University of Iowa
Iowa Research Online
Theses and Dissertations
Summer 2016
Genomic and phenotypic consequences of
asexuality
Joel Sharbrough
University of Iowa
Copyright 2016 Joel Sharbrough
This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/2140
Recommended Citation
Sharbrough, Joel. "Genomic and phenotypic consequences of asexuality." PhD (Doctor of Philosophy) thesis, University of Iowa,
2016.
http://ir.uiowa.edu/etd/2140.
Follow this and additional works at: http://ir.uiowa.edu/etd
Part of the Biology Commons
GENOMIC AND PHENOTYPIC CONSEQUENCES OF ASEXUALITY
by
Joel Sharbrough
A thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Biology
in the Graduate College of
The University of Iowa
August 2016
Thesis Supervisor: Associate Professor Maurine Neiman
Copyright by
Joel Sharbrough
2016
All Rights Reserved
Graduate College
The University of Iowa
Iowa City, Iowa
CERTIFICATE OF APPROVAL
____________________________
PH.D. THESIS
_________________
This is to certify that the Ph.D. thesis of:
Joel Sharbrough
has been approved by the Examining Committee for
the thesis requirement for the Doctor of Philosophy degree
in Biology at the August 2016 graduation.
Thesis Committee:
____________________________________________
Maurine Neiman, Thesis Supervisor
____________________________________________
John M. Logsdon, Jr.
____________________________________________
John H. Fingert
____________________________________________
Andrew Forbes
____________________________________________
Sarit Smolikove
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my wife Natalie. She has put up with me for
the past 7 years with grace and humor, and I could not have made it through graduate school
without her. Secondly, I would like to thank my advisor, Maurine Neiman. She has been as
supportive of an advisor as I could ever imagine and I am lucky to have been a part of the
Neiman Lab. A special thanks to all my fellow graduate students at the University of Iowa,
especially Emily Beck, Kyle Flippo, Kyle McElroy, Laura Bankers, Katelyn Larkin, Claire
Tucci, Andrew Adrian, Chris Rice, Nick Stewart, and Setu Vora, who have all shared failure,
success, and (more importantly) beer with me. I would also like to thank my committee John
Logsdon, John Fingert, Andrew Forbes, and Sarit Smolikove for all their help in navigating
this process. Getting a PhD takes a village, and I could not have figured it out by myself. Cindy
Toll deserves a special place in every Neiman/Logsdon lab member’s heart and we all miss
having her around. Last, but definitely not least, I want to thank the army of undergraduates
that I have gotten a chance to work with over the past six years: Michelle Sullivan, Nicole
Enright, Meagan Luse, JD Woodell, Emma Griemann, Madeline Peters, Jorge Moreno, Collin
Thatcher, Samantha Hennessey, Ashley Urra, Michael Kline, Praakruti Cherukuri, Nikhil
Puttagunta, and Michelle Zhang.
ii
ABSTRACT
Sexual reproduction is expected to facilitate the removal of deleterious mutations from
populations because biparental inheritance (i.e., segregation) and recombination during meiosis
break down linkage disequilibria (LD), allowing mutations to be selected independently from
their genetic background. Accordingly, the absence of recombination and segregation is
expected to increase selective interference between loci, translating into reduced efficacy of
natural selection. While there now exist multiple lines of evidence demonstrating that asexual
lineages do experience accelerated accumulation of putatively harmful mutations, whether
these mutations influence phenotype in a manner that could contribute to the maintenance of
sex remains almost entirely unevaluated. Here, I use the New Zealand freshwater snail,
Potamopyrgus antipodarum, to address these questions. In particular, I take advantage of the
fact that the mitochondrial genome is expected to suffer from these mutational effects and
interacts extensively with the nuclear genome to evaluate potential harmful effects of mutation
accumulation in asexuals on a genome-wide scale. I present evidence that harmful mutations
remain extant longer in asexual populations than in sexual populations, that the degree of
functional constraint determines the extent of mutation accumulation in asexuals, that there is
genetic variation for mitochondrial function in asexual lineages of P. antipodarum, and that
phenotypic variation for mitochondrial function is mediated by both genetic and environmental
variation. Together, these analyses provide strong evidence that asexual lineages are
accumulating deleterious mutations, and that there is genetic variation, structured by lake, for
mitochondrial function.
iii
PUBLIC ABSTRACT
Why sexual reproduction is so overwhelmingly common in nature despite its profound
costs remains an open question in evolutionary biology. One potential explanation for the
predominance of sex resides in the benefits of genetic recombination. A likely important
positive outcome of recombination is that it allows organisms to produce offspring that do not
inevitably inherit harmful mutations bourn by their parent(s). The mutation accumulation
expected to occur unchecked in asexual lineages has been put forth as an explanation for the
predominance of sex and the generally short lives of asexual lineages. Empirical evidence in
support of this mutational hypothesis remains scarce, in large part because very few organisms
can be used to directly test whether sex facilitates the removal of harmful mutations. I have
focused my thesis on one such organism, Potamopyrgus antipodarum, a New Zealand
freshwater snail featuring coexisting and competing but otherwise similar sexual and asexual
lineages to test 1) whether asexual lineages show signs of harmful mutation accumulation in
their mitochondrial genomes and 2) whether these accumulated mutations translate into
decreased function. I show that asexual P. antipodarum exhibit a lower rate of removal of
harmful mutations relative to sexuals, and that the extent of mutation accumulation in asexuals
depends upon the intensity of natural selection. I also found extensive variation for metabolic
function in P. antipodarum, stemming from both genetic and environmental factors, and that
this variation is population specific. Together this work provides a valuable test of a key
hypothesis for sex by showing that sex does facilitate the removal of harmful mutations in a
system where sex is actively maintained, but that the advantage gained by this mutational
clearance may not provide enough of a benefit to explain the persistence of sexual P.
antipodarum.
iv
TABLE OF CONTENTS
LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES ....................................................................................................................... ix
LIST OF ABBREVIATIONS ........................................................................................................ xi
CHAPTER 1: INTRODUCTION ................................................................................................... 1
Sexual reproduction facilitates the removal of harmful mutations .......................................................1
Delineating the individual fitness effects of mutations .........................................................................2
Mitochondria and the loss of sex ...........................................................................................................4
Potamopyrgus antipodarum: an emerging model snail system for the evolutionary
maintenance of sex ................................................................................................................................5
REFERENCES ..........................................................................................................................................8
CHAPTER 2: RADICAL AMINO ACID CHANGES PERSIST LONGER IN THE
ABSENCE OF SEX...................................................................................................................... 13
ABSTRACT ............................................................................................................................................13
INTRODUCTION ...................................................................................................................................14
Evolution of mitochondrial genomes in the absence of sex ................................................................14
Differential fitness effects of mutations and the efficacy of selection ................................................16
Potamopyrgus antipodarum: a snail model for the evolutionary maintenance of sex ........................17
MATERIALS & METHODS ..................................................................................................................19
Sequencing ..........................................................................................................................................19
Phylogenetic analysis ..........................................................................................................................20
Quantifying the rate of radical vs. conservative mutation and substitution ........................................20
Comparisons between sexual and asexual lineages ............................................................................24
RESULTS & DISCUSSION ...................................................................................................................26
Relative harm conferred by conservative vs. radical mutations ..........................................................26
Estimating the efficacy of purifying selection in sexuals vs. asexuals ...............................................29
TABLES ..................................................................................................................................................39
REFERENCES ........................................................................................................................................40
FIGURES .................................................................................................................................................50
SUPPLEMENTARY MATERIAL .........................................................................................................56
CHAPTER 3: INEFFICIENT PURIFYING SELECTION AND VARIATION IN
FUNCTIONAL CONSTRAINT DRIVES ACCELERATED BUT HETEROGENEOUS
ACCUMULATION OF HARMFUL MUTATIONS IN ASEXUAL LINEAGES OF A
FRESHWATER SNAIL ............................................................................................................... 63
ABSTRACT ............................................................................................................................................63
INTRODUCTION ...................................................................................................................................65
MATERIALS & METHODS ..................................................................................................................69
Evolutionary rate heterogeneity in the mitochondrial genome of P. antipodarum .............................69
Effect of rate heterogeneity on mutation accumulation in the absence of sex ....................................74
RESULTS ................................................................................................................................................76
Evolutionary rate heterogeneity in mitochondrial genome of P. antipodarum ...................................76
Effect of rate heterogeneity on mutation accumulation in the absence of sex ....................................78
DISCUSSION ..........................................................................................................................................80
Rate heterogeneity in mitochondrial genome of P. antipodarum .......................................................80
Effect of rate heterogeneity on mutation accumulation in the absence of sex ....................................82
TABLES ..................................................................................................................................................86
REFERENCES ........................................................................................................................................88
v
FIGURES .................................................................................................................................................93
CHAPTER 4: GENETIC VARIATION FOR MITOCHONDRIAL FUNCTION IN THE
NEW ZEALAND FRESHWATER SNAIL, POTAMOPYRGUS ANTIPODARUM ................. 103
ABSTRACT ..........................................................................................................................................103
INTRODUCTION .................................................................................................................................104
METHODS ............................................................................................................................................108
Snail husbandry .................................................................................................................................108
Mitochondrial function at the genomic level ....................................................................................108
Mitochondrial function at the organellar level ..................................................................................110
Mitochondrial function at the organismal level ................................................................................114
Comparison of mitochondrial functional assays ...............................................................................116
RESULTS ..............................................................................................................................................117
Mitochondrial function at the genomic level ....................................................................................117
Mitochondrial function at the organellar level ..................................................................................118
Mitochondrial function at the organismal level ................................................................................119
Comparison of mitochondrial functional assays ...............................................................................120
DISCUSSION ........................................................................................................................................122
Genetic variation for mt function at three different levels of biological organization ......................122
Relationship between mitochondrial functional assays.....................................................................124
Implications for the loss of sex..........................................................................................................125
REFERENCES ......................................................................................................................................128
FIGURES ...............................................................................................................................................137
SUPPLEMENTARY MATERIAL .......................................................................................................142
CHAPTER 5: INTRASPECIFIC DIVERGENCE FOR METABOLIC FUNCTION IN
POTAMOPYRGUS ANTIPODARUM, AN EMERGING MODEL FOR THE
EVOLUTIONARY MAINTENANCE OF SEX ........................................................................ 147
ABSTRACT ..........................................................................................................................................147
INTRODUCTION .................................................................................................................................148
MATERIALS & METHODS ................................................................................................................151
Field collections of P. antipodarum ..................................................................................................151
Aquatic respirometry .........................................................................................................................152
Determination of reproductive mode ................................................................................................152
Statistical analyses .............................................................................................................................153
RESULTS ..............................................................................................................................................154
DISCUSSION ........................................................................................................................................156
Phenotypic variation for mitochondrial function depends upon lake of origin and environment .....156
Implications for the loss of sex..........................................................................................................158
REFERENCES ......................................................................................................................................161
FIGURES ...............................................................................................................................................166
SUPPLEMENTARY MATERIAL .......................................................................................................168
CHAPTER 6: CONCLUSIONS AND FUTURE DIRECTIONS .............................................. 175
SUMMARY OF FINDINGS .................................................................................................................175
Genomic consequences of asexuality ................................................................................................175
Phenotypic consequences of asexuality ............................................................................................177
CONTRIBUTIONS OF FINDINGS TO THE FIELD ..........................................................................178
Population genetics and molecular evolution ....................................................................................178
Development of phenotypic markers for P. antipodarum .................................................................180
NEW AND OPEN QUESTIONS REGARDING THE MAINTENANCE OF SEX,
MITOCHONDRIAL FUNCTION, AND P. ANTIPODARUM ............................................................182
vi
Do asexual lineages accumulate harmful mutations more rapidly than related sexual lineages
in their nuclear genomes? ..................................................................................................................182
Does mitochondrial performance predict fitness outcomes in P. antipodarum? ..............................183
Do asexuals experience a reduction in fitness as a consequence of mutation accumulation? ..........184
Is genomic and phenotypic variation between lake populations a consequence of local
adaptation or population structure? ...................................................................................................185
CONCLUSIONS AND FUTURE PROSPECTS ..................................................................................187
REFERENCES ......................................................................................................................................188
vii
LIST OF TABLES
Table 2-1. Amino acid classification schemes. ............................................................................. 39
Table 2-S1. Summary of source populations of Potamopyrgus antipodarum.............................. 56
Table 2-S2. Primers used to amplify and sequence whole mtDNA in P. antipodarum. .............. 58
Table 2-S3. Number of sites for different mutational types across seven amino acid
classification schemes in the invertebrate mitochondrial genetic code. ....................................... 59
Table 3-1. Rates and patterns of natural selection across the 13 protein-coding genes and
OXPHOS complexes encoded by the mitochondrial genome in P. antipodarum. ....................... 86
Table 3-2. Log likelihood of PAML models of molecular evolution in each of the 13 proteincoding genes of the mitochondrial genome .................................................................................. 87
Table 4-S1. Lineages and samples sizes used to compare mitochondrial function in P.
antipodarum. ............................................................................................................................... 145
Table 5-S1. Field collected samples from seven New Zealand lakes. ........................................ 171
Table 5-S2. ANOVA Tables ....................................................................................................... 172
viii
LIST OF FIGURES
Figure 2-1. Whole-mt genome maximum likelihood phylogeny for P. antipodarum and an
outgroup, P. estuarinus. ..............................................................................................................50
Figure 2-2. Molecular evolution of conservative and radical changes in mt genomes of P.
antipodarum. ...............................................................................................................................52
Figure 2-3. Ratios of polymorphism to divergence for conservative and radical changes in
sexual vs. asexual lineages of P. antipodarum. ..........................................................................54
Figure 2-4. Mean number of unique polymorphisms per genome (θU) for sexual vs.
asexual lineages of P. antipodarum. ...........................................................................................55
Figure 3-1. Sliding window estimates of Jukes-Cantor-corrected dN/dS across the P.
antipodarum mitochondrial genome estimated using a window size of 100 bp and a step
size of 3 bp. .................................................................................................................................93
Figure 3-2. The number of substitutions per gene as a function of the number of
polymorphisms per gene. ............................................................................................................94
Figure 3-3. Relationship between dS and dN in mitochondrial genes of P. antipodarum as
estimated by the best-fit model M0 (species-wide rate; atp8), M1a (branch-specific rate;
all other 12 genes), or M2a (sexual-specific rate, asexual specific rate; 0 genes) in PAML. .....96
Figure 3-4. Complex identity governs molecular evolution of mitochondrial genes in P.
antipodarum at both interspecific and intraspecific levels. ........................................................98
Figure 3-5. Relationship between gene expression (FPKM) and OXPHOS complex. ..............99
Figure 3-6. Scatterplots depicting relationship between species-wide dN/dS (estimated by
species-wide model M0 in PAML) and the difference between mean branch-specific
estimates of dN/dS (estimated by model M1a in PAML) in asexuals vs. sexuals. ....................101
Figure 3-7. The per-mitochondrial gene extent of mutation accumulation in asexuals due
to reduced Ne. ............................................................................................................................102
Figure 4-1. MtDNA copy number variation in six asexual lineages of P. antipodarum,
rank ordered in terms of increasing mean. ................................................................................137
Figure 4-2. Mitochondrial membrane potential in mitochondrial extracts of six lineages of
P. antipodarum. ........................................................................................................................138
Figure 4-3. Interaction plot between O2 consumption residuals, temperature, and snail
lineage. ......................................................................................................................................139
ix
Figure 4-4. Comparisons of mitochondrial functional assays. .................................................141
Figure 4-S1. Variation in righting time in response to heat stress in P. antipodarum..............142
Figure 4-S2. Relationship between snail wet mass and oxygen consumption in P.
antipodarum. .............................................................................................................................144
Figure 5-1. Map of New Zealand lakes from which snails used in this study were
collected. ...................................................................................................................................166
Figure 5-2. Interaction plot depicting significant interaction between temperature and lake
of origin.....................................................................................................................................167
Figure 5-S1. Relationship between log-transformed mass and O2 consumption in fieldcollected snails. .........................................................................................................................170
x
LIST OF ABBREVIATIONS
ADP
ADENOSINE DI-PHOSPHATE
ATP
ADENOSINE TRIPHOSPHATE
ANOVA
ANALYSIS OF VARIANCE
BLAST
BASIC LOCAL ALIGNMENT SEARCH TOOL
bp
BASE PAIRS
CQ
QUANTITATION CYCLE
DFE
DIFFERENTIAL FITNESS EFFECTS OF MUTATIONS
DNA
DEOXYRIBONUCLEIC ACID
FSC
FORWARD SCATTER
ETC
ELECTRON TRANSPORT CHAIN
FET
FISHER’S EXACT TEST
FPKM
FRAGMENTS PER KILOBASE MAPPED
KS
JUKES-CANTOR-CORRECTED SYNONYMOUS SUBSTITUTION RATE
KA
KC
JUKES-CANTOR-CORRECTED NONSYNONYMOUS SUBSTITUTION
RATE
JUKES-CANTOR-CORRECTED CONSERVATIVE SUBSTITUTION RATE
KR
JUKES-CANTOR-CORRECTED RADICAL SUBSTITUTION RATE
LD
LINKAGE DISEQUILIBRIUM
LRT
LIKELIHOOD RATIO TEST
MK
MCDONALD-KREITMAN TEST OF SELECTION
ML
MAXIMUM LIKELIHOOD
mt
MITOCHONDRIAL
mtDNA
MITOCHONDRIAL DNA
Ne
EFFECTIVE POPULATION SIZE
NI
NEUTRALITY INDEX
NU-MTS
NUCLEAR-ENCODED MITOCHONDRIAL GENE
OXPHOS
OXIDATIVE PHOSPHORYLATION
PCR
POLYMERASE CHAIN REACTION
qPCR
QUANTITATIVE POLYMERASE CHAIN REACTION
RNA
RIBONUCLEIC ACID
xi
RNA-seq
RNA SEQUENCING
s
SELECTION COEFFICIENT
SSC
SIDE SCATTER
xii
CHAPTER 1: INTRODUCTION
Sexual reproduction facilitates the removal of harmful mutations
Sexual reproduction is expected to be extremely costly relative to asexual reproduction. In
particular, the investment of resources into the production of males incurs a substantial cost
because only females can directly produce offspring. The implications are that, all else being
equal, the number of females in an asexual population will rapidly outpace the number of
females in a sexual population. This differential rate of daughter production between sexuals and
asexuals means that asexual populations should grow at a higher rate than sexual populations,
ultimately driving the sexual competitors to extinction (Maynard Smith 1978). Sexual
reproduction is also associated with other potentially major costs including recombination load
(Charlesworth and Barton 1996), increased predation risk during copulation (reviewed in
Magnhagen 1991), and sexually transmitted diseases (Lockhart et al. 1996, Kaltz and Shykoff
2001), amongst others. Even so, sexual reproduction remains overwhelmingly common,
indicating that all else must not be equal and that sex must confer profound benefits in spite of its
costs.
At least part of the explanation for the predominance of sex likely resides in benefits of
meiotic recombination, a fundamental mechanism by which genetic diversity is generated
(reviewed in Neiman and Schwander 2011). In particular, recombination increases the efficacy
of selection by allowing distinct loci to be selected independently from their genetic background
(i.e., Hill-Robertson effect; Hill and Robertson 1966, Birky and Walsh 1988). Recombination
also allows organisms to produce offspring that do not inevitably inherit harmful mutations
bourn by their parent(s) (i.e., Muller’s Ratchet, Muller 1964). This logic leads to the expectation
1
that asexual lineages should be characterized by accelerated accumulation of harmful mutations
relative to sexual counterparts, and that this mutation accumulation should result in severe
declines in fitness for asexual lineages, eventually driving them extinct (Lynch et al. 1993). The
typically “twiggy” phylogenetic distribution of most asexual lineages is consistent with this
prediction, suggesting that asexual lineages are short lived relative to sexual taxa (e.g., Williams
1975, Maynard Smith 1978, Bell 1982, but see Schwander and Crespi 2009). There are also data
from a variety of plant (Horandl and Hojsgaard 2012, Marriage and Orive 2012, Voigt-Zielinski
et al. 2012, Hollister et al. 2015) and animal (Cutter and Payseur 2003, Neiman et al. 2010,
Tucker et al. 2013) taxa that do suggest that asexual lineages are subject to relatively high rates
of mutation accumulation. What remains largely uncharacterized is the extent to which these
“deleterious” mutations actually decrease fitness in asexual lineages. As such, the seemingly
inevitable extinction of asexual lineages remains an unresolved question both empirically (i.e.,
are asexually reproducing organisms inevitably doomed to extinction?) and mechanistically (i.e.,
how do asexual lineages go extinct?). Answering these questions requires a more complete
understanding of both the individual fitness effects of mutations as well as the cumulative effects
of genome-wide mutation accumulation on fitness and function. Together, these multiple lines of
investigation will provide powerful insight into whether and how deleterious mutation
accumulation affects competition between sexual and asexual lineages and the ultimate fate of
asexual lineages.
Delineating the individual fitness effects of mutations
Understanding the individual fitness effects of mutations requires that one examine the mutations
within functional context in which they occur. A good example of this context dependence is
2
provided by nonsynonymous mutations, which alter protein sequence, a primary predictor of
protein function (Lesk and Chothia 1980, Geisler and Weber 1982, Doms et al. 1988, Rumbley
et al. 2001). Deleterious nonsynonymous changes should therefore decrease protein function, and
potentially, fitness. Some evidence for this link between mutation accumulation and fitness
comes from evidence that these mutations are associated with protein dysfunction in a variety of
contexts including aging cell lineages (Hirsch 1980), tumorigenesis (Hafner et al. 2010, Morton
et al. 2010), mitochondrial function decline in mammals (Das 2003), and cellular senescence in
vegetative plant tissue (Ally et al. 2010). Together, these studies suggest that, on average,
accumulation of nonsynonymous mutations tend to be harmful. What these studies leave
unresolved is the types of mutations that tend to be harmful and how these mutations influence
otherwise similar asexual vs. sexual organisms and in natural contexts.
A powerful framework with which to address these questions of connections between
mutation and phenotype is provided by considering the consequences of different types of amino
acid substitutions: amino acid substitutions with similar biochemical properties to the ancestral
amino acid – so-called “conservative” mutations – are expected to exhibit less severe effects on
phenotype than amino acids with very different biochemical properties compared to the ancestral
amino acid – so called “radical” mutations. Incorporating the severity of mutation into analyses
of mutation accumulation can shed light upon the types of mutations that asexuals tend to
accumulate, thereby providing insight into the mechanisms governing asexual lineage fitness
decline.
3
Mitochondria and the loss of sex
Perhaps even more important to our understanding of deleterious mutation accumulation is the
cumulative effects mutations have on complex phenotypes (i.e., phenotypes determined by
multiple genes) that themselves influence relative fitness. Because genotype-phenotype
connections are often complex, determining how mutations influence phenotypic quality is not
straightforward. A particularly striking example of this complexity is provided by mitochondria:
mitochondrial function (the result of which is ATP production, the primary energy source in
eukaryotes) is determined by >100 genes encoded by two completely different genomes (i.e.,
nuclear and mitochondrial genomes). The implications are that the extent to which mitochondrial
function declines due to mitochondrial mutation accumulation depends upon both the individual
fitness effects of mutations in mitochondrial genes as well as the nuclear-encoded genes directly
interacting with those mitochondrial mutations. Declines in mitochondrial function have been
linked to organismal fitness in a number of animal systems, including Drosophila simulans
(James and Ballard 2003) and the copepod Tigriopus californicus (Ellison and Burton 2006). In
both of these cases, mitonuclear discordance (i.e., incompatibilities between interacting
mutations) appears to be driving decreased mitochondrial function and organismal fitness.
Maintaining proper mitochondrial function thus requires that harmful mutations in the
mitochondrial genome must be compensated by mutations in the nuclear genome and vice versa
(Howe and Denver 2008). The ability of asexual lineages to maintain mitochondrial function in
the face of mutation accumulation is therefore likely to be dependent upon both the nature of
these mitochondrial mutations as well as the extent to which linkage between nuclear and
mitochondrial genomes interferes with compensatory evolution. Together, the unique pattern of
bi-genomic inheritance for mitochondrial function and its close relationship to organismal fitness
4
makes mitochondria-related phenotypes very well suited to apply to investigation of the
consequences of lost sex, and all in a setting that is experimentally tractable and well understood.
As such, measuring mitochondrial function and organismal respiration in otherwise similar
sexual and asexual organisms will provide a novel and important test of how mutation
accumulation affects phenotype- and fitness-related traits in asexual lineages.
Potamopyrgus antipodarum: an emerging model snail system for the evolutionary
maintenance of sex
Rigorous and direct study of the benefits and maintenance of sex requires comparisons between
sexual and asexual populations that coexist and compete in nature and are otherwise
phenotypically and ecologically similar (Maynard Smith 1978, Bell 1982). Potamopyrgus
antipodarum is one of the only systems that satisfy these criteria (Maynard Smith 1978, Lively
1987, Jokela et al. 1997), making it an ideal model organism in which to study the evolutionary
maintenance of sex. Potamopyrgus antipodarum also features multiple separate transitions to
asexuality (Dybdahl and Lively 1995, Neiman and Lively 2004, Paczesniak et al. 2013), yielding
elegant natural experiments in which it is possible to compare a diverse array of closely related
sexual and asexual lineages.
Potamopyrgus antipodarum is the focus of a large body of research aimed at
understanding the evolutionary maintenance of sex. This research has demonstrated that asexual
P. antipodarum accumulate putatively harmful mitochondrial DNA (mtDNA) mutations more
rapidly than sexual conspecifics (Neiman et al. 2010). Similar patterns of accelerated mutation
accumulation have been documented in a variety of asexual systems including bdelloid rotifers
(Mark Welch and Meselson 2000), Daphnia pulex (Paland and Lynch 2006, Tucker et al. 2013),
5
Timema stick insects (Henry et al. 2012), yeast (Gray and Goddard 2012), Tetrahymena protists
(Brito et al. 2010), and the neo-Y (nonrecombining) chromosomes of Drosophila miranda
(Kaiser and Charlesworth 2010). In all of these cases, whether such mutation accumulation
affects fitness in a manner in which these fitness reductions could be relevant to the maintenance
of sex or asexual lineage extinction has not been evaluated.
For my dissertation research, I attempted to address these important knowledge gaps by
evaluating whether rates and patterns of molecular evolution in the mitochondrial genomes of
asexual P. antipodarum display evidence for reduced efficacy of selection and increased rates of
harmful mutation accumulation compared to their sexual counterparts (Chapters 2, 3) and
whether traits associated with the genes encoded in the mitochondrial genome exhibit evidence
of decline in asexual vs. sexual lineages of P. antipodarum (Chapters 4, 5). Specifically, I tested
whether different types of amino acid changes accumulate at different rates in the mitochondrial
genomes of sexual and asexual P. antipodarum (Chapter 2) and compared rates of evolution
among individual mitochondrial genes across sexuals and asexuals and as a function of gene
expression (Chapter 3). To test whether traits associated with mitochondrial genes exhibit
functional decline in asexual relative to sexual P. antipodarum, I first developed new methods to
assay mitochondrial function in P. antipodarum and then used these methods to evaluate whether
P. antipodarum exhibited genetic variation for mitochondrial function (Chapter 4). I then applied
these newly developed methods to attempt to determine whether sexuals and asexuals differ in
mitochondrial function in terms of organismal oxygen consumption (Chapter 5). Altogether,
these analyses indicate that deleterious mutations remain polymorphic longer in asexual lineages
than in sexual lineages, that some genes contribute to mutation accumulation in asexuals more
than others, that there exists substantial genetic variation for mitochondrial function in P.
6
antipodarum, and that this variation in mitochondrial function appears to be population specific.
This thesis provides an important empirical test of a key hypothesis for the maintenance of sex:
that sexual reproduction facilitates the removal of mutations that have negative phenotypic
consequences. More broadly, my combinatorial use of genetic, molecular, and organismal
techniques represents a novel and integrative approach to understand how mutations affect (and
are removed from) natural populations.
7
REFERENCES
Ally, D., Ritland, K., & Otto, S. P. (2010). Aging in a long-lived clonal tree. PLoS Biology, 8,
e1000454.
Bell, G. (1982). The masterpiece of nature: the evolution and genetics of sexuality. Berkeley:
University of California Press.
Birky, C. W., Jr., & Walsh, J. B. (1988). Effects of linkage on rates of molecular evolution.
Proceedings of the National Academy of Sciences USA, 85, 6414-6418.
Brito, P. H., Guilherme, E., Soares, H., & Gordo, I. (2010). Mutation accumulation in
Tetrahymena. BMC Evolutionary Biology, 10, 354.
Charlesworth, B., & Barton, N. H. (1996). Recombination load associated with selection for
increased recombination. Genetical Research, 67, 27-41.
Cutter, A. D., & Payseur, B. A. (2003). Rates of deleterious mutation and the evolution of sex in
Caenorhabditis. Journal of Evolutionary Biology, 16, 812-822.
Das, A. M. (2003). Regulation of the mitochondrial ATP-synthase in health and disease.
Molecular Genetics and Metabolism, 79, 71-82.
Doms, R. W., Ruusala, A., Machamer, C., Helenius, J., Helenius, A., & Rose, J. K. (1988).
Differential effects of mutations in 3 domains on folding, quaternary structure, and
intracellular-transport of vesicular stomatitis virus G-protein. Journal of Cell Biology,
107, 89-99.
Dybdahl, M. F., & Lively, C. M. (1995). Host-parasite interactions: infection of common clones
in natural populations of a freshwater snail (Potamopyrgus antipodarum). Proceedings of
the Royal Society of London Series B: Biological Sciences, 260, 99-103.
8
Ellison, C. K., & Burton, R. S. (2006). Disruption of mitochondrial function in interpopulation
hybrids of Tigriopus californicus. Evolution, 60, 1382-1391.
Geisler, N., & Weber, K. (1982). The amino acid sequence of chicken muscle desmin provides
a common structural model for intermediate filament proteins. EMBO Journal, 1, 16491656.
Gray, J. C., & Goddard, M. R. (2012). Sex enhances adaptation by unlinking beneficial from
detrimental mutations in experimental yeast populations. BMC Evolutionary Biology, 12,
43.
Hafner, C., Toll, A., Fernandez-Casado, A., Earl, J., Marques, M., Acquadro, F., . . . Real, F. X.
(2010). Multiple oncogenic mutations and clonal relationship in spatially distinct benign
human epidermal tumors. Proceedings of the National Academy of Sciences USA, 107,
20780-20785.
Henry, L., Schwander, T., & Crespi, B. J. (2012). Deleterious mutation accumulation in asexual
Timema stick insects. Molecular Biology and Evolution, 29, 401-408.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection.
Genetical Research, 8, 269-294.
Hirsch, H. R. (1980). Commitment theory of cellular aging: possibility of an immortal diploid
cell strain. Mechanisms of Ageing and Development 12, 25-30.
Hollister, J. D., Greiner, S., Wang, W., Wang, J., Zhang, Y., Wong, G. K., . . . Johnson, M. T.
(2015). Recurrent loss of sex is associated with accumulation of deleterious mutations in
Oenothera. Molecular Biology and Evolution, 32, 896-905.
Horandl, E., & Hojsgaard, D. (2012). The evolution of apomixis in angiosperms: A reappraisal.
Plant Biosystems, 146, 681-693.
9
Howe, D. K., & Denver, D. R. (2008). Muller's Ratchet and compensatory mutation in
Caenorhabditis briggsae mitochondrial genome evolution. BMC Evolutionary Biology, 8,
62.
James, A. C., & Ballard, J. W. O. (2003). Mitochondrial genotype affects fitness in Drosophila
simulans. Genetics, 164, 187-194.
Jokela, J., Lively, C. M., Dybdahl, M. F., & Fox, J. A. (1997). Evidence for a cost of sex in the
freshwater snail Potamopyrgus antipodarum. Ecology, 78, 452-460.
Kaiser, V. B., & Charlesworth, B. (2010). Muller's ratchet and the degeneration of the
Drosophila miranda neo-Y chromosome. Genetics, 185, 339-348.
Kaltz, O., & Shykoff, J. A. (2001). Male and female Silene latifolia plants differ in per-contact
risk of infection by a sexually transmitted disease. Journal of Ecology, 89, 99-109.
Lesk, A. M., & Chothia, C. (1980). How different amino acid sequences determine similar
protein structures - structure and evolutionary dynamics of the globins. Journal of
Molecular Biology, 136, 225-270.
Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by
parasitism. Nature, 328, 519-521.
Lockhart, A. B., Thrall, P. H., & Antonovics, J. (1996). Sexually transmitted diseases in animals:
ecological and evolutionary implications. Biological Reviews, 71, 415-471.
Lynch, M., Bürger, R., Butcher, D., & Gabriel, W. (1993). The mutational meltdown in asexual
populations. Journal of Heredity, 84, 339-344.
Magnhagen, C. (1991). Predation risk as a cost of reproduction. Trends in Ecology and Evolution
6, 183-186.
10
Mark Welch, D., & Meselson, M. (2000). Evidence for the evolution of bdelloid rotifers without
sexual reproduction or genetic exchange. Science, 288, 1211-1215.
Marriage, T. N., & Orive, M. E. (2012). Mutation-selection balance and mixed mating with
asexual reproduction. Journal of Theoretical Biology, 308, 25-35.
Maynard Smith, J. (1978). The evolution of sex. Cambridge: Cambridge University Press.
Morton, J. P., Timpson, P., Karim, S. A., Ridgway, R. A., Athineos, D., Doyle, B., . . . Sansom,
O. J. (2010). Mutant p53 drives metastasis and overcomes growth arrest/senescence in
pancreatic cancer. Proceedings of the National Academy of Sciences USA, 107, 246-251.
Muller, H. J. (1964). The relation of recombination to mutational advance. Mutational Research,
106, 2-9.
Neiman, M., Hehman, G., Miller, J. T., Logsdon, J. M., Jr., & Taylor, D. R. (2010). Accelerated
mutation accumulation in asexual lineages of a freshwater snail. Molecular Biology and
Evolution, 27, 954-963.
Neiman, M., & Lively, C. M. (2004). Pleistocene glaciation is implicated in the
phylogeographical structure of Potamopyrgus antipodarum, a New Zealand snail.
Molecular Ecology, 13, 3085-3098.
Neiman, M., & Schwander, T. (2011). Using parthenogenetic lineages to identify advantages of
sex. Evolutionary Biology, 3, 115-123.
Paland, S., & Lynch, M. (2006). Transitions to asexuality result in excess amino acid
substitutions. Science, 311, 990-992.
Rumbley, J., Hoang, L., Mayne, L., & Englander, S. W. (2001). An amino acid code for protein
folding. Proceedings of the National Academy of Sciences USA, 98, 105-112.
11
Schwander, T., & Crespi, B. J. (2009). Twigs on the tree of life? Neutral and selective models for
integrating macroevolutionary patterns with microevolutionary processes in the analysis
of asexuality. Molecular Ecology, 18, 28-42.
Tucker, A. E., Ackerman, M. S., Eads, B. D., Xu, S., & Lynch, M. (2013). Population-genomic
insights into the evolutionary origin and fate of obligately asexual Daphnia pulex.
Proceedings of the National Academy of Sciences USA, 110, 15740-15745.
Voigt-Zielinski, M. L., Piwczynski, M., & Sharbel, T. F. (2012). Differential effects of
polyploidy and diploidy on fitness of apomictic Boechera. Sexual Plant Reproduction,
25(2), 97-109.
Williams, G. C. (1975). Sex and evolution. Princeton: Princeton University Press.
12
CHAPTER 2: RADICAL AMINO ACID CHANGES PERSIST
LONGER IN THE ABSENCE OF SEX
ABSTRACT
Harmful mutations are ubiquitous and inevitable, and the rate at which these mutations are
removed from populations is a critical determinant of evolutionary fate. Closely related and
otherwise similar sexual and asexual taxa provide a particularly powerful setting in which to
study deleterious mutation elimination because sex should facilitate mutational clearance by
reducing selective interference between sites. Here, I compared the rate of removal of
conservative and radical nonsynonymous mutations in sexual vs. asexual populations of
Potamopyrgus antipodarum, a New Zealand freshwater snail species featuring coexisting and
ecologically similar sexual and asexual lineages. My analyses revealed that radical changes are
removed from populations at significantly higher rates than conservative changes and that sexual
lineages eliminate these radical changes more rapidly than asexual counterparts, especially over
relatively short time scales. Taken together, these results indicate that reduced efficacy of
purifying selection in asexual lineages allows harmful mutations to remain polymorphic longer
than in sexual lineages, potentially influencing the outcome of competition between sexual and
asexual lineages. That the ability to detect differential patterns of mutational clearance in sexual
vs. asexual individuals required polymorphism data emphasizes the critical importance of
population-level sampling for characterizing evolutionary phenomena.
13
INTRODUCTION
One of the primary hypothesized advantages for sexual reproduction is the clearance of harmful
mutations, which is expected to be much more effective when linkage disequilibria (LD) are
disrupted by sex (Hill and Robertson 1966). A particularly striking and important example of
how sex facilitates LD breakdown is provided by the transfer of mitochondrial (mt) genomes to
new and potentially divergent nuclear genomic backgrounds from parents (usually but not
always mothers; Barr et al. 2005) to sexually produced offspring. The close interactions between
nuclear and mt gene products mean that the consequences of breakdown of mitonuclear LD are
likely to be substantial. A good example of these consequences is provided by the common
observation that changes in nuclear genomic background can substantially decrease
mitochondrial function (Ellison and Burton 2006, Meiklejohn et al. 2013, Pichaud et al. 2013),
likely a result of coevolution between the mt genome and the nuclear genes that encode the
interacting protein subunits of the oxidative phosphorylation (OXPHOS) pathway. Indeed,
proper function of these subunits appears to be an important determinant of eukaryotic health
(e.g., Chen et al. 2007, Pike et al. 2007, Barreto and Burton 2013, Muir et al. 2016).
Evolution of mitochondrial genomes in the absence of sex
With respect to coevolved mitonuclear protein complexes like those found in the OXPHOS
pathway, the loss of sex in the nuclear genome is expected to result in at least one of two nonmutually exclusive evolutionary consequences. First, the biparental inheritance and meiotic
recombination that the nuclear genome experiences during canonical sexual reproduction will
decrease linkage disequilibrium (LD) in both the nuclear and mt genomes, increasing the
efficacy of natural selection by decreasing interference from linked sites (Hill and Robertson
14
1966, Neiman and Taylor 2009). By contrast, the uniparental (e.g., maternal) inheritance and
reduced or absent meiotic recombination that is a feature of asexual reproduction will decrease
effective population size (Ne) and increase LD across both genomes, decreasing the efficacy of
selection by increasing selective interference (i.e., Hill-Robertson effect, Hill and Robertson
1966). More specifically, nuclear and mt genomes “trapped” in asexual lineages are cotransmitted as a single genetic unit, such that the two genomes are effectively linked (Normark
and Moran 2000). This scenario underlies the expectation that asexual lineages will accumulate
deleterious mutations more rapidly than otherwise similar sexual lineages in both the nuclear
(Birky and Walsh 1988, Charlesworth 1993, Lynch et al. 1993) and mt genomes (Normark and
Moran 2000, Neiman and Taylor 2009). Because mitonuclear incompatibilities are expected to
increase with divergence (Burton and Barreto 2012) and mitonuclear linkage decreases the
efficacy of selection (Normark and Moran 2000, Neiman and Taylor 2009), accelerated
accumulation of mildly deleterious mutations within asexual lineages should also allow novel
mitonuclear combinations, untested by nature, to arise over time.
Second, the absence of sex may allow for elevated rates of mitonuclear coevolution if the
co-transmission of mt and nuclear genomes in asexuals allows selection to act more effectively
on multilocus (e.g., mitonuclear) genotypes that interact epistatically but tend to be disrupted by
recombination (Neiman and Linksvayer 2006). Thus, one important potential consequence of cotransmittance of nuclear and mt genomes in the context of mutation accumulation is that the
permanent linkage of these two genomes may allow for relatively strong and effective selection
for compensatory mutations in response to accumulated deleterious mutations in OXPHOS genes.
The first scenario (increased mutation accumulation in asexuals) is expected to favor
sexual lineages that can effectively remove deleterious mutations, while the second scenario
15
(tighter coevolution in asexuals) should increase the cost of sex (i.e., recombination load, see
Maynard Smith 1978), thereby favoring asexual lineages. Both scenarios are expected to result in
more rapid accumulation of nonsynonymous mutations in the mt genomes of asexual lineages
than sexual lineages, making it essential to evaluate the relative fitness effects (i.e., deleterious,
neutral, or beneficial, and to what extent) of the mutations that accumulate in asexual lineages.
Differential fitness effects of mutations and the efficacy of selection
The prediction that the mt genomes of asexuals should experience a higher rate of accumulation
of nonsynonymous mutations has found support from animal (Cutter and Payseur 2003, Neiman
et al. 2010, Henry et al. 2012) and plant (Horandl and Hojsgaard 2012, Voigt-Zielinsji et al.
2012, Hollister et al. 2015) taxa. While these results represent important steps towards
understanding the genomic consequences of asexuality, the evolutionary mechanisms underlying
these observations remain unclear, in large part because the extent to which accumulated
mutations are actually deleterious in asexuals has not been evaluated. Here, I depart from
previous studies assessing mutational load in asexuals that treat all nonsynonymous mutations as
a monolithic “deleterious” class (but see Henry et al. 2012) despite evidence that
nonsynonymous mutations are likely to vary widely in fitness effects (Keightley and EyreWalker 2007). Predicting and characterizing the fitness effects of these nonsynonymous
mutations remains a major challenge in evolutionary biology (Keightley and Charlesworth 2005,
Xue et al. 2008, Eyre-Walker and Keightley 2009, Halligan et al. 2011). Here, I show that
partitioning nonsynonymous changes into “conservative” changes (amino acid changes in which
the derived amino acid has similar biochemical properties to the ancestral amino acid) and
“radical” changes (amino acid changes in which the derived amino acid has markedly different
16
biochemical properties compared to the ancestral amino acid) represents a relatively
straightforward method for predicting the relative harm conferred by different types of amino
acid changes (also see Zhang 2000, Smith 2003, Hanada et al. 2007, Popadin et al. 2007).
Because mutational severity largely governs the efficacy with which selection can act
upon any given mutation (Fisher 1930), partitioning nonsynonymous changes into conservative
and radical mutational types also allows for an intuitive means of comparing the efficacy of
selection in sexual vs. asexual genomes. At present, commonly used methods for inferring
selection on DNA sequence data (e.g., π, θ, Tajima’s D, dN/dS, etc.) do not typically incorporate
information from the distribution of fitness effects (DFE) of mutations (but see Kryazhimskiy
and Plotkin 2008, Schneider et al. 2011). Here, I provide a proof of principle for using
mutational information (i.e., radical vs. conservative changes) to infer the type, intensity, and
efficacy of selection using traditional tests of selection (e.g., McDonald-Kreitman test of
selection, π, θ, KA/KS), applied to the mt genomes of sexual and asexual snails.
Potamopyrgus antipodarum: a snail model for the evolutionary maintenance of sex
The removal of deleterious mutations (and the fixation of beneficial mutations) depends upon the
efficacy of selection as well as the fitness effect of the mutation(s), such that elevated Ne and
reduced LD in sexual vs. asexual populations should result in more rapid removal of deleterious
mutations for the former (Birky and Walsh 1988). As such, comparing rates and patterns of
evolution across reproductive modes while incorporating DFE information will provide a unique
and powerful glimpse into the evolutionary dynamics governing the removal of deleterious
mutations. The New Zealand freshwater snail Potamopyrgus antipodarum is ideally suited to
evaluate this critically important evolutionary process because otherwise similar obligately
17
sexual and obligately asexual P. antipodarum frequently coexist within New Zealand lake
populations (Lively 1987, Jokela et al. 1997), enabling direct comparisons across reproductive
modes, and, thereby, across lineages that vary in the efficacy of selection. Asexual lineages of P.
antipodarum are the product of multiple distinct transitions from sexual P. antipodarum (Neiman
and Lively 2004, Neiman et al. 2011), meaning that these asexual lineages represent separate
natural experiments into the consequences of the absence of sex. Neiman et al. (2010) showed
that asexual lineages of P. antipodarum experience a higher rate of nonsynonymous substitution
in their mt genomes than sexual lineages. Here, I use an expanded mt genomic dataset to
evaluate whether sexual lineages distinguish between radical and conservative changes more
effectively than asexual lineages at interspecific and intraspecific levels. This approach allowed
me to evaluate for the first time whether harmful mutation accumulation is visible at the
polymorphic level and, if so, whether this phenomenon is driven by more effective selection in
sexual lineages vs. relatively rapid mitonuclear coevolution in asexual lineages. The outcome of
these analyses emphasizes fundamental differences in the rate of accumulation of conservative vs.
radical nonsynonymous mutations and suggests that radical mutations persist longer in mt
genomes of asexual lineages of P. antipodarum compared to sexual counterparts, likely a
consequence of reduced efficacy of purifying selection.
18
MATERIALS & METHODS
Sequencing
I analyzed 31 whole P. antipodarum mt genomes from 8 sexual lineages and 23 asexual lineages,
representing the natural range of this species in New Zealand along with several invasive
lineages (European, North American, see Table 2-S1, Figure 2-1). Eighteen of these genomes (4
sexual, 14 asexual) were obtained from Genbank (Accession Nos.: GQ996416 – GQ996433),
along with the whole mt genome of an outgroup species, Potamopyrgus estuarinus (Accession
No.: GQ996415.1) (Neiman et al. 2010). Five mt genomes (2 sexual lineages, 3 asexual lineages)
were assembled from the DNA sequence data generated via Illumina technology from the
ongoing P. antipodarum nuclear genome project. The remaining eight mt genomes (1 sexual
lineage, 7 asexual lineages) were newly sequenced via bi-directional Sanger sequencing on an
ABI 3730. DNA for the Sanger-sequenced lineages was extracted with a mollusk-adapted
phenol-chloroform extraction protocol (Fukami et al. 2004). Mt genomes were amplified in four
overlapping fragments using primers and programs designed in Neiman et al (2010). PCR
products were cleaned with Shrimp Exo shrimp alkaline phosphatase (Werle et al. 1994) and
directly sequenced with internal sequencing primers (Table 2-S2). The newly generated mt
genome sequence data were assembled and manually edited in Sequencher 5.0. For these eight
newly generated mt genome sequences, only unambiguous sites with ≥ 2 Sanger reads were used
in my analyses. All new sequences will be deposited in GenBank.
I used flow cytometry (following the protocol outlined in Neiman et al. 2011, Neiman et
al. 2012, Paczesniak et al. 2013, Krist et al. 2014) to assign ploidy and thus reproductive mode
(diploid – sexual; polyploid – asexual) to the newly sequenced lineages for which ploidy had not
already been determined.
19
Phylogenetic analysis
Concatenated mt genome sequences have been shown to produce more accurate tree topologies
than single gene trees (Rokas et al. 2003, Gadagkar et al. 2005). Accordingly, I concatenated
nucleotide sequences from ~11 kbp protein-coding nucleotides from each of the 31 P.
antipodarum lineages and from P. estuarinus, for a total of 32 concatenated sequences. The
concatenated sequences were aligned in the correct reading frame using the ClustalW package
implemented in MEGA 5.2.2 (Kumar et al. 2008) and manually edited (alignment available upon
request). To minimize the effects of selection on tree topology, I used only 3rd-position sites to
infer the mt genome phylogeny using the maximum likelihood (ML) methods implemented in
MEGA 5.2.2 software (Kumar et al. 2008), using the ML model selection tool in MEGA 5.2.2 to
select the Tamura-Nei model of molecular evolution with gamma-distributed sites (Tamura and
Nei 1993). Tree topology was assessed using 1,000 bootstrap replicates and visualized using
FigTree v1.4 (Raumbaut 2007): only nodes with bootstrap support > 60 were relied upon for
tests of molecular evolution. The tree topology that I obtained (Figure 2-1) is qualitatively
identical to previously published mitochondrial trees for P. antipodarum (Neiman et al. 2004,
Neiman et al. 2010, Paczesniak et al. 2013).
Quantifying the rate of radical vs. conservative mutation and substitution
I used seven different amino acid classification schemes drawn from Zhang (2000), Hanada et al.
(2007), and a modified Grantham scheme based on amino acid composition, polarity, and
volume (Grantham 1974) to evaluate patterns of radical and conservative nonsynonymous
polymorphism and substitution in the mt genome of P. antipodarum (Table 2-1). I defined
radical mutations as mutations in which the derived amino acid was from a different category
20
than the ancestral amino acid, while conservative mutations were defined as mutations in which
the derived amino acid and the ancestral amino acid were in the same category. While there is
some overlap between different classification schemes, each scheme highlights different amino
acid properties that are likely to shape protein evolution. To wit, amino acid charge is a major
determinant of protein folding (Perutz et al. 1965, Anfinsen 1973, Nakashima et al. 1986,
Bashford et al. 1987, Wright et al. 2005) and three-dimensional structure (Lesk and Chothia
1980, Geisler and Weber 1982, Doms et al. 1988, Rumbley et al. 2001), and uncharged to
charged amino acid changes (and vice versa) are rarely maintained. Amino acid polarity is
particularly important for proper membrane integration, as phospholipid membranes are highly
hydrophobic, and changes between polar and non-polar amino acids may expose or bury key
interaction residues (Von Heijne 1992). Volume and aromaticity can both affect protein folding
(e.g., proline is a structure breaker) and can play a role in protein-protein interactions (Burley
and Petsko 1985). Classification schemes 4 and 7 are unique in that they are based on
evolutionary information (although classification scheme 7 largely fits with charge and polarity
classifications), meaning that these schemes incorporate aspects of other amino acid
characteristics into their classifications. All mutational types were defined relative to the
invertebrate mt genetic code (http://www.ncbi.nlm.nih.gov/Taxonomy/Utils/wprintgc.cgi#SG5).
The number of synonymous, conservative nonsynonymous, and radical nonsynonymous sites per
codon are detailed in Table 2-S3.
Because the number of substitutions per site can be used as an estimate of the rate of
substitution (Li et al. 1985), I counted the number of nucleotide substitutions and determined the
type of substitution (i.e., synonymous, conservative nonsynonymous, and radical
nonsynonymous) in the protein-coding regions of the mt genomes (excluding codons with >1
21
change) of each lineage relative to P. estuarinus. I then calculated the number of mutational
target sites per lineage for a given type of mutation (see Table 2-S3). To confirm that the number
of each of type of site was properly calculated, I checked that the number of nonsynonymous
sites and the number of synonymous sites per codon summed to three and that the number of
conservative nonsynonymous sites and the number of radical nonsynonymous sites per codon
summed to the number of nonsynonymous sites per codon (Hanada et al. 2007). Of particular
note is that the “GTG”, “TTG”, “ATT”, “ATC”, and “ATA” codons can all be used as
alternative start codons in invertebrate mt genomes (only the GTG alternative start codon was
observed in the present dataset), which I accounted for in my site calculations. I then used
custom-built Python scripts (available upon request) to calculate substitution rates for each
mutational type and used the Jukes-Cantor correction to account for multiple hits (Figure 2-2 A,
Jukes and Cantor 1969).
Substitutions between species are generally older than polymorphisms within species
(McDonald 1996), such that different analyses implicitly assume particular and distinct time
scales (i.e., substitutions: relatively old time scale, hereafter “long time scale”; polymorphisms:
relatively recent time scale, hereafter “short time scale”; ratio of polymorphism to divergence –
composite of recent and old time scales, hereafter “composite time scale”). I compared the mean
rate of conservative nonsynonymous substitution (Jukes-Cantor corrected; number of
conservative nonsynonymous substitutions/conservative nonsynonymous site; KC) to the mean
rate of radical nonsynonymous substitution (Jukes-Cantor corrected; number of radical
nonsynonymous substitutions/radical nonsynonymous site; KR) from each of the seven amino
acid classification schemes to the Jukes-Cantor-corrected mean synonymous substitution rate
(number of synonymous substitutions/synonymous site; KS) and to one another using pairwise
22
Mann-Whitney U (MWU) tests and the Holm procedure for modified Bonferroni correction for
multiple comparisons (Holm 1979). All statistical tests were performed in R (R Core Team
2012). In subsequent analyses, I corrected for systematic across-lineage differences in underlying
mutation rate by dividing the estimate of KR and KC for each lineage by KS.
Because polymorphisms within species give rise to divergences between species, and
because higher ratios of polymorphism to divergence are indicative of more intense purifying
selection (McDonald and Kreitman 1991), I used McDonald Kreitman (MK) tests of selection to
evaluate whether conservative changes were more likely than radical changes to reach fixation
using a Fisher’s Exact Test (FET) with the Holm modification to the Bonferroni correction for
multiple comparisons for each of the seven amino acid categories. As an additional test of the
intensity of selection acting on conservative vs. radical changes, I compared the ratios of
polymorphism to divergence for conservative vs. radical changes by performing 10,000
bootstrap replicates. P-values were inferred by comparing the resulting bootstrap distributions
(Efron and Tibshirani 1993) and significance was evaluated using the Holm modification to the
Bonferroni correction for multiple comparisons for each of the seven amino acid categories
(Figure 2-2 B).
Finally, I evaluated nucleotide diversity, π (Tajima 1989), nucleotide heterozygosity, θ
(Watterson 1975), and Tajima’s D (Tajima 1989) in conservative vs. radical sites to test whether
conservative and radical sites differ at the intraspecific (i.e., relatively recent) level.
Polymorphism at sites under more stringent purifying selection will tend to exist at lower relative
frequencies than at sites under less stringent purifying selection. To compare these intraspecific
measures of molecular evolution in all seven amino acid classification schemes I performed
10,000 bootstrap replicates, inferred p-values from the resulting distributions, and evaluated
23
statistical significance using the Holm modification to the Bonferroni correction for multiple
comparisons (Figure 2-2 C).
Comparisons between sexual and asexual lineages
Although my sampling scheme provided a comprehensive picture of the mitochondrial diversity
present in P. antipodarum (see Figure 2-1, Neiman and Lively 2004, Neiman et al. 2011,
Paczesniak et al. 2013), I lacked the statistical power to use standard Phylogenetic Independent
Contrast methods (Felsenstein 1985) to compare sexuals and asexuals while accounting for
phylogenetic non-independence. Instead, I employed custom Python scripts to randomly sample
(with replacement) sexual and asexual lineages to compare rates and patterns of conservative and
radical nonsynonymous evolution at short, long, and composite time scales. I used 10,000
bootstrap replicates to estimate means and 95% CIs for KC/KS, KR/KS, PC/DC, PR/DR, πC/πS, πR/πS,
θC/θS, and θR/θS in sexual vs. asexual lineages in all seven amino acid classification schemes. I
estimated two-tailed p-values by determining the probability of overlap of the bootstrap
distributions and determined statistical significance using the Holm procedure to perform a
modified Bonferroni correction for multiple comparisons.
To account for possible effects of differential sampling of sexual (n = 8) vs. asexual (n =
23) lineages, I used the methods described above to compare rates and patterns of conservative
vs. radical molecular evolution across reproductive modes by generating 10,000 bootstrap
replicates using equal (n = 8) and randomly chosen with replacement samples for sexual and
asexual lineages. The results of this equivalent sampling analysis are depicted in Figure 2-3.
Because population demography can influence patterns of molecular evolution (Tajima
1989), and because transitions to asexuality and/or the rapid growth expected for a newly-
24
generated asexual lineage might influence important demographic parameters (e.g., Ne; Kaiser
and Charlesworth 2009), it is important to account for potential effects of demography when
evaluating molecular evolution in sexuals vs. asexuals (Wright and Charlesworth 2001). I dealt
with this issue by modifying the Watterson estimator of nucleotide heterozygosity θ (Watterson
1975) by calculating the mean number of unique mutations per mt genome in sexuals vs.
asexuals, θU:
(1)
θU =
SU
! an
where SU is the number of unique polymorphisms (i.e., changes that appear only within a single
lineage), L is the number of mutational target sites, and an is:
(2)
an =
n-1 !
i =1 !
in which n is equal to the number of lineages per reproductive mode. Fu and Li (1993) have
shown that θU is an estimate of nucleotide heterozygosity, θ, that only uses the external branch
lengths of a population (see equation 18, Fu and Li 1993). I compared the mean number of
unique polymorphisms per site for each mutational type across reproductive modes using 95%
CIs generated by 10,000 bootstrap replicates (Figure 2-4).
One group of asexual P. antipodarum that I included in these analyses is quite genetically
distinct (mean pairwise distance between clade A and (clade B, clade C) = 0.035) from other
lineages (mean pairwise distance within (clade B, clade C) = 0.013) (Figure 2-1). Therefore,
including clade A may cause an overestimation of the extent of mutation accumulation in asexual
P. antipodarum as a whole. In order to exclude the possibility that this divergent group of
asexual lineages may be driving the observed pattern of amino acid evolution, I repeated the
comparisons of molecular evolution with clade A excluded, using the clade B sexuals as an
outgroup to clade C (see Figure 2-1).
25
RESULTS & DISCUSSION
Relative harm conferred by conservative vs. radical mutations
I used ~11 kbp of mt protein-coding sequence representing ~8520 nonsynonymous sites
generated from eight sexual and 23 asexual P. antipodarum lineages (Table 2-S1, Figure 2-1)
and seven distinct amino acid classification schemes (Table 2-1) to evaluate the relative
harmfulness of radical vs. conservative mutations and address whether reproductive mode
influences the rate of accumulation of these two types of mutations in mt genomes. I quantified
the number of each mutation type present across these 31 P. antipodarum lineages relative to an
outgroup, P. estuarinus (Table 2-S1, Figure 2-1). I detected a total of 35 nonsynonymous fixed
differences between these two species (mean conservative nonsynonymous fixed differences =
27.14, SD = 4.49; mean radical nonsynonymous fixed differences = 6.71, SD = 3.59).
Before comparing mutation accumulation across reproductive modes, I established the
type and intensity of selection acting on conservative vs. radical changes in the P. antipodarum
mitochondrial genome. To infer the type and intensity of selection acting on conservative vs.
radical changes at a relatively long time scale, I compared the Jukes-Cantor corrected rates of
conservative nonsynonymous substitution (KC) and radical nonsynonymous substitution (KR) to
the Jukes-Cantor corrected synonymous substitution rate (KS). I found that conservative and
radical nonsynonymous substitutions accumulate at significantly lower rates than synonymous
substitutions in all amino acid classification schemes (Figure 2-2 A). Additionally, conservative
amino acid changes contribute to divergence between P. antipodarum and P. estuarinus
significantly more than radical changes (Figure 2-2 A) in each of the seven amino acid
classification schemes (MWU: p < 0.0001 for all seven schemes). These results indicate that
while both conservative and radical mutations are fixed at discernably lower rates than
26
synonymous changes, radical amino acid changes are fixed significantly less often than
conservative amino acid changes.
I next compared the ratio of polymorphism to divergence for synonymous,
nonsynonymous, conservative, and radical differences by performing MK tests of selection
(Figure 2-2 B). I calculated the mean number of polymorphisms and divergences across amino
acid classification schemes and found that conservative polymorphisms (mean number of
conservative polymorphisms = 122.14, SD = 27.25) and radical polymorphisms (mean number
of radical polymorphisms = 63.86, SD = 27.25) were significantly less likely to reach fixation
than synonymous polymorphisms (number of synonymous polymorphisms = 619) (FET: p < 2.2
x 10-16 for both mutational types). I also found that conservative polymorphisms were
significantly more likely to contribute to divergence than radical polymorphisms (FET: p =
0.049). I also compared the mean ratio of polymorphism to divergence for conservative (PC/DC)
vs. radical (PR/DR) changes using 10,000 bootstrap replicates. I found that mean PR/DR was
significantly higher than mean PC/DC by comparing bootstrap distributions (p < 0.0002). These
findings reveal that while both conservative and radical changes appear to be evolving under
strong purifying selection in P. antipodarum mt genomes, radical nonsynonymous changes are
eliminated even more rapidly than conservative nonsynonymous changes.
I next compared nucleotide diversity, π, at synonymous, nonsynonymous, conservative,
and radical sites, which provides a picture of the strength and efficacy of selection at a relatively
short time scale. Sites under purifying selection are expected to exhibit lower levels of nucleotide
diversity than relatively neutral sites (e.g., nonsynonymous vs. synonymous sites, respectively)
(Nei and Gojobori 1986). Consistent with this prediction, I used 10,000 bootstrap replicates to
compare nucleotide diversity across site types and found that mean nonsynonymous nucleotide
27
diversity (πA = 0.0050) was significantly lower than mean synonymous nucleotide diversity (πS =
0.066) in P. antipodarum (p < 0.0002). Similarly, mean conservative nonsynonymous nucleotide
diversity (πC = 0.0060) and mean radical nonsynonymous nucleotide diversity (πR = 0.0027)
were significantly lower than mean πS (p < 0.0002 for πC and πR). When correcting for πS (e.g.,
πA/πS), I found that mean synonymous-corrected radical nucleotide diversity (πR/πS = 0.042) was
over two-fold lower than mean synonymous-corrected conservative nucleotide diversity (πC/πS =
0.090) (p < 0.0002), indicating that radical polymorphisms are maintained at lower frequencies
than conservative polymorphisms (Figure 2-2 C). Thus, even at a relatively short time scale,
radical amino acid changes appear to experience more stringent purifying selection than
conservative amino acid changes in P. antipodarum mt genomes.
Together, these results are consistent with the expectation that radical mutations are
usually more harmful than conservative mutations (Rand et al. 2000, Freudenberg-Hua et al.
2003, Smith 2003, Yampolsky et al. 2005). My results are important in demonstrating that
radical and conservative mutations appear to experience very different histories of selection in
natural populations. In particular, I provide some of the first evidence from a non-model system
featuring reproductive mode polymorphism (a primary determinant of Ne) that there exists more
stringent purifying selection on radical vs. conservative mutational types, a pattern detectable
even at the intraspecific level. While there is substantial overlap amongst classification schemes,
the consistent signal of more stringent purifying selection acting on radical changes is evidence
that the grouping strategies employed by each scheme (e.g., charge, polarity, volume, etc.) are
founded on important biological properties of amino acids. My results also emphasize that the
relative degree of amino acid change should be an important consideration in evaluating patterns
of selection and suggest that the standard grouping of radical and conservative mutational types
28
into a monolithic “nonsynonymous” class is often overly simplistic and may be positively
misleading (for an example, see Summary & Implications).
Comparisons across long, short, and composite time scales (e.g., substitution rates vs.
nucleotide diversity vs. ratio of polymorphism to divergence) reveal qualitatively similar results
across all three time scales (Figure 2-2). In particular, radical changes clearly experience a higher
intensity of purifying selection than do conservative changes, indicating that radical changes
usually impart substantially more severe fitness effects than conservative changes and should
thus return to mutation-selection-drift equilibrium more rapidly than their conservative
counterparts (Figure 2-2, Fisher 1930). As such, I can use comparisons of these two mutational
types across different time scales to compare the efficacy with which sexual vs. asexual lineages
remove relatively mild (i.e., conservative) and relatively severe (i.e., radical) deleterious
mutations.
Estimating the efficacy of purifying selection in sexuals vs. asexuals
To estimate the relative differences in rates of evolution in sexual vs. asexual lineages of P.
antipodarum over a relatively long time scale, I compared synonymous substitution rates (KS),
conservative nonsynonymous substitution rates (KC), and radical nonsynonymous substitution
rates (KR) for sexuals vs. asexuals using 10,000 bootstrap replicates. I found that mean KS of
asexual P. antipodarum is significantly higher than KS of sexual P. antipodarum (p < 0.0002),
indicating that asexual P. antipodarum experience a higher rate of synonymous substitution than
sexuals (Table 2-S2; see also Neiman et al. 2010). To account for this difference, I corrected
estimates of KC and KR by KS and then compared KC/KS (mean sexual = 0.021, mean asexual =
0.022) and KR/KS (mean sexual = 0.0059, mean asexual = 0.0063) across reproductive modes
29
using 10,000 bootstrap replicates. While I did not detect significant differences between sexual
and asexual KC/KS or KR/KS (p = 0.061 and p = 0.18, respectively), I did recapitulate Neiman et
al. (2010) in detecting significantly higher KA/KS in asexuals vs. sexuals (p = 0.0034). Thus, by
subdividing the number of substitutions into conservative and radical categories, I likely lost
substantial power in my ability to detect differences across reproductive modes. This explanation
is especially likely considering there are only a total of 35 nonsynonymous fixed differences
between the P. antipodarum and P. estuarinus mt genome sequences. Absence of evidence for
significant differences in the conservative and radical substitution rates might also be linked to
the relatively recent and multiple derivations of asexuality in P. antipodarum (Neiman and
Lively 2004, Paczesniak et al. 2013). Indeed, my results suggest that the resolution with which
two groups of organisms (e.g., sexual vs. asexual) can be differentiated with substitutions might
decrease with increasing divergence from the outgroup. In other words, a large class of
deleterious mutations that are still evident on the intraspecific level have long since disappeared
at the interspecific level, rendering them invisible to methods of quantifying mutation
accumulation that only focus on divergence.
To compare the efficacy of selection in different sexuals vs. asexuals at a composite time
scale, I compared the mean ratios of polymorphism to divergence for conservative (PC:DC) and
radical (PR:DR) changes across reproductive modes in all seven amino acid classification
schemes. Because my sample included an unequal number of sexual lineages (n = 8) compared
to the number of asexual lineages (n = 23), and because the number of polymorphisms increases
with sample size, I performed pairwise comparisons of PC:DC and PR:DR across reproductive
modes by performing 10,000 bootstrap replicates of eight randomly sampled (with replacement)
sexual lineages and eight randomly sampled (with replacement) asexual lineages. While sexual
30
and asexual lineages exhibited significantly lower mean PC:DC than PR:DR (p = 0.0080 and p =
0.0082, respectively), the fact that mean sexual PCDC is significantly lower than mean asexual
PRDR (p = 0.0010), but that mean asexual PCDC is statistically indistinguishable from mean
sexual PRDR (p = 0.11, Figure 2-3) indicates that the probability of radical polymorphisms
proceeding to fixation is higher in asexual lineages than in sexual lineages. These results from a
composite time scale are consistent with sexual lineages eliminating radical, but not conservative,
polymorphisms more rapidly than asexual lineages.
I next compared nucleotide diversity (π) and nucleotide heterozygosity (θ) across
reproductive modes using 10,000 bootstrap replicates. To account for differential effects of
neutral processes (e.g., demographic changes, mutation rate, etc.) on nucleotide diversity in
sexuals vs. asexuals (Charlesworth and Wright 2001, Kaiser and Charlesworth 2009), and
because sexual πS was significantly lower than asexual πS (p = 0.011), I corrected πC and πR by
dividing each value by πS. Similarly, because asexuals exhibited significantly higher θS than
sexual lineages (p = 0.013), I corrected estimates of θC and θR by dividing by θS. I did not find
any significant differences in mean πC/πS, πR/πS, θC/θS, or θR/θS between sexuals and asexuals,
indicating that sexual and asexual P. antipodarum harbor similar numbers of conservative and
radical polymorphisms and at similar relative frequencies. At face value, these results are
consistent with a scenario in which sexual and asexual lineages experience similar efficacies of
selection. What must be taken into account, however, is that at the time of asexual origin, only a
single founding mt genome is transmitted to a newly asexual lineage. The consequence of this
bottleneck at the transition to a new asexual lineage is that ancestrally polymorphic sites become
immediately monomorphic within the new asexual lineage. This phenomenon should cause rapid
changes in allele frequencies and thus, changes in frequency-dependent measures of
31
polymorphism (e.g., π) of the sampled population. The implications are that such comparisons of
nucleotide diversity and heterozygosity between sexuals and asexuals do not necessarily provide
a complete or accurate picture of evolutionary processes.
To account for changes in allele frequency caused by the transition to asexuality itself, I
compared the mean number of unique polymorphisms per genome (see Methods) across
reproductive modes. By sampling only unique polymorphisms (i.e., relatively new mutations), I
was able to estimate mutation accumulation since the transition to asexuality. Sexuals and
asexuals did not differ in terms of the mean number of unique synonymous polymorphisms per
site (θU-S), indicating that new mutations arise at similar rates across reproductive modes (Figure
2-4). Sexuals exhibited significantly lower mean θU-R/θU-S than θU-C/θU-S (p < 0.0002), while
θU-C/θU-S and θU-R/θU-S were statistically indistinguishable in asexual lineages (p = 0.75),
indicating that selection recognizes and removes deleterious changes more rapidly in sexual
lineages than in asexual lineages (Figure 2-4). Further, these results indicate that differences in
the efficacy of selection across reproductive modes become apparent relatively quickly after the
transition to asexuality, especially with respect to mutational changes with relatively large
selection coefficients (i.e., radical mutations).
One group of asexuals was particularly genetically distinct relative to the rest of the P.
antipodarum dataset (clade A, Figure 2-1), raising the question of whether this group might be
contributing disproportionately to my observations of slowed radical mutation elimination in
asexual P. antipodarum. I addressed this possibility by performing the same analyses described
above on a subsample of P. antipodarum lineages (clade C) but treated the clade B sexuals as an
outgroup (see Methods) and excluded clade A entirely. The results of this more limited analysis
are largely consistent with the outcomes of analyses from the whole dataset, indicating that the
32
inclusion of the relatively divergent asexual clade A did not substantively affect the original
analysis. Notably, sexual lineages in clade C maintain radical polymorphisms as significantly
lower frequencies than conservative polymorphisms (mean πC/πS = 0.11, mean πR/πS = 0.058, p <
0.0002), while conservative and radical polymorphisms were maintained at statistically
indistinguishable levels in clade C asexual lineages (mean πC/πS = 0.13, mean πR/πS = 0.082, p =
0.27), consistent with ineffective selection in asexual lineages contributing to the retention of
radical polymorphisms.
Summary & Implications
Asexual P. antipodarum have already been found to exhibit elevated accumulation of
nonsynonymous substitutions in their mt genomes relative to sexual P. antipodarum (Neiman et
al. 2010). Here, I provide evidence that asexual P. antipodarum exhibit elevated ratios of
polymorphism to divergence for radical changes and harbor more unique radical polymorphisms
than sexuals, a particularly harmful type of nonsynonymous mutation. My analyses of the
relative effects of conservative vs. radical changes provide a novel line of evidence that the
mutations these asexual lineages are accumulating are deleterious. Together, these findings
indicate that asexual lineages of P. antipodarum likely experience an increased rate of
accumulation of harmful mutations than sexual conspecifics, a pattern that is observable at both
relatively long (substitution) (Neiman et al. 2010, present study) and relatively short
(polymorphism) time scales (present study).
Radical mutations appear more likely to be deleterious than conservative mutations and
asexual P. antipodarum appear to be accumulating these mutations more rapidly than sexual P.
antipodarum, raising the intriguing possibility that asexual P. antipodarum might exhibit
33
decreased mitochondrial function compared to sexual counterparts. The presumed severity of
some mutations in these genomes (e.g., a nonsense mutation in nd2 of one asexual lineage that
would truncate ND2 by three amino acids) suggests that either mitochondrial function is
decreased in at least some asexual lineages or that asexuals possess one or more mechanisms to
compensate for deleterious mutation load (e.g., RNA editing). RNA editing of mt-encoded
transcripts has been observed in a variety of plant taxa (Covello and Gray 1989, Gualberto et al.
1989) and in land snail mt-encoded tRNAs (Yokobori and Paabo 1995), but it is unclear whether
P. antipodarum employs similar strategies. Future work evaluating mitochondrial function at the
organelle and organismal levels in P. antipodarum will be essential to understanding how the
efficacy of selection influences the maintenance and distribution of sex in this system.
My study also provides a clear demonstration of a situation where inclusion of all
nonsynonymous changes in a monolithic “deleterious” category can obscure important
evolutionary dynamics, especially at the polymorphic level. A particularly illuminating example
of the potential for this type of grouping to result in misleading conclusions is that when
nonsynonymous changes are treated as a single group, the number of unique nonsynonymous
mutations per genome (θU-A/θU-S) in sexual vs. asexual lineages is statistically indistinguishable,
in stark contrast to the clear distinctions between sexual and asexual P. antipodarum that are
revealed by taking mutational effect into account. The implications are, that by partitioning
mutations into conservative and radical bins, I gain substantial resolution at the intraspecific
level. Similarly, I find that sexual and asexual P/D ratios are statistically indistinguishable if all
nonsynonymous mutations are grouped together. At a longer time scale (i.e., divergence), the
relative dearth of radical nonsynonymous changes appears to present temporal sampling issues,
34
such that differences between sexual and asexual P. antipodarum can no longer be detected. This
result highlights the importance of representative intraspecific sampling and analysis.
While I interpret these results as resulting from less effective selection in asexual lineages,
another possible (and non-mutually exclusive) explanation is that the co-transmission (and thus,
effective linkage) between the nuclear and mt genomes in asexuals has facilitated the persistence
and spread of beneficial nonsynonymous mutations via selection imposed by cooperation with
nuclear-encoded genes (Blier et al. 2001, Meiklejohn et al. 2007). Because asexuals co-transmit
their nuclear and mt genomes, mutations in either genome may cause decreases in mitochondrial
function. Therefore, long-term co-transmission of the nuclear and mt genomes may provide a
scenario in which asexuals experience relatively strong selection favoring compensatory
mutation(s). I have not detected any evidence of positive selection acting in the mt genome of P.
antipodarum (e.g., codon-by-codon dN/dS < 1, Neutrality Index > 1 for all 13 protein-coding
genes, sliding window πA/πS < 1 at all sites, data not shown), though indirect evidence that a
particular mt haplotype is spreading amongst asexual lineages hints that selection favoring
particular mt haplotypes or mitonuclear combinations might be involved (Paczesniak et al. 2013).
Evaluation of rates and patterns of evolution in the nuclear-encoded mt genes that make up
≥95% of the genes that influence mitochondrial function (Sardiello et al. 2003), coupled with the
functional analyses mentioned above, will ultimately be needed to determine whether
mitonuclear linkage in asexuals is at least in part responsible for elevated retention of apparently
harmful mutations in mt genomes.
My results taken from different time points in the P. antipodarum evolutionary history
lead me to conclude that patterns of protein evolution in this species are being driven both by
mutational severity and reproductive mode. In particular, at the shortest time scale (i.e., number
35
of unique polymorphisms), sexuals and asexuals differ in the number of radical but not
conservative changes. At an older though still population-level scale (i.e., nucleotide diversity for
clade C only), radical polymorphisms reside at higher frequencies in asexuals rather than sexuals.
Differences across reproductive modes disappear at even older time scales (i.e., species-wide
nucleotide diversity, rates of substitution), although the composite time scale revealed that
asexuals have higher ratios of polymorphism to divergence than sexuals, particularly with
respect to radical changes. While these observations might appear inconsistent, the implicit time
scale assumption of each respective measurement instead hints that the selective process plays
out at different rates in sexuals vs. asexuals. Namely, my data indicate that mutations in asexuals
experience smaller Ne|s| than in sexuals, meaning that the elimination of apparently deleterious
mutations occurs more slowly in asexual than in sexual lineages. Theoretical (Ohta 1987,
Charlesworth et al. 1993, Charlesworth and Wright 2001) and empirical (Wright et al. 2008,
Katju et al. 2015) work support this conclusion, in that populations with low Ne are expected to
harbor a larger proportion of “effectively neutral” mutations than populations with large Ne. My
data are consistent with this phenomenon, revealing that amino acid-changing mutations
(especially radical changes) in the mt genome remain polymorphic longer in asexual than in
sexual lineages, an observation only possible with a multiple time scales approach.
Ultimately, less efficient removal of deleterious mutations in asexual lineages is only
important to the maintenance of sex and/or the persistence of asexual lineages if those mutations
in fact negatively affect fitness. Recent empirical evidence suggests that harmful mutations
indeed play a role in asexual lineage deterioration: Tucker et al. (2013) found that obligately
asexual Daphnia suffer from gene conversion-type processes that decrease heterozygosity and
subsequently expose deleterious recessive alleles, leading to lineage deterioration (Tucker et al.
36
2013). By contrast, my data indicate that asexual P. antipodarum harbor elevated numbers of
deleterious mutations due to less efficient removal of existing mutations rather than accelerated
acquisition of new mutations. Because mt genotype is critically important to organismal function
and fitness (Ellison and Burton 2006, Meiklejohn et al. 2013, Pichaud et al. 2013, Muir et al.
2016), this increased load of likely harmful mutations could potentially contribute to negative
phenotypic consequences in asexuals, though slowed deleterious mutation removal in asexuals
would likely also need to be prevalent in the nuclear genome in order to provide the short-term
advantages necessary to maintain sexual reproduction within a population (Lynch et al. 1993).
Given that the nuclear genome is the site of the vast majority of gene content and recombination
in sexual lineages, My observation of elevated retention of deleterious mutations in mt genomes
of asexual P. antipodarum leads me to predict that the nuclear genome is also likely to exhibit
substantial differences in deleterious mutational load and efficacy of selection across
reproductive modes.
37
ACKNOWLEDGEMENTS
I thank Cindy Toll and Gery Hehman for their assistance with DNA sequencing. I also thank
Stephen I. Wright and Aneil F. Agrawal for helpful discussions regarding interpretation of
intraspecific data. I thank Samuel J. Fahrner for helpful discussions regarding non-parametric
bootstrapping. I thank anonymous reviewers who saw previous versions of this manuscript for
their helpful comments. The National Science Foundation (NSF: MCB – 1122176; DEB –
1310825) and the Iowa Academy of Sciences (ISF #13-10) funded this research.
38
TABLES
Table 2-1. Amino acid classification schemes.
39
REFERENCES
Anfinsen, C. B. (1973). Principles that govern the folding of protein chains. Science, 181,
223-230.
Barr, C. M., Neiman, M., & Taylor, D. R. (2005). Inheritance and recombination of
mitochondrial genomes in plants, fungi and animals. New Phytologist, 168, 39-50.
Barreto, F. S., & Burton, R. S. (2013). Elevated oxidative damage is correlated with reduced
fitness in interpopulation hybrids of a marine copepod. Proceedings of the Royal Society
of London B, 280, 20131521.
Bashford, D., Chothia, C., & Lesk, A. M. (1987). Determinants of a protein fold - unique
features of the globin amino acid sequences. Journal of Molecular Biology, 196, 199-216.
Birky, C. W., Jr., & Walsh, J. B. (1988). Effects of linkage on rates of molecular evolution.
Proceedings of the National Academy of Sciences USA, 85, 6414-6418.
Blier, P. U., Dufresne, F., & Burton, R. S. (2001). Natural selection and the evolution of
mtDNA-encoded peptides: evidence for intergenomic co-adaptation. Trends in Genetics,
17, 400-406.
Burley, S. K., & Petsko, G. A. (1985). Aromatic-aromatic interaction: a mechanism of protein
structure stabilization. Science, 229, 23-28.
Charlesworth, B. (1993). The evolution of sex and recombination in a varying environment.
Journal of Heredity, 84, 345-350.
Charlesworth, B., Morgan, M. T., & Charlesworth, D. (1993). The effect of deleterious
mutations on neutral molecular variation. Genetics, 134, 1289-1303.
Charlesworth, D., & Wright, S. I. (2001). Breeding systems and genome evolution. Current
Opinion in Genetics and Development, 11, 685-690.
40
Chen, J. H., Hales, C. N., & Ozanne, S. E. (2007). DNA damage, cellular senescence and
organismal ageing: causal or correlative? Nucleic Acids Research, 35, 7417-7428.
Covello, P. S., & Gray, M. W. (1989). RNA editing in plant mitochondria. Nature, 341,
662-666.
Cutter, A. D., & Payseur, B. A. (2003). Rates of deleterious mutation and the evolution of sex
in Caenorhabditis. Journal of Evolutionary Biology, 16, 812-822.
Das, J., Miller, S. T., & Stern, D. L. (2004). Comparison of diverse protein sequences of the
nuclear-encoded subunits of cytochrome C oxidase suggests conservation of structure
underlies evolving functional sites. Molecular Biology and Evolution, 21, 1572-1582.
Doms, R. W., Ruusala, A., Machamer, C., Helenius, J., Helenius, A., & Rose, J. K. (1988).
Differential effects of mutations in 3 domains on folding, quaternary structure, and
intracellular-transport of vesicular stomatitis virus G-protein. Journal of Cell Biology,
107, 89-99.
Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap, monographs on statistics
and applied probability, Vol. 57. New York and London: Chapman and Hall/CRC.
Ellison, C. K., & Burton, R. S. (2006). Disruption of mitochondrial function in interpopulation
hybrids of Tigriopus californicus. Evolution, 60, 1382-1391.
Eyre-Walker, A., & Keightley, P. D. (2009). Estimating the rate of adaptive molecular
evolution in the presence of slightly deleterious mutations and population size change.
Molecular Biology and Evolution, 26, 2097-2108.
Felsenstein, J. (1985). Phylogenies and the comparative method. American Naturalist, 125, 1-15.
Fisher, R. A. (1930). The genetical theory of natural selection. Oxford: The Clarendon Press.
41
Freudenberg-Hua, Y., Freudenberg, J., Kluck, N., Cichon, S., Propping, P., & Nothen, M. M.
(2003). Single nucleotide variation analysis in 65 candidate genes for CNS disorders in a
representative sample of the European population. Genetical Research, 13, 2271-2276.
Fu, Y. X., & Li, W-H. (1993). Statistical tests of neutrality of mutations. Genetics, 133, 693-709.
Fukami, H., Budd, A. F., Levitan, D. R., Jara, J., Kersanach, R., & Knowlton, N. (2004).
Geographic differences in species boundaries among members of the Montastraea
annularis complex based on molecular and morphological markers. Evolution, 58, 324337.
Gadagkar, S. R., Rosenberg, M. S., & Kumar, S. (2005). Inferring species phylogenies from
multiple genes: Concatenated sequence tree versus consensus gene tree. Journal of
Experimental Zoology Part B Molecular and Developmental Evolution, 304B, 64-74.
Geisler, N., & Weber, K. (1982). The amino acid sequence of chicken muscle desmin provides
a common structural model for intermediate filament proteins. EMBO Journal, 1, 16491656.
Grantham, R. (1974). Amino acid difference formula to help explain protein evolution. Science,
185, 862-864.
Gualberto, J. M., Lamattina, L., Bonnard, G., Weil, J. H., & Grienenberger, J. M. (1989). RNA
editing in wheat mitochondria results in the conservation of protein sequences. Nature,
341, 660-662.
Halligan, D. L., Peters, A. D., & Keightley, P. D. (2003). Estimating numbers of EMS-induced
mutations affecting life history traits in Caenorhabditis elegans in crosses between inbred
sublines. Genetical Research, 82, 191-205.
42
Hanada, K., Shiu, S. H., & Li, W.-H. (2007). The nonsynonymous/synonymous substitution
rate ratio versus the radical/conservative replacement rate ratio in the evolution of
mammalian genes. Molecular Biology and Evolution, 24, 2235-2241.
Henry, L., Schwander, T., & Crespi, B. J. (2012). Deleterious mutation accumulation in
asexual Timema stick insects. Molecular Biology and Evolution, 29, 401-408.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection.
Genetical Research, 8, 269-294.
Hollister, J. D., Greiner, S., Wang, W., Wang, J., Zhang, Y., Wong, G. K., Wright, S. I., &
Johnson, M. T. (2015). Recurrent loss of sex is associated with accumulation of
deleterious mutations in Oenothera. Molecular Biology and Evolution, 32, 896-905.
Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal
of Statistics, 6, 65-70.
Horandl, E., & Hojsgaard, D. (2012). The evolution of apomixis in angiosperms: A reappraisal.
Plant Biosystematics, 146, 681-693.
Jokela, J., Lively, C. M., Dybdahl, M. F., & Fox, J. A. (1997). Evidence for a cost of sex in the
freshwater snail Potamopyrgus antipodarum. Ecology, 78, 452-460.
Jukes T. H., & Cantor, C. R. (1969). Evolution of protein molecules. Mammalian Protein
Metabolism Vol. 3. New York: Academic Press, Inc.
Kaiser, V. B., & Charlesworth, B. (2009). The effects of deleterious mutations on evolution in
non-recombining genomes. Trends in Genetics, 25, 9-12.
Katju, V., Packard, L. B., Bu, L., Keightley, P. D., & Bergthorsson, U. (2015). Fitness decline
in spontaneous mutation accumulation lines of Caenorhabditis elegans with varying
effective population sizes. Evolution, 69, 104-116.
43
Keightley, P. D., & Charlesworth, B. (2005). Genetic instability of C. elegans comes naturally.
Trends in Genetics, 21, 67-70.
Keightley, P. D., & Eyre-Walker, A. (2007). Joint inference of the distribution of fitness effects
of deleterious mutations and population demography based on nucleotide polymorphism
frequencies. Genetics, 177, 2251-2261.
Krist, A. C., Kay, A. D., Larkin, K., & Neiman, M. (2014). Response to phosphorus limitation
varies among lake populations of the freshwater snail Potamopyrgus antipodarum. PLoS
ONE, 9, e85845.
Kryazhimskiy, S., & Plotkin, J. B. (2008). The population genetics of dN/dS. PLoS Genetics,
4, e1000304.
Kumar, S., Nei, M., Dudley, J., Tamura, K. (2008). MEGA: a biologist-centric software for
evolutionary analysis of DNA and protein sequences. Briefings in Bioinformatics, 9, 299306.
Lesk, A. M., & Chothia, C. (1980). How different amino acid sequences determine similar
protein structures - structure and evolutionary dynamics of the globins. Journal of
Molecular Biology, 136, 225-270.
Li, W. H., Wu, C.-I., & Luo, C. C. (1985). A new method for estimating synonymous and
nonsynonymous rates of nucleotide substitution considering the relative likelihood of
nucleotide and codon changes. Molecular Biology and Evolution, 2, 150-174.
Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by
parasitism. Nature, 328, 519-521.
Lynch, M., Bürger, R., Butcher, D., & Gabriel, W. (1993). The mutational meltdown in asexual
populations. Journal of Heredity, 84, 339-344.
44
Maynard Smith, J. (1978). The evolution of sex. London: Cambridge University Press.
McDonald, J. H. (1996). Detecting non-neutral heterogeneity across a region of DNA sequence
in the ratio of polymorphism to divergence. Molecular Biology and Evolution, 13, 253260.
McDonald, J. H., & Kreitman, M. (1991). Adaptive protein evolution at the Adh locus in
Drosophila. Nature, 351, 652-654.
Meiklejohn, C. D., Holmbeck, M. A., Siddiq, M. A., Abt, D. N., Rand, D. M. &
Montooth, K. L. (2013). An incompatibility between a mitochondrial tRNA and its
nuclear-encoded tRNA synthetase compromises development and fitness in Drosophila.
PLoS Genetics, 9, e1003238.
Meiklejohn, C. D., Montooth, K. L., Rand, D. M. (2007). Positive and negative selection on
the mitochondrial genome. Trends in Genetics, 23, 259-263.
Muir, R., Diot, A., & Poulton, J. (2016). Mitochondrial content is central to nuclear gene
expression: Profound implications for human health. BioEssays, 38, 150-156.
Nakashima, H., Nishikawa, K., Ooi, T. (1986). The folding type of a protein is relevant to
the amino acid composition. Journal of Biochemistry, 99, 153-162.
Nei, M., & Gojobori, T. (1986). Simple methods for estimating the numbers of synonymous
and nonsynonymous nucleotide substitutions. Molecular Biology and Evolution, 3, 418426.
Neiman, M., Hehman, G., Miller, J. T., Logsdon, J. M. Jr., & Taylor, D. R. (2010).
Accelerated mutation accumulation in asexual lineages of a freshwater snail. Molecular
Biology and Evolution, 27, 954-963.
Neiman, M., Larkin, K., Thompson, A. R., & Wilton, P. (2012). Male offspring production by
45
asexual Potamopyrgus antipodarum, a New Zealand snail. Heredity, 109, 57-62.
Neiman, M., & Linksvayer, T. A. (2006). The conversion of variance and the evolutionary
potential of restricted recombination. Heredity, 96, 111-121.
Neiman, M., & Lively, C. M. (2004). Pleistocene glaciation is implicated in the
phylogeographical structure of Potamopyrgus antipodarum, a New Zealand snail.
Molecular Ecology, 13, 3085-3098.
Neiman, M., Paczesniak, D., Soper, D. M., Baldwin, A. T., & Hehman, G. (2011). Wide
variation in ploidy level and genome size in a New Zealand freshwater snail with
coexisting sexual and asexual lineages. Evolution, 65, 3202-3216.
Neiman, M., & Taylor, D. R. 2009. The causes of mutation accumulation in mitochondrial
genomes. Proceedings of the Royal Society of London B, 276, 1201-1209.
Normark, B. B., & Moran, N. A. (2000). Opinion - Testing for the accumulation of deleterious
mutations in asexual eukaryote genomes using molecular sequences. Journal of Natural
History, 34, 1719-1729.
Ohta, T. (1987). Very slightly deleterious mutations and the molecular clock. Journal of
Molecular Evolution, 26, 1-6.
Paczesniak, D., Jokela, J., Larkin, K., & Neiman, M. (2013). Discordance between nuclear and
mitochondrial genomes in sexual and asexual lineages of the freshwater snail
Potamopyrgus antipodarum. Molecular Ecology, 22, 4695-4710.
Perutz, M. F., Kendrew, J. C., & Watson, H. C. (1965). Structure and function of haemoglobin
– Some relations between polypeptide chain configuration and amino acid sequence.
Journal of Molecular Biology, 13, 669-678.
46
Pichaud, N., Ballard, J. W., Tanguay, R. M., & Blier, P. U. (2013). Mitochondrial haplotype
divergences affect specific temperature sensitivity of mitochondrial respiration. Journal
of Bioenergetics and Biomembranes, 45, 25-35.
Pike, T. W., Blount, J. D., Bjerkeng, B., Lindstrom, J., & Metcalfe, N. B. (2007). Carotenoids,
oxidative stress and female mating preference for longer lived males. Proceedings of the
Royal Society of London B, 274, 1591-1596.
Popadin, K., Polishchuk, L. V., Mamirova, L., Knorre, D., & Gunbin, K. (2007). Accumulation
of slightly deleterious mutations in mitochondrial protein-coding genes of large versus
small mammals. Proceedings of the National Academy of Sciences USA, 104, 1339013395.
Rand, D. M., Haney, R. A., & Fry, A. J. (2004). Cytonuclear coevolution: the genomics of
cooperation. Trends in Ecology and Evolution, 19, 645-653.
Rand, D. M., Weinreich, D. M., & Cezairliyan, B. O. (2000). Neutrality tests of conservativeradical amino acid changes in nuclear- and mitochondrially-encoded proteins. Gene, 261,
115-125.
Rambaut, A. (2007). FigTree, a graphical viewer of phylogenetic trees. See
http://tree.bio.ed.ac.uk/software/figtree.
Rodriguez, F., Oliver, J. L., Marin, A., & Medina, J. R. (1990). The general stochastic model of
nucleotide substitution. Journal of Theoretical Biology, 142, 485–501.
Rokas, A., Williams, B. L., King, N., Carroll, S. B. (2003). Genome-scale approaches to
resolving incongruence in molecular phylogenies. Nature, 425, 798-804.
Rumbley, J., Hoang, L., Mayne, L., & Englander, S. W. (2001). An amino acid code for protein
folding. Proceedings of the National Academy of Sciences USA, 98, 105-112.
47
Sardiello, M., Licciulli, F., Catalano, D., Attimonelli, M., & Caggese, C. (2003). MitoDrome: a
database of Drosophila melanogaster nuclear genes encoding proteins targeted to the
mitochondrion. Nucleic Acids Research, 31, 322-324.
Schneider, A., Charlesworth, B., Eyre-Walker, A., & Keightley, P. D. (2011). A method for
inferring the rate of occurrence and fitness effects of advantageous mutations. Genetics,
189, 1427-1437.
Smith, N. G. (2003). Are radical and conservative substitution rates useful statistics in molecular
evolution? Journal of Molecular Evolution, 57, 467-478.
Tajima, F. (1983). Evolutionary relationship of DNA sequences in finite populations. Genetics,
105, 437-460.
Tajima, F. (1989). The effect of change in population size on DNA polymorphism. Genetics,
123, 597-601.
Tamura, K., & Nei, M. (1993). Estimation of the number of nucleotide substitutions in the
control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology
and Evolution, 10, 512-526.
Tucker, A. E., Ackerman, M. S., Eads, B. D., Xu, S., & Lynch, M. (2013). Population-genomic
insights into the evolutionary origin and fate of obligately asexual Daphnia pulex.
Proceedings of the National Academy of Sciences USA, 110, 15740-15745.
Voigt-Zielinski, M. L., Piwczynski, M., & Sharbel, T. F. (2012). Differential effects of
polyploidy and diploidy on fitness of apomictic Boechera. Sexual Plant Reproduction, 25,
97-109.
von Heijne, G. (1992). Membrane protein structure prediction. Hydrophobicity analysis and the
positive-inside rule. Journal of Molecular Biology, 225, 487-494.
48
Watterson, G. A. (1975). On the number of segregating sites in genetical models without
recombination. Theoretical and Population Biology, 7, 256-276.
Werle, E., Schneider, C., Renner, M., Volker, M., & Fiehn, W. (1994). Convenient single step,
one tube purification of PCR products for direct sequencing. Nucleic Acids Research, 22,
4354-4355.
Wright, C. F., Teichmann, S. A., Clarke, J., & Dobson, C. M. (2005). The importance of
sequence diversity in the aggregation and evolution of proteins. Nature, 438, 878881.
Wright, S. I., Nano, N., Foxe, J. P., & Dar, V. U. (2008). Effective population size and tests of
neutrality at cytoplasmic genes in Arabidopsis. Genetical Research, 90, 119-128.
Yampolsky, L. Y., Kondrashov, F. A., & Kondrashov, A. S. (2005). Distribution of the strength
of selection against amino acid replacements in human proteins. Human Molecular
Genetics, 14, 3191-3201.
Yokobori, S., & Paabo, S. (1995). Transfer RNA editing in land snail mitochondria. Proceedings
of the National Academy of Sciences USA, 92, 10432-10435.
Zhang, J. (2000). Rates of conservative and radical nonsynonymous nucleotide substitutions in
mammalian nuclear genes. Journal of Molecular Evolution, 50, 56-68.
49
FIGURES
Figure 2-1. Whole-mt genome maximum likelihood phylogeny for P. antipodarum and an
outgroup, P. estuarinus.
Maximum likelihood-based phylogeny of 23 asexual (red) and eight sexual (blue) Potamopyrgus
antipodarum lineages using only 3rd position nucleotides of the ~11kbp protein-coding region of
the mitochondrial genome. Branch support was estimated using 1,000 bootstrap replicates; only
bootstrap values >50 are shown. Because clade A represents a particularly distinct group of
asexuals (mean pairwise distance between clade A and (clade B, clade C) = 0.035, mean pairwise
distance within (clade B, clade C) = 0.013), I removed clade A from the re-analysis of rates and
50
patterns of amino acid sequence evolution; clade C was used as the ingroup and clade B as an
outgroup in this analysis.
51
Figure 2-2. Molecular evolution of conservative and radical changes in mt genomes of P.
antipodarum.
A) Comparison of the Jukes-Cantor-corrected substitution rates across different mutational types.
Left: Substitutions per site; KS – synonymous, KA – nonsynonymous, KC – conservative
nonsynonymous, and KR – radical nonsynonymous. Right: Inset of boxed-in region depicting
only KA, KC, and KR. Error bars represent inner-quartile ranges (IQR). Statistical significance
was assessed using a Mann-Whitney U test. B) Ratio of polymorphism to divergence in sites
with conservative (white) vs. radical (gray) changes. Ratios for synonymous (diagonal stripes)
52
and nonsynonymous (black) mutational types are shown for comparison. Error bars represent
95% confidence intervals generated using 10,000 bootstrap replicates. Statistical significance in
PC:DC vs. PR:DR was assessed using Fisher’s Exact Test. C) Mean synonymous-corrected
nucleotide diversity for conservative (white) vs. radical (gray) sites in P. antipodarum.
Nucleotide diversity at nonsynonymous sites is shown for comparison. Error bars represent 95%
confidence intervals generated using 10,000 bootstrap replicates. Asterisks indicate significant
differences (* = p < 0.05, *** = p < 0.0002).
53
Figure 2-3. Ratios of polymorphism to divergence for conservative and radical changes in
sexual vs. asexual lineages of P. antipodarum.
Comparison of ratios of polymorphism to divergence for sexual (blue) vs. asexual (red) lineages
for conservative (semi-transparent) vs. radical (solid) changes. Error bars indicate 95% CIs
generated using 10,000 bootstrap replicates. Lower-case letters indicate statistical groupings
determined using the Holm-modified Bonferroni correction for multiple comparisons.
54
Figure 2- 4. Mean number of unique polymorphisms per genome (θU) for sexual vs. asexual
lineages of P. antipodarum.
Mean number of conservative (semi-transparent) and radical (solid) unique polymorphisms per
site for sexual (blue) vs. asexual (red) lineages of P. antipodarum. Error bars indicate 95% CIs
generated using 10,000 bootstrap replicates. Lower-case letters indicate statistical groupings
determined using the Holm-modified Bonferroni correction for multiple comparisons..
55
SUPPLEMENTARY MATERIAL
Table 2-S1. Summary of source populations of Potamopyrgus antipodarum.
Reproductive
Lake of Origin
Latitude/Longitude
Ploidy Sequencing Platform Reference
Mode
Alexandrina
-35.440603, 139.083438
Sexual
2x
ABI 3730
Neiman et al. 2010
Alexandrina
-35.440603, 139.083438
Sexual
2x
ABI 3730
Neiman et al. 2010
Alexandrina
-35.440603, 139.083438
Sexual
2x
Illumina
This study
Alexandrina
-35.440603, 139.083438
Asexual
3x
ABI 3730
Neiman et al. 2010
Alexandrina
-35.440603, 139.083438
Asexual
3x
ABI 3730
Neiman et al. 2010
Alexandrina
-35.440603, 139.083438
Asexual
3x
ABI 3730
Neiman et al. 2010
Brunner
-42.607385, 171.439636
Asexual
3x
ABI 3730
This study
Brunner
-42.607385, 171.439636
Asexual
4x
ABI 3730
This study
Denmark
56.001158, 9.207968
Asexual
3x
ABI 3730
Neiman et al. 2010
Grasmere
-43.061036, 171.774569
Asexual
3x
ABI 3730
This study
Grasmere
-43.061036, 171.774569
Asexual
4x
ABI 3730
This study
Gunn
-44.874524, 168.090031
Asexual
3x
ABI 3730
Neiman et al. 2010
Heron
-43.481012, 171.169173
Asexual
3x
ABI 3730
Neiman et al. 2010
Heron
-43.481012, 171.169173
Asexual
3x
ABI 3730
Neiman et al. 2010
Ianthe
-43.481012, 171.169173
Sexual
2x
ABI 3730
Neiman et al. 2010
Ianthe
-43.481012, 171.169173
Sexual
2x
Illumina
This study
Kaniere
-43.481012, 171.169173
Sexual
2x
ABI 3730
This study
Kaniere
-43.481012, 171.169173
Asexual
3x
Illumina
This study
Lady
-42.599558, 171.573173
Sexual
2x
ABI 3730
Neiman et al. 2010
Lake Superior
47.917620, -86.953400
Asexual
3x
ABI 3730
Neiman et al. 2010
McGregor
-43.936019, 170.470227
Asexual
3x
ABI 3730
Neiman et al. 2010
56
Table 2-S1 – continued
Poerua
-42.702716, 171.495638
Asexual
3x
Illumina
This study
Poerua
-42.702716, 171.495638
Asexual
4x
ABI 3730, Illumina*
This study
Rotoiti
-38.037084, 176.345628
Asexual
4x
ABI 3730
This study
Rotoroa
-41.851912, 172.642727
Sexual
2x
ABI 3730
This study
Tarawera
-38.186357, 176.429540
Asexual
3x
ABI 3730
Neiman et al. 2010
Waikaremoana
-38.771011, 177.109841
Asexual
3x
ABI 3730
Neiman et al. 2010
Waikaremoana
-38.771011, 177.109841
Asexual
3x
ABI 3730
Neiman et al. 2010
Waikaremoana
-38.771011, 177.109841
Asexual
3x
ABI 3730
Neiman et al. 2010
Waikaremoana
-38.771011, 177.109841
Asexual
3x
Illumina
This study
Wales
52.321471, -3.703571
Asexual
3x
ABI 3730
Neiman et al. 2010
*
– Same DNA extraction was sequenced on both platforms
57
Table 2-S2. Primers used to amplify and sequence whole mtDNA in P. antipodarum.
Fragment
Forward
Reverse
1
5’-GAGGTAGGAGACTGTAGT-3’
5’-GAGTCCTAAGCCCAATGCA-3’
2
5’-GCTAGTATGAATGGTTTGACG-3’
5’-GTTATGGCAGCAATAGTAATTG-3’
3
5’-TCAGCTTGTGGATCTGA-3’
5’-GCCTAATCAGTATGAGGAAG-3’
4
5’-GGAGTGAACGGAAATCA-3’
5’-CTCTTGAGTATGCTGAGTACA-3’
PCR
Sequencing
1.1
5’TATGAATATTCAGATTTTTTAAATA-3’ 5’-CTGCCACCTTTATTATAAAG-3’
1.2
5’-GGCTCATAGTTTACTTAACTT-3’
5’-CGTCAAACCATTCATACTAGC-3’
1.3
5’-GCGGTTAGACCACGAAG-3’
5’-AACTAATAGATGTTTCTATG-3’
1.4
5’-CATAGAAACATCTATTAGTT-3’
5’-CTAATCCCAGTTTCCCTC-3’
2.1
5’-GTCTCTCTCTAATTTTATAG-3’
5’-GCTGATAGATGAAAGTCTC-3’
2.2
5’-CTCTCCTTATTTTTTAGCC-3’
5’-ATATTTGCAGGAATTCAGTG-3’
2.3
5’-GCATTGGAAGCTAAAGAC-3’
5’-GTTAACAGCTTCTGTTCG-3’
2.4
5’-CAGCACACTTGAAACATTG-3’
5’-CTCATATCTTGCTGCAAC-3’
3.1
5’-CTTCATCAATTAGCGCTTTATTT-3’
5’-GTGAAAGAAATCTTAGCCTA-3’
3.2
5’-CTTTCTACCTTAAGCCAGCTAG-3’
5’-GTACTAAGCCCCTAAAGGCAA-3’
3.3
5’-CAGCACAGCCTTTAACTAAG-3’
5’CTAAATGAAAGGGGTTACG-3’
3.4
5’-CCGCTAAATCCATTTGAAG-3’
5’-GTCATCCCTGTAGCTAG-3’
4.1
5’-CTATTGTAGTTATATTGTTGG-3’
5’-GAGATATTACAAGCGGTG-3’
4.2
5’-CAATTTCTTCCTCATACTGATT-3’
5’-CCTTAACTCCTAATCTTGGTAC-3’
4.3
5’-TTAGGGTGGATGCTATTTGC-3'
5’-GATACAAGAGCCTCTCATAC-3’
4.4
5’-GATTTAGCTATTTTTTCATTAC-3’
5’-CTGTTGTAATAAAGTTTACTG-3’
58
Table 2-S3. Number of sites for different mutational types across seven amino acid
classification schemes in the invertebrate mitochondrial genetic code.
1
2
3
4
5
6
7
Amino
Codon
S
NS
R
C
R
C
R
C
R
C
R
C
R
C
R
C
Acid
TTT
F
0.33 2.67 0.00 2.67 1.00 1.67 2.33 0.33 2.33 0.33 2.33 0.33 1.00 1.67 1.00 1.67
TTC
F
0.33 2.67 0.00 2.67 1.00 1.67 2.33 0.33 2.33 0.33 2.33 0.33 1.00 1.67 1.00 1.67
TTA
L
0.67 2.33 0.33 2.00 0.67 1.67 1.67 0.67 1.67 0.67 1.33 1.00 0.67 1.67 0.67 1.67
TTG
L
0.67 2.33 0.33 2.00 0.67 1.67 1.67 0.67 1.67 0.67 1.33 1.00 0.67 1.67 0.67 1.67
TTG
M†
0.67 2.33 0.33 2.00 0.67 1.67 1.67 0.67 1.67 0.67 1.33 1.00 0.67 1.67 0.67 1.67
TCT
S
1.00 2.00 0.00 2.00 1.00 1.00 1.00 1.00 0.67 1.33 0.67 1.33 1.00 1.00 1.00 1.00
TCC
S
1.00 2.00 0.00 2.00 1.00 1.00 1.00 1.00 0.67 1.33 0.67 1.33 1.00 1.00 1.00 1.00
TCA
S
1.00 2.00 0.33 1.67 1.67 0.33 1.00 1.00 1.00 1.00 0.67 1.33 1.67 0.33 1.67 0.33
TCG
S
1.00 2.00 0.33 1.67 1.67 0.33 1.00 1.00 1.00 1.00 0.67 1.33 1.67 0.33 1.67 0.33
TAT
Y
0.33 2.67 1.33 1.33 1.00 1.67 2.33 0.33 2.00 0.67 2.33 0.33 1.67 1.00 1.67 1.00
TAC
Y
0.33 2.67 1.33 1.33 1.00 1.67 2.33 0.33 2.00 0.67 2.33 0.33 1.67 1.00 1.67 1.00
TAA
*
0.33 2.67 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00
TAG
*
0.33 2.67 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00 2.67 0.00
TGT
C
0.33 2.67 0.33 2.33 1.00 1.67 2.67 0.00 1.67 1.00 1.67 1.00 1.67 1.00 1.33 1.33
TGC
C
0.33 2.67 0.33 2.33 1.00 1.67 2.67 0.00 1.67 1.00 1.67 1.00 1.67 1.00 1.33 1.33
TGA
W
0.33 2.67 0.67 2.00 2.33 0.33 2.67 0.00 2.33 0.33 2.67 0.00 2.00 0.67 2.33 0.33
TGG
W
0.33 2.67 0.67 2.00 2.33 0.33 2.67 0.00 2.33 0.33 2.67 0.00 2.00 0.67 2.33 0.33
59
Table 2-S3 – continued
CTT
L
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.33 0.67 1.00 1.00 0.67 1.33 0.67 1.33
CTC
L
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.33 0.67 1.00 1.00 0.67 1.33 0.67 1.33
CTA
L
1.33 1.67 0.33 1.33 0.67 1.00 1.00 0.67 1.00 0.67 0.33 1.33 0.67 1.00 0.67 1.00
CTG
L
1.33 1.67 0.33 1.33 0.67 1.00 1.00 0.67 1.00 0.67 0.33 1.33 0.67 1.00 0.67 1.00
CCT
P
1.00 2.00 0.67 1.33 1.33 0.67 1.00 1.00 1.00 1.00 0.67 1.33 1.33 0.67 1.33 0.67
CCC
P
1.00 2.00 0.67 1.33 1.33 0.67 1.00 1.00 1.00 1.00 0.67 1.33 1.33 0.67 1.33 0.67
CCA
P
1.00 2.00 0.33 1.67 1.33 0.67 1.00 1.00 1.00 1.00 0.33 1.67 1.33 0.67 1.33 0.67
CCG
P
1.00 2.00 0.33 1.67 1.33 0.67 1.00 1.00 1.00 1.00 0.33 1.67 1.33 0.67 1.33 0.67
CAT
H
0.33 2.67 2.33 0.33 0.67 2.00 2.33 0.33 1.33 1.33 2.33 0.33 2.33 0.33 2.33 0.33
CAC
H
0.33 2.67 2.33 0.33 0.67 2.00 2.33 0.33 1.33 1.33 2.33 0.33 2.33 0.33 2.33 0.33
CAA
Q
0.33 2.67 2.00 0.67 1.00 1.67 2.33 0.33 1.33 1.33 2.00 0.67 2.67 0.00 2.67 0.00
CAG
Q
0.33 2.67 2.00 0.67 1.00 1.67 2.33 0.33 1.33 1.33 2.00 0.67 2.67 0.00 2.67 0.00
CGT
R
1.00 2.00 1.67 0.33 0.67 1.33 1.67 0.33 1.67 0.33 1.67 0.33 1.67 0.33 1.67 0.33
CGC
R
0.67 2.33 2.00 0.33 0.67 1.33 2.00 0.33 1.67 0.33 1.67 0.33 2.00 0.33 1.67 0.33
CGA
R
1.00 2.00 2.00 0.00 1.00 1.00 2.00 0.00 1.33 0.67 2.00 0.00 2.00 0.00 2.00 0.00
CGG
R
1.00 2.00 2.00 0.00 1.00 1.00 2.00 0.00 1.33 0.67 2.00 0.00 2.00 0.00 2.00 0.00
ATT
I
0.33 2.67 0.00 2.67 1.00 1.67 1.33 1.33 1.33 1.33 0.33 2.33 1.00 1.67 1.00 1.67
ATT
M†
1.00 2.00 0.00 2.00 1.00 1.00 1.33 0.67 1.33 0.67 0.33 1.67 1.00 1.00 1.00 1.00
ATC
I
0.33 2.67 0.00 2.67 1.00 1.67 1.33 1.33 1.33 1.33 0.33 2.33 1.00 1.67 1.00 1.67
ATC
M†
1.00 2.00 0.00 2.00 1.00 1.00 1.33 0.67 1.33 0.67 0.33 1.67 1.00 1.00 1.00 1.00
60
Table 2-S3 – continued
ATA
M
0.33 2.67 0.33 2.33 1.00 1.67 1.00 1.67 1.00 1.67 0.33 2.33 1.00 1.67 1.00 1.67
ATA
M†
1.00 2.00 0.33 1.67 1.00 1.00 1.00 1.00 1.00 1.00 0.33 1.67 1.00 1.00 1.00 1.00
ATG
M
0.33 2.67 0.33 2.33 1.00 1.67 1.00 1.67 1.00 1.67 0.33 2.33 1.00 1.67 1.00 1.67
ATG
M†
1.67 1.33 0.33 1.00 1.00 0.33 1.00 0.33 1.00 0.33 0.33 1.00 1.00 0.33 1.00 0.33
ACT
T
1.00 2.00 0.00 2.00 1.00 1.00 0.67 1.33 0.33 1.67 0.00 2.00 1.00 1.00 1.00 1.00
ACC
T
1.00 2.00 0.00 2.00 1.00 1.00 0.67 1.33 0.33 1.67 0.00 2.00 1.00 1.00 1.00 1.00
ACA
T
1.00 2.00 0.33 1.67 1.00 1.00 0.67 1.33 0.67 1.33 0.33 1.67 1.33 0.67 1.33 0.67
ACG
T
1.00 2.00 0.33 1.67 1.00 1.00 0.67 1.33 0.67 1.33 0.33 1.67 1.33 0.67 1.33 0.67
AAT
N
0.33 2.67 1.33 1.33 0.33 2.33 2.33 0.33 2.00 0.67 1.67 1.00 1.67 1.00 1.67 1.00
AAC
N
0.33 2.67 1.33 1.33 0.33 2.33 2.33 0.33 2.00 0.67 1.67 1.00 1.67 1.00 1.67 1.00
AAA
K
0.33 2.67 2.67 0.00 0.67 2.00 2.67 0.00 2.33 1.30 2.67 0.00 2.67 0.00 2.67 0.00
AAG
K
0.33 2.67 2.67 0.00 0.67 2.00 2.67 0.00 2.33 1.30 2.67 0.00 2.67 0.00 2.67 0.00
AGT
S
1.00 2.00 0.33 1.67 0.33 1.67 1.33 0.67 0.67 1.33 0.33 1.67 1.00 1.00 0.67 1.33
AGC
S
1.00 2.00 0.33 1.67 0.33 1.67 1.33 0.67 0.67 1.33 0.33 1.67 1.00 1.00 0.67 1.33
AGA
S
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.33 0.67 1.00 1.00 1.67 0.33 1.33 0.67
AGG
S
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.33 0.67 1.00 1.00 1.67 0.33 1.33 0.67
GTT
V
1.00 2.00 0.33 1.67 0.67 1.33 1.33 0.67 1.33 0.67 0.67 1.33 0.33 1.67 0.67 1.33
GTC
V
1.00 2.00 0.33 1.67 0.67 1.33 1.33 0.67 1.33 0.67 0.67 1.33 0.33 1.67 0.67 1.33
GTA
V
1.00 2.00 0.33 1.67 0.67 1.33 1.00 1.00 1.00 1.00 0.33 1.67 0.33 1.67 0.67 1.33
GTG
V
1.00 2.00 0.33 1.67 0.67 1.33 1.00 1.00 1.00 1.00 0.33 1.67 0.33 1.67 0.67 1.33
61
Table 2-S3 – continued
GTG
M†
0.67 2.33 0.33 2.00 0.67 1.67 1.00 1.33 1.00 1.33 0.33 2.00 0.33 2.00 0.67 1.67
GCT
A
1.00 2.00 0.33 1.67 1.33 0.67 0.67 1.33 0.67 1.33 0.33 1.67 1.00 1.00 1.33 0.67
GCC
A
1.00 2.00 0.33 1.67 1.33 0.67 0.67 1.33 0.67 1.33 0.33 1.67 1.00 1.00 1.33 0.67
GCA
A
1.00 2.00 0.33 1.67 1.33 0.67 0.67 1.33 0.67 1.33 0.33 1.67 1.00 1.00 1.33 0.67
GCG
A
1.00 2.00 0.33 1.67 1.33 0.67 0.67 1.33 0.67 1.33 0.33 1.67 1.00 1.00 1.33 0.67
GAT
D
0.33 2.67 2.00 0.67 0.67 2.00 1.67 1.00 2.00 0.67 2.00 0.67 2.00 0.67 2.00 0.67
GAC
D
0.33 2.67 2.00 0.67 0.67 2.00 1.67 1.00 2.00 0.67 2.00 0.67 2.00 0.67 2.00 0.67
GAA
E
0.33 2.67 2.00 0.67 1.00 1.67 1.67 1.00 2.00 0.67 2.00 0.67 2.00 0.67 2.00 0.67
GAG
E
0.33 2.67 2.00 0.67 1.00 1.67 1.67 1.00 2.00 0.67 2.00 0.67 2.00 0.67 2.00 0.67
GGT
G
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.00 1.00 0.67 1.33 1.33 0.67 1.33 0.67
GGC
G
1.00 2.00 0.67 1.33 0.67 1.33 1.33 0.67 1.00 1.00 0.67 1.33 1.33 0.67 1.33 0.67
GGA
G
1.00 2.00 0.67 1.33 1.00 1.00 1.33 0.67 1.33 0.67 1.00 1.00 1.00 1.00 1.67 0.33
GGG
G
1.00 2.00 0.67 1.33 1.00 1.00 1.33 0.67 1.33 0.67 1.00 1.00 1.00 1.00 1.67 0.33
Mean
(SD)
Total
0.75
2.25
0.82
1.43
0.99
1.26
1.54
0.71
1.36
0.91
1.15
1.10
1.37
0.88
1.41
0.84
(0.35) (0.35) (0.82) (0.71) (0.47) (0.54) (0.65) (0.45) (0.57) (0.40) (0.85) (0.66) (0.64) (0.53) (0.61) (0.52)
52
158
58
100
69
88
108
50
95
64
80
77
96
62
99
59
* -- Indicates a stop codon
† -- Indicates case in which a codon is used as a start codon. Codons used as start codons have
fewer nonsynonymous sites (mean = 2.00 +/- 0.41) than when they are used at any other time in
a given amino acid sequence (mean = 2.50 +/- 0.28)
62
CHAPTER 3: INEFFICIENT PURIFYING SELECTION AND
VARIATION IN FUNCTIONAL CONSTRAINT DRIVES
ACCELERATED BUT HETEROGENEOUS ACCUMULATION OF
HARMFUL MUTATIONS IN ASEXUAL LINEAGES OF A
FRESHWATER SNAIL
ABSTRACT
Mitochondrial genomes exhibit extensive across-gene evolutionary rate heterogeneity at both
synonymous and nonsynonymous sites. The pattern of this heterogeneity both varies across
lineages and appears to be heavily influenced by effective population size (Ne), suggesting that
variation in Ne might be a primary driver of evolution in mtDNA. Because sexual reproduction is
expected to increase Ne, powerful and direct tests of the hypothesis linking Ne and mtDNA
evolution can come from evaluating patterns of mtDNA evolution in otherwise similar sexual vs.
asexual lineages. Here, I use Potamopyrgus antipodarum, a New Zealand freshwater snail
characterized by coexisting sexual and asexual lineages, to investigate the role of reproductive
mode (and thereby Ne) in mitochondrial evolutionary rate heterogeneity. I found that variation in
both mutation rate and the intensity of purifying selection on oxidative phosphorylation
(OXPHOS) complexes govern evolutionary rate heterogeneity across P. antipodarum
mitochondrial genes, suggesting that purifying selection is not particularly effective in P.
antipodarum mitochondrial genomes. I also found that genes exhibiting signs of relatively
intense purifying selection contribute relatively little to mutation accumulation in asexual P.
antipodarum compared to genes experiencing a lower intensity of purifying selection. These
findings also indicate that sexual reproduction increases Ne of the mitochondrial genome in P.
antipodarum, allowing more efficient mutational clearance in sexual vs. asexual lineages.
Together, these results suggest that understanding the evolutionary consequences of genome-
63
wide deleterious mutation accumulation in asexual lineages requires accounting for the intensity
of local functional constraint.
64
INTRODUCTION
Animal mitochondrial genomes typically display wide across-gene heterogeneity in rates and
patterns of DNA sequence evolution (Xia 1998, Ballard 2000, Marshall et al. 2009). These
across-gene differences also tend to be lineage specific (Gissi et al. 2000), which is particularly
surprising in light of the high conservation of proteins encoded by animal mitochondrial
genomes (Blier et al. 2001), evidence for an important role for purifying selection in
mitochondrial genome evolution in a wide variety of taxa (Dowling et al. 2008), and the fact that
proper mitochondrial function (i.e., production of ATP) also appears to be a critical component
to eukaryotic health. Clear support for a link between mitochondrial genetic variation and
organismal performance comes from the observation that mutations in mitochondrial DNA cause
decreased function and fitness in taxa ranging from mammals (Barja and Herrero 2000) and
copepods (Ellison and Burton 2006) to yeast (Zeyl et al. 2005).
While positive selection has been proposed as a potential explanation for some of the
across-gene variation in mitochondrial evolutionary rate (e.g., Bazin et al. 2006), many studies
also point instead towards an important role for variation in functional constraint. Evidence for
the relevance of variation in functional constraint is particularly apparent in subunits of
OXPHOS complexes shared amongst eukaryotes and proteobacteria, the so-called “core subunits”
(Zhang and Broughton 2013). In a situation where variation in functional constraint is a major
determinant of patterns of molecular evolution in OXPHOS genes, across-gene evolutionary rate
and pattern heterogeneity would be a function of variable intensity of purifying selection, with
more intense selection on some oxidative phosphorylation (OXPHOS) complexes than others.
A non-mutually exclusive but distinct hypothesis for the observed pattern of acrossmitochondrial gene evolutionary rate heterogeneity is rooted in data suggesting that across-gene
65
differences in mutation rate play a major role in contributing to the observed pattern of variable
rates and patterns of evolution across mitochondrial genes. Under this hypothesis, mitochondrial
evolution is predominantly governed by mutation-drift equilibrium rather than differences in
selective intensity. The unique biology of mitochondria lends itself toward relatively strong
effects mediated by neutral evolutionary forces. In particular, the recurrent bottlenecking
experienced by mitochondrial populations during oogenesis as well as the haploid and (typically)
uniparental inheritance of the mitochondrial genome are thought to reduce the effective
population size of mitochondria (Lynch et al. 1993, Normark and Moran 2000, Haig 2016),
decreasing the efficacy of selection on mitochondrial genes and allowing a relatively large
fraction of slightly deleterious mutations to contribute to divergence (reviewed in Neiman and
Taylor 2009). Stochastic differences in evolutionary rate across genes are a potential
consequence of this phenomenon, manifesting as across-gene differences in mutation
accumulation.
The relatively small effective population size of mitochondrial genes relative to their
nuclear-encoded subunits (Lynch et al. 2003, Neiman and Taylor 2009) means that organismal
effective population size (Ne) is also likely to play a major role in determining the relative
contributions of selection and mutation-drift equilibrium to rate heterogeneity. A particularly
useful setting in which to study the effect of Ne on selective vs. stochastic effects on
mitochondrial rate heterogeneity is that provided by comparing sexual and asexual lineages. In
particular, asexual lineages are expected to experience reduced efficacy of purifying selection in
their mitochondrial genomes relative to sexual lineages because the mitochondrial genome is
transmitted in complete linkage disequilibrium with the nuclear genome in asexuals. The
implications of this linkage disequilibrium are that selective interference should decrease the
66
ability of asexual lineages to purge deleterious mutations and/or fix beneficial mutations in their
mitochondrial genomes (Hill and Robertson 1966). By this logic, if the absence of sex in the
nuclear genome decreases the efficacy of selection in the mitochondrial genome, mitochondrial
genomes sequestered in asexual lineages should accumulate deleterious mutations more rapidly
than their sexual counterparts. The intensity of purifying selection should dictate the extent of
accelerated deleterious mutation accumulation: in particular, genes under less stringent selective
regimes are expected to be characterized by more marked departures between the trajectories of
sexual vs. asexual evolution than genes that experience a higher intensity of purifying selection
(Charlesworth et al. 1997). Whether these predictions are realized has important implications for
the maintenance of sex and/or the seemingly inevitable extinction of asexual lineages: if the
intensity of selection is inversely related to the degree of asexual deleterious mutation
accumulation, the genes that contribute the most to harmful mutation accumulation in asexual
lineages might contribute relatively little to important components of organismal fitness.
Direct evaluation of the role of the efficacy of selection on evolutionary rate
heterogeneity in sexual vs. asexual lineages requires that to the extent possible, these not
otherwise experience differing intensities of selection and are otherwise similar. From this
perspective, Potamopyrgus antipodarum, a New Zealand freshwater snail, provides a powerful
model by which to compare evolutionary processes across reproductive modes because multiple
separately derived and otherwise similar asexual lineages commonly coexist with sexual P.
antipodarum in natural populations (Lively 1987). Because asexual P. antipodarum have been
derived on multiple separate occasions (Neiman and Lively 2004, Neiman et al. 2011,
Paczesniak et al. 2013), each distinct asexual lineage represents a replicated natural experiment
into the consequences of asexuality. Sexual lineages of P. antipodarum eliminate putatively
67
deleterious mutations from their mitochondrial genomes more rapidly than asexual lineages,
likely due to less efficient purifying selection operating on mitochondrial genomes sequestered in
asexual lineages (Neiman et al. 2010, Chapter 2). What remains untested is the extent to which
this asexual deleterious mutation accumulation proceeds similarly across the different genes
encoded in mitochondrial genome. Population genomic methods allow answering these questions
in a powerful way, and, particularly considering the relatively small size of the mitochondrial
genome, gene-by-gene analysis is tractable and will illuminate heterogeneous patterns, if they
exist. Additionally, gene expression offers a useful phenotype for investigating this question
because genes that are broadly and/or highly expressed tend to experience a higher intensity of
purifying selection and relatively slow molecular evolution relative to genes that exhibit tissuespecific expression and/or are expressed at relatively low levels (Duret and Mouchiroud 2000,
Akashi 2001, Adrion et al. 2016). The implications are that I can use comparisons of gene-level
molecular evolution with gene-level expression to address the extent to which relatively high
expression (and, thus, the presumption of relatively high functional constraint) translates into
gene-level variation in molecular evolution. I take these approaches here, using mitogenomic and
transcriptomic data from a diverse array of sexual and asexual P. antipodarum to test whether
some genes and some OXPHOS complexes contribute to asexual mutation accumulation more
than other genes and whether reproductive mode (as a proxy for Ne) influences rate
heterogeneity in the mitochondrial genome.
68
MATERIALS & METHODS
Evolutionary rate heterogeneity in the mitochondrial genome of P. antipodarum
I used 31 whole mitochondrial genome sequences from P. antipodarum and one whole
mitochondrial genome sequence from Potamopyrgus estuarinus (a closely related congener,
Haase 2008) to evaluate rates and patterns of selection across the 13 protein-coding genes in the
P. antipodarum mitochondrial genome. Sequence data were obtained from Genbank (Accession
Nos.: GQ996416 – GQ996433, GQ996415.1, Table 2-S1) and from Sharbrough et al. (in review,
Chapter 2). I used the MUSCLE package in MEGA v 5.2 (Kumar et al. 2008) to build individual
alignments for each gene and then manually inspected and corrected each alignment. To
visualize evolutionary rate heterogeneity in the mitochondrial genome, I used DNAsp v5 to
perform a sliding window analysis of mean Jukes-Cantor-corrected substitution rates
(nonsynonymous substitution/nucleotide, dN; synonymous substitution/nucleotide, dS) with a
window size of 100bp and a step size of 3 bp along the 11215 protein-coding nucleotides in the P.
antipodarum mitochondrial genome (Neiman et al. 2010). Substitution rate provides a useful
metric for studying across-gene evolutionary rate variation via the emphasis on interspecific
divergence, which is affected to a much lesser extent by population demography than are
intraspecific polymorphism-based approaches (Tajima 1989). Because synonymous sites are
thought to have little to no effect on phenotype and thus to evolve under neutral (or close to
neutral) expectations, synonymous changes are expected to accumulate at a rate proportional to
their occurrence (i.e., the mutation rate). These expectations are in stark contrast to
nonsynonymous sites, which by definition affect amino acid sequence and often (but not always)
experience relatively strong purifying selection. By this logic, dN/dS provides an estimate of the
direction and intensity of selection on a given sequence such that dN/dS > 1 should reflect
69
positive selection, dN/dS < 1 is expected under purifying selection, and dN/dS ~ 1 indicates neutral
evolution at mutation-drift equilibrium (Li et al. 1985). I estimated 95% confidence intervals for
each window with 10,000 bootstrap replicates via a custom Python script (all scripts are
available at https://github.com/potamomics-scripts).
This sliding window analysis revealed substantial variation in dN/dS across the P.
antipodarum mitochondrial genome (Figure 3-1). I then used a series of analyses to determine
whether and to what extent mutation rate, differences in selection intensity, and/or differences in
selection efficacy were contributing to the observed variation. I began by establishing the extent
to which polymorphism leads to divergence for both synonymous and nonsynonymous sites by
using Spearman's rank correlations (implemented within R v.3.2.4, R Core Team 2016) to
compare the number of synonymous polymorphisms and substitutions and nonsynonymous
polymorphisms and substitutions, respectively, for each of the 13 genes. By establishing the
extent to which mutations (estimated by the number of polymorphisms per gene) contribute to
divergence between species (estimated by the number of substitutions per gene), this analysis
enabled me to evaluate whether mutation rate is the primary driver of evolutionary rate
heterogeneity in the protein-coding genes in the P. antipodarum mitochondrial genome. I then
used a Student’s t test to compare the slopes of the best-fit regression line for the relationship
between polymorphism and substitution for each of the two types of mutational change to
determine whether the relationship between polymorphisms and substitutions differed for
synonymous vs. nonsynonymous changes.
If protein sequence divergence is influenced by the underlying mutation rate, then the
number of nonsynonymous substitutions per site (dN) should be positively correlated with the
number of synonymous substitutions per site (dS). I addressed whether this prediction was met in
70
each of the 13 protein-coding genes in the P. antipodarum mitochondrial genome by using v 4.7
of the program Phylogenetic Analysis by Maximum Likelihood (PAML, Yang 2007) and loglikelihood ratio tests (LRT; chi-squared distribution) to compare the fit of 3 distinct models of
substitution for each gene, assuming the well-supported phylogenetic tree topology shown in
Figure 2-1: 1) M0 model (null hypothesis; – a single dN/dS value for a given gene for all branches
in the P. antipodarum mitochondrial phylogeny, 2) M1a model (alternative hypothesis 1; each
branch is allowed to vary in dN/dS value for a given gene), and 3) M2a model (alternative
hypothesis 2; sexual branches have a single dN/dS value and asexual branches have a single dN/dS
value for a given gene). If the M0 model was the best fit for the data or if neither of the other two
more complex models were a significantly better fit to the data than the M0 model, then a single,
species-wide dN/dS would be assigned to that gene. If M2a was the best-fit model for the data or
if the M1a Model was not a significantly better fit than the more complex M2 model, then two
estimates of dN/dS would be assigned to that gene; one for sexual lineages and one for asexual
lineages. If the M1a model was a significantly better fit to the data than either of the other two
models, then branch-specific estimates of dN/dS would be assigned to that gene. These PAML
analyses revealed that twelve genes were best explained by model M1a (branch-specific dN/dS)
and one gene, atp8, was equally well explained by M0 (species-wide dN/dS) as all alternative
models. No genes were best explained by model M2a (different sexual and asexual dN/dS, Table
3-2). Next, I used a Spearman's rank correlation to compare the PAML estimates of branchspecific dS and branch-specific dN for the 12 genes best explained by M2a and the species-wide
estimate from atp8 to determine if dN and dS were positively associated, as expected if mutation
rate is at least in part driving amino acid sequence evolution.
71
I then used DNAsp v5 and custom-built Python scripts to calculate mean species-wide
dN/dS for each of the 13 genes. I used these estimates to compare the relative strength and
efficacy of selection experienced by each gene by using a Kruskal-Wallis test to compare the
synonymous-corrected substitution rate across genes and complexes. Next, I estimated
synonymous-site-corrected nonsynonymous polymorphism per site using the Watterson
estimator, θ (Watterson 1975), for each gene and compared θA/θS across genes and complexes to
assess selection at the intraspecific level with Kruskal-Wallis tests. I then used a combination of
polymorphism and divergence data to perform McDonald-Kreitman tests of selection (McDonald
and Kreitman 1991; MK tests) and Fisher’s exact tests to generate additional insight into
presence and mode of selection acting on each mitochondrial gene (results displayed in Table 31).
Because mitochondrially encoded proteins are members of protein complexes and
therefore physically interact with other proteins from both the mitochondrial and nuclear genome,
I hypothesized that complex identity plays an important role in determining mitochondrial gene
evolutionary rate. Under this hypothesis, complexes under intense selection would be expected to
be composed of proteins exhibiting low substitution rates relative to complexes experiencing
relatively weak purifying selection. Natural selection on mitochondrial genes is thought to
operate to maintain OXPHOS complex function (Blier et al. 2001, Rand 2001), lending
additional support to the possibility that the intensity of selection will vary according to complex
identity. To test this hypothesis, I concatenated the sequence from the mitochondrially encoded
genes for each of the 4 OXPHOS complexes containing at least one of these genes (complex II is
encoded entirely by the nuclear genome). I then compared mean mitochondrial dN/dS and θA/θS
across each of the 4 OXPHOS complexes using Kruskal-Wallis tests. I also used McDonald-
72
Kreitman and Fisher's Exact test analyses to compare complex-wide ratios of synonymous and
nonsynonymous polymorphism to divergence across the four complexes, and I used a KruskalWallis test and Mann-Whitney U tests (applying the Holm procedure for Bonferroni correction
of multiple comparisons) to compare estimates of mean dN/dS (performed in PAML, see above)
of genes across complexes. This last analysis allowed me to address if substitution rate was
affected by complex identity as well as by complex-specific mutation rate.
These analyses indeed suggested that substitution rate was indeed affected by complex
identity (see Results). Because gene expression level is thought to influence substitution rate via
the expectation that higher expression should translate into more intense purifying selection
(Duret and Mourchiroud 2000), I next addressed whether at least some of the effect of complex
identity might be linked to similar within-complex expression levels. I evaluated this possibility
by determining whether complex identity influenced gene expression of nuclear-encoded
mitochondrial genes (nu-mts) across three different RNA-Seq datasets and their corresponding
transcriptomes obtained from the same inbred sexual lineage (Neiman et al. unpublished data). I
confined my analyses to nu-mts because the RNA-Seq data used here were obtained using polyA
enrichment, which does not adequately capture mitochondrial gene expression (Ozsolak and
Milos 2011). To identify nu-mts, I first used the KEGG database to identify mollusk homologs
(from Crassostrea gigas, Octopus bimaculoides, or Lottia gigantia) of nuclear-encoded
OXPHOS genes (http://www.genome.jp/dbget-bin/www_bget?map00190). I then used blastp to
identify nuclear-encoded mitochondrial homologs in the California sea hare (Aplysia californica)
non-redundant protein database and produced a custom Aplysia nu-mt protein database. I used
blastx to identify transcripts from each of the three RNAseq datasets that hit the Aplysia nu-mt
database with an e-value < 1 x 10-10 and a bit score > 100. Transcripts from each dataset passing
73
these filters were collected into three distinct P. antipodarum nu-mt transcriptomes. I then
mapped RNAseq reads from each RNAseq dataset to the corresponding nu-mt transcriptome
using Tophat v.2.0.13 (Trapnell et al. 2012), assuming the following parameters: library-type: frsecondstrand, read-mismatches: 3, read-edit-dist: 3, segment-mismatches: 3. Once reads were
mapped to nu-mt transcriptomes, I used Cufflinks v.2.2.1 (Trapnell et al. 2012) to calculate
Fragments Per Kilobase Mapped (FPKM) for each transcript assuming the following parameters:
multi-read-correct: True; upper-quartile-norm: True; no-effective-length-correction: True. I then
used Spearman’s rank correlation analyses to determine whether mean FPKM per transcript was
associated with the mean species-wide dN/dS in the corresponding mitochondrially encoded
subunits, predicting that complexes experiencing a higher intensity of purifying selection would
tend to show higher expression.
Effect of rate heterogeneity on mutation accumulation in the absence of sex
The efficacy of selection is the product of both the intensity of selection and Ne (reviewed in
Charlesworth 2009), with higher selection intensity and higher Ne both expected to translate into
more effective selection. The important roles of both the intensity of selection and Ne in
determining selection efficacy means that selection intensity must be taken into account when
evaluating the extent to which reduced Ne is contributing to the accelerated accumulation of
deleterious mutations in asexual lineages. By this logic, I can use the residuals of the relationship
between the degree of mutation accumulation in asexual lineages relative to sexual lineages at
the gene level (i.e., the extent to which dN/dS is elevated in asexuals relative to sexuals for genes
under purifying selection) and gene-specific selective intensity to establish the extent to which
mutation accumulation in asexuals is caused by reduced Ne; higher residuals represent relatively
74
high mutation accumulation for a given selection intensity. I addressed this possibility by using
Spearman's rank correlation analyses to determine whether the intensity of selection experienced
by each gene (estimated here using species-wide dN/dS; higher dN/dS reflects lower selection
intensity) was associated with the degree of difference between mean asexual and mean sexual
branch-specific estimates of dN/dS for the 12 genes best fit by M1a as well as for the 10 genes in
which mean branch-specific dN/dS was larger in asexual lineages than sexual lineages; the latter
analysis represents the genes that demonstrate the signature of accelerated mutation
accumulation in asexual lineages. I then calculated the best-fit linear regression line for the
difference in mean asexual and mean sexual branch-specific dN/dS and species-wide dN/dS as well
as the linear regression residuals for these data. Finally, I used a Kruskal-Wallis test to compare
the residuals across genes and complexes to test whether mutation accumulation in asexuals due
to reduced Ne affected some genes and/or complexes more than others.
75
RESULTS
Evolutionary rate heterogeneity in mitochondrial genome of P. antipodarum
The sliding window analysis indicated extensive heterogeneity in dN/dS across the 13 proteincoding genes of the mitochondrial genome. I also found that the numbers of synonymous
polymorphisms and nonsynonymous polymorphisms per gene, respectively, were significantly
and positively correlated with the number of synonymous and nonsynonymous substitutions per
gene, respectively (Synonymous: Spearman’s rho = 0.96, p < 0.0001; nonsynonymous:
Spearman’s rho = 0.58, p = 0.0361, Figure 3-2). Next, I used PAML to compare branch-specific
rates of synonymous (dS) and nonsynonymous (dN) substitution rates for the 12 genes best
explained by model M1a and the species-wide estimate of dS and dN for atp8. This analysis
revealed that dN is significantly and positively correlated with dS (Spearman’s rho = 0.54, p <
0.0001, Figure 3-3), indicating that across-gene variation in mutation rate (reflected by dS) likely
contributes to the observed variation in the rate of nonsynonymous mutation accumulation across
mitochondrial genes. These results indicate that the substitution rate is at least in part determined
by the mutation rate, suggesting that mutation-drift equilibrium plays a major role in shaping
rates and patterns of molecular evolution in P. antipodarum mitochondrial genomes, even for
nonsynonymous changes.
Although nonsynonymous variation is significantly affected by local mutation rate, the
slope of this relationship is significantly smaller for nonsynonymous changes than it is for
synonymous changes (Student's t = 8.154, p < 0.0001), suggesting that purifying selection is also
operating on nonsynonymous sites. Accordingly, I next used MK tests of selection on each gene
individually and on complexes collectively to test whether differential intensity of purifying
selection is likely to contribute to across-gene variation at evolutionary rate at nonsynonymous
76
sites in P. antipodarum mitochondrial genomes. This analysis revealed that seven out of 13
protein-coding genes exhibit significant signatures of purifying selection (i.e., Neutrality Index
(NI) > 1, see Table 3-1). I also found that all four complexes exhibited significant levels of
purifying selection (p < 0.0017, NI > 1 for all complexes). Together, these results implicate a
broad role for both mutation-drift equilibrium and purifying selection in the evolution of the
protein-coding region of the mitochondrial genome in P. antipodarum that is evident at both the
intraspecific (polymorphism) and the interspecific (substitution) level.
I next used Kruskal-Wallis tests to compare mean dN/dS and mean θA/θS across genes and
OXPHOS complexes and found that genes in different OXPHOS complexes exhibit significantly
different mean branch-specific dN/dS (Kruskal-Wallis χ2 = 82.115, p = 2.20 x 10-16, Figure 3-4 a).
Pairwise comparisons revealed that complex IV has a significantly lower substitution rate than
the other mitochondrially encoded complexes (p = 1.08 x 10-7 for all comparisons). I detected a
similar pattern at the intraspecific level, with θA/θS differing significantly across complexes
(Kruskal-Wallis χ2 = 9.2732, p = 0.0259, Figure 3-4 b) and with complex IV exhibiting a
significantly lower level of synonymous-corrected polymorphism than the other complexes
(Mann-Whitney U = 0.00, p = 0.00699). Together these results indicate that mitochondrial gene
evolutionary rate heterogeneity is influenced by complex membership at both the intraspecific
and the interspecific level, and that complex IV appears to have the highest degree of functional
constraint relative to complexes I, III, and V.
Two testable hypotheses arise from the observation that complex identity plays a role in
determining evolutionary rate: 1) genes from the same complex should be expressed at similar
levels to maintain biochemical stoichiometry (Akashi 2001), and 2) expression level should be
inversely related to evolutionary rate (e.g., broadly expressed genes are expected to evolve more
77
slowly than tissue-specific genes, Duret and Moouchiroud 2000). To address whether these
hypotheses were supported by my data, I estimated the number of Fragments Per Kilobase
Mapped (FPKM) for nuclear-encoded mitochondrial transcripts in three distinct P. antipodarum
transcriptomes (see Methods). I found that complex identity significantly affected FPKM
(Kruskal-Wallis χ2 = 12.33, p = 0.015, Figure 3-5 a), and that FPKM of nuclear-encoded
mitochondrial genes was significantly and positively correlated with mean complex dN/dS of
mitochondrial genes (Spearman’s rho = 0.21, p = 0.012, Figure 3-5 b).
Effect of rate heterogeneity on mutation accumulation in the absence of sex
To evaluate whether rate heterogeneity due to differences in functional constraint of genes
influences these patterns of differential mutation accumulation in asexual vs. sexual lineages of P.
antipodarum, I tested whether the PAML estimate of species-wide dN/dS for each gene was
related to the degree of difference in asexual vs. sexual branch-specific estimates of dN/dS. For
the 12 genes in which branch-specific values of dN/dS were estimated, I found that the intensity
of purifying selection (estimated via dN/dS) is not significantly correlated with the degree of
difference between sexual and asexual branch-specific estimates of dN/dS (Spearman’s rho =
0.404, p = 0.27, Figure 3-6 a). The outcome of this analysis was different when I included only
the 10 genes for which mean asexual dN/dS exceeds mean sexual dN/dS were included; here, dN/dS
and differential mutation accumulation in sexual vs. asexual lineages are significantly and
positively correlated (Spearman’s rho = 0.65, p = 0.0425, Figure 3-6 b). This result indicates that
for genes in which selective intensity is high (i.e., for which dN/dS is relatively low), asexuals
accumulate fewer deleterious changes relative to sexuals than in genes in which selective
intensity is relatively low.
78
I next addressed whether this signature of mutation accumulation in asexuals varied as a
result of differential reduction of Ne across genes and complexes by calculating the residuals of
the best-fit line for the relationship between selective intensity and the degree of mutation
accumulation in asexuals using the 10 gene dataset. I did not detect any differences across
mitochondrial genes or complexes in the extent of mutation accumulation due to reduced Ne
(Kruskal Wallis p > 0.05 for both tests, Figure 3-7), suggesting that the reduced Ne in asexuals
relative to sexuals affects different mitochondrial genes to a similar extent or that the effect of Ne
is stochastic across genes.
79
DISCUSSION
Rate heterogeneity in mitochondrial genome of P. antipodarum
I detected substantial across-gene heterogeneity in both the rate of accumulation of synonymous
changes as well as nonsynonymous (and likely harmful) changes in the mitochondrial genome of
P. antipodarum. This pattern was evident at both intraspecific (polymorphism) and interspecific
(substitution) time scales. These analyses also revealed that nonsynonymous polymorphism and
divergence are significantly and positively correlated with synonymous polymorphism and
divergence, suggesting that at least part of the across-gene variation that I detected is due to
variation in mutation rate. This result presents a stark contrast to the pattern of molecular
evolution found in species with known large effective population sizes (e.g., Drosophila
melanogaster) in which substitution rate at nonsynonymous sites is extremely low relative to
synonymous sites, likely due to relatively stringent and effective purifying selection (Montooth
et al. 2009).
Our results also suggest that variation in the intensity and efficacy of purifying selection
is a major contributor to patterns of molecular evolution in the P. antipodarum mitochondrial
genome. In particular, synonymous substitution rate-corrected estimates of nonsynonymous
divergence clearly indicate that OXPHOS complex identity and functional constraint plays a
major role in protein evolution. This observation is new for mollusks, and is similar to results
reported from mammals (Gissi et al. 2001, Nabholz et al. 2013, Zhang and Broughton 2013),
insects (Montooth et al. 2009, Nabholz et al. 2013), fish (Zhang and Broughton 2013), birds
(Nabholz et al. 2012), and nematodes (Nabholz et al. 2012). Complex identity also appears to
play a role in gene expression: I found significant across-complex variation in the expression
level of the nuclear-encoded OXPHOS components. Gene expression level of nuclear-encoded
80
mitochondrial genes was positively correlated with mitochondrial dN/dS, suggesting that genes
associated with OXPHOS complexes under relatively low functional constraint are expressed at
higher levels than genes associated with complexes under higher functional constraint. The
across-complex variation in expression levels is somewhat unexpected in light of the
expectations that highly expressed genes are usually thought to experience relatively effective
purifying selection (Duret and Mouchiroud 2001) and because OXPHOS complexes must
function in sequence to produce ATP, which would seem to predict similar levels of gene
expression across complexes. Differential expression of OXPHOS complexes could indicate that
electron flow and ATP production via oxidative phosphorylation can be regulated at the protein
complex level, potentially allowing for dynamic tuning of ATP production to environmental
conditions and/or, given the relatively large number of nonsynonymous polymorphisms present
in this species, a mechanism to compensate for mutational load. Such a mechanism has been
proposed to explain maintenance of mitochondrial function in spite of mtDNA damage in
mammalian aging (Das 2003), though whether P. antipodarum mitochondrial genomes with a
relatively high mutation load suffer negative consequences in terms of mitochondrial function
has yet to be evaluated.
More broadly, the combination of the influence of mutation rate and OXPHOS complex
identity on nonsynonymous variation suggests that while the mitochondrial genome is generally
undergoing purifying selection in P. antipodarum, that selection is not particularly effective,
even in sexual lineages. Altogether, this result suggests that P. antipodarum mitochondrial
genomes might experience small Ne relative to well-characterized model organisms like
Drosophila (Montooth et al. 2009) and nematodes (Nabholz et al. 2012).
81
Effect of rate heterogeneity on mutation accumulation in the absence of sex
A major class of hypotheses for the maintenance of sex is based on a scenario whereby the
decreased efficacy of selection in asexual lineages relative to sexual counterparts translates into
an increased rate of harmful mutation accumulation (Muller 1964, Hill and Robertson 1966,
Birky and Walsh 1988). Asexual P. antipodarum have already been shown to accumulate likely
harmful mutations in their mitochondrial genomes more rapidly than their sexual counterparts
(Neiman et al. 2010, Chapter 2). Here, I gained additional insight into the potential for harmful
phenotypic effects of these mutations by determining whether across-gene variation in DNA
sequence evolution in the protein-coding region of the P. antipodarum mitochondrial genome
affects the differential accumulation of harmful mutations in asexual vs. sexual lineages. Because
sex should facilitate the removal of deleterious mutations, even in the mitochondrial genome
(Normark and Moran 2000, Neiman and Taylor 2009), I expected more intense purifying
selection (estimated here using dN/dS) to reduce the cost of selective interference in asexual
lineages by increasing the proportion of harmful mutations that are filtered out by purifying
selection (Charlesworth et al. 1997). By this logic, because genes with higher functional
constraint should experience more intense purifying selection than genes under lower constraint,
I predicted that the proportion of mutations with negative fitness effects should be higher in
genes under relatively mild purifying selection. This logic in turn suggests that the asexuals
should show the most marked increase in harmful mutation accumulation for these relatively
unconstrained genes. This prediction was met by the comparison of the difference in sexual and
asexual dN/dS to my measure of the intensity of purifying selection realized by each gene, with
this sexual-asexual difference increasing with species-wide dN/dS. This result is novel and
important in suggesting that the types of genes that will tend to suffer the most marked mutation
82
accumulation in situations of reduced efficacy of purifying selection (e.g., in asexual lineages)
might tend to be genes that are not under particularly stringent purifying selection, with the
implication that this type of mutation accumulation might pose less of a threat to asexual lineage
success than generally presumed.
While this result would seem to contradict the results from Chapter 2, in which I
demonstrated that radical mutations persist longer in asexual than in sexual lineages, I suggest
that this pattern can be explained by a situation where these radical mutations are behaving
effectively neutral in asexuals but not sexuals as a consequence of reduced Ne in the former. To
expand, while the intensity of selection acting on these mutations does not differ across
reproductive modes, the lower Ne in asexuals reduces the efficacy of selection against radical
changes. The observation that conservative changes accumulate at a similar rate in sexuals and
asexuals suggests that the intensity of selection against these mutations is so low that they are
behaving effectively neutral in both sexuals and asexuals. One prediction stemming from this
hypothesis is that the nuclear genomes of sexual lineages should exhibit signs of more effective
clearance of conservative changes in addition to more rapid clearance of radical changes than the
nuclear genomes of asexuals because the difference in Ne across reproductive modes should be
larger in the nuclear genome than in the mitochondrial genome (Normark and Moran 2000).
Elucidating the consequences of reduced Ne in asexuals represents a critical next step in
understanding the evolution and maintenance of sex. Here, I show that the extent of mutation
accumulation in asexual lineages that is due to the reduction in Ne following transition to
asexuality does not vary across genes of the mitochondrial genome or among OXPHOS
complexes. Although the sample size is small (10 genes), this relationship (or lack thereof) is not
unexpected, as the reduction in Ne is expected to have a genome-wide effect on mutation
83
accumulation (Charlesworth et al. 1993, Neiman and Taylor 2009). This study is to my
knowledge the first to formally account for the local intensity of selection in testing for mutation
accumulation in asexuals. Given the substantially higher gene content and effects of sex in the
nuclear genome, exploring the relationship between local selective intensity, Ne, and the efficacy
of selection in the nuclear genome of P. antipodarum will provide an essential investigation into
the evolutionary maintenance of sex and the seemingly inevitable extinction of asexual lineages.
84
ACKNOWLEDGEMENTS
I thank Cindy Toll and Gery Hehman for their assistance with DNA sequencing. I also thank
Jeffrey L. Boore for assistance assembling mitochondrial genomes from Illumina sequence data.
I would like to thank Praakruti Cherukuri, Meagan Luse, and Michelle Zhang for assistance with
gene–by-gene information, and Emma Greimann and Rena Lin for building the P. antipodarum
nu-mt transcriptome. The National Science Foundation (NSF: MCB – 1122176; DEB –
1310825) and the Iowa Academy of Sciences (ISF #13-10) funded this research.
85
TABLES
Table 3-1. Rates and patterns of natural selection across the 13 proteincoding genes and OXPHOS complexes encoded by the mitochondrial
genome in P. antipodarum.
Gene
bp
dN/dS
θA/θS
NI
MK test FET p value
cox1
1536
0.00618 0.0236
3.346
0.159
cox2
687
0.00564 0.0248
3.600
0.337
atp8
159
0
0.134
∞
1.000
atp6
696
0.0113
0.154
11.500
0.0004
nd1
942
0
0.0763
∞
0.0006
nd6
507
0.0166
0.0778
3.889
0.151
cytB
1140
0.00301
0.143 38.0893
nd4l
297
0
0.112
∞
0.0437
nd4
1377
0.0175
0.0802
3.659
0.003
nd5
1719
0.0115
0.107
7.289
0.0001
cox3
779
0.00574 0.0623
9.409
0.0176
nd3
354
0
0.136
∞
0.0056
nd2
1054
0.0198
0.131
5.455
0.0004
Complex
bp
dN/dS
θA/θS
NI
MK test FET p value
I
6237
0.0907
0.123
6.413
< 0.0001
III
1140
0.103
0.145
38.089
< 0.0001
IV
3002
0.0304
0.0353
5.016
0.0017
V
855
0.0931
0.157
11.750
0.0003
< 0.0001
∞ – Neutrality index is undefined because no nonsynonymous substitutions appear
in this gene
86
Table 3-2. Log likelihood of PAML models of molecular evolution in each
of the 13 protein-coding genes of the mitochondrial genome.
Gene
M0
M1a
M2a
cox1
-3086.712
-3065.510*
-3083.782
cox2
-1300.031
-1285.467*
-1297.527
atp8
-222.035 *
-219.771
-221.075
atp6
-1412.946
-1380.626*
-1397.942
nd1
-2011.322
-1971.862*
-1987.721
nd6
-1009.203
-995.664*
-1009.203
cytB
-2483.568
-2424.952*
-2451.319
nd4l
-572.562
-561.956*
-569.267
nd4
-3167.522
-3125.200*
-3154.758
nd5
-3724.627
-3657.331*
-3690.886
cox3
-1677.282
-1643.888*
-1673.399
nd3
-781.306
-761.024*
-776.209
nd2
-2343.008
-2303.043*
-2326.126
* – Model was a significantly better fit to data according to LRT than all other models with
p < 0.05, using the Holm procedure for Bonferroni correction for multiple comparisons
87
REFERENCES
Adrion, J. R., White, P. S., Montooth, K. L. (2016). The roles of compensatory evolution and
constraint in aminoacyl tRNA synthetase evolution. Molecular Biology and Evolution, 33,
152-161.
Akashi, H. (2001). Gene expression and molecular evolution. Current Opinion in Genetics &
Development, 11, 660-666.
Ballard, J. W. O. (2000). Comparative genomics of mitochondrial DNA in members of the
Drosophila melanogaster subgroup. Journal of Molecular Evolution, 51, 48-63.
Barja, G., & Herrero, A. (2000). Oxidative damage to mitochondrial DNA is inversely related to
maximum life span in the heart and brain of mammals. Faseb Journal, 14, 312-318.
Bazin, E., Glemin, S., & Galtier, N. (2006). Population size does not influence mitochondrial
genetic diversity in animals. Science, 312, 570-572.
Birky, C. W., Jr., & Walsh, J. B. (1988). Effects of linkage on rates of molecular evolution.
Proceedings of the National Academy of Sciences USA, 85, 6414-6418.
Blier, P. U., Dufresne, F., & Burton, R. S. (2001). Natural selection and the evolution of
mtDNA-encoded peptides: evidence for intergenomic co-adaptation. Trends in Genetics,
17, 400-406.
Charlesworth, B. (2009). Fundamental concepts in genetics: effective population size and patterns of
molecular evolution and variation. Nature Reviews Genetics, 10, 195-205.
Charlesworth, B., Nordborg, M., & Charlesworth, D. (1997). The effects of local selection,
balanced polymorphism and background selection on equilibrium patterns of genetic
diversity in subdivided populations. Genetical Research, 70, 155-174.
88
Das, A. M. (2003). Regulation of the mitochondrial ATP-synthase in health and disease.
Molecular Genetics and Metabolism, 79, 71-82.
Dowling, D. K., Friberg, U., & Lindell, J. (2008). Evolutionary implications of non-neutral
mitochondrial genetic variation. Trends in Ecology and Evolution, 23, 546-554.
Duret, L., & Mouchiroud, D. (2000). Determinants of substitution rates in mammalian genes:
expression pattern affects selection intensity but not mutation rate. Molecular Biology
and Evolution, 17, 68-74.
Ellison, C. K., & Burton, R. S. (2006). Disruption of mitochondrial function in interpopulation
hybrids of Tigriopus californicus. Evolution, 60, 1382-1391.
Gissi, C., Reyes, A., Pesole, G., & Saccone, C. (2000). Lineage-specific evolutionary rate in
mammalian mtDNA. Molecular Biology and Evolution, 17, 1022-1031.
Haase, M. (2008). The radiation of hydrobiid gastropods in New Zealand: a revision including
the description of new species based on morphology and mtDNA sequence information.
Systematics and Biodiversity, 6, 99-159.
Haig, D. (2016). Intracellular evolution of mitochondrial DNA (mtDNA) and the tragedy of the
cytoplasmic commons. Bioessays.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection.
Genetical Research, 8, 269-294.
Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal
of Statistics, 6, 65-70.
Kumar, S., Nei, M., Dudley, J., & Tamura, K. (2008). MEGA: a biologist-centric software for
evolutionary analysis of DNA and protein sequences. Briefings in Bioinformatics, 9, 299306.
89
Li, W.-H., Wu, C-I., & Luo, C. C. (1985). A new method for estimating synonymous and
nonsynonymous rates of nucleotide substitution considering the relative likelihood of nucleotide
and codon changes. Molecular Biology and Evolution, 2, 150-174.
Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by
parasitism. Nature, 328, 519-521.
Lynch, M., Bürger, R., Butcher, D., & Gabriel, W. (1993). The mutational meltdown in asexual
populations. Journal of Heredity, 84, 339-344.
Marshall, H. D., Coulson, M. W., & Carr, S. M. (2009). Near neutrality, rate heterogeneity, and
linkage govern mitochondrial genome evolution in Atlantic cod (Gadus morhua) and
other gadine fish. Molecular Biology and Evolution, 26, 579-589.
McDonald, J. H., & Kreitman, M. (1991). Adaptive protein evolution at the Adh locus in
Drosophila. Nature, 351, 652-654.
Muller, H. J. (1964). The relation of recombination to mutational advance. Mutation Research,
106, 2-9.
Nabholz, B., Ellegren, H., & Wolf, J. B. (2013). High levels of gene expression explain the strong
evolutionary constraint of mitochondrial protein-coding genes. Molecular Biology and
Evolution, 30, 272-284.
Neiman, M., Hehman, G., Miller, J. T., Logsdon, J. M., Jr., & Taylor, D. R. (2010). Accelerated
mutation accumulation in asexual lineages of a freshwater snail. Molecular Biology and
Evolution, 27, 954-963.
Neiman, M., & Lively, C. M. (2004). Pleistocene glaciation is implicated in the
phylogeographical structure of Potamopyrgus antipodarum, a New Zealand snail.
Molecular Ecology, 13, 3085-3098.
90
Neiman, M., Paczesniak, D., Soper, D. M., Baldwin, A. T., & Hehman, G. (2011). Wide
variation in ploidy level and genome size in a New Zealand freshwater snail with
coexisting sexual and asexual lineages. Evolution, 65, 3202-3216.
Neiman, M., & Taylor, D. R. (2009). The causes of mutation accumulation in mitochondrial
genomes. Proceedings of the Royal Society of London B, 276, 1201-1209.
Normark, B. B., & Moran, N. A. (2000). Opinion - Testing for the accumulation of deleterious
mutations in asexual eukaryote genomes using molecular sequences. Journal of Natural
History, 34, 1719-1729.
Ozsolak, F., & Milos, P. M. (2011). RNA sequencing: advances, challenges and opportunities.
Nature Reviews Genetics, 12, 87-98.
Paczesniak, D., Jokela, J., Larkin, K., & Neiman, M. (2013). Discordance between nuclear and
mitochondrial genomes in sexual and asexual lineages of the freshwater snail
Potamopyrgus antipodarum. Molecular Ecology, 22, 4695-4710.
R Core Team (2016). R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.
Rand, D. M. (2001). The units of selection of mitochondrial DNA. Annual Review of Ecology
and Systematics, 2001, 415-448.
Tajima, F. (1989). The effect of change in population size on DNA polymorphism. Genetics, 123,
597-601.
Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D. R., . . . Pachter, L. (2012).
Differential gene and transcript expression analysis of RNA-seq experiments with
TopHat and Cufflinks. Nature Protocols 7, 562-578.
91
Watterson, G. A. (1975). On the number of segregating sites in genetical models without recombination.
Theoretical Population Biology 7, 256-276.
Yang, Z. (2007). PAML 4: phylogenetic analysis by maximum likelihood. Molecular Biology
and Evolution, 24, 1586-1591.
Zeyl, C., Andreson, B., & Weninck, E. (2005). Nuclear-mitochondrial epistasis for fitness in
Saccharomyces cerevisiae. Evolution, 59, 910-914.
Zhang, F., & Broughton, R. E. (2013). Mitochondrial-nuclear interactions: compensatory
evolution or variable functional constraint among vertebrate oxidative phosphorylation
genes? Genome Biology and Evolution, 5, 1781-1791.
Zhang, J. (2000). Rates of conservative and radical nonsynonymous nucleotide substitutions in
mammalian nuclear genes. Journal of Molecular Evolution, 50, 56-68.
92
FIGURES
Figure 3-1. Sliding window estimates of Jukes-Cantor-corrected dN/dS across the P.
antipodarum mitochondrial genome estimated using a window size of 100 bp and a step size
of 3 bp.
Red and blue lines represent the upper (red) and lower (blue) boundaries of 95% confidence
intervals generated using 10,000 bootstrap replicates for each 100 bp window. Breaks in lines
indicate regions of the mitochondrial genome that do not encode a protein (e.g., positions 3171 –
6147). Gene position and identity (blue bars) along the mitochondrial genome (orange bar) are
indicated below x axis for reference.
93
Figure 3-2. The number of substitutions per gene as a function of the number of
polymorphisms per gene.
Synonymous (S, red triangles) polymorphisms are significantly and positively correlated with
synonymous divergences (Spearman’s rho = 0.96, p < 0.0001). The number of nonsynonymous
(NS, black circles) polymorphisms is also positively correlated with the number of
nonsynonymous divergences (Spearman’s rho = 0.58, p = 0.0361, but the slope of the
relationship is significantly less than that of synonymous changes (t = 8.157, p < 0.0001). The
94
lines in the plot represent the best-fit linear regression for the data (red line: synonymous
changes, black line: nonsynonymous changes).
95
Figure 3-3. Relationship between dS and dN in mitochondrial genes of P. antipodarum as
estimated by the best-fit model M0 (species-wide rate; atp8), M1a (branch-specific rate; all
other 12 genes), or M2a (sexual-specific rate, asexual specific rate; 0 genes) in PAML.
Line depicts the best-fit linear regression model for the data. With the exception of atp8, 31
branch-specific estimates of dS and dN are shown for each gene as described in the key because a
branch-specific model was the best fit for the data. Only a single estimate of dS and dN is shown
for atp8 because model M0 (species-wide dN/dS) best explained the pattern of molecular
evolution for that gene.
96
97
Figure 3-4. Complex identity governs molecular evolution of mitochondrial genes in P.
antipodarum at both interspecific and intraspecific levels.
a) Median branch-specific dN/dS per complex as estimated by PAML assuming branch-specific
model M1a. Error bars denote standard deviations, and lower-case letters represent statistical
groupings based on Mann-Whitney U pairwise comparisons at the p < 0.05 level (following
Bonferroni correction using the Holm procedure). b) Mean species-wide θN/θS differs across
complexes (Kruskal-Wallis χ2 = 9.2732, p = 0.0259). Error bars represent standard deviation
within complexes.
98
Figure 3-5. Relationship between gene expression (FPKM) and OXPHOS complex.
a) Gene expression is significantly different across OXPHOS complexes (Kruskal-Wallis χ2 =
12.327, p = 0.0151). Boxes and error bars represent inner and outer quartile ranges, respectively.
Lower-case letters represent statistical groupings based on Mann-Whitney U pairwise
comparisons at the p < 0.05 level (following Bonferroni correction using the Holm procedure). b)
Quantile plot displaying significant and positive correlation between mean complex dN/dS from
mitochondrially encoded subunits and gene expression in terms of FPKM (Spearman’s rho =
0.21, p = 0.0122). Roman numerals indicate OXPHOS complex identity.
99
100
Figure 3-6. Scatterplots depicting relationship between species-wide dN/dS (estimated by
species-wide model M0 in PAML) and the difference between mean branch-specific
estimates of dN/dS (estimated by model M1a in PAML) in asexuals vs. sexuals.
a) The 12 genes that were best explained by model M1a in PAML. Positive y coordinates
indicate higher dN/dS ratios in asexuals; negative y coordinates indicate higher dN/dS ratios in
sexuals. Species-wide dN/dS is not correlated with the difference between mean branch-specific
estimates of dN/dS in asexuals vs. sexuals. b) There is a significant and positive correlation
between species-wide dN/dS and the difference between mean branch-specific estimates of dN/dS
in asexuals vs. sexuals when only the 10 genes in which asexuals exhibit signatures of
accelerated nonsynonymous mutation accumulation relative to sexuals. For both panels, the
black line in the plot represents the best-fit linear regression line.
101
Figure 3-7. The per-mitochondrial gene extent of mutation accumulation in asexuals due to
reduced Ne.
Bars represent residuals calculated from the best-fit linear regression for the relationship between
selective intensity (dN/dS) and the difference in branch-specific dN/dS in asexual vs. sexual
lineages (Figure 3-4 b) for each gene. Despite a wide range of values (-0.0032 – 0.0075), these
residuals do not significantly differ across genes or complexes (Kruskal- Wallis p > 0.05 for both
comparisons).
102
CHAPTER 4: GENETIC VARIATION FOR MITOCHONDRIAL
FUNCTION IN THE NEW ZEALAND FRESHWATER SNAIL,
POTAMOPYRGUS ANTIPODARUM
ABSTRACT
Mitochondrial function represents a complex and dynamically regulated phenotype, making
investigations into its function and evolution challenging. In particular, the proteins responsible
for mitochondrial function are encoded by two different genomes with distinct inheritance
regimes, rendering rigorous inference of genotype–phenotype connections intractable for all but
a few model systems. Asexual lineages provide a powerful means of addressing these challenges
by enabling comparisons among genetically identical individuals in which nuclear and
mitochondrial genomes are co-inherited. Here, I used Potamopyrgus antipodarum, a New
Zealand freshwater snail that is an emerging model system for the evolution of sexual
reproduction, to develop novel methods for measuring mitochondrial function in mollusks at
three distinct levels of biological organization: mitogenomic, organelle, and organismal. I
applied these methods to multiple asexual lineages of P. antipodarum and evaluated whether
there exists genetic variation (i.e., the raw material for adaptive evolution) for mitochondrial
genome copy number, mitochondrial membrane potential, and organismal oxygen consumption.
These analyses revealed that asexual lineages of P. antipodarum differed in terms of
mitochondrial function at all three levels of biological organization. These results both
demonstrate the existence of variation for mitochondrial function in P. antipodarum, the first
such report in a mollusk, and set the stage to use these methods to study phenomena such as
connections between mitochondrial mutation accumulation and the maintenance of sexual
reproduction.
103
INTRODUCTION
Mitochondrial function is of critical importance to eukaryotic health (Chen et al. 2007, Pike et al.
2007, Barreto and Burton 2013, Dowling 2014), and genetic variation for mitochondrial function
has been linked to adaptation (Rawson and Burton 2002, Bazin et al. 2006) and disease (Frank
and Hurst 1996, DiMauro and Schon 2001). The role of genetic variation for mitochondrial
function is further complicated by direct interaction between nuclear encoded and
mitochondrially encoded proteins, particularly with respect to oxidative phosphorylation
(OXPHOS) (Blier et al. 2001, Das et al. 2004, Rand et al. 2004). Mitonuclear cooperation is
critical to proper enzyme function in OXPHOS complexes I, III, IV, and V because these
complexes are typically comprised of subunits from both genomes. Accordingly, discordance
between mitochondrial and nuclear genomes has been demonstrated to have negative fitness
and/or functional consequences in copepods (Ellison and Burton 2006), Drosophila (Meiklejohn
et al. 2013, Pichaud et al. 2013), and salamanders (Lee-Yaw et al. 2014).
In sexually reproducing organisms, the maintenance of mitonuclear cooperation is
complicated by the expectation that the different mechanisms of nuclear vs. mitochondrial
genome inheritance will differentially affect the generation, maintenance, and distribution of
genetic variation. In particular, biparental inheritance and meiotic recombination in the nuclear
genome will increase effective population size (Charlesworth 2009) relative to the (typically)
uniparentally inherited and non-recombinant mitochondrial genome (reviewed in Barr et al. 2005,
Neiman and Taylor 2009). This logic is the basis for the expectation that mitochondrial genomes
will, when compared to the nuclear genome, experience reduced efficacy of selection and suffer
an increased rate of accumulation of mildly deleterious mutations (Muller 1964, Hill and
Robertson 1966, Neiman and Taylor 2009). This mechanism is also thought to generate selection
104
favoring compensatory changes in nuclear-encoded mitochondrial subunits. There is some
evidence that nuclear genomes are the primary site of these compensatory changes (Sloan et al.
2013, Zhang and Broughton 2013).
Otherwise similar sexual and asexual taxa provide a powerful means of addressing how
reproductive mode influences mitochondrial evolution and performance because the absence of
nuclear recombination in asexual organisms means that asexual lineages will transmit their
mitochondrial genomes in complete linkage disequilibrium (LD) with the nuclear genome. The
co-transmission of nuclear and mitochondrial genomes is expected to result in two non-mutually
exclusive evolutionary scenarios: (1) increased LD should decrease the efficacy of selection in
both genomes via the Hill-Robertson effect (Hill & Robertson 1966, Neiman and Taylor 2009),
and (2) tight linkage between nuclear and mitochondrial genomes may allow epistatic variation
to become available to natural selection in a manner that only asexuals can access (reviewed in
Neiman & Linksvayer 2006).
Potamopyrgus antipodarum, a New Zealand freshwater snail, is very well suited for
investigating mitochondrial function in the absence of sex because otherwise similar obligately
sexual and obligately asexual individuals frequently coexist in natural populations. The transition
from sexual ancestor to asexual descendent has occurred multiple separate times within this
species, providing many so-called “natural experiments” into the consequences of asexuality
(Neiman et al. 2011). Surveys of patterns of molecular evolution in the mitochondrial genome of
P. antipodarum have demonstrated that asexual lineages of P. antipodarum accumulate
nonsynonymous substitutions in their mitochondrial genomes more rapidly than their sexual
counterparts (Neiman et al. 2010, Chapter 2, 3) and retain radical amino acid-changing
polymorphisms longer than sexual lineages (Chapter 2). What remains unevaluated is whether
105
these putatively deleterious changes have phenotypic consequences – the key piece of
information required to evaluate the extent to which mutation accumulation can influence
asexual lineage success and the maintenance of sex.
Because natural selection can only operate on genetic variation that is linked to fitness,
the critical first step in addressing whether mitochondrial mutation accumulation might
contribute to the maintenance of sex in P. antipodarum is to determine whether there exists
genetic variation for mitochondrial function in this species. In particular, variation in
mitochondrial function amongst asexual lineages would indicate that genetic variation impacts
phenotypic variation, suggesting that the accumulation of deleterious mutations might well be
accompanied by a corresponding decrease in function and/or fitness.
Here, I use a common garden approach to test whether asexual lineages of P.
antipodarum vary in terms of mitochondrial function at three distinct levels of biological
organization: (1) mitochondrial genome copy number, (2) mitochondrial membrane potential and
electron transport, and (3) organismal oxygen (O2) consumption. All three traits have been
shown to be linked to mitochondrial function in other taxa. Mitochondrial genome copy number
is positively correlated with respiratory capacity, particularly with respect to Cytochrome C
Oxidase activity (Moraes et al. 1991). Elevated mitochondrial membrane potential, generated by
electron transport through OXPHOS, is positively correlated with the ratio of ATP:ADP,
indicating that stronger membrane potentials reflect healthier mitochondria (Nicholls 2004).
Finally, because O2 is the primary electron acceptor for the production of ATP (Chance and
Williams 1955), organismal O2 consumption represents an integrated measure of the ATP
production capacity of an organism. I took advantage of the fact that heat stress can increase
inefficiencies in oxidative phosphorylation (Heise et al. 2003) to use heat treatments to reveal
106
cryptic phenotypic variation for mitochondrial function in P. antipodarum. With respect to
mitochondrial genome copy number and membrane potential, I adapted well-established
mitochondrial functional assays that had previously been employed exclusively in model
organisms, probing the strength of the electrochemical gradient, and the flow of electrons
through the electron transport chain (ETC) to test for genetic variation in asexual lineages of P.
antipodarum. I used aquatic respirometry to quantify and compare O2 consumption under heat
stress across asexual lineages. Together, these analyses revealed substantial variation for
mitochondrial function across asexual lineages at all three levels of biological organization. The
development of these assays also provides novel and broadly useful tools for investigating
mitochondrial function in non-model organisms.
107
METHODS
Snail husbandry
I compared mitochondrial function across a diverse array of asexual P. antipodarum lineages
(Neiman and Lively 2004, Table 4-S1) reared under identical conditions for multiple generations,
enabling me to interpret across-lineage variation in my various measures of mitochondrial
function as representing genetic variation for these traits. Asexuality was established for each
lineage by determining ploidy using flow cytometry (sexual P. antipodarum are diploid, asexuals
are polyploid), as described in Neiman et al. (2012). Asexual lineages were chosen for functional
assays to represent the range of mitochondrial diversity found in New Zealand populations
(Neiman et al. 2011, Paczesniak et al. 2013) and to maximize my ability to compare different
functional assays. Adult female snails were selected arbitrarily from lineage populations for each
assay. Snails were housed at 16ºC on a 18 hr light/6 hr dark schedule and fed Spirulina algae 3x
per week, as described in Zachar and Neiman (2013).
Mitochondrial function at the genomic level
Mitochondrial genome copy number is known to affect respiratory capacity (Moraes et al. 1991,
Van den Bogart et al. 1993) and to be dynamically regulated in response to various cellular
environmental cues (Hori et al. 2009, Matsushima et al. 2010, Kelly et al 2012). As such,
mitochondrial genome copy number represents a phenotype that could respond to selection to
reduce the potentially negative consequences of mitochondrial mutation accumulation. To test
whether P. antipodarum exhibited phenotypic variation for mitochondrial genome copy number,
I used quantitative PCR (qPCR) to estimate mitochondrial genome copy number relative to a
putatively single copy nuclear gene in 6 asexual triploid lineages. A putatively single-copy gene
108
from the P. antipodarum nuclear genome (rad21; Neiman et al. in prep) and mitochondrial cytB
were chosen for analysis. Primer pairs were designed to generate a 264-bp rad21 product (F: 5’–
GATTCCAACAACTGATGTTTG –3’, R: 5’–CAAAACTTACTCTAAATCTGC–3’) product
and a 194-bp amplicon from cytB (F: 5’–TATGAATATTCAGATTTTTTAAATA–3’, R: 5’–
CCTTAACTCCTAATCTTGGTAC–3’). For measurement standards, each of these products was
cloned from total DNA from a single P. antipodarum lineage into the pGEM T-Easy plasmid
vector. Linearized plasmids were diluted in the presence of carrier human genomic DNA to
produce samples containing 300 to 300,000 copies of either the nuclear or mitochondrial
amplicon. Total DNA was isolated from three to five individual snails from each lineage using
the Qiagen DNEasy Plant Mini Kit, and nuclear and mitochondrial targets were amplified in
triplicate in separate reactions on the same plate, together with serial dilutions of cloned
standards. I converted quantitation cycle values (Cq), the PCR cycle at which fluorescence can
be detected, from snail samples into copy number using the curves generated from the cloned
standards as in Miller et al. (2003), and the ratio of mitochondrial to haploid nuclear genome
copies was verified for each sample; all lineages examined were triploid. Because the ratio of
cytB to rad21 fit the assumptions of normality (Shapiro-Wilks W = 0.956, p = 0.228) and
variances between lineages were not significantly different (Levene’s F = 0.632, p = 0.677), I
compared inferred mitochondrial genome copy number across lineages using a one-way
ANOVA and pairwise t tests as implemented in R (v 3.2.4, R Core Team 2016).
109
Mitochondrial function at the organellar level
JC-1 assay
Mitochondrial membrane potential represents the electrochemical gradient mitochondria use to
phosphorylate ADP into ATP. Accordingly, the amplitude of membrane potential is positively
correlated with the ratio of ATP to ADP in the cell (Nicholls 2004). Variation in this
mitochondrial membrane potential has been linked to cellular aging (Sugrue and Tatton 2001,
Nicholls 2004) and survival (Callegari et al. 2011), making mitochondrial membrane potential a
useful phenotype to gauge mitochondrial performance. The JC-1 assay (Garner and Thomas
1999) measures the strength of the electrochemical gradient in mitochondrial isolates using a
fluorescent dye, JC-1. Under UV light, JC-1 will fluoresce green when dispersed and red when
aggregated. In the absence of an electrochemical gradient, JC-1 will diffuse randomly across
membranes and will aggregate in proportion to its concentration, such that the ratio of red: green
fluorescence should be roughly equivalent. If isolated mitochondria establish and maintain
electrochemical gradients, JC-1, a small, positively charged molecule, will diffuse down the
gradient and accumulate inside mitochondrial matrices, yielding an elevated ratio of red: green.
A higher ratio of red: green is indicative of increased mitochondrial membrane potential, and
thereby, mitochondrial function (Garner and Thomas 1999). I used the uncoupler carbonyl
cyanide m-chlorophenyl hydrazone (CCCP) to abolish the electrochemical gradient across the
mitochondrial inner membrane, controlling for background fluorescence of JC-1 and of
mitochondrial membranes unrelated to mitochondrial function.
To test for genetic variation in mitochondrial function in P. antipodarum in terms of
mitochondrial membrane potential, I assayed 8-10 adult female snails from six distinct asexual
lineages representing a diverse subset of the natural mitochondrial haplotype diversity found in
110
New Zealand (Neiman and Lively 2004, Neiman et al. 2011, Paczesniak et al. 2013, Table 4-S1).
I first removed snail shells and briefly washed collected tissues by centrifugation at 600 x g in
extraction buffer (10.0 mM HEPES (pH 7.5), 0.2 M mannitol, 70.0 mM sucrose, 1.0 mM EGTA).
I rapidly homogenized snail tissues on ice in extraction buffer containing 2 mg/ml fatty acid-free
bovine serum albumin (fafBSA) using a micropestle and then centrifuged the homogenized
tissue at 4ºC for 5 minutes at 600 x g. The supernatant was pipetted off and held on ice
separately while I re-homogenized and centrifuged the pellet again, as above. I next centrifuged
the pooled mitochondrial-enriched supernatant at 12,000 x g for 10 minutes and then resuspended the pellet in buffer containing 10.0 mM HEPES (pH 7.5), 0.25 M sucrose, 1.0 mM
ATP, 0.08 mM ADP, 5.0 mM sodium succinate, 2.0 mM K2HPO4, and 1mM DTT. I divided
each sample into three subsamples of 30 µl each and added to each subsample to assay buffer
containing 20.0 mM MOPS (pH 7.5), 110.0 mM KCl, 10mM ATP, 10.0 mM MgCl2, 10.0 mM
sodium succinate, and 1.0 mM EGTA. The first subsample was incubated with 500 µl buffer
alone, to monitor background fluorescence; the second subsample was incubated with 2 µM JC-1
and 500 µl buffer, and the third subsample received 2 µM JC-1 and 30 µM CCCP with 500 µl
buffer. All three subsamples were incubated in the dark at 37ºC for 20 minutes, after which the
ratio of red: green fluorescence was determined using flow cytometry. For flow cytometry, I
collected forward scatter (FSC) and side scatter (SSC) at the following wavelengths: FL1: green
– 400 nm (log scale), FL2: red – 320 nm (log scale), and FL3: farther red – 300 nm (log scale). I
plotted FL1 vs. FL2 for each sample in FloJo v 10.0.8 and derived the ratio of red to green for
each particle. The median ratio of red to green for each sample run was used for all comparisons.
Because red: green fluorescence ratios were not normally distributed (Shapiro-Wilks W =
0.858, p = 7.131 x 10-6), I used a log transformation of the data to meet the normality assumption
111
of parametric analyses (Shapiro-Wilks W = 0.969, p = 0.15). I also found that ratios of red: green
across lineages did not have equal variances (Levene’s F = 3.093 p = 0.016), so I used a one-way
ANOVA with a White adjustment for unequal variances (McKinnon and White 1985) and
lineage as a random factor to determine whether lineage affected the dependent variable of red:
green ratio. I performed pairwise comparisons between lineages using Welch’s t tests (to reflect
unequal variances) and the Holm procedure for Bonferroni correction (Holm 1979). All
statistical analyses were conducted in R (v 3.2.4, R Core Team, 2016).
MTT assay
Isolated mitochondria can reduce MTT (3-(4,5-dimethylthizol-2-yl) diphenyltetrazolium
bromide). In the presence of succinate, MTT accepts electrons from the ETC (Liu et al. 1997),
forming a purple formazan product that is the basis of a colorimetric assay of electron transport
activity. The amount of purple formazan produced by reducing O2 is proportional to the electron
flow through the ETC. Electron flow provides the energy necessary to establish a proton gradient
(Liu et al. 1997), such that increased electron flow through the ETC should produce a
corresponding increase in mitochondrial membrane potential. The extent to which MTT accepts
electrons can be quantified by the MTT assays described by Liu et al. (1997), whose protocol I
adapted and implemented here. Briefly, I resuspended mitochondrial pellets (produced as
described in section 2.3.1) from snail tissue in buffer and added these resuspended mitochondrial
isolates to wells of a 96-well plate containing succinate as an energy substrate. The reaction was
initiated by adding MTT to each well, and the plate was incubated for 2 hours at 37°C to allow
electrons from the ETC to reduce MTT. Next, I applied a 20% SDS 50% dimethylformamide
solubilization solution to each well. The plate was left to incubate overnight, and the reduced
112
MTT product was measured by spectrophotometry (A570). I determined background absorbance
from duplicate mitochondrial samples incubated without an energy substrate and subtracted this
background value from all readings. Phosphate-buffered saline was used as a negative control
and dithiothreitol as a positive control for MTT reduction. A fraction of each mitochondrial
sample was lysed in SDS and then used in a bicinchoninic acid assay to determine the protein
concentration. MTT reduction was expressed as A570/ug protein.
While MTT values did not fit the assumptions of a normal distribution (Shapiro-Wilks W
= 0.840, p = 2.436 x 10-5), these values did meet the assumptions of normality following a log
transformation (Shapiro-Wilks W = 0.961, p = 0.140). Log-transformed MTT values did not
display unequal variances (Levene’s F = 0.964, p = 0.452), allowing me to compare logtransformed MTT values using a one-way ANOVA (as implemented in R v 3.2.4, R Core Team
2016) with lineage as a random factor to test whether different asexual lineages exhibited
different levels of electron flow through the ETC.
First, to determine whether P. antipodarum exhibit genetic variation for MTT reduction, I
used a one-way ANOVA to address whether the dependent variable of mean MTT reduction per
each of the six asexual lineages varied across the random factor of lineage. Next, to determine
the extent to which the MTT assay corresponded to results from the JC-1 assay, I compared
mean MTT reduction per lineage to mean red: green ratio per lineage from each of six asexual
lineages and two sexual lineages (one inbred, one field-collected, see Table 4-S1) using
Spearman’s rank correlation.
113
Mitochondrial function at the organismal level
Ectotherms respire at higher rates at elevated temperatures, such that O2 becomes limiting as
temperature increases (Abele et al. 2007). Accordingly, heat stress can reveal inefficiencies in
ATP production (Heise et al. 2003). By this logic, I hypothesized that measuring O2 consumption
under heat stress would reveal existing genetic variation for mitochondrial function. There is
some evidence from earlier studies that P. antipodarum are susceptible to heat stress: snails have
decreased fecundity (Dybdahl and Kane 2005) and tend to consume more O2 (Hudson 1975)
when raised at (Dybdahl and Kane 2005) and exposed to (Hudson 1975) elevated temperatures.
To further define the range of heat stress likely to alter mitochondrial function in P.
antipodarum, I used a behavioral assay that measures the amount of time that a snail takes to
right itself when placed upside down to compare righting ability in 13 asexual lineages (N = 10
per lineage) across three temperature treatments (16ºC, 22ºC, 30ºC). Righting ability is a
commonly used method to gauge levels of snail stress (see e.g., Orr et al. 2007, Lukowiak et al.
2008, Watson et al. 2014), and snails exposed to hypoxic conditions exhibit increased righting
time and elevated HSP70 gene expression compared to unexposed snails (Fei et al. 2008). I first
incubated adult female snails in water at the test temperature for 1 hour. Next, I placed snails
ventral side up in a petri dish and measured the number of seconds elapsed until the snail righted
itself, up to 180 seconds. I compared righting time across temperatures and lineages with a
Kruskal-Wallis test (as implemented in R) and found that both temperature (Kruskal Wallis, χ2 =
14.218, df = 2, p = 0.00082) and lineage (Kruskal Wallis, χ2 = 122.64, df = 12, p < 2.2 x 10-16)
significantly affected righting time. Righting time was ~24% faster at 22ºC than at 16ºC, but this
difference was not significant (Mann-Whitney, U = 9244.5, p = 0.0754). Conversely, righting
time took ~ 37% longer at 30ºC than at 16ºC (Mann-Whitney U = 8192, p = 0.0254) and ~85%
114
longer at 30ºC than at 22ºC (Mann-Whitney, U = 5996.5, p = 0.00014, Figure 4-S1). These
results are consistent with other righting time assays from other snail species in that some
amount of stress decreases righting time (e.g., presence of predators, Orr et al. 2007), and some
stimuli increase righting time (e.g., hypoxia, Fei et al. 2008), suggesting that my observed
responses indicate stress in P. antipodarum when exposed to elevated temperature (here, 30ºC).
After determining that snails do appear to be stressed by elevated temperatures, I tested
whether there exists genetic variation for O2 consumption under heat stress in P. antipodarum by
performing closed-system aquatic respirometry on seven asexual lineages of P. antipodarum (N
= 10 per lineage) at three different incubation temperatures (16ºC, 22ºC, 30ºC) using a
Strathkelvin Instruments RC200a respiration chamber, 892 Oxygen Meter, and a 1302 Clarktype oxygen electrode. I calibrated the electrode daily using the solubility of O2 at each
respective temperature (16ºC – 309.0 µmol/liter, 22ºC – 279.0 µmol/liter, 30ºC – 236.0
µmol/liter). A high calibration point was obtained by stirring carbon-filtered water vigorously for
30 minutes prior to calibration, while I used a 2% sodium sulfite solution as a low calibration
standard. I incubated each snail at the prescribed temperature for one hour prior to measurement,
placed snails into the cell chamber, and obtained O2 concentration readings for each snail every
second for 1 hour. I maintained constant temperature in the respiration chamber by pumping
temperature-controlled water into cell chamber’s water jacket. Upon completion of the 1 hour
test period, I blotted each snail dry and measured the wet mass of the snail on a Denver
Instruments Cubis Analytical Balance
I first determined that snail wet mass was significantly and positively correlated with O2
consumption (Spearman’s rho = 0.17, p = 0.016, Figure 4-S2). Because there is significant
variation for snail wet-mass between asexual lineages (Kruskal-Wallis χ2 = 76.82, df = 6, p <
115
0.0001), I calculated the residuals of wet mass vs. O2 consumption using a linear regression
model. Square-root-transformed O2 consumption residuals were not significantly different from a
normal distribution (Shapiro-Wilks, W = 0.9894, p = 0.14), allowing me to implement use a twoway ANOVA to address whether the fixed factor of temperature, the random factor of lineage,
and the interaction between temperature and lineage affected the dependent variable of wet-mass
corrected O2 consumption. Next, I tested whether lineage had an effect on O2 consumption
within each temperature by performing a one-way ANOVA with lineage as a random factor for
16ºC, 22ºC, and 30ºC temperature treatments.
Comparison of mitochondrial functional assays
Because each of the three assays that I developed and employed is designed to reflect some
aspect of mitochondrial function, the outcomes of different assays might be expected to be
correlated if the assays in question provide related measures of mitochondrial function. Lack of
association between assay outcomes will instead suggest that these different assays measure
independent components of mitochondrial phenotype. To test for associations between different
mitochondrial functional assays, I determined mean trait value for each lineage and assay and
used Spearman’s rank correlation to compare all 15 possible pairwise combinations between
assays. For these analyses, I included two additional sample populations that were not used in
any of the tests for genetic variation: one field-collected sample from a lake with a high
frequency of sexual individuals (“KnSF12”) and an inbred sexual lineage that has been
maintained in the lab for >20 generations (“Y2”, Table 4-S1). Because different individual snails
were used for each assay, it is only possible to compare lineage means for each assay, which
limited my ability to perform rigorous comparisons for some assay combinations.
116
RESULTS
Mitochondrial function at the genomic level
To test whether P. antipodarum exhibits heritable variation for mtDNA copy number, I
compared qPCR amplification of the mitochondrially encoded cytB locus to qPCR amplification
of a putatively single copy nuclear gene, rad21, in six asexual lineages of P. antipodarum reared
in a common garden setting (N = 3-8 per lineage). The mean (+/- SD) number of cytb copies to
the number of rad21 copies was 13.715 (+/- 3.571), though I also found significant differences in
the ratio of cytB copies to rad21 copies across asexual lineages (one-way ANOVA, F5, 25 = 2.72,
p < 0.0428, Figure 4-1), indicating that lineages differ in mitochondrial genome copy number. I
performed pairwise t-tests of copy number ratio among each pair of lineages and found that a
single lineage (Gr5; mean copy number ratio = 9.423, SD = 1.483) with very low mitochondrial
copy number appeared to be driving the significant among-lineage variance. To test if the lack of
differences amongst the remaining lineages could be an artifact of low statistical power, I
performed power analyses on each of the 28 pairwise comparisons. I found mean power (+/- SD)
= 0.402 (+/- 0.398), indicating both that power was low to moderate but also that power varied
across comparisons. For all comparisons involving the Gr5 lineage, mean power (+/- SD) =
0.881 (+/- 0.224), while for all comparisons not involving Gr5, mean power (+/- SD) was 0.162
(+/- 0.183) This result is likely driven by the low mitochondrial genome copy number of Gr5
relative to the other lineages and suggests that sample size was likely too small to detect the
more minor differences in copy number, if any, that might exist between other asexual lineages.
117
Mitochondrial function at the organellar level
JC-1
I next compared membrane potentials in mitochondria isolated from six asexual lineages of P.
antipodarum using the JC-1 assay, in which the ratio of red to green fluorescence is indicative of
membrane potential. I found significant differences in the log-transformed ratios of red: green
across asexual lineages (Welch’s one-way ANOVA, F5, 52 = 6.628, p = 7.671 x 10-5, Figure 4-2).
Pairwise posthoc comparisons revealed three significantly different groups (p < 0.05 after
correcting for multiple comparisons) amongst the six lineages (Figure 4-2). Unlike mitochondrial
genome copy number, this variation in mitochondrial membrane potential does not appear to be
driven by a single lineage. Because CCCP-treated samples did not vary across lineages (one-way
ANOVA, F5,52 = 1.367, df = 5, p = 0.252), the variation observed for red: green fluorescence
across lineages likely reflects variation in mitochondrial membrane potential. In particular, these
data suggest that under the current rearing and assay conditions, the Gr5 lineage appears to have
the strongest mitochondrial membrane potential (mean (SD) = 3.859 (1.546)) and the DenA
lineage (mean (SD) = 1.775 (0.468)) appears to have the weakest mitochondrial membrane
potential.
MTT
MTT reduction by isolated mitochondria is due to acceptance of electrons from the electron
transport chain of the inner membrane, meaning that the extent of MTT reduction should reflect
electron flow through oxidative phosphorylation. In general, the outcome of the MTT assay
would be expected to be correlated with the strength of the proton gradient measured in the JC-1
assay, although that gradient can also be affected by the presence of uncoupling proteins, which
118
reduce the gradient by letting protons leak back into the matrix, dissipating energy as heat.
Despite clear evidence for across-lineage variation in JC-1, I did not detect any differences
amongst asexual lineages in MTT reduction (ANOVA, F5, 38 = 0.751, p = 0.591), despite
relatively high power (power = 0.92).
Mitochondrial function at the organismal level
Organismal O2 consumption under stressful conditions is expected to reflect electron donor
turnover under maximal ATP production (Abele et al. 2007). To test for genetic variation in O2
consumption in heat-stressed P. antipodarum, I performed closed-system aquatic respirometry
for seven asexual lineages (N = 10 per lineage) at 16ºC, 22ºC, and 30ºC. I found that elevated
temperatures significantly affected mass-corrected O2 consumption residuals (two-way ANOVA,
F1, 191 = 79.118, p = 4.44 x 10-16) and that there was a significant interaction between lineage and
temperature (Figure 4-3, two-way ANOVA, F1, 191 = 3.919, p = 0.00102). Using a one-way
ANOVA within each temperature treatment, I found that lineage had a significant effect on O2
consumption at 22ºC (one-way ANOVA F 6, 58 = 3.002, p = 0.0127) and at 30ºC (one-way
ANOVA F6, 62 = 3.823, p = 0.00263), but not at 16ºC (one-way ANOVA, F6, 64 = 1.892, p =
0.096). In particular, lineages responded differently to elevated temperature, with some lineages
(e.g., Ta10) exhibiting relatively high O2 consumption at high temperatures and others
maintaining similar levels of O2 consumption (e.g., Gn5) across temperatures. This result
demonstrates that genetically distinct lineages of P. antipodarum consume different amounts of
O2 in response to heat stress.
119
Comparison of mitochondrial functional assays
To determine the extent to which my measures of mitochondrial function correlate with one
another, I performed Spearman’s rank correlation for each of the 15 possible pairwise
comparisons of the assays for mitochondrial genome copy number, log-transformed JC-1 red:
green ratios, log-transformed MTT values, and square-root-transformed O2 consumption at each
temperature. None of the correlations remained significant after correction for multiple
comparisons. The absence of significant relationships between assay results is not surprising in
light of the fact that I had power of ~ 0.35 – 0.65 to detect a relationship between two variables
at the p < 0.003 level, the alpha required by the Holm procedure. Power notwithstanding, there
were some interesting results that deserve further scrutiny. First, neither mitochondrial genome
copy number nor O2 consumption values for 16º C or 30º C were correlated with any other
phenotypic measure (Figure 4-4). While the MTT assay requires pooling of mitochondria from
3-4 individual snails per measurement and is thus somewhat less sensitive than JC-1, which only
requires mitochondria from one snail, I did find that MTT reduction was positively correlated
with JC-1 measures of the proton gradient in the six asexual and two sexual P. antipodarum
lineages examined (Spearman’s rho = 0.81, p = 0.0149, Figure 4-4 f). Mean log-transformed JC1 values were negatively correlated with O2 consumption at 22ºC for the six asexual lineages
(Spearman’s rho = -0.77, p = 0.0724, Figure 4-4 k), and mean log-transformed MTT values were
negatively correlated with O2 consumption at 22ºC for the six asexual lineages (Spearman’s rho
= -0.89, p = 0.0188, Figure 4-4 k). Together, these tentative relationships between the MTT
assay (reflecting OXPHOS electron flow) and O2 consumption and between the MTT assay and
the JC-1 assay (reflecting mitochondrial membrane potential) suggest that high mitochondrial
membrane potential is expected to result in low O2 consumption under O2 limiting conditions in
120
P. antipodarum. These results also suggest that genome copy number, JC-1, and organismal O2
consumption provide orthogonal lines of inquiry into mitochondrial function.
121
DISCUSSION
Genetic variation for mt function at three different levels of biological organization
I used obligately asexual lineages of Potamopyrgus antipodarum, a freshwater New Zealand
snail, to test for genetic variation in mitochondrial function in a common garden setting. I found
significant levels of variation at all three levels of biological organization that I assayed:
mitochondrial genome copy number, mitochondrial membrane potential, and variation in O2
consumption in response to heat stress. These results demonstrate that substantial levels of
variation for mitochondrial function exist in this species and suggest that mitochondrial function
at the organellar level is related to mitochondrial function at the organismal level. Importantly,
the methods described here can be easily adapted to other non-model organisms, especially
mollusks, providing a new means of quantifying genotype-phenotype relationships.
The extent to which this variation in mitochondrial phenotype contributes to fitness in P.
antipodarum remains to be directly evaluated. Even so, the substantial across-lineage variation
that I discovered provides functional evidence of high levels of asexual phenotypic diversity in P.
antipodarum, consistent with previous reports that asexual P. antipodarum harbor substantial
genetic (Dybdahl and Lively 1995, Jokela et al. 2003, Paczesniak et al. 2013) and phenotypic
(Kistner and Dybdahl 2013, Neiman et al. 2013) diversity. The mitochondrial phenotypes I
observed in one lineage, Gr5, were particularly distinct: snails from this lineage exhibited
relatively low mitochondrial copy number, high mitochondrial membrane potential and electron
flow, and low O2 consumption at 22ºC. Together, these phenotypic values suggest that the Gr5
lineage exhibits relatively high mitochondrial function, indicating that genetic dissection of its
mitochondrial haplotype may prove illuminating. Further comparisons with other P.
antipodarum lineages and in other conditions will provide substantial insight into the relative
122
fitness of this particular mitonuclear combination. Future studies should also focus on a
particular mitochondrial haplotype that appears to be especially common amongst asexual P.
antipodarum (Neiman and Lively 2004). Paczesniak et al. (2013) showed that this haplotype is
often found in divergent nuclear backgrounds, consistent with a scenario where this haplotype is
spreading into new populations and lineages. Evaluating mitochondrial function of the common
mitochondrial haplotype in P. antipodarum with divergent vs. similar nuclear backgrounds and
in various biologically relevant conditions would shed light onto intraspecific mitonuclear
coevolution in the context of asexuality and on whether selection is effective at favoring epistatic
combinations in the absence of sex.
That the phenotypic variation for mitochondrial function shown here resides in traits that
have been demonstrated to influence fitness in other organisms (Chen et al. 2007, Pike et al.
2007, Barreto and Burton 2013, Dowling 2014) means that this variation could have major
implications for asexual lineage success. In particular, asexual P. antipodarum are known to
harbor mitochondrial genomes with high mutational loads relative to sexual lineages (Neiman et
al. 2010, Chapter 2, 3), suggesting that accelerated mutation accumulation in asexual lineages
likely affects mitochondrial function. The phenotypic variation in mitochondrial function
combined with the putatively deleterious genetic variation present in asexual lineages of P.
antipodarum suggests the possibility that asexual lineages harbor largely deleterious phenotypic
variation relative to sexual phenotypic variation. This possibility can be addressed by
comparisons of mitochondrial function in diverse sets of sexual and asexual P. antipodarum.
123
Relationship between mitochondrial functional assays
All else being equal, stronger membrane potentials, greater electron flow, lower O2 consumption,
and higher mitochondrial genome copy number should all positively correlate with overall
mitochondrial function. Because all else is almost certainly not equal (e.g., variation in
temperature/resources in native range, phylogenetic constraint, etc.), it is possible that high
mitochondrial performance in common garden conditions does not reflect mitochondrial
performance in nature. Despite this caveat, the tools developed here will provide an important
starting point by which to interrogate mitochondrial function in a non-model system.
Although the number of asexual lineages included in the present study was relatively
small, I did observe some interesting relationships between mitochondrial genome copy number,
electron flux through the ETC, mitochondrial membrane potential, and organismal O2
consumption (Figure 4-4). In particular, I found an inverse relationship between O2 consumption
at 22ºC and electron flow through the ETC and a positive relationship between electron flow
through the ETC and mitochondrial membrane potential. These results are consistent with my
predictions that mitochondria with larger membrane potentials harvest electrons from O2 more
efficiently than mitochondria with smaller membrane potentials, such that fewer O2 molecules
need to be oxidized per ATP produced.
My data also provide a preliminary line of evidence that inefficient mitochondria might
be at least in part compensated for by elevated mitochondrial gene product expression:
transcriptomic data from the inbred sexual lineage used in this study indicate that OXPHOS
complexes experiencing less intense purifying selection are expressed at higher levels than those
experiencing more intense selection (Chapter 3), suggesting that complex expression may be
regulated in response to inefficiency stemming from mutational load. If this observation extends
124
species-wide, because asexual P. antipodarum harbor a high load of putatively harmful
mitochondrial mutations relative to sexual P. antipodarum (Neiman et al. 2010, Chapter 2, 3), I
predict that asexual P. antipodarum will tend to express OXPHOS genes at higher levels than
sexual counterparts.
Implications for the loss of sex
One of the primary hypotheses put forth as an explanation for the prevalence of sexual
reproduction in eukaryotes is the accumulation of slightly deleterious mutations in asexual
lineages, eventually causing mutational meltdown and lineage extinction (Lynch et al. 1993).
Empirical evidence in support of this hypothesis remains largely indirect: (1) the relatively
“twiggy” distribution of asexual taxa on the eukaryotic phylogeny compared to sexual taxa
(Williams 1975, Maynard Smith 1978, Bell 1982; but see Schwander and Crespi 2009), and (2)
the accumulation of putatively deleterious mutations in nuclear (Kaiser and Charlesworth 2010,
Tucker et al. 2013) and organelle genomes of asexual lineages (Cutter and Payseur 2003,
Neiman et al. 2010, Henry et al. 2012, Horandl and Hojsgaard 2012, Voigt-Zielinsji et al. 2012,
Hollister et al. 2015, Chapter 2, 3). What remains unresolved is whether the mutations being
accumulated by asexual lineages actually decrease function and/or fitness. Establishing the
presence of genetic variation for mitochondrial function represents the first step in determining
whether mitochondrial mutation accumulation in the absence of sex causes such a decline. More
specifically, if mitochondrial haplotype is not linked to mitochondrial phenotype, mutation
accumulation in mitochondrial genomes in asexual lineages would not cause decreases in
function, such that sexual reproduction must be associated with other benefits to persist in the
face of asexual competitors. Here, I show that snails with different mitochondrial genetic
125
backgrounds have different mitochondrial phenotypes is a shared environment, suggesting that
mutations present in certain lineages have some impact on function, and setting the stage for
direct comparisons between mitochondrial function in sexual and asexual P. antipodarum.
One alternative possibility to decreased mitochondrial function in asexual lineages is that
asexuals have access to genetic variation that sexuals do not. This scenario is not wholly
unexpected: because the nuclear and mitochondrial genomes are transmitted in complete LD in
asexuals, advantageous combinations of mutations (i.e., epistatic variation) can become visible to
selection (Neiman and Linksvayer 2006). By contrast, in sexual lineages, the mitochondrial
genome is swapped across diverse nuclear backgrounds each generation, such that selection
cannot act effectively on phenotypes linked to epistatic mitonuclear interactions. The
implications are that interacting nuclear and mitochondrial subunits may experience particularly
tight and effective coevolution. Such a phenomenon might exhibit a very similar pattern of
molecular evolution to that of accelerated deleterious mutation accumulation, meaning that
testing the functional consequences of asexual mutation accumulation is critical to determining
whether the absence of sex inevitably leads to mutation meltdown and lineage extinction.
126
ACKNOWLEDGEMENTS
The National Science Foundation (NSF: MCB – 1122176; DEB – 1310825) and the Iowa
Academy of Sciences (ISF #13-10) funded this research. I thank Nicole Enright ad JD Woodell
for help with aquatic respirometry. I thank Claire Tucci, Meagan Luse, Nikhil Puttagunta for
their help with the behavioral assay.
127
REFERENCES
Abele, D., Heise, K., Portner, H. O., & Puntarulo, S. (2002). Temperature-dependence of
mitochondrial function and production of reactive oxygen species in the intertidal mud
clam Mya arenaria. Journal of Experimental Biology, 205, 1831-1841.
Ballard, J. W. O. (2000). Comparative genomics of mitochondrial DNA in members of the
Drosophila melanogaster subgroup. Journal of Molecular Evolution, 51, 48-63.
Barja, G., & Herrero, A. (2000). Oxidative damage to mitochondrial DNA is inversely related to
maximum life span in the heart and brain of mammals. Faseb Journal, 14, 312-318.
Barr, C. M., Neiman, M., & Taylor, D. R. (2005). Inheritance and recombination of
mitochondrial genomes in plants, fungi and animals. New Phytologist, 168, 39-50.
Barreto, F. S., & Burton, R. S. (2013). Elevated oxidative damage is correlated with reduced
fitness in interpopulation hybrids of a marine copepod. Proceedings of the Royal Society
of London B, 280, 20131521.
Bazin, E., Glemin, S., & Galtier, N. (2006). Population size does not influence mitochondrial
genetic diversity in animals. Science, 312, 570-572.
Bell, G. (1982). The masterpiece of nature : the evolution and genetics of sexuality. Berkeley:
University of California Press.
Bergstrom, C. T., & Pritchard, J. (1998). Germline bottlenecks and the evolutionary maintenance
of mitochondrial genomes. Genetics, 149, 2135-2146.
Blier, P. U., Dufresne, F., & Burton, R. S. (2001). Natural selection and the evolution of
mtDNA-encoded peptides: evidence for intergenomic co-adaptation. Trends in Genetics,
17, 400-406.
128
Burton, R. S., & Barreto, F. S. (2012). A disproportionate role for mtDNA in DobzhanskyMuller incompatibilities? Molecular Ecology, 21, 4942-4957.
Callegari, S., McKinnon, R. A., Andrews, S., & de Barros Lopes, M. A. (2011). The MEF2 gene
is essential for yeast longevity, with a dual role in cell respiration and maintenance of
mitochondrial membrane potential. FEBS Lett, 585, 1140-1146.
Chance, B., & Williams, G. R. (1955). Respiratory enzymes in oxidative phosphorylation. I.
Kinetics of oxygen utilization. Journal of Biological Chemistry, 217, 383-393.
Charlesworth, B. (2009). Fundamental concepts in genetics: effective population size and
patterns of molecular evolution and variation. Nature Reviews Genetics, 10, 195-205.
Chen, J. H., Hales, C. N., & Ozanne, S. E. (2007). DNA damage, cellular senescence and
organismal ageing: causal or correlative? Nucleic Acids Research, 35, 7417-7428.
Crow, J. F., & Kimura, M. (1970). An introduction to population genetics theory. New York:
Harper & Row.
Cutter, A. D., & Payseur, B. A. (2003). Rates of deleterious mutation and the evolution of sex in
Caenorhabditis. Journal of Evolutionary Biology, 16, 812-822.
Das, J., Miller, S. T., & Stern, D. L. (2004). Comparison of diverse protein sequences of the
nuclear-encoded subunits of cytochrome C oxidase suggests conservation of structure
underlies evolving functional sites. Molecular Biology and Evolution, 21, 1572-1582.
DiMauro, S., & Schon, E. A. (2001). Mitochondrial DNA mutations in human disease. American
Journal of Medical Genetics, 106, 18-26.
Dowling, D. K. (2014). Evolutionary perspectives on the links between mitochondrial genotype
and disease phenotype. Biochimica Biophysica Acta, 1840, 1393-1403.
129
Dowling, D. K., Friberg, U., & Lindell, J. (2008). Evolutionary implications of non-neutral
mitochondrial genetic variation. Trends in Ecology and Evolution, 23, 546-554.
Dybdahl, M. F., & Kane, S. L. (2005). Adaptation vs. phenotypic plasticity in the success of a
clonal invader. Ecology, 86, 1592-1601.
Dybdahl, M. F., & Lively, C. M. (1995). Host-parasite interactions - infection of common clones
in natural populations of a freshwater snail (Potamopyrgus antipodarum). Proceedings of
the Royal Society of London Series B-Biological Sciences, 260, 99-103.
Ellison, C. K., & Burton, R. S. (2006). Disruption of mitochondrial function in interpopulation
hybrids of Tigriopus californicus. Evolution, 60, 1382-1391.
Fei, G., et al. (2008). HSP70 reduces chronic hypoxia-induced neural suppression via regulating
expression of syntaxin. Genes and Proteins Underlying Microbial Urinary Tract Virulence, 605,
35-40.
Frank, S. A., & Hurst, L. D. (1996). Mitochondria and male disease. Nature, 383, 224.
Garner, D. L., & Thomas, C. A. (1999). Organelle-specific probe JC-1 identifies membrane
potential differences in the mitochondrial function of bovine sperm. Molecular
Reproduction and Development, 53, 222-229.
Heise, K., Puntarulo, S., Portner, H. O., & Abele, D. (2003). Production of reactive oxygen
species by isolated mitochondria of the Antarctic bivalve Laternula elliptica (King and
Broderip) under heat stress. Comp Biochem Physiol C Toxicol Pharmacol, 134, 79-90.
Henry, L., Schwander, T., & Crespi, B. J. (2012). Deleterious mutation accumulation in asexual
Timema stick insects. Molecular Biology and Evolution, 29, 401-408.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection.
Genetical Research, 8, 269-294.
130
Hollister, J. D., Greiner, S., Wang, W., Wang, J., Zhang, Y., Wong, G. K., . . . Johnson, M. T.
(2015). Recurrent loss of sex is associated with accumulation of deleterious mutations in
Oenothera. Molecular Biology and Evolution, 32, 896-905.
Horandl, E., & Hojsgaard, D. (2012). The evolution of apomixis in angiosperms: A reappraisal.
Plant Biosystematics, 146, 681-693.
Hori, A., Yoshida, M., Shibata, T., & Ling, F. (2009). Reactive oxygen species regulate DNA
copy number in isolated yeast mitochondria by triggering recombination-mediated
replication. Nucleic Acids Research, 37, 749-761.
Hudson, J. (1983). Oxygen consumption of the mollusc Potamopyrgus antipodarum
in relation to habitat. Ph.D. Dissertation. University of Canterbury Library.
Jacobsen, R., & Forbes, V. E. (1997). Clonal variation in life­history traits and feeding
rates in the gastropod, Potamopyrgus antipodarum: performance across a salinity
gradient. Functional Ecology, 11, 260-267.
Jokela, J., Lively, C. M., Fox, J. A., & Dybdahl, M. F. (1997) Flat reaction norms and "frozen"
phenotypic variation in clonal snails (Potamopyrgus antipodarum). Evolution, 51, 11201129.
Kaiser, V. B., & Charlesworth, B. (2009). The effects of deleterious mutations on evolution in
non-recombining genomes. Trends in Genetics, 25, 9-12
Kaiser, V. B., & Charlesworth, B. (2010). Muller's ratchet and the degeneration of the
Drosophila miranda neo-Y chromosome. Genetics, 185, 339-348.
Kelly, R. D., Mahmud, A., McKenzie, M., Trounce, I. A., & St John, J. C. (2012). Mitochondrial
DNA copy number is regulated in a tissue specific manner by DNA methylation of the
nuclear-encoded DNA polymerase gamma A. Nucleic Acids Research, 40, 10124-10138.
131
Lee-Yaw, J. A., Jacobs, C. G., & Irwin, D. E. (2014). Individual performance in relation to
cytonuclear discordance in a northern contact zone between long-toed salamander
(Ambystoma macrodactylum) lineages. Molecular Ecology, 23, 4590-4602.
Liu, K. Z., Schultz, C. P., Johnston, J. B., Lee, K., & Mantsch, H. H. (1997). Comparison of
infrared spectra of CLL cells with their ex vivo sensitivity (MTT assay) to chlorambucil
and cladribine. Leukemia Research, 21, 1125-1133.
Lynch, M., Bürger, R., Butcher, D., & Gabriel, W. (1993). The mutational meltdown in asexual
populations. Journal of Heredity, 84, 339-344.
MacKinnon, J. G. & White H. (1985). Some heteroskedasticity-consistent covariance matrix
estimators with improved finite sample properties. Journal of Econometrics, 29, 305-325.
Marshall, H. D., Coulson, M. W., & Carr, S. M. (2009). Near neutrality, rate heterogeneity, and
linkage govern mitochondrial genome evolution in Atlantic cod (Gadus morhua) and
other gadine fish. Molecular Biology and Evolution, 26, 579-589.
Matsushima, Y., Goto, Y., & Kaguni, L. S. (2010). Mitochondrial Lon protease regulates
mitochondrial DNA copy number and transcription by selective degradation of
mitochondrial transcription factor A (TFAM). Proceedings of the National Academy of
Sciences USA, 107, 18410-18415.
Maynard Smith, J. (1978). The evolution of sex. Cambridge: Cambridge University Press.
Meiklejohn, C. D., Holmbeck, M. A., Siddiq, M. A., Abt, D. N., Rand, D. M., & Montooth, K. L.
(2013). An incompatibility between a mitochondrial tRNA and its nuclear-encoded tRNA
synthetase compromises development and fitness in Drosophila. PLoS Genetics, 9,
e1003238.
132
Miller, F. J., Rosenfeldt, F. L., Zhang, C., Linnane, A. W., & Nagley, P. (2003). Precise
determination of mitochondrial DNA copy number in human skeletal and cardiac muscle
by a PCR-based assay: lack of change of copy number with age. Nucleic Acids Research,
31, e61.
Moraes, C. T., Shanske, S., Tritschler, H. J., Aprille, J. R., Andreetta, F., Bonilla, E., . . .
DiMauro, S. (1991). mtDNA depletion with variable tissue expression: a novel genetic
abnormality in mitochondrial diseases. American Journal of Human Genetics, 48, 492501.
Muller, H. J. (1964). The relation of recombination to mutational advance. Mutational Research
106, 2-9.
Neiman, M., Hehman, G., Miller, J. T., Logsdon, J. M., Jr., & Taylor, D. R. (2010). Accelerated
mutation accumulation in asexual lineages of a freshwater snail. Molecular Biology and
Evolution, 27, 954-963.
Neiman, M., Larkin, K., Thompson, A. R., & Wilton, P. (2012). Male offspring production by
asexual Potamopyrgus antipodarum, a New Zealand snail. Heredity, 109, 57-62.
Neiman, M., & Linksvayer, T. A. (2006). The conversion of variance and the evolutionary
potential of restricted recombination. Heredity, 96, 111-121.
Neiman, M., & Lively, C. M. (2004). Pleistocene glaciation is implicated in the
phylogeographical structure of Potamopyrgus antipodarum, a New Zealand snail.
Molecular Ecology, 13, 3085-3098.
Neiman, M., Paczesniak, D., Soper, D. M., Baldwin, A. T., & Hehman, G. (2011). Wide
variation in ploidy level and genome size in a New Zealand freshwater snail with
coexisting sexual and asexual lineages. Evolution, 65, 3202-3216.
133
Neiman, M., & Taylor, D. R. (2009). The causes of mutation accumulation in mitochondrial
genomes. Proceedings of the Royal Society of London B, 276, 1201-1209.
Nicholls, D. G., & Ward, M. W. (2000). Mitochondrial membrane potential and neuronal
glutamate excitotoxicity: mortality and millivolts. Trends in Neuroscience, 23, 166-174.
Paczesniak, D., Jokela, J., Larkin, K. & Neiman, M. (2013). Discordance between nuclear and
mitochondrial genomes in sexual and asexual lineages of the freshwater snail Potamopyrgus
antipodarum. Molecular Ecology, 22, 4695-4710.
Pichaud, N., Ballard, J. W., Tanguay, R. M., & Blier, P. U. (2013). Mitochondrial haplotype
divergences affect specific temperature sensitivity of mitochondrial respiration. Journal
of Bioenergetics and Biomembranes, 45, 25-35.
Pike, T. W., Blount, J. D., Bjerkeng, B., Lindstrom, J., & Metcalfe, N. B. (2007). Carotenoids,
oxidative stress and female mating preference for longer lived males. Proceedings of the
Royal Society of London B, 274, 1591-1596.
Rand, D. M., Haney, R. A., & Fry, A. J. (2004). Cytonuclear coevolution: the genomics of
cooperation. Trends in Ecology and Evolution, 19, 645-653.
Rand, D. M., & Kann, L. M. (1996). Excess amino acid polymorphism in mitochondrial DNA:
contrasts among genes from Drosophila, mice, and humans. Molecular Biology and
Evolution, 13, 735-748.
Rawson, P. D., & Burton, R. S. (2002). Functional coadaptation between cytochrome c and
cytochrome c oxidase within allopatric populations of a marine copepod. Proceedings of
the National Academy of Sciences USA, 99, 12955-12958.
134
Sloan, D. B., Triant, D. A., Wu, M., & Taylor, D. R. (2014). Cytonuclear interactions and
relaxed selection accelerate sequence evolution in organelle ribosomes. Molecular
Biology and Evolution, 31, 673-682.
Sugrue, M. M., & Tatton, W. G. (2001). Mitochondrial membrane potential in aging cells.
Biological Signals and Receptors, 10, 176-188.
Tucker, A. E., Ackerman, M. S., Eads, B. D., Xu, S., & Lynch, M. (2013). Population-genomic
insights into the evolutionary origin and fate of obligately asexual Daphnia pulex.
Proceedings of the National Academy of Sciences USA,, 110, 15740-15745.
Van den Bogert, C., De Vries, H., Holtrop, M., Muus, P., Dekker, H. L., Van Galen, M. J., . . .
Taanman, J. W. (1993). Regulation of the expression of mitochondrial proteins:
relationship between mtDNA copy number and cytochrome-c oxidase activity in human
cells and tissues. Biochimica Biophysica Acta, 1144, 177-183.
Voigt-Zielinski, M. L., Piwczynski, M., & Sharbel, T. F. (2012). Differential effects of
polyploidy and diploidy on fitness of apomictic Boechera. Sexual Plant Reproduction, 25,
97-109.
Williams, G. C. (1975). Sex and evolution. Princeton: Princeton University Press.
Xia, X. (1998). The rate heterogeneity of nonsynonymous substitutions in mammalian
mitochondrial genes. Molecular Biology and Evolution, 15, 336-344.
Zachar, N., & Neiman, M. (2013). Profound effects of population density on fitness-related traits
in an invasive freshwater snail. PLoS ONE, 8, e80067.
Zeyl, C., Andreson, B., & Weninck, E. (2005). Nuclear-mitochondrial epistasis for fitness in
Saccharomyces cerevisiae. Evolution, 59, 910-914.
135
Zhang, F., & Broughton, R. E. (2013). Mitochondrial-nuclear interactions: compensatory
evolution or variable functional constraint among vertebrate oxidative phosphorylation
genes? Genome Biology and Evolution, 5, 1781-1791.
136
FIGURES
Figure 4-1. MtDNA copy number variation in six asexual lineages of P. antipodarum, rank
ordered in terms of increasing mean.
Quantitative PCR estimates of cytB copy number relative to a putatively single-copy nuclear
gene, rad21. MtDNA copy number was significantly different across complexes, which appears
to be largely driven by lineage Gr5. Shared lowercase letters indicate p > 0.05 for pairwise t-tests
corrected for multiple comparisons using the Holm procedure. N = 10 for all lineages.
137
Figure 4-2. Mitochondrial membrane potential in mitochondrial extracts of six lineages of
P. antipodarum.
Estimate of mitochondrial membrane potential using the ratio of red to green JC-1 dye
fluorescence. Shared lowercase letters indicate p > 0.05 for pairwise Welch’s t-tests, corrected
using the Holm procedure for multiple comparisons. N = 10 for all lineages.
138
Figure 4-3. Interaction plot between O2 consumption residuals, temperature, and snail
lineage.
Lines indicate best-fit linear regression of O2 consumption across temperature pairs (e.g., 16ºC –
22ºC). for seven asexual lineages of P. antipodarum. O2 consumption was measured for 10
individual snails at each temperature for each lineage. O2 consumption is affected by temperature
as well as by the interaction between temperature and lineage. Within temperatures, lineage had
a significant effect on O2 consumption at 22ºC (one-way ANOVA F (6, 58) = 3.002, p = 0.0127)
and at 30ºC (one-way ANOVA F (6, 62) = 3.823, p = 0.00263).
139
140
Figure 4-4. Comparisons of mitochondrial functional assays.
Spearman’s rank correlations are displayed for all pairwise comparisons of mitochondrial
functional assays. Best-fit linear regression lines (black) are only shown for those comparisons
that exhibited p < 0.05. a) mtDNA copy number vs. JC-1 assay (Spearman's rho = -0.1, p =
0.823). b) mtDNA copy number vs. log-transformed MTT reduction value (Spearman's rho = 0.33, p = 0.42). c) mtDNA copy number vs. square-root transformed O2 consumption residuals,
16ºC (Spearman's rho = 0.03, p = 0.957). d) mtDNA copy number vs. square-root transformed
O2 consumption residuals, 22ºC (Spearman's rho = 0.03, p = 0.957). e) mtDNA copy number vs.
square-root transformed O2 consumption residuals, 30ºC (Spearman's rho = 0.43, p = 0.396). f)
JC-1 assay vs. log-transformed MTT reduction value (Spearman's rho = 0.81, p = 0.0149). g) JC1 assay vs. square-root transformed O2 consumption residuals, 16ºC (Spearman's rho = -0.43, p =
0.397). h) JC-1 assay vs. square-root transformed O2 consumption residuals, 22ºC (Spearman's
rho = -0.77, p = 0.0724). i) JC-1 assay vs. square-root transformed O2 consumption residuals,
30ºC (Spearman's rho = 0.2, p = 0.704). j) log-transformed MTT reduction value vs. square-root
transformed O2 consumption residuals, 16ºC (Spearman's rho = 0.14, p = 0.787). k) logtransformed MTT reduction value vs. square-root transformed O2 consumption residuals, 22ºC
(Spearman's rho = -0.89, p = 0.0188). l) log-transformed MTT reduction value vs. square-root
transformed O2 consumption residuals, 30ºC (Spearman's rho = -0.09, p = 0.872). m) square-root
transformed O2 consumption residuals, 16ºC vs. square-root transformed O2 consumption
residuals, 22ºC (Spearman's rho = -0.36, p = 0.432). n) square-root transformed O2 consumption
residuals, vs. square-root transformed O2 consumption residuals, 30ºC (Spearman's rho = -0.14, p
= 0.76). o) square-root transformed O2 consumption residuals, 22ºC vs. square-root transformed
O2 consumption residuals, 30ºC (Spearman's rho = -0.04, p = 0.939).
141
SUPPLEMENTARY MATERIAL
Figure 4-S1. Variation in righting time in response to heat stress in P. antipodarum.
Ten asexual lineages (N = 10 snails per lineage) were incubated at either 16ºC, 22ºC, or 30ºC for
1 hour prior to assay to test for behavioral responses to elevated temperature. Temperature
(Kruskal-Wallis, χ2 = 14.218, df = 2, p = 0.00082) and lineage (Kruskal-Wallis, χ2 = 122.64, df =
12, p < 2.2 x 10-16) both significantly affected righting time, indicating that elevated temperature
can be used to induce stress in P. antipodarum.
142
143
Figure 4-S2. Relationship between snail wet mass and oxygen consumption in P.
antipodarum.
There was a significant positive correlation between snail wet mass and O2 consumption
(Spearman’s rho = 0.2, p = 0.0044).
144
Table 4-S1. Lineages and samples sizes used to compare mitochondrial function in P.
antipodarum.
Lineage
Reproductive
Lake of Origin
Name
mtDNA copy #
JC-1
MTT§
Mode
Righting
Oxygen
Ability
consumption
B52
Alexandrina
Asexual
3
10
8x3
–
30
C7
Alexandrina
Asexual
5
–
9x3
30
30
DenA
–
Asexual
-
10
–
–
–
Gn5
Gunn
Asexual
–
–
–
–
30
Gr5
Grasmere
Asexual
5
10
6x3
–
30
Hr4
Heron
Asexual
5
10
7x3
30
30
KnS12 *
Kaniere
Mixed
–
10
7x3
30
Ta10
Ta Anau
Asexual
8
8
7x3
30
30
Wk4
Waikeremoana
Asexual
5
10
7x3
–
30
Y2 †
Alexandrina
Sexual
–
10
6x3
30
–
Hr74
Heron
Asexual
–
–
–
30
–
Po74
Poerua
Asexual
–
–
–
30
–
Kn4
Kaniere
Asexual
–
–
–
30
–
Gn6
Gunn
Asexual
–
–
–
30
–
IaSF12*
Ianthe
Mixed
–
–
–
30
–
Po 95
Poerua
Asexual
–
–
–
30
–
SeF12*
Selfe
Mixed
–
–
–
30
–
145
Table 4-S1 – continued
Ri6
Rotoiti
Asexual
–
–
–
30
–
If present, numbers inside cells represent sample sizes for a particular assay; otherwise “–“ denotes missing data
from that lineage for the assay in question.
* – Field-collected population that was not used in tests for among-lineage variation in mitochondrial function.
† – Inbred sexual lineage that was not used in tests for among-lineage variation in mitochondrial function.
§ – MTT reduction assay required 3 whole snails per sample replicate.
146
CHAPTER 5: INTRASPECIFIC DIVERGENCE FOR METABOLIC
FUNCTION IN POTAMOPYRGUS ANTIPODARUM, AN EMERGING
MODEL FOR THE EVOLUTIONARY MAINTENANCE OF SEX
ABSTRACT
Cellular respiration is an essential component of eukaryotic health, but little is known about how
functional variation for respiratory capacity contributes to meaningful phenotypic variation in
natural populations. The mitochondrial and nuclear-encoded genes that together are responsible
for respiratory phenotype are generally expected to experience relatively intense purifying
selection, with the implications that most variation in these genes will be harmful. Because
sexual reproduction is thought to be required for effective clearance of harmful variation,
comparing respiratory-related phenotypes in otherwise similar sexual vs. asexual lineages offers
a unique glimpse into the evolutionary processes responsible for generating phenotypic variation.
Here, I used Potamopyrgus antipodarum, a New Zealand freshwater snail featuring multiple
separately derived asexual lineages that frequently coexist with closely related sexual lineages, to
quantify and compare phenotypic variation for mitochondrial function in sexual vs. asexual
lineages. My common garden study revealed extensive across-lineage variation in oxygen
consumption while exposed to heat stress. I also found evidence for a significant role of lake of
origin in this heat stress response, providing an initial line of evidence that population structure
and/or local adaptation influences variation in important traits in natural populations of P.
antipodarum.
147
INTRODUCTION
The production of ATP via cellular respiration is a critical component of eukaryotic function and
fitness (Chen et al. 2007, Pike et al. 2007, Barreto and Burton 2013, Dowling 2014), which likely
explains why the highly conserved components of the oxidative phosphorylation (OXPHOS)
pathway appear to be evolving under purifying selection in a wide variety of eukaryotic taxa
(Blier et al. 2001, Montooth et al. 2009, Zhang and Broughton 2013, Chapter 2, 3). By this logic,
one would expect that the genes underlying OXPHOS function would tend to harbor relatively
low intraspecific variation. Counter to these expectations, these genes are often polymorphic
within species, to the extent that divergent populations can rapidly evolve mitonuclear
incompatibilities (Ellison and Burton 2006), and explaining the wide application of the
mitochondrially encoded OXPHOS subunits as a means of delineating animal taxa (e.g.,
barcoding, Hebert et al. 2003). While the relatively high mutation rate (Denver et al. 2000) and
relatively low effective population size (Bergstron and Pritchard 1998, Normark and Moran
2000) expected to characterize animal mitochondrial genomes certainly contributes to this rapid
generation of intraspecific variation, how this genetic variation affects phenotypic variation for
metabolic function - and indeed, how mitochondrial function is maintained over evolutionary
time despite mutational pressure (Lynch et al. 1993)- remains unclear.
One plausible explanation for the continued maintenance of mitochondrial function is via
its genetic partners in the nuclear genome: because mitochondrial genes function as subunits of
massive protein complexes, interacting complex partners encoded by the nuclear genome may be
able to compensate for mutational changes in other subunits. Compensatory evolution is thought
to be particularly important for the maintenance of function in OXPHOS complexes I, III, IV,
and V, which all include both mitochondrial and nuclear-encoded subunits (Blier et al. 2001).
148
Cooperation between these subunits is thought to be critical to proper mitochondrial function and
ATP production (Rand et al. 2004), and co-introgression of nuclear- and mitochondrially
encoded allelic components of OXPHOS subunits from Drosophila yakuba into Drosophila
santomea (Beck et al. 2015) and hybrid breakdown mapped to the mitochondrial genome in
Tigriopus californicus (Ellison and Burton 2008) both support this hypothesis. The relatively
high mutation rate and small effective population size of animal mitochondrial genomes likely
shifts the burden of compensatory evolution to the nuclear-encoded subunits because selective
interference from physically linked sites (Hill and Robertson 1966) and Muller’s ratchet (Muller
1964, Lynch et al. 1993) are expected to decrease the efficacy of purifying selection in the
mitochondrial genome relative to the nuclear genome (reviewed in Neiman and Taylor 2009). By
this logic, if compensatory evolution is operating to maintain mitochondrial function, one should
expect to see high rates of evolution in the nuclear-encoded OXPHOS subunits that interact with
rapidly evolving mitochondrial counterparts (Sloan et al. 2013), leading to the downstream
prediction of rapid divergence in mitochondrial function between reproductively isolated
populations.
A non-mutually exclusive explanation for the maintenance of relatively high levels of
genetic variation in the mitochondrial genome is a scenario whereby mitochondrial function is
context specific. In this situation, across-population variation in environmental conditions as well
as fluctuating environments within populations might contribute to the intraspecific maintenance
of genetic variation (e.g., Anrqvist et al. 2010). In particular, spatial and temporal variation in
selection acting on mitochondrial variation could maintain allele frequencies at intermediate
frequencies, such that polymorphisms experiencing fluctuating selection are unlikely to rise to
fixation.
149
It is also important to consider the possibility that mitochondrial polymorphism might be
maintained simply because these variable sites do not in fact confer substantial phenotypic
consequences. Under this scenario, genetic and phenotypic polymorphism for mitochondrial
function should be especially high in taxa with relatively low metabolic requirements, and vice
versa. This hypothesis, under the presumption that flight is energetically expensive (Reinhold
1999), has been put forth as a potential explanation for evidence for surprisingly effective
purifying selection in the Drosophila melanogaster mitochondrial genomes (Montooth et al.
2009). By this logic, low metabolic requirements (i.e., relatively sedentary lifestyle, abundant
resources, etc.) in organisms like the New Zealand freshwater snail Potamopyrgus antipodarum
could explain the maintenance of high nonsynonymous mitochondrial polymorphism relative to
other taxa (Chapter 2, 3).
Taxa that are polymorphic for sexual reproduction provide a powerful means by which to
evaluate the relative contributions of these different evolutionary processes to the maintenance of
functional phenotypic variation because sex is expected to increase effective population size and,
thus, the efficacy of selection against harmful variants. There are at least two important
implications of these expected connections between reproductive mode and the efficacy of
selection for mitochondrial function in sexual vs. asexual lineages. First, asexual lineages should
experience a higher rate of harmful mutation accumulation in their mitochondrial and nuclear
genomes (Normark and Moran 2000). Second, the rate of compensatory nuclear evolution should
be slowed by the reduction of the efficacy of selection in asexuals (Neiman and Taylor 2009).
Together, the accelerated mutation accumulation and the reduced rate of compensatory mutation
should result in relatively low mitochondrial function in asexuals relative to sexual counterparts.
Potamopyrgus antipodarum, a New Zealand freshwater snail, offers a particularly powerful
150
means of addressing these questions, as it features multiple separately derived asexual lineages
that frequently coexist with otherwise similar sexual lineages (Lively 1987)
. Asexual P. antipodarum are typically locally derived from sympatric sexual populations
(Paczesniak et al. 2013), such that distinct asexual lineages of P. antipodarum represent so-called
“natural experiments” into the consequences of asexuality for mitochondrial function.
Here, I use these strengths of the P. antipodarum system to compare mitochondrial
function across New Zealand lake populations and reproductive modes using aquatic
respirometry with the goal of providing critical first steps towards testing a central hypothesis for
the commonness of sex and/or the seemingly inevitable extinction of asexual lineages: that
mutation accumulation in asexual lineages will translate into phenotypic decline.
MATERIALS & METHODS
Field collections of P. antipodarum
Snails were collected using kick nets from the shallow regions of seven South Island New
Zealand lakes in January 2015 (Figure 5-1). I chose these lakes to chosen to maximize the
potential for within- and between-lake comparisons across reproductive mode (i.e., all lakes were
known from previous studies to harbor both sexual and asexual P. antipodarum; Neiman et al.
2011, Paczesniak et al. 2013, Bankers et al. in review), although reproductive mode was not
determined until completion of aquatic respirometry. Upon arrival, snails were housed at 16ºC on
a 18hr light/6 hr dark schedule and fed Spirulina algae 3x per week as described in Zachar and
Neiman (2013). I arbitrarily selected 16 adult female snails from each lake collection and
individually isolated each female/sample in 0.5 L glass containers with 300ml carbon-filtered
151
H2O. I changed the water in each container once per week. Snails that died during the course of
the experiment were replaced to the extent possible from source collections.
Aquatic respirometry
Because oxygen becomes limiting to ectotherms under elevated temperatures (Abele et al. 2007)
and because P. antipodarum demonstrates signs of stress at elevated (~30ºC) temperatures (e.g.,
fecundity: Dybdahl and Kane 2005, elevated oxygen consumption and decreased righting ability:
Chapter 4), I measured oxygen consumption for each of the 16 individual snails from each of
these nine lakes at three different water temperatures that I chose to reflect ambient (16ºC;
temperature of the room in which snails were housed, set to reflect New Zealand lake water
temperature), intermediate (22ºC), and "high" (30ºC) thermal conditions. I followed the protocols
for aquatic respirometry as described in Chapter 3, including daily calibration of oxygen
concentration with a 1302 Clark-type electrode. I incubated individual snails at the
corresponding test temperature for 1 hour prior to measurement and measured oxygen
consumption for 1 hour following incubation. Following measurement, I blotted each snail dry
and obtained wet mass for each snail at each temperature using (Denver Instruments Cubis
Analytical Balance). Mean snail mass was calculated from the three different temperature trials
for each snail.
Determination of reproductive mode
Sexual P. antipodarum are diploid and asexual P. antipodarum are polyploid (triploid or
tetraploid; Wallace 1992, Neiman et al. 2011), allowing me to use flow cytometry to determine
nuclear DNA content and, thus, reproductive mode of individual P. antipodarum. After obtaining
152
oxygen consumption data for each snail at all three temperatures, I dissected head tissue from
each individual, flash froze this tissue in liquid nitrogen, and stored the frozen tissue at -80ºC
until flow cytometry. I then homogenized head tissue in DAPI solution, filtered this solution
through a 30 µm filter, and ran the solution on a Becton-Dickinson FACS Aria II following the
protocol (including a chicken red blood cell standard) outlined in Krist et al. (2014). I produced
flow cytometry data from 63 snails from seven lakes and found 34 diploid sexuals and 29
polyploid asexuals (Table 5-S1).
Statistical analyses
Because snail wet mass is correlated with O2 consumption in P. antipodarum (Chapter 4), and
because mean snail wet mass differed across lake populations (Kruskal-Wallis χ2 = 25.255, p =
0.00141), it was necessary to correct for snail wet mass prior to comparing O2 consumption
across treatments. To do so, I first tested the assumptions of normality for snail wet-mass
(Shapiro-Wilks W = 0.912, p = 8.814 x 10-9), O2 consumption at 16ºC (Shapiro-Wilks W =
0.943, p = 0.00686), 22ºC (Shapiro-Wilks W = 0.983, p = 0.535), and 30ºC (Shapiro-Wilks W =
0.94, p = 0.00951). For those parameters that did not fit the assumptions of normality, I
implemented the box-cox set of data transformations with lambda ranging from -2 to 2 (Box and
Cox 1964) and tested the normality of the transformed data set. Following transformation, logtransformed wet mass (lambda = 0, Shapiro-Wilks W = 0.9643, p = 0.0583), square-roottransformed O2 consumption at 16ºC (lambda = 0.5, Shapiro-Wilks W = 0.984, p = 0.588), and
power-transformed O2 consumption at 30ºC (lambda = -2, Shapiro-Wilks W = 0.961, p = 0.0727)
did not deviate from the assumptions of normality, along with O2 consumption at 22ºC. I next
calculated wet mass-corrected O2 consumption by calculating the residuals from a linear
153
regression with wet mass as the independent variable and O2 consumption as the dependent each
temperature treatment (regressions shown in Figure 5-S1). I then used these residuals as the
dependent variable in a three-way ANOVA to test whether the random factor of lake of origin or
the fixed factors of temperature and reproductive mode, or an interaction between any of these
factors. Because there was only a single lake for which both sexual and asexual individuals were
sampled, I performed a two-way ANOVA excluding reproductive mode as a factor. I also
performed a series of one-way ANOVAs within each temperature (with lake of origin as a fixed
factor and reproductive mode as a fixed factor) and to address lake-specific O2 consumption
responses. Because Lake Alexandrina contained both sexual and asexual samples, I was able to
perform a two-way ANOVA to test for any within-lake effects of reproductive mode and
temperature (with temperature and reproductive mode as fixed factors). For the other six lakes, I
performed one-way ANOVAs within each lake with temperature as a fixed factor. Results from
ANOVA comparisons are shown in Table 5-S2. All statistical comparisons were performed in R
v. 3.2.4 (R Core Team 2016).
RESULTS
To test whether organismal O2 consumption is affected by temperature, lake of origin, and
reproductive mode, I compared O2 residuals (see Methods) using a three-way ANOVA with lake
of origin as a random factor, temperature as a fixed factor, and reproductive mode as a fixed
factor. I found that there was a significant effect of the interaction between lake and temperature
on organismal oxygen consumption (F6, 161 = 2.433, p = 0.0281, Figure 5-1, Table 5-S2).
Because there was no effect of reproductive mode on O2 consumption in this test and only one
lake had both sexual and asexual samples, I performed a two-way ANOVA, excluding
154
reproductive mode as a factor. I found that the interaction of temperature and lake of origin again
had a significant effect on O2 consumption. To investigate this interaction effect, I compared O2
residuals within each temperature using a two-way ANOVA with lake of origin and reproductive
mode as fixed factors. I found that lake of origin significantly affected O2 residuals at 22ºC (F6, 53
= 3.355, p = 0.00697) but not at 16ºC (F 6, 54 = 1.122, p = 0.362) or 30ºC (F5, 47) = 1.555, p =
0.191). Reproductive mode did not affect O2 residuals at any temperature treatment (16ºC: F1, 53
= 1.530, p = 0.222, 22ºC: F 1, 54 = 0.013, p = 0.911, 30ºC: F1, 47 = 0.192, p = 0.663). Finally, I
tested the effect of temperature on O2 residuals within each lake using a two-way ANOVA with
temperature and reproductive more being fixed factors for Lake Alexandrina and one-way
ANOVAs for the remaining lakes. I found that temperature significantly altered O2 residuals in
Lake Alexandrina (F1, 38 = 53.711, p = 8.87 x 10-9), Lake Clearwater (F1, 9 = 5.337, p = 0.046),
and Lake Selfe F1, 5 = 9.723, p = 0.026). Reproductive mode did not affect O2 residuals in any
Lake Alexandrina (F1, 38 = 0.171, p = 0.681). The common garden method I employed here
indicates that even though snails are housed in identical conditions, they still exhibit lakespecific phenotypes in terms of organismal O2 consumption. The observed differences in O2
consumption across temperatures indicate that P. antipodarum possess some ability to regulate
O2 consumption in response to the current environment.
155
DISCUSSION
Phenotypic variation for mitochondrial function depends upon lake of origin and
environment
I employed a common garden approach to compare organismal O2 consumption in response to
heat stress in 64 individual snails collected from seven different New Zealand lakes (Figure 5-1).
I found that both temperature treatment and lake of origin significantly affected organismal O2
consumption in P. antipodarum. These results are consistent with previous experiments
demonstrating variation for mitochondrial function between lab-reared asexual lineages (Chapter
4), suggesting that there exists substantial genetic variation for mitochondrial function (measured
here as O2 consumption in response to heat stress) in P. antipodarum. The substantial differences
in O2 consumption across temperature treatments indicate that P. antipodarum is capable of
altering its metabolic rate in response to environmental stimuli. The dependence of
mitochondrial function on both genetic and environmental factors has been observed in other
systems (e.g., Willett and Burton 2003, Dowling et al. 2007, Arnqvist et al. 2010, Hoekstra et al.
2013), suggesting that genotype-by-environment effects on mitochondrial function may be
widespread amongst animals. Genotype x environment interactions are expected to reduce the
efficacy of selection, as variation is not reliably passed on from parent to offspring (Lande and
Shannon 1996). This pattern of phenotypic variation is consistent with the pattern of molecular
variation in P. antipodarum mitochondrial genomes: variation in the mitochondrial genome is
structured by phylogenetic history (Chapter 2, Figure 2-1), such that individuals from the same
lake tend to share mitochondrial haplotypes more often than individuals from different lakes
(Neiman and Lively 2004, Paczesniak et al. 2013). Potamopyrgus antipodarum mitochondrial
genomes tend to harbor substantial levels of nonsynonymous variation (Chapter 3), which is in
156
stark contrast to animals known to have large effective population sizes (e.g., Montooth et al.
2009). Together, these data suggest that phenotypic variation is structured by lake population,
meaning that future analyses must account for lake-effects when measuring metabolic function in
P. antipodarum.
While determining optimal mitochondrial function in laboratory experiments likely does
not wholly reflect how variation interfaces with the organism’s natural environment, these data
can still provide clues as to the source of variation. In particular, surveys of P. antipodarum
mitochondrial genome diversity indicate that one haplotype – the so-called haplotype “1A” – is
to be shared across many lakes and is especially common amongst New Zealand asexuals
(Neiman and Lively 2004, Paczesniak et al. 2013). Evaluating the mitochondrial function of this
haplotype under common garden conditions would likely provide valuable insight into
phenotypic variation for mitochondrial function in P. antipodarum.
These data might also help illuminate the evolutionary mechanisms underlying the global
invasion of P. antipodarum (Levri et al. 2007, Alonso and Castro- Díez 2012). Invasive P.
antipodarum all seem to share a more recent common ancestor with snails from the North Island
of New Zealand than with snails from other parts of the native range (Städler et al. 2005). This
observation suggests that some lineages of P. antipodarum are more well-equipped to colonize
novel environments than others, raising the question of whether what appears to be a nonrandom contribution of lineage source to invasion success could be related to the phenotypic
variation we observed here (Keller and Taylor 2008; e.g., Neiman and Krist 2016). In particular,
it would be relevant to determine whether invasive snails have distinct phenotypic profiles from
those lineages that remain in their native New Zealand lakes. Neiman and Krist (2016) found that
invasive P. antipodarum tended to be more resilient to nutrient limitation than their native New
157
Zealand counterparts, consistent with this possibility. Given the propensity of mitochondrial
function to respond to environmental cues, the suite of traits related to mitochondrial
performance are promising candidates for investigating why some lineages become invasive
while others do not.
Implications for the loss of sex
Asexual lineages of P. antipodarum harbor substantial levels of nonsynonymous variation at
both the intraspecific and interspecific levels (Neiman et al. 2010, Chapter 2, 3). What remains
unclear is whether and how this variation affects phenotype (see Chapter 2). Nonsynonymous
mutations that reach fixation within a species are thought to be either beneficial or behaving
neutrally (Ohta 1992). What "neutral behavior" implies is that mutations that may be deleterious
in other contexts and/or species may contribute to diversity when the efficacy of selection is low
(e.g., when Ne is small, Charlesworth 2009). Given the extent of phenotypic variation in P.
antipodarum for mitochondrial function and the inference that mitochondrial function is
responsive to environmental cues, it seems likely that selection against phenotypic variation for
mitochondrial function is either not particularly intense and/or that selection is ineffective due to
environmental contribution to phenotypic variation.
In this initial test of mitochondrial function in sexuals vs. asexuals, I was not able to
sample both reproductive modes for any lake other than Lake Alexandrina. Combined with the
large lake effect, this means that inferences regarding mitochondrial function in sexual vs.
asexual P. antipodarum are limited. The only firm conclusion for this dataset is that variation
between sexuals and asexuals from different lakes does not explain the pattern of phenotypic
variation as well as the lake of origin does. This conclusion is consistent with local derivation of
158
asexuals, because if asexual lineages in this data set traced back to a single transition to
asexuality and variation for mitochondrial function is shared within asexual lineages (Chapter 3),
we would expect the effect of reproductive mode to swamp out any lake effect. Because
reproductive mode cannot be assessed until after the phenotypic data is collected and the effect
size for within-lake samples appears to be small, future comparisons of sexual vs. asexual snails
will require extensive within-lake sampling to obtain the necessary power to rigorously test
whether sexual and asexual snails differ in metabolic function.
159
ACKNOWLEDGEMENTS
The ploidy identification data presented herein were obtained at the Flow Cytometry Facility,
which is a Carver College of Medicine/ Holden Comprehensive Cancer Center core research
facility at the University of Iowa. The Facility is funded through user fees and the generous
financial support of the Carver College of Medicine, Holden Comprehensive Cancer Center, and
Iowa City Veteran's Administration Medical Center. I acknowledge Laura Bankers, Kaitlin
Hatcher, Katelyn Larkin, and Curt Lively for snail collections. The National Science Foundation
(NSF: MCB – 1122176; DEB – 1310825), the National Geographic Society, and the Iowa
Academy of Science (ISF #13-10 and #14-03) supported this research
160
REFERENCES
Abele, D., Heise, K., Portner, H. O., & Puntarulo, S. (2002). Temperature-dependence of
mitochondrial function and production of reactive oxygen species in the intertidal mud
clam Mya arenaria. Journal of Experimental Biology, 205, 1831-1841.
Alonso, A., and P. Castro-Díez. 2012. The exotic aquatic mud snail Potamopyrgus antipodarum
(Hydrobiidae, Mollusca): state of the art of a worldwide invasion. Aquatic Sciences Research Across Boundaries 74, 375-383.
Arnqvist, G., Dowling, D. K., Eady, P., Gay, L., Tregenza, T., Tuda, M., & Hosken, D. J. (2010).
Genetic architecture of metabolic rate: environment specific epistasis between
mitochondrial and nuclear genes in an insect. Evolution, 64, 3354-3363.
Box, G. E., & Cox, D. R. (1964). An analysis of transformations. Journal of the Royal Statistical
Society. Series B (Methodological), 211-252.
Barreto, F. S., & Burton, R. S. (2013). Elevated oxidative damage is correlated with reduced
fitness in interpopulation hybrids of a marine copepod. Proceedings of the Royal Society
B, 280, 20131521.
Beck, E. A., Thompson, A. C., Sharbrough, J., Brud, E., & Llopart, A. (2015). Gene flow
between Drosophila yakuba and Drosophila santomea in subunit V of cytochrome c
oxidase: A potential case of cytonuclear cointrogression. Evolution, 69, 1973-1986.
Bergstrom, C. T., & Pritchard, J. (1998). Germline bottlenecks and the evolutionary maintenance
of mitochondrial genomes. Genetics, 149, 2135-2146.
Blier, P. U., Dufresne, F., & Burton, R. S. (2001). Natural selection and the evolution of
mtDNA-encoded peptides: evidence for intergenomic co-adaptation. Trends in Genetics,
17, 400-406.
161
Charlesworth, B. (2009). Fundamental concepts in genetics: effective population size and
patterns of molecular evolution and variation. Nature Reviews Genetics, 10, 195-205.
Chen, J. H., Hales, C. N., & Ozanne, S. E. (2007). DNA damage, cellular senescence and
organismal ageing: causal or correlative? Nucleic Acids Research, 35, 7417-7428.
Denver, D. R., Morris, K., Lynch, M., Vassilieva, L. L., & Thomas, W. K. (2000). High direct
estimate of the mutation rate in the mitochondrial genome of Caenorhabditis elegans.
Science, 289, 2342-2344.
Dowling, D. K. (2014). Evolutionary perspectives on the links between mitochondrial genotype
and disease phenotype. Biochimica Biophysica Acta, 1840, 1393-1403.
Dowling, D. K., Abiega, K. C., & Arnqvist, G. (2007). Temperature-specific outcomes of
cytoplasmic-nuclear interactions on egg-to-adult development time in seed beetles.
Evolution, 61(1), 194-201.
Dybdahl, M. F., & Kane, S. L. (2005). Adaptation vs. phenotypic plasticity in the success of a
clonal invader. Ecology, 86, 1592-1601.
Ellison, C. K., & Burton, R. S. (2006). Disruption of mitochondrial function in interpopulation
hybrids of Tigriopus californicus. Evolution, 60, 1382-1391.
Ellison, C. K., & Burton, R. S. (2008). Interpopulation hybrid breakdown maps to the
mitochondrial genome. Evolution, 62, 631-638.
Hebert, P. D., Ratnasingham, S., & deWaard, J. R. (2003). Barcoding animal life: cytochrome c
oxidase subunit 1 divergences among closely related species. Proceedings of the Royal
Society B, 270 Suppl 1, S96-99.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection.
Genetical Research, 8, 269-294.
162
Hoekstra, L. A., Siddiq, M. A., & Montooth, K. L. (2013). Pleiotropic effects of a mitochondrialnuclear incompatibility depend upon the accelerating effect of temperature in Drosophila.
Genetics, 195, 1129-1139.
Keller, S. R., and D. R. Taylor. 2008. History, chance and adaptation during biological invasion:
separating stochastic phenotypic evolution from response to selection. Ecology Letters,
11, 852-866.
Krist, A. C., Kay, A. D., Larkin, K., & Neiman, M. (2014). Response to phosphorus limitation
varies among lake populations of the freshwater snail Potamopyrgus antipodarum. PLoS
ONE, 9, e85845.
Lande, R., & Shannon, S. (1996). The role of genetic variation in adaptation and population
persistence in a changing environment. Evolution, 50, 434-437.
Levri, E. P., Lunnen, S. J., Itle, C. T., Mosquea, L., Kinkade, B. V., Martin, T. G., & DeLisser,
M. A. (2007). Parasite-induced alteration of diurnal rhythms in a freshwater snail.
Journal of Parasitology, 93, 231-237.
Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by
parasitism. Nature, 328, 519-521.
Lynch, M., Burger, R., Butcher, D., & Gabriel, W. (1993). The mutational meltdown in asexual
populations. Journal of Heredity, 84, 339-344.
Montooth, K. L., Abt, D. N., Hofmann, J. W., & Rand, D. M. (2009). Comparative genomics of
Drosophila mtDNA: Novel features of conservation and change across functional
domains and lineages. Journal of Molecular Evolution, 69, 94-114.
Muller, H. J. (1964). The relation of recombination to mutational advance. Mutat Res, 106, 2-9.
163
Neiman, M., and A. Krist. 2016. Sensitivity to dietary phosphorus limitation in native vs.
invasive lineages of a New Zealand snail. Ecological Applications, in press.
Neiman, M., & Taylor, D. R. (2009). The causes of mutation accumulation in mitochondrial
genomes. Proceedings of the Royal Society B, 276, 1201-1209.
Normark, B. B., & Moran, N. A. (2000). Opinion - Testing for the accumulation of deleterious
mutations in asexual eukaryote genomes using molecular sequences. Journal of Natural
History, 34, 1719-1729.
Ohta, T. (1995). Synonymous and nonsynonymous substitutions in mammalian genes and the
nearly neutral theory. Journal of Molecular Evolution, 40, 56-63.
Paczesniak, D., Jokela, J., Larkin, K., & Neiman, M. (2013). Discordance between nuclear and
mitochondrial genomes in sexual and asexual lineages of the freshwater snail
Potamopyrgus antipodarum. Molecular Ecology, 22, 4695-4710.
Pike, T. W., Blount, J. D., Bjerkeng, B., Lindstrom, J., & Metcalfe, N. B. (2007). Carotenoids,
oxidative stress and female mating preference for longer lived males. Proceedings of the
Royal Society of London B, 274, 1591-1596.
Rand, D. M., Haney, R. A., & Fry, A. J. (2004). Cytonuclear coevolution: the genomics of
cooperation. Trends in Ecology and Evololution, 19, 645-653.
Reinhold, K. (1999). Energetically costly behaviour and the evolution of resting metabolic rate in
insects. Functional Ecology, 13, 217-224.
Sloan, D. B., Triant, D. A., Wu, M., & Taylor, D. R. (2014). Cytonuclear interactions and
relaxed selection accelerate sequence evolution in organelle ribosomes. Molecular
Biology and Evolution, 31, 673-682.
164
Stadler, T., Frye, M., Neiman, M., & Lively, C. M. (2005). Mitochondrial haplotypes and the
New Zealand origin of clonal European Potamopyrgus, an invasive aquatic snail.
Molecular Ecology, 14, 2465-2473.
Wallace, C. (1992). Parthenogenesis, sex and chromosomes in Potamopyrgus. Journal of
Molluscan Studies, 58, 93-107.
Zachar, N., & Neiman, M. (2013). Profound effects of population density on fitness-related traits
in an invasive freshwater snail. PLoS ONE, 8, e80067.
Zhang, F., & Broughton, R. E. (2013). Mitochondrial-nuclear interactions: compensatory
evolution or variable functional constraint among vertebrate oxidative phosphorylation
genes? Genome Biology and Evolution, 5, 1781-1791.
165
FIGURES
Figure 5-1. Map of New Zealand lakes from which snails used in this study were collected.
166
Figure 5-2. Interaction plot depicting significant interaction between temperature and lake
of origin.
A two-way ANOVA revealed a significant interaction between lake of origin (N = 7), and
temperature in 63 field-collected P. antipodarum (F 6, 163 = 2.329, p = 0.0305). Lines represent
best-fit line fit to each pair of temperatures. Neither lake of origin nor temperature had a
significant effect on O2 consumption alone.
167
SUPPLEMENTARY MATERIAL
168
169
Figure 5-S1. Relationship between log-transformed mass and O2 consumption in fieldcollected snails.
a) Log-transformed mass was significantly and positively correlated with square-roottransformed O2 consumption (Pearson’s r = 0.27, p = 0.035). b) Log-transformed mass was not
correlated with O2 consumption at 22ºC (Pearson’s r = 0.09, p = 0.487). c) Log-transformed
mass was not correlated with power-transformed O2 consumption at 30ºC (Pearson’s r = -0.05, p
= 0.709). Residuals were calculated from the best-fit lines (black line within plot) and used to
compare O2 consumption across lake of origin, temperature, and reproductive mode.
170
Table 5-S1. Field collected samples from seven New
Zealand lakes.
Lake
Sexual
Asexual
Alexandrina
14
2
Clearwater
–
4
Heron
–
2
Kaniere
16
–
Paringa
5
–
Rotoroa
–
17
Selfe
–
3
Total
36
28
171
Table 5-S2. ANOVA Tables
Three-way ANOVA
Temperature
Lake of Origin
Reproductive Mode
Lake of Origin * Temperature
Temperature * Reproductive Mode
Residuals
Two-way ANOVA, Within Sex
Temperature
Lake of Origin
Temperature * Lake of Origin
Residuals
Two-way ANOVA, Within Asex
Temperature
Lake of Origin
Temperature * Lake of Origin
Residuals
Sum Sq
Df
F
p
4.81E+05
2
0.0023
0.99775
8.18E+08
6
1.2778
0.27064
2.20E+07
1
0.2059
0.65068
1.97E+09
11
1.6758
0.08365
4.51E+07
2
0.2116
0.80953
1.64E+10
154
Sum Sq
Df
F
p
6.73E+04
2
0.0003
0.9997
4.48E+08
2
1.778
0.175
1.01E+09
4
2.0015
0.1012
1.11E+10
88
Sum Sq
Df
F
p
1149423
2
0.0071
0.9929
369753916
4
1.1428
0.3442
957542487
7
1.6911
0.1264
5338544132
66
Sum Sq
Df
F
p
5.71E+05
2
0.0027
0.99729
8.01E+08
6
1.2705
0.27398
1.92E+09
11
1.6591
0.08751
1.65E+10
157
Sum Sq
Df
F
p
33.74
6
1.122
0.362
Two-way ANOVA, No Reproductive
Mode
Temperature
Lake of Origin
Temperature * Lake of Origin
Residuals
Two-way ANOVA within Temperature
16º C
Lake of Origin
172
Table 5-S2 – continued
Reproductive Mode
Residuals
22ºC
Lake of Origin
Reproductive Mode
Residuals
30ºC
Lake of Origin
Reproductive Mode
Residuals
7.67
1
1.53
0.222
265.72
53
Sum Sq
Df
F
p
34007
6
3.355
0.00697
22
1
0.013
0.91056
91218
54
Sum Sq
Df
F
p
2.72E+09
5
1.555
0.191
6.71E+07
1
0.192
0.663
1.64E+10
47
Sum Sq
Df
F
p
6973653779
2
46.8057
9.68E-11
18483218
1
0.2481
0.6214
37988900
2
0.255
0.7763
2681846865
36
Sum Sq
Df
F
p
467146810
2
3.3042
0.08994
565515470
8
Sum Sq
Df
F
p
1152
1
0.7047
0.4896
3269.5
2
Sum Sq
Df
F
p
481195883
2
1.5126
0.2309
7475837164
47
**
Two-way ANOVA within Lake of Origin
Alexandrina
Temperature
Reproductive Mode
Temperature * Reproductive Mode
Residuals
***
One-way ANOVA within Lake of Origin
Clearwater
Temperature
Residuals
Heron
Temperature
Residuals
Kaniere
Temperature
Residuals
173
Table 5-S2 – continued
Paringa
Temperature
Residuals
Rotoroa
Temperature
Residuals
Selfe
Temperature
Residuals
Sum Sq
Df
F
p
138794091
2
0.4103
0.6732
1860588393
11
Sum Sq
Df
F
p
49314633
2
0.3049
0.7387
3720163862
46
Sum Sq
Df
F
p
433570009
2
135272
2.19E-10
6410
4
***
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05
174
CHAPTER 6: CONCLUSIONS AND FUTURE DIRECTIONS
SUMMARY OF FINDINGS
Genomic consequences of asexuality
In Chapters 2 and 3, I showed that asexual lineages of P. antipodarum do not remove putatively
deleterious mutations from their genomes as efficiently as sexual lineages. In particular, asexual
lineages tended to harbor a larger number of radical polymorphisms than sexual lineages.
Superficially, this pattern is surprising in light of the expectation that mutation accumulation in
asexual lineages expected to be more apparent for mutations that are only slightly deleterious
(Hill and Robertson 1966, Birky and Walsh 1988, Charlesworth et al. 1993). Because the
efficacy of selection is the product of the effective population size and selection intensity
(Charlesworth 2009), a reduction in Ne will cause all mutations to behave closer to neutral
expectations. One implication of this effect of Ne reduction is that selection increasingly loses the
ability to distinguish between different types of mutations. The patterns I observed, that slightly
deleterious mutations accumulate in asexuals and that mutations of more severe consequence
take longer for selection to sort out compared to sexual populations, are consistent with a
scenario whereby the landscape of fitness variation becomes compressed amongst asexuals as
individual non-lethal mutations lose the capacity to alter fitness enough that selection can
remove them. Accordingly, genetic drift – the random fluctuation of allele frequencies over
generations due to random sub sampling of a population – becomes more effective in
determining mutational trajectory, and more broadly, lineage trajectory.
My data also suggest that selection does not appear to be particularly strong or effective
with respect to harmful mutations in the mitochondrial genomes of sexual lineages of P.
175
antipodarum. This pattern was evident in the patterns of evolutionary rate heterogeneity of
mitochondrial genes: mitochondrial genes vary widely in their selective intensities such that
some genes display signs of relatively intense purifying selection while others appear to
experience substantially reduced functional constraint. Because Ne in the mitochondrial genome
is already reduced due to uniparental inheritance, haploidy, and recurrent population
bottlenecking (Normark and Moran 2000, Neiman and Taylor 2009, Haig 2016), the difference
in Ne between mitochondrial genomes sequestered in asexual lineages compared to
mitochondrial genomes in sexual populations may not be substantial enough to cause fitness
deficits in asexuals, even in the face of accelerated mutation accumulation. Ultimately, the
nuclear genome, the site of the vast majority of gene content and recombination, will tell the
most important story regarding the consequences of asexuality for mutation accumulation for P.
antipodarum as well as for other asexual taxa. Nevertheless, my study of the entire proteincoding region of the mitochondrial genome showed that it is possible to infer the extent of
mutation accumulation due to a reduction in Ne by accounting for the intensity of selection acting
on individual genes. This relationship represents an important tool going forward, because the
reduction in Ne should be much more substantial in the nuclear genome. Consequently, I predict
that the variance per gene in the extent to which a reduction in Ne accelerates deleterious
mutation accumulation should be larger in the nuclear genome than the variance per gene in the
mitochondrial genome, considering that sex and recombination primarily exert their effects in the
nuclear genome.
176
Phenotypic consequences of asexuality
In Chapter 4, I developed several methods for assaying various elements of mitochondrial
function in snails. I then used these methods to show that P. antipodarum harbors extensive
phenotypic variation for mitochondrial function and that this variation has at least a partial
genetic basis, with evidence for lineage-specific patterns of my various metrics of mitochondrial
function (Chapters 4, 5). This pattern was especially evident at the organellar and organismal
levels.
The discovery of genetic variation for mitochondrial function sets the stage for a critical
test of the mutational hypotheses for sex: deleterious mutation accumulation in the absence of
sex should reduce function and fitness in asexual lineages compared to their sexual counterparts.
In the first of many ongoing tests of mitochondrial function in sexual vs. asexual lineages, I
could not detect any differences between sexual and asexual lineages from the different lakes. I
did, however, show that lake of origin has a large effect on mitochondrial phenotype, indicating
that distinguishing between sexual and asexual snails will require within-lake sampling of sexual
and asexual snails. More specifically, snails from a particular lake tended to show similar O2
consumption curves to other snails from the same lake and rather different O2 consumption
curves compared to snails from other lakes. The population structure I observed here reflects
patterns of haplotype structure at the mitochondrial genomic level (Neiman and Lively 2004,
Neiman et al. 2010, Paczesniak et al. 2013, Chapter 2), providing an orthogonal line of evidence
for genetic variation underlying phenotypic variation in this species. Evidence from other work
agrees with this conclusion: sympatric trematode parasites infect snail hosts at a higher rate than
allopatric parasites (Lively et al. 2004), and gene expression profiles (Bankers et al. 2016) and
177
responses to nutrient limitation (Krist et al. 2014) are lake specific, such that lake of origin
appears to be an important predictor of phenotype.
The lack of difference between sexual and asexual snails in terms of O2 consumption is
not particularly surprising, especially considering the dearth of within-lake comparisons. It is
also consistent with asexual lineages being young and locally derived, which Paczesniak et al.
(2013) suggests is the predominant pattern in P. antipodarum, although the extent to which this
is true may vary. More specifically, if asexual lineages are recent snapshots of sexual population
variation, there has likely not been a long enough time frame since the transition to accumulate
deleterious mutations and experience functional decline. Still, if mutation accumulation affects
the outcome of competition between sexual and asexual lineages, it likely has to work on an
ecological rather than evolutionary time scale, as the two-fold cost of males is expected to drive
sexual populations to extinction extremely quickly in evolutionary time (Lively 1996). The
current study is not powerful enough to test whether the effect of mutation accumulation on
mitochondrial function is relatively small, if asexuals are recently derived, or if mutation
accumulation is not causing a decline in asexual mitochondrial function. Future comparisons of
mitochondrial function in young vs. old asexuals will provide a powerful glimpse into the time
scale at which mutation accumulation contributes to the seemingly inevitable extinction in
asexual lineages.
CONTRIBUTIONS OF FINDINGS TO THE FIELD
Population genetics and molecular evolution
Perhaps the single most important contribution of my thesis work to the broader field of Biology,
especially population genetics and molecular evolution, is the evidence presented in Chapter 2
178
that radical nonsynonymous mutations are significantly more likely to be harmful than
conservative mutations, even at extremely short time scales. Indeed, the difference between
conservative and radical rates of evolution appear to be so different as to suggest that the
monolithic classification of all nonsynonymous mutations into a single category is a misleading
average of the rate of amino acid evolution. While others have investigated this topic in model
systems including Drosophila (Rand et al. 2000, Smith 2003), humans (Fay et al. 2001, Miller
and Kumar 2001, Eyre-Walker et al. 2002, Vitkup et al. 2003), mammals (Zhang 2000, Hanada
et al. 2007), and flowering plants (Rand et al. 2000), among others, the differential effects on
fitness and function are rarely incorporated to genomic datasets. In this work, I developed a
novel algorithm to compare rates and patterns of conservative and radical nonsynonymous
evolution and applied that algorithm to a whole mitochondrial genome dataset. Together, the
inference that radical mutations are more harmful than conservative mutations and the practical
tools that I have developed to help analyze rates and patterns of amino acid evolution will
provide an important tool in the genomics toolkit to analyze evolutionary processes that
contribute to the removal (or lack thereof) of harmful mutations.
The work described in Chapters 2 and 3 also provides an important advance to our
understanding of the consequences of sex on molecular evolution. In particular, I found that
sexual lineages experience a more rapid rate of removal of harmful mutations than asexual
lineages (Chapter 2) and that the intensity of selection on a given genomic region contributes to
the degree of harmful mutation retention in asexuals (Chapter 3). While the relationship between
the intensity of selection and the efficacy of selection has been hypothesized in theory
(Charlesworth 2009) and likely contributes to similar patterns of harmful mutation retention in
small vs. large effective populations (e.g., large-bodied vs. small-bodied mammals, Zhang 2000),
179
this work represents the first time that this relationship has been empirically demonstrated in a
system that is polymorphic for sexual reproduction. Importantly, dissecting the relationship
between the intensity and efficacy of selection represents the only way to directly measure the
extent to which the reduction in effective population size resulting from the loss of sex affects
harmful mutation accumulation and/or beneficial mutation fixation. The empirical evidence
presented here in support of this relationship will provide an important context in which to test
the mutational processes hypothesized to maintain sex in nature because it will allow
comparative approaches to account for the potentially confounding effects of natural selection
and reproductive mode.
Development of phenotypic markers for P. antipodarum
One of the major challenges facing non-model organisms is the dearth of phenotypic assays that
are connected to both genotypic and fitness variation. This is especially the case for mollusks,
whose genetic diversity is largely underrepresented, meaning that investigation into the
relationship between mutational information and phenotypic data cannot be easily translated into
fitness outcomes. One of the primary goals of this thesis was the development of several such
assays of mitochondrial performance, including mitochondrial genome copy number,
mitochondrial membrane potential, mitochondrial electron flux, and organismal oxygen
consumption (Chapter 4). All of these phenotypic measures have been connected to some
measure of fitness in other organisms (e.g., Moraes et al. 1991, Liu et al. 1997, Garner and
Thomas 1999, Abele et al. 2002); however, the relationship between variation in mitochondrial
function and variation in fitness differs across species (e.g., mammals vs. reptiles, Lane 2011),
making it necessary to establish whether variation in mitochondrial function leads to fitness
180
variation in P. antipodarum. The first step in assessing the extent to which this is true is testing
whether variation for mitochondrial performance is reliably inherited, as it represents one of four
prerequisites that must be satisfied for harmful mutations to be efficiently removed from
populations (i.e., there must exist phenotypic variation, there must exist variation in fitness,
variation in phenotype must be related to variation in fitness, and, my focal premise, phenotypic
variation must be reliably inherited). In Chapter 4, I showed that not only is there phenotypic
variation for mitochondrial performance at three different levels of biological organization in P.
antipodarum, but also that variation in mitochondrial function is more likely to be shared
between closely related individuals than between distantly related individuals. This, combined
with the observation of lake-specific mitochondrial function in Chapter 5, suggests that
phenotypic variation for mitochondrial function is indeed reliably inherited, meaning that there is
at least the potential for natural selection to remove harmful mitochondrial mutations in P.
antipodarum. While it is still not clear how closely mitochondrial performance is tied to snail
fitness, these newly developed measures of phenotypic variation in P. antipodarum offer novel
measures of genotypic and phenotypic variation in an emerging model organism.
The phenotypic assays developed here should also be easily transferrable to other
mollusk species, particularly the methods for assaying electron flux and mitochondrial
membrane potential. The MTT assay and the JC-1 assay are commonly used in model system
such as human (Liu et al. 1997, Okpalugo et al. 2004), mouse (Fiorentino et al. 1989, Wang et al.
2016), and Drosophila (Terzhaz et al. 2006, Xu et al. 2015), but until now have not been adapted
to suit measures of mitochondrial performance in mollusks. Of note, mollusks have several
differences from other eukaryotes in their mitochondrial dynamics including the use of opines as
alternate energy storage molecules (Harcet et al. 2013), such that developing useful tools from
181
other distantly related model organisms is not straightforward, despite a clear demand for
mitochondrial assays in studying neurobiology (e.g., Vallejo et al. 2014, Shomrat et al. 2011),
the biology of aging (Ungvari et al. 2013), and understanding the energetic demands of invasive
and/or problematic mollusk species (Baldwin et al. 1996, Pigneur et al. 2011, Levri et al. 2014).
Further development of similar tools and fine-tuning of these molecular assays of mitochondrial
performance will provide powerful tools for understanding fitness and function in organisms
lacking such previously developed infrastructure, especially considering the importance of
mitochondrial function in other species (Burton and Barreto 2012) and the genetic tractability of
the mitochondrial genome in animals.
NEW AND OPEN QUESTIONS REGARDING THE MAINTENANCE OF SEX,
MITOCHONDRIAL FUNCTION, AND P. ANTIPODARUM
Do asexual lineages accumulate harmful mutations more rapidly than related sexual
lineages in their nuclear genomes?
The data presented in Chapters 2 and 3 indicates that asexual lineages are clearly accumulating
putatively harmful mutations in their mitochondrial genomes at a higher rate than related sexual
lineages; however, it is not clear that mitochondrial mutation accumulation causes a large enough
decline in function and/or fitness of asexual lineages to affect the maintenance of sex or the
seemingly inevitable extinction of asexual lineages. In particular, theory suggests (Lynch et al.
1993, Charlesworth 2009) and my work corroborates (Chapter 3) that the mutations being
accumulated should be those of relatively small effect (e.g., higher rate of mutation accumulation
in mitochondrial genes that experience less intense selection). Furthermore, although the sample
size was extremely limited, differences in mitochondrial function across reproductive modes
182
appear to be slight if at all in field-collected snails, especially compared to the effects of other
important biological processes, such as restricted gene flow between lake populations (Chapter
5). Together, these data indicate that if harmful mutation accumulation is at least partially
responsible for the maintenance of sex in P. antipodarum, much of the contribution must be from
the nuclear genome. Currently, very few studies have been able to compare nuclear genomewide patterns of mutation accumulation in a species that varies in reproductive mode (but see
e.g., Tucker et al. 2013, Hollister et al. 2014), meaning that such comparisons in P. antipodarum
will prove essential tests of the mutational hypotheses for sex.
Does mitochondrial performance predict fitness outcomes in P. antipodarum?
One of the unique challenges of using P. antipodarum as a model system is the relatively long
generation time (6-9 months, Larkin et al. 2016), which makes connecting phenotypic variation
to fitness variation time consuming and expensive. There are, however, some aspects of P.
antipodarum that can be taken advantage of to help answer whether phenotypic variation is
associated with fitness variation. In particular, by leveraging the use of asexual lineages and the
observation that P. antipodarum is ovoviviparous (Schreiber et al. 1998), testing for an
association between mitochondrial function and fitness could be accomplished by testing
whether variation in brood size correlates with variation in mitochondrial performance. Because
P. antipodarum harbor smaller broods when under phosphorus limitation (Tibbets et al. 2010), it
seems likely that reproduction is energetically expensive in this species, meaning that variation
in mitochondrial function could influence variation in brood size. Additionally, time to
reproductive maturity is an important predictor of lifetime reproductive success in P.
antipodarum, as females that reach maturity sooner have offspring sooner (Larkin et al. 2016).
183
Testing whether time to reproductive maturity is correlated to mitochondrial function at
adulthood would provide a powerful test of whether and how variation in mitochondrial function
impacts fitness in P. antipodarum. Establishing the extent to which these relationships between
function and fitness exist is essential for understanding how mitochondrial mutation
accumulation in asexuals affects the maintenance of sex and the extinction of asexual lineages, a
central question discussed in this thesis.
Do asexuals experience a reduction in fitness as a consequence of mutation accumulation?
The mutations that asexual P. antipodarum lineages are accumulating exhibit all the hallmarks
of being harmful; however, there is very little evidence that these mutations are causing declines
in fitness. Firstly, snails from asexual lineages have a faster growth rate and reach reproductive
maturity sooner than snails from sexual lineages (Larkin et al. 2016). Mitochondrial performance
is not significantly different across reproductive modes (although limited sampling means that
this result has not been rigorously tested, Chapter 5). Anecdotally, asexual populations sampled
from the field tend to survive in lab conditions better than sexual populations sampled from the
field, and lakes typically harboring mixed assemblages of sexual and asexual lineages often
become asexual over the course of several years of lab rearing (e.g., Lake Selfe sample in
Chapter 5, Maurine Neiman, personal communication). Thus, if asexual lineages are
experiencing a decline in function and/or fitness, it has yet to be uncovered. This observation
implies that either asexual lineages are recently derived (a distinct possibility, see Figure 2.1,
Neiman et al. 2004, Neiman et al. 2011, Paczesniak et al. 2013), that the mutations being
accumulated are of so small effect as to render them inconsequential to the maintenance of sex
(but perhaps not to the extinction of asexual lineages), that asexual P. antipodarum are
184
particularly robust to harmful mutations, or some combination thereof. There is some evidence
that this final possibility is indeed the case. For example, one asexual lineage sampled in
Chapters 2 and 3 harbors a premature stop codon in its nd2 gene, despite no readily apparent
difference in fitness. Further, asexual P. antipodarum exhibit wide variation in one of the most
fundamental eukaryotic traits: genome copy number. In particular, the vast majority of
eukaryotes are diploid, and deviations from diploidy often have negative consequences on
organismal health, especially in mammals (Otto 2007). On the contrary, asexual lineages of P.
antipodarum not only exhibit rampant polyploidy, but also highly variable distributions of
genome size within ploidy classes. Increases in ploidy class (i.e., 2x to 3x, 3x to 4x, etc.) appear
to be associated with increases in variation in genome size (Neiman et al. 2011). One potential
explanation for this observation lies in the prospect of mutational masking. Because polyploidy
is often closely associated with asexuality (Otto and Whitton 2007), and it is possible that
elevated ploidy may allow for the temporary masking of deleterious mutations (Otto and Whitton
2000), polyploid asexual lineages may experience a slower rate of mutational meltdown than
diploid asexual lineages. Ultimately, uncovering the impacts that mutations have on function and
fitness in diploid, triploid, and tetraploid lineages of P. antipodarum represents a central question
to testing whether asexual lineages experience mutational meltdown in this species.
Is genomic and phenotypic variation between lake populations a consequence of local
adaptation or population structure?
One of the major outcomes of Chapter 5 is that phenotypic variation in P. antipodarum is highly
structured by lake, meaning that gene flow between New Zealand lake populations appears to be
restricted in this species. This is somewhat surprising in light of P. antipodarum’s successful
185
invasion and colonization of novel environments as distantly as Australia, Asia, North America,
South America, and Europe (Alonso and Castro-Díez 2012). Considering that genetic variation
in mitochondrial genomes is also structured across lakes (see Figure 2.1), this finding raises the
question as to the process(es) governing divergence between New Zealand lake populations:
local adaptation, mutation and drift, or some combination thereof. Local adaptation should be
mediated by environmental variables, meaning that lakes with similar environments should favor
similar phenotypes (e.g., snails with more tolerant heat stress responses are expected to inhabit
warmer lakes). By contrast, if neutral processes were governing phenotypic divergence amongst
lake populations, we would expect stochastic patterns of phenotypic divergence to emerge,
irrespective of environmental variables, and that these patterns of phenotypic divergence should
be shared across traits. Further, mutation and drift should exert genome/organism-wide effects on
variation, whereas the effects of selection should be gene/trait specific.
A similar pattern of lake-structured variation has emerged in the gene expression
response of P. antipodarum to trematode parasite infection (Bankers et al. 2016), in which
infected snails from the same lake share expression profiles to a greater degree than infected
snails from different lakes. Additionally, P. antipodarum snails are significantly more likely to
be infected by a trematode parasite when exposed to sympatric trematodes compared to exposure
to allopatric trematodes, suggesting that parasites are locally adapted to P. antipodarum (Lively
et al. 2004). Because the parasite in question, Microphallus livelii, eliminates reproductive
capacity in the snails that it infects (Koskella and Lively 2009), selection is likely an important
contributor to this pattern of across-lake variation. An interesting test of local adaptation would
therefore be to compare across-lake correlations in expression profile in response to trematode
infection to across-lake correlations in mitochondrial function because if local adaptation is
186
operating to increase phenotypic divergence between lake populations, then we would expect no
correlation across datasets. On the other hand, similar across-lake correlations in these two
phenotypes would likely implicate a larger role for more neutral process of divergence mediated
by geographic isolation. More broadly, unraveling the source of mitochondrial functional
divergence in P. antipodarum has major implications for our understanding of the evolutionary
forces operating on mitochondrial function in animals. There is some evidence from a variety of
taxa that mitochondria can evolve adaptively (e.g., Bazin et al. 2006, Das 2006); however, there
remains little evidence that local adaptation is a major contributor to the evolution of
mitochondrial function.
CONCLUSIONS AND FUTURE PROSPECTS
There is still no clear answer as to whether asexuality engenders negative phenotypic
consequences linked to harmful mutation accumulation. My thesis provides key steps forward by
outlining the genomic framework under which such a decline could happen. There are several
pressing issues to address on this front: first and foremost, does the nuclear genome, especially
nuclear-encoded mitochondrial genes, exhibit signs of deleterious mutation accumulation?
Second, are there signs of mitochondrial functional deterioration in asexuals at different levels of
biological organization? And finally, what is the time frame over which mutation accumulation
operates, and are these processes fast enough to impact the evolutionary maintenance of sex?
These questions represent major gaps in our understanding of evolution, but there is reason to
believe that these questions can be answered, especially in the (non-model) genomic era.
187
REFERENCES
Alonso, Á., & Castro-Díez, P. (2012). The exotic aquatic mud snail Potamopyrgus antipodarum
(Hydrobiidae, Mollusca): state of the art of a worldwide invasion. Aquat Sci, 74, 375-383.
Baldwin, B. S., Black, M., Sanjur, O., Gustafson, R., Lutz, R. A., & Vrijenhoek, R. C. (1996). A
diagnostic molecular marker for zebra mussels (Dreissena polymorpha) and potentially
co-occurring bivalves: mitochondrial COI. Mol Mar Biol Biotechnol, 5, 9-14.
Bankers, L., Fields, P., McElroy, K. E., Boore, J. L., Logsdon Jr., J. M., Neiman, M. (2016)
Genomic evidence for population-specific responses to coevolving parasites in a New
Zealand freshwater snail. In review. bioRxiv doi: http://dx.doi.org/10.1101/045674.
Bazin, E., Glemin, S., & Galtier, N. (2006). Population size does not influence mitochondrial
genetic diversity in animals. Science, 312, 570-572.
Birky, C. W., Jr., & Walsh, J. B. (1988). Effects of linkage on rates of molecular evolution. Proc
Natl Acad Sci U S A, 85, 6414-6418.
Burton, R. S., & Barreto, F. S. (2012). A disproportionate role for mtDNA in DobzhanskyMuller incompatibilities? Mol Ecol, 21, 4942-4957.
Charlesworth, B. (2009). Fundamental concepts in genetics: effective population size and
patterns of molecular evolution and variation. Nat Rev Genet, 10, 195-205.
Charlesworth, B., Morgan, M. T., & Charlesworth, D. (1993). The effect of deleterious
mutations on neutral molecular variation. Genetics, 134, 1289-1303.
Das, A. M. (2003). Regulation of the mitochondrial ATP-synthase in health and disease. Mol
Genet Metabol, 79, 71-82.
Eyre-Walker, A., Keightley, P. D., Smith, N. G., & Gaffney, D. (2002). Quantifying the slightly
deleterious mutation model of molecular evolution. Mol Biol Evol, 19, 2142-2149.
188
Fay, J. C., Wyckoff, G. J., & Wu, C. I. (2001). Positive and negative selection on the human
genome. Genetics, 158, 1227-1234.
Fiorentino, D. F., Bond, M. W., & Mosmann, T. R. (1989). Two types of mouse T helper cell. IV.
Th2 clones secrete a factor that inhibits cytokine production by Th1 clones. J Exp Med,
170, 2081-2095.
Garner, D. L., & Thomas, C. A. (1999). Organelle-specific probe JC-1 identifies membrane
potential differences in the mitochondrial function of bovine sperm. Mol Reprod Dev, 53,
222-229.
Haig, D. (2016). Intracellular evolution of mitochondrial DNA (mtDNA) and the tragedy of the
cytoplasmic commons. Bioessays.
Hanada, K., Shiu, S. H., & Li, W. H. (2007). The nonsynonymous/synonymous substitution rate
ratio versus the radical/conservative replacement rate ratio in the evolution of mammalian
genes. Mol Biol Evol, 24, 2235-2241.
Hill, W. G., & Robertson, A. (1966). The effect of linkage on limits to artificial selection. Genet
Res, 8, 269-294.
Hollister, J. D., Greiner, S., Wang, W., Wang, J., Zhang, Y., Wong, G. K., . . . Johnson, M. T.
(2015). Recurrent loss of sex is associated with accumulation of deleterious mutations in
Oenothera. Mol Biol Evol, 32, 896-905.
Koskella, B., & Lively, C. M. (2009). Evidence for negative frequency-dependent selection
during experimental coevolution of a freshwater snail and a sterilizing trematode.
Evolution, 63, 2213-2221.
189
Krist, A. C., Jokela, J., Wiehn, J., & Lively, C. M. (2004). Effects of host condition on
susceptibility to infection, parasite developmental rate, and parasite transmission in a
snail-trematode interaction. J Evol Biol, 17, 33-40.
Krist, A. C., Kay, A. D., Larkin, K., & Neiman, M. (2014). Response to phosphorus limitation
varies among lake populations of the freshwater snail Potamopyrgus antipodarum. PLoS
ONE, 9, e85845.
Lane, N. (2011). Mitonuclear match: optimizing fitness and fertility over generations drives
ageing within generations. Bioessays, 33, 860-869.
Larkin, K., Tucci, C., & Neiman, M. (2016). Effects of polyploidy and reproductive mode on life
history trait expression. Ecol Evol, 6, 765-778.
Levri, E. P., Krist, A. C., Bilka, R., & Dybdahl, M. F. (2014). Phenotypic plasticity of the
introduced New Zealand mud snail, Potamopyrgus antipodarum, compared to sympatric
native snails. PLoS ONE, 9, e93985.
Liu, K. Z., Schultz, C. P., Johnston, J. B., Lee, K., & Mantsch, H. H. (1997). Comparison of
infrared spectra of CLL cells with their ex vivo sensitivity (MTT assay) to chlorambucil
and cladribine. Leuk Res, 21, 1125-1133.
Lively, C. M. (1996). Host-parasite coevolution and sex. Bioscience, 46, 107–114.
Lively, C. M., Dybdahl, M. F., Jokela, J., Osnas, E. E., & Delph, L. F. (2004). Host sex and local
adaptation by parasites in a snail-trematode interaction. Am Nat, 164, S6-18.
Miller, M. P., & Kumar, S. (2001). Understanding human disease mutations through the use of
interspecific genetic variation. Hum Mol Genet, 10, 2319-2328.
190
Moraes, C. T., Shanske, S., Tritschler, H. J., Aprille, J. R., Andreetta, F., Bonilla, E., . . .
DiMauro, S. (1991). mtDNA depletion with variable tissue expression: a novel genetic
abnormality in mitochondrial diseases. Am J Hum Genet, 48, 492-501.
Neiman, M., & Lively, C. M. (2004). Pleistocene glaciation is implicated in the
phylogeographical structure of Potamopyrgus antipodarum, a New Zealand snail. Mol
Ecol, 13, 3085-3098.
Neiman, M., Paczesniak, D., Soper, D. M., Baldwin, A. T., & Hehman, G. (2011). Wide
variation in ploidy level and genome size in a New Zealand freshwater snail with
coexisting sexual and asexual Lineages. Evolution, 65, 3202-3216.
Neiman, M., & Taylor, D. R. (2009). The causes of mutation accumulation in mitochondrial
genomes. Proc Biol Sci, 276, 1201-1209.
Normark, B. B., & Moran, N. A. (2000). Opinion - Testing for the accumulation of deleterious
mutations in asexual eukaryote genomes using molecular sequences. J Nat Hist, 34,
1719-1729.
Okpalugo, T. I., McKenna, E., Magee, A. C., McLaughlin, J., & Brown, N. M. (2004). The MTT
assays of bovine retinal pericytes and human microvascular endothelial cells on DLC and
Si-DLC-coated TCPS. J Biomed Mater Res A, 71(2), 201-208.
Otto, S. P. (2007). The evolutionary consequences of polyploidy. Cell, 131, 452-462.
Otto, S. P., & Whitton, J. (2000). Polyploid incidence and evolution. Ann Rev Genet, 34, 401437.
Paczesniak, D., Jokela, J., Larkin, K., & Neiman, M. (2013). Discordance between nuclear and
mitochondrial genomes in sexual and asexual lineages of the freshwater snail
Potamopyrgus antipodarum. Mol Ecol, 22, 4695-4710.
191
Pigneur, L. M., Marescaux, J., Roland, K., Etoundi, E., Descy, J. P., & Van Doninck, K. (2011).
Phylogeny and androgenesis in the invasive Corbicula clams (Bivalvia, Corbiculidae) in
Western Europe. BMC Evol Biol, 11, 147.
Schreiber, E. S. G., Glaister, A., Quinn, G. P., & Lake, P. S. (1998). Life history and population
dynamics of the exotic snail Potamopyrgus antipodarum (Prosobranchia: Hydrobiidae) in
Lake Purrumbete, Victoria, Australia. Mar Freshw Res, 49, 73-78.
Shomrat, T., Graindorge, N., Bellanger, C., Fiorito, G., Loewenstein, Y., & Hochner, B. (2011).
Alternative sites of synaptic plasticity in two homologous "fan-out fan-in" learning and
memory networks. Curr Biol, 21, 1773-1782.
Smith, N. G. (2003). Are radical and conservative substitution rates useful statistics in molecular
evolution? J Mol Evol, 57, 467-478.
Terhzaz, S., Southall, T. D., Lilley, K. S., Kean, L., Allan, A. K., Davies, S. A., & Dow, J. A.
(2006). Differential gel electrophoresis and transgenic mitochondrial calcium reporters
demonstrate spatiotemporal filtering in calcium control of mitochondria. J Biol Chem,
281, 18849-18858.
Tibbets, T. M., Krist, A. C., Hall, R. O., Jr., & Riley, L. A. (2010). Phosphorus-mediated
changes in life history traits of the invasive New Zealand mudsnail (Potamopyrgus
antipodarum). Oecologia, 163, 549-559.
Tucker, A. E., Ackerman, M. S., Eads, B. D., Xu, S., & Lynch, M. (2013). Population-genomic
insights into the evolutionary origin and fate of obligately asexual Daphnia pulex. Proc
Natl Acad Sci USA, 110, 15740-15745.
192
Ungvari, Z., Csiszar, A., Sosnowska, D., Philipp, E. E., Campbell, C. M., McQuary, P. R., . . .
Ridgway, I. (2013). Testing predictions of the oxidative stress hypothesis of aging using a
novel invertebrate model of longevity: the giant clam (Tridacna derasa). J Gerontol A
Biol Sci Med Sci, 68, 359-367.
Vallejo, D., Habib, M. R., Delgado, N., Vaasjo, L. O., Croll, R. P., & Miller, M. W. (2014).
Localization of tyrosine hydroxylase-like immunoreactivity in the nervous systems of
Biomphalaria glabrata and Biomphalaria alexandrina, intermediate hosts for
schistosomiasis. J Comp Neurol, 522, 2532-2552.
Vitkup, D., Sander, C., & Church, G. M. (2003). The amino acid mutational spectrum of human
genetic disease. Gen Biol, 4, R72.
Wang, L., Xu, X., Kang, L., & Xiang, W. (2016). Bone marrow mesenchymal stem cells
attenuate mitochondria damage induced by hypoxia in mouse trophoblasts. PLoS ONE,
11, e0153729.
Xu, L., Li, S., Ran, X., Liu, C., Lin, R., & Wang, J. (2015). Apoptotic activity and gene
responses in Drosophila melanogaster S2 cells induced by azadirachtin A. Pest Manag
Sci.
Zhang, J. (2000). Rates of conservative and radical nonsynonymous nucleotide substitutions in
mammalian nuclear genes. J Mol Evol, 50, 56-68.
193