to vesicles (see of such mutants has revealed a set of proteins required for
Table 14-2).
formation of COPII vesicles, including the proteins which
A second set of GTP-binding proteins, the Rab proteins, comprise the COPII vesicle coat.
regulate docking of vesicles with the correct target mem- As described previously, formation of COPII vesicles is
brane. Each Rab appears to bind to a specific Rab effector triggered when Sec12, a guanine nucleotide-exchange factor
associated with the target membrane. in the ER membrane, catalyzes the exchange of bound GDP
Each v-SNARE in a vesicular membrane specifically binds for GTP on cytosolic Sarl . This exchange induces binding of
to a complex of cognate t-SNARE proteins in the target Sar 1 to the ER membrane followed by binding of a complex
of Sec23 and Scc24 proteins (see Figure 14-8). The resulting

640 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 cis-

l'tr 1
Golgi Sec23
network

¥1 J
\ ' -.....11
fJ ' \II
~~\
Transmembrane
segment
COPII
vesicle \l. )j
~-;:,.,""Coat
COPI
vesicle
of cargo protein

FIGURE 14- 12 Three-d imensional structure of the ternary

protein
complex comprising the COP II coat proteins Sec23 and Sec24 and
Sar1·GTP. Early in the formation of the COPII coat, Sec23 (orange)/
Sec24 (green) complexes are recruited to the ER membrane by Sarl
(red) in its GTP-bound state. In order to form a stable ternary complex
in solution for structural studies, the nonhydrolyzable GTP analog
GppNHp is used. A cargo protein in the ER membrane can be recruited
to COP II vesicles by interaction of a tripeptide di-acidic signal (purple)
in the cargo's cytosolic domain with Sec24. The likely position of the
Rough
ER

\
• -Soluble
cargo

~
._, Membrane
COPII vesicle membrane and the transmembrane segment of the cargo
protein are indicated. TheN-terminal segment of Sar1 that tethers it to
the membrane is not shown. [See X. Bi et al., 2002, Nature419:271;
interaction with peptide courtesy of J. Goldberg.]
Membrane receptor
cargo
FIGURE 14-11 Vesicle-mediated protein t rafficking b etween the
ER and cis-Golgi. Steps O-il Forward (anterograde) transport is
mediated by COPII vesicles, which are formed by polymerization of protein called Sec 16, which is hound to the cytosolic surface
soluble COPII coat protein complexes (green) on the ER membrane. of the ER, interacts with Sarl·GTP and the Sec13/31 and
v-SNAREs (orange) and other cargo proteins (blue) in the ER membrane Sec23/24 complexes and acts to organize the other coat pro-
are incorporated into the vesicle by interacting with coat proteins. teins, increasing the efficiency of coat polymerization. Simi-
Soluble cargo proteins (magenta) are recruited by binding to appropri- lar to Sec 13/31, clathrin also has this ability to self-assemble
ate receptors in the membrane of budding vesicles. Dissociation of the into a coatlike structure, as will he discussed in Section 14.4.
coat recycles free coat complexes and exposes v-SNARE proteins on Certain integral ER membrane proteins are specifically
the vesicle surface. After the uncoated vesicle becomes tethered to the recruited into COPII vesicles for transport to the Golgi. The
cis-Golgi membrane in a Rab-mediated process, pairing between the cytosolic segments of many of these proteins contain a di-
exposed v-SNAREs and cognate t-SNAREs in the Golgi membrane acidic sorting signal (the key residues in this sequence are
allows vesicle fusion, releasing the contents into the cis-Golgi compart- Asp-X-Giu, or DXE in the one-letter code) (see Table 14-2).
ment (see Figure 14-1 0). Steps D-lil: Reverse (retrograde) transport,
This sorting signal binds to the Sec24 subunit of the COPII
mediated by vesicles coated with CO PI proteins (purple). recycles the
coat and is essential for the selective export of certain mem-
membrane bilayer and certain proteins, such as v-SNAREs and
missorted ER-resident proteins (not shown), from the cis-Golgi to the
brane proteins from the ER (see Figure 14-12). Biochemical
ER. All SNARE proteins are shown in orange although v-SNAREs and
and genetic studies currently are under way to identify ad-
t-SNAREs are distinct proteins. ditional signals that help direct membrane cargo proteins
into COPII vesicles. Other ongoing studies seek to determine
how soluble cargo proteins are selectively loaded into COPII
vesicles. Although purified COPII vesicles from yeast cells
ternary complex formed between Sarl·GTP, Sec23, and have been found to contain a membrane protein that binds
Sec24 is shown m hgure 14-12. After this complex forms on the soluble a mating factor, the receptors for other solu ble
the ER membrane, a second complex comprising Sec13 and cargo proteins such as invertase are not yet known.
Sec31 proteins binds to complete the coat structure. Pure
Sec 13 and Sec31 proteins can spontaneously assemble into .. The inherited disease cystic fibrosis is characterized by
cagelike lattices. It is thought that Secl3 and Sec3l form the an imbalance in chloride and sodium ion transport in
structural scaffold for COPII vesicles. Finally, a large fibrous the epithelial cells of the lung, leading to fluid build-up and

14.3 Early Stages of the Secretory Pathway 641

difficulty breathing. Cystic fibrosis is caused by mutations in suggested that COPI vesicles mediate ER-to-Golgi transport,
a protein known as CFTR, which is synthesized as an integral subsequent experiments showed that their main function is
membrane protein in the ER and is transported to the Golgi retrograde transport, both between Golgi cisternae and from
before being transported to the plasma membranes of epithe- the cis-Golgi to the rough ER (see Figure 14-11, right). Be-
lial cells, where it functions as a chloride channel. Researchers cause COPI mutants cannot recycle key membrane proteins .·
have recently shown that the CFTR protein contains a di- back to the rough ER, the ER gradually becomes depleted of
acidic sorting signal that binds to the Sec24 subunit of the ER proteins such as v-SNAREs necessary for COPil vesicle
COPII coat and is necessary for transport of the CFTR protein function. Eventually, vesicle formation from the rough ER
out of the ER. The most common CFTR mutation is a dele- grinds to a halt; secretory proteins continue to be synthe-
tion of a phenylalanine at position 508 in the protein se- sized but accumulate in the ER, the defining characteristic of
quence (known as ~F508). This mutation prevents normal class B sec mutants. The general ability of sec mutants in-
transport of CFTR to the plasma membrane by blocking its volved in either COPI or COPII vesicle function to eventu-
packaging into COPII vesicles budding from the ER. Al- ally block both anterograde and retrograde transport
though the ~F508 mutation is not in the vicinity of the di- illustrates the fundamental interdependence of these two
acidic sorting signal, this mutation may change the transport processes.
conformation of the cytosolic portion of CFTR so that the As discussed in Chapter 13, the ER contains several sol-
di-acidic signal is unable to bind to Sec24. Interestingly, a uble proteins dedicated to the folding and modification of
folded CFTR with this mutation would still likely function newly synthesized secretory proteins. These include the
properly as a normal chloride channel. However, it never chaperone BiP and the enzyme protein disulfide isomerase,
reaches the membrane; the disease state is therefore caused by which are necessary for the ER to carry out its functions. .,
the absence of the channel rather than by a defective one. • Although such ER-resident luminal proteins are not specifi-
cally selected by COPII vesicles, their sheer abundance
The experiments described previously in which the transit causes them to be continuously loaded passively into vesicles
of VSVG-GFP in cultured mammalian cells is followed by destined for the cis-Golgi. The transport of these soluble
fluorescence microscopy (sec Figure 14-2) provided insight proteins back to the ER, mediated by COPl vesicles, pre-
into the intermediates in ER-to-Golgi transport. In some cells, vents their eventual depletion.
small fluorescent vesicles containing VSVG-GFP could be seen Most soluble ER-resident proteins carry a Lys-Asp-Glu-
to form from the ER, move less than 1 J-lm, and then fuse di- Leu (KDEL in the one-letter code) sequence at their C-terminus
rectly w1th the cis-Golgi. In other cells, in which the ER was (see Table 14-2). Several experiments demonstrated that this
located several micrometers from the Golgi complex, several KDI:L sorting signal is both necessary and sufficient to cause
ER-dcrived vesicles were seen to fuse with each other shortly a protein bearing this sequence to be located in the ER. For
after their formation, forming what is termed the ER-to-Golgt instance, when a mutant protein disulfide isomerase lacking
mtermedtate compartment or the cts-Golgi network. These these four residues is synthesized in cultured fibroblasts, the
larger structures then were transported along microtubules to protein is secreted. Moreover, if a protein that normally is
the cis-Golgi, mt!Ch in the way vesicles in nerve cells are trans- secreted is altered so that it contains the KDEL signal at its
ported from the cell body, where they are formed, down the C-terminus, the protein is located in the ER. The KDEL sorting
long axon to the axon terminus (Chapter 18). Microtubules signal is recognized and bound by the KDEL receptor, a trans-
function much as "railroad tracks," enabling these large ag- membrane protein found primarily on small transport vesicles
gregates of transport vesicles to move long distances to their shuttling between the ER and the cis-Golgi and on the cis-
cis-Golgi destination. At the time the ER-to-Golgi intermedi- Golgi reticulum. In addition, soluble ER-resident proteins that
ate compartment is formed, some COPT vesicles bud off from carry the KDEL signal have oligosaccharide chains with modi-
it, recycling some proteins back to the ER. fications that are catalyzed by enzymes found only in the cis-
Golgi or cis-Golgi network; thus at some time these proteins
must have left the ER and been transported at least as far as the
cis-Golgi network. These findings indicate that the KDEL re-
COPI Vesicles Mediate Retrograde Transport
ceptor acts mainly to retrieve soluble proteins containing the
Within the Golgi and from the Golgi to the ER KDEL sorting signal that have escaped to the cis-Golgi net-
COPI vesicles were first discovered when isolated Golgi frac- work and return them to the ER (figure 14-13 ). The KDEL
tions were incubated in a solution containing cytosol and a receptor binds more tightly to its ligand ar low pH, and it is
nonhydrolyzable analog of GTP (see Figure 14-9). Subse- thought that the receptor is able to bind KDEL peptides in the
quent analysis of these vesicles showed that the coat is cis-Golgi but to release th('o;e peptides in the ER because the pH
formed from large cytosolic complexes, called coatomers, of the Golgi is slightly lower than that of the ER.
composed of seven polypeptide subunits. Yeast cells contain- The KDEL receptor and other membrane proteins that
ing temperature-sensitive mutations in COPI protems accu- are transported back to the ER from the Golgi contain a Lys-
mulate proteins in the rough ER at the nonpermissive Lys-X-X sequence at the very end of their C-terminal seg-
temperature and thus are categorized as class B sec mutants ment, which faces the cytosol (see Table 14-2). This KKXX
(see Figure 14-4 ). Although discovery of these mutants initially sorting signal, which binds to a complex of the COPl a and

642 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

0 VIDEO: KDEL Receptor Trafficking

FIGURE 14- 13 Role of the KDEL receptor in retrieval of

Higher H+ ~ ---.._ ER-resident luminal proteins from t he Golgi. ER luminal proteins,
concentration r Missorted
(lower pH).
peptide binding
ER-resident
protein
v- KDEL peptide especially those present at high levels, can be passively incorporated
into COP II vesicles and transported to the Golgi (steps 0 and f)). Many

cis-Golgi ~ d?( 4rt..

such proteins bear a (-termina l KDEL (Lys-Asp-Giu-Leu) sequence (red)
that allows them to be retrieved. The KDEL receptor, located mainly in
network
-._. J "' \'

~~~..,;;;;.~
.;;;.~'
the cis-Golgi network and in both COP II and COP I vesiclE>~. binds
A
Y r KDEL
receptor
proteins bearing the KDEL sorting signal and returns them to the ER
(steps 0 and 9 ). This retrieval system prevents depletion of ER luminal
ER-to-Golgi~~ COPI coat \ 11 proteins such as those needed for proper folding of newly made
secretory proteins. The binding affinity of the KDEL receptor is very
~:~~ieort~~-r~
·:
·· ~
~~, -!- sensitive to pH. The small difference between the pH of the ER and that

Retrieval ,.~4rt.. of the Golgi favors binding of KDEL-bearing proteins to the receptor in
Golgi-derived vesicles and their release in the ER. [Adapted from
ofKDEL- \'

~~~{~~~s ~\. ~J
J. Semenza et al., 1990, Ce// 61:1349.]
\ L
II \ toER ~--~

COPII '1(~, j Q
This basic question concerning correct membrane parti-
Rough ~ c;~;....;:'#____,(s._J tioning has recently been answered for COPil vesicles. After
these vesicles form, the CO P![ coat proteins remain a~sem­
ER
Lower H+
concentration
t...,. bled long enough fo r the Sec23/Sec24 complex to interact
(higher pH). with a specific tethering factor attached to the c1s-Golgi mem-
peptide release brane. Vesicle uncoating to expose the v-SNAREs is com-
pleted only after the COPII vesicle is already closel} associated
with the cis-Golgi membrane and the COPII v-SNAREs are
in position to form complexes with their cognate t-S~AREs.
Al though COPII vesicles also carry COPl-spccific v-SNARE
proteins, which are being recycled back to the cis-Golgi, these
COPI v-SNARE proteins included in COPII vesicles never
13 subunits (two of th e seven polypeptide subunits in the have the opportunity to form SNARE complexes with cog-
COPI coatomer), is both necessary and sufficient to incorpo- nate ER-localized t-SNARE proteins.
rate membrane p r oteins into COPI vesicles for retrograde
transport to th e ER. Temperature-sensitive yeast mutants
Anterograde Transport Through the Golgi
lack ing COPia or COPI!3 not only are unable to bind the
KKXX signal but also are unable to retrieve proteins bearing Occurs by Cisternal Maturation
this signal back to the ER, indicating that COPI vesicles me- The Golgi complex is organized into three or four subcom-
diate retrograde Golgi-to-ER transport. partments, often arranged in a stacked set of flattened sacs,
A second sorting signal that will target proteins to COPI called cisternae. The subcompartments of the Golgi differ
vesicles and thus will enable recycling from the Golgi to the from one another accord ing to th e enzymes they contain.
ER is a d i-arginine seq uence. Unl ike the KKXX sorting sig- Many of the enzymes are g lycosidases and glycosyltransfer-
na l, which must be located at the cytoplasmically oriented ases that are involved in modifying N-linked or 0-linked car-
C-terminus of a protein, the di-arginine sorting signal can bohydrates attached to secretory proteins as they transit the
reside in any segment of a membrane protein that is on the Golgi stack. On the whole, the Golgi complex operates much
cytoplasmic face of the membrane. like an assemb ly line, with proteins moving in sequence
The partitioning of proteins between the ER and Golgi through the Golgi stack, the modified carbohydrate chains in
complex is a highl y dynamic process depending on both o ne compartment serving as the substrates for the modifying
COPII (anterograde) and COPI (retrograde) vesicles, with enzymes of the next compartment (see Figure 14-14 for a rep-
each type of vesicle responsible for recycling the components resentative sequence of modific~rion steps).
necessary for the function of the other type of vesicle. T he For many years it was thought that the Golgi complex was
organizatio n of this partitioning process raises an interesting an essentiall y static set of compartments with small transport
puzzle: how do vesicles preferentially use the v-SNAREs that vesicles carrying secretory p roteins forward, from the cis- to
will specify fusion w ith the correct target membrane instead the mediai-Golgi and from the medial- to the trans-Golgi. In-
of the v-SNAREs that are being recycled and wou ld have deed, electron microscopy reveals many small vesicles associ-
specificity for fusion with the donor membrane? ated with the Golgi complex that appear to move proteins

14.3 Early Stage s of the Secretory Pathway 643

FIGURE 14- 14 Processing of N-linked oligosaccharide chains on
glycoproteins w ithin cis-, medial-, and trans-Golgi cisternae in
Exit
vertebrate cells. The enzymes catalyzing each step are localized to the
---+
indicated compartments. After removal of three man nose residues in
the cis-Golgi (step D l. the protein moves by cisternal maturation to the
mediai-Golgi. Here, three N-acetylglucosamine (GicNAc) residues are
added (steps f) and r:J), two more man nose residues are removed
(step IJ), and a single fucose is added (step l'll. Processing is
completed in the trans-Golgi by addition of three galactose residues
(step ml and finally by linkage of an N-acetylneuraminic acid residue to
each of the galactose residues (step fJ). Specific transferase enzymes
add sugars to the oligosaccharide, one at a time, from sugar nucleotide
precursors imported from the cytosol. This pathway represents the
Golgi processing events for a typical mammalian glycoprotein.
Variations in the structure of N-linked oligosaccharides can result
from differences in processing steps in the Golgi. [SeeR. Kornfeld and
5. Kornfeld, 1985, Ann. Rev. Biochem. 45:631.)

fro m one Golgi compartment to another (Figure 14-15 ).

I lowever, these vesicles are now known to mediate retro-
grade transport, retrieving ER or Golgi enzymes from a later
compartment and tran~porting them to an earlier compart- Man 5 (GicNAc) 2
ment in th e secretory pathway. Thus the Golgi appears to Man8 (GicNAch ..
have a highly dynamic organization, continually forming
transport vesicles, though only in the retrograde direction.
Golgi j
To see the effect this retrograde transport has on the organi- I{! Transport vesicle
from ER
zation of the Golgi, consider the net effect on the medial-
Golgi compartment as enzymes from the trans-Golgi move • = N-Acetylglucosamine
• = Mannose •= Galactose
to the media/-Golgi while enzymes from the medial-Golgi
.a. = Fucose • = N-Acetylneuraminic acid

~ VIDEO: 3-D Model of a Golgi Complex

. XPERIMENT AL FIGURE 14-15 Electron
micrograph of t he Golgi complex in an exocrine
.. ·~

pan creatic cell reveals secretory and retrograde f.fi;:-;:-;:~~~~~.:.._,_L Forming secretory
tra nsport vesicles. A large secretory vesicle can vesicle
be seen form~ng from the trans-Golgi network.
Elements of the rough ER are on the bottom and
left in this micrograph. Adjacent to the rough ER are
transitional elements from which smooth protru-
sions appear to be budding. These buds form the
small vesicles that transport secretory proteins from
77Tt'i~~:=-:--:~s~r,.::,.l-.;:__
~~7"-i-~~~~40.-:::t-'--
:::a/}~olgi
.
CIS
Cisternae

the rough ER to the Golgi complex. Interspersed

among the Golgi cisternae are other small vesicles
now known to function in retrograde, not antero-
grade, transport. [Courtesy G. Palade.) ER-to-Golgi
t ransport vesicles
:---.:r~- Smooth prot rusion

644 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 1 1-1m
L..:.._j

EXPERIMENTAL FIGURE 14-16 Fluorescence-tagged fusion isolated by digital processing of the image. First only Vrg4-GFP is
proteins demonst rate Golgi cisternal maturation in a living yeast located in the isolated cisterna and then Sec7-DsRed alone is located
cell. Yeast cells expressing the early Golgi protein Vrg4 fused to GFP in the isolated cisterna, following a brief period in which both proteins
(green fluorescence) and the late Golgi protein Sec7 fused to DsRed are co-localized in this compartment. This experiment is a direct
(red fluorescence) are imaged by time-lapse microscopy. The top series demonstration of the cisternal maturation hypothesis, showing that
of images, taken approximately 1 minute apart, shows a collection of the composition of individual cisternae follow a process of maturation
Golgi cisternae, which at any one time are labeled either with Vrg4 or characterized by loss of early Golgi proteins and gain of late Golgi
Sec7. The bottom series of images show just one Golgi cisterna, proteins. (From Losev et al., 2006, Nature 441 :1002.]

are transported to the cis-Golgi. As this process continues, the they contain either the cis-Golgi protein or. the trans-Golgi
medial-Golgi acquires enzymes from the trans-Golgi whi le protein but only rarely contain both proteins. However, over
losing medial-Golgi enzymes to the cis-Golgi and thus pro- time an individual cisterna labeled with the cis-Golgt protein
gressively becomes a new trans-Golgi compartment. In this can be seen to progressively lose this protein and acquire the
way, secretory cargo proteins acquire carbohydrate modifica- trans-Golgi protein. This behavior is exactly that predicted for
tion in the proper sequential order without being moved from the cisternal maturation model, in which the composition of
one cisterna to another via anterograde vesicle transport. an individual cisterna changes as Golgi resident proteins move
The first evidence that the forward transport of cargo from later to earlier Golgi compartments.
proteins from the cis- to the trans-Golgi occurs by such a Although most protein traffic appears to move through
progressive mechanism, ca lled cisternal maturation, came the Golgi complex by a cisternal maturation mechanism,
from careful microscopic analysis of the synthesis of algal there is evidence that at least some of the COP! transport
scales. These cell-wall glycoproteins are assembled in the cis- vesicles that bud from Golgi membranes contain cargo pro-
Golgi into large. complexes visible in the electron micro- teins (rather than Golgi enzymes) and move in an antero-
scope. Like other secretory proteins, newly made scales move grade (rather than retrograde) direction.
from the cis- to the trans-Golgi, but they can be 20 times
larger than the usual transport vesicles that bud from Golgi
cisternae. Similarly, in the synthesis of collagen by fibro-
blasts, large aggregates of the procollagen precursor often
form in the lumen of the -cis-Golgi (see Figure 20-24). The KEY CONCEPTS of Se,.tion 14.3
procollagen aggregates are too large to be incorporated into Early Stages of the Secretory Pathway
small transport vesicles, and investigators could never find
CO PII vesicles tra nsport proteins from the rough ER to
such aggregates in transport vesicles. These observations show
the cis-Golgi; COPI vesicles transport proteins in the reverse
that the forward movement of these and perhaps all secre-
direction (see Figure 14-11 ).
tory proteins from one Golgi compartment to another does
not occur via small vesicles. COPII coats comprise three components: the small GTP-
A particularly elegant demonstration of cisternal matura- binding protein Sar l , a Sec23/Sec24 complex, and a Secl3/
tion in yeast takes advantage of different-colored fluorescent Sec3 1 complex.
labels to image two different Golgi proteins simultaneously. Components of the COPII coat bind to membrane cargo
Figure 14-16 shows how a cis-Golgi resident protein labeled proteins containi ng a di-acid ic o r other sorting signal in their
with a green fluorescent protein and a trans-Golgi protein la- cytosolic regions (see Figure 14-12). Soluble cargo proteins
beled with a red fluorescent protein behave in the same yeast probably are targeted to COPII vesicles by binding to a mem-
cell. At any given moment individual Golgi cisternae appear brane protein receptor.
to have a distinct compartmental identity, in the sense that

14.3 Early Stages of the Secretory Pathway 645

 ments maintains sufficient levels of these carbohydrate-
Many soluble ER-resident proteins contain a KDEL sort- modifying enzy mes in their functiona l compartments.
ing signal. Binding of this retrieval seq uence to a specific re- Eventually, properly processed cargo proteins reach the trans- ·.
ceptor protein in the cis-Golgi membrane recruits missorted Go!gi network, the most distal Golgi compartment. Here, they
ER proteins into retrograde COPI vesicles (see Figure 14-13). are sorted into one of a number of different kinds of vesicles for
Membrane proteins needed to form COPII vesicles can be delivery to their final destination. In this section we d iscuss the
retrieved from the cis-Golgi by COPI vesicles. One of the different kinds of vesicles that bud from the trans-Go!gi net-
sorting signals that directs membrane proteins into COP! work, the mechanisms that segregate cargo proteins among
vesicles is a KKXX sequence, w hich binds to subunits of the them, and key processing events that occur late in the secretory
COPI coat. A distinct di-arginine sorting signal operates by pathway. The vario us rypes of vesicles that bud from the trans-
a similar mechanism. Golgi are summarized in Figure 14-17.
COPI vesicles also carry Golgi-resident proteins from later
to earlier compartments in the Golgi stack. Vesicles Coated with Clathrin and/or Adapter
Soluble and membrane proteins advance through the Golgi Proteins Mediate Transport from the trans-Golgi
complex by cisternal maturation, a process of anterograde The best-characterized vesicles that bud from the trans-Golgi
transport that depends on resident Golgi enzymes moving by network (TGN) have a two-layered coat: an outer layer com-
COPI vesicular transport in a retrograde direction. posed of the fibrous protein clathrin and an inner layer com-
posed of adapter protein (AP) complexes. Purified clathrin
molecules, which have a three-limbed shape, are called triskel-
ions, from the Greek for "th ree-legged" (Figure 14-18a). Each
14.4 Later Stages of the Secretory limb contains o ne clathrin heavy chain (180,000 MW) and one
clathrin light chain (-35,000-40,000 MW). Triskelions po-
Pathway
lymerize to form a polygonal lattice with an intrinsic curvature
As cargo proteins move from the cis- to the trans-Golgi by cis- (Figure 14-18b). When clathrin polymerizes o n a donor mem-
ternal maturation, modifications to their oligosaccharide chains brane, it does so in association with AP complexes, which fill
are earned out by Golgi-resident enzymes. The retrograde traf- the space between the clathrin lattice and the membrane. Each
ficking of COPI vesicles from later to earlier Golgi compart- AP complex (340,000 MW) contains one copy each of four

~ Plasma

a( membrane

FIGURE 14-17 Vesicle-mediated protein

trafficking from the trans-Golgi network.

~
COPI (purple) vesicles mediate retrograde
transport within the Golgi (0 ). Proteins that
function in the lumen or in the membrane of
the lysosome are first transported from the Late
rrans-Golgi network via clathrin-coated (red) endosome
vesicles (0 ); after uncoating, these vesicles trans-Golgi
fuse with late endosomes, which deliver their
contents to the lysosome. The coat on most
clathrin vesicles contains additional proteins
(AP complexes) not indicated here. Some
vesicles from the rrans·Golgi carrying cargo
)
destined for the lysosome fuse with the Lysosome
lysosome directly (f)), bypassing the endo-
some. These vesicles are coated with a type of
AP complex (blue); it is unknown whether
these vesicles also contain clathrin. The coat
proteins surrounding constitutive (fill and
regulated (Ia) secretory vesicles are not yet
characterized; these vesicles carry secreted
c trans-Golgi
• = Clathrin
• =AP complex
• = COPI

proteins and plasma-membrane proteins from

the trans Golgi network to the cell surface.

646 CHAPTER 14 • Vesi cular Traffic, Secretion, and Endocytosis

(;) VIDEO: Birth of a Clathrin Coat

FIGURE 14- 18 Structure of clathrin coats. (a) A clathrin molecule, (a) Triskelion structure
called a triskelion, is composed of three heavy and three light chains.
It has an intrinsic curvature due to the bend in the heavy chains.
(b) Clathrin coats were formed in vitro by mixing purified clathrin Light
heavy and light chains with AP2 complexes in the absence of
;~oh•;o
membranes. Cryoelectron micrographs of more than 1000 assembled

0
hexagonal clathrin barrel particles were analyzed by digital image
processing to generate an dverage structural representatiOn. The

J
processed image shows only the clathrin heavy chains in a structure
composed of 36 triskelions. Three representative triskelions are Binding site
highlighted in red, yellow, and green. Part of the AP2 complexes for assembly
packed into the interior of the clathrin cage are also visible in this particles
representation . [See B. Pishvaee and G. Payne, 1998, Ce/1 95:443. Part (b)
from Fotin et al., 2004, Nature 432:573.]

different adapter subunit proteins. A specific association be- proteins include Asp-X-Leu-Leu and Asp-Phe-Giy-X-<P se-
tween the globular domain at the end of each clathrin heavy quences (where X and <Pare defined as above).
chain in a triskelion and one subunit of the AP complex both Some vesicles that bud from the trans-Golgi network
promotes the co-assembly of clathrin triskelions with AP com- have coats composed of the AP3 complex. Although the AP3
plexes and adds to the stability of the completed vesicle coat. complex does contain a binding site for clathrin Similar to
By binding to the cytosolic face of membrane proteins, the API and AP2 complexes, it is not clear 'whether clathrin
adapter proteins determine which cargo proreins are specifi- is necessary for functioning of AP3-containing vesicles si nce
cally included in (or excluded from) a budding transport vesi- mutant versions of AP 3 that lack the clathrin binding site
cle. Three different AP complexes are known (APl, AP2, AP3), appear to be fully functional. AP3-coated vesicles mediate
each with four subunits of different, though related, proteins. trafficking to the lysosome, but they appear to bypass the
Recently, a second general type of adapter protein known as late endosome and fuse directly with the lysosomal mem-
GGA has been shown to contain in a single 70,000 MW poly- brane (see Figure 14-17, step f)). In certain types of cells,
peptide both clathrin- and cargo-binding elements similar to such AP3 vesicles mediate protein transport to specialized
those found in the much larger hetero-tetrameric AP com- storage compartments related to the lysosome. For example,
plexes. Vesicles containing each type of adapter complex (AP AP3 is required for delivery of proteins to melanosomes,
or GGA) have been found to mediate specific transport steps which contain the black pigment melanin in skin cells, and
(see Table 14-1). All vesicles whose coats contain one of these to platelet storage vesicles in megakaryocytes, large cells that
complexes utilize ARF to initiate coat assembly onto the donor fragment into dozens of platelets. Mice with mutations in
membrane. As discussed previously, ARF also initiates assem- either of two different subunits of AP3 not only have abnor-
bly of COPI coats. The additional features of the membrane or mal skin pigmentation but also exhibit bleeding disorders.
protein factors that determine which type of coat will assemble The latter occur because tears in blood vessels cannot be re-
after ARF attachment are not well understood at this time. paired without platelets that contain normal storage vesicles.
Vesicles that bud from .the trans-Golgi network en route to
the lysosome by way of the late endosome (see Figure 14-17,
Dynamin Is Required for Pinching Off
step 0 ) have clathrin coats associated with either AP1 or GGA.
Both APl and GGA bind to the cytosolic domain of cargo of Clathrin Vesicles
proteins in the donor membrane. Membrane proteins contain- A fundamental step in the formation of a transport vesicle that
ing a Tyr-X-X-<P sequence, where X is any amino acid and <P we have not yet considered is how a vesicle bud is pinched off
is a bulky hydrophobic amino acid, are recruited into clathrinl from the donor membrane. In the case of clathrinlAP-coated
AP 1 vesicles budding from the trans-Golgi network. This vesicles, a cytosolic protein called dynamin is essential for re-
YXX<P sorting signal interacts with one of the APJ subunits in lease of complete vesicles. At the later stages of bud formation,
Lhe ve::sicle coat. As we discuss in the next section, vesicles with dynamin polymerizes around the neck portion and then hydro-
clathrin/AP2 coats, which bud from the plasma membrane lyzes GTP. The energy derived from GTP hydrolysis is thought
during endocytosis, also can recognize the YXX<P sorting sig- to drive a conformational change in dynamin that stretches
nal. Vesicles coated with GGA proteins and clathrin bind the vesicle neck until the vesicle pinches off (Figure 14-19).
cargo molecules with a different kind of sorting sequence. Interestingly, COPI and COPII vesicles appear to pinch off
Cytosolic sorting signals that specifically bind to GGA adapter from donor membranes without the aid of a GTPase such as

14.4 Later Stages of the Secretory Pathway 647

 Exoplasmic face

TP
Cytosolic face
Dyn amin
C GOP + P;

Solu b le
Fibrous cargo
clath rin protei n
coat

AP
complex
Integral
Integral
receptor
cargo protei n
protein
100 nm

EXPERIMENTAL FIGURE 14.. 20 G,TP hydrolysis by dynamin

is required for pinching off of clathrin-coated vesicles in cell-free
extracts. A preparation of nerve terminals, which undergo extensive
endocytosis, was lysed by treatment with distilled water and incubated
w ith GTP-"(-5, a non hydrolyzable derivative of GTP. After sectioning,
the preparation was treated with gold-tagged anti-dynamin antibody
and viewed in the electron microscope. This image, which shows a
long-necked clathrin/AP-coated bud with polymeri~ed dynamin lining
Clath rin-coated vesicle the neck, reveals that buds can form in the absence of GTP hydrolysis,
but vesicles cannot pinch off. The extensive polymerizat ion of dynamin
FIGURE 14- 19 Model for dynamin-mediated pinching off of
that occurs in the presence of GTP-"(-5 probably does not occur during
clathrin/AP-coated vesicles. After a vesicle bud forms, dynamin
the normal budding process. [From K. Takel et al., 1995, Nature 374:186;
polymerizes over the neck. By a mechanism that is not well under-
courtesy of Pietro De Camilli.]
stood, dynamin-catalyzed hydrolysis of GTP leads to release of the
vesicle from the donor membrane. Note that membrane proteins in the
donor membrane are incorporated into vesicles by interacting with AP
complexes in the coat. [Adapted from K. Takel et al., 1995, Nature 374:186.]
clathrin coat depolymerization. How the action of Hsc70
might he coupled to ARF switching is not well understood.

dynamin. In vitro budding experiments suggest that dimer-

Man nose 6-Phosphate Residues Target
ization of ARF proteins drives the pinching off of COP !
vesicles, but the mechanism is not understood. Soluble Proteins to Lysosomes
Incubation of ce ll ex tracts with a nonhydrolyzable de- As we have seen, many of the sorting signa ls that direct
rivative of GTP provides dramatic evidence for the impor- cargo protein trafficking in the secretory pathway are short
tance of dynamin in pinching off of clathrin/AP2 ves icles amino acid sequences in the targeted protein. In contrast, the
during endocytosis. Such treatment leads to accumulati on of sorting signal that directs soluble lysosomal enzymes from
clathrin-coatcd vesicle buds with excessively long necks that the trans-Golgi network to the late endosome is a carbohy-
are surrounded by polymeric dynamin but do not pinch off drate residue, mannose 6-phosphate (M6P), wh ich is formed
(Figure 14-20). Likewise, cells expressing mutant forms of in the cis-Golgi. The addition and initial processing of one or
dynamin that cannot bind GTP do not form clathrin-coatcd more preformed N-linked oligosaccharide precursors in the
vesicles and instead accumu late similar long- necked vesicle rough ER is the same for lysosomal enzymes as for membrane
buds encased with polymerized dynamin. and secreted proteins, yielding core Man 8 (GicNAch chains
As with COPI and COPII vesicles, clathrin/AP vesicles nor- (see figure 13-18). In the cis-Golgi, theN-linked o ligosaccha-
·' 0

mally lose their coat soon after their formation. Cytosolic rides present on most lysosomal enzymes undergo a rwo-step
Hsc70, a constitutive chaperone protein found in all eu kary- reaction sequence that generates M6P residues (Figure 14-21 ).
otic cells, is thought to usc energy derived from the hydrolysis The addition of M6P residues to the oligos:1.ccharidc chains of
of ATP to drive depolymerization of the clathrin coat into soluble lysosomal enzymes prevents these proteins from un- 0.
triskehons. Uncoating not only releases triskelions for reuse m dergoing the further processing reactions characteristic of se-
the formation of additional vesicles but also exposes v-SNAREs creted and membrane proteins (see Figure 14-14).
for usc in fusion with target membranes. Conforma tional As shown in Figure 14-22, the segregation of M6P-bearing
changes that occur when ARF switches from the GTP-bound lysosomal enzymes from secreted and membrane proteins
to GDP-bound state are thought to regulate the timing of occurs in the trans-Golgi network. Here, t ransmembrane

648 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 \
GlcNAc phosphotransferase Catalytic site Recognition site

FIGURE 14-21 Formation of mannose 6-phosphate (M6P) bound by th1s enzyme, phosphorylated GlcNAc groups are added
residues that target soluble enzymes to lysosomes. The M6P specifically to lysosomal enzymes. Step f) : After release of a modified
·. residues that direct proteins to lysosomes are generated in the cis-Golgi protein from the phosphotransferase, a phosphodiesterase removes
by two Golgi-resident enzymes. Step 0 : An N-acetylglucosamine the GlcNAc group, leaving a phosphorylated man nose residue on the
(GicNAc) phosphotransferase transfers a phosphorylated GlcNAc group lysosomal enzyme. [See A. B. Cantor et al., 1992, J. Bioi. Chem. 267:23349, and
to carbon atom 6 of one or more man nose residues. Because only 5. Kornfeld, 1987,FASEBJ. 1:462.]
lysosomal enzymes contain sequences (red) that are recognized and

mannose 6-phosphate receptors bind the M6P residues on lysosomal enzymes. As a result, undigested glycohpids and
lysosome-destined proteins very tightly and specifically. extracellular components that would normally be degraded
Clathrin/APl vesicles containing the M6P receptor and by lysosomal enzymes accumulate in lysosomes as large in-
bound lysosomal enzymes then bud from the trans-Golgi clusions. Patients with lysosomal storage diseases can have
network, lose their coats, and subsequently fuse with the late a variety of developmental, physiological, and neurological
endosome by mechan isms described previously. Because abnormalities depending on the type and severity of the
M6P receptors can bind M6P at the slightly acidic pH (- 6.5) storage defect. !-cell disease is a particularly severe type of
of the trans-Golgi network but not at a pH less than 6, the lysosomal storage disease in which multi.ple enzymes are
bound lysosomal enzymes are released within late endo- missing from the lysosomes. Cells from affected individuals
somes, which have an internal pH of 5.0-5.5. Furthermore, lack the N-acetylglucosamine phosphotransferase that is re-
a phosphatase within late endosomes usually removes the quired for formation of M6P residues on lysosomal enzymes
phosphate from M6P residues on lysosomal enzymes, pre- in the cis-Golgi (see Figure 14-21). Biochemical comparison
venting any rebinding to the M6P receptor that might occur of lysosomal enzymes from normal individuals w1th those
in spite of the low pH in endosomes. Vesicles budding from from patients with 1-cell disease led to the initial discovery
late endosomes recycle the M6P receptor back to the trans- of mannose 6-phosphate as the lysosomal sorting signal.
Golgi network or, on occasion, to the cell surface. Eventu- Lacking the M6P sorting signal, the lysosomal enzymes in
ally, mature late endosomes fuse with lysosomes, delivering I-cell patients are secreted rather than being sorted to and
the lysosomal enzymes to their final destination. sequestered in lysosomes.
The sorting 'of soluble lysosomal enzymes in the trans- When fibroblasts from patients with I-cell disease are
Golgi network (Figure 14-22, steps D-19) shares many of the grown in a medium containing lysosomal enzymes bearing
features of trafficking between the ER and cis-Golgi compart- M6P residues, the diseased cells acquire a nearly normal
ments mediated by COPII and COPI vesicles. First, mannose intracellular content of lysosomal enzymes. This finding
6-phosphate acts as a sorting signal by interacting with the indicates that the plasma membrane of these cells contains
luminal domain of a receptor protein in the donor membrane. M6P receptors, which can internalize extracellular phos-
Second, the membrane-embedded receptors with their bound phorylated lysosomal enzymes by receptor-mediated endo-
ligands are incorporated into the appropriate vesicles-in this cytosis. This process, used by many cell-surface receptors
case, either GGA or APt-containing clathrin vesicles-by in- to bring bound proteins or particles into the cell, is dis-
teracting with the vesicle coat. Third, these transport vesicles cussed in detail in the next section. It is now known that
fuse only with one specific organelle, here the late endosome, even in norma l cells, some M6P receptors are transported
as the result of interactions between specific v-SNAREs and to the plasma membrane and some phosphorylated lyso-
t-SNAREs. And finally, intracellular transport receptors are somal enzymes are secreted (see Figure 14-22 ). The se-
recycled after dissociating from their bound ligand. creted enzymes can be retrieved by receptor-media ted
endocytosis and directed to lysosomes. This pathway thus
Study of lysosomal Storage Diseases Revealed scavenges any lysosomal enzymes that escape the usual
M6P sorting pathway.
Key Components of the lysosomal
Hepatocytes from patients with I-cell disease contain a
Sorting Pathway normal complement of lysosomal enzymes and no inclusions,
IF.!j A group of genetic disorders termed lysosomal storage even though these cells are defective in man nose phosphoryla-
H diseases are caused by the absence of one or more tion. This finding implies that hepatocytes (the most abundant

14.4 Later Stages of the Secretory Pathway 649

 M6P
receptor
Exterior

Plasma membrane

Cytosol Clath rin-

coated pit

Clathrin-
Receptor- coated
mediated vesicle
endocytosis

Recycling of
M6P receptor
Uncoated
endocy tic
ves icle

Constitutive
secretion IIJ

'-
/
l )
Late
IJ endosome
(low pH)

transport
vesicle

trans-
Golgi
network
Clathrin-
coated
bud
Clathrin-coated
vesicle ( l
__,/ Lysosome

FIGURE 14-22 Trafficking of soluble lysosomal enzymes from receptors and are dephosphorylated, late endosomes subsequently
t he trans-Golgi network and cell surface to lysosomes. Newly fuse with a lysosome (step ~ ). Note that coat proteins and M6P
synthesized lysosomal enzymes, produced in the ER, acquire receptors are recycled (steps m and mJ. and some receptors are
man nose 6-phosphate (M6P) residues in the cis-Golgi (see Figure delivered to the cell surface (step ~ ). Phosphorylated lysosomal
14-21 ). For simplicity, only one phosphorylated oligosaccharide enzymes occasionally are sorted from the trans-Golgi to the cell
chain is depicted, although lysosomal enzymes typically have many surface and secreted. These secreted enzymes can be retrieved by
such chains. In the trans-Golgi network, proteins that bear the M6P receptor-mediated endocytosis (steps ~ -(lJ) , a process that closely
sorting signal interact with M6P recept ors in the membrane and parallels trafficking of lysosomal enzymes from the trans-Golgi
thereby are directed into clathrin/ AP1 vesicles (step 0 ). The coat network to lysosomes. [See G. Griffiths et al., 1988, Cell 52:329; S. Kornfeld,
surrounding released vesicles is rapidly depolymerized (step f) ), and 1992, Ann. Rev. Biochem. 61 :307; and G. Griffiths and J. Gruenberg, 1991 ,
the uncoated transport vesicles fuse with late endosomes (step ID). Trends Cell Bioi. 1 :5.]
After the phosphorylated enzymes dissociate from the M6P

650 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

type of liver cell) employ an M6P-independent pathway for do not associate with these proteins, and thus do not form
sorting lysosomal enzymes. The nature of this pathway, aggregates, would be sorted into unregulated transport vesi-
which also may operate in other cell types, is unknown. • cles by default.

Protein Aggregation in the trans-Golgi May Some Proteins Undergo Proteolytic Processing
Function in Sorting Proteins to Regulated After leaving the trans-Golgi
Secretory Vesicles For some ~ccretory proteins (e.g., growth hormone ) and cer-
As noted in the chapter introduction, all eukaryoric cells con- tain viral membrane proteins (e.g., the VSV glycoprotein), re-
tinuously secrete certain proteins, a process commonly called moval of the N-rerminal ER signal sequence from the nascent
constitutive secretion. Specialized secretory cells also store chain is the only known proteolytic cleavage required to con-
other proteins in vesicles and secrete them only when trig- Yert the polypeptide to rhe mature, active species (see Figure
gered by a specific stimulus. One example of such regulated 13-6). However, some membrane and many soluble secretory
secretion occurs in pancreatic f3 cells, which store newly made proteins initially are synthesized as relatively long-lived, inac-
insulin in special secretory vesicles and secrete insulin in re- tive precursors, termed proproteins, that require further pro-
sponse to an elevation in blood glucose (see Figure 16-38). teolytic processing to generate the mature, active proteins.
These and other secretory cells simultaneously utilize nvo dif- Examples of proteins that undergo such processing are soluble
ferent types of vesicles to move proteins from the trans-Golgi lysosomal enzymes, many membrane proteins such as influ-
network to the cell surface: regulated transport vesicles, often enza hemagglutinin (HA), and secreted proteins such as serum
simply called secretory vesicles, and unregulated transport albumin, insulin, glucagon, and the yeast a mating factor. In
vesicles, also called constitutive secretory vesicles. general, the proteolytic conversion of a proprotein to the cor-
A common mechanism appears to sort regulated proteins responding mature protein occurs after the proprotein has
as diverse as ACTH (adrenocorticotropic hormone), insulin, been sorted in the trans-Golgi network to appropriate vesicles.
and trypsinogen into regulated secretOf)' vesicles. Evidence In the case of soluble lysosomal enzymes~ the proproteins
for a common mechanism comes from experiments in which arc called proenzymes, which are sorted by the M6P recep-
recombinant DNA techniques are used to induce the synthe- tor as catalytically inactive enzymes. In the late endosome or
sis of insulin and trypsinogen in pituitary tumor cells already lysosome a proenzyme undergoes a proteolytic cleavage that
synthesizing ACTH. In these cells, which do not normally generates a smaller but enzymatically active polypeptide. De-
express insulin or trypsinogen, all three proteins segregate laying the activation of lysosomal proenzymes until thC)
into the same regulated secretory vesicles and are secreted reach the lysosome prevents them from digesting macromol-
together when a hormone binds to a receptor on the pitu- ecules in earlier compartments of the secretory pathway.
itary cells and causes a rise in cytosolic Ca 2+. Although these Normally, mature vesicles carrying secreted proteins to
three proteins share no identical amino acid sequences that the cell surface are formed by fusion of several immature
might serve as a sorting sequence, they must have some com- ones containing proprotein. Proteolytic cleavage of propro-
mon feature that signals their incorporation into regulated teins, such as proinsulin, occurs in vesicles after they move
secretory vesicles. away from the trans-Golgi network (Figure 14-23). The pro-
Morphologic evidence suggests that sorting into the regu- proteins of most constitutively secreted proteins (e.g., albumin)
lated pathway is controlled by selective protein aggregation. are cleaved only once at a site C-terminal to a dibasic recogni-
For instance, immature vesicles in this pathway-those that tion sequence such as Arg-Arg or Lys-Arg (Figure 14-24a).
have just budded from the trans-Golgi network-<:ontain dif- Proteolytic processing of proteins whose secretion is regu-
fuse aggregates of secreted protein that are visible in the elec- lated generally entails additional cleavages. In the case of
tron microscope. These aggregates also are found in vesicles proinsulin, multiple cleavages of the single polypeptide chain
that are in the process of budding, indicating that proteins yields the N-terminal B chain and the C-terminal A chain of
destined for regulated secretory vesicles selectively aggregate mature insulin, which arc linked by disulfide bonds, and th e
together before their incorporation into the vesicles. central C peptide, which is lost and subsequently degraded
Other studies have shown that regulated secretory vesi- (figure 14-24b).
cles from mammalian secretory cells contain three proteins, The breakthrough in identifying the proteases responsible
chromogranin A, chromogranin B, and secretogranin II, that for such processing of secreted proteins came from analysis of
together form aggregates when incubated at the ionic condi- yeast with a mutation in the KEX2 gene. These mutant cells
tions (pH 6.5 and 1 mM Cal+) thought to occur in the trans- synt hP~ized the precursor of the a mating factor but could
(jolgi network; such aggregates do not form at the neutral not proteolytically process it to the functional form and thus
pH of the ER. The selective aggregation of regulated secreted were unable to mate with cells of the opposite mating type
proteins together with chromogranin A, chromogranin B, or (see Figure 16-23). The wild-type KEX2 gene encodes an en-
secretogranin II could be the basis for sorting of these pro- doprotease that cleaves the a-factor precurso r at a site C-
teins into regulated secretory vesicles. Secreted proteins that terminal to Arg-Arg and Lys-Arg residues. Mammals contain

14.4 Later Stag es of the Secretory Pathway 651

EXPERIMENTAL FIGURE 14-23 Proteolytic cleavage of (a) Proinsulin antibody
proinsulin occurs in secretory vesicles after they have budded
from t he trans-Golgi network. Serial sections of the Golgi region of an
insulin-secreting cell were stained with (a) a monoclonal antibody that
recognizes proinsulin but not insulin, or (b) a different antibody that
recognizes insulin but not proinsulin. The antibodies, which were
bound to electron-opaque gold particles, appear as dark dots in these
electron micrographs (see Figure 9-29). Immature secretory vesicles
(closed arrowheads) and vesicles budding from the trans-Golgi (arrows)
stain with the proinsulin antibody but not with insulin antibody. These
...
vesicles contain diffuse protein aggregates that include proinsulin and
other regulated secreted proteins. Mature vesicles (open arrowheads)
stain with insulin antibody but not with proinsulin antibody and have a
dense core of almost crystalline insulin. Since budding and immature
secretory vesicles contain proinsulin (not insulin), the proteolytic
conversion of proinsulin to insulin must take place in these vesicles
after they bud from the trans-Golgi network. The inset in (a) shows a
proinsulin-rich secretory vesicle surrounded by a protein coat (dashed
..
...
G ,:· ..
. ··.
..
. .,
.
~ - ~*
-~

line). [From L. Orci et al., 1987, Ce/1 49:865; courtesy of L. Orci.) ,...

1·~···.-:·~~·:
..
..· · ··.,.
~. · : ~:
~
-~·
·~
. . ,,,
. .
. ~·· · '··....... I
'
a family of endoproteases homologous to the yeast KEX2 I 'l'o / : ....-.( "
protein, all of which cleave a protein chain on the C-terminal .. :. .·.:·
.. _.,..
side of an Arg-Arg or Lys-Arg sequence. One, called furin, is
found in all mammalian cells; it processes proteins such as (b) Insulin antibody
albumin that are secreted by the continuous pathway. In con-
trast, the PC2 and PCJ endoproteases are found only in cells
that exhibit regulated secretion; these enzymes are localized
to regulated secretory vesicles and proteolytically cleave the
precursors of many hormones at specific sites.

Several Pathways Sort Membrane Proteins

to the Apical or Basolateral Region ·~
.. -~
.
of Polarized Cells J.:_.
The plasma membrane of polarized epithelial cells is divided
into two domains, apical and basolateral; tight junctions lo-
cated between the two domains prevent the movement of
plasma-membrane proteins between the domains (see Figure
20-10). Several sorting mechanisms direct newly synthesized
membrane proteins to either the apical or basolateral do- G
main of epithelial cells, and any one protein may be sorted
by more than one mechanism. As a result of this sorting and
the restriction on protein movement within the plasma mem-
brane due to tight junctions, distinct sets of proteins are
found in the apical or basolateral doma in. This preferential
localization of certain transport proteins is critical to a vari-
ety of important physiological functions, such as absorption
of nutrients from the intestinal lumen and acidification of
the stomach lumen (see Figures 11-30 and 11-31 ).
Microscopic and cell-fractionation studies indicate that
proteins destined for either the apical or the basolateral and then move to the apical region, whereas proteins des-
membranes are initially transported together to the mem- tined for the basolateral membrane are sorted into other
branes of the trans-Golgi network. In some cases, proteins vesicles that move to the basolateral region. The different
destined for the apical membrane are sorted into their own vesicle types can be distinguished by their protein constitu-
transport vesicles that bud from the trans-Golgi network ents, including distinct Rab and v-SNARE proteins, which

652 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 (a) Constitutive secreted proteins FIGURE 14-24 Proteolytic processing of pro proteins in the
constitutive and regulated secretion pathways. The processing of
Proalbumin
proalbumin and proinsulin is typical of the constitutive and regulated
N H3+ I Arg ArgJ ICOO pathways, respectively. The endoproteases that function in such

1 1 endoprotease
processing cleave at the ( -terminal side of two consecutive amino
acids. (a) The endoprotease fu rin acts on the precursors of constitutive
secreted proteins. (b) Two endoproteases, PC2 and PC3, act on the
NH 3+ I A rg Arg j Albumin ICOO precursors of regulated secreted proteins. The final processing of many
such proteins is catalyzed by a carboxypeptidase that sequentially
removes two basic amino acid residues at the (-terminus of a
(b) Regulated secreted proteins
polypeptide. [See D. Steiner et al., 1992, J. Bioi. Chem. 267:23435.]
Proinsulin
s----------------------s
apparently target them to the appropriate plasma-membrane
domain. In this mechanism, segregation of proteins destined
s-s for either the apical or basolateral membranes occurs as
PCJ endoprotease ~ PC2 endoprotease cargo proteins are incorporated into particular types of ves-
!~ I c Lys Arg I icles budding from the trans-Golgi network.
Such direct basolatera l-apical sorting has been investi-
s s gated in cultured Madin-Darhy canine kidney (MDCK )

NH 3
s
I
B
II Arg A rg
"+'
I ¢Pcoo
s-s
cells, a line of cultured polarized epithelial cells (see Figure
9-4). In MDCK cells infected with the influenza virus, prog-
eny viruses bud only from the apical membrane, '"·hereas in
cells infected with vesicular stomatitis virus (VSV ), progeny
Carbo•ypeptida•

~I
'' iruses bud only from the basolateral membrane. This dif-
Arg II Arg I ference occurs because the HA glycoprotein of influenza
virus is transported from the Golgi complex exclusively to
s s the apical membrane and the VSV G protein is transported

~
s only to the hasolateral membrane (Figure 14-25). Further-
I I
B Insulin more, when the gene encoding HA protein is introduced
into uninfected cells by recombinant DNA techniques, all
s-s
the expressed HA accumulates in the apical membrane, in-
dicating that the sorting signal resides in the HA glycopro-
tein itself and not in other viral proteins produced during
viral infection.

Influenza virus HA glycoprotein

~b
{l./1 I
VSV G glycoprotein

Dire~t
sortmg
apical
r\,
-
FIGURE 14-25 Sorting of proteins desti ned for
the apical and basolateral plasma membranes of

.n:_ ~0·:~~~~~teral
polarized cells. When cultured MOCK cells are infected

~" c----t"" simultaneously with VSV and influenza virus, the VSV G
glycoprotein (pu rple) is found only on the basolateral
tt _ - 1 ~ GPi anchor membrane, whe reas the influenza HA glycoprotein
(green) is found only on the apical membrane. Some

t .Endocytosis '-- t cellular proteins (orange circle), especially those with

a GPI anchor, are likewise sorted directly to the apical
•~
~ ~~
trans-Golg i --_:) membrane and others to the basolateral membrane
2 . . . . ._Jb\
network
Transcytosis r{ (not shown) via specific transport vesicles that bud

~~ \_) ~ 0 __.. ~ '-....Apicalprotein

from the trans-Golgi network. In certain polarized cells,
some apical and basolateral pruleins are transported
Clathrin-./ Recycling of-'1"'\ \..../ 1 together to the basolateral surface; the apical proteins
coated p1t
Basolateral/
plasma membrane
~V

Tight
junction
----:1
~
o- ~ '-
Apical plasma
membrane
(yellow oval) then move selectively, by endocytosis and
t ranscytosis, to the apical membrane. [After K. Simons and
A. Wand1nger-Ness, 1990, Cell 62:207, and K. Mostovet al., 1992,
J. Cell Bioi. 116:577.]

14.4 Later Stages of the Secretory Pathway 653

 Among the cellular proteins that undergo similar apical-
basolateral sorting in the Golgi are those with a glyco- Many vesicles that bud from the trans-Golgi network as
sylphosphatidylinositol (GPT) membrane anchor. In MDCK well as endocytic vesicles bear a coat composed of AP (adapter
cells and most other types of epithelial cells, GPI-anchored protein) complexes and clathrin (see Figure 14-l8).
proteins are targeted to the apical membrane. In membranes • Pinching off of clathrin-coated vesicles requires dynamin,
GPI-anchored proteins are clustered into lipid rafts, which which forms a collar around the neck of the vesicle bud and
are rich in sphingolipids (see Chapter 10). This finding sug- hydrolyzes GTP (see Figure 14-19).
gests that lipid rafts are localized to the apical membrane
• Soluble enzymes destined for lysosomes arc modified in
along with proteins that preferentially partition them in
the cis-Golgi, yielding multiple mannose 6-phosphate (M6P)
many cells. However, the GPI anchor is not an apical sorting
residues on their oligosaccharide chains.
signal in all polarized cells; in thyroid cells, for example,
GPI-anchored proteins are targeted to the basolateral mem- • M6P receptors in the membrane of the trans-Golgi net-
brane. Other than GPI anchors, no unique sequences have work bind proteins bearing M6P residues and direct their
been identified that are both necessary and sufficient to tar- transfer to late endosomes, where receptors and their ligand
get proteins to either the apical or basolateral domain. In- proteins dissociate. The receptors then are recycled to the
stead, each membrane protein may contain multiple sorting Golgi or plasma membrane, and the lysosomal enzymes are
signals, any one of which can target it to the appropriate delivered to lysosomes (see Figure J4-22).
plasma-membrane domain. The identification of such com- • Regulated secreted proteins are concentrated and stored
plex signals and of the vesicle coat proteins that recognize in secretory vesicles to await a neural or hormonal signal
them is currently being pursued for a number of different for exocytosis. Protein aggregation within the trans-Golgi
proteins that are sorted to specific plasma-membrane do- network may play a role in sorting secreted proteins to the
mains of polarized epithelial cells. regulated pathway.
Another mechanism for sorting apical and basolateral
• Many proteins transported through the secretory pathway
proteins, also illustrated in Figure 14-25, operates in hepato-
undergo post-Golgi proteolytic cleavages that yield the mature,
cytes. The basolateral membranes of hepatocytes face the
active proteins. Generally, proteolytic maturation can occur in
blood (as in intestinal epithelial cells), and the apical mem-
vesicles carrying proteins from the trans-Golgi network to the
branes line the small intercellular channels into which bile is
cell surface, in the late endosome, or in the lysosome.
secreted. In hepatocytes, newly made apical and basolateral
proteins are first transported in vesicles from the trans-Golgi In polarized epithelial cells, membrane proteins destined
network to the basolateral region and incorporated into the for the apical or basolateral domains of the plasma mem-
plasma membrane by exocytosis (i.e., fusion of the vesicle brane are sorted in the trans-Golgi network into different
membrane with the plasma membrane). From there, both transport vesicles (see Figure 14-25 ). The GPI anchor is the
basolateral and apical proteins are endocytosed in the same only apical-basolateral sorting signal identified so far.
vesicles, but then their paths diverge. The endocytosed baso- • In hepatocytes and some other polarized cells, all plasma-
lateral proteins are sorted into transport vesicles that recycle membrane proteins are directed first to the basolateral mem-
them to the basolateral membrane. In contrast, the apically brane. Apically destined proteins then are endocytosed and
destined endocytosed proteins are sorted into transport ves- moved across the cell to the apical membrane (transcytosis ).
icles that move across the cell and fuse with the apical mem-
brane, a process called transcytosis. This process also is used
to move extracellular materials from one side of an epithe-
lium to another. Even in epithelial cells, such as MOCK cells,
in which apical-basolateral protein sorting occurs in the
1 4.5 Receptor-Mediated Endocytosis
Golgi, transcytosis may provide an editing function by which In previous sections we have explored the main pathways
an apical protein sorted incorrectly to the basolateral mem- whereby soluble and membrane secretory proteins synthe-
brane would be subjected to endocytosis and then correctly sized on the rough ER are delivered to the cell surface or
delivered to the apical membrane. other destinations. Cells also can internalize materials from
their surroundings and sort these to particular destinations. A
few cell types (e.g., macrophages) can take up whole bacteria
KfY CONCEPTS of Section 14.4 and other large particles by phagocytosis, a nonselective
actin-mediated process in which extensions of the plasma
later Stages of the Secretory Pathway membrane envelop the ingested material, forming large vesi-
• The trans-Golgi network (TGN) is a major branch point cles called phagosomes (sec Figure 17-19 ). In contrast, all
in the secretory pathway where soluble secreted proteins, eukaryotic cells continually engage in endocytosis, a process
lysosomal proteins, and in some cells membrane proteins in which a small region of the plasma membrane invaginates
destined for the basolateral or apical plasma membrane are to form a membrane-limited vesicle about 0.05-0.1 f.lm in di-
segregated into different transport vesicles. ameter. In one form of endocytosis, called pinocytosis, small
droplets of extracellular fluid and any material dissolved in it

654 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

are nonspecifically taken up. Our focus in this section, how- general, the transmembrane receptor proteins that function
ever, is on receptor-mediated endocytosis, in which a specific in the uptake of extracellular ligands are internalized from
receptor on the cell surface binds tightly to an extracellular the cell surface during endocytosis and are then sorted and
macromolecular ligand that it recognizes; the plasma mem- recycled back to the cell surface, much like the recycling of
brane region containing the receptor-ligand complex then M6P receptOrs to the plasma membrane and trans-Golgi. The
buds inward and pinches off, becoming a transport vesicle. rate at which a ligand is internalized is limited by the amount
Among the common macromolecules that vertebrate cells of its corresponding receptor on the cell surface.
internalize by receptor-mediated endocytosis are cholesterol- Clathrin/AP2 pits make up about 2 percent of the surface
containing particles called low-density lipoprotein (LDL), the of cell<; such as hepatocytes and fibroblasts. Many internal-
iron-carrying protein transferrin, many protein hormones ized ligands have been observed in these pits and vesicles,
(e.g., insulin), and certain glycoproteins. Receptor-mediated which are thought to function as intermediates in the endo-
endocytosis of such ligands generally occurs via clathrin/AP2- cytosis of most (though not all) ligands bound to cell-surface
coated pits and vesicles in a process similar to the packaging receptors (Figure 14-26). Some receptors arc clustered ()\er
of lysosomal enzymes by binding of mannose 6-phosphate clathrin-coated pits even in the absence of ligand. Other re-
(M6P) in the trans-Golgi network (see Figure 14-22). As ceptors diffuse freely in the plane of the plasma membrane
noted earlier, some M6P receptors are found on the cell sur- but undergo a conformational change when binding to li-
face, and these participate in the receptor-mediated endocy- gand, so that when the receptor-ligand complex diffuses into
tosis of lysosomal enzymes that are mistakenly secreted. In a clathrin-coated pit, it is retained there. Two or more types

(a) (b)

..
..
LDL-ferritin
LDL-ferritin
Clathrin-coated pit
(c) (d)

EXPERIMENTAL FIGURE 14·26 The initial stages of receptor- to 37 oc and then prepared for microscopy at periodic intervals. (a) A
mediated endocytosis of low-density lipoprotein (LDL) part icles coated pit, showing the clathrin coat on the inner (cytosolic) surface
are revealed by electron microscopy. Cultured human fibrobla~ts of the pit, soon after the temperature was raised. (b) A pit Lontaining
were incubated in a medium containing LDL particles covalently linked LDL apparently closing on itself to form a coated vesicle. (c) A coated
to the electron-dense, iron-containing protein ferritin; each small iron vesicle containing ferritin-tagged LDL particles. (d) Ferritin-tagged LDL
particle in ferritin is visible as a small dot under the electron micro- particles in a smooth-surfaced early endosome 6 minutes after internal-
scope. Cells initially were incubated at 4 °C; at this temperature LDL ization began. (Photographs courtesy of R. Anderson. Reprinted by permission
can bind to its receptor, but internalization does not occur. After excess from J. Goldstein et al., Nature 279:679. Copyright 1979, Macmillan Journals
LDL not bound to the cells was washed away, the cells were warmed Limited. See also M.S. Brown and J. Goldstein, 1986, Science 232:34.]

14.5 Receptor-Mediated Endocytosis 655

of receptor-bound ligands, such as LDL and transferrin, can core of the particle is packed with cholesterol in the form of
be seen in the same coated pit or vesicle. cholesteryl esters.
Two general experimental approaches have been used ro
study how LDL particles enter cells. The first method makes
Cells Take Up Lipids from the Blood
use of LDL that has been labeled by the covalent attachment
in the Form of Large, Well-Defined of radioactive 1251 to the side chains of tyrosine residues in
Lipoprotein Complexes apoB-100 on the surfaces of the LDL particles. After cul-
Lipids absorbed from the diet in the intestines or stored in tured cells are incubated for several hours with the labeled
adipose tissue can be distributed to cells throughout the LDL, it is possible to determine how much LDL is bound to
body. To facilitate the mass transfer of lipids between cells, the surfaces of cells, how much is internalized, and how
animals have evolved an efficient way to package from hun- much of the apoB-100 component of the LDL is degraded by
dreds to thousands of lipid molecules into water-soluble, enzymatic hydrolysis to individual amino acids. The degra-
macromolecular carriers, called lipoproteins, that cells can dation of apoB-100 can be detected by the release of 125 [-
take up from the circulation as an ensemble. A lipoprotein tyrosine into the culture medium. Figure 14-28 shows the
particle has a shell composed of proteins (apolipoproteins) time course of events in receptor-mediated cellular LDL pro-
and a cholesterol-containing phospholipid monolayer. The cessing, determined by pulse-chase experiments with a fixed
shell is amphipathic because its outer surface is hydrophilic, concentration of 1251-labeled LDL.,These experiments clearly
making these particles water soluble, and its inner surface is demonstrate the order of events: surface binding of LDL ~
hydrophobic. Adjacent to the hydrophobic inner surface of internalization~ degradation. The second approach in-
the shell is a core of neutral lipids containing mostly choles- volves tagging LDL particles with an electron-dense label
teryl esters, triglycerides, or both. Mammalian lipoproteins that can be detected by electron microscopy. Such studies
fall into different classes, defined by their differing buoyant can reveal the details of how LDL particles first bind to the
densities. The class we will consider here is low-density lipo- surface of cells at clathrin-coated endocytic pits and then re-
protein (LDL). A typical LDL particle, depicted in Figure main associated with the coated pits as the'y invaginate and
14-27, is a sphere 20-25 nm in diameter. The amphipathic bud off to form coated vesicles and finally are transported to
outer shell is composed of a phospholipid monolayer and a endosomes (see Figure 14-26).
single molecule of a large protein known as apoB-1 00; the

I' Phospho-
I
lipid '?-.,....---- - Polar
surface
Unesterified /.\\
cholesterol Degradation
;.:.;;,...,...."r:M.......:;..-.:-~ Apolar
core

~
Cholesteryl
ester

Apolipoprotein B

Time at 37 oc (min)
LDL
.l(PER II MENTAL FIGURE 14-28 Pulse-chase experi men t
FIGURE 14-27 M odel of low-density l ipoprotein (LDL). This class demonstrates precursor-product relations i n cellular uptake
and the other classes of lipoproteins have the same general structure: of LDL. Cultured normal human skin fibroblasts were incubated in
an amphipathic shell, composed of a phospholipid monolayer (not a medium containing 1251-LDL for 2 hours at 4 ° ( (the pulse). After
bilayer), cholesterol, and protein, and a hydrophobic core, composed excess 1251-LDL not bound to the cells was washed away, the cells were
mostly of cholesteryl esters or triglycerides or both but with minor incubated at 37 C for the indicated amounts of time in the absence of
amounts of other neutral lipids (e.g., some vitamins). This model of LDL external LDL (the chase). The amounts of surface-bound, internalized,
is based on PIPrtron microscopy and other low-resolution biophysical and degraded (hydrolyLed) ' " 1-LDL were measured. Bmding but not
methods. LDL is unique in that it contains only a single molecule of one internalization or hydrolysis of LDL apoB-1 00 occurs during the 4 ° (
type of apolipoprotein (apoB), which appears to wrap around the out- pulse. The data show the very rapid disappearance of bound 1251-LDL
side of the particle as a band of protein. The other lipoproteins contain from the surface as it is internalized after the cells have been warmed
multiple apolipoprotein molecules, often of different types. [Adapted to allow membrane movements. After a lag period of 15-20 minutes,
from M. Krieger, 1995, in E. Haber, ed., Molecular Cardiovascular Medicine, Scien- lysosomal degradation ofthe internalized 1251-LDL commences.
tific American Medicine, pp. 31-47.] [SeeM. 5. Brown and J. L. Goldstein, 1976, Cell 9:663.)

656 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 Receptors for Low-Density Lipoprotein domain. Seven cysteine-rich repeats form the ligand-binding
and Other Ligands Contain Sorting Signals domain, which mteracts with the apoB-100 molecule in an
LDL particle. Figure 14-29 shows how LDL receptor proteins
That Target Them for Endocytosis
facilitate internalization of LDL particles by receptor-mediated
The key to understanding how LDL particles bind to the cell endocytosis. After internalized LDL particles reach lysosomes,
surface and are then taken up into endocytic vesicles came lysosomal proteases hydrolyze their surface apolipoproteins
from discovery of the LDL receptor (LDLR). The LDL recep- and lysosomal cholesteryl esterases hydrolyze their core choles-
tor is an 839-residue glycoprotein with a single transmembrane teryl esters. The unesterificd cholesterol is then free to leave the
segment; it has a short C-terminal cytosolic segment and a long lysosome and be used as necessary by the cell in tht: synthesis
N-terminal exoplasmic segment that contains the LDL-binding of membranes or various cholesterol derivatives.

At neutral pH, ligand-binding

arm is free to bind another
LDL receptor LDL particle

Plasma membrane

AP2
complex

Coated
vesicle

Early Late endosome

endosome

Lysosome
pH 5.0

Cholesterol

·.
FIGURE 14-29 Endocytic pathway for i nternalizing low-density some) fuses with the late endosome. The acidic pH in this compart-
lipoprotein (LDL). Step 0 : Cell-surface LDL receptors bind to an apoB ment causes a conformational change in the LDL receptor that leads
protein embedded in the phospholipid outer layer of LDL particles. to release of the bound LDL particle. Step B: The late endosome
Interaction between the NPXY sorting signal in the cytosolic tail of the fuses with the lysosome, and the proteins and lipids of the free LDL
LDL receptor and the AP2 complex incorporates the receptor-ligand particle are broken down to their constituent parts by enzymes in the
complex into forming endocytic vesicles. Step 6 : Clathrin-coated lysosome. Step L'J : The LDL receptor recycles to the cell surface, where
pits (or buds) containing receptor-LDL complexes are pinched off by at the neutral pH of the exterior medium the receptor undergoes a
the same dynamin-mediated mechanism used to form clathrin/AP1 conformational change so that it can bind another LDL particle. [See
vesicles on the trans-Golgi network (see Figure 14- 19). Step J): After M. 5. Brown and J. L Goldstein, 1986, Science 232:34, and G. Rudenko et al.,
the vesicle coat is shed, the uncoated endocytic vesicle (early en do- 2002, Science 298:2353.]

14.5 Receptor-Mediated Endocytosis 657

 .. The discovery of the LDL receptor and an understand- In some cell-surface proteins, however, other sequences
ing of how it functions came from studying cells from (e.g., Leu-Leu) or covalently linked ubiquitin molecules sig-
patients with familial hypercholesterolemia (FH), a heredi- nal endocytosis. Among the proteins associated with clath-
tary disease that is marked by elevated plasma LDL choles- rin/AP2 vesicles, several contain domains that specifically
terol and is now known to be caused by mutations in the bind to ubiquitin, and it has been hypothesized that these
LDLR gene. In patients who have one normal and one de- vesicle-associated proteins mediate the selective incorpora-
fective copy of the LDLR gene (heterozygotes), LDL choles- tion of ubiquitinated membrane proteins into endocytic
terol in the blood is increased about twofold. Those with vesicles. As described later, the ubiquitin tag on endocy-
two defective LDLR genes (homut.ygotes) have LDL choles- tosed membrane protems is also recognized at a later stage
terol levels that are from fourfold to sixfold as high as nor- in the endocytic pathway and plays a role in delivering
mal. FH heterozygotes commonly develop cardiovascular these proteins into the interior of the lysosome, where they
disease about I 0 years earlier than normal people do, and are degraded.
FH homozygotes usually die of heart attacks before reaching
their late 20s.
A variety of mutations in the gene encoding the LDL re- The Acidic pH of late Endosomes
ceptor can cause familial hypercholesterolemia. Some muta-
Causes Most Receptor-ligan,d
tions prevent the synthesis of the LDLR protein; others
prevent proper folding of the receptor protein in the ER, Complexes to Dissociate
leading to its premature degradation (Chapter 13); still other The overall rate of endocytic internalization of the plasma
mutations reduce the ability of the LDL receptor to bind membrane is quite high: cultured fibroblasts regularly inter-
LDL tightly. A particularly informative group of mutant re- nalize 50 percent of their cell-surface proteins and phospho-
ceptors are expressed on the cell surface and bind LDL nor- lipids each hour. Most cell-surface receptors that undergo
mally but cannot mediate the internalization of bound LDL. endocytosis will repeatedly deposit their ligands within the
In individuals with this type of defect, plasma-membrane re- cell and then recycle to the plasma membrane, once again to
ceptors for other ligands arc internalized normally, but the mediate internalization of ligand molecules. For instance, the
mutant LDL receptor is not recruited into coated pits. Anal- LDL receptor makes one round trip into and out of the cell
ysis of this mutant receptor and other mutant LDL receptors interior every 10-20 minutes, for a total of several hundred
generated experimentally and expressed in fibroblasts identi- trips in its 20-hour life span.
fied a four-residue motif in the cytosolic segment of the re- Internalized receptor-ligand complexes commonly fol -
ceptor that is crucial for its internalization: Asn-Pro-X-Tyr, low the pathway depicted for the M6P receptor in Figure
where X can be any amino acid. This NPXY sorting signal 14-22 and the LDL receptor in Figure 14-29. Endocytosed
binds to the AP2 complex, linking the clathrin/AP2 coat to cell-surface receptors typically dissociate from their ligands
the cytosolic segment of the LDL receptor in coated pits. A within late endosomes, which appear as spherical vesicles
mutation in any .of the conserved residues of the NPXY sig- with tubular branching membranes located a few microm-
nal will abolish the ability of the LDL receptor to be incor- eters from the cell surface. The original experiments that
porated into coated pits. defined the late endosome sorting vesicle utilized the asialo-
A small number of individuals who exhibit the usual glycoprotein receptor. This liver-specific protein mediates
symptoms associated with familial hypercholesterolemia the binding and internalization of abnormal glycoproteins
produce normal LDL receptors. In these individuals, the whose oligosaccharides terminate in galactose rather than
gene encoding the AP2 subunit protein that binds the the normal sialic acid; hence the name asialoglycoprotein.
NPXY sorting signal is defective. As a result, LDL recep- Electron microscopy of liver cells perfused with asialoglyco-
tors arc not mcorporated into clathrin/AP2 vesicles and protein reveal that 5-10 minutes after internalization, li-
endocytosis of LDL particles is compromised. Analysis of gand molecules are found in the lumen of late endosomes,
patients with this genetic disorder highlights the impor- while the tubular membrane extensions are rich in receptor
tance of adapter proteins in protein trafficking mediated by and rarely contain ligand. These findings indicate that the
clathrin vesicles. • late endosome is the organelle in which receptors and li-
gands arc uncoupled.
The dissociation of receptor-ligand complexes in late en-
Mutational studies have shown that other cell-surface dosomes occurs not only in the endocytic pathway but also
receptors can be directed into forming clathrin/AP2 pits by a in the delivery of soluble lysosomal enzymes via the secretory
YXXI- sorting ~ignal. Recall from our earlier discussion that pathway (sec Figure 14-22). As discu~~cJ in Chapter 11, the
this same sorting signal recruits membrane proteins into membranes of late endosomes and lysosomes contain V-class
clathrin/APl vesicles that bud from the trans-Golgi network proton pumps that act in concert with Cl- channels to acid-
by binding to a subunit of APl (see Table 14-2). All these ify the vesicle lumen (see Figure 11-14 ). Most receptors, in-
observations indicate that YXXF is a widely used signal for cluding the M6P receptor and cell-surface receptors for LDL
sorting membrane proteins to clathrin-coated vesicles. particles and asialoglycoprotein, bind their ligands tightly at

658 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

..
(a) (b)
Released
Cell surface [pH 7.0) Endosome [pH 5] LDL particle
Ligand-binding
arm (R1- l
LDL LDL Surface of
receptor particle ~-propeller
domain becomes
NPXY positively charged,
ApoB protein
sortin and then binds
to the ligand-
binding arm

FIGURE 14-30 Model for pH-dependent binding of LDL particles

by the LDL receptor. Schematic depiction of an LDL receptor at the
neutral pH found at the cell surface (a) and at the acidic pH found in
.· the interior of the late endosome (b). (a) At the cell surface, apoB-1 00
on the surface of an LDL particle binds tightly to the receptor. Of the
~-propeller
seven repeats (R1-R7) in the ligand-binding arm, R4 and RS appear to
domain
be most critical for LDL binding. (b, top) Within the endosome, histidine
residues in the [3-propeller domain of the LDL receptor become pro-
tonated. The positively charged propeller can bind with high affinity to
the ligand-binding arm, which contains negatively charged residues,
causing release of the LDL particle. (b. bottom) Experimental electron
density and Cn trace model of the extracellular region of the LDL recep·
tor at pH 5.3 based on x-ray crystallographic analysis. In this conforma-
tion, extensive hydrophobic and Ionic interactions occur between the
f3 propeller and the R4 and RS repeats. [Part (b) from G. Rudenko et al.,
2002, Soence 298:2353.]

neutral pH but release their ligands if the pH is lowered to storage in the body) and from the intestine (the site of iron
6.0 or below. The late endosome is the first vesicle encoun- absorption). The iron-free form, apotransferrin, binds two
tered by receptor-ligand complexes whose luminal pH is suf- Fe3 ions very tightly to form ferrotransferriu. All mamma-
ficiently acidic to promote dissociation of most endocytosed lian cells contain cell-surface transferrin receptors that avidly
receptors from their tightly bound ligands. bind fcrrotransferrin at neutral pH, after which the receptor-
The mechanism by which the LDL receptor releases bound bound ferrotransferrin is subjected to endocytosis. Like the
LDL particles is now understood in detail (Figure 14-30). At components of an LDL particle, the two bound Fe ;+ atoms
the endosomal pH of 5 .0-5.5, histidine residues in a region remain in the cell, but the apotransferrin part of the ligand
known as the 13-propeller doma in of the receptor become does not dissociate from the receptor in the late endosome,
protonated, forming a site that can bind with high affinity to and within minutes after being endocytosed, apotransferrin
the negatively charged repeats in the LDL-binding domain. is returned to the cell surface and secreted from the cell .
This intramolecular interaction sequesters the repeats in a As depicted in Figure 14-31, the explanation for the be-
conformation that cannot simultaneously bind to apoB-100, havior of the transferrin receptor-ligand complex lies in the
thus causing release of the bound LDL particle. unique ability of apotransfe rrin to remain bound to the
transferrin receptor at the low pH (5 .0-5.5 ) of late endo-
somes. At a pH of less t han 6.0, the two bound Fe>+ atoms
The Endocytic Pathway Delivers Iron to Cells dissociate from ferrotransfcrrin, are reduced to Fe 2 - by an
unknown mechanism, and then are exported into the cytosol
Without Dissociation of the Receptor-Transferrin
by an endosomal transporter specific for divalent metal ions.
Complex in Endosomes The reccptor-apotransferrin complex remaining after disso-
The endocytic pathway involving the transferrin receptor ciation of the iron atoms is recycled back to the cell surface.
and its ligand d iffers from the LDL pathway in that the Although apotransferrin binds tightly to its receptor at a pH
receptor- liga nd complex does not dissociate in late endo- of 5.0 or 6.0, it does not bind at neutral pH. Hence the
somes. Nonetheless, changes in pH also med iate the sorting bound apotransferrin dissociates from the transferrin recep-
of receptors and ligands in the transferrin pathway, which tor w hen the recycling vesicles fuse with th e plasma mem-
functions to deliver iron to cells. brane and the receptor-ligand complex encounters the
A major glycoprotein in the blood, transferrin transports neutral pH of the extracellular interstitial fluid or growth
iron to all tissue cell s from the liver (the main site of iron medium. The recycled receptor is then free to bind another

14.5 Receptor-Mediated Endocytosis 659

.·
 .· ·.
\

Ferrotransferrin
Apotransferrin
Exterior dissociates from
{pH 7.0) receptor at
neutral pH

Adapter complex _ _. •• ,.
Clathrin

'11---·-~~ Fe3•
Low pH causes release of Fe3+ Fe 2+ Fe2+ ~
. J
, ~ ~ U ~ Apotransferrin
from ligand; ligand remains pH~
bound to receptor
Late endosome

FIGURE 14-31 The transferrin cycle, which operates in all with the membrane of the endosome. Fe3 is released from the
growing mammalfan cells. Step 0 : The transferrin dimer carrying two receptor-ferrotransferrin complex in the acidic late endosome com-
bound atoms of FeH, called ferrotransferrin, binds to the transferrin partment. Step til:The apotransferrin protein remains bound to its re-
receptor at the cell surface. Step FJ:Interaction between the tail of the ceptor at this pH, and they recycle to the cell surface together. Step [;!:
transferrin receptor and the AP2 adapter complex incorporates the The neutral pH of the exterior medium causes release of the iron-free
receptor-ligand complex into endocytic clathrin-coated vesicles. apotransferrin. [See A. Ciechanover et al., 1983, J. Bioi. Chem. 258:9681.]
Steps iJ and B :The vesicle coat is shed and the endocytic vesicles fuse

molecule of ferrotransferrin, and the released apotransferrin

is carried in the bloodstream to the liver or intestine to be • Sorting signals in the cytosolic domain of cell-surface re-
reloaded with tron. ceptors target them into clathrin/AP2-coated pits for inter-
nalization. Known signals include the Asn-Pro-X-Tyr, Tyr-
X-X-<D, and Leu-Leu sequences (see Table 14-2).
The endocytic pathway de livers some ligands (e.g., LDL
particles) to lysosomes, where they are degraded. Transport
KEY CONCEPTS of Section 14.5 vesicles from rhe cell surface first fuse with late endosomes,
Receptor-Mediated Endocytosis which subsequently fuse with the lysosome.
• Some extracellular ligands that bind to specific cell-surface Most receptor-ligand complexes dissociate in th e acidic
receptors are internalized, along with their receptors, in milieu of the late endosome; the receptors are recycled to the
clathrin-coated vesicles whose coats also contain AP2 com- plasma membrane, while the ligands are sorted to lysosomes
plexes (see Figure 14-26). (see Figure 14-29).

660 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 membrane by transport vesicles that bud from the trans-
• Iron is imported into cells by an endocytic pathway in Golgi network by the same basic mechanisms described in
which FeH ions are released from ferrotransferrin in the late earlier sections. In contrast, endocytosed membrane proteins
endosome. The receptor-apotransferrin complex is recycled such as receptor proteins that are to be degraded are trans-
to the cell surface, where the complex dissociates, releasing ferred in their entirety to the interior of the lysosome by a
both the receptor and apotransferrin for reuse. specialized delivery mechanism. Lysosomal degradation of
cell-surface receptors for extracellular signaling molecules is
a common mechanism for controlling the sensitivity of cells
to such signals (Chapter 15 ). Receptors that be1..ome Jam-

14.6 Directing Membrane Proteins aged also arc targeted for lysosomal degradation.
Early evidence that membranes can be delivered to the
and Cytosolic Materials to the Lysosome lumen of compartments came from electron micrographs
The major function of lysosomes is to degrade extracellular showing membrane vesicles and fragments of membranes
materials taken up by the cell and intracellular components within endosomes and lysosomes. Parallel experiments in
under certain conditions. Materials to be degraded must be yeast revealed that endocytosed receptor proteim targeted to
delivered to the lumen of the lysosome, where the various deg- the vacuole (the yeast organelle equivalent to the lysosome)
radative enzymes reside. As just discussed, endocytosed ligands were primarily associated w ith membrane fragments and
(e.g., LDL particles) that dissociate from their receptors in the small vesicles within the interior of the vacuole rather than
late endosome subsequently enter the lysosomal lumen when with the vacuole surface membrane.
the membrane of the late endosome fuses with the membrane These observations suggest that endocytosed membrane
of the lysosome (see Figure 14-29). Likewise, phagosomes car- proteins can be incorporated into specialized vesicles that form
rying bacteria or other particulate matter can fuse with lyso- at the endosomal membrane (figure 14-32). Although these
somes, releasing their contents into the lumen for degradation. vesicles are similar in size and appearance to transport vesicles,
It is apparent how the general vesicular trafficking mecha- they differ topologically. Transport vesicles bud outward from
nism discussed in this chapter can be used to deliver the lumi- the surface of a donor organelle into the cytosol, whereas vesi-
nal contents of an endosomal organelle to the lumen of the cles within the endosome bud inward from the surface into the
lysosome for degradation. H owever, membrane proteins de- lumen (away from the cytosol). Mature endosomes containing
livered to the lysosome by the typical vesicular trafficking pro- numerous vesicles in their interior are usually called multive-
cess we have discussed in this chapter should ultimately be stcular endosomes (or bodies). The surface membrane of a mul-
delivered to the membrane of the lysosome. How then are tivesicular endosome fuses with the membrane of a lysosome,
membrane proteins degraded by the lysosome? As we will see thereby delivering its internal vesicles and the membrane pro-
in this section, the cell has two different specialized pathways teins they contain into the lysosome interior for degradation.
for delivery of materials to the lysosomal lumen for degrada- Thus the sorting of proteins in the endosomal membrane deter-
tion, one for membrane proteins and one for cytosolic materi- mines which ones will remain on the lysosome surface (e.g.,
als. The first pathway, used to degrade endocytosed membrane pumps and transporters) and which ones will be incorporated
p roteins, utilize.s an unusua l type of vesicle that buds into the into internal vesicles and ultimately degraded in lysosomes.
lumen of the endosome to produce a multivesicular endo- Many of the proteins required for inward budding of the
some. The second pathway, known as au tophagy, involves the endosomal membrane were first identified by mutations in
de novo formation of a double membrane organelle known as yeast that blocked delivery of membrane proteins to the inte-
an autophagosome that envelops cytosolic material, such as rior of the vacuole. More than I 0 such "budding" proteins
soluble cyrosolic proteins or sometimes organelles such as per- have been identified in yeast, most with significant similarities
oxisomes or mitocho ndria. Both pathways lead to fusion of to mammalian proteins that evidently perform the same func-
either the multivesicular'endosome or autophagosome with tion in mamma lian cells. The current model of cndosomal
the lysosome, depositing the contents of these organelles into budding to form multivesicular cndosomes in mammalian
the lysosomal lumen fo r degradation. cells is based primarily on studies in yeast (Figure 14-33).
Most cargo proteins that enter the multivesicular endosome
are tagged with ubiquitin. Cargo proteins destined to enter the
Multivesicular Endosomes Segregate Membrane multivesicular endosome usua lly receive their ubiquitin tag at
Proteins Destined for the Lysosomal Membrane the plasma membrane, the TGN, or the endosomal mem-
brane. We have already seen how ubiquitin tagging can serve
from Proteins Destined for Lysosomal
as a signal for degradation of cytosolic or misfolded ER pro-
Degradation teins by the proteasomc (see Chapters 3 and 13). When used
Resident lysosomal proteins, such as V-class proton pumps as a signal for proteasomal degradation, the ubiquitin tag usu-
and amino acid transporters, can carry out their functions ally consists of a chain of covalently linked ubiquitin mole-
and remain in the lysosomal membrane, where they arc pro- cules (polyubiquitin), whereas ubiquitin used to tag proteins
tected from degradation by the soluble hyd rolytic enzymes in for entry into the multivesicular endosome usually takes the
the lumen. Such proteins are delivered to the lysosomal form of a single (monoubiquitin) molecule. In the membrane

14.6 Directing Membrane Proteins and Cytosolic Materials to the Lysosome 661
 Plasma membrane

-- Golgi
lysosome
late endosome/
multivesicular body

FIGURE 14-32 Delivery of plasma-membrane proteins to the multivesicular endosome contain ing many such internal vesicles
lysosomal i nterior for degradation. Early endosomes carrying (step 0 ). Fusion of a multivesicular endosome directly with a lysosome
endocytosed plasma-membrane proteins (blue) and vesicles carrying releases the internal vesicles into the lumen of the lys,osome, where
lysosomal membrane proteins (green) from the trans-Golgi network they can be degraded (step B ). Because proton pumps and other
fuse with the late endosome, transferring their membrane proteins to lysosomal membrane proteins normally are not incorporated into inter-
the endosomal membrane (steps D and f)). Proteins to be degraded, nal endosomal vesicles, they are delivered to the lysosomal membrane
such as those from the early endosome, are incorporated into vesicles and are protected from degradation. [See F. Reggiori and D. J. Klionsky, 2002,
that bud into the interior of the late endosome, eventually forming a Eukaryot. Cell 1:11, and D. J. Katzmann et al., 2002, Nature Rev. Mol. Cell Bioi. 3:893.]

of the endosome a ubiquitin-tagged peripheral membrane vesicle buds. Finally the ESCRT proteins pinch off the vesi-
protein, known as Hrs, facilitates recruitment of a set of three cle, releasing it and the specific membrane cargo proteins it .·
different protein complexes to the membrane. These ESCRT carries into the interior of the endosome. An ATPase, known
(endosomal sorting complexes required for transport) pro- as Vps4, uses the energy from ATP hydrolysis to disassemble
teins include the ubiquitin-binding protein TsglOl. The mem- the ESCRT proteins, releasing them into the cytosol for an-
brane-associated ESCRT proteins act to drive vesicle budding other round of budding. In the fusion event that pinches off
directed into the interior of the endosome as well as loading of a completed endosomal vesicle, the ESCRT proteins and
specific monoubiquitinated membrane cargo proteins into the Vps4 may function like SNAREs and NSF, respectively, in

Cytosol

FIGURE 14-33 Model of the mechanism Ubiquitin

for formation of multivesicular endosomes.
In endosomal budding, ubiquitinated Hrs on the
..
endosomal membrane directs loading of specific
membrane cargo proteins (blue) into vesicle
buds and then recruits cytosolic ESCRT proteins
to the membrane (step 0 ). Note that both Hrs
and the recruited cargo proteins are tagged with
ubiquitin. After the set of bound ESCRT com-
plexes mediate the completion and pinching
off of the inwardly budding vesicles (step f)),
they are disassembled by the ATPase Vps4 and
returned to the cytosol (step 0 ). See text for
discussion. [Adapted from 0. Pornillos et al., 2002,
Trends Cell Bioi. 1 2:569.]

662 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

the typical membrane-fusion process discussed previously binds to the plasma membrane of an infected cell and -4000
(see Figure 14-10). Gag molecules polymerize into a spherical shell, producing a
structure that looks like a vesicle bud protruding outward
Retroviruses Bud from the Plasma Membrane from the plasma membrane. Mutational studies with HIV
have revealed that the N-terminal segment of Gag protein is
by a Process Similar to Formation
required for association with the plasma membrane, whereas
of Multivesicular Endosomes the C-terminal segment is required for pinching off of com-
The vesicles that bud into the interior of endosomes have a plete HIV particles. For instance, if the portion of the viral
topology <>imilar to that of enveloped virus particles that bud genome encoding the C-termmus of Gag is removed, HIV
from the plasma membrane of virus-infected cells. Moreover, buds will form in infected cells, but pinching off does not
recent experiments demonstrate that a common set of pro- occur, and thus no free virus particles are released.
teins is required for both types of membrane-budding events. The first indication that HIV budding employs the same
In fact, the two processes so closely parallel each other in molecular machinery as vesicle budding into endosomes
mechanistic detail as to suggest that enveloped viruses have came from the observation that TsglO I, an ESCRT protein,
evolved mechanisms to recruit the cellular proteins used in binds to the C-terminus of Gag protein. Subsequent findings
inward endosomal budding for their own purposes. have clearly established the mechanistic parallels between
The human immunodeficiency virus (HIV) is an enveloped the two processes. For example, Gag is ubiquitinated as part
retrovirus that buds from the plasma membrane of infected of the process of virus budding, and in cells with mutations
cells in a process driven by viral Gag protein, the major struc- in Tsgl 01 or Vps4, H!V virus buds accumulate but cannot
tural component of completed virus particles. Gag protein pinch off from the membrane (Figure 14-34). Moreover,

@ PODCAST: HIV Budding from the Plasma Membrane

FIGURE 14-34 Mechanism for budd ing (a) HIV virus

of HIV f rom the plasma membrane. Proteins HIV envelope
required for formation of multivesicular '\.
endosomes are exploited by HIV for virus
budding from the plasma membrane.
(a) Budding of HIV particles from HIV-infected Extracellular
cells occurs by a similar mechanism as in Figure space
14-33, using the virally encoded Gag protein and
cellular ESCRT and Vps4 (steps 0 - IJ). Ubiqui- Plasma mem brane ,;;:.
tinated Gag near a budding particle functions
like Hrs. See text for discussion. (b) In wild-type
~;-~~~
cells infected with HIV, virus particles bud from HIV Gag ' \ Ubiquitin
the plasma membrane and are rapidly released protein
into the extracellular space. (c) In cells that lack
the functional ESCRT protein Tsg1 01, the viral
Gag protein forms dense viruslike structures, but
budding of these structures from the plasma
membrane cannot be completed and chains of
incomplete viral buds still attached to the
plasma membrane accumulate. [Wes Sundquist,
, (b) (c)
University of Utah.]

14.6 Directing Membrane Proteins and Cytosolic Materials to the Lysosome 663
when a segment from the cellular Hrs protein is added to a situation that occurs w hen the contents of multivesicular en-
truncated Gag protein by construction of the appropriate dosomes are delivered to the lysosome, lipases and proteascs
hybrid gene, proper budding and release of vi ru s particles is within the lysosome will degrade the autophagic vesicle and
restored. Taken together, these results indicate that Gag pro- its contents into thei r molecular components. Am ino acid
tein mimics the function of Hrs, redirecting ESCRT p roteins permeases in the lysosomal membrane then allow for trans-
to the plasma membrane, where they can function in the port of free amino acids hack into the cytosol for use in syn-
budding of virus particles. thesis of new proteins.
Other enveloped retroviruses such as murine leukemia By studying mutants defective in the autophagic path-
virus and Rous sarcoma vtrus also have been shown to require way, scientists have identified processes other than recycling
ESCRT complexes for their budding, although each virus ap- of cellular components during starvation that also depend
pears to have evolved a somewhat different mechanism tore- on autophagy. Experiments carried out principally in Dro-
cruit ESCRT complexes to the site of virus budding. sophila and mice have shown that autophagy participates in
a type of quality control that removes organelles that have
ceased to function properly. In particular, the autophagic
The Autophagic Pathway Delivers Cytosolic
pathway can target for destruction dysfunctional mitochon-
Proteins or Entire Organelles to lysosomes dria that have lost their integrity and no longer have an elec-
When cells arc placed under stress such as conditions of star- trochemical gradient across their inner membrane. In certain
vation, they have the capacity to recycle macromolecules for cell types, pathogenic bacteria and viruses that are multiply-
use as nutrients in a process of lysosomal degradation known ing in the cytosol of host cells can be targeted to the au-
as autophagy ("eating oneself"). The autophagic pathway tophagic pathway for destruction in the lysosome as part of
involves the formation of a flattened double-membrane cup- a host defense mechanism against infection.
shaped structure that envelops a region of the cytosol or an For each of these processes and in all eukaryotic organ-
entire organelle (e.g., mitochondrion), forming an autopha- isms the autophagic pathway takes place in three basic steps.
gosome, or autophagic vesicle (figure 14-35). The outer Although the underlying mechanisms for each of these steps
membrane of an autophagic vesicle can fuse with the lyso- are relatively poorly understood, they are thought to be re-
some, delivenng a large vesicle, bounded by a single mem- lated to the basic mechanisms for vesicular trafficking dis-
brane bilayer, to the interior of the lysosome. Similar to the cussed in this chapter.

Mitochondrion Atg12

. ""' .J.~ 1
Autophagic pathways Atg5, Atg 16 .,

,. ~ ~ Aut?phagic
• Mii'l.'f'llr." ves1cle
II ~ :'-.i.
El

Amino
acids

Lysosome
- \
FIGURE 14·35 The autophagic pathway. The autophagic pathway membrane of a lysosome releases a sinqle-layer vesicle and its contents
allows cytosolic proteins and organelles to be delivered to the into the lysosome interior (step f) ). After degradation of the protein
lysosomal interior for degradation. In the autophagic pathway, a and lipid components by hydrolases in the lysosome interior, the
cup-shaped structure forms around portions of the cytosol (right) or an released amino acids are transported across the lysosomal membrane
organelle such as a mitochondrion as shown here (left). Continued into the cytosol. Proteins known to participate in the autophagic
addition of membrane eventually leads to the formation of an pathway include Atg8, which forms a coat structure around the
autophagosome vesicle that envelops its contents by two complete autophagosome.
membranes (step OJ. Fusion of the outer membrane with the

664 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

Autophagic Vesicle Nucleation The autophagic vesicle is
thought to originate from a fragment of a membrane-bounded KEY CONCEPTS of Section 14.6
organelle. The origin of this membrane has been difficult to
Directing Membrane Proteins and Cytosolic
trace because no known integral membrane proteins, which
Materials to the Lysosome
might serve to identify the source of this membrane, are
known to be required for the formation of the autophagic • Endocytosed membrane proteins destined for degradation
vesicle. Studies in yeast have shown that some mutants defec- in the lysosome are incorporated into vesicles that bud into
tive in Golgi trafficking are also defective in autophagy, sug- the interior of the endosome. M ultivesicular endosomes,
gesting that the ~ utophagic vesicle is initially derived from a which contain many of these internal vesicles, can fuse with
fragment of the Golgi. Autophagy that is induced by starva- the lysosome to deliver the vesicles to the interior of the lyso-
tion appears to be a nonspecific process in which a random some (see Figure 14-32).
portion of the cytoplasm, including organelles, becomes envel- • Some of the cellular components (e.g., ESCRT) that medi
oped by an autophagosome. In these cases, the site of nucle- ate inward budding of endosomal membranes are used in the
ation is probably random. In cases in which defective budding and pinching off of enveloped viruses such as HIV
organelles are enveloped by the autophagosome, some type of from the plasma membrane of virus-infected cells (see Fig-
signal or binding site must be present on the surface of the ures 14-33 and 14-34 ).
organelle to target nucleation of rhe autophagic vesicle.
• A portion of the cytoplasm or an entire organelle (e.g., a
mitochondrion) can be enveloped in a flattened membrane
Autophagic Vesicle Growth and Completion New mem-
and eventually incorporated into a double-membrane au-
brane must be delivered to the autophagosome membrane in
tophagic vesicle. Fusion of the outer vesicle membrane with
order for this cup-shaped organelle to grow. This growth is
the lysosome delivers the enveloped contents to the interior
likely to occur by the fusion of transport vesicles with the
of the lysosome for degradation (see Figure 14-35).
membrane of the autophagosome. About 30 proteins that
participate in the formation of autophagosomes have been
identified in generic screens for yeast mutants rhat are defec-
tive in autophagy. One of these proteins is Atg8, shown in
Perspectives for the Future
Figure 14-35, which is covalently linked to the lipid phospha-
tidylethanolamine and thus becomes attached to the cyto- The biochemical, genetic, and structural information presented
plasmic leaflet of the autophagic vesicle. Association of Atg8 in this chapter shows that we now have a basic understanding
with a membrane vesicle appears to be the key step in en- of how protein traffic flows from one membrane-bounded eel
abling a vesicle to fuse with the growing autophagosome. lular compartment to another. Our understanding of these
Fusion of Atg8-containing vesicles with the autophago- processes has come largely from experiments on the function
some involves the formation of a cyrosolic assembly of Atg12, of various types of transport vesicles. These studies have led to
Atg5, and Atg16. Atg12 is similar in structure to ubiquitin, the identification of many vesicle components and the discov-
and a set of proteins related to ubiquitin-conjugating enzymes ery of how these components work together to drive vesicle
are responsible for covalently joining Atg12 to Atg5, by a budding, to incorporate the correct set of cargo molecules
process similar to that used for covalently joining ubiquitin to from the donor organelle, and then to mediate fusion of a com-
a target protein (see Figure 3-29). The covalently linked pleted vesicle with the membrane of a target organelle.
Atg12-Atg5 dimer then co-assembles with Atg16 to form a Despite these advances, there is still much to learn about
polymeric complex localized to the site of growing autopha- important stages of the secretory and endocytic pathways. For
gosomes. By an unknown mechanism this cytosolic complex example, we do not yet know what types of proteins form the
is thought to bring about the fusion of Atg8-containing vesi- coats of either the regulated or the constitutive secretory vesi-
cles into a cup-shaped autophagosome. cles that bud from the trans-Golgi network. In the same vein,
we do not know what feature of the Golgi membrane deter
Autophagic Vesicle Targeting and Fusion The outer mem- mines whether a COPI-coated vesicle or a clathrin/AP-coated
brane of the completed autophagosome is thought to con- vesicle will bud from it. In both cases, binding of ARF protein
tain a set of proteins that target fusion with the membrane of to the Golgi membrane appears to initiate vesicle budding.
the lysosome. Two vesicle-tethering proteins have been Moreover, the types of signals on cargo proteins that might
found to be required for autophagosome fusion with the target them for packaging into secretory vesicles have not yet
lysosome, but the corresponding SNARE proteins have not been defined.
been identified. Fusion of the autophagosome with the ly<;o- Another baffling process is the formation of vesicles that
some occurs after Atg8 has been released from the mem- bud away from the cytosol, such as the vesicles that enter
brane by proteolytic cleavage, and this proteolysis step only multivesicular endosomes. Although some of the proteins
occurs once the autophagic vesicle has completely formed a that participate in formation of these "internal" endosome
sealed double-membrane system. Thus Atg8 protein appears vesicles are known, we do not know what determines their
to mask fusion proteins and to prevent premature fusion of shape or what type of process causes them to pinch off from
the autophagosome with the lysosome. the donor membrane. Similarly, the origin and growth of the

Perspectives for the Future 665

membrane of the autophagic vesicle is also poorly under- 2. Vesicle budding is associated with coat proteins. What is
stood. In the future, it should be possible for these and other the role of coat proteins in vesicle budding? How are coat
poorly understood vesicle-trafficking steps to be dissected proteins recruited to membranes? What kinds of molecules
through the use of the same powerful combination of bio- are likely to be included or excluded from newly formed
chemical and genetic methods that have delineated the work- vesicles? What is the best-known example of a protein likely
ing parts of COPI, COPII, and clathrin/AP vesicles. to be involved in vesicle pinching off?
In addition to understanding the basic mechanisms that 3. Treatment of cells with the drug brefeldin A (BfA) has
direct the trafficking of cargo proteins in the secretory and the effect of decoating Golgi apparatus membranes, resulting
endocytic palltway~, a major goal of research in this area is to in a cell in which the vast majority of Golgi proteins arc
define all of the signals that direct proteins to specific intra- found in the ER. What inferences can be made from this
cellular locations. Although a number of such sequences are observation regarding roles of coat proteins other than pro-
known (sec Table 14-2), we are only beginning to compile a moting vesicle formation? Predict what type of mutation in
catalog of these targeting signals and the context in which Arfl might have the same effect as treating cells with BFA.
they are read. The ultimate goal will be to deduce, starting
4. Microinjection of an antibody known as EAGE, which
with just the primary coding sequence of any given gene, the
reacts with the "hinge" region of the 13 subunit of COPI,
trafficking pattern and intracellular location of the gene's
causes accumulation of Golgi enzymes in transport vesicles
protein product. Our ability to fully extract biological infor-
and inhibits anterograde transport' of newly synthesized ves-
mation from genomic sequences will be realized only when
icles from the ER to the plasma membrane. What effect does
we have the capability of reading trafficking information
the antibody have on COPI activity? Explain the results.
from primary protein sequences.
5. Specificity in fusion between vesicles involves two discrete
and sequential processes. Describe the first of the two processes
Key Terms and its regulation by GTPase switch proteins. What effect on
the size of early endosomes might result from Overexpression of
AP (adapter protein) multi vesicular
a mutant form of Rab5 that is stuck in the GTP-bound state?
complexes 646 endosomes 661
6. Sec18 is a yeast gene that encodes NSF. It is a class C mu-
anterograde transport 640 Rab proteins 638
tant in the yeast secretory pathway. What is the mechanistic
ARF protein 635 receptor-mediated role of NSF in membrane trafficking? As indicated by its
autophagy 661 endocytosis 655 class C phenotype, why does an NSF mutation produce ac-
cisternal maturation 629 regulated secretion 651 cumulation of vesicles at what appears to be only one stage
clathrin 634 retrograde of the secretory pathway?
constitutive secretion 651 transport 640 7. What feature of procollagen synthesis provided early evi-
COPI634 sec mutants 632 dence for the Golgi cisternal maturation model?
COPII 634 secretory 8. Sorting signals that cause retrograde transport of a pro-
pathway 627 tein in the secretory pathway are sometimes known as re-
dynamin 647
sorting signals 637 trieval sequences. List the two known examples of retrieval
endocytic pathway 627
transcytosis 654 sequences for soluble and membrane proteins of the ER.
ESCRT proteins 000 How does the presence of a retrieval sequence on a soluble
trans-Golgi network
late endosome 629 ER protein result in its retrieval from the cis-Golgi complex?
(TGN) 629
low-density Describe how the concept of a retrieval sequence is essential
transport vesicles 627
lipoprotein (LDL) 656 to the cisternal-maturation model.
t-SNAREs 634
mannose 6-phosphate 9. Clathrin adapter protein (AP) complexes bind directly to
(M6P) 648 v-SNAREs 634 the cytosolic face of membrane proteins and also interact
with clathrin. What are the four known adapter protein
complexes? What observation regarding AP3 suggests that
Review the Concepts clathrin is an accessory protein to a core coat composed of
adapter proteins?
1. The studies of Palade and colleagues used pulse-chase la-
beling with radioactively labeled amino acids and autoradi- 10. 1-cell disease is a classic example of an inherited human
ography to visualize the location of newly synthesized defect in protein targeting that affects an entire class of pro-
proteins in p:~ncreatic acinar cells. These early experiments teins, soluble enzyme~ of the lysosome. What is the molecu-
provided invaluable information on protein synthesis and in- lar defect in !-cell disease? Why does it affect the targeting of
tercompartmental transport. New methods have replaced an entire class of proteins? What other types of mutations
these early approaches, but two basic requirements are still might produce the same phenotype?
necessary for any assay to study this type of protein trans- 11. The TGN, trans-Golgi network, is the site of multiple
port. What are they and how do recent experimental ap- sorting processes as proteins and lipids exit the Golgi complex.
proaches meet these criteria? Compare and contrast the sorting of proteins to lysosomes

666 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 versus the packaging of proteins into regulated secretory gran- pared using yeast t-SNARE complexes: those containing
ules such as those containing insulin. Compare and contrast plasma membrane t-SNAREs, Golgi t-SNAREs, or vacuolar
the sorting of proteins to the basolateral versus apical cell sur- t-SNAREs. Each of these was mixed with fluorescent lipo-
faces in MDCK cells versus hepatocytes. somes containing one of three different yeast v-SNAREs. The
12. What does the budding of influenza virus and vesicular following data were obtained.
stomatitis virus (VSV ) from polarized MDCK cells revea l
about the sorting of newly synthesized cel l plasma mem-
brane proteins to the apical or basolateral domains? Now v-SNARE 1 v-SNARE 2 v-SNARE 3
comidcr the following result: A peptide with a ~equem:e
tSNARE =
identical to that of the VSV G protein cytoplasmic domain plasma
inhibits targeting of the G protein to the hasolateral surface ~ membrane
·v;
and has no effect on HA targeting to the apical membrane, c
Q)
but a peptide in which the si ngl e tyrosine residue is mutated c
to an alani ne has no effect on G protein basolateral target- Q)
<.>
ing. What does this tell you about the sorting process? c
Q)
tSNARE = Golgi
<.>
(/)
13. Describe how pH plays a key role in regulating the inter- ~
0
action between mannose 6-phosphate and the mannose :J

6-phosphate receptor. Why does elevating endosoma l pH u:

tSNARE =
lead to t he secretion of newly synthesized lysosomal enzymes vacuolar
into the extracellular medium?
14. What mechanistic features are shared by (a) the forma- Time after mixing
tion of multivesicular endosomes by budding into the inte-
rior of the endosome and (b) the outward budding of HIV
virus at the cell surface? You w ish to design a peptide in- a. What can be deduced from these data about the spec-
hibitor/competitor of HIV budding and decide to mimic in a ificity of membrane fusion events?
synthetic peptide a portion of the HIV Gag protein. Which b. Where might you expect to find v-SNAREs I, 2, and 3
portion of the HIV Gag protein would be a logica l choice? in yeast?
W hat normal cellular process might this inhibitor block? c. What kind of experiment could be designed to deter-
15. The phagocytic and autophagic pathways serve two fun- mine where in the secretory pathway a gi\Cn v-SNARE is
damental ro les, but both deliver their vesicles to the lyso- required in vivo?
some. What arc the fundamental differences between the d. The cytoplasmic domain of \-SNARE 2 has been ex-
two pathways? Describe the three basic steps in the forma - pressed and purified from E. coli. Various amounts of this
tion and fusion of autophagic vesicles. domain are incubated either with the Golgi t-SNARE lipo-
'· 16. Compare and contrast the location and pH sensitivity of somes or with v-SNARE 2 liposomes. The liposomes are then
receptor-ligand interaction in the LDL and transferrin receptor- washed free of unbound protein. The various liposomes are
mediated endocytosis (RME) pathways. then mixed, as indicated below, and the fluorescence of each
sample is measured 1 hour after mixing. How can the data be
.. 17. What do LDL receptor (LDLR ) cytoplasmic domain muta-
tions that cause familial hypercholesterolemia revea l about explained? What would you predict the outcome to be if yeast
the receptor-mediated endocytosis (RME) pathway? were to overexpress the cytoplasmic domain of v-SNARE 2?

~
·~ ~-------- = v-SNARE 2 liposomes incubated with the
Analyze the Data ' ~ ""..., cytoplasmic domain of v- SNARE 2 and
c ............ ..., then mixed with Golgi t·SNARE liposomes
1. In order to examine the specificity of membrane fusion con- Q) ........

ferred by specific v-SNAREs and t-SNAREs, researchers recon- g ........

~ ...,..., = Golgi t-SNARE liposomes incubated with
stituted liposomes (artificial lipid membranes) with specific ~ the cytoplasmic domain of v-SNARE 2 and
t-SNARE complexes or with v-SNAREs (see McNew et al., o
:J
then mixed with v·SNARE 2 liposomes

2000, Nature 407:1 53-159). To measure fusion, the v-SNARE u: ~------------------

. liposomes also contained a fluorescent lipid at a relatively Amount of v-SNARE 2
high concent ration such that its fluorescence is quenched. cytoplasmic domain added
(Quenching is reduced fluorescence relative to that expected.
In this case, quenching occurs because the fluorescent lipids 2. You have genetically engineered green fluorescent protein
are too concentrated and interfere with each other's ability to (GFP) containing a KDEL sequence. When the construct is
become excited.) On fusion of these liposomes with those transfected into normal human fibroblasts and examined
lacking the fluorescent lipid, the fluorescent lipids are diluted, using fluorescence microscopy, the fluorescence appears
and quenching is alleviated. Three sets of liposomes were pre- throughout the cytoplasm, as drawn below.

Analyze the Data 667

 skin fibroblasts, is compared to protein samples from fibro-
blasts of his healthy parents (lanes 1 and 2) and siblings
(lanes 4-6) using Western blot analysis and antibodies against
Nucleus N-acetylglucosamine phosphotransfcrase (-145 kDa ) and
.___..., «
actin (loading control, - 43 kDa ), the following is seen:
• :::::::::> *
~ RER ? ~~
cz,'
* q ,-r;
0~<?' TGN
q,,C$~
* l/~ ~
'/}
~

* * 0 * cz,v
~(f 2 3* 4 5 6

150 kDa
-- - - - - - -
--
a. How would you explain this pattern given that KDEL 102 kDa
is supposed to be an ER-specific sorting sequence?
86 kDa

- - - --- -
b. To analyze the results further, fractions of different
organelles and the cytoplasm were collected from cells 49 kDa
expressing this KDEL-containing GFP construct and then 31 kDa
examined on Western blots using antibodies against GFP (27
kDa) and protein disulfide isomerase (PDI), a resident-rough
ER (RER ) protein of approximately 55 kDa.
In a second set of experiments, N-acetylglucosamine phos-
photransferase was isolated from cells from the afflicted
child and from his healthy parents, and used in an assay with
12
P to measure enzyme activity and the production of man-
nose 6-phosphate. The assay yielded the following results:

90 kDa
-- 150

--
72 kDa 125

66 kDa

40 kDa

- - -
Q)
iO
..c.
a. 100

-
"'0
..c.~
a.E 0 Father
21 kDa ch-a,
Q) ::1. 75 X Child
"'c0
c D Mother
~

~ 50
The blot confirms the presence of GfP exclusively in the cy-
toplasm, and as expected a PDI signal in the RER fraction. 25
How would you explain the PDI band, albeit weak, in the
Golgi fraction? Given the function of PDI proteins, what
would you expect if both alleles of a PDI gene were knocked 0
0 5 15 30 60 120 240
out in mice? Time (min) at 37 •c
c. The antibodies used above do not detect any signals in
the nuclear fraction, which indicates their specificity or the
fact that no proteins were isolated and loaded in the nuclear a. Using fibroblasts cultured from the child, design an ex-
fraction lane. What antibody could be used to show there periment using the N-acctylglucosamine phosphotransferase
were nuclear proteins present in this sample? antibody and fluorescence microscopy and draw the results
3. A child appears to be suffering from I-cell disease, but that could explain why the child presents symptoms similar to
when a sample of his proteins (lane 3* below), isolated from 1-cell disease.

668 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

.·
 b. Given the results from these three different experi- Ostermann, J., er al. 1993. Stepwise assembly of functionally
ments, how woul<.l you explain the !-cell symptoms seen in active transport vesicles. Cell75: 1015-1025.
the child and what experiment would you propose to test Sch1mmoller, F., I. S1mon, and S. Pfeffer. 1998. Rab GTPases,
directors of vesicle docking.]. Bioi. Chem. 273:22161-22164.
your hypothesis?
Weber, T., et al. 1998. SNARI:.pins: mimmal machinery for
c. A laser scanning confocal micrograph of MDCK cells membrane fusion. Cell92:759-772.
labeled with an antibody against the mannose 6-phosphate Wickner, W., and A. Haas. 2000. Yeast homotypic vacuole
receptor shows the following: fus1on: a window on organelle trafficking mecham~ms. Ann. Rcu.
Bwchem. 69:247-275.
Zerial, M., and H ..\kBride. 2001. Rab proteins as membrane
Apical surface orgamzers. Nature Rev. Mol. Cell Btol. 2: I 07-117.

Early Stages of the Secretory Pathway

Barlowe, C.. 2003. S1gnals for COPII-dependent export from the
ER: what's the ticket our? Trends Cell Bioi. 13:295-300.
Behnia, R., and S. Munro. 2005. Organelle 1denmy and the
signposts for membrane traffic. Nature 438:597-604.
Bi, X., et al. 2002. Structure of the Sec23/24-Sarl pre-buddmg
complex of the COPII vesicle coat. Nature 419:271-2 7 -.
Gurkan, C., et al. 2006. The COPll cage: unifyrng principles of
vesicle coat assembl). Nat. Reu. Mol. Cell Btol. 7:727-7 38.
Lee, .\1. C., et al. 2004. B1-direcrional protein transport between
Bas a I surface the ER and Golgi. Ann. Rev. Cell Deu. B10l. 20:87-123.
Letourneur, r., er al. 1994. Coaromer IS essential for retrieval
of dilysine-tagged protem'> to the endoplasmic reticulum. Cell
How do you explain the labeling on the apical and basolat-
79:1199-1207.
eral surfaces, for a receptor whose function is to target en-
Lose,·, E., et al. 2006. Golgi maturation v1suali1ed in li' ing
zymes from the trans-Golgi network (TGN) to the lysosome? yeast. Nature 441: I 002-1006.
Likewise, what explains the labeling seen at the RER? Pelham, H. R. 1995. Sorting and retrieval between rhe endoplas-
mic renculum and Golgi apparatus. Czm. Opi11. Cell Bioi. 7:530-535.
Later Stages of the Secretory Pathway
Bonifacino, J. S. 2004. The GGA proteins: adaptors on the
References
move. Nat. Rev. Mol. Cell Bwl. 5:13-32.
Techniques for Studying the Secretory Pathway Bonifacino, J. S., and E. C. Dell'Angclica. 1999. Molecular
Beckers, C.]., er al. 1987 . Sem1-mracr cells permeable to hases for the recogmrion of tyrosine-based sorrmg signals. J. Cell
macromolecules: use in reconstitution of protein transport from the Bwl. 145:923-926.
endoplasmic renculum to the Golgi complex. Cell 50:523-534. Edehng, M. A., C. Smith, and D. Owen. 2006. Life of a clathrin
Kaiser, C. A., and R. Schekman. 1990. Disnnct sets of SEC coat: insights from clarhrin and AP structures. "Jat. Rev. VIol. Cell
genes govern transnorr vesicle formation and fu~ion early 111 the Btol. 7:32-44.
secretory pathway. Cel/61:723-733. Form, A., et al. 2004. :\lolecular model for a complete clathrin
Novick, P., ct al. 1981. Order of events in the yeast secretor} lawce from electron cryomicroscopy. Nature 432:573-579.
pathway. Cel/25:461-469. Ghosh, P., et al. 2003. Mannose 6-phosphare receptors: new
Lippincott-Schwartz, J., et ,1!. 2001. Studying protein dynam1cs twists 111 the tale. Nature Rev. Mol. Cell Bio. 4:202-21.3.
in living cells . .1\Jature Reu. Mol. Cell Bioi. 2:444-456. t-,.1osrov, K. E., M. Verges, andY. Altschuler. 2000 . .\1embrane
Orc1, L., er al. 1989. Dissection of a s1ngle round of vesicular traffic 1n polarized epirhel1al cells. Curr. Opin. Cell Bwl. 12:483-490.
transport: sequennal intermediates for inrercisrcrnal movement 111 Schmid, S. 1997. Clarhrin-coared vesicle formation and protem
the Golg1 stack. Cel/56:35 7-3'68. sorting: an integrated process. Ann. Rev. Biochem. 66:511-548.
Palade, G. 197.S. Intracellular aspects of the process of protein Simons, K., and E. Ikonen. 1997 . Functional rafts in cell
synthesis. Scw1tce 189:.347-358. membranes. Nature 387:.569-572.
Song, B. D., and S. L. Schm1d. 2003. A molecular motor or a
Molecular Mechanisms ofVesicle Budding and Fusion regulator? Dynamin's in a class of its own. Bmchemistrv 42: I 369-1376.
Sterner, D. F., et al. 1996. The role of prohorma'ne convertases
Bonifacino, .J. S., and B.S. Glick. 2004. The mechanisms of
in msulin biosynthesis: evidence for inherited defects in their action
vesicle budding and fusion. Cell116: 153-166.
in man and expenmcntal ammals. Dwbetes Metab. 22:94-104.
Grosshans, B. 1., D. Ortiz, and P. Novick. 2006. Rabs and their
Tooze, S. A., er al. 2001. Secretory granule biogenesi;: rafting to
effectors: achieving specificity in membrane traffic. Proc. Nat/.
the SNARE. Trc11ds Cell B10l. 11:116-122.
Acad. Set. USA 103:11821-1182-.
Jahn, R., er al. 200.3. Membrane fusion. Ce//112:519-5033.
Kirchhausen, T. 2000. Three ways to make a ves1cle. Nature Receptor-Mediated Endocytosis
Reu. Mol. Cell Bioi. 1:187-198. Brown, M. S., and J. 1.. Goldsrem. I 986. Receptor-mediated
.\llc;\!ew, J. A., er al. 2000. Compartmental speetficity of cellular pathwa> for chole>terol homeosta;is. Nobel Prrze Lecture. SCience
membrane fusion encoded in SNARE proteins. Nature 407:153-159. 232:34-4-.

References 669
 Kaksonen, VI., C. P. Toret, and D. G. Drubin. 2006. Harnessing !Jenne, W. M., N.J. Buchkovich, and S. D. Emr. 2011. The
acttn dynam1cs for clathnn-mediated endocytosis. Nat. Rev. Mol. ESC RT pathway. Dev. Cell11:77-9!.
Cell Bioi. 7:404-414. Katzmann, D. J., et a!. 2002. Receptor downregulation and
Rudenko, G., era!. 2002. Structure of the LDL receptor mulnvesicular-hody sorting. Nature Rev. Mol. Cell Bioi. 3:893-905.
extracellular domain at endosomal pH. Science 298:2353-2358. Lemmon, S. K., and L. .\1. Traub. 2000. Sorting in the endosomal
system in yeast and animal cells. Curr. Opm. Cell R10l. 12:457-466.
Directing Membrane Proteins and Cytosolic Materials Pornillos, 0., et al. 2002. Mechanisms of enveloped RNA v1rus
to the Lysosome budding. Trends Cell Biol.12:569-579.
Geng, J., and D. J. Klionsky. 2008. The Atg8 and Argl2 ubiquitin- Shinrani, T., and D. J. Klionsky. 2004. Autophagy in health and
likc conjugarion systems in maLwautuplmg}. EMBO Rep. 9:8.59-864. disease: a double-edged sword. Sctence 306:990-99 5.

..

670 CHAPTER 14 • Vesicular Traffic, Secretion, and Endocytosis

 CLASSIC EXPERIMENT 14.1

Following a Protein Out of the Cell

J. Jamieson and G. Palade, 1966, Proc. Nat/. Acad. Sci. USA 55(2):424-431

he advent of electron microscopy from the ER to the Golgi complex and of this organelle in protein secretion,
Tallowed researchers see the to ~.:ell then to the plasma membrane. Palade turned to in vitro pulse-chase
and its structures at an unprecedented experiments, which permitted more
level of detail. George Palade utilized precise monitoring of the fate of la-
this tool not only to look at the fine
The Experiment beled proteins. In this labeling tech-
details of the cell but also to analyze Paladc wanted to identify which cell nique, cells are exposed to radiolabeled
the process of secretion. By combining structures and organelles participate in precursor, in this case [3 Hj-leucine, for
electron microscopy with pulse-chase protein secretion. To study such a com- a short period known as the pulse. The
experiments, Palade uncovered the plex process, he carefully chose an ap- radioactive precursor is then replaced
path proteins follow to leave the cell. propriate model system for his studies, with its nonlabeled form for a subse-
the pancreatic exocrine cell, which is quent chase period. Proteins synthe-
responsible for producing and secreting sized during the pulse period will be
Background
large amounts of digestive enzymes. Be- labeled and detected by autoradiogra-
In addition to synthesizing proteins to cause these cells have the unusual prop- phy, whereas those synthesized during
carry out cellular functions, many cells erty of expressing only secretory the chase period, which are non labeled,
must also produce and secrete addi- proteins, a general label for newly syn- will not be detected. Palade began by
tional proteins that perform their du- thesized protein, such as radioactively cutting guinea pig pancreas into thick
ties outside the cell. Cell biologists, labeled leucine, will only be incorpo- slices, which were then incubated for 3
including Palade, wondered how se- rated into protein molecules that are minutes in media containing [~H)-leu­
creted proteins make their passage following the secretory pathway. cine. At the end of the pulse, he added
from the inside to the outside of the Palade first examined the protein se- excess unlabeled leucine. The tissue
cell. Early experiments suggesting that cretion pathway in vivo by injecting live slices were then either fixed for autora-
proteins destined for secretion are syn- guinea pigs with ['H)-leucine, which diography or used for cell fraction-
thesized in a particular intracellular lo- was incorporated into newly made pro- ation. To ensure that his results were
cation and then follow a pathway to teins, thereby radioactively labeling an accurate reflection of protein secre-
the cell surface employed methods to them. At time points from 4 minutes to tion in vivo, Palade meticulousl y char-
disrupt cells synthesizing a particular 15 hours, the animals were sacrificed, acterized the system. Once convinced
secreted protein and to separate their and the pancreatic tissue was fixed. By that his in vitro system accurately mim-
various organelles by centrifugation. subjecting the specimens to autoradiog- icked protein secretion in vivo, he pro-
These cell-fractionation studies showed raphy and viewing them in an electron ceeded to the critical experiment. He
that secreted proteins can be found in microscope, Palade could trace where pulse-labeled tissue slices with [' H)-
membrane-bounded vesicles derived the labeled proteins were in cells at vari- leucine for 3 minutes, then chased the
from the endoplasmic reticulum (ER), ous times. As expected, the radioactivity label for 7, 17, 37, 57, and 117 minutes
where they are synthesized, and within localized in vesicles at the ER at time with unlabeled leucine. Radioactivity,
zymogen granules, from which they are points immediately following the [JH]- again confined in vesicles, began at the
eventually released from' the cell. Un- Ieucine injection and at the plasma ER, then traveled in vesicles to the Golgi
fortunately, results from these studies membrane at the later time points. The complex and remained in the vesicles
were hard to interpret due to difficul- surprise came in the middle time points. as they passed through the Golgi and
ties in obtaining clean separation of all Rather than traveling straight from the onto the plasma membrane (see Fig-
of the different organelles that contain ER to the plasma membrane, the radio- ure 1 ). As the vesicles tra veled farther
secretory proteins. To further clarify actively labeled proteins appeared to along the pathway, they became more
the pathway, Palade turned to a newly stop off at the Golgi complex in the densely packed with radioactive pro-
developed technique, high-resolution middle of their journey. In addition, tein. From his remarkable series of
autoradiography, that allowed him to there never was a time point where the autoradiogram~ at different chase times,
detect the position of radioactively la- radioactively labeled proteins were not Palade concluded that secreted proteins
beled proteins in thin cell sections that confined to vesicles. travel in vesicles from the ER to the
had been prepared for electron micros- The observation that the Golgi Golgi and onto the plasma membrane
copy of intracellular organelles. His complex was involved in protein secre- and that throughout this process, they
work led to the seminal finding that se- tion was both surprising and intrigu- remain in vesicles and do not mix with
creted proteins travel within vesicles ing. To thoroughly address the role the rest of the cell.

Following a Protein Out of the Cell 671

FIGURE 1 The synthesis and movement of guinea pig pancreatic (a) (b)
secretory proteins as revealed by electron microscope autoradiog-
raphy. After a period of labeling with eHJ-Ieucine, the tissue is fixed,
sectioned for electron microscopy, and subjected to autoradiography.
The radioactive decay of [3 H) in newly synthesized proteins produces
autoradiographic grains in an emulsion placed over the cell section
(which appear in the micrograph as dense, wormlike granules) that
mark the position of newly synthesized proteins. (a) At the end of a
3-minute labeling period autoradiographic grains are over the rough
ER. (b) Following a 7-minute chase period with unlabeled leucine, most
of the labeled proteins have moved to the Golgi vesicles. (c) After a
37-minute chase, most of the proteins are over immature secretory
vesicles. (d) After a 11?-minute chase, the majority of the proteins are
over mature zymogen granules. [Courtesy of J. Jamieson and G. Palade.]

(c) (d)

Discussion
throughout the pathway. These find - during the labeling, obscuring the fate
Paladc's experiments gave biologists the ings were predicated from two impor- of secretory proteins in particular.
first clear look at the stages of the secre- tant aspects of the experimental design. Palade's work set the stage for
tory pathway. Hi's studies on pancreatic Palade's careful use of electron micros- more detailed studies. Once the secre-
exocrine cells yielded two fundamental copy and autoradiography allowed him tory pathway was clearly described,
observations. First, that secreted pro- to look at the fine detai ls of the path- entire fields of research were opened
teins pass through the Golgi complex way. Of equal importance was th e up to investigation of the synthesis and
on their way out of the cell. This was choice of a cell type devoted to secre- movement of both secreted and mem-
the first function assigned to the Golgi tion, the pancreatic exocrine cell, as a brane proteins. for this groundbreak-
complex. Second, that secreted proteins model system. In a different cell type, ing work, Paladc was awarded t he
never mix with cellular proteins in the significant amounts of nonsecrcted pro- Nobel Prize for Physiology and Medi-
cytosol; they arc segregated into vesicles teins would have also been produced cine in 19 74.

·.
672 CHAPTER 14 • Vesicu lar Traffic, Secret ion, and Endocytosis
 CHAPTER

·.·

SIGNAL TRANSDUCTION
AND G PROTEIN-
COUPLED RECEPTORS

The mouse retina contains ph otoreceptors (purple) that sense light

using G protein-coupled receptors and four other types of neurons
stained yellow, green, pink, and blue, which connect the photoreceptor
cells to the brain. [Rachel Wong, University of Washington]

o cell lives in isolation . Cellular communication is a some receptors, this signal is a physical stimulus such as

N fundamental property of all cells and shapes the devel-

opment and function of every living organism. Even
single-celled eukaryotic microorganisms, such as yeasts, slime
light, touch, or heat. For others, it is a chemical molecule.
Many types of chemicals are used as signals: small molecules
(e.g., amino acid or lipid derivatives, acetylcholine), gases
molds, and protozoans, communicate through extracellular (nitric oxide), peptides (e.g., ACTH and vasopressin), solu-
signals: secreted molecules called p heromones coordinate the ble proteins (e.g., insulin and growth hormone), and pro-
aggregation of free-living cells for sexual mating or differentia- teins that are tethered to the surface of a cell or bound to the
tion under certain environmental conditions. More important extracellular mat rix. Many of these extracellular signaling
in plants and animals are hormones and other extracellular molecules are synthesized and released by specialized signal-
signaling molecules that function within an organism to con- ing cells within multicellular organisms. Most receptors bind
trol a variety of processes, including the metabolism of sugars, a single molecule or a group of closely related molecules.
fats, and amino acids; the growth and differentiation of tissues; Some signaling molecules, especially hydrophobic mole-
the synthesis and secretion of proteins; and the composition of cules such as steroids, retinoids, and thyroxine, spontane-
intracellular and extracellular fluids. Many types of cells also ously diffuse through the plasma membrane and bind to
respond to signals from the external environment, including intracellular receptors; signaling from such intracellular re-
light, oxygen, odorants, anJi tastants in food. ceptors is discussed in detail in Chapter 7.
In any system, for a signal to have an effect on a target, Most extracellular signaling molecules, however, are too
it has to be received. In cells, a signal produce~ a specific re- large and too hydrophilic to penetrate through the plasma
sponse only in target cells with receptors for that signal. For membrane. How, then, can they affect intracellular processes?

OUTLINE

15.1 Signal Transduction: From Extracellular 15.4 G Protein- Coupled Receptors That Reg ulate
Signal to Cellular Response 675 lon Channels 693

1 5.2 Studying Cell-Surface Receptors and Signal 15.5 G Protein- Coupled Receptors That Activate
Transduction Proteins 681 or Inhibit Adenylyl Cyclase 699

15.3 G Protein-Coupled Receptors: Structure 15.6 G Protein-Coupled Receptors That Trigger

and Mechanism 687 Elevations in Cytosolic Ca2+ 707
0 OVERVIEW ANIMATION: Extracellular Signaling
FIGURE 1 5-1 Overview of signaling by cell-surface receptors.
Communication by extracellular signals usually involves the following
steps: synthesis of the signaling molecule by the signaling cell and its
incorporation into small intracellular vesicles (step Ol. its release into the
extracellular space by exocytosis (step 8), and transport of the signal to
the target cell (step iJ). Binding of the signaling molecule to a specific
cell-surface receptor protein triggers a conformational change in the 0
receptor, thus activating it (step B). The activated receptor then activates
one or more downstream signal transduction proteins or small-molecule ~!r'';"• :o~o o o o /mo
~
second messengers (step B), eventually leading to activation of one or
more effector proteins (step mJ.The end result of a signaling cascade can . 0 /_ • Active cell-surface
lnact1ve cell-surface .,. receptor
be either a short-term change in cellular function, metabolism, or
receptor
movement (step m or a long-term change in gene expression or

~~~~
development (step fm). Termination or down-modulation of the cellular

{(J
response is caused by negative feedback from intracellular signaling
molecules (step [i)) and by removal of the extracellular signal (step m).
Responding
cell 1
t
These signaling molecules bind to cell-surface receptors that
are integral membrane proteins embedded in the plasma mem-
Signal tra nsduction
proteins and second
messengers
Q
brane. Cell-surface receptors generally consist of three discrete
domains, or segments: an extracellular domain facing the
m~
extracellular fluid, a membrane-spanning (transmembrane)
domain that spans the plasma membrane, and the intracellular
Modification of
ce llular metabolism,
fl! 0
___ ,. ',
Effector protein
. -<- -- \
funct1on, movement ~
domain segment facing the cytosol. The signaling molecule I
acts as a ligand, which binds to a structurally complementary I
I

site on the extracellular or the membrane-spanning domains of

the receptor. Binding of the ligand to its site on the receptor
flD '',• M 0 d·t· .
I !CatiOn 0
f
Nucleus gene expression,
induces a conformational change in the receptor that is trans- development
mitted through the membrane-spanning domain to the cyto-
solic domain, resulting in binding to and subsequent activation
or inhibition of other proteins in the cytosol or attached to the
plasma membr~ne. In many cases, these activated proteins 900 G protein-coupled receptors, including receptors in the
catalyze the synthesis of certain small molecules or change the visual, olfactory (smell), and gustatory (taste) systems, many
concentration of an intracellular ion such as Ca2 • These intra- neurotransmitter receptors, and most of the receptors for
cellular proteins or small molecule second messengers then hormones that control carbohydrate, amino acid, and fat
carry the signal to one or more effector proteins. The overall metabolism and even behavior. Signal transduction through
process of converting extracellular signals into intracellular re- GPCRs usua ll y induces short-term changes in cell function,
sponses, as well as the individual steps in this process, is termed such as a change in metabolism or movement. In contrast,
signal transduction (Figure 15-1). activation of other cell-surface receptors primarily alters a
In eukaryotes, there are about a dozen classes of cell- cell's pattern of gene expression, leading to cell differentia-
surface receptors, which activate several types of intracellu- tion or division and other long-term consequences. These
lar signal transduction pathways. Our knowledge of signal latter receptors and the intracellular signaling pathways they
transduction has advanced greatly in recent years, in large activate are explored in Chapter 16.
measure because these receptors and pathways are highly In this chapter, we first review some general principles of
conserved and function in essentially the same way in organ- signal transduction, such as the molecular basis for ligand-
tsms as diverse as worms, flies, mice, and humans. Genetic receptor binding, and certain evolutionarily conserved com-
studies combined with biochemical analyses have enabled ponents of signal transduction pathways. Next we describe
researchers to trace many entire signaling pathways from how cell-surface receptors and signal transduction proteins
binding of ligand to final cellular responsr. are identified and characterized biochemically. We then turn
Perhaps the most numerous class of receptors-found in to an in-depth discussion of G protein-coupled receptors,
organisms from yeasts to humans-are G protein-coupled focusing first on their structure and mechanism of action
receptors (GPCRs). As their name implies, G protein-coupled and then on the signaling pathways activated by them. We
receptors consist of an integral membrane receptor protein show how these pathways affect many aspects of cell func-
coupled to an intracellular G protein that transmits signals tion, including glucose metabolism, muscle contraction, per-
to the interior of the cell. The human genome encodes about ception of light, and gene expression.

674 CHAPTER 15 • Signal Transduction and G Protein - Coupled Receptors

 15.1 Signal Transduction: From ln paracrine signaling, the signaling molecules released by
a cell affect only those target cells in close proximity. A nerve
Extracellular Signal to Cellular Response cell releasing a neurotransmitter (e.g., acetylcholine) that acts
As shown in Figure 15-1, signal transduction begins when on an atljacent nerve cell or on a muscle cell (inducmg or in-
extracellular signaling molecules bind to cell-surface recep- hibiting muscle contraction) is an example of paracrine sig-
tors. Binding of signaling molecules to their receptors induces naling. In addition to neurotransmitters, many protein
two major types of cellular responses: ( l) changes in the ac- growth factors regulating development in multicellular or-
tivity or function of specific enzymes and other proteins that ganisms act at short range. Some of these growth-factor pro-
preexist in rhe cell and (2) changes in the amounts of specific teins bind tightly to components of the extracellular matrix
proteins produced by a cell, most commonly by modification
of transcription factors that stimulate or repress gene expres-
sion (see Figure 15-1, steps fl1 and fl!l). In general, the first (a) Endocrine signaling
type of response occurs more rapidly than the second. Tran-
scription factors activated in the cytosol by these pathways
0 0 0
move into the nucleus, where they stimulate (or occasionally 0 () () 0
repress) transcription of specific target genes.
The connection between an activated receptor and a cel-
/ oJo\
lular response is not direct and generally involves several in-
Hormone secretion
0 0 0
termediate proteins or small molecules. Collectively, this into blood by endocrine gland Distant target cells
chain of intermediates is called a signal transduction path-
way because it transduces, or converts, information from
one form into another as a signal is relayed from a receptor (b) Paracrine signaling
to its targets. Some signal transduction pathways contain
just two or three intermediates; others can involve over a
0
dozen. Regardless, most pathways contain members of cer-
tain classes of signal transduction proteins that have been
o oQ
0
0
highly conserved throughout evolution.
In this section, we provide an overview of the major steps
in signal transduction, starting with the signaling molecules Secretory cell Adjacent target cell
( themselves. We explore the molecular basis for ligand-receptor
binding and the chain of events initiated in the target cell by
binding of the signal to its receptor, focusing on a few compo- (c) Autocrine signaling
nents that are central to many signal transduction pathways. Key:

o Extracellular signal

Signaling Molecules Can Act Locally Receptor

or at a Distance
T Membrane-attached
Cells respond to many different types of signals-some orig- signal
inating from outside the organism, some internally gener-
ated. Those that are generated internally can be described by Target sites on same cell
how they reach their target. Some signaling molecules are
transported long distancflS by the blood; others have more
(d) Signaling by plasma-membrane-attached proteins
local effects. In animals, signaling by extracellular molecules
can be classified into three types-endocrine, paracrine, or
autocrine-based on the distance over which the signal acts
(Figure 15-2a-c). In addition, certain membrane-bound pro-
teins on one cell can directly signal an adjacent cell.
\)
In endocrine signaling, the signaling molecules are synthe-
sized and secreted by signaling cells (for example, those found
in endocrine glands), transported through the circulatory sys- Signaling cell Adjacent target cell
tem of the organism, and tinally act on target cells Jistant FIGURE 1 S-2 Types of extracellular signaling. (a-c) Cell-to-cell
from their site of synthesis. The term hormone generally re- signaling by extracellular chemicals occurs over distances from a
fers to signaling molecules that mediate endocrine signaling. few micrometers in autocrine and paracrine signaling to several
Insulin secreted by the pancreas and epinephrine secreted by meters in endocrine signaling. (d) Proteins attached to the plasma
the adrenal glands are examples of hormones that travel membrane of one cell can interact directly with cell-surface receptors
through the blood and thus mediate endocrine signaling. on adjacent cells.

15.1 Signal Transduction: From Extracellular Signal to Cellular Response 675

and are unable to signal to adjacent cells; subsequent degra- can bind to receptors on an adjacent cell. In addition, cleavage
dation of these matrix components, triggered by injury or by an extracellular protease releases a soluble form of EGF,
infection, will release the active growth factor and enable it to which can signal in either an autocrine or a paracrine manner.
signal. Many developmentally important signaling proteins
diffuse away from the signaling cell, forming a concentration
Binding of Signaling Molecules Activates
gradient and inducing different cellular responses depending
on the concentration of the signaling protein. Receptors on Target Cells
In autocrine signaling, cells respond to substances that Receptor proteins for all hydrophilic extracellular small mole-
they themselves release. Some growth factors act in this fash- cules and protein ~ignaling molecules a re located on the sur-
ion, and cultured cells often secrete growth factors that stim- face of the target cell. The signaling molecule, or ligand, binds
ulate their own growth and proliferation. This type of to a site on the extracellular domain of the receptor with high
signaling is particularly characteristic of tumor cells, many specificity and affinity. Each receptor generally binds o nly a
of which overproduce and release growth factors that stimu- single signaling molecule or a group of structurally very closely
late inappropriate, unregulated self-proliferation, a process related molecules. The binding specificity of a receptor refers
that may lead to formation of a tumor. to its abi lity to bind or not bind closely related substances.
Integral membrane proteins located on the cell surface Ligand binding depends on weak, multiple noncovalent
also play an important role in signaling (Figure 15-2d). In forces (i.e., ionic, van der Waals, and hydrophobic interac-
some cases, such membrane-bound signals on one cell bind tions) and molecular complementarity between the interacting
receptors on the surface of an adjacent target cell, triggering surfaces of a receptor and ligand (see Figure 2-12). For exam-
its differentiation. In other cases, proteolytic cleavage of a ple, the growth hormone receptor (Figure 15-3) binds to
membrane-bound signaling protein releases the extracellular growth hormone but not to other hormones with very similar,
segment, which functions as a soluble signaling molecule. though not identical, structures. Similarly, acetylcholine recep-
Some signaling molecules can act at both short and long tors bind only this small molecule and not others that differ
ranges. For example, epinephrine (also known as adrena line) only slightly in chemical structure, while the insulin receptor
functions as a systemic hormone (endocrine signaling) and as binds insulin and related hormones called insulin-like growth
a neurotransmitter (paracrine signaling). Another example is factors 1 and 2 (TGF-1 and IGF-2), but no other hormones.
epidermal growth factor (EGF), which is synthesized as an Binding of ligand to receptor causes a conformational
integral plasma membrane protein. Membrane-bound EGF change in the receptor that initiates a sequence of reactions

(a) (b) (c)

Growth
hormone

.··

EXPERIMENTAL FIGURE 15-3 Growth hormone binds to its folded protein. Similar studies showed that two tryptophan residues
receptor through molecular complementary. (a) As determined from (blue) in the receptor contribute most of the energy responsible for tight
the three-dimensional structure of the growth hormone-growth binding of growth hormone, although other amino acids at the interface
hormone receptor complex, 28 amino acids in the hormont' dre at the with the hormone (yellow) are also important. (b) Binding of growth
binding interface with one receptor. To determine which amino acids are hormone to one receptor molecule is followed by (c) binding of a second
important in ligand-receptor binding, researchers mutated each of these receptor (purple) to the opposing side of the hormone; this involves the
amino acids one at a time, to alanine, and measured the effect on same set of yellow and blue amino acids on the receptor but different
receptor binding. From this study, it was found that only eight amino residues on the hormone. As we see in the next chapter, such hormone-
acids on growth hormone (pink) contribute 85 percent of the energy induced receptor dimerization is a common mechanism for activation of
that is responsible for tight receptor binding; these amino acids are receptors for protein hormones. [After B. Cunningham and J. Wells, 1993,
distant from each other in the primary sequence but adjacent in the J. Mol. Bioi. 234:554, and T. C1ackson and J. Wells, 1995, Science 267:383.]

676 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

leading to a specific response inside the cell. Organisms have 0
evolved to be able to use a single ligand to stimulate different I
0 -P=O
cells to respond in distinct ways. For example, different cell I
0
types may have different sets of receptors for the same ligand,
each of which induces a different intracellular signal response
pathway. Alternatively, the same receptor can be found on
various cell types in an organism, but binding of a particular
ligand to the receptor triggers a different response in each type
of cell, given the unique complement of proteins expressed by
the cell. In these ways, the same ligand can induce different
cells to respond in a variety of ways. This is what's known as kinase V
the effector specificity of the receptor-ligand complex.
For instance, the surfaces of skeletal muscle cells, heart
Protein~
v
( \ Protein
phosphatast

muscle cells, and the pancreatic acinar cells that produce hy-
drolytic digestive enzymes each have different types of recep-
tors for acetylcholine. In a skeletal muscle cell, release of
acetylcholine from a motor neuron innervating the cell triggers
muscle contraction by activating an acetylcholine-gated ion Inactive
channel. In heart muscle, the release of acetylcholine by certain
FIGURE 15-4 Regulation of protein activity by a kinase/
neurons activates a G protein-coupled receptor and slows the
phosphatase switch. The cyclic phosphorylation and dephosphorylation
rate of contraction and thus the heart rate. Acetylcholine stim- of a protein is a common cellular mechanism for regulating protein
ulation of pancreatic acinar cells triggers a rise in cytosolic activity. In this example, the target, or substrate, protein is inactive
lCa2 J that induces exocytosis of the digestive enzymes stored (light green) when not phosphorylated and active (dark green)
in secretory granules to facilitate digestion of a meal. Thus for- when phosphorylated; some proteins have the opposite pattern. Both
mation of different acetylcholine-receptor complexes in differ- the protein kinase and the phosphatase act only on .specific target
ent cell types leads to different cellular responses. proteins, and their activities are usually highly regulated.

Protein Kinases and Phosphatases Are Employed

leading to activation of the appended kinase. The kinase
in Virtually All Signaling Pathways then phosphorylates the monomeric, inactive form of a spe-
Activation of virtually all cell-surface receptors leads directly cific transcription factor, leading to its dimerization and
or indirectly to changes in protein phosphorylation through movement from the cytosol into the nucleus, where it acti-
the activation of protein kinases, which add phosphate vates transcription of target genes. A phosphatase in the
groups to specific residues of specific target proteins. Some nucleus subsequently removes the phosphate group from the
receptors activate protein phosphatases, which remove phos- transcription factor, causing it to form two inactive mono-
phate groups frqm specific residues on target proteins. Phos- mers and then move back into the cytosol, where it can be
phatases act in concert with kinases to switch the function of reactivated by a receptor-associated kinase.
various proteins on or off (Figure 15-4). As this example illustrates, the activity of all protein ki-
At last count, the human genome encodes about 600 pro- nases is opposed by the activity of protein phosphatases,
tein kinases and 100 different phosphatases. In general, each some of which are themselves regulated by extracellular stg-
protein kinase phosphorylates specific amino acid residues in nals. Thus the activity of a protein in a cell can be a complex
a set of target, or substrate, proteins whose patterns of expres- function of the activities of the usually multiple kinases and
sion generally differ in different cell types. Animal cells con- phosphatases that act on it, either directly or indirectly
tain two types of protein kinases: those that add phosphate to through phosphorylation of another protein. Several exam-
the hydroxyl group on tyrosine residues and those that add ples of this phenomenon occur in regulation of the cell cycle
phosphate to the hydroxyl group on serine or threonine (or and are described in Chapter 19.
both) residues. All kinases also bind to specific amino acid Many proteins are substrates for multiple kinases, each of
sequences surrounding the phosphorylated residue, and thus which phosphorylates different amino acids. Each phosphory-
one can look at amino acid sequences surrounding tyrosine, lation event can modify the activity of a particular target pro-
serine, and threonine residues in a protein and make a good tein in different ways, some activating its function, others
guess as to which kinases might phosphorylate this re~.;idue. inhibiting it. An example we encounter later is glycogen pho~­
In some signaling pathways, the receptor itself possesses phorylase kinase, a key regulatory enzyme in glucose metabo-
intrinsic kinase activity or the receptor is tightly bound to a lism. ln many cases, addition of a phosphate group to an amino
cytosolic kinase. Figure 15-5 illustrates a simple signal trans- acid creates a binding surface that allows a second protein to
duction pathway involving one kinase tightly bound to a bind; in the following chapter we will encounter many exam-
receptor and one predominant target protein. In the absence of ples of such kinase-driven assembly of multiprotein complexes.
a bound ligand the kinase is held in the inactive state. Ligand Commonly the catalytic activity of a protein kinase itself
binding triggers a conformational change in the receptor, is modulated by phosphorylation by other kinases, by the

15.1 Signal Transduction: From Extracellular Signal to Cellular Response 677

FIGURE 15-5 A simple signal transduction Ligand-
Ligand binding sites Bound ligand
pathway involving one kinase and one target
protein. The receptor is tightly bound to a
protein kinase that, in the absence of a bound
ligand, is held in the inactive state. Ligand
binding triggers a conformational change in the
receptor, leading to activation of the appended
Cytosotic
kinase (0 ). The kinase then phosphorylates the inactive
monomeric, inactive form of a specific transcrip- non phosphorylated
tion factor (f)), leading to its dimerization (IJ) transcription factor
and movement from the cytosol into the
nucleus (m ). where it activates transcription
Act~<:
oftarget genes. A phosphatase in the nucleus
will remove the phosphate group from the
transcription factor (if). causing it to form the
Cytosol
Inactive
kinase \f)
kinase

inactive monomer and then move back into the

m ,;?)
cytosol (1";1).

Dimeriza~ion Active
~
~ \1 phosphorylated
transcription factor

Into nucleus; / .
~
binds DNA
and activates
transcription
m
N uclear
phosphatase

binding of other proteins to it, and by changes in the levels mediated by a GTPase, which slowly hydrolyzes the bound
of various small intracellular signaling molecules and me- GTP to GDP and P., thus altering the conformation of the
tabolites. The resulting cascades of kinase activity are a com- switch I and switch II segments so that they are unable to
mon feature of many signaling pathways. bind to the effector protein. The GTPase can be an intrinsic
part of the G protein or a separate protein.
The rate of GTP hydrolysis reg ulates the length of time
GTP-Binding Proteins Are Frequently Used
the switch protein remains in the active conformation and is
in Signal Transduction as On/Off Switches
Many signal transduction pathways utilize intracellular
"switch" proteins that turn downstream proteins on or off. Active ("on")
The most important group of intracellular switch proteins is
the GTPase super fa m ily. All the GTPase switch proteins
exist in two forms (Figure 15-6): (1 ) an active ("on") form
with bound GTP (guanosine triphosphate) that modulates
the activity of specific target proteins and (2) an inactive
("off") form with bound GDP (guanosine diphosphate).
Conversion of the inactive to active state is triggered by a
signal (e.g., a hormone binding to a receptor) and is mediated
by a guanine nucleotide exchange factor (GEF), which causes
release of GDP from the switch protein. Subsequent binding p
of GTP, favored by its high intracellular concentration rela- FIGURE 15-6 GTPase switch proteins cycle between active and
tive to its binding affinity, induces a conformational change inactive forms. The switch protein is active when it has bound GTP
to the active form. The principal conformational changes in- and inactive when it has bound GOP. Conversion of the active into the
volve two highly conserved segments of the protein, termed inactive form by hydrolysis of the bound GTP is accelerated by GAPs
switch I and switch II, that allow t he protein to bind to and (GTPase-accelerating proteins) and other proteins. Reactivation is
activate other downstream signaling p roteins (figure 15-7). promoted by GEFs (guanine nucleotide exchange factors) t hat catalyze
Conversion of the active form back to the inactive state is the dissociation of the bound GOP and its replacement by GTP.

678 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 FIGURE 15-7 Switching mechanism of G

l
(a) GTP-bound "on" state (b) GOP-bound "off" state
proteins. The ability of a G protein to interact with
other proteins and thus transduce a signal differs
Gly-60 Thr-35
in the GTP-bound "on" state and GOP-bound "off"
I state. (a) In the active "on" state, two domains,
I
Switch II termed switch I (green) and switch II (blue), are
bound to the terminal gamma phosphate of GTP
through interactions with the backbone amide
groups of a conserved threonine and glycine
y residue. When bound to GTP in this way, lh~ two
GOP switch domains are in a conformation such that
they can bind to and thus activate specific
downstream effector proteins. (b) Release of the
gamma phosphate by GTPase-catalyzed hydrolysis
causes switch I and switch II to relax into a
different conformation, the inactive "off" state; in
this state they are unable to bind to effector
proteins. The ribbon models shown here represent
both conformations of Ras, a monomeric G
protein. A similar spring-loaded mechanism
switches the alpha subunit in trimeric G proteins
between the active and inactive conformations by
movement of three switch segments. [Adapted from
I. Vetter and A. Wittinghofer, 2001. SCience 294:1299.]

able to signal its downstream target proteins: the slower the signaling molecules termed second messengers. These, in turn,
rate of GTP hydrolysis, the longer the protein rema ins in the bind to other proteins, modifying their activity.
active state. The rate of GTP hydrolysis is often modulated One second messenger used in virtually all metazoan
by other proteins. for instance, both GTPase-activating pro- cells is Ca 1 + ions. We noted in Chapter 11 that the concen-
tems (GAP) and regulator of G protein signaling (RGS) pro- tration of free Ca 2 + in the cytosol is kept very low (< 10 "M)
teins accelerate GTP hydrolysis. Many regulators of G protein by ATP-powered pumps that continually transport Ca 2 out
activity are themselves controlled by extracellular signals. of the cell or into the endoplasmic reticulum (ER). The cyto-
Two large classes of GTPase switch proteins are used in solic Ca2+ level can increase from 10- to 100-fold by a signal-
signaling. Trimeric (large) G proteins directly bind to and are induced release of Ca 1 from ER stores or by its import
activated by cettain cell-surface receptors. As we will see in through calcium channels from the extracellular environment;
Section 15.3, G protein-coupled receptors function as guanine this change can be detected by fluorescent dyes introduced
nucleotide-exchange factors (GEFs )-triggering release of into the cell (see Figure 9-11 ). In muscle, a signal-induced rise
GDP and binding of GTP, thus activating the G protein. in cytosolic Cah triggers contraction (see Figure 17-35) . In
Monomeric (small) G proteins, such as Ras and various Ras- endocrine cells, a similar increase in Ca 2 + induces exocytosis
like proteins, are not bound to receptors but play crucial roles of secretory vesicles containing hormones, which are thus
in many pathways that re.gulate cell division and cell motility, released into the circulation. In nerve cells, an increase in
as is evidenced by the fact that mutations in genes encoding cytosolic Ca2+ leads to the exocytosis of neurotransmitter-
these G proteins frequently lead to cancer. Other members of conta ining vesicles (see Chapter 22). In all cells, this rise in
both GTPase classes, by switching between GTP-bound "on" cytosolic Ca2+ is sensed by Ca2+ -binding proteins, particu-
and GDP-bound "off" forms, function in protein synthesis, the larly those of the EF hand family, such as calmodulin, all of
transport of proteins between the nucleus and the cytoplasm, which contain the helix-loop-helix motif (see Figure 3-9b).
the formation of coated vesicles and their fusion with target The binding of Ca 1 to calmodulin and other EF hand pro-
membranes, and rearrangements of the actin cytoskeleton. teins causes a conformationa l change that permits the pro-
tein to bind various target proteins, thereby switchrng their
activities on or off (see Figure 3-31 ).
Intracellular "Second Messengers" Transmit
Another nearly universal second messenger is cyclic AMP
and Amplify Signals from Many Receptors (cAMP). In many eukaryotic cells, a rise in cAMP triggers
The binding of ligands ("first messengers") to many cell-surface activation of a particular protei n kinase, protein kinase A,
receptors leads to a short-lived increase (or decrease) in the that in turn phosphorylates specific target proteins to induce
concentration of certain low-molecular-weight intracellular specific changes in cell metabolism. In some cells, cAMP

1S.1 Signal Transduction: From Extracellular Signal to Cellular Response 679

0 FOCUS ANIMATION: Second Messengers in Signaling Pathways

NH 2 0

N~N t ~NH
~NH
2
1 PO3

~
5 N__. ' N-:7 N---lN I o;03 2
0 1;>'), O- CH2

~0~1
2 CH3 -(CH2 )n - C

CH3 -{CH 2 )n
0
c
0 CH 2

- CH
2
1
OH
6

. '

O=P--0
r-i·
OH O=P - -0
H' OH Fatty acyl groups
0
3
'h---3-r
OH HO/ opo 2
3

CH 20H
0 0 Glyc: rol Inositol
3',5'-CyclicAMP 3',5 '-Cyclic GMP 1,2-Diacylglycerol 1.4,5 -trisphosphate
(cAMP) (cGMP) (DAG) (IP3 )

Activates protein kinase A {PKA) Activates protein kinase G {PKG) Activates protein kinase C Opens Ca2+ channels in
and opens cation channels in {PKC) the endoplasmic reticu lu m
rod cells
FIGURE 15-8 Four common intracellular second messengers. The structural formula. Calcium ions (Ca2+) a~d several membrane-bound
major direct effect or effects of each compound are indicated below its phosphatidylinositol derivatives also act as second messengers.

regulates the activity of certain ion channels. The structures number of hormones to the ava ilable rece ptors often require
of cAMP and three other common second messengers are production of tens o r hundreds of thousand's of activated ef-
shown in Figure 15-8. Later in this chapter, we examine the fector molecules per cell. In the case of G protein-coupled
specific roles of second messengers in signa ling pathways ac- hormone receptors, signal amplification is possible in part be-
tivated by various G protein-cou pled receptors. cause a single receptor can activa te multiple G proteins, each
Because second messengers such as Ca 2 and cAMP dif- of w hich in turn activates an effector protein. For example, a
fuse through the cytosol much faster than do proteins, they single epinephrine-GPCR complex causes activation of up to
arc employed in pathways where the d ownstream ta rget is 100 adenylyl cyclase molecu les, each o f wh ich in turn cata-
located in an intracellular organelle (such as a secreto ry ves- lyzes syn thesis of many cAMP molecules during t he time it
icle or the nucleus) dista nt from the plasma membra ne re- remains in the active state. Two cAMP molecules activate one
ceptor where the messenger is generated. molecule of protein kinase A that in turn p h osph oryl ate~ and
Another advantage of second messengers is that they fa- activates multiple target product molecules (Figure 15-9).
cilitate amplification of an extracellular signal. Activation of a Later in th is chapter, we see how th is amplificati on cascade
single cell-surface receptor molecule can result in an increase allows blood levels of epinephrine as low as I o-to M to stim-
in perhaps thousand!> of cAMP molecules o r Ca2 ions in the ulate glycogenolysis (conversion of glycogen to glucose) by the
cytosol. Each of these, in turn, by activating its target protein liver and release of glucose into the blood.
affects the activity of multiple downstream proteins. In many
signal transduction pathways, amplification is necessary be-
cause cell surface receptors are typica lly low-abundance pro-
teins, present in only a thousand or so copies per cell. Yet the {1o -1o M)
• Epinephrine
~t --....
cellular responses induced by the bindi ng of a relatively small Amplification
~
.6 A /.:::,. Adenylyl
cycla se

FIGURE 15-9 Amplification of an extracellular signal. In this

example, binding of a sing le epinephrine molecule to one G protein-
Amplification

••
~1~
•• •• •• •• cAMP {10- 6 M)
coupled receptor molecule induces activation of several molecules of
adenylyl cyclase, the enzyme that catalyzes the synthesis of cyclic AMP,
and each of these enzyme synthesizes a large number of cAMP
l l l l l
0 0 0 0 0
Protein
kinase A

~1~
molecules, the first level of amplification. Two molecules of cAMP
activate one molecule of protein kinase A (PKA), but each activated Amplification
PKA phosphorylates and activates multip le target proteins. This second
level of amplification may involve severa l sequential reactions in which • • • • 0
Activated
enzyme
the product of one reaction activates the enzyme catalyzing the next
reaction. The more steps in such a cascade, the greater the signal
amplification possible.
Amplification

•
~1~
• • • • Product

680 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

 activity of signal transduction components, such as kinases
KEY CONCEPTS of Section 15.1 and GTP-binding "switch" proteins.
Signal Transduction: From Extracellular Signal
to Cellular Response
The Dissociation Constant Is a Measure
• All cells communicate through extracellular signals. In
of the Affinity of a Receptor for Its Ligand
unicellular organisms, extracellular signaling molecules reg-
ulate interactions between individuals, while in multicellular Ligand binding to a receptor usually can be viewed as a sim-
organisms, they regulate physiology and development. ple reversible reaction, where the receptor is represented as
R, the ligand as L, and the receptor-ligand complex as RL:
• External signals include membrane-anchored and secreted
proteins or peptides (e.g., vasopressin and insulin), small hy-
drophobic molecules (e.g., steroid hormones and thyroxine),
small hydrophilic molecules (e.g., epinephrine), gases (e.g., R+L RL (15-1)
0 2, nitric oxide), and physical stimuli (e.g., light).
• Binding of extracellular signaling molecules to cell-surface koff is the rate constant for dissociation of a ligand from its re-
receptors triggers a conformational change in the receptor, ceptor, and konis the rate constant for formation of a receptor-
which in turn leads to activation of intracellular signal trans- ligand complex from free ligand and receptor.
duction pathways that ultimately modulate cellular metabo- At equilibrium, the rate of formation of the receptor-ligand
lism, function, or gene expression (see Figure 15-1 ). complex is equal to the rate of its dissociation and can be de-
.··· • Signals from one cell act on distant cells in endocrine sig- scribed by the simple equilibrium-binding equation
naling, o n nearby cells in paracrine signaling, or on the sig-
naling cell itself in autocrine signaling (see Figure 15-2). [R LJ
K - -
• Protein phosphorylation and de-phosphorylation, cata-
u- RLl- (15-2)

lyzed by protein kinases and phosphatases, are employed in

virtually all signaling pathways. The activities of kinases and where [R] and [L] are the concentrations of free receptor
phosphatases are highly regulated by many receptors and (that is, receptor without bound ligand) and ligand, respec-
signal transduction proteins (see Figures 15-4 and 15-5). tively, at equilibrium, and [RL] is the concentration of the
receptor-ligand complex. Kd, the dissociation constant, is a
• GTP-binding proteins of the GTPase superfamil y act as
measure of the affinity (or tightness of binding) of the recep-
switches regulating many signal transduction pathways (see
tor for its ligand (see also Chapter 2 ). For a simple binding
Figures 15-6 and 15-7).
reaction, Kd = korrlkon· The lower k off is relative to ko"' the
• Ca 2 + , cAMP, and other nonprotein, low-molecular-weight more stable the RL complex-the tighter the binding-and
intracellular molecules (see Figure 15-8) act as "second mes- thus the lower the value of Kd. Another way of seeing this
sengers," relaying and often amplifying the signa l of the key point is that Ku equals the concentration of ligand at
.· "first messenger," that is, the ligand. Binding of ligand to which half of the receptors have a ligand bound when the
cell-surface receptors often results in a rapid increase (or, system is at equilibrium; at this ligand concentration [R] =
occasionally, decrease) in the intracellular concentration of [RL] and thus, from Equation 15-2, Kd = [L]. The lower the
these ions or molecules. Kd, the lower the ligand concentration required to bind 50
percent of the cell-surface receptors. The Kd for a binding
reaction here is essentially equivalent to the Michaelis con-
stant Km, which reflects the affinity of an enzyme for its sub-
' strate (see Chapter 3). Like all equilibrium constants,
15.2 Studying Cell-Surface Receptors however, the value of Ku does not depend on the absolute
values of koff and k0 n, only on their ratio. In the next section,
and Signal Transduction Proteins
we learn how Kd values are experimentally determined.
The response of a cell to an external signal depends on the
cell's complement of receptors that recognize the signal and Hormone receptors are characterized by their high af-
the signal transduction pathways activated by those recep- finity and specificity for their ligands. Because of their
tors. In this section, we explore the biochemical basis for the high affinity and great specificity for their target hormone,
specificity of receptor-ligand binding, as well as the ability of the extracellular, ligand-binding domains of cell c;urface re-
different concentrations of hgand to activate a pathway. We ceptors can be converted into powerful drugs. Consider the
also examine experimental techniques used to characterize hormone tumor necrosis factor alpha (TNFa ), which is se-
receptor proteins. Many of these methods are also applicable creted by a number of immune system cells. TNfa induces
to receptors that mediate endocytosis (see Chapter 14) or inflammation by recruiting various immune cells to a site of
cell adhesion (see Chapter 20). We conclude the section with inj ury or infection; abnorma l levels of TN Fa cause the exces-
a discussion of techniques commonly used to measure the sive inflammation seen in patients with autoimmune diseases

15.2 Studying Cell-Surface Receptors and Signal Transduction Proteins 681

such as the hlistering skin disease psoriasis or the joint disease 5.0
rheumatoid arthritis. These diseases are being treated with a 4.5
chimeric "fusion" protein, generated by recombinant DNA,
that contains the extracellular domain of a TNFa. receptor Iii 4.0
fused to the constant (Fe) region of a human immunoglohin
(see Figures 3-19 and 23-8). The drug binds tightly to free
Qi
u
<0
0
3.5 --------------------c Specific
(i;
binding
TNFa. and prevents it from binding to its cell-surface recep- 3.0
C1l

tors and causing inflammation; the fused Fe domain causes 0

2.5
the protein to be stable when injected into the body. • ~
8. 2.0
LU
"0
Binding Assays Are Used to Detect Receptors c
~
1.5
B
and Determine Their Affinity and Specificity 0
OJ 1.0
for ligands 0.5
Usually receptors are detected and measured hy their ahil ity
to bind radioactive or fluorescent ligands to intact cells or to 2 4 6
cell fragments. Figure 15-10 illustrates such a binding assay [Free 125 1-EIIlol (nM)
for interaction of the red-cell-forming hormone erythropoie- EXPERIMENTAL FIGURE 15-10 Binding assays can determine
tin (Epo) with Epo receptors that are expressed by recombi- the Kd and the number of receptors per cell. Shown here are data for
nant DNA in a line of cultured cells. The amounts of erythropoietin-specific receptors on the surface of a cultured mouse
radioactive Epo bound to its receptor on growi ng cells (verti- cell line that expresses a recombinant human erythropoietin (Epo)
cal axis) were measured as a function of increasing concen- receptor compared to control cells that do not normally express the
tration of 12 'I-Iabeled Epo added to the extracellular fluid receptor. A suspension of cells is incubated for 1 hour at 4 °( with
12
(horizontal axis). Both the number of ligand-binding sites per increasing concentrations of sl-labeled Epo; the low temperature is
cell and the Kd va lue are easily determined from the specific used to prevent endocytosis of the cell-surface receptors. The cells are
hinding curve (curve C). Assuming each receptor hinds just separated from unbound 12sl Epo, usually by centrifugation, and the
amount of radioactivity bound to them is measured. The total binding
one ligand molecule, the total number of ligand-binding sites
curve A represents Epo specifically bound to high-affinity receptors as
on a cell equals the number of active receptors per cell. In the
well as Epo nonspecifically bound with low affinity to other molecules
example shown in Figure 15-10, the value of Kd is about
on the cell surface. The contribution of nonspecific binding to total
1.1 X 1 0 10 M, or 0.1 nM. In other words, an Epo concen- binding is determined by repeating the binding assay with the control
tration of 1.1 X 10 10 M in the extracellular fluid is required cell line, where Epo binds only to nonspecific sites, yielding curve B.
for 50 percent of a cell's Epo receptors to have a bound Epo. The specific binding curve Cis calculated as the difference between
Direct binding assays like the one in rigure 15-10 are curves A and B. As determined by the maximum of the specific binding
feasible with receptors that have a high affinity for their li- curve C, the number of specific Epa-binding sites (surface receptors)
gands, such as t.he erythropoietin receptor and t he insulin per cell is about 2200 (3.7 X 10 1s moles X 6.02 X 1023 molecules/
receptor on liver cells (Kd = 1.4 X 10 10 M). However, mole/106 cells = 2227 molecules/cell). The Kd is the concentration of
many ligands, such as epinephrine and other catecholamines, Epo required to bind to 50 percent of the surface Epo receptors (in this
hind to their receptors with much lower affinity. If the Kd for case about 1050 receptors/cell). Thus the Kd is about 1.1 X 10 10 M, or
binding is greater than -1 X 10 ., M, a case when the rate 0.1 nM. (Courtesy Alec Gross; after A. Gross and H. Lodish, 2006, J. Bioi. Chern.
constant koff is relatively large compared to k""' then it is 281:2024.]
likely that during the seconds to minutes required to mea-
sure the amount of bound ligand, some of the receptor-
hound ligand will dissociate and thus the observed binding of the high-affinity ligand hound in this assay because little
values will be systematically too low. dissociates during the experimental manipulations required
One way to measure relatively weak binding of a ligand for the measurement (relatively low kotfl·
to its receptor is in a comfJetition assay with another ligand
that binds to the same receptor with high affinity (low Kd B Competitive binding is often used tO study synthetic ana-
value). In this type of assay, increasing amounts of an unla- U logs of natural hormones that activate or inhibit recep-
beled, low-affinity ligand (the competitor) arc added to a cell tors. These analogs, which are widely used in research on
sample with a constant amount of the radio labe led, high- cell-surface receptors and as drugs, fall into two classes: ago-
affinity ligand (Figure 15-11 ). Binding of unlabeled competi- nists, wh ich mimic the function of a natural hormone by bind
tor to the receptor blocks binding of the radioactive ligand ing to its receptor and inducing the normal response, and
to the receptor. The concentration dependence of this com- antagonists, which bind to the receptor but induce no response.
petition can be used together with the Kd value of the radio- By occupying ligand-binding sites on a receptor, an antagonist
active ligand to calculate the inhibitory constant, K., which can block binding of the natural hormone (or agonist) and
is very close to the Kd value for binding of the competitor to thus reduce the usual physiological activity of the hormone. In
the receptor. It is possible to accurately measure the amount other words, antagonists inhibit receptor signaling. •

682 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 100

~
Ol
c 80
:0
c
:0
0 60
0c
....Q)
0.

0"' 40
c
0
·.= CH 2 OH CH
:0
.r= 20 II
CH
I +
I 3
E I O-CH2-CH-CH2-NH2-CH

CD2~ ~H 3

Competitor concentration (M) ~

I Alprenol ol (API

EXPERIMENTAL FIGURE 15· 11 For low-affinity ligands, the inhibition of [3H) alprenolol binding versus epinephrine or
binding can be detected in competition assays. In this example, the isoproterenol concentration, such as shown here, the concentration of
synthetic ligand alprenolol, which binds with high affinity to the the competitor that inhibits alprenolol binding by SO percent approxi-
epinephrine receptor on liver cells (Kd- 3 x 10 9 M), is used to detect mates the Kd value for competitor binding. Note that the concentra-
the binding of two low-affinity ligands, the natural hormone epineph- tions of competitors are plotted on a logarithmic scale. The Kd for
rine (EP) and a synthetic ligand called isoproterenol (IP). Assays are binding of epinephrine to its receptor on liver cells is only- 5 X 10 5 M
performed as described in Figure 15-10 but in reactions containing a and would not be measurable by a direct binding assay with eHJ
constant amount of [3 H) alprenolol and increasing amounts of epinephrine. The Kd for binding of isoproterenol, which induces the
unlabeled epinephrine or isoproterenol. At each competitor concentra- normal cellular response, is more than tenfold lower.
tion, the amount of bound labeled alprenolol is determined. In a plot of

Consider for instance the drug isoproterenol, used to treat level of that molecule in the extracellular fluids or blood. We
asthma. Isoproterenol is made by the chemical addition of can see this principle in practice by comparing the levels of
two methyl groups to epinephrine (see Figure 15-ll, right). insulin present in the body and the Kd for binding of insulin
Isoproterenol, an agonist of the epinephrine-responsive G to its receptor on liver cells, 1 .4 X l 0 10 M. Suppose, for
protein-coupled receptors on bronchial smooth muscle cells, instance, that the norma l concentration of insulin in the
binds about tenfold more strongly (tenfold lower Kd) than blood is 5 X 10 12 M. By substituting this value and the in-
does epinephrine"(see Figure 15-11, left). Because activation of sulin K..1 into Equation 15-2, we can calculate the fraction of
these receptors promotes relaxation of bronchial smooth mus- insulin receptors with bound insulin
cle and thus opening of the air passages in the lungs, isopro-
terenol is used in treating bronchial asthma, chronic bronchitis, RL/( RL + Rj)
and emphysema. ln contrast, activation of a different type of
epinephrine-responsive G protein-coupled receptors on car- at equilibrium as 0.0344; that is, about 3 percent of the total
diac muscle cells (called [3-,adrenergic receptors) increases the insulin receptors will be bound with insulin. If the insulin
heart contraction rate. Antagonists of this receptor, such as concentration rises fivefold to 2.5 X 10 11 M, the number of
alprenolol and related compounds, are referred to as beta- receptor-hormone complexes will rise proportionately, al-
blockers; such antagonists are used to slow heart contractions most fivefold, so that about 15 percent of the total receptors
in the treatment of cardiac arrhythmias and angina. will have bound insulin. If the extent of the induced cellular
response parallels the number of insulin-receptor complexes,
Maximal Cellular Response to a Signaling [RLj, as is often the case, then the cellular responses also will
increase by abou t fivefold.
Molecule Usually Does Not Require
On the other hand, suppose that the normal concentra-
Activation of All Receptors tion of insulin in the blood were the same as the Kd value of
All signaling systems evolved such that a rise in the level of 1.4 X 10 10 M; in this case, 50 percent of the total receptors
extracellular signaling molecules induces a proportional re- would have a bound insulin. A fivefold increase in the insu-
sponse in the responding cell. For this to happen, the binding lin concentration to 7 X 10- 10 M would result in 83 percent
affinity (Kd value) of a cell-surface receptor for a signaling of all insulin receptors having insulin bound (a 66 percent
molecule must be greater than the norma l (unstimulated) increase). Thus, in order for a rise in hormone concentration

15.2 Studying Cell -Su rface Receptors and Signal Transduction Proteins 683
 Cl
typical erythroid progenitor cell. The Kd for binding of
c 1.0 Physiological response
erythropoietin (Epo) to its receptor is about 10- 10 M. As we
~Q)
·- (/) noted above, only 10 percent of the -1000 erythropoietin
~ § 0.8
roo. receptors on the surface of a cell must be bound to ligand to
-~ ~ induce the maximal cellular response. We can determine the

·; 0
u
~ ~ 0.6
E~

0~
c u
0

~
:J

~
0.4

0.2
:yt:
:
I
1

I
I
Ligand concentration
for 50% physiological response
ligand concentration, [L], needed to indllce the maximal re-
sponse by rewriting Equation 15-2 as follows:

LL
1
I
1/ Kd
I

I
for ligand binding (15-3)

0 1 2 3 4
Relative concentration of ligand

EXPERIMENTAL FIGURE 15-12 The maximal physiological where RT = [R] + [RL], the total number of receptors per
response to an external signal occurs when only a fraction of the cell. If the total number of Epo receptors per cell, RT, is
receptors are occupied by ligand. For signaling pathways that exhibit 1000, Kd is 10 10 M, and [RL] is 100 (the number of Epa-
this behavior, plots of the extent of ligand binding to the receptor and occupied receptors needed to indoce the maximal response),
of physiological response at different ligand concentrations differ. In then an Epo concentration ([L]) of 1.1 X 10-ll M will elicit
the example shown here, SO percent of the maximal physiological the maximal response. If the total number of Epo receptors
response is induced at a ligand concentration at which only 18 percent (RT) is reduced to 200 per cell, then a ninefold-higher Epo
of the receptors are occupied. Likewise, 80 percent of the maximal
concentration (10 10 M) is required to occupy 100 receptors
response is induced when the ligand concentration equals the Kd value,
and induce the maximal response. Clearly, therefore, a cell's
at which SO percent of the receptors are occupied.
sensitivity to a signaling molecule is heav,ily influenced by
the number of receptors for that ligand that are present as
to cause a proportional increase in the fraction of receptors well as the Kd.
with bound ligand, the normal concentration of the hor-
mone must be well below the Kd value. A Epithelial growth factor (EGF), as its name implies,
In general, the maximal cellular response to a particular 11..11 stimulates the proliferation of many types of epithelial
ligand is induced when much less than 100 percent of its re- cells, including those that line the ducts of the mammary
ceptors are bound to the ligand. This phenomenon can be gland. In about 25 percent of breast cancers, the tumor cells
revealed by determining the extent of the response and of produce elevated levels of one particular EGF receptor called
receptor-ligand binding at different concentrations of ligand HER2. The overproduction of HER2 makes the cells hyper-
(Figure 15-12). For example, a typical red blood (erythroid) sensitive to ambient levels of EGF that normally are too low
progenitor cell has - 1000 surface receptors for erythropoi- to stimulate cell proliferation; as a consequence, growth of
etin, the protein hormone that induces these cells to prolifer- these tumor cells is inappropriately stimulated by EGF. We
ate and differentiate into red blood cells. Because only 100 will see in Chapter 16 that an understanding of the role of
of these receptors need to bind erythropoietin to induce divi- HER2 in certain breast cancers led to development of mono-
sion of a progenitor cell, the ligand concentration needed to clonal antibodies that bind HER2 and thereby block signal-
induce 50 percent of the maximal cellular response is pro- ing by EGF; these antibodies have proved useful in treatment
portionally lower than the Kd value for binding. In such of these breast cancer patients. •
cases, a plot of the percentage of maximal binding versus li-
gand concentration differs from a plot of the percentage of The HER2-breast cancer connection vividly demon-
maximal cellular response versus ligand concentration. strates that regulation of the number of receptors for a given
signaling molecule expressed by a cell plays a key role in di-
Sensitivity of a Cell to External Signals recting physiological and developmental events. Such regula-
tion can occur at the levels of transcription, translation, and
Is Determined by the Number of Surface
post-translational processing or by controlling the rate of re-
Receptors and Their Affinity for Ligand ceptor degradation. Alternatively, endocytosis of receptors
Because the cellular response to a particular signaling mole- on the cell surface can sufficiently reduce the number present
cule depends on the number of receptor-ligand complexes, the such that the cellular response is term ina ted. As we discuss
fewer receptors present on the surface of a cell, the less sensi- in later sections, other mechanisms can reduce a receptor's
tive the cell is to that ligand. As a consequence, a higher ligand affinity for ligand and so reduce the cell's response to a given
concentration is necessary to induce the physiological re- concentration of ligand. Thus reduction in a cell's sensitivity
sponse than would be the case if more receptors were present. to a particular ligand, called desensitization, can result from
To illustrate the important relationship between receptor various mechanisms and is critica l to the ability of cells to
number and ligand sensitivity, let's extend our example of a respond appropriately to external signals.

684 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 Receptors Can Be Purified is highly regulated and can phosphorylate many target pro-
by Affinity Techniques teins. Immunoprecipitation assays are frequently used to
measure the activity of a particular kinase in a cell extract. In
In order to fully understand how receptors function, it IS one version of the method, an antibody specific for the de-
necessary to purify them and analyze their biochemical sired kinase is first reacted with small beads coated with Pro-
properties. Determining their molecular structures with and tein A; this causes the antibody to bind to the beads via its Fe
without a bound ligand, for instance, can elucidate the con- segment (see Figure 9-29). The beads are then mixed with a
formational changes that occur on ligand binding that acti- preparation of cell cytosol or nucleus, then recovered by cen-
vate downstream signal transduction proteins. But this can trifugation and washed extensively with a salt solution to
be challenging. A "typical" mammalian cell has 1000 to remove weakly bound proteins that arc unlikely to be bind-
50,000 copies of a single type of cell-surface receptor. This ing specifically to the antibody. Thus only cell proteins that
may seem like a large number, but when you consider that specifically bind to the antibody-the kinase itself and pro-
this same cell contains -10 10 total protein molecules and teins tightly bound to the kinase-are present on the beads.
-106 proteins in the plasma membrane alone, you realize The beads are then incubated in a buffered solution with a
that these receptors constitute only 0.1 to 5 percent of substrate protein and -y-e2 Pl ATP, where only the-y phos-
plasma-membrane proteins. This low abundance compli- phate is labeled. The amount of [uP] transferred to the sub-
cates the isolation and purification of cell-surface receptors. strate protein is a measure of kinase activity and can be
Purification of receptors is also difficult because these inte- quantified either by polyacrylamide gel electrophoresis fol-
gral membrane proteins first must be solubilized from the lowed by autoradiography (see Figure 3-36) or by tmmuno-
membrane with a non-ionic detergent (see Figure 10-23) and precipitation with an antibody specific for the substrate
then separated from other cellular proteins. followed by counting the radioactivity in the immunopre-
As we saw with the Epo receptor discussed earlier, recom- cipitate. By comparing extracts from cells before and after
binant DNA techn iques can be used to generate cells that ex- ligand addition, for example, one can readily determine
press large amounts of these proteins. But even when whether or not a particular kinase is activated in the signal
recombinant DNA techniques are used to generate cells that transduction pathway triggered by that ligand.
express receptors in large amounts, special techniques arc nec- We noted that many proteins can be phosphorylated by
essary to isolate and purify them from other membrane pro- ~everal different kinases, usually on different serine, threo-
teins . One technique often used in purifying cell-surface nine, or tyrosine residues. Thus it is important to measure
receptors that retain their ligand-binding ability when solubi- the extent of phosphorylation of a single amino acid side
lized by detergents is similar to affinity chromatography using chain in a specific protein, say before and after hormone
antibodies (see Figure 3-38c). To purify a receptor by this tech- stimulation. Antibodies play a crucial role in detecting such
nique, a ligand for the receptor of interest, rather than an anti- phosphorylation events. To generate an antibody that can
body, is chemicall y linked to the beads used to form a column. recognize a specific phosphorylated amino acid in a specific
A crude, detergent-solubilized preparation of membrane pro- protein, one first chemically synthesizes an approximately
·. teins is passed through the column; only the receptor binds, 15 amino acid peptide that has the amino acid sequence sur-
while other proteins are washed away. Passage of an excess of rounding the phosphorylated amino acid of the specific pro-
the soluble ligand through the column causes the bound re- tein but where a phosphate group has been chemically linked
ceptor to be displaced from the beads and eluted from the to the desired serine, threonine, or tyrosine. After coupling
column. In some cases, a receptor can be purified as much as this peptide to an adjuvant to increase its immunogeniciry, it
100,000-fold in a single affinity-chromatographic step. is used to generate a set of monoclonal antibodies (see Figure
9-6). One then selects a particular monoclonal antibody that
lmmunoprecipitation Assays and Affinity reacts only with the phosphorylated, but nor the nonphos-
phorylated peptide; such an antibody generally will bind to
Techniques Can Be Used to Study the Activity
the parent protein only when this specific amino acid is
of Signal Transduction Proteins phosphorylated. This specificity is possible because the anti-
Following ligand binding, receptors activate one or more sig- body binds simultaneously to the phosphorylated amino
na l transduction proteins that, in turn, can affect the activity acid and to side chains of adjacent amino acids. As an example
of multiple effector proteins (see Figure 15-1 ); to understand of the use of such antibodies, Figure 15-13 shows that three
a signaling cascade requires the researcher to be able to signal transduction proteins in red-cell progenitors become
quantify the activity of these signal transduction proteins. phosphorylated on specific amino acid residues within 10
Kinases and GTP-binding proteins are found in many signal- minutes of stimulation by varying concentrations of the hor-
ing cascades, and in this section we describe several assays mone erythropoietin; phosphorylation increases with Epo
used for measuring their activities. concentration and is the first step in triggering the differen-
tiation of these cells into red blood cells.
lmmunoprecipitation of Kinases Kinases function in virtu-
ally all signaling pathways, and typical mammalian cells Pulldown Assays of GTP-Binding Proteins We've seen that
contain a hundred or more different kinases, each of which the GTPase superfamily of intracellular-switch proteins cycle

15.2 Studying Cell-Surface Receptors and Signal Transduction Proteins 685

 which it binds only when it has a bound GTP; the target pro-
Epo (U/ml) tein usua lly has a specific binding domain that binds to the
switch segments of the GTP-binding protein. Pull-down as-
anti-® Stat5 says used to quantify the activation of a specific GTP-bindi ng
protein are similar to immunoprecipi tations except that the

I•. Ill • •• anti-Stat5

specific binding domain of the target protein is immobilized
o n small beads (Figure 15-14). T he beads are mixed with a
cell extract and then recovered by centrifugatio n; the amount
of the GTP-binding protein on the bea J ~ i~ quantified by
anti-®Akt Western blotting. The example in Figure 15-14 shows tha t

....., __ ......,,
I--- ,......, .
anti-Akt
{a) Assay Priciple
Lysate # 1 Lysate# 1

I -~-- 1 anti-® p42/p44 (Low GTP-bound

Rae content)
{High GTP-bound
Rae content)

1--- :nil
anti-p42/p44

Activation of three signal

GOP-bound Rae
• GTP-boun d Rae
C. PAK1 PB D aga rose
~ •
•• ·.I•
•
• •I
transduction proteins by phosphoryl ation. Mouse erythrocyte
progenitor cells were treated for 10 min with different concentrations J • •I
•
of the hormone erythropoietin (Epo). Extracts of the cell were analyzed
by Western blotting with three different antibodies specific for the
phosphorylated forms of three signal transduction proteins and three
1
-=::::.
D PAKl PBD agarose
is added !
that recognize a nonphosphorylated segment of amino acids in the I
~"{
same protein. The data show that with increasing concentration of Epo,
'=J:·' ';'
~• =•
the three proteins become phosphorylated. Treatment with 1 unit Epo
per ml is sufficient to maximally phosphorylate and t hus activate all
three pathways. Stat 5 = transcription factor phosphorylated on
tyrosine 694; Akt = kinase phosphorylated on serine 473; p42/p44 =
I.· l
)"
•• •
·l
)•
p42/p44 MAP kinase phosphorylated on threonine 202 and tyrosine
204. [Courtesy Jing Zhang; Zhang et al., 2003, Blood 102:3938.]

between an active ("on") form with bound GTP that modu-

1
--:-
f) M ixi ng and
centifugat ion
{pull-down of
1
GTP-bound Rae)
lates the activity of specific target proteins and an inactive
("off") form with bound GDP. The principal assay for mea- 1•:·~• •
suring activation of this class of proteins takes advantage of ••
••
the fact that each such protein has one or more targets to
,p ·~
TA u A pull-down assay shows that
t he small GTP-binding protein Rac1 is activated by platelet-
derived growth factor (PDGF). Like other small GTPases, Racl regulates
molecular events by cycling between an inactive GOP-bound form and
an active GTP·bound form. In its active (GTP-bound) state, Racl binds
specifically to the p21-binding domain (PBD) of p21-activated protein
kinase (PAK) to control downstream signaling cascades. (a) Assay
principle: the Rae-binding PBD domain is generated by recombinant
DNA techniques and attached to agarose beads, then mixed with cell
extracts (step 0 ). The beads are recovered by centrifugation (step f)) {b) Western blot of hematopoietic stem cells
before and after treatment w ith PDGF
and the amount of GTP-bound Racl is quantified by Western blotting
using an anti-Racl antibody (step Ill. (b) Western blot showing PDGF 0 1'
activation of Rac1 after treatment of hematopoietic stem cells for 1 min
with the hormone platelet-derived growth factor (PDGF). A Western Rae GTP (visualized
w ith ant i-Rae antibody)
blot for actin serves as a control that the same amount of total protein
is loaded on each lane of the gel. [(a) After Cell Biolabs Inc.; (b) from G. . . . . . . . . . . . . ~ actin {visua lized with
Ghiaur et al., 2006, Blood 108:2087-2094.] ~ anti-actin antibody)

686 CHAPTER 15 • Signal Transduction and G Protein - Coupled Receptors

the fraction of the small GTPase Racl that has a bound GTP In this section, we discuss the basic structure and mecha-
increases markedly after stimulation by the hormone platelet- nism of GPCRs and their associated trimeric G proteins. In
derived growth factor (PDGF), indicating that Racl is a sig- Sections 15.4 through 15.6, we describe GPCR pathways
nal transduction protein activated by the PDGF receptor. that activate several different effector proteins.

KEY ("ni\JriEPTS nf c;.oction 1 5-2 All G Protein-Coupled Receptors Share

the Same Basic Structure
Studying Cell-Surface Receptors and Signal
All G protein-coupled receptors have the same orientation in
Transduction Proteins
the membrane and contain seven transmembrane a-helical re-
The concentration of ligand at which half the ligand's re- gions (H1-H7), four extracellular segments, and four cyto-
ceptors are occupied, the Kd, can be determined experimen- solic segments (Figure 15-15). Invariably theN-terminus is on
tally and is a measure of the affinity of the receptor for the the exoplasmic face and the C-terminus is on the cytosolic
ligand (see Figure 15-1 0) . face of the plasma membrane. The carboxyl-terminal segment
Because of receptors' high affinity for their target ligand, (C4 ), the C3 loop, and, in some receptors, also the C2 loop
the extracellular domain of receptors can be used as a drug are involved in interactions with a coupled trimeric G protein.
to reduce the level of free hormone. Many subfamilies of G protein-coupled receptors have been
conserved through evolution; members of these subfamilies
The maximal response of a cell to a particular ligand gener-
are especially similar in amino acid sequence and structure.
ally occurs at ligand concentrations at which less than 100
G protein-coupled receptors are stably anchored in the
percent of its receptors are bound to ligand (see Figure 15-12).
hydrophobic core of the plasma membrane by many hy-
Affinity chromatography techniques can be used to purify drophobic amino acids on the outer surfaces of the seven
receptors even when they are present in low abundance. membrane-spanning segments. One group of G protein-
Immunoprecipitation assays using antibodies specific for coupled receptors whose structure is known in molecular
protein kinases can measure kinase activity. lmmunoprecipi- detail is the P-adrenergic receptors, which· bind hormones
tation assays using antibodies specific for phosphorylated such as epinephrine and norepinephrine (Figure 15-16). In
peptides can measure phosphorylation of a specific amino these and many other receptors, segments of several membrane-
acid on any desired protein within a cell (see Figure 15-13). embedded a helices and extracellular loops form the ligand
binding site that is open to the exoplasmic surface. The an-
Pull-down assays using the protein-binding domain of a tar-
tagonist cyanopindolol, shown in Figure 15-16, binds with a
get protein can be used to quantify activation of a GTP-binding
much higher affinity to the receptor than most agomsts, and
protein within a cell (see Figure 15-14).
the receptor-ligand complex has been crystallized and its
structure determined. Side chains of 15 amino acids located in
four transmembrane a helices and extracellular loop 2 make

15.3 G Protein-Coupled Receptors:

Structure and Mechanism
As noted above, perhaps the most numerous class of recep-
tors are the G protein-coupled receptors (GPCRs). In hu-
mans, GPCRs are used to detect and respond to many Exterior
different types of signals,)ncluding neurotransmitters, hor-
mones involved in glycogen and fat metabolism, and even
photons of light. All GPCR signal transduction pathways
share the following common elements: ( 1) a recepror that
contains seven membrane-spanning a helixes; (2) a coupled Cytosol
trimeric G protein, which functions as a switch by cycling
between active and inactive forms; (3) a membrane-bound
effector protein; and (4) proteins that participate in feedback
regulation and desensitization of the signaling pathway. A FIGURE 15-15 General struct ure of G protein-coupled receptors.
second messenger also occurs in many GPCR pathways. All receptors of this type have the same orientation in the membrane
GPCR pathways usually have short-term effects in the cell by and contain seven transmembrane a-helical regions (H l-H7), four
quickly modifying existing proteins, either enzymes or ion extracellular segments (El-E4), and four cytosolic segments (Cl-(4).
channels. Thus these pathways allow cells to respond rapidly The carboxyl-terminal segment (C4), the C3 loop, and, in some
to a variety of signals, whether they are environmental stim- receptors, also the C2 loop are involved in interactions with a coupled
uli such as light or hormonal stimuli such as epinephrine. trim eric G protein.

15.3 G Protein-Coupled Recepto rs : Structure and Mechanism 687

FIGURE 15· 16 St ructure of the turkey Jl 1 -adrenergic receptor (a) ~-adrenergic receptor
complexed with the antagonist cyanopindolol. (a) Side view
showing the approximate location of the membrane phospholipid Exoplasmic face
bilayer. A ribbon representation of the receptor structure is in rainbow
coloration (N-terminus, blue; (-terminus, red), with cyanopindolol as a
gray space-filling model. The extracellular loop 2 (E2) and cytoplasmic
loops 1 and 2 ((1, C2) are labeled. (b) View from external face showing
a close-up of the ligand-binding pocket that is formed by amino acids
in helices 3, 5, 6, and 7, as well as extracellular loop 2, located between
helite~ 4 and 5. Cyanopindolol atoms are colored grey (carbon), blue
(nitrogen), and red (oxygen). The ligand-binding pocket comprises 15
side chains from amino acid residues in four transmembrane a-helices
and extracellular loop 2. As examples of specific binding interactions,
the positively charged N atom in the amino group found both in
cyanopindolol and in epinephrine forms an ionic bond with the
carboxylate side chain of aspartate 121 (0 121 ) in helix 3 and the
carboxylate of asparagine 329 (N 32 ~ in helix 7. [From T. Wayne et al.,
-ooc
2008, Nature 454:486.]
Cytosolic face

C2
noncova lent contacts with the ligand. The amino acids that
form the interior of different G protein-coupled receptors
(b) View from exoplasmic face
are diverse, allowing different receptors to bind very different
small molecu les, whether they are hydrophilic such as epi-
nephrine or hydrophobtc such as many odorants.
While all G protein-coupled receptors share the same
basic structure, different subtypes of GCPRs can bind the
same hormone, with different cell ular effects. To illustrate
the versatility of these receptors, we will consider the set of G
protein-coupled receptors for epinephrine found in different
types of mammalian cells. The hormone epinephrine is par-
ticularly important in mediating the body's response to stress,
also known as the fight-or-flight response. During moments
of fear or heavy exercise, when tissues may have an increased
need to catabofize glucose and fatty acids to produce ATP,
epinephrine signals the rapid breakdown of glycogen to glu-
cose in the liver and of triacylglycerols to fatty acids in adi-
pose (fat) cells; within seconds these principal metabolic fuels
are supplied to the blood. In mammals, the li beration of glu-
cose and fatty acids is triggered by binding of epinephrine (or
its derivative norepinephrine) to [3-adrenergic receptors on
the surface of hepatic (liver) and adipose cells.
Epinephrine has other bodily effects as well. Epinephrine
bound to [3-adrenergic receptors on heart muscle cells, for
example, increases the contraction rate, which increases the
blood supply to the tissues. In contrast, epinephrine stimula-
tion of [3-adrenergic receptors on smooth muscle cells of the
intestine causes them to relax. Another type of epinephrine
GPCR, the a-adrenergtc receptor, is found on smooth mus-
cle cells lining the blood vessels in the intestinal tract, skin,
and kidneys. Binding of epinephrine to these receptors causes
the arteries to constrict, cutting off circulation to these or-
gans. These diverse effects of epinephrine help o rchestrate
integrated responses throughout the body all directed to a
common end: supplying energy to major locomotor muscles,
while at the same time diverting it from other organs not as
crucial in executing a response to bodily stress.

688 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

Ligand-Activated G Protein- Coupled Receptors enables the receptor to bind to the Ga subunit (Figure 15-1 7,
Catalyze Exchange of GTP for GOP on steps D and f)). This binding releases the bound GDP; thus
the activated ligand-bound receptor functions as a guanine
the a Subunit of a Trimeric G Protein
nucleotide exchange factor (GEF) for the Ga subunit (step
Trimeric G proteins contain three subunits designated a , 13, 1)). Next, GTP rapidly binds to the "empty" guanine nucleo-
and 'Y· Both the G" and G'Y subunits are linked to the mem- tide site in the G" subunit, causing a change in the conforma-
brane by covalently attached lipids. The 13 and 'Y subunits are tion of its switch segments (see Figure 15-7). These changes
always bound together and are usually referred to as the G~h weaken the binding of Ga with both the receptor and the Gfi'Y
subunit. In the resting state, when no ligand is bound to the su bunit (step a ). In most cases, Gn·GTP, which remains an-
receptor, the<..;" subunit has a bound GDP and is complexed chored in the membrane, then interacts with and activates an
with Gfi'Y· Binding of a ligand (e.g., epinephrine) or an ago- effector protein, as depicted in Figure 15-17 (step ~). In
nist (e.g., isoproterenol) to a G protein-coupled receptor some cases, Ga·GTP inhibits the effector. Moreover, depend-
changes the conformation of its cytosol-facing loops and ing on the type of cell and G protein, the Gfi'Y subunit, freed

G) OVERVIEW ANIMATION: Extracellular Signaling

0 Hormone
Exterior~
i Inactive
receptor
Cytosol

RESTING
STAT E

D Binding of hormone induc~s

a conformational change H
in receptor 0 ormone

fJ Activated receptor {
~ P
[]
binds toGa subunit
1!1 Hormone dissociates )
from receptor; Ga binds
to effector, activating it

T
p

7'
G
T IJ Binding of GTP to Gu
P triggers dissociation of Ga
both from the receptor and
fl fromGpy

~
FIGURE 15-17 General mechanism of the activation of effector and leads to reassembly of the trimeric G protein, returning the system
proteins associated w ith G protein- coupled receptors. The G., and to the resting state (step ~ ).Binding of another ligand molecule
G~'Y subunits of trimeric G proteins are tethered to the membrane by causes repetition of the cycle. In some pathways, the effector protein is
covalently attached lipid molecules (wiggly black lines). Following activated by the free G 11~ subunit. The sin trimeric G, protein stands for
ligand binding, exchange of GOP with GTP, and dissociation of the "stimulatory." [After W. Oldham and H. Hamm, 2006, Quart. Rev. Biophys.
G protein subunits (steps 0--EJ ). the free Gu·GTP binds to and activates 39:1 17.)
an effector protein (step 1!.1 ). Hydrolysis of GTP terminates signaling

15.3 G Protein-Co upled Receptors: Structure and Mechanism 689

from its a subunit, will sometimes transduce a signal by in- as well as GTP does but cannot be hydrolyzed by the intrinsic
teracting with an effector protein. GTPase. In some of these compounds, the P-0-P phosphodies-
The active G":GTP state is short-lived because the bound ter linkage connecting the [3 and 'Y phosphates of GTP is re-
GTP is hydrolyzed to GDP in minutes, catalyzed by the intrinsic placed by a nonhydrolyzable P-CHrP or P-NH-P linkage.
GTPase activity of the G., subunit (see Figure 15-17, step DJ). Addition of such a GTP analog to a plasma membrane prepara-
The conformation of the G" thus switches back to the inactive tion in the presence of an agonist for a particular receptor re-
G,:GDP state, blocking any further activation of effector pro- sults in a much longer-lived activation of the G protein and its
teins. The rate of GTP hydrolysis is sometimes further en- associated effector protein than occurs with GTP. In this ex-
hanced by binding of the G"·GTP complex to the effector; the periment, once the nonhydrulyzable GTP analog is exchanged
effector thus functions as a GTPase-activating protein (GAP). for GDP bound to G", it remains permanently bound to Ga.
This mechanism significantly reduces the duration of effector Because the Ga·GTP-analog complex is as functional as the
activation and avoids a cellular overreaction. In many cases, a normal G.,·GTP complex in activating the effector protein, the
second type of GAP protein called a regulator of G protein effector remains permanently active.
signaling (RGS) also accelerates GTP hydrolysis by the G" sub- GPCR-mediated dissociation of trimeric G proteins can be
unit, further reducing the time during which the effector re- detected in living cells. These studies have exploited the phe-
mains activated. The resulting G.,·GDP quickly reassociates nomenon of fluorescence energy transfer, which changes the
with G~"Y and the complex becomes ready to interact with an wavelength of emitted fluorescence,when two fluorescent pro-
activated receptor and start the process all over again. Thus the teins interact (see Figure 9-22). Figure 15-18 shows how this
GPCR signal transduction system contains a built-in feedback experimental approach has demonstrated the dissociation of
mechanism that ensures the effector protein becomes activated the G"·G~"Y complex within a few seconds of ligand addition,
only for a few seconds or minutes following receptor activa- providing further evidence for the model of G protein cycling.
tion; continual activation of receptors via ligand binding to- This general experimental approach can be used to follow the
gether with subsequent activation of the corresponding G formation and dissociation of other protein-protein complexes
protein is essential for prolonged activation of the effector. in living cells. ·
Early evidence supporting the model shown in figure 15-17 ror many years, it was impossible to determine the struc-
came from studies with compounds called GTP analogs that ture of the same GPCR in the active and inactive states. This
are structurally similar to GTP and so can bmd to Gu subunits has now been accomplished with the [3-adrenergic receptor (as

0 PODCAST: Activation of G Proteins Measured by Fluorescence Resonance Energy Transfer (FRET)

(a) (b)
cAMP
0

(lf1 .
Inactive

D /~U~( ~
Fluorescence
527 nm / l-
~ ·
P
gL_)
/ F
( ellow) Fluorescence
Y energy
transfer Excitation light
440 nm
Fluorescence
490 nm
(cyan)
l
Excitation light
440 nm
0
Time (s)

EXPERIMENTAL FIGURE 15-18 Activation of G proteins emission of 527-nm (yellow) light. characteristic ofYFP. However, if
occurs within seconds of ligand binding in amoeba cells. in the ligand binding leads to dissociation of the G" and G~1 subunits, then
amoeba Dictyostelium discoideum cell, cAMP acts as an extracellular fluorescence energy transfer cannot occur. In this case, irradiation
signaling molecule and binds to a G protein- coupled receptor; it is of cells at 440 nm causes emission of 490-nm light (cyan) characteristic
not a second messenger. Amoeba cells were transfected with genes of CFP (right). (b) Plot of the emission of yellow light (527 nm) from a
encoding two fusion proteins: a Ga fused to cyan fluorescent protein single transfected amoeba cell before and after addition of extracellu-
(CFP), a mutant form of green fluorescent protein (GFP), and a G ~ fused lar cyclic AMP (arrow), the ligand for the G protem-coupled receptor
to another GFP variant, yellow fluorescent protein (YFP). CFP normally in these cells. The drop in yellow fluorescence, which results from the
fluoresces 490-nm light; YFP, 527-nm light. (a) When CFP and YFP are dissociation of the G,.-CFP fusion protein from the G~-YFP fusion
nearby, as in the resting G.. ·G~~ complex, fluorescence energy transfer protein, occurs within seconds of cAMP addition. [Adapted from
can occur between CFP and YFP (left). As a result, irradiation of resting C. Janetopoulos et al., 2001, Science 291 :2408.]
cells with 440-nm light (which directly excites CFP but not YFP) causes

690 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

.'•
 (a) Side view p-adrenergic receptor FIGURE 15-19 Structure of the j3-adrenergic receptor in the
·. inactive and active states and with its associated trim eric G
protein, G us• (a) Comparison of the three-dimensional structures of the
activated [3-adrenergic receptor (gold) bound to a strong agonist and
Exterior
the inactive receptor (purple) bound to an antagonist. (b) View from
the cytosolic surface. Note the major changes seen in the conforma
Membrane tions of the intracellular domains of transmembrane helices 5 (TMS)
and 6 (TM6). In the active state, TMS is extended by two helical turns,
whereas TM6 is moved outward by 1.4 nm. (c) The overall structure of
the active receptor complex shows the adrenergic receptor (gold)
bound to an agonist (black and red spheres) and engaged in extensive
interactions with a segment of G0 , (purple). G.,, together with G~
(g reen) and Gy (red) constitute the heterotrimeric G protein G,. [After
Cytosol S. Rasmussen et al., 2011, Nature 476:387-390.]

TM5
~-
cha nge (Figure 15-19a) in which there are substantial move-
(b) View from cytosolic surface ments of transmembrane helices 5 and 6 and changes in the
structure of the C3 loop; together these c reate a surface that
can now bind to a segment of the Gm subunit (Figure 15-19b).
X-ray crystallographic studies of the complex of activated
receptor and G, have also revealed how the subunits of a G
protein interact with each other and provided clues about
how binding of GTP leads to dissociation of the G('l from the
Gp-y subunit. As revealed in the structural model in Figure
15-19b, a large surface of Ga ·GDP interacts with the G 13 sub-
un it; part of this surface is located in the aN alpha helix in
theN-terminal segment of G('I·GDP. Note that G" directl y
contacts Gil but not G'Y. Binding of theN-terminal alpha-helical
segments aN a nd aS of the Ga, protein to transmembrane
(c) P-adrenergic receptor
helices 5 and 6 of the activated receptor (figure 15- J9b) will,
as with other G proteins, be followed by opening of the G"
Exterior subunit, eviction of the bound GDP, and its replacement with
GTP; this is immediately followed by conformational changes
Membrane within switches I and II that disrupt the molecular interac-
tions between Ga and G 13-y, leading to their dissociation.

Cytosol Different G Proteins Are Activated

by Different GPCRs and In Turn Regulate
Different Effector Proteins
Gas All effector proteins in GPCR pathways are either mem-
brane-bound ion c hannels or membrane-bound enzymes
that catalyze formation of the second messengers shown in
Figure 15-8. The variations on the theme of GPCR signaling

GTP ~
J that we exami ne in Sections 15.4 through 15.6 arise because
multiple G proteins are encoded in eukaryotic genomes. At
last count, humans have 2 1 different G('l subunits encoded by
16 genes, several of which undergo alternative spl icing; six
G 13 subunits; and 12 G'Y subunits. So far as is known, the dif-
ferent Gll-v sub units a re essentia ll y interchangeable in their
functi ons, while the different G('l subunits afford the various
well as with rhodopsin, discussed in Section 15.4 ). The seven G proteins their specificity. Thus we can refer to the entire
membrane-embedded a helices of the ~-adrenergic receptor three-subunit G protein by the name of its alpha subunit.
completely surround a central segment to which an agonist or Table 15-"1 summarizes the functions of the major classes
antagonist is noncovalently bound (Figu re 15- 19). Binding of o f G proteins with different Gu subunits. For example, the
an agonist to the receptor induces a major conformational different types of epinephrine receptors mentioned previously

15.3 G Protein-Coupled Receptors: Structure and Mechanism 691

 TABLE 1 S- 1 Major Classes of Mammalian Trimeric G Proteins and Their Effectors·

Ga Class Associated Effecto r 2nd Messenger Receptor Exam ples

Adenylyl cyclase cAMP (increased) 13-Adrenergic (epinephrine) receptor;

receptors for glucagon, serotonin, vasopressin

GQI Adenylyl cyclase cA.\1P (decreased) arAdrenergic receptor

K+ channel (G~1 Change in membrane potential Muscarinic acetylcholine receptor
activates effector)

Adenylyl cyclase cAMP (increased) Odorant receptors in nose

Phospholipase C IP3 , DAG (increased) a 1-Adrenergic receptor

G"o Phospholipase C IP3, DAG (increased ) Acetylcholine receptor in endothelial cells

cGMP phosphodiesterase cGMP (decreased) Rhodopsin (light rece'ptor) in rod cells

,. A g1ven c. subclass may be assoc1ared w1rh more rhan one cffecror protein. To dare, only one major C,,_ has been identified, bur multiple C.4 and
C,. proteins have been described. Effector proteins commonly are regula red by C. bur in some cases by C~) or rhe combined acrion of G. and G~)·
IP3 = inositol
1,4,5-rnsphosphare; DAG = 1,2-dmcylglycerol.
SOURCE~: See L. B1rnbaumer, 1992, Cell 71 :1069; Z. Farfel er al., 1999, New F11g. ]. Med. 340: 1012; and K. Pierce cr al., 2002, Nature Reu. Mol. Cell
Bioi. 3:639.

are coupled to different G" subunits that influence effector reducing the inhibition of adenylyl cyclase. The resulting in-
proteins differently and so have distinct effects on cell behav- crease in cAMP in epithelial cells of the airways promotes
ior in a target cell. Both subtypes of 13-adrenergic receptors, loss of fluids and electrolytes and mucus secretion. •
termed 13 1 and 13 2, arc coupled to a stimulatory G protein (G.)
whose alpha subunit (Gnsl activates a membrane-bound ef-
fector enzyme called adenylyl cyclase. Once activated, this
enzyme catalyzes synthesis of the second messenger cAMP. In KEY CONCEPTS of Section 1 5.3
contrast, the a 2 subtype of 13-adrenergic receptor is coupled to
an inhibitory G protein (G,) whose alpha subunit G"' inhibits G Protein-Coupled Receptors: Structure
adenylyl cyclase, the same effector enzyme associated with and Mechanism
13-adrenergic receptors. The Guq subunit, which is coupled to • G protein-coupled receptors (GPCRs) are a large and di-
the a 1-adrenergic receptor, activates a different effector en- verse family with a common structure of seven membrane-
zyme, phospholipase C, which generates two other second spanning a helices and an internal ligand-binding pocket
messengers, DAG and IP3 (see Figure 15-8 ). Examples of sig- that is specific for ligands (see Figures 15-15 and 15-16).
naling pathways that use each of the Gn subunits listed in
GPCRs can have a range of cellular effects depending on
Table 15-1 are described in the following three sections.
the subtype of receptor that binds ligand. The hormone epi-
nephrine, for example, which mediates the fight-or-flight
.. Some bacterial toxins contain a subunit that penetrates
response, binds to multiple subtypes of GPCRs in multiple
the plasma membrane of target mammalian cells and in
cell types, with varying physiological effects.
the cytosol catalyzes a chemical modification on Gn proteins
that prevents hydrolysis of bound GTP to GDP. For example, GPCRs are coupled to trimeric G proteins, which contain
toxins produced by the bacterium Vibrio cholera, which three subunits designated a, 13, and -y. The Ga subunit is a
causes cholera, or certain strains of E. coli, modify the G,,. GTPase switch protein that alternates between an active
protein in intestinal epithelial cells. As a result, Ga, remains in ("on") state with bound GTP and inactive ("off" ) state with
the active state, continuously activating the effector adenylyl GDP. The "on" form separates from the 13 and -y subunits
cyclase m the absence of hormonal stimu lation. The resulting and activates a membrane-bound effector. The 13 and-y sub-
excessive rise in intracellular cAMP leads to the loss of elec- units remain bound together and only occasionally trans-
trolytes and water into the intestinal lumen, producing the duce signals (see Figure 15-17) .
watery diarrhea characteristic of infection by these bacteria. Ligand binding causes a conformational change in certain
The toxin produced by Bordetella pertussis, a bacterium that membrane-spann ing helices and intracellular loops of the
commonly infects the respiratory tract and causes whooping GPCR, allowing it to bind to and function as a guanine nu-
cough, catalyzes a modification of Gni that prevents release of cleotide exchange factor (GEF) for its coupled Ga subunit,
bound GDP. As a result, G ... i is locked in the inactive state,

692 CHAPTER 15 • Signal Transduction and G Protein - Coupled Receptors

 Acetylcholine
K- channel
catalyzing dissociation of GDP and allowing GTP to bind.
The resulting change in conformation of switch regions in Exterior t +
Ga causes it to dissociate from the G 13'Y subunit and the recep-
tor and interact with an effector protein (see Figure 15-17).
Cytosol
• Fluorescence energy-transfer experiments demonstrate
receptor-mediated dissociation of coupled Ga and G 13'Y sub- Active muscarinic
units in living cells (see Figure 15-18). acetylcholine receptor
The effector proteins activated (or ina~tivated) by trimeric
G proteins are either enzymes that form second messengers K•
(e.g., adenylyl cyclase, phospholipase C) or ion channels (see

J]~
Table 15-1 ). In each case, it is the Ga subunit that determines + ... + + +
the function of the G protein and affords its specificity.

FIGURE 1 5 -20 Activation of the muscarinic acetylcholine

15.4 G Protein-Coupled Receptors
receptor and its effector K+ channel in heart muscle. Binding of
That Regulate lon Channels acetylcholine triggers activation of theGn, subunit and its dissociation
from the G~~ subunit in the usual way (see Figure 15-1 7). In this case,
One of the simplest cellular responses to a signal is the open-
the released Gil~ subunit (rather than Gu1·GTP) binds to and opens the
ing of ion channels essential for transmission of nerve im- associated effector protein, a K- channel. The increase in K permeability
pulses. Nerve impulses are essential to the sensory perception hyperpolarizes the membrane, which reduces the frequency of heart
of environmental stimuli such as light and odors, to transmis- muscle contraction. Though not shown here, activation is terminated
sion of information to and from the brain, and to the stimula- when the GTP bound to G., 1 is hydrolyzed (by a GAP enzyme that is an
tion of muscle movement. During transmission of nerve intrinsic part of the G,., subunit) to GDP and G0 , ·GDP recombines with
impulses, the opening and closing of ion channels cause!> G~) · [See K. Ho et al., 1993, Nature 362:31, and Y. Kubo et al.. 1993, Nature
changes in the membrane potential. Many neurotransmitter 362:127.]
receptors are ligand-gated ion channels, which open in re-
sponse to binding of a ligand. Such receptors include some
types of glutamate, serotonin, and acetylcholine receptors,
including the acetylcholine receptor found at nerve-muscle
synapses. Ligand-gated ion channels that function as neu-
rotransmitter receptors are covered in Chapter 22. receptor is coupled to a Gai subunit, and ligand binding leads
Many neurotransmitter receptors, however, are G protein- to opening of associated K+ channels (the effector protein) in
coupled receptors whose effector proteins are Na.,. or K'" chan- the plasma membrane (Figure 15-20). The subsequent efflux
nels. Neurotransmitter binding to these receptors causes the of K ions from the cytosol causes an increase in the magni-
associated ion channel to open or close, leading to changes in tude of the usual inside-negative potential across the plasma
the membrane potential. Still other neurotransmitter receptors, membrane that lasts for several seconds. This state of the
as well as odorant receptors in the nose and photoreceptors in membrane, called hyperpolarization, reduces the frequency
the eye, are G protein-coupled receptors that indirectly modu- of muscle contraction. This effect can he shown experimen-
late the activity of ion chapnels via the action of second mes- tally by adding acetylcholine to isolated heart muscle cells
sengers. In this section, we consider two G protein-coupled and measuring the membrane potential using a microelec-
receptors that illustrate the direct and indirect mechanisms for trode inserted into the cell (see Figure 11-19).
regulating ion channels: the muscarinic acetylcholine receptor As shown in Figure 15-20, the signal from activated mus-
of the heart and the light-activated rhodopsin protein in the eye. carinic acetylcholine receptors is transduced to the effector
channel protein by the released G 13) subunit rather than by
Ga,·GTP. That G 13'Y directly activates the K• channel was dem-
onstrated by patch-clamping experiments, which can measure
Acetylcholine Receptors in the Heart Muscle
ion flow through a single ion channel in a small p;:~tch of
Activate a G Protein That Opens K+ Channels membrane (see Figure ll-22). When purified G 13'Y protein was
Muscarinic acetylcholine receptors are a type of GPCR added to the cytosolic face of a patch of heart muscle plasma
found in cardiac muscle. When activated, these receptors membrane, K channels opened immediately, even in the ab-
slow the rate of heart muscle contraction. Because musca- sence of acetylcholine or other neurotransmitters--clearly in-
rine, an acetylcholine analog, also activates these receptors, dicating that it is the G 13'Y protein that is responsible for
they are termed "muscarinic." This type of acetylcholine opening the effector K channels and not G,;GTP.

15.4 G Protein-Coupled Receptors That Regulate lon Channels 693

Light Activates G Protein-Coupled Rhodopsins flattened membrane disks that make up the outer segment of
in Rod Cells of the Eye these rod-shaped cells (Figure 15-21 ). A human rod cell con-
tains about 4 X 10 7 molecules of rhodopsin. The trimeric G
The human retina contains two types of photoreceptor cells, protein coupled to rhodopsin, called transducin (Gt), con-
rods and cones, which are the primary recipients of visual tains a G 0 unit referred to as Gat; like rhodopsin, Gat is found
stimulation. Cones are involved in color vision, while rods are only in rod cells.
stimulated by weak light such as moonlight over a range of Rhodopsin differs from other GPCRs in that binding of a
wavelengths. The photoreceptor cells synapse on layer upon ligand is not what activates the receptor. Rather, absorption of
layer of interneurons that are innervated by different combina- a photon uf light by the bound retinal is the activating stgnal.
tions of photoreceptor cells. All these signals are processed and On absorption of a photon, the retinal moiety of rhodopsin is
interpreted by the part of the brain called the visual cortex. immediately converted from the cis form (known as 11-cis-
Rod cells sense light with the aid of a light-sensitive retinal) to the aU-trans isomer, causing a conformational change
GPCR known as rhodopsin. Rhodopsin consists of the pro- in the opsin protein (Figure 15-22). This is equivalent to the
tein opsin, which has the usual GPCR structure, covalently activating conformational change that occurs on ligand binding
linked to a light-absorbing pigment called retinal. Rhodop- by other G protein-coupled receptors; this conformational
sin, found only in rod cells, is localized to the thousand or so

(a) (b)

Outer
segment

Disks
containing -='-:,.,.r--1
rhodopsin

Rough
Golgi endoplasmic
reticulum

Inner
FIGURE 15-21 Human rod cell.
segment
(a) Schematic diagram of an entire rod
cell. At the synaptic body, the rod cell
forms synapses with one or more
interneurons. Rhodopsin, a light-
sensitive G protein-coupled receptor,
Nucleus
is located in the flattened membrane
disks of the cell's outer segment. (b)
Electron micrograph of the region of
the rod cell indicated by the bracket in
(a). This region includes the junction
of the inner and outer segments. [Part
(b) from R. G. Kessel and R. H. Kardon, 1979, '\Synaptic
Ttssues and Organs: A Text-Atlas ofScanning
body
Electron Microscopy, W. H. Freeman and
Human rod cell 0.5,um
Company, p. 91.)

694 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 11-cis-retinal moiety membrane of resting rod cells is due to the presence of a
large number of open nonselectiue ion channels that admit
Na+ and Ca 2 +; reca ll from Chapter 11 that movement of
positively charged ions such as Na + and Ca 2 ' from the out-
side of the cell to the inside will reduce the magnitude of the
Lysine side chain inside-negative membrane potential. Absorption of light by
H3C + I rhodopsin leads to closing of these channels, causing the
membrane potential to become more inside negative.
C=N-(CH 2 )4 - Opsin
I I The more photons absorbed by rhodopsin, the more
H H channels arc closed, the fewer Na + and Ca 2 + ions cross the
Rhodopsin membrane from the outside, the more negative the mem-
brane potential becomes, and the less neurotransmitter is
Light-i~du~ed
released. The reduction in neurotransmitter release is trans-
tsomenzatton
W (<10 2 s) mitted to the brain by a series of neurons, where it is per-
ceived as light.
Unlike the muscarinic acetylcholine receptor discussed
.. All-trans-retinal moiety earlier, G proteins activated by rhodopsin do not act directly
on ion channels. The closing of cation channels in the rod-

CH3
11
I
trans

CH3
cell plasma membrane requires changes in the concentration
of the second messenger cyclic GMP, or cGMP (see Figure
15-8). Rod cell outer segments contain an unusually high
'-'::: '-'::: '-':::
12 concentration (-0.07 mM) of cGMP, which is continu-
ously formed from GTP in a reaction catalyzed by guanylyl
Meta-rhodopsin II cyclase. However, ligh t absorption by rhodopsin induces ac-
(activated opsin ) tivation of a cGMP phosphodiesterase (PDE), which hydro-
lyzes cGMP to 5' -GMP. As a result, the cGMP concentration
FIGURE 15-22 The light-triggered step in vision. The light-
decreases on ill umination. The high level of cGMP present in
absorbing pigment 1 1-cis-retinal is covalently bound to the amino
the dark acts to keep cGMP-gated cation channels open; the
group of a lysine residue in opsin, the protein part of rhodopsin.
light-induced decrease in cGMP leads to channel closing, mem-
Absorption of light causes rapid photo isomerization of the bound
cis-retinal to the all-trans isomer. This triggers a conformational change in
brane hyperpolarization, and reduced neurotransmitter release.
the opsin protein, forming the unstable intermediate meta-rhodopsin II, As depicted in Figure 15-23, cGMP phosphodiesterase is
.· or activated opsin (see Figure 15-23), w h ich activates G, proteins . the effector protein for Gao both of which are localized to
Within seconds, all-trans-retinal dissociates from opsin and is con- the disk membrane of the rod cell. The free G"',·GTP com-
verted by an enzyme back to the cis isomer, which then rebinds to plex that is generated after light absorption by rhodopsin
another opsin molecule. [See J. Nat hans, 1992, Biochemistry 31 :4923.] binds to the two inhibitory 'Y subunits of cGMP phosphodi-
esterase, releasing the active catalytic a and [3 subunits, which
then convert cGMP to GMP. This is a clear example of how
signal-induced removal of an inhibitor can quickly activate
change allows rhodopsin to bind an adjacent Gar subunit of the
coupled G protein, triggering exchange of GTP for GDP. Acti- an enzyme, a common mechanism in signaling pathways. In
vated rhodopsin, R ,, , is unstable and spontaneously dissociates turn, the lowered concentra tion of cGMP leads to closing of
into its component parts, releasing the covalently attached the cGMP-gated ion channel in the rod cell plasma membrane,
t hereby reducing neurotransmitter release.
opsin, which can no longer bind to a G,, subunit and all-trans-
retinal, thereby terminating visual signaling. In t he dark, free Direct support for the role of cGMP in rod-cell activity has
all-trans-retinal is converted back to 11-cis-retinal, which can been obtained in patch-clamping studies using isolated patches
then rebind to opsin, re-forming rhodopsin. of rod outer-segment plasma membrane, which contains
abundant cGMP-gated cation channels. When cGMP is added
to the cytosolic surface of these patches, there is a rapid in-
crease in the number of open ion channels; cGMP binds di-
Activation of Rhodopsin by light leads
rectly to a site on the channel protein to keep them open. Like
to Closing of cGMP-Gated Cation Channels the K • channels discussed in Chapter 11, the cGMP-gated
In the dark, the membrane potential of a rod cell is about channel protein contains four subuniL~ (~ec Figure 11-20). In
-3 0 mV, considerably less than the resting potential (-60 t his case, each of the subunits is able to bind a cGMP mole-
to - 90 mY) typica l of neurons and other electrically active cule. Three or four cGMP molecules must bind per channel in
cells. This state of the membrane, called depolarization, causes order to open it; this allosteric interaction makes channel
rod cells in t he dark to constantly secrete neurotransmitters, opening very sensitive to small changes in cGMP levels.
and thus the neurons w ith which they synapse are continu- Conversion of active G"',·GTP back to inactive Ga,·GDP
all y being stimulated. The depolarized state of the plasma is accelerated by a specific GTPase-activating protein (GAP).

15.4 G Protein-Coup led Receptors That Reg ulate ion Channels 695
 Light
Cytosol

\ Disk membrane

Disk lumen

IJ

~
(;' - -D- + II Inactive
PDE
Active
PDE Rod

.Ji- ~ ~l~
plasma
u membrane

p
(' P~
G
T
p
G
0
p
.m cGMP GMP

Inactive
~ II

w
Low
POE cy\Osolic ~ Na•
cGMP( e ) ~ caz•
Closed cGMP-gated
ion channel
(less neurotransmitter released)

High Dark-adapted
cytosolic state
cGMP
(• ) ""i:5 rl Na•
C2 5 caz•
Open cGMP-gated
ion channel
(more neurotransmitter released)

FIGURE 15·2 3 Light-act ivated rhodopsin pathway and t he and (3 subunits of PDE hydrolyze cGMP to GMP (step ml. The resulting
closing of cation channels i n rod cells. In dark-adapted rod cells, a decrease in cytosolic cGMP leads to dissociation of cGMP from the
high level of cGMP keeps nucleotide-gated nonselective cation nucleotide-gated channels in the plasma membrane and closing of the
channels open, leading to depolarization of the plasma membrane channels (step [a ). The membrane then becomes transiently hyperpo·
and neurotransmitter release. Light absorption generates activated larized, and neurotransmitter release is reduced. The complex of
rhodopsin, R· (step 0 ), which binds inactive GDP-bound G.,, protein G"1·GTP and the PDE 'Y subunits binds a GTPase activating complex
and mediates replacement of GDP with GTP (step f)). The free G"' ·GTP termed RGS9-Gf35 (step fJ ); by hydrolyzing the bound GTP, this
generated then activates cGMP phosphodiesterase (PDE) by binding triggers the physiologically rapid inactivation of the phosphodiesterase.
m>
to its inhibitory 'Y subunits (step and dissociating them from the [Adapted from V. Arshavsky and E. Pugh, 1998, Neuron 20:11 , and V. Arshavsky,
catalytic a and (3 subunits (step 0 ). Relieved of their inhibition, the a 2002, Trends Neurosci. 25:124.]

In mammals, G"' normally remains in the active GTP-bound signal transduction pathway. Each activated opsin in the disk
state for only a fraction of a second. Thus cGMP phospho- membrane of the rod cell can activate 500 Gn, molecules,
diesterase rapidly becomes inactivated, and the cGMP level each of which in turn activates a cGMP phosphodiesterase.
gradually rises to its original level when the light stimulus is Each molecule of phosphodiesterase hydrolyzes hundreds of
removed. This allows rapid responses of the eye toward cGMP molecules during the fraction of a second it remains
moving or changing objects. active. Thus absorbance of a single photon-yielding a single
activated opsin molecule-can trigger closing of thousands of
ion channels in the plasma membrane and a measurable
Signal Amplification Makes the Rhodopsin change in the membrane potential of the cell.
Signal Transduction Pathway
Exquisitely Sensitive
Rapid Termination of the Rhodopsin Signal
Remarkably, a single photon absorbed by a resting rod cell
Transduction Pathway Is Essential
produces a measurable response, a more inside-negative
change in the membrane potential of about I mV, which in for Acute Vision
amphibians lasts a second or two. Humans arc able to detect As in all G protein-coupled signaling pathways, timely termi-
a flash of as few as five photons. The light-detecting system is nation of the rhodopsin signaling pathway requires that all
so sensitive because the signal is greatly amplified during the the activated intermediates be inactivated rapidly, restoring

696 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

the system to its basal state, ready for signaling again. Thus since Cal-t- is continuously pumped out of the cell indepen-
the three protem intermediates, activated rhodopsin (R ''), dently of the state of these channels. The fall in intracellular
Gar'GTP, and activated cGMP phosphodiesterase (POE), Ca 2 + is sensed by Cal-t- -binding proteins called guanylate-
must all be inactivated, and the concentration of cytoplasmic cyclase-activating proteins, or GCAPs. This results in a rapid
messenger cGMP must be restored to its dark level by guany- stimulation of cGMP synthesis by guanylate cyclase, causmg
lyl cyclase. During a single photon response of a mammalian the ion channels to reopen.
rod cell, the entire process of rhodopsin activation and inac-
tivation is completed within -50 milliseconds, enabling the Rh odopsin Phosphorylatio n an d Bind ing of Arrestin A
eye to detect rapid movements or other changes in objects in major process that down-modulates and terminates the v1sual
our surroundings. Several mechanisms act together to make response involves phosphorylation of rhodopsin in its active
possible this very rapid response. conformation (R *) but not in its inactive, or dark form (R) by
rhodopsin kinase (Figure 15-24), a member of a class of GPCR
GAP Proteins That Inactivate Gat·GTP The complex of the kinases. Each opsin molecule has three principal serine phos-
inhibitory 'Y phosphodiesterase subunit and Gur·GTP recruits phorylation sites on its cytosol-facing C-terminal C4 segment;
a complex of two proteins, RGS9 and G[35, that together act the more sites that are phosphorylated, the less able R * is to
as a GAP protein and hydrolyze the bound GTP to GOP. This activate G.,, and so induce closing of cGMP-gated cation chan-
releases the inhibitory 'Y subunit and terminates phosphodies- nels. The protein arrestin binds to three phosphorylated serine
terase activation. Experiments with mice that have the RGS9 residues on the C-terminal opsin segment. Bound arrestin com-
gene knocked out showed that this protein is essential for pletely prevents interaction of Gar with phosphorylated R *,
normal inactivation of the cascade in vivo. In single mouse totally blocking formation of the active Ga,·GTP complex and
rod cells, the time for recovery from a single flash increased stopping additional activation of cGMP phosphodiesterase.
from the normal 0.2 s to about 9 s in the mutant, a 45-fold During a single photon response of a mammalian rod cell, the
increase, arresting to the importance of this GAP protein. entire process of multiple rhodopsin phosphorylation and ar-
restin binding is completed within -50 milliseconds. The phos-
Ca 2 + -Sensing Proteins That Activate Guanylate Cyclase phates linked to rhodopsin are being continuo~sly removed by
Light-triggered closing of the cGMP-gated Na and Ca 2 + phosphodiesterase enzymes, causing dissociation of arrestin
channel will cause a drop in the cytosolic Ca2 concentration and rapid restoration of rhodopsin to its original, native state.

Cytosol

Disk membrane

Disk lumen

Cytosol

Amount of
activation of Ga1:

Full No
activation activ<~tion

FIGURE 15-24 Inhibition of rhodopsin signaling by rhodopsin R* to activate transducin. Arrest in binds to the completely phosphory-
kinase. Light-activated rhodopsin (R*), but not dark-adapted rhodop- lated opsin, forming a complex that cannot activate transducin at all.
sin, is a substrate for rhodopsin kinase. The extent of rhodopsin [See A. Mendez et al., 2000, Neuron 28:153, and V. Arshavsky, 2002, Trends
phosphorylation is proportional to the amount of time each rhodopsin Neurosci. 2S: 124.]
molecule spends in the light-activated form and reduces the ability of

15.4 G Protein -Co upled Receptors That Regulate ion Channels 697
Rod Cells Adapt to Varying Levels of Ambient light levels, from very dim to bright sunlight. This wide range
Light by Intracellular Trafficking of Arrestin of sensitivity is possible because differences in light levels in the
visual field, rather than the abso lute amount of absorbed
and Transducin light, are ultimately sensed by the brain and used to form visual
Cone cells are insensitive to low levels of illumination, and images. A mechanism for light-dependent regulation of the
the activity of rod cells is inhibited at high light levels. Thus rhodopsin-signaling pathway that involves subcellular traf-
when we move from bright daylight into a dimly lighted ficking of two key signal transduction proteins (Figure 15-25 )
room, we are initially blinded. Slowly, however, the rod cells is responsible for this extraordinarily wide sensitivity range.
become sensitive to the dim light, i!nd we gradually are able In dark-adapted rod celb, -80 to 90 percent of the G,"
to see and distinguish objects. During this interval, the rod and Gfl-y transducin subunits are in the outer segments, while
cell has "turned up" its sensitivity to flashes of light, or con- less than 10 percent of arrestin is localized there (Figure 15-25).
trast. Through this process of visual adaptation, a rod cell This allows maximal activation of the downstream effector
can perceive contrast over a 100,000-fold range of ambient cGMP phosphodiesterase and thus maximal sensitivity to ,•
small changes in light. But exposure for I 0 minutes to mod-
(a) Transducin Arrestin erate daytime intensities of light causes a complete redistri-
oi--,J...
bution of these proteins: over 80 percent of the Gar and Gp-.,
Dark
~ I~ subunits move out of the outer s~gment into other parts of
~I the cell while over 80 percent of the inhibitor arrestin moves
,••
- 1 ~~
~~
Outer
segment into the outer segment. The mechanism by which these pro-
~ ~~ \~ I
~~
teins move is not yet known but probably involves microtu-
~~ I~ bule-attached motors that move attached proteins and
I ~
.. ffi
~~ . . I
particles outward and inward (see Chapter 18 ). The reduc-
tion in Gar and Gfl-y in the outer segment means that Gar pro-
I , .,·~·
\ /
.,~~
~~~ -
teins are physically unable to bind activated rhodopsin and
so activate cGMP phosphodiesterase. At the same time, the
[ )I increase in arrestin in the outer segment means that any ac-

\
tivated rhodopsin will become inactivated more rapidly. To-
\; __.../ /) gether, the drop in transducin and the increase in arrestin
\ (
greatly reduce the ability of small increases in light levels to
.·.

( V'JI.
)
\[ l•IJI!
activate the downstream effector cGMP phosphodiesterase;
thus only large changes in light levels will be sensed by the
(b)
-~ .!~
rod cells. These protein movements are reversed when the
.:;.~ .il :';i: ambient light level is lowered .
Bright ~~
light
~
~ Jfal
,~
~
'\~
~I
'~I \~
~

....,.
,~...... ~
I ~"'"
}
I
~I
•
\ .,/
KEY CONCEPTS of Section 1 5.4
G Protein-Coupled Receptors That Regulate
lon Channels
•~"
The cardiac muscarinic acetylcholine receptor is a GPCR

( ,\ ~ )
whose effector protein is a K" channel. Receptor activation
releases the G~'Y subunit, which binds to and opens K ~ chan-

.,.
''
)
>~(
l
\ nels (see Figure 15-20). The resulting hyperpolarization of the
cell membrane slows the rate of heart muscle contraction.
• Rhodopsin, the photosensitive GPCR in rod cells, com-
prises the protein opsin linked to 11-cis-retinal. Light-induced
FIGURE 15-25 Schematic illustration of transducin and arrestin isomerization of the 11-cis-retinal moiety produces activated
distribution in dark-adapted and light-adapted rod cells. (a) In the
opsin, which then activates the coupled trimeric G protein
dark, most transducin is localized to the outer segment, while most
transducin (G,) by catalyzing exchange of free GTP for hound
arrestin is found in other purts of the cell; in this condition vision is
GDP on the Gar subunit (see Figures 15-22 and 15-23).
most sensitive to very low light levels. (b) In bright light, little transdu-
cin is found in the outer segment and abundant arrestin is found there; • The effector protein in the rhodopsin pathway is cGMP
in this condition vision is relatively insensitive to small changes in light. phosphodiesterase, which is activated by the GurGTP-
Coordinated movement of these proteins contributes to our ability to mediated release of inhibitory subunits. Reduction in t he
perceive images over a 100,000-fold range of ambient light levels. [After cGMP level by this enzyme leads to closing of cGMP-gated
P. Calvert et aL, 2006, Trends Cell Bioi. 16:560.]

698 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 NH 2
Na !Ca 1 channels, hyperpolarization of the membrane, and
0 N~'N
decreased release of neurotransmitter (see Figure 15-23 ). II
- o-P-O-CH 2 0
'
'N
I NJ
• Several mechanisms act to terminate visual signaling: GAP
proteins inactivate Gcxc·GTP, Ca 1 + -sensing proteins activate 0 ~·~
guanylate cyclase, and rhodopsin phosphorylation and bind- -o-P=o H)--f"H
ing of arrestin inhibits activation of transducin. 6 HO OH
• Adaptation to a wide range of ambient light levels is medi- I
-0-P=O
ated by movements of transducm and arrestin into and out
of the rod-celt outer segment, which together modulate the ATP
ability of small increases in light levels to activate the down- J Adenylyl
stream effector cGMP phosphodiesterase and thus the sensi- PP; ""! cyclase
tivity of the rod cell in different ambient levels of light. NH 2

Nl lN
0- CH2 0 (N I NJ
15.5 G Protein- Coupled Receptors That
Activate or Inhibit Adenylyl Cyclase -O=P--
~ 0 OH
GPCR pathways that utilize adenylyl cyclase as an effector
protein and cAMP as the second messenger are found in cAMP
most mammalian cell s, where they regulate cellular func-
H 2 0~ cAMP
tions as diverse as metabolism of fats and sugars, synthesis phosphodiesterase
H+
and secretion of hormones, and muscle contraction. These
pathways follow the general GPCR mechanism outlined in NH 2
Figure 15-17: ligand binding to the receptor activates a cou-
0 Nl f ~N
-o-~-O-C~H2
pled trimeric G protein that activates an effector protein-in
this case, adenylyl cyclase, which synthesizes the diffusible 0 N- - l N-9'
second messenger cAMP from ATP (Figure 15-26). cAMP, -o H H ·
in turn, activates a cAMP-dependent protein kinase that H H
phosphorylates specific target proteins.
To explore this GPCR/cAMP pathway, we focus on the HO OH AMP
first such pathway discovered: the hormone-stimulated gen- FIGURE 15- 26 Synthesis and hydrolysis of cAMP by adenylyl
eration of glucose-1-phosphate from glycogen, a storage cyclase and cAMP phosphodiesterase. Similar reactions occur for
polymer of glucose. The breakdown of glycogen (glycoge- production of cGMP from GTP and hydrolysis of cGMP.
nolysis) occurs in muscle and liver cells in response to hor-
mones such as epinephrine and glucagon and is a principal
way that glucose is made available to cells in need of energy.
This example shows how activation of a GPCR can stimulate the total concentration of Ga,-GTP resulting from binding of
the activity of a host of intracellular enzymes, all involved in both hormones to their respective receptors.
a physiologically important task: glycogen metabolism. Positive (activation) and negative (inhibition) regulation
of adenylyl cyclase activity occurs in many cell types, provid-
ing fine-tuned control of the cAMP level and so of the down-
Adenylyl Cyclase Is Stimulated and Inhibited
stream cellular response (Figure 15-27). For example, in
by Different Receptor-Ligand Complexes adipose cells the breakdown of triacylglycerols to fatty acids
Under conditions where demand for glucose is high because of and glycerol (lipolysis) is stimulated by binding of epineph-
low blood sugar, glucagon is released by the alpha cells in the rine, glucagon, or adrenocorticotropic hormone (ACTH) to
pancreatic islets; in case of sudden stress, epinephrine is re- separate receptors that activate adenylyl cyclase. Conversely,
leased by the adrenal glands. Both glucagon and epinephrine binding of two other hormones, prostaglandin £ 1 (PGEtl or
signal liver and muscle cells to depolymerize glycogen, releas- adeno~ine, to their respective G protein-coupled receptors
ing individual glucose molecules. In the liver, glucagon and inhibits adenylyl cyclase. The prostaglandin and adenosine
epinephrine bind to d ifferent G protein-coupled receptors, receptors activate an inhibitory G, protein that contains the
but both receptors interact with and activate the same stimu- same f3 and -y subunits as the stimulatory G, protein but a
latory G, protein that activates adenylyl cyclase. Hence both different a subunit (G 01 ). After the active Gu,·GTP complex
hormones induce the same metabolic responses. Activation of dissociates from G~'Y' it binds to but inhibits (rather than
adenylyl cyclase, and thus the cAMP level, is proportional to stimulates) adenylyl cyclase, resulting in lower cAMP levels.

15.5 G Protein-Coupled Receptors That Activate or Inhibit Adenylyl Cyclase 699

 Stimu latory { Epinephrine Inhi bitory { PGE 1
h ormone Glu cagon h o rmon e A denosin e
IACTH
I
0 ~
Exterior
r Adenylyll
· cyc lase

Cytosol
~
Recept o r for
sti m ulatory
horm one St imulat o ry
G protein
!!
cAM P
Inh ibitory
G p rote in
Receptor fo r
inhibito ry
h ormon e

com plex com plex

FIGURE 1 S-27 Hormone-induced activation and inhibition of and their corresponding receptors differ. Ligand-stimulated formation of
adenylyl cyclase in adipose cells. Ligand binding to Gas-coupled active G0 ·GTP complexes occurs by the sarne mechanism in both G.,, and
receptors causes activation of adenylyl cyclase, whereas ligand binding to Gu; proteins (see Figure 15-17). However, Gn,·GTP and G,.;·GTP interact
Gaccoupled receptors causes inhibition of the enzyme. The G~~ subunit in differently with adenylyl cyclase, so that one stimulates and the other
both stimulatory and inhibitory G proteins is identical; the Ga subunits inhibits its catalytic activity. [See A. G. Gilman, 1984, Ce// 36:577.]

Structural Studies Established How Gas·GTP G0 ;GTP-the switch II helix and the a3-[35 loop-contact
Binds to and Activates Adenylyl Cyclase the adenylyl cyclase fragments (Figure 15-28b). These con-
tacts are tho ught to be responsible for the activation of the
X-ray crystallographic ana lysis has pi npointed the regions in enzyme by Ga;GTP. Recall that switch II is one of the seg-
Gn;GTP that interact with adenylyl cyclase. This enzyme is a ments of a G" subunit whose conformation ts different in the
multispanning transmembrane protein with two large cyto- GTP-bound and GDP-bound states (see Figure 15-7). The
solic segments containing the catalytic domains that convert GTP-induced conformation of G"' that favors its dissociation
ATP to cAMP (Figure 15-28a). Because such transmembrane from Gll"Y is precisely the conformation essential for binding
proteins are notoriously difficult to crystallize, scientists pre- of G 0 , to adenylyl cyclase.
pared rwo protein fragments encompassing the two catalytic
domains of adenylyl cyclase that rightly associate with each
other in a heterodimer. When these catalytic fragments are
(a)
allowed to associate in the presence of Ga;GTP and forskolin Adenylyl cyc lase
(a plant chemical that binds to and activates adeny lyl cy-
clase), they arc stabilized in their active conformations.
Tbe resulting water-soluble complex (two adenylyl cy-
clase domain fragments/G".-GTP/fo rskolin) had an enzy-
matic activity synthesizing cAMP similar to that of intact
full-length adenylyl cyclase. In this complex, two regions of Cytosol

FIGURE 1 S- 28 Structure of mammalian adenylyl cyclase and its

interaction with G0 ,·GTP. (a) Schematic diagram of mammalian
adenylyl cyclase. The membrane-bound enzyme contains two similar
catalytic domains, which convert ATP to cAMP, on the cytosolic face of
the membrane, and two integral memb rane domains, each of which
is thought to contain six transmembrane Ci helices. (b) Model of the
three-dimensional structure of GasGTP complexed with two fragments
encompassing one cata lytic domain of adenylyl cyclase determined by
x-ray crystallography. The a3-f35 loop (g ray) and the helix in the switch
II region (blue) of G.,,GTP interact simultaneously with a specific region
of adenylyl cyclase. The darker-colored part of Gus is the GTP-binding
domain, which is similar in structure to Ras (see Figure 15-7); t he lighter
part is a helical domain. The two adenylyl cyclase fragments are shown
in orange and yellow. Forskolin (green) locks the cyclase fragments in
their active conformations. [Part (a), see W.-J. Tang and A. G. Gilman, 1992,
Ce// 70:869: part (b) adapted from J. J. G. Tesmer et al., 1997, Science 278:1907.]

700 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

cAMP Activates Protein Kinase A by Releasing binding of cAMP. Each R subunit has two distinct cAMP-
Inhibitory Subunits binding sites, called CNB-A and CNB-B (Figure 15-29b).
Binding of cAMP to both sites on an R subunit causes a con-
The second messenger cAMP, synthesized by adenylyl cy- formational change in the R subunit that leads to release of
clase, transduces a wide variety of physiological signals in the associated C subunit, unmasking its catalytic site and
different cell types in multicellular animals. Virtually all of activating its kinase activity (figure 15-29c).
the diverse effects of cAMP are mediated through activation Binding of cAMP by an R subunit of protein kinase A
of protein kinase A (PKA), also called cAMP-dependent pro- occurs in a cooperative fashion; that is, binding of the first
tein kinase, which phosphorylates different intracellular tar- cAMP molecule to CNB-B lowers the Kd for binding of the
get proteins expressed in different cell types. Inactive PKA is second cAMP to CNB-A. Thus small changes in the level of
a tetra mer consisting of two regulatory (R) subunits and two cytosolic cAMP can cause proportionately large changes in
catalytic (C) subunits (Figure 15-29a). Each R subunit binds the number of dissociated C subunits and, hence, in cellular
to the active site in a catalytic domain and inhibits the activ- kinase activity. Rapid activation of enzymes by hormone-
ity of the catalytic subunits. Inactive PKA is turned on by triggered dissociation of an inhibitor is a common feature of
many signaling pathways.
(a) Inact ive PKA Active PKA
Catalytic subunits Glycogen Metabolism Is Regulated

00 by Hormone-Induced Activation
of Protein Kinase A

A
.t. .t.
.t. .t. .t. Glycogen, a large glucose polymer, is the major storage form
+ .t. "'•
.t. •.. .t. - - + of glucose in animals. Like all biopolymers, glycogen is syn-
cAMP thesized by one set of enzymes and degraded by another
Regulatory (Figure 15-30). Degradation of glycogen, or glycogenolysis,
subunits involves the stepwise removal of glucose res.idues from one
(b) Regulatory (R) subunit structure end of the polymer by a phosphorolysis reaction, catalyzed
by glycogen phosphorylase, yielding glucose !-phosphate.
AKAP Di merization/docki ng In both muscle and liver cells, glucose-1-phosphate pro-
binding / domain
site duced from glycogen is converted to glucose-6-phosphate. In
muscle cells, this metabolite enters the glycolytic pathway and
is metabolized to generate ATP for use in powering muscle
Catalytic
subunit
contraction (see Chapter 12). Unlike muscle cells, liver cells
binding contain a phosphatase that hydrolyzes glucose-6-phosphate to
site glucose, which is exported from these cells mainly by a glucose
I transporter (GLUT2) in the plasma membrane (see Chapter
11 ). Thus glycogen stores in the liver are primarily broken
down to glucose, which is immediately released into the blood
and transported to other tissues, particularly the muscles and
brain, to nourish them.

FIGURE 15-29 Struct ure of protein kinase A (PKA) and its

activation by cAMP. (a) Protein kinase A (PKA) consists of two
regulatory (R) subunits (green) and two catalytic (C) subunits. When
cAMP (red triangle) binds to the regulatory subunit, the catalytic
subunit is released, thus activating PKA. (b) The two regulatory
subunits form a dimer, joined by a dimerization/docking domain and a
flexible linker to which A-kinase-activating protein (AKAP; see Figure
15-33) can bind. Each R subunit has two cAMP-binding domains, CNB-A
cAMP
bound and CNB-8, and a binding site for a catalytic subunit (arrow). (c) Binding
cAMP of cAMP to the CNB-A domain causes a subtle conformational change
that displaces the catalytic subunit from R, leading to its activation.
cAMP-binding Without bound cAMP, one loop of the CNB-A domain (purple) is in a
domains conformation that can bind the catalytic (C) subunit. A glutamate
(E200) and arginine (R209) residue participate in binding of cAMP (red),
which causes a conformational change (green) in the loop that
prevents binding of the loop to the C subunit. [Part (b) after S. S. Taylor
et al., 2005, Biochim. Biophys. Acto 1754 :25; part (c) after C. Kim, N.H. Xuong, and
5. 5. Taylor, 2005, Science 307:690.}

15.5 G Protein - Coupled Receptors That Activate or Inhibit Adenylyl Cyclase 701
 0 0 +
II II
0- P- 0 - P-0-jUridinel 0 0 ...
I
OH 0 0 OH OH
UDP-glucose Glycogen (n residues)

1
Glycoge'l
synt h~•A

0 0 +
II
0-P-0-P-O-jUridinel 0
I I
o o- OH
Glycogen In + 1 residues)

i
UDP

Pi Glycogen
r >sphorylase

+
0 o ...
OH OH
Glucose-1-phosphate Glycogen (n residues)

FIGURE 15- 30 Synthesis and degradation of glycogen. Incorpo- catalyzed by glycogen phosphorylase. Because two different enzymes
ration of glucose from UDP-glucose into glycogen is catalyzed by catalyze the formation and degradation of glycogen, the two reactions
glycogen synthase. Removal of glucose units from glycogen is can be independently regulated.

The epinephrine-stimulated activation of adenylyl cy- phorylase (GP). As a consequence, the synthesis of glycogen
clase, resulting increase in cAMP, and subsequent activation by glycogen synthase is enhanced and the degradation of gly-
of protein kinase A (PKA) enhances the conversion of glyco- cogen by glycogen phosphorylase is inhibited.
gen to glucose-1-phosphate in two ways: by inhibiting glyco- Epinephrine-induced glycogenolysis thus exhibits dual
gen synthesis and by stimulating glycogen degradation regulation: activa tion of the enzymes catalyzing glycogen
(Figure 15-31 a). PKA phosphorylates and in so doing inacti- degradation and inhibition of enzymes promoting glycogen
vates glycogen synthase (GS), the enzyme that synthesizes synthesis. Such dual regulation provides an efficient mecha-
glycogen. PKA promotes glycogen degradation indirectly by nism for regulating a particular cellular response and is a
phosphorylating and thus activating an intermediate kinase, common phenomenon in cell biology.
glycogen phosphorylase kinase (G PK), that in turn phos-
phorylates and activates glycogen phosphorylase (GP), the cAMP-Mediated Activation of Protein
enzyme that degrades glycogen. These kinases are counter-
Kinase A Produces Diverse Responses
acted by a phosphatase called phosphoprotein phosphatase
(P P). At high cAMP levels, PKA phosphorylates an inhibitor in Different Cell Types
of phosphoprotein phosphatase (IP), which keeps this phos- In adipose cells, epinephrine-induced activation of protein
phatase in ItS inactive state (see Figure 15-3Ja, right). kinase A (PKA) promotes phosphorylation and activation of
The entire process is reversed when epinephrine i~ re- the lipase that hydrolyzes stored triglycerides to yield free
moved and the level of cAMP drops, in activating protein fatty acids and glycerol. These fatty acids are released into the
kinase A (PKA). When PKA is inactive, it can no longer blood and taken up as an energy source by cells in other tis-
phosphorylate the inhibitor of phosphoprotein phosphatase sues such as the kidney, heart, and muscles (see Chapter 12).
(IP), so this phosphatase becomes active (figure 15-31 b). Therefore, activation of PKA by epinephrine in two different
Phosphoprotein phosphatase (PP) removes the phosphate cell types, liver and adipose, has different effects. Indeed,
residues previously added by PKA to glycogen synthase (GS), cAMP and PKA mediate a large array of hormone-induced
glycogen phosphorylase kinase (GPK), and glycogen phos- cellular responses in numerous tissues (Table 15-2).

702 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 (a) Increased cAMP
Stimulation of Inhibition of
glycogen breakdown glycogen synthesis

GPK a0 GS {f) Inhibition of

phosphoprotein

_
l ,~
phosphatase

Glycogen+ n Pi
r n Glucose-1-phosphate

(b) Decreased cAMP

Stimulation of
glycogen synthesis

Abbreviations:

PKA Protein kinase A

PP Phosphoprotein phosphatase
GPK Glycogen phosphorylase kinase
GP Glycogen phosphorylase
GS Glycogen synthase
IP Inhibitor of phosphoprotein
UDP-glucose ---~ Glycogen+ UDP phosphatase

FIGURE 15-31 Regulation of glycogen metabolism by cAMP and phosphatase (PP). Binding of the phosphorylated inbibitor to PP
PKA. Active enzymes are highlighted in darker shades; inactive forms, prevents this phosphatase from dephosphorylating the activated
in lighter shades. (a) An increase in cytosolic cAMP activates protein enzymes in the kinase cascade or the inactive glycogen synthase.
kinase A (PKA), wh ich inhibit s glycogen synthesis directly and (b) A decrease in cAMP inactivates PKA, leading to release of the active
promotes glycogen degradation via a protein kinase cascade. At high form of PP. The action of this enzyme promotes glycogen synthesis
cAMP, PKA also phosphorylates an inhibitor of phosphoprotein and inhibits glycogen degradation.

Although protein kinase A acts on different substrates in version of up to I 00 inactive Gas molecules tO the active form
different types of cells, it always phosphorylates a serine or before epinephrine dissociates from the receptor. Each active
threonine residue that occurs within the same sequence Ga;GTP, in turn, actiYates a single adenylyl cyclase molecule,
motif: X-Arg-(Arg!Lys)-X-(Serrfhr)-<1>, where X denotes any which then catalyzes synthesis of many cAMP molecules dur-
amino acid and <P denotes a hydrophobic amino acid. Other ing the time G,,.·GTP is bound to it.
serine/threonine kinases phosphorylate target residues within The amplification that occurs in such a signal transduction
other sequence motifs. cascade depends on the number of steps in it and the relative
concentrations of the various components. In the epinephrine-
Signal Amplification Occurs in the cAMP-Protein induced cascade shown in Figure 15-9, for example, blood lev-
Kinase A Pathway els of epinephrine as low as 10 10 M can stimulate liver
glycogenolysis and release of glucose. An epinephrine stimulus
We've seen that receptOrs such as the [3-adrenergic receptor are of this magnitude generates an intracellular cAMP concentra-
low-abundance proteins, typically present in only a few thou- tion of 10- 6 M, an amplification of 104 -fold. Because three
sand copies per cell. Yet the cellular responses induced by a more catalytic steps precede the release of glucose, another 10 4
hormone such as epinephrine can require production of large amplification can occur, resulting in a 108 amplification of the
numbers of cAMP and activated enzyme molecules per cell. As epinephrine signal. In striated muscle, the amplification is less
an example, following activation of G 11,-coupled receptors, the dramatic because the concentrations of the three successive en-
intracellular concentration of cAMP will rise to about 10 6 M; zymes in the glycogenolytic cascade-protein kinase A, glyco-
in a typical cell that is a cube - 15 IJ-m on a side, this comes to gen phosphorylase kinase, and glycogen phosphorylase-arc in
- 2 million molecules of cAMP produced per cell. Thus sub- a 1:10:240 ratio (a potential 240-foiJ maximal amplification).
stantial amplification of the signal is necessary in order for the
hormone to induce a significant cellular response. We have
CREB Links cAMP and Protein Kinase A
already seen how signal amplification occurs following photon
absorbance in rod cells. In the case of G protein-coupled hor- to Activation of Gene Transcription
mone receptOrs, signal amplification is possible in part because Activation of protein kinase A also stimulates the expresston of
both receptors and G proteins can diffuse rapidly in the plasma many genes, leading to long-term effects on the cells that often
membrane. A single epinephrine-GPCR complex causes con- enhance the short-term effects of activated protein kinase A.

15.5 G Protein-Coupled Receptors That Activate or Inhibit Adenylyl Cyclase 703

 .·

TABLE 15-2 Cellular Responses to Hormone-Induced Rise in cAMP in Various Tissues*

Tissue Horm one Inducing Rise in cAMP Cellular Respon se

Adipose Epinephrine; ACTH; glucagon Increase in hydrolysis of triglyceride; decrease in ammo acid uptake

L1ver Epinephrine; Increase in conversion of glycogen to glucose; inhibition

norepinephrine; glucagon of glycogen synthesis; increase in amino acid uptake; increase
in gluconeogenesis (synthesis of glucose from amino acids)

Ovarian follicle FSH; LH Increase in synthesiS of estrogen, progesterone

Adrenal cortex ACTH Increase in synthesis of aldosterone, cortisol

Cardiac muscle Epinephnne Increase in contraction rare

Thyroid gland TSH Secretion of thyroxine

Bone Parathyroid hormone Increase in resorption of calciUm from bone

Skeletal muscle Epinephrine Conversion of glycogen to glucose-1-phosphate

Intestine Fpinephrine Fluid secretion

Kidney Vasopressin Resorption of water

Blood platelets Prostaglandin I Inhibinon of aggregation and secretion

"Nearly all the effects of cAMP are mediated through prorein kinase A (PKA ), which is activated b) binding of cAMP.
\OURCE: E. \X. Sutherland, 19~2, SCience 177:40 I.

For instance, in liver cells, protein kinase A induces expression localizes isoforms of protein kinase A (PKA) to specific subcel-
of several enzymes involved in gluconeogenesis-the conversion lular locations, thereby restricting cAMP-dependent responses
of three-carbon compounds such as pyruvate (sec Figure 12-3) to these locations. These proteins, referred to as A kinase-
to glucose-thus. increasing the level of glucose in the blood. associated proteins (AKAPs), have a two-domain structure
All genes regulated by protein kinase A contain a cis-acting with one domain conferring a specific subcellular location and
DNA sequence, the cAM P-res{Jonse element ( CRE), that another that binds to the regu latory (R) subunit of protein ki-
binds the phosphorylated form of a transcription factor called nase A (see Figure 15-29b).
CRE-binding (CREE) protein, which is found only in the nu- One such anchoring protein (AKAP15) is tethered to the
cleus. Following the elevation of cAMP levels and the release cytosolic face of the plasma membrane near a particular type
of the active protein kinase A catalytic subunit, some of the of gated Ca 2 channel in certain heart musc le cells. In the
catalytic subunits then translocate to the nucleus. There they heart, activation of 13-adrenergic receptors by epinephrine
phosphorylate serine-133 on the CREB protein. Phosphory- (as parr of the fight-or-flight response) leads to PKA-caralyzed
lated CREB protein binds ro CRE-containing target genes and phosphorylation of these Cal+ channels, causing them to
also binds to a co-activator termed CBP/300. CBP/300 links open; the resulting influx of Ca 2 ' increases the rate of heart
CREB to RNA polymerase 2 and other gene regulatory pro- muscle contraction. The binding of AKAP15 to protein ki-
teins, thereby stimulating gene transcription (Figure 15-32). nase A localizes the kinase next to these channels, thereby ,'
Thus protein kinase A phosphorylates multiple types of reducing the time that otherwise wou ld be requ ired for dif-
proteins: some have relatively short-term effects on cellular fusion of PKA catalytic subunits from their sites of genera-
metabolism, lasting seconds to minutes; other substrates tion to their Ca 2 -channel substrates.
such as CREB, by activating expression of specific genes, af- A different AKAP in heart muscle anchors both protein
fect cellular metabolism over hours and days. kinase A and cAMP phosphodiesterase (POE) the enzyme
that hydrolyzes cAMP to AMP (see figure 15-26)-to the
outer nuclear membrane. Because of the close proximity of
Anchoring Proteins Localize Effects of cAMP
PDE to protein kinase A, negative feedback provides tight
to Specific Regions of the Cell local control of the cAMP concentration and hence loca l
In many cell types, a rise in the cAMP level may produce a re- PKA activity (Figure 15-33). As cAMP levels rise in response
sponse that is required in one part of the cell but is unneeded, to hormone stimulation, PKA is activated. Activated PKA
perhaps deleterious, in another. A family of anchoring proteins phosphorylates PDE, which in turn becomes more active and

704 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

·.
~ OVERVIEW ANIMATION: Extracellular Signalinq

FIGURE 1 5- 32 Activation of CREB transcription factor following Adenylyl

ligand binding to G. protein-coupled receptors. Receptor stimula- cyclase
tion (0 ) leads to activation of protein kinase A (PKA) (f) ). Catalytic
subunits of PKA translocate to the nucleus (ID) and there phosphorylate Ext erior
and activate the CREB transcription factor(!)). Phosphorylated CREB
associates with the co-activator CBP/P300 ( ~ )and other proteins to ---+~D _____..
stimulate transcription of the various target genes controlled by the
Cytosol
CRE regulatory element. [SeeK. A. Lee and N. Masson, 1993, Biochim. Biophys.
Acta 1 174 :221, and D. Parker et al., 1996, Mol. Cell Bioi. 1 6(2):694.]
.. ... ...
\
cAMP

hydrolyzes cAMP, thu~ returning PKA to its inactive state.

The localization of protein kinase A near the nuclear mem-
brane also facilitates entry of its catalytic subunits into the
nucleus, where they phosphorylate and activate the CREB
transcription factor (see Figure 15-32).

Multiple Mechanisms Down-Regulate Signaling

from the GPCR/cAMP/PKA Pathway Nucleus

For cells to respond effectively to changes in their environ-

ment, they must not only activate a signaling pathway but
also down-modulate or terminate the response once it is no
longer needed; otherwise signal transduction pathways would
remain ''on" too long or at too high a level and the cell would
become overstimulated. Abnormal regulation of signaling
pathways is very common in cancer cells, where mutant pro-
teins that stimulate cell proliferation or prevent programmed
cell death remain active even when the signals that normally
activate them are absent.

D El II
Basal PDE activity= Increased cAMP: PDE phosphorylation
resting state PKA activation and activation; reduction
in cAMP level

~:~")ill}·~
Cytosol
..
.. .. .
..to nucleus

Outer
nuclear
membrane

IJ Return to resting state

FIGURE 15-33 Localization of protein k i nase A (PKA) to t he in excess of that which can be degraded by POE. The resulting binding
nuclear membrane in heart muscle by an A-kinase-associated of cAMP to the regulatory (R) subunits of PKA releases the active
protein (AKAP) . This member of the J\KAP family, designated mAKAP, catalytic (C) subunits 1nto the cytosol. Some C subunits enter into the
anchors both cAMP phosphodiesterase (POE) and the regulatory nucleus, where they phosphorylate and thus activate certain transcrip-
subunit (R, see Figure 1 5-29b) of PKA to the nuclear membrane, tion factors (see Figure 15-32). Other C subunits phosphorylate POE,
maintaining them in a negative feedback loop that provides close local stimulating its catalytic activity. Active POE hydrolyzes cAMP, thereby
control of t he cAMP level and PKA activity. Step 0 :The basal level of driving cAMP levels back to basal levels and causing re-formation of
POE activity in the absence of hormone (resting state) keeps cAMP the inactive PKA-R complex. Step !): Subsequent dephosphorylation of
levels below those necessary for PKA activation. Steps f) and ID: POE returns the complex to the resting state. [Adapted from K. L. Dodge
Activation of 13-adrenergic receptors causes an increase in cAMP level et al., 2001, EMBO J. 20:1921.]

15.5 G Protein-Coupled Receptors That Activate or Inhibit Adenylyl Cyclase 705

 Earlier we saw the multiple mechanisms that rapidly ter- conformation. This process is called homologous desensitiza-
minate the rhodopsin signal transduction pathway, includ- tion, because only those receptors that are in their active con-
ing GAP proteins that stimulate the hydrolysis of GTP bound formations are subject to deactivation by phosphorylation.
to G,m Ca 2 ~-sensing proteins that activate guanylate cyclase, We noted that binding of arrestin to extensively phosphor-
and phosphorylation of active rhodopsin by rhodopsin ki- ylated opsin completely inhibits activation of coupled G pro-
nase followed by binding of arrestin (see Figure 15-24). In teins by activated opsin (see Figure 15-24 ). In fact, a related
fact, most G protein-coupled receptors are modu lated by protein termed {3-arrestin plays a similar role in desensitizing
multiple mechanisms that down-regulate their activity, as is other G protein-coupled receptors, including (3-adrenergic re-
exemplified by j3-adrenergic receptors and others coupled to ceptors. An additional function of ~-arrestin in regulating cell-
G"' that activate adenylyl cyclase. surface receptors initially was suggested by the observation
that disappearance of (3-adrenergic receptors from the cell sur-
• First, the affinity of the receptor for its ligand decreases
face in response to ligand binding is stimulated by overexprcs-
when the GDP bound to G.., is replaced with GTP. This in-
sion of BARK and (3-arrestin. Subsequent studies revealed that
crease in the Kd of the receptor-hormone complex enhances
(3-arrestin binds not only to phosphorylated receptors but also
dissociation of the ligand from the receptor and thereby lim-
to clathrin and an associated protein termed AP2, two key
its the number of Gcx, proteins that are activated.
components of the coated vesicles that are involved in one
• Second, the intrinsic GTPase activity of G"' converts the type of endocytosis from the plasma membrane (Figure 15-34).
bound GTP to GDP, resu lting in inactivation of Gcx, and de- These interactions promote the formation of coated pits and
creased activation of its downstream target adenylyl cyclase. endocytosis of the associated receptors, thereby decreasi ng the
Importantly, the rate of hydrolysis of GTP bound to G"', is number of receptors exposed on the cell surface. Eventually
enhanced when G.,, binds to adcnylyl cyclase, lessening the some of the internalized receptors are degraded intracellularly,
duration of cAMP production; thus adenylyl cyclase func- and some are dephosphorylated in endosomes. Following dis-
tions as a GAP for G 0 , . More generally, hinding of most if sociation of (3-arrestin, the resensitized (dep~osphorylated) re-
not all G"·GTP complexes to their respective effector pro- ceptors recycle to the cell surface, similar to recycling of the
teins accelerates the rate of GTP hydrolysis. LDL receptor (see Chapter 14).
• Finally, cAMP phosphodiesterase acts to hydrolyze cAMP
to 5'-AMP, terminating the cellular response . Thus the con-
tinuous presence of hormone at a high enough concentration
NH 3+ .
is required for continuous activation of adenylyl cyclase and G protem-coupled receptor
maintenance of an elevated cAMP level. Once the hormone Ext erior
concentration falls sufficiently, the cellular response quickly
terminates.
Most GPCRs are also down-regulated by feedback re-
pression, a term describing the situation in which the end
product of a signaling pathway blocks an early step in the Cytosol
pathway. For instance, when a Gcx, protein-coupled receptor
is exposed to hormonal stimulation for several hours, several
serine and threonine residues in the cytosolic domain of the
receptor become phosphorylated by protein kinase A (PKA),
the end product of the Gcx, signaling pathway. The phosphor-
ylated receptor can bind its ligand but cannot efficiently acti-
vate Gu,; thus ligand hinding to the phosphorylated receptor
leads to reduced activation of adenylyl cyclase compared Activation of MAP
kinase cascade
with a nonphosphorylated receptor. Because the activity of
PKA is enhanced hy the high cAMP level induced by any FIGURE 15-34 Role of 13-arrestin in GPCR desensitization and
hormone that activates Gu., prolonged exposure to one such signal transduction.j3-Arrestin binds to phosphorylated serine and
threonine residues in the (-terminal segment of G protein-coupled
hormone, say, epinephrine, desensitizes not only (3-adrenergic
receptors (GPCRs). Clathrin and AP2, two other proteins bound by
receptors hut also other G"' protein-coupled receptors that
~-arrestin, promote endocytosis of the receptor.l3-arrestin also
bind different ligands (e.g., glucagon receptor in liver). This
functions in transducing signals from activated receptors by binding to
cross-regulation is called heterologous desensitization. and activating several cytosolic protein kinases. c-Src activates the MAP
Similar to phosphorylation of activated rhodopsin by rho- kinase pathway, leading to phosphorylation of key transcription factors
dopsin kinase, particular residues in the cytosolic domain of the (see Chapter 16). Interaction of fj-arrestin with three other proteins,
(3-adrenergic receptor, not those phosphorylated by PKA, can including JNK-3 (a Jun N-terminal kinase), results in phosphorylation
be phosphorylated by the related enzyme {3-adrenergic receptor and activation of another transcription factor, c-Jun. [Adapted from W.
kinase (BARK), but only when epinephrine or an agonist is Miller and R. J. Lefkowitz, 2001, Curr. Opin. Cell Bioi. 13:1 39, and K. Pierce et al.,
bound to the receptor and thus the receptor is in its active 2002, Nature Rev. Mol. Cell Bioi. 3:639.]

706 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

 In addition to its role in regulating receptor activity, [3-
arrestin functions as an adapter protein in transducing signals Localization of PKA to specific regions of the cell by an-
from G protein-coupled receptors to the nucleus (see Chapter choring proteins restricts the effects of cAMP to particular
16). The GPCR-arresrin complex acts as a scaffold for bind- subcellular locations (see Figure 15-33).
ing and activating several cytosolic kinases (see Figure 15-34), Signaling from G,-couplcd receptors is down-regulated by
which we discuss in detail in subsequent chapters. These in- multiple mechanisms: ( 1) the affinity of the receptor for its
clude c-Src, a cytosolic protein tyrosine kinase that activates ligand decreases when the GDP bound to G,, is replaced
the MAP kinase pathway and other pathways leading to with GTP; (2) the intrinsic GTPase activity of Gu, that con-
transcription of genes needed for cell divi.,ion (sec Chapter verts the bound GTP ro GDP is enhanced when G"' bmds to
19). A complex of three arrcstin-bound proteins, including a adenylyl cyclase (this occurs when many Ga·GTP complexes
Jun N-terminal kinase (JNK-3), initiates a kinase cascade that bind to their respective effector proteins); and (3) cAMP
ultimately activates the c-Jun transcription factor, which pro- fJhosphodiesterase acts to hydrolyze cAMP to 5'-AMP, ter-
mores expression of certain growth-promoting enzymes and minating the cellular response.
other proteins that help cells respond to stresses. Thus the
Most GPCRs are also down-regulated by feedback re-
BARK-[3-arrestin pathway, originally just thought ro sup-
pression, in which the end product of a pathway (e.g., PKA)
press signaling by GPCRs, actually functions as a switch,
blocks an early step in the pathway. As with opsin, binding of
turning off signaling by G proteins and turning on other sig-
[3-arresrin to phosphorylated [3-adrenergic receptors completely
naling pathways. The multiple functions of [3-arrestin illus-
inhibits activation of coupled G proteins (see Figure 15-24 ).
trate the importance of adapter proteins in both regulating
signaling and transducing signals from cell-surface receptors. • [3-adrenergic receptors are deactivated by [3-adrenergic ki-
nase (BARK), which phosphorylates cytosolic residues of the
receptor in irs active conformation. BARK phosphorylation
of ligand-bound [3-adrenergic receptors also leads to the bind-
KEY CONCEPTS of Section 15.5
ing of [3-arrestin and endocytosis of the receptors. The conse-
G Protein-Coupled Receptors That Activate quent reduction in the number of cell-surface receptors ren-
or Inhibit Adenylyl Cyclase ders the cell less sensitive to additional horrrione.
Ligand binding of G protein-coupled receptors that acti- The GPCR-arrestin complex functions as a scaffold that
vate Gas results in the activation of the membrane-bound activates several cytosolic kinases, initiating cascades that
enzyme adenylyl cyclase, which converts ATP to the second lead to transcriptional activation of many genes controlling
messenger cyclic AMP (cAMP; see Figure 15-26). Ligand cell growth (see Figure 15-34).
binding of G protein-coupled receptors that activate Ga., re-
sults in the inhibition of adenylyl cyclase and lower levels of
cAMP (see Figure 15-27).
• GasGTP and G 01 · GTP bind to the heterodimeric active
site domains in adenylyl cyclase to activate or inhihit the
15.6 G Protein-Coupled Receptors That
enzyme, respectively (sec Figure 15-28). Trigger Elevations in Cytosolic Ca2+
cAMP binds cooperative!} to a regulatory subunit of pro- Calcium ions play an essential role in regulating cellular re-
tein kinase A (PKA), releasing the active kinase catalytic sub- sponses to many signals, and many GCPRs and other types
unit (see Figure 15-29). of receptors exert their effects on cells by influencing the cy-
• PKA med iates the diverse effects of cAMP in most cells tosolic concentration of Ca 2 ... . As we saw in Chapter 11, the
(sec Table 15-2). The substrates for PKA, and thus the cel- level of Cal+ in the cytosol is maintained at a submicromolar
lu lar response to hormon~-induced activation of PKA, vary level (- 0.2 f..l-M) by the continuous action of ATP-powered
among cell types. Ca 2 pumps, which transport Cal+ ions across the plasma
membrane to the cell exterior or into the lumen of the endo-
• In liver and muscle cells, activation of PKA induced by
plasmic reticulum and other vesicles. Much intracellular
epinephrine and other hormones exerts a dual effect, inhibit-
Ca 2 is also sequestered in mitochondria.
ing glycogen synthesis and stimulating glycogen breakdown
A small rise in cytosolic Cal+ induces a variety of cellular
via a kinase cascade (see Figure 15-31 ).
responses, including hormone secretion by endocrine cells,
• The signal that activates the GPCR/adcnylyl cyclase/ secretion of digestive enzymes by pancreatic exocrine cells,
cAMP/PKA signaling pathway is amplified tremendously by and contraction of muscle (Tahle 15-3). For example, acetyl-
~t:wnd messengers and kinase cascades (see Figure 15-9). choline stimulation of GPCRs in secretory cells of the pan-
• Activation of PKA often leads to phosphorylation of nuclear creas and parotid (salivary) gland induces a rise in cytosolic
CREB protein, which together with the CBP/300 co-activator Ca 2 - that triggers the fusion of secretory vesicles w1rh the
stimulates transcription of genes, thus initiating a long-term plasma membrane and release of their protein contents into
change in the cell's protein composition (see Figure 15-32). the extracellular space. Thrombin, an enzyme in the blood-
clotting cascade, binds to a GPCR on blood platelets and

15.6 G Protein-Coupled Receptors That Trigger Elevations in Cytosolic Ca 2 • 707

 TABLE 15-3 Cellular Responses to Hormone-Induced Rise in Cytosolic CaZ+ in Various Tissues*

Tissue Ho rmone Inducing Rise in Ca2+ Cellular Response

Pancreas (acinar cells) Acetylcholine Secretion of digestive enzymes, such as amylase and trypsinogen

Parotid (salivary) gland Acetylcholine Secretion of amylase

Vascular or stomach Acetylcholine Contraction

smooth muscle

Liver Vasopressin Conversion of glycogen to glucose

Blood platelets Thrombin Aggregation, shape change, secretion of hormones

Mast cells Antigen Histamine secretion

Fibroblasts Peptide growth factors DNA synthesis, cell division (e.g., bombesin and PDGF)

• Hormone stimulauon leads to production of inositol 1,4,5-tnsphosphate (IP,). a second messenger that promotes release of Ca2 stored m the
endoplasmic reticulum.
~OURCE: M. J. Berridge, 1987, Allll. Rev. Bmchem. 56:159, and M. J. Berridge and R. 1-. Irvine, 1984, Nature 312:315.

triggers a rise in cytosolic Ca 2 ~ that, in turn, causes a confor- combined actions of various kinases and phosphatases dis-
mational change in the platelets that leads to their aggrega- cussed in Chapter 16. One derivative of PI, the lipid phospha-
tion, an important step in blood clotting to prevent leakage tidyl inositol 4,5-bisphosphate (PIP2 ), is cleaved by activated
of blood out of damaged blood vessels. phospholipase C into two important second messengers:
In this section, we first discuss an important signal trans- 1,2-diacylglycerol (DAG), a lipophilic molecule that remains
duction pathway that results in an elevation of cytosolic Ca2 + associated with the membrane, and inositol 1,4,5-trisphosphate
ions: the GPCR-stimulated activation of a phospholipase C (IP3 ), which can freely diffuse in the cytosol (Figure 15-35).
(PLC). Phospholipases C (PLCs) are a family of enzymes that We refer to downstream events involving these two second
hydrolyzes a phosphoester bond in certain phospholipids, messengers collectively as the IP/ DAG pathway.
yielding two second messengers that function in elevating the Phospholipase C is activated by G proteins containing
cytosolic Ca 2 + level and activating a family of kinases known either Gno or Guq subunits. In response to hormone activa-
as protein kinases C (PKCs); PK Cs in turn affect many im- tion of the GPCR, the Gno or G.,q subunits bound to GTP
portant cellular processes such as growth and differentiation. separate from Gp-y and bind to and activate phospholipase C
Some PLCs are activated by GPCRs, as we describe here; oth- in the membrane (Figure 15-36a, step 0 ). In turn, activated
ers, covered in the following chapter, are activated by other phospholipase C cleaves PIP 2 into DAG, which remains as-
types of receptors. Phospholipase Cs also produce second sociated with the membrane, and IP3, which freely diffuses in
messengers that are important for remodeling the actin cyto- the cytosol (Figure 15-36a, step 6 ). The two second mes-
skeleton (see Chapter 17) and for binding of proteins impor- sengers trigger separate downstream effects.
tant for endocytosis and vesicle fusions (see Chapter 14).
later in this section, we see how one PLC pathway leads to Ca2 + Release from the ER Triggered by IP 3 G protein-coupled
the synthesis of a gas, nitric oxide (NO), that in turn signals receptors that activate phospholipase C induce an elevation
adjacent cells. In the final part of the section, we will see how in cytosolic Ca 2 even when Ca 2 ions are absent from the
second messengers such as Ca2+ are used to help cells inte- surrounding extracellular fluid. In this case, Cal+ is released
grate their responses to more than one extracellular signal. into the cytosol from the ER lumen through operation of the
IP r gated Ca 2 channel in the ER membrane, as depicted in
Activated Phospholipase C Generates Two Key Figure 15-36a (steps IJ and 19). This large-channel protein is
composed of four identical subunits, each of which contains
Second Messengers Derived from the Membrane
an IPr binding site in the N-rerminal cyrosolic domain. IP 3
Lipid Phosphatidylinositol binding induces opening of the channel, allowing Ca 2 to
A number of important second messengers, used in several flow down its concentration gradient from the ER into the
signal transduction pathways, are derived from the mem- cytosol. When various phosphorylated inositols found in cells
brane lipid phosphatidy/inositol (PI). The inositol group in are added to preparations of ER vesicles, only IP 3 causes re-
this phospholipid, which always faces the cytosol, can be lease of Ca 2 + ions from the vesicles. This simple experiment
reversibly phosphorylated at one or more positions by the demonstrates the specificity of the IP3 effect.

708 CHAPTER 15 • Signal Transduction and G Protein - Coupled Receptors

 1 ,2-Diacylglycerol
(OAG)

C=O C=O c-o c=o C=O C=O C=O C=O

0
I
\
I
0
I ATP ADP
6\ 6I ATP ADP
I
0
\
I
0
I
I
0
\
I
0
I

_____;;;
CH-CH-CH CH-CH-CH CH-CH-CH CH-CH-CH
I 2 2 \. 2. I 2 2 \. 2 1 I 2 2
0 0 I'll"~ k• . . . . 0
I I I '
0-:~=0
, OH o-:~=o
~
{
0-P~=O
OH
0
o OH OH OH
~ 4 4

~
OH HO OH HO p OH HO
E OH
,'· ' I

Phosphatidylinositol PI 4-phosphate PI 4,5-bisphosphate

(PI) (PIP) (PIP2l

Inositol 1.4.5-
trisphosphate
IIP3)

FIGURE 1 S-35 Synthesis of second messengers DAG and IP3 Cleavage of PIP2 by phospholipase C yields the two important second
from phosphatidylinositol (PI). Each membrane-bound PI kinase messengers DAG and IP 3. [See A. Toker and L. C. Cantley, 1997, Nature
places a phosphate (yellow circles) on a specific hydroxyl group on the 387:673, and C. L. Carpenter and L. C. Cantley, 1996, Curr. Opin. Cell Bial. 8:1 53.]
inositol ring, producing the phosphorylated derivatives PIP and PIP2 .

The lPr mediated rise in the cytosolic Ca 2 level is tran- increase in Ca 2 + influx, establishing that these two protems
sient because Ca2 pumps located in the plasma membrane arc the key components of the store-operated Ca1 pathway.
and ER membrane actively transport CaH from the cytosol Continuous activation of certain G protein-coupled re-
to the cell exterior and ER lumen, respectively. Furthermore, ceptors induces rapid, repeated spikes in the level of cyto-
within a second of its generation, the phosphate linked to the solic Ca 2 .... These bursts in cyrosolic Ca 2 ... levels are caused
carbon-S of lPl (see Figure 15-35) is hydrolyzed, yielding by a complex interaction between the cytosolic Ca 2 concen-
inositol I ,4-bisphosphate. This compound cannot bind to tration and the IPrgated Ca 2 + -channel protein. The submi-
the IPrgated Ca 2• channel protein and thus does not stimu- cromolar level of cyrosolic Ca 2 + in the resting state potentiates
J
late ca~ release from the ER. opening of these channels by IP,, thus facilitating the rapid
Without some means for replenishing depleted stores of rise in cytosolic Ca 2 " following hormone stimulation of the
intracellular Ca 2 •, a cell would soon be unable to increase cell-surface G protein-coupled receptor. However, the higher
the cytosolic Ca 2 + level in response to hormone-induced IPl. cytosolic Ca 1 levels reached at the peak of the spike inhibit
Patch-clamping studies (see Figure 11-22) have revealed that IPr induced release of Ca 2 + from intracellular stores by de-
a plasma membrane Cal+ channel, called the store-operated creasing the affinity of the Ca 2 .. channels for IP,. As a result,
channel, opens in response to depletion of ER Ca 2 stores. the channels close, and the cytosolic Ca 2 ~ level drops rap-
Studies in which each potential channel protein was knocked idly. Thus cytosolic Ca 2 is a feedback inhibitor of the pro-
down one at a time with shRNAs established the identity of tein, the IP,-gated Ca 2 - channel, that when open triggers
this channel protein as Orai l. The Ca 1 '-sensing protein is elevation in cytosolic Ca2 • Calcium ion spikes occur in the
STI.M, a transmembrane protein in the endoplasmic reticu- pituitary gland cells that secrete luteinizing hormone (LH),
lum membrane (Figure 15-36b). An EF hand, similar to that which plays an important role in controlling ovulation and
in calmodulin (see Figure 3-31), on the luminal side of the thus female fertility. LH secretion is induced by binding of
ER membrane binds Cah when its level in the lumen is high. luteinizing hormone-releasing hormone (LHRH) to its G
As endoplasmic reticulum Ca 1 stores are depleted, the protein-coupled receptor<; on these cells; LHRH binding in-
STIM proteins lose their bound Ca 2 \ o ligomerize, and in an duces repeated Ca2+ spikes. Each Ca 2 spike induces exocy-
unknown manner relocalize to areas of the ER membrane tosis of a few LH-containing secretory vesicles, presumably
near the plasma membrane (Figure 15-36b, right). There the those close to the plasma membrane.
STlM CAD domains bind to and trigger opening of Orai 1,
allowing influx of extracellular Ca1 . Combined overexpres- DAG Activati on of Protein Kinase C After its formation by
sion of Orai and STlM in cultured cells leads to a marked phospholipase C-catalyzed hydrolysis of PIP2 , DAG remains

15.6 G Protein-Co upled Receptors That Trigger Elevations in Cytosolic Ca 2 • 709

0 FOCUS ANIMATION: Second Messengers in Signaling Pathways

(a) G protein-coupled recepto r (GPCR) Phospholipase C

0
0 0 0
Ext erior 00 0

Plasma
membrane
Cytosol

IP3-gated
Ca2· channel

ER membrane

ER lumen

Endoplasmic reticulum

(b) 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0
0 Ca2·· 0 0
0
0 0
0 0
0 0 0 0 0 0 0
0 0 0
Exterior
Plasma
membrane
Orai 1

Cytosol

ER lumen

0
Bound Ca2•

High [Ca2•] in ER lumen Low [Ca2•J in ER lumen

FIGURE 15-36 IP3/ DAG pathway and the elevation of cytosolic receptors, thereby altering their activity (step fJ ). (b) Opening of
Ca2+. (a) Opening of endoplasmic reticulum Ca 1 channels. This plasma membrane Ca 2 channels. Left: In the resting cell, Ca 1 levels in
pathway can be triggered by ligand binding to GPCRs that activate the endoplasmic reticulum lumen are high, and Ca 1 • ions (blue ci rcles)
either the G.,0 or G.,q alpha subunit leading to activation of phospholi- bind to the EF hand domains of the transmembrane STIM p roteins.
pase C (step 0 ). Cleavage of PIP2 by phosphol ipd~e C yields IP3 and Right: As endoplasmic reticulum CaH stores are depleted and Ca 2+
DAG (step f) ). After diffusing through the cytosol, IP3 interacts with and ions dissociate from the EF hands. STIMs undergo oligomerization and
opens Ca 2 • channels in the membrane of the endoplasmic reticulum relocalization to areas of the ER membrane near the plasma mem-
(step i) ), causing release of stored Ca 2 · ions into the cytosol (step [1 ). brane. There the STIM CRAC-activating domains (CAD, green) bind to
One of several cellular responses induced by a rise in cytosolic Ca 2 is and trigger opening of Orail, the store-operated Ca 1 channels in the
recruitment of protein kinase C (PKC) to the plasma membrane (step 1,'1). p lasma membrane, allowing influx of extracellular Ca 2· • [Adapted from
where it is activated by DAG (step rlt). The activated membrane- J. W. Putney, 1999, Proc. Nat'/. Acod. Sci. USA 96:14669; Y. Zhou, 2010, Proc. Not'/.
associated kinase can phosphorylate various cellular enzymes and Acad. Sci. USA 107:4896; and M. Cahalan, 2010, Science 130:43.]

710 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 associated with the plasma membrane. The principal func- surrounding the blood vessels that "feed" the heart muscle
tion of DAG is to activate a family of protein kinases col- itself, thereby increasing the diameter of the blood vessels
lectively termed protein kinase C (PKC). In the absence of and increasing the flow of oxygen-bearing blood to the heart
hormone stimulation, protein kinase Cis present as a soluble muscle. One of the most intriguing discoveries in modern
cytosolic protein that is catalytically inactive. A rise in the medicine is that NO, a toxic gas found in car exhaust, is in
cytosolic Ca 1 · level causes protein kinase C to translocate to fact a natural signaling molecule. •
the cytosolic leaflet of the plasma membrane, where it can
interact with membrane-associated DAG (see Figure 15-36a, Definitive evidence for the role of NO in inducing relax-
. ' steps l1 and m ). Activation of protein kinase C thus depend~ ation of smooth muscle came from a set of experiments in
on an increase of both Ca 1+ ions and DAG, suggesting an in- which acetylcholine was added to experimental preparations
teraction between the two branches of the IP/DAG pathway. of the smooth muscle cells that surround blood vessels. Direct
The activation of protein kinase C in different cells re- application of acetylcholine to these cells caused them to con-
sults in a varied array of cellular responses, indicating that it tract, the expected effect of acetylcholine on these muscle cells.
plays a key role in many aspects of cellular growth and me- But addition of acetylcholine to the lumen of small isolated
tabolism. In liver cells, for instance, protein kinase C helps blood vessels caused the underlying smooth muscles to relax,
regulate glycogen metabolism by phosphorylating and so in- not contract. Subsequent studies showed that in response to
hibiting glycogen synthase. Protein kinase C also phosphory- acetylcholine, the endothelial cells that line the lumen of blood
lates various transcription factors that in some cells activate vessels were releasing some substance that in turn triggered
genes necessary for cell division. muscle-cell relaxation. That substance turned out to be NO.
We now know that endothelial cells contain a G 0 protein-
The Ca2+ -Calmodulin Complex Mediates Many coupled receptor that binds acetylcholine and activates
phospholipase C, leading to an increase in the level of cyto-
Cellular Responses to External Signals solic Cal+. After Ca 2 + binds to calmodulin, the resulting
The ubiquitous small cytosolic protein calmodulin functions as complex stimulates the activity of NO synthase, an enzyme
a multipurpose switch protein that mediates many cellular ef- that catalyzes formation of NO from 0 2 and the amino acid
fects of Ca2 + ions. Binding of Cal+ to four sites on calmodulin arginine. Because NO has a short half-life (2-30 seconds), it
yields a complex that interacts with and modulates the activity can diffuse only locally in tissues from its site of synthesis.
of many enzymes and other proteins (see Figure 3-31). Because In particular, NO diffuses from the endothelial cell into
four Ca2 bind to calmodulin in a cooperative fashion, a small neighboring smooth muscle cells, where it triggers muscle
change in the level of cytosolic Ca2 leads to a large change in relaxation (Figure 15-37).
the level of active calmodulin. One well-studied enzyme acti- The effect of NO on smooth muscle is mediated by the
vated by the Ca2 • -calmodulin complex is myosin light-chain second messenger cGMP, which is formed by an intracellular
kinase, which regulates the activity of myosin and thus contrac- NO receptor expressed by smooth muscle cells. Binding of
tion in muscle cells (see Chapter 17). Another is cAMP phos- NO to the heme group in this receptor leads to a conforma-
phodiesterase, the enzyme that degrades cAMP to 5'-AMP tional change that increases its intrinsic guanylyl cyclase activ-
and terminates its effects. This reaction thus links Ca 2+ and ity, leading to a rise in the cytosolic cGMP level. Most of the
cAMP, one of many examples in which two second-messenger- effects of cGMP are mediated by a cGMP-dependent protein
mediated pathways interact to fine-tune a cellular response. kinase, also known as protein kinase G (PKG). In vascular
In many cells, the rise in cytosolic Cal+ following recep- smooth muscle, protein kinase G activates a signaling path
tor signaling via phospholipase C-generated IP 3 leads to ac- way that results in inhibition of the actin-myosin complex,
tivation of specific transcription factors. In some cases, relaxation of the cell, and so dilation of the blood vessel. In
Ca 2 + -calmodulin activates protein kinases that, in turn, this case, cGMP acts indirectly via protein kinase G, whereas
phosphorylate transcription factors, thereby modifying their in rod cells cGMP acts directly by binding to and opening
activity and regulating gene expression. In other cases, Ca 2 + - cation channels in the plasma membrane (see Figure 15-23).
calmodulin activates a phosphatase that removes phosphate
groups from a transcription factor, thus activating it. An im-
portant example of this mechanism involves T cells of the
Integration of Ca2+ and cAMP Second
immune system (see Chapter 23). Messengers Regulates Glycogenolysis
Just as no cell lives in isolation, no intracellular signaling path-
Signal-Induced Relaxation of Vascular Smooth way functions alone. All cells constantly receive multiple sig-
nals from their environment, including changes in hormone
Muscle Is Mediated by a Ca 2 + -Nitric Oxide-
levels, metabolites, and gases such as NO and oxygen; these
cGMP-Activated Protein Kinase G Pathway signals must be integrated. The breakdown of glycogen to glu-
1!;1 Nitroglycerin has been used for over a century as a cose (glycogenolysis) provides an excellent example of how
H treatment for the intense chest pain of angina. It was cells can integrate their responses to more than one signal. As
known to slowly decompose in the body to nitric oxide discussed in Section 15 .5, epinephrine stimulation of muscle
(NO), which causes relaxation of the smooth muscle cells and liver cells leads to a rise in the second messenger cAMP,

15.6 G Protein-Coupled Receptors That Trigger Elevations in Cytosolic Ca 2 • 711

FIGURE 1 5- 3 7 The Ca2+ /nitric oxi de (NO)/ lumen of
cGMP pathway and the relaxation of arterial blood vessel
smooth muscle. Nitric oxide is synthesized in
endothelial cells in response to activation of
Phospholipase fJ
acetylcholine GPCRs, phospholipase C, and the .____...:::c'-----' ~

c
subsequent elevation in cytosolic Ca 2 ~ Endothelial ilt!" 3

0
(steps 0--9). NO diffuses locally through tissues cells
and activates an intracellular NO receptor with
guanylyl cyclase activity in nearby smooth
muscle cells (1,1). The resulting rise in cGMP (nil
activates protein kinase G (fJ), leading to
relaxation of the muscle and thus vasodilation
(llJ). PP, = pyrophosphate. [See C. 5. Lowenstein
et al., 1994, Ann. Intern. Med. 120:227.)

Smooth muscle cells RELAXATION

OF MUSCLE CELL

which promotes glycogen breakdown (see Figure 15-31a). In prolonged contraction. Recall that phosphorylation by cAMP-
both muscle and liver cells, other second messengers also pro- dependent protein kinase A also activates glycogen phosphor-
duce the same cellular response. ylase kinase (see Figure 15-31 ). Thus this key regulatory
In muscle cells, stimulation by nerve impulses causes the enzyme in glycogenolysis is subject to both neural and hor-
release of Ca 1 + ions from the sarcoplasmic reticulum and an monal regulation in muscle (Figure 15-38a).
increase in the cytosolic Ca 1 ' concentration, which triggers In liver cells, hormone-induced activation of the effector
muscle contraction. The rise in cytosolic Ca 2 "'" also activates
glycogen phosphorylase kinase (GPK), thereby stimulating the
protein phospholipase C also regulates glycogen breakdown
by generating the second messengers DAG and IP 3 . As we
.·.
degradation of glycogen to glucose-1-phosphate, which fuels just learned, IP 1 induces an increase in cyrosolic Ca 1 , which

(a) Muscle cells (b) liver cells

Neural Hormonal
stimulation stimulation

t
Caz+ cAMP
t
I I
+ +
+ +
JGPKJ ._+- ~
J !
t I
FIGURE 1 5- 38 Integrated regulation of glycogenolysis ~ §]
by Ca2+ and cAMP/ PKA pathways. (a) Neuronal stimulation

l l
of striated muscle cells or epinephrine bind ing to {3-adrenergic
receptors on their surfaces leads to increased cytosolic
concentration of the second messengers Ca 2 or cAMP, respec-
tively. The key regulatory enzyme glycogen phosphorylase
kinase (GPK) is artivated by binding Ca 2 + ions and by phos-
phorylation by cAMP-dependent protein kinase A (PKA). (b) In
liver cells, hormonal stimulation of 13-ad renergic receptors
leads to increased cytosolic concentrations of cAMP and two
other second messengers, diacylglycerol (OAG) and inositol
1,4,5-trisphosphate (IP3). Enzymes are marked by white boxes. Protein kinase A GP Glycogen phosphorylase
GPK Glycogen phosphorylase kinase GS Glycogen synthase
( +) activation of enzyme activity; (-) inhibition.

712 CHAPTER 15 • Signal Transduction and G Protem-Coupled Receptors

 activates glycogen phosphorylase kinase as in muscle cells, lead-
ing to glycogen degradation. Moreover, the combined effect of • The Ca2+ -calmodulin complex regulates the acnv1ty of
DAG and increased Ca2 + activates protein kinase C (see Figure many different proteins, including cAMP phosphodiesterase
15-36). This kinase can phosphorylate glycogen synthase, and protein kinases and phosphatases that control the activ
thereby inhibiting the enzyme and reducing the rate of glycogen iry of various transcription factors.
synthesis. In this case, multiple intracellular signal transduction • Stimulation of acetylcholine G protein-coupled receptors
pathways are activated by the same signal (Figure 15-38b). on endothelial cells induces an increase in cytosolic Cal+ and
The dual regulation of glycogen phosphorylase kinase by subsequent synthesis of NO. After diffusing into surround-
Ca 1 ~ and protein kinase A in horh muscle and liver results ing smooth muscle cells, NO activates an intracellular gua-
from its multimeric subunit structure (al3)'8) 4 • The)' subunit nylate cyclase to synthesize cGMP. The resulting increase in
is the catalytic enzyme; the regulatory a and 13 subunits, cGMP leads to activation of protein kinase G, which triggers
which are similar in structure, are phosphorylated by protein a pathway resulting in muscle relaxation and vasodilation
kinase A; and the 8 subunit is the calcium sensor calmodulin. (see Figure 15-37).
Glycogen phosphorylase kinase is maximally active when
• Glycogen breakdown and synthesis is coordinately regu-
Ca2 + ions are bound to the calmodulin subunit and the a
lated by the second messengers Caz... and cAMP, whose lev-
subunit has been phosphorylated by protein kinase A. In
els are regulated by neural and hormonal stimulation (see
fact, binding of Ca 1 to the calmodulin subunit may bees-
Figure 15-38).
sential to the enzymatic activity of glycogen phosphorylase
kinase. Phosphorylation of the a and also the 13 subunits by
protein kinase A increases the affinity of the calmodulin sub-
unit for Ca 2 , allowing Cal+ ions to bind to the enzyme at
the submicromolar Ca 1 • concentrations found in noncon-
tracting cells. Thus increases in the cytosolic concentration
Perspectives for the Future
of Ca 2 · or of cAMP (or both) induce incremental increases
in the activity of glycogen phosphorylase kinase. As a result In this chapter, we focused primarily on signal transduction
of the elevated level of cytosolic Ca2+ after neuronal stimula- pathways activated by individual G protein-coupled receptors.
tion of muscle cells, glycogen phosphorylase kinase will be However, even these relatively simple pathways presage the
active even if it is unphosphorylated; therefore glycogen can more complex situation within living cells. Many G protein-
be hydrolyzed to fuel continued muscle contraction even in coupled receptors form homodimers or heterodimers with
the absence of hormone stimulation. other G protein-coupled receptors that bind ligands with
different specificities and affinities. Much current research is
focused on determining the functions of these dimenc recep-
·. tors in the body.
With -900 members in total, the G protein-coupled re-
KEY CONCEPTS of Section 1 5.6 ceptors represent the largest protein family in the human ge-
nome. Approximately half of these genes are thought to
G Protein-Coupled Receptors That Trigger
encode sensory receptors; of these the majority are in the
Elevations in Cytosolic Ca 2 +
olfactory system and bind odorants. Of the remaining G
• A small rise in cytosolic Ca2 .,. induces a variety of re- protein receptors, the natural ligand has not been identified
sponses in different cells, including hormone secretion, con- for many so-called orphan GPCRs-that is, putative GPCRs
traction of muscle, and platelet aggregation (see Table 15-3 ). without known cognate ligands. Many of these orphan re-
• Many hormones bind GPCRs coupled to G proteins con- ceptors are likely to bind heretofore unidentified s1gnaling
taining a Gao or Gaq subunit. The effector protein activated molecules, including new peptide hormones. G protein-
by GTP-bound Gao or Gaq is a phospholipase C enzyme. coupled receptors already comprise the targets of more than
30 percent of all approved therapeutic drugs, and therefore
• Phospholipase C cleaves a phospholipid known as PIP2 , gen-
orphan GPCRs represent a fruitful resource for drug discovery
erating two second messengers: diffusible IP 3 and membrane-
by the pharmaceutical industry.
bound DAG (see Figure 15-35).
One approach that has proved fruitful in identifying li-
• IP3 triggers opening of IPrgated Ca2+ channels in the en- gands of orphan GPCRs involves expressing the receptor genes
doplasmic reticulum and elevation of cytosolic free Ca2 • In in transfected cells and using them as a reporter system to de-
response to elevated cytosolic Ca 2 ·, protein kinase C is re- tect substances in tissue extracts that activate signal transduc-
cruited to the plasma membrane, where it is activated by tion pathways in these cells. This approach has already led to
DAG (see Figure 15-36a). stunning insights into human behavior. One example is two
• Depletion of ER Ca 2 + stores leads to opening of plasma novel peptides termed orexin-A and orexin-B (from the Greek
membrane store-operated Ca2 • channels and an influx of orexis, meaning "appetite") that were identified as the ligands
Ca2+ from the extracellular medium (see Figure 15-36b). for two orphan GPCRs. Further research showed that the
orexin gene is expressed only in the hypothalamus, the part

Perspectives for the Future 713

of the brain that regulates feeding. Injection of orexin into base of the brain and acts through growth hormone recep-
the brain ventricles caused animals to eat more, and expres- tors located on the liver. Is this an example of endocrine,
sion of the orexin gene increased markedly during fasting. paracrine, or aurocrine signaling? Why?
Both of these findings are consistent with orexin's role in 3. A ligand binds two different receptors with a Kd value of
increasing appetite. Strikingly, mice deficient for orexins suf- 10 ~ M for receptor l and a Kd value of 10 9 M for receptor 2.
fer from narcolepsy, a disorder characterized in humans by For which receptor does the ligand show the greater affinity?
excessive daytime sleepiness (for mice, nighttime sleepiness). Calculate the fraction of receptors that have a bound ligand
Moreover, very recent reports suggest that the orexin system (IRL]/Rr) in the case of receptor 1 and rt'ceptor 2 if the con-
is dysfunctional in a majority of human narcolepsy patients: centration of free ligand is 10 8 M.
orexin peptides cannot be detected in their cerebrospinal
4. To understand how a signaling pathway works, it often is
fluid (although there is no evidence of mutation in their
useful to isolate the cell-surface receptor and to measure the
orexin genes). These findings firmly link orexin neuropep-
activity of downstream effector proteins under different condi-
tides and their receptors ro both feeding behavior and sleep
tions. How could you use affinity chromatography to isolate a
in both animals and humans.
cell-surface receptor? With what technique could you measure
More recently a new neuropeptide, neuropeptide S, was
the amount of activated G protein {the GTP-bound form) in
identified as the ligand for another previously orphaned GPCR.
ligand-stimulated cells? Describe thJ! approach you would take.
Researchers then showed that this neuropeptide modulates a
number of biological functions, including anxiety, arousal, lo- 5. How do seven transmembrane domain G protein-coupled
comotion, and memory. One can only wonder about what receptors transmit a signal across the plasma membrane? In
other peptJJes and small-molecule hormones remain to be dis- your answer, include the conformational changes that occur
covered and the insights that study of these will provide for our in the receptor in response to ligand binding.
understandmg of human metabolism, growth, and behavior. 6. Signal-transducing trimeric G proteins consist of three
subunits designated a, ~'and 'Y· The G" su~>Unit is a GTPase
switch protein that cycles between active and inactive states
Key Terms depending on whether it is bound to GTP or to GDP. Review
the steps for ligand-induced activation of effector proteins
adenylyl cyclase 692 IP 3/DAG pathway 708 mediated by the trimeric G proteins. Suppose that you have
agonist 682 kinase 677 isolated a mutant G ... subunit that has an increased GTPase
antagonist 682 muscarinic acetylcholine activity. What effect would this mutation have on the G pro-
arrestin 697 receptors 693 tein and the effector protein?
autocrine 676 nitric oxide 71 I 7. Explain how FRET could be used to monitor the associa-
~-adrenergic receptors 687 paracrine 675 tion of G"' and adenylyl cyclase following activation of the
epinephrine receptor.
calmodulin 679 phosphatase 677
phospholipase C (PLC) 708 8. Which of the following steps amplify the epinephrine sig-
competition assay 682
nal response in cells: receptor activation of G protein, G pro-
cyclic AMP (cAMP) 679 protein kinase A (PKA) 701
tein activation of adenylyl cyclase (AC), cAMP activation of
desensitization 6 84 protein kinase C (PKC) 711 PKA, or PKA phosphorylation of glycogen phosphorylase
endocrine 675 protein kinase G (PKG) 711 kinase (GPK)? Which change will have a greater effect on
epinephrine 688 rhodopsin 694 signal amplification: an increase in the number of epineph-
glucagon 699 second messengers 6 7 4 rine receptors or an increase in the number of G"' proteins?
glycogenolysis 699 signal amplification 680 9. The cholera toxin, produced by the bacterium Vibrio
signal transduction 674 cholera, causes a watery diarrhea in infected individuals.
G protein-coupled receptors
What is the molecular basis for this effect of cholera toxin?
(GPCRs) 674 transducin 694
10. Both rhodopsin in vision and the muscarinic acetylcho-
GTPase superfamily 678 trimcric G proteins 679
line receptor system in cardiac muscle arc coupled to ion
hormone 673 channels via G proteins. Describe the similarities and differ-
ences between these two systems.
11. Epinephrine binds to both ~-adrenergic and a-adrenergic
Review the Concepts
receptors. Describe the opposite actions on the effector pro-
1. What common features are shared by most cell signaling tein, adcnylyl L)cla!>e, elicited by the binding of epinephnne
systems? to these two types of receptors. Describe the effect of adding
2. Signaling by soluble extracellular molecules can be classi- an agonist or antagonist to a ~-adrenergic receptor on the
fied as endocrine, paracrine, or autocrine. Describe how activity of adenylyl cyclase.
these three types of cellular signaling differ. Growth hor- 12. In liver and muscle, epinephrine stimulation of the
mone is secreted from the pituitary, which is located at the cAMP pathway activates glycogen breakdown and inhibits

714 CHAPTER 15 • Signal Transduction and G Protein-Coupled Receptors

 glycogen synthesis, whereas in adipose tiss ue, epinephrine 250
activates hydrolysis of triglycerides and in other cells causes ?:
a diversity of other responses. What step in the cAMP signal- :2:t; 200
ing pathways in these cells specifies the cell response? "'~ 150
13. Continuous exposure of a G,., protein-coupled receptor u"'
>
to its ligand leads to a phenomenon known as desemitiza- (.) 100
tion. Describe severa l molecular mechanisms for receptor >
>c::
desensitization . How can a receptor be reset to its original Q) 50
"0
4:
sensitized state? What effect would a mutant receptor lack-
ing serine or threonine phosphorylation sites have on a cell? GTP GTP+iso GTPyS GTP GTP+iso GTPyS
14. What is the purpose of A kinase-associated proteins Wild type Mutant
(AKAPs)? Describe how AKAPs work in heart muscle cells.
15. lnositoll,4,5-trisphosphate (IP 1) and diacylglycerol (DAG)
are second messenger molecules derived from the cleavage of b. In the transfected cells described in part (a), what
the phosphatidylinositol 4,5-bisphosphate (PlP2 ) by activated would you predict would be the cAMP levels in cells trans-
phospholipase C. Describe the role of IP3 in causing a rise in fccted with the wild-type G .., and the mutant G0 ,? What ef-
cytosolic Ca 2 concentration. How do cells restore resting lev- fect might this have on the cells?
els of cytosolic Cal-? What is the principal function of DAG? c. To further characterize the molecular defect caused by
16. In Chapter 3, the Kd of calmodulin's EF hands for bind- this mutation, the intrinsic GTPase activity present in both
ing Ca 2 is given as - 10 6 M. Many proteins have much wild-type and mutant Gu, was assayed. Assays for GTPase ac-
higher affinities for their respective ligands. Why is the spe- tivity showed that the mutation reduced the kmr-GTI' (catalysis
cific affini ty of calmodulin important for Ca 2 • signaling pro- rate constant for GTP hydrolysis) from a wild-type value of
cesses such as that initiated by production of IP1 ? 4.1 min 1 to the mutant value of 0.1 min 1• What do you
17. Most of the short-term physiological responses of cells conclude about the effect of the mutation on the GTPase activ-
to cAMP are mediated by activation of PKA. cGMP is an- ity present in the mutant G 0 , subunit? How tlo these GTPase
other common second messenger. What are the targets of results explain the adenylyl cyclase results shown in part (a)?
cGMP in rod and smooth muscle cells? 2. The phosphorylation of a protein can influence its abilil) to
interact with other proteins. These protein-protein interactions
play a fundamental role in signal transduction pathways, and
these interactions can be identified using numerous techniques,
including fluorescence energy transfer (see Figure 15-18). Pro-
tein kinase A (PKA ) has many substrates, one of which is gly-
cogen phosphorylase kinase, which has a multimeric (al3-yo)4
Analyze the Data
structure containing two regulatory subunits (cr_ and 13), the
1. Mutations in .trimcric G proteins can cause many diseases catalytic 'Y subunit, and the calcium sensor o subunit.
in humans. Patients with acromegaly often have pituitary tu - a. You are using fluorescence energy transfer to investi-
mors that oversecrete growth hormone (GH). GH-releasing gate PKA interactions with glycogen phosphorylase kinase
hormone (GHRH) stimulates GH release from the pituitary and have cloned eDNA fusion constructs for three of its four
by binding to GHRH receptors and stimulating adenylyl cy- different subunits ('Y, 13, and o), all containing a fluorescent
clase. Researchers wanted to know whether mutations in G.,, tag that when expressed excites at 480 nm and emits fluores-
played a roll in this condition. Cloning and sequencing of the cence at 535 nm. You also have eDNA encoding the catalytic
wild-type and mutant Gus'gene from normal individuals and domain of PKA fused to a tag that when expressed excites at
patients with the pituitary tumors revealed a missense muta- 440 nm and emits at 480 nm . ln the assay, if the PKA fusion
tion in the G., 5 gene sequence. protein interacts with one or more of the ragged glycogen
a. To investigate the effect of the mutation on G.., activity, phosphorylase kinase substrates, the transfer of energy from
.· wild-type and mutant G.,, cDNAs were transfected into cells the PKA tag excites the tag on the substrate, causing it to emit
that lack the G.,, gene. These cells express a 13radrenergic fluorescence at 535 nm, and this can be detected.
receptor, which can be activated by isoproterenol, a 13 2 - Liver cells arc transfected with the PKA fusion construct
adrenergic receptor agonist. Membranes were isolated from alone (control) or with the PKA fusion construct plus one of
transfected cells and assayed for adcnylyl cyclase activity in the three tagged glycogen pho~phorylasc kinase constructs
the presence of GTP or the hydrolysis-resistant GTP analog, and th en later treated with epinephrine. The fluorescence
GTP--yS. from the figure below, what do you conclude emissions at 535 nm, resulting from the four different trans-
about the effect of the mutation on G.,, activity in the pres- fection experiments, repeated three different times, are
ence of GTP alone compared with GTP--yS alone or GTP shown in the graph below. !.abel the four bars on the graph,
plus isoproterenol ( iso )? showing the emission of PKA by itself, PKA + the 'Y subunit,

Analyze the Data 715

PKA + the (3 subunit, and PKA + the 8 subunit. Explain
why there is only one major peak and why the values repre- c::
--- -----
.g
sented by the other three bars are not significantly different
~
from each other. c::
Q)
(.)
c::
0
(.)

0..
~
<X:
(.)

Time

·.
b. Would you expect lower or higher levels of PKA in
b. Which combination above would produce emission at
535 nm if the experiment was repeated but instead of epi- cells treated with cholera toxin? Explain how you came to
these conclusions.
nephrine you used dibutyryl-cAMP, which freely crosses the
plasma membrane?
c. As described above, there are two regulatory subunits
of glycogen phosphorylase kinase, both of which are subjected References
to post-translational modifications. If the gene encoding the a.
Signal-Transduction: From Extracellular Signal to Cellular Response
subunit contained missense mutations whereby during trans-
Cabrera-Vera, T . ..\1., et al. 2003. Insights Into G protein
lation all the serine, threonine and tyrosine residues were con-
structure, function, and regulation. Endocr. Rev. '24:765-781.
verted to some other amino acid, how would this affect the Grecco, H., M. Schmick, and P. Bastiaens. 2011. Signaling from
calcium sensor subunit of glycogen phosphorylase kinase? the living plasma membrane. Cell 144:897-909.
The activity of glycogen phosphorylase kinase in cells treated Kornev, A., and S. S. Taylor. 2010. Definmg the conserved
with epinephrine is shown in the following graph. internal architecture of a protein kinase. Biochim. Biophys. Acta
1804:440-444.
:-.-tanning, G., et al. 2002. Evolunon of protem kinase signaling
Q) from yeast to man. Trends Bwchem. Sci. 27:514-520.
(/)
Manning, G., et al. 2002. The protein kinase complement of the
"'
~> human genome. Science 298:1912-1934.
o.~
.s::: > Taylor, S. S., and A. Kornev. 2011. Protem kinases: evolution of
a.·~
(/) (.)
o<x: dynamic regulatory proteins. Trends Biochem. Sci. 36:65-77.
.s::: Q)
0.. (/) Vetter, I. R., and A. Wittinghofer. 2001. The guanine nucleotide-
c "'
a>.!: binding switch in three dimensions. Setence 294:1299-1304.
~:..!
~
6 Studying Cell-Surface Receptors and Signal Transduction Proteins
Gross, A., and H. F. l.odish. 2006. Cellular trafficking and
degradation of erythropoietin and NESP.]. Bioi. Chem. 281:
2024-2032.
Lauffenburger, D., and J. Lmderman. 1993. Receptors: models for
bindmg, traffickmg, and signaling. New York: Oxford Umversity Pres~.
Selinger, Z. 2008. Discovery of G protein signaling. Ann. Rev.
Draw what the activity would look like in comparison in Biochem. 77:1-13.
epinephrine-treated cells expressing the a. subunit containing Tarrant, ;\I., and P. Cole. 2009. The chem1cal biology of protem
the missense mutations described above. phosphorylation. Ann. Rev. Biochem. 78:797-825.
3. cAMP is a second messenger that regulates many diverse
cellular functions. In the intestinal lumen, cAMP is respon- G Protein-Coupled Receptors: Structure and Mechanism
sible for maintaining electrolyte and water balance. Certain B1rnbaumer, L. 2007. The discovery of signal transduction by
bacterial toxins, including one produced by Vibrio cholera, G proteins: a personal account and an overview of the initial
can upset the levels of cAMP, leading to fatal dehydr:ltion. findmgs and contrihurions thar led to our present under!>tanding.
Biochrm. Biophys. Acta 1768:756-Tl.
a. Given what you know about the mechanism of Vibrio
Oldham, W. M., and H. E. Hamm. 2008. Heterotrimeric G
cholera toxin, label the graph below showing cAMP concen-
protem activation by G-prorein-coupled receprors. Nat. Rev. Mol.
tration in ( 1) normal intestinal epithelial cells treated with a Cell Brol. 9:60-71.
GPCR agonist to activate G.., and (2) cholera toxin-treated Ro,enbaum, D., S. Ra.,mussen, and B. Kobilka. 2008. The
cells stimulated with the same GPCR agonist. Explain how structure and function of G-protein-coupled receptors Nature
you came to these conclusions. 459:356-363.

716 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

 Schwartz, T., and W. Hubbell. 2008. Structural biology: a Lefkowitz, R. J., and S. K. Shcnoy. 2005. Transduction of recep-
movmg story of receptors. Nature 454:4-3. tor stgnals by 13-arrestms. Scwrce 308:512-517.
Sprang, S. 2011. Cell signaling: binding the receptor at both Rajagopal, S., K. Ra,agopal, and R. j. Lcfkowttz. 20 I 0.
ends. Nature 469:1-2-1-3. Teachmg old receptors new tricks: btasing seven -transmembrane
Tesmer, J. 2010. The quest to understand heterommeric G receptors. Nat. Reu. Drug Disco!!. 9:373-386.
protem signalling. Nat. Struct. Mol. 810/. 17:650-652. Somsak, L., et al. 2008. New mhibirors of glycogen phosphorylase
Warne, T., et. a!. 2008 Strucrure of a 131-adrenergic G-protein- as potential antidiabetic agents. Curr. Med. Chem. 15:2933-2983.
coupled receptor. Nature 454: 486-491. Taylor, S. S., et al. 2005. Dynamics of s1gnalmg by PKA.
Brochim. Biophys. Acta 1754:25-37.
G Protein-Coupled Receptors That Regulate Ion Channels T~ylor, S S., et al. 2008. Signaling through cA\.1P and cA:V1P-
Burns, M., and V. Arshavsky. 2005. Beyond counting photons: dependenr protein kmase: diverse strategies for drug design.
trials and trends in vertebrate visual transduction. Neuron 48:38"'-40 I. Rtochrm. Bwphys. Acta 1784:16-26.
Calvert, P., et al. 2006. I ight-dnven translocation of signaling
proteins in vertebrate photoreceptors. Trends Cell Bioi. 16:560-568.
Hofmann, K. P., et al. 2009. A G protein-coupled recepror at G Protein-Coupled Receptors That Trigger Elevations
work: the rhodopsin model. Trends Biochem. Set. 34:540-552. in Cytosolic Ca 2 ~
Smith, S. 0. 2010. Structure and acnvation of the 'isual Cahalan, M. 2010. How to STIMulate calcium channels.
pigment rhodopsm. Amr. Rev. Biophys. 39:309-328. Science 130:43.
Chtn, D., and \. R. ~1eans. 2000. Calmodulin: a prororyptcal
G Protein-Coupled Receptors That Activate calcium sensor. Trends Cell Bioi. 10:322-328.
or Inhibit Adenylyl Cyclase Duda, T. 2009. Atrial natnuretic factor-receptor guanylate
Agius, L. 2010. Physiological control of liver glycogen metabo- cyclase signal transduction mechanism. Mol. Cell Biochem.
lism: lessons from novel glycogen phosphorylase inhibitors. 334:37-51.
Mint-Rev. Med. Chem. 10:11 75-1187. Hoeflich, K. P., and M. 1kura. 2002. Calmodulm in action:
Carnegie, G., C. Means, and J. Scott. 2009. A-kinase anchoring diversity in target recogmtion and activation mechanisms. Cell
protems: from protein complexes to physiology and disease. TUBMB 108:739-742.
I.ife 61(4):394-406. Hogan, P. G., R. S. Lewis, and A. Rao. 2010. Molecular basis
Dessauer, C. 2009. Adenylyl cyclase-A-kinase anchoring protein of calcmm stgnaling in lymphocytes: STI.\1 and ORAl. Ann. Rev.
complexes: the ne.x t dimension in cAMP signaling. Mol. Pharmacal. lmmrmol. 28:491-533.
76:935-941. Parekh, A. 2011. Decoding cyrosoltc Ca ' oscillanons. Trends
DeWire, S., et al. 2007. 13-Arresrins and cell signali ng. Ann. ReL•. Btochem. Sci. 36:78-87.
Physrol. 69:483-510. Zhou, Y., eta!. 20 I 0. Pore architecture of the ORA II store·
Johnson, L. N. 1992. Glycogen phosphorylase: control by operated calcium channel. Proc. Nat'/ Acad. Sci. USA 107:4896-
phosphorylation and allosteric effectors. FASEB]. 6:2274-2282. 4901.

References 717
-'----'
 CLASSIC EXPERIMENT 15.1

The Infancy of Signal Transduction-

GTP Stimulation of cAMP Synthesis
M. Rodbell et al., 1971,J. Bioi. Chem. 246:1877

n the late 1960s, the study of hormone The Experiment glucagon. This decreased affinity both
I action blossomed following the dis-
covery that cyclic adenosine monophos- One of Rodhell's first goals was to
affects the ability of glucagon to bind
to the receptor and encourages the dis-
phate (cAMP) functioned as a second characterize the binding of glucagon to sociation of bound glucagon.
messenger, coupling the hormone- the glucagon receptor in the cell-free The observation that GTP was in-
mediated activation of a receptor to a rat liver membrane system. First, puri- volved in the action of glucagon led to
cellular response. In setting up an ex- fied rat liver membranes were incu- a second key question: Can GTP also
perimental system to investigate the bated with glucagon labeled with the exert an effect on adenyl cyclase? Ad-
hormone-induced synthesis of cAMP, radioactive isotope of iodine ( 125 1). dressing this question experimentally
Membranes were then separated from required the addition of both ATP, as a
Martin Rod bell discovered an important
new player in intracellular signaling- e
the unbound 2 'Il glucagon by centrif- substrate for adenyl cyclase, and GTP,
guanosine triphosphate (GTP). ugation. Once it was established that as the factor being examined, to the
labeled glucagon would indeed bind to purified rat liver membranes. How-
the purified rat liver cell membranes, ever, the previous study had shown
the study went on to determine if this thar the concentration of ATP required
Background
binding led directly to activation of ad- as a substrate for adenyl cyclase could
The discovery of GTP's role in regulat- enyl cyclase and production of cAMP affect glucagon bindirig. Might it also
ing signal transduction began with in the purified rat liver cell membranes. stimulate adenyl cyclase? The concen-
studies on how glucagon and other hor- The production of cAMP in the tration of ATP used in the experiment
mones send a signal across the plasma cell-free system required the addition of could not be reduced because ATP was
membrane that eventually evokes a cel- ATP; the substrate for adenyl cyclase, readily hydrolyzed by ATPases present
lular response. At the outset of Rod- Mg2 +; and an ATP-regenerating system in the rat liver membrane. To get
bell's studies, it was known that binding consisting of creatine kinase and phos- around this dilemma, Rodbell replaced
of glucagon to specific receptor proteins phocreatine. Surprisingly, when the ATP with an AMP analog, 59-adenyl-
embedded in the membrane stimulates glucagon-binding experiment was re- imidodiphosphate (AMP-PNP), which
production of cAMP. The formation of peated in the presence of these addi- can be converted to cAMP by adenyl
cAMP from ATP is catalyzed by a tional factors, Rodbell observed a 50 cyclase yet is resistant to hydrolysis by
membrane-bound enzyme called adenyl percent decrease in glucagon binding. membrane ATPases. The critical ex-
cyclase. It had been proposed that the Full binding could be restored only periment now could be performed. Pu-
action of glucagon, and other cAMP- when ATP was omitted from the reac- rified rat liver membranes were treated
stimulating hormones, relied on addi- tion. This observation inspired an in- with glucagon both in the presence and
tional molecular components that vestigation of the effect of nucleoside absence of GTP, and the production of
couple receptor activation to the pro- triphosphates on the binding of gluca- cAMP from AMP-PNP was measured.
duction of cAMP. However, in studies gon to its receptor. It was shown that The addition of GTP clearly stimulated
with isolated fat-cell membranes known relatively high (i.e., millimolar) concen- the production of cAMP when com-
as "ghosts," Rod bell and his coworkers trations of not only ATP but also uri - pared to the addition of glucagon
were unable to provide any further in- dine triphosphate (UTP) and cytidine alone (Figure l ), indicating that GTP
sight into how glucagon binding leads triphosphate (CTP) reduced the bind- affects not only the binding of gluca-
to an increase in production of cAMP. ing of labeled glucagon. In contrast, the gon to its receptor but also stimulates
Rodbell then began a series of studies reduction of glucagon binding in the the activation of adenylyl cyclase.
with a newly developed cell-free system, presence of GTP occurred at far lower
purified rat liver membranes, which re- (micromolar) concentrations. More-
Discussion
tained both membrane-bound and over, low concentrations of GTP were
membrane-associated proteins. These found to stimulate the dissociation of Two key factors led Rodbell and his col-
experiments eventually led to the find- bound glucagon from the receptor. leagues to detect the role of GTP in sig-
ing that GTP is required for the gluca- Taken together, these studies suggested nal transduction, whereas previous
gon-induced stimulation of adenyl that GTP alters the glucagon receptor studies had failed to do so. First, by
cyclase. in a manner that lowers its affinity for switching from fat-cell ghosts to the rat

The Infancy of Signal Transduction-GTP Stimulation of cAMP Synthesis 719

FIGURE 1 Effect of GTP on glucagon-stimulated cAMP produc- 1000
tion from AMP-PNP by purified rat liver membranes. In the absence
of GTP, glucagon stimulates cAMP formation about twofold over the
basal level in the absence of added hormone. When GTP also is added,
cAMP production increases another fivefold. [Adapted from M. Rodbell
800
et al., 1971,). Bioi. Chern. 246:1877.]

a.. 600
~
<X:
0

0"'
E
0. 400
Glucagon

200 Basal

5 10 15 20
Minutes

liver membrane system, the Rodbell re- tions of other nucleoside triphosphates. ceptors are involved in the action of
searchers avoided contamination of their The possibility of contamination sug- many hormones as well as in a number
cell-free system with GTP, a problem as- gested to him that small concentrations of other biological activities, including
sociated with the procedure for isolating of GTP might exert large effects on glu- neurotransmission and the immune re-
ghosts. Such contamination would mask cagon binding and the stimulation of sponse. It is now known that binding
the effects of GTP on glucagon binding adenyl cyclase. of ligands to their cognate G protein-
and activation of adenyl cyclase. Second, This critica l series of experiments coupled receptors stimulates the associ-
when ATP was ·first shown to influence stimulated a large number of studies on ated G proteins to bind GTP. This
glucagon binding, Rodbell did not sim- t he role of GTP in hormone action, binding causes transduction of a signal
ply accept the plausible explanation that eventuall y leading to the discovery of G that stimulates adenyl cyclase to produce
ATP, the substrate for adenyl cyclase, proteins, the GTP-binding proteins that cAM.P and also desensitization of the re-
also affects binding of glucagon. Instead, couple certain receptors to the adenyl ceptor, which then releases its ligand.
he chose to test the effects on binding of cyclase. Subsequently, an enormous Both of these effects were observed in
the other common nucleoside triphos- family of receptors that require G pro- Rodbell's experiments on glucagon ac-
phates. Rod bell later noted that he knew teins to transduce their signa ls were tion. For these seminal observations,
commercial preparations of ATP often identified in eukaryotes from yeast to Rodbell was awarded the Nobel Prize in
are contaminated with low concentra- humans . These G protein-coupled re- Physiology or Medicine in 1994.

720 CHAPTER 15 • Signal Transduction and G Protein- Coupled Receptors

 CHAPTER

Signaling Pathways
That Control Gene
Expression

A molecular valentine-dimerized extracellular domain of the

epidermal growth factor receptor (red, yellow, and green) bound to
two molecules of epidermal growth factor (magenta). [Courtesy
Jiahai Shi.)

xtracellular signals can have both short- and long-term mediating critical aspects of development, metabolism, and

E effects on cells. Short-term effects are usually triggered

by modification of existing proteins or enzymes, as we
saw in Chapter 15 . Many extracellular signals also affect
movement, it is not surprising that mutations in such signal-
ing pathways cause many human diseases, including cancer,
diabetes, and immune disorders.
gene expression and thus induce long-term changes in cell Transcription of genes is influenced by chromatin structure,
function. Long-term changes include alterations in cell divi- epigenetic modifications to histones and other nuclear proteins,
sion and differentiation, such as occur during development and the cell's complement of transcription factors and other
and cell fate determination. The body's production of red proteins (see Chapter 7). These properties determine which
blood cells, white blood cells, and platelets in response to genes the cell can potentially transcribe at any given time; we
cytokines is a good example of signal-induced changes in think of these properties as the cell's "memory," determined by
gene expression that influence cell proliferation and differen- its history and response to previous signals. Importantly, many
tiation. Changes in gene expression also enable differenti- key regulatory transcription factors are held in an inactive state
ated cells to respond to their environment by changing their in the cytosol or nucleus and become activated only in response
shape, metabolism, or movement. In immune system cells, to external signals, thus inducing expression of a set of genes
for example, several hormones activate one type of tran- that are specific to this cell type.
scription factor (NF-KB) that ultimately impacts expression In this chapter, we explore the main signaling pathways
of more than 150 genes involved in the immune response to that cells use to influence gene expression. Jn eukaryotes,
infection. Given the extensive role of gene transcription in there are about a dozen classes of highly conserved cell-surface

OUTLINE

16.1 Receptors That Activate Protein 16.5 Signaling Pathways Controlled by

Tyrosine Kinases 723 Ubiquitinati on: Wnt, Hedge hog, and NF-KB 752

16.2 The Ras/MAP Kinase Pathway 734 16.6 Signaling Pathways Controlled by Protein
Cleavage: Notch/Delta, SREBP 760
16.3 Phosphoinositide Signaling Pathways 745
16.7 Integration of Cellular Responses to Multiple
16.4 Receptor Serine Kinases That Activate Smads 748 Signaling Pathways 765
receptors, and these activate several types of highly con- including transcription factors located in the cytosol (Figure
served intracellular signal transduction pathways. Many of 16-la, 0 ). Some receptor kinases also activate small GTP-
these pathways consist of multiple proteins, small intracel- binding "switch" proteins such as Ras (Figure 16-la, f)). Other
lular molecules, and ions such as Ca 2 T, which together form receptors, mainly the seven spanning receptors introduced in
a complex cascade. Given this complexity, cell signaling Chapter 15, activate the larger GTP binding G'-' proteins (fig-
can seem a daunting subject to learn for the first time; the ure 16-1 b). Both types of GTP-binding proteins can activate
many names and abbreviations of molecules found in each protein kinases that in turn phosphorylate multiple target pro-
pathway can indeed be challenging. The subject repays teins, including transcription factors. Many signal transduc-
careful study, however: when one becomes famlltar with tion pathways, such as those activated by Ras, involve several
these pathways, one understands in a profound way the kinases in which one kinase phosphorylates and thus activates
regulatory mechanisms that control a vast array of bio- (or occasionally inhibits) the activity of another kinase.
logical processes. ln yet other signaling pathways, binding of a ligand to a
For simplicity, signal transduction pathways can be receptor triggers disassembl y of a multiprotein complex in
grouped into several basic types, based on the sequence of in- the cytosol, releasing a transcription factor that then translo-
tracellular events. In one very common type of signal transduc- cates into the nucleus (Figure 16-lc). Finally, in the last com-
tion pathway (figure 16-la), ligand binding to a receptor mon type, proteolytic cleavage of an inhibitor or the receptor
triggers activation of a receptor-associated kmase. This kinase itself releases an active transcripti6n factor, which then trav-
may be an intrinsic part of the receptor protein or be tightly els into the nucleus (Figure 16-1 d). While every signaling
bound to the receptor. These kinases often directly phosphory- pathway has its own subtleties and distinctions, nearly every
late and activate a variety of signal transduction proteins, one can be grouped into one of these basic types.

(a) Receptor- (b) Cytosolic (c) Protein subunit (d) Protein

associated kinase dissociation cleavage
kinase

Exterior

~· fJ ~p
oW:_
II e§~r;
i" u~
~TP
l
~
\
~ I
I
I

•
I

Transcription
factor
Q~
V"t' 8 goa ~
Other
target
proteins
~-"
Oe~ \7'1-
+
IDP
I
\ I
\
\
\
I
I
I
I
I
I
I
I
I
I

I
I
I

==~..(:=-- I: -
I
I I
Cytosol I I I
I
I I
\ I I I
I I I I
I

,,
\ I I
\ I I
\ I
T
Nucleus '....... ...... ,"' ¥,

Gene activation
or repression: •
,r"r ~p.("'l'
~- ~ ~ ' .~J"~v~'
4
Representat ive RTKs GPCRs Wnt Notch/Delta
receptors and TGF-P receptors cAMP/PKA!CREB Hedgehog
pathways
Cytokine receptors NF-kB
JAK-STAT
Ras/MAP Kinase
FIGURE 16-1 Several common cell-surface receptors and signal activity of another kinase. Many of the kinases in these pathways
transduction pathways. (a) The cytosolic domains of many receptors phosphorylate multiple protein targets that can be different in
contain protein kinase domains or are tightly associated with a different cells, including transcription factors. (b) Other receptors,
cytosoltc ktnase; commonly the kinases are activated by ligand binding mainly the seven-spanning receptors, activate the larger GTP-binding
followed by receptor dimerization. Some of these kinases directly Ga proteins, which in turn activate specific kinases or other signaling
phosphorylate and activate transcription factors (0 ) or other signaling proteins. (c) Several signaling pathways involve disassembly of a
proteins. Many of these receptors also activate small GTP·binding multiprotein complex in the cytosol, releasing a transcription factor
"switch" proteins such as Ras (f) ). Many signal transduction pathways, that then translocates into the nucleus. (d) Some signaling pathways
such as those activated by Ras, involve several kinases in which one are irreversible; in many cases proteolytic cleavage of a receptor
kinase phosphorylates and thus activates (or occasionally inhibits) the releases an active transcription factor.

722 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 The pathways we discuss in this chapter have been con- phorylate specific tyrosine residues on target proteins, usu-
served throughout evolution and operate in much the same ally in the context of a specific linear sequence of amino
manner in flies, worms, and humans. The substantial homology acids in which the tyrosine is embedded. The phosphory-
exhibited among proteins in these pathways has enabled re- lated target proteins can then activate one or more signaling
searchers to study them in a variety of experimental systems. pathways. These pathways arc noteworthy because they
for instance, the secreted signaling protein Hedgehog (Hh) and regulate most aspects of cell proliferation, differentiation,
its receptor were first identified in Drosophila mutants. Subse- survival, and metabolism.
quently, the human and mouse homologs of these proteins There are two broad categories of receptors that activate
were cloned and shown to participate in a number of important tyrosine kinases: (1) those in which the tyrosine kinase en-
signaling events during differentiation, resulting in the discov- zyme is an intrinsic part of the receptor's polypeptide chain
ery that abnormal activation of the Hh pathway occurs in sev- (encoded by the same gene), called the receptor tyrosine ki-
eral human tumors. Such discoveries illustrate the importance nases (RTKs), and (2) those, such as cytokine receptors, in
of studying signaling pathways both genetically-in tlies, mice, which the receptor and kinase are encoded by different genes
worms, yeasts, and other organisms-and biochemically. yet bound tightly together. for cytokine receptors, the tightly
... No signaling pathway acts in isolation. Many cells re- hound kinase is known as a JAK kinase. Both classes of re-
spond to multiple types of hormones and other signaling mol- ceptors activate similar intracellular signal transduction
ecules; some mammalian cells express - 100 different types of pathways, and we therefore con!>ider them together in this
cell-surface receprors, each of which binds a different ligand. section (Figure 16-2).
Since many genes are regu lated by multiple transcription fac- We will explore each of these important pathways in sub-
tors that in turn are activated or repressed by different intra- sequent sections of the chapter. In this section, we focus on
cellular signaling pathway<;, expression of any one gene can be the receptors themselves, showing how ligand bindmg leads
regulated by multiple extracellular signals. Especially during to kinase activation. We start with RTKs and then turn to
early development, such "cross talk" between signaling path- cytokine receptors. After discussing how both types of recep-
ways and the resultant sequential alterations in the pattern of tors are activated, we explore some of the downstream mol-
gene expression eventually can become so extensive that the ecules that are enlisted on activation. In th~ last part of the
cell assumes a different developmental fate. In this chapter, we section, we discuss how signaling from RTKs and cytokine
will see how multiple signaling pathways interact to regulate receptors is down-regulated.
crucial aspects of metabolism, such as the level of glucose in
the blood and the formation of adipose cells. Numerous Factors Regulating Cell Division
.• and Metabolism Are Ligands for Receptor
Tyrosine Kinases
16.1 Receptors That Activate Protein The signaling molecules that activate RTKs are soluble or
memhrane-bound peptide or protein hormones, including
Tyrosine Kinases
many that were initially identified as growth factors for spe-
We begin with a discussion of two large classes of receptors cific types of cells. These RTK ligands include many, such as
that activate protein tyrosine kinascs. Protein tyrosine kinases, nerve growth factor (NGF), platelet-derived growth factor
of which there are about 90 in the human genome, phos- (PDGF), fibroblast growth factor (FGf ), and cp1dermal

(al STAT - - - - - - - - - --+ Transcriptional activation

------~ ---------~

lbl GRB2 or She - + Ras - + MAP kinase - + Transcriptional activation or repression

H --+ Receptor
ormone tyrosine kinase

Transcriptional activation or repression

(cl Phospholipase c. - + Elevation of Ca2· - - - - +
M od ification of other cellular proteins
Cytokine-+ Cytokine- + JAK
receptor kinase
Transcriptional activation or repression
ldl Pl-3 kinase - + Protei n kinase B - - -+
M odification of other cellular proteins

FIGURE 16-2 Overview of signal transduction pathways t riggered (b) Binding of one type of adapter protein (GRB2 or She) to an activated
by receptors that activate protein tyrosine kinases. Both RTKs and receptor leads to activation of the Ras/MAP kinase pathway (see Section
cytokine receptors activate multiple signal transduction pathways that 16.2). (c, d) Two phosphoinositide pathways are triggered by recruitment
ultimately regulate transcription of genes. (a) In the most direct pathway, of phospholipase Cy and PI-3 kinase to the membrane (see Section 16.3).
mainly employed by cytokine receptors, a STAT transcription factor binds Elevated levels of Ca 2 and activated protein kinase B modulate the
to the activated receptor, becomes phosphorylated, moves to the activity of transcription factors as well as of cytosolic proteins that are
nucleus, and directly activates transcription (see Section 16.1 ). involved in metabolic pathways or cell movement or shape.

16.1 Receptors That Activate Protein Tyrosine Kinases 723

growth factor (EGF), that stimulate proliferation and dif- activity of an RTK is very low (see Figure 16-3, step 0 ). Like
ferentiation of specific cell types. Others, such as insulin, most other kinases, RTKs contain a flexible domain termed the
regulate expression of multiple genes that control sugar and activation lip. In the resting state, the activation lip is unphos-
lipid metabolism in liver, muscle, and ad ipose (fat) cells. phorylated and assumes a conformation that blocks kinase ac-
Many RTKs and their ligands were identified in studies of tivity. In some receptors (e.g., the insulin receptor), it prevents
human cancers associated with mutant forms of growth-factor binding of ATP. In others, (e.g., the FGF receptor), it prevents
receptors that stimulate proliferation even in the absence of binding of substrate. Binding of ligand causes a conformational
growth factor. The mutation "tricks" the receptor into be- change that promotes dimerization of the extracellular domains
having as though the ligand is present at all times anJ su the of RTKs, which brings their transmembrane segments-and
receptor is constantly in an active state (constitutively acti ve). therefore their cyrosolic domains-dose together. The kinase in
Other RTKs have been uncovered during analysis of develop- one subunit then phosphorylates a particular tyrosine residue in
mental mutations that lead to blocks in differentiation of cer- the activation lip in the other subunit (Figure 16-3, step f)).
tain cell types in C. elegans, Drosophila, and the mouse. This phosphorylation leads to a conformational change in the
activation lip that unblocks and thus activates kinase activity by
reducing the Km for ATP or the substrate to be phosphorylated.
Binding of Ligand Promotes Dimerization The resulting enhanced kinase activity can then phosphorylate
additional tyrosine residues in the cytosolic domain of the re-
of an RTK and Leads to Activation
ceptor (Figure 16-3, step IJ) as well as phosphorylate other tar-
of Its Intrinsic Kinase get proteins, leading to intracellular signaling.
All RTKs have three essential components: an extracellular do- Although dimerization is a necessary step in the activa-
main containing a ligand-binding site, a single hydrophobic tion of all RTKs, functional dimers can be formed in multiple
transmembrane a helix, and a cyrosolic segment that includes ways. Binding of EGF, for example, to its RTK triggers a
a domain with protein tyrosine kinase activity (Figure 16-3). conformational change in the receptor extracellular domain
Most RTKs are monomeric, and ligand binding to the extra- so that it "clamps" down on the ligand. This action pushes
cellular domain induces formation of receptor dimers. The for- out a loop located between the two EGF-binding domains,
mation of functional dimers is a necessary step in activation of and interactions between the two extended ("activated" )
all RTKs. We term this process of two (or more) receptors loop segments allow formation of the functional receptor
joining together "activation by receptor oligomerization." dimer (Figure 16-4 ). In other cases, such as the fibroblast
Such oligomerization of cell-surface receptors is, we will see, a growth factor (FGF) receptor, each of the two ligands binds
common mechanism for activating multiple types of receptors. simultaneously to the extracellular domains of two receptor
RTK activation can be summarized as follows: in the rest- subunits. FGF also binds tightly to heparan sulfate, a nega-
ing, unstimulated (no ligand bound) state, the intrinsic kinase tively charged polysaccharide component of some cell-surface

Ligand-
binding sites Bound ligand

Activation
lip

Cytosol Poorly active

protein tyrosine
kinase
D 6 ID
Receptor tyrosine Dimerization and Phosphorylation
kinases (RTKs) without phosphorylation of of additional
bound ligand activation lip tyrosines tyrosine residues

FIGURE 16-3 General structure and activation of receptor that then phosphorylate each other on a tyrosine residue in the activation
tyrosine kinases (RTKs). The cytosolic domain of RTKs contains an lip (f) ). Phosphorylation causes the lip to move out of the kinase catalytic
intrinsic protein tyrosine kinase catalytic site. In the absence of ligand (0 ), site, thus increasing the ability of ATP and the protein substrate to bind.
RTKs generally exist as monomers with poorly active kinases. Ligand The activated kinase then phosphorylates several tyrosine residues in the
binding causes a conformational change that promotes formation of a receptor's cytosolic domain (10). The resulting phosphotyrosines function
functional dimeric receptor, bringing together two poorly active kinases as docking sites for various signal transduction proteins.

724 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 (a) Exterior (a) Side view

EGF
EGF- 0
binding __L.
domains Heparan
sulfate

(b)

Membrane Membrane surface

(b) Top-down view

FIGURE 16-4 Ligand-induced dimerization of HER1, a human

receptor for epidermal growth factor (EGF). (a) Schematic depiction
ofthe extracellular and transmembrane domains of HER1, which is a
receptor tyrosine kinase. Binding of one EGF molecule to a monomeric
receptor causes an alteration in the structure of a loop between the
two EGF-binding domains. Dimerization of two identical ligand-bound
receptor monomers in the plane of the membrane occurs primarily
through interactions between the two "activated" loop segments.
(b) Structure ofthe dimeric HER1 protein bound to transforming growth
factor a (TGF-a), a member of the EGF family. The receptor's extracel- Heparan
lular domains are shown in blue; the transmembrane domain is shown sulfate
in red as an alpha helix, but its structure is not known in detail. The two
smaller TGFa molecules are colored green. Note the interaction FIGURE 16-S Structure of the fibroblast growth factor (FGF)
between the "activated" loop segments in the two receptor monomers. receptor, stabilized by heparan sulfate. Shown here are side
[Part (a) adapted from J. Schlessinger, 2002, Ce// 11 0:669; part (b) from T. Garrett and top-down views of the complex comprising the extracellular
et al., 2002, Ce// 11 0:763.) domains of two FGF receptor (FGFR) monomers (green and blue),
two bound FGF molecules (white), and two short heparan sulfate
chains (purple), which bind tightly to FGF. (a) In the side view, the
proteins and of the extracellular matrix (sec Chapter 20); this upper domain of one receptor monomer (blue) is seen situated
association enhances ligand binding and formation of a di- behind that of the other (green); the plane of the plasma membrane
meric receptor-ligand complex (Figure 16-5). The participa- is at the bottom. A small segment of the extracellular domain whose
tion of the heparan sulfate is essential for efficient receptor structure is not known connects to the membrane-spanning
activation. The ligands for some RTKs are dimeric, and their a-helical segment of each of the two receptor monomers (not shown)
binding brings two receptor monomers together directly. Yet that protrude downward into the membrane. (b) In the top view,
other RTKs, !.uch as the insulin receptor, form disulfide- the heparan sulfate chains are seen threading between and making
linked dimcrs even 111 the absence of hormone; binding of li- numerous contacts with the upper domains of both receptor
gand to this type of RTK alters its conformation in such a monomers. These interactions promote binding of the ligand to the
receptor and receptor dimerization. [Adapted from J. Schlessinger et al.,
way that the receptor kinase becomes activated. This last ex-
2000, Mol. Ce// 6:743.)
ample highlights that simply having two receptor monomers
in close contact is not sufficient for receptor activation-the

16.1 Receptors That Activate Protein Tyrosine Kinases 725

 Inactive (monomer or dimer) Asymmetric dimer Active dimer
EGF
0
Extracellular
EGF-binding
domains

M embrane
Juxtamembrane { JM-A
segment j
N lobe
Activation
Kinase~
domain L lip p
D Receiver Activator p
C-terminal tail

Autophospl'lorylation sites

FIGURE 16-6 Activation of EGF receptor by EGF results in t he activator kinase binds the juxtamembrane segment of the receiver
formation of an asym metric kinase d o main d imer.ln the inactive, kinase, causing a conformational change that removes the activation
monomeric state (0 ) the unstructured segment of the juxtamembrane lip from the kinase site of the receiver kinase, activating its kinase
domain (JM-B; green) binds to the upper, or N lobe of the kinase activity. (10) The active kinase then phosphorylates tyrosine residues
domain, causing a conformational change that positions the activation (yellow circles) in the (-terminal segments of the receptor cytosolic
lip in the kinase active site and thus inhibits kinase activation. Receptor domain. [After N. Jura et al., 2009, Cell 137:1293.]
dimerization generates an asymmetric kinase dimer (f)) such that the

proper conformational changes must accompany receptor Homo- and Hetero-oligomers of Epidermal
dimerization to lead to tyrosine kinase activation. Once an
Growth Factor Receptors Bind Members
RTK is locked into a functional dimeric state, its associated
of the Epidermal Growth Factor Superfamily
...
tyrosine kinase becomes activated.
Exactly how dimerization leads to kinase activation is un- Four receptor tyrosine kinases (RTKs) participate in signal-
derstood only for members of the EGF receptor family and ing by the many members of the epidermal growth factor
was uncovered through structural studies of the receptor cyto- (EGF) family of signaling molecules. ln humans, the four
solic domains in both active and inactive states. The kinase members of the HER (human epidermal growth factor re-
domains are separated from the transmembrane segment by a ceptor) family are denoted HERJ, 2, 3, and 4. HERl directly
so-called juxtamembrane segment, whose two parts are col- binds three EGF family members: EGF, heparin-binding
ored red and green in Figure 16-6. In the inactive, monomeric EGF (HB-EGF), and tumor-derived growth factor alpha
state, one part of the juxtamembrane segment binds to the (TGF-ct). Binding of any of these ligands to the extracellular
upper, or N, lobe of the adjacent kinase domain in the same domain of a HERl monomer leads to homodimerization of
molecule. This causes a conformational change such that the the HERl extracellular domain (Figure 16-7).
acti,·ation lip is localized in the active site of the kinase, block- Two other members of the EGF family, neuregulins I
ing its activity. In this way the kinase is maintained in the and 2 (NRGl and NRG2), bind to both HER3 and HER4;
"off" state (Figure 16-6, step 0). Receptor dimerization gen- HB-EGF also binds to HER4. Importantly, HER2 does not
erates an asymmetric kinase dimer (Figure 16-6, step f)) such directly bind a ligand but exists on the membrane in a preac-
that one kinase domain-termed the activator-binds the jux- tivated conformation with the loop segment protruding out-
ta membrane segment of the second kinase domain-the re- ward and the ligand-binding domains in close proximity
ceiver. This changes the conformation of theN lobe of the (Figure 16-7a). HER2, however, cannot form homodimers.
receiver, causing the activation lip to move out of the kinase It can signal only by forming heterocomplexes with ligand-
active site and allowing the kinase to function (step D ). In a bound HERl, HER3, or HER4. Thus it facilitates signaling
sense, an RTK can be thought of as an allosteric enzyme by all EGF family members (Figure 16-7b); an increase in
whose active site is inside the cell and whose allosteric effec- HER2 on the cell surface will make the cell more sensitive to
tor-the ligand-binds to an extracellular regulatory site on signaling by many EGF family members because the rate at
the enzyme. Evolution has produced many variations on the which the signaling heterodimers are formed after ligand
theme of this simple ligand-RTK mechanism, as is exemplified binding will be enhanced. Even though HER3 lacks a func-
by the families of EFG ligands and receptors discussed below. tional kinase domain, it can still participate in signaling;

726 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 {a) ... EGF
... HB-EGF
... TGF-a

L
Exterior

Cytosol

HER1 HER2 HER3 HER4

HER1 HER1 HER1 HER2 HER3 HER2 HER4 HER2

·. FIGURE 16-7 The HER family of receptors and their ligands. {b) Ligand-bound HER1 can form activated homodimers bound
Humans express four receptor tyrosine kinases-denoted HER1 , 2, 3, together by loop segments (red hooks), as detailed in Figure 16-4.
and 4-that bind epidermal growth factor {EGF) and other EGF family HER2 forms heterodimers with ligand-bound HER1, HER3, and HER4
members. (a) As shown, the HER proteins differentially bind EGF, and facilitates signaling by all EGF family members. HER3 has a very
heparin-binding EGf (HB-EGF), tumor-derived growth factor alpha poorly active kinase domain and can signal only when complexed with
(TGF-a), and neuregulins 1 and 2 {NRG 1 and NRG2). Note that HER2, HER2. [After N. E. Hynes and H. A. Lane, 2005, Nature Rev. Cancer 5:341 (erratum
which does not directly bind a ligand, exists in the plasma surface in Nature Rev. Cancer 5:580), and A. B. Singh and R. C. Harris, 2005, Cell Signal
membrane in a preactivated state indicated by a red hook. 17(0ct.):1183.]

after binding a ligand, it dlmerizes with HER2 and becomes gene occurs in approximately 25 percent of breast cancers,
phosphorylated by the HER2 kinase. This activates down- resulting in overexpression of HER2 protein in the tumor
stream signal transduction pathways as indicated below. cells. Breast cancer patients w ith HER2 overexpression have a
worse prognosis, including shortened survival, than do pa-
8 Understanding the HERs has helped explain why a par- tients without this abnormality. As Figure 16-7 emphasizes,
D ticular form of breast cancer is so dangerous and has overexprcssion of HER2 makes the tumor cells sensitive to
led to an important drug therapy. Breast cancer can involve growth stimulation by low levels of any member of the EGF
the abnormal growth of breast epithelial cells. Normal epi- family of growth factors, levels that would not stimulate pro-
thelial cells express a small amount of HER2 protein on their liferation of celb with normal HER2 levels. Discovery of the
plasma membranes in a tissue-specific pattern, and they do role of HER2 overexpression in certain breast cancers led re-
not grow inappropriately. In tumor cells, errors in DNA rep- searchers to develop monoclonal antibodies specific for the
lication often result in formation of multiple copies of a given HER2 protein. These have proved to be effective therapies for
gene on a single chromosome, an alteration known as gene those breast cancer patients in which HER2 is overcxprcssed,
amplification (see Chapter 24). Amplification of the HER2 reducing recurrence by about 50 percent in these patients. •

16.1 Receptors That Activate Protein Tyrosine Kinases 727

 ----,1
Hematopoietic stem cell functioning ofT cells and antibody-producing B cells of the
immune system. Another family of cytokines, the interferons,
! .___
I are produced and secreted by certain cell types following virus
infection and act on nearby cells to induce enzymes that ren-
der these cells more resistant to virus infection.
Progenitors of other
types of blood cells
•• Many cytokines induce formation of important types
of blood cells. All blood cells are derived from a com-
mon stem cell, which forms a series of progenitor cells that
then differentiate into the mature blood cells (Figure 16-8;
Erythroid progenitor (CFU-E)
see also Figure 21-18 ). For instance, the cytokine granulo-
No Epo I I...____,_!------,! Epo cyte colony stimulating factor (G-CSF) induces a granulocyte
~ progenitor cell in the hone marrow to divide several times
and then differentiate into granulocytes, the type of white

® 00
Apoptosis
1\ 1\
blood cell that inactivates bacteria and other pathogens. An-
other cytokine, erythropoietin (Epo), triggers production of
erythrocytes (red blood cells) by Inducing the proliferation
and differentiation of erythroid progenitor cells in the bone

aoaa
(cell death) marrow (Figure 16-8 ). Erythropoietin is synthesized by cer-
tain kidney cells. A drop in blood oxygen, such as caused by
loss of blood from a large wound, signifies a lower than op-

1\/\1\/\ timal level of erythrocytes, whose major function is to trans-

port oxygen complexed to hemoglobin. By means of the

00000000 oxygen-sensitive transcription factor HIF-1 a, the kidney

cells respond to low oxygen by synthesizing more erythro-
/1/111 1\1\ t'\~~ poietin and secreting it into the blood. As the level of eryth-

••••••••••••••••
~ ~ ~ ~ ~ ~ ~ i ~ ~ ~ ~ ~ ~ ~ ~
ropoietin rises, more and more erythroid progenitors are
induced to divide and differentiate; each progenitor pro-
duces -50 or so erythrocytes in a period of only a few days.
1111111111111111 Mature red cells
In this way, the body can respond to the loss of blood by
accelerating the production of erythrocytes. Both Epo and
GCSI- are produced commercially by recombinant expres-
FIGURE 16-8 Erythropoietin and format ion of red blood cells sion in cultured mammalian cells. Patients with kidney dis-
(erythrocytes). Er~throid progenitor cells, called colony-forming units
ease, especially those undergoing dialysis, frequently are
erythroid (CFU-E), are derived from hematopoietic stem cells, which
anemic (have a low red blood cell count) and therefore are
also give rise to progenitors of other blood cell types (see Figure 21-18).
treated with recombinant Epo to boost red cell levels. Epo
In the absence of erythropoietin (Epo). CFU-E cells undergo apoptosis.
Binding of Epo to its receptors on a CFU-E induces transcription of
and GCSF are used as adjuncts to certa in cancer therapies
several genes whose encoded proteins prevent programmed cell
since many cancer treatments affect the bone marrow and
death (apoptosis). allowing the cell to survive. Other Epo-induced reduce production of red cells and granulocytes. •
proteins trigger the developmental program of three to five terminal
cell divisions. If CFU-E cells are cultured with Epo in a semisolid Binding of a Cytokine to Its Receptor
medium (e.g., containing methylcellulose), daughter cells cannot move
Activates a Tightly Bound JAK Protein
away, and thus each CFU-E produces a colony of 30-100 erythroid cells;
hence its name. [SeeM. Socolovsky et al., 2001, 8/ood98:3261.) Tyrosine Kinase
All cytokines evolved from a common ancestral protein and
have a similar tertiary structure consisting of four long con-
served a helices folded together. Likewise, the various cyto-
Cytokines Influence Development
kine receptors undoubtedly evolved from a single common
of Many Cell Types ancestor since all cytokine receptors have similar structures.
The cytokines form a family of relatively small, secreted sig- Their extracellular domains are constructed of two subdo-
naling molecules (genera lly containing about 160 amino mains, each of which contains seven conserved 13 strands
acids) that control growth and differentiation of specific types folded together in a characteristic fashion. The interaction of
of cells. During pregnancy, for example, the cytokine prolac- one erythropoietin molecule with two identical erythropoie-
tm induces epithelial cells lining the immature ductules of the tin receptor (EpoR) proteins, depicted in 1-"igurc 16-9, exem-
mammary gland to differentiate into the acinar cells that pro- plifies the binding of a cytokine to its receptor.
duce milk proteins and secrete them into the ducts. Other cy- Cytokine receptors do not possess intrinsic enzymatic
tokines, the interleukins, are essential for proliferation and activit). Rather, a JAK kinase is tightly bound to the cytosolic

728 CHAPTER 16 • Signaling Pathways That Control Gene Expression

.·

Eryth ro po ieti n FIGURE 16-9 Structure of erythropoietin bound to an erythro-

poietin receptor. Erythropoietin (Epo) contains four conserved long
a helices that are folded in a particular arrangement. The activated
erythropoietin receptor (EpoR) is a dimer of identical subunits; the
extracellular domain of each monomer is constructed oftwo subdo-
mains each containing seven conserved 13 strands folded in a charac-
teristic fashion. Side chains of residues on two of the a helices in Epo,
termed site 1, contact loops on one EpoR monomer, while residues on
the two other Epo a helices, site 2. bind to the same loop segments in
a second receptor monomer, thereby stabilizing the dimeric receptor
in a specific conformation. The structures of other cytokines and their
receptors are similar to Epo and EpoR. [Courtesy Lucy Zhang; adapted from
R. 5. Syed et al., 1998, Nature 395:511 , and L. Zhang et al., 2009, Mol. Cell
33:266-274.]

late the other on a critical tyrosine in the activation lip (Figure

16-10, step f)). As with many other kinases, phosphorylation
coo coo- of the activation lip leads to a conformational change that
enhances the affinity for ATP or the substrate to be phos-
Membrane surface
phorylated, thereby increasing kinase activity (Figure 16-10,
step D ). One piece of evidence for this activation mechanism
domain of all cyrokine receptors (Figure 16-1 0). The four comes from study of a mutant JAK2 in which the critical
members of the JAK family of kinases contain an N-termi- tyrosine is mutated to phenylalanine. The mutant JAK2
nal receptor-binding domain, a C-terminal kinase domain binds normally to the EpoR but cannot be phosphorylated
that is normally poorly active catalytically, and a middle and is catalytically inactive. ln erythroid cells, expression of
domain that regulates kinase activity by an unknown mech- this mutant JAK2 in greater than normal amounts totall}
anism. (JAKs are so named because when they were cloned blocks EpoR signaling because the mutant JAK2 binds to the
and characterized, their function was unknown; they were majority of cytokine receptors, preventing binding and func-
termed just another kinase.) As in RTKs, thi s kinase be- tioning of the wild-type JAK2 protein. This type of muta-
comes activated after ligand binding and receptor dimeriza- tion, referred to as dominant-negative, causes loss of function
tion (Figure 16-10, step 0 ). even in cells that carry copies of the wild-type gene because
As a result of receptor dimerization, the associated JAKs the mutant protein prevents the normal protein from func-
are brought close enough together so that one can phosphory- tioning (see Chapter 5).

Ligand-
Lig~nd binding sites Bound ligand

Activation
lip
JAK Kinase
Cytosol

D f) !I
Cytokine receptors Dimerization and Phosphorylation
without bound ligand phosphorylation of of additional
activation lip tyrosines tyrosine residues

FIGURE 16-10 General structure and activation of cytokine brings together the associated JAK kinase domains, which then
receptors. The cytosolic domain of cytokine receptors binds tightly phosphorylate each other on a tyrosine residue in the activation lip (fl ).
and irreversibly to a JAK protein tyrosine kinase. In the absence of Downstream signaling !D l then proceeds in a manner similar to that
ligand (0 ), the receptors form a homodimer but the JAK kinases are from receptor tyrosine kinases.
poorly active. Ligand binding causes a conformational change that

16.1 Receptors That Activate Protein Tyrosine Kinases 729

Phosphotyrosine Residues Are Binding Surfaces the peptide-into a two-pronged "socket" in the SH2 do-
for Multiple Proteins with Conserved Domains main. The tvvo glutamic acids fit snugly onto the surface of the
SH2 domain between the phosphotyrosine socket and the hy-
Once the RTK kinases or JAK kinases become activated, they drophobic socket that accepts the isoleucine resid ue. This
first phosphorylate several tyrosi ne residues on the cytosolic specificity plays an important role in determini ng which signal
domain of the receptor (see Figures 16-3 and 16-1 0). Several transduction proteins bind to which receptors and so what
of these phosphotyrosine residues then serve as binding sites pathways are activated.
for proteins that have conserved phosphotyrosine-binding do- There are other small protein domains besides SH 2 that
mains. One such phosphotyrosine binding domain is called can recognize and bind to phosphotyrosine-conraining pep-
the SH2 domain. The SH2 domain derived its full name, the tides. One such domain is called the PTB (phosphotyrosine-
Src homology 2 domain, from its homology with a region in binding) domain. PTB domains are often found on so-called
the prototypical Src cytosolic tyrosine kinase encoded by the multidocking proteins, which serve as docking sites for other
src gene. (Src is an acronym for sarcoma, and a mutant form signal transduction proteins. For example, when several
of the cellular src gene was found in chickens with sarcomas, ·.
RTKs (e.g., the insulin receptor) and cytokine receptors (e.g.,
as Chapter 24 details.) The three-dimensional structures of the IL-4 receptor) are activated and tyrosine phosphorylated,
SH2 domains in different proteins are very similar, but each they bind a multidocking protein called IRS-1 (discovered
hinds to a distinct sequence of amino acids surrounding a because it is an insulin receptor Substrate) (Figure 16- 12 ).
phosphotyrosine residue. The unique amino acid sequence of The activated receptor then phosphorylates the bound dock-
each SH2 domain determines the specific phosphotyrosine ing protein, forming many phosphotyrosines that in turn
residues it binds (Figure 16-11). Variations in the hydropho- serve as docking sites for SH2-containing signaling proteins.
bic socket in the SH 2 domains of different signal transduction Some of these proteins in turn may also be phosphorylated
proteins allow them to bind to phosphotyrosines adjacent to by the activated receptor, and thus these multidocking pro-
different sequences, accounting for differences in their binding teins expand the number of intracellular signaling pathways
partners. The SH2 domain of the Src tyrosine kinase, for ex- that can be activated by the receptor.
ample, binds strongly to any peptide containing a critical four-
residue core sequence: phosphotyrosine-glutamic acid-glutamic
SH2 Domains in Action: JAK Kinases Activate
acid-isoleucine (Figure 16-11 ). These four amino acids make
intimate contact with the peptide-binding site in the Src SH2 STAT Transcription Factors
domain. Binding resembles the insertion of a two-pronged To illustrate how binding of SH 2 domains to specific phos-
"plug"-the phosphotyrosine and isoleucine side chains of phoryrosine residues induces specific signaling pathways, here

Top view Front view

FIGURE 16· 11 Surface model of an SH2 domain bound to a glutamic acid (Giu2)-isoleucine (lle3). Binding resembles the insertion
phosphotyrosine-containing peptide. The peptide bound by this of a two-pronged "plug·-the phosphotyrosine and isoleucine side
SH2 domain from Src tyrosine kinase (blue backbone with red oxygen chains of the peptide-into a two-pronged "socket" in the SH2 domain.
atoms) is shown in stick form. The SH2 domain binds strongly to The two glutamate residues are bound to sites on the surface of the
shorttarget peptides containing a critical four-residue core SH2 domain between the two sockets. [See G. Waksman et al., 1993,
sequence: phosphotyrosine (TyrO and OP03 )-glutamic acid (Giul )- Ce//72:779.]

730 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 Activated RTK ter introduction. Because different cell types have unique
complements of transcription factors and unique epigenetic
modifications on their chromatin, the genes that are available
to he actiYated by any STAT are also different. For example,
in mammary gland cells STAT5, the same STAT activated b)
the Epo receptor in erythroid cells, becomes activated follow-
ing prolactin binding to the prolactin receptor and induces
transcription of genes encoding certain milk proteins. In con-
trast, when STAT5 becomes auivated in erythroid progenitor
cells following binding of Epo to the Epo receptor, it induces
transcription of the Bcl-x 1 gene. Bcl-x1 prevents the pro-
grammed cell death, or apoptosis, of these progenitors, al-
lowing them to proliferate and differentiate into red blood
cells. Here we have a case of different cytokine receptors in
different cells activating the same intermediate signaling mol-
ecule, STAT5, yet leading to the activation of different genes.
Signaling proteins Combinatorial diversity allows a relatively limited set of sig-
FIGURE 16-12 Recruitment of intracellular signal transduction
naling pathways to control a vast array of cellular activities.
proteins to the cell membrane by binding to phosphotyrosine
residues in receptors or receptor-associated proteins. Cytosolic Multiple Mechanisms Down-Regulate Signaling
proteins with SH2 (purple) or PTB (maroon) domains can bind to
from RTKs and Cytokine Receptors
specific phosphotyrosine residues in activated RTKs (shown here) or
cytokine receptors. In some cases, these signal transduction proteins In the last chapter, we saw several ways in which signaling
then are phosphorylated by the receptor's intrinsic or associated from G protein-coupled receptors is terminated. For in-
protein tyrosine kinase, enhancing their activity. Certain RTKs and stance, phosphorylation of receptors and downstream sig-
cytokine receptors utilize multidocking proteins such as IRS-1 to naling proteins suppress signaling and this suppression can
increase the number of signaling proteins that are recruited and be reversed by the controlled action of phosphatases. Here
activated. Subsequent phosphorylation of a receptor-bound IRS-1 by we discuss several mechanisms by which RTK and cytokinc
the receptor kinase creates additional docking sites for SH2-containing
receptor signaling is regulated.
signaling proteins.

Receptor-M ediated Endocytosis Prolonged treatment of cells

with ligand often reduces the number of available cell-surface
we'll discuss the straightforward mechanism by which all JAK receptors such that the cells will have a less robust response
kinases and some RTKs directly activate members of the to exposure to a given concentration of ligand than they did
STAT family of transcription factors. All STAT proteins con- before the treatment. This desensitization response helps
tain an N-terminal DNA-binding domain, an SH2 domain prevent inappropriately prolonged receptor activity. In the
that binds to one or more specific phosphotyrosines in a cyto- absence of epidermal growth factor (EGF), for instance, cell-
kine receptor's cytosolic domain, and a C-terminal domain surface HER 1 receptors for this ligand are relatively long-
with a critical tyrosine residue. Once a monomeric STAT is lived, with an average half-life of 10 to 15 hours. Unbound
bound to the receptor via its SH2 domain, the C-terminal receptors arc internalized via clathrin-coated pits into cndo-
tyrosine is phosphorylated by an associated JAK kinase (Fig- somes at a relatively slow rate, on average once every 30 min-
ure 16-13a). This a rrangement ensures that in a particular utes, and often are returned rapidly to the plasma membrane
cell, only those STAT proteins with an SH2 domain that can so that there is little reduction in total surface receptor num-
bind to a particular receptor protein will be activated and only bers. Following binding of an EGF ligand, the rate of endocy-
when that receptor is activated. The erythropoietin receptor, tosis of HER 1 is increased ~ 10-fold, and only a fraction of
for example, activates STAT5 but not STATs 1, 2, 3, or 4; the internaliL.cd receptors return to the plasma membrane; the
these are activated by other receptors. A phosphorylated rest are degraded in lysosomes. Each time a HERl-EGF com-
STAT dissociates spontaneously from the receptor, and two plex is internalized, via the process termed receptor-mediated
phosphorylated STAT proteins form a dimer in which the endocytosis (see Figure 14-29), the receptor has about a 20 to
SH2 domain on each binds to the phosphotyrosine in the 80 percent chance of being degraded, depending on the cell
other. Because dimerization involves conformational changes type. Exposure of a fibroblast cell to high levels of EGF for
that expose the nuclear-localization signal (NLS), STAT di- several hours induces several rounds of endocytosis, resulting
mers move into the nucleus, where they bind to specific en- in degradation of most cell-surface receptor molecules and
hancers (DNA regulatory sequences) controlling target genes thus a reduction in the cell's sensitivity to EGF. In this way,
(Figure 16-1 3b) and thus alter gene expression. prolonged treatment with a given concentration of EGF de-
A given STAT can activate different genes in different sensitizes the cell to that level of hormone, though the cell may
cells depending on the "cell memory" discussed in the chap- respond if the level of EGF is increased.

16.1 Receptors That Activate Protein Tyrosine Kinases 731

 (a) Epo FIGURE 16- 13 Activation and structure of STAT proteins.
(a) Phosphorylation and dimerization of STAT proteins. Step 0 :
Following activation of a cytokine receptor (see Figure 16-1 0), an
inactive monomeric STAT transcription factor binds to a phosphotyro-
sine in the receptor, bringing the STAT close to the active JAK
associated with the receptor. The JAK then phosphorylates the
(-terminal tyrosine in the STAT. Steps D and ID: Phosphorylated STATs
spontaneously dissociate from the receptor and spontaneously
dimerize. Because the STAT homodimer has two phosphotyrosine-SH2
domain interactions, whereas the receptor-STAT complex is stabilized
by only one such interaction, phosphorylated STATs tend not to rebind
to the receptor. Step B :The STAT dimer moves into the nucleus, where
it can bind to promoter sequences and activate transcription of target
genes. (b) Ribbon diagram of the STATl dimer bound to DNA (black).
The STAT1 dimer forms a (-shaped clamp around DNA that is stabilized
by reciprocal and highly specific interactions between the SH2 domain
(purple) of one monomer and the phosphorylated tyrosine residue
(yellow w ith red oxygens) on the C-terminal segment of the other. The
phosphotyrosine-binding site of the SH2 domain in each monomer is
coupled structurally to the DNA-binding domain (magenta), suggest-
ing a potential role for the SH2-phosphotyrosine interaction in the
stabilization of DNA interacting elements. [Part (b) after X. Chen et al.,
El 1998, Ce/1 93:827.]

Into nucleus; /
binds DNA
and activates
transcription

(b)

SH2 domain
Ty rosine
PO.

,•
HERl mutants that lack kinase activity do not undergo nalized receptors can continue to signal from endosomes or
accelerated endocytosis in the presence of ligand. It is likely other intracellular compartments before their degradation,
that ligand-induced activation of the kinase activity in nor- as evidenced by their binding to signaling proteins such as
mal HER 1 induces a conformational change in the cytosolic Grb-2 and Sos, which arc discussed in the next section.
tail, exposing a sorting motif that facilitates receptor recruit-
ment into clathrin-coated pits and subsequent interna liza- lysosomal Degradation After internalization, some cell-
tion of the receptor-ligand complex. Despite extensive study surface receptors (e.g., the LDL receptor) are efficiently re-
of mutant HER 1 cytosolic domains, the identity of these cycled to the surface (see Figure 14-29). As noted above, the
"sorting motifs" is controversial, and most likely multiple fraction of activated HER 1 receptors that are sorted to lyse-
motifs function to enhance endocytosis. Interestingly, inter- somes can vary from 20 to 80 percent in different cell types.

732 CHAPTER 16 • Signaling Pathways That Control Gene Expression

There are several potential processes that can influence the (a) Short-term regulation: JAK2 deactivation
recycling versus lysosomal degradation fates of surface re- by SHP1 phosphatase
ceptors. One is covalent modification by the small protein Epo
ubiquitin (see Chapter 3). There is a strong correlation be-
tween monoubiquitination (addition of a single ubiquitin to
a given lysine of a protein) of the HERl cytosolic domain
and HER 1 degradation. The monubiquitination is mediated
by the enzyme c-Cbl. An E3 ubiquitin ligase (see Figure
3-29), c-Cbl contains an EGFR-binding domain, which
binds directly to phosphorylated EGF receptors, and a RING
finger domain, which recruits ubiquitin-conjugating enzymes
Active
and mediates transfer of ubiquitin to the receptor. The ubiq-
SHP1
uitin functions as a "tag" on the receptor that stimulates its
incorporation from endosomes into multivesicular bodies
(see Figure 14-33) that ultimately are degraded inside lyse-
somes. A role for c-Cbl in EGF receptor trafficking emerged Inactive
from genetic studies in C. elegans, which established that SHP1
c-Cbl negatively regulates the function of the nematode EGF
SH2 Phosphatase
receptor (Let-23), probably by inducing its degradation. domains domain
Similarly, knockout mice lacking c-Cbl show hyperprolifera-
tion of mammary gland epithelia, consistent with a role of (b) Long-term regulation: signal blocking and protein
c-Cbl as a negative regulator of EGF signaling. degradation by sacs proteins
Experiments with mutant cell lines demonstrate that inter-
nalization of RTKs plays an important role in regulating cel-
lular responses to EGF and other growth factors. For instance,
a mutation in the EGF receptor (HER I) that prevents it from
being incorporated into coated pits makes it resistant to
receptor-mediated (ligand-induced) endocytosis. As a result,
this mutation leads to substantially above-normal numbers of
EGF receptors on cells and thus increased sensitivity of cells
to EGF as a mitogenic signal. Such mutant cells are prone to
EGF-induced transformation into tumor cells (sec Chapter 24).
Interestingly, the other EGF family receptors-HER2, HER3, ... ...
...
and HER4-do not undergo ligand-induced internalization, sacs a;>
an observation that emphasizes how each receptor evolved to protein T \ ' ... , '..,. Recruitment
' , '..,. of E3
be regulated in its own appropriate manner. SH2 sacs ',..,. ubiquitin
domain box ligase
Phosphotyrosine Phosphatases These dephosphorylating
FIGURE 16-14 Two mechanisms for terminating signal transduc-
enzymes specifically hydrolyze phosphotyrosine linkages on
tion from the erythropoietin receptor (EpoR). (a) Short-term
specific target proteins. An excellen t example of how phos-
regulation: SHPl, a phosphotyrosine phosphatase, is present in an
photyrosinc phosphatase enzymes function to suppress the
inactive form in unstimulated cells. Binding of an SH2 domain in SHPl
activity of protein tyrosine kinases is provided by SHPl, the to a particular phosphotyrosine in the activated receptor unmasks its
phosphatase that negatively regulates signaling from several phosphatase catalytic site and positions it near the phosphorylated
types of cytokine receptors. Its role was first identified from tyrosine in the lip region of JAK2. Removal of the phosphate from
ana lysis of mice lacking this protein; they died because of ex- this tyrosine inactivates the JAK kinase. (b) Long-term regulation:
cess production of erythrocytes and several other types of SaCS proteins, whose expression is induced by STAT proteins in
blood cells. erythropoietin-stimulated erythroid cells, inhibit or permanently
SHPl dampens cytokine signaling by binding to a cyto- terminate signaling over longer time periods. Binding of SaCS to
kine receptor and inactivating the associated JAK protein, as phosphotyrosine residues on EpoR or JAK2 blocks binding of other
is depicted in Figure 16- 14a. In addition to a phosphatase signaling proteins (left). The sacs box can also target proteins such as
catalytic domain, SHP 1 has two SH2 domains. When cells JAK2 for degradation by the ubiquitin-proteasome pathway (right).
arc in the resting ~tate, unstimulated by a cytokine, one of Similar mechanisms regulate signaling from other cytokine receptors.
[Part (a) adapted from S. Constantinescu et al., 1999, Trends Endocrin. Metabol.
the SH2 domains in SHPl physically binds to and inactivates
10:1 8; part (b) adapted from B. T. Kile and W. S. Alexander, 2001, Cell. Mol.
the catalytic site in the phosphatase domain. In the stimu-
Life Sci. 58:1 .]
lated state, however, this blocking SH2 domain binds to a
specific phosphotyrosine residue in the activated receptor.
The conformational change that accompanies this binding

16.1 Receptors That Activate Protein Tyrosi ne Kinases 733

 ...

unmasks the SHPl catalytic site and also brings it adjacent

to the phosphotyrosine residue in the activation li p of the • Activation of an RTK leads to phosphorylation of the acti-
JAK associated with the receptor. By removing this phos- vation lip in the protein tyrosine kinases that are an intrinsic
phate, SHPl inactivates the JAK, so that it can no longer part of their cytoplasmic domains, enhancing their catalytic
phosphorylate the receptor or other substrates (e.g. STATs) activity (see Figure 16-3 ). The activated kinase then phos-
unless additional cytokine molecules bind to cell-surface re- phorylates tyrosine residues in the receptor cytosolic domain
ceptors, initiating a new round of signaling. and in other protein substrates.
• Humans express many RTKS, four of which (HERl-4)
SOCS Proteins In a classic example of negative feedback,
define the epidermal growth factor receptor family that
among the genes whose transcription is induced by STAT
mediates signaling from different members of the epidermal
proteins arc those encoding a class of small proteins termed
growth factor family of signaling molecules (see Figure 16-7).
SOCS proteins, which terminate signaling from cytokine re-
One of these receptors, HER2, does not bind ligand; it forms
ceptors. These negative regulators act in two ways (Figure
active heterodimers with ligand-bound monomers of the
16-14b). First, the SH2 domain in several SOCS proteins
other three HER proteins. Overexpression of HER2 is impli-
binds to phosphotyrosines on an activated receptor, prevent-
cated in about 25 percent of breast cancers.
ing binding of other SH2-containing signaling proteins (e.g.,
STATs) and thus competitively inhibiting receptor signaling. • Cytokines play numerous role~ in development. Erythro-
One SOCS protein, SOCS-1, also binds to the critical phos- poietin, a cytokine secreted by kidney cells, promotes prolif-
photyrosine in the activation lip of activated JAK2 kinase, eration and differentiation of erythroid progenitor cells in
thereby inhibiting its catalytic activity. Second, all SOCS the bone marrow (see Figure 16-8) to increase the number of
proteins contain a domain, called the SOCS box, that re- mature red cells in the blood.
cruits components of E3 ubiquitin ligases (sec Figure 3-29). • All cytokine receptors have similar structures, and their
As a result of binding SOCS-1, for instance, JAK2 becomes cytosolic domains are tightly bound to a JAK protein tyro-
polyubiquitinated (a polymer of ubiquitins covalently at- sine kinase, which becomes activated after' cytokine binding
tached to the side chain of a lysine) and is then degraded in and receptor dimerization (see Figure 16-1 0).
proteasomes, thereby permanentl y turning off all JAK2-
• In both RTKs and cyrokine receptors, short amino acid
mediated signaling pathways until new JAK2 proteins can be
sequences containing a phosphotyrosine residue are bound
made. The observation that proteasome inhibitors prolong
by proteins with conserved SH2 or PTB domains, which are
JAK2 signal transduction supports this mechanism.
found in many signal-transducing proteins. The sequence of
Studies with cultured mammalian cells have shown that
amino acids surrounding the phosphorylated tyrosine deter-
the receptor for growth hormone, which belongs to the cyto-
mines wh ich domain will bind to it. Such protein-protein
kine receptor superfamily, is down-regulated by another
interactions are important in many signaling pathways (sec
SOCS protein, SOCS-2. Strikingly, mice deficient in SOCS-2
Figures 16-11 and 16-12).
grow significa~tly larger than their wild-type counterparts
and have long bone lengths and proportionate enlargement • The]A KISTAT pathway operates downstream from all cyto-
of most organs. Thus SOCS proteins play an essential nega- kine receptors and some RTKs. STAT monomers bound to
tive role in regulating intracellular signaling from the recep- phosphotyrosines on receptors are phosphorylated by receptor-
tors for erythropoietin, growth hormone, and other cytokines. associated JAKs, then dimerize and move to the nucleus, where
they activate transcription (see Figure 16-13).
• Endocytosis of receptor-hormone complexes and their
KEY CONCEPTS of Section 16.1 degradation in lysosomes is a principal way of reducing the
number of receptor tyrosine kinases and cytokine receptors
Receptors That Activate Protein Tyrosine Kinases on the cell surface, thus decreasing the sensitivity of cells ro
• Two broad classes of receptors activate tyrosine kinases: (1) many peptide hormones.
receptor t)rosine kinasel. (RTKs), in which the kinase is an in- • Signaling from cytokine receptors is terminated by the
trinsic part of the receptor, and (2) cytokine receptors, in which phosphotyrosine phosphatase SHPl and several SOCS pro-
the kinase is bound tightly to the cytosolic domain of the re- teins (see Figure 16-14 ).
ceptor. Signaling from receptor tyrosine kinases and cytokine
receptors activate similar downstream signaling pathways (see
figure 16-2).
• Receptor tyrosine kinases, which bind to peptides and sig- 16.2 The Ras/MAP Kinase Pathway
naling proteins such as growth factors and insulin, may exist
as preformed dimers or dimerize during binding to ligands. Almost all receptor tyrosine kinascs and cytokine receptors
Ligand binding triggers formation of functional dimeric recep- activate the Ras!MAP kinase pathway (see Figure 16-2b ).
tors, a necessary step in activation of the receptor-associated The Ras protein, a monomeric (small) G protein, belongs to
kinase. the GTPase superfamily of intracellular switch proteins (see
figure 15 -7). Activated Ras promotes formation, at the

734 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 membrane, of signal transduction complexes containing a hundredfold; the actual hydrolysis of GTP is catalyzed by
three sequentially acting protein kinases. This kinase cascade amino acids from both Ras and GAP. In particular, insertion
culminates in activation of certain members of the MAP kinase of one of GAP's arginine side chains into the Ras active site
family, which can translocate into the nucleus and phos- stabilizes an intermediate in the hydrolysis reaction.
phorylate many different proteins. Among the target pro-
teins for MAP kinase are transcription factors that regulate Mammalian Ras proteins have been studied in great
expression of proteins with important roles in the cell cycle ·· detail because mutant Ras proteins are associated \Vith
and in differentiation. Importantly, different types of extra- many types of human cancer. These mutant proteins, which
cellular sign::~ Is often activate different signaling pathways bind but cannot hydrolyze GTP, are permanently in the
that result in activation of different members of the MAP "on" state and contribute to oncogenic transformation (see
kinase family. Chapter 24 ). Determination of the three-dimensional struc-
Because an activating mutation in a RTK, Ras, or a pro- ture of the Ras-GAP complex and tests of mutant forms of
tein in the MAP kinase cascade is found in almost all types of Ras explained the puzzling observation that most oncogenic,
human tumors, the RTK/Ras/MAP kinase pathway has been constitutively active Ras proteins (Ras 0 ) contain a mutation
subjected to extensive study and a great deal is known about at position 12. Replacement of the normal glycine-12 with
the components of this pathway. We begin our discussion by any other amino acid (except proline) blocks the functional
reviewing how Ras cycles between the active and inactive binding of GAP and in essence "locks" Ras in the active
state. We then describe how Ras is activated and passes a GTP-bound state. •
signal to the MAP kinase pathway. Finally we examine recent
studies indicating that both yeasts and cells of higher eukary- The first indication that Ras functions downstream from
otes contain multiple MAP kinase pathways and consider the RTKs in a common signaling pathway came from experi
ways cells keep different MAP kinase pathways separate ments in which cultured fibroblast cells were induced to pro-
from one another through the use of scaffold proteins. liferate by treatment with a mixture of two protein hormones:
platelet-derived growth factor (PDGF) and epidermal growth
Ras, a GTPase Switch Protein, Operates factor (EGF). Microinjection of anti-Ras.antibodies into
these cells blocked cell proliferation. Conversely, injection of
Downstream of Most RTKs
Ras 0 , a constitutively active mutant Ras protein that hydro-
and Cytokine Receptors lyzes GTP very inefficiently and thus persists in the active
Like the G 0 subunits in trimeric G proteins discussed in state, caused the cells to proliferate in the absence of the
Chapter 15, the monomeric G protein known as Ras alter- growth factors. These findings are consistent with studies,
nates between an active "on" state with a bound GTP and using the pull-down assay method detailed in Figure 15-14,
an inactive "off" state with a bound GDP (see Figure 15-6 tO showing that addition of FGF to fibroblasts leads to a rapid
review this concept). Unlike trimeric G proteins, Ras is not increase in the proportion of Ras present in the GTP-bound
directly linked to cell-surface receptors. Ras (-170 amino active form. However, as we will see, an activated RTK (or
acids) is smaller than G" proteins (- 300 amino acids), but cytokine receptor) cannot directly activate Ras. Instead,
the GTP-binding domains of the two proteins have a similar other proteins must first be recruited to the activated recep-
structure (~ee Figure 15-7 to review the structure of Ras). tor and serve as adapters.
Structural and biochemical studies show that Ga also con -
tains a GTPase-activating protein (GAP) domain that in-
creases the intrinsic rate of GTP hydrolysis by G". Because Genetic Studies in Drosophila Identified
this domain is not present in Ras, it has an intrinsically
Key Signal-Transducing Proteins
slower rate of GTP hydrolysis. Thus the average lifetime of a
GTP bound to Ras is aboot 1 minute, which is much longer in the Ras/MAP Kinase Pathway
.· than the average lifetime of a Ga·GTP complex . Our knowledge of the proteins involved in the Ras/MAP kinase
The activity of the Ras protein is regulated by several pathway came principally from genetic analyses of mutant fruit
factors. Ras activation is accelerated by a guanine nucleotide flies (Drosophila) and worms (C. elegans) that were blocked at
exchange factor (GEF), which binds to the Ras·GDP com- particular stages of differentiation. To illustrate the power of
plex, causing dissociation of the bound GDP (see Figure 15-6). this experimental approach, we consider development of a par-
Because GTP is present in cells at a higher concentration ticular type of cell in the compound eye of Drosophila.
than GDP, GTP binds spontaneously to "empty" Ras mole- The compound eye of the fly is composed of some 800 in-
cules, with release of GEF and formation of the active dividual eyes called ommatidia (Figure 16-15a). f.ach omma-
Ras·GTP. Subsequent hydrolysis of the bound GTP to GDP tidium consists of 22 cells, eight of which are photosensitive
deactivates Ras. Because the intrinsic GTPase activity of neurons called retinula, or R cells, designated R 1-R8 (Figure
Ras·GTP is low compared to that of Ga·GTP, Ras·GTP re- 16-15b). An RTK called Sevenless (Sev) specifically regulates
quires the assistance of another protein, a GTPase-activating development of the R7 cell and is not essential for any other
protein (GAP), to deactivate it. Binding of GAP to Ras·GTP known function. In flies with a mutam sevenless (set' ) gene, the
accelerates the intrinsic GTPase activity of Ras by more than R7 cell in each ommatidium does not form (Figure 16-15c,

16.2 The Ras/MAP Kinase Pathway 735

 (a) (b) (c)

Axo nsto RS
I Toward
b rain eye
surf ace

FIGURE 16-15 The compound eye of Drosophila melanogaster. technique that can distinguish the photoreceptors in an ommat idium.
(a) Scanning electron micrograph showing individual ommatidia that The plane of sectioning is indicated by the blue arrows in (b), and the
compose the fruit fly eye. (b) Longitudinal and cutaway views of a R8 cell is out of the plane of these images. The seven photoreceptors
single ommatidium. Each of these tubu lar structu res conta ins eight in th is pla ne are easily seen in the w ild-type ommat idia (top), whereas
photoreceptors, designated R1-R8, w hich are long, cylindrically shaped only six are visible in the mutant ommatidia (bottom). Flies w ith t he
light-sensitive cells. Rl-R6 (yellow) extend throughout the depth of t he sevenless mutation lack the R7 cell in their eyes. [Part (a) from E. Hafen
retina, whereas R7 (brown) is located toward the surface of the eye and and K. Basler, 1991, Development 1(suppl.):123; part (b) adapted from R. Reinke
RS (blue) toward the back side, where the axon s exit. (c) Comparison o f and S. L. Zipursky, 1988, Cell 55:321; part (c) courtesy of U. Banerjee.]
eyes from w ild-type and sevenless mutant flies viewed by a special

bottom). Since the R7 photoreceptor is necessary only for fl ies occur, and no R7 cells develop (Figure 16-16b); this is the origin
to see in ultraviolet light, mutants that lack functional R7 cells of the name "Sevenless" for the RTK in the R7 cells.
bur are otherwise normal arc easily isolated. Therefore, fly R7 To identify int racellular signa l-transducing proteins in
cells are an ideal genetic system for studying cell development. the Sev RTK pathway, investigators produced mutant flies
During development of each ommatidium, a protein called expressing a temperature-sensitive Sev protein. When these
Boss (Bride of Sevenless) is expressed on the surface of the R8 flies were maintained at a permissive temperature, all their
cell. This membrane-tethered protein is the ligand for the Sev ommatidia contained R7 cells; when they were maintained
RTK on the surface of the neighboring R7 precursor cell, signal- at a nonpermissive temperature, no R7 cells developed. At a
ing it to develop i!ltO a photosensitive neuron (Figure 16-16a). In particu lar intermediate temperature, however, just enough
mutant flies that do not express a functional Boss protein or Sev of the Sev RTK was functional to mediate normal R7 devel-
RTK, interaction between the Boss and Sev proteins cannot opment. The investigators reasoned that at this intermediate

(a) Wild type (b) Single mutant (c) Double mutant

(sev-) (sev-; RasD)

:XPERIMENTAL FIGURE 16-16 Genetic studies reveal that

activation of Ras induces development of R7 photo receptors in the
Drosophila eye. (a) During larval development of w ild-type flies, the RS
cell in each developing ommatidium expresses a cell-surface protein,
called Boss, which binds to the Sev RTK on the surface of its neighbor-
ing R7 precursor cell. This interaction induces changes in gene
expression that result in differentiation of the precursor cell into a
functional R7 neuron. (b) In fly embryos with a mutation in the
R7 p recursor
sevenless (sev) gene, R7 precursor cells cannot bind Boss and therefore
do not differentiate normally into R7 cells. Rather, the p recursor cell
enters an alternative developmental pathway and eventually becomes
l lod"o<loo
a cone cell. (c) Double-mutant larvae (sev ; Ras 0 ) express a constitu-
tively active Ras (Ras 0 ) in the R7 precursor cell, which induces differen-
tiation of R7 precursor cells in the absence of the Boss-mediated signal.
This finding shows that activated Ras is sufficient to mediate induction
of an R7 cell. [SeeM. A. Simonet al., 1991, Ce/1 67:701, and M. E. Fortini et al.,
1992, Nature 355:559.]

736 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 FIGURE 16-17 Activation of Ras following ligand binding to
receptor tyrosine kinases (RTKs) or cytokine receptors. The
EGF receptors for epidermal growth factor (EGF) and many other growth
monomers factors are RTKs. The cytosolic adapter protein GRB2 binds to a specific
phosphotyrosine on an activated, ligand-bound receptor and to the
Exterior
cytosolic Sos protein, bringing it near the plasma membrane and to its
Cytosol --- ~ substrate, the inactive Ras·GOP. The guanine nucleotide exchange
factor (GEF) activity of Sos then promotes formation of active Ras·GTP.
Inactive Ras -W p
Note t hat Ras is tethered to the cytosolic surface ofthe plasma
membrane by a hydrophobic farnesyl anchor (see Figure 10-19).
[See J. Schlessinger, 2000, Ce// 103:211, and M.A. Simon, 2000, Ce//1 03:13.)

Binding of hormone ca uses receptor

dimerization, kinase activation, and
phosphorylation of cytosolic receptor
D 1 temperature, the signaling pathway would become defective
(and thus no R7 cells would develop) if the level of another
ty rosine residues protein involved in the pathway was reduced, thereby reduc-
1, ' I

ing the activity of the overall pathway below the level re-
Active quired to form an R7 cell. A recessive mutation affecting
EGF
such a protein would have this effect because, in diploid or-
ganisms such as Drosophila, a heterozygote containing one
wild-type and one mutant allele of a gene will produce half

Q p -~
the normal amount of the gene product; hence, even if such
a recessive mutation is in an essential gene, the organism will
usually be viable. However, a fly carrying a temperature-
sensitive mutation in the sev gene and a second mutation
~ affecting another protein in the signaling pa~hway would be
SH3
expected to lack R 7 cells at the intermediate temperature.
Binding of GRB2 and Sos couples By use of this screen, researchers identified three genes
receptor to inactive Ra s encoding important proteins in the Sev pathway: an SH2-
containing adapter protein exhibiting 64 percent amino acid
sequence identity to human GRB2 (growth factor receptor-
.· bound protein 2), a g uanine nucleotide exchange factor
called Sos (Son of Sevenless) exhibiting 45 percent identity with
its mouse counterpart, and a Ras protein exhibiting 80 percent
identity with its mammalian counterparts. These three pro-
teins later were found to fu nction in other signaling path-
ways initiated by ligand binding to different RTKs and used
at different times and places in the developing fl y.
In subsequent studies, researchers introduced a mutant
ras 0 gene into fly embryos carrying the sevenless mutation. As
noted earlier, the ras0 gene encodes a constituti vely active Ras

GOP ~ IJ Sos promotes dissociation of GOP

from Ras; GTP binds and active
protein that is present in the active GTP-bound form even in
the absence of a hormone signal. Although no functional Sev
GTP i R11s dissociates from Sos RTK was expressed in these double mutants (sev ; ras 0 ), R7
cells formed normally, indicating that presence of an activated
Ras protein is sufficient fo r induction of R7-cell development
(Figure 16-16c). This finding, which is consistent with the re-
sults with cultured fibroblasts described earlier, supports the
conclusion that activation of Ras is a principal step in intracel-
lular signaling by most if not all RTKs and cytokine receptors.

Receptor Tyrosine Kinases and JAK Kinases

Are Linked to Ras by Adapter Proteins
In order for activated RTKs and cytokine receptors to acti-
vate Ras, two cytosolic proteins-GRB2 and Sos-must first
Signal ing be recruited to provide a link between the receptor and Ras
(Figure 16-17). GRB2 is an adapter protein, meaning that it

16.2 The Ras/MAP Kinase Pathway 737

has no enzymatic activity and serves as a link, or scaffold, to SH2 domains follows a similar strategy: certain residues
between two other proteins-in this case between the acti- provide the key structural motif necessary for binding, and
vated receptor and Sos. Sos is a guanine nucleotide exchange neighboring residues confer specificity to the binding.
protein (GEF), which catalyzes conversion of inactive GDP-
bound Ras to the active GTP-bound form.
GRB2 is able to serve as an adapter protein because of Binding of Sos to Inactive Ras Causes
its SH2 domain, which binds to a specific phosphotyrosine a Conformational Change That Triggers
residue in the activated RTK (or cytokine receptor ). In addi- an Exchange of GTP for GOP
tion to its SH2 Jumain, the GRB2 adapter protein contains
Following activation of an RTK (e.g., the EGF receptor), a
two SH3 domains, which bind to Sos, the Ras guanine nu-
complex containing the activated receptor, GRB2, and Sos is
cleotide exchange factor (Figure 16-17). Like phosphoryrosine-
binding SH2 and PTB domains, SH3 domains are present in formed on the cytosolic face of the plasma membrane (see Fig-
a large number of proteins involved in intracellular signal- ure 16-17). Complex formation depends on the ability of
GRB2 to bind simultaneously to the receptor and to Sos. Thus
ing. Although the three-dimensional structures of various
receptor activation leads to reloca lization of Sos from the cyto-
SH3 domains are similar, their specific amino acid se-
sol to the membrane, bringing Sos near to its substrate, namely,
quences differ. The SH3 domains in GRB2 selectively bind
Ras·GDP that is already bound tq the plasma membrane by
to proline-rich sequences in Sos; different SH3 domains in
means of a cova len tl y attached lipid. Binding of Sos to
other proteins bind to proline-r ich sequences distinct from
Ras·GDP leads to conformational changes in the Switch I and
those in Sos.
Switch li segments of Ras, thereby opening the binding pocket
Proline residues play two roles in the interaction between
for GDP so it can diffuse out (Figure 16-19). In other words,
an SH3 domain in an adapter protein (e.g., GRB2) and a
proline-rich sequence in another protein (e.g., Sos). First, rhe Sos is functions as a GEF for Ras. GTP then binds to and acti-
proline-rich sequence assumes an extended conformation vates Ras. Binding of GTP to Ras, in turn, induces a specific
conformation of Switch I and Switch II that a·! lows Ras·GTP to
that permits extensive contacts with the SH3 domain, thereby
activate the next protein in the Ras/MAP kinase pathway.
facilitating interaction. Second, a subset of these pralines fit
into binding "pockets" on the surface of the SH3 domain
(Figure 16-18). Several nonproline residues also interact with
Signals Pass from Activated Ras to a Cascade
the SH3 domain and are responsib le for determin ing the
binding specificity. Hence the binding of proteins to SH3 and of Protein Kinases, Ending with MAP Kinase
Biochemica l and genetic studies in yeast, C. elegans, Dro-
sophila, and mammals have revealed that downstream of
Ras is a highly conserved cascade of three protein kinases,
culminating in MAP kinase. Although activation of the ki-
nase cascade does not yield the same biological results in all
cells, a common set of sequentially acting kinases defines the
Ras/MAP kinase pathway, as outlined in Figure 16-20. Ras
is activated by exchange of GDP for GTP (step 0 ). Active
Ras·GTP binds to the N-terminal regulatory domain of Raf,
a serine/threonine (not tyrosine) kinase, thereby activating it
(step f)). In unstimulated cell s, Raf is phosphorylated and
bound in an inactive state to the phosphoserine-binding pro-
tein 14-3-3. Hydrolysis of Ras·GTP to Ras·GDP releases ac-
tive Raf from its complex with 14-3-3 (step D ), and Raf
subsequently phosphorylates and thereby activates MEK
(step 0 ). (A dua l-specificity protein kinase, MEK phosphor-
ylates its target proteins on both tyrosine and serine/threonine
residues.) Active MEK then phosphorylates and activates
MAP kinase, another serine/threonine k inase also known as
ERK (step Ia). MAP kinase phosphorylates many different
proteins, including nuclear transcription factors that mediate
FIG URE 16 - 18 Surface model of an SH3 domain bound to a
cellular responses (step 1':!).
target pept ide. The short, proline-rich target peptide is shown as a
Several types of experiments have demonstrated that Raf,
space-filling model. In this target peptide, two prolines (Pro4 and Pro7, MEK, and MAP kinase lie downstream from Ras and have
dark blue) fit into binding pockets on the surface of the SH3 domain. revealed the sequential order of these proteins in the pathway.
Interactions involving an arginine (Argl, red), two other prolines (light For example, mutant Raf proteins missing the N-terminal
blue). and other residues in the target peptide (green) determine the regulatory domain are constitutively active and induce quies-
specificity of binding. [After H. Yu et al., 1994, Ce//76:933.] cent cultured cells to proliferate in the absence of stimulation

738 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 (a) Ras·GOP (b) Ras-Sos (c) Ras·GTP

..
• Switch I

• Switch II

GTP a, p
phosphates

FIGURE 16-19 Structures of Ras bound to GOP, Sos protein, and binding of the GTP phosphates completes the interaction. The
GTP. (a) In Ras·GOP, the Switch I (green) and Switch II (blue) segments resulting conformational change in Switch I and Switch II segments of
do not directly interact with GOP. (b) One a helix (brown) in Sos binds Ras, allowing both to bind to the GTP -y phosphate, displaces Sos and
to both switch regions of Ras·GOP, leading to a massive conformational promotes interaction of Ras·GTP with its effectors (discussed later). See
change in Ras. In effect, Sos pries Ras open by displacing the Switch I Figure 15-8 for another depiction of Ras·GOP and Ras·GTP. [Adapted
region, thereby allowing GOP to diffuse out. (c) GTP is thought to bind from P. A. Boriack-Sjodin and J. Kuriyan, 1998, Nature 394:341.]
to the Ras-Sos complex first through its base (guanine); subsequent

Ras activated Active Ras recruits, GTP hydrolysis

by exchange of binds, and activates leads to dissociation
Exterior GOP for GTP Raf of Ras from Raf
II
cj p
- \o::--o-...,.
GOP GTP
r ---+•
Inactive Ras

Cytosol
Inactive Raf

N-terminal
regulatory - - C-terminal
domain kinase
domain

FIGURE 16-20 Ras/MAP kinase pathway.ln unstimulated cells,

most Ras is in the inactive form with bound GOP; binding of a ligand to
its RTK or cytokine receptor leads to formation of the active Ras·GTP
complex (step 0 ; see also Figure 16-17). Activated Ras triggers the
downstream kinase cascade depicted in steps f) -lit culminating in
activation of MAP kinase (MAPK). In unstimulated cells, binding of a
dimer of the 14-3-3 protein to Raf stab ilizes it in an inactive conforma-
tion (the 14-3-3 protein binds phosphoserine residues in a number of
important signaling proteins). Each 14-3-3 monomer binds to a
phosphoserine residue in Raf, one to phosphoserine-259 in the Active MAP kinase translocates
N-terminal domain and the other to phosphoserine-621 in the kinase to nucleus; activates many
transcription factors
domain. Interaction of the Raf N-terminal regulatory domain with
Ras·GTP results in dephosphorylation of one of the serines that bind
Raf to 14-3-3, phosphorylation of other residues, and activation of Raf
kinase activity. After inactive Ras·GOP dissociates from Raf, it presum-
ably can be reactivated by signals from activated receptors, thereby
recru iting additional Raf molecules to the membrane. [See E. Kerkhoff
and U. Rapp, 2001,Adv. EnzymeRegul. 41:261; J. Avruch et al., 2001, Recent Prog.
Hormone Res. 56:127; and M. Yip-Schneider et al., 2000, Biochem. J. 351 :151.]

16.2 The Ras/MAP Kinase Pathway 739

by growth factors. These mutant Raf proteins were initially selective inhibitors of the 8-Raf kinase have recently entered
identified in tumor cells; like the constitutively active Ras 0 the clinic and are producing excellent responses in patients
protein, such mutant Raf proteins are said to be encoded by with B-Ra{ mutant melanoma. 8
oncogenes, whose encoded proteins promote transformation
of the cells in which they are expressed (see Chapter 24). Con-
versely, cultured mammalian cells that express a mutant, non-
Phosphorylation of MAP Kinase Results
functional Raf protein cannot be stimulated to proliferate in a Conformational Change That
uncontrollably by a constitutively active Ras 0 protein. This Enhances Its Catalytic Activity
finding established a link between the Raf and Ras proteins and Promotes Kinase Dimerization
and that Raf lies downstream of Ras in the signaling pathway.
Biochemical and x-ray crystallographic studies have provided
In vitro binding studies further showed that the purified
a detailed picture of how phosphorylation activates MAP ki-
Ras·GTP protein binds directly to theN-terminal regulatory
nase. As in JAK kinases and receptor tyrosine kinases, the
domain of Raf and activates its catalytic activity.
catalytic site in the inactive, unphosphorylated form of MAP
That MAP kinase is activated in response to Ras activa-
kinase is blocked by a stretch of amino acids, the activation
tion was demonstrated in quiescent cultured cells expressing
lip (Figure 16-21a). Binding of MEK to MAP kinase destabi-
a constitutively active Ras 0 protein. In these cells, activated
lizes the lip structure, resulting in exposure of tyrosine-185,
MAP kinase is generated in the absence of stimulation by
which is buried in the inactive conformation. Following
growth-promoting hormones. More importantly, R7 photo-
phosphorylation of this critical tyrosine, MEK phosphory-
receptors develop normally in the developing eye of Dro-
lates the neighboring threonine-183 (Figure 16-2 1b).
sophila mutants that lack a functional Ras or Raf protein
Both the phosphorylated tyrosine and the phosphorylated
but express a constitutively active MAP kinase. This finding
threonine residue in MAP kinase interact with additional
indicates that activation of MAP kinase is sufficient to trans-
amino acids, thereby conferring an altered conformation to
mit a proliferation or differentiation signal normally initi-
the lip region, which in turn permits binding of ATP to the
ated by ligand binding to a receptor tyrosine kinase such as
Sevenless (see Figure 16-16). Biochemical studies showed,
however, that Raf cannot directly phosphorylate MAP ki-
nase or otherwise activate its activity. Kinase catalytic site:
The final link in the kinase cascade activated by Ras·GTP Inactive Active
emerged from studies in which scientists fractionated ex-
tracts of cultured cells searching for a kinase activity that (a) Inactive MAP kinase
.'•
could phosphorylate MAP kinase and that was present only
in cells stimulated with growth factors, not unstimulated
cells. This work led to identification of MEK, a kinase that
specifically phosphorylates one threonine and one tyrosine
residue on the activation lip of MAP kinase, thereby activat-
mg irs catalytic activity. (The acronym MEK comes from
MAP and ERK kinase.) Later studies showed that MEK
binds to the C-terminal catalytic domain of Raf and is phos-
phorylated by the Raf serinelrhreonine kinase; this phos-
phorylation activates the catalytic activity of MEK.
Hence activation of Ras induces a kinase cascade that
includes Raf, MEK, and MAP kinase: activated RTK ~ Ras
~ Raf ~ MEK ~ MAP kinase. Although we will not em-
phasize this here, the complexity of this pathway is increased
by the multiple isoforms of each of its components. In hu-
mans, there are three RAS, three Raf, two MEK, and two
Erk proteins, and each of these has overlapping but also
nonredundant functions.
FIGURE 16- 21 Structures of inactive, unphosphorylated MAP
kinase and the active, phosphorylated form. (a) In inactive MAP
Activating mutations in the 8-Raf gene occur in over
kinase, the activation lip is in a conformation such that it blocks the
40 percent of melanomas, a skin cancer that is often kinase active site. (b) Phosphorylation by MEK at tyrosine-185 (Y185)
caused by exposure to ultraviolet radiation in sunlight. Of and threonine-183 (T183) leads to a marked conformational change
these melanomas, one particular mutation, a glutamic acid in the activation lip. This activating change both promotes binding
substitution for the valine at position 600, accounts for over of its substrates-AlP and its target proteins-to the kinase and MAP
90 percent. This mutant B-Ra( stimulates MEK-ERK signal- kinase dimerization. A similar phosphorylation-dependent mechanism
ing in cells in the absence of growth factors, and mutant B- activates JAK kinases and the intrinsic kinase activity of RTKs.
Ra( transgenes induce melanoma in mice. Very potent and [After B. J. Canagarajah et al., 1997, Ce//90:859.)

740 CHAPTER 16 • Signaling Pathways That Control Gene Expression

catalytic site, which, as in all kinases, is in the groove between of many genes encoding proteins necessary for cells to pro-
the upper and lower kinase domains. The phosphotyrosine gress through the cell cycle. Most RTKs that bind growth
residue (pYl85) also plays a key role in binding specific sub- factors utilize the MAP kinase pathway to activate genes en-
strate proteins to the surface of MAP kinase. Phosphoryla- coding proteins such as c-Fos, which in turn propel the cell
tion promotes not only the catalytic activity of MAP kinase through the cell cycle.
but also its dimerization. The dimeric form of MAP kinase is The enhancer that regulates the c-fos gene contains a
translocated to the nucleus, where it regulates the activity of serum response element (SRE), so named because it is acti-
many nuclear transcription factors. vated by many growth factors in serum. This complex en-
hancer contains DNA sequences that bind multiple
MAP Kinase Regulates the Activity of Many transcription factors. As depicted in Figure 16-22, activated
(phosphorylated) dimeric MAP kinase induces transcription
Transcription Factors Controlling Early
of the c-fos gene by direct activation of one transcription
Response Genes factor, ternary complex factor (TCF), and indirect activation
Addition of a growth factor (e.g., EGF or PDGF) to quies- of another, serum response factor (SRF). In the cytosol,
cent (non-growing) cultured mammalian cells causes a rapid MAP kinase phosphorylates and activates a kinase called
increase in the expression of as many as 100 different genes. p90RSK, which translocates to the nucleus, where it phos-
These are called early response genes because they are in- phorylates a specific serine in SRF. After translocating to the
duced well before cells enter the S phase and replicate their nucleus, MAP kinase directly phosphorylates specific serines
DNA (see Chapter 20). One important early response gene in TCF. Association of phosphorylated TCF with two mol
encodes the transcription factor c-Fos. Together with other ecules of phosphorylated SRF forms an active trimeric factor
transcription factors, such as c-Jun, c-Fos induces expression that activates gene transcription.

Active, dimeric MAP kinase

Transcription
Coding sequence Coding sequence
c-tos gene c fos gene
Inactive gene

FIGURE 16-22 Induction of gene transcription by MAP kinase. factor TCF that is already bound to the promoter of the c-fos gene.
Steps 0--IJ: In the cytosol, MAP kinase phosphorylates and activates Step l';J: Phosphorylated TCF and SRF act together to stimulate
the kinase p90RsK, which then moves into the nucleus and phosphory- transcription of genes (e.g., c-fos) that contain an SRE sequence in their
m
lates the SRF transcription factor. Steps and ~ : After translocating promoter. See the text for details. [SeeR. Marais et al., 1993, Ce//73:381, and
into the nucleus, MAP kinase directly phosphorylates the transcription V. M. Rivera et al., 1993, Mol. Cell Bioi. 13:6260.]

16.2 The Ras/ MAP Kinase Pathway 741

 "'.. .. • • • •
a factor •
• ~("' a factor
G Protein-Coupled Receptors Transmit Signals
to MAP Kinase in Yeast Mating Pathways

-~: • . @ .• Although in multicellular animals MAP kinase is often acti-

.• . . / ..
vated by RTKs or cytokine receptors, signaling from other

·~··
/
a-factor • a-factor receptor
receptors can activate MAP kinase in different eukaryotic
cells (see Figure 15-34). To illustrate, we consider the mating
pathway in S. cerevisiae, a well-studied example of a MAP
receptor
kinase cascade linked to G protein-coupled receptors
(GPCRs), in this case for two secreted peptide pheromones,
the a and a factors.
Haploid yeast cells are either of the a or a mating type
and secrete protein signals known as pheromones, which in-
duce mating between haploid yeast cells of the opposite mat-
ing type, a or a. An a haploid cell secretes the a mating factor
and has cell-surface receptors for the a factor; an a cell se-
cretes the a factor and has cell-~urface receptors for the a
factor (see Figure 16-23 ). Thus each type of cell recognizes
Cell-cycle arrest the mating factor produced by the opposite type. Activation

1 Morphogenesis
(shmoo formation)
of the MAP kinase pathway by either the a or a receptors
induces transcription of genes that inhibit progression of the
cell cycle and others that enable cells of opposite mating type
to fuse together and ultimately form a diploid cell.
Ligand binding to either of the two yeast pheromone
+Nutrients GPCRs triggers the exchange of GTP for GDP on the Ga sub-

1 Coli f"'loo
unit and dissociation of Ga·GTP from the Gfl-v complex. This
activation process is identical to that for the GPCRs discussed
in the previous chapter (see Figure 15-17). In many mamma-
lian GPCR-initiated pathways, the active G" transduces the

~ signal. In contrast, mutant and biochemical studies have

shown that the dissociated Gp-y complex mediates all the phys-

1 ""'''"'"''00
iological responses induced by activation of the yeast phero-
mone receptors (Figure 16-24a). For instance, in yeast cells
that lack G"' the Gp-y subunit is always free. Such cells can
mate in the absence of mating factors; that is, the mating re-
~
~ Diploidcell
sponse is constitutively on. However, in cells defective for the
Gfl or G-v subunit, the mating pathway cannot be induced at
all. If dissociated Ga were the transducer, in these mutant cells
+ Nutrien~ ~Nutrients the pathway would be expected to be constitutively active.
In yeast mating pathways, Gp'l functions by triggering a
Mitotic growth Meiosis
kinase cascade that is ana logous to the one downstream

(I)@ ~
from Ras; each protein has a yeast-specific name but shares
sequences with and is ana logous in structure and function to
Sporulation the corresponding mammalian protein shown in Figure 16-20.

80 The components of this cascade were uncovered mainly

through analyses of mutants that possess functional a and a

80
FIGURE 16-23 Pheromone-induced mating of haploid yeast
receptors and G proteins but are sterile (Ste), or defective in
mating responses. The physical interactions between the
components were assessed through immunoprecipitation ex-
periments with extracts of yeast cells and other types of stud-
cells. The a cells produce a mating factor and a-factor receptor; the
a cells produce a factor and a-factor receptor. Both r~ceptors are
ies. Based on these studies, scientists have proposed the
G protein-coupled receptors. Binding of the mating factors to their kinase cascade shown in Figure 16-24a. Free GP-v• which is
cognate receptors on cells of the opposite type leads to gene tethered to the membrane via the lipid bound to the-y sub-
activation, resulting in mating and production of diploid cells. In the unit, binds the Ste5 protein, thus recruiting it and its bound
presence of sufficient nutrients, these cells will grow as diploids. kinases to the plasma membrane. Stc5 has no obvious cata-
Without sufficient nutrients, the cells will undergo meiosis and form lytic function and acts as a scaffold for assembling other
four haploid spores. components in the cascade (Stell, Ste7, and Fus3). Gpy also

742 CHAPTER 16 • Sig naling Pathways That Control Gene Expression

(;) PODCAST: Scaffold Proteins in Yeast Map Kinase Cascades
(a) Mating pathway (b) Osmoregulatory pathway
Exterior
Activation by high
Mating osmotic strength
factor
1 Receptor
Activation
Q I
6
of G protein
1"7-----.
l!

Cytosol

Ste5 Pbs2
scaffold scaffold

Other
targets

Transcription
factors

FIGURE 16-24 Y~ast MAP kinase cascades in the mating and factor, allowing it to bind to DNA and initiate transcription of genes
osmoregulatory pathways. In yeast, different receptors activate that inhibit progression of the cell cycle and others that enable cells of
different MAP kinase pathways, two of which are outlined here. The opposite mating type to fuse together and ultimately form a diploid
two MEKs depicted, like all MEKs, are dual specificity threonine/tyrosine cell. (b) Osmoregulatory pathway: Two plasma membrane proteins,
kinases; all of the others are serine/threonine kinases. (a) Mating Shol and Msbl, are activated in an unknown manner by exposure of
pathway: The receptors for yeast~ and a mating factors are coupled to yeast cells to media of high osmotic strength. Activated Shol recruits
the same trimeric G protein. Following ligand binding and dissociation the Pbs2 scaffold protein, which contains a MEK domain, to the plasma
of the G protein subunits, the rl)embrane-tethered G~y subunit binds membrane. Similar to the mating pathway, at the plasma membrane
the SteS scaffold to the plasma membrane. G~y also activates Cdc24, a the Shol Msbl complex also activates Cdc42, which in turn activates
GEF for the Ras-like protein Cdc42; the active GTP-bound Cdc42 in the resident Ste20 kinase. Ste20 in turn phosphorylates and activates
turn binds to and activates the resident Ste20 kinase. Ste 20 then Stell, initiating a kinase cascade that activates Hogl, a MAP kinase. In
phosphorylates and activates Stell, which is analogous to Raf and the cytosol, Hogl phosphorylates specific protein targets, including ion
other mammalian MEK kinase (MEKK) proteins. Ste20 thus serves as a channels; after translocating to the nucleus, Hogl phosphorylates
MAPKKK kinase. Stell initiates a kinase cascade in which the final several transcription factors and chromatin-modifying enzymes. Hogl
component, Fus3, is functionally equivalent to MAP kinase (MAPK) in appears also to promote transcriptional elongation. Together, the
higher eukaryotes. Like other MAP kinases, activated Fus3 then newly synthesized and modified proteins support survival in high-
translocates into the nucleus. There it phosphorylates two proteins, osmotic-strength media. [After N. Dard and M. Peter, 2006, Oio[ssays 28:146,
Digl and Dig2, relieving their inhibition of the Ste12 transcription and R. Chen and J. Thorner 2007, Biochim. Biophys. Acta 1773:1311.]

16.2 The Ras/ MAP Kinase Pathway 743

activates cdc24, a GEF for the Ras-like protein cdc 42; Once the sharing of components among different MAP
GTP·cdc42 in turn activates the Ste20 protein kinase. Ste20 kinase pathways was recognized, researchers wondered how
phosphorylates and activates Stell, a serine/threonine ki- the specificity of the cellular responses to particular signals is
nase analogous to Raf and other mammalian MEKK pro- achieved. Studies with yeast provided the initial evidence
teins. Activated Stell then phosphorylates Ste7, a dual- that pathway-specific scaffold proteins enable the signa l-
specificity MEK that then phosphorylates and activates transducing kinases in a particular pathway to interact with
Fus3, a serine/threonine kinase equivalent to MAP kinase. one another but not with kinases in other pathways. For ex-
After translocation to the nucleus, Fus3 phosphorylates two ample, the scaffold protein Ste5 stabilizes a large complex
proteins, Digl and Dig2, relieving their inhibition of the that includes the kinases in the maring pathway; similarly,
Ste 12 transcription factor. Activated Ste 12 in turn induces the Pbs2 scaffold is used for the kinase cascade in the osmo-
expression of proteins involved in mating-specific cellular regulatory pathway (see Figure 16-24). In each pathway in
responses. Fus3 also affects gene expression through phos- which Stell participates, it is constrained within a large
phorylating other proteins. complex that forms in response to a specific extracellular
signal, and signaling downstream from Stell is restricted to
the complex in which it is localized. As a result, exposure of
yeast cells to mating factors induces activation of a single
Scaffold Proteins Separate Multiple MAP Kinase
MAP kinase, Fus3, whereas exposure to a high osmolarity
Pathways in Eukaryotic Cells induces activation of a different MAP kinase, Hogl.
Thus both yeasts and higher eukaryotic cells contain a Ras/ Scaffolds for MAP kinase pathways are well documented
MAP kinase signaling pathway that is activated by extra- in yeast, fly, and worm cells, but their presence in mamma-
cellular protein signals and culminates in the MAP kinase- lian cells has been difficult to demonstrate. Perhaps the best-
mediated phosphorylation of transcription factors and other documented scaffold protein in metazoans is Ksr (kinase
signaling proteins that together trigger specific changes in suppressor of Ras), which binds both MEK and MAP ki-
cell behavior. Importantly, all eukaryotes possess multiple nase. In Drosophila, loss of the Ksr homolog blocks signal-
highly conserved MAP kinase pathways that are activated by ing by a constitutively active Ras protein, suggesting a
different extracellular signals and that activate different positive role for Ksr in the Ras/MAP kinase pathway in fly
MAP kinase proteins that phosphorylate different transcrip- cells. Although knockout mice that lack Ksr are phenotypi-
tion factors; these in turn trigger different changes in cell cally normal, activation of MAP kinase by growth factors or
division, differentiation, or function. Mammalian MAP kinases cytokines is lower than normal in several types of cells in
include fun N-terminal kinases (]NKs) and p38 kinases, these animals. This finding suggests that Ksr functions as a
which become activated by signaling pathways in response scaffold that enhances but is not essential for Ras/MAP ki-
to various types of stresses and which phosphorylate different nase signaling in mammalian cells. Thus the signal specificity
transcription factors and other types of signaling proteins of different MAP kinases in animal cells may arise from their
that affect cell division. association with various scaffold-like proteins, but much ad-
Current genetic and biochemical studies in the mouse ditional research is needed to test this possibility.
and Drosophila are aimed at determining which MAP ki-
nases mediate which responses to which signals in higher
eukaryotes. Thts has already been accomplished in large part
for the simpler organism S. cerevisiae. Each of the six MAP
kinases encoded in the S. cerevisiae genome has been as-
signed by genetic analyses to specific signaling pathways KEY CONCEPTS of Section 16.2
triggered by various extracellular signals, such as phero- The Ras/MAP Kinase Pathway
mones, high osmolanty, starvation, hypotonic shock, and
• Ras is an intracellular GTPase switch protein that acts
carbon/nitrogen deprivation. A second yeast MAP kinase
downstream from most RTKs and cytokine receptors. Like
cascade, known as the osmoregulatory pathway, is shown in
G"' Ras cycles between an inactive GDP-bound form and an
Figure 16-24b. Each yeast MAP kinase mediates very spe-
active GTP-bound form. Ras cycling requires the assistance
cific cellular responses, as exemplified by Fus3 in the mating
of two proteins: a guanine nucleotide exchange factor (GEF)
pathway and Hogl in the osmoregulatory pathway.
and a GTPase-activating protein (GAP).
A complication arises because in both yeasts and higher
eukaryotic cells, different MAP kinase cascades share some • RTKs are linked indirectly to Ras via two proteins: GRB2,
common components. For instance, the MEKK Stell func- an adapter protein, and Sos, which has GFF activity (see
tions in three yeast signaling pathways: the mating pathway, Figure 16-17).
the osmoregulatory pathway, and the filamentous growth • The SH2 domain in GRB2 binds to a phosphotyrosine in
pathway, which is induced by starvation. Nevertheless, each activated RTKs, while its two SH3 domains bind Sos,
pathway activates a distinct MAP kinase. Similarly, in mam- thereby bringing Sos close to membrane-bound Ras·GDP
mahan cells, common upstream signal-transducing proteins and activating its nucleotide-exchange activity.
participate in activating multiple JNK kinases.

744 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 phosphatidylinositol 4,5-bisphosphate (PIP2 ) to generate two
• Binding of Sos to inactive Ras causes a large conforma- important second messengers: 1,2-diacylglycerol (DAG ) and
tional change that permits release of GOP and binding of inositol 1,4,5-trisphosphate (IP3). Signaling via the IP/DAG
GTP, forming active Ras (see Figure 16-19). pathway leads to an increase in cytosolic Ca 2 " and to activa-
• Activated Ras triggers a kinase cascade in which Raf, tion of protein kinase C (see Figure 15-36).
MEK, and MAP kinase are sequentially phosphorylated and Although we did not mention it during our discussion of
thus activated. Activated MAP kinase then translocates ro phospholipase C in Chapter 15, it is specifically the 13 isoform
the nucleus (see Figure 16-20). of this enzyme (PLCa) that is activated by GPCRs. Many RTKs
and cytokine receptors also can initiate the fPi DAG pathway
• Activation of MAP kinase following stimulation of a
by activating another isoform of phospholipase C, the -y iso-
growth-factor receptor leads to phosphorylation and activa-
form (PLC.y), an isoform that contains SH2 domains. The SH2
tion of two transcription factors, which associate into a tri-
domains of PLC'Y bind to specific phosphotyrosines on the
meric complex that promotes transcription of various early
activated receptors, thus positioning the enzyme close to its
response genes (see Figure 16-22).
membrane-bound substrate, phosphatidyl inositol 4,5-
• Different extracellular signals induce activation of different bisphosphate (PIP2). In addition, the kinase activity associated
MAP kinase pathways, which regulate diverse cellular pro- with receptor activation phosphorylates tyrosine residues on
cesses by phosphorylating different sets of transcription factors. the bound PLC-y, enhancing its hydrolase activity. Thus acti-
• The kinase components of each MAP kinase cascade as- vated RTKs and cytokine receptors promote PLC-y activity in
semble into a large pathway-specific complex stabilized by a two ways: by localizing the enzyme to the membrane and by
scaffold protein (see Figure 16-24). This ensures that activa- phosphorylating it. As seen in Chapter 15, the IP3/DAG path-
tion of one MAP kinase pathway by a particular extracellu- way initiated by PLC has multiple physiological effects.
lar signal does not lead to activation of other pathways con-
taining shared components.
Recruitment of Pl-3 Kinase to Activated
Receptors Leads to Synthesis of Thr~e
Phosphorylated Phosphatidylinositols
16.3 Phosphoinositide Signaling Pathways
Besides the IP 3/DAG pathway, many activated RTKs and cyto-
In previous sections, we have seen how signal transduction kine receptors initiate another phosphoinositide pathway by
from receptor tyrosine kinases (RTKs) and cytokine receptors recruiting the enzyme phosphatidy/inositol-3 (Pl-3) kinase to
begins with formation of multiprotein complexes associated the membrane. PI-3 kinase is recruited to the plasma mem-
with the plasma membrane (see Figures 16-12 and 16-13) and brane by binding of its SH2 domain to phosphoryrosines on
how these complexes initiate the Ras/MAP kinase pathway. the cytosolic domain of many activated RTKs and cytokine
Here we discuss how these same receptors initiate signaling receptors. This recruitment positions the catalytic domain of
pathways that involve as intermediates special phosphorylated PI-3 kinase near its phosphoinositide substrates on the cyto-
phospholipids derived from phosphatidyl inositol. As discussed solic face of the plasma membrane. Unlike kinases we have
in Chapter 15, these membrane-bound lipids are collectively encountered earlier that phosphorylate proteins, PI-3 kinase
referred to as phosphoinositides. These phosphoinositide sig- adds a phosphate to the 3' carbon in the lipid phosphatidylino-
naling pathways include several enzymes that synthesize differ- sitol, leading to formation of t\vo separate phosphatidyl inosi-
ent phosphoinositides and proteins with domains that can bind tol 3-phosphates: PI 3,4-bisphosphate or PI 3,4,5-trisphosphate
to these molecules and are thus recruited to the cytosolic sur- (Figure 16-25). By acting as docking sites for various signal-
face of the plasma membrane. In addition to the short-term ef- transducing proteins, these membrane-bound PI 3-phosphatc
fects on cell metabolism Wf! encountered in Chapter 15, these products of the PI-3 kinase reactions in turn transduce signals
phosphoinositide pathways have long-term effects on the pat- downstream in several important pathways.
tern of gene expression. We will see that phosphoinositide path- In some cells, thts Pl-3 kinase pathway can trigger cell
ways end with a variety of kinases, including protein kinase C division and prevent programmed cell death (apoptosis),
(PKC) and protein kinase B (PKB), that play key roles in cell thus ensuring cell survival. In other cells, this pathway in-
growth and metabolism. As an example, later in the chapter we duces specific changes in cell metabolism.
see how insulin activation of PKB plays a key role in stimulating Pl-3 kinase was first identified in studies of the polyoma
glucose import into muscle. virus, a DNA virus that transforms certain mammalian cells
to uncontrolled growth. Transformation requires several
vtral-encoded oncoproteins, including one termed "middle T."
Phospholipase C.y Is Activated by Some RTKs
In an attempt to discover how middle T functions, investiga-
and Cytokine Receptors tors uncovered PI-3 kinase protein in partially purified prep-
As discussed in Chapter 15, hormonal stimulation of some G arations of middle T, suggesting a specific interaction
protein-coupled receptors leads to activation of phospholipase between the two. Then they set out to determine how PI-3
C (PLC). This membrane-associated enzyme then cleaves kinase might affect cell behavior.

16.3 Phosphoinositide Signaling Pathways 745

FIGURE 16-25 Generation of phosphatidylinositol 3-phosphates.
The enzyme phosphatidylinositol-3 kinase (PI-3 kinase) is recruited to
the membrane by many activated receptor tyrosine kinases (RTKs) and
cytokine receptors. The 3-phosphate added by this enzyme, to yield
PI 3,4-bisphosphate or PI 3,4,5-trisphosphate, is a binding site for
various signal-transduction proteins, such as the PH domain of protein
kinase B. Pl4,5-bisphosphate also is the substrate of phospholipase C
(see Figure 15-35). [See L. Rameh and L. C. Cantley, 1999, J. Bioi. Chem.
274:8347.)
c-o c=o C=O C=O
I I I I

When an inactive, dominant-negative version of PI-3 ki -

nase was expressed in polyoma virus-transformed cells, it
I
0
I
0
CH-CH-CH
2
\
0
I
2
ATP
\.
ADP
./'
Pl-5 kinase
. I
0
I
0
CH-CH-CH
2
\
0
I
2

inhibited the uncontrolled cell proliferation characteristic of o -o-P=O OH 0-P=O

virus-transformed cells. This fi nding suggested that the nor- ·~0 6 6
I
0
mal kinase is important in certain signaling pathways essen- .E
tial for cell proliferation or for the prevention of apoptosis.
Subsequent work showed that PI-3 kinases participate in
many signaling pathways related to cell growth and apopto-
sis. Of the nine PI-3 kinase homologs encoded by the human PI 4-phosphate PI 4,5-bisphosphate
genome, the best characterized contains a pllO subunit with (PIP) (PIP2l
catalytic activity and a p85 subunit with an SH2 phosphoty-

t t
rosine-binding domain. ATP ATP
Pl-3 kina' Pl-3 kin se

ADP ADP

Accumulation of PI 3-Phosphates
in the Plasma Membrane leads
to Activation of Several Kinases
Many protein kinases become activated by binding to phos-
pharidyl inositol 3-phosphates in the plasma membrane. In
turn, these kinases affect the activity of many cellular proteins.
One important kinase that binds to PI 3-phosphates is protein
kinase B (PKB), a serine/threonine kinase that is also called
Akt. Besides its kinase domain, protein kinase B also contains C=O C=O C=O C=O
a PH domain, a conserved protein domain present in a wide I I I I
variety of signaling proteins that binds with high affinity to
the 3-phosphares in both PI 3,4-bisphosphare and PI 3,4,5-
0
\
CH-CH-CH
I
0
I ATP
\.
ADP
./' . 0
CH-CH-CH
\
0
I

0-%
2 2 I 2 2
trisphosphare. Since these inositol phosphates are present on 0 Pl-5 kinase 0
I I
the cytosolic face of the plasma membrane, binding recruits -o-P=O OH
the entire protein to the cell membrane. In unstimulated, rest- 6 6
ing cells, the level of these phosphoinositides (collectively
called PI 3-phosphates) is low, and protein kinase B is present
in the cytosol in an inactive form (Figure 16-26). Following
hormone stimulation and the resulting rise in Pl 3-phosphares,
protein kinase B binds to these membrane-bound molecules PI 3,4-bisphosphate PI 3,4,5-trisphosphate
via its PH domain and becomes localized at the plasma mem-
brane. Binding of protein kinase B to PI 3-phosphates not only
recruits the enzyme to the plasma membrane but also releases
inhibition of the catalytic site by the PH domain. However, kinase Bon a critical threonine residue in its activation lip-
maximal activation of protein kinase B depends on recruit- yet another example of kinase activation by phosphoryla-
ment of two other kinases, named PDKl and PDK2. tion. Phosphorylation of a second ser ine, not in the lip
PDK 1 is recruited to the plasma membrane via binding segment, by PDK2 is necessary for maximal protein kinase B
of irs own PH domain to PI 3-phosphates. Both membrane- activity (Figu re 16-26 ). Similar to the regulation of Raf ac-
associated protein kinase Band PDKl diffuse randomly in tivity (see Figure 16-20 ), release of an inhibitory domain and
the plane of the membrane, eventually bringing them close phosphorylation by other kinases regulate the activi ty of
enough together so that PDK 1 can phosphorylate protein protein kinase B.

746 CHAPTER 16 • Signaling Pathways That Control Gene Expression

 ---~ . ___ ..__..

·.
Exterior PI 3,4-bisphosphate FIGURE 16-26 Recruitment and
activation of protein kinase B (PKB) in
Pl-3 kinase pathways. In unstimulated
Cytosol cells (0 ), PKB is in the cytosol with its PH
Pl-3 kinase domain bound to the catalytic kinase
PI 4-phosphate domain, inhibiting its activity. Hormone
stimulation leads to activation of PI-3
kinase and subsequent formation of
phosphatidylinositol (PI} 3-phosphates
PH domain
(see Figure 16-25). The 3-phosphate group
Kinase Activation serves as docking sites on the plasma
domain lip
Partially membrane for the PH domain of PKB
active PKB (f) ) and another kinase, PDKl. Full
Inactive PKB activation of PKB requires phosphorylation
both in the activation lip by PDKl and at
D Inactive PKB in cytosol Ill Formation of IJ Fully active PKB the (-terminus by a second kinase, PDK2
of unst imulated cell PI 3-phosphates, (IJ). [Adapted from A. Toker and A. Newton,
recruitment and partial PDK2 2000, Ce// 103:185, and S. Sarbassov et al., 2005,
activation of PKB
Curr. Opin. Cell Bioi. 17:596.)

Activated Protein Kinase B Induces The Pl-3 Kinase Pathway Is Negatively

Many Cellular Responses Regulated by PTEN Phosphatase
Once fully activated, protein kinase B can dissociate from Like virtually all intracellular signaling events, phosphoryla-
the plasma membrane and phosphorylate its many target tion by PI-3 kinase is reversible. The relevant phosphatase,
proteins throughout the cell, which have a wide range of ef- termed PTEN phosphatase, has an unusuall'y broad specific-
fects on cell behavior. Activation of PKB takes only 5 tO 10 ity. Although PTEN can remove phosphate groups attached
minutes, yet its effects can last as long as several hours. to serine, threonine, and tyrosine residues in proteins, its abil-
In many cells, activated protein kinase B directly phos- ity to remove the 3-phosphate from PI 3,4,5-trisphosphate is
phorylates and inactivates pro-apoptotic proteins such as thought to be its major function in cells. Overexpression of
Bad, a short-term effect that prevents activation of an apop- PTEN in cultured mammalian cells promotes apoptosis by
totic pathway leading to cell death (see figure 2 1-38). Acti- reducing the level of PI 3,4,5-trisphosphate and hence the ac-
vated protein kinase B also promotes survival of many tivation and anti-apoptotic effect of protein kinase B.
cultured cells by phosphorylating the Forkhead transcription
factor FOX03a on multiple serine/threonine residues, The PTEN gene is deleted in multiple types of ad-
thereby reducing its ability tO induce expression of several vanced human cancers. The resulting loss of PTEN
pro-apoptotic genes. protein contributes to the uncontrolled growth of cells.
In the absence of growth factors, FOX03a is unphosphor- Indeed, cells lacking PTEN have elevated levels of PI
ylated and mainly localizes to the nucleus, where it activates 3,4,5-trisphosphate and PKB activity. Since protein kinase B
transcription of several genes encoding pro-apoptotic pro- exerts an anti-apoptotic effect, loss of PTEN indirectly re-
teins. When growth factOrs are added ro the cells, protein duces the programmed cell death that is the normal fate of
kinase B becomes active and phosphorylates FOX03a. This many cells. In certain cells, such as neuronal stem cells, ab-
allows the cytosolic phosphoserine-binding protein 14-3-3 to sence of PTEN not only prevents apoptosis but also leads to
bind FOX03a and thus sequester it in the cytosol. (Recall that stimulation of cell cycle progression and an enhanced rate of
14-3-3 also retains phosphorylated Raf protein in an inactive