University of Tennessee, Knoxville

TRACE: Tennessee Research and Creative

Exchange

Doctoral Dissertations Graduate School

5-2023

Investigating Drivers of Algal Bloom Succession in Lake Erie

Brittany Zepernick
University of Tennessee Knoxville, bzeperni@vols.utk.edu

Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss

Part of the Environmental Microbiology and Microbial Ecology Commons

Recommended Citation
Zepernick, Brittany, "Investigating Drivers of Algal Bloom Succession in Lake Erie. " PhD diss., University of
Tennessee, 2023.
https://trace.tennessee.edu/utk_graddiss/8127

This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee
Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized
administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact
trace@utk.edu.
To the Graduate Council:

I am submitting herewith a dissertation written by Brittany Zepernick entitled "Investigating

Drivers of Algal Bloom Succession in Lake Erie." I have examined the final electronic copy of this
dissertation for form and content and recommend that it be accepted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy, with a major in Microbiology.

Steven W. Wilhelm, Major Professor

We have read this dissertation and recommend its acceptance:

Erik R. Zinser, Jill A. Mikucki, George S. Bullerjahn

Accepted for the Council:

Dixie L. Thompson

Vice Provost and Dean of the Graduate School

(Original signatures are on file with official student records.)

INVESTIGATING DRIVERS OF ALGAL BLOOM SUCCESSION IN LAKE ERIE

A Dissertation Presented for the

Doctor of Philosophy
Degree
The University of Tennessee Knoxville

Brittany Noel Zepernick

May 2023
Copyright © 2023 by Brittany Noel Zepernick
All rights reserved.

ii
 DEDICATION

To my brilliant and fearless mentor • Steven Wilhelm,

my academic dads • George Bullerjahn, Mike McKay, Hans Paerl, & Jeff Krause,

my unwaveringly supportive parents • Nanette & Randy Zepernick,

my hard-working grandparents • Clyde, Ida Zepernick & Richard, Pauline Wilcox

my ambitious counterpart and loving husband • Joe Caniglia

my encouraging father-in-law • Dave Caniglia

my coven of cats & dogs.

iii
 ACKNOWLEDGEMENTS

I must acknowledge my undergraduate professors, Dr. R. Michael L. McKay and Dr.

George S. Bullerjahn, who sparked my passion for harmful algal bloom research and encouraged

me to pursue a graduate degree. I want to thank George for recommending Dr. Wilhelm as a

graduate mentor and telling me to “put a star next to that one”. I also wish to thank Dr. Jeffrey

Krause for mentoring me as an undergraduate NSF REU student in 2016, despite the fact I had

never touched a pipette prior. I thank Dr. Becca Pickering for being my first graduate student

mentor and for teaching me how to use said aforementioned pipette (among other useful things).

I also want to thank Dr. Hans Paerl, whose guidance, input and candor has been critical in

shaping my career and successes to date. These individuals, whom I often refer to as my

“academic dads”, have consistently cheered me on throughout my career.

Dr. Wilhelm is responsible for the scientist I am today. He is an exceedingly gifted

scientist who encourages us to push the boundaries of science, challenge ourselves, and cultivate

a passion for our work. Beyond his countless accomplishments as a scientist, he is an exemplary

mentor. From my first day as a graduate student in his lab, he has never failed to encourage,

guide, support and inspire me. A compassionate mentor, he has repeatedly placed the needs of

his students above his own, including sometimes a need for sleep as he is prone to answering our

frantic science emails sent at all hours of the night. He has fostered a collaborative, supportive,

inclusive laboratory family to which I am eternally grateful to be a part of. In short, all of my

academic successes and accomplishments I owe to him.

The Wilhelm lab family is just that, my second family. Over the past 5 years, these

individuals have seen me through exuberant successes and rattling failures. While there are too

many to name them all, I specifically want to thank Robbie Martin, Lauren Krausfeldt, Lena

iv
Pound, Eric Gann, Naomi Gilbert, Alex Truchon, Emily Chase, and Katelyn Houghton, for your

inspiration, collaboration, and unwavering support throughout the years. I would not have been

able to accomplish half of what I did without you, and you have made these past five years some

of the best of my life. Beyond the lab family, I want to thank my cohort (Alexandra, Liz, Elise,

Katie and Sean) for your love, laughter, and support. I must also thank the Knox establishments

which offered us a relaxed respite and liquid encouragement.

I want to thank my grandparents and those who paved the way before me, whose

shoulders I stand on today. I thank my dad’s side (the Zepernick family), who for generations

earned their keep farming. I thank my mom’s side (the Wilcox family), who came from humble

coal mining and fishing families. Notably, I thank my grandfather Richard Wilcox, who had to

drop out of college and never complete his degree, but always spoke of his love for science, the

sea, and “bacteriology”. I thank him for embodying the act of being “confident yet humble” and

encouraging me throughout this journey until he had to move on. I finished this degree not only

for myself, but for you. This degree belongs to my bloodlines as much as it does to me.

I owe this degree to my family and to my parents, who have shown me what true hard

work and sacrifice look like. To my mom and dad, I thank you for your relentless

encouragement, love and unwavering support. I thank my father in law, Dave, for taking me on

as the daughter you never had and never failing to offer fatherly wisdom, albeit usually

unsolicited. To my husband Joe, this degree would not have been possible without you. The

sacrifices that you’ve made so that I could pursue my goals are many, and you have been my

biggest supporter. You have been steadfastly committed to this journey with me, riding shotgun

through every high and low. I am eternally grateful for your love, support and your (sometimes

annoying) willingness to challenge me. Thank you for being my counterpart in life.

v
 ABSTRACT

Harmful Algal Blooms (HABs) are algae undergoing prolific, unregulated growth. A

well-documented HAB taxa is the cyanobacterium Microcystis spp., which induces

anthropogenic, ecological, and economic consequences due to the production of toxins and

biomass. Microcystis spp. blooms are globally distributed in freshwater systems, with climate

change and the aquatic continuum serving to further exacerbate bloom distribution, duration, and

frequency. Lake Erie of the Laurentian Great Lakes experiences annual summer Microcystis spp.

blooms which have rendered the drinking water unfit for human consumption (i.e., the Toledo

Water Crisis of 2014). As a result, these events have incited a concerted effort within the field at

large to elucidate the anthropogenic and ecological drivers of these detrimental summer

cyanobacterial blooms. Thus, there exists an expansive amount of research concerning summer

Microcystis spp. blooms in Lake Erie. Yet, gaps in the field exist, and there remains a need to

elucidate the factors driving the ecological success of summer Microcystis blooms and the

ecological “failures” of their competitors (i.e., diatoms). In contrast, diatom blooms dominate the

winter water column of Lake Erie. While these blooms are non-toxic, the dense biomass of these

blooms has been found to induce massive summer hypoxia events and dead zones. Yet, despite

the magnitude of these winter diatom blooms, they remain widely uncharacterized and unstudied

to date. Hence, a seasonal pattern of algal bloom succession occurs within Lake Erie: toxic

Microcystis spp. blooms dominate summer-fall, and nuisance diatom blooms dominate winter-

spring. My dissertation utilizes physical assays and molecular approaches to conduct in vitro, in

situ, and in silico experiments to investigate the factors contributing to the ecological success

and succession of these respective bloom-forming classes. Hence, this work serves to assess

Lake Erie algal bloom success and succession across spatial, temporal and climatic scales.

vi
 TABLE OF CONTENTS
CHAPTER I: A TALE OF TWO BLOOMS: REFINING ECOLOGICAL PARADIGMS
OF LAKE ERIE ALGAL BLOOM SUCCESS AND SUCCESSION ..................................... 1
PUBLICATION NOTE ..................................................................................................................... 2
ABSTRACT ................................................................................................................................... 3
ECOLOGICAL SUCCESS OF LAKE ERIE CYANOBACTERIAL BLOOMS............................................... 4
ECOLOGICAL SUCCESS OF LAKE ERIE DIATOM BLOOMS ............................................................... 7
SUCCESSIONAL DYNAMICS BETWEEN CYANOBACTERIAL AND DIATOM BLOOMS .......................... 8
REFINING THE TEMPERATURE PARADIGM OF ALGAL BLOOM SUCCESSION .................................. 11
REFINING NUTRIENT PARADIGMS OF ALGAL BLOOM SUCCESSION .............................................. 14
PH A NOVEL PARADIGM OF ALGAL BLOOM SUCCESSION? ........................................................... 16
CONCLUSION - CAVEAT BIOLOGUS ............................................................................................ 19
DISSERTATION OVERVIEW.......................................................................................................... 20
REFERENCES: ............................................................................................................................. 24
CHAPTER II: RE-ASSESSING CONSEQUENCES OF CLASSICAL LABORATORY
TECHNIQUES USING MICROCYSTIS AERUGINOSA AS A MODEL SYSTEM............ 35
PUBLICATION NOTE ................................................................................................................... 36
ABSTRACT ................................................................................................................................. 37
INTRODUCTION .......................................................................................................................... 38
METHODS .................................................................................................................................. 39
RESULTS .................................................................................................................................... 46
DISCUSSION ............................................................................................................................... 52
COMMENTS AND RECOMMENDATIONS ....................................................................................... 56
ACKNOWLEDGMENTS ................................................................................................................ 57
REFERENCES: ............................................................................................................................. 58
APPENDIX .................................................................................................................................. 60
CHAPTER III: ELUCIDATING THE ROLE OF PH ON MICROCYSTIS-DIATOM
COMPETITION DYNAMICS IN LAKE ERIE...................................................................... 67
PUBLICATION NOTE ................................................................................................................... 68
ABSTRACT ................................................................................................................................. 69
INTRODUCTION .......................................................................................................................... 70
RESULTS .................................................................................................................................... 79
DISCUSSION ............................................................................................................................... 88
ACKNOWLEDGEMENTS............................................................................................................... 94
REFERENCES: ............................................................................................................................. 95
APPENDIX .................................................................................................................................. 99
CHAPTER IV: INVESTIGATING ELEVATED PH EFFECTS ON FRESHWATER
DIATOM TRANSCRIPTION, MORPHOLOGY AND PHYSIOLOGY ........................... 120
PUBLICATION NOTE ................................................................................................................. 121
ABSTRACT ............................................................................................................................... 122
INTRODUCTION ........................................................................................................................ 123
METHODS ................................................................................................................................ 124
RESULTS .................................................................................................................................. 129

vii
 DISCUSSION ............................................................................................................................. 138
CONCLUSION............................................................................................................................ 151
ACKNOWLEDGEMENTS............................................................................................................. 152
REFERENCES: ........................................................................................................................... 153
APPENDIX ................................................................................................................................ 161
CHAPTER V: DIATOM RESPONSES TO DECREASING ICE COVER IN LAKE ERIE
..................................................................................................................................................... 173
PUBLICATION NOTE ................................................................................................................. 174
ABSTRACT ............................................................................................................................... 175
INTRODUCTION ........................................................................................................................ 176
METHODS ................................................................................................................................ 178
RESULTS .................................................................................................................................. 181
ACKNOWLEDGEMENTS............................................................................................................. 202
REFERENCES: ........................................................................................................................... 203
APPENDIX ................................................................................................................................ 209
CHAPTER VI: CONCLUSIONS ............................................................................................ 243
VITA........................................................................................................................................... 249

viii
 LIST OF TABLES
Table 2.1. Buffering characteristics of five freshwater culture media…………………………….40
Table 3.1. Statistical analysis of F. crotonensis growth rates as a function of pH………………...99
Table 3.2. Statistical analysis of M. aeruginosa growth rates as a function of pH……………….100
Table 3.3. Statistical analysis of in vitro F. crotonensis Si deposition as a function of pH………101
Table 3.4. Statistical analysis of in situ diatom Si deposition as a function of pH……………….102
Table 4.1. Statistical analysis of F. crotonensis FlowCAM data………………………………...161
Table 4.2. Statistical analysis of F. crotonensis abundance and Chl a autofluorescence………...161
Table 4.3. Statistical analysis of F. crotonensis photopigment data……………………………..161
Table 4.4. Statistical analysis of F. crotonensis PhytoPAM data………………………………..162
Table 5.1. Metatranscriptomic libraries with respect to study variables………………………...218

ix
 LIST OF FIGURES
Figure 1.1. Cyanobacterial blooms in the Lake Erie summer water column (2012-2017)…………6
Figure 1.2. Diatom blooms in the Lake Erie winter water column (2007-2010)…………………...9
Figure 2.1. Flame-induced declines in pH in control and flamed freshwater media ……………...48
Figure 2.2. Net change in pH observed in each media after 10 days of aseptic flaming…………...48
Figure 2.3. Total carbon in control and flamed CT freshwater media replicates………………….51
Figure 2.4. Aseptic flaming effects on CT media and axenic M. aeruginosa cultures…………….53
Figure 2.5. Effects of flaming on heterotrophic culturing media…………………………………60
Figure 2.6. Effects of environmental conditions on pH of CT media……………………………..61
Figure 2.7. Determination of CO2 as the primary driver of pH decline in freshwater media………62
Figure 2.8. Mitigation of pH decline due to flaming of CT media using TAPS amendments……..63
Figure 2.9. Effects of tube inversion on control and flame replicate pH decline……………...…..64
Figure 2.10. Correlation between Chlorophyll a autofluorescence and cell density……………...65
Figure 2.11. Comparison of M. aeruginosa biomass accumulation in CT vs. amended CT……....66
Figure 3.1. Environmental data corresponding to a 2015 Lake Erie Microcystis bloom………….80
Figure 3.2. In vitro F. crotonensis monoculture growth curves as a function of pH……………....82
Figure 3.3. In vitro F. crotonensis co-culture growth curves……………………………………..83
Figure 3.4. Epifluorescent microscopy of F. crotonensis filaments labeled with PDMPO….........85
Figure 3.5. Si deposited per F. crotonensis filament as a function of pH………………………….87
Figure 3.6. Si deposited per Chl a concentration in environmental diatoms as a function of pH….87
Figure 3.7. Dissolved Si profiles corresponding to a 2015 Lake Erie Microcystis bloom……….103
Figure 3.8. Chl a concentrations corresponding to a 2015 Lake Erie Microcystis bloom……......104
Figure 3.9. FlowJo graph depicting a gated population of F. crotonensis filaments………….....105
Figure 3.10. Standard curve for in vitro F. crotonensis Si deposition assay…………………......106
Figure 3.11. Standard curve for in situ Lake Erie diatom Si deposition assay…………………...107
Figure 3.12. In vitro M. aeruginosa monoculture growth curves as a function of pH…………...108
Figure 3.13. In vitro M. aeruginosa co-culture growth curves…………………………………..109
Figure 3.14. pH drift in F. crotonensis monocultures…………………………………………...110
Figure 3.15. pH drift in M. aeruginosa monocultures…………………………………………...111
Figure 3.16. pH drift in F. crotonensis and M. aeruginosa co-cultures………………………….112
Figure 3.17. Final pH of F. crotonensis and M. aeruginosa monoculture and co-cultures………113
Figure 3.18. Simple linear regressions of M. aeruginosa concentration and pH drift…………...114
Figure 3.19. Epifluorescent microscopy of in vitro F. crotonensis filaments + PDMPO……......115
Figure 3.20. Epifluorescent microscopy data of F. crotonensis PDMPO assay…………………116
Figure 3.21. Total Si deposited per F. crotonensis culture as a function of pH………………….117
Figure 3.22. Total Si deposited per replicate of in situ diatoms as a function of pH…………......118
Figure 3.23. Water column profiles collected during a 2009 Lake Erie winter survey………......119
Figure 4.1. Heat maps depicting DE genes relating to F. crotonensis morphology/growth…......131
Figure 4.2. FlowCAM data of F. crotonensis pH assay…………………………………………133
Figure 4.3. Heat maps depicting DE genes relating to F. crotonensis energy production……….135

x
Figure 4.4. Photopigment data of F. crotonensis pH assay……………………………………...137
Figure 4.5. PhytoPAM data of F. crotonensis pH assay…………………………………………139
Figure 4.6. Heat map depicting DE genes relating to the F. crotonensis mobilome…………......140
Figure 4.7. nMDS of similarity between F. crotonensis pH transcriptomes………………….....163
Figure 4.8. Top 50 genes contributing to F. crotonensis pH transcriptome dissimilarity……......164
Figure 4.9. Volcano plots of DE F. crotonensis genes as a function of pH………………………165
Figure 4.10. Heat map depicting DE genes relating to F. crotonensis Carbon metabolism……...166
Figure 4.11. Flow cytometry and fluorometer data for photopigment pH assay………………...167
Figure 4.12. Total Chl a concentrations from photopigment pH assay……………………….....168
Figure 4.13. Fucoxanthin and Neoxanthin concentrations from photopigment pH assay…….....169
Figure 4.14. Ratio of total carotenoids: total Chl a from photopigment pH assay……………….170
Figure 4.15. Ratios of carotenoids: total Chl a from photopigment pH assay…………………...171
Figure 4.16. Ratio of Chl a: Chl c1c2 from photopigment pH assay……………………………..172
Figure 5.1. Spatial and climatic variability across metatranscriptome samples…………………182
Figure 5.2. Characterization of the biotic community across 12 winter Lake Erie sites…............184
Figure 5.3. Relative transcript abundance of MEPT and diatom classes………………………...186
Figure 5.4. nMDS of 20winter-spring Lake Erie metatranscriptomic libraries (TPM)……….....187
Figure 5.5. Diatom transcript abundance patterns in response to ice cover of COG C………......190
Figure 5.6. Diatom transcript abundance patterns in response to ice cover of COG M………….191
Figure 5.7. Phylogenetic tree of fasciclin gene distribution within diatoms…………………......193
Figure 5.8. Temperature/nutrient profiles across the 12 sample sites organized by season….......219
Figure 5.9. Contribution of Chl a >20 µM to total Chl a across sample sites……………………220
Figure 5.10. Cell abundance of MEPT across sample sites……………………………………...221
Figure 5.11. Cell abundance of centric filamentous diatoms across winter sample sites………...222
Figure 5.12. Cell abundance of small centric diatom taxa (5-20µm) across sample sites……......223
Figure 5.13. Mean percent contribution of three centric diatoms to total diatom abundance……224
Figure 5.14. Relative transcript abundance of domains across 20 metatranscriptomes………….225
Figure 5.15. Relative transcript abundance of major eukaryota groups…………………………226
Figure 5.16. Relative transcript abundance of Mediophyceae genera…………………………...227
Figure 5.17. Relative transcript abundance of Coscinodiscophyceae genera……………………228
Figure 5.18. Relative transcript abundance of Bacillariophyceae genera……………………….229
Figure 5.19. Relative transcript abundance of Fragilariophyceae genera……………………….230
Figure 5.20. Simple linear regression of diatom cell and transcript abundance…………………231
Figure 5.21. ANOSIM of whole community and diatom expression (TPM)……………………232
Figure 5.22. Taxonomic distributions of DE diatom genes by ice cover………………………...233
Figure 5.23. Taxonomic distribution of DE Mediophyceae genes of COG C…………………...234
Figure 5.24. Diatom transcript abundance patterns in response to ice cover of COG P………….235
Figure 5.25. Taxonomic distribution of DE Mediophyceae genes of COG P……………………236
Figure 5.26. Diatom transcript abundance patterns in response to ice cover of COG G…………237
Figure 5.27. Normalized expression (VST) of diatom rhodopsin genes DE by ice cover……….238

xi
Figure 5.28. Distribution of DE genes by COG category comparing season and ice cover……...239
Figure 5.29. Taxonomic distribution of DE Mediophyceae genes of COG M…………………..240
Figure 5.30. MODIS satellite image of ice-free and turbid Lake Erie (2-12-2023)……………...241
Figure 5.31. Plankton net tows from a 2007 Lake Erie winter survey…………………………...242

xii
 LIST OF ATTACHMENTS

Appendix Sheet 4.1

Sheet 4.1A. Transcriptome mapping statistics
Sheet 4.1B. Similarity Percentage (SIMPER) analysis on pH transcriptome genes
Sheet 4.1C. DE genes organized by COG category.
Sheet 4.1D. FlowCAM statistics

Appendix Sheet 5.1-5.2

Sheet 5.1A. JGI and manuscript metatranscriptome library IDs, metadata
Sheet 5.1B. Metatranscriptome coassembly statistics
Sheet 5.1C. Metatranscritpome library mapping statistics
Sheet 5.1D. Water column parameters, metadata
Sheet 5.1E. Water column Chlorophyll a, metadata
Sheet 5.1F. Water column nutrients, metadata
Sheet 5.1G. Water column phytoplankton ID and enumeration, metadata
Sheet 5.1H. Relative transcript abundance of mapped reads for domains
Sheet 5.1I. Relative transcript abundance of mapped reads for Eukaryota groups
Sheet 5.1J. Relative transcript abundance of mapped reads for MEPT
Sheet 5.1K. Relative transcript abundance of mapped reads for diatoms by class
Sheet 5.1L. Relative transcript abundance of mapped reads for Mediophyceae genera
Sheet 5.1M. Relative transcript abundance of mapped reads for Coscinodiscophyceae
Sheet 5.1N. Relative transcript abundance of mapped reads for Bacillariophyceae genera
Sheet 5.1O. Relative transcript abundance of mapped reads for Fragilariophyceae genera
Sheet 5.1P. SIMPER analysis of whole community expression (TPM) as a function of ice
cover (winter libraries 1-14)
Sheet 5.1Q. PRIMER Statistical results. ANOSIM Analysis of whole community
expression (TPM) as a function of ice cover (winter libraries 1-14)
Sheet 5.1R. PRIMER Statistical results. SIMPER Analysis of whole community
expression (TPM) as a function of season (libraries 1-20)
Sheet 5.1S. PRIMER Statistical results. ANOSIM Analysis of whole community
expression (TPM) as a function of season (libraries 1-20)
Sheet 5.1T. PRIMER Statistical results. SIMPER Analysis of Bacillariophyta community
expression (TPM) as a function of ice cover (winter libraries 1-14)
Sheet 5.1U. PRIMER Statistical results. ANOSIM Analysis of Bacillariophyta
community expression (TPM) as a function of ice cover (winter libraries 1-14)
Sheet 5.1V. PRIMER Statistical results. SIMPER Analysis of Bacillariophyta community
expression (TPM_ as a function of season (libraries 1-20)
Sheet 5.1W. PRIMER Statistical results. ANOSIM Analysis of Bacillariophyta
community expression (TPM) as a function of season (libraries 1-20)

xiii
 Sheet 5.1X. DESeq2 results for comparing winter diatom community expression by ice
cover (L1-4 vs. L5-14)
Sheet 5.1Y. DESeq2 results for comparing diatom community expression by season (L1-
14 vs. L15-20)
Sheet 5.1Z. DE genes by ice cover in winter libraries within COG C category
Sheet 5.2A. DE genes by ice cover in winter libraries within COG M category
Sheet 5.2B. DE genes by ice cover in winter libraries within COG P category
Sheet 5.2C. DE genes by ice cover in winter libraries within COG G category
Sheet 5.2D. The eighteen diatom fasciclin protein sequences from this study
Sheet e 5.2E. Fasciclins within the DE dataset and the proteins used to functionally
annotate them from eggNOG
Sheet 5.2F. All diatoms found to contain fasciclins during phylogenetic search

xiv
CHAPTER I: A TALE OF TWO BLOOMS: REFINING ECOLOGICAL PARADIGMS
OF LAKE ERIE ALGAL BLOOM SUCCESS AND SUCCESSION

1
Publication Note
This chapter contains content adapted from a peer-reviewed, published review article in
Environmental Microbiology Reports by Brittany N. Zepernick, Steven W. Wilhelm, George S.
Bullerjahn, and Hans W. Paerl and content from a manuscript in preparation by Brittany N.
Zepernick and Steven W. Wilhelm to be submitted to the Journal of Great Lakes Research.

All authors contributed to the drafting and final version of the manuscripts.

2
Abstract
The popular press and most scientists view Lake Erie as a summertime monoculture of

cyanobacteria propelled by anthropogenic nutrients. While cyanobacterial blooms (often

dominated by Microcystis spp.) deserve attention, the rediscovery of large-scale winter-spring

diatom blooms has challenged our precepts. Lake Erie algal blooms do not pertain to

cyanobacteria alone, as diatom blooms have been historically and consistently reported in the

literature spanning all four seasons. While history also tells us cyanobacterial blooms “like it

hot”, recent data suggests they “also like it cold”: this challenges the classical paradigm that

cyanobacterial and diatom bloom succession is an artefact of differing temperature optima.

Recently, cyanobacterial blooms have emerged in the oligotrophic Lake Superior, further

contesting paradigms (i.e., that cyanobacteria are a symptom of eutrophication while diatoms are

favored in meso-oligotrophic systems). In addition, the recent rediscovery of cyanobacterial

bloom-induced elevated pH in the water column (i.e., lake basification) highlights the pivotal

role of other drivers of algal success and succession. Considering these observations, there is an

acute need to re-assess the ecological successes of bloom-forming phyla within Lake Erie and

the factors which constrain their succession. Here, we evaluate the historical and present status of

Lake Erie algal blooms and their implicit interlinkage. We further suggest potential

reassessments of classic paradigms of algal bloom succession, such as temperature and nutrient

status, and propose a novel successional paradigm: pH. Given lakes are sentinels of climate

change, there is a pressing need to identify how algal blooms “succeed” across seasonal water

columns before climate change further alters the algal bloom cycle. Cumulatively, this work

serves as a cautionary tale that Lake Erie algal blooms cannot be studied solitarily, and

traditional paradigms cannot be applied sweepingly.

3
Ecological success of Lake Erie cyanobacterial blooms
Lake Erie (US/Canada) is a North American treasure: it provides potable water and

serves as an economic fulcrum to over 30 million lake basin residents (Canada and Agency,

1995; Field et al., 1998; Buttle et al., 2004; Millerd, 2005). On a global scale, the Great Lakes

cumulatively serve as the largest freshwater body on Earth containing 20% of the globe’s

freshwater (Ashworth, 1987; Gronewold et al., 2013; Huisman et al., 2018). Yet, Lake Erie has a

long-standing problem with summer cyanobacterial blooms, with “blooms” defined as algae

undergoing unregulated growth which invoke detrimental consequences due to high biomass

and/or the production of toxins (Anderson et al., 2002). Cyanobacterial genera such as

Microcystis, Aphanizomenon and Dolichospermum were initially reported in Lake Erie

throughout the early 1900’s as nutrient loads to Lake Erie increased (Davis, 1954; Allinger and

Reavie, 2013b; Steffen et al., 2014a; Watson et al., 2016). Yet, it was not until the 1950’s, when

nitrogen-fixing Dolichospermum and Aphanizomenon began to form blooms, that the lake

received enhanced public attention (Davis, 1954; 1964; Steffen et al., 2014a; Huisman et al.,

2018). These dense blooms shifted Lake Erie’s trophic status to “hypereutrophic”(Verduin,

1964; Sweeney, 1995; McKindles et al., 2020), a major impetus for the Great Lakes Water

Quality Agreement (Ijc, 1989) instituted to decrease algal blooms. This solution worked for a

time and Lake Erie was pronounced “rejuvenated” (Sweeney, 1995), with cyanobacterial blooms

substantially decreased throughout the 1970’s-1980’s (Makarewicz and Bertram, 1991; Nicholls

and Hopkins, 1993; Makarewicz et al., 1999; Huisman et al., 2018). Yet, cyanobacterial blooms

returned in the mid-1990’s (now dominated by non-nitrogen fixing Microcystis spp. and

Planktothrix spp.), coinciding with the re-eutrophication of Lake Erie (Brittain et al., 2000;

Conroy and Culver, 2005; Bridgeman et al., 2013; Watson et al., 2016). Since then, Microcystis

spp. blooms have dominated the Lake Erie summer season, invoking detrimental consequences

4
for ecosystem and human health (Figure 1.1) (Rinta-Kanto et al., 2009; Bridgeman et al., 2013;

Steffen et al., 2017) While Microcystis spp. blooms have proliferated throughout the western

basin of Lake Erie, the water column is far from a Microcystis monoculture. Planktothrix spp.

blooms have dominated the Sandusky Bay area (Rinta-Kanto and Wilhelm, 2006; McKindles et

al., 2022) and blooms of Dolichospermum spp. and Aphanizomenom spp. manifest throughout

the western-central basins (Wynne and Stumpf, 2015; Chaffin et al., 2019; Yancey et al., 2023).

In contrast to their summer notoriety, cyanobacteria are thought to remain largely sub-

dominant throughout the winter-spring period (though this may be due to a disproportionate

number of limnological surveys focused on the summer water column rather than the winter-

spring). Beyond the lack of ecological studies concerning major cyanobacterial genera in the

winter, we note there are populations of other cyanobacteria that are persistently abundant yet

rarely receive attention. Genera such as Synechococcus and Cyanobium are but a couple

examples of the many picocyanobacteria that can be abundant components of fresh waters

(Wilhelm et al., 2014). These genera contribute to the water column algal biomass (Fahnenstiel

and Carrick, 1992; Carrick and Schelske, 1997) and reach densities exceeding 100,000 cells • L-
1
in both summer (Wilhelm et al., 2006c) and winter periods (Twiss et al., 2012). While these

cells types are persistently present, they are often overlooked and rarely included in assessments

of “the cyanobacteria” in lake systems: given the potential for this size fraction to make up

significant components of water column chlorophyll, it is clear that their assessment needs to be

included by ecosystems ecologists to fully understand system processes (Wilhelm et al., 2003).

Nonetheless, the ecological success of certain cyanobacteria in the summer months has remained

uncontested and well-documented within the Lake Erie literature.

5
Figure 1.1: Cyanobacterial blooms (mainly comprised of Microcystis) in the Lake Erie summer
water column (2012-2017). Pictures were taken across the western basin by Steven W. Wilhelm.

6
Ecological success of Lake Erie diatom blooms
One could argue diatoms lack notoriety due to the fact Lake Erie diatoms are not known

to produce toxins. That said, they are not without ecological consequence, and Lake Erie has a

long history of seasonally abundant diatom communities. According to the paleolimnetic record,

oligotrophic diatoms (Aulacoseira distans and Cocconeis disculus) dominated the water column

along with eutrophic diatoms (Stephanodiscus, Aulacoseira granulata, Cyclotella bodanica)

prior to 1850 (Stoermer et al., 1987; Stoermer et al., 1989; Stoermer et al., 1993; Stoermer et al.,

1996). Cumulatively, these findings indicated Lake Erie has always contained eutrophic diatom

taxa tolerant of high-nutrient conditions (Allinger and Reavie, 2013b). Subsequent nutrient

loading throughout the first half of the twentieth century led to a regime shift within the diatom

community, resulting in dominance by eutrophic diatom taxa such as Fragilaria spp. and

Stephanodiscus spp. throughout the eutrophic western basin (Britt, 1955; Verduin, 1964; Hohn,

1969). In turn, while diatoms have been suggested to have dominated the Lake Erie summer

water column for the first half of the twentieth century (Region and Davis, 1958; Nicholls et al.,

1977), net diatom summer abundance declined by the 1960’s (Britt, 1955; Casper, 1965). Yet,

there remains a forgotten anomaly to this trend; prolific summer Fragilaria crotonensis blooms

were reported throughout the western basin of Lake Erie throughout the late 1960’s to the early

2000’s, with abundances reported as high as 950,000 cells • L-1 (Beeton, 1965; Munawar and

Munawar, 1976; Gladish and Munawar, 1980; Hartig, 1987; Munawar et al., 2008). To date, this

diatom remains a prominent and metabolically active member of the Lake Erie summer water

column (Saxton et al., 2012b). Indeed, in a 2019 Lake Erie study, Fragilaria spp. filaments were

widely abundant in samples from the western basin (unpublished data associated with Zepernick

et al. (2021)). Hence, this overlooked member of the planktonic community remains ecologically

pertinent and present today, though the factors which constrain its ecological success are limited

7
to a few reports (Hartig and Wallen, 1986; Hartig, 1987; Zepernick et al., 2021; Dengg et al.,

2022; Zepernick et al., 2022b). Beyond the historical summer diatom blooms, there are

prominent winter-spring diatom blooms within Lake Erie (Saxton et al., 2012b; Twiss et al.,

2012). Diatom blooms (comprised of Asterionella, Synedra, Stephanodiscus, and Cyclotella)

were reported under Lake Erie ice in the 1930’s (Chandler, 1940; Chandler, 1942; 1944;

Chandler and Weeks, 1945), and remained acknowledged (Stoermer, 1975; Munawar and

Munawar, 1982) yet unstudied until recently. Dense blooms of eutrophic diatoms Aulacoseira

islandica (Coscinodiscophyceae) and Stephanodiscus binderanus (Mediopyceae) were

rediscovered within and underneath Lake Erie ice cover by Twiss et al. (2012) (Figure 1.2).

Subsequent studies determined these diatoms were metabolically active and reached Chl a

concentrations and biovolumes that rivaled summer cyanobacterial blooms. Indeed, the large

biovolumes and associated organic carbon from winter diatoms have been linked to summer

hypoxia (Saxton et al., 2012b; Wilhelm et al., 2014; Reavie et al., 2016). Yet, these studies also

demonstrated the winter diatom community has significantly changed since the foundational

winter Chandler surveys, invoking a call to revisit comprehensive seasonal studies within the

Great Lakes (Ozersky et al., 2021). Cumulatively, these studies support the critical and

longstanding importance of diatom blooms in Lake Erie.

Successional dynamics between cyanobacterial and diatom blooms

As we have noted, Lake Erie literature over the last two decades has focused on summer

cyanobacterial blooms, although recently this has been expanded to include winter-spring diatom

blooms. Cyanobacteria and Bacillariophyta are the two most prominent Lake Erie bloom-

8
Figure 1.2: Diatom blooms (mainly comprised of Aulacoseira islandica and Stephanodiscus) in
the Lake Erie winter water column (2007-2010). Pictures were taken across the western-central
basin by Steven W. Wilhelm.

9
forming phyla which co-exist in a successional cycle: winter-spring diatom blooms are followed

by summer cyanobacterial blooms, with the activity of one bloom setting the stage for the next

(Wilhelm et al., 2020). Succession from one alga to another is a widely observed paradigm in

itself, being reported in eutrophic freshwater systems such as Lake Tai (Taihu, China), Lake

Victoria (Kenya), and the English Windermere South Basin amongst others (Canale and Vogel,

1974; Talling, 1976; Galat and Verdin, 1989; Krivtsov et al., 2000; Ke et al., 2008; Goldenberg

and Lehman, 2012; Sitoki et al., 2012). Such studies have focused on one phylum in isolation,

with relatively few studies considering both algal blooms in tandem. This is an important gap, as

summer cyanobacterial and winter-spring diatom blooms are intrinsically linked (Leflaive and

Ten‐Hage, 2007; Niu et al., 2011; Reavie et al., 2016). Further, many studies (and models)

attribute seasonal algal succession as a preference of phyla for temperature optima, relying on

the classical paradigm that diatoms “like it cold” and cyanobacteria “like it hot” (Reynolds,

1997; 2006; Paerl and Huisman, 2008; Berry et al., 2017; Shatwell and Köhler, 2019). Yet,

historical summer diatom blooms (Hartig, 1987) and recently addressed winter-spring

cyanobacterial blooms (Reinl et al., 2023) suggest this successional trend may not be so black

and white. Indeed, while there exists ample literature concerning the prevalence and causes of

algal blooms in Lake Erie (Munawar and Munawar, 1976; Munawar and Heath, 2008; Allinger

and Reavie, 2013b; Reavie et al., 2014; Steffen et al., 2014a; Bullerjahn et al., 2016; O'Donnell

et al., 2023), there remains the need to re-visit the ecological constraints which dictate algal

bloom success and succession in the face of new and shifting paradigms. Here, we refine two

major paradigms (temperature and nutrients) which have been shown to constrain algal bloom

succession in Lake Erie. We further propose a novel paradigm which serves to constrain algal

10
bloom succession: pH. Cumulatively, this work serves as a cautionary tale that Lake Erie algal

blooms cannot be studied solitarily, and traditional paradigms cannot be applied sweepingly.

Refining the temperature paradigm of algal bloom succession

There exists a widely recognized temperature paradigm: cyanobacteria are adept to

warm, summer temperatures and diatoms are restricted to cold, winter temperatures. Indeed, in

the literature cyanobacteria “like it hot” (Paerl and Huisman, 2008). Prior studies demonstrated

cyanobacteria achieve higher growth rates at higher temperatures (Robarts and Zohary, 1987;

Reynolds, 2006; Joehnk et al., 2008; Lürling et al., 2013) and possess gas vesicles which allow

them to benefit from temperature-induced stratification of the water column (Huisman et al.,

2005; Paerl et al., 2006; Reynolds, 2006; Wagner and Adrian, 2009). Indeed Microcystis spp.

peak abundances coincide with high temperatures in Lake Erie (Davis et al., 2009; Rinta-Kanto

et al., 2009; Zepernick et al., 2021). While it has been previously suggested many cyanobacteria

(such as Microcystis) “disappear” from the planktonic water column at temperatures <10°C

(Reavie et al., 2016; Visser et al., 2016; Cao et al., 2022; Ming et al., 2022), emerging evidence

suggests there are exceptions to this paradigm. For example Reinl et al. (2023) recently cited 37

observations of freshwater cyanobacterial “blooms” at temperatures <15° C across multiple

freshwater systems. These reports suggest a need to revisit the school of thought that winter

cyanobacteria are simply vegetative “overwintering” cells which are cryotolerant rather than

cryophilic (Bridgeman and Penamon, 2010; Kutovaya et al., 2012; Cirés et al., 2013; Kitchens et

al., 2018). In contrast, Reinl et al. (2023) found many common summer cyanobacterial bloom

formers such as Dolichospermum, Aphanizomenon, Microcystis, Planktothrix, etc. can also grow

in colder waters (Mankiewicz‐Boczek et al., 2011; Babanazarova et al., 2013; Bižić-Ionescu et

al., 2014; Persaud et al., 2015; Ma et al., 2016; Wejnerowski et al., 2018). Yet, while cold

11
tolerance in cyanobacteria is not by any means a novel concept (Dietlicher, 1974; Tang et al.,

1997; Los and Murata, 1999; Vincent, 2007) it is seldomly investigated in the environment.

Further, oftentimes in vitro studies contradict field observations. For example, in vitro assays

demonstrated Microcystis elicits a cold stress response to temperature drops from 26° C to 18-

19° C (Peng et al., 2018; Martin et al., 2020), which seemingly contradicts the report of a winter

Microcystis bloom in Lake Tai (Ma et al., 2016). Further, discrepancies exist amongst in vitro

studies investigating cold stress in Microcystis, with Ming et al. (2022) reporting cold

temperatures of 4° C and 10° C permitted cell growth in Microcystis cultures in contrast to prior

reports of the opposite (Peng et al., 2018; Martin et al., 2020). These incongruencies suggest a

need to further elucidate cold temperature physiology of cyanobacterial bloom formers within

controlled in vitro laboratory settings and the in situ environment. They also stress there may be

strong degrees of variability between different strains (a phenomenon noted in the recent

examination of toxin production by Microcystis (Hellweger et al., 2022). Surprisingly, despite

Lake Erie’s notoriety for cyanobacterial blooms, Reinl et al. (2023) only reported one

documented instance of a cold cyanobacterial bloom within the Lake Erie water column. This

study reported Synechococcus spp. abundances of 106 - 108 cells • L-1 during a 2009-2012 Lake

Erie winter survey, with high abundances associated with ice-cover (Twiss et al., 2012). Yet

data suggest this is not as much a bloom but instead a typical abundance for the often overlooked

picoplankton (Wilhelm et al., 2006b). To this end, it may prove important to define “blooms”

based on the species under discussion.

Building upon the observed winter abundances of Synechococcus in the winter Lake Erie

water column, we observed what are considered high abundances of Dolichospermum (~1,800

cells • L-1) during a 2020 spring survey in Lake Erie where water temperatures were ~10° C

12
(Bullerjahn et al., 2022; Zepernick et al., 2022a). Additionally Mckay et al. (2018) reported

Planktothrix agardhii abundances > 1 x 107 • L-1 during early May of 2016 within the Maumee

River, when water temperatures were < 15° C. Yet, beyond these studies there has been a lack of

reports concerning winter/cold cyanobacterial blooms in Lake Erie. We note the absence of

evidence cannot be interpreted as evidence of absence, as this may very well stem from a lack of

winter surveys (Ozersky et al., 2021). Hence, there is a need to investigate the potential for

psychrotolerant or psychrophilic cyanobacteria within the winter-spring Lake Erie water column

and characterize cold temperature physiology in cyanobacterial bloom-forming taxa. Indeed, it

may be argued these shoulder seasons could be strongly affected by projected future climates,

creating a disconnect between the important biological queues of light and temperature.

There also exists the need to re-assess the paradigm that diatom blooms are constrained to

cold temperatures. Shatwell et al. (2008) embodied this paradigm by stating “It is well known

that cyanobacteria prefer warmer temperatures than diatoms”. While temperature does constrain

particular Bacillariophyta (e.g., psychrophilic bloom-forming diatoms Stephanodiscus hantzschii

and Aulacoseira islandica are confined to cold temperatures (Jung et al., 2009; D'souza, 2012;

Saxton et al., 2012b)), there exist prominent exceptions to this paradigm which must be

considered. For example, Hartig (1987) reported prolific Fragilaria crotonensis blooms

throughout the Lake Erie summer basin. These blooms persisted throughout the latter part of the

twentieth century (Beeton, 1965; Munawar and Munawar, 1976; Gladish and Munawar, 1980),

and this genus remains a prominent member of the summer Lake Erie water column today

(Munawar et al., 2008; Saxton et al., 2012b). Indeed, Hartig and Wallen (1986) found in vitro F.

crotonensis cultures reached maximum growth rates at 17-23° C compared 5-11° C, and recent

in vitro studies demonstrated F. crotonensis exhibits high growth rates at 26° C (Zepernick et al.,

13
2021; Zepernick et al., 2022b). Returning to in situ ecological observations, large populations of

diatoms have been reported in the Nyanza Gulf and Rusinga Channel throughout the warm, dry

season in Lake Victoria, Kenya which is situated directly on the Equator (Sitoki et al., 2012;

Sitoki et al., 2013). Diatoms were previously reported to dominate the summer water column of

Lake Constance, Germany, which reaches summer temperatures comparable to Lake Erie

(Sommer and Stabel, 1983). In addition, Mancuso et al. (2021) reported diatoms were the

dominant planktonic taxa throughout April-October in the Muskegon Estuary (Michigan), with

summer temperatures reported at an average of ~23° C, albeit these temperatures were cooler

than prior summers. Further, Stoermer (1993) reported abundances of diatoms such as Cyclotella

meneghiniana and Cyclotella cryptica within the recent Lake Erie paleolimnological record,

citing these diatoms have exceedingly high temperature optima conducive to the summer Lake

Erie water column. Cumulatively, there is ample evidence to suggest diatoms have the potential

to bloom in the warm, summer water column. Yet, Lake Erie diatoms broadly remain widely

uncharacterized and unsequenced to date (Edgar et al., 2016; Zepernick et al., 2022c) ,with novel

diatom species still being discovered within Lake Erie as of this year (Reavie, 2023). Thus,

further research is also required regarding the potential presence of summer diatom blooms and

their physiological response to warm temperatures. Broadly, this synthesis indicates the role of

temperature in algal bloom success and succession requires revisiting.

Refining nutrient paradigms of algal bloom succession

There exists the widely accepted paradigm: cyanobacteria are a symptom of

eutrophication and diatoms are an indicator of meso-oligotrophic systems. Indeed,

anthropogenic nutrient loading is considered the primary driver of cyanobacterial blooms. The

emergence of Microcystis spp. summer blooms has been attributed to the re-eutrophication of

14
Lake Erie (Scavia et al., 2014). Yet, the role of nitrogen and phosphorus in bloom constraint

remains a topic of debate (Paerl et al., 2016; Schindler et al., 2016; Hellweger et al., 2022;

Huisman et al., 2022; Stow et al., 2022; Wilhelm et al., 2022). Absent from the general

discussion is the process being limited: in marine sciences “limitation” generally refers to the

accumulation of biomass by phototrophs. In the present case, it appears that in Lake Erie (and

other fresh waters) researchers have conflated ideas associated with both “how much” is there

(biomass) and “who” is there (species composition), even though evidence suggested 20 years

ago that biomass and diversity of phototrophs need to be decoupled with respect to nutrient

effects (Wilhelm et al., 2003).

For example, research has indicated shifts from ammonium and nitrate to urea, which

have occurred on a broad scale in North America (Paerl et al., 2016), generally promote

cyanobacterial blooms (Chaffin and Bridgeman, 2014; Glibert et al., 2016; Krausfeldt et al.,

2019). This phenomenon has been suggested to result from a superior ability of certain

cyanobacteria to assimilate urea (relative to diatoms) as well as an ability to metabolize urea as a

carbon source during periods of elevated pH (Steffen et al., 2017; Krausfeldt et al., 2019).

Nonetheless, studies assert cyanobacterial blooms are constrained to eutrophic systems and

exacerbated by nutrient loadings (Almanza et al., 2019). However, in the past few years this

paradigm has also been called into question. The emergence of cyanobacterial blooms within

oligotrophic Lake Superior was most recently suggested to “shift the high-nutrient paradigm”

(Reinl et al., 2021). These blooms, comprised of Dolichospermum, which is typically associated

with eutrophic lake conditions (Bruun, 2012), were hypothesized to be driven by temperature

and precipitation rather than nutrient loading (Sterner et al., 2020). Indeed, cyanobacterial

blooms have been reported in other oligotrophic freshwater systems across the Northeast U.S.

15
(Carey et al., 2012) and Ontario, Canada (Winter et al., 2011). Cumulatively this suggests the

“high nutrient cyanobacterial paradigm” requires revisiting, and additional studies are required

regarding cyanobacterial prominence and physiology in oligotrophic conditions.

In turn, there is a need to re-assess the notion that diatoms are at a disadvantage in

eutrophic conditions. Throughout the decades there are various reports of diatom blooms in

eutrophic freshwaters such as Lake Erie, Lake Michigan, Lake Victoria, and others (Schelske,

1975; Hartig, 1987; Middelboe et al., 1995; Sitoki et al., 2012). Most recently, Asterionella

formosa blooms were reported in eutrophic Lake Taihu (Liu et al., 2022). In addition, numerous

diatom genera serve as eutrophic ecological indicators (Kitner and Poulícková, 2003; Bellinger

et al., 2006; Vilmi et al., 2015). Indeed, Stoermer (1993) indicated diatoms such as Aulacosiera

islandica, Fragilaria, and Stephanodiscus serve as eutrophication indicators in the Lake Erie

paleolimnological record. Further, increases in Fragilaria crotonensis and Asterionella formosa

have been considered markers for when nitrogen levels have surpassed a trophic threshold in

oligotrophic lakes throughout the western U.S. (Saros et al., 2005; Wolfe et al., 2006; Spaulding

et al., 2015). More recently, a regime shift in the Lake Victoria diatom community was reported,

with eutrophic conditions spurring a shift from endemic Aulacosiera species to eutrophic

Nitzschia (Sitoki et al., 2013; Simiyu and Kurmayer, 2022). Hence, diatom prominence and

physiology within eutrophic systems merits further attention within the “nutrient paradigm”. In

general, there exists a need to refine the role of nutrients and trophic status in algal bloom

success and succession.

pH a novel paradigm of algal bloom succession?

Recent work has demonstrated diatoms are successful during periods of circumneutral pH

while cyanobacteria proliferate at higher pH conditions. It is widely accepted pH constrains

16
phytoplankton phylogeny and physiology within the global oceans (Lomas et al., 2012; Collins

et al., 2014; Das and Mangwani, 2015; Gao et al., 2019), with these works ignited by the

revelation of ocean acidification (Doney et al., 2009). In contrast, there is limited literature

regarding the role of pH in freshwater phytoplankton ecophysiology. This is a limnetic

knowledge gap of significant importance in the face of present and future climatic changes.

Looking to future climates, the Great Lakes were found to be experiencing increases in pCO2

(and thus declines in pH) concomitantly with the global oceans (Phillips et al., 2015) due to

atmospheric CO2 accumulation with projected declines of 0.3 - 0.5 pH units by 2100. The

acidification of Lake Erie has profound implications for cyanobacterial blooms: research

indicates cyanobacteria are not favored by acidic conditions and prefer slightly alkaline pHs >

7.7 (Wicks and Thiel, 1990). In contrast, acidifying surface waters may benefit diatoms (Guillard

and Lorenzen, 1972; Arzet et al., 1986; Hervé et al., 2012), as a body of marine literature

suggests acidic conditions benefit diatoms (Wu et al., 2014; Bach and Taucher, 2019). Overall,

there remains a lack of information regarding how Lake Erie algal blooms will respond to a low

level, decadal increase in the acid load to the water column and this must be considered with

respect to the ecological success and successional patterns of algal blooms in Lake Erie.

In contrast to atmospheric driven “lake acidification” is the biologically driven "lake

basification” (Zepernick et al., 2021; Zepernick et al., 2022b; Zepernick et al., 2023). Microcystis

blooms increase the pH of the water column as CO2 is rapidly depleted by photosynthesis

(Verspagen et al., 2014; Bullerjahn et al., 2016; Ji et al., 2020). As a result, water column pH

levels can range from 9-11 during peak cyanobacterial bloom periods (Krausfeldt et al., 2019).

For example, during a 2015 Microcystis bloom in Lake Erie the mean water column pH

remained at ~9.2 for a month (Zepernick et al., 2021). Further, Microcystis blooms have been

17
documented to drive water column pH to ~10 in Lake Tai (Taihu, China) (Van Dam et al., 2018)

and Lake Kennermermeer (The Netherlands) (Sandrini et al., 2016). During these basification

events, the pH fluctuates on a diel cycle by as much as 0.5 pH units, with the highest pH levels

coinciding with peak photosynthetic periods of the late afternoon as observed during a 2015

Lake Erie Microcystis bloom (Krausfeldt et al., 2019). Beyond episodic pH spikes, extended

basification events have been observed in systems such as Lake Santa Olalla (Spain), which had

a mean pH of 9.5 for 2 consecutive years due to persistent cyanobacterial blooms (Lopez-

Archilla et al., 2004). Overall, this is a trend within cyanobacterial blooms which can be traced in

the literature for decades (Talling, 1976; Booker and Walsby, 1981; Klemer et al., 1982; Paerl

and Ustach, 1982). Yet, despite the thorough documentation of this phenomenon, few studies

have directly assessed how cyanobacterial bloom-induced basification effects freshwater

phytoplankton physiology.

In contrast to evidence of cyanobacterial success at high pH levels, diatoms are at a

decided disadvantage. Marine diatoms are not found within the water column at pHs > 8.7,

directly affecting species succession in marine environments (Hansen, 2002). Recent

observations suggest this paradigm also applies to fresh waters. Alkaline pH conditions decrease

growth and silica deposition within the model diatom F. crotonensis (in vitro) and environmental

Lake Erie diatom communities (in situ). Cumulatively, observations suggest diatoms are

“unsuccessful” during cyanobacterial-bloom-induced pH levels (Zepernick et al., 2021;

Zepernick et al., 2022b; Zepernick et al., 2023). Further, these studies imply a pivotal role of pH

in Lake Erie algal bloom succession, as prolonged basification likely suppresses diatoms

throughout the summer and delays fall diatom succession (Wilhelm et al., 2020; Zepernick et al.,

2021; Zepernick et al., 2022b). The recent literature lends further support for the case of a “pH

18
paradigm”. A large-scale survey of 464 North American lakes demonstrated pH was the only

variable to significantly relate to cyanobacterial biomass: with temperature, trophic status, and

other factors falling short of influence (Bonilla et al., 2023). Further, Zhang (2023) demonstrated

cyanobacterial-induced pH levels caused calcium carbonate precipitation and silica dissolution in

a Florida subtropical lagoon. Burdick et al. (2020) reported cyanobacterial blooms in the Upper

Klamath Lake (Oregon) drove the water column pH to levels of 10, which likely invokes

sublethal stress on endangered Lost River and Short Nose suckers. Hence, there is a need to

determine how lake basification alters the physiology of those inducing these events

(cyanobacteria) and those which are also affected (diatoms and other alga). Cumulatively, the

role of pH in freshwater algal bloom success and succession requires additional attention.

Conclusion - Caveat Biologus

The paradigms discussed herein undoubtedly maintain their efficacy. Yet, traditional

concepts require revisiting and expansion in the face of emerging knowledge. The success of

cyanobacteria such as Microcystis spp. has been recognized as “a complicated and confusing

story” (Wilhelm et al., 2020). Undoubtedly, no one factor, or paradigm, is responsible for algal

bloom success or succession at all times or in all places. In turn, no single algal bloom occurs

solitarily, rather each bloom shapes the water column for the next. The observations accumulated

here provide a cautionary tale that paradigms are often too liberally applied to algal blooms,

which places future bloom projections and models at a significant fault. For the limnologist, a

return to many ecological principles (e.g., competitive exclusion theory (Hardin, 1960)) is ripe

for examination in the context of fresh waters. Often the devil is in the ecological details, and

this cautionary tale must be kept in mind as the field increasingly relies on models to predict

bloom magnitude and severity.

19
Dissertation overview
The research presented in this dissertation employs in vitro, in situ and in silico

approaches to investigate the drivers which constrain algal bloom success and succession in Lake

Erie (US/Canada). This work primarily focuses on the often unsung and understudied diatom

blooms of Lake Erie. Hence, the early Chapters of this dissertation entailed the establishment of

an in vitro model diatom system which was subsequently used to investigate ecologically

pertinent hypotheses. In turn, summer in situ assays and winter-spring water column samples

culminate into a comprehensive snapshot of the seasonal dynamics of the Lake Erie diatom

community. In total, the research delineated here assesses algal bloom success and succession

across spatial, temporal and climatic gradients.

Chapter II describes the establishment of an in vitro pH assay developed to investigate

Microcystis bloom-induced pH effects on freshwater diatoms. Early optimizations of the assay

revealed the pH of freshwater media varied widely from the initial pH levels it was titrated to

(Zepernick et al., 2020). Further investigations indicated flaming of the culture tube openings as

part of aseptic technique significantly reduced the pH of freshwater culture media via the

incorporation of CO2 (g). Indeed, the pH pf freshwater culture media dropped by 2.0 full pH units

after mimicking a 10 day growth study with daily subsampling events for cell enumeration.

Further, it was discovered aseptic flaming altered Microcystis growth dynamics as CO2 (g) serves

as a carbon source for the cyanobacteria, fueling growth and variability amongst biological

replicates. As a result of this work, optimizations were made to our freshwater media to ensure

the pH remained constant throughout pH growth assays and aseptic flaming was markedly

reduced during growth studies. In total, this work serves as a cautionary tale of the unintended

consequences of classical laboratory techniques.

20
 Following the establishment of the pH assay in Chapter II, in vitro pH assays were

conducted using the freshwater diatom F. crotonensis and Microcystis aeruginosa to deduce to

effects of pH on species success and succession in Chapter III. While the bloom induced pH of

9.2 significantly reduced F. crotonensis growth rates by ~50% compared to pH 7.7 controls, it

did not significantly affect Microcystis growth rates (Zepernick et al., 2021). Further

investigation revealed silica (Si) deposition (i.e., the process of diatom cell wall formation) was

significantly decreased within in vitro F. crotonensis cultures at pH 9.2; serving as one potential

mechanism behind high pH induced declines in growth rates. Subsequent in situ pH assays

performed within the Lake Erie water column indicated environmental communities (largely

dominated by diatoms F. crotonensis and Asterionella formosa) were significantly decreased at

the pH of 9.2 compared the controls. Cumulatively, this Chapter demonstrated the Microcystis

bloom-induced pH of 9.2 decreased growth and Si deposition within the in vitro model diatom F.

crotonensis and in situ Lake Erie diatom communities. In particular, this work suggested

Microcystis induced pH levels likely suppress summer diatom populations and further delay

diatom fall succession of the summer cyanobacterial blooms.

While Chapter III served as a preliminary investigation into the effects of pH on

freshwater algal bloom taxa, it fell short of describing the intracellular mechanism behind the

observed declines in diatom growth and Si deposition. To address this gap, the genome of F.

crotonensis was sequenced (Zepernick et al., 2022b) to facilitate transcriptomic studies with the

aim of generating hypotheses. Following, Chapter IV delineated results of an in vitro pH

transcriptomic study of the model diatom F. crotonensis. Transcriptome-derived hypotheses

were informed with in vitro physiological assays which revealed three central findings: 1) high

pH results in smaller, rougher and “browner” F. crotonensis filaments, 2) high pH decreases

21
photosynthetic processes and increases the expression of cell cycle arrest genes in F. crotonensis

filaments, and 3) high pH increases the expression of F. crotonensis transposable elements

resulting in the coining of the “Genomic Roulette” hypothesis (Zepernick et al., 2022d).

In contrast to Chapters II-IV which focused on summer diatom success and succession

dynamics, Chapter V pivots to the winter-spring Lake Erie diatom community which remains

widely unstudied and uncharacterized to date. In efforts to address this knowledge gap, samples

were throughout winter and spring 2019-202 to investigate how Lake Erie winter diatom

communities respond to ice cover vs. ice-free conditions. To our knowledge, this was the first

large-scale bioinformatic assessment of the winter Lake Erie community (Zepernick et al.

2022a). Transcriptomic analyses and physiochemical data indicated the winter diatom

community of 2019-2020 was markedly different from prior reports of 2007-2012 (Zepernick et

al., 2023 in preparation). Yet, the winter water column remained dominated by diatoms,

particularly the centric filamentous diatoms A. islandica and Stephanodiscus. Despite this

continued dominance, ice-free conditions reduced diatom abundance and increased heterogeneity

within the diatom community. Notably, diatoms of the class Mediophyceae (polar centric)

significantly increased the expression of fasciclin and photosynthesis genes within the ice-free

water column, leading to the “fasciclin rafting hypothesis” which suggests diatoms may raft

together into colonies to optimize their location in the ice-free turbid water column of Lake Erie.

While this Chapter serves as the first large-scale bioinformatic analysis of the Lake Erie winter

community, it illuminated a need for comprehensive fresh water taxonomic annotation tools and

physiological assessments of the winter community.

Cumulatively, the work described above advances our knowledge of Lake Erie algal

blooms and the drivers which constrain their success and succession. Further, this work

22
emphasizes the importance of diatoms within the Lake Erie water column while serving as a

foundational basis for the establishment of a novel algal bloom successional paradigm (pH) and

the emergence of a new field (lake basification).

23
References:
Allinger, L.E., and Reavie, E.D. (2013). The ecological history of Lake Erie as recorded by the
phytoplankton community. Journal of Great Lakes Research 39, 365-382.
Almanza, V., Pedreros, P., Laughinghouse Iv, H.D., Félez, J., Parra, O., Azócar, M., and Urrutia,
R. (2019). Association between trophic state, watershed use, and blooms of cyanobacteria
in south-central Chile. Limnologica 75, 30-41.
Anderson, D.M., Glibert, P.M., and Burkholder, J.M. (2002). Harmful algal blooms and
eutrophication: nutrient sources, composition, and consequences. Estuaries 25, 704-726.
Arzet, K., Steinberg, C., Psenner, R., and Schulz, N. (1986). Diatom distribution and diatom
inferred pH in the sediment of four alpine lakes. Hydrobiologia 143, 247-254.
Ashworth, W. (1987). The late, Great Lakes: an environmental history. Wayne State University
Press.
Babanazarova, O., Sidelev, S., and Schischeleva, S. (2013). The structure of winter
phytoplankton in Lake Nero, Russia, a hypertrophic lake dominated by Planktothrix-like
Cyanobacteria. Aquatic Biosystems 9, 1-11.
Bach, L.T., and Taucher, J. (2019). CO 2 effects on diatoms: a synthesis of more than a decade
of ocean acidification experiments with natural communities. Ocean Science 15, 1159-
1175.
Beeton, A.M. (1965). EUTROPHICATION OF THE ST. LAWRENCE GREAT LAKES 1.
Limnology and Oceanography 10, 240-254.
Bellinger, B.J., Cocquyt, C., and O’reilly, C.M. (2006). Benthic diatoms as indicators of
eutrophication in tropical streams. Hydrobiologia 573, 75-87.
Berry, M.A., Davis, T.W., Cory, R.M., Duhaime, M.B., Johengen, T.H., Kling, G.W., Marino,
J.A., Den Uyl, P.A., Gossiaux, D., and Dick, G.J. (2017). Cyanobacterial harmful algal
blooms are a biological disturbance to Western Lake Erie bacterial communities.
Environmental microbiology 19, 1149-1162.
Bižić-Ionescu, M., Amann, R., and Grossart, H.-P. (2014). Massive regime shifts and high
activity of heterotrophic bacteria in an ice-covered lake. PloS one 9, e113611.
Bonilla, S., Aguilera, A., Aubriot, L., Huszar, V., Almanza, V., Haakonsson, S., Izaguirre, I.,
O'farrell, I., Salazar, A., and Becker, V. (2023). Nutrients and not temperature are the key
drivers for cyanobacterial biomass in the Americas. Harmful Algae 121, 102367.
Booker, M., and Walsby, A. (1981). Bloom formation and stratification by a planktonic blue-
green alga in an experimental water column. British phycological journal 16, 411-421.
Bridgeman, T.B., Chaffin, J.D., and Filbrun, J.E. (2013). A novel method for tracking western
Lake Erie Microcystis blooms, 2002–2011. Journal of Great Lakes Research 39, 83-89.
Bridgeman, T.B., and Penamon, W.A. (2010). Lyngbya wollei in western Lake Erie. Journal of
Great Lakes Research 36, 167-171.
Britt, N.W. (1955). Stratification in western Lake Erie in summer of 1953: effects on the
Hexagenia (Ephemeroptera) population. Ecology 36, 239-244.
Brittain, S.M., Wang, J., Babcock-Jackson, L., Carmichael, W.W., Rinehart, K.L., and Culver,
D.A. (2000). Isolation and characterization of microcystins, cyclic heptapeptide
hepatotoxins from a Lake Erie strain of Microcystis aeruginosa. Journal of Great Lakes
Research 26, 241-249.
Bruun, K. (2012). Algae can function as indicators of water pollution. Nostoca Algae
Laboratory, Washington State Lake Protection Association. Available online: www.
nostoca. com.

24
Bullerjahn, G.S., Anderson, J.T., and Mckay, R.M. (2022). "Winter survey data from Lake Erie
from 2018-2020". 3 ed. (Biological and Chemical Oceanography Data Management
Office (BCO-DMO)).
Bullerjahn, G.S., Mckay, R.M., Davis, T.W., Baker, D.B., Boyer, G.L., D’anglada, L.V.,
Doucette, G.J., Ho, J.C., Irwin, E.G., and Kling, C.L. (2016). Global solutions to regional
problems: Collecting global expertise to address the problem of harmful cyanobacterial
blooms. A Lake Erie case study. Harmful Algae 54, 223-238.
Burdick, S.M., Hewitt, D.A., Martin, B.A., Schenk, L., and Rounds, S.A. (2020). Effects of
harmful algal blooms and associated water-quality on endangered Lost River and
shortnose suckers. Harmful Algae 97, 101847.
Buttle, J., Muir, T., and Frain, J. (2004). Economic impacts of climate change on the Canadian
Great Lakes hydroelectric power producers: a supply analysis. Canadian Water
Resources Journal/Revue canadienne des ressources hydriques 29, 89-110.
Canada, G.O., and Agency, U.E.P. (1995). "The Great Lakes: An environmental atlas and
resource book". Great Lakes National Program Office Chicago, Illinois).
Canale, R.P., and Vogel, A.H. (1974). Effects of temperature on phytoplankton growth. Journal
of the Environmental Engineering Division 100, 231-241.
Cao, H., Han, L., and Li, L. (2022). A deep learning method for cyanobacterial harmful algae
blooms prediction in Taihu Lake, China. Harmful Algae 113, 102189.
Carey, C.C., Ewing, H.A., Cottingham, K.L., Weathers, K.C., Thomas, R.Q., and Haney, J.F.
(2012). Occurrence and toxicity of the cyanobacterium Gloeotrichia echinulata in low-
nutrient lakes in the northeastern United States. Aquatic Ecology 46, 395-409.
Carrick, H.J., and Schelske, C.L. (1997). Have we overlooked the importance of small
phytoplankton in productive waters? Limnology and Oceanography 42, 1613-1621.
Casper, V.L. (1965). A phytoplankton bloom in western Lake Erie. University of Michigan Great
Lakes Res Div 13, 29-35.
Chaffin, J.D., and Bridgeman, T.B. (2014). Organic and inorganic nitrogen utilization by
nitrogen-stressed cyanobacteria during bloom conditions. Journal of applied phycology
26, 299-309.
Chaffin, J.D., Mishra, S., Kane, D.D., Bade, D.L., Stanislawczyk, K., Slodysko, K.N., Jones,
K.W., Parker, E.M., and Fox, E.L. (2019). Cyanobacterial blooms in the central basin of
Lake Erie: Potentials for cyanotoxins and environmental drivers. Journal of Great Lakes
Research 45, 277-289.
Chandler, D.C. (1940). Limnological studies of western Lake Erie: I. Plankton and certain
physical-chemical data of the Bass Islands region, from Septbember 1938, to November,
1939. Ohio Journal of Science 40, 291-336.
Chandler, D.C. (1942). Limnological Studies of Western Lake Erie. III, Phytoplankton and
Physical-Chemical Data from November, 1939, to November 1940.
Chandler, D.C. (1944). Limnological studies of western Lake Erie IV. Relation of limnological
and climatic factors to the phytoplankton of 1941. Transactions of the American
Microscopical Society 63, 203-236.
Chandler, D.C., and Weeks, O.B. (1945). Limnological studies of western Lake Erie: V. Relation
of limnological and meteorological conditions to the production of phytoplankton in
1942. Ecological Monographs 15, 435-457.

25
Cirés, S., Wörmer, L., Agha, R., and Quesada, A. (2013). Overwintering populations of
Anabaena, Aphanizomenon and Microcystis as potential inocula for summer blooms.
Journal of Plankton Research 35, 1254-1266.
Collins, S., Rost, B., and Rynearson, T.A. (2014). Evolutionary potential of marine
phytoplankton under ocean acidification. Evolutionary applications 7, 140-155.
Conroy, J.D., and Culver, D.A. (2005). Do dreissenid mussels affect Lake Erie ecosystem
stability processes? The American midland naturalist 153, 20-32.
D'souza, N.A. (2012). Psychrophilic diatoms in ice-covered Lake Erie. Bowling Green State
University.
Das, S., and Mangwani, N. (2015). Ocean acidification and marine microorganisms: responses
and consequences. Oceanologia 57, 349-361.
Davis, C.C. (1954). A Preliminary Study of the Plankton of the Cleveland Harbor Area, Ohio: II.
The Distribution and Quantity of the Phytoplankton. Ecological Monographs 24, 322-
347.
Davis, C.C. (1964). Evidence for the eutrophication of Lake Erie from phytoplankton records.
Limnology and Oceanography 9, 275-283.
Davis, T.W., Berry, D.L., Boyer, G.L., and Gobler, C.J. (2009). The effects of temperature and
nutrients on the growth and dynamics of toxic and non-toxic strains of Microcystis during
cyanobacteria blooms. Harmful Algae 8, 715-725.
Dengg, M., Stirling, C.H., Reid, M.R., Verburg, P., Armstrong, E., Kelly, L.T., and Wood, S.A.
(2022). Growth at the limits: comparing trace metal limitation of a freshwater
cyanobacterium (Dolichospermum lemmermannii) and a freshwater diatom (Fragilaria
crotonensis). Scientific Reports 12, 467.
Dietlicher, K. (1974). The Water Quality of the Lakes of Zurich and" Walensee": An Essay of
Description. Zurich Waterworks.
Doney, S.C., Fabry, V.J., Feely, R.A., and Kleypas, J.A. (2009). Ocean acidification: the other
CO2 problem. Annual Review of Marine Science 1, 169-192.
Edgar, R., Morris, P., Rozmarynowycz, M., D'souza, N., Moniruzzaman, M., Bourbonniere, R.,
Bullerjahn, G., Phuntumart, V., Wilhelm, S., and Mckay, R. (2016). Adaptations to
photoautotrophy associated with seasonal ice cover in a large lake revealed by
metatranscriptome analysis of a winter diatom bloom. Journal of Great Lakes Research
42, 1007-1015.
Fahnenstiel, G.L., and Carrick, H.J. (1992). Phototrophic picoplankton in Lakes Huron and
Michigan: Abundance, distribution, composition, and contribution to biomass and
production. Canadian Journal of Fisheries and Aquatic Sciences 49, 379-388.
Field, C.B., Behrenfeld, M.J., Randerson, J.T., and Falkowski, P. (1998). Primary production of
the biosphere: integrating terrestrial and oceanic components. Science 281, 237-240.
Galat, D.L., and Verdin, J.P. (1989). Patchiness, collapse and succession of a cyanobacterial
bloom evaluated by synoptic sampling and remote sensing. Journal of Plankton Research
11, 925-948.
Gao, K., Beardall, J., Häder, D.-P., Hall-Spencer, J.M., Gao, G., and Hutchins, D.A. (2019).
Effects of ocean acidification on marine photosynthetic organisms under the concurrent
influences of warming, UV radiation, and deoxygenation. Frontiers in Marine Science 6,
322.

26
Gladish, D., and Munawar, M. (1980). The phytoplankton biomass and species composition at
two stations in western Lake Erie, 1975/76. Internationale Revue der gesamten
Hydrobiologie und Hydrographie 65, 691-708.
Glibert, P.M., Wilkerson, F.P., Dugdale, R.C., Raven, J.A., Dupont, C.L., Leavitt, P.R., Parker,
A.E., Burkholder, J.M., and Kana, T.M. (2016). Pluses and minuses of ammonium and
nitrate uptake and assimilation by phytoplankton and implications for productivity and
community composition, with emphasis on nitrogen‐enriched conditions. Limnology and
Oceanography 61, 165-197.
Goldenberg, S.Z., and Lehman, J.T. (2012). Diatom response to the whole lake manipulation of a
eutrophic urban impoundment. Hydrobiologia 691, 71-80.
Gronewold, A.D., Fortin, V., Lofgren, B., Clites, A., Stow, C.A., and Quinn, F. (2013). Coasts,
water levels, and climate change: A Great Lakes perspective. Climatic Change 120, 697-
711.
Guillard, R.R., and Lorenzen, C.J. (1972). Yellow-green algae with chlorophyllide C 1,2 Journal
of Phycology 8, 10-14.
Hansen, P.J. (2002). Effect of high pH on the growth and survival of marine phytoplankton:
implications for species succession. Aquatic Microbial Ecology 28, 279-288.
Hardin, G. (1960). The competitive exclusion principle. Science 131, 1292-1298.
Hartig, J.H. (1987). Factors contributing to development of Fragilaria crontonensis Kitton
Pulses in Pigeon Bay waters of western Lake Erie. Journal of Great Lakes Research 13,
65-77.
Hartig, J.H., and Wallen, D.G. (1986). The influence of light and temperature on growth and
photosynthesis of Fragilaria crotonensis Kitton. Journal of Freshwater Ecology 3, 371-
382.
Hellweger, F.L., Martin, R.M., Eigemann, F., Smith, D.J., Dick, G.J., and Wilhelm, S.W. (2022).
Models predict planned phosphorus load reduction will make Lake Erie more toxic.
Science 376, 1001-1005.
Hervé, V., Derr, J., Douady, S., Quinet, M., Moisan, L., and Lopez, P.J. (2012). Multiparametric
analyses reveal the pH-dependence of silicon biomineralization in diatoms. PloS One 7,
e46722.
Hohn, M.H. (1969). Qualitative and Quantitative Analyses of Plankton Diatoms: Bass Island
Area, Lake Erie, 1938-1965, Including Synoptic Surveys of 1960-1963. Ohio State
University.
Huisman, J., Codd, G.A., Paerl, H.W., Ibelings, B.W., Verspagen, J.M., and Visser, P.M. (2018).
Cyanobacterial blooms. Nature Reviews Microbiology 16, 471-483.
Huisman, J., Dittmann, E., Fastner, J., Schuurmans, J.M., Scott, J.T., Van De Waal, D.B., Visser,
P.M., Welker, M., and Chorus, I. (2022). Comment on “Models predict planned
phosphorus load reduction will make Lake Erie more toxic”. Science 378, eadd9959.
Huisman, J., Matthijs, H., and Visser, P. (2005). Harmful Cyanobacteria Springer Aquatic
Ecology Series 3. Dordrecht, The Netheralands.
Ijc (1989). International Joint Commission (IJC) United States and Canada 1989. Great Lakes
Water Quality Agreement of 1978. http://www.ijc.org/agree/quality.html [Online].
[Accessed].
Ji, X., Verspagen, J.M., Van De Waal, D.B., Rost, B., and Huisman, J. (2020). Phenotypic
plasticity of carbon fixation stimulates cyanobacterial blooms at elevated CO2. Science
Advances 6.

27
Joehnk, K.D., Huisman, J., Sharples, J., Sommeijer, B., Visser, P.M., and Stroom, J.M. (2008).
Summer heatwaves promote blooms of harmful cyanobacteria. Global Change Biology
14, 495-512.
Jung, S.W., Kwon, O.Y., Lee, J.H., and Han, M.-S. (2009). Effects of water temperature and
silicate on the winter blooming diatom Stephanodiscus hantzschii (Bacillariophyceae)
growing in eutrophic conditions in the lower Han River, South Korea. Journal of
Freshwater Ecology 24, 219-226.
Ke, Z., Xie, P., and Guo, L. (2008). Controlling factors of spring–summer phytoplankton
succession in Lake Taihu (Meiliang Bay, China). Hydrobiologia 607, 41-49.
Kitchens, C.M., Johengen, T.H., and Davis, T.W. (2018). Establishing spatial and temporal
patterns in Microcystis sediment seed stock viability and their relationship to subsequent
bloom development in Western Lake Erie. PloS one 13, e0206821.
Kitner, M., and Poulícková, A. (2003). Littoral diatoms as indicators for the eutrophication of
shallow lakes. Hydrobiologia 506, 519-524.
Klemer, A., Feuillade, J., and Feuillade, M. (1982). Cyanobacterial blooms: carbon and nitrogen
limitation have opposite effects on the buoyancy of Oscillatoria. Science 215, 1629-1631.
Krausfeldt, L.E., Farmer, A.T., Castro Gonzalez, H., Zepernick, B.N., Campagna, S.R., and
Wilhelm, S.W. (2019). Urea is both a carbon and nitrogen source for Microcystis
aeruginosa: tracking 13C incorporation at bloom pH conditions. Frontiers in
Microbiology 10, 1064.
Krivtsov, V., Corliss, J., Bellinger, E., and Sigee, D. (2000). Indirect regulation rule for
consecutive stages of ecological succession. Ecological Modelling 133, 73-81.
Kutovaya, O.A., Mckay, R.M.L., Beall, B.F., Wilhelm, S.W., Kane, D.D., Chaffin, J.D.,
Bridgeman, T.B., and Bullerjahn, G.S. (2012). Evidence against fluvial seeding of
recurrent toxic blooms of Microcystis spp. in Lake Erie's western basin. Harmful Algae
15, 71-77.
Leflaive, J.P., and Ten‐Hage, L. (2007). Algal and cyanobacterial secondary metabolites in
freshwaters: a comparison of allelopathic compounds and toxins. Freshwater Biology 52,
199-214.
Liu, X., Li, Y., Shen, R., Jeppesen, E., Liu, Z., and Chen, F. (2022). A trophic cascade triggers
blooms of Asterionella formosa in subtropical eutrophic Lake Taihu, China. Freshwater
Biology 67, 1938-1948.
Lomas, M.W., Hopkinson, B.M., Ryan, J.L.D., Shi, D., Xu, Y., and Morel, F. (2012). Effect of
ocean acidification on cyanobacteria in the subtropical North Atlantic. Aquatic Microbial
Ecology 66, 211-222.
Lopez-Archilla, A.I., Moreira, D., López-García, P., and Guerrero, C. (2004). Phytoplankton
diversity and cyanobacterial dominance in a hypereutrophic shallow lake with
biologically produced alkaline pH. Extremophiles 8, 109-115.
Los, D.A., and Murata, N. (1999). Responses to cold shock in cyanobacteria. J. Mol. Microbiol.
Biotechnol 1, 221-230.
Lürling, M., Eshetu, F., Faassen, E.J., Kosten, S., and Huszar, V.L. (2013). Comparison of
cyanobacterial and green algal growth rates at different temperatures. Freshwater Biology
58, 552-559.
Ma, J., Qin, B., Paerl, H.W., Brookes, J.D., Hall, N.S., Shi, K., Zhou, Y., Guo, J., Li, Z., and Xu,
H. (2016). The persistence of cyanobacterial Microcystis spp. blooms throughout winter
in Lake Taihu, China. Limnology and Oceanography 61, 711-722.

28
Makarewicz, J.C., and Bertram, P. (1991). Evidence for the restoration of the Lake Erie
ecosystem. Bioscience 41, 216-223.
Makarewicz, J.C., Lewis, T.W., and Bertram, P. (1999). Phytoplankton composition and biomass
in the offshore waters of Lake Erie: pre-and post-Dreissena introduction (1983–1993).
Journal of Great Lakes Research 25, 135-148.
Mancuso, J.L., Weinke, A.D., Stone, I.P., Hamsher, S.E., Villar-Argaiz, M., and Biddanda, B.A.
(2021). Cold and wet: Diatoms dominate the phytoplankton community during a year of
anomalous weather in a Great Lakes estuary. Journal of Great Lakes Research 47, 1305-
1315.
Mankiewicz‐Boczek, J., Gągała, I., Kokociński, M., Jurczak, T., and Stefaniak, K. (2011).
Perennial toxigenic Planktothrix agardhii bloom in selected lakes of Western Poland.
Environmental toxicology 26, 10-20.
Martin, R.M., Moniruzzaman, M., Stark, G.F., Gann, E.R., Derminio, D.S., Wei, B., Hellweger,
F.L., Pinto, A., Boyer, G.L., and Wilhelm, S.W. (2020). Episodic decrease in temperature
increases mcy gene transcription and cellular microcystin in continuous cultures of
Microcystis aeruginosa PCC 7806. Frontiers in Microbiology 11, 601864.
Mckay, R.M.L., Tuttle, T., Reitz, L.A., Bullerjahn, G.S., Cody, W.R., Mcdowell, A.J., and
Davis, T.W. (2018). Early onset of a microcystin-producing cyanobacterial bloom in an
agriculturally-influenced Great Lakes tributary. Journal of Oceanology and Limnology
36, 1112-1125.
Mckindles, K., Frenken, T., Mckay, R.M.L., and Bullerjahn, G.S. (2020). Binational efforts
addressing cyanobacterial harmful algal blooms in the Great Lakes. Contaminants of the
Great Lakes, 109-133.
Mckindles, K.M., Mckay, R.M., and Bullerjahn, G.S. (2022). Genomic comparison of
Planktothrix agardhii isolates from a Lake Erie embayment. Plos one 17, e0273454.
Middelboe, M., Søndergaard, M., Letarte, Y., and Borch, N. (1995). Attached and free-living
bacteria: production and polymer hydrolysis during a diatom bloom. Microbial Ecology,
231-248.
Millerd, F. (2005). The economic impact of climate change on Canadian commercial navigation
on the Great Lake. Canadian Water Resources Journal 30, 269-280.
Ming, H., Yan, G., Zhang, X., Pei, X., Fu, L., and Zhou, D. (2022). Harsh temperature induces
Microcystis aeruginosa growth enhancement and water deterioration during
vernalization. Water Research 223, 118956.
Munawar, M., and Heath, R. (2008). Checking the pulse of Lake Erie.
Munawar, M., Munawar, I., Fitzpatrick, M., Niblock, H., Bowen, K., Lorimer, J., and Heath, R.
(2008). An intensive assessment of planktonic communities in the Canadian waters of
Lake Erie, 1998. Checking the pulse of Lake Erie. Aquatic Ecosystem Health and
Management Society, Burlington, Ontario, 297-346.
Munawar, M., and Munawar, I.F. (1976). A lakewide study of phytoplankton biomass and its
species composition in Lake Erie, April–December 1970. Journal of the Fisheries Board
of Canada 33, 581-600.
Munawar, M.U., and Munawar, I. (1982). Phycological studies in Lakes Ontario, Erie, Huron,
and Superior. Canadian Journal of Botany 60, 1837-1858.
Nicholls, K., Standen, D., Hopkins, G., and Carney, E. (1977). Declines in the near-shore
phytoplankton of Lake Erie's western basin since 1971. Journal of Great Lakes Research
3, 72-78.

29
Nicholls, K.H., and Hopkins, G.J. (1993). Recent changes in Lake Erie (north shore)
phytoplankton: cumulative impacts of phosphorus loading reductions and the zebra
mussel introduction. Journal of Great Lakes Research 19, 637-647.
Niu, Y., Shen, H., Chen, J., Xie, P., Yang, X., Tao, M., Ma, Z., and Qi, M. (2011).
Phytoplankton community succession shaping bacterioplankton community composition
in Lake Taihu, China. Water research 45, 4169-4182.
O'donnell, D.R., Briland, R., Budnik, R.R., Ludsin, S.A., and Hood, J.M. (2023). Trends in Lake
Erie phytoplankton biomass and community structure during a 20-year period of rapid
environmental change. Journal of Great Lakes Research.
Ozersky, T., Bramburger, A.J., Elgin, A.K., Vanderploeg, H.A., Wang, J., Austin, J.A., Carrick,
H.J., Chavarie, L., Depew, D.C., and Fisk, A.T. (2021). "The changing face of winter:
lessons and questions from the Laurentian Great Lakes". (Journal of Geophysical
Research: Biogeosciences: Wiley Online Library).
Paerl, H., Fulton, R., Graneli, E., and Turner, J. (2006). Ecology of harmful marine algae.
Ecology of harmful algae, 95-107.
Paerl, H.W., and Huisman, J. (2008). Blooms like it hot. Science 320, 57-58.
Paerl, H.W., Scott, J.T., Mccarthy, M.J., Newell, S.E., Gardner, W.S., Havens, K.E., Hoffman,
D.K., Wilhelm, S.W., and Wurtsbaugh, W.A. (2016). It Takes Two to Tango: When and
Where Dual Nutrient (N & P) Reductions Are Needed to Protect Lakes and Downstream
Ecosystems. Environmental Science & Technology 50, 10805-10813.
Paerl, H.W., and Ustach, J.F. (1982). Blue‐green algal scums: An explanation for their
occurrence during freshwater blooms 1. Limnology and Oceanography 27, 212-217.
Peng, G., Martin, R.M., Dearth, S.P., Sun, X., Boyer, G.L., Campagna, S.R., Lin, S., and
Wilhelm, S.W. (2018). Seasonally relevant cool temperatures interact with N chemistry
to increase microcystins produced in lab cultures of Microcystis aeruginosa NIES-843.
Environmental Science & Technology 52, 4127-4136.
Persaud, A.D., Paterson, A.M., Dillon, P.J., Winter, J.G., Palmer, M., and Somers, K.M. (2015).
Forecasting cyanobacteria dominance in Canadian temperate lakes. Journal of
environmental management 151, 343-352.
Phillips, J.C., Mckinley, G.A., Bennington, V., Bootsma, H.A., Pilcher, D.J., Sterner, R.W., and
Urban, N.R. (2015). The potential for CO₂-induced acidification in freshwater: A Great
Lakes case study. Oceanography 28, 136-145.
Reavie, E. (2023). Asymmetric, biraphid diatoms from the Laurentian Great Lakes. PeerJ
Aquatic Biology.
Reavie, E.D., Barbiero, R.P., Allinger, L.E., and Warren, G.J. (2014). Phytoplankton trends in
the Great Lakes, 2001–2011. Journal of Great Lakes Research 40, 618-639.
Reavie, E.D., Cai, M., Twiss, M.R., Carrick, H.J., Davis, T.W., Johengen, T.H., Gossiaux, D.,
Smith, D.E., Palladino, D., and Burtner, A. (2016). Winter–spring diatom production in
Lake Erie is an important driver of summer hypoxia. Journal of Great Lakes Research
42, 608-618.
Region, L.E., and Davis, C.C. (1958). An Approach to Some Problems of Secondary Production
in the Western. Limnology and Oceanography 3, 15-28.
Reinl, K.L., Brookes, J.D., Carey, C.C., Harris, T.D., Ibelings, B.W., Morales‐Williams, A.M.,
De Senerpont Domis, L.N., Atkins, K.S., Isles, P.D., and Mesman, J. (2021).
Cyanobacterial blooms in oligotrophic lakes: Shifting the high‐nutrient paradigm.
Freshwater Biology.

30
Reinl, K.L., Harris, T.D., North, R.L., Almela, P., Berger, S.A., Bizic, M., Burnet, S.H.,
Grossart, H.P., Ibelings, B.W., and Jakobsson, E. (2023). Blooms also like it cold.
Limnology and Oceanography Letters.
Reynolds, C.S. (1997). Vegetation processes in the pelagic: a model for ecosystem theory.
Reynolds, C.S. (2006). The Ecology of Phytoplankton. Cambridge University Press.
Rinta-Kanto, J.M., Konopko, E.A., Debruyn, J.M., Bourbonniere, R.A., Boyer, G.L., and
Wilhelm, S.W. (2009). Lake Erie Microcystis: relationship between microcystin
production, dynamics of genotypes and environmental parameters in a large lake.
Harmful algae 8, 665-673.
Rinta-Kanto, J.M., and Wilhelm, S.W. (2006). Diversity of microcystin-producing cyanobacteria
in spatially isolated regions of Lake Erie. Applied and Environmental Microbiology 72,
5083-5085.
Robarts, R.D., and Zohary, T. (1987). Temperature effects on photosynthetic capacity,
respiration, and growth rates of bloom‐forming cyanobacteria. New Zealand journal of
marine and freshwater research 21, 391-399.
Sandrini, G., Tann, R.P., Schuurmans, J.M., Van Beusekom, S.A., Matthijs, H.C., and Huisman,
J. (2016). Diel variation in gene expression of the CO2-concentrating mechanism during a
harmful cyanobacterial bloom. Frontiers in Microbiology 7, 551.
Saros, J.E., Michel, T.J., Interlandi, S.J., and Wolfe, A.P. (2005). Resource requirements of
Asterionella formosa and Fragilaria crotonensis in oligotrophic alpine lakes: implications
for recent phytoplankton community reorganizations. Canadian Journal of Fisheries and
Aquatic Sciences 62, 1681-1689.
Saxton, M.A., D'souza, N.A., Bourbonniere, R.A., Mckay, R.M.L., and Wilhelm, S.W. (2012).
Seasonal Si:C ratios in Lake Erie diatoms - Evidence of an active winter diatom
community. Journal of Great Lakes Research 38, 206-211.
Scavia, D., Allan, J.D., Arend, K.K., Bartell, S., Beletsky, D., Bosch, N.S., Brandt, S.B., Briland,
R.D., Daloğlu, I., and Depinto, J.V. (2014). Assessing and addressing the re-
eutrophication of Lake Erie: Central basin hypoxia. Journal of Great Lakes Research 40,
226-246.
Schelske, C.L. (1975). Silica and nitrate depletion as related to rate of eutrophication in Lakes
Michigan, Huron, and Superior. Coupling of land and water systems, 277-298.
Schindler, D.W., Carpenter, S.R., Chapra, S.C., Hecky, R.E., and Orihel, D.M. (2016). Reducing
phosphorus to curb lake eutrophication is a success. Environmental Science &
Technology 50, 8923-8929.
Shatwell, T., and Köhler, J. (2019). Decreased nitrogen loading controls summer cyanobacterial
blooms without promoting nitrogen‐fixing taxa: Long‐term response of a shallow lake.
Limnology and Oceanography 64, S166-S178.
Shatwell, T., Köhler, J., and Nicklisch, A. (2008). Warming promotes cold‐adapted
phytoplankton in temperate lakes and opens a loophole for Oscillatoriales in spring.
Global Change Biology 14, 2194-2200.
Simiyu, B.M., and Kurmayer, R. (2022). Response of planktonic diatoms to eutrophication in
Nyanza Gulf of Lake Victoria, Kenya. Limnologica 93, 125958.
Sitoki, L., Kofler, W., and Rott, E. (2013). Planktonic needle-shaped Nitzschia species from
Lake Victoria, Africa, revisited. Diatom research 28, 165-174.

31
Sitoki, L., Kurmayer, R., and Rott, E. (2012). Spatial variation of phytoplankton composition,
biovolume, and resulting microcystin concentrations in the Nyanza Gulf (Lake Victoria,
Kenya). Hydrobiologia 691, 109-122.
Sommer, U., and Stabel, H.-H. (1983). Silicon consumption and population density changes of
dominant planktonic diatoms in Lake Constance. The Journal of Ecology, 119-130.
Spaulding, S.A., Otu, M.K., Wolfe, A.P., and Baron, J.S. (2015). Paleolimnological records of
nitrogen deposition in shallow, high-elevation lakes of Grand Teton National Park,
Wyoming, USA. Arctic, Antarctic, and Alpine Research 47, 703-717.
Steffen, M.M., Belisle, B.S., Watson, S.B., Boyer, G.L., and Wilhelm, S.W. (2014). Status,
causes and controls of cyanobacterial blooms in Lake Erie. Journal of Great Lakes
Research 40, 215-225.
Steffen, M.M., Davis, T.W., Mckay, R.M.L., Bullerjahn, G.S., Krausfeldt, L.E., Stough, J.M.,
Neitzey, M.L., Gilbert, N.E., Boyer, G.L., and Johengen, T.H. (2017). Ecophysiological
examination of the Lake Erie Microcystis bloom in 2014: linkages between biology and
the water supply shutdown of Toledo, OH. Environmental Science & Technology 51,
6745-6755.
Sterner, R.W., Reinl, K.L., Lafrancois, B.M., Brovold, S., and Miller, T.R. (2020). A first
assessment of cyanobacterial blooms in oligotrophic Lake Superior. Limnology and
Oceanography 65, 2984-2998.
Stoermer, E., Kociolek, J., Schelske, C., and Conley, D. (1987). Quantitative analysis of
siliceous microfossils in the sediments of Lake Erie's central basin. Diatom Research 2,
113-134.
Stoermer, E.F. (1975). Phytoplankton composition and abundance in Lake Ontario during
IFYGL. National Environmental Research Center.
Stoermer, E.F. (1993). Evaluating diatom succession: some pecularities of the Great Lakes case.
Journal of Paleolimnology 8, 71-83.
Stoermer, E.F., Emmert, G., Julius, M.L., and Schelske, C.L. (1996). Paleolimnologic evidence
of rapid recent change in Lake Erie's trophic status. Canadian Journal of Fisheries and
Aquatic Sciences 53, 1451-1458.
Stoermer, E.F., Emmert, G., and Schelske, C.L. (1989). Morphological variation of
Stephanodiscus niagarae Ehrenb.(Bacillariophyta) in a Lake Ontario sediment core.
Journal of Paleolimnology 2, 227-236.
Stoermer, E.F., Wozin, J., and Schelske, C. (1993). Paleolimnological comparison of the
Laurentian Great Lakes based on diatoms. Limnology and oceanography 38, 1311-1316.
Stow, C.A., Stumpf, R.P., Rowe, M.D., Johnson, L.T., Carrick, H.J., and Yerubandi, R. (2022).
Model assumptions limit implications for nitrogen and phosphorus management. Journal
of Great Lakes Research 48, 1735-1737.
Sweeney, R.A. (1995). Rejuvenation of Lake Erie. GeoJournal 35, 65-66.
Talling, J. (1976). The depletion of carbon dioxide from lake water by phytoplankton. The
Journal of Ecology, 79-121.
Tang, E.P., Tremblay, R., and Vincent, W.F. (1997). CYANOBACTERIAL DOMINANCE OF
POLAR FRESHWATER ECOSYSTEMS: ARE HIGH‐LATITUDE MAT‐FORMERS
ADAPTED TO LOW TEMPERATURE? 1. Journal of Phycology 33, 171-181.
Twiss, M., Mckay, R., Bourbonniere, R., Bullerjahn, G., Carrick, H., Smith, R., Winter, J.,
D'souza, N., Furey, P., and Lashaway, A. (2012). Diatoms abound in ice-covered Lake

32
 Erie: An investigation of offshore winter limnology in Lake Erie over the period 2007 to
2010. Journal of Great Lakes Research 38, 18-30.
Van Dam, B.R., Tobias, C., Holbach, A., Paerl, H.W., and Zhu, G. (2018). CO2 limited
conditions favor cyanobacteria in a hypereutrophic lake: an empirical and theoretical
stable isotope study. Limnology and Oceanography 63, 1643-1659.
Verduin, J. (1964). Changes in western Lake Erie during the period 1948–1962. Verh Internat
Verein Limnol, 639-644.
Verspagen, J.M., Van De Waal, D.B., Finke, J.F., Visser, P.M., Van Donk, E., and Huisman, J.
(2014). Rising CO2 levels will intensify phytoplankton blooms in eutrophic and
hypertrophic lakes. PloS One 9.
Vilmi, A., Karjalainen, S.M., Landeiro, V.L., and Heino, J. (2015). Freshwater diatoms as
environmental indicators: evaluating the effects of eutrophication using species
morphology and biological indices. Environmental monitoring and assessment 187, 1-10.
Vincent, W.F. (2007). Cold tolerance in cyanobacteria and life in the cryosphere. Algae and
cyanobacteria in extreme environments, 287-301.
Visser, P.M., Verspagen, J.M., Sandrini, G., Stal, L.J., Matthijs, H.C., Davis, T.W., Paerl, H.W.,
and Huisman, J. (2016). How rising CO2 and global warming may stimulate harmful
cyanobacterial blooms. Harmful Algae 54, 145-159.
Wagner, C., and Adrian, R. (2009). Cyanobacteria dominance: quantifying the effects of climate
change. Limnology and Oceanography 54, 2460-2468.
Watson, S.B., Miller, C., Arhonditsis, G., Boyer, G.L., Carmichael, W., Charlton, M.N.,
Confesor, R., Depew, D.C., Höök, T.O., and Ludsin, S.A. (2016). The re-eutrophication
of Lake Erie: Harmful algal blooms and hypoxia. Harmful algae 56, 44-66.
Wejnerowski, Ł., Rzymski, P., Kokociński, M., and Meriluoto, J. (2018). The structure and
toxicity of winter cyanobacterial bloom in a eutrophic lake of the temperate zone.
Ecotoxicology 27, 752-760.
Wicks, R.J., and Thiel, P.G. (1990). Environmental factors affecting the production of peptide
toxins in floating scums of the cyanobacterium Microcystis aeruginosa in a hypertrophic
African reservoir. Environmental Science & Technology 24, 1413-1418.
Wilhelm, S., Bullerjahn, G., and Rlm, M. (2020). The complicated and confusing ecology of
Microcystis blooms. mBio 11, e00529-00520.
Wilhelm, S.W., Bullerjahn, G.S., Eldridge, M.L., Rinta-Kanto, J.M., Poorvin, L., and
Bourbonniere, R.A. (2006a). Seasonal hypoxia and the genetic diversity of prokaryote
populations in the central basin hypolimnion of Lake Erie: evidence for abundant
cyanobacteria and photosynthesis. Journal of Great Lakes Research 32, 657-671.
Wilhelm, S.W., Carberry, M.J., Eldridge, M.L., Poorvin, L., Saxton, M.A., and Doblin, M.A.
(2006b). Marine and freshwater cyanophages in a Laurentian Great Lake: evidence from
infectivity assays and molecular analyses of g20 genes. Applied and Environmental
Microbiology 72, 4957-4963.
Wilhelm, S.W., Debruyn, J.M., Gillor, O., Twiss, M.R., Livingston, K., Bourbonniere, R.A.,
Pickell, L.D., Trick, C.G., Dean, A.L., and Mckay, R.M.L. (2003). Effect of phosphorus
amendments on present day plankton communitites in pelagic Lake Erie. Aquatic
Microbial Ecology 32, 275-285.
Wilhelm, S.W., Hellweger, F.L., Martin, R.M., Schampera, C., Eigeman, F., Smith, D.J., and
Dick, G.J. (2022). Response to "Model assumptions limit implications for nitrogent and

33
 phosphorus management": the need to move beyond the phosphorus = biomass = toxin
doctrine. Journal of Great Lakes Research 48.
Wilhelm, S.W., Lecleir, G.R., Bullerjahn, G.S., Mckay, R.M., Saxton, M.A., Twiss, M.R., and
Bourbonniere, R.A. (2014). Seasonal changes in microbial community structure and
activity imply winter production is linked to summer hypoxia in a large lake. FEMS
microbiology ecology 87, 475-485.
Winter, J.G., Desellas, A.M., Fletcher, R., Heintsch, L., Morley, A., Nakamoto, L., and Utsumi,
K. (2011). Algal blooms in Ontario, Canada: increases in reports since 1994. Lake and
Reservoir Management 27, 107-114.
Wolfe, A.P., Cooke, C.A., and Hobbs, W.O. (2006). Are current rates of atmospheric nitrogen
deposition influencing lakes in the eastern Canadian Arctic? Arctic, Antarctic, and Alpine
Research 38, 465-476.
Wu, Y., Campbell, D.A., Irwin, A.J., Suggett, D.J., and Finkel, Z.V. (2014). Ocean acidification
enhances the growth rate of larger diatoms. Limnology and Oceanography 59, 1027-
1034.
Wynne, T.T., and Stumpf, R.P. (2015). Spatial and temporal patterns in the seasonal distribution
of toxic cyanobacteria in western Lake Erie from 2002–2014. Toxins 7, 1649-1663.
Yancey, C.E., Mathiesen, O., and Dick, G.J. (2023). Transcriptionally active nitrogen fixation
and biosynthesis of diverse secondary metabolites by Dolichospermum and
Aphanizomenom-like Cyanobacteria in western Lake Erie Microcystis blooms. Harmful
Algae, 102408.
Zepernick, B.N., Denison, E.R., Chaffin, J.D., Bullerjahn, G.S., Pennacchio, C.P., Frenken, T.,
Peck, D.H., Anderson, J.T., Niles, D., and Zastepa, A. (2022a). Metatranscriptomic
Sequencing of Winter and Spring Planktonic Communities from Lake Erie, a Laurentian
Great Lake. Microbiology Resource Announcements, e00351-00322.
Zepernick, B.N., Gann, E.R., Poiund, H.L., Martin, R.M., Krausfeldt, L.E., Chaffin, J.D., and
Wilhelm, S.W. (2021). Elevated pH conditions associated with Microcystis spp. blooms
decrease viability of the cultured diatom Fragilaria crotonensis and natural diatoms in
Lake Erie. Frontiers in Microbiology 12, 598736.
Zepernick, B.N., Niknejad, D.J., Stark, G.F., Truchon, A.R., Martin, R.M., Rossignol, K.L.,
Paerl, H.W., and Wilhelm, S.W. (2022b). Morphological, physiological, and
transcriptional responses of the freshwater diatom Fragilaria crotonensis to elevated pH
conditions. Frontiers in Microbiology 13.
Zepernick, B.N., Truchon, A.R., Gann, E.R., and Wilhelm, S.W. (2022c). Draft Genome
Sequence of the Freshwater Diatom Fragilaria crotonensis SAG 28.96. Microbiology
Resource Announcements, e00289-00222.
Zepernick, B.N., Wilhelm, S.W., Bullerjahn, G.S., and Paerl, H.W. (2023). Climate change and
the aquatic continuum: A cyanobacterial comeback story. Environmental Microbiology
Reports 15, 3-12.
Zhang, J.-Z. (2023). Cyanobacteria blooms induced precipitation of calcium carbonate and
dissolution of silica in a subtropical lagoon, Florida Bay, USA. Scientific Reports 13,
4071.

34
CHAPTER II: RE-ASSESSING CONSEQUENCES OF CLASSICAL LABORATORY
TECHNIQUES USING MICROCYSTIS AERUGINOSA AS A MODEL SYSTEM

35
Publication Note
This chapter is a version of a peer-reviewed, published article in Limnology and Oceanography:
Methods 18(5) by Brittany N. Zepernick, Lauren E. Krausfeldt, and Steven W. Wilhelm.

BNZ identified initial discrepancies in media pH, and performed all media preparation, flaming
assays, culture work, growth curves, and statistical analyses along with LEK. BNZ, LEK, and
SWW designed the experiments. TC analyses were performed by Adrian Gonzalez at the Water
Quality Core Facility at the University of Tennessee Knoxville. BNZ, LEK, and SWW
contributed to manuscript editing and writing.

36
Abstract
Aseptic technique has historically served as a fundamental practice in microbiology,

helping to maintain culture purity and integrity. This technique has been widely encouraged and

employed for use with cultures of heterotrophic bacteria as well as freshwater and marine algae.

Yet, recent observations have suggested these approaches may bring their own influences. We

observed variations in growth amongst replicate experimental cyanobacterial cultures upon

flaming of the culture tube opening during sample transfer and collection. Investigation revealed

the pH of culture media had decreased from the initial pH established during media preparation.

Flaming of sterile culture media alone confirmed a significant decrease, by as much as 1.7 pH

units, and correlated with increased flaming events over time. We hypothesized that the

causative factor was the introduction of carbon dioxide (CO2) into the media. To test this

hypothesis, qualitative and quantitative analyses were used to determine the primary driver of pH

decline. We further assessed the direct effects of flaming and subsequent pH changes on

Microcystis aeruginosa cultures, showing flame-driven pH changes and/or the introduction of

CO2 influenced experimental results. Our observations provide a cautionary tale of how classic

and well-accepted approaches may have unintended consequences, suggesting new approaches

may be necessary in research areas assessing pH or carbon-related effects on microbial

communities.

37
Introduction
Aseptic technique has served as a foundational method of microbiological research for

decades (Harrigan and McCance, 1966). Defined as steps to prevent contamination during

manipulation of microbial cultures or sterile culture media (Madigan, 2015), aseptic technique

commonly includes flame sterilization of the opening of culture vessels (Bykowski and

Stevenson, 2008). Flame sterilization involves passing the opening of a culture vessel through a

flame to prevent the introduction of microbial contaminants to a sample and is an approach used

in a variety of laboratories, hospitals, and industrial facilities. Yet, upon review of the literature,

we found a general lack of consensus within the scientific community as to the details of the

method. While flaming immediately before and after sub-sampling or culture transfer is specified

in a variety of publications (Andersen, 2005; Sanders, 2012; Madigan, 2015), the exact number

of times required for the mouth of the vessel to pass through a flame is generally not specified.

Additionally, while some argue flaming must be performed by passing the tube through the inner

blue section of the flame (Bykowski and Stevenson, 2008), others instruct the tube to be passed

above the flame (Coté, 1998). Further variances in flaming technique arise in the

recommendation of a wait time prior to re-capping the culture post-flaming (Coté, 1998), while

others recommend immediate recapping to minimize introduction of airborne contaminants

(Harrigan and McCance, 1966). In fact, the primary purpose of flaming in itself is often disputed

within the literature, as some sources claim it is to combust any contaminants located on the

mouth of the vessel (Coté, 1998), while others indicate it is to create an upwards convection

current to prevent atmospheric contaminants from entering the tube (Sanders, 2012; Madigan,

2015). These variances amongst techniques create a lapse in methodological consistency within

the scientific community. Additionally, the discrepancies may be exacerbated by confounding

factors that are unintentionally applied because of flaming. For example, the practice of flaming

38
in aseptic technique is mandated in the handling of freshwater media and phytoplankton cultures

(Andersen, 2005). Yet, previous studies have demonstrated aseptic flaming may have detrimental

implications on algal cultures, such as decreased growth rates and increased cell death in

cyanobacterial Prochlorococcus spp. cultures linked to flame-generated peroxides (Morris and

Zinser, 2013).

In the present study, we assessed the effects of flaming as part of aseptic technique on the

pH of freshwater culture media and subsequently cyanobacterial cell growth. Effects were

assessed as follows: (1) aseptic flaming on different freshwater growth media from the scientific

literature, (2) aseptic flaming on the growth of M. aeruginosa cultures. Based on our

observations, recommendations are made in the form of media modifications to effectively

sterilize while eliminating the potential introduction of confounding variables.

Methods
Freshwater media selection and preparation

Four freshwater culture media commonly used for phytoplankton culturing (C, CB, CT,

and CSi) were selected: recipes were taken from a publication of the NIES-Collection Microbial

Culture Collection (Watanabe, 1997). The components and concentrations for each medium are

identical except for the buffer type and concentration, allowing for direct comparisons of the

specific buffers in response to flaming (Table 2.1). Briefly, the major constituents of C medium

include 1.5x10-4 g/mL Ca(NO3)2⋅4H2O, 1.0x10-4 g/mL KNO3, 5.0x10-5 g/mL Na2⋅𝛽𝛽-

glycerophosphate, 4.0x10-5 g/mL MgSO4⋅7H2O, and trace metals (Steffen et al., 2015a). To

maintain comparability, silica was not included in CSi media used in this study. Beyond these

39
Table 2.1: Buffer concentration (mM), buffering pH range, and pKa for each of the five
freshwater culture media used in this study (CT, C, CSi, CB, BG-11).

Media [Buffer] / L pH Range pKa at 20°C

CT 1.64 mM TAPS 7.7-9.1 8.49
C 4.13 mM Tris 7.0-9.0 8.20
CSi 2.10 mM HEPES 6.8-8.2 7.55
CB 3.06 mM Bicine 7.6-9.0 8.35
BG-11 1.75mM K2HPO4 •3H20 8.2-9.6 7.21

40
media, BG-11 medium (Andersen, 2005), one of the most-commonly used for cyanobacterial

culture work, was also tested. We also assessed the effects of flaming on 3 heterotrophic

bacterial growth media: Lysogeny broth (LB) (Bertani, 1951), Nutrient broth (Sambrook et al.,

1989) and Minimal media (M9) (Miller, 1972). All media were prepared in 1 L volumes, titrated

to a pH of 8.2 with 1 M NaOH and autoclaved. Samples were cooled at room temperature (20.5

°C) and gases allowed to equilibrate (for 1 day). The pH of each medium was confirmed via the

sub-sampling a 10 mL volume and analyzed using a temperature compensating laboratory pH

meter (Mettler Toledo Seven CompactTM pH/Ion meter S220, fitted with a Mettler InLab Expert

Pro-ISM electrode with a temperature range of up to 100ºC). The pH meter was calibrated daily

using 3 pH standards (pH 4.0, 7.0, and 10.0) to yield an efficiency of 97% or higher prior to all

measurements. Standard vitamin additions for freshwater media in the forms of vitamin B12,

thiamine HCl, and biotin were omitted from all media.

Effects of flaming on media types

Twenty-five mL of each medium were aliquoted into 50 mL sterile, acid-washed glass

culture tubes (sterilized, acid-washed tubes were used in all following experiments). For each

growth medium, triplicate control and flame-treated replicates were generated. The flaming

technique used in this experiment was performed in accordance with protocols expected during

standard lab-based sub-sampling (for cell number, etc.) during a typical cyanobacteria growth

study. Sub-sampling consisted of the removal of a small volume of culture daily for cell

enumeration, though subsampling was not performed in all of the experiments of this study to

decrease risk of heterotrophic contamination. Flame replicates were inverted 3-times (normally

done to ensure cell resuspension) and held at a 45° angle while passed above the blue cone of the

41
flame for 4 passes through the flame. After a brief pause, during which one would collect a small

volume for determining cell number or inoculating cells into fresh media, the mouth of the

culture tube was again passed through the flame 4-times (thus a total of 8 passes through the

flame corresponds to a single sub-sampling event). The cap was then immediately replaced, and

the tube inverted 3-times before it was replaced in the rack (see aseptic flaming demonstration

video, https://youtu.be/Bvn6OvM8JpM). In this study, the treatment of replicates using an open

flame will be denoted as “flamed”. The controls in this study are subject to the same aseptic

procedure as denoted above, with the exception of the presence of a flame: they have been

assigned the terms “unflamed or air flamed”. All samples were stored at 20.5° C and

approximately 15-20 µmol photons m-2 s-1. This mock sampling of uninoculated growth media

was performed daily for 10-d. Prior to the daily flaming, pH was measured for each replicate

using a pH probe, with sterilization of the probe performed prior to sampling using 70% EtOH.

Effects of buffer age on media pH

We assessed whether buffer age had any effect on the buffering capacity. Twenty-five

mL volumes of CT media containing aged TAPS buffer stock (TAPS stock solution aged

approximately 3.5 months and stored at 4° C) and CT media containing fresh TAPS buffer

(TAPS stock solution prepared 3 days prior and stored at 4° C) were aliquoted into 50 mL glass

culture tubes in triplicate manner for the control treatments. The culture tubes were air-flamed in

the same manner as mentioned prior. All replicates were stored on a lab bench at ~20.5° C. The

pH of replicates was monitored daily for 10-d.

42
Influence of gas-exchange on media pH

The potential implications of gas exchange with the atmosphere were analyzed, with 25

mL volumes of CT media aliquoted into 50 mL glass culture tubes in triplicate. The lids were

closed firmly for one treatment group, while the lids were left loose for the second treatment.

Replicates were stored on a lab bench at 20.5° C and 15-20 µmol photons m-2 s-1. The pH of all

replicates was monitored daily for 10-d. Efforts to explore potential mitigating-practices in

reducing flaming effects on media pH were performed by abstaining from any type of inversion

or simulated shaking of the media. 25 mL volumes of CT media were aliquoted in triplicate into

50 mL glass culture tubes. Control and flamed treatments were applied daily for 10-d, without

inversion of the tubes or any other form of disturbance. All replicates were stored on a lab bench

at 20.5° C and approximately 15-20 µmol photons m-2 s-1.

Consequences of photo-oxidation on media pH

It has been demonstrated that light exposure causes photooxidation (and the inherent

generation of reactive oxygen species, such as the weak acid HOOH) within a variety of media

utilizing buffers such as HEPES, TAPS, Bicine, and TRIS (Morris and Zinser, 2013). To

determine the effects of photooxidation on media pH, 25 mL volumes of CT media were

aliquoted into 50 mL glass culture tubes in triplicate for air-flame treatments. All replicates were

stored on a lab bench at 20.5° C and approximately 15-20 µmol photons m-2 s-1. The “non-

photooxidation” treatment replicates were wrapped in foil (kept in dark conditions) to prevent

superoxide generation. The pH of all replicates was monitored daily for 10-d.

43
Effects of increased buffer concentration on media pH

Attempts to minimize the effects of flaming upon media pH were made by increasing the

TAPS buffer concentration in CT media by 10-fold and 100-fold (the later serving as the positive

control) and monitoring the effects of flaming and air-flaming upon the triplicated experimental

groups. All replicates were stored on a lab bench at 20.5° C and 15-20 µmol photons m-2 s-1. The

pH of all replicates was monitored daily for 10-d.

Testing drivers of pH decline as a result of flaming

In a preliminary investigation into the potential of flame-generated CO2 as the primary

driver of pH decline, limewater (Ca(OH)2 was prepared following standard protocol

(Shakhashiri, 1983). Aliquots (25 mL) were dispensed in triplicate to test the effects of flaming

across the following treatments: flamed 8-times, 16-times, 24-times, and 32-times. Note 8 passes

through the flame corresponded to 1 sub-sampling event, or day 1 in a growth study, with 16

passes corresponding to 2 sub-samplings, or day 2 in a growth study, and so forth. Negative

controls included 3 tubes which were air-flamed, and positive controls were exposed to 10

seconds of CO2 bubbling (administered via exhalation into straw submerged in limewater).

Qualitative assessment of turbidity due to the formation of CaCO3 as an indirect proxy for CO2

introduction was made using a spectrophotometer (Thermo Spectronic Genesys 20) at 600 nm.

Quantification of the net CO2 (total carbon-TC) incorporated into media due to flaming

was measured by the Water Quality Core Facility at the University of Tennessee Knoxville

utilizing a Shimadzu Carbon/Nitrogen analyzer (TOC-L Shimadzu, detection limit of 4 µg / L

according to the manufacturer’s specifications). Volumes of CT media (25 mL) were aliquoted

into 50 mL acid washed, sterilized tubes in triplicate per treatment. Control (unflamed), flamed

44
8-times, flamed 16-times, flamed 24-times, and flamed 32-times treatments were administered,

then immediately analyzed for TC with technical replicates of 3 performed per sample.

Effects of flaming on freshwater cyanobacterial cultures

To determine the consequences that flaming may have on freshwater cyanobacterial

cultures, axenic Microcystis aeruginosa NIES 843 was inoculated in triplicate into CT media at a

pH of 8.2 and flamed daily for 10-d to mimic sub-sampling during growth assays. Note one

mock-sampling event corresponds to 8 passes through the flame according to standard aseptic

protocol performed in the video demonstration. Cultures were incubated at 26° C and

approximately 55-60 µmol photons m-2 s-1 in incubators on a diel cycle (VWR low temperature

diurnal illumination incubator). Chlorophyll a auto fluorescence (fluorescence signal units, FSU)

was quantified daily as a proxy for relative biomass and cell health utilizing a fluorometer

(Turner Designs TD-700), equipped with a blue mercury bulb, a #10-050R excitation filter (340

– 500 nm) and a #10-115 (680 nm) emission filter. The instrument was standardized before each

use with Turner’s solid standard (#7000-994). Flaming/air-flaming was performed post-FSU

measurements. Final pH values were measured on day 10 after the completion of the experiment.

Effects of increased buffer concentration on freshwater cyanobacterial cultures

We assessed the effects of a 10-fold increase in the concentration of TAPS buffer in CT

media on M. aeruginosa NIES 843 growth. Cultures were inoculated into standard TAPS-

buffered CT medium as well as a 10X TAPS-buffered CT medium for comparison. Post-

inoculation, the cultures were incubated at 26° C and approximately 55-60 µmol photons m-2 s-1

as mentioned previously for 10-d. FSU (Tuner Designs TD-700) and cell number (BD

45
FACSCalibur flow cytometer) were assessed every 2-d for additional comparison of the

techniques themselves.

Statistical analyses:

Multiple T-tests corrected for multiple comparisons (Holm-Sidak method) were used to

statistically assess the following: flame assays of freshwater media (Figure 2.1), M. aeruginosa

growth rates (Figure 2.4b), flame assays of heterotrophic media (Appendix Figure 2.5), external

influences on media pH (Appendix Figure 2.6), buffer concentration effects on media pH

(Appendix Figure 2.8), effects of shaking on media pH (Appendix Figure 2.9), and the

comparison of M. aeruginosa biomass accumulation in standard CT vs. 10X TAPS-buffered CT

medium (Appendix Figure 2.11). One-way ANOVAs corrected for multiple comparisons (Tukey

method) were used to statistically assess the following: net pH decline by buffer type (Figure

2.2), TC (mM) incorporation as a result of flaming (Figure 2.3), and limewater turbidity as an

indicator for CO2 incorporation (Appendix Figure 2.7).

Results
Effects of flaming on pH of freshwater media

After flaming the mouths of the culture tubes, a decline in pH was observed in all the

freshwater media analyzed in this study compared to the controls (Figure 2.1). After 10 days, CT

(p = 0.01) and BG-11 (p = 0.02) experienced the greatest decline in pH after flaming, dropping

from 8.2 to average final pH values of 6.47 and 6.45, respectively (Figure 2.1d, e). TRIS (p =

0.03), HEPES (p = 0.01), and Bicine buffered- C media (p = 0.01) had average final pH values of

7.02, 6.86, and 6.87, respectively (Figure 2.1a, b, c). Differences between the controls and

flamed replicates manifested after only 2-d of daily flame sterilization (or 16 passes through the

46
flame resulting from 2 standard aseptic flaming events). A comparison of net decline in pH of

each media/buffer indicated there were no differences between the pH declines in CT and BG-11

(p = 0.99) when flamed, and neither were different from the control (p = 0.27) (Figure 2.2). This

data confirms CT (TAPS buffer) and BG-11 (inorganic phosphate buffer) demonstrated the

greatest decline in pH when flamed (1.5 - 1.7 pH units). TRIS, HEPES, and Bicine-buffered

media that was flamed exhibited similar trends as in the unflamed control (Figure 2.2). The data

demonstrated a consistent trend amongst the 5 freshwater media/buffers tested, in which flaming

resulted in a decline in pH after just 2-d (16 passes through the flame) and continuing up to 10-d

(80 passes through the flame). This trend was observed (albeit to a lesser extent) when the

experiment was replicated by flaming 3 growth media used for heterotrophic bacteria culture

work: LB, Nutrient broth, and M9 (Appendix Figure 2.5). In the case of the unbuffered media

(LB and Nutrient broth) the declining pH trend was observed on a smaller scale, with flamed LB

decreasing in pH from 8.2 to an average of 7.5 after 10-d of flaming and Nutrient broth to an

average pH of 7.42 (Appendix Figure 2.5a, b). M9 media demonstrated a similar decline in pH

upon flaming (Appendix Figure 2.5c). However, all time points of the M9 flamed replicates (T2-

T10) were significantly different from the controls (p < 0.01) whereas this was not the case for

LB and Nutrient broth. The data demonstrated that flame sterilization resulted in pH declines of

lower magnitude in unbuffered media compared to buffered media, with freshwater media

experiencing the highest decline in pH due to flaming.

Influence of external conditions on the pH of freshwater media

A pH decline was observed in the control replicates in all 5 freshwater media,

though to a much lesser extent than the flamed replicates (Figure 2.2). Factors associated with

47
Figure 2.1: Flame-induced declines in pH in control (black) and flamed replicates (orange) of
freshwater media. Data present mean ± SEM. Where not shown, error bars are within the
symbol. (a) C media/Tris buffer (indicated by squares). (b) CSi/HEPES media (triangles). (c)
CB/Bicine media (diamonds). (d) CT media/TAPS buffer (circles). (e) BG-11/inorganic
phosphate buffer (hexagons).

Figure 2.2: Net change in pH (T0 − T10 pH) observed in each media after 10 d of aseptic
flaming. Net change in pH observed in the control replicates of each media indicated in black.
Net change in pH of the flamed replicates for each media indicated in orange.
Analysis of carbon dioxide as driver of flame-induced pH declines

48
media handling and storage were analyzed to identify potential confounding variables or

contributing factors to this pH decline observed in controls. Buffer age appeared to have no

immediate implications corresponding to pH (Appendix Figure 2.6a). The effects of light

exposure (i.e. photo-oxidation) on pH revealed no differences (p = 0.99) between CT subjected

to photo-oxidation and those exposed to photo-oxidation inhibition (Appendix Figure 2.6c).

Influence of gas exchange (i.e. tight vs. loose caps) revealed no effect on pH (p = 0.90)

throughout the 10-d sampling (Appendix Figure 2.6b). Efforts to minimize pH decline in CT

media were made by forming a CT -TAPS buffer concentration gradient. Daily flaming events

over the course of 10-d revealed the largest pH declines in the control (protocol-standard

concentration of TAPS), with 10-fold and 100-fold increases in TAPS resulting in more stable

pH (Appendix Figure 2.8a, b). While a 10-fold increase in TAPS buffer did not result in

statistical differences between the un-flamed and flamed treatments (p = 0.07) (p = 0.35), the

decline in pH was observed to a lesser extent compared to the standard buffer media. Additional

means of minimizing flame-induced pH decline was accomplished by abstaining from tube

inversion/disturbance throughout 10-d and revealed a substantially lower net decline in pH in CT

media (Appendix Figure 2.9).

Analysis of carbon dioxide as driver of flame-induced pH declines

Preliminary qualitative means for determining the absence/presence of CO2 in flamed

media were conducted utilizing a Limewater turbidity assay (Shakhashiri, 1983). After flaming,

evidence of CO2 incorporation was observed via the visible formation of a Ca(CO)3 precipitate in

the corresponding manner: negative controls or “air-flamed” replicates depicted no change in

turbidity formation of precipitates, while 8, 16, and 24 passes through the flame each resulted in

increased turbidity and precipitate formation, serving as a visual proxy for CO2 introduction to

49
the media (Appendix Figure 2.7a). Spectrophotometric analyses further support the limewater

analyses (Appendix Figure 2.7b). Although variable, as the number of passes through the flame

increased, a parallel rise in turbidity and precipitate formation of Ca(OH)2 occurred. Increased

variability of Ca(OH)2 formation amongst replicates was observed as the number of passes

through the flame increased. Quantification of TC input revealed an average of 170 +/- 6.4 mM

C (or 4.26 mmol C per 25mL) present in the unflamed controls, 193 +/- 3.60 mM C present in

the 8 passes through the flame replicates, 196 +/- 2.0 mM C present in the 16 passes through the

flame replicates, 204 +/- 3.6 mM C present in the 24 passes through the flame replicates, and 216

+/- 2.0 mM C present in the 32 passes through the flame replicates (Figure 2.3a). It was

determined all flaming treatments resulted in an increase of C to the media compared to the

control (p=0.002, p=0.0002, p=0.0001, p=0.000, with p values listed corresponding to tubes

passed through the flame 8 to 32 times in consecutive order), and increased flaming highly

correlated with the increase in TC mmol (R2 = 0.93, slope = 0.06 Figure 2.3b).

Implications of flaming for freshwater cyanobacteria cultures

The M. aeruginosa in the CT control consistently had the lowest FSU. Replicates flamed

24-times had the highest FSU (with FSU interpreted to be indirectly indicative of cell number

and biomass, while directly suggestive of cell health (Figure 2.4a). Estimates of the relationship

between FSU and cell density (cells/mL) confirmed the presence of a strong and significant

relationship between the two (R2 = 0.92, p = 0.0001 Appendix Figure 2.10). Replicates that were

flamed exhibited higher instances of culture death in comparison to the controls (death of

replicates that were flamed 8-times graphed separately, Figure 2.4a). The pH of the dead

replicates was considerably lower than replicates that survived, all of which remained close to

50
Figure 2.3: TC (mmol L−1) in control and flamed CT media replicates. (a) Net TC (mmol L−1)
incorporated into CT media control and flame treatments. Eight passes through the flame
indicated by orange squares, 16 passes through the flame indicated by purple triangles, 24 passes
through the flame indicated by blue inverted triangles, and 32 passes through the flame indicated
by green diamonds. (b) Linear trend of mean TC (mmol L−1) introduced per flame treatment (y =
0.06323 [number of passes through flame] + 4.391, R2 = 0.93).

51
the initial pH of 8.2 (Figure 2.4c). One replicate that was flamed 16-times had a correspondingly

low pH, though there was no indication of death. The M. aeruginosa cultures that were flamed 8-

times and 24-times had higher growth rates than the control (p ≥ 0.001), with replicates flamed

16-times falling short of statistical difference (p = 0.072) (Figure 2.4b).

Discussion
There is little doubt of the importance of aseptic technique in microbiology and microbial

ecology. In the current study, our efforts to monitor microscale changes in pH as drivers of

phytoplankton community composition in lakes revealed a Heisenberg-like moment (Wheeler et

al., 1983) in biology – where the process of making the measurement shapes the actual measure

one is making. In short, our use of aseptic technique used during experiments examining pH

effects was itself affecting pH of our culture media. These observations are critical in face of the

massive number of ongoing climate change studies in aquatic systems where pH is a major

variable. The data we collected suggested flaming as a part of aseptic technique had effects on 5

freshwater culture media as well as effects on the growth of the cyanobacterium M. aeruginosa.

While flaming is a sterilization method which has been endorsed for decades (Andersen, 2005;

Madigan, 2015), our results indicate media pH was altered by as much as 1.7 units via the

introduction of CO2, serving as a confounding variable. In contrast, the assessment of other

variables including buffer age, photo-oxidation effects, and gas exchange were found to have no

effect on pH. These observations effectively ruled out additional contributors to the pH decline

and indicated flaming was solely responsible for the pH change. Previous studies have identified

pH as a parameter of importance when culturing algae, with broad growth rate optima falling

52
Figure 2.4: Aseptic flaming effects on CT media + axenic M. aeruginosa. (a) M. aeruginosa
growth treated with flaming: control replicates (0 daily passes through the flame) indicated by
solid black dots, 8 daily passes through the flame indicated by solid orange squares, 16 daily
passes through the flame indicated by solid purple triangles, 24 daily passes through the flame
indicated by solid blue inverted triangles, and dead replicates indicated in gray open symbols. (b)
Growth rate of each replicate + mean growth rate. (c) pH of replicates at the completion of the
experiment (after 10 daily treatments): control (0 daily passes through the flame), 8 daily passes
through the flame, 16 daily passes through the flame, or 24 daily passes through the flame,
respectively.

53
within pH values of 7 to 9, though many organisms exhibit growth outside of these ranges

(Lavens and Sorgeloos, 1996; Berberoglu et al., 2008; Huang et al., 2017) . Changes in pH can

effect algal cultures, with many algal species having growth optima at narrow pH ranges (Lavens

and Sorgeloos, 1996; Tsaloglou, 2016). The cyanobacterium used in this study, M. aeruginosa,

has an optimal pH for growth between 7.5 - 10 (Fang et al., 2018) with an optimal photosynthetic

potential at pH > 8.0 (Bano and Siddiqui, 2004). Flaming of culture vessel openings, specifically

during the process of aliquoting media for inoculation, had the capacity to drive down the pH

(Figure 2.1a-e). Additionally, flaming open culture tubes post-inoculation could further decrease

pH levels if CO2 is not assimilated by the cyanobacteria. BG-11 and CT media exhibited the

highest net declines in pH (Figure 2.2), which raises concern, as these are frequently utilized

freshwater growth media. Random culture death in certain flamed cultures (Figure 2.4) may be

due to the variability of CO2 incorporation from flaming, as some cultures may have experienced

larger-scale CO2 inputs that resulted in marked pH decline if the CO2 is not consumed by the

culture. Previous studies have also found that up to 2 µM of HOOH may be generated as a result

of flaming (Morris and Zinser, 2013). While it is recognized that mM concentrations of HOOH

are needed to cause cell death (Palenik et al., 1991; Alam et al., 2001), the 2 µM of HOOH

generated from flaming results in physiological changes during the upregulation of intracellular

peroxidases (Price and Harrison, 1988). There may also be a role here for incomplete combustion

products (although we did not investigate this).

Dissolved inorganic carbon (DIC) plays a significant role in the growth and development

of phytoplankton species. CO2 and HCO3- are the major chemical forms of carbon assimilated

during photosynthesis (Schindler et al., 1971; Wetzel, 2001). Previous studies have indicated free

CO2 is labile and readily accessible to most algae and aquatic photoautotrophs (Wetzel, 2001).

54
M. aeruginosa isolates have been shown to have a high affinity for DIC which is reflective in its

enzymatic low half-saturation constant, allowing it to outcompete other algal species due to its

efficient carbon acquisition (Yamamoto and Nakahara, 2005).

It appeared that cultures which received CO2 via flaming benefited substantially and

exhibited higher growth rates in our study. This suggested C availability may have been growth

limiting in CT medium when cultures were not bubbled or shaken (Figure 2.4). CO2 is

extremely soluble in aqueous solutions, demonstrating 200-times higher solubility relative to

oxygen (Wetzel, 2001). H2CO3 is a weak acid which readily and rapidly dissociates, losing its 2

protons in a 2-step process (pKa1 = 6.43, pKa2 = 10.43 at 15° C) (Schindler, 1971), (Wetzel,

2001). Hence, CO2 that is not readily taken-up by M. aeruginosa in its initial gaseous form

readily dissolves into the media and hydrates into H2CO3, as demonstrated in the flaming of

freshwater media alone (Figure 2.1). In addition, M. aeruginosa’s direct uptake of flame-

generated gaseous CO2 is further supported in the consistent pH values amongst control and

flamed replicates, as healthy cultures did not experience any significant declines in pH, save for

1 replicate in the flamed 16-times treatment (Figure 2.4c). Yet, when culture death did occur, the

final pH was in the same range as CT media subjected to 10-d of flaming (Figure 2.1d),

indicating CO2 had dissolved into the media. Should the concentration of dissolved CO2 become

too high, H2CO3 formation would yield low culture pH, which would inhibit growth and

development in a variety of algal species, including M. aeruginosa (Weisse and Stadler, 2006).

Thus, while CO2 influxes via flaming increased growth rate and biomass in carbon-limited

cultures, if the CO2 dissolution rate becomes too high, significant drops in pH will invoke

detrimental effects on the culture. In total it appears likely that the availability of CO2 coupled to

pH decline shaped the inconsistency observed amongst replicates in previous experiments.

55
Comments and recommendations
We have demonstrated that flame sterilization of culture medium vessel openings can

result in a significant decline in the pH of freshwater media due to the dissolution and hydration

of CO2. This decline in pH has several implications, as optimal pH ranges for algal

photosynthesis and growth are often narrow and well-defined (Huang et al., 2018). Moreover,

under current climate scenarios, research into the effects of small pH shifts is common yet could

be compromised by this observation. This pH shift would also effect the bioavailability and

accessibility of macro and micronutrients, including crucial trace metals within the media, many

of which are associated with photosynthetic processes (Wetzel, 2001). Low pH levels that result

from CO2 introduction, coupled with HOOH generation via flaming and excessive trace metal

bioavailability, may result in randomized spontaneous cell and culture death, as seen in our

previous flamed growth assays (Figure 2.4). Flaming further effects cultures by serving as a

variable yet direct source of CO2, which is labile and thus readily accessible for uptake by M.

aeruginosa. This CO2 is likely consumed by M. aeruginosa cells, influencing cell growth rate

and biomass accumulation (Figure 2.4). This mechanism is further supported by the relatively

constant pH levels observed in all M. aeruginosa cultures (flamed and non-flamed) as after 10

days the pH remains constant (excluding the “dead” replicates who demonstrated lower pH).

These CO2 inputs may result in a higher degree of replicate variability in unshaken M.

aeruginosa cultures in contrast to shaken cultures, as unshaken cultures in potentially C-limiting

media such as CT are likely to be C-limited.

There is a need for updated aseptic alternatives to flame-induced sterilization. It is

advisable to abstain from flaming when feasible, and alter flaming practices when flaming is

necessary. Refraining from inverting, shaking, or otherwise disturbing the media post-flaming

may decrease the dissolution of CO2 into the media and mitigate pH decline, though this is not

56
always feasible. When CT media was subject to a 10-d flaming study without any

inversion/shaking, it was found the pH decline was drastically diminished (Appendix Figure 2.9)

and not significantly different (p=0.30) from the unflamed controls. Additionally, exercising

caution in media selections is advised, as the commonly utilized medium (BG-11) was shown to

have the highest decline in pH overall. Ten-fold increases in TAPS buffer concentration within

CT media were shown to mitigate pH shifts within unflamed replicates (Appendix Figure 2.8)

with no observed biological consequences (Appendix Figure 2.11). However, the approach of

simply increasing buffer concentrations must be pursued with caution: e.g., TRIS is toxic to cells

in high concentrations and HEPES generates higher amounts of reactive oxygen species when

present in higher concentrations (Morris and Zinser, 2013). Our results indicate a 10-fold TAPS

increase in CT may serve as a suitable medium in future studies concerning pH (climate change

research), but each specific case should be individually examined.

This study serves as a cautionary of the unintended effects of flaming of culture vessel

openings upon the media and microorganisms. While flaming has served as an aseptic technique

for decades, it may be time to put aside the Bunsen burner and pursue further alternatives to this

classic practice.

Acknowledgments
We thank Dr. Gary LeCleir, Eric Gann, Robbie Martin, Lena Pound, Naomi Gilbert and

Professor George Bullerjahn for comments. We also thank Dr. Adrian Gonzalez with the Water

Quality Core Facility at the University of Tennessee Knoxville for TC analyses. This work was

supported by grants from the National Science Foundation (DEB-1240870; IOS-1451528) to

SWW and by funding from the NIH (1P01ES028939-01) and NSF (OCE-1840715) to the

Bowling Green State University Great Lakes Center for Fresh Waters and Human Health.

57
References:
Alam, M. Z. B., M. Otaki, H. Fufami, and S. Ohgaki. 2001. Direct and indirect inactivation of
Microcystis aeruginosa by UV radiation. Water Res. 35: 1008-1014.
Andersen, R. A. 2005. Algal culturing techniques. Elsevier/Academic Press.
Bano, A., and P. J. Siddiqui. 2004. Characterization of five marine cyanobacterial species with
respect to their pH and salinity requirements. Pakistan J. Bot. 36: 133-144.
Berberoglu, H., L. Pilon, and A. Melis. 2008. Radiation characteristics of Chlamydomonas
reinhardtii CC125 and its truncated chlorophyll antenna transformants tla1, tlaX and
tla1-CW+. Int. J. Hydrogen Energy 33: 6467-6483.
Bertani, G. 1951. Studies on lysogenesis I.: the mode of phage liberation by lysogenic
Escherichia coli1. J. Bacteriol. 62: 293.
Bykowski, T., and B. Stevenson. 2008. Aseptic technique. Curr. Prot. Microbiol. 11: A. 4D. 1-
11.
Coté, R. J. 1998. Aseptic technique for cell culture. Curr. Prot. Microbiol. 1.3.1 - 1.3.10.
Fang, F., Y. Gao, L. Gan, X. He, and L. Yang. 2018. Effects of different initial pH and irradiance
levels on cyanobacterial colonies from Lake Taihu, China. J. Appl. Phycol. 30: 1777-
1793.
Ferreira, C. M., I. S. Pinto, E. V. Soares, and H. M. Soares. 2015. (Un) suitability of the use of
pH buffers in biological, biochemical and environmental studies and their interaction
with metal ions–a review. RSC Advances 5: 30989-31003.
Harrigan, W., and M. E. McCance. 1966. Laboratory methods in microbiology. Academic Press.
London and New York. 374 pp.
Huang, J. and others 2018. New perspectives on CO2, temperature, and light effects on BVOC
emissions using online measurements by PTR-MS and cavity ring-down spectroscopy.
Environ. Sci. Technol. 52: 13811-13823.
Huang, J. and others 2017. Influence of pH on heavy metal speciation and removal from
wastewater using micellar-enhanced ultrafiltration. Chemosphere 173: 199-206.
Lavens, P; Sorgeloos, P. (eds.) Manual on the production and use of live food for aquaculture
FAO Fisheries Technical Paper. No. 361. Rome, FAO. 1996. 295 pp.
Madigan, M. T., J. M. Martinko, K. S. Bender, D. H. Buckley, D.A. Stahl, and T. Brock. 2015.
Brock Biology of Microorganisms. Fourteenth edition. ed. Pearson.
Miller, J. 1972. Experiments in molecular biology. Cold Spring Harbor Laboratory, Cold Spring
Harbor, NY.
Morris, J., and E. R. Zinser. 2013. Continuous hydrogen peroxide production by organic buffers
in phytoplankton culture media. J. Phycol. 49: 1223-1228.
Palenik, B., N. M. Price, and F. M. Morel. 1991. Potential effects of UV-B on the chemical
environment of marine organisms: a review. Environ. Pollution 70: 117-130.

58
Price, N. M., and P. J. Harrison. 1988. Specific selenium-containing macromolecules in the
marine diatom Thalassiosira pseudonana. Plant Physiol. 86: 192-199.
Sambrook, J., E. Fritsch, and T. Maniatis. 1989. Molecular cloning. Cold Spring Harbor
Laboratory Press. 1659 pp.
Sanders, E. R. 2012. Aseptic laboratory techniques: volume transfers with serological pipettes
and micropipettors. J Vis. Exp. e2754.
Schindler, D. 1971. A hypothesis to explain differences and similarities among lakes in the
Experimental Lakes Area, northwestern Ontario. Jour. Fish. Res. Bd. Canada 28: 295-
301.
Schindler, D. W., F. Armstrong, S. Holmgren, and G. Brunskill. 1971. Eutrophication of Lake
227, experimental lakes area, northwestern Ontario, by addition of phosphate and nitrate.
Jour. Fish. Res. Bd. Canada 28: 1763-1782.
Steffen, M. M., B. S. Belisle, S. B. Watson, R. A. Bourbonniere, G. L. Boyer, and S. W.
Wilhelm. 2015. Metatranscriptomic evidence for co-occurring top-down and bottom-up
controls on toxic cyanobacterial communities. Appl. Environ. Microbiol. 81: 3268-3275.
Talling, J. 2010. pH, the CO2 system and freshwater science. Freshwater Rev. 3: 133-147.
Tsaloglou, M.-N. 2016. Microalgae. Caister Academic Press, Cambridge, MA. 152 pp.
Watanabe, M. M., M. Kawachi, M. Kiroki and F. Kasai 2000. NIES - Collection. List of Strains,
Sixteh Edition 2000. Microalgae and Protozoa. Microbial Culture Collections, National
Institute for Environmental Studies, Tsukuba, Japan, 159 pp.
Weisse, T., and P. Stadler. 2006. Effect of pH on growth, cell volume, and production of
freshwater ciliates, and implications for their distribution. Limnology and Oceanography
51: 1708-1715.
Weitz, J. S. and others 2015. A multitrophic model to quantify the effects of marine viruses on
microbial food webs and ecosystem processes. ISME J. 9: 1352-1364.
Wetzel, R. G. 2001. Limnology: lake and river ecosystems. gulf professional publishing.
Wheeler, J. A., H. Zurek, J. A. Wheeler, and H. Zurek. 1983. p. 62-84. Quantum Theory and
Measurement. Princeton University Press.
Yamamoto, Y., and H. Nakahara. 2005. Competitive dominance of the cyanobacterium
Microcystis aeruginosa in nutrient‐rich culture conditions with special reference to
dissolved inorganic carbon uptake. Phycological Research 53: 201-208.

59
Appendix

Figure 2.5 Effects of flaming on heterotrophic media. A.) Effects on pH of LB medium, control
indicated by black circles, flaming indicated by orange circles. B.) Effects on pH of Nutrient broth,
control replicates indicated by black diamonds, flamed replicates indicated by orange diamonds.
C.) Effects on pH of M9, control replicates indicated by black squares, flamed replicates indicated
by orange squares. Treatments that demonstrate statistical difference denoted as an x (p < 0.01).

60
Figure 2.6 Effects of environmental conditions on pH of CT media. A.) Effects of buffer age on
pH, aged buffer indicated by solid black circles, new buffer indicated by open grey circles B.)
Effects of gas exchange on pH, closed caps indicated by solid black circles, loosened caps
indicated by open grey circles C.) Effects of photooxidation on pH, light-exposed samples
indicated with solid black circles, dark-treatment samples indicated with open grey circles. Where
not visible, error bars are within symbols. Treatments that demonstrate statistical difference
denoted as an x (p < 0.01).

61
Figure 2.7 Determination of CO2 as the primary driver of pH decline in freshwater media. A.)
Limewater turbidity assay demonstrating increased turbidity with increased flaming. Treatments
from left to right: control, flamed x8, flamed x16, flamed x24, and positive control (+). B.)
Turbidity (600 nm) of CaCO3 precipitation as indirect proxy for CO2 incorporation.

62
Figure 2.8 Mitigation of pH decline due to flaming of CT medium with TAPS ent. A.) Effects of
increased buffer concentration on pH in control (standard TAPS concentration, indicated by solid
black circles), TAPSx10 (open dotted circles), and TAPSx100 (open crossed circles) B.) Effects
of increased buffer concentration on pH in flamed replicates. Standard TAPS concentration
indicated by solid orange circles, TAPSx10 indicated by open dotted circles, TAPSx100
indicated by open crossed circles.

63
Figure 2.9 Effects of tube inversion (shaking) on control and flame replicate pH decline. Data for
non-inverted CT medium control replicates indicated by solid black circles, with non-inverted
flamed CT medium replicates indicated by solid orange circles.

64
Figure 2.10 Correlation between chlorophyll a auto fluorescence (FSU) and cell density
(cells/mL) from samples collected during M. aeruginosa growth curve in standard CT media
(closed black circles). The data indicate a strong and significant relationship (R2 = 0.92, p =
0.0001) suggesting fluorescence is a good predictor of cell density.

65
Figure 2.11 Comparison of M. aeruginosa biomass accumulation when grown in CT medium
with standard TAPS concentration (black circles), and CT media with a ten-fold increase in
TAPS concentration (green squares).

66
CHAPTER III: ELUCIDATING THE ROLE OF PH ON MICROCYSTIS-DIATOM
COMPETITION DYNAMICS IN LAKE ERIE

67
Publication Note
This chapter is a version of a peer-reviewed, published article in Frontiers in Microbiology
12(2021):598736 by Brittany N. Zepernick, Eric R. Gann, Robbie M. Martin, Helena L. Pound,
Lauren E. Krausfeldt, Justin D. Chaffin, and Steven W. Wilhelm.

BNZ and SWW designed these experiments. BNZ and LEK performed preliminary culture
optimizations and experimental planning. In vitro co-culture assays and in vitro silica deposition
assays conducted by BNZ. Epifluorescence microscopy performed by ERG and BNZ. In situ Lake
Erie silica deposition assays conducted by BNZ, HLP and RMM with logistical support from JDC.
JDC performed data collection and analyses corresponding to the 2015 M. aeruginosa bloom in
Figure 1. Statistical analyses performed by BNZ. All authors contributed to the drafting of the
manuscript.

68
Abstract
Cyanobacterial Harmful Algal Blooms (CyanoHABs) commonly increase water column

pH to alkaline levels ≥ 9.2, and to as high as 11. This elevated pH has been suggested to confer a

competitive advantage to cyanobacteria such as Microcystis aeruginosa. Yet, there is limited

information regarding the restrictive effects bloom-induced pH levels may impose on this

cyanobacterium’s competitors. Due to the pH-dependency of biosilicification processes, diatoms

(which seasonally both precede and proceed Microcystis blooms in many fresh waters), may be

unable to synthesize frustules at these pH levels. We assessed the effects of pH on the

ecologically relevant diatom Fragilaria crotonensis in vitro, and on a Lake Erie diatom

community in situ. In vitro assays revealed F. crotonensis monocultures exhibited lower growth

rates and abundances when cultivated at a starting pH of 9.2 in comparison to pH 7.7. The

suppressed growth trends in F. crotonensis were exacerbated when co-cultured with M.

aeruginosa at pH conditions and cell densities that simulated a cyanobacteria bloom. Estimates

demonstrated a significant decrease in silica (Si) deposition at alkaline pH in both in vitro F.

crotonensis cultures and in situ Lake Erie diatom assemblages, after as little as 48 h of alkaline

pH-exposure. These observations indicate elevated pH negatively affected growth rate and

diatom silica deposition; in total providing a competitive disadvantage for diatoms. Our

observations demonstrate pH likely plays a significant role in bloom succession, creating a

potential to prolong summer Microcystis blooms and constrain diatom fall resurgence.

69
Introduction
Toxin-producing cyanobacteria of the genus Microcystis have inundated fresh waters in

recent decades (Steffen et al., 2014a). Blooms have detrimental ecological and economic effects

due to the production of secondary metabolites and the formation of extensive biomass that,

upon bloom termination, can drive anoxia (Anderson, 2009). To this end, there is a crucial need

to determine the factors responsible for the ecological success of Microcystis.

The mechanisms by which Microcystis displaces other phytoplankton in fresh waters,

including Lake Erie (USA/Canada), Lake Okeechobee (USA) and Lake Taihu (China) remain

unclear, but are likely multifaceted. In these lakes, a seasonal pattern of phytoplankton taxa

succession has emerged. Non-toxic diatoms and other algae dominate throughout fall, winter,

and spring, only to be displaced by Microcystis blooms mid-summer into fall (Ke et al., 2008;

Reavie et al., 2014). This successional trend has been evidenced in Erie’s paleolimnological

record, which traces the emergence of eutrophication back to the 1930’s (Allinger and Reavie,

2013a). Monitoring efforts of a 2015 Microcystis bloom in Lake Erie’s western basin further

confirmed this succession, demonstrating diatoms dominated the early summer period prior to

their succession by cyanobacteria in mid-summer (Figure 3.1a).

Several factors contribute to the ecological success of Microcystis. Summer dominance in

Lake Erie’s western basin has been attributed to nutrient loading (Michalak et al., 2013; Paerl et

al., 2016), predation (Vanderploeg et al., 2001; Steffen et al., 2015b), and temperature

(Andersson et al., 1994; Peng et al., 2018). Likewise, spring diatom decline has been linked to

silica limitation and temperature intolerance (Twiss et al., 2012; Reavie et al., 2016). While these

factors each contribute to Microcystis growth during cyanobacterial bloom years, non-

cyanobacterial bloom years have demonstrated that diatoms, such as the temperature tolerant

Fragilaria crotonensis, can persist and even dominate the summer water column in Lake Erie

70
(Hartig and Wallen, 1986; Saxton et al., 2012a; Reavie et al., 2014). Indeed, F. crotonensis

summer blooms were a frequent occurrence in the western basin of Lake Erie throughout the

1960-1980’s during lake remediation efforts (Hartig, 1987). Furthermore, monitoring data from

the 2015 Lake Erie Microcystis bloom indicates dissolved silica concentrations, though lowest

during the peak diatom bloom, were non-limiting during Microcystis succession (Appendix

Figure 3.7). These observations suggest there are additional and multiple factors contributing to

Microcystis succession of spring-summer diatoms (Wilhelm et al., 2020) Amongst these factors

playing a potential role in succession dynamics is pH. For example, during the 2015 M.

aeruginosa bloom monitoring efforts, a sharp rise in water column pH was found to co-occur

with cyanobacterial bloom formation (Figure 3.1a, b). While pH can have multiple effects on

cellular physiology and biogeochemistry, in the present study, we investigated the response of

one physiological aspect of diatoms – silicification - to the shifts in pH that occur during

Microcystis blooms.

Microcystis blooms increase water column pH above 9.2 as CO2 is consumed during

photosynthesis (Verspagen et al., 2014; Bullerjahn et al., 2016; Krausfeldt et al., 2019). This

alkaline pH is considered advantageous to cyanobacteria (Wilson et al., 2010; Shruthi and

Rajashekhar, 2014), due to their unique carbon concentrating mechanisms (CCMs) which confer

a competitive advantage during growth at low CO2 / high pH conditions (Shapiro, 1990; Raven,

2010; Sandrini et al., 2016). Yet freshwater and marine diatoms have been shown to possess a

competitive array of CCMs themselves which optimize CO2 and HCO3- acquisition (Clement et

al., 2017). While this may allow diatoms to evade pH-induced carbon-limitation, elevated pH has

been shown to decrease carbon uptake, growth rate and metabolic processes in various marine

diatoms (Raven, 1981; Taraldsvik and Myklestad, 2000). While these effects of pH on diatom

71
carbon acquisition have been well characterized, pH-induced effects on other metabolic

processes have been widely unstudied to date, particularly in freshwater diatoms.

One metabolic process that serves as a distinctive metric for diatom viability is deposition.

Diatoms possess siliceous cell walls (i.e., frustules) which may pose a unique disadvantage in

alkaline bloom conditions. Silica deposition relies on the uptake of dissolved silica (dSi) in the

form of silicic acid (Si[OH]4) to synthesize biogenic silica (bSi) frustules (Vrieling et al., 1999;

Otzen, 2012; Hildebrand et al., 2018). In marine and estuarine systems, diatom viability has been

strongly correlated to pH, with studies demonstrating marine diatoms are unable to survive at pH

> 8.7 due to silica solubility dynamics and the inhibition of biosilicification (Hansen, 2002;

Hervé et al., 2012). Yet, to our knowledge, the effect of pH in freshwater diatom Si deposition

remains unassessed.

In this study, we combined lab and field-based experiments to assess the effect of pH on

diatom growth rate and silica deposition. As part of this effort, we assessed the effect of pH on F.

crotonensis growth rate in monoculture and ecologically relevant co-cultures with M.

aeruginosa. Effects of pH on silica deposition were assessed using a fluorescent dye (PDMPO)

which intercalates into newly formed frustules. Laboratory and field-based results indicate pH

conditions consistent with M. aeruginosa blooms (i.e., pH ≥ 9.2) decreased diatom growth rate,

abundance, and silica deposition. In total the pH shift reduces diatom viability and the ability to

compete for valuable niche-space with cyanobacteria.

72
Methods

Assessing successional trends of a 2015 Lake Erie Microcystis bloom

A temporal dataset collected at the Ohio State Stone Laboratory was used to preliminarily

evaluate the dynamics of cyanobacteria, diatoms, and pH during the summer of 2015, which was

the largest M. aeruginosa bloom observed to date (Davis et al., 2019). Water column pH was

recorded every 30 min via a Yellow Spring Instruments 6600v2 multiprobe sonde suspended at 1

m depth from a buoy located between South Bass and Gibraltar Islands (N 41.66°, W 82.92°).

Water samples for phytoplankton community composition were collected next to the buoy

several times a week. Diatom-specific and cyanobacteria-specific Chl a concentrations were

recorded via a bbe Moldaenke FluoroProbe (Beutler et al., 2002). Total chlorophyll a

concentrations corresponding to the sampling period have been provided (Appendix Figure 3.8),

with the complete details of this dataset found in the original publication (Chaffin et al., 2018).

Effect of pH on growth in F. crotonensis and M. aeruginosa monocultures

To assess the effects of this environmentally observed pH on diatom and cyanobacteria

growth, in vitro monoalgal experiments were performed using 2 model taxa. F. crotonensis

SAG 28.96 (acquired from the Culture Collection of Algae at the University of Göttingen,

Germany) and M. aeruginosa NIES 843 (acquired from the National Institute for Environmental

Studies, Japan) were maintained in batch cultures using CT medium (Wilhelm, 2017) at

respective optimal pH levels of pH 7.7 (Guillard and Lorenzen, 1972; Hervé et al., 2012) and 8.2

(Watanabe et al., 2000; Krausfeldt et al., 2019). To initiate in vitro monoculture experiments, F.

crotonensis and M. aeruginosa cultures were filter-concentrated respectively using a 2.0-µm and

1.0-µm nominal pore-size 47-mm diameter polycarbonate filter and inoculated into sterile 250

73
mL filter-vented, baffled polycarbonate flasks (Corning) at a starting concentration of ~700

cells/mL. Monocultures were maintained in 125 mL CT medium (Steffen et al., 2015b)

containing a non-limiting concentration of silicic acid (176 µM Na2SiO3 ∙ 9H2O)(Hervé et al.,

2012) and adjusted to an initial pH of 7.7 (optimal pH for diatom growth) or 9.2 (pH observed

during Microcystis blooms). pH conditions in the lab study were maintained by adding TAPS

buffer as described previously (Zepernick et al., 2020). Cultures were monitored for 30 d at 26°

C, with orbital shaking at 70 rpm, and a light intensity of approximately 55-60 µmol photons m-2

s-1 on a 12:12 light: dark photoperiod cycle.

Abundances were measured every two days via flow cytometry (BD FACSCalibur).

Populations of each species were gated and counted based on forward scatter (FSC), a proxy for

size, and chlorophyll a fluorescence (FL3) using FlowJoTM software (Becton, Dickinson and

Company). Due to the filamentous nature of F. crotonensis, direct estimates of individual cell

abundance are challenging (Bramburger et al., 2017). In this study, F. crotonensis abundances

are estimated based on filaments/mL, which form a tight cluster (Appendix Figure 3.9).

Exponential growth rates (𝜇𝜇) were calculated as the slope of log -scaled data and were reported

in filaments/mL for F. crotonensis, and cells/mL for M. aeruginosa. Specifically, log-scaled

growth data for each replicate were fitted with a linear regression to select time points to be used

for 𝜇𝜇 calculations. Time points demonstrating the logarithmic growth phase with a linear

regression R2 value of ≥ 0.95 (i.e., the most linear data points) were subsequently used to

calculate average growth rate). Culture pH was checked every 10-d using a sterilized pH probe

(Mettler Toledo Seven Compact TM pH/Ion meter S220 fitted with a Mettler InLab Expert Pro-

ISM electrode with a temperature range and correction of up to 100°C). Growth experiments

74
were performed in biological triplicate. We note all results will be referred to in this study based

on the initial pH condition of the treatment.

Effect of pH on growth in F. crotonensis and M. aeruginosa co-cultures

To evaluate the effects of ecologically relevant pH conditions on diatom growth, in vitro

co-culture assays were performed. Concurrent with the monoculture assays, co-cultures of F.

crotonensis SAG 28.96 and M. aeruginosa NIES 843 were inoculated. To initiate in vitro co-

culture experiments, M. aeruginosa and F. crotonensis batch cultures were filter-concentrated

and inoculated into the same experimental media and initial pH levels as previously described.

Taxa were inoculated at 3 ratios (reported as F. crotonensis: M. aeruginosa) based on the

succession patterns observed (Figure 3.1): 10:1 ratio simulating a spring diatom bloom, 1:1 ratio

simulating the onset of the summer M. aeruginosa bloom, and a 1:10 ratio simulating the peak

M. aeruginosa bloom. All co-cultures were inoculated at net starting concentrations of ~7,000

cells/mL. Hereafter, co-culture treatments will be referred to by the F. crotonensis: M.

aeruginosa ratio. Co-cultures were subjected to the same incubation conditions and procedures

as described above. All experiments were performed in biological triplicate.

Effect of pH on in vitro F. crotonensis silica deposition

To determine the effect of pH on silica deposition in vitro, batch cultures of F.

crotonensis SAG 28.96 were inoculated with the fluorescent dye PDMPO [2‐(4‐pyridyl) ‐5‐((4‐

(2‐dimethylaminoethylaminocarbamoyl) methoxy) phenyl) oxazole] (Lysosensor DND 160

Yellow/Blue; Invitrogen, Carlsbad, CA). Diatom cultures were acclimated to pH conditions of

7.7 and 9.2 for a 6-d period (i.e., approximately 2 doubling times). Acclimated cultures were

75
filter-concentrated using a 2.0-µm nominal pore-size 47-mm diameter polycarbonate filter and

inoculated in acid-clean, sterilized 50 mL glass culture tubes containing 25 mL of CT medium

with 176 μM Na2SiO3 ∙ 9H2O. Tubes were inoculated at an initial concentration of ~1500

filaments/mL. PDMPO was added at a final concentration of 0.125 µM (Leblanc and Hutchins,

2005). Cultures were incubated at 26° C and approximately 55-60 µmol photons m-2 s-1 on a

12:12 light: dark photoperiod cycle for 48 h. Abundances were determined via flow cytometry as

described above.

Si deposition was assessed using microscopic and fluorometric approaches that detect

freshly incorporated PDMPO. Bulk Si deposition into individual cells was assessed via

epifluorescence microscopy. After 48 h, 2 mL of each culture was filtered onto 0.2-µm nominal

pore-size 25 mm diameter black polycarbonate filters (Millipore), mounted onto glass slides,

treated with anti-fade (Suttle and Fuhrman, 2010), and a coverslip applied prior to storage (-80°

C). F. crotonensis killed controls (0.5% glutaraldehyde-fixed) were performed according to

previous studies (Saxton et al., 2012a) to assess abiotic incorporation. Slides were viewed on a

Leica DM5500 (Wetzlar, Germany) epifluorescence microscope equipped with a Hamamatsu

ORCA-ER camera (Sewickley PA) according to previous methods (Saxton et al., 2012a). A

“Texas red” filter cube set (λex = 520-600nm; λem = 570-720 nm) was used to view chlorophyll a

autofluorescence, and a “DAPI filter” cube set (λex = 340-380 nm; λem > 425 nm) to view

PDMPO fluorescence, indicative of Si deposition during the experimental period. Quantitative

assessment of epifluorescence microscopy data was achieved by randomized scoring of 100 F.

crotonensis filaments per pH treatment: analyses included chlorophyll a fluorescing cells per

filament, PDMPO fluorescing cells per filament, and the proportion of filaments demonstrating

≥ one PDMPO fluorescing cells. Total Si deposition was quantitatively measured

76
fluorometrically after HCl-Milli Q lysis to remove unincorporated PDMPO from the silica

deposition vesicle (SDV), followed by frustule digestion with hot-NaOH (Saxton et al., 2012a;

Zepernick et al., 2019). After 48 h of growth, 20 mL of culture was collected on a 0.2-µm

nominal pore-size 47-mm diameter polycarbonate filter and subjected to HCl-Milli Q lysis.

Filters were flash frozen and stored at -80° C until hot NaOH digestion. After frustule digestion,

samples were cooled in an ice bath and neutralized using 1M HCl. PDMPO fluorescence was

quantitatively determined using a Turner Designs TD-700 fluorometer fitted with a specialized

filter set (𝜆𝜆ex = 360-380nm: 𝜆𝜆em = 522-542 nm, Andover Corporation, Salem, NH). A PDMPO

standard curve was generated using PDMPO and NaOH-HCl matrix (Appendix Figure 3.10),

with the PDMPO concentration converted to Si using a conversion factor of 3230:1 for Si:

PDMPO (mol: mol) (Saxton et al. 2012). Total silica deposited into frustules after 48 h (μmol)

was normalized to final abundance(filaments/mL). PDMPO experiments were performed with 5

biological replicates.

Effect of pH on in situ Lake Erie diatom community silica deposition

To evaluate the effects of pH on silica deposition in natural populations, we queried

diatom-enriched communities from Lake Erie with PDMPO under varying pH conditions.

Samples were collected in late July of 2019 from the western basin of Lake Erie near the Ohio

State University Stone Laboratory on South Bass Island (N 41.69; W -82.79). Water column

physiochemistry (temperature = 25.3° C; dissolved oxygen = 7.60 mg/L; pH = 8.63; turbidity =

1.46 NTFU; chlorophyll a = 0.13 µg/L) was recorded prior to sampling using an EXO

multiparameter sonde (YSI xylem TM). Experiments were initiated by enriching for diatoms

using a 64-µm mesh phytoplankton net, which was lowered to a depth of ~7 m. Equal volumes

77
of concentrated seston were diluted with lake water and inoculated into acid washed, rinsed 500

mL polycarbonate bottles. Lake water was buffered using TRIS (4.13 mM final concentration) in

accordance with the protocol for freshwater “C medium” (Watanabe et al., 2000). The

experiment consisted of three pH treatments: 7.7, 9.2, and an in situ pH control for the sample

collection site (pH 8.6). To achieve these pH conditions, samples were incrementally titrated

using 1 M HCl or NaOH. PDMPO dye was added at a final concentration of 0.125 µM (Leblanc

and Hutchins, 2005), and bottles were placed into an in situ mesh incubation chamber for 48 h.

Sample pH and chlorophyll a concentration were determined at the initiation (T0) and

termination (Tf; 48 h) of the incubation. pH was assessed via immediate readings of 15 mL

subsamples using a pH probe (Mettler Toledo Seven Compact TM pH/Ion meter S220, fitted with

a Mettler InLab Expert Pro-ISM electrode with a temperature range of up to 100° C).

Chlorophyll a concentration was determined from filtration of 100 mL onto 0.2-µm nominal

pore-size 47-mm diameter polycarbonate filters. Samples were extracted in 90% acetone for 24

hrs at 4° C and assessed on a Turner Designs 10-AU Field Fluorometer (Welschmeyer, 1994).

To measure silica deposition, samples were collected by filtering 100 mL of sample onto 0.2-µm

nominal pore-size 47-mm diameter polycarbonate filters, followed by the HCl-Milli Q lysis

method as described above. Samples were flash frozen in liquid N2 and stored at -80° C until

further processing. Quantitative Si deposition analyses were performed using hot-NaOH frustule

digestion, fluorometry, and subsequent calculations using a fresh standard curve (Appendix

Figure 3.11) (Zepernick et al., 2019). Total silica deposited per bottle after 48 h (μmol) was

normalized to chlorophyll a concentration (μg/L) (Saxton et al., 2012a). Field experiments were

performed with 4 biological replicates.

78
Statistical analyses

Statistical comparisons were made using unpaired two-tailed t-tests, ordinary one-way

ANOVAs, or ordinary two-way ANOVAs, depending on experimental design. One-way and

two-way ANOVA post-hoc multiple comparisons were adjusted using Tukey’s HSD. While F.

crotonenis and M. aeruginosa monoculture and co-culture growth rates are presented separately

in this text, all experiments were performed concurrently in the same conditions, and thus have

been statistically analyzed using ordinary two-way ANOVAs to compare both pH and abundance

(Appendix Tables 3.1 and 3.2). All analyses were performed using GraphPad’s Prism software

(Version 8). For this study, we consider a p value < 0.05 to be significant but have reported all

values so the reader may decide (Appendix Tables 3.1, 3.2, 3.3 and 3.4).

Results
Role of pH in 2015 Lake Erie bloom succession trends

Monitoring data from a 2015 M. aeruginosa-dominated bloom demonstrated that total chl

a concentration across the season varied from ~ 1-3 µg/L in June to ~20-120 µg/L in July and

August (Chaffin et al. 2018; Appendix Figure 3.8). The pre-cyanobacterial bloom period (June

through early July) was dominated by diatoms which form ~50%-80% of the total chl a

concentration, whereas cyanobacteria were less than 10% (Figure 3.1a). The mean daily pH

during the corresponding diatom bloom period was between 8.08 and 8.56 (Figure 3.1b).

Conversely, during the cyanobacterial bloom period cyanobacteria dominate, forming ~56% -

84% of the chl a concentration, whereas diatoms were less than 7% and frequently not detected

(Figure 3.1a). The mean daily pH during the Microcystis bloom peaked at ~9.27 and remained

higher than 9 throughout most of August (Figure 3.1b).

79
Figure 3.1 Environmental data corresponding to a 2015 Lake Erie M. aeruginosa bloom. (A)
Relative abundance (reported as percentage of total chlorophyll a) of diatoms (blue squares) and
cyanobacteria (blue circles) within the Lake Erie water column. (B) Average daily pH of the
Lake Erie water column (closed blue circles).

80
Alkaline pH decreases growth rate of F. crotonensis monocultures

Monoculture experiments demonstrated F. crotonensis growth was suppressed at high

pH. F. crotonensis cultures inoculated at pH 9.2 attained lower abundances throughout the 30-d

experiment compared to their pH 7.7 counterparts (Figure 3.2a). F. crotonensis mean growth rate

at pH 7.7 was 𝜇𝜇 = 0.34, with pH 9.2 monocultures exhibiting a significantly lower mean growth

rate of 𝜇𝜇 = 0.22 (p= 0.0002) (Figure 3.2b). Overall, F. crotonensis monocultures inoculated at pH

9.2 had a 1.5-fold lower mean growth rate compared to its pH 7.7 equivalents.

Effects of pH on M. aeruginosa monoculture were less pronounced. M. aeruginosa reached

higher cell abundances at pH 7.7 compared to pH 9.2 equivalents (Appendix Figure 3.12a). Yet,

M. aeruginosa growth rates were unaffected by pH overall (p= 0.503) (Appendix Figure 3.12b).

M. aeruginosa modulates the effect of pH on F. crotonensis in co-culture

F. crotonensis reached higher abundances at pH 9.2 than pH 7.7 when co-cultured with

non-dominant M. aeruginosa concentrations of 10:1 and 1:1 (Figure 3.3a, c). Additionally, F.

crotonensis growth rates at the designated pH treatments were not significantly different in the

10:1 ratio (p= 0.999) and 1:1 ratio (p= 0.206). (Figure 3.3b, d). Conversely, at the dominant M.

aeruginosa co-culture ratio of 1:10, F. crotonensis growth was substantially suppressed at pH 9.2

(Figure 3.3e). F. crotonensis mean growth rate at pH 7.7 in the 1:10 co-culture was 𝜇𝜇 = 0.35,

with pH 9.2 co-cultures exhibiting a significantly lower mean growth rate of 𝜇𝜇 = 0.23 (p=

0.0002) (Figure 3.3f). Overall, at the 10:1 ratio and pH 9.2, F. crotonensis has a 1.5 times lower

mean growth rate compared to its pH 7.7 equivalents. As in the monocultures, the effects of pH

on M. aeruginosa growth in the co-culture replicates were less pronounced. M. aeruginosa

81
Figure 3.2 (A) In vitro F. crotonensis monoculture growth curves at pH 7.7 (black squares) and
pH 9.2 (green squares). (B) F. crotonensis growth rate at pH 7.7 (black squares) and pH 9.2
(green squares). Statistically significant differences between pH treatments are denoted by p
values generated by Two-way ANOVAs. Standard error of the mean reported by error bars.

82
Figure 3.3 (A) In vitro F. crotonensis co-culture growth curves in a 10:1 ratio (F. crotonensis:
M. aeruginosa) at pH 7.7 (black squares) and pH 9.2 (green squares). (B) F. crotonensis growth
rate at 10:1 ratio (C) F. crotonensis growth curves in a 1:1 ratio (D) F. crotonensis growth rate in
1:1 ratio (E) F. crotonensis growth curves in a 1:10 ratio (F) F. crotonensis growth rate in a 1:10
ratio. Statistically significant differences between pH treatments are denoted by p values
generated by Two-way ANOVAs. Standard error of the mean reported by error bars.

83
reached higher cell concentrations at pH 7.7 in all co- culture ratios compared to pH 9.2

equivalents (Appendix Figure 3.13a, c, e). Yet, M. aeruginosa culture ratios compared to pH 9.2

equivalents (Appendix Figure 3.13a, c, e). Yet, M. aeruginosa growth rates were unaffected by

pH overall (p > 0.503) (Appendix Figure 13b, d, f).

M. aeruginosa concentrations correspond with pH increases

While F. crotonensis monocultures inoculated at pH 7.7 remained at this pH throughout

the 30-d experiment (Appendix Figure 3.14), M. aeruginosa monocultures demonstrated a steady

climb in pH, reaching ~8.0 by 30-d (Appendix Figure 3.15). Similarly, all 3 co-culture ratios

inoculated at pH 7.7 demonstrated continual increases in pH throughout the 30-d experiment,

reaching final pH levels of ~8.10 (Appendix Figure 3.16). In total, pH 7.7 inoculated M.

aeruginosa monocultures and co-cultures all experienced increases in pH of ~0.30, coinciding

with increases in M. aeruginosa concentrations throughout the 30-d experiment (Appendix

Figure 3.13) (Appendix Figure 3.17a). Upon further analysis, M. aeruginosa concentrations

demonstrated a strong linear relationship with culture pH increases observed in the mono and co-

cultures (Simple linear regression R2 ≥ 0.8460) (Appendix Figure 3.18). Collectively, pH was

maintained within a range of approximately +/- 0.40 pH units throughout the 30-d experiment

(Appendix Figure 3.17).

Silica deposition decreases at alkaline pH in F. crotonensis monocultures

In vitro PDMPO incubations demonstrated a pronounced effect of alkaline pH on silica

deposition. Epifluorescence microscopy revealed pH 7.7 acclimated cultures deposited more Si

after 48-h PDMPO incubations (Figure 3.4a, c, e; Appendix Figure 3.19a, c) compared to pH 9.2

acclimated cultures (Figure 3.4b, d, f; Appendix Figure 3.19b, d). Quantitative counts of these

84
Figure 3.4 Epifluorescent microscopy images (40x magnification) of F. crotonensis filaments
after 48 h PDMPO incubations. Scale bar represents 25 μm. Chlorophyll a autofluorescence is
depicted in red, and PDMPO fluorescence is in blue. (A, C, E) F. crotonensis cultures acclimated
to pH 7.7. (B, D, F) F. crotonensis cultures acclimated to pH 9.2.

85
images also demonstrated pH 9.2 acclimated cultures formed significantly smaller filaments than

pH 7.7 acclimated cultures (p<0.0001; unpaired two-tailed t-test t=4.057, df=197, n=100)

(Appendix Figure 3.20a). In total, ~66% of cells in each filament deposited silica after 48 h in

the pH 7.7 treatments, while only ~30% of the cells in each filament had deposited Si at pH 9.2

(p< 0.0001; unpaired two-tailed t-test t=9.457, df=197, n=100) (Appendix Figure 3.20b). 100%

of F. crotonensis filaments incubated at pH 7.7 exhibited at least one diatom cell depositing Si,

while only 66% of pH 9.2 F. crotonensis filaments demonstrated at least one instance of Si

deposition per filaments.

Fluorometric data revealed pH 7.7 acclimated cultures deposited a mean of 25.28 μmol

Si total, while pH 9.2 acclimated cultures deposited a significantly lower mean of 15.81 μmol Si

total (p< 0.0001; unpaired two-tailed t-test t=8.544, df=8, n=5) (Appendix Figure 3.21).

Normalization of this data to abundance (final filament concentration) reflected a similar trend.

F. crotonensis cultures acclimated to pH 7.7 deposited a mean of 1.17 nmol Si/filament, while

cultures acclimated to pH 9.2 deposited a significantly lower mean of 0.59 nmol Si/filament (p<

0.0001; unpaired two-tailed t-test t=9.446, df=8, n=5) (Figure 3.5). Overall, diatoms acclimated

to pH 9.2 deposited ~50% less silica in comparison to their pH 7.7 counterparts.

Silica deposition decreases at alkaline pH in Lake Erie diatom communities

Elevating the pH negatively influenced Si deposition in the Lake Erie diatom community

(Figure 3.6). Samples incubated at pH 7.7, control pH (8.6), and pH 9.2 deposited a mean of

219.07 μmol Si total, 214.52 μmol Si total, and 194.36 μmol Si total, respectively (Appendix

Figure 3.22). Total Si deposited in pH 9.2 treatments was less than pH 7.7 treatments, though not

statistically significant (p = 0.127). Normalization of this data to chlorophyll a concentration

86
Figure 3.5 Si deposited per filament after 48 h PDMPO incubations in F. crotonensis cultures
acclimated to pH 7.7 treatments (black squares) and pH 9.2 (green squares). Statistically
significant differences are denoted by respective p values generated by unpaired two-tailed t-
tests. Standard error of the mean reported by error bars.

Figure 3.6 Si deposited per chl a concentration in pH 7.7 treatments (black squares), control pH
8.6 (grey squares), and pH 9.2 (green squares) after 48 h incubations. Statistically significant
differences are denoted by respective p values generated by One-way ANOVAs. Standard error
of the mean reported by error bars.

87
upheld this observation, with pH 7.7 treatments depositing a mean of 27.61 μmol Si/ Chl a,

control treatments depositing 22.45 μmol Si/Chl a, and pH 9.2 treatments depositing 18.16 μmol

Si/ Chl a, respectively (Figure 3.6). The pH 9.2 treated community deposited significantly less Si

per chl a concentration after 48 h compared to the pH 7.7 treated community (p = 0.037).

Overall, Lake Erie diatom communities incubated at pH 9.2 deposited ~1.5 times less Si per chl

a concentration than their pH 7.7 counterparts.

Discussion
Seasonal succession drivers associated with Microcystis blooms are complicated. While

it remains clear that nutrient-loading results in the planktonic biomass observed during toxic

cyanobacterial blooms, the environmental conditions that allow specific organisms to

outcompete others are more nuanced (Wilhelm et al., 2020). Here we build on the idea that pH

serves as a contributing piece to this puzzle. Previous analyses have suggested a correlation

between pH and diatom-Microcystis succession in Lake Taihu, China (Ke et al., 2008) and Lake

Erie (Krausfeldt et al., 2019). In these and other cases, authors have suggested that the effects of

pH on carbon acquisition and the superior carbon concentrating mechanisms of cyanobacteria

were the major mechanistic drivers of these observations. Additionally, previous studies have

indicated nutrient speciation at alkaline pH may favor Microcystis, such as the discovery that

urea serves as both a carbon and nitrogen source to M. aeruginosa at alkaline pH levels

(Krausfeldt et al., 2019). While the direct and indirect effects of pH on freshwater diatom carbon

and nutrient acquisition cannot be discounted or ruled out, our data demonstrated a previously

uncharacterized effect of pH on freshwater diatoms, which may serve to depress them beyond, or

in addition to, their ability to acquire CO2. We present this information as a factor that likely

enhances the exclusion of Si depositing phytoplankton observed during heated summer

88
competition. These observations lead to a take-away message from this study: sometimes it is not

the ability of Microcystis but the inability of its competitors that results in the taxa succession.

Effect of pH on freshwater diatom growth

We used pH manipulation in mono- and co-culture experiments to demonstrate that an

elevated pH, consistent with Microcystis-bloom conditions, negatively affected the diatom F.

crotonensis. F. crotonensis monocultures inoculated at pH 9.2 exhibited lower growth rates and

failed to establish a substantial population, demonstrating alkaline pH alone decreases the

viability of this model freshwater diatom. Likewise, when co-cultured with dominant

concentrations of M. aeruginosa at the 1:10 ratio, these alkaline pH growth effects were

exacerbated. This data is consistent with freshwater diatom decline at the alkaline pH levels

observed during summer Microcystis blooms. Interestingly, when F. crotonensis was co-cultured

in the 1:10 ratio at pH 7.7, it was able to maintain growth rates resembling those observed in the

pH 7.7 monocultures, suggesting alkaline pH may have a larger role in diatom viability than

previously thought. Surprisingly, when F. crotonensis was co-cultured with M. aeruginosa at

10:1 and 1:1 (i.e., where the diatom biomass dominated) it did not exhibit significant declines in

growth rate at pH 9.2. F. crotonensis cell abundances in the 10:1 and 1:1 co-culture were higher

at pH 9.2 than their pH 7.7 counterparts, though statistical significance was lacking. While the

underlying mechanisms of these results remain unelucidated, this data suggests that while pH is a

factor, it alone is likely not the sole driver of diatom exclusion. Another important observation is

that in all M. aeruginosa mono and co-cultures inoculated at pH 7.7, pH increases in tandem

with M. aeruginosa cell concentration. This data demonstrates M. aeruginosa is indeed capable

of driving the pH up substantially despite increased buffer use, and mimics environmental data

89
previously observed during a 2015 Microcystis bloom. Cumulatively, in vitro co-cultures suggest

diatoms may be able to persist in the water column during the spring diatom blooms and onset

Microcystis blooms regardless of water column pH. Yet, during peak Microcystis bloom

conditions when the pH is driven to alkaline levels, diatoms are at a disadvantage. This data

further suggests these persisting alkaline pH levels may prolong the Microcystis bloom period by

preventing diatom fall resurgence as a result of decreased diatom growth and viability.

Effects of pH on diatom silica deposition

Though previous studies have investigated the effects of pH on marine diatom

biosilicification (Vrieling et al., 1999; Martin‐Jézéquel et al., 2000; Hansen, 2002), this study

builds on these observations through an assessment of pH effects on freshwater diatoms. We

used PDMPO assays to demonstrate that pH conditions consistent with Microcystis blooms

significantly decrease silica deposition in both cultured and environmental freshwater diatoms.

When interpreting this data it is important to note, flow cytometry analyses of filamentous

microorganisms such as F. crotonensis count “filaments per volume” rather than “cells per

volume”. As a result, an estimate for average number of cells per chain is often used to calculate

cells/mL (Bramburger et al., 2017). Our data demonstrate the average number of cells per

filament differs significantly in response to culture pH (Appendix Figure 3.20), which has the

potential to introduce additional error in estimates of cell/mL and biovolume. Indeed,

epifluorescence microscopy revealed pH 9.2-acclimated F. crotonensis had ~ 1.5 times shorter

filaments and ~2 times fewer silica depositing cells per filament. Fluorometric data from the Si

deposition assays revealed a similar trend, with F. crotonensis cultures acclimated to pH 9.2

depositing ~50% less silica per filament in comparison to their pH 7.7 counterparts. This trend

90
was further observed in Lake Erie diatom communities, with communities incubated at pH 9.2

depositing ~1.5 times less Si per chl a concentration than their pH 7.7 counterparts.

Cumulatively, this data suggests pH-induced decreases in silica deposition may serve as an

important contributor to the freshwater diatom decline observed during Microcystis blooms.

Furthermore, these results also bring to light a need to further optimize detection and

normalization techniques in studies concerning filamentous phytoplankton.

While we have demonstrated a decrease in silica deposition at pH 9.2, the underlying

mechanisms remain unclear. Part of our limitation comes from the lack of knowledge concerning

functions in the organelle responsible for silica deposition, known as the silica deposition vesicle

(SDV). Despite decades of research, the SDV has yet to be isolated or characterized (Martin‐

Jézéquel et al., 2000; Hildebrand et al., 2018). Additionally, intracellular proteins and pathways

associated with diatom biosilicification remain elusive (Thamatrakoln and Hildebrand, 2008;

Vardi et al., 2009; Otzen, 2012). External alkaline pH may negatively affect intracellular

metabolism in the SDV, which relies on acidic conditions and an undisturbed pH gradient

(Vrieling et al., 1999; Hervé et al., 2012). Previous studies have also demonstrated high pH

levels may shape intra-cellular diatom silica storage pools (Werner, 1966; Azam et al., 1974;

Sullivan, 1977; Martin‐Jézéquel et al., 2000). Alkaline pH has previously been shown to limit

silica deposition by altering the chemical species of silicic acid or decreasing diatom-uptake rates

of dissolved silica (dSi) (Riedel and Nelson, 1985; Amo and Brzezinski, 1999). However, this is

unlikely in our study due to the non-limiting concentration of silicic acid in our media (176 µM

Na2SiO3 ∙ 9H2O).

In this study, we observed both a decrease in growth rate and silica deposition in response

to alkaline pH. Though evidence of a causal link between these two physiological processes is

91
lacking in this study, prior research has established that diatom silica uptake and deposition are

tightly coupled with the cell cycle, thus exerting a dependency of silica metabolism on the

growth rate (Martin‐Jézéquel et al., 2000; Hildebrand et al., 2018). Yet, while silica deposition is

essential to diatom viability and their ability to reproduce, diatoms can downregulate silica

deposition (i.e., form thinner frustules) to maintain optimal growth rates (Brzezinski et al., 1990;

McNair et al., 2018). Alternatively, previous research has also demonstrated that as pH increases,

growth rates decrease and intracellular silicic acid increases in marine diatoms, potentially

indicating a decoupling to silica deposition (Hervé et al., 2012). Hence, pH-associated effects on

alternate metabolic processes such as cellular respiration or photosynthesis (i.e., disruptions in

normal metabolic regulators) may also contribute to a decline in silica deposition. Furthermore,

thinner-frustules have been shown to increase the potential for viral infection and mortality in

marine diatoms, exacerbating population declines in the environment (Kranzler et al., 2019).

Ultimately, several of these underlying mechanisms may contribute to the decreased silica

deposition observed in this study, and further research is needed before any definitive

relationship between growth rate, silica deposition, and alkaline pH can be established.

A growing influence of pH in future phytoplankton diversity

Climate change continues to pose a threat to freshwater and marine systems alike. As a

result, there is a need to elucidate its effects on factors constraining the ecological success of

phytoplankton, such as pH. A recent study has demonstrated ocean acidification has the potential

to decrease marine diatom biosilicification rates (Petrou et al., 2019). Conversely, freshwater

systems are experiencing a basification attributed to increases in the frequency and duration of

HAB events (Wells et al., 2020), which has the potential to decrease freshwater diatom

biosilicification. In this manner, the effects of projected pH shifts on phytoplankton succession

92
serve as a critical point of study for ensuring the integrity of global aquatic systems (Flynn et al.,

2015; Wells et al., 2020). Our results build on these previous studies, demonstrating pH may

play a pivotal role not only in cyanobacterium-driven diatom decline, but phytoplankton taxa

diversity in general. While previous studies have demonstrated alkaline bloom-induced pH can

serve as a positive feedback mechanism for M. aeruginosa (Krausfeldt et al., 2019), this work

builds on these efforts by demonstrating these same conditions can facilitate the exclusion of

siliceous algae (diatoms). Furthermore, while Lake Erie summer cyanobacterial blooms drive up

the western basin pH to an average of ≥ 9.2, previous winter surveys demonstrate the diatom-

dominated water column remains at an average of ~7.8-8.2 despite fluctuations in chlorophyll a

and sampling location (Appendix Figure 3.23), though additional surveys are needed concerning

winter diatom blooms. Cumulatively, this data suggests a role of pH on both the inter and intra-

season shifts of phytoplankton taxa within Lake Erie and demonstrate the need to further assess

the role of pH in phytoplankton succession.

We noted our pH co-cultures of 10:1 and 1:1 yield higher peak diatom abundance at pH

9.2 in comparison to pH 7.7, despite the diatom monoculture yielding markedly lower

abundances at the same elevated pH. In this manner, there may be a window of opportunity for

diatoms to persist, and even benefit at low densities of M. aeruginosa if the cyanobacterial

populations do not become dominant. Many other biological / biogeochemical processes (e.g.,

inorganic carbon cycling, nitrogen speciation, trace metal chemistry) are pH sensitive and likely

play a role in shaping the outcomes of competition for niche space between phototrophs in fresh

waters. Our observations serve as a salient reminder that competition in aquatic systems is

condition dependent and often complicated by a mix of factors (Wilhelm et al., 2020).

93
 In this study, we confirmed that pH levels of 9.2 decreased diatom growth rate in the

filamentous diatom F. crotonensis. Our data further demonstrated that silica deposition in lab

cultures and environmental diatom communities declined at alkaline pH levels. Cumulatively,

these effects reduce diatom viability and fitness in the competition against Microcystis blooms.

While the pH shift itself may not be sufficient to exclude the diatoms from this (or any) system,

the resulting decrease in competitive ability for carbon and nutrients, in addition to pressure from

other factors including top-down regulators such as predators and viruses (Kranzler et al., 2019;

Pound et al., 2020) appear to tip the scale to favor the cyanobacteria. What remains to be

determined beyond this study is how these pressures allow a specific genus of cyanobacteria to

proliferate while in competition with many others.

Acknowledgements
We thank Dr. Gary LeCleir, Dr. Matthew Saxton, Dr. Robert McKay, Dr. George Bullerjahn, and

Naomi Gilbert for comments and suggestions. We also thank Keara Stanislawczyk for facilitating

field work at OSU Stone Lab.

94
References:
Allinger, L.E., and Reavie, E.D. (2013). The ecological history of Lake Erie as recorded by the
phytoplankton community. Journal of Great Lakes Research 39, 365-382.
Amo, Y.D., and Brzezinski, M.A. (1999). The chemical form of dissolved Si taken up by marine
diatoms. Journal of Phycology 35, 1162-1170.
Anderson, D.M. (2009). Approaches to monitoring, control and management of harmful algal
blooms (HABs). Ocean & Coastal Management 52, 342-347.
Andersson, A., Haecky, P., and Hagström, Å. (1994). Effect of temperature and light on the
growth of micro-nano-and pico-plankton: impact on algal succession. Marine Biology
120, 511-520.
Azam, F., Hemmingsen, B.B., and Volcani, B.E. (1974). Role of silicon in diatom metabolism.
Archives of Microbiology 97, 103-114.
Beutler, M., Wiltshire, K.H., Meyer, B., Moldaenke, C., Lüring, C., Meyerhöfer, M., Hansen, U.-
P., and Dau, H. (2002). A fluorometric method for the differentiation of algal populations
in vivo and in situ. Photosynthesis Research 72, 39-53.
Bramburger, A.J., Reavie, E.D., Sgro, G., Estepp, L., Chraïbi, V.S., and Pillsbury, R. (2017).
Decreases in diatom cell size during the 20th century in the Laurentian Great Lakes: a
response to warming waters? Journal of Plankton Research 39, 199-210.
Brzezinski, M.A., Olson, R.J., and Chisholm, S.W. (1990). Silicon availability and cell-cycle
progression in marine diatoms. Marine Ecology Progress Series, 83-96.
Bullerjahn, G.S., Mckay, R.M., Davis, T.W., Baker, D.B., Boyer, G.L., D’anglada, L.V.,
Doucette, G.J., Ho, J.C., Irwin, E.G., and Kling, C.L. (2016). Global solutions to regional
problems: Collecting global expertise to address the problem of harmful cyanobacterial
blooms. A Lake Erie case study. Harmful Algae 54, 223-238.
Chaffin, J.D., Kane, D.D., Stanislawczyk, K., and Parker, E.M. (2018). Accuracy of data buoys
for measurement of cyanobacteria, chlorophyll, and turbidity in a large lake (Lake Erie,
North America): implications for estimation of cyanobacterial bloom parameters from
water quality sonde measurements. Environmental Science and Pollution Research 25,
25175-25189.
Clement, R., Jensen, E., Prioretti, L., Maberly, S.C., and Gontero, B. (2017). Diversity of CO2-
concentrating mechanisms and responses to CO2 concentration in marine and freshwater
diatoms. Journal of Experimental Botany 68, 3925-3935.
Davis, T.W., Stumpf, R., Bullerjahn, G.S., Mckay, R.M.L., Chaffin, J.D., Bridgeman, T.B., and
Winslow, C. (2019). Science meets policy: a framework for determining impairment
designation criteria for large waterbodies affected by cyanobacterial harmful algal
blooms. Harmful Algae 81, 59-64.
Flynn, K.J., Clark, D.R., Mitra, A., Fabian, H., Hansen, P.J., Glibert, P.M., Wheeler, G.L.,
Stoecker, D.K., Blackford, J.C., and Brownlee, C. (2015). Ocean acidification with (de)
eutrophication will alter future phytoplankton growth and succession. Proceedings of the
Royal Society B: Biological Sciences 282, 20142604.
Guillard, R.R., and Lorenzen, C.J. (1972). Yellow-green algae with chlorophyllide C 1,2 Journal
of Phycology 8, 10-14.
Hansen, P.J. (2002). Effect of high pH on the growth and survival of marine phytoplankton:
implications for species succession. Aquatic Microbial Ecology 28, 279-288.

95
Hartig, J.H. (1987). Factors contributing to development of Fragilaria crontonensis Kitton
Pulses in Pigeon Bay waters of western Lake Erie. Journal of Great Lakes Research 13,
65-77.
Hartig, J.H., and Wallen, D.G. (1986). The influence of light and temperature on growth and
photosynthesis of Fragilaria crotonensis Kitton. Journal of Freshwater Ecology 3, 371-
382.
Hervé, V., Derr, J., Douady, S., Quinet, M., Moisan, L., and Lopez, P.J. (2012). Multiparametric
analyses reveal the pH-dependence of silicon biomineralization in diatoms. PloS One 7,
e46722.
Hildebrand, M., Lerch, S.J., and Shrestha, R.P. (2018). Understanding diatom cell wall
silicification—moving forward. Frontiers in Marine Science 5, 125.
Ke, Z., Xie, P., and Guo, L. (2008). Controlling factors of spring–summer phytoplankton
succession in Lake Taihu (Meiliang Bay, China). Hydrobiologia 607, 41-49.
Kranzler, C.F., Krause, J.W., Brzezinski, M.A., Edwards, B.R., Biggs, W.P., Maniscalco, M.,
Mccrow, J.P., Van Mooy, B.A., Bidle, K.D., and Allen, A.E. (2019). Silicon limitation
facilitates virus infection and mortality of marine diatoms. Nature Microbiology 4, 1790-
1797.
Krausfeldt, L.E., Farmer, A.T., Castro Gonzalez, H., Zepernick, B.N., Campagna, S.R., and
Wilhelm, S.W. (2019). Urea is both a carbon and nitrogen source for Microcystis
aeruginosa: tracking 13C incorporation at bloom pH conditions. Frontiers in
Microbiology 10, 1064.
Leblanc, K., and Hutchins, D.A. (2005). New applications of a biogenic silica deposition
fluorophore in the study of oceanic diatoms. Limnology and Oceanography: Methods 3,
462-476.
Martin‐Jézéquel, V., Hildebrand, M., and Brzezinski, M.A. (2000). Silicon metabolism in
diatoms: implications for growth. Journal of Phycology 36, 821-840.
Mcnair, H.M., Brzezinski, M.A., and Krause, J.W. (2018). Diatom populations in an upwelling
environment decrease silica content to avoid growth limitation. Environmental
Microbiology 20, 4184-4193.
Michalak, A.M., Anderson, E.J., Beletsky, D., Boland, S., Bosch, N.S., Bridgeman, T.B.,
Chaffin, J.D., Cho, K., Confesor, R., and Daloğlu, I. (2013). Record-setting algal bloom
in Lake Erie caused by agricultural and meteorological trends consistent with expected
future conditions. Proceedings of the National Academy of Sciences 110, 6448-6452.
Otzen, D. (2012). The role of proteins in biosilicification. Scientifica 2012.
Paerl, H.W., Scott, J.T., Mccarthy, M.J., Newell, S.E., Gardner, W.S., Havens, K.E., Hoffman,
D.K., Wilhelm, S.W., and Wurtsbaugh, W.A. (2016). It takes two to tango: When and
where dual nutrient (N & P) reductions are needed to protect lakes and downstream
ecosystems. Environmental Science & Technology 50, 10805-10813.
Peng, G., Martin, R.M., Dearth, S.P., Sun, X., Boyer, G.L., Campagna, S.R., Lin, S., and
Wilhelm, S.W. (2018). Seasonally relevant cool temperatures interact with n chemistry to
increase microcystins produced in lab cultures of Microcystis aeruginosa NIES-843.
Environmental Science & Technology 52, 4127-4136.
Petrou, K., Baker, K.G., Nielsen, D.A., Hancock, A.M., Schulz, K.G., and Davidson, A.T.
(2019). Acidification diminishes diatom silica production in the Southern Ocean. Nature
Climate Change 9, 781-786.

96
Pound, H.L., Gann, E.R., Tang, X., Krausfeldt, L.E., Huff, M., Staton, M., Talmy, D., and
Wilhelm, S. (2020). The “neglected viruses” of Taihu: Abundant transcripts for viruses
infecting eukaryotes and their potenital role in phytoplankton succession Frontiers in
Microbiology 11, 338.
Raven, J. (1981). "Nutrient transport in microalgae," in Advances in microbial physiology.
Elsevier), 47-226.
Raven, J.A. (2010). Inorganic carbon acquisition by eukaryotic algae: four current questions.
Photosynthesis research 106, 123-134.
Reavie, E.D., Barbiero, R.P., Allinger, L.E., and Warren, G.J. (2014). Phytoplankton trends in
the Great Lakes, 2001–2011. Journal of Great Lakes Research 40, 618-639.
Reavie, E.D., Cai, M., Twiss, M.R., Carrick, H.J., Davis, T.W., Johengen, T.H., Gossiaux, D.,
Smith, D.E., Palladino, D., and Burtner, A. (2016). Winter–spring diatom production in
Lake Erie is an important driver of summer hypoxia. Journal of Great Lakes Research
42, 608-618.
Riedel, G.F., and Nelson, D.M. (1985). Silicon uptake by algae with no kown Si requirement. II.
Strong pH dependence of uptake kinetic parameters in Phaeodactylum tricornutum
(Bacillariophyceae) 1. Journal of Phycology 21, 168-171.
Sandrini, G., Tann, R.P., Schuurmans, J.M., Van Beusekom, S.A., Matthijs, H.C., and Huisman,
J. (2016). Diel variation in gene expression of the CO2-concentrating mechanism during
a harmful cyanobacterial bloom. Frontiers in Microbiology 7, 551.
Saxton, M.A., D'souza, N.A., Bourbonniere, R.A., Mckay, R.M.L., and Wilhelm, S.W. (2012).
Seasonal Si: C ratios in Lake Erie diatoms—evidence of an active winter diatom
community. Journal of Great Lakes Research 38, 206-211.
Shapiro, J. (1990). Current beliefs regarding dominance by blue-greens: the case for the
importance of CO2 and pH. Internationale Vereinigung für theoretische und angewandte
Limnologie: Verhandlungen 24, 38-54.
Shruthi, M., and Rajashekhar, M. (2014). Effect of salinity and pH on the growth and biomass
production in the four species of estuarine cyanobacteria. J. Algal Biomass Utln 5, 29-36.
Steffen, M.M., Belisle, B.S., Watson, S.B., Boyer, G.L., Bourbonniere, R.A., and Wilhelm, S.W.
(2015). Metatranscriptomic evidence for co-occurring top-down and bottom-up controls
on toxic cyanobacterial communities. Appl. Environ. Microbiol. 81, 3268-3276.
Steffen, M.M., Belisle, B.S., Watson, S.B., Boyer, G.L., and Wilhelm, S.W. (2014). Status,
causes and controls of cyanobacterial blooms in Lake Erie. Journal of Great Lakes
Research 40, 215-225.
Sullivan, C. (1977). Diatom mineralization of silicic acid. II. Regulation of Si (OH) 4 transport
rates during the cell cycle of Navicula pelliculosa 1. Journal of Phycology 13, 86-91.
Suttle, C.A., and Fuhrman, J.A. (2010). Aquatic Viral Ecology. Manual of Aquatic Viral
Ecology, 145.
Taraldsvik, M., and Myklestad, S. (2000). The effect of pH on growth rate, biochemical
composition and extracellular carbohydrate production of the marine diatom Skeletonema
costatum. European Journal of Phycology 35, 189-194.
Thamatrakoln, K., and Hildebrand, M. (2008). Silicon uptake in diatoms revisited: a model for
saturable and nonsaturable uptake kinetics and the role of silicon transporters. Plant
Physiology 146, 1397-1407.
Twiss, M., Mckay, R., Bourbonniere, R., Bullerjahn, G., Carrick, H., Smith, R., Winter, J.,
D'souza, N., Furey, P., and Lashaway, A. (2012). Diatoms abound in ice-covered Lake

97
 Erie: An investigation of offshore winter limnology in Lake Erie over the period 2007 to
2010. Journal of Great Lakes Research 38, 18-30.
Vanderploeg, H.A., Liebig, J.R., Carmichael, W.W., Agy, M.A., Johengen, T.H., Fahnenstiel,
G.L., and Nalepa, T.F. (2001). Zebra mussel (Dreissena polymorpha) selective filtration
promoted toxic Microcystis blooms in Saginaw Bay (Lake Huron) and Lake Erie.
Canadian Journal of Fisheries and Aquatic Sciences 58, 1208-1221.
Vardi, A., Thamatrakoln, K., Bidle, K.D., and Falkowski, P.G. (2009). Diatom genomes come of
age. Genome Biology 9, 245.
Verspagen, J.M., Van De Waal, D.B., Finke, J.F., Visser, P.M., Van Donk, E., and Huisman, J.
(2014). Rising CO2 levels will intensify phytoplankton blooms in eutrophic and
hypertrophic lakes. PloS One 9.
Vrieling, E.G., Gieskes, W., and Beelen, T.P. (1999). Silicon deposition in diatoms: control by
the pH inside the silicon deposition vesicle. Journal of Phycology 35, 548-559.
Watanabe, M., Kawachi, M., Hiroki, M., and Kasai, F. (2000). NIES-Collection List of strains.
Microalgae and protozoa. 159 pp. National Institute for Environmental Studies,
Environment Agency, Tsukuba.
Wells, M.L., Karlson, B., Wulff, A., Kudela, R., Trick, C., Asnaghi, V., Berdalet, E., Cochlan,
W., Davidson, K., and De Rijcke, M. (2020). Future HAB science: Directions and
challenges in a changing climate. Harmful algae 91, 101632.
Welschmeyer, N.A. (1994). Fluorometric analysis of chlorophyll a in the presence of chlorophyll
b and pheopigments. Limnology and Oceanography 39, 1985-1992.
Werner, D. (1966). Die kieselsäure im stoffwechsel von Cyclotella cryptica Reimann, Lewin und
Guillard. Archiv für Mikrobiologie 55, 278-308.
Wilhelm, S., Bullerjahn, G., and Rlm, M. (2020). The complicated and confusing ecology of
Microcystis blooms. mBio.
Wilhelm, S.W. (2017). CT media. protocols.io.
Wilson, A., Kinney, J.N., Zwart, P.H., Punginelli, C., D'haene, S., Perreau, F., Klein, M.G.,
Kirilovsky, D., and Kerfeld, C.A. (2010). Structural determinants underlying
photoprotection in the photoactive orange carotenoid protein of cyanobacteria. Journal of
Biological Chemistry 285, 18364-18375.
Zepernick, B.N., Krausfeldt, L.E., and Wilhelm, S.W. (2020). Flaming as part of aseptic
technique increases CO2(g) and decreases pH in freshwater culture media. Limnology
and Oceanography: Methods.
Zepernick, B.N., Saxton, M.A., and Wilhelm, S.W. (2019). "Quantifying biogenic silica (bSi)
deposition rates adapted method", in: Protocols.io ).

98
Appendix
Table 3.1: Statistical analysis of F. crotonensis growth rate in pH mono and co-culture assays.
Statistical analyses: ordinary two-way ANOVA with Tukey’s HSD multiple comparisons test.

Treatment comparison Summary P value

pH 7.7 10:1 vs. pH 7.7 1:1 ** 0.0098
pH 7.7 10:1 vs. pH 7.7 1:10 ns 0.0815
pH 7.7 10:1 vs. pH 7.7. mc ns 0.1696
pH 7.7 10:1 vs. pH 9.2 10:1 ns 0.9994
pH 7.7 10:1 vs. pH 9.2 1:1 ns 0.7142
pH 7.7 10:1 vs. pH 9.2 1:10 ns 0.1053
pH 7.7 10:1 vs. pH 9.2 mc * 0.0378
pH 7.7 1:1 vs. pH 7.7 1:10 ns 0.9450
pH 7.7 1:1 vs. pH 7.7 mc ns 0.7800
pH 7.7 1:1 vs. pH 9.2 10:1 ** 0.0037
pH 7.7 vs. 1:1 vs. pH 9.2 1:1 ns 0.2063
pH 7.7 1:1 vs. pH 9.2 1:10 **** <0.0001
pH 7.7 1:1 vs. pH 9.2 mc **** <0.0001
pH 7.7 1:10 vs. pH 7.7 mc ns 0.9998
pH 7.7 1:10 vs. pH 9.2 10:1 * 0.0316
pH 7.7 1:10 vs. pH 9.2 1:1 ns 0.7745
pH 7.7 1:10 vs. pH 9.2 1:10 *** 0.0002
pH 7.7 1:10 vs. pH 9.2 mc **** <0.0001
pH 7.7 mc vs. pH 9.2 10:1 ns 0.0698
pH 7.7 mc vs. pH 9.2 1:1 ns 0.9423
pH 7.7 mc vs. pH 9.2 1:10 *** 0.0005
pH 7.7 mc vs. pH 9.2 mc *** 0.0002
pH 9.2 10:1 vs. pH 9.2 1:1 ns 0.4212
pH 9.2 10:1 vs. pH 9.2 1:10 ns 0.2443
pH 9.2 10:1 vs. pH 9.2 mc ns 0.0965
pH 9.2 1:1 vs. pH 9.2 1:10 ** 0.0045
pH 9.2 1:1 vs. pH 9.2 mc ** 0.0015
pH 9.2 1:10 vs. pH 9.2 mc ns 0.9990

99
Table 3.2: Statistical analysis of M. aeruginosa growth rate in pH mono and co-culture assays.
Statistical analyses: ordinary two-way ANOVA with Tukey’s HSD multiple comparisons test.

Treatment comparison Summary P value

pH 7.7 10:1 vs. pH 7.7 1:1 ns 0.8850
pH 7.7 10:1 vs. pH 7.7 1:10 ns 0.3390
pH 7.7 10:1 vs. pH 7.7 mc * 0.0175
pH 7.7 10:1 vs. pH 9.2 10:1 ns 0.5031
pH 7.7 10:1 vs. pH 9.2 1:1 ns 0.1299
pH 7.7 10:1 vs. pH 9.2 1:10 ns 0.1243
pH 7.7 10:1 vs. pH 9.2 mc ** 0.0018
pH 7.7 1:1 vs. pH 7.7 1:10 ns 0.9643
pH 7.7 1:1 vs. pH 7.7 mc ns 0.1893
pH 7.7 1:1 vs. pH 9.2 10:1 ns 0.9954
pH 7.7 1:1 vs. pH 9.2 1:1 ns 0.7273
pH 7.7 1:1 vs. pH 9.2 1:10 ns 0.7129
pH 7.7 1:1 vs. pH 9.2 mc * 0.0231
pH 7.7 1:10 vs. pH 7.7 mc ns 0.6945
pH 7.7 1:10 vs. pH 9.2 10:1 ns >0.9999
pH 7.7 1:10 vs. pH 9.2 1:1 ns 0.9982
pH 7.7 1:10 vs. pH 9.2 1:10 ns 0.9977
pH 7.7 1:10 vs. pH 9.2 mc ns 0.1504
pH 7.7 mc vs. pH 9.2 10:1 ns 0.5118
pH 7.7 mc vs. pH 9.2 1:1 ns 0.9527
pH 7.7 mc vs. pH 9.2 1:10 ns 0.9581
pH 7.7 mc vs. pH 9.2 mc ns 0.9325
pH 9.2 10:1 vs. pH 9.2 1:1 ns 0.9794
pH 9.2 10:1 vs. pH 9.2 1:10 ns 0.9761
pH 9.2 10:1 vs. pH 9.2 mc ns 0.0876
pH 9.2 1:1 vs. pH 9.2 1:10 ns >0.9999
pH 9.2 1:1 vs. pH 9.2 mc ns 0.3813
pH 9.2 1:10 vs. pH 9.2 mc ns 0.3943

100
Table 3.3: Statistical analysis of in vitro F. crotonensis Si deposition assay. (A) Comparison of F.
crotonensis cell concentration as a function of pH. Statistical analyses performed using unpaired
two-tailed t-test (t=6.705, df=8, n=5). (B) Comparison of total Si deposited as a function of pH.
Statistical analyses performed using unpaired two-tailed t-test (t=8.544, df=8, n=5). (C)
Comparison of Si deposited per filament as a function of pH. Statistical analyses performed using
unpaired two-tailed t-test (t=9.446, df=8, n=5). (D) Comparison of epifluorescence microscopy F.
crotonensis cell number per filament and % of cells fluorescing PDMPO per filament as a function
of pH. Statistical analyses performed using unpaired two-tailed t-test: #cells/filament: (t=4.057,
df=197, n=100), %PDMPO cells/filament: (t=9.457, df=197, n=100).

Treatment comparison Significance? Summary P value

pH 7.7 1:10 vs. pH 7.7 1:1 No ns 0.9996
pH 7.7 1:10 vs. pH 7.7 10:1 No ns 0.1853
pH 7.7 1:10 vs. pH 7.7. 0:1 Yes * 0.0300
pH 7.7 1:10 vs. pH 9.2 1:10 No ns 0.6941
pH 7.7 1:10 vs. pH 9.2 1:1 Yes * 0.0160
pH 7.7 1:10 vs. pH 9.2 10:1 Yes **** <0.0001
pH 7.7 1:10 vs. pH 9.2 0:1 No ns 0.0819
pH 7.7 1:1 vs. pH 7.7 10:1 No ns 0.3801
pH 7.7 1:1 vs. pH 7.7 0:1 Yes * 0.0118
pH 7.7 1:1 vs. pH 9.2 1:10 No ns 0.4156
pH 7.7 vs. 1:1 vs. pH 9.2 1:1 Yes ** 0.0063
pH 7.7 1:1 vs. pH 9.2 10:1 Yes **** <0.0001
pH 7.7 1:1 vs. pH 9.2 0:1 No ns 0.1881
pH 7.7 10:1 vs. pH 7.7 0:1 Yes *** 0.0002
pH 7.7 10:1 vs. pH 9.2 1:10 Yes ** 0.0080
pH 7.7 10:1 vs. pH 9.2 1:1 Yes **** <0.0001
pH 7.7 10:1 vs. pH 9.2 10:1 Yes ** 0.0012
pH 7.7 10:1 vs. pH 9.2 0:1 No ns 0.9996
pH 7.7 0:1 vs. pH 9.2 1:10 No ns 0.4860
pH 7.7 0:1 vs. pH 9.2 1:1 No ns >0.9999
pH 7.7 0:1 vs. pH 9.2 10:1 Yes **** <0.0001
pH 7.7 0:1 vs. pH 9.2 0:1 Yes **** <0.0001
pH 9.2 1:10 vs. pH 9.2 1:1 No ns 0.3180
pH 9.2 1:10 vs. pH 9.2 10:1 Yes **** <0.0001
pH 9.2 1:10 vs. pH 9.2 0:1 Yes ** 0.0032
pH 9.2 1:1 vs. pH 9.2 10:1 Yes **** <0.0001
pH 9.2 1:1 vs. pH 9.2 0:1 Yes **** <0.0001
pH 9.2 10:1 vs. pH 9.2 0:1 Yes ** 0.0030

101
Table 3.4: Statistical analysis of in situ Lake Erie Si deposition assay. (A) Comparison of
community chlorophyll a concentration as a function of pH. Statistical analyses performed using
ordinary one-way ANOVA with Tukey’s multiple comparisons test. (B) Comparison of total Si
deposited as a function of pH. Statistical analyses performed using ordinary one-way ANOVA
with Tukey’s multiple comparisons test. (C) Comparison of Si deposited per chlorophyll a
concentration as a function of pH. Statistical analyses performed using ordinary one-way
ANOVA with Tukey’s multiple comparisons test.

A.) Treatment comparison Summary P value

control 48 hr vs. pH 7.7 48 hr chlorophyll a ns 0.7187
pH 7.7 48 hr vs. pH 9.2 48hr chlorophyll a ns 0.4464
pH 9.2 48 hr vs. control 48hr chlorophyll a ns 0.9614

B.) Treatment comparison Summary P value

Lake Erie diatom control vs. pH 7.7 Net Si ns 0.916
Lake Erie diatom pH 7.7 vs. pH 9.2 Net Si ns 0.127
Lake Erie diatom pH 9.2 vs. control Net Si ns 0.229

C.) Treatment comparison Summary P value

Lake Erie diatom control vs. pH 7.7 Si/chl a ns 0.285
Lake Erie diatom pH 7.7 vs. pH 9.2 Si/chl a * 0.037
Lake Erie diatom 9.2 vs. control Si/chl a ns 0.404

102
Figure 3.7 Dissolved silica profiles corresponding to a 2015 Lake Erie M. aeruginosa bloom in
the western basin. Silica-limiting concentrations (defined in this study as 10µM) are indicated by
the dotted black line. Complete details of this dataset can be found in (Chaffin et al., 2018).

103
Figure 3.8 Total chlorophyll a concentration corresponding to a 2015 Lake Erie M. aeruginosa
bloom in the western basin. Complete details of this dataset can be found in (Chaffin et al.,
2018).

104
Figure 3.9 FlowJo graph depicting a gated population of F. crotonensis filaments run on the
flow cytometer. Sample analyzed was a pH 7.7 F. crotonensis monoculture culture from 20-d.
Background noise visible in lower graph quadrant results from CT media.

105
Figure 3.10 Standard curve for in vitro F. crotonensis Si deposition assay used to obtain silica
deposited per cell reported in Figure 5. Simple linear regression (R2=0.9999, F=53417,
p<0.0001, n=5).

106
Figure 3.11 Standard curve for in situ Lake Erie Si deposition assay used to obtain silica
deposited per chlorophyll a concentration reported in Figure 6. Simple linear regression
(R2=0.9995, F=7368, p<0.0001, n=6).

107
Figure 3.12 (A) In vitro M. aeruginosa monoculture growth curves at pH 7.7 (black circles) and
pH 9.2 (green circles). (B) M. aeruginosa growth rate at pH 7.7 (black circles) and pH 9.2 (green
circles). Statistically significant differences between pH treatments are denoted by p values
generated by Two-way ANOVAs. Standard error of the mean reported by error bars.

108
Figure 3.13 (A) In vitro M. aeruginosa growth curves in a 10:1 ratio (F. crotonensis:M.
aeruginosa) at pH 7.7 (black circles) and pH 9.2 (green circles). (B) M. aeruginosa growth rate
in a 10:1 ratio (C) M. aeruginosa growth curves in a 1:1 ratio (D) M. aeruginosa growth rate in a
1:1 ratio (E) M. aeruginosa growth curves in a 1:10 ratio (F) M. aeruginosa growth rate in a 1:10
ratio. Statistically significant differences between pH treatments are denoted by p values
generated by Two-way ANOVAs. Standard error of the mean reported by error bars.

109
Figure 3.14 pH drift in F. crotonensis monocultures inoculated at pH 7.7 (black inverted
triangles) and pH 9.2 (green inverted triangles). Standard error of the mean reported by error
bars.

110
Figure 3.15 pH drift from initial pH in M. aeruginosa monocultures inoculated at pH 7.7 (black
inverted triangles) and pH 9.2 (green inverted triangles). Standard error of the mean reported by
error bars.

111
Figure 3.16 pH drift from initial pH in co-cultures inoculated at pH 7.7 (black inverted triangles)
and pH 9.2 (green inverted triangles). Standard error of the mean reported by error bars.

112
Figure 3.17 Final pH of all mono and co-cultures after 30-d. (A) Final pH of pH 7.7. treatment
cultures (black) of F. crotonensis monocultures (squares), M. aeruginosa monocultures (circles)
and co-cultures (triangles). (B) Final pH of pH 9.2 treatment cultures (green) of F. crotonensis
monocultures (squares), M. aeruginosa monocultures (circles) and co-cultures (triangles). Initial
pH levels at 0-d are indicated by dotted lines. Standard error of the mean reported by error bars.

113
Figure 3.18 Simple linear regressions of M. aeruginosa concentration and pH drift of mono (A)
and co-cultures (B, C, D) inoculated at pH 7.7. Cell concentrations and pH levels were sampled
at 0, 10, 20, and 30 d. (A) Simple linear regression: R2=0.9732, F=363.4,p<0.0001, n=12, (B)
Simple linear regression: R2=0.9497, F=188.9,p<0.0001,n=12, (C) Simple linear regression:
R2=0.8943, F=84.60, p<0.0001, n=12, (D) Simple linear regression: R2=0.8577, F=60.29,
p<0.0001, n=12).

114
 A

C D

Figure 3.19 Epifluorescent microscopy images (10x magnification) of in vitro F. crotonensis

filaments after 48 h PDMPO incubations. Scale bar represents 75 𝜇𝜇m (depicted in panel A).
Chlorophyll a autofluorescence is depicted in red, and PDMPO fluorescence is indicated in blue.
(A, C) F. crotonensis chains acclimated to pH 7.7. (B, D) F. crotonensis filaments acclimated pH
9. 2.

115
Figure 3.20 Epifluorescent microscopy data of F. crotonensis Si deposition assay after 48 h
PDMPO incubations. (A) Number of cells per filament in pH 7.7 acclimated (black squares) and
pH 9.2 acclimated (green squares) F. crotonensis cultures. (B) Percentage of cells fluorescing
PDMPO per filament in pH 7.7 acclimated (black squares) and pH 9.2 acclimated (green
squares) F. crotonensis cultures. Statistically significant differences are denoted by respective p
values generated by unpaired two-tailed t-tests. Standard error of the mean reported by error
bars.

116
Figure 3.21 Total silica deposited per F. crotonensis culture after 48 h PDMPO incubations in
pH 7.7 (black squares) and pH 9.2 (green squares) acclimated cultures. Data corresponds to in
vitro F. crotonensis Si deposition assay. Statistically significant differences are denoted by
respective p values generated by unpaired two-tailed t-tests. Standard error of the mean reported
by error bars.

117
Figure 3.22 Total silica deposited per sample after 48 h PDMPO incubations in pH 7.7 (black
squares), pH 8.6 (grey squares) and pH 9.2 (green squares) treatments. Data corresponds to in
situ Lake Erie Si deposition assay. Statistically significant differences are denoted by respective
p values generated by One-way ANOVAs. Standard error of the mean reported by error bars.

118
Figure 3.23: Water column profiles collected during a 2-day period (February 17-19) of a 2009
winter cruise in the western-central basins of Lake Erie. Sampling stations correspond to the
following: Station 357, 341, 452 (western basin), station 1053, 84 (central basin). (A) Mean
temperature per station. (B) Mean pH per station. (C) Mean chlorophyll a concentration per
station. Standard error of the mean reported by error bars.

119
CHAPTER IV: INVESTIGATING ELEVATED PH EFFECTS ON FRESHWATER
DIATOM TRANSCRIPTION, MORPHOLOGY AND PHYSIOLOGY

120
Publication Note
This chapter is a version of a peer-reviewed, published article in Frontiers in Microbiology
13(2022) by Brittany N. Zepernick, David J. Niknejad, Gwendolyn F. Stark, Alexander R.
Truchon, Robbie M. Martin, Karen L. Rossignol, Hans W. Paerl, and Steven W. Wilhelm.

All culture work, pH assays, and sample collection performed by BZ and DN. FlowCAM
analyses performed by BZ, DN. PhytoPAM measurements performed by GS. Photopigment
extraction and HPLC analysis performed by KR, HP. RNA extractions and QC performed by
BZ. Transcriptome analyses performed by BZ, AT, RM. PRIMER analyses performed by BZ.

121
Abstract
Harmful algal blooms (HABs) caused by the toxin-producing cyanobacteria Microcystis

spp., can increase water column pH. While the effect(s) of these basified conditions on the bloom

formers are of a high research priority, how these pH shifts affect other biota remains

understudied. Recently, it was shown these high pH levels decrease growth and Si deposition

rates in the freshwater diatom Fragilaria crotonensis and natural Lake Erie diatom populations.

However, the physiological mechanisms and transcriptional responses of diatoms associated with

these observations remain to be documented. Here, we examined F. crotonensis with a set of

morphological, physiological, and transcriptomic tools to identify cellular responses to high pH.

We suggest 2 potential mechanisms that may contribute to morphological and physiological pH

effects observed in F. crotonensis. Moreover, we identified a significant upregulation of mobile

genetic elements in the F. crotonensis genome which appears to be an extreme transcriptional

response to this abiotic stress to enhance cellular evolution rates– a process we have termed

“genomic roulette”. We discuss the biological and biogeochemical effects high pH conditions

impose on fresh waters and suggest a means by which freshwater diatoms such as F. crotonensis

may evade high pH stress to survive in a “basified” future.

122
Introduction
Algal blooms are a symptom of an imbalanced ecosystem (Heisler et al., 2008; Watson et

al., 2015), where both the biotic and abiotic characteristics of a lake’s water column are altered

(Anderson, 2009; Gobler and Sunda, 2012). In the case of freshwater Microcystis spp.-dominated

harmful algal blooms (HABs), nutrient drawdown, oxygen depletion and increased light

attenuation are well-documented consequences (Paerl et al., 2001; Verspagen et al., 2004;

Lehman et al., 2013; Zepernick et al., 2022d). The effects of Microcystis blooms on water

column pH serve as a recent addition to this growing list of consequences (Van Dam et al.,

2018; Krausfeldt et al., 2019; Turner et al., 2021; Zepernick et al., 2021), raising the question

how elevated pH levels associated with blooms influence other biota. Research has demonstrated

Microcystis spp. blooms can increase lake pH to well above 9.0 via the photosynthetic depletion

of CO2 (Verspagen et al., 2014; Ji et al., 2020), a phenomenon recently termed “lake

basification” (Zagarese et al., 2021; Zepernick et al., 2021). Basification events have been

recorded in fresh waters including Lake Taihu, China and Lake Erie, U.S./Canada (Su et al.,

2015; Wilhelm et al., 2020). In Lake Erie, the mean daily water column pH remained ≥ 9.2 for

~30 days during a record-breaking 2015 Microcystis spp. bloom (Zepernick et al., 2021).

Further, these pH spikes oscillated on a diel cycle, with the highest pH levels (as much as 0.5

units above ambient) coinciding with peak photosynthetic periods in the late afternoon

(Krausfeldt et al., 2019). While an increase in pH may benefit Microcystis spp. (Sandrini et al.,

2016; Krausfeldt et al., 2019) and function as a positive feedback loop for late-stage bloom

maintenance (Tang et al., 2018), it serves as a potential detriment to other organisms.

A recent study suggested elevated pH conditions have the potential to negatively affect

algal communities beyond pH-induced carbon limitation (Zagarese et al., 2021). In a previous

investigation, we demonstrated that the freshwater diatom Fragilaria crotonensis, which

123
historically bloomed during the summer in Lake Erie (Hartig, 1987), exhibited lower growth

rates and silica (Si) deposition rates at pH 9.2 in both monoculture and co-culture with M.

aeruginosa (Zepernick et al., 2021). That study identified factors which likely contributed to

observed declines of diatom populations during HAB events, yet there remains a need for a

comprehensive assessment of how these high pH conditions influence freshwater diatom

morphology and physiology beyond growth and Si deposition rates.

In the present study, we used transcriptomics to generate hypotheses concerning the

physiological response and mechanistic changes behind high-pH-induced effects in this

freshwater diatom. We then performed morphological and physiological measurements to

validate observed transcriptomic responses and quantitatively assess the effects of elevated pH

conditions on F. crotonensis.

Methods
Culture conditions

To investigate the morphological and physiological effects high pH may impose on

freshwater diatoms, non-axenic monocultures of F. crotonensis SAG 28.96 were acclimated for

6 d to either their optimal growth pH of 7.7 (Hervé et al., 2012) or the simulated Lake Erie

Microcystis bloom-induced basification pH of 9.2 (Krausfeldt et al., 2019), as described

previously (Zepernick et al., 2021). After a 6 d acclimation, samples were filter-concentrated and

inoculated (T initial - Ti) into their respective pH treatments for the pH assay. All pH assays

were inoculated at ~1,500 filaments mL-1 in this study except for the photopigment assay, as this

method required higher biomass. These samples were incubated for an additional 2 d at

conditions consistent with the Lake Erie summer water column (26° C; light intensity ~55-60

µmol photons m-2 s-1 on a 12:12 h light dark cycle) prior to sample collection on day 8 of pH

124
exposure (T final - Tf). In this study, the treatment pH of 9.2 will be referred to as “high pH”

while the control treatment pH of 7.7 will be referred to as “low pH”.

pH effects on F. crotonensis transcription

To assess how transcriptional activity was affected at high pH and generate preliminary

hypotheses based on transcriptional findings, F. crotonensis cultures were inoculated at ~1,500

filaments mL-1 into the respective pH treatments (n = 3). At Tf, each replicate was collected on a

2.0-𝜇𝜇m nominal pore-size 47-mm diameter polycarbonate filter to concentrate diatom biomass.

Samples were flash frozen in liquid nitrogen and stored at -80° C until extraction. RNA

extractions were performed using acid phenol-chloroform methods with ethanol precipitation

(Martin and Wilhelm, 2020). Residual DNA in samples was digested using a modified version of

the Turbo DNase protocol and the Turbo DNA-free kit (Ambion). Removal of genomic DNA

was confirmed via the absence of an amplicon band in an agarose gel after 30 cycles of PCR

amplification using 519F/785R 16s rRNA primers as reported previously (Zepernick et al.,

2022a). Final RNA concentrations were determined using the HS Qubit RNA assay

(Invitrogen). Sample library prep (Poly-A selection) and Illumina NovaSeq 6000 platform

sequencing (~25 million reads, 100bp, paired end) were performed at Hudson Alpha (Discovery

Life Sciences). Sequencing data was interleaved, filtered, and trimmed using CLC Genomics

Workbench default settings (v.20) (Qiagen Digital Insights). The quality of trimmed reads was

confirmed using FastQC (v.0.11.9) (Babraham Institute). BBMap.sh (default settings) was used

to remove residual rRNA reads and BBMap.repair (default settings) was used to validate paired-

end reads. (Bushnell, 2014a). Sorted reads were mapped to the annotated F. crotonensis

reference genome (Zepernick et al., 2022c) in CLC (default settings: length fraction: 0.5,

125
similarity fraction 0.8) (Appendix Attachment 4.1a), and normalized to transcripts per million

(TPM). To calculate similarity (Bray-Curtis) and identify contributors to gene-expression

differences between samples, non-metric multi-dimensional scaling (nMDS) and Similarity

Percentage (SIMPER) analyses were performed on normalized expression values (TPM) using

PRIMER (v.7) (Clarke and Gorley, 2015). Differential Expression (DE) analyses were

performed in CLC, and the results were stringently filtered by significance (FDR-corrected p-

value ≤ 0.05, log2 |fold-change| > 2), with predicted genes of hypothetical or unknown functions

omitted from downstream analyses. Heat maps were constructed via heatmapper.ca (Clustering

method: Average linkage, Distance measurement method: Pearson) (Babicki et al., 2016) using

standardized expression scores in which genes were grouped using gene descriptions based on

EggNOG and COG categories pre-assigned by the EggNOG annotation database (Huerta-Cepas

et al., 2019). Manual categorization of genes was further performed based on KEGG Mapper and

the KEGG Orthology (KO) database (Kanehisa et al., 2016).

pH effects on F. crotonensis morphology

To directly test the morphological effects of high pH hinted at within our transcriptomes,

F. crotonensis cultures were inoculated at ~1,500 filaments mL-1 into their respective pH

treatments (n = 5). At Tf, filament morphology was assessed with a FlowCAM 8000 imaging

system using the 10x objective with a particle per used image < 1.7 (PPUI) (FlowCAM 8000,

Fluid Imaging Technologies, Yokogawa Electric Corporation). Briefly, ~1,000 F. crotonensis

filaments per biological replicate were individually analyzed using the automated functions of

the instrument with the following parameters measured: area (µm2), biovolume (µm3), length

(µm), width (µm), roughness, and average green content.

126
pH effects on F. crotonensis photopigment composition

To test for photopigment changes hypothesized from the transcriptome, F. crotonensis

cultures were inoculated at ~20,000 filaments mL-1 into their respective pH treatments (n = 5).

At Tf, measurements were made for total filaments mL-1 by flow cytometry (Beckman Coulter

CytoFLEX S, equipped with the blue laser (488nm, 50 mW), and red laser (638 nm, 50 mW)

with populations gated on PerCP and FSC-H. Chlorophyll a (Chl a) autofluorescence was

measured using a Turner Designs TD-700 fluorometer, equipped with a “blue” mercury bulb, a

#10-050R excitation filter (λ = 340-500nm), and a #10-115 (λ = 680 nm) emission filter. A solid

standard (Turner Designs #7000-994) was used to calibrate the fluorometer prior to each day’s

measurements. Subsequently, 20 mL of each culture was collected on 47-mm diameter GF/F

(Whatman) filters for pigment extraction. Samples were stored at -20° C prior to pigment

extraction and high-performance liquid chromatography (HPLC) analysis at University of North

Carolina Chapel Hill-Institute of Marine Sciences as described previously (Paerl et al., 2014).

Briefly, photopigment samples were extracted in 100% acetone, sonicated, and stored at -20° C

for ~24 h. The extracts (200 µL) were next assessed via HPLC as described previously (Van

Heukelem et al., 1994; Pinckney et al., 1996; Pinckney et al., 1998; Pinckney et al., 2001).

Photopigments were identified based on absorption spectra, which were determined from

commercially obtained pigment standards (DHI, Denmark). Net pigment concentrations (µg L-1)

were normalized to Tf filament counts (filaments L-1) and mean filament length (µm).

pH effects on F. crotonensis photosynthetic physiology

To further evaluate the potential effects of pH on photosynthetic physiology derived from

our transcriptomes, F. crotonensis cultures were inoculated at ~1,500 filaments mL-1 into the

127
respective pH treatments (n = 5). At Tf, photosynthetic efficiency metrics were assessed via

Pulse-Amplitude-Modulation (PAM) fluorometry with the brown algae (diatom) taxa setting

(Phyto-PAM-II Compact Version, WALZ). Samples were dark-acclimated for 20 min prior to

Phyto-PAM readings as described previously (Ritchie, 2008; Torres et al., 2014; Gleich et al.,

2020). After dark-acclimation, replicates were exposed to a white saturating light pulse (5,000

µmol m-2 s-1 PAR) prior to determination of the maximum theoretical photochemical quantum

yield of photosystem II (PSII) (Fv • Fm -1). Subsequently, samples were exposed to rapid light

curve measurements to determine the relative maximum rate of electron transport through PSII

(rETRmax, µmol electrons/m2 s1). Rapid light curves were run in fourteen steps. For each step

cultures were exposed to increasing actinic irradiances starting from 1 µmol m-2 s-1 PAR, until a

maximum of 1,257 µmol m-2 s-1 PAR was reached. The steps were run in 20 second intervals and

a saturating pulse of 5,000 µmol m-2 s-1 PAR was run after each step. The light intensity at

which saturation of PSII occurs (IK, µmol photons m-2 • s-1) was also determined.

Statistical analyses

Statistical analyses of data (Figures 4.2, 4.4, and 4.5) were performed in GraphPad Prism

(v.9.3.1) using unpaired, two-tailed t-tests. For this study, we consider a p-value < 0.05 to be

significant but report all values so the reader may decide their level of risk (Appendix

Attachment 4.1-4.4). Statistical comparisons of gene expression (Figure 4.1, 4.3, and 4.6) were

performed in CLC Genomics Workbench. All z-scores reported in heat maps were calculated by

Heatmapper.ca. Variability in expression between replicates was assessed via PRIMER.

128
Results
In this study, we performed transcriptomics to generate preliminary hypotheses regarding

how pH may affect the physiology and morphology of F. crotonensis. Our transcriptomes

identified 3 COG categories of genes which formed most of the differentially expressed (DE)

genes between pH treatments: 1.) Cell cycle control, cell division and chromosome partitioning,

2.) Energy production and conversion, 3.) Mobilome: transposons, prophages. Following,

laboratory assays were used to test hypotheses regarding morphology and physiology for these

main COG categories. Hence, our results are organized by COG category, with the transcriptome

data presented first, followed by the physical laboratory assay used to test the transcriptome-

derived hypotheses.

pH induced differential gene expression between treatments

nMDS revealed an 86% similarity among low pH replicates. In contrast, high pH

replicates exhibited more variability, with replicates 2 and 3 sharing 76% similarity, but replicate

1 appearing more like low pH replicates (Appendix Figure 4.7). SIMPER analyses determined a

mean dissimilarity of ~27% between low and high pH replicates and identified the contribution

of each gene to this variation. Notably, ~60% of the top 50 genes driving dissimilarity were

related to photosynthesis, including 17 copies of the chlorophyl a/b binding protein (LHCB3)

and 8 copies of fucoxanthin-chlorophyll ac binding protein (FCP) (Appendix Attachment 1b)

(Appendix Figure 4.8A). Of the ~26,000 predicted genes in the F. crotonensis genome, a total of

713 were differentially expressed (FDR-corrected p-value ≤ 0.05, log2 |fold-change| > 2)

(Appendix Figure 4.9A). Of these, 435 genes are annotated as either hypothetical or of unknown

function (Appendix Figure 4.9B). All downstream analyses focused on the 278 DE genes with

129
predicted function (Appendix Figure 4.9C) (Appendix Attachment 1c), of which 193 were

increased in representation at pH 9.2 and 85 decreased in representation.

High pH decreased carbohydrate transport and metabolism gene expression

Overall, genes involved in the COG category “Carbohydrate transport and metabolism”

decreased in representation at high pH relative to low pH (Appendix Figure 4.10). Genes

involved in the Calvin-Benson-Basham (CBB) cycle decreased in expression, including

glyceraldehyde 3-phosphate dehydrogenase (GAPDH), which catalyzes the single reductive step

of the CBB cycle during photosynthetic carbon fixation. Carbon metabolism genes involved in

cellular respiration also decreased in relative expression at high pH. Genes involved in the Krebs

cycle, such as malate/L-lactate dehydrogenase (MLDH) and divalent anion/Na+symporter

(DASS)(Takahashi-Íñiguez et al., 2016; Lu, 2019), decreased in representation at high pH.

Carbonic anhydrase (CA), which is responsible for the interconversion/acquisition of bicarbonate

and CO2 (Sültemeyer, 1998; Burkhardt et al., 2001), decreased in expression at high pH.

High pH increased expression of cell cycle control & cell wall biogenesis genes

The relative expression of genes categorized in “Cell cycle control, cell division and

chromosome partitioning” increased at pH 9.2 (Figure 4.1A). Proportional transcript abundance

increased in two copies of the ribosomal RNA processing protein (RRP15) at high pH, which has

been found to activate the G1 /S checkpoint in cancer cells and thus inducing cellular arrest in G1

130
Figure 4.1 Heat maps depicting differentially expressed genes relating to F. crotonensis growth
and cell wall morphology. (A) Genes in the COG category “Cell cycle control, cell division and
chromosome partitioning”. (B) Genes in the COG category “Cell wall, membrane, envelope
biogenesis”. Cladogram clustering is to demonstrate similarity in expression. All TPM values
were row z-scored, with increases in proportional transcript abundance indicated in yellow, and
decreases in proportional transcript abundance indicated in blue. The sum of transcripts across all
treatments (LogTPM) is indicated for each gene.

131
stage of interphase (Dong et al., 2017). Expression levels of RI0 kinase 1 (RIOK1), which is

required to enter S phase (LaRonde-LeBlanc and Wlodawer, 2005), also increased at pH 9.2.

High pH increased the transcript abundance of butyrate response factor (ZFP36L), involved in

cellular senescence and shown to induce cell cycle arrest at the G1 phase (Saini et al., 2020).

Genes classified within the “Cell wall, membrane and envelope biogenesis” category also

increased in overall expression (Figure 4.1B). A decrease in the representation of

Tetratricopeptide repeat (TPR) was observed at high pH, a gene that has previously been shown

to be positively correlated with silaffin expression (Frigeri et al., 2006). Further, we observed

decreases in the relative expression of N-acetylgalactosamine 3-beta-galactosyltransferase

(C1GALT1) and UDP glucuronate decarboxylase (UXS1) at high pH, both involved in

glycosaminoglycan biosynthesis which has been implicated in diatom biosilicification processes

(Alexander et al., 2015).

High pH shaped F. crotonensis filament morphology

FlowCAM analyses showed filaments grown at high pH had ~20% lower mean filament

area (p = 0.007) (Figure 4.2A) and mean biovolume (p = 0.031) (Figure 4.2B) than those grown

at low pH. These reductions in area and biovolume were likely due to decreased filament length

at high pH. Specifically, pH 9.2 F. crotonensis filaments had a 25% lower mean filament length

(p = 0.002) (Figure 4.2C), yet mean filament width was not significantly different (p = 0.055)

(Figure 4.2D). Additionally, pH 9.2 filaments were significantly rougher on the surface (p =

0.002) (Figure 4.2E) and less green in color (p < 0.0001) (Figure 4.2F). The exact number of

filaments assessed per biological replicate and. intra-variation of each replicate are reported in

Appendix Attachment 4.2.

132
Figure 4.2 FlowCAM pH assay results collected at (Tf). pH 7.7 replicates denoted by open,
black squares. pH 9.2 replicates denoted by open, green squares. Each data point represents the
mean of ~1000 filaments per biological replicate, with the group mean of treatment replicates
indicated by the central bar, and variability indicated by error bars representing the standard error
of the mean (SEM). Variance for individual dots is reported is supplemental dataset 2.
(A) Mean area (µm2) of F. crotonensis filaments. (B) Mean biovolume (µm3) of F. crotonensis
filaments. (C) Mean length (µm) of F. crotonensis filaments. (D) Mean width (µm) of F.
crotonensis filaments. (E) Mean roughness (dimensionless) of F. crotonensis filaments. (F)
Mean green coloration of F. crotonensis filaments (dimensionless). P-values are for
comparisons between treatments.

133
High pH decreased expression of energy production and conversion genes

Genes classified in the COG category “Energy production and conversion” were

analyzed within the subcategories “Photosynthesis” and “Cellular respiration” that were created

based on KEGG Mapper KO’s (Figure 4.3). Photosynthesis-related genes decreased in

representation at high pH (Figure 4.3A). All 5 copies of the PSII light harvesting complex III

chlorophyll a/b binding genes (LHCB) and 3 copies of the fucoxanthin-Chl binding complex

genes (FCP), both involved in light absorption and energy delivery during the first step of

photosynthesis (Ballottari et al., 2012), decreased in relative expression at high pH. In the

chlorophyll biosynthesis pathway, magnesium chelatase subunit D (chlD) and

NADPH:protochlorophyllide oxidoreductase (POR) increased in relative expression at pH 9.2,

while divinyl chlorophyllide a 8-vinyl-reductase (DVR) and bifunctional glutamyl/prolyl tRNA

synthetase (EPRS) decreased. In contrast, genes within the carotenoid biosynthesis pathway such

as Zeaxanthin epoxidase (ZEP) and violaxanthin de-epoxidase (VDE) decreased in

representation at high pH. Regarding photosynthetic metabolism, ferredoxin NADP+ reductase

(petH) and photosystem I subunit VII (psaC), which are both involved in the final step of

electron transfer from ferredoxin PSI to NADPH to fuel the Calvin cycle (Fischer et al., 1998;

Nguyen et al., 2021), decreased in relative expression at pH 9.2. Genes involved in oxidative

stress such as redoxin (PRDX5) (Hopkins and Neumann, 2019) and the catalase peroxidase gene

(katG.1), which quenches photosynthetically-produced ROS species (Nishiyama et al., 2001),

appeared to decrease overall at high pH.

In contrast, genes classified within “cellular respiration” exhibited an overall increase in

relative expression at pH 9.2 (Figure 4.3B). Representation of genes involved in oxidative

phosphorylation increased at pH 9.2, including NADH quinone oxidoreductase chain 5 and 6

134
Figure 4.3 Heat map depicting differentially expressed genes relating to energy production and
conversion. (A) Genes further sorted into photosynthesis and (B) cellular respiration categories.
All TPM values were row z-scored, with increases in proportional transcript abundance indicated
in yellow, and decreases in proportional transcript abundance indicated in blue. Cladogram
clustering is to demonstrate similarity in expression. The sum of transcripts across all treatments
(LogTPM) is indicated for each gene.

135
genes (ND5) (ND6) (Melo et al., 2004) and ATP synthase subunit α gene (ATP6) (Vázquez-

Acevedo et al., 2016). In contrast, mitochondrial transport genes decreased in relative expression

at high pH, including carrier gene (SLC25) (Ruprecht and Kunji, 2020).

High pH modified photopigment composition

After 8 d of pH exposure, there was no significant difference in filament concentration as

a function of pH treatment (p = 0.1745) (Appendix Figure 4.11A). However, high pH replicates

had significantly decreased in Chl a autofluorescence (p < 0.0001) (Appendix Figure 4.11B).

Normalized photopigment concentrations of Chlorophyllide a, Chlorophyll a, Chlorophyll c1c2

(fg • µm fil-1) (p ≥ 0.3263) (Figure 4.4A, B, C), and Total Chl a (fg • µm fil-1) (p = 0.3503)

(Appendix Figure 4.12) were not significantly different as a function of pH. In contrast, high pH

cultures had ~40% more β-carotene (p = 0.0015) (Figure 4.4D) and ~20% more Diadinoxanthin

(p = 0.0189) (Figure 4.4F) per cell compared to low pH replicates. Further, high pH incubated

cultures demonstrated ~30% less Violaxanthin (fg • µm fil-1) compared to pH 7.7 counterparts (p

= 0.0090) (Figure 4.4E), yet Fucoxanthin and Neoxanthin normalized pigments did not

significantly vary per cell as a function of pH (p ≥ 0.337) (Appendix Figure 4.13).

Ratios of total carotenoids:total Chl a were higher at pH 9.2 but fell short of significance

(p = 0.2207) (Appendix Figure 4.14). However, β-carotene:total Chl a (p = 0.0002) and

Diadinoxanthin:total Chl a (p = 0.0008) ratios were both significantly higher at pH 9.2

(Appendix Figure 4.15A, C). Mean Chl a/Chl c1c2 ratios, which serve as a proxy for the size of

the light harvesting antenna complex (Lamote et al., 2003; Nguyen-Deroche et al., 2012;

Heydarizadeh et al., 2019), demonstrated a consistent downward trend at high pH, though these

results were not significant (p = 0.4335) (Appendix Figure 4.16).

136
Figure 4.4 Photopigment pH assay results (n = 5) (fg • µm fil-1) collected at (Tf). pH 7.7
replicates denoted by open, black squares. pH 9.2 replicates denoted by open, green squares. The
group mean of treatment replicates indicated by the central bar and variability is indicated by
error bars representing the standard error of the mean (SEM). (A) Chlorophyllida a of F.
crotonensis filaments. (B) Chlorophyll a. (C) Chlorophyll c1c2. (D) β-Carotene. (E)
Violaxanthin. (F) Diadinoxanthin. P-values are for comparisons between treatments.

137
In summary, pH did not significantly affect chlorophyll pigment concentration per cell, but a

significant effect of pH 9.2 was observed on carotenoids (β-carotene and xanthophylls) in this

study.

High pH altered F. crotonensis photophysiology

High pH of 9.2 did not significantly alter optimal photochemical quantum yields of

photosystem II (Fv • Fm-1) (p = 0.3089) (Figure 4.5A). However, the relative mean maximum

electron transport rate through photosystem II (rETRmax) was ~50% lower at pH 9.2 (p < 0.0001)

(Figure 4.5B). Additionally, the photon flux at which light saturation of photosynthesis occurs

(Ik) was ~50% lower at pH 9.2 (p < 0.0001) (Figure 4.5C).

High pH increased expression of transposon genes

Out of the 278 DE genes in the dataset, 193 were overrepresented at the high pH

treatment. Further, of these 193 DE genes which increased in relative expression at pH 9.2, 25%

belong to the “Mobilome: transposons, prophages” COG category (Figure 4.6). All 48 genes

within the Mobilome COG category increased in representation at high pH. Notably, 9 copies of

the plant transposon gene (PTRP) were overrepresented at high pH. No Mobilome categorized

genes were decreased in relative expression at the high pH treatment within our DE dataset.

Discussion
HAB-induced increases in pH have been shown in the literature for decades (Talling,

1976; Jeppesen et al., 1990; Lopez-Archilla et al., 2004; Sandrini et al., 2016), yet these studies

largely focus on carbon chemistry. Few assess the consequences on other members of the

138
Figure 4.5 PhytoPAM pH assay results (n = 5) collected at (Tf). pH 7.7 replicates denoted by
open, black squares. pH 9.2 replicates denoted by open, green squares. The group mean of
treatment replicates indicated by the central bar and variability is indicated by error bars
representing the standard error of the mean (SEM). (A) Fv • Fm-1 of F. crotonensis filaments.
(B) rETRmax (µmol electrons/m2 • s1). (C) Photon flux at which light saturation of photosynthesis
occurs Ik (µmol photons m-2 • s-1). P-values are for comparisons between treatments.

139
Figure 4.6 Heat map depicting differentially expressed genes relating to the “Mobilome:
prophages and transposons. All TPM values were row z-scored, with increases in proportional
transcript abundance indicated in yellow, and decreases in proportional transcript abundance
indicated in blue. Cladogram clustering is to demonstrate similarity in expression. The sum of
transcripts across all treatments (LogTPM) is indicated for each gene.

140
biotic community. Previously, we demonstrated that growth of F. crotonensis decreased at pH

9.2 and discovered Si deposition declined at high pH in diatom cultures as well as natural Lake

Erie communities. Here, we built on this using transcriptomics to identify physiological

processes and resulting morphological changes that may be altered by high pH. Then, we

employed a variety of in vitro lab assays to better clarify and validate our transcriptomic results

and identified potential mechanisms by which these changes may be occurring. Finally, we

discuss these observations within the broader ecological scope of lake basification, and potential

future implications on global freshwater diatom communities.

Evidence of carbon limitation is lacking in high pH transcriptomes

Declines in diatom viability during bloom-induced basification have historically been

thought to be due to inorganic carbon-limitation. However, evidence of carbon limitation was

lacking in our transcriptomes. Recent studies investigating carbon-limitation in the model diatom

Phaeodactylm tricorntum observed increases in the expression of the CO2 concentrating

mechanism (CCM) genes (specifically CAs) in response to low CO2 availability (Burkhardt et

al., 2001; Heydarizadeh et al., 2019). However, there was an absence of biophysical CCM genes

in our DE transcript list except for one CA, which was decreased at pH 9.2. Additionally,

Heydarizadeh et al. (2019) concluded high light and carbon limitation increased the expression

of genes associated with the biochemical CCM, yet we saw no evidence of this occurring within

our transcriptomes. In addition, a prior study determined that environmental F. crotonensis

bloom samples incubated at carbon-limited pH 9.4 did not recover photosynthetic rates after CO2

enrichment compared to carbon-limited controls (Talling, 1976), suggesting F. crotonensis may

not have been exclusively carbon limited. Further, the F. crotonensis photosynthesis rates at pH

141
9.4 in both carbon-limited and carbon-enriched samples were lower when compared to general

phytoplankton photosynthetic rates (Talling, 1976). In contrast, when Talling (1976) replicated

the experiment with natural Microcystis spp. bloom samples, a significant recovery of

photosynthetic rate was observed after CO2 enrichment. These results from (Talling, 1976) and

others are consistent with our own; they suggest F. crotonensis is exhibiting decreased

photosynthetic metabolism not due to pH-induced carbon-limitation alone, but due directly to

alkaline pH. In summary, prior studies have induced C-limitation in diatoms and observed

alterations in carbon-related genes within their transcriptomes. In our study, we did not observe

these gene trends found to coincide with carbon-limitation, yet further research is needed

concerning this phenomenon.

Photostress is a likely physiological consequence of high pH in F. crotonensis

Previously, we demonstrated that growth and Si deposition rates decline at high pH. Yet,

the mechanisms driving these observations remained unclear. The present study indicated one of

the primary processes driving these effects may be photostress. Surprisingly, ~60% of the top 50

genes contributing to pH transcriptome dissimilarity are involved in photosynthesis, with PSII

and light antennae complex components forming the majority. Further, PhytoPAM data

suggested a significant decrease in PSII electron flow and light saturation threshold at high pH,

while pigment analysis indicated a significant increase in the photoprotective carotenoids β-

carotene and diadinoxanthin. Cumulatively, this data suggests F. crotonensis experienced

photostress at high pH.

Prior studies demonstrated that short-term photoacclimation strategies decrease

photosynthetic processes via LHC modifications (Horton et al., 1996; Bassi and Caffarri, 2000),

142
reductions in the size of the LHC (Perry et al., 1981), and initiation of nonphotochemical

quenching (NPQ) (Wilhelm et al., 2006a; Bertrand, 2010). In our study, F. crotonensis decreased

electron flow through PSII at high pH as indicated by significantly lower PSII rETRmax and

decreased expression of genes encoding for the LHC of PSII (LHCBs and FCPS) (Ballottari et

al., 2012). Additional transcriptomic data implied a corresponding decrease in PSI

photosynthetic capacity at high pH, as evidenced by decreases in ferredoxin NADP+

oxidoreductase (FNR) enzyme gene expression (petH and psaC). Cumulatively, these data

suggest there are alterations to photosynthetic capacity at high pH. In addition, decreases in

LHCB expression have been found to serve as a photoprotective response to excessive light

intensities and light saturation to prevent damage to PSII and reduce ROS generation (Thomas,

2016). In our study, this strategy appeared to be successful, as PhytoPAM data demonstrated no

significant differences in (Fv • Fm-1) as a function of pH. Regarding LHC modifications,

significant increases in diadinoxanthin photopigment composition at high pH, decreases in

carotenoid biosynthesis gene expression, and increases in expression of the key regulatory gene

of the chlorophyll biosynthesis pathway (chlD), collectively suggest F. crotonensis may be

restructuring the pigment composition of its LHCs in response to alkaline pH. Additionally,

mean Chl a/Chl c1c2 ratios, which serve as a size proxy for the LHC, exhibited a downward

trend at pH 9.2, though this was not significant. This suggested that after 8 d of alkaline pH

exposure, F. crotonensis may have begun reducing the size of its LHC as a photoacclimation

strategy. However, this would likely result in an increase in rETRmax as more light would then be

required to drive rETR saturation, which we do not observe in this study. Hence, it is probable

while F. crotonensis appears to be restructuring the photopigment constituents of its LHC, this

does not appear to alter the LHC size overall. Moreover, we observed significant increases in

143
diadinoxanthin pigment at pH 9.2 and increased representative expression of thylakoid

translocase subunit SecA gene (cpSecA.1), suggesting the activation of a NPQ energy dissipation

pathway which is controlled by diadinoxanthin and a trans-thylakoid proton gradient (Muller et

al., 2001; Bertrand, 2010). Indeed, the diadinoxanthin cycle has been described as the most

important short-term photoprotective mechanism in diatoms (Bertrand, 2010). Cumulatively,

these results indicate that F. crotonensis protects the photosynthetic membrane at high pH by

employing short-term photoacclimation mechanisms such as modifying LHC composition and

increasing the concentration of pigments involved in NPQ.

The photostress and photoresponse findings are supported by prior diatom studies. Park

et al. (2010) determined that Chaetoceros neogracile alters diadinoxanthin levels and FCP

expression in response to increased light intensity, and a recent study determined the diatom P.

tricornutum alters the photopigment composition of its FCP binding complexes in response to

varying light sources (Oka et al., 2020). Collectively, these prior studies suggest the high pH

effects we observed in our study mirror diatom responses to high light intensity. Yet, the light

intensity was held constant throughout the entirety of our experiments, implying that growth at

pH 9.2 affects F. crotonensis’s phototolerance. Indeed, in our study F. crotonensis filaments may

be light saturated at pH 9.2 as evidenced by a 50% lower PSII saturation threshold (Ik), with

saturation setting in at 60 µmol photons m-2 s-1. This suggests that F. crotonensis is experiencing

the light saturation of downstream metabolisms, with potential alterations to the light antennae

serving as a secondary effect of this phenomenon. Further, photoprotective pigments increased at

high pH while light harvesting pigments did not, implying filaments are prioritizing

photoprotection over photon acquisition. Cumulatively, transcriptomic, morphological, and

144
physiological evidence in our study imply F. crotonensis may have a lower phototolerance of

light at high pH, suggesting basification pH conditions alter diatom photophysiology.

Cell cycle arrest another physiological consequence of high pH?

High-pH-induced photostress explains altered photopigment composition and

photosynthetic physiology but falls short of directly accounting for morphological changes

observed in this study (i.e., differences in filament length and frustule roughness) and prior

observed physiological effects (i.e., decreased growth and Si deposition rates). Our

transcriptomic analyses indicate another factor may be driving these high pH effects: arrest of the

cell cycle at the G1/S checkpoint (Jang et al., 2005). Indeed, a prior study determined this

checkpoint is typically regulated by light in photosynthetic eukaryotes (Moulager et al., 2010),

thus photostress may be contributing to its arrest in the diatom F. crotonensis. In support of this,

we observed an increase in the expression of multiple genes associated with the G1/S checkpoint

at high pH. Moreover, a recent study demonstrated cell cycle arrest in G1 does not lead to

changes in Fv • Fm-1, but does result in lower rETRmax, higher NPQ, and gene expression

patterns consistent with LHC and ROS scavenging in the diatom P. tricornutum (Kim et al.,

2017), all of which are observed in our study.

However, if photostress is inducing cell cycle arrest, how does this result in lower growth

and Si deposition rates? With regard to growth, a prior diatom study indicated photosynthetic

metabolism and the cell cycle are closely related, with photosynthetic capacity at its highest

during the main growth phase of G1 (Claquin et al., 2004). Biosilicification is also linked to the

cell cycle and cell growth (Hildebrand et al., 2007; Shrestha et al., 2012), with the G1 phase

serving as the phase where diatoms reach their full size and where girdle band formation of the

145
frustules occurs (Javaheri et al., 2014). Additional studies suggest cell growth cannot occur

without girdle band formation (Crawford, 1981; Volcani, 1981; Claquin et al., 2002). Hence, we

may observe decreased growth and Si deposition rates, smaller filaments, and malformed

frustules due in part to this arrest and disruption in the G1 cell phase. It is well established that

eukaryotic phototrophs use cell cycle arrest as a means to combat stress, with this mechanism

induced by a variety of abiotic stressors (Eekhout and De Veylder, 2019; Takahashi et al., 2019).

We suggest the photostress observed in pH 9.2 treatments may be responsible for this apparent

cell cycle arrest hinted at in our transcriptomic data, and therefore contributing to morphological

and physiological effects observed in F. crotonensis.

High pH results in smaller, rougher & browner F. crotonensis morphologies

In this study, growth at pH 9.2 resulted in significant morphological changes within our

model freshwater diatom. Filaments maintained at high pH for 8 d exhibited significantly shorter

lengths, resulting in lower biovolume and area. Despite this, individual diatom cells did not

significantly differ in width. A prior study demonstrated warming temperatures results in smaller

diatom cells across the Laurentian Great Lakes (Bramburger et al., 2017). Considering lake

basification events will likely coincide with warming temperatures, future freshwater diatom

communities may exhibit both smaller filaments and smaller individual cells.

FlowCAM analyses further indicated filaments had rougher exterior surfaces at high pH.

We previously examined the pH dependence of Si deposition, demonstrating that it decreased at

pH 9.2. Here, we build upon this with FlowCAM and transcriptomic data which suggest diatoms

are struggling to deposit Si, and the diatoms that do successfully form frustules may have

malformations which are evidenced by this “roughness”. We observed a decrease in the

146
expression of a gene (TPR) whose expression is positively correlated with silaffin expression (a

gene directly involved in Si deposition) (Frigeri et al., 2006), suggesting F. crotonensis may be

downregulating the biosilicification process (Shrestha et al., 2012). Genes suggested to encode

for diatom cell wall components such as CLPP and ANKR3 (Frigeri et al., 2006), and C1GALT

and UXS1 (Alexander et al., 2015), were also decreased in expression at high pH. Alterations in

the expression of biosilicification genes, and/or those involved in forming the cell wall structures

which the frustules sit upon, will produce morphological malformities (Round et al., 1990;

Hildebrand et al., 2006). Cumulatively, these data suggest diatoms struggle to deposit Si

frustules at pH 9.2, likely leading to cell walls with rougher and malformed phenotypes.

In addition to smaller and “rougher” filaments, high pH results in “browner” diatoms due

to significantly increased concentrations of β-carotene and diadinoxanthin but ~constant

chlorophyll concentrations. Prior studies demonstrate both fucoxanthin (involved in light

harvesting) and diadinoxanthin (involved in photoprotection) are the main carotenoids of

diatoms, serving as components of the LHCs (Ballottari et al., 2012). Indeed, total

carotenoid:total Chl a ratios were higher at pH 9.2, though these findings fell short of

significance due to ~constant fucoxanthin concentrations. Nonetheless, there were significant

increases in the diadinoxanthin:total Chl a and β-carotene:total Chl a ratios observed in pH 9.2

filaments. This contradicts our transcriptomic data, as carotenoid biosynthesis genes (ZEP and

VDE) were decreased in expression at pH 9.2. However, changes in pigment levels often occur

without detectable changes in gene expression (Kuczynska et al., 2015). Regardless, significantly

higher β-carotene and diadinoxanthin photopigment concentrations lead to “less green” filaments

and significantly lower Chl a autofluorescence, all suggesting “browner” diatoms at pH 9.2.

147
Cumulatively, this data demonstrates lake basification induces significant changes to F.

crotonensis morphology, resulting in smaller, rougher, and browner filaments.

Lake basification has ecological & biogeochemical implications

While acidification has been suggested to benefit the growth and ecological resilience of

marine diatoms (Wu et al., 2014; Valenzuela et al., 2018), we suggest here that basification poses

a detriment. Diatoms are responsible for ~20% of global primary production (Nelson et al.,

1995), acting as an integral component of the aquatic primary producers. Hence, altering diatom

populations will evoke significant biotic and biogeochemical implications within the aquatic

ecosystem (Rühland et al., 2015). For example, here we demonstrate high pH results in smaller,

browner, rougher diatoms. Based on prior research, we postulate these morphological alterations

will modify grazing patterns of secondary consumers and selective filtration of higher trophic

consumers (Vanderploeg et al., 2001; Baranowska et al., 2013; Vanderploeg et al., 2013). This

decrease in size will also likely increase diatom predation by Dreissenids, as smaller diatoms

seem to be selectively filtered (Reavie and Barbiero, 2013). Additionally, declines in Si

deposition rates combined with rougher frustules suggests thinner and malformed cell walls

which may reduce their effectiveness as a defense mechanism against zooplankton grazing

(Pančić et al., 2019; Ryderheim et al., 2022) and viral infection (Kranzler et al., 2019).

Due in part to their heavy Si frustules, diatoms also play an instrumental role in global

biogeochemical cycles and nutrient export to the benthos (Struyf et al., 2009; Benoiston et al.,

2017), with studies demonstrating the Si cycle is more strongly inter-related with the carbon,

nitrogen and phosphorus cycles than previously thought (Tréguer and De La Rocha, 2013;

Tréguer et al., 2021). For example, Lake Erie undergoes such a high degree of diatom deposition

148
to the benthos in the winter-spring, that it serves as a substantial driver of summer hypoxia

(Reavie et al., 2016). Another study demonstrated that even when an estuary water column was

dominated by flagellates, the labile organic matter settling to the sediment was derived from

diatoms (Haese et al., 2007). Thus, we postulate lake basification-induced changes to diatoms

will have significant downstream effects within the lacustrine ecosystem and biogeochemical

cycles, particularly with respect to the benthos which experience the highest diatom-nutrient

deposition.

“Genomic roulette”: the key to freshwater diatom success in a basified future?

Abundant transcripts from transposable elements within the F. crotonensis genome

provide insight into a mechanism these diatoms may use to adapt and persist at high pH or other

severe environmental stressors that could be otherwise detrimental. Approximately 25% of the

DE genes that were increased in expression at high pH belonged within the “Mobilome:

prophages, transposons” COG classification. Mobile elements like these change the architecture

of an organism’s genome by rearranging themselves to new locations or (in some cases) moving

pieces of the genome itself. Such rearrangements can, for example, insert into a gene and disrupt

it, or insert into a regulatory site and change how genes are expressed (Pennisi, 1998; Lisch,

2013; Schrader and Schmitz, 2019). Extensive genomic variation has been attributed to similar

rearrangements in green alga like Chlamydomonas reinhardtii (Flowers et al., 2015).

It appears that a potential response to environmental conditions stressful enough to

heavily damage an organism – like a shift to extreme pH conditions - is an attempt to make a

significant, microevolutionary leap by re-arranging one’s own genomic architecture at random.

Diatoms have been described as “one of the most rapidly evolving eukaryotic taxa on Earth”

149
(Oliver et al., 2007; Vardi et al., 2009). This speed has been attributed to their high proportion of

retrotransposons, long terminal repeats, and transposable elements (Bowler et al., 2008; Vardi et

al., 2009; Rastogi et al., 2018). This type of win or lose strategy, which we have termed

“genomic roulette”, is perhaps an extreme interpretation of the role of “jumping genes” in

organismal adaptation to ecological stress (McClintock, 1956; Capy et al., 2000; Horváth et al.,

2017). We define genomic roulette as an upregulation of mobile elements in response to an

environmental stressor to rapidly increase mutations within an organism, and thus the genetic

diversity within a population, over a short period of time. While most random mutations of this

nature can be hypothesized to prove disadvantageous, there remains the likelihood a small subset

will confer fitness to the organism and facilitate survival.

Stress-induced alterations to genomic architecture have been suggested in other

phototrophs ranging from single-celled cyanobacteria (Lin et al., 2010; Hu et al., 2018) to

multicellular plants (Negi et al., 2016; Roquis et al., 2021). For example, Microcystis aeruginosa

in lab cultures has been shown to massively upregulate transposases when shifted to urea as a

nitrogen source for growth (Steffen et al., 2014b). Returning to diatoms, recent studies have

suggested genome evolution was responsible for cold- climate adaptations in the polar marine

diatom Fragilariopsis cylindrus (Mock et al., 2017) and warmer- climate adaptations in tropical

marine diatoms in response to ocean warming (Jin and Agustí, 2018). Our observations provide a

potential mechanism for how our freshwater diatom F. crotonensis may be gambling in the game

of genomic roulette at high pH, with the potential to facilitate a rapid adaptation to these high pH

conditions at the likely cost of many in the cohort. Going forward, there is a need to better

elucidate this hypothesis within diatoms, determine if this strategy of genomic roulette

150
extrapolates to other organisms beyond photoautotrophs, and a necessity to develop an

understanding of its rate of success.

Conclusion
In this study, transcriptomic analyses revealed genes associated with photosynthesis, not

carbohydrate metabolism, were driving dissimilarity between high pH vs. low pH expression.

We demonstrated that a pH of 9.2, which is routinely reached during lake basification events,

significantly alters F. crotonensis filament morphology resulting in smaller, browner, and

rougher diatoms after just 8 days of exposure. This pH further modifies F. crotonensis

photophysiology by significantly decreasing both maximum electron transport rates through PSII

and light saturation thresholds. Cumulatively, these transcriptomic, morphological, and

physiological findings imply F. crotonensis experienced photostress in the high pH treatment,

with evidence suggesting filaments invoke photoacclimative strategies in response. In turn,

transcriptomic evidence suggests this photostress is inducing cell cycle arrest at the G1 /S

checkpoint, which would explain the decreased growth and Si deposition rates observed

previously (Zepernick et al., 2021).

Increased sensitivity to light stress is likely exacerbated in the environment, as prior

findings indicate the highest diel pH spikes coincide with peak photosynthesis rates and light

levels during the afternoon (Krausfeldt et al., 2019), a phenomenon likely intensified by

prolonged basification events of up to ~40 d in the environment (Zepernick et al., 2021). Yet,

this stress may be partially alleviated by the thick Microcystis spp. scum associated with bloom

events, as colonies regulate their own buoyancy and shade the water column. Thus, further

research is needed to understand responses of freshwater diatoms to both diel variations in pH on

a short-term scale, and the long-term effects basification will impose on freshwater diatom

151
communities. As climate change serves to increase cyanobacterial bloom distribution, duration

and frequency (Wells et al., 2020), there is a need to elucidate how freshwater diatom

communities will respond (and adapt) to these lake basification events.

Acknowledgements
We thank Dr. Gary LeCleir, Dr. R. Michael L. McKay, Dr. George Bullerjahn, Dr. Erik Zinser,

Naomi Gilbert and Elizabeth Dennison for their comments and suggestions.

152
References:
Alexander, H., Jenkins, B.D., Rynearson, T.A., and Dyhrman, S.T. (2015). Metatranscriptome
analyses indicate resource partitioning between diatoms in the field. Proceedings of the
National Academy of Sciences 112, E2182-E2190.
Anderson, D.M. (2009). Approaches to monitoring, control and management of harmful algal
blooms (HABs). Ocean & Coastal Management 52, 342-347.
Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J.R., Maciejewski, A., and Wishart, D.S.
(2016). Heatmapper: web-enabled heat mapping for all. Nucleic acids research 44,
W147-W153.
Ballottari, M., Girardon, J., Dall'osto, L., and Bassi, R. (2012). Evolution and functional
properties of photosystem II light harvesting complexes in eukaryotes. Biochimica et
Biophysica Acta (BBA)-Bioenergetics 1817, 143-157.
Baranowska, K., North, R.L., Winter, J., and Dillon, P. (2013). Long-term seasonal effects of
dreissenid mussels on phytoplankton in Lake Simcoe, Ontario, Canada. Inland Waters 3,
285-295.
Bassi, R., and Caffarri, S. (2000). Lhc proteins and the regulation of photosynthetic light
harvesting function by xanthophylls. Photosynthesis Research 64, 243-256.
Benoiston, A.-S., Ibarbalz, F.M., Bittner, L., Guidi, L., Jahn, O., Dutkiewicz, S., and Bowler, C.
(2017). The evolution of diatoms and their biogeochemical functions. Philosophical
Transactions of the Royal Society B: Biological Sciences 372, 20160397.
Bertrand, M. (2010). Carotenoid biosynthesis in diatoms. Photosynthesis research 106, 89-102.
Bowler, C., Allen, A.E., Badger, J.H., Grimwood, J., Jabbari, K., Kuo, A., Maheswari, U.,
Martens, C., Maumus, F., and Otillar, R.P. (2008). The Phaeodactylum genome reveals
the evolutionary history of diatom genomes. Nature 456, 239.
Bramburger, A.J., Reavie, E.D., Sgro, G., Estepp, L., Chraïbi, V.S., and Pillsbury, R. (2017).
Decreases in diatom cell size during the 20th century in the Laurentian Great Lakes: a
response to warming waters? Journal of Plankton Research 39, 199-210.
Burkhardt, S., Amoroso, G., Riebesell, U., and Sültemeyer, D. (2001). CO2 and HCO3 ߚ uptake
in marine diatoms acclimated to different CO2 concentrations. Limnology and
Oceanography 46, 1378-1391.
Bushnell, B. (2014). "BBMap: a fast, accurate, splice-aware aligner". Lawrence Berkeley
National Lab.(LBNL), Berkeley, CA (United States)).
Capy, P., Gasperi, G., Biémont, C., and Bazin, C. (2000). Stress and transposable elements: co-
evolution or useful parasites? Heredity 85, 101-106.
Claquin, P., Kromkamp, J.C., and Martin-Jezequel, V. (2004). Relationship between
photosynthetic metabolism and cell cycle in a synchronized culture of the marine alga
Cylindrotheca fusiformis (Bacillariophyceae). European Journal of Phycology 39, 33-41.
Claquin, P., Martin‐Jézéquel, V., Kromkamp, J.C., Veldhuis, M.J., and Kraay, G.W. (2002).
Uncoupling of silicon compared with carbon and nitrogen metabolisms and the role of
the cell cycle in continuous cultures of Thalassiosira pseudonana (Bacillariophyceae)
under light, nitrogen, and phosphorus control1. Journal of Phycology 38, 922-930.
Clarke, K., and Gorley, R. (2015). Getting started with PRIMER v7. PRIMER-E: Plymouth,
Plymouth Marine Laboratory 20.
Crawford, R. (1981). "The siliceous components of the diatom cell wall and their morphological
variation," in Silicon and siliceous structures in biological systems. Springer), 129-156.

153
Dong, Z., Zhu, C., Zhan, Q., and Jiang, W. (2017). The roles of RRP15 in nucleolar formation,
ribosome biogenesis and checkpoint control in human cells. Oncotarget 8, 13240.
Eekhout, T., and De Veylder, L. (2019). Plant stress: hitting pause on the cell cycle. Elife 8,
e46781.
Fischer, N., Hippler, M., Sétif, P., Jacquot, J.-P., and Rochaix, J.-D. (1998). The psaC subunit of
photosystem I provides an essential lysine residue for fast electron transfer to ferredoxin.
The EMBO Journal 17, 849-858.
Flowers, J.M., Hazzouri, K.M., Pham, G.M., Rosas, U., Bahmani, T., Khraiwesh, B., Nelson,
D.R., Jijakli, K., Abdrabu, R., and Harris, E.H. (2015). Whole-genome resequencing
reveals extensive natural variation in the model green alga Chlamydomonas reinhardtii.
The Plant Cell 27, 2353-2369.
Frigeri, L.G., Radabaugh, T.R., Haynes, P.A., and Hildebrand, M. (2006). Identification of
Proteins from a Cell Wall Fraction of the Diatom Thalassiosira pseudonana: Insights into
Silica Structure Formation. Molecular & Cellular Proteomics 5, 182-193.
Gleich, S.J., Plough, L.V., and Glibert, P.M. (2020). Photosynthetic efficiency and nutrient
physiology of the diatom Thalassiosira pseudonana at three growth temperatures. Marine
Biology 167, 1-13.
Gobler, C.J., and Sunda, W.G. (2012). Ecosystem disruptive algal blooms of the brown tide
species, Aureococcus anophagefferens and Aureoumbra lagunensis. Harmful Algae 14,
36-45.
Haese, R.R., Murray, E.J., Smith, C.S., Smith, J., Clementson, L., and Heggie, D.T. (2007).
Diatoms control nutrient cycles in a temperate, wave‐dominated estuary (southeast
Australia). Limnology and Oceanography 52, 2686-2700.
Hartig, J.H. (1987). Factors contributing to development of Fragilaria crontonensis Kitton
Pulses in Pigeon Bay waters of western Lake Erie. Journal of Great Lakes Research 13,
65-77.
Heisler, J., Glibert, P.M., Burkholder, J.M., Anderson, D.M., Cochlan, W., Dennison, W.C.,
Dortch, Q., Gobler, C.J., Heil, C.A., and Humphries, E. (2008). Eutrophication and
harmful algal blooms: a scientific consensus. Harmful Algae 8, 3-13.
Hervé, V., Derr, J., Douady, S., Quinet, M., Moisan, L., and Lopez, P.J. (2012). Multiparametric
analyses reveal the pH-dependence of silicon biomineralization in diatoms. PloS One 7,
e46722.
Heydarizadeh, P., Veidl, B., Huang, B., Lukomska, E., Wielgosz-Collin, G., Couzinet-Mossion,
A., Bougaran, G., Marchand, J., and Schoefs, B. (2019). Carbon orientation in the diatom
Phaeodactylum tricornutum: the effects of carbon limitation and photon flux density.
Frontiers in plant science 10, 471.
Hildebrand, M., Frigeri, L.G., and Davis, A.K. (2007). Synchronized growth of Thalassiosira
psuedonana (Bacillariophyceae) provides novel insights into cell-wall synthesis
processes in relation to the cell cycle. Journal of Phycology 43, 730-740.
Hildebrand, M., York, E., Kelz, J.I., Davis, A.K., Frigeri, L.G., Allison, D.P., and Doktycz, M.J.
(2006). Nanoscale control of silica morphology and three-dimensional structure during
diatom cell wall formation. Journal of Materials Research 21, 2689-2698.
Hopkins, B.L., and Neumann, C.A. (2019). Redoxins as gatekeepers of the transcriptional
oxidative stress response. Redox Biology 21, 101104.
Horton, P., Ruban, A., and Walters, R. (1996). Regulation of light harvesting in green plants.
Annual Review of Plant Biology 47, 655-684.

154
Horváth, V., Merenciano, M., and González, J. (2017). Revisiting the relationship between
transposable elements and the eukaryotic stress response. Trends in Genetics 33, 832-
841.
Hu, L., Xiao, P., Jiang, Y., Dong, M., Chen, Z., Li, H., Hu, Z., Lei, A., and Wang, J. (2018).
Transgenerational epigenetic inheritance under environmental stress by genome-wide
DNA methylation profiling in cyanobacterium. Frontiers in Microbiology 9, 1479.
Huerta-Cepas, J., Szklarczyk, D., Heller, D., Hernández-Plaza, A., Forslund, S.K., Cook, H.,
Mende, D.R., Letunic, I., Rattei, T., and Jensen, L.J. (2019). eggNOG 5.0: a hierarchical,
functionally and phylogenetically annotated orthology resource based on 5090 organisms
and 2502 viruses. Nucleic Acids Research 47, D309-D314.
Jang, S.J., Shin, S.H., Yee, S.T., Hwang, B., Im, K.H., and Park, K.Y. (2005). Effects of abiotic
stresses on cell cycle progression in tobacco BY-2 cells. Molecules & Cells (Springer
Science & Business Media BV) 20.
Javaheri, N., Dries, R., and Kaandorp, J. (2014). Understanding the sub-cellular dynamics of
silicon transportation and synthesis in diatoms using population-level data and
computational optimization. PLoS Computational Biology 10, e1003687.
Jeppesen, E., Søndergaard, M., Sortkjær, O., Mortensen, E., and Kristensen, P. (1990).
"Interactions between phytoplankton, zooplankton and fish in a shallow, hypertrophic
lake: a study of phytoplankton collapses in Lake Søbygård, Denmark," in Trophic
Relationships in Inland Waters. Springer), 149-164.
Ji, X., Verspagen, J.M., Van De Waal, D.B., Rost, B., and Huisman, J. (2020). Phenotypic
plasticity of carbon fixation stimulates cyanobacterial blooms at elevated CO2. Science
Advances 6.
Jin, P., and Agustí, S. (2018). Fast adaptation of tropical diatoms to increased warming with
trade-offs. Scientific Reports 8, 1-10.
Kanehisa, M., Sato, Y., Kawashima, M., Furumichi, M., and Tanabe, M. (2016). KEGG as a
reference resource for gene and protein annotation. Nucleic Acids Research 44, D457-
D462.
Kim, J., Brown, C.M., Kim, M.K., Burrows, E.H., Bach, S., Lun, D.S., and Falkowski, P.G.
(2017). Effect of cell cycle arrest on intermediate metabolism in the marine diatom
Phaeodactylum tricornutum. Proceedings of the National Academy of Sciences 114,
E8007-E8016.
Kranzler, C.F., Krause, J.W., Brzezinski, M.A., Edwards, B.R., Biggs, W.P., Maniscalco, M.,
Mccrow, J.P., Van Mooy, B.A., Bidle, K.D., and Allen, A.E. (2019). Silicon limitation
facilitates virus infection and mortality of marine diatoms. Nature Microbiology 4, 1790-
1797.
Krausfeldt, L.E., Farmer, A.T., Castro Gonzalez, H., Zepernick, B.N., Campagna, S.R., and
Wilhelm, S.W. (2019). Urea is both a carbon and nitrogen source for Microcystis
aeruginosa: tracking 13C incorporation at bloom pH conditions. Frontiers in
Microbiology 10, 1064.
Kuczynska, P., Jemiola-Rzeminska, M., and Strzalka, K. (2015). Photosynthetic pigments in
diatoms. Marine Drugs 13, 5847-5881.
Lamote, M., Darko, E., Schoefs, B., and Lemoine, Y. (2003). Assembly of the photosynthetic
apparatus in embryos from Fucus serratus. Photosynthesis Research 77, 45-52.

155
Laronde-Leblanc, N., and Wlodawer, A. (2005). The RIO kinases: an atypical protein kinase
family required for ribosome biogenesis and cell cycle progression. Biochimica et
Biophysica Acta (BBA)-Proteins and Proteomics 1754, 14-24.
Lehman, P., Marr, K., Boyer, G., Acuna, S., and Teh, S.J. (2013). Long-term trends and causal
factors associated with Microcystis abundance and toxicity in San Francisco Estuary and
implications for climate change impacts. Hydrobiologia 718, 141-158.
Lin, S., Haas, S., Zemojtel, T., Xiao, P., Vingron, M., and Li, R. (2010). Genome-wide
comparison of cyanobacterial transposable elements, potential genetic diversity
indicators. Nature Precedings, 1-1.
Lisch, D. (2013). How important are transposons for plant evolution? Nature Reviews Genetics
14, 49-61.
Lopez-Archilla, A.I., Moreira, D., López-García, P., and Guerrero, C. (2004). Phytoplankton
diversity and cyanobacterial dominance in a hypereutrophic shallow lake with
biologically produced alkaline pH. Extremophiles 8, 109-115.
Lu, M. (2019). Structure and mechanism of the divalent anion/Na+ symporter. International
Journal of Molecular Sciences 20, 440.
Martin, R.M., and Wilhelm, S.W. (2020). Phenol-based RNA Extraction from Polycarbonate
Filters. protocols.io.
Mcclintock, B. (Year). "Controlling elements and the gene", in: Cold Spring Harbor symposia
on quantitative biology: Cold Spring Harbor Laboratory Press), 197-216.
Melo, A.M., Bandeiras, T.M., and Teixeira, M. (2004). New insights into type II NAD (P) H:
quinone oxidoreductases. Microbiology and Molecular Biology Reviews 68, 603-616.
Mock, T., Otillar, R.P., Strauss, J., Mcmullan, M., Paajanen, P., Schmutz, J., Salamov, A.,
Sanges, R., Toseland, A., and Ward, B.J. (2017). Evolutionary genomics of the cold-
adapted diatom Fragilariopsis cylindrus. Nature 541, 536-540.
Moulager, M., Corellou, F., Vergé, V., Escande, M.-L., and Bouget, F.-Y. (2010). Integration of
light signals by the retinoblastoma pathway in the control of S phase entry in the
picophytoplanktonic cell Ostreococcus. PLoS Genetics 6, e1000957.
Muller, P., Li, X.-P., and Niyogi, K.K. (2001). Non-photochemical quenching. A response to
excess light energy. Plant Physiology 125, 1558-1566.
Negi, P., Rai, A.N., and Suprasanna, P. (2016). Moving through the stressed genome: emerging
regulatory roles for transposons in plant stress response. Frontiers in Plant Science 7,
1448.
Nelson, D.M., Tréguer, P., Brzezinski, M.A., Leynaert, A., and Quéguiner, B. (1995). Production
and dissolution of biogenic silica in the ocean: revised global estimates, comparison with
regional data and relationship to biogenic sedimentation. Global Biogeochemical Cycles
9, 359-372.
Nguyen, M.K., Shih, T.-H., Lin, S.-H., Lin, J.-W., Nguyen, H.C., Yang, Z.-W., and Yang, C.-M.
(2021). Transcription profile analysis of chlorophyll biosynthesis in leaves of wild-type
and chlorophyll b-deficient rice (Oryza sativa L.). Agriculture 11, 401.
Nguyen-Deroche, T.L.N., Caruso, A., Le, T.T., Bui, T.V., Schoefs, B., Tremblin, G., and
Morant-Manceau, A. (2012). Zinc affects differently growth, photosynthesis, antioxidant
enzyme activities and phytochelatin synthase expression of four marine diatoms. The
Scientific World Journal 2012.

156
Nishiyama, Y., Yamamoto, H., Allakhverdiev, S.I., Inaba, M., Yokota, A., and Murata, N.
(2001). Oxidative stress inhibits the repair of photodamage to the photosynthetic
machinery. The EMBO Journal 20, 5587-5594.
Oka, K., Ueno, Y., Yokono, M., Shen, J.-R., Nagao, R., and Akimoto, S. (2020). Adaptation of
light-harvesting and energy-transfer processes of a diatom Phaeodactylum tricornutum to
different light qualities. Photosynthesis Research 146, 227-234.
Oliver, M.J., Petrov, D., Ackerly, D., Falkowski, P., and Schofield, O.M. (2007). The mode and
tempo of genome size evolution in eukaryotes. Genome Research 17, 594-601.
Paerl, H.W., Fulton, R.S., Moisander, P.H., and Dyble, J. (2001). Harmful freshwater algal
blooms, with an emphasis on cyanobacteria. The Scientific World Journal 1, 76-113.
Paerl, H.W., Hall, N.S., Peierls, B.L., Rossignol, K.L., and Joyner, A.R. (2014). Hydrologic
variability and its control of phytoplankton community structure and function in two
shallow, coastal, lagoonal ecosystems: the Neuse and New River Estuaries, North
Carolina, USA. Estuaries and Coasts 37, 31-45.
Pančić, M., Torres, R.R., Almeda, R., and Kiørboe, T. (2019). Silicified cell walls as a defensive
trait in diatoms. Proceedings of the Royal Society B 286, 20190184.
Park, S., Jung, G., Hwang, Y.-S., and Jin, E. (2010). Dynamic response of the transcriptome of a
psychrophilic diatom, Chaetoceros neogracile, to high irradiance. Planta 231, 349-360.
Pennisi, E. (1998). "How the genome readies itself for evolution". American Association for the
Advancement of Science).
Perry, M., Talbot, M., and Alberte, R. (1981). Photoadaption in marine phytoplankton: response
of the photosynthetic unit. Marine Biology 62, 91-101.
Pinckney, J., Millie, D., Howe, K., Paerl, H., and Hurley, J. (1996). Flow scintillation counting
of 14C-labeled microalgal photosynthetic pigments. Journal of Plankton Research 18,
1867-1880.
Pinckney, J., Paerl, H., Harrington, M., and Howe, K. (1998). Annual cycles of phytoplankton
community-structure and bloom dynamics in the Neuse River Estuary, North Carolina.
Marine Biology 131, 371-381.
Pinckney, J.L., Richardson, T.L., Millie, D.F., and Paerl, H.W. (2001). Application of
photopigment biomarkers for quantifying microalgal community composition and in situ
growth rates. Organic Geochemistry 32, 585-595.
Rastogi, A., Maheswari, U., Dorrell, R.G., Vieira, F.R.J., Maumus, F., Kustka, A., Mccarthy, J.,
Allen, A.E., Kersey, P., and Bowler, C. (2018). Integrative analysis of large scale
transcriptome data draws a comprehensive landscape of Phaeodactylum tricornutum
genome and evolutionary origin of diatoms. Scientific Reports 8, 1-14.
Reavie, E.D., and Barbiero, R.P. (2013). Recent changes in abundance and cell size of pelagic
diatoms in the North American Great Lakes. Phytotaxa 127, 150-162.
Reavie, E.D., Cai, M., Twiss, M.R., Carrick, H.J., Davis, T.W., Johengen, T.H., Gossiaux, D.,
Smith, D.E., Palladino, D., and Burtner, A. (2016). Winter–spring diatom production in
Lake Erie is an important driver of summer hypoxia. Journal of Great Lakes Research
42, 608-618.
Ritchie, R.J. (2008). Fitting light saturation curves measured using modulated fluorometry.
Photosynthesis Research 96, 201-215.
Roquis, D., Robertson, M., Yu, L., Thieme, M., Julkowska, M., and Bucher, E. (2021). Genomic
impact of stress-induced transposable element mobility in Arabidopsis. Nucleic Acids
Research 49, 10431-10447.

157
Round, F.E., Crawford, R.M., and Mann, D.G. (1990). Diatoms: biology and morphology of the
genera. Cambridge university press.
Rühland, K.M., Paterson, A.M., and Smol, J.P. (2015). Lake diatom responses to warming:
reviewing the evidence. Journal of Paleolimnology 54, 1-35.
Ruprecht, J.J., and Kunji, E.R. (2020). The SLC25 mitochondrial carrier family: structure and
mechanism. Trends in Biochemical Sciences 45, 244-258.
Ryderheim, F., Grønning, J., and Kiørboe, T. (2022). Thicker shells reduce copepod grazing on
diatoms. Limnology and Oceanography Letters.
Saini, Y., Chen, J., and Patial, S. (2020). The tristetraprolin family of RNA-binding proteins in
cancer: Progress and future prospects. Cancers 12, 1539.
Sandrini, G., Tann, R.P., Schuurmans, J.M., Van Beusekom, S.A., Matthijs, H.C., and Huisman,
J. (2016). Diel variation in gene expression of the CO2-concentrating mechanism during a
harmful cyanobacterial bloom. Frontiers in Microbiology 7, 551.
Schrader, L., and Schmitz, J. (2019). The impact of transposable elements in adaptive evolution.
Molecular Ecology 28, 1537-1549.
Shrestha, R.P., Tesson, B., Norden-Krichmar, T., Federowicz, S., Hildebrand, M., and Allen,
A.E. (2012). Whole transcriptome analysis of the silicon response of the diatom
Thalassiosira pseudonana. BMC Genomics 13, 499.
Steffen, M.M., Dearth, S.P., Dill, B.D., Li, Z., Larsen, K.M., Campagna, S.R., and Wilhelm,
S.W. (2014). Nutrients drive transcriptional changes that maintain metabolic homeostasis
but alter genome architecture in Microcystis. The ISME Journal 8, 2080-2092.
Struyf, E., Smis, A., Van Damme, S., Meire, P., and Conley, D.J. (2009). The global
biogeochemical silicon cycle. Silicon 1, 207-213.
Su, X., Xue, Q., Steinman, A.D., Zhao, Y., and Xie, L. (2015). Spatiotemporal dynamics of
microcystin variants and relationships with environmental parameters in Lake Taihu,
China. Toxins 7, 3224-3244.
Sültemeyer, D. (1998). Carbonic anhydrase in eukaryotic algae: characterization, regulation, and
possible function during photosynthesis. Canadian Journal of Botany 76, 962-972.
Takahashi, N., Ogita, N., Takahashi, T., Taniguchi, S., Tanaka, M., Seki, M., and Umeda, M.
(2019). A regulatory module controlling stress-induced cell cycle arrest in Arabidopsis.
Elife 8, e43944.
Takahashi-Íñiguez, T., Aburto-Rodríguez, N., Vilchis-González, A.L., and Flores, M.E. (2016).
Function, kinetic properties, crystallization, and regulation of microbial malate
dehydrogenase. Journal of Zhejiang University-SCIENCE B 17, 247-261.
Talling, J. (1976). The depletion of carbon dioxide from lake water by phytoplankton. The
Journal of Ecology, 79-121.
Tang, X., Krausfeldt, L.E., Shao, K., Lecleir, G.R., Stough, J.M., Gao, G., Boyer, G.L., Zhang,
Y., Paerl, H.W., and Qin, B. (2018). Seasonal gene expression and the ecophysiological
implications of toxic Microcystis aeruginosa blooms in Lake Taihu. Environmental
Science & Technology 52, 11049-11059.
Thomas, B. (2016). Encyclopedia of applied plant sciences. Academic Press.
Torres, M., Ritchie, R., Lilley, R., Grillet, C., and Larkum, A. (2014). Measurement of
photosynthesis and photosynthetic efficiency in two diatoms. New Zealand Journal of
Botany 52, 6-27.
Tréguer, P.J., and De La Rocha, C.L. (2013). The world ocean silica cycle. Annual Review of
Marine Science 5, 477-501.

158
Tréguer, P.J., Sutton, J.N., Brzezinski, M., Charette, M.A., Devries, T., Dutkiewicz, S., Ehlert,
C., Hawkings, J., Leynaert, A., and Liu, S.M. (2021). Reviews and syntheses: The
biogeochemical cycle of silicon in the modern ocean. Biogeosciences 18, 1269-1289.
Turner, E., Ombres, E., Bennett-Mintz, J., Dortch, Q., Broadwater, M., Berger, H., and Harris,
M. (2021). "Ocean acidification program: Harmful algal blooms and ocean acidification
workshop: Defining a research agenda". NOAA NCCOS, NOAA OAP).
Valenzuela, J.J., López García De Lomana, A., Lee, A., Armbrust, E., Orellana, M.V., and
Baliga, N.S. (2018). Ocean acidification conditions increase resilience of marine diatoms.
Nature Communications 9, 1-10.
Van Dam, B.R., Tobias, C., Holbach, A., Paerl, H.W., and Zhu, G. (2018). CO2 limited
conditions favor cyanobacteria in a hypereutrophic lake: an empirical and theoretical
stable isotope study. Limnology and Oceanography 63, 1643-1659.
Van Heukelem, L., Lewitus, A.J., Kana, T.M., and Craft, N.E. (1994). Improved separations of
phytoplankton pigments using temperature-controlled high performance liquid
chromatography. Marine Ecology Progress Series, 303-313.
Vanderploeg, H., Wilson, A., Johengen, T., Bressie, J.D., Sarnelle, O., Liebig, J., Robinson, S.,
and Horst, G. (2013). Role of selective grazing by dreissenid mussels in promoting toxic
Microcystis blooms and other changes in phytoplankton composition in the Great Lakes.
Quagga and Zebra Mussels: Biology, Impacts, and Control, 509-523.
Vanderploeg, H.A., Liebig, J.R., Carmichael, W.W., Agy, M.A., Johengen, T.H., Fahnenstiel,
G.L., and Nalepa, T.F. (2001). Zebra mussel (Dreissena polymorpha) selective filtration
promoted toxic Microcystis blooms in Saginaw Bay (Lake Huron) and Lake Erie.
Canadian Journal of Fisheries and Aquatic Sciences 58, 1208-1221.
Vardi, A., Thamatrakoln, K., Bidle, K.D., and Falkowski, P.G. (2009). Diatom genomes come of
age. Genome Biology 9, 245.
Vázquez-Acevedo, M., Vega-Deluna, F., Sánchez-Vásquez, L., Colina-Tenorio, L., Remacle, C.,
Cardol, P., Miranda-Astudillo, H., and González-Halphen, D. (2016). Dissecting the
peripheral stalk of the mitochondrial ATP synthase of chlorophycean algae. Biochimica
et Biophysica Acta (BBA)-Bioenergetics 1857, 1183-1190.
Verspagen, J.M., Snelder, E.O., Visser, P.M., Huisman, J., Mur, L.R., and Ibelings, B.W. (2004).
Recruitment of Benthic Microcystis (cyanophyceae) to the Water Column: Internal
Buoyancy Changes or Resuspension? 1. Journal of Phycology 40, 260-270.
Verspagen, J.M., Van De Waal, D.B., Finke, J.F., Visser, P.M., Van Donk, E., and Huisman, J.
(2014). Rising CO2 levels will intensify phytoplankton blooms in eutrophic and
hypertrophic lakes. PloS One 9.
Volcani, B. (1981). "Cell wall formation in diatoms: morphogenesis and biochemistry," in
Silicon and siliceous structures in biological systems. Springer), 157-200.
Watson, S.B., Whitton, B.A., Higgins, S.N., Paerl, H.W., Brooks, B.W., and Wehr, J.D. (2015).
"Harmful Algal Blooms," in Aquatic Ecology, Freshwater Algae of North America, eds.
J.D. Wehr, R.G. Sheath & P.J. Kociolek. Second ed (Aquatic Ecology: Academic Press),
873-920.
Wells, M.L., Karlson, B., Wulff, A., Kudela, R., Trick, C., Asnaghi, V., Berdalet, E., Cochlan,
W., Davidson, K., and De Rijcke, M. (2020). Future HAB science: Directions and
challenges in a changing climate. Harmful Algae 91, 101632.

159
Wilhelm, C., Büchel, C., Fisahn, J., Goss, R., Jakob, T., Laroche, J., Lavaud, J., Lohr, M.,
Riebesell, U., and Stehfest, K. (2006). The regulation of carbon and nutrient assimilation
in diatoms is significantly different from green algae. Protist 157, 91-124.
Wilhelm, S., Bullerjahn, G., and Rlm, M. (2020). The complicated and confusing ecology of
Microcystis blooms. mBio.
Wu, Y., Campbell, D.A., Irwin, A.J., Suggett, D.J., and Finkel, Z.V. (2014). Ocean acidification
enhances the growth rate of larger diatoms. Limnology and Oceanography 59, 1027-
1034.
Zagarese, H.E., Sagrario, M.D.L.Á.G., Wolf-Gladrow, D., Nõges, P., Nõges, T., Kangur, K.,
Matsuzaki, S.-I.S., Kohzu, A., Vanni, M.J., and Özkundakci, D. (2021). Patterns of CO2
concentration and inorganic carbon limitation of phytoplankton biomass in agriculturally
eutrophic lakes. Water Research 190, 116715.
Zepernick, B.N., Denison, E.R., Chaffin, J.D., Bullerjahn, G.S., Pennacchio, C.P., Frenken, T.,
Peck, D.H., Anderson, J.T., Niles, D., and Zastepa, A. (2022a). Metatranscriptomic
Sequencing of Winter and Spring Planktonic Communities from Lake Erie, a Laurentian
Great Lake. Microbiology Resource Announcements, e00351-00322.
Zepernick, B.N., Gann, E.R., Poiund, H.L., Martin, R.M., Krausfeldt, L.E., Chaffin, J.D., and
Wilhelm, S.W. (2021). Elevated pH conditions associated with Microcystis spp. blooms
decrease viability of the cultured diatom Fragilaria crotonensis and natural diatoms in
Lake Erie. Frontiers in Microbiology 12, 598736.
Zepernick, B.N., Truchon, A.R., Gann, E.R., and Wilhelm, S.W. (2022b). Draft Genome
Sequence of the Freshwater Diatom Fragilaria crotonensis SAG 28.96. Microbiology
Resource Announcements, e00289-00222.
Zepernick, B.N., Wilhelm, S.W., Bullerjahn, G.S., and Paerl, H.W. (2022c). Climate change and
the aquatic continuum: A cyanobacterial comeback story. Environmental Microbiology
Reports.

160
Appendix
Table 4.1 Statistical analysis of F. crotonensis morphological features (mean area, mean
biovolume, mean length, mean width, mean roughness, and mean avg. green) collected at Tf of
the morphology assay as a function of pH. Statistical analyses performed using unpaired two-
tailed t-tests.

Treatment comparison Summary P value

pH 7.7 vs. pH 9.2 Area (µm2) ** p=0.0074
pH 7.7 vs. pH 9.2 Biovolume (µm3) * p=0.0310
pH 7.7 vs. pH 9.2 Length (µm) ** p=0.0017
pH 7.7 vs. pH 9.2 Width (µm) ns p=0.0553
pH 7.7 vs. pH 9.2 Roughness *** p=0.0002
pH 7.7 vs. pH 9.2 Avg. Green Content **** p<0.0001

Table 4.2 Statistical analysis of F. crotonensis filament concentrations and Chlorophyll a

autofluorescence collected at Tf of the photopigment assay as a function of pH. Statistical
analyses performed using unpaired two-tailed t-tests

Treatment comparison Summary P value

pH 7.7 vs. pH 9.2 filament conc. (filaments·mL-1) ns p=0.1745
pH 7.7 vs. pH 9.2 Chlorophyll a (fsu) **** p<0.0001

Table 4.3 Statistical analysis of F. crotonensis normalized photopigment concentrations

(Chlorophyllide a, Chlorophyll a, Chlorophyll c1c2, Total Chlorophyll a, β-carotene,
Violaxanthin, Diadinoxanthin, Fucoxanthin, Neoxanthin) and photopigment ratios (Total
Carotenoids/Total Chl a, β-car/ Total Chl a, Viola/ Total Chl a, Diadino/Total Chl a, Total Chl
a/Chl c1c2) collected at Tf of the photopigment assay as a function of pH. Statistical analyses
performed using unpaired two-tailed t-tests.

Treatment comparison Summary P value

pH 7.7 vs. pH 9.2 Chlorophyllide a (fg·µm fil-1) ns p=0.8657
pH 7.7 vs. pH 9.2 Chlorophyll a (fg·µm fil-1) ns p=0.3263
pH 7.7 vs. pH 9.2 Chlorophyll c1c2 (fg·µm fil-1) ns p=0.9484
pH 7.7 vs. pH 9.2 Total Chlorophyll a (fg·µm fil-1) ns p=0.3503
pH 7.7 vs. pH 9.2 β-carotene (fg·µm fil-1) ** p=0.0015
pH 7.7 vs. pH 9.2 Violaxanthin (fg·µm fil-1) ** p=0.0090
pH 7.7 vs. pH 9.2 Diadinoxanthin (fg·µm fil-1) * p=0.0189
pH 7.7 vs. pH 9.2 Fucoxanthin (fg·µm fil-1) ns p=0.4594
pH 7.7 vs. pH 9.2 Neoxanthin (fg·µm fil-1) ns p=0.3307

pH 7.7 vs. pH 9.2 Total Carotenoids/Total Chl a ns p=0.2207

pH 7.7 vs. pH 9.2 β-car/ Total Chl a *** p=0.0002
pH 7.7 vs. pH 9.2 Viola/ Total Chl a ** p=0.0093
pH 7.7 vs. pH 9.2 Diadino/Total Chl a *** p=0.0008
pH 7.7 vs. pH 9.2 Total Chl a/Chl c1c2 ns p=0.4335

161
Table 4.4 Statistical analysis of F. crotonensis PhytoPAM analyses (Fv ⋅ Fm-1, ETRmax, Ik)
collected at Tf of the photo physiology assay as a function of pH. Statistical analyses performed
using unpaired two-tailed t-tests.

Table 4.4 Statistical analysis of F. crotonensis PhytoPAM analyses (Fv ⋅ Fm-1, ETRmax, Ik)
collected at Tf of the photo physiology assay as a function of pH. Statistical analyses
performed using unpaired two-tailed t-tests.
Treatment comparison Summary P value
-1
pH 7.7 vs. pH 9.2 Fv · Fm ns p=0.3089
- 2
pH 7.7 vs. pH 9.2 ETRmax (µmol e /m ·s) **** p<0.0001
pH 7.7 vs. pH 9.2 Ik (µmol photons/ m2·s) **** p<0.0001

162
Figure 4.7 Non-metric Multidimensional analysis (nMDS) of similarity between pH 7.7 and pH
9.2 transcriptomes (TPM). pH 7.7 replicates are indicated by black triangles, pH 9.2 replicates
are indicated by inverted green triangles.

163
Figure 4.8 Top 50 genes contributing to mean pH treatment dissimilarity (~27%) determined by
Similarity Percentages (SIMPER). Genes categorized in COG category photosynthesis depicted
in green, genes categorized in cell cycle control, cell division, chromosome partitioning depicted
in blue, genes categorized in cell wall, membrane, envelope biogenesis depicted in orange, and
genes categorized in mobilome: transposons; prophages depicted in yellow. Only annotated
genes were included in final reports (A) Contribution to dissimilarity (%) and cumulative
dissimilarity (%) of the top 50 annotated genes driving mean dissimilarity between pH 7.7 and
pH 9.2 normalized (TPM) expression values. (B) Mean abundance (TPM) of top 50 genes
driving dissimilarity within the pH 7.7 replicates. (C) Mean abundance (TPM) of top 50 genes
driving dissimilarity within pH 9.2 replicates.

164
Figure 4.9 Volcano plots of differentially expressed genes at pH 9.2 compared to pH 7.7. (A) All
genes within the pH transcriptome (B) All 713 DE genes according to the statistical cutoff (FDR-
corrected p-value ≤ 0.05, log2 |fold-change| > 2) (C) All 435 DE and annotated genes with an
assigned function used for all downstream analyses in this study.

165
Figure 4.10 Heat map depicting differentially expressed genes relating to the “Carbohydrate
transport and metabolism” COG category. All TPM values were row z-scored, with increases in
proportional transcript abundance indicated in yellow, and decreases in proportional transcript
abundance indicated in blue. The sum of transcripts across all treatments (LogTPM) is indicated
for each gene.

166
Figure 4.11 Flow cytometry and fluorometer analyses corresponding to the pigment pH assay.
(A) F. crotonensis concentration (filaments • mL-1) at the beginning (Ti) and end of the
experiment (Tf). pH 7.7 replicates denoted by open, black squares. pH 9.2 replicates denoted by
open, green squares. (B) F. crotonensis Chl a autofluorescence at (Ti) and (Tf). pH 7.7 replicates
denoted by open, black squares. pH 9.2 replicates denoted by open, green squares. Mean values
indicated by solid vertical bars, with variability (SEM) indicated.

167
Figure 4.12 Total Chlorophyll a photopigment data concentrations collected at (Tf). pH 7.7
replicates denoted by open, black squares. pH 9.2 replicates denoted by open, green squares. (A)
Net Chlorophyll a pigment concentration (µg • L-1) of F. crotonensis filaments. (B) Normalized
Total Chlorophyll a pigment concentration (fg • µm fil-1) of F. crotonensis filaments.

168
Figure 4.13 Fucoxanthin and Neoxanthin photopigment pH assay results collected at (Tf). pH
7.7 replicates denoted by open, black squares. pH 9.2 replicates denoted by open, green squares.
(A) Fucoxanthin pigment concentration (fg • µm fil-1) of F. crotonensis filaments. (B)
Neoxanthin pigment concentration (fg • µm fil-1) of F. crotonensis filaments.

169
Figure 4.14 Ratio of total carotenoids:total Chlorophyll a pigment in F. crotonensis filaments.
pH 7.7 replicates denoted by open, black squares. pH 9.2 replicates denoted by open, green
squares.

170
Figure 4.15 Ratios of significantly altered carotenoid pigments:total Chlorophyll a pigments in
F. crotonenis filaments. pH 7.7 replicates denoted by open, black squares. pH 9.2 replicates
denoted by open, green squares. (A) Ratio of 𝛽𝛽-carotene: total Chlorophyll a pigment
concentration in filaments. (B) Ratio of Violaxanthin:total Chlorophyll a pigment concentration
in filaments. (C) Ratio of Diadinoxanthin:total Chlorophyll a pigment concentration in filaments.

171
Figure 4.16 Ratio of Chlorophyll a:Chlorophyll c1c2 pigment concentration in F. crotonensis
filaments.

172
CHAPTER V: DIATOM RESPONSES TO DECREASING ICE COVER IN LAKE ERIE

173
Publication Note
This chapter is a draft version of a manuscript to be submitted as a peer-reviewed, published
article in ISME Communications by Brittany N. Zepernick, Elizabeth R. Dension, Naomi E.
Gilbert, Emily E. Chase, Alexander R. Truchon, Robbie M. Martin, Thijs Frenken, William R.
Cody, Justin D. Chaffin, George S. Bullerjahn, Robert Michael L. McKay, and Steven W.
Wilhelm.

Samples were collected by RMLM, GB, TF and JC. RNA extractions and quality assessment
performed by BZ. Metatranscriptomic processing performed by BZ, NG, EC and LD using a
pipeline established by NG. Python scripts associated with metatranscriptomic pipeline were
written by AT. Statistical analyses and figures were made by RM and BZ. The first draft was
written by BZ. All authors contributed to the drafting of the current manuscript.

174
Abstract
The ecophysiology of phytoplankton communities beneath the ice has been a “black box”

of limnology for decades. Winter surveys conducted throughout 2007-2012 discovered prolific

winter diatom blooms embedded within and under the Lake Erie ice. These studies concluded ice

plays a vital role in winter diatom dynamics, and concluded prolific winter-spring diatom blooms

are a significant driver of summer hypoxia. Yet, due to logistical constraints the ecophysiology

of winter diatom blooms and the role of ice in their bloom cycle remains widely unstudied. We

opened this “black box” by using the first wide scale metatranscriptomic survey of the winter

Lake Erie water column. We demonstrated ice cover alters diatom bloom magnitude and

phylogeny. Further, we discovered polar centric diatoms significantly increased the expression of

fasciclin genes, which we hypothesize is a means to “raft” together and optimize light acquisition

during turbid conditions of the ice-free water column. The winter of 2023 was the lowest mean

annual ice cover for Lake Erie on record, indicating large-scale climatic changes are already

underway. Hence, it is necessary to investigate how diatoms respond to ice conditions of today to

predict how they will fare in a climatically altered tomorrow.

175
Introduction
Winter has historically been considered “ecologically unimportant” with winter

limnology only recently defined as a “New Frontier” (Powers and Hampton, 2016). Indeed,

winter was often considered a period of planktonic death and persistence rather than growth

(Sommer et al., 1986; Sommer et al., 2012). Yet, we now know this is far from the truth

(Ozersky et al., 2021). Active planktonic growth under the ice was first reported in Lake Erie in

the 1930’s (Chandler, 1940; Chandler, 1942), but has since remained largely unexplored. A

winter limnological survey conducted between 2007-2010 discovered dense blooms primarily

comprised of the centric, filamentous diatom Aulacoseira islandica (Coscinodiscophyceae)

distributed within and under the ice of Lake Erie (Twiss et al., 2012). This finding ignited

interest in the Lake Erie winter water column, with subsequent studies demonstrating ice-

associated communities were dominated by centric colonial diatoms such as A. islandica and

Stephanodiscus spp. (Mediophyceae), while pennate diatoms often formed < 1% of the diatom

community (D'souza, 2012; Twiss et al., 2012; Wilhelm et al., 2014; Beall et al., 2016; Edgar et

al., 2016). In fact, chlorophyll a (Chl a) concentrations during winter surpassed those of spring

(Twiss et al., 2012) and examinations of silica deposition in frustules demonstrated cells were

metabolically active (Saxton et al., 2012). Additional studies determined winter-spring diatom

biovolumes can surpass summer cyanobacterial biovolumes by 1.5- to 6-fold (Reavie et al.,

2016), with diatom blooms driving recurrent summer hypoxia (i.e., dead zones) within the

central basin of Lake Erie (Wilhelm et al., 2014; Reavie et al., 2016; Ozersky et al., 2021).

One proposed physiological contributor to the ecologic success of winter diatoms is their

ability to attach to ice cover via symbiotic interactions with ice-nucleating bacteria, which allows

diatoms to co-locate themselves to the under-ice surface to maintain an optimal light climate for

photosynthesis (D'souza, 2012; D'souza et al., 2013). Another contributor to success includes

176
psychrophilic adaptations that increase membrane fluidity and enhance light-harvesting under

low-light conditions (Edgar et al., 2016). However, to date there is little known of winter diatom

bloom ecophysiology. This critical gap is largely attributed to the inability to maintain these taxa

(e.g. A. islandica and Stephanodiscus binderanus) in culture (D'souza, 2012). To circumvent

this, recent studies have turned to molecular approaches to characterize in situ winter diatom

communities (Beall et al., 2016; Edgar et al., 2016). These studies offer further support that ice

cover plays a critical role in shaping winter diatom ecophysiology, with Beall et al., (Beall et al.,

2016) noting diatom abundances of ice-associated A. islandica declined during the low-ice

winter of 2012. While this study indicated centric, filamentous diatom taxa significantly decline

during years of low ice cover due to wind-induced turbidity/mixing, it did not identify species-

specific intracellular mechanisms behind this trend. Elucidating the ecophysiological response of

the winter diatom community to ice coverage is of present-day importance, as Lake Erie is

experiencing unprecedented declines in ice cover due to climate change (Mason et al., 2016;

Wang et al., 2018; Ozersky et al., 2021). Indeed, projections suggest ice cover may disappear

entirely by the end of the century (Filazzola et al., 2020). Given the keystone role of diatoms in

freshwater food webs and global biogeochemical cycles (Nelson et al., 1995; Struyf et al., 2009;

Rühland et al., 2015; Benoiston et al., 2017), there is a pressing need to elucidate current winter

diatom ecophysiology in order to better project how they will respond to a climatically altered

future.

In this study, we employed metatranscriptomics to investigate how Lake Erie winter

diatom communities respond to ice cover vs. ice-free conditions. To our knowledge, this is the

first large-scale bioinformatic assessment of the winter Lake Erie community. Driven by

collaborative efforts with the U.S. and Canadian Coast Guards (McKay et al., 2011),

177
opportunistic samples were collected throughout 2019 and 2020 yielding winter samples

collected from both the ice-covered (2019) and ice-free (2020) winter water column (Zepernick

et al., 2022a). The survey included spring samples that serve as an outgroup. From this we offer

new hypotheses on the dynamics of winter diatom communities and their physiological response

to climate-induced variations in lake surface ice.

Methods
Lake Erie winter-spring water column sampling

Samples of opportunity (n = 77) from the Lake Erie planktonic community were

collected across temporal, spatial, and climatic gradients throughout the winter of 2019 and

2020. This large-scale collaborative effort included multiple surveys conducted by USCGC Neah

Bay, CCGS Limnos and M/V Orange Apex, resulting in a large metatranscriptomic dataset

(Zepernick et al., 2022a). Prior to sample collection, water column physiochemical parameters

were recorded along with meteorological conditions and ice cover observations. Briefly, water

samples were collected from 0.5 m below the surface and processed for analyses of dissolved

and particulate nutrients (mg L-1), size-fractionated (<0.22-𝜇𝜇m and <20 𝜇𝜇m) Chl a biomass (𝜇𝜇g

L-1), phytoplankton taxonomy and enumeration (Cells L-1), and total community RNA. Class-

level taxonomic assignments (i.e., Mediophyceae, Coscinodiscophyceae) of diatom microscopy

results were performed according to the Barcode of Life Data System (BOLD) (Ratnasingham

and Hebert, 2007). Metadata are available online at the Biological and Chemical Oceanography

Data Management Office (BCO-DMO) (Bullerjahn et al., 2022). Refer to Appendix

Methods/Results for further detail.

178
RNA extraction and sequencing

RNA extractions were performed using previously described acid phenol-chloroform

methods with ethanol precipitation (Martin and Wilhelm, 2020). Residual DNA in samples was

digested via a modified version of the Turbo DNase protocol using the Turbo DNA-free kit

(Ambion, Austin, TX, USA). Samples were determined to be DNA-free via the absence of a

band in the agarose gel after PCR amplification using 16S rRNA primers as previously reported

(Zepernick et al., 2022a). Samples were quantified using the Qubit RNA HS Assay Kit

(Invitrogen, Waltham, MA, USA) and sent to the Department of Energy Joint Genome Institute

(DOE JGI) for ribosomal RNA reduction and sequencing using an Illumina NovaSeq S4 2 ×

151-nucleotide indexed run protocol (15 million 150-bp paired-end reads per library) as reported

previously (Zepernick et al., 2022a).

Metatranscriptomic analysis

Filtering and trimming of raw reads was performed by DOE JGI using BBDuk (v.38.92)

and BBMap (v.38.86) (Bushnell, 2014a; Clum et al., 2021). Bioinformatic processing was

conducted using a prior-established metatranscriptomic workflow (Gilbert et al., 2022). Trimmed

and filtered libraries (n = 77) were concatenated and assembled (co-assembled) using MEGAHIT

(v.1.2.9) (Li et al., 2016). Co-assembly statistics were determined via QUAST QC (v.5.0.2)

(Gurevich et al., 2013). Trimmed reads were mapped to the co-assembly using BBMap (default

settings) (v.38.90) (Bushnell, 2014a). Gene predictions within the co-assembly were called using

MetaGeneMark (v.3.38) (Zhu et al., 2010) using the metagenome style model. Taxonomic

annotations of predicted genes were determined using the MetaGeneMark protein file, EUKulele

(v.1.0.6) (Krinos et al., 2020) and the PhyloDB database (v.1.076). Genes were functionally

179
annotated using eggNOG-mapper using a specified e-value of 1e-10 (v.2.1.7) (Cantalapiedra et

al., 2021). Following, featureCounts (Liao et al., 2014) within the subread (v.2.0.1) package was

used to tabulate read counts to predicted genes. Mapped reads were normalized to TPM

(Transcripts Per Million), representing relative “expression” values.

To investigate transcriptional patterns of the winter diatom bloom community, we

focused on a subset of libraries (n=20) selected for consistency in sample collection methods

(whole water filtration) and diatom abundances (Appendix Table 5.1). Thus, all data reported

hereafter pertains to these 20 libraries. Raw data for all 77 transcriptomic libraries are available

at the JGI Data Portal (https://data.jgi.doe.gov) under Proposal ID 503851 (Zepernick et al.,

2022a). Refer to Appendix Methods/Results and Appendix Attachments for further detail.

Phylogenetic analysis

A phylogenetic tree of fasciclin containing domains (proteins of interest) was produced

using differentially expressed (DE) putative proteins of this study (n=18), domains recovered

from the eggNOG orthology database and publicly available domains from NCBI (Wheeler et

al., 2007). A custom database was curated using all NCBI diatom proteins. A DIAMOND

(v.2.0.15) (Buchfink et al., 2015) blastp alignment was performed with putative fasciclin proteins

and eggNOG domains against the diatom database to recover all putative diatom fasciclin

domains. The recovered domains were then aligned (DIAMOND blastp) against the NCBI non-

redundant database. These results were compiled and collapsed to 80% similarity using CD-HIT

(v.4.7) (Fu et al., 2012) and a multiple sequence alignment was performed using MAFFT

(v.7.310) (Katoh and Standley, 2013) with 500 iterations. Gaps were closed using trimAl with

gappyout (v.1.4.rev15) (Capella-Gutiérrez et al., 2009) and examined using AliView (v.1.28)

180
(Larsson, 2014). A phylogenetic tree (1000 bootstraps) was constructed using a model test

selecting for a general non-reversible Q matrix model estimated from Pfam database (v. 31) (El-

Gebali et al., 2019) with a gamma rate heterogeneity. The consensus tree was visualized using

iTOL (Letunic and Bork, 2019). Refer to Appendix Methods/Results and Appendix Attachments

for further detail.

Statistical analyses

Comparisons of water column physiochemical features by ice cover were made in Prism

(v. 9.3.1) via two-tailed unpaired t-tests. Variability in expression (TPM) between transcriptomic

libraries was assessed via ANalysis Of Similarities (ANOSIM) and Similarity Percentage

(SIMPER) analyses using PRIMER (v.7) (Clarke and Gorley, 2015). Bray-Curtis similarity and

non-metric multi-dimensional scaling (nMDS) were performed in R. Differential expression

(DE) of transcript abundance was performed using DESeq2 in R (v.1.28.1) (Love et al., 2014).

Genes with an absolute log2 fold change (Log2FC) >2 and adjusted p-value of < 0.05 were

considered differentially expressed. Z-scores reported in heat maps were calculated by

heatmapper.ca (Clustering method: Average linkage, Distance measurement method: Pearson)

(Babicki et al., 2016)) using the DESeq2 variance stabilizing transformed values (VST) (Babicki

et al., 2016). Refer to Appendix Methods/Results for further detail.

Results

Physiochemical profiles and winter community characterization

Samples were collected across 12 sites throughout the central basin of Lake Erie with true

biological replication at a subset of stations (Figure 5.1A) (Appendix Table 5.1). Temporally,

181
Figure 5.1: Spatial and climatic variability across samples. (A) Sample sites across Lake Erie
visited throughout winter-spring 2019 and 2020. (B) Historical trends in Lake Erie mean annual
maximum ice cover (%). Open circles are years that (to our knowledge) do not have peer-
reviewed published survey data. Solid black circles are years that were previously surveyed in
prior published studies. Solid blue circles are years sampled in this study. Figure adapted from
data retrieved from NOAA GLERL database (NOAA-GLERL).

182
the samples span February-March 2019 and February-June 2020, yielding 14 winter and 6 spring

libraries. Climatically, the winter of 2019 was a year of high ice cover (mean maximum ice cover

of 80.9%), whereas winter 2020 was a year of negligible ice cover (mean maximum ice cover of

19.5%) (NOAA, 2021) (Figure 5.1B). Libraries 1-4 were collected during ice cover (ranging

from 2.54 - 15.24 cm in thickness) while winter libraries 5-14 were collected during no ice cover

conditions. Winter lake surface temperatures ranged from ~0-6 °C across sample sites (Appendix

Figure 5.8A). Overall, nutrient concentrations at ice-covered sites were not significantly different

from ice-free sites save for nitrate (Appendix Figure 5.8B-H). While not significant (p ≥ 0.13),

the highest total Chl a concentrations (> 0.22 𝜇𝜇m) coincided with ice cover (Figure 5.2A, B).

The larger sized-fraction of phytoplankton contributed an average of 70% (+/- 27%) to total Chl

a during ice cover and 50% (+/- 13%) in ice-free winter sites (Appendix Figure 5.9), but the

differences were not significant (p = 0.22). Cell concentrations of diatoms (Bacillariophyta)

dominated the winter water column regardless of ice conditions, with other eukaryotic

phytoplankton (e.g., Chlorophyta, Cryptophyta, and Dinophyta) present at concentrations 1-2

orders of magnitude lower (Appendix Figure 5.10). While Bacillariophyta concentrations

decreased slightly at ice-free sites (p = 0.33), Dinophyta concentrations significantly increased (p

= 0.03), with Cryptophyta and Chlorophyta exhibiting similar trends (p ≥ 0.05). Overall, centric

diatoms dominated the winter diatom community while pennate diatoms were found at

concentrations an order of magnitude lower (Figure 5.2C, D). Despite this dominance, centric

diatoms demonstrated a decreasing trend in ice-free samples while pennate diatoms exhibited

significant increases in ice-free samples (p = 0.03) albeit remaining at low abundances. Cell

concentrations of the centric bloom formers Stephanodiscus spp. (Mediophyceae) and A.

islandica (Coscinodiscophyceae) were generally highest during ice cover (Figure 5.2E, F).

183
Figure 5.2: Characterization of the biotic community across the 12 Lake Erie sample sites.
Samples are organied on the x-axis by season (W = winter, S = spring) and year. Solid shapes
indicate the sample was collected during ice cover (2019) open shapes indicate the sample was
collected during no ice cover (2020). Ice cover samples are indicated by a blue asterisk. (A)
Total Chlorophyll a concentration of the whole water column community (i.e., >0.22 μm in size)
(μg L-1) (B) Chlorophyll a concentration of the large size fractioned community (i.e., >20
μm in size) (μg L-1). (C) Cell abundances (Cells ⋅L-1) of centric diatoms (Stephanodiscus spp. +
A. islandica + Small centric diatoms of 5-20 μm). (D) Cell abundances of pennate diatoms
(Fragilaria spp. + Asterionella formosa + Nitzschia spp). (E) Cell abundances (Cells ⋅L-1) of
Stephanodiscus spp., a common winter bloom-forming taxon (Mediophyceae class). (F) Cell
abundances (Cells ⋅L-1) of A. islandica, a common winter diatom bloom-forming taxon
(Coscinodiscophyceae class).

184
Notably, Stephanodiscus spp. concentrations were significantly higher than A. islandica in ice-

covered samples (p = 0.03), yet not significantly greater than A. islandica in ice-free samples (p

= 0.19) (Appendix Figure 5.11). Concentrations of Stephanodiscus spp., were significantly

reduced by ~50% in ice-free samples, (p = 0.005) (Figure 5.2E), while A. islandica abundances

were not (p = 0.14) (Figure 5.2F). Further, while concentrations of small centric diatoms (5-20

µm size) were not detected in ice covered samples, they were found to range from ~300-3,000

cells L-1 in ice-free samples (Appendix Figure 5.12). While small centric diatom taxa accounted

for ~83% of the winter diatom community at site 8, they otherwise contributed an average of

26% to the total diatom community in ice-free samples (Appendix Figure 5.13). Refer to the

Appendix Material for further detail (Appendix Attachments 5.1D-G).

Transcriptomic response of winter diatom community to ice cover

While the number of raw reads mapping to Eukaryota decreased in ice-free samples,

every other domain increased in proportional read abundance (Appendix Figure 5.17). Notably,

the percentage of reads mapping to Eukaryota were higher than Prokaryotes within ice-cover but

not ice-free samples. Diatoms dominated the winter transcriptional pool across major eukaryotic

phytoplankton communities regardless of ice cover (Figure 5.3A). In turn, polar centric diatoms

(Mediophyceae) dominated diatom community transcription regardless of ice cover (Figure

5.3B). Reference the Appendix Material (Appendix Methods/Results, Appendix Figures 5.15-

5.20, Appendix Attachments 5.1G-O) for further detail on taxonomically resolved transcript

distributions.

Normalized expression (TPM) profiles of the total water column community displayed

clustering by ice cover (Figure 5.4A), with SIMPER analyses demonstrating an average

185
Figure 5.3: Relative transcript abundance of major eukaryotic phytoplankton taxa and diatom
classes. Libraries/samples are listed in chronological order of sample date on x-axes, with
biological replicates joined by grey horizontal bars. Ice cover samples are indicated by a blue
asterisk. (A) Relative transcript abundance of MEPT. All groups which formed <5% of the total
mapped reads are included within “Other” (Amoebozoa, Hilomonadea, Excavata, Rhizaria, NA).
(B) Relative transcript abundance of Bacillariophyta classes Mediophyceae,
Coscinodiscophyceae, Bacillariophyceae, Fragilariophyceae and those not annotated (NA).

186
Figure 5.4: Dissimilarity (Bray-Curtis Matrix) clustering of the 20 metatranscriptomic libraries
normalized expression values (TPM). (A) nMDS of the entire water column community
expression, stress value = 0.0633. (B) nMDS of the Bacillariophyta community expression,
stress value= 0.0497. Samples are presented as follows: February = squares, March = triangles.
May = diamonds, June = circles. Blue indicates the sample was collected during the winter,
black indicates the sample was collected during the spring. Solid shapes indicate the sample was
collected during ice cover (2019) open shapes indicate the sample was collected during no ice
cover (2020).

187
dissimilarity of 64% between ice cover and ice-free winter libraries (Appendix Attachment

5.1P). ANOSIM tests confirmed ice strongly affected winter community expression (R = 0.87, p

= 0.002) (Appendix Figure 5.21A) (Appendix Attachment 5.1Q). Surprisingly, diatom

community expression did not strongly cluster by ice cover (Figure 5.4B), with SIMPER

analyses indicating an average dissimilarity of 47% between ice cover and ice-free libraries

(Appendix Attachment 5.1T). ANOSIM tests confirmed ice cover exerts a lesser influence on

winter diatom community expression overall compared to the full water column community (R =

0.282, p = 0.059) (Appendix Figure 5.21B). In contrast, season had a strong effect on diatom

expression (SIMPER Average dissimilarity = 77%; ANOSIM R = 0.927, p = 0.001) (Appendix

Attachments 1V, W).

To investigate how ice cover contributed to the ~50% dissimilarity in winter diatom

expression, differential expression analyses were performed. These results indicated 354 genes

belonging to putative Bacillariophyta were differentially expressed (|Log2FC| ≥ 2, padj < 0.05),

with 311 of these genes increasing in relative expression in ice free samples and 43 decreasing

(Appendix Attachment 5.1X). The Mediophyceae class had the highest representation within the

DE genes, comprising ~50% of DE genes while other classes formed a net total of ~10%

(NA~40%) (Appendix Figure 5.22A). Further analysis revealed 33% of the Mediophyceae DE

genes were annotated as Chaetoceros-like (Appendix Figure 5.22B), despite Chaetoceros-like

genes forming ≤ 10% of mapped reads throughout the winter libraries (Appendix Figure 5.16).

Here, “Chaetoceros-like” indicates the transcriptomes were annotated with marine databases due

to a lack of comprehensive freshwater taxonomic databases. Interestingly, genes categorized in

the COG category C (Energy production and conversion) were the second highest represented

COG category within the DE dataset, with most genes exhibiting increased transcript

188
representation within ice-free diatom communities (Figure 5.5). Of these genes, 64% belonged to

Mediophyceae (Figure 5.5, Appendix Figure 5.23). Notably, the expression of genes encoding

for iron-containing photosynthetic proteins such as cytochromes (CoxN, CtaC, Cob, NuoB,

Cox1, Cox2) increased in relative expression in ice-free communities (Figure 5.5) along with

other photosynthetic genes (ferrodoxin-petF, flavoprotein-etfA, ferritin-ftnA, and photosystem II-

psbA). Likewise, relative expression of genes within COG category P (Inorganic ion transport

and metabolism) increased in ice-free samples (Appendix Figure 5.24), with expression of 6

putative iron transporting genes (tonB_1-3 and OMFeT_1-3) increasing in ice-free communities.

DE genes within COG category P largely belong to the Mediophyceae class, comprising ~40%

of the annotated genes (Appendix Figure 5.24, 5.25). COG category G (Carbohydrate transport

and metabolism) genes also increased in expression in ice-free communities (Appendix Figure

5.26). Notably, 4 genes encoding for pectinesterase (pecT, FBX011_1-3), which has a structural

role in plant cell walls and has been implicated with intercellular communication (Shin et al.,

2021), increased in expression. Further, two proton-pumping rhodopsin genes, which were

recently found to be an alternative to classical phototrophy in a cold-adapted freshwater

photosynthetic bacterium (Kopejtka et al., 2022), were increased in expression in ice-free diatom

communities (Appendix Figure 5.27).

DE analyses in response to season were performed with diatom libraries to identify trends

unique to the ice cover DE dataset. The top 10 COG categories represented in each dataset

overlapped except for COG category M, which was the third most abundant in ice cover analyses

compared to the twelfth most abundant in season analyses (Appendix Figure 5.28). Further

analysis of these COG M (Cell wall, membrane, and envelope biogenesis) genes revealed 58%

belonged to Mediophyceae (Figure 5.6A, Appendix Figure 5.29). Surprisingly, despite the large

189
Figure 5.5: Bacillariophyta transcript abundance patterns in response to ice cover-COG C. (A)
Taxonomic distribution of DE genes categorized within COG category C (Energy Production
and Conversion). (B) COG assignments for all 354 DE genes in response to ice cover, with COG
category C indicated in blue. (C) Heatmap depicting COG category C differentially expressed
gene expression (VST) in response to ice cover across the 14 winter libraries.

190
Figure 5.6: Bacillariophyta transcript abundance patterns in response to ice cover-COG M. (A)
Taxonomic distribution of DE genes categorized within COG category M (Cell wall, membrane,
envelope biogenesis). (B) COG assignments for all 354 DE genes in response to ice cover, with
COG category M indicated in blue. (C) Heatmap depicting COG category M differentially
expressed gene expression (VST) in response to ice cover across the 14 winter libraries.

191
abundance of A. islandica (Cosinodiscophyceae) according to cell counts, transcripts belonging

to the Coscinodiscophyceae class were not represented within DE COG M genes. Intriguingly,

50% of the DE COG M genes encode for fasciclins (FASDP, TGFBI, FASR, FASC) (Figure

5.6B, C) also called the FAS1 domain. Fasciclins are secreted glycoproteins involved in diatom

cell-cell adhesion and cell-extracellular matrix adhesion (Willis et al., 2014; Lachnit et al., 2019).

Expression of fasciclins increased during ice-free periods and decreased during ice cover, with

all 18 DE fasciclin genes taxonomically assigned to Mediophyceae or unresolved. Phylogenetic

analyses indicated diatoms horizontally acquired FAS1 from bacteria, as there is evidence for at

least 6 instances of horizontal gene transfer within our analysis (Figure 5.7). Broadly, the FAS1

domain is widely distributed in diatoms, with ~140 marine and freshwater diatoms found to

contain this protein domain including cold-adapted diatom such as Fragilariopsis cylindrus and

Thalassiosira antarctica. Refer to Appendix Methods/Results and Appendix Attachments for

further details.

Discussion

The winter period in lakes remains a limnological “blind spot” due to a lack of research

(Ozersky et al., 2021). Our investigations into this Lake Erie “black box” revealed another gap:

Lake Erie winter surveys conducted since the turn of the millennium were during (or directly

following) consecutive periods of high ice cover (mean maximum ice cover >80%) (Saxton et

al., 2012a; Twiss et al., 2012; Twiss et al., 2014; Wilhelm et al., 2014; Beall et al., 2016; Edgar

et al., 2016) (Figure 5.1B). Thus, foundations established by these studies do not necessarily

represent current or future communities or conditions. Historically, Lake Erie experienced 4-5

192
Figure 5.7: Phylogenetic tree of fasciclin distribution within diatoms. Bootstrap values above 70
are indicated with black lines. The FAS1 domain was found in 141 marine and freshwater
diatoms of diverse ecological habitats (indicated in blue). The 18 DE diatom fasciclins noted in
this studies metatranscriptomes are indicated in purple with asterisks. Fasciclins are most widely
distributed in bacteria (indicated in dark green). Notably, they were also found within
cyanobacteria (indicated in pink), chytrids (indicated in red), other algae (indicated in purple),
and other eukaryotes (indicated in light green), but at a much lesser extent.

193
consecutive years of high ice cover followed by one year of low ice cover (Figure 5.1B). Thus,

foundations established by these studies do not necessarily represent current or future

communities or conditions. Historically, Lake Erie experienced 4-5 consecutive years of high ice

cover followed by one year of low ice cover (Figure 5.1B). Yet since 2012, there has been a lack

of consecutive periods of ice cover. Notably, the winter of 2023 has unprecedented low ice

coverage (mean maximum ice cover 5% as of March 7th 2023), indicating large scale climatic

changes are underway (NOAA, 2023). In light of this, the present study investigated how winter

diatoms respond to ice cover. We demonstrate the winter diatom community has shifted from

observations made during the period 2007-2011. Comparisons between ice-cover and ice-free

samples demonstrate shifts in winter diatom abundance and phylogeny. Transcriptional trends

suggest key diatom genera elicit a physiological response to ice cover, primarily within the

diatom class Mediophyceae (polar centric). Notably, diatoms exhibited increased relative

expression of photosynthetic and fasciclin genes during ice-free conditions. These observations

lead to a new hypothesis regarding the ecophysiological role of fasciclins within ice-free winter

diatom communities. We provide this information couched within the context of the ecological

implications within a climatically altered future.

Declining ice cover alters winter diatom bloom magnitude and phylogeny

Our findings are in juxtaposition to prior Lake Erie winter studies (2007-2011) which

reported winter diatom communities overwhelmingly dominated by A. islandica, with

Stephanodiscus spp. present in lesser concentrations (Saxton et al., 2012a; Twiss et al., 2012;

Twiss et al., 2014; Wilhelm et al., 2014; Edgar et al., 2016). Our study demonstrated the inverse,

with cell abundances of Stephanodiscus spp. significantly higher than A. islandica in the ice-

194
covered community. We hypothesize a decrease in consecutive years of high ice cover may drive

this decline of A. islandica dominance. The breakdown of this historical trend appears to favor a

shift to Stephanodiscus spp. dominance within the larger, centric filamentous diatom community.

Broadly, abundances of these taxa were 1-2 magnitudes lower than previously reported by Twiss

et al., (Twiss et al., 2012), with Chl a concentrations supporting this trend. Hence, we suggest

the collective seasonal decline of ice cover may have contributed to this decrease in winter

diatom bloom magnitude. Yet, an alternative explanation exists for this decline in diatom

abundance observed in our study. We acknowledge our surface-sample grabs of the winter water

column do not reflect the full depth of the water column. Thus, we speculate our study may

underestimate the abundances of centric diatom taxa within the winter water column, noting the

well-mixed, isothermal conditions which come with ice-free winters would evenly distribute

algae throughout (Appendix Figure 5.30). In support of this, a 2007 survey noted concentrated

diatom communities were found ten meters below the surface of an ice-free area, while very

dilute seston was found at a nearby ice-covered area at six meters depth with the diatoms

concentrated within the ice cover (Appendix Figure 5.31). Cumulatively, this suggests ice cover

may not alter winter diatom abundance as previously suggested but alter diatom distribution

within the water column.

Surprisingly, Chl a concentrations did not significantly decrease relative to ice cover in

our study, though they exhibited a declining trend. Cell counts confirmed centric diatoms were

the primary eukaryotic phototrophs within winter samples regardless of ice cover, though

abundances were lower in ice-free samples. Indeed, a prior study reported significant declines in

A. islandica abundance during a low-ice winter in Lake Erie and suggested this niche would be

filled by cryptophytes and dinoflagellates better suited for the turbid water column (Beall et al.,

195
2016; Ozersky et al., 2021). While we observed higher abundances of these groups in ice-free

samples within our study, their cellular and transcriptional abundances remained below centric

diatoms by an order of magnitude. Hence, our results demonstrated that low ice cover during this

season did not induce significant large-scale phyla-level shifts in major eukaryotic phytoplankton

community composition. This suggests that future ice-free winter communities may remain

dominated by centric diatoms as observed in this study, but at a lesser magnitude.

While centric diatom abundances did not significantly differ by ice conditions, diatom

community composition exhibited significant changes. Cell abundances of Stephanodiscus spp.,

were ~50% lower in ice-free samples, resulting in ~equal abundances of Stephanodiscus spp. and

A. islandica within the ice-free water column. Further, small centric diatom taxa (5-20 µm size)

were absent in samples from ice-covered sites yet formed 10-82% of total diatom counts in ice-

free sites, with a bloom of these taxa noted at site 8. This suggests ice-free conditions may

increase populations of smaller, centric diatoms in future warmer and ice-free winters. The trend

is supported by prior studies which demonstrate warming temperatures decrease phytoplankton

cell size (Bramburger et al., 2017) and select for smaller taxa (Daufresne et al., 2009; Winder et

al., 2009). If these observations represent longer-term trends, future ice-free diatom communities

will be more diverse with lesser biomass. In addition, significant increases in pennate diatoms in

ice-free samples suggest lake warming/ice-free conditions will likely increase pennate diatoms

throughout the winter and further drive winter diatom community heterogeneity.

Key diatom genera transcriptionally respond to declining ice cover

Despite the effect of ice cover on winter diatom phylogeny, the transcriptionally active

diatom community did not exhibit substantial trends. Transcript abundances of diatoms were

196
lower in ice-free libraries yet reads mapping to diatom classes and the genera within did not

definitively change. Diatoms of the class Mediophyceae (polar centric) dominated transcript

abundances regardless of ice cover and formed ~50% of genes differentially expressed by ice

cover. In contrast, only ~10% of differential represented transcripts belonged to the other 3

diatom classes combined (NA = ~40%). Cumulatively, this suggests diatoms within

Mediophyceae elicit a strong physiological response to changes in ice cover. In contrast, though

A. islandica (Coscinodiscophyceae) was strongly present in cell counts, Coscinodiscophyceae

transcripts formed only ~10% of reads across winter diatom libraries and was nearly absent from

DE genes, suggesting this species did not respond to ice conditions.

One cautionary note arising from this study is the lack of sequenced freshwater diatom

taxa and an absence of freshwater taxonomic annotation databases. Edgar et al., (Edgar et al.,

2016) noted only 23% of the taxonomic diatom annotations within their Lake Erie

metatranscriptome could be tied to genera present within the Great Lakes. Hence, there may be

transcriptional changes within the winter diatom community which have gone undetected within

this study. Many of the annotations we generated were best aligned to marine counterparts (i.e.,

where reads are annotated as belonging to a genus “-like” genome). Further, sequence data is

often not coincident with classic taxonomy. We note numerous different taxonomic databases

conflict regarding diatom taxonomy, especially concerning which genera belong to

Mediophyceae (polar centric) vs. Coscinodiscophyceae (centric) classes. More broadly, Reavie

(2023) discussed the lack of studies regarding Great Lakes diatoms in general, citing there are

various undescribed and unclassified diatoms to date. This lack of clarity and consensus within

freshwater taxonomy is a knowledge gap which requires further attention. Beyond the taxonomic

gap, we observed a disconnect between A. islandica cell counts and transcript abundance in our

197
study. This suggests A. islandica cells were not transcriptionally active, and is in accordance

with other studies which report a stark juxtaposition between “who is there” vs. “who is

metabolically active” (Xie et al., 2020).

The most definitive trend to emerge from differential expression analyses by ice cover

was the increase in expression of putative fasciclins (COG M) within ice-free diatom

communities. Though these genes remain widely uncharacterized in diatoms, prior studies have

described fasciclins within the Bacillariophyceae genera Amphora coffeaeformis (Lachnit et al.,

2019) and Phaeodactylum tricornutum (Willis et al., 2014). Notably, both studies identified

fasciclin proteins within diatom-secreted exopolymer substance adhesion trails and concluded

these molecules facilitate diatom motility, adhesion, and aggregation. These studies built upon a

prior publication which first described algal fasciclins and referred to them as “cell adhesion

molecules” (Huber and Sumper, 1994). In this study, 58% of the DE diatom fasciclins belonged

to the class Mediophyceae (27% to the NA group). We also observed increased transcripts for

outer membrane proteins involved in cell adhesion and pectinesterases involved in cell signaling.

These observations have led us to hypothesize that diatoms within Mediophyceae increased

expression of cell adhesion and signaling genes to form large colonial “mats” that facilitate

buoyancy. Indeed, a similar “rafting” strategy has been well-documented in centric marine

diatoms Rhizosolenia spp. (Villareal et al., 1993; Villareal and Lipschultz, 1995; Villareal et al.,

1996; Villareal et al., 1999b) and Ethmodiscus spp. (Villareal and Carpenter, 1994; Villareal et

al., 1999a). These studies demonstrated rafts can be positively and negatively buoyant in

response to physiological stressors (Villareal et al., 1996; McKay et al., 2000; Villareal,

unpublished). Here, we propose Mediophyceae diatoms form colonial mats to optimize their

location within the ice-free water column and maintain photosynthetic quotas.

198
The rafting hypothesis: A role of fasciclins in the ice-free turbid water column

Ice cover has been suggested to create a niche that promotes diatom photosynthetic

processes via its high light transmittance, especially through snow-free ice which is common to

Lake Erie (Bolsenga and Vanderploeg, 1992; Twiss et al., 2012; D'souza et al., 2013). Indeed,

prior research has demonstrated diatom and bacterial communities promote the formation of

frazil ice, which “rafts” them throughout the water column to partition to the surface ice cover

where the light climate is enhanced (D'souza et al., 2013). In the absence of ice cover, winter

storms resuspend sediment (Chandler, 1944; Valipour et al., 2017) (Supplemental Figure 23),

with increased turbidity and deep-mixing posed to negatively affecting diatom light harvesting

(Vanderploeg et al., 1992; Vanderploeg et al., 2007; Beall et al., 2016; Ozersky et al., 2021). We

observed increased relative expression of photosynthetic genes and iron transporters within ice-

free diatom communities, suggesting efforts to optimize photosynthesis within the turbid water

column. Of these transcripts, 64% phylogenetically belonged to Mediophyceae (with NA =

27%). Phylogenetic analyses demonstrated diatoms acquired fasciclins from horizontal gene

transfer events with bacteria, and fasciclins were identified in ~140 marine and freshwater

diatoms including the model polar marine diatom Fragilariopsis cylindrus (Otte et al., 2023).

Hence, this further suggests a physiological role for these proteins in the frigid water column. As

a result, we hypothesize Mediophyceae diatoms were rafting via cell-adhesion fasciclins to

optimize their location within the turbid Lake Erie water column.

In a supporting twist to this theory, we also observed increases in the expression of genes

encoding proton-pumping rhodopsins within the ice-free diatom community. Proton-pumping

rhodoposins were recently found to be an alternative to classical phototrophy in a freshwater

photosynthetic bacterium (Kopejtka et al., 2022), and they have been characterized within

199
marine diatoms (Marchetti et al., 2015) and dinoflagellates (Slamovits et al., 2011). Globally, it

is thought rhodopsin driven, retinal-based phototrophy is a major marine light harvesting process

(Gomez-Consarnau et al., 2019). Our observations suggest that the role of these proton-pumping

rhodopsins within fresh waters demands more attention (Sharma et al., 2009), as it is possible

that in future scenarios (less ice cover, more turbidity) they may serve as important evolutionary

selectors.

Implications of an ice-free future for the Lake Erie winter community

Our results suggest declining ice cover will have ecological and biogeochemical

consequences for future winter communities. Using relative transcript abundance as a proxy for

activity, our results suggest eukaryotes were more active than bacteria in ice covered samples but

not ice-free samples. This confirms observations made by Beall et al. (2016), which suggested

the Lake Erie winter water column directly contradicts the well-established paradigm that

bacteria dominate eukaryotes in terms of sheer cell abundance (Oh et al., 2011; Eiler et al., 2013;

Mou et al., 2013; Eiler et al., 2014). We build upon this central finding by demonstrating bacteria

appear to be less transcriptionally active in summa in the ice-covered winter water column:

observations in concurrence with reduced bacterial production rates that have been previously

measured in winter-collected samples (Wilhelm et al., 2014). Indeed, a decline in eukaryotic

phototrophic winter dominance could induce substantial biogeochemical and physiochemical

alterations within the water column (i.e., higher CO2 and lower O2 concentrations, etc.).

Beyond the effect of ice cover on community structure, we observed alterations to winter

diatom bloom magnitude and phylogeny. We noted consistently lower Chl a concentration, total

diatom cell abundance, and total diatom transcript abundance within the ice-free water column

200
(albeit at p ≥ 0.1285). Considering summer hypoxia is in large part fueled by winter-spring

diatom blooms (Wilhelm et al., 2014; Reavie et al., 2016), the lower diatom abundance of ice-

free winters could be favorable by decreasing oxygen depletion. Moreover, the associated change

in diatom community composition has ecological implications as grazers and filter-feeders

preferentially select for smaller and less silicified diatom taxa (Reavie and Barbiero, 2013;

Pančić et al., 2019). Thus, if ice decline in the future leads to a shift to small centric and pennate

diatoms, we may observe food web alterations at its base.

During periods of ice cover, large filamentous winter diatoms embed themselves within

this ice cover and rely on frazil ice to optimize their location in the water column (Twiss et al.,

2012; D'souza et al., 2013). Currently, we hypothesize diatoms are increasing the expression of

fasciclin-encoding genes to raft and evade light limitation in response to declining ice cover.

Ozersky et al. (2021) suggested warmer winters serve to induce a change in the Great Lakes

mixing regime, shifting from dimictic mixing patterns to a warm monomictic mixing pattern

characterized by continuous isothermal conditions throughout winter. Hence, this fasciclin-

induced rafting would be of increasing importance in future winter diatom survival within the

warming and ice-free water column. Nevertheless, with diatoms previously described as “one of

the most rapidly evolving eukaryotic taxa on Earth” (Oliver et al., 2007; Vardi et al., 2009) and

prone to promiscuous horizontal gene transfer events (Dorrell et al., 2021), it would be surprising

if they failed to adapt to an ice-free future.

Ultimately, we cannot place the consequences of the observations we describe in a

quantitative framework. As noted, diatoms are a key component of the food web and lake

biogeochemical function, and thus it is anticipated that changes to this group will reshape the

entire ecosystem. To this end, our observations, which demonstrate variability associated with

201
conditions consistent with projected future climate scenarios, carve out a critical path forward

and provide cautionary insight of what may be yet to come in lakes such as Lake Erie.

Acknowledgements
We are grateful to the command and members of the U.S. Coast Guard Cutter Neah Bay,

Canadian Coast Guard Ship Limnos and M/V Orange Apex for their help with sample collection

and the generation of hydrochemistry data. We thank Daniel H. Peck, James T. Anderson, Derek

Niles and Arthur Zastepa for their help with sample collection and pre-processing. We thank

Christa Pennacchio with the JGI for her coordination and sequencing expertise. We thank Dr.

Gary LeCleir, Katelyn Houghton, Dr. Erik Zinser, and Dr. Jill Mikucki for their comments and

suggestions. This work was funded by the National Institutes of Health, NIEHS grant

1P01ES023-28939-01, National Science Foundation grant OCE-1840715 (GB, RMLM, JC,

SW), NSERC grant RGPN-2019-03943 (RMLM), and JGI project 503851 (SW and RMLM).

The work conducted by the U.S. DOE JGI is under contract DE-AC02-05CH11231.

202
References:
Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J.R., Maciejewski, A., and Wishart, D.S.
(2016). Heatmapper: web-enabled heat mapping for all. Nucleic acids research 44,
W147-W153.
Beall, B., Twiss, M., Smith, D., Oyserman, B., Rozmarynowycz, M., Binding, C., Bourbonniere,
R., Bullerjahn, G., Palmer, M., and Reavie, E. (2016). Ice cover extent drives
phytoplankton and bacterial community structure in a large north‐temperate lake:
implications for a warming climate. Environmental microbiology 18, 1704-1719.
Benoiston, A.-S., Ibarbalz, F.M., Bittner, L., Guidi, L., Jahn, O., Dutkiewicz, S., and Bowler, C.
(2017). The evolution of diatoms and their biogeochemical functions. Philosophical
Transactions of the Royal Society B: Biological Sciences 372, 20160397.
Bolsenga, S., and Vanderploeg, H. (1992). Estimating photosynthetically available radiation into
open and ice-covered freshwater lakes from surface characteristics; a high transmittance
case study. Hydrobiologia 243, 95-104.
Bramburger, A.J., Reavie, E.D., Sgro, G., Estepp, L., Chraïbi, V.S., and Pillsbury, R. (2017).
Decreases in diatom cell size during the 20th century in the Laurentian Great Lakes: a
response to warming waters? Journal of Plankton Research 39, 199-210.
Buchfink, B., Xie, C., and Huson, D.H. (2015). Fast and sensitive protein alignment using
DIAMOND. Nature methods 12, 59-60.
Bullerjahn, G.S., Anderson, J.T., and Mckay, R.M. (2022). "Winter survey data from Lake Erie
from 2018-2020". 3 ed. (Biological and Chemical Oceanography Data Management
Office (BCO-DMO)).
Bushnell, B. (2014). "BBMap: a fast, accurate, splice-aware aligner". Lawrence Berkeley
National Lab.(LBNL), Berkeley, CA (United States)).
Cantalapiedra, C.P., Hernández-Plaza, A., Letunic, I., Bork, P., and Huerta-Cepas, J. (2021).
eggNOG-mapper v2: functional annotation, orthology assignments, and domain
prediction at the metagenomic scale. Molecular biology and evolution 38, 5825-5829.
Capella-Gutiérrez, S., Silla-Martínez, J.M., and Gabaldón, T. (2009). trimAl: a tool for
automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25,
1972-1973.
Chandler, D.C. (1940). Limnological studies of western Lake Erie: I. Plankton and certain
physical-chemical data of the Bass Islands region, from Septbember 1938, to November,
1939. Ohio Journal of Science 40, 291-336.
Chandler, D.C. (1942). Limnological Studies of Western Lake Erie. III, Phytoplankton and
Physical-Chemical Data from November, 1939, to November 1940.
Chandler, D.C. (1944). Limnological studies of western Lake Erie IV. Relation of limnological
and climatic factors to the phytoplankton of 1941. Transactions of the American
Microscopical Society 63, 203-236.
Clarke, K., and Gorley, R. (2015). Getting started with PRIMER v7. PRIMER-E: Plymouth,
Plymouth Marine Laboratory 20.
Clum, A., Huntemann, M., Bushnell, B., Foster, B., Foster, B., Roux, S., Hajek, P.P., Varghese,
N., Mukherjee, S., and Reddy, T. (2021). DOE JGI metagenome workflow. Msystems 6,
e00804-00820.
Confer, A.W., and Ayalew, S. (2013). The OmpA family of proteins: roles in bacterial
pathogenesis and immunity. Veterinary Microbiology 163, 207-222.

203
D'souza, N., Kawarasaki, Y., Gantz, J., Lee, R., Beall, B., Shtarkman, Y., Koçer, Z., Rogers, S.,
Wildschutte, H., and Bullerjahn, G. (2013). Diatom assemblages promote ice formation
in large lakes. The ISME journal 7, 1632-1640.
D'souza, N.A. (2012). Psychrophilic diatoms in ice-covered Lake Erie. Bowling Green State
University.
Daufresne, M., Lengfellner, K., and Sommer, U. (2009). Global warming benefits the small in
aquatic ecosystems. Proceedings of the National Academy of Sciences 106, 12788-12793.
Dorrell, R.G., Villain, A., Perez-Lamarque, B., Audren De Kerdrel, G., Mccallum, G., Watson,
A.K., Ait-Mohamed, O., Alberti, A., Corre, E., and Frischkorn, K.R. (2021).
Phylogenomic fingerprinting of tempo and functions of horizontal gene transfer within
ochrophytes. Proceedings of the National Academy of Sciences 118, e2009974118.
Edgar, R., Morris, P., Rozmarynowycz, M., D'souza, N., Moniruzzaman, M., Bourbonniere, R.,
Bullerjahn, G., Phuntumart, V., Wilhelm, S., and Mckay, R. (2016). Adaptations to
photoautotrophy associated with seasonal ice cover in a large lake revealed by
metatranscriptome analysis of a winter diatom bloom. Journal of Great Lakes Research
42, 1007-1015.
Eiler, A., Drakare, S., Bertilsson, S., Pernthaler, J., Peura, S., Rofner, C., Simek, K., Yang, Y.,
Znachor, P., and Lindström, E.S. (2013). Unveiling distribution patterns of freshwater
phytoplankton by a next generation sequencing based approach. PLoS One 8, e53516.
Eiler, A., Zaremba‐Niedzwiedzka, K., Martínez‐García, M., Mcmahon, K.D., Stepanauskas, R.,
Andersson, S.G., and Bertilsson, S. (2014). Productivity and salinity structuring of the
microplankton revealed by comparative freshwater metagenomics. Environmental
Microbiology 16, 2682-2698.
El-Gebali, S., Mistry, J., Bateman, A., Eddy, S.R., Luciani, A., Potter, S.C., Qureshi, M.,
Richardson, L.J., Salazar, G.A., and Smart, A. (2019). The Pfam protein families
database in 2019. Nucleic acids research 47, D427-D432.
Filazzola, A., Blagrave, K., Imrit, M.A., and Sharma, S. (2020). Climate change drives increases
in extreme events for lake ice in the Northern Hemisphere. Geophysical Research Letters
47, e2020GL089608.
Fu, L., Niu, B., Zhu, Z., Wu, S., and Li, W. (2012). CD-HIT: accelerated for clustering the next-
generation sequencing data. Bioinformatics 28, 3150-3152.
Gilbert, N.E., Lecleir, G.R., Strzepek, R.F., Ellwood, M.J., Twining, B.S., Roux, S., Pennacchio,
C., Boyd, P.W., and Wilhelm, S.W. (2022). Bioavailable iron titrations reveal oceanic
Synechococcus ecotypes optimized for different iron availabilities. ISME
Communications 2, 1-12.
Gomez-Consarnau, L., Raven, J.A., Levine, N.M., Cutter, L.S., Wang, D., Seegers, B., Aristegui,
J., Fuhrman, J.A., Gasol, J.M., and A., S.-W.-S. (2019). Microbial rhodopsins are major
contributors to the solar energy captured in the sea. Sciecne Advances 5, eaaw8855.
Gurevich, A., Saveliev, V., Vyahhi, N., and Tesler, G. (2013). QUAST: quality assessment tool
for genome assemblies. Bioinformatics 29, 1072-1075.
Huber, O., and Sumper, M. (1994). Algal‐CAMs: isoforms of a cell adhesion molecule in
embryos of the alga Volvox with homology to Drosophila fasciclin I. The EMBO journal
13, 4212-4222.
Katoh, K., and Standley, D.M. (2013). MAFFT multiple sequence alignment software version 7:
improvements in performance and usability. Molecular biology and evolution 30, 772-
780.

204
Kopejtka, K., Tomasch, J., Kaftan, D., Gardiner, A.T., Bína, D., Gardian, Z., Bellas, C., Dröge,
A., Geffers, R., and Sommaruga, R. (2022). A bacterium from a mountain lake harvests
light using both proton-pumping xanthorhodopsins and bacteriochlorophyll-based
photosystems. Proceedings of the National Academy of Sciences 119, e2211018119.
Krinos, A.I., Hu, S.K., Cohen, N.R., and Alexander, H. (2020). EUKulele: Taxonomic
annotation of the unsung eukaryotic microbes. arXiv preprint arXiv:2011.00089.
Lachnit, M., Buhmann, M.T., Klemm, J., Kröger, N., and Poulsen, N. (2019). Identification of
proteins in the adhesive trails of the diatom Amphora coffeaeformis. Philosophical
Transactions of the Royal Society B 374, 20190196.
Larsson, A. (2014). AliView: a fast and lightweight alignment viewer and editor for large
datasets. Bioinformatics 30, 3276-3278.
Letunic, I., and Bork, P. (2019). Interactive Tree Of Life (iTOL) v4: recent updates and new
developments. Nucleic acids research 47, W256-W259.
Li, D., Luo, R., Liu, C.-M., Leung, C.-M., Ting, H.-F., Sadakane, K., Yamashita, H., and Lam,
T.-W. (2016). MEGAHIT v1. 0: a fast and scalable metagenome assembler driven by
advanced methodologies and community practices. Methods 102, 3-11.
Liao, Y., Smyth, G.K., and Shi, W. (2014). featureCounts: an efficient general purpose program
for assigning sequence reads to genomic features. Bioinformatics 30, 923-930.
Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and
dispersion for RNA-seq data with DESeq2. Genome Biology 15, 1-21.
Marchetti, A., Catlett, D., Hopkinson, B.M., Ellis, K., and Cassar, N. (2015). Marine diatom
proteorhodopsins and their potential role in coping with low iron availability. The ISME
Journal 9, 2745-2748.
Martin, R.M., and Wilhelm, S.W. (2020). Phenol-based RNA Extraction from Polycarbonate
Filters. protocols.io.
Mason, L.A., Riseng, C.M., Gronewold, A.D., Rutherford, E.S., Wang, J., Clites, A., Smith,
S.D., and Mcintyre, P.B. (2016). Fine-scale spatial variation in ice cover and surface
temperature trends across the surface of the Laurentian Great Lakes. Climatic Change
138, 71-83.
Mckay, R.M.L., Beall, B.F., Bullerjahn, G.S., and Woityra, L.W.C. (2011). Winter limnology on
the Great Lakes: The role of the US Coast Guard. Journal of Great Lakes Research 37,
207-210.
Mckay, R.M.L., Villareal, T.A., and La Roche, J. (2000). Vertical migration by Rhizosolenia
spp.(Bacillariophyceae): Implications for Fe acquisition. Journal of Phycology 36, 669-
674.
Mou, X., Jacob, J., Lu, X., Robbins, S., Sun, S., and Ortiz, J.D. (2013). Diversity and distribution
of free-living and particle-associated bacterioplankton in Sandusky Bay and adjacent
waters of Lake Erie Western Basin. Journal of Great Lakes Research 39, 352-357.
Nelson, D.M., Tréguer, P., Brzezinski, M.A., Leynaert, A., and Quéguiner, B. (1995). Production
and dissolution of biogenic silica in the ocean: revised global estimates, comparison with
regional data and relationship to biogenic sedimentation. Global Biogeochemical Cycles
9, 359-372.
Noaa. 2021. NOAA projects 30% maximum Great Lakes ice cover for winter. Available:
https://www.workboat.com/government/noaa-projects-30-maximum-great-lakes-ice-
cover-for-winter [Accessed Febraury 8m 2023].
Noaa (2023). Low ice on the Great Lakes this winter. NOAA Research News.

205
Noaa-Glerl Historical Ice Cover [Online]. NOAA-GLERL. Available:
https://www.glerl.noaa.gov/data/ice/#historical [Accessed 2-12-2023 2023].
Oh, S., Caro-Quintero, A., Tsementzi, D., Deleon-Rodriguez, N., Luo, C., Poretsky, R., and
Konstantinidis, K.T. (2011). Metagenomic insights into the evolution, function, and
complexity of the planktonic microbial community of Lake Lanier, a temperate
freshwater ecosystem. Applied and environmental microbiology 77, 6000-6011.
Oliver, M.J., Petrov, D., Ackerly, D., Falkowski, P., and Schofield, O.M. (2007). The mode and
tempo of genome size evolution in eukaryotes. Genome Research 17, 594-601.
Otte, A., Winder, J.C., Deng, L., Schmutz, J., Jenkins, J., Grigoriev, I.V., Hopes, A., and Mock,
T. (2023). The diatom Fragilariopsis cylindrus: A model alga to understand cold‐adapted
life. Journal of Phycology.
Ozersky, T., Bramburger, A.J., Elgin, A.K., Vanderploeg, H.A., Wang, J., Austin, J.A., Carrick,
H.J., Chavarie, L., Depew, D.C., and Fisk, A.T. (2021). "The changing face of winter:
lessons and questions from the Laurentian Great Lakes". (Journal of Geophysical
Research: Biogeosciences: Wiley Online Library).
Pančić, M., Torres, R.R., Almeda, R., and Kiørboe, T. (2019). Silicified cell walls as a defensive
trait in diatoms. Proceedings of the Royal Society B 286, 20190184.
Powers, S.M., and Hampton, S.E. (2016). Winter limnology as a new frontier. Limnology and
Oceanography Bulletin 25, 103-108.
Ratnasingham, S., and Hebert, P.D. (2007). BOLD: The Barcode of Life Data System
(http://www. barcodinglife. org). Molecular ecology notes 7, 355-364.
Reavie, E. (2023). Asymmetric, biraphid diatoms from the Laurentian Great Lakes. PeerJ
Aquatic Biology.
Reavie, E.D., and Barbiero, R.P. (2013). Recent changes in abundance and cell size of pelagic
diatoms in the North American Great Lakes. Phytotaxa 127, 150-162.
Reavie, E.D., Cai, M., Twiss, M.R., Carrick, H.J., Davis, T.W., Johengen, T.H., Gossiaux, D.,
Smith, D.E., Palladino, D., and Burtner, A. (2016). Winter–spring diatom production in
Lake Erie is an important driver of summer hypoxia. Journal of Great Lakes Research
42, 608-618.
Rühland, K.M., Paterson, A.M., and Smol, J.P. (2015). Lake diatom responses to warming:
reviewing the evidence. Journal of Paleolimnology 54, 1-35.
Saxton, M.A., D'souza, N.A., Bourbonniere, R.A., Mckay, R.M.L., and Wilhelm, S.W. (2012).
Seasonal Si: C ratios in Lake Erie diatoms—evidence of an active winter diatom
community. Journal of Great Lakes Research 38, 206-211.
Sharma, A.K., Sommerfeld, K., Bullerjahn, G.S., Matteson, A.R., Wilhelm, S.W., Jezbera, J.,
Brandt, U., Doolittle, W.F., and Hahn, M.W. (2009). Actinorhodopsin genes discovered
in diverse freshwater habitats and among cultivated freshwater Actinobacteria. The ISME
journal 3, 726-737.
Shin, Y., Chane, A., Jung, M., and Lee, Y. (2021). Recent advances in understanding the roles of
pectin as an active participant in plant signaling networks. Plants 10, 1712.
Slamovits, C.H., Okamato, N., Burri, L., James, E.R., and J., K.P. (2011). A bacterial
proteorhodopsin proton pump in marine eukaryotes. Nature Communications 2, 183.
Sommer, U., Adrian, R., De Senerpont Domis, L., Elser, J.J., Gaedke, U., Ibelings, B., Jeppesen,
E., Lürling, M., Molinero, J.C., and Mooij, W.M. (2012). Beyond the Plankton Ecology
Group (PEG) model: mechanisms driving plankton succession. Annual review of ecology,
evolution, and systematics 43, 429-448.

206
Sommer, U., Gliwicz, Z.M., Lampert, W., and Duncan, A. (1986). The PEG-model of seasonal
succession of planktonic events in fresh waters. Archiv für Hydrobiologie 106, 433-471.
Struyf, E., Smis, A., Van Damme, S., Meire, P., and Conley, D.J. (2009). The global
biogeochemical silicon cycle. Silicon 1, 207-213.
Twiss, M., Mckay, R., Bourbonniere, R., Bullerjahn, G., Carrick, H., Smith, R., Winter, J.,
D'souza, N., Furey, P., and Lashaway, A. (2012). Diatoms abound in ice-covered Lake
Erie: An investigation of offshore winter limnology in Lake Erie over the period 2007 to
2010. Journal of Great Lakes Research 38, 18-30.
Twiss, M.R., Smith, D.E., Cafferty, E.M., and Carrick, H.J. (2014). Phytoplankton growth
dynamics in offshore Lake Erie during mid-winter. Journal of Great Lakes Research 40,
449-454.
Valipour, R., Boegman, L., Bouffard, D., and Rao, Y.R. (2017). Sediment resuspension
mechanisms and their contributions to high‐turbidity events in a large lake. Limnology
and Oceanography 62, 1045-1065.
Vanderploeg, H., Johengen, T., Lavrentyev, P.J., Chen, C., Lang, G., Agy, M., Bundy, M.,
Cavaletto, J., Eadie, B., and Liebig, J. (2007). Anatomy of the recurrent coastal sediment
plume in Lake Michigan and its impacts on light climate, nutrients, and plankton. Journal
of Geophysical Research: Oceans 112.
Vanderploeg, H.A., Bolsenga, S.J., Fahnenstiel, G.L., Liebig, J.R., and Gardner, W.S. (1992).
Plankton ecology in an ice-covered bay of Lake Michigan: Utilization of a winter
phytoplankton bloom by reproducing copepods. Hydrobiologia 243, 175-183.
Vardi, A., Thamatrakoln, K., Bidle, K.D., and Falkowski, P.G. (2009). Diatom genomes come of
age. Genome Biology 9, 245.
Villareal, T.A. (unpublished).
Villareal, T.A., Altabet, M.A., and Culver-Rymsza, K. (1993). Nitrogen transport by vertically
migrating diatom mats in the North Pacific Ocean. Nature 363, 709-712.
Villareal, T.A., and Carpenter, E.J. (1994). CHEMICAL COMPOSITION AND
PHOTOSYNTHETIC CHARACTERISTICS OF ETHMODISCUS REX
(BACILLARIOPHYCEAE): EVIDENCE FOR VERTICAL MIGRATION 1. Journal of
Phycology 30, 1-8.
Villareal, T.A., Joseph, L., Brzezinski, M.A., Shipe, R.F., Lipschultz, F., and Altabet, M.A.
(1999a). Biological and chemical characteristics of the giant diatom Ethmodiscus
(Bacillariophyceae) in the central North Pacific gyre. Journal of Phycology 35, 896-902.
Villareal, T.A., and Lipschultz, F. (1995). Internal nitrate concentrations in single cells of large
phytoplankton from the sargasso sea 1. Journal of Phycology 31, 689-696.
Villareal, T.A., Pilskaln, C., Brzezinski, M., Lipschultz, F., Dennett, M., and Gardner, G.B.
(1999b). Upward transport of oceanic nitrate by migrating diatom mats. Nature 397, 423-
425.
Villareal, T.A., Woods, S., Moore, J.K., and Culverrymsza, K. (1996). Vertical migration of
Rhizosolenia mats and their significance to NO3− fluxes in the central North Pacific
gyre. Journal of Plankton Research 18, 1103-1121.
Wang, J., Kessler, J., Bai, X., Clites, A., Lofgren, B., Assuncao, A., Bratton, J., Chu, P., and
Leshkevich, G. (2018). Decadal variability of Great Lakes ice cover in response to AMO
and PDO, 1963–2017. Journal of Climate 31, 7249-7268.

207
Wheeler, D.L., Barrett, T., Benson, D.A., Bryant, S.H., Canese, K., Chetvernin, V., Church,
D.M., Dicuccio, M., Edgar, R., and Federhen, S. (2007). Database resources of the
national center for biotechnology information. Nucleic acids research 35, D5-D12.
Wilhelm, S.W., Lecleir, G.R., Bullerjahn, G.S., Mckay, R.M., Saxton, M.A., Twiss, M.R., and
Bourbonniere, R.A. (2014). Seasonal changes in microbial community structure and
activity imply winter production is linked to summer hypoxia in a large lake. FEMS
microbiology ecology 87, 475-485.
Willis, A., Eason‐Hubbard, M., Hodson, O., Maheswari, U., Bowler, C., and Wetherbee, R.
(2014). Adhesion molecules from the diatom Phaeodactylum tricornutum
(Bacillariophyceae): genomic identification by amino‐acid profiling and in vivo analysis.
Journal of Phycology 50, 837-849.
Winder, M., Reuter, J.E., and Schladow, S.G. (2009). Lake warming favours small-sized
planktonic diatom species. Proceedings of the Royal Society B: Biological Sciences 276,
427-435.
Xie, G., Tang, X., Shao, K., Hu, Y., Liu, H., Martin, R.M., and Gao, G. (2020). Spatiotemporal
patterns and environmental drivers of total and active bacterial abundances in Lake
Taihu, China. Ecological Indicators 114, 106335.
Zepernick, B.N., Denison, E.R., Chaffin, J.D., Bullerjahn, G.S., Pennacchio, C.P., Frenken, T.,
Peck, D.H., Anderson, J.T., Niles, D., and Zastepa, A. (2022). Metatranscriptomic
Sequencing of Winter and Spring Planktonic Communities from Lake Erie, a Laurentian
Great Lake. Microbiology Resource Announcements, e00351-00322.
Zhu, W., Lomsadze, A., and Borodovsky, M. (2010). Ab initio gene identification in
metagenomic sequences. Nucleic acids research 38, e132-e132.

208
Appendix
Appendix Methods

Lake Erie winter-spring water column sampling

To assess Chl a concentration, 50 mL volumes of sample water were collected on 0.22-

𝜇𝜇m nominal pore-size polycarbonate filters and stored at -20°C until Chl a extraction at Bowling

Green State University. In addition, 200 mL volumes of sample water were collected on 20 𝜇𝜇m

nominal pore-size polycarbonate filters and stored at -20°C. Subsequently, samples were subject

to a 24-hour extraction in 90% acetone at 4°C and quantified on a fluorometer (Turner Designs

TD-700) equipped with a blue mercury bulb, a #10-050R excitation filter (340–500 nm) and a

#10–115 (680 nm) emission filter (Welschmeyer, 1994). Samples for RNA isolation were

filtered through 0.22-𝜇𝜇m nominal pore-size filters until the filter was saturated, flash frozen, and

stored at -80 °C until extraction at The University of Tennessee Knoxville. For dissolved nutrients

(SiO2, NO3, NO2, NH3, SRP, SO4, Cl), 50 mL volumes of filtrate from the sterivex were

collected and stored at -20°C until processing. Whole water samples for particulate nutrients

(TN, TP) were also stored at -20°C until processing at The Ohio State University Stone

Laboratory. There, samples were subsequently processed on a SEAL Analytical QuAAtro 5-

Chaneel continuous segments flow auto-analyzer (Chaffin et al., 2019). Fifty milliliter samples

of whole water were preserved with Lugol’s iodine and stored at room temperature until

phytoplankton identification and enumeration at Aquatic Taxonomy Specialists, (Malinta OH).

Briefly, samples were analyzed by counting phytoplankton in a measured aliquot using a

modified inverted microscope for the Utermöhl method plus a small magnification modification

of the stratified counting technique of Munawar and Munawar (Munawar and Munawar, 1976).

A measured aliquot of mixed sample was placed into an inverted microscope counting chamber

209
and allowed to settle a minimum of 4 h per centimeter of overlying water depth. Larger and

recognizable rare cells were counted at 400× along a minimum of one transect across the entire

counting chamber. Smaller algae were counted at 1000× along a measured transect until a

minimum of 300 cells were enumerated. Phytoplankton were counted as individual cells. All

water column data was plotted in prism (v.9.3.1) by Julian Day to account for the interannual

variability (2019-2020). Visualization of the 12 sample sites was performed in R using tmap

(Tennekes, 2018) and open source shapefiles of Lake Erie shoreline (https://gis-

michighan.opendata.arcgis.com/datasets) and bathymetry

(https://www.ngdc.noaa.gov/mgg/fliers).

Metatranscriptomic analysis

The bioinformatic workflow used to process the libraries within this manuscript was

initially compiled and published by Gilbert et al. (2022). Trimming and filtering of raw reads to

remove adaptors, contaminants, and sequence spike in-ins was performed by JGI via BBDuk

(v.38.92) within the BBtools package (Bushnell, 2014b). Following, ribosomal RNA was

removed using BBMap (default settings) (v.38.86) in line with the DOE JGI pipeline (Clum et

al., 2021). Filtered, trimmed reads from all 77 metatranscriptomic libraries were concatenated

and assembled with MEGAHIT (v.1.2.9) (Li et al., 2016), with the concatenated file serving as

the input as performed previously (Gilbert et al., 2022). Quality assessment of the coassembly

was performed using QUAST (v.5.0.2) (Mikheenko et al., 2018) with all statistics included in

(Supplemental Dataset 1). Trimmed, filtered reads from each library (n=20) were mapped to the

coassembly via BBMap (default settings) (v.38.86) with mapping statistics included in

(Supplemental Dataset 2). Following, calling of the open reading frames (ORFs) within the

210
coassembly was performed via MetaGeneMark (v.3.38) (Besemer and Borodovsky, 1999; Zhu et

al., 2010) using the gene finding algorithm. Taxonomic annotation of the coassembly genes was

performed using EUKulele (Krinos et al., 2020) to perform a blast against the PhyloDB (v.1.076)

database. Functional annotation of coassembly genes was conducted via eggNOG-mapper using

orthology data and a specified e-value of 1e-10 (v.2.1.7) (Cantalapiedra et al., 2021). In this study,

genes that were not assigned a “Preferred Name” by eggNOG were given one which we made

based on the gene description provided by eggNOG-mapper. A list of the genes and their

eggNOG preferred name (or the name we assigned in this study) is provided in the Appendix

Materials (Appendix Attachments 5.1Z-5.2C). Trimmed, mapped reads were then tabulated

according to ORF coordinates using featureCounts (Liao et al., 2014) within the subread

(v.2.0.1) package. . After, the tabulated read counts, EUKulele annotations, and eggNOG-

mapper annotations were merged into a single file in R (v.4.0.0). Read mappings were

normalized to transcripts-per-million (TPM) prior to statistical analyses (ANOSIM, SIMPER,

nMDS etc.). Raw reads were used for DESeq2 analysis as required by the program.

Relative transcript abundance and taxonomy

To investigate taxonomic trends inferred from proportional raw transcript abundance,

coassembly genes were sorted according to the 7 classification levels of taxonomy assigned by

EUKulele (domain-genus/species). To focus on diatoms specifically, analysis of raw reads

mapped to domains (classification level 1), eukaryota supergroups (classification level 2), major

eukaryotic phytoplankton phyla (classification level 3 and 4), bacillariophyta classes

(classification level 6), and genus (classification level 7) were performed to determine relative

proportions of the transcriptionally active community as performed previously (Kranzler et al.,

211
2021). Major eukaryotic phytoplankton taxa common to Lake Erie were selected based on prior

long-term environmental surveys and manual sorting (Reavie et al., 2014). In the analysis of the

relative transcript abundance, groups that were <5% of the total mapped reads were classified as

“Other” as performed previously (Kranzler et al., 2021). Due to the low resolution of taxonomic

annotation at the genus level, “not annotated” reads were not included in the calculations at the

genus level (Appendix Figures 5.16-5.19). Rather, the NA reads were included within the

“Other” group to allow for the visualization of taxonomic trends of annotated taxa. Relative

abundances were graphically visualized in R using ggplot2 (Wickham, 2016) and further

modified in Adobe Illustrator (Adobe Inc., 2022).

In this study, we used a marine taxonomic annotation database (EUKulele and PhyloDB)

due to their unmatched depth and diversity. Hence, at the genus level all genera are reported as

“genus-like” to reflect that these are annotations made based on marine taxonomy which we have

extrapolated to our freshwater system. For example, Aulacoseira-like annotated genes within our

dataset can be confidently suggested to be Aulacosiera based on freshwater cell abundances and

taxonomy, yet we report it as Aulacoseira-like for consistency. Unfortunately, freshwater

systems have not been as comprehensively sequenced/studied as marine systems. Indeed, TARA

Oceans propelled the marine sequencing and taxonomic curation far beyond freshwater

capacities (Sunagawa et al., 2015). Further, there is a lack of eukaryotic representatives within

available freshwater sequencing datasets, which further exacerbates taxonomic annotation

difficulties associated with freshwater communities such as Lake Erie diatom blooms. In turn,

this study illuminates a critical gap and limitation to freshwater meta-omics work;

comprehensive taxonomic databases are required for more direct and fine-scale community

identification. Diatom classes (Coscinodiscophyceae, Mediophyceae, Bacillariophyceae,

212
Fragilariophyceae) for diatom genera including A. islandica and Stephanodiscus spp. were

assigned based on the BOLD taxonomic database (Ratnasingham and Hebert, 2007).

Phylogenetic analysis

The phylogenetic tree was made using fasciclin containing domains recovered from this

study (Appendix Attachment 5.2D), domains recovered from the eggNOG orthology database

(Appendix Attachment 5.2E), and publicly available domains downloaded from NCBI. The

resulting diatoms found to contain fasciclins (n=141) are provided in the Appendix Attachment

5.2F.

Appendix Results

Taxonomically resolved winter diatom community transcriptional response to ice cover

While the number of reads mapping to Eukaryota did not vary as a function of ice-cover,

every other domain demonstrated an increasing trend in read abundance in ice-free samples

(Appendix Figure 5.14). Eukaryota contributed an average of 66.99% (+/- 6.48) of transcripts in

ice cover samples and 47.99% (+/- 9.06) of reads in ice-free winter samples. Notably, the

number of reads mapping to Eukaryota were higher than bacteria within ice-cover samples but

not ice-free samples. Stramenopiles contributed an average of 56.79% (+/-8.04) of the Eukaryota

transcripts during ice cover and 44.72% (+/- 12.33) during ice-free conditions (Appendix Figure

5.15). Bacillariophyta reads were dominated by the class Mediophyceae (polar centric diatoms),

which remained the most abundant of the transcriptionally active winter diatom community

regardless of ice cover (Figure 5.3B). Coscinodiscophyceae (centric) was the second most

abundant, while pennate classes Bacillariophyceae and Fragilariophyceae remained at very low

213
transcriptional abundances. Approximately 25% of the winter Mediophyceae community were

classified as Thalassiosira-like (Appendix Figure 5.16), while 50% of the Coscinodiscophyceae

winter community was annotated as Aulacoseira-like (Appendix Figure 5.17). In contrast,

pennate diatom classes Bacillariophyceae and Fragilariophyceae exhibited small trends in genera

transcript abundance as a function of ice cover (Appendix Figures 5.18-5.19). Overall, diatom

transcriptional abundance did not correlate with diatom cell abundances (R2 = 0.02) (Appendix

Figure 5.20).

Phylogenetic analysis

The FAS1 domain appears to primarily originate from bacteria and is widely distributed

throughout the bacterial domain. The FAS1 domain was found in 141 diatoms which span

freshwater and marine distributions. Notably, the FAS1 domain was found in various strains of

the polar, cold-adapted marine diatom Fragilariopsis cylindrus, which has been recently

designated as a “model for understanding cold-adapted alga” (Otte et al., 2023). Further,

phylogenetic analysis demonstrates multiple instances of horizontal gene transfer (HGT), thus

the FAS1 domain within diatoms appears to have been horizontally acquired from bacteria. Two

of the putative HGT branches are quite long. This is likely due to lack of available

genomes/sequencing for the respective diatom species. One long HGT branch corresponds to the

diatom Mayamaea pseudoterrestris, the only member of genus Mayamaea with a genome

sequenced (many of the individuals in this group are newly defined–e.g., 2020, 2021, 2022). The

second is from a transcriptomic dataset, and labeled as an unculturered Nitzschia, four genomes

are sequenced for this group, however given that it is an environmental sample it could likely be

considered as Nitzschia because that is the closest relative/hit to current databases. Therefore, it

214
could be a group not defined or not well defined to date. Generally, the branch lengths are long

for the “diatom section”, also likely related to incomplete taxonomy. Notably, the FAS1 domain

is well represented in diatoms compared to other alga and cyanobacteria, yet most of the FAS1

diatom domains were annotated as “hypothetical, unnamed, or predicted protein” (Appendix

Attachment 5.2F). Hence, the FAS1 domain is not well-annotated or characterized within

diatoms. Cumulatively, this data suggests the FAS1 domain distribution within diatoms is a

result of HGT events with bacteria. Further, this data suggests the FAS1 domain is well-

distributed across at least 141 diatom taxa as found in this study yet remains unannotated and

undefined in most publicly available diatom sequencing data.

215
Appendix References:
Besemer, J., and Borodovsky, M. (1999). Heuristic approach to deriving models for gene
finding. Nucleic acids research 27, 3911-3920.
Bushnell, B. (2014). BBTools software package. URL http://sourceforge. net/projects/bbmap
578, 579.
Cantalapiedra, C.P., Hernández-Plaza, A., Letunic, I., Bork, P., and Huerta-Cepas, J. (2021).
eggNOG-mapper v2: functional annotation, orthology assignments, and domain
prediction at the metagenomic scale. Molecular biology and evolution 38, 5825-5829.
Chaffin, J.D., Mishra, S., Kane, D.D., Bade, D.L., Stanislawczyk, K., Slodysko, K.N., Jones,
K.W., Parker, E.M., and Fox, E.L. (2019). Cyanobacterial blooms in the central basin of
Lake Erie: Potentials for cyanotoxins and environmental drivers. Journal of Great Lakes
Research 45, 277-289.
Clum, A., Huntemann, M., Bushnell, B., Foster, B., Foster, B., Roux, S., Hajek, P.P., Varghese,
N., Mukherjee, S., and Reddy, T. (2021). DOE JGI metagenome workflow. Msystems 6,
e00804-00820.
Gilbert, N.E., Lecleir, G.R., Strzepek, R.F., Ellwood, M.J., Twining, B.S., Roux, S., Pennacchio,
C., Boyd, P.W., and Wilhelm, S.W. (2022). Bioavailable iron titrations reveal oceanic
Synechococcus ecotypes optimized for different iron availabilities. ISME
Communications 2, 1-12.
Kranzler, C.F., Brzezinski, M.A., Cohen, N.R., Lampe, R.H., Maniscalco, M., Till, C.P., Mack,
J., Latham, J.R., Bruland, K.W., and Twining, B.S. (2021). Impaired viral infection and
reduced mortality of diatoms in iron-limited oceanic regions. Nature Geoscience 14, 231-
237.
Krinos, A.I., Hu, S.K., Cohen, N.R., and Alexander, H. (2020). EUKulele: Taxonomic
annotation of the unsung eukaryotic microbes. arXiv preprint arXiv:2011.00089.
Li, D., Luo, R., Liu, C.-M., Leung, C.-M., Ting, H.-F., Sadakane, K., Yamashita, H., and Lam,
T.-W. (2016). MEGAHIT v1. 0: a fast and scalable metagenome assembler driven by
advanced methodologies and community practices. Methods 102, 3-11.
Liao, Y., Smyth, G.K., and Shi, W. (2014). featureCounts: an efficient general purpose program
for assigning sequence reads to genomic features. Bioinformatics 30, 923-930.
Mikheenko, A., Prjibelski, A., Saveliev, V., Antipov, D., and Gurevich, A. (2018). Versatile
genome assembly evaluation with QUAST-LG. Bioinformatics 34, i142-i150.
Munawar, M., and Munawar, I.F. (1976). A lakewide study of phytoplankton biomass and its
species composition in Lake Erie, April–December 1970. Journal of the Fisheries Board
of Canada 33, 581-600.
Otte, A., Winder, J.C., Deng, L., Schmutz, J., Jenkins, J., Grigoriev, I.V., Hopes, A., and Mock,
T. (2023). The diatom Fragilariopsis cylindrus: A model alga to understand cold‐adapted
life. Journal of Phycology.
Ratnasingham, S., and Hebert, P.D. (2007). BOLD: The Barcode of Life Data System
(http://www. barcodinglife. org). Molecular ecology notes 7, 355-364.
Reavie, E.D., Barbiero, R.P., Allinger, L.E., and Warren, G.J. (2014). Phytoplankton trends in
the Great Lakes, 2001–2011. Journal of Great Lakes Research 40, 618-639.
Sunagawa, S., Coelho, L.P., Chaffron, S., Kultima, J.R., Labadie, K., Salazar, G., Djahanschiri,
B., Zeller, G., Mende, D.R., and Alberti, A. (2015). Structure and function of the global
ocean microbiome. Science 348, 1261359.
Tennekes, M. (2018). tmap: Thematic Maps in R. Journal of Statistical Software 84, 1-39.

216
Welschmeyer, N.A. (1994). Fluorometric analysis of chlorophyll a in the presence of chlorophyll
b and pheopigments. Limnology and Oceanography 39, 1985-1992.
Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New York.
ISBN 978-3-319-24277-4.
Zhu, W., Lomsadze, A., and Borodovsky, M. (2010). Ab initio gene identification in
metagenomic sequences. Nucleic acids research 38, e132-e132.

217
Appendix Tables/Figures

Table 5.1. Metatranscriptomic libraries with respect to spatial, temporal, and climatic
variables. Libraries listed within the same row are biological replicates. Sample sites are listed
in chronological order of sampling (S1-12). Samples that were collected from under the ice are
denoted with “Yes” and samples collected during a no ice cover are denoted with “No”. Season
is listed as either “Winter” or “Spring”. Sample day is presented in Julian day, to account for
interannual variability. Samples collected during 2019, a year of high ice cover overall
throughout Lake Erie, are indicated with an asterisk. Samples collected during 2020, a year of
low ice cover throughout Lake Erie, do not have an asterisk.

Library Site Ice cover Season Julian day Month Year

L1 S1 Yes Winter 57 February 2019*
L2 S2 Yes Winter 57 February 2019*
L3 S3 Yes Winter 70 March 2019*
L4 S4 Yes Winter 70 March 2019*
L5, L6 S5 No Winter 45 February 2020
L7, L8 S6 No Winter 45 February 2020
L9, L10 S7 No Winter 45 February 2020
L11, L12 S8 No Winter 62 March 2020
L13, L14 S9 No Winter 62 March 2020
L15, L16, L17 S10 No Spring 122 May 2020
L18 S11 No Spring 143 May 2020
L19, L20 S12 No Spring 160 June 2020

218
Figure 5.8: Temperature and nutrient profiles (µM) across the 12 sample sites organized by
season. W = winter (February-March), S = spring (May-June). Solid shapes indicate the sample
was collected during ice cover (2019), open shapes indicate the sample was collected during no
ice cover (2020). X’s indicate there was no data reported. (A) Water temperature. (B) Dissolved
Chlorine. (C) Dissolved Ammonia. (D) Dissolved Silicate. (E) Dissolved Nitrate. (F) Dissolved
Soluble Reactive Phosphorus (G) Silica: Nitrate ratios. (H) Total Particulate Nitrogen: Total
Particulate Phosphorus. Statistical comparisons as a function of ice cover were made using
unpaired, two-tailed t-tests.

219
Figure 5.9: Mean percent contribution of Chl a >20 µm to total Chl a across sample sites
organized by season. W = winter (February-March), S = spring (May-June). Solid shapes
indicate the sample was collected during ice cover (2019) open shapes indicate the sample was
collected during no ice cover (2020). Statistical comparisons as a function of ice cover were
made using unpaired, two-tailed t-tests.

220
Figure 5.10: Abundance (Cells•L-1) of major eukaryotic phytoplankton taxa across sample sites
organized by season. W = winter (February-March), S = spring (May-June). Solid shapes
indicate the sample was collected during ice cover (2019) open shapes indicate the sample was
collected during no ice cover (2020). Sites were the taxa was not detected are indicated with
“n.d.”. (A) Sum diatom counts (A. islandica + Stephanodiscus spp. + centric diatoms of 5-20 µm
+ Fragilaria spp. + Asterionella formosa. + Nitzschia spp.). (B) Abundance of cryptophytes
across sample sites. (C) Abundance of dinoflagellates across sample sites. (D) Abundance of
chlorophytes. Statistical comparisons as a function of ice cover were made using unpaired, two-
tailed t-tests. The y-axis scales are different between panels.

221
Supplemental Figure 5.11: Abundance (Cells•L-1) of major centric, filamentous bloom-forming
diatom taxa (S = Stephanodiscus spp., A = A. islandica) across winter sample sites organized by
ice cover. Solid shapes indicate the sample was collected during ice cover (2019) open shapes
indicate the sample was collected during no ice cover (2020). (A) Abundances of Stephanodiscus
spp and A. islandica in winter ice-cover samples. (B) Abundances of Stephanodiscus spp and A.
islandica in winter ice-free samples. Statistical comparisons as a function of ice cover were made
using unpaired, two-tailed t-tests.

222
Figure 5.12: Abundance (Cells•L-1) small centric diatom taxa (5-20 µm size) across sample sites
organized by season. W = winter (February-March), S = spring (May-June). Solid shapes
indicate the sample was collected during ice cover (2019) open shapes indicate the sample was
collected during no ice cover (2020). Sites were the taxa was not detected are indicated with
“n.d.”. Statistical comparisons as a function of ice cover were made using unpaired, two-tailed t-
tests.

223
Figure 5.13: Mean percent contribution of three centric diatom abundance to total diatom
abundance across sample sites organized by season. W = winter (February-March), S = spring
(May-June). Solid shapes indicate the sample was collected during ice cover (2019) open shapes
indicate the sample was collected during no ice cover (2020). (A) Percent contribution of
Stephanodiscus spp. genera abundance to total diatom abundance. (B) Percent contribution of A.
Islandica abundance to total diatom abundance. (C) Percent contribution of small centric diatom
taxa (5-20 µm size) to total diatom abundance. Statistical comparisons as a function of ice cover
were made using unpaired, two-tailed t-tests.

224
Figure 5.14: Relative transcript abundance of domains across the 20 libraries (listed in
chronological order of sample date on the x-axis). NA = not annotated.

225
Figure 5.15: Relative transcript abundance of major eukaryota across the 20 libraries (listed in
chronological order of sample date on the x-axis). All groups which formed <5% of the total
mapped reads are included within “Other” (Amoebozoa, Hilomonadea, Excavata, Rhizaria, NA).

226
5.16: Relative transcript abundance of Mediophyceae genera across the 20 libraries (listed in
chronological order of sample date on the x-axis). All groups which formed <5% of the total
mapped reads are included within “Other” (Attheya spp., Cyclotella spp., Ditylum spp.,
Eucampia spp., Extubocellulus spp., Helicotheca spp., Minutocellus spp., Triceratium spp). At
the genus level, NA counts were not included within total % read mapping calculations due to
low-resolution at the genus level. They have been included within “Other” despite accounting for
59.04% of the total mapped reads to emphasize trends within annotated genera.

227
Figure 5.17: Relative transcript abundance of Coscinodiscophyceae genera across the 20
libraries (listed in chronological order of sample date on the x-axis). All groups which formed
<5% of the total mapped reads are included within “Other” (Coscinodiscus spp., Dactyliosolen
spp., Lepticylindrus spp., Proboscia spp., Rhizosolenia spp., Stephanopyxis spp). At the genus
level, NA counts were not included within total % read mapping calculations due to low-
resolution at the genus level. They have been included within “Other” despite accounting for
40.98% of the total mapped reads to emphasize trends within annotated genera.

228
Figure 5.18: Relative transcript abundance of Bacillariophyceae genera across the 20 libraries
(listed in chronological order of sample date on the x-axis). All groups which formed <5% of the
total mapped reads are included within “Other” (Craspedostauris spp., Cylindrotheca spp.,
Entomoneis spp., Phaeodactylum spp., Psuedo-nitzschia spp., Stauroneis spp). At the genus
level, NA counts were not included within total % read mapping calculations due to low-
resolution at the genus level. They have been included within “Other” despite accounting for
66.00% of the total mapped reads to emphasize trends within annotated genera.

229
Figure 5.19: Relative transcript abundance of Fragilariophyceae genera across the 20 libraries
(listed in chronological order of sample date on the x-axis). All groups which formed <5% of the
total mapped reads are included within “Other” (Astrosyne spp., Cyclophora spp.,
Grammatophora spp., Licmophora spp., Striatella spp., and Syndropsis spp.). At the genus level,
NA counts were not included within total % read mapping calculations due to low-resolution at
the genus level. They have been included within “Other” despite accounting for 50.27% of the
total mapped reads to emphasize trends within annotated genera.

230
Figure 5.20: Simple linear regression of diatom transcriptional abundances and diatom cell
abundances.

231
Figure 5.21: ANOSIM tests plotted by R statistic and p-value. An R statistic close to 1.00 and a
p-value < 0.05 indicates there is a significant difference between the 2 variables. Ice cover is
indicated by a filled circle. (A) ANOSIM tests on the relative expression profiles (TPM) of the
whole winter community comparing months, season (winter vs. spring), and ice cover (ice cover
vs. no ice cover). (B) ANOSIM tests on the relative expression profiles (TPM) of the
Bacillariophyta community comparing months, season (winter vs. spring), and ice cover (ice
cover vs. no ice cover).

232
Figure 5.22: Taxonomic distributions of diatom genes that were differentially expressed by ice
cover. (A) Taxonomic distribution of all 354 DE genes by diatom class. (B) Taxonomic
distribution of DE genes belonging to the Mediophyceae class.

233
Figure 5.23: Taxonomic distributions of Mediophyceae diatom genes that were differentially
expressed by ice cover within COG Category C.

234
Figure 5.24: Bacillariophyta transcript abundance patterns in response to ice cover-COG P
(Inorganic ion transport and metabolism). (A) Taxonomic distribution of DE genes categorized
within COG category P. (B) COG assignments for all 354 DE genes in response to ice cover,
with COG category P indicated in blue. (C) Heatmap depicting COG category P differentially
expressed gene expression (VST) in response to ice cover across the 14 winter libraries.

235
Figure 5.25: Taxonomic distributions of Mediophyceae diatom genes that were differentially
expressed by ice cover within COG Category P.

236
Figure 5.26: Bacillariophyta transcript abundance patterns in response to ice cover-COG G
(Carbohydrate transport and metabolism). (A) Taxonomic distribution of DE genes categorized
within COG category G. All Mediophyceae genes were not annotated at the genus level. (B)
COG assignments for all 354 DE genes in response to ice cover, with COG category G indicated
in blue. (C) Heatmap depicting COG category G differentially expressed gene expression (VST)
in response to ice cover across the 14 winter libraries.

237
Figure 5.27: Normalized expression (VST) of two genes functionally annotated as rhodopsins
(PFAM Rhodopsin indicated by circles, Bacteriorhodopsin-like protein indicated by squares).
Solid shapes indicate the sample was collected during ice cover (2019) open shapes indicate the
sample was collected during no ice cover (2020).

238
Figure 5.28: Distribution of DE genes by COG category. (A) COG assignments for all DE genes
(n = 4,854) in response to season (winter vs. spring). (B) COG assignments for all DE genes (n =
354) in response to ice cover, with COG category M indicated in blue.

239
Figure 5.29: Taxonomic distributions of Mediophyceae diatom genes that were differentially
expressed by ice cover within COG Category M.

240
Figure 5.30: MODIS satellite image (photo takes February 12th, 2023) demonstrating a lack of
ice cover across the Great Lakes. Sediment plumes can be observed throughout Lake Erie, the
shallowest of the Great Lakes. Photo Credit: NOAA GLERL/NOAA Great Lakes
CoastWatchNode.

241
Figure 5.31: Plankton net tows from a 2007 Lake Erie winter survey. (A) Concentrated diatom
seston from a net-tow conducted in ice free waters at station 340 at 10 meters depth. (B) Dilute
seston from a net-tow conducted in ice-covered waters at station 357 at 6 meters depth. Both
samples were collected on February 2, 2007. Dr. R. Michael M. McKay is holding the samples,
Dr. Steven W. Wilhelm took the photographs.

242
CHAPTER VI: CONCLUSIONS

243
 Lakes are sentinels of climate change, serving as a window into how anthropogenically

driven influences alter ecological and biogeochemical phenomena. The body of this work

investigates the ecological success and succession of algal bloom communities within Lake Erie:

a Great Laurentian Lake which is undergoing rapid large-scale climatic changes. In turn, a

myriad of recent advancements coupled with climatic alterations suggest a need to re-assess

many of the classic paradigms which are thought to constrain algal bloom success and

succession. Here, we address these paradigms by investigating the role of bloom-induced pH

levels and alterations in winter ice cover on the ecological success and succession of algal bloom

taxa common to Lake Erie. Notably, this works focuses primarily on the widely unstudied

diatom bloom forms of Lake Erie such as the summer bloom-forming F. crotonensis and winter

diatom bloom formers A. islandica and Stephanodiscus spp. By employing a combination of in

vitro, in situ, and in silico approaches we demonstrate the ecological importance of bloom-

induced pH, ice cover, and diatom communities at the intersection of limnology, climate change,

and harmful algal blooms.

The first contribution of this work (Chapter II) serves as a cautionary tale of the

unintended consequences of classical laboratory techniques. Preliminary optimizations of the pH

assays revealed aseptic flaming of algal cultures introduced CO2(g) into the cultures which

significantly decreased the pH. In turn, fluctuations in culture pH differentially altered growth

dynamics of Microcystis aeruginosa cultures, with the CO2(g) serving as a carbon source. In total,

this project has widescale implications for climate change work (i.e., ocean acidification and lake

basification studies) which require precise pH/CO2 levels. As a result of this work, the buffer

capacity of our freshwater media was increased and aseptic flaming was markedly reduced,

facilitating the establishment of the in vitro pH assay used in Chapters III and IV.

244
 The next contribution of this work (Chapter III) demonstrates Microcystis induced pH

levels significantly decrease growth and silica deposition in the model diatom F. crotonensis and

in situ Lake Erie diatom communities. We discovered Microcystis increases the water column

pH during photosynthetic CO2 drawdown, driving pH levels up to an average daily of 9.2 for ~30

days during a 2015 Microcystis bloom in Lake Erie. While this elevated pH did not significantly

affect Microcystis growth, it invoked detrimental consequences for in vitro F. crotonensis and in

situ diatom viability. Cumulatively, this suggests pH plays a role in Microcystis’s suppression

and out competition of summer diatom communities. Further, considering elevated pH levels

were found to persist well into September, we hypothesize pH likely plays a role in delaying fall

diatom succession. While this work is foundational in establishing the emerging field of lake

basification, it falls short of identifying the intracellular mechanisms responsible for the observed

declines in diatom growth and silica deposition at high pH levels. Hence, Chapter III sought to

investigate this knowledge gap using in vitro pH assays and transcriptomics.

Transcriptomic analyses conducted in Chapter IV generated hypotheses concerning how

the elevated pH level of 9.2 alters F. crotonensis intracellular physiology. Notably, the

expression of genes involved in cell cycle control and cell wall biogenesis increased at pH 9.2,

and follow-up FlowCAM analyses demonstrated alkaline pH levels cause smaller, browner, and

rougher F. crotonensis filaments. In contrast, at pH 9.2 genes involved in photosynthesis

decreased in expression, with follow-up PhytoPAM and photopigment analyses indicating

declines in maximum electron transport rates of PSII and light saturation thresholds of PSII

while photopigments of diadinoxanthin and B-carotene significantly increased. Further, 25% of

DE genes which increased in expression at pH 9.2 belonged to the “Mobilome” COG category,

suggesting F. crotonensis filaments were increasing instances for potential mutations.

245
Cumulatively, these results suggested F. crotonensis filaments were undergoing photostress and

potentially cell cycle arrest at elevated pH levels. Further, this work led to the “genomic roulette

hypothesis” that F. crotonensis filaments are upregulating mobile elements in response to an

environmental stressor to rapidly increase mutations within an organism over a short period of

time. Hence, despite the various physiological consequences high pH imposes on diatoms, there

exists the opportunity for diatoms to adapt to a “basified future”. Nonetheless, Chapter II and III

provide contextual explanations for the ecological “lack of success” of summer diatoms in the

Lake Erie water column.

In juxtaposition to Chapters III and IV which focus on the ecologically “unsuccessful”

summer diatom community, Chapter V pivots to the ecologically “successful” winter diatom

blooms. Despite reports of winter diatom blooms under the Lake Erie ice dating back to the

1930’s, they remained widely unstudied save for a large-scale winter survey conducted from

2007-2012. Our study addressed this gap by serving as the first wide scale bioinformatic survey

of the winter Lake Erie water column. Notably, our 2 year survey spanned a year of high ice

cover (2019) and a year of low ice cover (2020), thus serving as a window into the “ice-free”

future of Lake Erie. Indeed, Lake Erie is observing a rapid decline in ice cover, with the winter

of 2023 demonstrating unprecedentedly negligible ice cover. Our study demonstrates the winter

diatom community has shifted from prior observations made in 2007-2012. Diatom abundances

and Chl a concentrations were significantly lower, and Stephandiscus (rather than A. islandica)

dominated the centric diatom assemblage. Despite these contrasting findings, diatoms still

dominated the ice-free water column, contradicting prior speculations that a phyla-level regime

shift would be observed in the ice-free water column to favor cryptophytes and dinoflagellates.

Most surprising, the metatranscriptomes from this dataset led to the formation of the “fasciclin

246
rafting hypothesis” which suggests polar centric diatoms are employing fasciclins to raft together

and co-locate in the turbid ice-free Lake Erie water column. Hence, further physiological

observations are required to investigate this potentially critical phenomenon, which may prove

pivotal in the future ice-free and climatically altered Lake Erie water column.

In summary, these collective works aim to tease apart the complicated and confusing

story of algal bloom success and succession. Here, we re-assess classical paradigms of ecological

succession and propose novel paradigms such as a role of pH in algal niche competition. We

identify mechanisms which facilitate the ecological successes (or failures) of diatoms within the

basified summer water column and ice-free winter water column. Further, we offer hypotheses

which serve to facilitate freshwater diatom adaptations to the climatically altered water column

including the “genomic roulette” and “fasciclin rafting” hypotheses. While algal bloom

succession remains a complicated and confusing phenomenon, this work serves to identify a few

novel pieces of the puzzle.

While this cumulative work has made contributions to the limnological field at large, it

has also revealed there are many questions which remain to be answered. For example, the role

of pH in algal bloom success and succession remains widely unexplored to date. While this work

made advancements regarding how freshwater diatoms are affected by Microcystis bloom-

induced pH levels, it has yet to be ascertained how Microcystis itself responds to elevated pH

levels. Hence, elucidating the intracellular and ecophysiological response of Microcystis to high

pH remains a research priority. Yet more generally, this work provides a foundation for which

the emerging field of “Lake Basification” can build upon. Another future direction of this work

is a comprehensive bioinformatic characterization of the winter diatom bloom community.

Indeed, our metatranscriptomic work presented in Chapter V illuminated the need for freshwater

247
diatom sequencing efforts and further taxonomic resolution. In contrast to the large-scale

sequencing efforts concerning the marine environment such as Tara Oceans, the limnological

field lacks sequencing depth and comprehensive annotation databases. Hence, obtaining

taxonomic resolution within freshwater diatom bioinformatic datasets is greatly hindered. More

broadly, diatoms within Lake Erie are widely overlooked within the literature at large. Indeed,

novel diatom taxa are continuously reported within the Great Lakes in as recent as the past few

months. Hence, further attention is required not just pertaining to freshwater diatom sequencing

efforts or the winter diatom community, but diatoms across the Great Lakes in general. Beyond

the gap in freshwater diatom sequencing, this study reported that fasciclins remain widely

unstudied within diatoms. Indeed, at the time of this dissertation only two studies were found

mentioning fasciclins and diatoms. Yet, we located fasciclins within ~140 different freshwater

and marine diatom genera. Hence, there is a need to elucidate the functional role these proteins

play within diverse diatom genera and the ecophysiological role they may have as a function of

environment. Another future direction which stems from this work is the need to functionally

and taxonomically characterize proton pumping rhodopsins within freshwater diatoms. To our

knowledge, the work presented in Chapter V is the first report of proton pumping rhodopsins in

freshwater diatoms. While proton pumping rhodopsins have been identified across the marine

diatom lineage, the distribution of these proteins within freshwater diatoms remains largely

unstudied. This serves as a significant gap in the freshwater field, especially considering proton

pumping rhodopsins have been suggested to generate as much energy as classical

photosynthesis. Hence, while the chapters within this dissertation present a few novel pieces of

the puzzle of algal bloom success and succession, many more remain to be discovered.

248
 VITA

Brittany Noel Zepernick was born in Cleveland, Ohio and raised in Aurora, Ohio. She

attended Aurora High School and graduated in 2014. Here, she was broadly interested in

Biology, Chemistry, and English literature. She received her Honors Bachelor of Science Degree

in Biology with a Specialization in Marine and Aquatic Science from Bowling Green State

University in 2017. Here, she discovered her passion for Limnology and freshwater Harmful

Algal Blooms (HABs). She worked with Dr. George S. Bullerjahn and Dr. R. Michael L. McKay

researching the freshwater cyanobacteria Microcystis aeruginosa. This work included 2 field

courses abroad focused on assessing anthropogenic influences on HABs in the Baltic Sea

(Zingst, Germany) and Lake Balaton (Balaton, Hungary). During the summer of 2016, she

conducted a National Science Foundation Research Experience for Undergraduates at Dauphin

Island Sea Laboratory in the lab of Dr. Jeffrey Krause and Dr. Behzad Mortazavi, where she

worked on a marine ecology project investigating how dissolved organic matter influences

coastal phytoplankton communities. In 2018, she joined the lab of Dr. Steven Wilhelm at the

University of Tennessee Knoxville Department of Microbiology. Here, she worked on a

multitude of topics within freshwater systems and HABs, developed a model freshwater diatom

system in the lab, and partook in national (Lake Erie, United States) and international (Lake

Victoria, Kenya) research expeditions concerning aquatic ecosystem health and HABs. She

defended her PhD at the University of Tennessee Knoxville in April 2023 and was honored to

receive the Jimmy and Ileen Graduate Student Medal of Excellence.

249