Detecting alternative splicing from mRNA-seq data

mRNA-seq data will provide information of alternatively spliced transcripts of all the genes

and generate cDNA sequencing

library

Nat Methods
. 2009 Nov;6(11 Suppl):S22-32.
doi: 10.1038/nmeth.1371
 Alternative splicing revealed from mRNA-seq data
5’CAP AAAAA
mRNA isoforms
5’CAP AAAAA

RNA-seq reads
mapped onto
genome Exon reads Exon reads

Genomic
sequence
Exon-junction read
3 ACTTTGAC---------------------------------------------------------------ATGAGTCAA
Short read
Exon-junction read Exon-junction read
RNA-seq
1 ACTTTGAC---------------------GAAGGAATT 2 ACTCGTTGA--------------------ATGAGTCAA

Exon reads: reads that map entirely to one exon only

Exon-junction reads: reads with one end mapping to one exon and the other end mapping to another axon
 Alternative splicing revealed from mRNA-seq data
mRNA isoforms TTGCTGATTT TAGCTGATAT TAAAATCACT CCACATTGTC TAAACAAACT CAAAAACGCT

TTGCTGATTT TAGCTGATA TAAACAAACT CAAAAACGCT

TAGCTGATAT-------------------------------------------------------------------------------------------TAAACAAACT Read 3

CCACATTGTC------------------- TAAACAAACT Read 2

Read 1 TAGCTGATAT------------------------------------ TAAAATCACT
TTGCTGATTT TAGCTGATAT GGTTGATGTA CATTTGCTCT TAAAATCACT CCACATTGTC GAAGATGATC TAAACAAACT CAAAAACGCT
Genomic sequence

Exon read: TTGCTGATTT

Exon read: TAAACAAACT

Exon-junction read 1: TAGCTGATAT TAAAATCACT

PSI= 2/3= 0.66 (middle exon is included in 66% of the transcripts)
Exon-junction read 2: CCACATTGTC TAAACAAACT

Exon junction read 3: TAGCTGATAT TAAACAAACT 3

Alternative splicing can be quantified based on the PSI metric

Percent spliced-in (PSI)= Inclusion reads

Inclusion reads + exclusion reads

PSI (for exon 2) = 5/10= 0.5 (Middle exon is included in 50% of the transcripts)
Week 11: Post-transcriptional gene silencing

miRNAs, siRNAs, and RNAi

5
 Post-transcriptional gene silencing (PTGS)
PTGS: form of gene silencing that takes place after mRNA transcripts Small RNAs make a big splash
have been synthesized. 2002

MODULE: small regulatory RNAs

• Micro-RNAs (miRNAs) biogenesis and function

• Small interfering RNAs (siRNAs) biogenesis and function

• Role of miRNAs in development and disease

• Techniques: Soft agar colony formation assay and CLASH

MODULE: The discovery of RNAi

6
• MicroRNAs (miRNAs): These are derived from primary transcripts which fold into a stem-loop
structure (containing some unpaired nucleotides) and downregulate cytoplasmic mRNAs through
reduction of translational efficiency and mRNA degradation.

• Small interfering RNAs (siRNAs). These are derived from longer double stranded RNAs (endogenous
or exogenous e.g., viral).
Sources of endogenous RNAs include:
Ø RNA transcripts that have fold-back structures
Ø Both sense and antisense RNA are transcribed from a gene
siRNAs target RNAs for cleavage in the cytoplasm and are though to act as a cellular defense mechanism
against exogenous RNAs.

7
 miRNA: key features

• Small non-coding regulatory eukaryotic RNAs that direct post-transcriptional gene silencing through
reduction of translation efficiency and mRNA degradation.

• Are encoded in the genome as specific genes and bind to complementary sequences in their mRNA targets.

• miRNAs have a 5’ cap and are polyadenylated.

mRNA 3’UTR
• >60% of human protein-coding genes are regulated by miRNAs

• A single miRNA can downregulate translation from 100-200 different mRNAs

• A single mRNA can be targeted by multiple miRNAs (amplification of translational repression)

• miRNA genes: 147 in C.elegans, 164 in fruit flies, 475 in mouse, and 519 in human (2019 data).

• 277 miRNA genes are conserved between humans and rodents

• Regulate development. Altered miRNA expression can contribute to diseases such as cancer.

8
miRNA and siRNA biogenesis

9
 miRNA biogenesis and function

miRNAs are transcribed by RNA pol II to

produce a primary miRNA (pri-mRNA)
transcript.

Pri-miRNA folds back onto itself

to form a hairpin structure
Pri-miRNA is cleaved by a microprocessor complex:
Drosha + DGRC8
Drosha has 2 RNase domains that each cut a
strand of the Pri-mRNA to form a Pre-miRNA hairpin

~60 nucleotide long

Pre-miRNA hairpin
with 5’ monophosphate and 3’-OH

Pre-miRNA is exported to the cytoplasm.

10
DOI: 10.1016/j.cell.2018.03.006
 miRNA biogenesis and function (continued)

In the cytoplasm, the Pre-miRNA loop is removed by an RNase called Dicer

The resulting product is called miRNA duplex or

miR-miR* duplex which has 2-nucleotide 3’overhang at each end.
miR:miR* molecules are usually not fully complementary
Red strand=guide strand
Blue strand= passenger strand miR:miR* loaded into an Argonaute (Ago) protein, which
miRISC undergoes conformational change and expels the passenger
strand while keeping the guide strand. This complex is called
Extensive pairing between mRNA miRISC (RNA induced silencing complex). The strand with a 5’ U or A
and miRNA- silencing complex slices is preferably retained by Ago.
the mRNA (seen in plants).
Partial base-pairing between miRNA and target
mRNA=> The silencing complex binds to 3’UTR sites
within mRNA and directs other types of
repression (seen in humans and other
mammals).

Perfect match required 11

DOI: 10.1016/j.cell.2018.03.006
between seed positions 2-7
 miRISC represses gene expression through multiple mechanisms
Note: you do not have to
memorize protein names
on this slide

1) Ago recruits an adaptor protein called TNRC6 ,which binds to PABP and recruits CCR4-NOT deadenylase complex,
which shortens the poly (A) tail. This results in destabilization of mRNA which is then degraded through decapping and
5’-to-3’ decay .
2) Additionally, other complexes are recruited that prevent formation of the pre-initiation complex that is necessary
for translational initiation .
Note: miRNA activity results in knockdown of gene function but it is not a complete knockout
DOI: 10.1016/j.cell.2018.03.006 12
 siRNA biogenesis and function

Note: 1) binding of siRNA-RISC complex to the target

mRNA results in mRNA slicing
2) siRNA pathway causes knockdown of a gene
but it is not a knockout.

~22 nt long
siR::siR*are usually fully complementary

AAAAAA
siRISC 6

Plants use siRNA pathway also called RNAi (RNA

interference) to silence transposable elements AAAAAA
as well as in antiviral defense. 13
https://doi.org/10.1038/nature02874 14
 Post-transcriptional gene silencing (PTGS)

• Micro-RNAs (miRNAs) biogenesis and function

• Small interfering RNAs (siRNAs) biogenesis and function

• Role of miRNAs in development and disease

• The discovery of RNAi and it’s broad application to genetic analysis

15
 miRNAs promote maternal-to-zygotic transition (MZT)

MZT
MBT
zygote syncytial blastoderm cellular blastoderm
A P gastrulation

fast slow cell

nuclear nuclear divisions
divisions divisions
cell cycle 11 cell cycle 14
M M
G2 G1
S S
Zygotic gene expression

Maternal gene
products
16
 miRNAs promote maternal-to-zygotic transition (MZT)
miR-309 cluster contains 8 miRNAs: miR-3, miR-4, miR-5, three copies of miR-6, miR-286, and
miR-309

MZT
MBT
zygote syncytial blastoderm cellular blastoderm
A P gastrulation

fast slow cell

nuclear nuclear divisions
divisions divisions
cell cycle 11 cell cycle 14
M M
miR-309 G2
cluster G1
S degrades maternalS
transcripts Zygotic gene expression

Maternal gene
products
17
 miRNAs and disease
Proto-oncogene and oncogene: Proto-oncogenes code for proteins that promote cell proliferation and inhibit apoptosis.
An oncogene is defined as an altered proto-oncogene that can induce malignant transformation of cells in culture or in
model organisms. A proto-oncogene could become an oncogene due to mutations that lead to increased protein
expression.

One of the ways by which protein levels of proto-oncogenes are controlled is through regulation by miRNAs. A recently
recognized mechanism of oncogene activation is the loss of miRNA binding sites.

3’UTR 3’UTR

GENE GENE

18
Nature Reviews Genetics volume 14, pages 496–506 (2013)
 Proto-oncogene IMP-1

Proximal PAS 1 2 3 4 5 Distal PAS

IMP-1
let-7 binding sites

• IMP-1, Insulin like growth factor 2 mRNA binding protein 1, is a proto-

oncogene.
Proto-oncogenes upon acquiring mutations act as oncogenes and thus contribute to tumorigenesis.
• IMP-1 is overexpressed in a several human cancers.
• let-7 is a miRNA.

Q: Are levels of IMP-1 protein regulated by let-7 miRNA?

19
Cell (2009) 138: 673-684 DOI: 10.1016/j.cell.2009.06.016
 Luciferase activity increased significantly upon loss of let-7 binding sites
long-mut-let-7 PAS 1 2 3 4 5 PAS
Loss of let-7 binding sites increased protein
levels by 1.4- to 1.8-fold Luciferase
let-7 binding sites mutated

long-wt
PAS 1 2 3 4 5 PAS
Luciferase
let-7 binding sites

Protein levels measured using

Luciferase reporter assay.

>>

Lung cancer
Breast cancer 20
Sarcoma
Q2: Is the loss of regulation by miRNA in an oncogene sufficient to induce cancer cell
behavior? In other words, can expression of IMP-1 lacking let-7 miRNA binding sites induce
malignant transformation of cells?

21
 Assay: Soft Agar Colony Formation which is indicative of malignant transformation of cells. Cell line used: mouse
fibroblasts NIH3T3.
Rationale behind the technique: Cells are grown in soft agar which acts as a viscous gel. Normal cells are unable to grow
since they need to attach to a solid substrate through the extracellular matrix (anchorage-dependent growth).
Transformed cells can grow without a substrate and form colonies (anchorage-independent growth)

Cells are grown in a layer of soft agar mixed with cell culture medium that rests on another layer of soft agar, also
mixed with cell culture medium, but containing a higher concentration of agar. This prevents cells from adhering
to the culture plate.

https://www.creative-bioarray.com/services/soft-agar-colony-formation-assay-service.htm 22
 Assay: Soft Agar Colony Formation

Normal cells do not form colonies Transformed cells form colonies

• Considered most stringent in vitro tests for transformation, as only malignant cells would have the ability to proliferate
under the anchorage-independent condition.
• The ability of cancer cells to form colonies correlates closely with that of in vivo tumorigenicity as measured by the
formation of tumors in animal models.”

23
 Loss of let-7 miRNA binding sites in IMP-1 is sufficient to induce cancer cell behavior
NIH 3T3 cells transfected with IMP-1 mutant variant
long-wt PAS
PAS 1 2 3 4 5
oncogenic IMP-1
2
let-7 binding sites
3 PAS
long-mut-let-7 PAS 1 2 3 4 5
IMP-1
3

mutated
let7 site
let-7 binding sites mutated

1. Negative control
2. AIM: to examine the effect of long 3’UTR isoform of IMP-1
OBSERVATION: colony formation comparable to the negative control
3. AIM: to examine the effect of increased protein levels of IMP-1
2 (via loss of let-7 miRNA binding sites) on malignant transformation of cells
1 OBSERVATION: significant increase in fraction of cells forming colonies
(compare 2 & 3)

Data not shown: ~3-fold increase in IMP-1 mRNA

and protein expression was seen in case 3
24
Cell (2009) 138: 673-684
Summary :

• The 3’UTR is an important site for translational control.

• miRNAs play an important role in fine tuning of gene expression

• Escape from miRNA-mediated repression is one of the mechanisms by which oncogene
activation occurs in cancer cells.

25
 miRNAs play an important role in regulating tissue growth
FYI
Red: oncogenic miRNAs are found
to be overexpressed in breast cancer

Green: tumor suppressor miRNAs

known to be downregulated in
breast cancer

26
https://doi.org/10.1016/j.phrs.2015.04.015
 miRNAs as a therapy for cancer

MiRNA therapeutics FYI &

posted on Quercus

miRNA drugs- considered as “next-generation” therapeutics.

The common miRNA market size was estimated at $ 160.5
million in 2017. Several biopharmaceutical companies are
involved in RNA therapeutics development.

27
 CLASH (Cross-linking, ligation, and sequencing of hybrids)
AIM: To identify in vivo targets of miRNAs
1
to cross-link
Ago and RNA

e
7

e nc
2

u
Seq
6 ‘UNAfold’
T4 RNA ligase
3 8

and
CR
RT-P
Proteinase K
4 to degrade Ago
+ RNA isolation
5

miRNA mRNA
28
Data from paper
 CLASH (Cross-linking, ligation, and sequencing of hybrids)
STEP 1: UV crosslink Ago-RNA complex

STEP 2: Use anti-Ago antibody to immuno-precipitate Ago-RNA complex on beads

STEP 3: Use T4 RNA ligase to ligate RNA to miRNA

STEP 4: Proteinase K treatment to degrade Ago. Isolate RNA.

STEP 5: RT-PCR to generate cDNA library and perform NGS (Illumina sequencing)

Processing sequencing data

STEP 1: Map the reads to the transcriptome and identify reads that map to both miRNAs and mRNAs

STEP 2: Use software tools (e.g. UNAfold) to identify region of miRNA that interacts with mRNA

Negative controls: (1) miRNA knockout cells (if the experiment is being conducted to detect mRNA targets of a specific
miRNA)

(2) No antibody control to detect non-specific binding to the beads

Positive control: check the data to see if you detect known miRNA-mRNA interactions 29
Post-transcriptional gene silencing (PTGS)

MODULE: The discovery of RNAi

30
 The discovery of RNA interference

A great paper that gave rise to a new field of research

Potent and specific genetic interference by double-stranded

RNA in Caenorhabditis elegans
A Fire , S Xu, M K Montgomery, S A Kostas, S E Driver, C C Mello
Nature . 1998 Feb 19;391(6669):806-11. doi: 10.1038/35888

Unc-22 gene: loss-of-function mutation caused a twitching

phenotype.

FYI: https://www.nobelprize.org/prizes/medicine/2006/fire/lecture/

31
 (Twitching phenotype in larvae)

Slide from the Nobel prize lecture by Andrew Fire 32

 (Twitching phenotype in larvae)

Slide from the Nobel prize lecture by Andrew Fire 33

 The discovery of RNAi

Sense RNA

Parent

Worms exhibit motion abnormalities

Offspring

Cell Death and Differentiation(2007)14,1998–2012; doi:10.1038/sj.cdd.4402253

34
 The discovery of RNAi

Sense RNA
RNA preparations were not pure and had contaminating bands.
Antisense RNA

Cell Death and Differentiation(2007)14,1998–2012; doi:10.1038/sj.cdd.4402253 35

 The discovery of RNAi
The twitching effect was lost upon injection of purified sense or antisense strands suggesting that something else in the
unpurified samples was silencing Unc-22 gene
Purified antisense RNA
Purified sense RNA

Cell Death and Differentiation(2007)14,1998–2012; 36

doi:10.1038/sj.cdd.4402253
 The discovery of RNAi
1. Interestingly, injection of double-stranded RNA produced the twitching effect. This form of gene silencing was termed RNAi.
2. They also found that a few molecules of dsRNA per cell (~ 30/cell) were effective at destroying a much larger amount of mRNA
suggesting that the process was catalytic in nature
3. Silencing effect was transmitted to several generations suggesting an amplification mechanism.

Purified antisense RNA

Purified sense RNA

37
Cell Death and Differentiation(2007)14,1998–2012; doi:10.1038/sj.cdd.4402253
 The discovery of RNAi
4. The authors also showed that interference targeted against GFP
L1 drastically reduced the amount of GFP seen in a GFP-expressing strain
(left).

Fluorescence images of the progeny of injected animals carrying a GFP

Adult reporter construct:
(a) and (b) GFP expression is not lost in progeny of animals injected with
control dsRNA indicating that RNAi is sequence-specific.

L1
(c) and (d) GFP expression is lost in animals injected with dsRNA against GFP.
Some cells are resistant to interference effect.

Adult

Box Figure 13-02-02

DOI: 10.1038/35888 Craig et al., Molecular Biology 38

Oxford University press, 2021
 The discovery of RNAi
5. mex-3 transcripts were not detected after treatment with dsRNA => RNA-dependent RNA degradation

DOI: 10.1038/35888 39
 The discovery of RNAi
6. RNAi is effective only when the dsRNA is made to target the exonic regions of the mRNA suggesting that the interference target
is a mature mRNA molecule and that RNAi causes gene silencing post-transcriptionally. Additionally, RNAi approach could be
used to silence any gene => RNAi effect is not limited to a few genes

(dsRNA)

(dsRNA)
(dsRNA)
(dsRNA)
(dsRNA)
(dsRNA)

DOI: 10.1038/35888 40
 Mechanism of RNAi

~22 nt long

5
siRISC
AAAAAA
6

AAAAAA

41
The Nature paper is considered as a landmark paper, but several other key papers helped define RNAi.
https://www.nature.com/articles/4402253.pdf?origin=ppub

42
 Summary:

§ Fire, Mello and colleagues found that the best trigger of interference was double-stranded RNA
- Injection with double-stranded unc-22 (sense plus antisense; S+AS) led to twitching while the single stranded
ones did not.

§ Non-specific RNAs did not elicit a response, so interference was sequence-specific.

§They further explored the mechanism of interference by testing introns which did not induce interference. This
suggested that the interference target has already been processed to mature mRNA,

§ Observation of interference in progeny, where the amounts of RNA would be very small, suggested an
amplification mechanism

Many of the mechanistic features in this landmark publication have turned out to be broadly accurate in many
organisms. By simply delivering dsRNA to a given organism, large-scale screens can be carried to determine
gene function in a matter of days.

Craig et al., Molecular Biology FYI RNAi drug: https://www.nature.com/articles/d41586-018-05867-7 43

Oxford University press, 2021
 How do you experimentally introduce dsRNA to carry out large-scale screens?

dsRNA constructs are

available against 94%
of C.elegans genes.

Figure 16.5
Elliott D and Ladomery M,
Molecular Biology of RNA 44
 Method for expressing dsRNA in bacteria dsRNA
dsRNA synthesis is induced in E. coli by using a vector that contains two T7 RNA polymerase promoters flanking
the cDNA sequence of your gene of interest.

Phillip A. Newmark et al. PNAS 2003;100:suppl 1:11861-

11865

45
https://doi.org/10.1073/pnas.1834205100
Clone cDNA of your gene of interest within the Multiple cloning site. T7 RNA polymerase will bind to the T7 promoters and
synthesize Sense and antisense RNA from the gene.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5396541/figure/F5/?report=objectonly

46
 Method for feeding bacteria expressing dsRNA to planarians
dsRNA synthesis is induced in E. coli by using a vector that
contains two T7 RNA polymerase promoters flanking the cDNA
sequence of your gene of interest.

Phillip A. Newmark et al. PNAS 2003;100:suppl 1:11861-

11865

47
https://doi.org/10.1073/pnas.1834205100
RNAi has enabled large scale screens to identify genes involved in regeneration in planarian flatworms

1065 genes were screened, 240 genes were identified that play an important role in regeneration
FYI

48
10.1016/j.devcel.2005.02.014
 RNAi controls and limitations

Ø Negative control- dsRNA against a gene that is not expressed e.g., GFP

Ø Positive control- dsRNA against a gene that is ubiquitously expressed e.g., house-keeping genes

Limitations:

Ø RNAi results in gene knockdown and not knockout

Ø siRNAs (or dsRNA) can have off-target effects. Best strategy is to knockdown a gene using two or three independent siRNAs to
eliminate the possibility of non-specific effects

49
 shRNA is a useful tool for spatial and temporal knockdown of gene expression

shRNA (short hairpin RNA) construct is cloned into a suitable vector. The plasmid is introduced into organism/ cell lines of
interest. Stable transgenics/ cell lines can thus be established.

Promoterr shRNAr

• Processed into siRNAs

• Inducible, can be expressed in a spatial and temporal manner Vector
• Useful especially in the case of essential genes where knockdown in the entire
animal would result in lethality.
• Can design shRNA against a specific mRNA splice isoform

Figure 16.8 Elliott D and Ladomery M,

Molecular Biology of RNA 50
51