Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Environmental epigenomics in stickleback : plasticity of DNA methylation and gene expression patterns… Metzger, David Calin Hartley 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_may_metzger_david.pdf [ 8.9MB ]
Metadata
JSON: 24-1.0365265.json
JSON-LD: 24-1.0365265-ld.json
RDF/XML (Pretty): 24-1.0365265-rdf.xml
RDF/JSON: 24-1.0365265-rdf.json
Turtle: 24-1.0365265-turtle.txt
N-Triples: 24-1.0365265-rdf-ntriples.txt
Original Record: 24-1.0365265-source.json
Full Text
24-1.0365265-fulltext.txt
Citation
24-1.0365265.ris

Full Text

ENVIRONMENTAL EPIGENOMICS IN STICKLEBACK: PLASTICITY OF DNA METHYLATION AND GENE EXPRESSION PATTERNS ACROSS TIME SCALES  by  David Calin Hartley Metzger  B.A., Ithaca College 2005 M.Sc., The University of Washington 2012  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2018  © David Calin Hartley Metzger, 2018 ii  Abstract  Epigenetic mechanisms such as DNA methylation have been proposed as an important source of variation that can influence phenotypic plasticity and adaptive evolutionary processes, yet little is known about the role of DNA methylation in an ecological or evolutionary context in vertebrates. In this thesis I examine the effects of the environment and sex on DNA methylation and gene expression patterns in the threespine stickleback fish (Gasterosteus aculeatus), an ecological and evolutionary model system that has been used to study mechanisms involved in the evolution of adaptive phenotypes in novel environments.   The dynamic regulation of DNA methylation and gene expression patterns during early developmental periods plays an important role in cell differentiation and establishing adult phenotypes. Here I demonstrate that adult DNA methylation and gene expression patterns are modified in response to the temperature and salinity experienced during development. Similarly, maternal stress can have long-term effects on neurodevelopment and the behavior of offspring that can influence offspring performance and population evolutionary trajectories. I demonstrate that the effects of maternal stress on the brain transcriptome differ between adult male and female stickleback offspring. These sex-specific effects of maternal stress suggest that male and female offspring may respond differently to maternal stress exposure, which could have important implications when assessing the long-term ecological and evolutionary impacts of stress across generations.  DNA methylation has also been proposed to play a key role in regulating sexually dimorphic phenotypes and in the evolution of sex determination mechanisms. I compare genome-wide DNA methylation patterns between male and female stickleback and identify apparent differential methylation on the stickleback sex chromosome that correspond to the iii  regions of genetic divergence between the X and Y chromosome. These data provide evidence of a potential role of DNA methylation in the evolution of sex chromosomes in vertebrates.   Taken together, these data demonstrate that there is a complex relationship between genetic, epigenetic, and transcriptomic processes that are dynamically regulated during development and in response to environmental cues, and that epigenetic processes may be involved in regulating evolutionary processes.    iv  Lay Summary  Epigenetic mechanisms such as DNA methylation are thought to play key roles in an organism’s to response to changes in the environment by regulating the expression of genes in different environmental conditions. However, little is known about how DNA methylation patterns are modified in response to the environment. In this thesis I demonstrate that differences in environmental temperature, salinity, and maternal stress have widespread effects on DNA methylation levels and gene expression in threespine stickleback (Gasterosteus aculeatus), and demonstrate that sex is an important factor in determining the effects of maternal stress. I also identified sex-specific DNA methylation patterns on the stickleback sex chromosome that are associated with the evolution of this chromosome. These data suggest that DNA methylation is likely involved in both the dynamic regulation of gene expression in response to environmental change, and in regulating evolutionary processes such as the evolution of sex chromosomes. v  Preface   Sections of the introduction (chapter one) and discussion (chapter six) of this thesis have been published as: Metzger DCH, Schulte PM (2016) Epigenomics in marine fishes. Marine Genomics, 30, 43–54. Dr. Schulte and I contributed equally to the writing of this manuscript.  A version of chapter two has been published as: Metzger DCH, Schulte PM (2017) Persistent and plastic effects of temperature on DNA methylation across the genome of threespine stickleback (Gasterosteus aculeatus). Proceedings of the Royal Society B: Biological Sciences, 284, 20171667. I designed and conducted the experiments for this study, performed the data analysis, and wrote the manuscript under the supervision of Dr. Schulte.  A version of chapter three has been published as Metzger DCH, Schulte PM (2018) Similarities in temperature-dependent gene expression plasticity across time-scales in threespine stickleback (Gasterosteus aculeatus). Molecular Ecology. I designed and conducted the experiments for this study, performed the data analysis, and wrote the manuscript under the supervision of Dr. Schulte.  A version of chapter four has been accepted for publication as: Metzger DCH, Schulte PM (2018) The DNA methylation landscape of stickleback reveals patterns of sex chromosome evolution and effects of environmental salinity. Genome Biology and Evolution, 10, 775-785. I designed and conducted the experiments for this study, performed the data analysis, and wrote the manuscript under the supervision of Dr. Schulte.  A version of chapter five has been published as: Metzger DCH, Schulte PM (2016) Maternal stress has divergent effects on gene expression patterns in the brains of male and female threespine stickleback. Proceedings of the Royal Society B: Biological Sciences, 283, vi  20161734. I designed and conducted the experiments for this study, performed the data analysis, and wrote the manuscript under the supervision of Dr. Schulte.  All animal use was conducted under approved University of British Columbia animal care protocols A11-0372 and A10-0285, and adhered to all guidelines and policies of the Canadian Council on Animal Care. vii  Table of Contents Abstract .......................................................................................................................................... ii	Lay Summary ............................................................................................................................... iv	Preface .............................................................................................................................................v	Table of Contents ........................................................................................................................ vii	List of Tables .................................................................................................................................. x	List of Figures ............................................................................................................................... xi	List of Abbreviations ................................................................................................................. xvi	Acknowledgements .................................................................................................................. xviii	Chapter 1: Introduction ................................................................................................................1	1.1	 Goals of this thesis ............................................................................................................. 1	1.2	 Epigenetics and plasticity .................................................................................................. 2	1.2.1	 What is epigenetics? ...................................................................................................... 3	1.3	 Ecological epigenetics ....................................................................................................... 4	1.4	 DNA methylation as an epigenetic mark ........................................................................... 6	1.4.1	 Methods for assessing DNA methylation ................................................................... 10	1.5	 DNA methylation in fish .................................................................................................. 15	1.5.1	 Effects of environmental stressors on DNA methylation in fish ................................ 17	1.5.2	 DNA methylation and sex determination in fish ........................................................ 18	1.6	 Research organism: threespine stickleback (Gasterosteus aculeatus) ............................. 22	1.6.1	 Threespine stickleback taxonomy and evolution ........................................................ 23	1.6.2	 Threespine stickleback genomic tools ........................................................................ 24	1.6.3	 Physiological divergence of halotolerance between stickleback populations ............ 25	1.6.4	 Evolution of thermal tolerance in threespine stickleback ........................................... 27	1.6.5	 Evolution of heteromorphic sex chromosomes in threespine stickleback .................. 29	1.6.6	 Effects of maternal stress on offspring phenotypes in threespine stickleback ............ 30	1.6.7	 Epigenomics in stickleback ......................................................................................... 31	1.7	 Thesis organization .......................................................................................................... 34	1.7.1	 Chapter two ................................................................................................................. 34	1.7.2	 Chapter three ............................................................................................................... 34	1.7.3	 Chapter four ................................................................................................................ 35	1.7.4	 Chapter five ................................................................................................................. 35	Chapter 2: Persistent and plastic effects of temperature on DNA methylation across the genome of threespine stickleback (Gasterosteus aculeatus) ......................................................36	2.1	 Introduction ...................................................................................................................... 36	2.2	 Materials and methods ..................................................................................................... 38	2.2.1	 Experimental design .................................................................................................... 38	2.2.2	 Reduced representation bisulfite sequencing .............................................................. 41	2.2.3	 Quantification and statistical analysis ......................................................................... 41	2.3	 Results and discussion ..................................................................................................... 44	2.3.1	 Temperature change induces genomic hypermethylation ........................................... 44	2.4	 Conclusions ...................................................................................................................... 49	viii  Chapter 3: Similarities in temperature-dependent gene expression patterns between developmental plasticity and phenotypic flexibility in threespine stickleback (Gasterosteus aculeatus) ......................................................................................................................................50	3.1	 Introduction ...................................................................................................................... 50	3.2	 Materials and Methods ..................................................................................................... 54	3.2.1	 Gasterosteus aculeatus rearing conditions ................................................................. 54	3.2.2	 RNA isolation and sequencing .................................................................................... 58	3.2.3	 Sequence alignment and expression analysis .............................................................. 59	3.2.4	 Differential methylation analysis ................................................................................ 62	3.3	 Results .............................................................................................................................. 62	3.3.1	 Temperature dependent gene expression plasticity at different life stages ................. 62	3.3.2	 Identification of candidate biological processes affected by gene expression plasticity across timescales ................................................................................................................... 69	3.3.3	 Comparison to patterns of DNA methylation ............................................................. 72	3.4	 Discussion ........................................................................................................................ 74	3.4.1	 Conserved effects of cold and warm temperature on gene expression ....................... 74	3.4.2	 Relationship to patterns of DNA methylation ............................................................. 78	3.5	 Conclusions ...................................................................................................................... 80	Chapter 4: The DNA methylation landscape of stickleback reveals patterns of sex chromosome evolution and effects of environmental salinity. .................................................81	4.1	 Introduction ...................................................................................................................... 81	4.2	 Material and methods ....................................................................................................... 83	4.2.1	 Fish collection ............................................................................................................. 83	4.2.2	 Fertilization and rearing procedure ............................................................................. 84	4.2.3	 Whole genome bisulfite sequencing (WGBS) ............................................................ 85	4.2.4	 WGBS data analysis .................................................................................................... 85	4.2.5	 Candidate gene analysis .............................................................................................. 87	4.3	 Results and discussion ..................................................................................................... 88	4.3.1	 Characterization of the stickleback methylome .......................................................... 88	4.3.2	 Sex-biased DNA methylation patterns ........................................................................ 90	4.3.3	 Identifying differential methylation on the sex chromosome ..................................... 94	4.3.4	 DNA methylation and sex chromosome evolution ..................................................... 96	4.3.5	 DNA methylation and sex-biased gene expression ..................................................... 99	4.3.6	 Effects of environmental salinity on DNA methylation ........................................... 100	4.4	 Conclusions .................................................................................................................... 102	Chapter 5: Maternal stress has divergent effects on gene expression patterns in the brains of male and female threespine stickleback ..............................................................................103	5.1	 Introduction .................................................................................................................... 103	5.2	 Materials and methods ................................................................................................... 104	5.2.1	 Stress treatment and animal rearing .......................................................................... 104	5.2.2	 RNA isolation and sequencing .................................................................................. 105	5.2.3	 Sequence alignment and expression analysis ............................................................ 106	5.3	 Results ............................................................................................................................ 107	5.3.1	 Sexual dimorphism in gene expression ..................................................................... 108	5.3.2	 Sex-specific effects of maternal stress ...................................................................... 109	ix  5.3.3	 Chromosomal location of differentially expressed genes ......................................... 112	5.4	 Discussion ...................................................................................................................... 113	5.4.1	 Sex-specific effects of maternal stress ...................................................................... 114	5.4.2	 Sources of sexual dimorphism in brain gene expression .......................................... 115	5.4.3	 Effects of maternal stress on stress-related genes ..................................................... 117	5.4.4	 Effects of maternal stress on epigenetic processes ................................................... 118	5.5	 Conclusions .................................................................................................................... 119	Chapter 6: Conclusion ...............................................................................................................120	6.1	 Overview ........................................................................................................................ 120	6.2	 Ecological epigenetics and threespine stickleback ........................................................ 121	6.3	 Sex specific effects on DNA methylation and gene expression .................................... 123	6.3.1	 Differential methylation on the stickleback sex chromosome .................................. 124	6.3.2	 Divergent effects of maternal stress in male and female stickleback ....................... 125	6.4	 Methodological considerations ...................................................................................... 128	6.5	 Future epigenomic research in fishes ............................................................................. 130	6.6	 Conclusions .................................................................................................................... 133	Bibliography ...............................................................................................................................135	Appendices ..................................................................................................................................163	Appendix A Supporting information for chapter two ............................................................. 163	Appendix B Supporting information for chapter three ........................................................... 167	Appendix C Supporting information for chapter four ............................................................ 170	C.1	 Effects of environmental salinity on fertilization and hatching success ................... 170	C.2	 Characterizing family effects on DNA methylation patterns in stickleback ............. 170	C.3	 Supporting figures and tables for chapter four .......................................................... 171	Appendix D Supporting information for chapter five ............................................................. 179	D.1	 Fish collection and acclimation ................................................................................. 179	D.2	 Experimental stress treatment, crossing design and rearing ..................................... 179	D.3	 DNA Isolation and Sex Identification ....................................................................... 181	D.4	 Plasma cortisol .......................................................................................................... 181	D.5	 RNA isolation ........................................................................................................... 181	D.6	 Gene Ontology and pathway analysis ....................................................................... 182	D.7	 Differential expression between males and females ................................................. 183	D.8	 Effects of increased stringency thresholds on detection of sexually-dimorphic effects of maternal stress ................................................................................................................ 183	D.9	 Candidate genes involved in the stress response and epigenetic regulation ............. 184	D.10	 Supporting tables and figures for chapter five .......................................................... 186	 x  List of Tables  Table 1.1: Advantages and disadvantages of selected methods for assessing DNA methylation 14	Table 3.1: Genes that are differentially expressed in response to changes in both developmental temperatures and adult acclimation temperatures. ............................................................ 68	Table 3.2: Significantly enriched biological process gene ontologies for genes that are up- or down-regulated in response to warm or cold thermal acclimation in adult stickleback muscle tissue. .................................................................................................................... 70	 xi  List of Figures  Figure 1.1: Flow diagram illustrating parallel nature of ecological genetics (black lines) and ecological epigenetics (grey lines) (adapted from Bossdorf et al. 2008). .......................... 5	Figure 1.2: Patterns of CpG island methylation in proximal promoter regions. CpG islands are regions with high concentrations of CpG motifs. Shore and shelf regions are located at immediately flanking CpG islands. A high level of methylation in a CpG island associated with the transcription start site of a gene is associated with inactivation of transcription. Bent arrow indicates transcription start site; empty circles indicate umethylated CpG sites; filled circles indicate methylated CpG sites (reproduced from Metzger & Schulte 2016). ................................................................................................... 8	Figure 1.3: Bisulfite sequencing. Treatment of genomic DNA with sodium bisulfite converts unmethylated cytosines (C) into uracil (U), while methylated cytosines are protected. During PCR and sequencing uracil is converted to thymine (T). Sequence comparison of untreated and bisulfite converted DNA sequences allows the detection of methylated sites (reproduced from Metzger & Schulte 2016). .................................................................... 11	Figure 2.1: Experimental design. Solid lines represent stickleback that were developed and acclimated to 18 °C (black) or that developed at 18 °C and were acclimated to 25 °C (red) or 5 °C (dark blue). Dashed lines represent the developmental temperature treatments of 24 °C (pink), or 12 °C (light blue). Note that altered development temperature changes time to hatch by several days. Post-hatch, stickleback were transferred to 18 °C where they developed to adults for the remainder of the experiment (10 months). .................... 40	Figure 2.2: (A-D) Manhattan plots of the chromosomal positions of methylated CpG loci that differed significantly between stickleback that were developed at and acclimated to 18 °C, stickleback developed at 12 °C (A) or at 24 °C (B) then reared at 18 °C, or that were developed at 18 °C then acclimated to 5 °C (C) or 25 °C (D). Each point represents a single differentially methylated cytosine (DMC). The y-axis presents the difference in percent methylation for that DMC relative to the stickleback developed and acclimated at 18 °C. Only DMCs with >10% change in methylation are shown. Points above and below the horizontal dashed line are hypermethylated and hypomethylated loci, respectively. Points indicate chromosomal locations. Vertical dashed lines and colors indicate different chromosomal regions. (E-H) Chromosomal frequency distribution of hypermethylated (dark grey) and hypomethylated (light grey) DMCs in stickleback that developed at 12 °C (E) developed at 24 °C (F) or that were acclimated to 5 °C (G) or acclimated to 25 °C (H). The vertical axis represents the percent of CpGs that were sequenced that were differentially methylated. .................................................................................................. 43	 xii  Figure 2.3: (A) Violin plot of genomic DNA methylation levels for each treatment condition and control condition. Colors depict different treatments. Stickleback developed and acclimated to 18 °C (grey/black), stickleback that developed at 18 °C and were acclimated to 25 °C (red) or acclimated to 5 °C (dark blue), and stickleback with a developmental temperature of 24 °C (pink), or 12 °C (light blue). Width indicates the pooled distribution density of percent methylation of CpG loci in a given treatment. Embedded box plots summarize variation in the median methylation level across the six samples in each treatment. The line indicates the median of these medians, the box defines the interquartile range (IQR), and the whiskers represent the maximum and minimum values, excluding values greater than 1.5x IQR (which are shown as individual points). (B) Cumulative distribution frequency plot of pooled DNA methylation levels for each treatment. Colors as in panel A and Figure 2.1. ....................................................... 46	Figure 2.4: Venn Diagram depicting the number of common and unique differentially methylated regions identified between stickleback that were developed at and acclimated to 18 °C and stickleback that developed at 12 °C (dashed light blue line) or at 24 °C (dashed pink line) and then were acclimated to 18 °C, or that developed at 18 °C and were then acclimated to 5 °C (solid dark blue line) or 25 °C (solid red line). .................................. 48	Figure 3.1: Experimental design. Six stickleback families were generated for use in this experiment. Each clutch was split between three developmental temperatures of 12 °C (blue), 18 °C (black), and 24 °C (red). Following hatch, all stickleback were transferred to aquaria at 18 °C and reared for 8 months. At this point, fish from the 18 °C development treatment were acclimated to either 5 °C (blue shaded aquaria), 18 °C (grey shaded aquaria), or 25 °C (red shaded aquaria) for 4 weeks, while fish from the other developmental treatments were maintained at 18 °C. This experimental design resulted in stickleback with 5 different thermal histories. At an age of nine months, muscle tissue was sampled for RNA-seq analysis (n=6 per treatment). ................................................. 55	Figure 3.2: Differential expression in stickleback acclimated to either 5 °C (A/C) or 25 °C (B/D) relative to stickleback that were kept at 18 °C for the duration of the experiment. Panels A and B show the –log10 FDR corrected p-value (q-value) against log2 fold change for significantly (q<0.05) up-regulated (orange), down-regulated (blue) genes and non- differentially expressed genes (grey). Embedded plots are frequency histograms of log2 fold change (e.g. a bar above a log2 fold change of 2 represents the genes that range in a log2 fold change between 1 and 2). Panels C and D are heatmaps of differential expression in stickleback acclimated to 5 °C and 25 °C, respectively. Each row represents the expression value (log2 counts per million) for a single gene relative to the mean expression value for that gene across all individuals (orange representing higher expression and blue representing lower expression). Each column represents an individual stickleback. Columns 1-6 are stickleback that were kept at 18 °C for the duration of the experiment. Columns 7-12 are either stickleback that were kept at 18 °C for 8 months and then acclimated to 5 °C (C) or 25 °C (D) for four weeks. .................... 63	 xiii  Figure 3.3: Differential expression in stickleback exposed to either 12 °C (A/C) or 24 °C (B/D) during development and then kept at 18 °C until 9 months of age relative to stickleback that were kept at 18 °C for the duration of the experiment. Panels A and B show the –log10 FDR corrected p-value (q-value) against log2 fold change for significantly (q<0.05) up-regulated (orange), down-regulated (blue) genes and non- differentially expressed genes (grey). Embedded plots are frequency histograms of log2 fold change (e.g. a bar above a log2 fold change of 2 represents the genes that range in a log2 fold change between 1 and 2). Panels C and D are heatmaps of differential expression in stickleback exposed to 12 °C and 24 °C during development, respectively. Each row represents the expression value (log2 counts per million) for a single gene relative to the mean expression value for that gene across all individuals (orange representing higher expression and blue representing lower expression). Each column represents an individual stickleback. Columns 1-6 are stickleback that were kept at 18 °C for the duration of the experiment. Columns 7-12 are either stickleback that were kept at 18 °C for 8 months and then acclimated to 5 °C (C) or 25 °C (D) for four weeks. .................... 65	Figure 3.4: (A) Comparison of up- and down-regulated genes in stickleback muscle tissue from adult stickleback acclimated as adults to 5 °C (blue) or 25 °C (red) for four weeks. (B) Comparison of up- and down-regulated genes in stickleback muscle tissue from stickleback exposed to 12 °C (blue) or 24 °C (red) during development and then kept at 18 °C until 9 months of age. (C) Comparison of differentially expressed genes in stickleback muscle tissue. Cold acclimation = fish developed and reared at 18 °C for eight months and then acclimated to 5 °C for four weeks. Warm acclimation = fish developed and reared at 18 °C for eight months and then acclimated to 25 °C for four weeks. Cold development = fish developed at 12 °C until hatch, and then held at 18 °C for nine months. Warm development = fish developed at 24 °C until hatch, and then held at 18 °C for nine months. All differential expression was identified relative to fish held at 18 °C for the duration of the experiment. ......................................................................... 67	Figure 3.5: Genomic distribution of differentially methylated cytosines associated with differentially expressed genes (within 5 kilobase pairs) in stickleback muscle tissue from adult stickleback acclimated to 5 °C (cold acclimation) or 25 °C (warm acclimation). ... 73	Figure 4.1: Mean CpG methylation level across chromosomes 4, 10, 11, 12, and 16. Each point represents the mean methylation level across all twelve individuals for a single 10 kb window. The solid line represents the smoothed spline fit to these data. Position along the x-axis represents the base position along the chromosome. The y-axis is the average DNA methylation level. .............................................................................................................. 89	Figure 4.2: Differentially methylated CpG loci between male and female stickleback. .............. 91	  xiv  Figure 4.3: Differential methylation between sexes on chromosome 19 (chr19). (A) Mean DNA methylation levels for CpG loci along chromosome chr19. Each point represents the mean DNA methylation level in a 10 kb window for six individual stickleback that were either male (blue) or female (red). Solid lines represent the smooth spline fit for the DNA methylation levels in males (blue) and females (red). (B-C) Proportion of DMCs along chr 19 that are hypermethylated (B) or hypomethylated (C) in female stickleback compared to male stickleback. Values on the y-axis represent the total number of DMCs in a 10 kb window relative to the number of CpG loci in that same 10 kb window. (D) Ratio of mapped read counts for males relative to females along chr19. Each point represents the ratio of mean counts for a 10kb window in males compared to females. The solid black line is the smooth spline fit. Vertical dashed lines represent the boundaries between the three evolutionary strata on chr19: the pseudo autosomal region (PAR), stratum two, and stratum one. ............................................................................... 93	Figure 4.4: Differentially methylated CpG loci between stickleback reared a low and high salinity. (A) Differentially methylated CpG (DMC) loci between stickleback reared at a salinity of 2 ppt compared to 21 ppt. Each point represents an individual DMC. The y-axis indicates the percent difference in methylation between salinity rearing treatments. A positive value on the y-axis indicates a DMC this is hypermethylated in stickleback reared at 21 ppt relative 2 ppt. A negative value on the y-axis indicates a DMC that is hypomethylated in stickleback reared at 21 ppt relative to 2 ppt. The x-axis indicates the position of the DMC in the stickleback genome. Chromosome boundaries are represented by vertical dashed lines. Only DMCs for which a change in methylation of > 10 % are presented. (B) The percentage of CpG loci on a given chromosome that were detected as being differentially methylated between stickleback reared at 21 ppt and 2 ppt. The light shading represents DMCs that are hypermethylated and dark shading represents DMCs that are hypomethylated in stickleback reared at 21 ppt compared to 2 ppt respectively.......................................................................................................................................... 101	Figure 5.1: Differential expression and principal component analyses reveal sexually dimorphic effects of maternal stress in threespine stickleback offspring. (A) Venn diagram depicting the total number of differentially expressed genes for which a main effect of sex (purple), maternal stress (orange), or an interaction between sex and maternal stress (green) was detected. (B) Principal components 1 and 2 of expressed genes in the brain tissue of males (squares) and females (circles) from unstressed (open symbol) and stressed (closed symbol) mothers. ............................................................................................................. 108	Figure 5.2: Heat map of all differentially expressed (DE) genes for which a significant interaction between sex and maternal stress was identified. Expression results are displayed a log2 counts per million normalized to the mean expression for all samples. Blue indicates lower expression and yellow indicates higher expression. ..................... 109	  xv  Figure 5.3: Cluster analysis of the significantly enriched gene ontology (GO) terms for biological process (A) and cellular component (B) over-represented among genes that increased in males (blue circles) or in females (red circles) from stressed mothers compared to unstressed mothers. Size of the circle is representative of the total number of DE genes in that GO category. Note that some names have been omitted from the figure due to semantic similarity. ......................................................................................................... 111	Figure 5.4: Histograms representing the total number of differentially expressed genes localized to each of the 21 stickleback chromosomes and the mitochondrial genome. (A) Genes for which a significant effect of offspring sex was detected. (B) Genes for which significant interaction between offspring sex and maternal stress was detected. Blue bars represent genes that are up-regulated in males (A) or are expressed at higher levels in males from stressed mothers compared to females from stressed mothers (B). Red bars represent genes that are up-regulated females (A) or are expressed at higher levels in females from stressed mothers compared to males from stressed mothers (B). ................................... 113	   xvi  List of Abbreviations  14:10 h 14 hours of light followed by 10 hours of dark 5caC  5-carboxyl-cytosine 5fC  5-formyl-cytosine 5hmC  5-hydroxy-methyl cytosine 5mC  methylated cytosine nuceotide bp  base pairs BSradSeq bisulfite restriction site associated sequencing C  cytosine nucleotide °C  degrees celsius cDNA  complementary DNA sequence CH3  methyl group chr  chromosome cm  centimeter CpG  cytosine-phosphate-guanine dinucleotide CpHpG cytosine-phosphate-(A,C or T nucleotide)-phosophate-Guanine nucleotide CpHpH cytosine-phosphate-(A,C or T nucleotide)-phosophate--(A,C or T      nucleotide) nucleotide DE  differentially expressed DMC  differentially methylated cytosine DMR  differentially methylated region DNA  deoxyribonucleic acid DNMT DNA methyltransferase E2  estradiol e.g.  exempli gratia endo-siRNA endogenous small interfering RNA EpiRADseq methylation-sensitive restriction site associated DNA sequencing ESD  environmental sex determination et al.  et alia F  filial generation FDR  false discovery rate G  guanine nucleotide gDNA  genomic DNA Gb  giga base pairs GO  gene ontology GPS  global positioning system GSD  genotypic sex determination HBCD  hexa-bromo-cyclododecane HPI  hypothalamic-pituitary-interrenal  HPLC  high-performance liquid chromatography i.e.  id est IQR  interquartile range xvii  K+  potassium kb  kilobase pair KEGG  Kyoto encyclopedia of genes and genomes K-S  Kolmogorov-Smirnov L  litre Ma  mega-annum MBDSeq Methyl-CpG binding domain protein sequencing MeDIP-seq methylated DNA immunoprecipitation sequencing mg  milligrams miRNA micro RNA mL  milliliter µL  microliter mOsm  milliosmoles mRNA  messenger RNA MS-AFLP methylation-sensitive amplified fragment length polymorphism n  number Na+  sodium NIH  National Institute of Health NKA  sodium potassium ATPase PAR  pseudo autosomal region PC  principal component PCR  polymerase chain reaction piRNA  PIWI-interacting ribonucleic acids ppt  parts per thousand RIN  RNA integrity number RLE  relative log expression RNA  ribonucleic acid RNA-Seq RNA-sequencing RP-HPLC reverse phase high-performance liquid chromatography RRBS  reduced representation bisulfite sequencing s  seconds SD  standard deviation SLIM  sliding linear model SNP  single nucleotide polymorphism T  thymine nucleotide TET  ten-eleven translocation enzyme TpG  cytosine-phosphate-thymine dinucleotide TSD  temperature dependent sex determination U  uracil nucleotide UBC  University of British Columbia W  W chromosome WGBS  whole genome bisulfite sequencing X  X chromosome Y  Y chromosome Z  Z chromosome xviii  Acknowledgements   I would like to thank my supervisor Dr. Patricia (Trish) Schulte for her mentorship, patience, and support. This PhD project has exposed me to many new and unexpected areas of science and Trish was a beacon of enthusiasm as we learned about the world of epigenomics together. I would also like to thank my committee members Drs. Jeffery Richards, Dolph Schluter, and Bob Devlin for their positive encouragement and thoughtful advice. I am also grateful to the members of Schulte lab for enduring the hours of lab meetings spent discussing epigenetics and bioinformatics, for helping to keep my fish alive, and for providing a fun and diverse environment in which to work.  I would also like to take this opportunity to thank not only those that have supported me during my PhD but to also thank all of my teachers and mentors that have helped me get to this point. I am fortunate to have experienced working in many different laboratories with many excellent people who have all played important roles in shaping my views on scientific research.   Finally, completing this degree would not have been possible without the support of my family and friends. The trips back east, weekends at Galiano, time in the mountains, and nights spent sharing bottles filled with the wonders of fermentation kept my batteries fully charged. I am, and will forever be, grateful for their unwavering support.     1  Chapter 1: Introduction  1.1 Goals of this thesis Changes in gene expression regulate many of the cellular processes that allow organisms to respond to environmental change (Gracey 2007; Somero 2010). Epigenetic mechanisms, such as DNA methylation, have a well characterized role in regulating gene expression activity (Jaenisch & Bird 2003) yet we know very little about how these mechanisms are involved in responses to environmental change, particularly in an ecologically relevant context (Bossdorf et al. 2008; Verhoeven et al. 2016). The goal of my thesis was to examine plasticity of DNA methylation patterns and gene expression across different timescales in response to environmental change. Using the threespine stickleback (Gasterosteus aculeatus) as a model system, I examined how temperature, salinity, and maternal stress affect DNA methylation and gene expression patterns because these are ecologically relevant stressors in this species. By examining the effects of these environmental factors on DNA methylation and gene expression, the studies outlined in this thesis contribute to our fundamental understanding of epigenetic processes in fish by 1) characterizing how DNA methylation and gene expression are regulated in response to changes in the environment, 2) investigating the persistent effects of developmental conditions on DNA methylation and gene transcript levels, and 3) assessing the effects of sex on differences in DNA methylation patterns among individuals. In the remainder of this chapter, I provide the background required to place my studies in the context of the biological literature and the current state of understanding of the role of epigenetic mechanisms in responses to environmental change, with particular reference to our knowledge of these processes in fish. 2   1.2 Epigenetics and plasticity  Environmentally induced phenotypic plasticity is a critical component of organismal responses to a changing environment (Ghalambor et al. 2007; Donelson et al. 2017). Understanding the molecular mechanisms underlying this plasticity is an important question in basic biology that also has applications in areas as diverse as aquaculture and environmental monitoring (Somero 2012; Pittman et al. 2013). Much of the work on the mechanisms of environmentally-induced plasticity in fishes has focused on the reversible phenotypic plasticity that is characteristic of acclimation/acclimatization responses, with many studies using techniques such as cDNA microarray to examine changes in RNA levels in response to environmental change (Gracey et al. 2004; Cossins 2006; Kalujnaia et al. 2007; Logan & Buckley 2015). Less well understood are the longer-lasting phenotypic changes that occur in response to environmental changes that are experienced during development, a phenomenon known as developmental plasticity (Pfennig et al. 2010; Scott & Johnston 2012; Alvarado et al. 2015). Some of these acquired phenotypes can even last for multiple generations, in what has been termed transgenerational plasticity (Agrawal et al. 1999; Salinas et al. 2013).   The mechanisms that regulate these longer-lasting forms of plasticity are often grouped together under the umbrella term of “epigenetics”. Much of the research on epigenetic mechanisms has been performed in model systems (Bird 2002; Li & Zhang 2014), but recent epigenetic studies in ecologically relevant non-model organisms have begun to demonstrate the promise of epigenetic and epigenomic approaches in fields such as ecology, evolutionary biology and environmental biology (Bossdorf et al. 2008; Ledon-Rettig 2013; Bonasio 2015).   3   1.2.1 What is epigenetics?  Conrad Waddington was first credited with using the term “epigenetics” in the 1940s to describe the developmental mechanisms that can give rise to alternative phenotypes (Waddington 1942), but since this time there has been almost continuous refinement of this concept (Deans & Maggert 2015). More contemporary definitions are intrinsically tied to aspects of phenotypic plasticity, or how a single genotype can produce multiple phenotypes in response to changes in the environment. Thus, the broadest definitions of epigenetics include any process that results in heritable phenotypic changes such as variation in gene expression. Several attempts have been made, however, to refine the definition of epigenetics in a way that distinguishes epigenetic mechanisms from broader processes that regulate gene expression and phenotypic plasticity. The most widely accepted contemporary definitions of epigenetics echo the one put forward by Deans and Maggert (2015) who define epigenetics as “The study of phenomena and mechanisms that cause chromosome-bound, heritable changes to gene expression that are not dependent on changes to DNA sequence.” In this context, the word heritable is defined as involving both meiotic and mitotic inheritance (Wu & Morris 2001), and thus according to these definitions, epigenetic mechanisms need not be confined to processes that are inherited across generations, but must at least be inherited across cell divisions. However, not all epigeneticists subscribe to this definition. For example, the field of epigenomics (or the study of epigenetics at a whole-genome level) tends to follow a broader definition that includes additional forms of phenotypic plasticity. For example, the NIH epigenomic roadmap program uses the following definition: “epigenetics refers to both heritable changes in gene activity and expression (in the progeny of cells or of individuals) and also stable, long-term alterations in the 4  transcriptional potential of a cell that are not necessarily heritable” (http://www.roadmapepigenomics.org/overview).   Because there is still disagreement about how to most appropriately define the field of epigenetics, the mechanisms that constitute an epigenetic signal or process are also controversial (Schaefer & Nadeau 2015; Deans & Maggert 2015). Depending on the author, the list of potential epigenetic mechanisms may include covalent modification of DNA through methylation of cytosines (Lister et al. 2009), covalent modification of histones (Bannister & Kouzarides 2011) via acetylation, phosphorylation, methylation and other mechanisms, as well as gene regulation by non-coding RNAs (as microRNAs (miRNA), endogenous small interfering RNAs (endo-siRNA), and PIWI-interacting RNAs (piRNA) (Costa 2008).   1.3 Ecological epigenetics  The field of “ecological epigenetics” aims to address the potential role of epigenetic mechanisms in ecological and evolutionary processes (Bossdorf et al. 2008; Kilvitis et al. 2014). In an ecological context, epigenetic mechanisms could be involved in regulating phenotypic plasticity in response to changes in the environment (Bossdorf et al. 2008; Castonguay & Angers 2012; Zhang et al. 2013) and these effects can persist across a wide range of timescales. The persistence of epigenetic marks could have heritable effects on fitness that are not entirely dependent on the underlying DNA sequence, and thus epigenetic changes have the potential to influence microevolutionary processes (Bossdorf et al. 2008).   The field of ecological epigenetics is closely parallel to that of ecological genomics (Figure 1.1). A key objective of both fields is to understand how heritable variation (i.e. epigenetic or genetic) translates into phenotypic variation and fitness differences on which 5  selection can act and influence evolutionary trajectories (Bossdorf et al. 2008; Kilvitis et al. 2014; Verhoeven et al. 2016). The major difference between epigenetic and genetic processes is that epigenetic processes can be directly modified in response to environmental signals (e.g. Verhoeven et al. 2010; Dowen et al. 2012), which adds an additional level of complexity to understanding the role epigenetic processes in ecology and evolution. While genetic variation can also accumulate in populations from cellular and environmental processes (i.e. genetic recombination and DNA damage), changes to epigenetic processes can occur much more rapidly and could therefore provide an accelerated pathway for evolutionary change (Kilvitis et al. 2014). For example, epigenetic variation could facilitate the exploration of phenotypic space and maintain a particular phenotype as it becomes integrated into the genome through genetic assimilation (Verhoeven et al. 2016). However, ecological epigenetic and ecological genomic processes are not mutually exclusive. Broadly speaking, epigenetic mechanisms have the potential to regulate the rate at which genetic mutations accumulate in the genome and can regulate the expression of genetic variation while the DNA sequence can determine which regions of the genome are affected by epigenetic processes (Verhoeven et al. 2016).  Figure 1.1: Flow diagram illustrating parallel nature of ecological genetics (black lines) and ecological epigenetics (grey lines) (adapted from Bossdorf et al. 2008). Phenotypic variation Gene expression Natural selection Genetic variation Ecological interactions Ecological Genetics Epigenetic variation Ecological Epigenetics Through Alters Affects Affects Affects Alter Alters Affects 6   Despite the widely acknowledged potential for epigenetic mechanisms to be important in an ecological and evolutionary context (Jablonka & Lamb 2005; Bossdorf et al. 2008; Ledon-Rettig 2013; Burggren & Crews 2014; Verhoeven et al. 2016; Hu & Barrett 2017), there is very little empirical evidence to support the ecological and evolutionary roles of epigenetic mechanisms. Most of what is currently known about how epigenetic processes are regulated and how epigenetic processes are involved in regulating cellular functions have come from studies in model systems investigating the role of epigenetics in the development of chronic diseases such as cancer. Given the limited amount of empirical evidence regarding the role epigenetic mechanisms in ecology and evolution, there are several key questions that should be addressed in order to begin to understand the potential ecological and evolutionary role of epigenetic processes including how different environments effect epigenetic variation, how epigenetic variation is influenced by genetic variation and vice versa, and how epigenetic variation influences phenotypes (from gene expression to whole animal phenotypes). My thesis aims to begin to address these issues.  1.4 DNA methylation as an epigenetic mark  DNA methylation, which fulfills the criterion of being a chromosome-bound, potentially heritable change that is not dependent on a change to DNA sequence, was the first epigenetic mark to be discovered and has become one of the most widely studied of the potential epigenetic mechanisms (Deans & Maggert 2015). DNA methylation is a covalent modification in which a methyl group (CH3) is added to position 5 of the pyrimidine ring of a cytosine (5mC). In eukaryotic cells, the majority of methylated cytosines occur on cytosine-phosphate-guanine 7  (CpG) dinucleotides resulting in complementary methylation of CpG motifs on both strands of DNA. Methylated cytosines can also occur on non-CpG sites such as CpHpG and CpHpH motifs (where H = A, C or T). However, these epigenetic marks are fundamentally different from CpG methylation because they are strand-specific. These strand-specific marks are rare in animals, although they are prominent in the epigenetic landscape of plants (Law & Jacobsen 2010).   The addition or removal of methyl groups can regulate the transcriptional activity of neighbouring genes by altering the structure and function of chromatin. High levels of methylation at the promoter of a gene are associated with repression of transcription (Li & Zhang 2014). This repression of transcription by promoter methylation is thought to be due to the formation of inactive heterochromatin (Razin 1998; Martinowich et al. 2003). In addition, methylation of enhancer elements can also repress transcription by preventing the binding of the transcription factors that would otherwise induce expression. Alternatively, methylated DNA can recruit methyl-binding proteins that repress transcription (Klose & Bird 2006). DNA methylation of intragenic regions, however, is associated with active expression of nearby genes (Zemach et al. 2010; Jjingo et al. 2012) and the regulation of alternative splice variants (Laurent et al. 2010; Lyko et al. 2010; Park et al. 2011).   There is dramatic variation in the prevalence and patterns of DNA methylation among taxa (Feng et al. 2010). For example, in Arabidopsis only ~22% of CpG sites are methylated, while the CpG sites in vertebrate genomes are generally methylated to high levels (>80%). Invertebrates exhibit a wide range of CpG methylation levels, with the “model” invertebrates Drosophila melanogaster and Caenorhabditis elegans essentially lacking DNA methylation and the honeybee having very low CpG methylation levels (~0.7%; Lyko et al. 2010). Many other invertebrates have intermediate levels of CpG methylation (around 50%) with a “mosaic” pattern 8  of methylation such that the genome consists of interspersed regions of methylated and unmethylated DNA of similar length (Feng et al. 2010). In contrast, vertebrate genomes consist of long tracts in which the CpG dinucleotides are methylated to high levels (>80%) punctuated by short unmethylated regions. These unmethylated regions typically fall in areas called CpG islands (CGI), which have a GC-rich base composition with a CpG dinucleotide approximately every 10 base pairs. CpG islands tend to be preferentially located in the proximal promoter regions of genes (Figure 1.2; Deaton & Bird 2011), and this accounts for an observed decrease in the levels of DNA methylation around the transcription start sites of actively transcribed genes (Zemach et al. 2010; Jones 2012).   Figure 1.2: Patterns of CpG island methylation in proximal promoter regions. CpG islands are regions with high concentrations of CpG motifs. Shore and shelf regions are located at immediately flanking CpG islands. A high level of methylation in a CpG island associated with the transcription start site of a gene is associated with inactivation of transcription. Bent arrow indicates transcription start site; empty circles indicate umethylated CpG sites; filled circles indicate methylated CpG sites (reproduced from Metzger & Schulte 2016).   There is also epigenetic variation among individuals within a species in DNA methylation patterns (e.g. Massicotte et al. 2011; Liebl et al. 2013). This individual-level variation in DNA methylation can be maintained throughout the lifespan of an organism and can even be transmitted across generations in at least some species (e.g. Johannes et al. 2009; Verhoeven et al. 2010), providing an added source of heritable variation that is independent of genetic variation. Thus, the potential impact of epigenetic variation has become an important topic in the fields of ecology and evolution (Bossdorf et al. 2008; Ledon-Rettig 2013), as it has CpG	Island	Promoter	CpG	Island	Inac2ve	Shore	Shelf	Ac2ve	9  been suggested to be important for evolutionary adaptation to altered environments (Flores et al. 2013).  DNA methylation patterns can also vary within a single individual across life stages or among cell types (Richardson 2003; Flanagan et al. 2006; Massicotte et al. 2011). DNA methylation during development is thought to play an important role in cellular differentiation and is involved in maintaining cell-type specific transcriptional activity in the genome (Monk et al. 1987; Li 2002). Variation in DNA methylation patterns that are induced by environmental factors can alter the developmental trajectory of an individual resulting in the expression of alternative phenotypes from a single genotype (i.e. developmental plasticity) (e.g. Wolff et al. 1998; Kucharski et al. 2008; Shao et al. 2014).   Most of what is known about the mechanisms that establish and maintain DNA methylation patterns has been derived from studies in mammals. In mammals, DNA methylation patterns are established by the de novo DNA methylation activity of the DNA methyltransferases, DNMT3A and DNMT3B, which add methyl groups to previously unmethylated cytosines (Law & Jacobsen 2010). Maintenance of DNA methylation across cell divisions is achieved by the addition of methyl groups to the newly synthesized strand of DNA by the DNA methyltransferase DNMT1 which is recruited to DNA by the SET- and RING-associated domain contain protein UHRF1 (Sharif et al. 2007; Bostick et al. 2007). Two other genes in the mammalian genome are related to the described DNA methyltransferase genes but do not possess DNA methyltransferase activity. DNMT3L is a methylation cofactor that has been associated with genomic imprinting (Suetake et al. 2004). Another methyltransferase, DNMT2, predominantly methylates RNA molecules, specifically tRNA, as opposed to DNA (Goll et al. 2006). 10   While the addition of DNA methylation marks in the genome is an active process, de-methylation can occur via both active and passive processes. Passive de-methylation occurs when DNA methylation is not maintained through DNA replication and cell division as a result of the lack of DNMT1 activity (Li & Zhang 2014). Active de-methylation requires the recruitment of protein complexes that convert methylated cytosines to the methylation intermediates that are then removed by base excision repair mechanisms. The intermediate forms of methylated cytosines include 5-hydroxy-methyl cytosine (5hmC), 5-formyl-cytosine (5fC), and 5-carboxyl-cytosine (5caC) (Shen et al. 2014). A family of enzymes known as the ten-eleven translocation (TET) enzymes catalyzes the stepwise oxidation of 5mC to 5fC and then 5caC, which can then be converted to unmethylated cytosines (Shen et al. 2014). These oxidized derivatives were originally thought to be transient demethylation intermediates, but recent studies suggest that these marks are more persistent in the genome than originally thought, and may have functional significance in their own right (Branco et al. 2011; Véron & Peters 2011; Kroeze et al. 2015).  1.4.1 Methods for assessing DNA methylation  There are a number of reviews summarizing the major techniques available for assessing levels of DNA methylation (Shen & Waterland 2007; Laird 2010; Harrison & Parle-McDermott 2011; Umer & Herceg 2013; Plongthongkum et al. 2014; Metzger & Schulte 2016b). In general, different techniques offer different levels of resolution from low-resolution information about whole-genome methylation levels to techniques that provide information about methylation status at the single base-pair level. Each of these approaches has advantages and disadvantages which are summarized in Table 1.1. Choosing the method that is best suited for a particular study 11  generally comes down to making a complex trade-off that weighs the cost, the resolution of the assay, and the extent of the genome that is monitored. In this thesis I use techniques that provide DNA methylation information at single base-pair resolution. In the following section I briefly describe different methods that can be used to obtain DNA methylation information at the level of individual nucleotides and discuss the advantages and disadvantages of the different methods.  Most methods for measuring DNA methylation with single base-pair resolution rely on bisulfite conversion of genomic DNA. When DNA is treated with sodium bisulfite or metabisulfite, the unmethylated cytosine bases (C) in the DNA are chemically converted into the base uracil (U), whereas methylated cytosines (5mC) are protected from this bisulfite conversion. During subsequent PCR amplification, a thymine (T) base is incorporated at any position that has been converted to a U, which changes the original sequence from a C to a T in unmethylated positions, whereas methylated positions remain as C’s (Figure 1.3).   Figure 1.3: Bisulfite sequencing. Treatment of genomic DNA with sodium bisulfite converts unmethylated cytosines (C) into uracil (U), while methylated cytosines are protected. During PCR and sequencing uracil is converted to thymine (T). Sequence comparison of untreated and bisulfite converted DNA sequences allows the detection of methylated sites (reproduced from Metzger & Schulte 2016).  By comparing the bisulfite converted and untreated genomic DNA sequences it is possible to detect the location of methylated cytosines. Individual cells within a tissue can vary in whether a G A TC GT G ATCme	Bisulfite	conversion	Sequence	 Sequence	G A TC GT G ATUme	Compare	Sequences	G A T C G T C T A G G A T C G T T T A G Methylated	 Unmethylated	G A TC GT G ATTG A TC GT G ATCOriginal	Bisulfite	converted	12  particular position is methylated or not, and the initial DNA isolation step that is the starting point of bisulfite conversion is usually a sample that contains multiple cells. Thus, results from these assays provide information on the percentage of cells that have a methylated cytosine at a particular position.   The most comprehensive method for assessing patterns of DNA methylation at single base-pair resolution is bisulfite conversion of genomic DNA followed by whole-genome shotgun sequencing (whole-genome bisulfite sequencing WGBS; Laird 2010). However, the cost of comprehensive sequencing remains prohibitive for organisms with large genomes, such as most fishes, and thus WGBS is usually only applied to questions that can be answered using small numbers of biological replicates. For example, current recommendations are that 10X genome coverage and at least two biological replicates should be used at a minimum for bisulfite sequencing (Ziller et al. 2015). As the costs of genome sequencing decrease, however, WGBS has become more accessible, and this technique is likely to have much wider application. An additional barrier to applying this method stems from the challenges of the associated bioinformatics, which can be difficult even in model organisms (Bock 2012; Adusumalli et al. 2015) and are a particular barrier in species that lack a sequenced genome, such as many fishes. However, there has been some recent development of bioinformatics methods that should allow the application of these techniques even in the absence of a sequenced genome (Bewick et al. 2015).  A variety of methods have been devised to mitigate the high cost of WGBS. In general, these methods reduce costs by focusing sequencing effort on the methylated portions of the genome. This enrichment is performed either using affinity-based or restriction-enzyme based approaches. Reduced representation bisulfite sequencing (RRBS; Jeddeloh et al. 2008; Gu et al. 13  2011) is a restriction-enzyme based approach where genomic DNA is first digested using a methylation-insensitive restriction enzyme such as MspI that specifically target CpG motifs. These fragments are then sequenced. This allows sequencing efforts to be concentrated primarily on regions of the genome that have the potential to be methylated (Jeddeloh et al. 2008; Gu et al. 2010). Although the RRBS technique helps to mitigate the cost of sequencing per individual, the downstream analysis of RRBS data in the absence of a reference genome can be complicated. However, there are currently a variety of tools that are available to help perform these analyses in species without a reference genome (e.g. Chen et al. 2010; Stockwell et al. 2014).                                14   Table 1.1: Advantages and disadvantages of selected methods for assessing DNA methylation Method Resolution Advantages Disadvantages Reverse Phase HPLC (RP-HPLC) Low (whole genome) • No sequence information required • Inexpensive • Only very large differences can be detected • Very low resolution ELISA Low (whole genome) • No sequence information required • High throughput • Inexpensive  • Only very large differences can be detected • Very low resolution MS-AFLP Low-Moderate • No sequence information required • Inexpensive • Some information about location of methylated sites can be obtained (if candidate loci sequenced) • Low resolution • Provides +/- information (qualitative) • Screens anonymous loci (candidates can be sequenced) • Reproducibility can be poor Bisulfite sequencing of a candidate locus Single base pair • All CpG sites in targeted region can be analyzed • Generates information about location of methylated sites • Inexpensive   • Low coverage of the genome  • Low throughput • Identifying appropriate candidate regions is challenging • Sequence information is required for primer design EpiRADseq Single base pair • High resolution • Quantitative • Genome-wide (~ 2 million CpG sites examined per sample; ~6% of total sites)*  • Expensive • Requires genomic sequence information • Only examines methylation at HpaII sites • Fully methylated sites are not captured Reduced Representation bisulfite sequencing (RRBS) Single base pair • High resolution • Quantitative • Genome-wide (~3 million CpG examined per sample; ~10% of total sites)* • Expensive • Requires genomic sequence information • Only examines sites within ~100bp of an enzyme recognition site MeDIP-Seq/MBDSeq 100-300bp (single base pair when combined with bisulfite sequencing) • Fairly high resolution • Quantitative • Genome-wide (~6 million CpG examined per sample; ~20% of total sites)* • Expensive • Requires genomic sequence information • Quantification complex  Whole-genome bisulfite sequencing (WGBS) Single base pair • High resolution • Quantitative • Genome-wide (all CpG sites examined) • Extremely expensive  • Requires sophisticated bioinformatics • Requires assembled genome sequence *Based on a 3Gbp genome size  15  1.5 DNA methylation in fish  Much of what is currently understood about mechanisms and patterns of DNA methylation in fish comes from studies conducted in zebrafish (Goll & Halpern 2011; Kamstra et al. 2015). Proteins involved in DNA methylation in mammals are also present in fish including the DNA methyltransferases (DNMTs), CXXC finger protein 1 (cfp1), methyl binding proteins mecp2 and mbd2, the Piwi-related proteins ziwi and zili, and the dnmt1 cofactor uhrf1, which is required for the maintenance of DNA methylation patterns (Goll & Halpern 2011).  Overall levels of DNA methylation appear to be similar among vertebrates, with 80% of CpG dinucleotides in the larval zebrafish genome being methylated compared to about 74% in the mouse (Feng et al. 2010). Similarly, as is the case for other vertebrates, zebrafish genomes have prominent unmethylated CpG islands located near the transcription start sites of actively transcribed genes. Differential methylation of distal enhancer elements has also been shown to play a biologically important role in zebrafish (Lee et al. 2015). The de-methylation intermediate 5hmC has also been observed in zebrafish, with the highest levels being observed in brain tissue (Kamstra et al. 2015). Non-CpG methylation, such as CpHpG and CpHpH, has also been detected in zebrafish but at low levels (2.13% in embryos) and these sites are evenly distributed throughout the genome. By comparison, non-CpG methylation in mouse embryos is 0.59% (Feng et al. 2010).  In mammals, methylation patterns are thought to be globally erased and reset soon after fertilization. Whether or not this occurs in zebrafish is not clear. The earliest studies of this phenomenon in zebrafish yielded contradictory results (Macleod et al. 1999; Mhanni & McGowan 2004), whereas the most recent work suggests that intermediate levels of erasure occur (Jiang et al. 2013; Potok et al. 2013). Zebrafish sperm DNA is hypermethylated (95%) 16  relative to oocyte DNA (75%). Following fertilization, 16-cell embryo DNA methylation levels are intermediate to those of sperm and oocytes, but whether this is the result of erasure is unclear (Jiang et al. 2013; Potok et al. 2013). The lack of clarity regarding the level of erasure of methylation may be due, at least in part, to the asynchronous development of zebrafish oocytes (Kimmel et al. 1995) and the associated difficulty of obtaining enough developmentally synchronized embryos to obtain the quantity of genomic DNA necessary for this type of analysis. Perhaps using a fish species in which more synchronous fertilization can be achieved would allow for a more detailed analysis of whole genome DNA methylation levels earlier in development.   After the 16-cell stage, whole-embryo DNA methylation levels increase at each cell division, approaching the methylation levels observed in sperm DNA by the time the embryo reaches gastrulation. It has been proposed that patterns of DNA methylation in developing embryos are reset to the patterns observed in sperm. However, parthenogenic fertilization of zebrafish oocytes results in the same whole-genome DNA methylation patterns as traditional fertilization (Potok et al. 2013), suggesting that components within the oocyte provide all of the information required to establish methylation patterns in the offspring. It is possible that the demethylated state of the oocyte is indicative of a more totipotent state of the oocyte compared to the more differentiated state of the embryo, with increased whole genome methylation levels. Alternatively, piRNAs could survive the UV irradiation treatment of sperm DNA prior to gynogenic fertilization acting as the primary regulators of DNA methylation reprograming to a paternal pattern (Potok et al. 2013). Whether the changes in methylation levels and patterns during development are conserved across fishes or are unique to zebrafish remains largely unexplored. 17  1.5.1 Effects of environmental stressors on DNA methylation in fish  Epigenomic changes have the potential to play an important role in the response of fishes to environmental change, and may influence adaptation to new environments (Franks & Hoffmann 2012). Understanding responses to environmental change is of increasing importance in the context of global environmental change and other human impacts on aquatic environments. For example, multiple studies have shown that the DNA methylation patterns of fishes change following exposure to environmental toxicants (Timme-Laragy et al. 2005; Aniagu et al. 2008; Wang et al. 2009; Pierron et al. 2014), highlighting an underappreciated impact of environmental toxicants on organismal biology. In the remainder of this section, I focus on what is known about the effects of temperature and salinity on DNA methylation in fishes because these environmental factors are directly assessed in this thesis.  Changes in environmental temperature has pervasive effects on the biology of fishes, yet there are surprisingly few studies investigating its effects on the epigenome (Varriale & Bernardi 2006; Campos et al. 2013; Shao et al. 2014). A particularly intriguing study (Varriale & Bernardi 2006) found that whole-genome methylation levels (as assessed using RP-HPLC) were inversely correlated with temperature, with the genomes of fish from cooler habitats exhibiting higher levels of whole-genome DNA methylation compared to fish from warmer habitats. Several subsequent studies have also identified temperature as an important factor that may modify methylation levels in fish (Campos et al. 2013; Shao et al. 2014; Anastasiadi et al. 2017). For example, temperature has been shown to affect methylation levels in the promoter of the myogenin gene in Senegalese Sole (Solea senegalensis), as detected using candidate-gene bisulfite sequencing (Campos et al. 2013). These changes were correlated with changes in myogenin expression and muscle cellularity, clearly demonstrating the linkages between changes 18  in DNA methylation and important organismal phenotypes. In Atlantic salmon, changes in the expression of genes known to be involved in DNA de-methylation have also been demonstrated in response to changes in embryonic temperature and myogenin expression (Burgerhout et al. 2017).   Similar to studies investigating the effects of environmental temperature on DNA methylation patterns, most of the studies conducted to date that have investigated the effects of environmental salinity on DNA methylation patterns have used low resolution techniques (e.g. Morán et al. 2013; Li et al. 2017). From these studies, it is clear that fish modify genomic DNA methylation levels in response to changes in environmental salinity; however, it is unknown which regions of the genome are affected by these changes in DNA methylation levels. A candidate gene study that measured DNA methylation levels in the igf1 gene in the liver tissue of half smooth tongue sole (Li et al. 2017) found an inverse correlation between DNA methylation levels of the igf1 gene and expression of igf1 7 days after transfer from 30 ppt to 15 ppt and no differences in DNA methylation levels were detected after 30 days suggesting that DNA methylation levels are dynamically regulated in response to changes in environmental salinity.  1.5.2 DNA methylation and sex determination in fish  Gorelick (2003) presents a hypothesis for how differences in DNA methylation patterns could provide a mechanism of sex determination in both environmental sex determination (ESD) and genotypic sex determination (GSD). In ESD, DNA methylation is involved in the sex-specific expression of key genes involved in sex determination such as genes that regulate male and female gonad development. In GSD, DNA methylation is hypothesized to be involved in both suppressing recombination between sex chromosomes by regulation the formation of 19  heterochromatin, and in increasing mutation rates via the deamination of methylated cytosines to thymines. Together, recombination suppression and increased mutation rates could initiate and accelerate the speed of Muller’s ratchet thus leading to the degradation and divergence of the heterogametic sex chromosome pairs (Gorelick 2003).  Fishes demonstrate a wide range of sex-determining mechanisms, including both ESD and GSD, but the details of these sex determining mechanisms remain poorly understood (Devlin & Nagahama 2002; Mank et al. 2006). There is accumulating evidence that DNA methylation likely plays a key role in temperature dependent sex determination (TSD) mechanisms, a form of ESD (e.g. Navarro-Martín et al. 2011; Chen et al. 2014; Shao et al. 2014; Wen et al. 2014). For example, the cyp19a gene is an aromatase gene that converts androgens into estrogens and is thought to play an important role in gonad differentiation. Hypermethylation of the cyp19a gene in male gonad tissue compared to female gonad tissue in the European Sea Bass reduces expression of cyp19a and causes a reduction in the synthesis of estrogen in male gonad tissue (Navarro-Martín et al. 2011). Interestingly, exposure of European Sea Bass larvae to elevated temperatures results in increased methylation of the cyp19a promoter and a masculinization of female gonad tissue.  The role of epigenetics in sex determination has also been investigated in the Half-Smooth Tongue Sole (Cynoglossus semilaevis), which exhibits TSD and GSD. This species has a ZZ/ZW sex determination system in which it has been previously demonstrated that male-specific expression of the Z-linked dmrt1 gene is associated with male gonad development (Chen et al. 2014). Shao et al. (2014) used WGBS to determine whole-genome DNA methylation patterns in male, female, and pseudo-male gonads. They found that genes associated with sex-determination pathways are differentially methylated between male and female gonad tissue. 20  Specifically, the authors discovered that the dmrt1 gene has high levels of DNA methylation in female gonads, whereas ZZ males and ZW pseudo-males were hypomethylated at the dmrt1 locus suggesting that demethylation of dmrt1 is necessary for male gonad development. Interestingly, although ZW pseudo-males can be produced by exposing ZW females to high environmental temperatures, pseudo-males can also give rise to ZW pseudo-male offspring in the absence of a high-temperature environmental cue (Chen et al. 2014). This suggests that there may be a transgenerational epigenetic mechanism of sex determination in this species.   A study by Wen et al. (2014) measured DNA methylation and gene expression levels of both cyp19a and dmrt1 in Japanese Flounder (Paralichthys olivaceus). They found that high levels of dmrt1 expression in male gonad tissue are correlated with low levels of DNA methylation in the dmrt1 promoter, whereas low levels of expression are correlated with high levels of DNA methylation. In contrast, the promoter of cyp19a was hypermethylated in testes compared to ovaries, which corresponded to lower levels of cyp19a mRNA in testis tissue compared to ovaries. Subsequent studies in Japanese flounder have observed similar patterns in which DNA methylation patterns of cyp19a1a and foxl2 are inversely correlated with mRNA expression levels during ovarian development (Si et al. 2016). Taken together, these studies clearly demonstrate that the methylation of the dmrt1 and cyp19a genes is important for sex determination in a variety of fish species.   For species that use GSD, sex is determined by the inheritance of sex specific factors such as male and female specific alleles or heteromorphic sex chromosomes. Different forms of GSD appear to have repeatedly evolved in different fish species, which makes it difficult to conclusively determine the ancestral GSD system in fish (Mank et al. 2006); however, some studies suggest that autosomal sex determination is the ancestral state (Ohno 1967; Traut & 21  Winking 2001). Sex chromosome evolution from autosomes is thought to involve recombination suppression of sex determining regions during meiosis followed by the accumulation of deleterious mutations in these regions leading to degeneration of the chromosome in one sex but not the other (the male Y in an XY sex chromosome system or the female W in a ZW sex chromosome system; Wright et al. 2016; Graves 2016).   In more derived heteromorphic sex chromosome systems (such as those found in mammals), DNA methylation is thought to be involved in regulating gene expression of dosage sensitive genes on heteromorphic sex chromosomes (Graves 2016), and it is also thought to play a role in the initial establishment of heteromorphic sex chromosomes (Gorelick 2003). Differences in DNA methylation levels between sex chromosomes have been hypothesized to play a role in this initial recombination suppression, and there are some data supporting this idea from studies in plants (Zhang et al. 2008). However, nothing is known about the potential role of DNA methylation in the establishment of heteromorphic sex chromosomes in vertebrates. Fish are an excellent model system for studying the evolution of heteromorphic sex chromosomes because various fish species have heterogametic sex chromosomes that are at different evolutionary stages of differentiation. For example, sex determination in the tiger pufferfish (Takifugu rupripes) is controlled by a single missense mutation in the amhr2 gene (Kamiya et al. 2012) with no evidence of suppressed recombination or additional genetic divergence between the proto-X and proto-Y chromosomes. In contrast, threespine stickleback have a heteromorphic XY sex chromosome pair at the early stages of differentiation (Peichel et al. 2004; White et al. 2015). However, to date, no studies have investigated patterns of DNA methylation in young heteromorphic sex chromosome systems.   22  1.6 Research organism: threespine stickleback (Gasterosteus aculeatus)  The development of genomic resources for threespine stickleback has created a powerful system in which to study how genetic variation and phenotypic plasticity contribute to the evolution of adaptive phenotypes, and the evolution of heteromorphic sex chromosomes and sex determination in vertebrates. The utility of stickleback as a model system is due in large part to the ability to capitalize on the extensive efforts to study the natural history of this species and characterize the behavioral and morphological variation among marine and freshwater populations. In this thesis, I examine how three environmental factors that differ between marine and freshwater environments (salinity, temperature, and maternal stress) affected DNA methylation and gene expression patterns in stickleback form a marine population, and explore the relationship between DNA methylation and the evolution of heteromorphic sex chromosomes. Previous studies in stickleback have demonstrated a great deal of plasticity with respect to coping with changes in these environmental factors. In addition, differences in these environmental factors between marine and freshwater environments are thought to impose different selective pressures, which have contributed to the rapid divergence of marine and freshwater stickleback populations. Therefore, examining DNA methylation patterns in stickleback and how transcriptomic and epigenomic mechanisms respond to environmental changes will provide insight into the relationship between epigenomics and phenotypic plasticity in the context of ecologically relevant evolutionary processes. In this section I discuss how threespine stickleback have been used as a model system for studying mechanisms of adaptive evolution and how our current understanding of stickleback evolution can be used as a powerful foundation from which to investigate the ecological and evolutionary roles of epigenetic processes. 23   1.6.1 Threespine stickleback taxonomy and evolution Populations of threespine stickleback (Gasterosteus aculeatus) are found in both marine and freshwater environments throughout their Holarctic distribution. Analysis of mitochondrial DNA, allozyme and microsatellite data indicates that current stickleback populations can be divided into three distinct clades termed the Pacific, Euro-North American, and Japan sea clades (Haglund et al. 1992; Higuchi & Goto 1996; Yamada et al. 2001; Kitano et al. 2007). Post glacial rebound following the retreat of the Pleistocene glaciers ~10-12 thousand years ago resulted in the reproductive isolation of marine stickleback in newly formed freshwater habitats across the northern species range (Haglund et al. 1992; Bell & Foster 1994; Orti et al. 1994; Taylor & McPhail 1999). With the exception of the anadromous Japan sea populations (Orti et al. 1994; Higuchi & Goto 1996; Kitano et al. 2007), much of the phenotypic diversity that has been observed among threespine stickleback populations is found among the freshwater populations that have evolved post glacially (Colosimo et al. 2005; Marchinko 2009).   Reproductive isolation of freshwater populations combined with different selective pressures in freshwater environments compared to the ancestral marine environment has resulted in the rapid, parallel divergence of numerous morphological and behavioral phenotypes between freshwater populations and the ancestral marine populations (Boughman 2007). The amount of phenotypic diversity observed among threespine stickleback populations initially led to the characterization of as many as 30 species of threespine stickleback (Mattern 2007). More recent classifications based on studies using current molecular and phylogenetic tools now classify threespine stickleback as a single species, Gasterosteus aculeatus (Kawahara et al. 2009) stemming from the greek etymology combining “gaster”, meaning stomach, “osteon” meaning 24  bone, and “aculeatus” meaning spiny or prickly, an apt description of their boney lateral plates that are present in various numbers depending on the population, and three prominent dorsal spines. While only a single species of threespine stickleback is officially recognized, populations are often distinguished from one another based on their lateral plate morphology and the environments in which they are found, and these differences were often the basis of earlier classification schemes (eg. G. trachurus for fully plated marine populations and G. leiurus for low plated freshwater populations; Mattern 2007). Reproductively isolated populations of freshwater stickleback have been further resolved into benthic and limnetic sympatric species pairs, which inhabit the littoral and pelagic zones respectively (Boughman 2007). While a comprehensive review of stickleback phylogeny is beyond the scope of this thesis, several excellent books and review papers that have been written on this topic which illustrate the diversity of morphological and behavioral traits that are contained within the threespine stickleback species complex and which have been used to study the parallel evolution of adaptive morphological and behavioral traits (e.g. Bell & Foster 1994; Östlund-Nilsson et al. 2007).   1.6.2 Threespine stickleback genomic tools  One of the most significant contributions that has catapulted threespine stickleback into a prominent vertebrate model systems for studying the molecular basis of adaptive evolutionary processes was the completion of the stickleback genome sequence in 2006 by the Broad Institute. The stickleback genome is ~446 megabases encoding 20, 787 genes assembled into 21 chromosomes and several unassembled scaffolds and it is publically available at the Ensembl genome database (Zerbino et al. 2018). Due to the intense research interest in stickleback genomics and genome evolution, the stickleback genome assembly continues to be updated 25  (Glazer et al. 2015; Peichel et al. 2017). This wealth of genomic information, combined with the extensive literature on the parallel evolution of adaptive phenotypes among threespine stickleback populations resulted in a rapid increase in the number of studies investigating the genetic mechanisms of adaptive evolution in stickleback (e.g. Jones et al. 2012).   1.6.3 Physiological divergence of halotolerance between stickleback populations  Teleosts are osmoregulators and maintain an internal osmolarity of approximately 250-350 mOsm. Transitioning from marine to freshwater environments requires a major shift in osmoregulatory physiology to cope with the challenges posed the hyper-osmotic marine environment with an osmolarity of approximately 1,000 mOsm where stickleback must actively expel ions and drink water in order to balance water lost by diffusion, to a hypo-osmotic freshwater environment with an osmolarity of approximately 0.5-15 mOsm where they must actively uptake ions and expel water (Hill et al. 2008). Threespine stickleback are euryhaline and can tolerate a wide range of both hyper and hypo-osmotic conditions by restructuring the proteins involved in ion transport in the mitochondrial-rich ionocytes of the gill epithelium from conditions that favor ion uptake and inhibit diffusive ion losses in fresh water to ion excretion and water retention in marine environments (Edwards & Marshall 2013). The energy required to actively uptake ions from fresh water or actively expel them in seawater comes from the electrochemical gradient that is established by the basolateral Na+/K+ - ATPase (NKA). Ion co-transporters located on the apical and basolateral membranes of the ionocytes utilize the potential energy created by this electrochemical gradient to transport ions against their electrochemical gradient. The composition of ion transporters and ion transporter isoforms that are expressed by ionocytes can vary among species and can be modified by the individual depending on whether 26  the primary objective is ion-uptake in freshwater environments or ion excretion in marine environments (Richards et al. 2003).  Diffusive ion movement across the large surface area of the gill epithelium creates another potential challenge to ionoregulation in teleosts. To cope with this challenge, fish can alter the expression of proteins involved in the formation of tight junctions to modify the paracellular ion permeability of the gill epithelium. Fish can also reduce the amount of “leakiness” of the gill by modifying the size of the gill interlamellar cell mass, which effectively increases the diffusive distance across the gill and reduces the amount of ion loss due to passive ion flux. Recent studies have shown that this type of gill remodeling occurs in response to a range of environmental stressors including salinity, temperature, oxygen levels, and exercise (Sollid 2005; Mitrovic & Perry 2009; LeBlanc et al. 2010; Brauner et al. 2011; Fu et al. 2011; Nilsson et al. 2012; Perry et al. 2012; Barnes et al. 2014).   Stickleback from marine populations reared in freshwater conditions for several generations exhibit a similar divergence rate in halotolerance (0.569 haldanes; Divino et al. 2016) and cold tolerance (0.63 haldanes; Barrett et al. 2011) that are more rapid than the divergence in many of the morphological traits between marine and freshwater stickleback (Bell & Aguirre 2013). The physiological divergence in halotolerance and thermal tolerance suggests that these abiotic factors impose a strong selective pressure on the physiological mechanisms underlying these traits.  Salinity transfer experiments of marine and resident freshwater stickleback have demonstrated reduced tolerance to hyper-saline conditions in derived freshwater populations (Heuts 1947; Marchinko & Schluter 2007; McCairns & Bernatchez 2010; DeFaveri & Merilä 2014). Consistent with these findings, analysis of marine stickleback that had been reared in 27  freshwater for two generations exhibit increased tolerance to freshwater suggesting that rapid evolution of physiological processes involved in ion regulation (Divino et al. 2016). Alternatively, epigenetic mechanisms and transgenerational plasticity could also be playing an important role in modulating physiological adaptions to freshwater environments.  Recent studies have investigated how changes in environmental salinity between marine populations and resident freshwater populations of stickleback have shaped the physiological evolution of stickleback. These studies have identified regions of genetic divergence that are associated with osmoregulatory genes (Shimada et al. 2011; DeFaveri et al. 2011, 2013; Jones et al. 2012a; Kusakabe et al. 2017), as well as divergence in the regulation of the expression of osmoregulatory genes (McCairns & Bernatchez 2010; Taugbøl et al. 2014; Gibbons et al. 2017; Kusakabe et al. 2017).   1.6.4 Evolution of thermal tolerance in threespine stickleback  Temperature, a measure of average kinetic energy, has profound effects on the performance and distribution of organisms (Fry 1947; Hochachka & Somero 2002; Somero 2005; Sunday et al. 2012). Increases in temperature can reduce the stability of biological structures such as proteins, and increase the rate at which biological reactions occur (Hochachka & Somero 2002), which in turn can influence higher order processes such as metabolism and growth. Therefore, organisms must actively regulate biochemical processes in order to maintain performance in response to changes in temperature.  Biochemical processes of ectothermic poikilotherms, such as most fishes, are particularly vulnerable to changes in environmental temperatures as their body temperature closely matches that of the surrounding environment (Hochachka & Somero 2002; Schulte 2011). Observed 28  variation in thermal tolerance among different species and among different populations of the same species is the result of variation in plasticity and genetic divergence that impact lower level processes such as gene expression and protein function (Dalziel et al. 2009). Stenotherms have a fairly narrow thermal tolerance window while eurytherms can tolerate a wider range of temperatures. Differences in the thermal limits among organisms is often associated with variation in the thermal environment such that organisms that inhabit regions that experience fairly stable thermal conditions, such as polar and equatorial regions, are often stenothermic, while organisms that experience larger variation in environmental temperatures, such as large seasonal changes, are often eurythermic (Stillman 2002; Chown et al. 2004, 2010; Sunday et al. 2011; Clusella-Trullas et al. 2011).  Threespine stickleback are eurytherms, but freshwater populations in temperate regions experience more variation in environmental temperature compared to marine populations (Lee & Bell 1999). Adaptation of marine stickleback to freshwater habitats is associated with a rapid divergence in thermal performance at low temperatures, with freshwater populations having increased tolerance to cold conditions (Barrett et al. 2011). The rapid acquisition of thermal tolerance traits in stickleback may be associated with transgenerational plasticity, as the maternal and grand-maternal environment has been observed to impact thermal performance of offspring (Shama & Wegner 2014; Shama et al. 2016). However, the mechanism(s) through which information from the parental and grand-parental thermal environment is transmitted to subsequent generations is not well understood.   29  1.6.5 Evolution of heteromorphic sex chromosomes in threespine stickleback  In the family Gasterosteidae, in which at least six species are recognized, at least five different sex-determining systems have been identified, two of which are found among populations of G. aculeatus. Threespine stickleback (G. aculeatus) from Pacific and Euro-North American clades which have a relatively young XY sex chromosome pair that has evolved since the species first arose at least ~13-16 Ma (Kawahara et al. 2009; Ross et al. 2009; Bell et al. 2009). In contrast, G. aculeatus from the Sea of Japan have an X1X2Y sex chromosome system which is thought to have been generated from a fusion between chr19, an ancestral Y, with an autosomal copy of chr09 and arose independently from the X1X2Y sex chromosome system in black-spotted stickleback, G. wheatlandi (Ross et al. 2009) which is thought to have been generated from a fusion between chr19 and chr12 (Kawahara et al. 2009).  Divergence between the X and Y chromosome in G. aculeatus is thought to have involved at least three pericentric inversions on the Y followed by a large deletion (Ross & Peichel 2008) generating three evolutionary strata consisting of the pseudo-autosomal region (PAR), a younger evolutionary strata (stratum two), and an older evolutionary strata (stratum one; White et al. 2015). However, relatively little is known about the mechanisms that are involved in the recombination suppression on sex chromosome pairs in vertebrates. In plants, epigenetic modifications such as DNA methylation are thought play a role in regulating the formation of heterochromatin (Zhang et al. 2008), which is thought to regulate sex chromosome recombination (Gorelick 2003). The diversity of sex chromosome systems found among closely related members of the Gasterosteidae family, combined with the relatively recent evolutionary history of these systems makes G. aculeatus a particularly interesting system in which to study the evolution of sex chromosomes and sex determining systems. 30   1.6.6 Effects of maternal stress on offspring phenotypes in threespine stickleback  Much of what is known regarding developmental effects of maternal stress comes from mammals where the placenta is involved in regulating the fetal environment. Whether maternal stress has similar effects on offspring development in non-placental organisms is not well understood. Unlike mammalian systems, where complex maternal-fetal interactions via the placenta are thought to regulate the sexually dimorphic effects of maternal stress (Nugent & Bale 2015; Bronson & Bale 2016), oviparous embryos develop outside the body, separated from the further influence of maternal conditioning. In oviparous fish, females deposit nutrients and proteins into developing eggs during a process known as vitellogenesis (Tyler & Sumpter 1996). Variation in maternal condition such as nutritional state or circulating stress hormone levels can affect the nutrient and hormone composition of the eggs, which can have persistent effects on adult offspring phenotypes (e.g. Marteinsdottir & Steinarsson 1998; Mccormick 1998; Giesing et al. 2011). Despite differences in the developmental environment of mammalian and teleost embryos, several studies have observed effects of maternal stress on offspring behavior and the regulation of the hypothalamic-pituitary-interrenal (HPI) axis in teleost fishes (Mommer & Bell 2013; Bell et al. 2016; Sopinka et al. 2017).  Threespine stickleback are an interesting system in which to investigate the effects of parental stress on offspring phenotypes and the mechanisms underlying the transmission and development of these traits. Variation in the number and type of predators present between marine and freshwater environments, and variation among freshwater environments is thought to exhibit a strong selective pressure in stickleback (Boughman 2007). This divergent selection has 31  resulted in differentiation of behavioral traits such as variation in schooling behavior between benthic and limnetic stickleback (Boughman 2007).  Changes in predatory stressors among different environments have the potential to result in changes in maternal responses to the novel stressors, which could affect their offspring. For example, exposure of female stickleback to simulated predatory stressors increases the amount cortisol levels and the size of their eggs (Giesing et al. 2011). Cortisol is a glucocorticoid stress hormone that is part of the HPI stress response axis in (Bonga 1997). While the concentration of maternally derived cortisol has been shown to decrease rapidly following fertilization (Alsop & Vijayan 2009; Paitz et al. 2015), reducing the concentration of maternally derived cortisol prior to fertilization has been shown to disrupt development of the HPI axis in zebrafish (Nesan & Vijayan 2016). Increased levels of cortisol are often implicated in modulating the effects of maternal stress on offspring phenotypes, but the artificial elevation of exogenous cortisol levels often has little to no effect on offspring phenotypes (Paitz et al. 2016) suggesting that cortisol levels may not be directly responsible for inducing many of developmental effects on offspring phenotypes. Thus, the mechanisms through which maternal stress impacts offspring development is complex and likely includes other factors in addition to components of the HPI axis (Beijers et al. 2014).   1.6.7 Epigenomics in stickleback  At the time I started the experiments described in the following chapters of this thesis, only a single study had examined DNA methylation patterns in stickleback and this study identified significant changes in whole-genome DNA methylation levels in gonad tissue 32  following environmental exposure to estradiol (E2) were detected using HPLC, but no changes in response to hexa-bromo-cyclododecane (HBCD) exposure in liver (Aniagu et al. 2008).   Over the past five years, three studies have been published that used reduced representation sequencing techniques (RRBS and BSradSeq) to characterize DNA methylation patterns in threespine stickleback. These studies investigated differences in DNA methylation between stickleback that differed in the number of lateral plates (Smith et al. 2015), between stickleback from a marine and a freshwater population (Trucchi et al. 2016), and between stickleback from a marine and a freshwater population at different salinities (Artemov et al. 2017). These papers revealed substantial differences in putative methylation patterns between stickleback populations or morphs (Smith et al. 2015) and suggest that DNA methylation patterns are modified following transfer between high and low salinity environments (Artemov et al. 2017). However, each of these studies has limitations that reduce our ability to fully interpret the reported data, in part because of the limitations on our current understanding of epigenomics in an environmental and ecological context.  The first limitation on the interpretation of these studies is that each study involved aligning the bisulfite treated DNA sequences from multiple wild populations or morphs to the single reference genome available from the Ensembl genome database. This creates a potential issue in the interpretation of the datasets if there are DNA polymorphisms between the populations or morphs. Unmethylated CpG loci are identified when a CpG dinucleotide is converted to a TpG dinucleotide following treatment with sodium bisulfite. Therefore, naturally occurring C to T polymorphisms between bisulfite sequences and the reference sequence can affect the calculation of DNA methylation levels. This complicates the interpretation of these data because one population might have the same CpG sequence as the reference genome while 33  the other population has a TpG at that location. These would then be interpreted as methylated and unmethylated loci respectively, when in fact the detected difference is a genetic polymorphism, not a difference in methylation. Because each of these studies investigated DNA methylation patterns in stickleback between genetically divergent populations, it is possible that the patterns described in these data are confounded by genetic polymorphisms both between populations and between the populations and the reference sequence.  A second potential limitation of these studies is that they used wild-caught individuals. Therefore, it cannot be determined if the differential methylation described in these studies are fixed differences between populations or changes in response to environmental differences. This is particularly true for the study by Artemov et al. (2017) because DNA methylation patterns were compared between stickleback from a marine and a freshwater population acclimated to their native salinity. Therefore, the differential methylation patterns described in this study could represent DNA methylation patterns in stickleback that are caused by exposure to different salinities instead of evolved differences between marine and freshwater populations as suggested in the manuscript.  However, despite their limitations, these studies illustrate how modern advances in sequencing technology have enabled the use of previously cost-prohibitive techniques to address questions in ecology and evolution, highlighting the interest in using threespine stickleback as a model system for studying epigenetic processes in an ecological and evolutionary context. To fully interpret these studies, however, it is critical to understand how DNA methylation patterns in stickleback are affected by changes in the environment, and how plasticity at different stages in development influences DNA methylation patterns. Therefore, a first step towards understanding whether DNA methylation has an important role in regulating the expression of 34  the phenotypic diversity observed among divergent populations of stickleback is to understand the plasticity of DNA methylation patterns in response to different environments and whether exposure to different environments at different developmental stages are temporary or if they have persistent effects on DNA methylation patterns. My thesis aims to fill this knowledge gap.  1.7 Thesis organization  My thesis is organized around four data chapters addressing the key questions that I posed that the beginning of this introductory chapter. Below I outline the methodological approaches and questions addressed in each of these chapters.   1.7.1 Chapter two  Chapter two of my thesis investigates the persistent effects of warm and cold developmental temperature on DNA methylation patterns in adult stickleback muscle tissue using an RRBS approach. I then compare these effects to changes in the DNA methylation patterns in the muscle tissue of adults acclimated to warm and cold temperatures. Comparing changes in DNA methylation patterns across these two timescales provides insight into the potential effects of the environment on epigenetic variation and into how epigenetic responses might influence the capacity of organisms to cope with environmental change.  1.7.2 Chapter three  Chapter three of this thesis uses the same experimental design and individual organisms used in chapter two to investigate the effects of developmental temperature and adult thermal acclimation on gene expression in adult muscle tissue using an RNA-Seq approach. By 35  comparing developmental effects on adult gene expression to adult thermal acclimation, I investigate whether there are potential mechanistic connections between gene expression plasticity across timescales. Because the same individuals used in chapter two to investigate temperature effects on DNA methylation patterns were also used in chapter three, I was able to directly compare temperature dependent changes in gene expression to changes in DNA methylation.   1.7.3 Chapter four  Chapter four examines DNA methylation patterns in the young XY heteromorphic sex chromosome system of threespine stickleback using a whole genome bisulfite sequencing approach to determine whether DNA methylation patterns differ between male and female threespine stickleback. In this chapter I also characterize the effects of salinity on DNA methylation in stickleback and determine whether regions known to be under differential selection between marine and freshwater environments and genes whose expression is salinity responsive are differentially methylated in response to environmental salinity.  1.7.4 Chapter five  In chapter five I investigate whether stress experienced during vitellogenesis in female stickleback has a detectable effect on gene expression patterns in the brain tissue of adult stickleback offspring. This experiment also examined whether maternal stress has similar effects on male and female offspring.   36  Chapter 2: Persistent and plastic effects of temperature on DNA methylation across the genome of threespine stickleback (Gasterosteus aculeatus)  2.1 Introduction  Environmental temperature is a critical factor that determines the distribution and abundance of organisms (Angilleta 2009; Somero 2010; Schulte et al. 2011; Sunday et al. 2012), and the rapidly changing thermal environment due to global climate change is projected to have a particularly pronounced effect on ectothermic species such as fishes (Somero 2010; Sunday et al. 2012). Phenotypic plasticity plays an important role in the responses of organisms to rapid changes in the environment and could be critical in determining the ability of species to cope with climate change (Charmantier et al. 2008; Somero 2010; Seebacher et al. 2015). Plastic responses can persist across multiple time scales, from short-term and largely reversible acclimation responses, to longer-term developmental plasticity, to transgenerational responses (Beaman et al. 2016), but plasticity at these different timescales often impacts similar physiological processes. For example, in fishes, temperature changes during development can have persistent effects on sex determination, metabolism, thermal performance, muscle phenotypes, and gene expression (Devlin & Nagahama 2002; Schaefer & Ryan 2006; Scott & Johnston 2012; Schnurr et al. 2014). While exposure of adults to short term temperature change can induce temporary plastic responses in many of these traits (Guderley 2004; Schulte et al. 2011; Scott & Johnston 2012; Morris et al. 2014b). Physiological plasticity is often modulated at the level of gene expression, thus investigating regulatory mechanisms controlling gene 37  expression is fundamental to understanding how species cope with climate change (Somero 2010).  Epigenetic modifications can influence the plasticity of organisms by regulating gene expression without modifying the DNA sequence (Deans & Maggert 2015). DNA methylation, the addition of a methyl group to a cytosine nucleotide, is one of the best-characterized epigenetic processes. The addition or removal of DNA methylation can be dynamic, occurring rapidly in response to environmental cues (Bossdorf et al. 2008; Angers et al. 2010; Dowen et al. 2012). DNA methylation patterns can also persist through cell division and be passed on through multiple generations potentially influencing evolutionary processes and fitness (Jablonka & Raz 2009; Flores et al. 2013). The intra and intergenerational effects of DNA methylation and the role of DNA methylation in regulating plasticity is an intriguing system in which to examine how epigenetic processes modulate plastic and persistent phenotypic variation in response to changing environments. How this epigenetic system of “soft inheritance” (Mayr 1982; Dickins & Rahman 2012) functions as a response to changes in temperature has been highlighted by studies in plants (Dubin et al. 2015; Nicotra et al. 2015; Song et al. 2015; Keller et al. 2016; Rakei et al. 2016; Gugger et al. 2016; Kawakatsu et al. 2016; Li et al. 2016), yet little is known about these processes in ectothermic vertebrates (Hofmann 2017).  To understand how DNA methylation is modified in response to environmental temperature in vertebrates, we used threespine stickleback (Gasterosteus aculeatus), an important vertebrate model for studies of ecological adaptation (Bell & Foster 1994; Jones et al. 2012b), in which phenotypic effects of both developmental temperature and adult acclimation temperature have been clearly demonstrated (Jordan & Garside 1972; Guderley et al. 2001; Shama et al. 2014; Ramler et al. 2014; Morris et al. 2014b; Shama & Wegner 2014; Teigen et al. 38  2015). By comparing DNA methylation patterns between threespine stickleback with different thermal histories during development to methylation patterns in adult threespine stickleback acclimated to differing temperatures, we demonstrate that epigenetic mechanisms are a component of both persistent and plastic responses to environmental change.   2.2 Materials and methods 2.2.1 Experimental design Adult threespine stickleback (G. aculeatus) were collected from a fully plated “marine” population in Oyster Lagoon (British Columbia, Canada, GPS: 49.6121-124.0314) in June 2014. Stickleback were separated into six 110-litre glass tanks (20 stickleback per tank) and acclimated for three weeks to 20 ppt salt water (dechlorinated Vancouver municipal tap water supplemented with Instant Ocean Sea Salt), 18 °C and 14:10 h light:dark photoperiod, which mimics the natural environmental conditions at the collection location at the time of collection. Stickleback were fed daily to satiation with Hakari Bio-Pure frozen Mysis Shrimp.  After the three weeks of laboratory acclimation, eggs were collected from six gravid females and testes were dissected from six males and individually macerated in a 1.75 mL microcentrifuge tube containing 150 µL Ginzberg’s fish Ringer’s solution. Eggs collected from a single female were arranged as a monolayer in petri dishes containing 5 mL of 20 ppt salt water. 50 µL of the sperm solution from a single male was applied directly on the egg mass and left for 30 min at the acclimation conditions to allow fertilization to occur. This process was repeated six times using six different females and six different males to generate six independent families. Following fertilization, an additional 10 mL of 20 ppt saltwater was added to each petri dish. Each clutch was then split across three separate petri dishes. A single petri dish from each family 39  was transitioned to a developmental temperature of 12, 18, or 24 °C (such that all six families were represented in each developmental temperature treatment). Petri dishes were partially covered to prevent water loss via evaporation and to allow for surface gas exchange. Eggs were monitored twice daily during which time any unfertilized eggs were removed and 10 mL of water was changed to prevent mold growth. Hatching time differed by several days between stickleback reared at different development temperatures (Appendix A, Figure A.1). After hatching, embryos were moved to 110 L glass aquaria at the original acclimation conditions of 18 °C with a 14:10 h light:dark photoperiod. Each family was held in a separate mesh breeding box, with two breeding boxes (i.e. two families) per glass aquaria. Aquaria were equipped with hanging box filters (Aquaclear) and sponge filters for filtration and aeration. Larvae were fed live brine shrimp nauplii twice daily ad libitum until they were large enough to feed on frozen Mysis shrimp. Once juvenile stickleback reached a size of approximately 1 cm they were released from the breeding boxes into the aquaria. For the experiment on the effects of developmental temperature on DNA methylation, families were kept separate (two families per split aquarium) until sampling, which occurred at 10 months of age. A total of six stickleback from each development temperature treatment were euthanized and muscle tissue samples (from behind the dorsal fin to the base of the tail) were taken from each stickleback (four females and two males from each treatment except for the 12 °C development group which had five females and one male), with sampling distributed across families (Appendix A, Table A.1). This experimental design minimizes the potential for effects of genetic variation and sex on DNA methylation patterns (Liu et al. 2010; Gertz et al. 2011). For the experiment on the effects of adult thermal acclimation on DNA methylation patterns, after nine months of development, a random sample of stickleback from the six families 40  that had developed at 18 °C were mixed together and then split between three different acclimation temperatures (5, 18, or 25 °C). Following four weeks of thermal acclimation, a total of six sticklebacks from each acclimation temperature treatment were euthanized and muscle tissue was sampled as described above (from four females and two males from each treatment). Muscle tissue was immediately snap frozen in liquid nitrogen and stored at -80 °C until further use. The developmental and acclimation temperature protocols described above created 5 different groups (4 treatments and 1 control) as follows: stickleback developed at 12 °C and acclimated to 18 °C, stickleback developed at 24 °C and acclimated to 18 °C, stickleback developed at 18 °C and acclimated to 5 °C, and stickleback developed at 18 °C and acclimated to 25 °C and stickleback developed at 18 °C and acclimated to 18 °C that were treated as the control group (Figure 2.1).   Figure 2.1: Experimental design. Solid lines represent stickleback that were developed and acclimated to 18 °C (black) or that developed at 18 °C and were acclimated to 25 °C (red) or 5 °C (dark blue). Dashed lines represent the developmental temperature treatments of 24 °C (pink), or 12 °C (light blue). Note that altered development temperature changes time to hatch by several days. Post-hatch, stickleback were transferred to 18 °C where they developed to adults for the remainder of the experiment (10 months).  051015202530Temperature (°C)Embryonic development Post−embryonic development Adult acclimation41  2.2.2 Reduced representation bisulfite sequencing  We chose to examine methylation patterns in muscle tissue because previous work (Scott & Johnston 2012; Morris et al. 2014b) has indicated that both developmental temperature and acclimation temperature alter muscle gene expression in fish. Genomic DNA was isolated from muscle tissue using a Quick-DNATM Miniprep kit (Zymo Research). Bisulfite-treated reduced representation genomic DNA libraries were prepared by the UBC Nucleic Acid and Protein Service core facility and sequenced at the UBC Biodiversity Centre sequencing facility. Libraries were created using a Bio-O NEXTflex Bisulfite-Seq Kit on MspI digested gDNA. Samples were barcoded using NEXTflex Bifulfite-seq Barcodes. Purified, adapter-ligated DNA was then bisulfite treated using an EZ DNA Methylation Gold Kit (Zymo Research). Samples were split between two lanes (3 samples per treatment per lane for 15 samples total per lane) of an Illumina HiSeq 2000. Average sequencing library size was 19,900,578 ± 3,880,665 (mean ± SD) million reads and covered an average of 12,901,548 CpG sites per sample.   2.2.3 Quantification and statistical analysis  Sequences from the bisulfite treated samples were aligned to the stickleback genome (Ensembl release 87) and annotated using CLC genomics workbench v9.5 with the bisulfite sequencing plugin v1.1.1. Average mapping efficiency was 88.3%. Total coverage and methylated cytosine coverage data were exported and analyzed for differential methylation in R v3.3.1 using methylKit package v1.1.3 (Akalin et al. 2012). Prior to global DNA methylation analysis, CpG loci were filtered so that only sites with at least 8 reads were retained in each sample. Sites that were in the 99.9th percentile of coverage were also removed from the analysis 42  to account for potential PCR bias. After filtering, an average of 137,954 CpG sites (1% of all CpG sites in the genome) were retained per library with a mean coverage depth of 15 reads.  To test for differential methylation at specific loci an additional filter was applied such that only CpG loci with at least 8 reads in at least 4 of the 6 samples within a treatment were kept. Threespine stickleback that were kept at 18 °C for the duration of the experiment were used as a reference group for a pairwise comparison between each of the developmental and acclimation temperature treatments.  Mean and median DNA methylation values for each individual were compared using a non-parametric Kruskal-Wallis Test. A Kolmogorov-Smirnov test was performed to test for differences in the shape of the cumulative methylation distributions between treatments. All statistical analyses were performed in R v3.3.1. To identify differentially methylated regions (DMRs), the genome was divided into 100bp regions using the tileMethylCounts() function in methylKit v 1.1.3 with a window size of 100 and a step size of 100. Logistic regression was implemented using the calculateDiffMeth() function to identify differentially methylated loci and DMRs. P-values were false discovery rate corrected to Q-values using the sliding linear model (SLIM) method (Wang et al. 2011). To determine the common and unique DMRs among treatment groups, the data were first filtered to include only those regions that were present in the reduced representation sequencing data with at least 8 reads in 4 of the 6 individuals in every group. Unique and overlapping DMRs were then visualized using the R-package VennDiagram. Annotations for DMRs were obtained using the annotatePeakInBatch() function in the R package ChIPpeakAnno v3.6.5 from Bioconductor (Zhu et al. 2010). GO annotations were obtained from a previous study (Metzger & Schulte 2016a). Enrichment analysis was performed using goseq v1.24 (Young et al. 2010) 43  and p-values were false discovery corrected for multiple comparisons (Benjamini & Hochberg 1995).  Figure 2.2: (A-D) Manhattan plots of the chromosomal positions of methylated CpG loci that differed significantly between stickleback that were developed at and acclimated to 18 °C, stickleback developed at 12 °C (A) or at 24 °C (B) then reared at 18 °C, or that were developed at 18 °C then acclimated to 5 °C (C) or 25 °C (D). Each point represents a single differentially methylated cytosine (DMC). The y-axis presents the difference in percent methylation for that DMC relative to the stickleback developed and acclimated at 18 °C. Only DMCs with >10% change in methylation are shown. Points above and below the horizontal dashed line are hypermethylated and hypomethylated loci, respectively. Points indicate chromosomal locations. Vertical dashed lines and colors indicate different chromosomal regions. (E-H) Chromosomal frequency distribution of hypermethylated (dark grey) and hypomethylated (light grey) DMCs in stickleback that developed at 12 °C (E) developed at 24 °C (F) or that were acclimated to 5 °C (G) or acclimated to 25 °C (H). The vertical axis represents the percent of CpGs that were sequenced that were differentially methylated. ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●−60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21Change in % methylationAchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21% CpG per chromosome0246810hypomethylatedhypermethyatedE●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●−60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21BChange in % methylationchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21% CpG per chromosome0246810F●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●−60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21CChange in % methylationchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21% CpG per chromosome0246810G●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●−60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21DChange in % methylationchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21% CpG per chromosome0246810H44  2.3 Results and discussion 2.3.1 Temperature change induces genomic hypermethylation  We identified a total of 2,130 CpG loci in the stickleback genome that were differentially methylated in response to temperature in at least one treatment. Pairwise comparison of DNA methylation levels at individual CpG loci in stickleback from developmental and adult acclimation treatments to the control (18 °C) group identified 554 differentially methylated cytosines (DMCs) in stickleback that developed at 12 °C, 480 DMCs in stickleback that developed at 24 °C, 1150 DMCs in stickleback that were acclimated to 5 °C and 778 DMCs in stickleback acclimated to 25 °C. There was no apparent clustering of DMCs on a specific chromosome or chromosomal region (Figure 2.2A-D), and DMCs were distributed across promoters, introns, exons, and intergenic regions proportionally to the distribution of these features within the genome (Appendix A, Figure A.2).   Approximately 2-8% of the sequenced CpG sites on each chromosome exhibited differential methylation in response to altered temperature (Figure 2.2E-F). More DMCs were hypermethylated than were hypomethylated in all treatments suggesting that an increase in genomic DNA methylation levels is a general response to changes in environmental temperature (Figure 2.2E-F).   An increase in genomic DNA methylation levels was expected in stickleback exposed to cooler temperatures based on previous reports of latitudinal variation in genomic DNA methylation levels in fishes, in which polar fishes exhibit higher global DNA methylation levels than do equatorial fish (Varriale	&	Bernardi	2006).	However, the observed increase in global DNA methylation levels in stickleback exposed to higher temperatures is in contrast to the inter-45  specific correlations between DNA methylation levels and ambient temperature	(Varriale	&	Bernardi	2006).  To further investigate the effects of temperature on genomic DNA methylation levels we examined the mean DNA methylation level for all individuals. Mean genomic methylation values ranged from 72.03-73.96%, and median methylation levels ranged from 80.3%-84.12% (Appendix A, Table A.2), but the treatment groups were not detected as significantly different from the controls. We also compared the cumulative distribution curves of genomic DNA methylation levels for each of the treatment groups to those of the 18 °C control group using the Kolmogorov-Smirnov (K-S) test. Results from the K-S test indicate that the distribution of genomic DNA methylation levels for all treatments was significantly different compared to stickleback from control conditions (p < 2.2e-16, Figure 2.3 A/B) and that the distribution was the most different in stickleback acclimated to 25°C, which had the largest K-S test D statistic (0.05126, Appendix A, Table A.3). Visual inspection of the violin plots and cumulative distribution curves for DNA methylation (Figure 2.3A/B) suggest that, in general, these differences are driven by shifts towards increased proportions of highly methylated CpG loci relative to control conditions, which is consistent with the biases towards hypermethylation that we observed among DMCs for all treatments. Taken together these data clearly illustrate that both developmental temperature and acclimation temperature alter the methylome of threespine stickleback. Changes in DNA methylation have the potential to causes changes in transcriptional regulation, and thus in cellular and organismal phenotypes, but changes in DNA methylation can also be the result of changes in cellular phenotype, which makes interpreting the physiological significance of changes in DNA methylation challenging (Lappalainen & Greally 2017). For example, both developmental 46  temperature and thermal acclimation have been shown to cause changes in the proportion of oxidative and glycolytic muscle fibres in fish (Egginton & Sidell 1989; Scott & Johnston 2012). Thus, the changes in DNA methylation that we observe could be either a cause or a consequence of changes in cell type.   Figure 2.3: (A) Violin plot of genomic DNA methylation levels for each treatment condition and control condition. Colors depict different treatments. Stickleback developed and acclimated to 18 °C (grey/black), stickleback that developed at 18 °C and were acclimated to 25 °C (red) or acclimated to 5 °C (dark blue), and stickleback with a developmental temperature of 24 °C (pink), or 12 °C (light blue). Width indicates the pooled distribution density of percent methylation of CpG loci in a given treatment. Embedded box plots summarize variation in the median methylation level across the six samples in each treatment. The line indicates the median of these medians, the box defines the interquartile range (IQR), and the whiskers represent the maximum and minimum values, excluding values greater than 1.5x IQR (which are shown as individual points). (B) Cumulative distribution frequency plot of pooled DNA methylation levels for each treatment. Colors as in panel A and Figure 2.1.  Inter-specific variation in genomic DNA methylation levels with body temperature has also been hypothesized to be the result of temperature effects on the rate of deamination of methylated cytosines (Varriale & Bernardi 2006). This “methylation-temperature-deamination hypothesis” posits that the relatively low methylation levels of endotherms and tropical fishes compared to ectotherms in cooler climates could be the result of reduced genome GC content in animals with higher body temperatures, and further suggests that this reduced GC content is the 47  result of increased rates of deamination of methylated cytosines (causing increased mutation rate from C to T) at higher body temperatures (Varriale 2014). In the context of this hypothesis, our observation of increased genomic DNA methylation levels in stickleback exposed to elevated temperatures could ultimately result in accelerated mutation rates over evolutionary timescales, due to the potential for a positive feedback between the increased rate of deamination due to elevated environmental temperatures and the increased genomic DNA methylation levels.   2.3.2 Conserved differential methylation between temperature treatments  To determine whether there is a common core response of the epigenome to thermal change during both development and adult thermal acclimation, we summarized DNA methylation information across 100bp windows, and identified 1,206 differentially methylated regions (DMRs). In order to assess the potential functional significance of these DMRs we identified the nearest neighbor genes for each DMR. We then performed functional (Gene Ontology; GO) enrichment analysis for annotated genes located within 2kb of a DMR. There was no significant enrichment of GO-categories for any treatment group. Approximately 25% of the DMRs associated with variation in developmental temperature were also differentially methylated as part of the thermal acclimation response in adults (Figure 2.4). We identified 172 DMRs that overlapped between stickleback that developed at cold temperatures and adult stickleback acclimated to cold temperatures, and 146 DMRs that overlapped between stickleback that developed at warm temperatures and adult stickleback that were acclimated to warm temperatures. Comparison of all DMRs identified 50 DMRs that were differentially methylated in response to all four treatments (Figure 2.4), and only two of these DMRs showed different directions of response to temperature across treatment while the other 48  48 regions all exhibited the same direction of change in DNA methylation level in all treatments (Appendix A, Table A.4). These regions may represent candidates for a generalized response to thermal change. One of these common DMRs was located in the TNF receptor-associated factor 7 gene, an E3 ubiquitin ligase that has been previously identified as a target for natural selection with latitude or temperature in sea urchins (Pespeni et al. 2012). These data suggest that there is a common core response of the epigenome to thermal change, and highlights the possibility that developmentally induced variation in DNA methylation patterns could influence plasticity in adult acclimation responses.  Figure 2.4: Venn Diagram depicting the number of common and unique differentially methylated regions identified between stickleback that were developed at and acclimated to 18 °C and stickleback that developed at 12 °C (dashed light blue line) or at 24 °C (dashed pink line) and then were acclimated to 18 °C, or that developed at 18 °C and were then acclimated to 5 °C (solid dark blue line) or 25 °C (solid red line).  The changes in methylation patterns that we detect in response to thermal change may point toward mechanisms underlying the transgenerational effects of temperature in stickleback (Shama et al. 2014, 2016; Ramler et al. 2014; Shama & Wegner 2014; Shama 2015), which have 1103210260245027382532931 3849272915°C Acclimation25°C Acclimation12°C Development24°C Development49  been suggested to buffer the short-term effects of climate warming. Whether the effects of temperature on stickleback methylation levels persist through multiple generations requires further investigation; however, the abundance of distinct differentially methylated loci in each treatment suggests that variability in environmental temperatures could contribute to the observed accumulation of epigenetic variation between stickleback morphotypes (Smith et al. 2015).               2.4 Conclusions  Understanding the mechanisms regulating the capacity of organisms to respond to environmental variation is a critical aspect of determining the impacts of environmental change on populations (Somero 2010). In this study we have demonstrated that altered temperature during development has prolonged effects on DNA methylation levels in an ectothermic vertebrate and that modifications to DNA methylation levels are also associated with the plastic adult acclimation response to environmental temperatures. In addition, we have demonstrated that the persistent effects of developmental plasticity on DNA methylation patterns affect regions of the genome where DNA methylation patterns are also modified during adult acclimation. These data illustrate the profound effect of temperature on DNA methylation patterns across multiple time scales, which has important implications for elucidating the underlying mechanisms that may modulate the capacity of organisms to cope with environmental change. 50  Chapter 3: Similarities in temperature-dependent gene expression patterns between developmental plasticity and phenotypic flexibility in threespine stickleback (Gasterosteus aculeatus)  3.1 Introduction  Many organisms are able to respond to changes in environmental conditions by altering their phenotype, a phenomenon known as phenotypic plasticity. Phenotypic plasticity can be adaptive, maladaptive, or neutral (Ghalambor et al. 2007), and both maladaptive and adaptive phenotypic plasticity are thought to be important factors that can influence evolutionary trajectories (Ghalambor et al. 2015; Hendry 2016). However, there remains substantial debate as to whether phenotypic plasticity typically impedes or accelerates evolutionary change (Hendry 2016). The majority of phenotypic plasticity is ultimately the result of the differential regulation of gene expression (Schlichting & Smith 2002), and it has been suggested that understanding the molecular processes underlying this phenotypic plasticity is an important step in evaluating its effects in an evolutionary context (Schneider et al. 2014; Pfennig & Ehrenreich 2014).   Plastic responses can occur over a wide range of time-scales both across generations and within an individual organism’s lifetime. Within-individual plasticity can, in turn, be divided into two broad classes acting at different life-history stages (Beaman et al. 2016). In the first class, which is typically termed developmental plasticity, the environment encountered during early life alters developmental trajectories. This type of plasticity is considered to result in a stable change in phenotype that lasts for the duration of the organism’s lifetime. The second class of phenotypic plasticity, which is often termed phenotypic flexibility (Piersma & Drent 2003), 51  involves rapid and reversible changes in phenotype. This rapid and reversible plasticity includes processes such as physiological acclimation. There is substantial debate in the literature as to whether plasticity operating at different life stages operates via similar or different mechanisms (Kingsolver et al. 2002; Shintani & Ishikawa 2007; Kristensen et al. 2008; Colinet & Hoffmann 2012; Teets & Denlinger 2013). Thus, it is unclear whether plastic responses at different life stages are mechanistically linked and whether plastic responses across different life stages are under similar selective constraints (Gerken et al. 2015; Beaman et al. 2016). The ability to comprehensively measure changes in gene expression through the use of high throughput sequencing technologies provides an unbiased and powerful approach to better understand the mechanistic relationship between plasticity at different life stages (Aubin-Horth & Renn 2009).  In ectothermic organisms such as fishes, body temperature closely mimics that of the environmental temperature. Consequently, changes in environmental temperature have been shown to have pervasive effects on biochemical and physiological processes, including profound changes in gene expression. In fishes, gene expression plasticity in response to changes in environmental temperature has been predominately investigated by manipulating juvenile or adult thermal environments and measuring gene expression patterns to detect thermal acclimation responses (Gracey et al. 2004; Scott & Johnston 2012; Morris et al. 2014b; Shama et al. 2016; Kim et al. 2017a; Healy et al. 2017). A few studies have also examined the transgenerational effects of thermal exposure on the transcriptome of offspring, detecting effects of maternal or grandmaternal thermal exposure on offspring gene expression (Veilleux et al. 2015; Shama et al. 2016).  Similarly, persistent effects of the temperatures experienced during early development (prior to hatch) on gene expression have been detected in adult fish held under common conditions (Scott & Johnston 2012; Oomen & Hutchings 2017). However, very 52  little is known about whether plasticity at different life stages affects similar or different processes.    Epigenetic processes, such as DNA methylation, which result in chromosome bound, heritable changes to gene expression patterns that are not dependent on changes to the underlying DNA sequence (Deans & Maggert 2015), are thought to be important mechanisms regulating gene expression plasticity (Hu & Barrett 2017). For example, changes in DNA methylation patterns during development are thought to play an important role in cellular differentiation and in maintaining cell-type specific transcriptional activity through mitosis (Monk et al. 1987; Li 2002). Although epigenetic effects are often considered to be relatively stable across the lifespan, or even heritable, recent studies (Baránek et al. 2015; Viggiano & de Pinto 2017; Metzger & Schulte 2017) indicate that DNA methylation can be rapidly altered by environmental change, and could be implicated in modulating plasticity at different life-history stages (Bird 2007; Deans & Maggert 2015; Huang et al. 2017a). Understanding the relationship between plasticity in response to thermal change at different life-history stages has important implications for predicting the resilience of populations to anticipated environmental change (Somero 2010). Within the next century climate change is expected to result in an increase in mean temperatures as well as an increase in the magnitude, frequency, and duration of extreme temperature events and these changes in environmental thermal regimes are predicted to impact the distribution and abundance of ectothermic organisms (Sunday et al. 2012; Bauerfeind & Fischer 2014; Thornton et al. 2014; Seebacher et al. 2015; Frainer et al. 2017). Understanding the relationship between gene expression plasticity at different life stages has the potential to be important in determining the capacity of organisms to cope with these changes (Donelson et al. 2017)  53   Here, we examine changes in gene expression in response to altered developmental temperature and adult thermal acclimation in the threespine stickleback fish (Gasterosteus aculeatus). Stickleback populations are found in both marine and freshwater environments throughout their Holarctic distribution. Post-glacial rebound following the retreat of the Pleistocene glaciers ~10-12 thousand years ago resulted colonization of newly formed freshwater habitats (followed by adaptation and reproductive isolation). Differences in the selective pressures of freshwater environments have resulted in the rapid, parallel morphological and behavioral divergence of freshwater populations from ancestral marine populations (eg. Jones et al. 2012b), including variation in DNA methylation patterns (Smith et al. 2015; Artemov et al. 2017). Characterization of the underlying genetic divergence of marine and freshwater stickleback populations has established stickleback as a powerful system in which to investigate the genetic basis of adaptive evolution (Jones et al. 2012b). In addition, there is accumulating evidence of phenotypic plasticity in stickleback in ecologically relevant traits. For example, gene expression plasticity in response to temperature has been investigated in adult stickleback (Morris et al. 2014b), and temperature-dependent transgenerational and developmental plasticity in body shape (Ramler et al. 2014), hatching success (Shama et al. 2014), growth (Shama & Wegner 2014), DNA methylation (Metzger & Schulte 2017), and reproductive strategies (Kim et al. 2017b; Hovel et al. 2017) have been observed. Thus, stickleback not only present an ideal system in which to investigate the mechanisms underlying plasticity a different life stages, but to also help better understand the effects of plasticity on evolutionary processes.    Specifically, the objectives of this study were to 1) examine the persistent effects of temperature manipulations during development (from fertilization until hatch) on gene expression patterns in the muscle tissue of adults held under common conditions, 2) assess 54  mechanisms associated with phenotypic flexibility by acclimating adult stickleback to warmer and colder temperatures and measuring the temperature-dependent effects of thermal acclimation on gene expression patterns in muscle tissue, 3) examine whether the persistent effects of thermal manipulations during development on gene expression are similar to adult acclimation effects on gene expression patterns, and 4) assess whether gene expression plasticity in response to developmental and adult temperature manipulations are associated previously characterized changes in DNA methylation patterns (Metzger & Schulte 2017; Chapter two of this thesis).  3.2 Materials and Methods 3.2.1 Gasterosteus aculeatus rearing conditions  Adult threespine stickleback (G. aculeatus) of the fully plated “marine” ecotype were collected from Oyster Lagoon (British Columbia, Canada, GPS: 49.6121, -124.0314) in June 2014. In the lab, stickleback were separated into six 110-litre glass aquarium tanks (20 stickleback/tank) and acclimated to 20 ppt salt water (instant ocean), 18 °C and 14:10h light:dark photoperiod, and this photoperiod was held constant throughout the experimental period. These conditions are similar to summer conditions in Oyster Lagoon (20-30 ppt) at the time of collection. Stickleback were fed daily to satiation with Hakari Bio-Pure frozen Mysis Shrimp for three weeks prior to initiation of the breeding protocol.   Eggs were collected from a total of six different females as they became gravid. Testes were dissected from six different males, and used to generate six independent (unrelated) families. Fertilization and stickleback rearing were conducted as previously described (Metzger & Schulte 2017; Chapter two of this thesis) and are summarized in Figure 3.1.   55    Figure 3.1: Experimental design. Six stickleback families were generated for use in this experiment. Each clutch was split between three developmental temperatures of 12 °C (blue), 18 °C (black), and 24 °C (red). Following hatch, all stickleback were transferred to aquaria at 18 °C and reared for 8 months. At this point, fish from the 18 °C development treatment were acclimated to either 5 °C (blue shaded aquaria), 18 °C (grey shaded aquaria), or 25 °C (red shaded aquaria) for 4 weeks, while fish from the other developmental treatments were maintained at 18 °C. This experimental design resulted in stickleback with 5 different thermal histories. At an age of nine months, muscle tissue was sampled for RNA-seq analysis (n=6 per treatment).  Briefly each clutch was split across three separate 10 cm petri dishes. A single petri dish from each of the six families was held at 12, 18, or 24 °C until hatch. We chose these temperatures (hereafter, developmental temperatures) because they span the temperature range that might be 24 °C 18 °C 12 °C 18 °C 5-23 days (until hatch) 8 months 4 weeks 18 °C 18 °C 18 °C 5 °C 25 °C 18 °C 18 °C 56  experienced during the peak of breeding season of this population which is typically May through July (Saimoto 1993; Barrett et al. 2011).  Fertilized eggs were kept in a monolayer and submerged in 15 mL of 20 ppt seawater. Petri dishes were partially covered to decrease water loss from evaporation while still allowing for surface gas exchange to insure adequate oxygenation. Eggs were monitored twice daily during which time unfertilized eggs and mortalities were removed and 10 mL of water was changed to prevent mold growth. Hatching success was determined based on the number of fertilized embryos that hatched compared to the total number of fertilized embryos.  Developmental temperature did not have a significant effect on survival until hatch (Appendix B; Figure B.1). Embryos that developed at the coolest temperature took approximately 13 days longer to hatch than embryos that developed at the warmest temperature (Metzger & Schulte 2017; Chapter two of this thesis). Once all the embryos in a given family had hatched they were transferred to 110-litre glass aquarium tanks and maintained at 18 °C at a salinity of 20 ppt. At eight months post-hatch, a random sample of stickleback that developed at 18 °C were mixed together from the six different families and then split between three different acclimation temperatures (5, 18, or 25 °C) and held at these temperatures for four weeks.   These acclimation temperatures were chosen because they represent the ecologically relevant extremes that populations in this region would experience as adults in the wild in the winter and summer respectively (Barrett et al. 2011), and because they are close to the maximum and minimum temperatures to which threespine stickleback can be acclimated for extended periods (Wootton 1984; Lefébure et al. 2011). We selected 18 °C as the “control” temperature because it is the typical temperature at which stickleback are held in the lab. The two “experimental” acclimation temperatures differ in their magnitude of temperature change from 57  the “control” group of 18 °C because thermal performance curves are typically asymmetric, increasing gradually until reaching a peak and then rapidly decreasing (Dowd et al. 2015). Thus increasing and decreasing temperature by the same magnitude does not result in the same shift in thermal performance, a phenomenon known as Jensen’s inequality (Denny 2017). Therefore, these acclimation temperatures were chosen to minimize the effects of Jensen’s inequality by accounting for differences in slope along a thermal performance curve and choosing temperatures that would result in similar effects on performance rather than similar changes in temperature. We chose to use different thermal ranges at the two different life stages because these stages have different thermal sensitivities, and also encounter different temperatures in nature. The selected temperatures represent similar extents of the thermal tolerance breadth at each life stage.  Following four weeks of acclimation to these temperatures, a random sample of six stickleback from each acclimation temperature were euthanized and muscle tissue samples were snap frozen in liquid nitrogen and stored at -80 °C until further use. A total of six stickleback from each development temperature treatment were also euthanized at nine months post hatch, and muscle tissue samples were taken from each stickleback with sampling distributed across families (Appendix A; Table A.1). One-way ANOVAs were used to determine whether there was a main effect of either developmental temperature or adult acclimation temperature on the stickleback length and weight. A Tukey’s post-hoc analysis was performed to test for significant differences between temperatures. There was no difference in the length of the individuals from different temperature treatments (Appendix B; Figure B2 A/B). Developmental temperature had a significant effect on wet weight (ANOVA p-value = 0.0223). Stickleback that were exposed to 12 °C during development and had a significantly higher wet weight compared to stickleback 58  exposed to 18 °C (p-value = .0459) or 24 °C (p-value = 0.0307) during development (Figure B.2 C/D). However, because some individuals within a treatment are from the same family, we cannot rule out the possibility that the observed effects of developmental temperature on adult wet weight are heritable transgenerational temperature effects of the parental and grand-parental environment (Shama & Wegner 2014; Shama et al. 2016), or that the analysis is affected by pseudo-replication at the family level in some treatments but not others (Appendix A; Table A.1).  We elected to examine gene expression in muscle tissue because previous work in a variety of species of fish suggests that both developmental temperature and adult thermal acclimation have substantial effects on muscle phenotype (Johnston 2006; Macqueen et al. 2008; Johnston et al. 2009; Finstad & Jonsson 2012; Salinas & Munch 2012; Scott & Johnston 2012; Schnurr et al. 2014). A muscle sample from the other side of the same individuals was used in a previous study that investigated the effects of developmental temperature and adult temperature acclimation on DNA methylation levels (Metzger & Schulte 2017; Chapter two of this thesis).  3.2.2 RNA isolation and sequencing  Total RNA was prepared from stickleback muscle tissue using TRIzol Reagent (Invitrogen Life Technologies). Approximately 20 mg of muscle tissue was homogenized in 1 mL of TRIzol in 1.5 mL Eppendorf® Safe-Lock micro centrifuge tubes containing approximately ten 1.0 mm ceria stabilized zirconium oxide beads (Next Advance) using a Bullet Blender24 (Next Advance). Total RNA was DNase treated using the Qiagen RNeasy DNase I on-column DNA digestion protocol. Total RNA was quantified using a QBit® RNA broad range assay kit (product # Q10210; ThermoFisher Scientific) and an Invitrogen™ Qubit® 2.0 Fluorometer. 59  RNA quality was assessed using an Agilent RNA 6000 Pico Kit (product # 5067-1514) and an Agilent 2100 Bioanalyzer (Agilent Technologies). RNA integrity numbers (RIN) were between 7.9 - 9.1 (mean = 8.6 ± 0.4 SD). Preparation of cDNA libraries and 100 base-pair paired end sequencing was performed at the UBC Nucleic Acid Protein Service Unit (NAPS) and UBC Biodiversity Research Center’s next generation sequencing facility. Briefly, mRNA was purified using BIO-O NEXTflex® Poly-A beads. Sequencing libraries were prepared using the BIO-O NEXTflex® Rapid RNA-Seq kit. Each sample was individually barcoded and samples from different treatments were evenly distributed across 3 sequencing lanes of an Illumina HiSeq 2000 flow cell (10 samples/lane, two samples from each treatment/lane). Mean sequenced library size was 39,737,041 reads (± 7,227,925 SD; Appendix B, Table B.1).   3.2.3 Sequence alignment and expression analysis  Reads were aligned to the stickleback genome (http://www.ensembl.org) using CLC genomics workbench v9.5. Average mapping efficiency of paired and broken reads was 88 %. Analysis of total read counts was performed in R v3.3.1 with edgeR v3.14.0 (Robinson et al. 2010; McCarthy et al. 2012b) based on the recommended guidelines in Lin et al. (2016). Genes with no reads were removed from the datasets. Counts were normalized using the relative log expression (RLE) method. However, temperature acclimation has generally been shown to induce the up-regulation of a large proportion of genes in fish (Gracey et al. 2004; Healy et al. 2017). Many methods for normalization of RNA-seq data (e.g. TMM and RLE) assume that the majority of genes in an RNA-seq dataset are not differentially expressed (Dillies et al. 2013). Thus, normalizing factors can become problematic when a large proportion expressed genes are differentially expressed between treatments, particularly when the direction of change is biased 60  in one direction (Dillies et al. 2013; Evans et al. 2017). Ideally, only those genes that are not differentially expressed should be used to calculate the normalization factors. Therefore, we applied the method describe in Healy et al. (2017), which utilizes a preliminary analysis of the dataset to identify and remove genes that are likely to be differentially expressed and then calculates standard RLE normalizing factors using the remaining dataset. To identify and remove putatively differentially expressed genes, two separate preliminary analyses were performed.   In the first preliminary analysis, putatively differentially expressed genes were identified without library normalization. Genes with low expression were filtered from the dataset. The minimum criterion for retaining a gene was at least 0.5 counts per million (~10 counts in the smallest library) in each of the six samples of each temperature. Tagwise dispersions were calculated using the robust method in edgeR. The data were then fit to a negative binomial generalized linear model using glmFit().  In the second preliminary analysis, sequencing libraries were normalized using the RLE method. Genes with low expression were filtered from the dataset using the same criteria as previously described. Tagwise dispersions were calculated using the robust method in edgeR. The data were then fit to a negative binomial generalized linear model using glmFit().  Genes that were identified as differentially expressed in each of these two preliminary analyses were then removed from datasets for the purposes of normalization. The majority of genes that remained following these steps are less likely to be differentially expressed in response to temperature stimuli and are thus suitable to calculate normalizing factors for the rest of the dataset. Normalizing factors using the RLE method were then calculated for the dataset that contained these remaining genes. These normalizing factors were then used in a final analysis of the data. Separate preliminary analyses and normalizing factors were calculated in 61  this way for each pairwise comparison of either a developmental or acclimation temperature treatment using stickleback that were held at 18 °C for the entire duration of the experiment as the control group.   For the final analysis of the data, differential expression was assessed using pairwise comparisons of the gene expression data for each developmental or acclimation treatment to stickleback that were held at 18 °C for the duration of the experiment. The minimum criterion for retaining a gene, following RLE normalization as described above, was for a gene to have at least 0.5 counts per million (CPM; ~10 counts in the smallest library) in each of the six samples of within a temperature treatment. If a gene had a read count less than 0.5 CPM in at least one sample within a treatment then it was discarded from the analysis. After normalization using this method and filtering of the dataset a total of 12,199 genes remained in the 5 °C acclimation dataset, 12,097 genes remained in the 25 °C acclimation dataset, 11,507 genes remained in the 12 °C development dataset, and 11,661 genes remained in the 24 °C development dataset for differential expression analysis. Tagwise dispersions were calculated using the robust method in edgeR. Differentially expressed (DE) genes were identified using the glmFit() function from edgeR to fit a negative binomial generalized linear model followed by a likelihood ratio test, glmLRT(). The resulting p-values were adjusted based on a false discovery rate (FDR) correction, and the threshold for significance of these adjusted p-values (q-value) was set at 0.05. Gene ontology (GO) pathway enrichment analyses were conducted using the goseq (v1.22.0) R package (Young et al. 2010), with FDR correction as previously described.   62  3.2.4 Differential methylation analysis Differentially methylated cytosines (DMCs) associated with differentially expressed genes were identified by filtering previously identified DMCs (Metzger & Schulte 2017; Chapter two of this thesis) for those located within 5 kilobase pairs (kb) upstream or downstream of genes that were differentially expressed in stickleback from the same temperature treatment. Analysis of the genomic distribution of DMCs was conducted using the annotateWithGeneParts() function in the genomation v1.10 R package.  3.3 Results 3.3.1 Temperature dependent gene expression plasticity at different life stages  The expression levels of 10,140 genes were responsive to thermal acclimation (Figure 3.2). A total of 7,940 genes were differentially expressed in stickleback that were acclimated to 5 °C (Figure 3.2A/C) and 7,015 genes were differentially expressed in stickleback that were acclimated to 25 °C (Figure 3.2B/D). The majority of genes that responded to thermal acclimation changed by a log2 fold of less than two (i.e. an absolute fold change of less than four) for both cold (86 % of differentially expressed genes, Figure 3.2A) and warm (95 % of differentially expressed genes, Figure 3.2B) acclimated stickleback. In both cold and warm acclimation, substantially more genes were up-regulated than were down-regulated (cold-acclimated: 72% of all differentially expressed genes were up-regulated; warm-acclimated: 80% of all differentially expressed genes were up-regulated) (Figure 3.2C/D). 63   Figure 3.2: Differential expression in stickleback acclimated to either 5 °C (A/C) or 25 °C (B/D) relative to stickleback that were kept at 18 °C for the duration of the experiment. Panels A and B show the –log10 FDR corrected p-value (q-value) against log2 fold change for significantly (q<0.05) up-regulated (orange), down-regulated (blue) genes and non- differentially expressed genes (grey). Embedded plots are frequency histograms of log2 fold change (e.g. a bar above a log2 fold change of 2 represents the genes that range in a log2 fold change between 1 and 2). Panels C and D are heatmaps of differential expression in stickleback acclimated to 5 °C and 25 °C, respectively. Each row represents the expression value (log2 counts per million) for a single gene relative to the mean expression value for that gene across all individuals (orange representing higher expression and blue representing lower expression). Each column represents an individual stickleback. Columns 1-6 are stickleback that were kept at 18 °C for the duration of the experiment. Columns 7-12 are either stickleback that were kept at 18 °C for 8 months and then acclimated to 5 °C (C) or 25 °C (D) for four weeks. ●●● ●●●● ●● ●●●●●●● ●● ● ●●●● ●●● ●● ●●● ●● ●● ●●●● ●●● ●● ● ●●●● ● ●●● ●●●● ●●● ● ●● ● ●●● ●●●●● ●●●●● ●●● ●●● ●●● ●●● ● ●●● ●●●●●●● ●●● ● ●●●● ●● ●●● ●● ●●●● ●●●●● ●● ●● ●●● ●●● ●● ●● ●● ●●● ●●●● ●●● ●●●●● ●●● ●●● ●●●● ● ●●● ●●● ●●●●● ●●●●● ●● ●● ●●●● ●● ●●● ●●●●●● ●●●●●●● ●● ●●● ● ●●●●●● ● ●●● ● ●●●● ●● ●● ●●● ● ●● ● ●●●●● ●●● ●● ●●● ●●● ●●● ● ●●● ●●●● ●● ●●● ●●●● ●● ● ●●●●● ●●● ●● ●●●● ●●●●● ●●●●●●●●● ●●●●● ●●● ●● ● ●●● ●●●● ●●●● ●●● ●●●●●● ●●●●●● ●●●● ●●●●● ●●● ●● ●●● ●−10 −5 0 5 10020406080100120log2 Fold Change●●●●●●● ●●●● ● ●●● ●●● ●●●● ● ●● ●●●●● ●● ●● ●● ● ● ●●●● ●●●● ●●● ●●● ●● ●●●●●●● ●●● ●● ●●●● ●●●●● ● ●●●●● ●● ●● ● ●● ●●● ●● ●● ●● ● ●●●● ●●● ●● ●● ●● ●●●●● ●●●●● ●● ●●●● ●●● ●●●● ● ●●●●● ●●●● ●● ●● ●●●●●●● ●● ●● ●●●●●● ●●●● ●●●●● ●● ●● ●●●● ● ●● ●● ●●●● ● ●● ●● ● ● ●●● ●●●● ●● ●● ●● ●● ●● ●●●● ●● ●●● ●●●●● ●●● ●●●● ●● ●●●●● ●●● ●●●●● ●●● ●● ●●−log10 Q−value●●●●●●●●●●●●● ●●●●●●●●●●● ●●●●●● ●● ●●●●●● ●●● ●● ● ●●● ●●● ●●●● ● ●●● ●● ●●●●● ●●●● ●●● ●●●● ● ● ●● ●● ●● ●●● ●●●● ● ●●●●●●● ● ●●●●● ● ●● ●● ●●● ●● ●● ●● ●●●● ●●● ●●●● ●● ●●● ●●● ●● ●● ●●● ●●●●●● ●● ●● ● ●●●●●● ●● ●● ●●● ●● ● ● ●●●●● ●● ●●●●● ●● ●● ●● ●●● ●●●● ●●● ● ●● ●●● ●●● ● ●●●●●●●●● ● ●● ●● ●● ●●●● ●● ● ●● ●●● ●● ●● ●● ●●● ●●●● ●●● ● ●●● ●●● ●● ●●●● ●● ●● ●●●● ●● ●●●●● ●● ●● ●● ●●● ●● ●●● ● ●●●● ●●●● ●●●●● ●●● ●●●●● ●● ● ●●● ●●● ● ●●●●● ●● ●● ●● ●● ●●●●● ●●● ●● ●●● ●● ●●● ●● ●●●●● ●● ●● ●● ● ●●●●● ● ●●●● ● ●●●● ●● ●●●● ● ●● ●● ● ● ●●● ●●●●● ●● ●● ●●●● ●● ●●●● ●●● ●●●● ●● ● ●● ●●● ●●● ●● ●●●●● ● ●●●● ●●● ●●● ●●●● ●● ● ●●● ● ●● ● ● ●● ●● ●● ●●●●●● ●●● ●● ●●● ●● ● ●● ●● ●● ● ●●● ● ●●● ●●● ● ●●● ●● ●● ●●●●● ●● ●●● ●● ●●●● ●●● ● ●● ●●●● ●●●● ●●● ●● ●●● ●● ●●● ●● ●●●● ●●●●●● ●● ● ●●● ●●●● ●●● ●●● ●● ●●● ●● ●● ●●●●● ●●● ●●●●●● ●●● ●● ●●●●● ●● ●● ●●●●●●● ●● ●●●●●●● ●● ● ●● ●●●●●● ●●● ●● ●●●−10 −5 0 5 10020406080100120log2 Fold Change●● ●● ●● ● ●●●●●● ●● ● ●● ●●● ●● ●●●● ●●● ●●● ●●● ●● ●● ●● ●● ●● ●●● ● ●●● ●● ●● ●●●● ●●● ●●●●● ●●● ●●● ●● ● ●●● ●●● ●●●●● ●●●●● ●● ●●● ●●●● ● ●●● ●●●● ● ●● ● ●●●● ●● ●● ●● ●● ●●● ●●●●● ●●●● ●● ●● ●● ●●●● ●● ●●● ●●●● ●●● ●●●●●● ●● ●●●●● ●●● ● ● ●●●●● ●●● ●● ●● ●● ●●● ●●●●●● ●●●● ●● ● ●●● ●●● ●●● ●●● ● ●●● ●●●●● ●●● ● ●●●●●●● ●●●● ●●●●● ●●●●●● ●●● ●● ●● ●●●● ●● ●●●●●●● ● ●●●● ●●● ●●● ●●● ●● ●● ●●● ●●−log10 Q−value−10 −5 0 5 100500100015002000250030003500log2 Fold ChangeFrequency−10 −5 0 5 100500100015002000250030003500log2 Fold ChangeFrequency18 °C 5°C 18 °C 25°C C D A B 0 5 10 -5 -10 0 5 10 -5 -10 log2 Fold Change log2 Fold Change      1  1  -log 10 Q-value 0  10 -5 -10 log2 Fold Change   1  -5 -10 log2 Fold Change 30  35  25  20  15  10  5  Frequency 30  35  25  20  15  10  5  Frequency 0 0.5 1 -0.5 -1 64   Only 57 genes were differentially expressed in the muscle tissue of adult stickleback that were exposed to different temperatures during development (Figure 3.3). A total of 33 genes were differentially expressed in stickleback that had developed at 12 °C (Figure 3.3A/C) and 29 genes were differentially expressed in stickleback that had developed at 24 °C (Figure 3.3B/D). Similar to the patterns of differential expression in response to adult thermal acclimation, the majority of these genes had a log2 fold change less than two for stickleback that were exposed to colder (94 % of differentially expressed genes, Figure 3.3A) or warmer (90 % of differentially expressed genes, Figure 3.3B) temperatures during development. However, unlike the pattern that was observed in response to adult acclimation temperature, none of the genes affected by developmental temperature had a log2 fold change greater than five. As was the case for adult thermal acclimation, the majority of the differentially expressed genes in response to altered developmental temperature were up-regulated (cold-development: 64 % of all differentially expressed genes were up-regulated; warm-development: 83 % of all differentially expressed genes were up-regulated)(Figure 3.3C/D).  65   Figure 3.3: Differential expression in stickleback exposed to either 12 °C (A/C) or 24 °C (B/D) during development and then kept at 18 °C until 9 months of age relative to stickleback that were kept at 18 °C for the duration of the experiment. Panels A and B show the –log10 FDR corrected p-value (q-value) against log2 fold change for significantly (q<0.05) up-regulated (orange), down-regulated (blue) genes and non- differentially expressed genes (grey). Embedded plots are frequency histograms of log2 fold change (e.g. a bar above a log2 fold change of 2 represents the genes that range in a log2 fold change between 1 and 2). Panels C and D are heatmaps of differential expression in stickleback exposed to 12 °C and 24 °C during development, respectively. Each row represents the expression value (log2 counts per million) for a single gene relative to the mean expression value for that gene across all individuals (orange representing higher expression and blue representing lower expression). Each column represents an individual stickleback. Columns 1-6 are stickleback that were kept at 18 °C for the duration of the experiment. Columns 7-12 are either stickleback that were kept at 18 °C for 8 months and then acclimated to 5 °C (C) or 25 °C (D) for four weeks.   Of the 10,140 genes that were differentially expressed in response to adult thermal acclimation, 4,851 were differentially expressed in both cold- and warm-acclimated stickleback (Figure 3.4A), and the majority of these genes (4,235 genes) were differentially expressed in the same direction between acclimation temperatures (Figure 3.4A). In contrast, of the 57 genes ●●●●●●●● ●● ● ●●● ● ●● ●● ●● ●●●−4 −2 0 2 4051015log2 Fold Change●●● ●● ● ●●● ●●● ● ● ●● ●●● ● ●●●●●● ●● ●●●● ●●●● ●● ●●● ● ●●● ● ● ●●● ●● ●●● ● ● ●● ●●● ●● ●● ●● ● ●● ●●● ●●●●●● ●●● ● ●●● ●● ● ●● ●●● ● ●● ● ●● ●●● ● ● ● ●●● ●● ●●● ●● ●●●●●● ● ● ● ●●● ●● ●● ●● ● ● ● ●●● ●●● ● ●●● ●●● ● ●● ●● ● ●● ●●●● ●● ●● ● ●●●● ● ●●● ●● ●● ●● ●●● ● ●● ●● ● ●●●● ● ●● ●●● ●● ●●●●−log10 Q−value●●●●●●●● ●● ●●●●●● ● ●●● ●−4 −2 0 2 4051015log2 Fold Change●●●●●● ● ●●● ●● ● ●● ●●●●● ●● ●● ●● ●●●● ● ●●● ●●● ● ●●●●● ● ●●● ● ●● ● ●●●● ●● ●● ●● ●● ●● ●●●● ● ●● ●● ● ● ●●●● ●● ●● ●●●● ●● ●● ● ●● ●● ●● ●●●●● ●● ● ●● ●● ●●● ●●●●●● ●●●● ● ●●●● ●●● ● ●● ●● ●● ● ●● ●●● ●● ●●●●● ●●● ●● ●● ●●● ●●●●● ●●●● ●● ●● ● ●●●●●● ●● ●● ●●● ●●●● ●● ● ●●●● ● ●●● ●●●●−log10 Q−value18 °C 12°C 18 °C 24 °C C D A B −4 −2 0 2 4051015log2 Fold ChangeFrequency−4 −2 0 2 4051015log2 Fold ChangeFrequency0 2 4 -2 -4 log2 Fold Change 0 2 4 -2 -4 log2 Fold Change   1  1  -log 10 Q-value 0 2 4 -2 -4 log2 Fold Change 0 2 4 -2 -4 log2 Fold Change         Frequency Frequency 0 0.5 1 -0.5 -1 66  identified in stickleback from different developmental temperatures, only five were differentially expressed in stickleback from both warm and cold developmental treatment temperatures (Figure 3.4B). The expression of all five of these genes changed in the same direction in stickleback exposed to both cold and warm developmental temperatures.  To further assess the degree to which developmentally plastic responses to environmental temperature are consistent with adult thermal acclimation responses, we compared the list of genes that were differentially expressed between developmental treatments to those that were differentially expressed in adult stickleback acclimated to different temperatures. From this analysis, we identified 27 genes that were differentially expressed both in response to development at cold temperatures and in response to cold temperature acclimation (Figure 3.4C, Table 3.1). The direction of differential expression for all 27 genes was conserved across developmental and adult treatments. Similarly, we identified 18 genes that were differentially expressed in response to development at warm temperatures and in response to warm temperature acclimation (Figure 3.4C). The direction of differential expression for 17 of these genes was conserved between developmental and acclimation treatments (Table 3.1). Comparison of the differentially expressed genes from all four analyses identified four genes (irs2b, klhl38b, gadd45ga, and slc3a2a) that were differentially expressed in all treatments and each of these genes was up-regulated in each treatment (Table 3.1).  In addition to comparisons of gene expression of the 12 °C and 24 °C developed groups to the group developed at 18 °C, we also examined the list of expressed transcripts in the stickleback that developed at 12 °C and 24 °C to determine whether there were any novel genes that were expressed at one temperature but not the other, because these genes would not 67  necessarily be revealed by comparison to stickleback developed at 18 °C. However, there were no genes that fell into this category.  Figure 3.4: (A) Comparison of up- and down-regulated genes in stickleback muscle tissue from adult stickleback acclimated as adults to 5 °C (blue) or 25 °C (red) for four weeks. (B) Comparison of up- and down-regulated genes in stickleback muscle tissue from stickleback exposed to 12 °C (blue) or 24 °C (red) during development and then kept at 18 °C until 9 months of age. (C) Comparison of differentially expressed genes in stickleback muscle tissue. Cold acclimation = fish developed and reared at 18 °C for eight months and then acclimated to 5 °C for four weeks. Warm acclimation = fish developed and reared at 18 °C for eight months and then acclimated to 25 °C for four weeks. Cold development = fish developed at 12 °C until hatch, and then held at 18 °C for nine months. Warm development = fish developed at 24 °C until hatch, and then held at 18 °C for nine months. All differential expression was identified relative to fish held at 18 °C for the duration of the experiment.  A	B	 C	A	B	 C	A B C 68  Table 3.1: Genes that are differentially expressed in response to changes in both developmental temperatures and adult acclimation temperatures.     Log2 Fold Change Ensembl Gene ID Gene Name Cold Development Cold Acclimation Warm Development Warm Acclimation ENSGACG00000003564 irs2b 0.80 1.06 1.03 1.76 ENSGACG00000006793 gadd45ga 0.95 0.89 1.52 2.75 ENSGACG00000006167 klhl38b 1.08 2.60 1.25 1.15 ENSGACG00000019745 slc3a2a 1.48 1.28 1.23 2.02 ENSGACG00000013368 bcl2b -1.31 -1.04 - - ENSGACG00000007797 vkorc1 -1.27 -2.56 - - ENSGACG00000014656 ifi35 -1.24 -1.00 - - ENSGACG00000008237 EVC -1.11 -0.97 - - ENSGACG00000004133 pim1 -0.92 -1.33 - - ENSGACG00000003021 klf2b -0.81 -1.71 - - ENSGACG00000019700 DDIT4L -0.81 -3.73 - - ENSGACG00000012452 IL16 -0.77 -2.23 - - ENSGACG00000009575 id3 -0.66 -1.04 - - ENSGACG00000005010 tob1a 0.57 0.70 - - ENSGACG00000010010 ezrb 0.89 0.61 - - ENSGACG00000008895 sesn1 0.94 1.79 - - ENSGACG00000006997 alas1 (1 of 2) 1.00 1.66 - - ENSGACG00000015298 ENSGACG00000015298 1.28 -1.53 - - ENSGACG00000015297 METTL21C (2 of 2) 1.33 -3.44 - - ENSGACG00000013859 zc3h12a 1.40 0.80 - - ENSGACG00000005398 tfr1b 1.56 3.73 - - ENSGACG00000017927 cebpd 1.65 1.15 - - ENSGACG00000002379 ddit4 1.70 1.54 - - ENSGACG00000016373 tfr1a 1.72 2.60 - - ENSGACG00000010739 klf13 1.79 1.42 - - ENSGACG00000001466 fkbp5 1.89 3.16 - - ENSGACG00000001632 samhd1 (2 of 3) 2.11 2.72 - - ENSGACG00000000049 rasgef1ba - - 1.20 2.52 ENSGACG00000001607 trim63b - - 0.96 1.96 ENSGACG00000006161 fbxo32 - - 1.52 2.15 ENSGACG00000006480 ddit3 - - 0.58 0.42 ENSGACG00000006908 ENSGACG00000006908 - - 1.94 -1.01 ENSGACG00000008429 ENSGACG00000008429 - - 3.48 1.10 ENSGACG00000010788 nr4a1 - - -0.70 -1.97 ENSGACG00000010861 RASGRF1 (1 of 2) - - -1.89 -3.88 ENSGACG00000011050 DUSP8 - - 0.85 0.51 ENSGACG00000011743 HIVEP2 (1 of 2) - - 0.74 0.75 ENSGACG00000013469 pptc7a - - 1.50 1.57 ENSGACG00000014133 irs2a - - 0.72 0.88 ENSGACG00000015066 camk2n1a - - 1.49 2.36 ENSGACG00000016438 ALPK3 (2 of 2) - - 1.83 1.08  69  3.3.2 Identification of candidate biological processes affected by gene expression plasticity across timescales Comparison of the most significantly enriched biological processes associated with genes that were differentially expressed in response to acclimation to low or high temperature revealed that acclimation to both low and high temperature induced changes in the expression of genes involved in a common set of biological processes (Table 3.2). Enriched biological processes for the up-regulated genes were generally associated with cell division, mRNA splicing, and protein degradation. Enriched biological processes for the down-regulated genes were generally associated with extracellular matrix organization and cell adhesion in both warm and cold-acclimated individuals. There were no significantly enriched terms for genes that were uniquely up-regulated in cold-acclimated stickleback. Genes that were uniquely up-regulated in warm-acclimated stickleback were enriched for biological processes involved in protein translation and amino acid metabolism. Genes that were uniquely down-regulated in cold-acclimated stickleback were enriched for biological processes involved in muscle filament sliding, muscle contraction, oxidation-reduction, angiogenesis, and epidermis development. There were no significantly enriched processes associated with genes that were uniquely down-regulated in warm-acclimated stickleback.         70  Table 3.2: Significantly enriched biological process gene ontologies for genes that are up- or down-regulated in response to warm or cold thermal acclimation in adult stickleback muscle tissue.   Cold Acclimation   Warm Acclimation   GO ID GO term Over represented p-value   GO ID GO term Over represented p-value Up-regulated genes GO:0000278 mitotic cell cycle 4.06E-21   GO:0010467 gene expression 1.41E-34 GO:0010467 gene expression 1.53E-18   GO:0016032 viral process 4.78E-16 GO:0000398 mRNA splicing, via spliceosome 1.08E-17   GO:0000278 mitotic cell cycle 2.47E-15 GO:0008380 RNA splicing 3.51E-17   GO:0006364 rRNA processing 3.47E-15 GO:0008033 tRNA processing 5.51E-13   GO:0008033 tRNA processing 4.40E-15 GO:0006521 regulation of cellular amino acid metabolic process 1.03E-12   GO:0031145 anaphase-promoting complex-dependent proteasomal ubiquitin-dependent protein catabolic process 5.97E-15 GO:0043687 post-translational protein modification 4.59E-10   GO:0051439 regulation of ubiquitin-protein ligase activity involved in mitotic cell cycle 1.16E-14 GO:0006364 rRNA processing 1.51E-09   GO:0051437 positive regulation of ubiquitin-protein ligase activity involved in regulation of mitotic cell cycle transition 2.24E-14 GO:0018279 protein N-linked glycosylation via asparagine 1.57E-09   GO:0051436 negative regulation of ubiquitin-protein ligase activity involved in mitotic cell cycle 2.96E-14 GO:0031145 anaphase-promoting complex-dependent proteasomal ubiquitin-dependent protein catabolic process 1.76E-09   GO:0043488 regulation of mRNA stability 5.12E-14               Down-regulated genes GO:0030198 extracellular matrix organization 8.88E-28   GO:0030198 extracellular matrix organization 1.19E-32 GO:0030574 collagen catabolic process 5.56E-17   GO:0030574 collagen catabolic process 6.14E-22 GO:0007155 cell adhesion 9.80E-15   GO:0022617 extracellular matrix disassembly 1.30E-18 GO:0030049 muscle filament sliding 2.11E-14   GO:0007155 cell adhesion 5.68E-15 GO:0006936 muscle contraction 1.28E-13   GO:0030199 collagen fibril organization 4.14E-14 GO:0022617 extracellular matrix disassembly 3.78E-13   GO:0007160 cell-matrix adhesion 7.29E-09 GO:0001525 angiogenesis 1.95E-09   GO:0001501 skeletal system development 1.95E-08 GO:0030199 collagen fibril organization 2.55E-09   GO:0007156 homophilic cell adhesion via plasma membrane adhesion molecules 1.95E-07 GO:0001501 skeletal system development 7.58E-09   GO:0051056 regulation of small GTPase mediated signal transduction 3.40E-07 GO:0007156 homophilic cell adhesion via plasma membrane adhesion molecules 2.87E-07   GO:0005980 glycogen catabolic process 1.28E-06   71  There was no significant enrichment of biological processes for genes differentially expressed between stickleback that experienced different temperatures during development. This is likely due to the relatively small number of genes affected by developmental temperature; however, many of these genes are known to be involved in the same processes that were enriched among genes differentially expressed in adult stickleback acclimated to different temperatures. For example, transcripts encoding genes for the DNA damage inducible transcript 4 (ddit4/redd1) and ddit4-like (ddit4l/redd2) were differentially expressed in stickleback reared at 12 °C and are thought to be involved in the attenuation of the mTORC1 protein synthesis which can result in muscle atrophy (Kelleher et al. 2013). In addition, several genes involved in muscle cell development, growth, aging and metabolism were also differentially expressed in stickleback that developed at 12 °C, including methyltransferase like 21C (mettl21c), CCAAT/enhancer binding protein delta (cebpd), transferrin receptor 1a and 1b (tfr1a, tfr1b/tfrc), pim-1 proto-oncogene, serine/threonine kinase (pim1), B-cell CLL/lymphoma 2b (bcl2b), FK506 binding protein 5 (fkbp5), Kruppel-like factor 2b and 13 (klf2b, klf13), transducer of ERBB2, 1 (tob1), and DNA-binding protein inhibitor ID-3 (id3).   A different set of genes that have been implicated in muscle development, growth and metabolism were differentially expressed in stickleback that developed at 24 °C including Ras protein specific guanine nucleotide releasing factor 1 (rasgrf1), alpha kinase 3 (alpk3), f-box protein 32 (fbxo32), dual specificity phosphatase 8 (dusp8), tripartite motif containing 63b (trim63b/murf1b), endothelial lipase G (lipg), nuclear receptor subfamily 4 group A member 1 (nr4a1/nur77), and plexin A2 (plxna2).  Comparison of genes that were differentially expressed by either warm or cold development identified five genes whose expression was affected by both developmental 72  temperatures (Figure 3.4B, 3.4C), including the solute carrier family 3 member 2a (slc3a2a), growth arrest and DNA-damage-inducible, gamma a (gadd45ga), insulin receptor substrate 2b (irs2b), kelch-like family member 38b (klhl38b), and one unannotated gene ENSGACG00000008429. All but the last of these genes were also affected by thermal acclimation.  3.3.3 Comparison to patterns of DNA methylation  Because the samples used in this experiment were derived from the same individuals used to demonstrate that both altered developmental temperature and altered adult acclimation temperature result in changes in DNA methylation patterns in threespine stickleback muscle tissue (Metzger & Schulte 2017; Chapter two of this thesis), it is possible to directly compare changes in DNA methylation patterns to changes in gene expression. In this analysis we compared the genes that we identified in this study as being differentially expressed to DMCs that were located near a gene (within 5 kb of either the transcription start site or the 3’ end of the gene) identified in our previous study. Using this identification cutoff, none of the genes that were differentially expressed between developmental temperature treatments were associated with previously reported differential methylation, whereas 125 genes that responded to warm acclimation were associated with differentially methylated loci, and 199 genes that responded to cold acclimation were associated with differentially methylated loci. Of these DMCs, the majority were located in intergenic regions within 5 kb upstream or downstream of the differentially expressed genes (Figure 3.5), and only 5-6% (depending on the acclimation temperature) were located in promoter regions (within 2 kb upstream of the transcription start site). When considering the complete reduced representation bisulfite sequencing (RRBS) 73  dataset, approximately 9% of sequence data were from promoter regions. In the subset of the data analyzed here, which excludes intergenic sequences located more than 5kb upstream or downstream of a gene (thus effectively excluding much of the intergenic sequence), the proportion of promoter sequences in the background sequence data is higher. Thus, the fact that we observe a lower percent of DMCs in promoter regions suggests that there is no evidence for enrichment of differential methylation in the promoter regions of differentially expressed genes.   Figure 3.5: Genomic distribution of differentially methylated cytosines associated with differentially expressed genes (within 5 kilobase pairs) in stickleback muscle tissue from adult stickleback acclimated to 5 °C (cold acclimation) or 25 °C (warm acclimation).  We did not detect significant enrichment of gene ontologies among the differentially expressed genes associated with DMCs, but a number of these genes are involved in processes 6%13%26%56%Cold Acclimationpromoterexonintronintergenic5%11%27%57%Warm Acclimation74  that are likely affected by thermal acclimation based on the GO enrichment analysis of the DE genes, including processes such as proper formation of the sarcomere (eg. sh3bgr and ttn.2), mRNA splicing (aqr, sf3b6, and pus10), the ubiquitin proteasome pathway (eg. psmd1, psmd13, ube2e2), muscle cell growth and development (eg. col12a1a and relb), mitochondrial proliferation (eg. mdh1, hpc2/elac2 and ugp2b), and myogenesis (eg. foxk1i).  3.4 Discussion 3.4.1 Conserved effects of cold and warm temperature on gene expression  In this study we present evidence of similarities in temperature-induced gene expression plasticity at different life stages on adult gene expression patterns. In both the developmental and adult acclimation treatments, colder temperatures resulted in the differential expression of more genes compared to warmer temperatures, and there were more differentially expressed genes that were up-regulated compared to down-regulated in all treatments. While the effects of adult thermal acclimation on gene expression were much more extensive than the effects of developmental temperature on gene expression in adult stickleback muscle, the majority of gene expression patterns that were affected by developmental temperature were also responsive to thermal acclimation, and these changes were generally in the same direction. Taken together these data suggest that there could be similar mechanisms that regulate plastic responses at these different life stages.  The effects of thermal acclimation on muscle gene expression have been examined in many fish species (eg. Gracey et al. 2004; Scott & Johnston 2012; Healy et al. 2016) including threespine stickleback (Morris et al. 2014b); however, studies that contrast the gene expression patterns of cold- or warm-acclimated fish are less common (eg. Ikeda et al. 2017). In this study 75  we characterized the transcriptional response to cold (5 °C) and warm (25 °C) temperature acclimation independently by comparing muscle gene expression patterns to those in individuals at a common control temperature (18 °C). By using this experimental design we are able to identify the common set of genes (Figure 3.4C) and biological processes (Table 3.2) that are differentially regulated in response to both warm and cold temperature acclimation, determine whether direction of regulation of these genes is similar at both temperatures (Figure 3.4B), identify genes that are uniquely regulated in one temperature but not the other (Figure 3.4A and 3.4C), and compare the overall magnitude of the transcriptomic response to warm and cold temperature acclimation (Figure 3.2). In contrast, experiments that compare differentially expressed between two temperature extremes could potentially lead to an oversimplification of patterns of thermal acclimation, if (as demonstrated here) both warm and cold acclimation lead to similar changes in gene expression. For example, a previous study of variation in thermal acclimation responses between marine and freshwater populations of stickleback compared gene expression patterns in stickleback acclimated to 7 °C compared to stickleback acclimated to 22 °C (Morris et al. 2014b). Many of the enriched biological processes that we identified for genes that were up-regulated in both cold- and warm-acclimated stickleback, such as gene expression and protein ubiquitination, were previously reported as being up-regulated in stickleback acclimated to warm temperatures (Morris et al. 2014b). The advantages of comparing thermal extremes to an intermediate temperature can be illustrated with an example – the gene encoding the peroxisome proliferator activated receptor alpha (pparaa). This gene has been previously associated with a genetically divergent region of the stickleback genome that is under positive selection between marine and freshwater stickleback populations (Jones et al. 2012a), and has been shown to be up-regulated in stickleback from a freshwater population acclimated to cold 76  conditions but not in a stickleback from a marine population acclimated to cold conditions (Morris et al. 2014b). Here, we detected up-regulation of pparaa transcripts in stickleback from a marine population in response to both warm and cold acclimation. Thus, if only stickleback acclimated at warm and cold temperatures are compared, patterns of differential expression can be missed.  Disruption in the timing or magnitude of gene expression during development can result in permanent changes to adult phenotypes (West-Eberhard 2005). These effects have been previously demonstrated in zebrafish where warm temperatures during development have been shown to impact muscle cell fiber composition, swimming performance, and metabolism (Scott & Johnston 2012). Interestingly, in zebrafish these large changes in muscle phenotype and function were not associated with large changes in the transcriptional program of muscle tissue in adults. The magnitude of the transcriptional response in zebrafish muscle was modest (26 DEGs with a q-value < 0.05) and similar to what we observed in stickleback (33 and 29 DEGs with a q-value < 0.05 in stickleback exposed to either cold or warm temperatures during development, respectively). Genes that were differentially expressed between stickleback that developed at different temperatures are know to be involved in metabolism and muscle cell development (eg. cebpd, tfr1, fkbp5, and klf2 in stickleback that developed at 12 °C, nr4a1 in stickleback that developed at 24 °C, and slc3a2 in both developmental treatment temperatures). Among this relatively small set of genes there were two genes, fkbp5 and slc3a2, that were also affected by developmental temperature in zebrafish, These data suggest that altered developmental temperatures induce changes in muscle phenotype that could impact the performance of adults, and that the effect of developmental temperature on the regulation of these genes may be conserved across distantly related fish species.  77    Plasticity is thought to influence evolutionary processes by revealing previously hidden sources of variation and by facilitating phenotypic variation and divergence among populations that inhabit different environments (Schlichting 2008; Schneider & Meyer 2017). Natural selection can then act on this variation to refine the plastic phenotype closer to the optimum (Levis & Pfennig 2016). For example, in a “plasticity-first” hypothesis, developmental plasticity can induce phenotypes that increase fitness in stressful environments (Levis & Pfennig 2016). However, this type of analysis assumes the presence of beneficial plasticity, whereas maladaptive or non-adaptive plasticity can also affect evolutionary trajectories (Ghalambor et al. 2015). The differential expression of genes in stickleback reared at colder or warmer temperatures is consistent with the potential effects of developmental temperature on metabolism and muscle cell fiber composition that could impact adult performance, but whether these changes are likely to be adaptive or maladaptive is unknown.  Examination of the genes differentially regulated in response to all four of our temperature treatments strongly suggests that changes in the regulation of muscle growth are a core phenomenon uniting phenotypic plasticity in response to temperature change at different life stages and in response to both low and high temperature. Although there have been relatively few studies that directly compare plastic responses at the molecular level between life stages, one study comparing rapid cold hardening and exposure to altered developmental temperatures in Drosophila melanogaster (Gerken et al. 2015) detected effects on similar functional classes of genes at different life stages, although the genes themselves differed. This commonality in responses across life stages and temperatures could potentially be interpreted in two different ways. First, this pattern could reflect a common underlying mechanism regulating plasticity. An alternative, but not mutually exclusive, interpretation of this pattern is that these plastic 78  responses at different life stages represent beneficial or adaptive responses to thermal change. For example, the persistent effects of developmental temperature on gene expression patterns could be adaptive by shifting gene expression levels closer to those observed in fish at lower and higher temperatures or by reducing the cost of thermal acclimation.   Although the phenotypic consequences of the gene expression patterns observed here are unknown, a previous study in zebrafish (Schaefer & Ryan 2006) found that acclimation to increased temperature increased whole-organism thermal tolerance, and increased developmental temperature had a similar, but much smaller, effect. Taken together, this suggests that changes in gene expression may have effects on a variety of traits at the whole-organism level.   3.4.2 Relationship to patterns of DNA methylation Epigenetic mechanisms have been proposed to be important processes through which environmentally induced variation in phenotypic expression in one generation can impact subsequent generations and through which environmental conditions during development can have persistent phenotypic effects later in life. To explore the relationship between DNA methylation patterns and the effects of changes environmental temperature during development and in adults on gene expression, we compared the differentially expressed genes identified in this study to differentially methylated loci from the same stickleback reported in our previous study. Differentially expressed genes in stickleback from different developmental temperatures were not closely associated with any of the differentially methylated loci that were previously described, suggesting a potential role of DNA methylation on trans-acting factors that regulate gene expression. Alternatively, differentially methylated regions could be located on regulatory elements that are not within 5 kb of a gene, or the persistent effects of developmental 79  temperature on gene expression may not be regulated by DNA methylation and instead could be due to effects such as changes in histone acetylation or changes in miRNA activity. However, we cannot firmly rule out a potential role for changes in DNA methylation in regulating the changes in gene expression that we observed because important regulatory regions located near differentially expressed genes may not have been assessed in the reduced representation analysis of DNA methylation (RRBS). For example, only ~10% of the differentially expressed genes were represented within the RRBS dataset, highlighting a potential limitation when deducing functional relationships that are based on a correlation between a reduced representation approach and a more comprehensive technique such as RNA-seq.  Analysis of differentially expressed genes and differentially methylated loci in adult stickleback acclimated to warm and cold temperatures identified genes involved in several processes that are known to be differentially regulated in marine ectotherms in response to changes in environmental temperatures, including the ubiquitin-proteasome pathway, aerobic metabolism and mitochondrial proliferation, mRNA splicing, myogenesis, proper formation of the sarcomere, and muscle cell growth and development. These data suggest that modified DNA methylation levels are likely involved in many of the transcriptional responses of stickleback to variation in environmental temperatures. One potentially interesting candidate gene that was both differentially expressed and associated with at least one DMC in adult stickleback acclimated to different temperatures is phosphodiesterase 4B (pde4ba). Epigenetic mechanisms are thought to play an important role in an organism’s capacity to adapt to environmental changes, particularly over shorter timescales, but there is little empirical evidence to support this hypothesis. The pde4ba gene has been previously associated with the genetic divergence of Baltic sea stickleback populations along a thermal gradient (Guo et al. 2015) and thus may play an adaptive role to 80  changes in environmental thermal regimes. It is therefore a strong candidate for subsequent analysis of the effects of DNA methylation on adaptive evolutionary processes.  Taken together, these data demonstrate that changes in DNA methylation patterns are likely implicated in short-term, potentially reversible transcriptional response of adults to changes in environmental temperature.  3.5 Conclusions  The data presented here demonstrate that the temperature experienced during early stages of development (before hatch) can have persistent effects on gene expression patterns in adult stickleback muscle tissue. In addition, we demonstrate that the majority of differentially expressed genes in stickleback from different developmental temperatures are also differentially expressed as part of the adult stickleback thermal acclimation response. This pattern suggests that developmental plasticity and phenotypic flexibility in gene expression in response to temperature change share some common underlying mechanisms and may have similar functional consequences. However, adult acclimation resulted in a much larger overall change in the transcriptome than did developmental temperature exposure. Some genes that were differentially expressed as a result of the adult acclimation treatments were also associated with previously identified temperature-dependent effects on DNA methylation patterns, suggesting a potential role for epigenetic mechanisms in regulating plastic responses during acclimation. Overall, these results emphasize both the similarities and differences between developmental plasticity and phenotypic flexibility in adults and highlight the relationships between plasticity acting across different time scales. 81  Chapter 4: The DNA methylation landscape of stickleback reveals patterns of sex chromosome evolution and effects of environmental salinity.  4.1 Introduction  Epigenetic variation has the potential to impact ecological and evolutionary processes, and thus affect species distributions and evolutionary trajectories (Bossdorf et al. 2008; Flores et al. 2013; Jablonka and Raz 2009; Varriale 2014; Franks and Hoffmann 2012). Currently, one of the best-studied mechanisms underlying epigenetic variation is DNA methylation, a heritable epigenetic modification in which a methyl group is added to position 5 of the pyrimidine ring on a cytosine (5mC), most commonly found on cytosine-phosphate-guanine (CpG) dinucleotides in vertebrates (Heard and Martienssen 2014). Changes in DNA methylation can have profound effects on chromatin structure, which can in turn alter gene expression (Klose and Bird 2006; Jaenisch and Bird 2003). The addition or removal of these methyl groups can be dynamically regulated in response to changes in the environment (Cooney et al. 2002; Kucharski et al. 2008; Boyko et al. 2010). Variation in DNA methylation levels have therefore been hypothesized to play a key role in mediating phenotypic responses to environmental change (Bossdorf et al. 2008; Flores et al. 2013; Hofmann 2017), and may represent a dynamic source of heritable variation that can respond to changes in the environment and influence phenotypic variation over multiple time-scales (Richards 2006).  In addition to its potential role in regulating gene expression in response to environmental change, DNA methylation is also critical in regulating gene expression in dosage compensation systems that have evolved to minimize the unequal expression of genes on heteromorphic sex 82  chromosomes (Graves 2016). In older XY sex chromosome systems, such as those found in most mammalian species, DNA methylation is involved in the global silencing of one of the two X chromosome in females (Graves 2016). In ZW sex chromosome systems (which have a female-specific W chromosome) such as those in birds and some reptiles and fishes, DNA methylation is involved in gene-specific dosage compensation via the activation or suppression of particular dosage sensitive genes (Graves 2016). While epigenetic silencing has been well established as a mechanism involved in dosage compensation of older heteromorphic sex chromosome systems, patterns of DNA methylation in young sex chromosome systems are less well understood, but have been hypothesized to play a key role sex chromosome evolution (Gorelick 2003).   The threespine stickleback (Gasterosteus aculeatus) has been extensively used to investigate the genetic basis of adaptive evolution to novel environments (Jones et al. 2012a; Jones et al. 2012b). Following the last glaciation, ancestral marine populations of stickleback colonized and adapted to newly available freshwater habitats in the north-temperate zone (Bell and Foster 1994). Adaptation to these novel environments drove the rapid parallel evolution of divergent phenotypes including changes in body shape, armor plate number, gene expression levels, and gene expression plasticity (Jones et al. 2012a; Gibbons et al. 2017; Morris et al. 2014; McCairns & Bernatchez 2010; Ishikawa et al. 2017). Several studies have used reduced representation approaches to characterize variation in DNA methylation patterns between stickleback that vary in their lateral plate morphology, and have suggested that variation in DNA methylation patterns may contribute to the phenotypic divergence observed between marine and freshwater populations (Smith et al. 2015; Trucchi et al. 2016; Artemov et al. 2017). Threespine stickleback also have a relatively young XY sex chromosome pair that has evolved since the species first arose ~13-16 Ma (Kawahara et al. 2009; Ross et al. 2009; Bell et al. 2009), and this 83  species has become a powerful model system to explore the evolution of heteromorphic sex chromosomes and dosage compensation mechanisms (Schultheiß et al. 2015; White et al. 2015). Thus, the threespine stickleback is an ideal model in which to investigate the complementary roles of DNA methylation in both environmental adaptation and the evolution of sex chromosome systems.  In this study we present the first high-resolution analysis of DNA methylation patterns in the stickleback genome using whole genome bisulfite sequencing (WGBS). This approach allowed us to characterize prominent features in the DNA methylation landscape of stickleback, including variation in DNA methylation patterns between males and females along the entire stickleback sex chromosome, which provides insight into the relationship between epigenetic mechanisms and sex chromosome evolution. By rearing putatively ancestral marine stickleback at both low and high salinity, we also describe the effects of environmental salinity on genomic DNA methylation patterns, and highlight potential salinity responsive genes that may be differentially regulated by DNA methylation.  4.2 Material and methods 4.2.1 Fish collection  All animal experimentation was conducted according to University of British Columbia approved animal care protocols (A10-0285 and A11-0372). Adult threespine stickleback (G. aculeatus) of the fully plated “marine” ecotype were collected at the beginning of their natural spawning season in May 2013 from Oyster Lagoon, British Columbia in Canada (GPS: 49.6121,-124.0314). Fish were separated into six 110-litre glass tanks (20 fish per tank) and acclimated to 21 ppt salt water (dechlorinated Vancouver municipal tap water supplemented with Instant 84  Ocean Sea Salt), 18 °C and 14:10 h light:dark photoperiod. These conditions mimic the natural environmental conditions at the collection location at the time of collection. Fish were fed daily to satiation with Hakari Bio-Pure frozen Mysis Shrimp and were acclimated to laboratory conditions for four weeks.  4.2.2 Fertilization and rearing procedure  To determine the impact of salinity on fertilization and hatching, eggs were collected from gravid female stickleback and immediately divided into six different petri dishes containing 5 ml of 2, 7, 14, 21, 28, or 35 ppt saltwater. Testes were collected from males displaying sexually mature characteristics and individually macerated in a 1.75 mL microcentrifuge tube containing 300 mL Ginzberg’s fish Ringer’s solution. Eggs from a single clutch were fertilized with sperm solution from a single male across all salinities (50 mL of sperm solution for each petri dish at each different salinity). Following fertilization, an additional 10mL of water at the appropriate salinity was added to each petri dish. This process was repeated ten times creating a total of ten different families, each fertilized at all salinities. Petri dishes were partially covered to prevent water loss via evaporation and to allow for surface gas exchange. Eggs were monitored twice daily during which time any unfertilized eggs were removed and 10 mL of water was changed with sterilized water of the appropriate salinity to prevent mold growth. Percent fertilization and percent hatch were recorded. Percent hatch is recorded as the proportion of fertilized embryos that hatched. The effect of salinity on fertilization and hatching success was analyzed using a logistic regression (Warton & Hui 2011) in the R v3.3.2 base stats package. Tukey post hoc analysis was performed using the glht() function in the multicomp v1.4-6 R package.  85   After all fish in a petri dish had hatched and the yolks had been absorbed (~15 days post fertilization), larvae were transferred to hanging net boxes (Aquaclear) in 110 L glass aquaria containing water at the fertilization salinity. Sponge filters were used for filtration and aeration. Each family was kept separate throughout the experiment. At one-month post hatch whole animals were snap frozen in liquid nitrogen.  4.2.3 Whole genome bisulfite sequencing (WGBS)  Genomic DNA was isolated from one-month old whole fish samples from the 2 ppt and the 21 ppt salinity treatments using a Qiagen DNeasy Blood and Tissue Kit following the manufacturer’s recommended protocol for RNA-free genomic DNA using RNAase A. The sex of each sample was identified by PCR analysis using primers designed to idh, gasm6, and stn190 following previously described methods (Toli et al. 2016). Genomic DNA samples from three males and three females from 2 ppt and 21 ppt (twelve samples total) were sent to the McGill University and Genome Quebec Innovation Center for DNA quality assessment, library preparation, bisulfite treatment, and 150 base pair paired-end sequencing using an Illumina HiSeqX. The 12 samples were split evenly across 3 sequencing lanes (4 samples/lane) such that one male and female sample from each of the salinity treatments were represented on each sequencing lane. Average sequencing library size was 102,011,555 reads (± 13,147,873 SD).   4.2.4 WGBS data analysis  Reads were mapped to a revised assembly of the stickleback genome (Glazer et al. 2015) obtained from the Dryad Digital Repository (http://datadryad.org/resource/doi:10.5061/dryad.q018v) and DNA methylation levels were 86  calculated using the bisulfite sequencing plugin v1.2 in CLC Genomics Workbench v10.0. Average mapping efficiency was 89.5 % (± 1% SD). DNA methylation data were exported from CLC and analyzed using in R v3.3.2 using methylKit package v3.5 (Akalin et al. 2012) following previously recommended guidelines for bisulfite sequence analysis (Ziller et al. 2015; Wreczycka et al. 2017). Sequenced CpG loci were filtered so that only sites with at least 10 reads in each of the 12 samples were retained. Sites that were in the 99.9th percentile of coverage were also removed from the analysis to account for potential PCR bias.   Hierarchical cluster analysis was conducted using Ward’s method and was implemented using the clusterSamples() function. Pairwise comparisons between groups were performed using a logistic regression model with a correction for overdispersion using the calculateDiffMeth() function followed by a Chi-square test to identify significantly differentially methylated cytosines (DMCs) between groups. The p-values for DMCs were false discovery rate (FDR) corrected using the sliding linear model method (SLIM) with a maximum q-value threshold of 0.05 and a minimum change in percent methylation of 10%. For the comparison between males and females, salinity and family were included as covariates. For the comparison between salinity rearing treatments, sex and family were included as covariates. For comparisons between families, sex and salinity were included as covariates. To calculate mean methylation levels across 10 kilobase (kb) genomic regions, the tileMethylCounts() function in methylKit was used to calculate DNA methylation values across sequential 10 kb windows of the genome. All figures were generated in R.  To obtain nearest neighboring gene annotations, the gene coordinates in the annotation file (.gtf) provided by Glazer et al. (Glazer et al. 2015) in the dryad digital repository (which contains gene coordinates that correspond to the stickleback genome available in Ensembl) were 87  converted to the coordinates in the updated assembly using the R script convertCoordinate.R that is provided by the authors of the revised assembly. Distances to nearest neighboring genes were calculated using the annotatePeakInBatch() function in the R package ChIPpeakAnno v3.6.5.  CpG islands for the revised stickleback genome assembly were calculated using python scripts (https://github.com/lucasnell/TaJoCGI) that apply an algorithm based on the methods described by Takai and Jones (Takai & Jones 2002). The observed distribution of DMCs was compared to the distribution of CpGs across the genome using a Chi-square test.  4.2.5 Candidate gene analysis  Previous RNA-seq studies have identified many genes that respond to changes in salinity in stickleback (2, 771 in gill tissue (Gibbons et al. 2017) and 1,844 in kidney tissue (Wang et al. 2014)). To determine whether DNA methylation could be involved in the differential regulation of these candidate genes we compiled a list of salinity responsive genes from previous studies (4,615 candidate genes) and compared them to genes within 2 kb of DMCs in stickleback reared at different salinities.  Similarly, several studies have also characterized sex-biased gene expression patterns in stickleback. We therefore compiled a list of 2,282 genes that display sex-biased gene expression patterns in brain (1,255 genes (Metzger & Schulte 2016a)) and liver tissue (1,268 genes (Leder et al. 2010)), and compared these to genes within 2 kb of DMCs that were identified between male and female stickleback.  We also compared genes within 2 kb of DMCs between individuals reared a different salinities to genes associated with single nucleotide polymorphisms (SNPs) under positive 88  selection in threespine stickleback from freshwater environments compared to marine environments (Jones et al. 2012a).   4.3 Results and discussion 4.3.1 Characterization of the stickleback methylome  We performed WGBS on fish from a marine population of threespine stickleback reared from fertilization to the age of 1 month at a salinity of either 2 ppt or 21 ppt. These salinities represent the widest range that still allows good fertilization and hatching success in this population (Appendix C, Figure C.1). In this study we utilized a balanced design with WGBS performed on one male and one female from each of three families and each of the salinity rearing treatments. This design was chosen because genetic variation has been shown to have substantial effects on the divergence of DNA methylation patterns among individuals in both plants and animals (Gertz et al. 2011; McRae et al. 2014; Vidalis et al. 2016). Consistent with this observation, we detected strong effects of family on DNA methylation (Appendix C, Figure C.2 and C.3). However, family-level variation in DNA methylation levels could also be indicative of transgenerational environmental or maternal effects (Jablonka and Raz 2009).   CpG loci in the stickleback genome had an average methylation level of 70.3% (Appendix C, Figure C.4), which is consistent with whole genome assessments of methylation in other fish species (Feng et al. 2010; Zemach et al. 2010; Shao et al. 2014). However, there were several hypomethylated regions (<40% methylation) that are indicative of DNA methylation canyons or valleys (Xie et al. 2013; Jeong et al. 2014) with the most prominent of these located on chromosomes 4, 10, 11, 12, and 16 (Figure 4.1 and A4.4).  89    Figure 4.1: Mean CpG methylation level across chromosomes 4, 10, 11, 12, and 16. Each point represents the mean methylation level across all twelve individuals for a single 10 kb window. The solid line represents the smoothed spline fit to these data. Position along the x-axis represents the base position along the chromosome. The y-axis is the average DNA methylation level.  0 500 1000 1500 2000 2500 3000 3500020406080100 chr 40 500 1000 1500020406080100 chr 100 500 1000 1500020406080100 chr 11Percent methylation0 500 1000 1500 2000020406080100 chr 120 500 1000 1500 2000020406080100 chr 16Chromosome position (10kb)90  While the factors that determine the size of DNA methylation canyons remains unknown, larger hypomethylated canyons such as those described here in stickleback have been shown to be under strong transcriptional suppression due to the increased abundance of repressive histone H3 lysine 27 methylation that interacts with hypomethylated DNA (Nakamura et al., 2014). This mechanisms of transcriptional repression is thought to maintain these regions in a “poised” transcriptional state to allow rapid activation of gene transcription in these regions at specific times during embryonic development (Nakamura et al. 2014), but whether these hypomethylated canyons play a functional role in adults is unknown. However, genes that are essential for proper development typically dominate these regions (Xie et al. 2013; Jeong et al. 2014; Nakamura et al. 2014). Consistent with this pattern, genes located in the hypomethylated canyons in the stickleback genome include protocadherins on chromosome 4 and homeobox genes on chromosomes 10, 11, 12 and 16 (Appendix C, Table C.1), suggesting a conserved role of hypomethylated canyons across vertebrates.  4.3.2 Sex-biased DNA methylation patterns  We identified a total of 18,564 DMCs between males and females (Figure 4.2). Although relatively few studies have examined differential methylation patterns between males and females at the whole genome level in fishes, a study in tilapia detected a similar number of DMCs between males and females in muscle tissue (17,112 DMCs; Wan et al. 2016), whereas a study of sex-specific differential methylation in zebrafish brain detected only 914 DMCs (Chatterjee et al. 2016). These data suggest that the extent of sexually dimorphic methylation may be highly variable among teleosts, consistent with the wide range of sex-determining mechanisms in this group (Devlin & Nagahama 2002). 91  The distribution of DMCs across genomic features (e.g. promoters, exons, CpG islands) did not differ from the relative proportions of these features in the genome (Appendix C, Table C.2). No DMCs were identified between males and females in the mitochondrial genome. The majority of DMCs (90 %; 16,626 DMCs) between males and females showed a bias towards higher methylation in females relative to males suggesting that female stickleback genome is hypermethylated relative to male stickleback genome.    Figure 4.2: Differentially methylated CpG loci between male and female stickleback.  (A) Differentially methylated CpG (DMC) loci between male and female threespine stickleback. Each point represents an individual DMC. The y-axis indicates the percent difference in methylation between males and females. A positive value on the y-axis indicates a DMC that is hypermethylated in females relative to males. A negative value on the y-axis indicates a DMC that is hypomethylated in females relative to males. The x-axis indicates the position of the DMC in the stickleback genome. Chromosome boundaries are represented by vertical dashed lines. Only DMCs for which a change in methylation of > 10 % are presented. (B) The percentage of CpG loci on a given chromosome that were differentially methylated between male and female threespine stickleback. The light shading represents DMCs that are hypermethylated and the dark shading represents DMCs that are hypomethylated in female stickleback compared to males.   The most striking pattern in these data is the apparent hypermethylation of chr19 (the threespine stickleback sex chromosome) in females relative to males, which is where 65 % of the putative DMCs identified between males and females are located (12,112 DMCs). Chr19 also had the highest proportion of DMCs relative to the number of CpG sites on the chromosome (5 −60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21Change in % methylationAchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 210123456B% CpG per chromosomehypomethylatedhypermethyated92  %) compared to the rest of the autosomes where ~0.07 % of CpG loci were differentially methylated between the sexes (Figure 4.2).   Three distinct regions (strata) have been characterized on chr19 based on the extent of divergence in these regions between the X and Y chromosome (Ross & Peichel 2008; White et al. 2015), and two of these strata no longer recombine between the X and Y: stratum two (the younger evolutionary stratum located between ~2.5 Mb and 12 Mb), and stratum one (the older evolutionary stratum located from ~12 Mb to the end of the chromosome). There is also a pseudoautosomal region (PAR) located in the first ~2.5 Mb of chr19 that is thought to still recombine between the X and Y chromosomes (White et al. 2015). To assess whether these evolutionary strata are also associated with sex-specific DNA methylation, we divided chr19 into 10 kb consecutive non-overlapping bins and calculated the frequency of CpG loci that were putatively identified as hypermethylated or hypomethylated in female stickleback relative to male stickleback (Figure 4.3). This analysis revealed clear patterns that correspond to the evolutionary strata of chr19 for loci that were hypermethylated in females relative to males, with the greatest apparent hypermethylation in stratum two and the least in the PAR (Figure 4.3B).  93   Figure 4.3: Differential methylation between sexes on chromosome 19 (chr19). (A) Mean DNA methylation levels for CpG loci along chromosome chr19. Each point represents the mean DNA methylation level in a 10 kb window for six individual stickleback that were either male (blue) or female (red). Solid lines represent the smooth spline fit for the DNA methylation levels in males (blue) and females (red). (B-C) Proportion of DMCs along chr 19 that are hypermethylated (B) or hypomethylated (C) in female stickleback compared to male stickleback. Values on the y-axis represent the total number of DMCs in a 10 kb window relative to the number of CpG loci in that same 10 kb window. (D) Ratio of mapped read counts for males relative to females along chr19. Each point represents the ratio of mean counts for a 10kb window in males compared to females. The solid black line is the smooth spline fit. Vertical dashed lines represent the boundaries between the three evolutionary strata on chr19: the pseudo autosomal region (PAR), stratum two, and stratum one. 0 500 1000 1500 2000chr 19 (10kb)Percent methylation020406080100 PAR Stratum Two Stratum OneA0 500 1000 1500 20000.000.050.100.150.20hypermethylated CpG/ total CpGchr 19 (10kb)PAR Stratum Two Stratum OneB0 500 1000 1500 20000.000.050.100.150.20hypomethylated CpG/ total CpGchr 19 (10kb)PAR Stratum Two Stratum OneC0 500 1000 1500 2000chr 19 (10kb)Male/Female read counts0.00.51.01.52.02.5 PAR Stratum Two Stratum OneD94  4.3.3 Identifying differential methylation on the sex chromosome  One of the challenges for unambiguously determining levels of sex-specific DNA methylation on stickleback sex chromosomes is that there is currently no publically available sequence for the Y chromosome and the published sequence for chr19 is predominately derived from X chromosome sequence. Because DNA methylation is detected as sequence differences between bisulfite-treated DNA and the reference sequence at CpG sites, both divergence between the X and Y chromosome at CpG sites and differential methylation can result in the same patterns in the sequence data. Thus, the signal of differential methylation that we observe could be attributed to one of three possible mechanisms: 1) identical sequences but differential methylation between the X and Y chromosome, 2) differential methylation between X chromosomes in males and females, or 3) sequence divergence between the X and Y chromosomes resulting in alteration of CpG sites. To address this issue, we again divided chr19 in to 10kb consecutive non-overlapping segments and compared the number of reads that mapped to chr19 in males and females for each 10kb segment. If chr19 reads map uniquely to the published X chromosome then we would expect half the number of reads to map to chr19 in males compared to females. Given that the PAR is known to recombine between the X and Y chromosomes, suggesting low levels of divergence in this region, we predicted that sequencing reads derived from both the X and Y chromosomes would map to the reference sequence, resulting in a ratio of one for the number of reads mapped in males and females in the PAR. The results from the read coverage analysis matched this prediction (Figure 4.3D).  For the younger, less diverged stratum (stratum two), we predicted a read count ratio between 0.5 and 1 because sequence similarity between the X and Y chromosome would result 95  in reads from both chromosomes mapping to the X chromosome reference sequence. The results from the read coverage analysis matched this prediction (Figure 4.3D). Thus, we cannot unambiguously determine whether differential methylation or X Y polymorphism is the cause of the apparent hypermethylation in females in stratum two. Stratum one is thought to be the most divergent region of the sex chromosome, and thus we predicted that few reads from the Y chromosome would map to the chr19 reference, resulting in a ratio of 0.5 for the number of reads mapped in males and females. Again, the results from the read coverage analysis mostly matched this prediction (Figure 4.3D), although there were specific regions where the read count ratio was close to one. These regions in stratum one may correspond to regions that are thought to be under purifying selection to maintain dosage sensitive genes from being lost on the Y chromosome (White et al. 2015). We also detected apparent hypermethylation in females in these regions (Figure 4.3B and 4.3D). Thus, we cannot conclusively determine whether this apparent hypermethylation of chr19 in females is due to differential methylation or from the accumulation of TG polymorphisms on the Y that are being interpreted as unmethylated loci.  While we are unable to unambiguously determine the ultimate cause of the apparent DNA methylation differences between males and females on chr19, whether the patterns we observe are the result of sequence polymorphism between the X and Y chromosome that alters CpG sites, or are due to differential methylation of conserved sequences between males and females in chr19, the ultimate effect would be differences in methylation between the sex chromosomes. Thus, taken together, the patterns of putative differential methylation that we observe suggest that divergence in DNA methylation patterns between males and females on the 96  stickleback sex chromosome are closely associated with the known evolutionary history of this chromosome.  4.3.4 DNA methylation and sex chromosome evolution Sex chromosome evolution from autosomes is thought to involve recombination suppression in sex determining regions, followed by the accumulation of deleterious mutations and the degeneration of the sex-specific (e.g. Y) chromosome (Graves 2016). Degradation of the sex-specific heteromorphic sex chromosome following recombination suppression has the potential to cause imbalances in gene expression. Many taxa with heteromorphic sex chromosome pairs have evolved dosage compensation mechanisms to resolve this effect, but the nature and extent of these dosage compensation mechanisms varies greatly among taxa (Graves 2016). DNA methylation has been proposed as a key mechanism responsible for regulating every step of the evolution of sex chromosomes from recombination suppression in the early stages of sex chromosome evolution to dosage compensation in more derived sex chromosome systems (Gorelick 2003); however, there has been little empirical evidence to test this hypothesis. Taxa with relatively “young” heteromorphic sex chromosomes, such as the threespine stickleback, provide an opportunity to investigate the potential role of DNA methylation in regulating sex chromosome recombination and the evolution of dosage compensation mechanisms. In the following section, we discuss the apparent differential methylation between males and females on chr19 in the context of the different stages of sex chromosome evolution in stickleback.  DNA methylation promotes the formation of heterochromatin (Melamed-Bessudo & Levy 2012; Mirouze et al. 2012; Yelina et al. 2015), and it is thought to play a role in suppressing recombination of sex chromosomes in plants (Zhang et al. 2008). In stickleback we 97  observed apparent hypermethylation of the younger evolutionary stratum in females (stratum two), and relatively less differential methylation between males and females along the older evolutionary stratum (stratum one) and the PAR. The apparent hypermethylation of stratum two on the X chromosome in females (hypomethylated in males) corresponds to the region hypothesized to have undergone the first chromosomal inversion during the evolution of the Y chromosome (Ross & Peichel 2008). The apparent differential methylation in stratum two between males and females (hypomethylated in males and hypermethylated in females) could have played a role in suppressing recombination between male and female sex chromosomes and in establishing the boundaries in which this inversion first occurred.  The next stage in sex chromosome evolution, following recombination suppression, is thought to be the accumulation of genetic variation and degradation in the non-recombining region(s). If methylated cytosines on the female X chromosome correspond to thymines on the male Y chromosome as previously discussed, then it is possible that the accelerated mutation rate of methylated cytosines, which can be deaminated to become thymines (Coulondre et al. 1978; Shen et al. 1994), could play an important role in the divergence between X and Y chromosomes. Alternatively, instead of being a direct result of C to T polymorphisms, the observed increase in hypermethylated loci in females could be closely linked to the accumulation of genetic polymorphisms. The frequency at which single nucleotide polymorphisms (SNPs) occur in genomes has been shown to be higher near methylated CpG loci (Qu et al. 2012). The CGCG motif has been identified as a candidate cis-element associated with this observation and is enriched in hypomethylated regions (Qu et al. 2012). Therefore, individuals with higher DNA methylation levels at particular loci would be predicted to have a higher degree of sequence divergence near those loci relative to individuals with lower methylation levels. The PAR had 98  the highest frequency of the CGCG motif (1.98 / 1kb) while stratum one and stratum two had a lower frequency of the CGCG motif (1.22 / 1kb and 1.18 / 1 kb respectively). These data suggest that DNA methylation may also be associated with the accumulation of genetic variation between the non-recombining regions of the male and female sex chromosomes.  We next explored whether the apparent differential methylation patterns between males and females on chr19 are consistent with the regulation of dosage sensitive genes. In stickleback, there are two conflicting hypotheses regarding the existence of a dosage compensation system. One hypothesis is that there is locally confined partial dosage compensation in stratum one in males that is also associated with a hypertranscription of genes in stratum one in females (Schultheiß et al. 2015). The second hypothesis suggests that dosage compensation has not evolved in the stickleback but that there is purifying selection to maintain dosage sensitive genes in stratum one of the Y chromosome (White et al. 2015). The differential methylation patterns between sexes along the X chromosome that we observe are not entirely consistent with either of these prevailing hypotheses. We observed apparent hypermethylation in stratum two in females. This might be expected to result in reduced transcription or partial silencing of genes in this region, which has not been observed in stickleback (Schultheiß et al. 2015; White et al. 2015). The less extensive and highly localized pattern of hypermethylation in stratum one that we observe is suggestive of gene-specific regulation, which is not consistent with a generalized hypertranscription of genes in stratum one in females (Schultheiß et al. 2015). Because the male to female coverage ratio is similar in these localized regions in stratum one, this localized pattern may be more consistent with the potential preservation of dosage sensitive genes in these regions (White et al. 2015); however, it is also possible that the differential methylation in these regions 99  is caused by the accumulation of C to T polymorphisms which is less consistent with purifying selection acting in these regions (White et al. 2015).  Taken together, the apparent differential methylation between male and female stickleback described in this study is consistent with the proposed role of DNA methylation in the evolution of sex chromosomes (Gorelick 2003). Thus differential DNA methylation could be playing a role in the evolution of this “young” heteromorphic sex chromosome system, either through influencing patterns of recombination or potentially through mediating the early stages of the development of dosage compensation.  4.3.5 DNA methylation and sex-biased gene expression  To determine whether the apparent variation in DNA methylation patterns between males and females could be influencing previously described sex-biased gene expression patterns in stickleback, we compared the list of genes near DMCs between males and females to previously identified genes that exhibit sex-biased gene expression patterns (Leder et al. 2010; Metzger & Schulte 2016a). Of the 2,282 genes that have been shown to exhibit sex biased gene expression patterns from these studies, 490 overlapped with genes near DMCs in our study of which 269 are on chr19 including genes located in the region considered to be tightly linked to sex determination in stickleback such as sema4ba and idh2 (Peichel et al. 2004). This pattern is consistent with differential methylation between males and females playing a role in regulating sex-biased patterns of gene expression.  100  4.3.6 Effects of environmental salinity on DNA methylation  Variation in DNA methylation has also been suggested to be an important component of an organism’s response to environmental change (Bossdorf et al. 2008; Flores et al. 2013; Hofmann 2017). Changes in environmental salinity are known to cause substantial changes in gene expression in many species of fish, including stickleback (Wang et al. 2014; Gibbons et al. 2017; Zhang et al. 2017). In order to explore whether changes in DNA methylation may be involved in environmental regulation of gene expression, we identified differentially methylated loci in stickleback reared at two salinities (2 and 21 ppt), and we compared genes near DMCs identified in fish reared at different salinities to genes that have been previously identified as salinity-responsive using RNA-seq (Wang et al. 2014; Gibbons et al. 2017).  1,259 CpG loci were differentially methylated between salinity treatments, and these DMCs were distributed across all chromosomes, with an average of 0.01 % of the CpG loci on each chromosome being differentially methylated (Figure 4.4). No DMCs were identified between individuals from low and high salinities in the mitochondrial genome. The distribution of DMCs across genomic features (e.g. promoters, exons, CpG islands) did not differ from the relative proportions of these features in the genome (Appendix C, Table C.2). The majority of DMCs (1,051) were hypomethylated in individuals from high salinity relative to low salinity. Analysis of the genes located close to these DMCs revealed several genes known to be involved in the response to salinity in fish. However, GO enrichment analysis did not detect significant enrichment for any GO categories following FDR correction. The ten GO terms with the lowest p-values are listed in tables A4.3-4.5.   101   Figure 4.4: Differentially methylated CpG loci between stickleback reared a low and high salinity. (A) Differentially methylated CpG (DMC) loci between stickleback reared at a salinity of 2 ppt compared to 21 ppt. Each point represents an individual DMC. The y-axis indicates the percent difference in methylation between salinity rearing treatments. A positive value on the y-axis indicates a DMC this is hypermethylated in stickleback reared at 21 ppt relative 2 ppt. A negative value on the y-axis indicates a DMC that is hypomethylated in stickleback reared at 21 ppt relative to 2 ppt. The x-axis indicates the position of the DMC in the stickleback genome. Chromosome boundaries are represented by vertical dashed lines. Only DMCs for which a change in methylation of > 10 % are presented. (B) The percentage of CpG loci on a given chromosome that were detected as being differentially methylated between stickleback reared at 21 ppt and 2 ppt. The light shading represents DMCs that are hypermethylated and dark shading represents DMCs that are hypomethylated in stickleback reared at 21 ppt compared to 2 ppt respectively.   Comparison of genes located near DMCs in fish from reared at different salinities to previously identified as salinity-responsive using RNA-seq (Wang et al. 2014; Gibbons et al. 2017) identified 126 candidate genes with changes in both expression and methylation in response to salinity. Among the candidate genes that we identified are ion channels that are important for regulating cellular ion concentrations in hyper and hypo-osmotic conditions such as the calcium pump atp2b4, the sodium/chloride co-transporter slc12a3, and the sodium/potassium/2 chloride co-transporter slc12a1. Taken together, these data suggest that changes in DNA methylation could play a role in facilitating the transition between marine and freshwater environments.  In stickleback, a variety of genomic regions have been identified as having been subject to positive selection following colonization of freshwater habitats by ancestral marine fish (Jones −60−40−200204060chr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 21Change in % methylationAchr 1chr 2chr 3chr 4chr 5chr 6chr 7chr 8chr 9chr 10chr 11chr 12chr 13chr 14chr 15chr 16chr 17chr 18chr 19chr 20chr 210.000.010.020.030.04B% CpG per chromosomehypomethylatedhypermethyated102  et al. 2012a). Because epigenetic variation has been suggested to be a driver of adaptive evolution (Flores et al. 2013), we screened our dataset of salinity responsive DMCs in marine fish to identify those associated with genes found in regions under positive selection in freshwater populations (Jones et al. 2012a). Very few of the DMCs identified in our study were near these genes, suggesting that salinity-responsive changes in DNA methylation are unlikely to have played a role in driving genetic divergence in these regions between marine and freshwater populations of stickleback.   4.4 Conclusions  In this study we used whole-genome bisulfite sequencing to identify novel DNA methylation features in the stickleback epigenome. Apparent hypermethylation of stratum two on the female X chromosome compared to levels in males suggests that DNA methylation could play an important role in the suppressing recombination between the X and Y chromosome, and potentially in regulating sex-biased gene expression patterns. We also detected significant changes in DNA methylation in response to rearing salinity, some of which were associated with genes known to be differentially regulated in response to changes in environmental salinity. This epigenetic change reflects a response to environmental salinity that could facilitate the accumulation of epigenetic variation between natural populations, and thus be implicated in long-term responses to environmental change. 103  Chapter 5: Maternal stress has divergent effects on gene expression patterns in the brains of male and female threespine stickleback   5.1 Introduction  Maternal exposure to stress has been shown to have persistent effects on the neurodevelopmental processes of offspring across a wide range of taxa, resulting in profound changes to the health and behavior of subsequent generations (Storm & Lima 2010; Bale 2015; Bell et al. 2016). For example, in humans, epidemiological studies suggest that prenatal exposure to maternal stress increases the prevalence of neurodevelopmental diseases such as schizophrenia and autism spectrum disorder in adult offspring (Babenko et al. 2015). In contrast, in a variety of ecological contexts the effects of maternal stress on offspring phenotypes have been suggested to be positive, conferring increased fitness in stressful environments (Sheriff & Love 2013). Whether maladaptive or adaptive, these effects are likely the result of a complex interaction between genes and the environment, mediated through a variety of epigenetic processes (Weaver et al. 2004; Murgatroyd et al. 2009).   In mammals, the effects of maternal stress on adult offspring have been repeatedly demonstrated to differ between males and females (Weinstock 2007; Jazin & Cahill 2010; McCarthy et al. 2012a). These sex-specific effects are thought to be regulated by complex maternal-fetal interactions via the placenta (Nugent & Bale 2015). Thus, it is unclear whether conclusions derived from studies in mammals can be broadly applied across non-placental taxa. Studies in birds have detected sexually dimorphic effects of maternal stress on offspring behavior (Holloway & Clayton 2001; Wade et al. 2004), and the regulation of the hypothalamic-104  pituitary axis (HPA) (Schmidt et al. 2014), but the generality of these findings across oviparous vertebrates remains unknown. This is particularly important in the context of ongoing debates regarding the evolutionary implications of maternal effects in natural populations (Räsänen & Kruuk 2007; Badyaev 2008; Sheriff & Love 2013).  In this study we use threespine stickleback (Gasterosteus aculeatus) as a model to study the effects of maternal stress on the brain transcriptome of male and female offspring. There is compelling evidence from studies in this species that maternally derived stressors have profound and persistent effects on developmental, morphological and behavioral phenotypes of offspring (Bell et al. 2016) that can persist across generations (Shama et al. 2014, 2016; Shama & Wegner 2014; Shama 2015) making them an important model in which to study the adaptive potential of transgenerational plasticity, and an ideal candidate in which to explore the sexually dimorphic effects of maternal stress on the offspring of an oviparous vertebrate.   5.2 Materials and methods 5.2.1 Stress treatment and animal rearing  Adult threespine stickleback (G. aculeatus) of the fully plated “marine” ecotype were collected in May 2013 in Oyster lagoon (British Columbia, Canada, GPS: 49.6121,-124.0314) and acclimated to laboratory conditions for three weeks (see appendix C for details). Following the three-week acclimation period, fish were divided into two treatment groups (Unstressed and Stressed) with three replicate tanks per treatment. Fish in the stressed treatment were chased once daily with a fish net for 30 s before being captured and held out of water for a further 30 sec and then returned to their tank. This was intended to mimic similar predatory and handling stressors that have been used in investigate the effects of maternal stress on offspring behavior in 105  fish (Giesing et al. 2011; Sopinka et al. 2017). Timing of the stress treatment was randomized during each day to avoid conditioning. The stress treatment was applied for two weeks, and significantly increased plasma cortisol levels in mothers (ANOVA p = 0.0014; Appendix D, Figure D.1). Testes were then harvested from unstressed males and used to generate half-sib crosses, with the same male being used to fertilize the eggs of one stressed female and one unstressed female. A total of 6 pairs of half-sib crosses were produced.   After one year post hatch, and prior to reaching sexual maturity, brain tissue was dissected from the offspring and immediately frozen in liquid nitrogen and stored at -80°C until further use. Sex of the 1-year-old offspring was determined by PCR amplification of Idh, Gasm6, and Stn190 as previously described (Toli et al. 2016). Table S1 summarizes the family information for the fish that were selected for RNA-seq. Because not all families had the required number of male and female offspring at one year post-hatch, we were unable to use a completely balanced design with male and female half sibs represented in both the unstressed and stressed treatments. However, we attempted to maximize the genetic diversity within the sample by limiting the use of full-sibs where possible (Appendix D, Table D.1).  5.2.2 RNA isolation and sequencing  Total RNA was isolated from stickleback whole brain tissue (3 fish per sex per stress treatment; 12 fish total) using TRIzol Reagent (Invitrogen Life Technologies) as previously described (Metzger et al. 2016) followed by DNase treatment (Qiagen Rneasy). RNA integrity numbers (RIN) were between 7.9 – 9.1 (mean = 8.6 ± 0.4 SD). Preparation of Illumina TruSeq cDNA libraries and 100 base-pair paired end sequencing was performed at the UBC Nucleic Acid Protein Service Unit (NAPS) and UBC Biodiversity Research Center’s next generation 106  sequencing facility, and at the McGill University and Genome Quebec Innovation Center. Sequencing depth and sample multiplexing were devised to optimize statistical power given the total sequence generated (Ching et al. 2014; Wu et al. 2015). Mean library size was 19.8 million reads (Appendix D, Table D.2).   5.2.3 Sequence alignment and expression analysis  Reads were aligned to the stickleback genome (http://www.ensembl.org) using CLC genomics workbench v8.5 (Qiagen). Average mapping efficiency of paired and broken reads was 88.5 ± 0.9%. Analysis of total read counts was performed in R v3.2.2 with edgeR v3.12.0. Recommended RNA-seq analysis guidelines were followed (Lin et al. 2016) with the addition of RUVr from the RUVSeq protocol v1.4.0 (Risso et al. 2014) to account for batch effects. Genes with no reads were removed from the dataset. Counts were then normalized using the relative log expression (RLE) method (Anders et al. 2010). Following normalization, genes with low expression were filtered from the dataset. The minimum criterion for retaining a gene was at least 1 count per million (~10 counts in the smallest library) in each of the three samples of at least one sex by stress treatment group. A total of 16,477 genes were retained after normalization and filtering for subsequent expression analysis. Robust dispersions were calculated using the robust method in edgeR (Zhou et al. 2014). The data were then fit to a negative binomial generalized linear model using glmFit(). Likelihood ratio tests were run to assess effects of sex, maternal stress and the interaction of these main effects with factors calculated from RUVseq included in the model. The resulting p-values were adjusted based on FDR correction (Benjamini & Hochberg 1995), and the threshold for significance of these adjusted p-values was set at 0.05. Principal component analysis (PCA) was performed using the prcomp() function from the base 107  package in R on log 2 counts per million expression values. GO and KEGG pathway enrichment analyses were conducted using the goseq (v1.22.0) R package (Young et al. 2010), with FDR correction (see Appendix D for details). Cluster analysis of differentially expressed gene ontologies was performed using REVIGO (http://revigo.irb.hr), and the resulting source code was used to generate figures using the ggplot package in R.   5.3 Results  Analysis of RNA-seq data from the brains of one-year-old male and female threespine stickleback offspring from stressed and unstressed mothers revealed over 2,900 genes with either a significant main effect of sex (1,255 genes) or a significant interaction between maternal stress and sex (1,650 genes), but very few genes with a main effect of maternal stress (Figure 5.1A). Principal component (PC) analysis of all expressed genes distinguished four groups (Figure 5.1B). PC1 explained ~30% of the variation in the data and separated males and females that came from stressed mothers indicating that maternal stress has different effects on the brain transcriptome in male and female stickleback. PC2 explained ~11.5% of the variation in the data and separated males and females that came from unstressed mothers. Combined, PC1 and PC2 accounted for ~41% of the variation in the RNA-seq data.  108   Figure 5.1: Differential expression and principal component analyses reveal sexually dimorphic effects of maternal stress in threespine stickleback offspring. (A) Venn diagram depicting the total number of differentially expressed genes for which a main effect of sex (purple), maternal stress (orange), or an interaction between sex and maternal stress (green) was detected. (B) Principal components 1 and 2 of expressed genes in the brain tissue of males (squares) and females (circles) from unstressed (open symbol) and stressed (closed symbol) mothers.  5.3.1 Sexual dimorphism in gene expression  Independent of the effects of maternal stress, there were differences in brain gene expression patterns between male and female threespine stickleback (Appendix D, Figure D.2). 614 genes were expressed at higher levels in females whereas 641 genes were expressed at higher levels in males (Appendix D, Figure D.2). Enrichment analysis (Gene-Ontology (GO)-enrichment, KEGG pathway enrichment) did not reveal any terms or pathways that were significantly enriched among the genes with a significant main effect of sex. However, GO enrichment analyses do not provide comprehensive categorization of the GO terms associated with a differentially expressed gene list (Primmer et al. 2013). Genes associated with metabolic processes were expressed at higher levels in females whereas genes associated with cytoskeletal organization, cell adhesion and developmental processes including nervous system development were expressed at higher levels in males (Appendix D, Figure D.3).   ●●●●●●−1.5 −1.0 −0.5 0.0 0.5 1.0 1.5−2−1012PC1 (30%)PC2 (11.5%)AMaternalStressSexInteraction33 112551015634PC1 (30%)-1.5 -1.0 -0.5 0.0 0.5 1.0 1.51.02.00.-1.0-2.0PC2 (11.5%)BFig 1.109  5.3.2 Sex-specific effects of maternal stress  The effects of maternal stress on brain gene expression were strikingly different in male and female offspring (Figure 5.2) and these effects were generally in opposing directions between the sexes such that genes that increased in expression in response to maternal stress in female offspring decreased in expression or did not change in response to maternal stress in male offspring and vice versa. This is in direct contrast to the small number of genes that showed a conserved effect of maternal stress in both males and females, with only three genes demonstrating a maternal stress effect without an effect of offspring sex or an interaction (Figure 5.1A).   Figure 5.2: Heat map of all differentially expressed (DE) genes for which a significant interaction between sex and maternal stress was identified. Expression results are displayed a log2 counts per million normalized to the mean expression for all samples. Blue indicates lower expression and yellow indicates higher expression.  Unstressed Mothers Stressed MothersMale MaleFemale FemaleUp-regulated by maternal stress in:FemalesMalesFig 2.10.50-0.5-1110   GO-enrichment analysis of genes with a significant interaction between maternal stress and offspring sex revealed several enriched GO categories that suggest differential expression of genes involved in protein translation, metabolism, and synapse organization and assembly (Appendix A, Table A5.6). Enrichment analysis of KEGG pathways identified those involved in protein synthesis, and several neurodegenerative diseases such as Huntington’s, Parkinson’s, and Alzheimer’s disease, as well as pathways associated with synaptic function such as nicotine addiction and glutamatergic synapse (Table S7).   To further explore functional enrichment in the genes exhibiting significant interactions, we also performed separate GO enrichment analysis genes that were up-regulated by maternal stress in female offspring and genes that were up-regulated by maternal stress in male offspring. This analysis divided the previously identified enriched GO categories into two distinct groups (Figure 5.3). The majority of enriched GO categories for genes that were up-regulated by maternal stress in female offspring were involved in protein translation and regulation of metabolic processes, and this was reflected in both the biological process (Figure 5.3A) and cellular component (Figure 5.3B) ontologies. Enriched KEGG pathways up-regulated by maternal stress in female offspring were those associated with protein translation and diseases such as Parkinson’s, Huntington’s, and Alzheimer’s disease, all of which have a metabolic component (Beal 1998; Schapira 1999; Appendix D, Table D.5 ). A different set of processes was identified as enriched for genes that were up-regulated by maternal stress in male offspring. The majority of enriched GO categories for genes up-regulated by maternal stress in males were involved in nervous system development and synapse formation, and again this was reflected in both the biological process (Figure 5.3A) and cellular component (Figure 5.3B) ontologies. Enriched KEGG pathways up-regulated by maternal stress in males were those associated with 111  synaptic function and organization including glutamatergic synapse, GABAergic synapse, long-term depression, and nicotine addiction (Appendix D, Table D.5).   Figure 5.3: Cluster analysis of the significantly enriched gene ontology (GO) terms for biological process (A) and cellular component (B) over-represented among genes that increased in males (blue circles) or in females (red circles) from stressed mothers compared to unstressed mothers. Size of the circle is representative of the total number of DE genes in that GO category. Note that some names have been omitted from the figure due to semantic similarity.   ABFig 3.Synapse organizationSynapse assemblyNervous systemdevelopmentSteroid hormone mediatedsignaling pathwaySynaptic transmissionNeuron cell-cell adhesionRespiratory electron transport chainCellular metabolic processCellular nitrogen compoundmetabolic processCellular protein metabolic processTranslationPosttranscriptional gene silencing by RNAExtracellular vesicular exosomeSynaptic vesicleMitochondrionMitochondrial inner membraneNucleusRibosomeAxonDendriteSynapseCell junctionPlasma membranePostsynaptic membraneSemantic space x0-5 5Semantic space x0-5 50-44Semantic space y0-55Semantic space y112  5.3.3 Chromosomal location of differentially expressed genes  Of the 1,255 genes that exhibited a main effect of sex, 1170 could be localized to chromosomes, and of these, 585 were expressed at higher levels in females, and 585 were expressed at higher levels in males. A disproportionate number of these genes were located on the sex chromosome (chromosome 19; 441 genes, 38%), and 87% of these sex-chromosome localized genes (383 genes) were expressed at higher levels in females than in males (Figure 5.4A). In contrast, genes that were expressed at higher levels in males than in females were distributed fairly uniformly across the chromosomes. Because similar numbers of genes were expressed at higher levels in females and in males, genes with autosomal locations tended to be more highly expressed in males.  A strikingly different pattern of chromosomal location was observed for genes that differed in their response to maternal stress between male and female offspring (Figure 4B). In this case, 1,477 genes of the 1,650 genes with a significant interaction between offspring sex and maternal could be localized to chromosomes. Of these, 504 genes were expressed at higher levels in female offspring from stressed mothers, whereas 973 genes were expressed at higher levels in male offspring from stressed mothers. These genes were uniformly distributed across the 21 stickleback chromosomes in both sexes (Figure 4B), with each chromosome harboring 4.8 ± 1% of the genes that demonstrated a significant interaction between maternal stress and offspring sex.   113   Figure 5.4: Histograms representing the total number of differentially expressed genes localized to each of the 21 stickleback chromosomes and the mitochondrial genome. (A) Genes for which a significant effect of offspring sex was detected. (B) Genes for which significant interaction between offspring sex and maternal stress was detected. Blue bars represent genes that are up-regulated in males (A) or are expressed at higher levels in males from stressed mothers compared to females from stressed mothers (B). Red bars represent genes that are up-regulated females (A) or are expressed at higher levels in females from stressed mothers compared to males from stressed mothers (B).  5.4 Discussion  The results presented here clearly demonstrate two distinct classes of sexual dimorphism in brain gene expression in threespine stickleback: 1) differences between males and females that are present under all conditions, and 2) differences in the response of the male and female brain to maternal stress. Sexual dimorphism in brain gene expression is known for a wide variety of taxa (Jazin & Cahill 2010), and has been previously observed in several fish species including threespine stickleback (Santos et al. 2008; Sreenivasan et al. 2008; Manousaki et al. 2014; Sharma et al. 2014; Schultheiß et al. 2015; Lu et al. 2015). In contrast, the critical importance of the interaction between maternal stress and offspring sex has been pointed out in mammals (McCarthy et al. 2012a), but has been poorly studied in other taxa and is often neglected in ecological and evolutionary studies. Our finding that maternal stress affects the brains of male and female stickleback offspring differently emphasizes the importance of taking sex into account when studying the effects of maternal stress on gene expression, neurodevelopment and 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 MT050100150200250300350400450ChromosomeNumber of Genes1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 MT050100150200250300350400450ChromosomeNumber of GenesA0 1 5 10 15 20 TChromosome Number1 5 10 15 20 TChromosome Number4504003503002502001501050Number of DE Genes 4504035030250201501050Number of DE GenesBFig 4.114  behavior and has important implications for the assessment of the ecological and evolutionary impacts of stress across generations.   5.4.1 Sex-specific effects of maternal stress  In this study we detected 1,650 genes that demonstrated a significant interaction between maternal stress and offspring sex, while only 6 genes demonstrated a significant effect of maternal stress with no interaction. This pattern is surprising because the power to detect differential expression is greater for the main effect of maternal stress (n=6 individuals per group) than it is for the interactive effects of maternal stress and offspring sex (n=3 individuals per group). Simulation studies suggest that six individuals per group should be sufficient to detect differential expression given the sequencing depth per sample used here (Wu et al. 2015), and thus we should be able to detect a general effect of maternal stress, if present. The lack of significant main effects of maternal stress is likely driven by the observation that genes up-regulated by maternal stress in male offspring were generally down-regulated in female offspring, and vice versa.   We had relatively low power to detect differences between males and females in the effects of maternal stress (i.e. the interaction between offspring sex and maternal stress), suggesting that the 1,650 genes for which we detected significant interactions may represent an underestimate of the sex-specific effects of maternal stress. Another key question is the extent of false positives within this dataset, as rates of false positives increase at low samples sizes (Wu et al. 2015). Two possible approaches to reduce this “false discovery cost” are to apply a very stringent FDR threshold, or to apply an effect-size filter and retain only genes above a certain fold-change cutoff as differentially expressed, which greatly reduces the probability of detecting 115  false positives. We applied each of these approaches (Appendix D, Figure D.4) and these analyses support the conclusion of sex-specific effects of maternal stress on brain gene expression patterns.   We also detected clear divergence in the pathways identified as enriched for genes up-regulated by maternal stress in male versus female offspring of stressed mothers. This pattern suggests that maternal stress has physiologically distinct effects on the brains of male and female offspring of stressed mothers. Enrichment analysis indicated that maternal stress up-regulates genes involved in mitochondrial respiration in females, while up-regulating genes involved in synaptic function in males, and these patterns were supported when very stringent effect size and FDR filters were applied to the data (see Appendix D for details). Although differences in gene expression do not necessarily result in changes in protein activity, the striking differences in expression patterns between the sexes and the sexual dimorphism in the biological processes identified and the enrichment of different biological processes between the sexes in response to maternal stress strongly suggests important differences in the physiological impacts of maternal stress between male and female offspring in threespine stickleback. In combination with similar patterns that have been observed in mammals (Weinstock 2007), these data suggest that sexually-dimorphic effects of maternal stress may be common across a wide range of taxa.  5.4.2 Sources of sexual dimorphism in brain gene expression  We also detected significant sexual dimorphism in brain gene expression independent of maternal stress. Although the specific functional consequences of the sexually dimorphic patterns of brain gene expression that we and others (Schultheiß et al. 2015; White et al. 2015) observe in threespine stickleback remain unknown, they may be either causes or consequences of 116  the many morphological and behavioral differences between male and female threespine stickleback (Bell & Foster 1994; Shaw et al. 2007; Kotrschal et al. 2012). For example, male stickleback perform several complex behaviors that females do not, such as nest building, courtship behaviors, and paternal care of offspring (Bell & Foster 1994). Associated with this sexual dimorphism in behavior, there is striking sexual dimorphism in brain size, with males having larger brains compared to females (Kotrschal et al. 2012; Samuk et al. 2014; Herczeg et al. 2015), and similar patterns have been observed in other fish species (Kolm et al. 2009; Herczeg et al. 2014). It has been proposed that the larger brain size of male stickleback may be due to the increased cognitive demand on males compared to females (Jacobs 1996; Kotrschal et al. 2012; Herczeg et al. 2015). In addition to differences in brain size and behavior between males and females, another possible cause of differences in brain gene expression between the sexes could be differences in the relative sizes of different brain regions. Such a pattern has been observed in a variety of species (Cooke et al. 1998), including stickleback (Park & Bell 2010).   In addition to differing in size and morphology between the sexes, the vertebrate brain also has substantial capacity for plasticity in many species (Kolb & Whishaw 1998; Kaslin et al. 2008), including stickleback (Park et al. 2012). Interestingly, male stickleback have been shown to have increased brain plasticity relative to female stickleback (Herczeg et al. 2015). The effects of maternal stress on brain gene expression represent a form of transgenerational plasticity. Here we observed increased expression of genes associated with neurodevelopment processes in males from stressed mothers but not in females from stressed mothers. This observation raises the intriguing possibility that the differential effects of maternal stress on the male and female brain could be due to underlying differences in brain plasticity between the sexes. Indeed, previous studies have detected a strong maternal effect on brain size and the relative sizes of brain regions 117  in stickleback (Noreikiene et al. 2015), and taken together with the known sexual dimorphism in brain plasticity in stickleback (Gonda et al. 2009; Park et al. 2012; Herczeg et al. 2015), could account for the differences in the impacts of maternal stress that we observe in brain gene expression between male and female offspring.  Another potential source of sexual dimorphism in gene expression is the biased genomic localization of these genes on divergent sex chromosomes (Rice 1984; Rinn & Snyder 2005; Ellegren & Parsch 2007; Leder et al. 2010; Chatterjee et al. 2016). Consistent with previous studies in threespine stickleback (Schultheiß et al. 2015) we observed that genes that were more highly expressed in females, independent of maternal stress, were heavily biased towards localization on the sex chromosome (Figure 5.4). In contrast, the genes that exhibited divergent expression patterns in the response to maternal stress between male and female offspring (i.e. those that showed a significant interaction between maternal stress and offspring sex) did not exhibit biases in genomic localization. This pattern strongly suggests that mechanisms acting in trans, such as regulation by sexually dimorphic genes localized on the sex chromosome, or via sexually dimorphic hormones, are likely to be responsible for the different responses to maternal stress observed in male and female brains.   5.4.3 Effects of maternal stress on stress-related genes  One of the clear conclusions emerging from previous studies in mammals is that stress experienced during early development, such as during prenatal development or during childhood, can induce changes in the hypothalamic-pituitary-adrenal stress response (Liu 1997; Francis et al. 1999), which can lead to the sensitization of adults to stress-related disorders (Oberlander et al. 2008). Similar effects have been observed in a variety of fish species, including threespine 118  stickleback (Mommer & Bell 2013). To explore the effects of maternal stress on the expression of genes associated with the glucocorticoid stress response, we examined the expression patterns of a suite of candidate genes known to be involved in this pathway.  Both of the isoforms of nr3c1, which encode glucocorticoid receptors, were up-regulated by maternal stress in male offspring and were down-regulated by maternal stress in female offspring. In addition, there was evidence for sexual dimorphism and sex differences in the response to maternal stress in the expression of several of the key regulators of the glucocorticoid receptor (Appendix D, Table D.4). Brain glucocorticoid receptors are important for the negative feedback regulation of the stress hormone axis (Liu 1997). If the changes in mRNA levels that we detect are predictive of protein levels, this implies that female stickleback from stressed mothers may show reduced ability to down regulate the stress-hormone axis compared to male offspring, although current evidence suggests that these effects are likely to be subtle, if present (Mommer & Bell 2013). However, taken together, these data suggest that maternal stress re-shapes the stress hormone axis, and that these effects differ between male and female stickleback.  5.4.4 Effects of maternal stress on epigenetic processes  Epigenetic mechanisms such as DNA methylation are known to be involved in modulating the effects of maternal stress on offspring phenotypes (Bale 2015) including changes in the expression of the glucocorticoid receptor (Palma-Gudiel et al. 2015). We identified multiple sex-specific effects of maternal stress on the expression of DNA-demethylases (e.g. TET) suggesting a potential role for epigenetic mechanisms influencing the dimorphic expression of genes in males and females from stressed mothers. Similarly, GO enrichment 119  analysis identified post-translational gene silencing by RNA (GO: 0035194, Figure 5.3A) as a process that is over-represented in the set of genes that are up-regulated by maternal stress in male, but not female, offspring. The overall up-regulation of genes involved in a variety of epigenetic processes by maternal stress in male offspring and not in female offspring suggests a potential difference in the mechanisms via which maternal stress is transduced between the sexes.  5.5 Conclusions  Our data demonstrate that the effects of maternal stress on brain gene expression are sex-specific in threespine stickleback. Maternal stress has the potential to result in adaptive alterations in offspring physiology and behavior, enabling mothers to prepare their offspring for the environment they will encounter as adults. Alternatively, this phenomenon can also be maladaptive when temporary stressors induce permanent developmental changes that are suboptimal for the environment or when they lead to the development of disease (Hales & Barker 2001; Ghalambor et al. 2007; Sheriff & Love 2013). Determining whether a specific offspring phenotype is the result of adaptive or maladaptive plasticity is a significant challenge in ecology and evolutionary biology. Our observation of sex-specific effects of maternal stress in stickleback thus has important implications for the analysis of the evolutionary significance of maternal effects on transgenerational plasticity.    120  Chapter 6: Conclusion  6.1 Overview   The three main objectives of this thesis were 1) to determine how DNA methylation and gene expression are regulated in response to changes in environmental conditions, 2) to investigate the persistent effects of developmental conditions on DNA methylation and gene expression levels, and 3) to assess the effects of sex on DNA methylation patterns. Addressing the first objective, I demonstrated that DNA methylation patterns in stickleback are modified in response to both temperature (chapter two) and salinity (chapter four). Addressing the second objective, I demonstrated that temperature (chapters two and three) salinity (chapter four) and maternal factors (chapter five) influence DNA methylation and gene expression patterns across a range of timescales within an organism’s lifetime. Through this work I also found that maternal stress has divergent effects on gene expression patterns in the brain tissue between male and female offspring. Addressing the third objective, I identified genome wide differences in DNA methylation patterns between male and female stickleback, including apparent differential methylation on the stickleback sex chromosome that is associated with the known evolutionary history of this chromosome.  By addressing these objectives, this thesis provides insight into the inheritance of DNA methylation and gene expression patterns through mitotic cell division, provides evidence for a role of epigenetic mechanisms in modulating phenotypic responses to environmental stressors, and provides empirical evidence for a role of DNA methylation in the evolution of heteromorphic sex chromosomes. In this chapter I discuss how the studies described in this thesis have contributed to our understanding of the ecological and evolutionary role of DNA 121  methylation and discuss several potential future directions that have been inspired by these results.  6.2 Ecological epigenetics and threespine stickleback  One of the key questions in ecological epigenetic studies is to understand how the environment affects epigenetic variation. Several studies have demonstrated that variation in DNA methylation patterns can accumulate among individuals that inhabit different environments (Cervera et al. 2002; Riddle & Richards 2002; Keyte et al. 2006; Vaughn et al. 2007; Zhang et al. 2011). For example, a study of Kenyan house sparrows observed a negative correlation between DNA methylation diversity and genetic diversity but a positive correlation between epigenetic diversity and inbreeding among individuals from a rapidly expanding population (Liebl et al. 2013). Several studies have also demonstrated positive correlations between phenotypic diversity and variation in DNA methylation patterns (Lira-Medeiros et al. 2010; Lyko et al. 2010; Liebl et al. 2013) which could play a role the spread of invasive species or the colonization of new habitats (Liebl et al. 2013). However, our understanding of how environmental cues are involved in establishing or modulating this observed epigenetic variation remains largely unknown.   In this thesis I demonstrate that different environmental temperatures (chapter two) and salinities (chapter four) can have both persistent and plastic effects on DNA methylation patterns, suggesting that DNA methylation patterns are sensitive to environmental conditions and epigenetic variation between populations could be highly dependent on environmental differences. Thus, observed differences in DNA methylation patterns between natural populations may not necessarily be indicative of fixed population differences in DNA 122  methylation patterns. Therefore, differences in the environment conditions of populations should be considered when interpreting apparent differences in DNA methylation patterns between natural populations.   Many authors have suggested that in order to be evolutionarily relevant, epigenetic variation must be able to persist across multiple generations unaffected by subsequent environmental changes (Bossdorf et al. 2008; Jablonka & Raz 2009). Based on this idea of evolutionarily relevant epigenetic variation, one potential experiment that could be conducted in order to determine whether the effects of environmental temperature or salinity on DNA methylation patterns in stickleback are adaptive responses to environmental cues would be to compare DNA methylation patterns among wild populations that span a thermal or salinity gradient but that have been reared in controlled laboratory conditions for multiple generations. Differences in DNA methylation patterns that persist following acclimation to a common environment that are consistent the effects described in either chapter two or four would then be considered candidates for adaptive evolutionary changes.   However, this experimental design seems somewhat contradictory in that epigenetic variation must be both environmentally sensitive in one generation and fixed (i.e. environmentally independent) in the next. In this scenario, epigenetic mechanisms can act as a type of heritable “memory” about past environmental conditions. However, it is not clear why there should be a requirement for these epigenetic changes to stop being plastic in subsequent generations. It is possible for epigenetic states to persist in a population because the environmental condition that originally induced the epigenetic change has also persisted, or for epigenetic mechanisms to influence developmental trajectories but to still remain plastic in response to environmental cues in later stages of development. For example, the difference in 123  DNA methylation between stickleback from marine and freshwater populations when kept at their native salinity (Artemov et al. 2017) might be more representative of DNA methylation patterns in stickleback reared at two salinities and not evolved difference in DNA methylation patterns between populations. Differences in DNA methylation patterns between marine and freshwater populations could also be due to environmental effects on epigenetic drift. Epigenetic drift is an age-dependent change in DNA methylation patterns over time due to the accumulation of errors in maintaining DNA methylation states through successive rounds of cell division (Fraga et al. 2005). Epigenetic drift has been shown to be affected by both genetic and environmental factors (Shah et al. 2014), thus epigenetic variation between populations may not be due to the active regulation of DNA methylation patterns but on the passive accumulation of DNA methylation variation over the lifetime of the individual. Environmentally dependent and persistent DNA methylation patterns could still be considered adaptive, or maladaptive, and could influence evolutionary trajectories by manipulation the expression of different phenotypes in different environments. Therefore, instead of searching for fixed epigenetic differences between populations that are no longer responsive to environmental cues, selection might act on the plasticity of epigenetic processes. Thus, a more productive approach to understanding the adaptive, or maladaptive consequences of epigenetic variation would be for future studies to investigate differences in the plasticity of the reaction norms of epigenetic processes (e.g. DNA methylation levels) in response to environmental cues.  6.3 Sex specific effects on DNA methylation and gene expression  Another important result that emerged from the data in this thesis is that sex has a large effect on responses to environmental stress and on DNA methylation patterns in the stickleback 124  genome. These effects are clearly demonstrated in the whole genome DNA methylation patterns described in chapter four and in the effects of maternal stress on gene expression patterns in the brain tissue of offspring in chapter five.   6.3.1 Differential methylation on the stickleback sex chromosome  The major finding from chapter four is that the majority of DMCs identified between males and females were located on chr19, which is the threespine stickleback sex chromosome. Furthermore, the differential methylation pattern patterns on chr19 were closely associated with the known evolutionary strata of the stickleback sex chromosome.   These results appear to be consistent with the theory proposed by Gorelick (2003) who suggested a role for DNA methylation in accelerating the divergence of heteromorphic sex chromosomes by accelerating the effects of Muller’s ratchet on chromosome degradation. Muller’s ratchet is a process that describes chromosomal degradation resulting from suppressed recombination between chromosome pairs followed by the irreversible accumulation of deleterious mutations (Muller 1964; Felsenstein 1974). High levels of DNA methylation are associated with the formation of heterochromatin, and heterochromatin suppresses chromosome recombination. Thus, differences in the DNA methylation levels of chr19 between males and females could have been responsible for suppressing recombination. Methylated cytosines are hypermutable and deaminate to become thymines at a faster rate than unmethylated cytosines (Duncan & Miller 1980; Bulmer 1986; Britten et al. 1988; Sved & Bird 1990). Therefore, differences in DNA methylation levels between males and females could also result in differences in the rate at which genetic polymorphisms accumulated on one sex chromosome but not the other. Data from chapter 4 strongly implicate variation in DNA methylation patterns 125  between males and females in the evolution of heteromorphic sex chromosomes, and future analysis of DNA methylation levels on the Y chromosome will provide further insight into the role of DNA methylation in sex chromosome evolution.  6.3.2 Divergent effects of maternal stress in male and female stickleback  Maternal exposure to stress has been shown to have persistent effects on the neurodevelopmental processes of their offspring across a wide range of taxa resulting in profound changes to the health and behavior of subsequent generations (Storm & Lima 2010; Bale 2015; Bell et al. 2016). These effects are thought to be the result of a complex interaction between genes and the environment, mediated through a range of epigenetic processes (Darnaudéry & Maccari 2008; Bale 2011; Mychasiuk et al. 2011; Zucchi et al. 2013). Unlike mammalian systems where complex maternal-fetal interactions via the placenta are thought to regulate the sexually dimorphic effects of maternal stress (Nugent & Bale 2015; Bronson & Bale 2016), oviparous embryos develop outside the body, separated from the further influence of maternal conditioning. Despite differences in the developmental environment of mammalian and teleost embryos, several studies have observed effects of maternal stress on offspring behavior and the regulation of the hypothalamic-pituitary-interrenal (HPI) axis in teleost fishes (Mommer & Bell 2013; Bell et al. 2016; Sopinka et al. 2017). The results from chapter five add to this growing body of research by demonstrating that the persistent effects of maternal stress are different depending on whether the offspring is male or female. For example, the divergent effects of maternal stress on gene expression patterns between males and females may result in differences in neural development and metabolic processes. The importance of considering sexually dimorphic effects of maternal stress on neurodevelopmental processes has been 126  repeatedly highlighted and discussed in the mammalian literature (McCarthy et al. 2012a; Cahill & Aswad 2015). The apparent similarity in the sexually dimorphic effects of maternal stress in placental and oviparous organisms suggests that there may be more similarities in the mechanisms that regulate maternal effects among placental and oviparous organisms than previously thought.   One of the unresolved questions in studies addressing maternal stress effects on offspring phenotypes is what mechanism or mechanisms are responsible for transferring information about the maternal environment to the offspring. In this thesis I measured circulating levels of cortisol, a glucocorticoid hormone involved in the HPI stress response, as a proxy for determining whether the stress treatment applied in this study was adequate to elicit a stress response. Previous studies in stickleback have shown that similar stress protocols resulting in elevated levels of circulating cortisol can also result in elevated cortisol levels in the eggs (Giesing et al. 2011). While glucocorticoids, such as cortisol, are known to regulate many developmental processes, it is unclear what effects maternally contributed cortisol has on developmental processes because maternal cortisol is rapidly metabolized in the embryo shortly after fertilization (Paitz et al. 2016). However, maternal stress can also affect other biological processes. For example, maternal stress has been shown to affect the nutritional content of eggs from stressed song sparrows and these changes in egg nutrient content have been shown to influence developmental processes of their offspring (Schmidt et al. 2012, 2014, 2015). Epigenetic processes have also been hypothesized to mediate the effects of maternal stress on offspring phenotypes (Bale 2015). While I did not measure epigenetics processes directly in chapter five, one of the significantly enriched biological processes that was identified as differentially expressed was the regulation of micro RNA expression. Micro RNAs are thought 127  to be one of the mechanisms that could play a role in epigenetic changes (Chuang & Jones 2007; Sætrom et al. 2007). It is possible that maternal stress could influence the composition of maternally derived micro RNAs that are deposited into the developing egg and these micro RNAs could influence developmental processes.  The results described in chapter 5 are particularly important in the context of ongoing debates regarding the evolutionary implications of maternal effects in natural populations (Räsänen & Kruuk 2007; Badyaev 2008; Sheriff & Love 2013). The environmental/maternal-matching hypothesis (Gluckman & Hanson 2004; Monaghan 2008; Love & Williams 2008; Sheriff & Love 2013; Sheriff et al. 2017) describes the adaptive potential of the effects of maternal stress on offspring phenotypes when the maternal environment is predictive of the environment experienced by the offspring. In the context of this hypothesis, divergent effects of maternal stress in males and females could suggest that the optimal phenotype in a stressful environment could differ depending on the sex of the offspring. In contrast, divergent effects in males and females could also suggest that the effects of maternal programming may be adaptive for one sex and maladaptive for the other in a stressful environment. However, both scenarios suggest that generalized effects of maternal stress on offspring performance should take the sex of the offspring into account. Whether maternal stress effects are adaptive or maladaptive will require additional studies that consider the fitness consequences of these effects in different environmental stress scenarios.  Stickleback exhibit paternal care of their offspring in which the male stickleback protects the developing embryos in a nest. The paternal duties of a stickleback include fanning the eggs to provide sufficiently oxygenated water to the clutch as well as defending the clutch from predators (Van Iersel 1953). Studies of adult offspring from predator-exposed fathers have 128  observed a decrease in offspring condition and offspring activity (Stein & Bell 2014; McGhee & Bell 2014; Bell et al. 2016). Interestingly reduced parental care in mammals (generally maternal care) is associated with the epigenetic modification of gene regulatory networks (Weaver et al. 2004). Whether the underlying mechanisms impacted by variation in paternal care in stickleback are similar to those impacted by altered maternal care in mammalian systems remains unknown.  6.4 Methodological considerations  The current gold-standard for characterizing DNA methylation levels at high resolution is the application of bisulfite sequencing, which can be applied to single candidate loci, subsets of the whole genome (e.g. RRBS), or as WGBS. Bisulfite sequencing of candidate loci is cost-effective and straightforward, but is critically dependent on the appropriate choice of candidate gene. This can be difficult in the absence of prior information suggesting that a genomic region is likely to be differentially methylated. In chapters two and four of this thesis I identify several candidate loci that are differentially methylated in adults from different developmental temperatures, in adults acclimated to different temperatures, and in stickleback reared at different salinities. In chapter five I describe the effects of maternal stress on gene expression patterns and many of the genes identified are known be regulated by DNA methylation (e.g. nr3c1). Taken together, these data chapters highlight several candidate loci that could be studied using a cost effective candidate approach that could further investigate environmentally dependent relationships between DNA methylation and gene expression.  In contrast, WGBS provides complete information about the methylation status of every cytosine in the genome, but this technique remains prohibitively expensive for many studies. RRBS is thought to provide cost effective alternative that captures a subset of the methylated 129  regions of the genome at a more reasonable cost. Similarly, the recently developed non-bisulfite-dependent method of EpiRADseq (Schield et al. 2016) represents another approach to increase the tractability of high-resolution epigenomic techniques. However, based on results described in chapters two and three, it is difficult to determine the extent to which the effects of developmental temperature on adult gene expression patterns may be associated with changes to DNA methylation patterns when using a reduced representation approach such as RRBS or EpiRADseq. RRBS covers a very small subset of the CpG loci in the genome making it very difficult to correlate changes in DNA methylation in specific regions of interest (ie. regions near specific genes), or to datasets with a small number of differentially expressed genes. Instead, RRBS may be particularly useful for situations where changes in transcript abundance for a large number of genes are observed, such as in the adult thermal acclimation response, so that the odds of obtaining DNA methylation information near differentially expressed genes are better. Even under these conditions, RRBS does not provide a comprehensive analysis of the DNA methylation state of differentially expressed genes, but it can provide evidence of a role for DNA methylation in the regulation of genes involved in key biological processes.  One key challenge for epigenomic studies of DNA methylation is that these studies examine correlations between changes in methylation patterns and phenotypic changes, which does not directly establish causation. To demonstrate that changes DNA methylation levels are causing the observed changes in phenotype requires direct functional assays such as mutagenesis of the putatively important methylated sites. New technologies such as the CRISPR/Cas9 endonuclease system (Hsu et al. 2014) should allow these causal links to be established. This technique has been proposed to allow targeted genetic manipulation of essentially any species and has already been successfully applied in a marine fish species (Aluru et al. 2015), and recent 130  technological advances (eg CRISP-Cas SunTag-directed DNMT3A; Huang et al. 2017b) have modified the CRISPR/Cas9 system such that is now possible to modify the DNA methylation level of specific target regions of the genome.   6.5 Future epigenomic research in fishes  One of the biggest unresolved issues in epigenomic research is determining if, and understanding how, epigenetic marks such as DNA methylation are transferred from one generation to the next. Studies in mammalian systems suggest that DNA methylation patterns are largely erased and replaced during early development, thus the efficacy of transgenerational epigenetic inheritance in vertebrates has been in doubt. Earlier in this chapter I argued that ecologically and evolutionarily relevant epigenetic changes do not necessary have to be heritable from one generation to the next to be important for adaptive evolution and that the environment may play a large role in establishing epigenetic patterns in a population. However, my intent was not to suggest that heritable epigenetic variation does not exist, or that heritable epigenetic variation is not an important form of heritable variation. Epidemiological studies in humans, for example, clearly show that maternal condition can have transgenerational effects on DNA methylation patterns that are associated with increased susceptibility of developing chronic diseases (Babenko et al. 2015), yet how these epigenetic effects are maintained during development is not well understood. Results from chapter five of this thesis suggest that many of these maternal effects may be acting through conserved pathways in mammals and fish, yet fish have many unique attributes compared to other model systems that make them a powerful system in which to investigate the mechanisms of transgenerational epigenetic inheritance. 131   In mammals, an environmentally induced phenotype must be maintained at least to the F3 generation to convincingly demonstrate a heritable epigenetic effect. This is because exposure of the gestating female (F0) simultaneously exposes the developing embryo (F1) and the developing germ line of the embryo (F2). In contrast, in fishes only the mother (FO) and her eggs (F1) are exposed. Thus, in fishes it is possible to conclusively demonstrate an epigenetic effect by the F2 generation. In addition, because many fishes develop externally, it is relatively straightforward to manipulate the environment during development, especially in comparison to the challenges of these approaches in mammals. Many species also have large clutch sizes, which allows for full-sib comparisons by using a split-clutch design in which multiple members of the same family are exposed to different treatments. Similarly, a split clutch design can be used to generate multiple half-sib families, which could help to disentangle paternal and maternal effects.   Fishes are particularly well suited for future research on the role of DNA methylation in sex-determination and the evolution of sex chromosomes. Differential methylation has been shown to be involved in the sex-specific regulation of key genes involved in ESD mechanisms (Navarro-Martín et al. 2011; Chen et al. 2014; Shao et al. 2014; Wen et al. 2014), establishing a foundation in which to build on our understanding of the role of DNA methylation in the evolution of ESD processes. The data presented in chapter 4 describe an apparent association between sex-specific methylation patterns and the evolution of heteromorphic sex chromosomes. While the X and Y chromosomes in stickleback are at a relatively early stage of differentiation, sex chromosome systems in other fish species are independently derived and are thought to be at various stages of differentiation (Mank et al. 2006; Wright et al. 2016), providing a model framework to investigate the theories described by Gorelick (2003) and the roles of DNA 132  methylation as the underlying mechanism regulating the evolution of sex determination mechanisms.  Although fish and mammals share many common features of their methylation patterns, which are distinct from the patterns in other major taxonomic groups such as the invertebrates, fungi and plants (Zemach et al. 2010), there are also fundamental differences in methylation patterns and mechanisms between fish and mammals (Goll & Halpern 2011) that are interesting for future research. For example, studies in zebrafish suggest that fish and mammals may differ in the extent to which methylation patterns are erased early in development, with current studies suggesting that fish retain a higher proportion of methylated sites (Jiang et al. 2013; Potok et al. 2013). Whether other species of fish have similar developmental methylation levels and what the mechanisms are that either retain or re-establish epigenetic patterns is unknown.  Fish genomes also contain additional DNA methyltransferase genes not present in the genomes of other organisms. As in mammals, the major proteins involved in DNA methylation in fishes are the DNA methyltransferases (DNMTs). The zebrafish genome contains 8 DNA methyltransferase orthologs: one dnmt1 ortholog, one dnmt2 ortholog, two dnmt3a orthologs, and four dnmt3b orthologs, although they lack an ortholog of the mammalian dnmt3L isoform (Wu et al. 2011). The teleosts are known to have undergone a whole-genome duplication relative to mammals (Jaillon et al. 2004), thus zebrafish might be expected to have two copies of each mammalian Dnmt family member. This suggests that the duplicates of dnmt1 and dnmt2 were lost during the rediploidization following the whole genome-duplication event, and that dnmt3b has undergone an additional round or rounds of duplication in zebrafish following the teleost whole-genome duplication. Whether these genes facilitate a unique functional role of DNA methylation, and the evolution of this gene family in fish remains to be studied. For example, the 133  lack of a dnmt3l gene in zebrafish is consistent with the reduced importance of imprinting in fishes. For example, in contrast to mammals, homozygous diploid clones are viable in fishes demonstrating that imprinting must be fundamentally different in fishes compared to mammals (Streisinger et al. 1981). DNMT3L is also absent from amphibian and bird genomes, suggesting that this gene may have arisen in mammals along with the evolution of X-inactivation (Yokomine et al. 2006).     6.6 Conclusions  The threespine stickleback is an effective model system for studying the mechanisms of natural selection and the evolution of adaptive phenotypes. In this thesis I have demonstrated a relationship between DNA methylation and several key ecological and evolutionary processes, which builds on the rich body of research in stickleback and establishes a new base from which to continue to examine and develop the ecological and evolutionary of role(s) of DNA methylation.  Because there was very little empirical data that demonstrates a relationship between environmental conditions and epigenetic processes in stickleback, I started exploring this relationship by investigating plasticity of DNA methylation patterns and gene expression at multiple timescales in response to changes in temperature, salinity, and maternal stress. From these studies, we now know that DNA methylation patterns are dynamically regulated in response to several key environmental factors that are thought to drive directional selection and divergence of many phenotypic traits in stickleback. I also show that the dynamic regulation of DNA methylation patterns is controlled by both developmental processes and in adult acclimation responses, which suggests that manipulating DNA methylation patterns is a key 134  component of developmental and adult phenotypic responses to environmental change. These results also suggest that this dynamic regulation of DNA methylation could contribute to much of the observed epigenetic variation observed between distinct populations of stickleback and should be considered in the interpretations of environmental epigenetic datasets.    I have also demonstrated that DNA methylation patterns along the sex chromosome in stickleback are different between males and females and that these differences correspond to the known evolutionary history of the stickleback sex chromosome. To my knowledge, this is the first time that an epigenetic mechanism has been empirically tested and correlated with sex chromosome evolution in vertebrates. This result not only provides the foundation for future experimental work on the role of DNA methylation in sex chromosome evolution, but also to addresses fundamental questions about the relationship between DNA methylation and genetic divergence using sex chromosomes as a model.  Furthermore, I have demonstrated that the effects of maternal stress differ between male and female offspring in stickleback. This result is particularly important because it suggests that the adaptive potential of transgenerational environmental effects might depend on the sex of the offspring. These data also suggest that the effects of maternal stress gene expression in offspring likely impact similar cellular mechanisms in both stickleback and mammalian systems suggesting that future research in this area may have broader implications beyond the ecology and evolution of stickleback.  Taken together, the results of my research clearly demonstrate the high degree of plasticity in DNA methylation and gene expression in response to environmental change across multiple time-scales, and highlight potential role of variation between sexes in these epigenetic processes. 135  Bibliography  Adusumalli S, Mohd Omar MF, Soong R, Benoukraf T (2015) Methodological aspects of whole-genome bisulfite sequencing analysis. Briefings in Bioinformatics, 16, 369–379. Agrawal AA, Laforsch C, Tollrian R (1999) Transgenerational induction of defences in animals and plants. Nature, 401, 60–63. Akalin A, Kormaksson M, Li S et al. (2012) methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. Genome Biology, 13, R87. Alsop D, Vijayan MM (2009) Molecular programming of the corticosteroid stress axis during zebrafish development. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 153, 49–54. Aluru N, Karchner SI, Franks DG et al. (2015) Targeted mutagenesis of aryl hydrocarbon receptor 2a and 2b genes in Atlantic killifish (Fundulus heteroclitus). Aquatic Toxicology, 158, 192–201. Alvarado S, Rajakumar R, Abouheif E, Szyf M (2015) Epigenetic variation in the Egfr gene generates quantitative variation in a complex trait in ants. Nature Communications, 6, 6513. Anastasiadi D, Díaz N, Piferrer F (2017) Small ocean temperature increases elicit stage-dependent changes in DNA methylation and gene expression in a fish, the European sea bass. Scientific Reports, 7, 1–12. Anders S, Huber W, Nagalakshmi U et al. (2010) Differential expression analysis for sequence count data. Genome Biology, 11, R106. Andersen B, Rosenfeld MG (2001) POU domain factors in the neuroendocrine system: lessons from developmental biology provide insights into human disease. Endocrine reviews, 22, 2–35. Angers B, Castonguay E, Massicotte R (2010) Environmentally induced phenotypes and DNA methylation: how to deal with unpredictable conditions until the next generation and after. Molecular Ecology, 19, 1283–1295. Angilleta M (2009) Thermal adaptation: a theoretical and empirical synthesis. Oxford. Aniagu SO, Williams TD, Allen Y, Katsiadaki I, Chipman JK (2008) Global genomic methylation levels in the liver and gonads of the three-spine stickleback (Gasterosteus aculeatus) after exposure to hexabromocyclododecane and 17-β oestradiol. Environment International, 34, 310–317. Artemov A V, Mugue NS, Rastorguev SM et al. (2017) Genome-wide DNA methylation profiling reveals epigenetic adaptation of stickleback to marine and freshwater conditions. Molecular Biology and Evolution, 34, 2203–2213. Aubin-Horth N, Renn SCP (2009) Genomic reaction norms: using integrative biology to 136  understand molecular mechanisms of phenotypic plasticity. Molecular Ecology, 18, 3763–3780. Babenko O, Kovalchuk I, Metz GAS (2015) Stress-induced perinatal and transgenerational epigenetic programming of brain development and mental health. Neuroscience and Biobehavioral Reviews, 48, 70–91. Badyaev A V. (2008) Maternal effects as generators of evolutionary change: a reassessment. Annals of the New York Academy of Sciences, 1133, 151–61. Baillat D, Shiekhattar R (2009) Functional dissection of the human TNRC6 (GW182-related) family of proteins. Molecular and cellular biology, 29, 4144–4155. Bale TL (2011) Sex differences in prenatal epigenetic programming of stress pathways. Stress, 14, 348–56. Bale TL (2015) Epigenetic and transgenerational reprogramming of brain development. Nature reviews. Neuroscience, 16, 332–344. Bannister AJ, Kouzarides T (2011) Regulation of chromatin by histone modifications. Cell Research, 21, 381–395. Baránek M, Čechová J, Raddová J et al. (2015) Dynamics and reversibility of the DNA methylation landscape of grapevine plants (Vitis vinifera) stressed by in vitro cultivation and thermotherapy. PLoS ONE, 10, e0126638. Barnes KR, Cozzi RRF, Robertson G, Marshall WS (2014) Cold acclimation of NaCl secretion in a eurythermic teleost: mitochondrial function and gill remodeling. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 168, 50–62. Barrett RDH, Paccard A, Healy TM et al. (2011) Rapid evolution of cold tolerance in stickleback. Proceedings of the Royal Society B: Biological Sciences, 278, 233–238. Barrett RDH, Rogers SM, Schluter D (2008) Natural selection on a major armor gene in threespine stickleback. Science, 322, 255–7. Bauerfeind SS, Fischer K (2014) Simulating climate change: temperature extremes but not means diminish performance in a widespread butterfly. Population Ecology, 56, 239–250. Beal MF (1998) Mitochondrial dysfunction in neurodegenerative diseases. Biochim Biophys Acta, 1366, 211–223. Beaman JE, White CR, Seebacher F (2016) Evolution of plasticity: mechanistic link between development and reversible acclimation. Trends in Ecology and Evolution, 31, 237–249. Beijers R, Buitelaar JK, de Weerth C (2014) Mechanisms underlying the effects of prenatal psychosocial stress on child outcomes: beyond the HPA axis. European Child & Adolescent Psychiatry, 23, 943–956. Bell MA, Aguirre WE (2013) Contemporary evolution, allelic recycling, and adaptive radiation of the threespine stickleback. Evolutionary Ecology Research, 15, 377–411. 137  Bell MA, Foster SA (1994) The Evolutionary Biology of the Threespine Stickleback. Oxford University Press, New York. Bell AM, McGhee KE, Stein LR (2016) Effects of mothers’ and fathers’ experience with predation risk on the behavioral development of their offspring in threespined sticklebacks. Current Opinion in Behavioral Sciences, 7, 28–32. Bell MA, Stewart JD, Park PJ (2009) The world’s oldest fossil threespine stickleback fish. Copeia, 2, 256–265. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate : a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B, 57, 289–300. Bewick AJ, Hofmeister BT, Lee K et al. (2015) FASTmC: a suite of predictive models for non-reference-based estimations of DNA methylation. G3; Genes|Genomes|Genetics, 49, 2–15. Bird A (2002) DNA methylation patterns and epigenetic memory. Genes & Development, 16, 6–21. Bird A (2007) Perceptions of epigenetics. Nature, 447, 396–8. Bock C (2012) Analysing and interpreting DNA methylation data. Nature reviews. Genetics, 13, 705–19. Bonasio R (2015) The expanding epigenetic landscape of non-model organisms. Journal of Experimental Biology, 218, 114–122. Bonga SEW (1997) The stress response in fish. Physiological Reviews, 77, 591–625. Bossdorf O, Richards CL, Pigliucci M (2008) Epigenetics for ecologists. Ecology letters, 11, 106–15. Bostick M, Kim JK, Estève P-O et al. (2007) UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science, 317, 1760–4. Boughman JW (2007) Speciation in sticklebacks. In: Biology of the three-spined stickleback (eds Östlund-Nilson S, Mayer I, Huntingford FA), pp. 83–126. CRC Press, New York. Boyko A, Blevins T, Yao Y et al. (2010) Transgenerational adaptation of Arabidopsis to stress requires DNA methylation and the function of Dicer-Like proteins (S-H Shiu, Ed,). PLoS ONE, 5, e9514. Branco MR, Ficz G, Reik W (2011) Uncovering the role of 5-hydroxymethylcytosine in the epigenome. Nature Reviews Genetics, 13, 7–13. Brauner CJ, Matey V, Zhang W et al. (2011) Gill remodeling in crucian carp during sustained exercise and the effect on subsequent swimming performance. Physiological and Biochemical Zoology, 84, 535–542. Britten RJ, Baron WF, Stout DB, Davidson EH (1988) Sources and evolution of human Alu repeated sequences. Proceedings of the National Academy of Sciences of the United States of America, 85, 4770–4774. 138  Bronson SL, Bale TL (2016) The placenta as a mediator of stress effects on neurodevelopmental reprogramming. Neuropsychopharmacology, 41, 207–218. Bulmer M (1986) Neighboring base effects on substitution rates in pseudogenes. Molecular biology and evolution, 3, 322–329. Burgerhout E, Mommens M, Johnsen H et al. (2017) Genetic background and embryonic temperature affect DNA methylation and expression of myogenin and muscle development in Atlantic salmon (Salmo salar). PLoS ONE, 12, e0179918. Burggren WW, Crews D (2014) Epigenetics in comparative biology: why we should pay attention. Integrative and Comparative Biology, 54, 7–20. Cahill L, Aswad D (2015) Sex influences on the brain: an issue whose time has come. Neuron, 88, 1084–1085. Campos C, Valente L, Conceição L, Engrola S, Fernandes J (2013) Temperature affects methylation of the myogenin putative promoter, its expression and muscle cellularity in Senegalese sole larvae. Epigenetics, 8, 389–397. Castonguay E, Angers B (2012) The key role of epigenetics in the persistence of asexual lineages. Genetics research international, 2012, 534289. Cervera MT, Ruiz-García L, Martínez-Zapater JM (2002) Analysis of DNA methylation in Arabidopsis thaliana based on methylation-sensitive AFLP markers. Molecular Genetics and Genomics, 268, 543–552. Charmantier A, McCleery RH, Cole LR et al. (2008) Adaptive phenotypic plasticity in response to climate change in a wild bird population. Science, 320, 800–803. Chatterjee A, Lagisz M, Rodger EJ et al. (2016) Sex differences in DNA methylation and expression in zebrafish brain: a test of an extended “male sex drive” hypothesis. Gene, 590, 307–316. Chen P-Y, Cokus SJ, Pellegrini M (2010) BS Seeker: precise mapping for bisulfite sequencing. BMC bioinformatics, 11, 203. Chen S, Zhang G, Shao C et al. (2014) Whole-genome sequence of a flatfish provides insights into ZW sex chromosome evolution and adaptation to a benthic lifestyle. Nature genetics, 46, 253–60. Ching T, Huang S, Garmire LX (2014) Power analysis and sample size estimation for RNA-Seq differential expression. RNA, 20, 1684–1696. Choi W Il, Jeon BN, Yoon JH et al. (2013) The proto-oncoprotein FBI-1 interacts with MBD3 to recruit the Mi-2/NuRD-HDAC complex and BCoR and to silence p21WAF/CDKN1A by DNA methylation. Nucleic Acids Research, 41, 6403–6420. Chown SL, Gaston KJ, Robinson D (2004) Macrophysiology: large-scale patterns in physiological traits and their ecological implications. Functional Ecology, 18, 159–167. Chown S, Hoffmann A, Kristensen T et al. (2010) Adapting to climate change: a perspective from evolutionary physiology. Climate Research, 43, 3–15. 139  Chuang JC, Jones PA (2007) Epigenetics and microRNAs. Pediatric Research, 61, 24–29. Clusella-Trullas S, Blackburn TM, Chown SL (2011) Climatic predictors of temperature performance curve parameters in ectotherms imply complex responses to climate change. The American Naturalist, 177, 738–751. Colinet H, Hoffmann AA (2012) Comparing phenotypic effects and molecular correlates of developmental, gradual and rapid cold acclimation responses in Drosophila melanogaster. Functional Ecology, 26, 84–93. Colosimo PF, Hosemann KE, Balabhadra S et al. (2005) Widespread parallel evolution in sticklebacks by repeated fixation of Ectodysplasin alleles. Science, 307, 1928–33. Cooke B, Hegstrom CD, Villeneuve LS, Breedlove SM (1998) Sexual differentiation of the vertebrate brain: principles and mechanisms. Frontiers in neuroendocrinology, 19, 323–362. Cooney CA, Dave AA, Wolff GL (2002) Maternal methyl supplements in mice affect epigenetic variation and DNA methylation of offspring. The Journal of nutrition, 132, 2393S–2400S. Cossins A (2006) Post-genomic approaches to understanding the mechanisms of environmentally induced phenotypic plasticity. Journal of Experimental Biology, 209, 2328–2336. Costa FF (2008) Non-coding RNAs, epigenetics and complexity. Gene, 410, 9–17. Coulondre C, Miller JH, Farabaugh PJ, Gilbert W (1978) Molecular basis of base substitution hotspots in Escherichia coli. Nature, 274, 775–780. Dalziel AC, Rogers SM, Schulte PM (2009) Linking genotypes to phenotypes and fitness: how mechanistic biology can inform molecular ecology. Molecular ecology, 18, 4997–5017. Darnaudéry M, Maccari S (2008) Epigenetic programming of the stress response in male and female rats by prenatal restraint stress. Brain Research Reviews, 57, 571–585. Deans C, Maggert KA (2015) What do you mean, “epigenetic”? Genetics, 199, 887–896. Deaton AM, Bird A (2011) CpG islands and the regulation of transcription. Genes & development, 25, 1010–22. DeFaveri J, Jonsson PR, Merilä J (2013) Heterogeneous genomic differentiation in marine threespine sticklebacks: adaptation along an environmental gradient. Evolution, 67, 2530–2546. DeFaveri J, Merilä J (2014) Local adaptation to salinity in the three-spined stickleback? Journal of Evolutionary Biology, 27, 290–302. DeFaveri J, Shikano T, Shimada Y, Goto A, Merilä J (2011) Global analysis of genes involved in freshwater adaptation in threespine sticklebacks (Gasterosteus aculeatus). Evolution, 65, 1800–1807. Denny M (2017) The fallacy of the average: on the ubiquity, utility and continuing 140  novelty of Jensen’s inequality. The Journal of Experimental Biology, 220, 139–146. Devlin RH, Nagahama Y (2002) Sex determination and sex differentiation in fish: an overview of genetic, physiological, and environmental influences. Aquaculture, 208, 191–364. Dickins TE, Rahman Q (2012) The extended evolutionary synthesis and the role of soft inheritance in evolution. Proceedings of the Royal Society B: Biological Sciences, 279, 2913–2921. Dillies M-A, Rau A, Aubert J et al. (2013) A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis. Briefings in Bioinformatics, 14, 671–683. Divino JN, Monette MY, McCormick SD et al. (2016) Osmoregulatory physiology and rapid evolution of salinity tolerance in threespine stickleback recently introduced to fresh water. Evolutionary Ecology Research, 17, 179–201. Donelson JM, Salinas S, Munday PL, Shama LNS (2017) Transgenerational plasticity and climate change experiments: where do we go from here? Global Change Biology, 1–22. Dowd WW, King FA, Denny MW (2015) Thermal variation, thermal extremes and the physiological performance of individuals. Journal of Experimental Biology, 218, 1956–1967. Dowen RH, Pelizzola M, Schmitz RJ et al. (2012) Widespread dynamic DNA methylation in response to biotic stress. Proceedings of the National Academy of Sciences, 109, E2183–E2191. Du Y, Liu B, Guo F et al. (2012) The essential role of Mbd5 in the regulation of somatic growth and glucose homeostasis in mice. PLoS ONE, 7, e47358. Dubin MJ, Zhang P, Meng D et al. (2015) DNA methylation in Arabidopsis has a genetic basis and shows evidence of local adaptation. eLife, 4, e05255. Duncan BK, Miller JH (1980) Mutagenic deamination of cytosine residues in DNA. Nature, 287, 560–561. Edwards SL, Marshall WS (2013) Principles and patterns of osmoregulation and euryhalinity in fishes. In: Fish Physiology: Euryhaline Fishes (eds McCormick SD, Farrell AP, Brauner CJ), pp. 1–44. Academic Press, New York. Egginton S, Sidell BD (1989) Thermal acclimation induces adaptive changes in subcellular structure of fish skeletal muscle. The American journal of physiology, 256, R1-9. Ellegren H, Parsch J (2007) The evolution of sex-biased genes and sex-biased gene expression. Nature reviews. Genetics, 8, 689–98. Evans C, Hardin J, Stoebel D (2017) Selecting between-sample RNA-Seq normalization methods from the perspective of their assumptions. Briefings In Bioinformatics, 1–17. Felsenstein J (1974) The evolutionary advantage of recombination. Genetics, 78, 737–56. 141  Feng S, Cokus SJ, Zhang X et al. (2010) Conservation and divergence of methylation patterning in plants and animals. Proceedings of the National Academy of Sciences of the United States of America, 107, 8689–94. Finstad A, Jonsson B (2012) Effect of incubation temperature on growth performance in Atlantic salmon. Marine Ecology Progress Series, 454, 75–82. Flanagan JM, Popendikyte V, Pozdniakovaite N et al. (2006) Intra- and interindividual epigenetic variation in human germ cells. American journal of human genetics, 79, 67–84. Flores KB, Wolschin F, Amdam G V (2013) The role of methylation of DNA in environmental adaptation. Integrative and comparative biology, 53, 359–372. Fraga MF, Ballestar E, Paz MF et al. (2005) Epigenetic differences arise during the lifetime of monozygotic twins. Proceedings of the National Academy of Sciences of the United States of America, 102, 10604–9. Frainer A, Primicerio R, Kortsch S et al. (2017) Climate-driven changes in functional biogeography of Arctic marine fish communities. Proceedings of the National Academy of Sciences, 201706080. Francis D, Diorio J, Liu D, Meaney MJ (1999) Nongenomic transmission across generations of maternal behavior and stress responses in the rat. Science, 286, 1155–8. Franks SJ, Hoffmann AA (2012) Genetics of climate change adaptation. Annual review of genetics, 46, 185–208. Fry FEJ (1947) Effects of the environment on animal activity. Publications of the Ontario Fisheries Research Laboratory. Fu S-J, Brauner CJ, Cao Z-D et al. (2011) The effect of acclimation to hypoxia and sustained exercise on subsequent hypoxia tolerance and swimming performance in goldfish (Carassius auratus). Journal of Experimental Biology, 214, 2080–2088. Gerken AR, Eller OC, Hahn DA, Morgan TJ (2015) Constraints, independence, and evolution of thermal plasticity: probing genetic architecture of long- and short-term thermal acclimation. Proceedings of the National Academy of Sciences, 112, 4399–4404. Gertz J, Varley KE, Reddy TE et al. (2011) Analysis of DNA methylation in a three-generation family reveals widespread genetic influence on epigenetic regulation. PLoS Genetics, 7, e1002228. Ghalambor CK, Hoke KL, Ruell EW et al. (2015) Non-adaptive plasticity potentiates rapid adaptive evolution of gene expression in nature. Nature, 525, 372–375. Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non-adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Functional Ecology, 21, 394–407. Gibbons TC, Metzger DCH, Healy TM, Schulte PM (2017) Gene expression plasticity in response to salinity acclimation in threespine stickleback ecotypes from different 142  salinity habitats. Molecular Ecology, 26, 2711–2725. Giesing ER, Suski CD, Warner RE, Bell AM (2011) Female sticklebacks transfer information via eggs: effects of maternal experience with predators on offspring. Proceedings of the Royal Society B: Biological Sciences, 278, 1753–1759. Gigek CO, Chen ES, Smith MAC (2016) Methyl-CpG-binding protein (MBD) family: epigenomic read-outs functions and roles in tumorigenesis and psychiatric diseases. Journal of cellular biochemistry, 117, 29–38. Glazer AM, Killingbeck EE, Mitros T, Rokhsar DS, Miller CT (2015) Genome assembly improvement and mapping convergently evolved skeletal traits in sticklebacks with genotyping-by-sequencing. G3, 5, 1463–72. Gluckman PD, Hanson MA (2004) Developmental origins of disease paradigm: a mechanistic and evolutionary perspective. Pediatric Research, 56, 311–317. Goll MG, Halpern ME (2011) DNA methylation in zebrafish. Progress in molecular biology and translational science, 101, 193–218. Goll MG, Kirpekar F, Maggert K a et al. (2006) Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2. Science, 311, 395–8. Gonda A, Herczeg G, Merila J (2009) Habitat-dependent and -independent plastic responses to social environment in the nine-spined stickleback (Pungitius puntitius) brain. Proceedings of the Royal Society B: Biological Sciences, 276, 2085–2092. Gorelick R (2003) Evolution of dioecy and sex chromosomes via methylation driving Muller’s ratchet. Biological Journal of the Linnean Society, 80, 353–368. Gracey AY (2007) Interpreting physiological responses to environmental change through gene expression profiling. Journal of Experimental Biology, 210, 1584–1592. Gracey AY, Fraser EJ, Li W et al. (2004) Coping with cold: an integrative, multitissue analysis of the transcriptome of a poikilothermic vertebrate. Proceedings of the National Academy of Sciences of the United States of America, 101, 16970–16975. Graves JAM (2016) Evolution of vertebrate sex chromosomes and dosage compensation. Nature reviews. Genetics, 17, 33–46. Gu H, Bock C, Mikkelsen TS et al. (2010) Genome-scale DNA methylation mapping of clinical samples at single-nucleotide resolution. Nature Methods, 7, 133–136. Gu L, Frommel SC, Oakes CC et al. (2015) BAZ2A (TIP5) is involved in epigenetic alterations in prostate cancer and its overexpression predicts disease recurrence. Nat Genet, 47, 22–30. Gu H, Smith ZD, Bock C et al. (2011) Preparation of reduced representation bisulfite sequencing libraries for genome-scale DNA methylation profiling. Nature protocols, 6, 468–481. Guderley H (2004) Metabolic responses to low temperature in fish muscle. Biological reviews of the Cambridge Philosophical Society, 79, 409–427. Guderley H, Leroy PH, Gagné A (2001) Thermal acclimation, growth, and burst 143  swimming of threespine stickleback: enzymatic correlates and influence of photoperiod. Physiological and Biochemical Zoology, 74, 66–74. Gugger PF, Fitz-Gibbon S, PellEgrini M, Sork VL (2016) Species-wide patterns of DNA methylation variation in Quercus lobata and their association with climate gradients. Molecular Ecology, 25, 1665–1680. Guo B, DeFaveri J, Sotelo G, Nair A, Merilä J (2015) Population genomic evidence for adaptive differentiation in Baltic Sea three-spined sticklebacks. BMC Biology, 13, 19. Haglund TR, Buth DG, Lawson R (1992) Allozyme variation and phylogenetic relationships of Asian , North American , and European populations of the threespine stickleback, Gasterosteus aculeatus. Copeia, 1992, 432–443. Hales CN, Barker DJ (2001) The thrifty phenotype hypothesis. British medical bulletin, 60, 5–20. Harrison A, Parle-McDermott A (2011) DNA methylation: a timeline of methods and applications. Frontiers in genetics, 2, 74. Healy TM, Bryant HJ, Schulte PM (2017) Mitochondrial genotype and phenotypic plasticity of gene expression in response to cold acclimation in killifish. Molecular Ecology, 26, 814–830. Heard E, Martienssen RA (2014) Transgenerational epigenetic inheritance: myths and mechanisms. Cell, 157, 95–109. Heitzer MD, Wolf IM, Sanchez ER, Witchel SF, DeFranco DB (2007) Glucocorticoid receptor physiology. Reviews in Endocrine and Metabolic Disorders, 8, 321–330. Hellstrom IC, Dhir SK, Diorio JC, Meaney MJ (2012) Maternal licking regulates hippocampal glucocorticoid receptor transcription through a thyroid hormone-serotonin-NGFI-A signalling cascade. Philosophical Transactions of the Royal Society B: Biological Sciences, 367, 2495–2510. Hendry AP (2016) Key questions on the role of phenotypic plasticity in eco-evolutionary dynamics. Journal of Heredity, 107, 25–41. Herczeg G, Gonda A, Balázs G, Noreikiene K, Merilä J (2015) Experimental evidence for sex-specific plasticity in adult brain. Frontiers in Zoology, 12, 38. Herczeg G, Välimäki K, Gonda A, Merilä J (2014) Evidence for sex-specific selection in brain: a case study of the nine-spined stickleback. Journal of Evolutionary Biology, 27, 1604–1612. Heuts MJ (1947) Experimental studies on adaptive evolution in Gasterosteus aculeatus L. Evolution, 1, 89–102. Higuchi M, Goto A (1996) Genetic evidence supporting the existence of two distinct species in the genus Gasterosteus around Japan. Environmental Biology of Fishes, 47, 1–16. Hill RW, Wyse GA, Anderson M (2008) Animal Physiology. Sinauer Associates, Sunderland, Massachusetts. 144  Hochachka P, Somero G (2002) Biochemical adaptation: mechanism and process in physiological evolution. New York. Hofmann GE (2017) Ecological epigenetics in marine metazoans. Frontiers in Marine Science, 4, 4. Holloway CC, Clayton DF (2001) Estrogen synthesis in the male brain triggers development of the avian song control pathway in vitro. Nature neuroscience, 4, 170–175. Hovel RA, Carlson SM, Quinn TP (2017) Climate change alters the reproductive phenology and investment of a lacustrine fish, the three-spine stickleback. Global Change Biology, 23, 2308–2320. Hsu PD, Lander ES, Zhang F (2014) Development and applications of CRISPR-Cas9 for genome engineering. Cell, 157, 1262–1278. Hu J, Barrett RDH (2017) Epigenetics in natural animal populations. Journal of Evolutionary Biology, 30, 1612–1632. Huang X, Li S, Ni P et al. (2017a) Rapid response to changing environments during biological invasions: DNA methylation perspectives. Molecular Ecology. Huang Y-H, Su J, Lei Y et al. (2017b) DNA epigenome editing using CRISPR-Cas SunTag-directed DNMT3A. Genome Biology, 18, 176. Van Iersel JJA (1953) An analysis of the parental behaviour of the male three-spined stickleback (Gasterosteus aculeatus L.). Behaviour. Supplement., 1–159. Ikeda D, Koyama H, Mizusawa N et al. (2017) Global gene expression analysis of the muscle tissues of medaka acclimated to low and high environmental temperatures. Comparative Biochemistry and Physiology - Part D: Genomics and Proteomics, 24, 19–28. Ishikawa A, Kusakabe M, Yoshida K et al. (2017) Different contributions of local- and distant-regulatory changes to transcriptome divergence between stickleback ecotypes. Evolution, 71, 565–581. Jablonka E, Lamb MJ (2005) Evolution in four dimensions: genetic, epigenetic, behavioral, and symbolic variation in the history of life. MIT Press, Cambridge, Massachusetts. Jablonka E, Raz G (2009) Transgenerational epigenetic inheritance: prevalence, mechanisms, and implications for the study of heredity and evolution. The Quarterly review of biology, 84, 131–76. Jacobs LF (1996) Sexual selection and the brain. Trends in Ecology and Evolution, 11, 82–86. Jaenisch R, Bird A (2003) Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nature Genetics, 33, 245–254. Jaillon O, Aury J-M, Brunet F et al. (2004) Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature, 431, 946–957. 145  Jazin E, Cahill L (2010) Sex differences in molecular neuroscience: from fruit flies to humans. Nature reviews. Neuroscience, 11, 9–17. Jeddeloh JA, Greally JM, Rando OJ (2008) Reduced-representation methylation mapping. Genome Biology, 9, 231. Jeong M, Sun D, Luo M et al. (2014) Large conserved domains of low DNA methylation maintained by Dnmt3a. Nature genetics, 46, 17–23. Jiang L, Zhang J, Wang J-J et al. (2013) Sperm, but not oocyte, DNA methylome is inherited by zebrafish early embryos. Cell, 153, 773–84. Jjingo D, Conley A, Yi S, Lunyak V, Jordan I (2012) On the presence and role of human gene-body DNA methylation. Oncotarget, 3, 462–474. Johannes F, Porcher E, Teixeira FK et al. (2009) Assessing the Impact of Transgenerational Epigenetic Variation on Complex Traits (PM Visscher, Ed,). PLoS Genetics, 5, e1000530. Johnston IA (2006) Environment and plasticity of myogenesis in teleost fish. Journal of Experimental Biology, 209, 2249–2264. Johnston IA, Lee H-T, Macqueen DJ et al. (2009) Embryonic temperature affects muscle fibre recruitment in adult zebrafish: genome-wide changes in gene and microRNA expression associated with the transition from hyperplastic to hypertrophic growth phenotypes. The Journal of experimental biology, 212, 1781–1793. Jones PA (2012) Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nature Reviews Genetics, 13, 484–492. Jones FC, Chan YF, Schmutz J et al. (2012a) A genome-wide SNP genotyping array reveals patterns of global and repeated species-pair divergence in sticklebacks. Current Biology, 22, 83–90. Jones FC, Grabherr MG, Chan YF et al. (2012b) The genomic basis of adaptive evolution in threespine sticklebacks. Nature, 484, 55–61. Jordan CM, Garside ET (1972) Upper lethal temperatures of threespine stickleback, Gasterosteus aculeatus (L.), in relation to thermal and osmotic acclimation, ambient salinity, and size. Canadian Journal of Zoology, 50, 1405–1411. Kalujnaia S, McWilliam IS, Zaguinaiko VA et al. (2007) Salinity adaptation and gene profiling analysis in the European eel (Anguilla anguilla) using microarray technology. General and Comparative Endocrinology, 152, 274–280. Kamiya T, Kai W, Tasumi S et al. (2012) A trans-species missense SNP in Amhr2 is associated with sex determination in the tiger pufferfish, Takifugu rubripes (Fugu). PLoS Genetics, 8, e1002798. Kamstra JH, Aleström P, Kooter JM, Legler J (2015) Zebrafish as a model to study the role of DNA methylation in environmental toxicology. Environmental Science and Pollution Research, 22, 16262–16276. Kaslin J, Ganz J, Brand M (2008) Proliferation, neurogenesis and regeneration in the non-mammalian vertebrate brain. Philosophical transactions of the Royal Society of 146  London. Series B, Biological sciences, 363, 101–22. Kawahara R, Miya M, Mabuchi K, Near TJ, Nishida M (2009) Stickleback phylogenies resolved: evidence from mitochondrial genomes and 11 nuclear genes. Molecular phylogenetics and evolution, 50, 401–4. Kawakatsu T, Huang SC, Jupe F et al. (2016) Epigenomic diversity in a global collection of Arabidopsis thaliana accessions. Cell, 166, 492–505. Kelleher AR, Kimball SR, Dennis MD, Schilder RJ, Jefferson LS (2013) The mTORC1 signaling repressors REDD1/2 are rapidly induced and activation of p70S6K1 by leucine is defective in skeletal muscle of an immobilized rat hindlimb. AJP: Endocrinology and Metabolism, 304, E229–E236. Keller TE, Lasky JR, Yi S V. (2016) The multivariate association between genomewide DNA methylation and climate across the range of Arabidopsis thaliana. Molecular Ecology, 25, 1823–1837. Keyte AL, Percifield R, Liu B, Wendel JF (2006) Infraspecific DNA methylation polymorphism in cotton (Gossypium hirsutum L.). Journal of Heredity, 97, 444–450. Kilvitis HJ, Alvarez M, Foust CM et al. (2014) Ecological Epigentics. In: Ecological Genomics Advances in Experimental Medicine and Biology. (eds Landry CR, Aubin-Horth N), pp. 191–210. Springer Netherlands, Netherlands. Kim S-Y, Costa MM, Esteve-Codina A, Velando A (2017a) Transcriptional mechanisms underlying life-history responses to climate change in the three-spined stickleback. Evolutionary Applications, 718–730. Kim S-Y, Metcalfe NB, da Silva A, Velando A (2017b) Thermal conditions during early life influence seasonal maternal strategies in the three-spined stickleback. BMC Ecology, 17, 34. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF (1995) Stages of embryonic development of the zebrafish. Developmental Dynamics, 203, 253–310. Kingsolver JG, Pfennig DW, Servedio MR (2002) Migration, local adaptation and the evolution of plasticity. Trends in Ecology and Evolution, 17, 540–541. Kitano J, Mori S, Peichel CL (2007) Phenotypic divergence and reproductive isolation between sympatric forms of Japanese threespine sticklebacks. Biological Journal of the Linnean Society, 91, 671–685. Klose RJ, Bird AP (2006) Genomic DNA methylation: the mark and its mediators. Trends in biochemical sciences, 31, 89–97. Kolb B, Whishaw IQ (1998) Brain plasticity and behavior. Annual Review of Psychology, 49, 43–64. Kolm N, Gonzalez-Voyer A, Brelin D, Winberg S (2009) Evidence for small scale variation in the vertebrate brain: mating strategy and sex affect brain size and structure in wild brown trout (Salmo trutta). Journal of Evolutionary Biology, 22, 2524–2531. 147  Kotrschal A, Räsänen K, Kristjánsson BK, Senn M, Kolm N (2012) Extreme sexual brain size dimorphism in sticklebacks: a consequence of the cognitive challenges of sex and parenting? PLoS ONE, 7, e30055. Kristensen TN, Hoffmann AA, Overgaard J et al. (2008) Costs and benefits of cold acclimation in field-released Drosophila. Proceedings of the National Academy of Sciences, 105, 216–221. Kroeze LI, van der Reijden BA, Jansen JH (2015) 5-Hydroxymethylcytosine: an epigenetic mark frequently deregulated in cancer. Biochimica et Biophysica Acta - Reviews on Cancer, 1855, 144–154. Kucharski R, Maleszka J, Foret S, Maleszka R (2008) Nutritional control of reproductive status in honeybees via DNA methylation. Science, 319, 1827–1830. Kusakabe M, Ishikawa A, Ravinet M et al. (2017) Genetic basis for variation in salinity tolerance between stickleback ecotypes. Molecular Ecology, 26, 304–319. Laget S, Joulie M, Le Masson F et al. (2010) The human proteins MBD5 and MBD6 associate with heterochromatin but they do not bind methylated DNA. PLoS ONE, 5. Laird PW (2010) Principles and challenges of genomewide DNA methylation analysis. Nature reviews. Genetics, 11, 191–203. Lappalainen T, Greally JM (2017) Associating cellular epigenetic models with human phenotypes. Nature Reviews Genetics, 18, 441–451. Laurent L, Wong E, Li G et al. (2010) Dynamic changes in the human methylome during differentiation Dynamic changes in the human methylome during differentiation. Genome research, 20, 320–331. Law JA, Jacobsen SE (2010) Establishing, maintaining and modifying DNA methylation patterns in plants and animals. Nature Reviews Genetics, 11, 204–220. LeBlanc DM, Wood CM, Fudge DS, Wright PA (2010) A fish out of water: gill and skin remodeling promotes osmo- and ionoregulation in the mangrove killifish Kryptolebias marmoratus. Physiological and Biochemical Zoology, 83, 932–949. Leder EH, Cano JM, Leinonen T et al. (2010) Female-biased expression on the X chromosome as a key step in sex chromosome evolution in threespine sticklebacks. Molecular biology and evolution, 27, 1495–503. Ledon-Rettig CC (2013) Ecological epigenetics: an introduction to the symposium. Integrative and Comparative Biology, 53, 307–318. Lee CE, Bell MA (1999) Causes and consequences of recent freshwater invasions by saltwater animals. Trends in Ecology & Evolution, 14, 284–288. Lee HJ, Lowdon RF, Maricque B et al. (2015) Developmental enhancers revealed by extensive DNA methylome maps of zebrafish early embryos. Nature Communications, 6, 6315. Lefébure R, Larsson S, Byström P (2011) A temperature-dependent growth model for the three-spined stickleback Gasterosteus aculeatus. Journal of Fish Biology, 79, 1815–148  1827. Levis NA, Pfennig DW (2016) Evaluating “plasticity-first” evolution in Nature: key criteria and empirical approaches. Trends in Ecology & Evolution, 31, 563–574. Li E (2002) Chromatin modification and epigenetic reprogramming in mammalian development. Nature Reviews Genetics, 3, 662–673. Li S, He F, Wen H et al. (2017) Analysis of DNA methylation level by methylation-sensitive amplification polymorphism in half smooth tongue sole (Cynoglossus semilaevis) subjected to salinity stress. Journal of Ocean University of China, 16, 269–278. Li J, Huang Q, Sun M et al. (2016) Global DNA methylation variations after short-term heat shock treatment in cultured microspores of Brassica napus cv. Topas. Scientific Reports, 6, 38401. Li E, Zhang Y (2014) DNA methylation in mammals. Cold Spring Harbor Perspectives in Biology, 6, a019133–a019133. Liebl AL, Schrey AW, Richards CL, Martin LB (2013) Patterns of DNA methylation throughout a range expansion of an introduced songbird. Integrative and Comparative Biology, 53, 351–358. Lin Y, Golovnina K, Chen Z-X et al. (2016) Comparison of normalization and differential expression analyses using RNA-Seq data from 726 individual Drosophila melanogaster. BMC Genomics, 17, 28. Lira-Medeiros CF, Parisod C, Fernandes RA et al. (2010) Epigenetic variation in mangrove plants occurring in contrasting natural environment. PLoS ONE, 5, e10326. Lister R, Pelizzola M, Dowen RH et al. (2009) Human DNA methylomes at base resolution show widespread epigenomic differences. Nature, 462, 315–322. Liu D (1997) Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal responses to stress. Science, 277, 1659–1662. Liu J, Carmell MA, Rivas F V et al. (2004) Argonaute2 is the catalytic engine of mammalian RNAi. Science, 305, 1437–41. Liu J, Morgan M, Hutchison K, Calhoun VD (2010) A study of the influence of sex on genome wide methylation. PLoS ONE, 5, e10028. Logan CA, Buckley BA (2015) Transcriptomic responses to environmental temperature in eurythermal and stenothermal fishes. Journal of Experimental Biology, 218, 1915–1924. Love OP, Williams TD (2008) The adaptive value of stress‐induced phenotypes: effects of maternally derived corticosterone on sex‐biased investment, cost of reproduction, and maternal fitness. The American Naturalist, 172, E135–E149. Lu J, Zheng M, Zheng J et al. (2015) Transcriptomic analyses reveal novel genes with sexually dimorphic expression in yellow catfish (Pelteobagrus fulvidraco) brain. Marine Biotechnology, 17, 613–623. 149  Lyko F, Foret S, Kucharski R et al. (2010) The honey bee epigenomes: differential methylation of brain DNA in queens and workers. PLoS Biology, 8, e1000506. Macleod D, Clark VH, Bird A (1999) Absence of genome-wide changes in DNA methylation during development of the zebrafish. Nature genetics, 23, 139–140. Macqueen DJ, Robb DHF, Olsen T et al. (2008) Temperature until the “eyed stage” of embryogenesis programmes the growth trajectory and muscle phenotype of adult Atlantic salmon. Biology letters, 4, 294–298. Mank JE, Promislow D, Avise JC (2006) Evolution of alternative sex-determining mechanisms in teleost fishes. Biological Journal of the Linnean Society, 87, 83–93. Manousaki T, Tsakogiannis A, Lagnel J et al. (2014) The sex-specific transcriptome of the hermaphrodite sparid sharpsnout seabream (Diplodus puntazzo). BMC Genomics, 15, 655. Marchinko KB (2009) Predation’s role in the repeated phenotypic and genetic divergence of armor in threespine stickleback. Evolution, 63, 127–138. Marchinko KB, Schluter D (2007) Parallel evolution by correlated response: lateral plate reduction in threespine stickleback. Evolution, 61, 1084–1090. Marteinsdottir G, Steinarsson A (1998) Maternal influence on the size and viability of Iceland cod Gadus morhua eggs and larvae. Journal of fish biology, 52, 1241–1258. Martinowich K, Hattori D, Wu H et al. (2003) DNA methylation-related chromatin remodeling in activity-dependent BDNF gene regulation. Science, 302, 890–3. Massicotte R, Whitelaw E, Angers B (2011) DNA methylation: A source of random variation in natural populations. Epigenetics, 6, 421–427. Mattern MY (2007) Phylogeny, systematics, and taxonomy of sticklebacks. In: Biology of the three-spined stickleback (eds Östlund-Nilsson S, Mayer I, Huntingford FA), pp. 1–40. CRC Press, New York. Mayr E (1982) The growth of biological thought: diversity , evolution , and inheritance. Harvard University Press, Cambridge, MA. McCairns RJS, Bernatchez L (2010) Adaptive divergence between freshwater and marine sticklebacks: insights into the role of phenotypic plasticity from an integrated analysis of candidate gene expression. Evolution, 64, 1029–1047. McCarthy MM, Arnold AP, Ball GF, Blaustein JD, De Vries GJ (2012a) Sex differences in the brain: the not so inconvenient truth. Journal of Neuroscience, 32, 2241–2247. McCarthy DJ, Chen Y, Smyth GK (2012b) Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Research, 40, 4288–4297. Mccormick MI (1998) Behaviorally induced maternal stress in a fish influences progeny quality by a hormonal mechanism. Ecology, 79, 1873–1883. McGhee KE, Bell AM (2014) Paternal care in a fish: epigenetics and fitness enhancing effects on offspring anxiety. Proceedings of the Royal Society B: Biological 150  Sciences, 281, 20141146–20141146. McGowan PO, Sasaki A, D’Alessio AC et al. (2009) Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat Neurosci, 12, 342–348. McRae AF, Powell JE, Henders AK et al. (2014) Contribution of genetic variation to transgenerational inheritance of DNA methylation. Genome Biology, 15, R73. Melamed-Bessudo C, Levy AA (2012) Deficiency in DNA methylation increases meiotic crossover rates in euchromatic but not in heterochromatic regions in Arabidopsis. Proceedings of the National Academy of Sciences of the United States of America, 109, E981–E988. Metzger DCH, Hemmer-Hansen J, Schulte PM (2016) Conserved structure and expression of hsp70 paralogs in teleost fishes. Comparative Biochemistry and Physiology - Part D: Genomics and Proteomics, 18, 10–20. Metzger DCH, Schulte PM (2016a) Maternal stress has divergent effects on gene expression patterns in the brains of male and female threespine stickleback. Proceedings of the Royal Society B: Biological Sciences, 283, 20161734. Metzger DCH, Schulte PM (2016b) Epigenomics in marine fishes. Marine Genomics, 30, 43–54. Metzger DCH, Schulte PM (2017) Persistent and plastic effects of temperature on DNA methylation across the genome of threespine stickleback (Gasterosteus aculeatus). Proceedings of the Royal Society B: Biological Sciences, 284, 20171667. Mhanni AA, McGowan RA (2004) Global changes in genomic methylation levels during early development of the zebrafish embryo. Development Genes and Evolution, 214, 412–7. Mirouze M, Lieberman-Lazarovich M, Aversano R et al. (2012) Loss of DNA methylation affects the recombination landscape in Arabidopsis. Proceedings of the National Academy of Sciences, 109, 5880–5885. Mitrovic D, Perry SF (2009) The effects of thermally induced gill remodeling on ionocyte distribution and branchial chloride fluxes in goldfish (Carassius auratus). Journal of Experimental Biology, 212, 843–852. Mommer BC, Bell AM (2013) A test of maternal programming of offspring stress response to predation risk in threespine sticklebacks. Physiology & Behavior, 122, 222–227. Monaghan P (2008) Early growth conditions, phenotypic development and environmental change. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 1635–1645. Monk M, Boubelik M, Lehnert S (1987) Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development. Development, 99, 371–82. Morán P, Marco-Rius F, Megías M, Covelo-Soto L, Pérez-Figueroa A (2013) 151  Environmental induced methylation changes associated with seawater adaptation in brown trout. Aquaculture, 392–395, 77–83. Morris MRJ, Richard R, Leder EH et al. (2014a) Gene expression plasticity evolves in response to colonization of freshwater lakes in threespine stickleback. Molecular Ecology, 23, 3226–3240. Morris MRJ, Richard R, Leder EH et al. (2014b) Gene expression plasticity evolves in response to colonization of freshwater lakes in threespine stickleback. Molecular Ecology, 23, 3226–3240. Muller HJ (1964) The relation of recombination to mutational advance. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 1, 2–9. Murgatroyd C, Patchev A V, Wu Y et al. (2009) Dynamic DNA methylation programs persistent adverse effects of early-life stress. Nature neuroscience, 12, 1559–66. Mychasiuk R, Ilnytskyy S, Kovalchuk O, Kolb B, Gibb R (2011) Intensity matters: Brain, behaviour and the epigenome of prenatally stressed rats. Neuroscience, 180, 105–110. Nakamura R, Tsukahara T, Qu W et al. (2014) Large hypomethylated domains serve as strong repressive machinery for key developmental genes in vertebrates. Development, 141, 2568–2580. Navarro-Martín L, Viñas J, Ribas L et al. (2011) DNA methylation of the gonadal aromatase (cyp19a) promoter is involved in temperature-dependent sex ratio shifts in the European sea bass. PLoS Genetics, 7, e1002447. Nesan D, Vijayan MM (2016) Maternal cortisol mediates hypothalamus-pituitary-interrenal axis development in zebrafish. Scientific Reports, 6, 22582. Nicotra AB, Segal DL, Hoyle GL et al. (2015) Adaptive plasticity and epigenetic variation in response to warming in an Alpine plant. Ecology and Evolution, 5, 634–647. Nilsson GE, Dymowska A, Stecyk JAW (2012) New insights into the plasticity of gill structure. Respiratory Physiology and Neurobiology, 184, 214–222. Noreikiene K, Herczeg G, Gonda A et al. (2015) Quantitative genetic analysis of brain size variation in sticklebacks: support for the mosaic model of brain evolution. Proceedings of the Royal Society B: Biological Sciences, 282, 20151008. Nugent BM, Bale TL (2015) The omniscient placenta: metabolic and epigenetic regulation of fetal programming. Frontiers in neuroendocrinology, 39, 28–37. Oberlander TF, Weinberg J, Papsdorf M et al. (2008) Prenatal exposure to maternal depression, neonatal methylation of human glucocorticoid receptor gene (NR3C1) and infant cortisol stress responses. Epigenetics, 3, 97–106. Ohno S (1967) Sex chromosomes and sex-linked genes (Ohno, Ed,). Springer-Verlag, New York. Oomen RA, Hutchings JA (2017) Transcriptomic responses to environmental change in fishes: insights from RNA sequencing. FACETS, 2, 610–641. 152  Orti G, Bell MA, Reimchen TE, Meyer A (1994) Global survey of mitochondrial DNA sequences in the threespine stickleback: evidence for recent migrations. Evolution, 48, 608–622. Östlund-Nilsson S, Mayer I, Huntingford FA (2007) Biology of the three-spined stickleback (S Östlund-Nilsson, I Mayer, FA Huntingford, Eds,). CRC Press, New York. Paitz RT, Bukhari SA, Bell AM (2016) Stickleback embryos use ATP-binding cassette transporters as a buffer against exposure to maternally derived cortisol. Proceedings of the Royal Society B: Biological Sciences, 283, 20152838. Paitz RT, Mommer BC, Suhr E, Bell AM (2015) Changes in the concentrations of four maternal steroids during embryonic development in the threespined stickleback (Gasterosteus aculeatus). Journal of Experimental Zoology Part A: Ecological Genetics and Physiology, 323, 422–429. Palma-Gudiel H, Córdova-Palomera A, Leza JC, Fañanás L (2015) Glucocorticoid receptor gene (NR3C1) methylation processes as mediators of early adversity in stress-related disorders causality: a critical review. Neuroscience & Biobehavioral Reviews, 55, 520–535. Park PJ, Bell MA (2010) Variation of telencephalon morphology of the threespine stickleback (Gasterosteus aculeatus) in relation to inferred ecology. Journal of Evolutionary Biology, 23, 1261–1277. Park PJ, Chase I, Bell MA (2012) Phenotypic plasticity of the threespine stickleback Gasterosteus aculeatus telencephalon in response to experience in captivity. Current Zoology, 58, 189–210. Park J, Peng Z, Zeng J et al. (2011) Comparative analyses of DNA methylation and sequence evolution using Nasonia genomes. Molecular Biology and Evolution, 28, 3345–3354. Peichel CL, Ross JA, Matson CK et al. (2004) The master sex-determination locus in threespine sticklebacks is on a nascent Y chromosome. Current Biology, 14, 1416–1424. Peichel CL, Sullivan ST, Liachko I, White MA (2017) Improvement of the threespine stickleback genome using a Hi-C-based proximity-guided assembly. Journal of Heredity, 108, 693–700. Perry SF, Fletcher C, Bailey S et al. (2012) The interactive effects of exercise and gill remodeling in goldfish (Carassius auratus). Journal of Comparative Physiology B, 182, 935–945. Pespeni MH, Garfield DA, Manier MK, Palumbi SR (2012) Genome-wide polymorphisms show unexpected targets of natural selection. Proceedings of the Royal Society B: Biological Sciences, 279, 1412–1420. Pfennig DW, Ehrenreich IM (2014) Towards a gene regulatory network perspective on phenotypic plasticity, genetic accommodation and genetic assimilation. Molecular Ecology, 23, 4438–4440. 153  Pfennig DW, Wund MA, Snell-Rood EC et al. (2010) Phenotypic plasticity’s impacts on diversification and speciation. Trends in Ecology & Evolution, 25, 459–467. Pierron F, Baillon L, Sow M, Gotreau S, Gonzalez P (2014) Effect of low-dose cadmium exposure on DNA methylation in the endangered European eel. Environmental science & technology, 48, 797–803. Piersma T, Drent J (2003) Phenotypic flexibility and the evolution of organismal design. Trends in Ecology and Evolution, 18, 228–233. Pittman K, Yúfera M, Pavlidis M et al. (2013) Fantastically plastic: fish larvae equipped for a new world. Reviews in Aquaculture, 5, S224–S267. Plongthongkum N, Diep DH, Zhang K (2014) Advances in the profiling of DNA modifications: cytosine methylation and beyond. Nature Reviews Genetics, 15, 647–661. Potok ME, Nix DA, Parnell TJ, Cairns BR (2013) Reprogramming the maternal zebrafish genome after fertilization to match the paternal methylation pattern. Cell, 153, 759–772. Préfontaine GG, Lemieux ME, Giffin W et al. (1998) Recruitment of octamer transcription factors to DNA by glucocorticoid receptor. Molecular and cellular biology, 18, 3416–3430. Primmer CR, Papakostas S, Leder EH, Davis MJ, Ragan MA (2013) Annotated genes and nonannotated genomes: cross-species use of Gene Ontology in ecology and evolution research. Molecular Ecology, 22, 3216–3241. Qu W, Hashimoto S, Shimada A et al. (2012) Genome-wide genetic variations are highly correlated with proximal DNA methylation patterns. Genome research, 22, 1419–25. Rakei A, Maali-Amiri R, Zeinali H, Ranjbar M (2016) DNA methylation and physio-biochemical analysis of chickpea in response to cold stress. Protoplasma, 253, 61–76. Ramler D, Mitteroecker P, Shama LNS, Wegner KM, Ahnelt H (2014) Nonlinear effects of temperature on body form and developmental canalization in the threespine stickleback. Journal of Evolutionary Biology, 27, 497–507. Räsänen K, Kruuk LEB (2007) Maternal effects and evolution at ecological time-scales. Functional Ecology, 21, 408–421. Razin A (1998) CpG methylation, chromatin structure and gene silencing-a three-way connection. The EMBO journal, 17, 4905–8. Rehwinkel J, Behm-Ansmant I, Gatfield D, Izaurralde E (2005) A crucial role for GW182 and the DCP1:DCP2 decapping complex in miRNA-mediated gene silencing. RNA (New York, N.Y.), 11, 1640–7. Rice WR (1984) Sex chromosomes and the evolution of sexual dimorphism. Evolution, 38, 735. Richards EJ (2006) Inherited epigenetic variation — revisiting soft inheritance. Nature 154  Reviews Genetics, 7, 395–401. Richards JG, Semple JW, Bystriansky JS, Schulte PM (2003) Na+/K+-ATPase alpha-isoform switching in gills of rainbow trout (Oncorhynchus mykiss) during salinity transfer. The Journal of experimental biology, 206, 4475–86. Richardson B (2003) Impact of aging on DNA methylation. Ageing Research Reviews, 2, 245–261. Riddle NC, Richards EJ (2002) The control of natural variation in cytosine methylation in Arabidopsis. Genetics, 162, 355–363. Rinn JL, Snyder M (2005) Sexual dimorphism in mammalian gene expression. Trends in Genetics, 21, 298–305. Risso D, Ngai J, Speed TP, Dudoit S (2014) Normalization of RNA-seq data using factor analysis of control genes or samples (RUVSeq). Nature Biotechnology, 32, 896–902. Robinson MD, McCarthy DJ, Smyth GK (2010) edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26, 139–140. Ross JA, Peichel CL (2008) Molecular cytogenetic evidence of rearrangements on the Y chromosome of the threespine stickleback fish. Genetics, 179, 2173–2182. Ross JA, Urton JR, Boland J, Shapiro MD, Peichel CL (2009) Turnover of sex chromosomes in the stickleback fishes (Gasterosteidae). PLoS Genetics, 5, e1000391. Sætrom P, Snøve O, Rossi JJ (2007) Epigenetics and MicroRNAs. Pediatric Research, 61, 17R–23R. Saimoto RK (1993) Life history of marine threespine stickleback in Oyster Lagoon, British Columbia. The University of British Columbia. Salinas S, Brown SC, Mangel M, Munch SB (2013) Non-genetic inheritance and changing environments. Non-Genetic Inheritance, 1. Salinas S, Munch SB (2012) Thermal legacies: transgenerational effects of temperature on growth in a vertebrate. Ecology Letters, 15, 159–163. Samuk K, Iritani D, Schluter D (2014) Reversed brain size sexual dimorphism accompanies loss of parental care in white sticklebacks. Ecology and Evolution, 4, 3236–3243. Santos EM, Kille P, Workman VL, Paull GC, Tyler CR (2008) Sexually dimorphic gene expression in the brains of mature zebrafish. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 149, 314–324. Schaefer S, Nadeau JH (2015) The genetics of epigentic inheritance: modes, molecules, and mechanisms. The Quarterly Review of Biology, 90, 381–415. Schaefer J, Ryan A (2006) Developmental plasticity in the thermal tolerance of zebrafish Danio rerio. Journal of Fish Biology, 69, 722–734. 155  Schapira AHV (1999) Mitochondrial involvement in Parkinson’s disease, Huntington’s disease, hereditary spastic paraplegia and Friedreich’s ataxia. Biochimica et Biophysica Acta (BBA) - Bioenergetics, 1410, 159–170. Schield DR, Walsh MR, Card DC et al. (2016) EpiRADseq: scalable analysis of genomewide patterns of methylation using next-generation sequencing (M Bunce, Ed,). Methods in Ecology and Evolution, 7, 60–69. Schlichting CD (2008) Hidden reaction norms, cryptic genetic variation, and evolvability. Annals of the New York Academy of Sciences, 1133, 187–203. Schlichting CD, Smith H (2002) Phenotypic plasticity: linking molecular mechanisms with evolutionary outcomes. Evolutionary Ecology, 16, 189–211. Schmidt KL, Kubli SP, MacDougall-Shackleton EA, MacDougall-Shackleton SA (2015) Early-life stress has sex-specific effects on immune function in adult song sparrows. Physiological and Biochemical Zoology, 88, 183–194. Schmidt KL, MacDougall-Shackleton EA, MacDougall-Shackleton SA (2012) Developmental stress has sex-specific effects on nestling growth and adult metabolic rates but no effect on adult body size or body composition in song sparrows. Journal of Experimental Biology, 215, 3207–3217. Schmidt KL, MacDougall-Shackleton EA, Soma KK, MacDougall-Shackleton SA (2014) Developmental programming of the HPA and HPG axes by early-life stress in male and female song sparrows. General and Comparative Endocrinology, 196, 72–80. Schneider RF, Li Y, Meyer A, Gunter HM (2014) Regulatory gene networks that shape the development of adaptive phenotypic plasticity in a cichlid fish. Molecular Ecology, 23, 4511–4526. Schneider RF, Meyer A (2017) How plasticity, genetic assimilation and cryptic genetic variation may contribute to adaptive radiations. Molecular Ecology, 26, 330–350. Schnurr ME, Yin Y, Scott GR (2014) Temperature during embryonic development has persistent effects on thermal acclimation capacity in zebrafish. The Journal of experimental biology, 217, 1370–1380. Schulte PM (2011) Effects of temperature: an introduction. In: Encyclopedia of Fish Physiology, pp. 1688–1694. Elsevier. Schulte PM, Healy TM, Fangue NA (2011) Thermal performance curves, phenotypic plasticity, and the time scales of temperature exposure. Integrative and Comparative Biology, 51, 691–702. Schultheiß R, Viitaniemi HM, Leder EH (2015) Spatial dynamics of evolving dosage compensation in a young sex chromosome system. Genome biology and evolution, 7, 581–90. Scott GR, Johnston IA (2012) Temperature during embryonic development has persistent effects on thermal acclimation capacity in zebrafish. Proceedings of the National Academy of Sciences, 109, 14247–14252. Seebacher F, White CR, Franklin CE (2015) Physiological plasticity increases resilience 156  of ectothermic animals to climate change. Nature Climate Change, 5, 61–66. Shah S, McRae AF, Marioni RE et al. (2014) Genetic and environmental exposures constrain epigenetic drift over the human life course. Genome Research, 24, 1725–1733. Shama LNS (2015) Bet hedging in a warming ocean: predictability of maternal environment shapes offspring size variation in marine sticklebacks. Global Change Biology, 21, 4387–4400. Shama LNS, Mark FC, Strobel A et al. (2016) Transgenerational effects persist down the maternal line in marine sticklebacks: gene expression matches physiology in a warming ocean. Evolutionary Applications, 9, 1096–1111. Shama LNS, Strobel A, Mark FC, Wegner KM (2014) Transgenerational plasticity in marine sticklebacks: maternal effects mediate impacts of a warming ocean. Functional Ecology, 28, 1482–1493. Shama LNS, Wegner KM (2014) Grandparental effects in marine sticklebacks: transgenerational plasticity across multiple generations. Journal of Evolutionary Biology, 27, 2297–2307. Shao C, Li Q, Chen S et al. (2014) Epigenetic modification and inheritance in sexual reversal of fish. Genome Research, 24, 604–615. Sharif J, Muto M, Takebayashi S et al. (2007) The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature, 450, 908–912. Sharma E, Künstner A, Fraser BA et al. (2014) Transcriptome assemblies for studying sex-biased gene expression in the guppy, Poecilia reticulata. BMC Genomics, 15, 400. Shaw KA, Scotti ML, Foster SA (2007) Ancestral plasticity and the evolutionary diversification of courtship behaviour in threespine sticklebacks. Animal Behaviour, 73, 415–422. Shen J cheng, Rideout WM, Jones PA (1994) The rate of hydrolytic deamination of 5-methylcytosine in double-stranded DNA. Nucleic Acids Research, 22, 972–976. Shen L, Song C-X, He C, Zhang Y (2014) Mechanism and function of oxidative reversal of DNA and RNA methylation. Annual Review of Biochemistry, 83, 585–614. Shen L, Waterland RA (2007) Methods of DNA methylation analysis. Curr Opin Clin Nutr Metab Care, 10, 576–81. Sheriff MJ, Bell A, Boonstra R et al. (2017) Integrating ecological and evolutionary context in the study of maternal stress. Integrative and Comparative Biology, 57, 437–449. Sheriff MJ, Love OP (2013) Determining the adaptive potential of maternal stress. Ecology letters, 16, 271–80. Shimada Y, Shikano T, Merilä J (2011) A high incidence of selection on physiologically important genes in the three-spined stickleback, gasterosteus aculeatus. Molecular 157  Biology and Evolution, 28, 181–193. Shintani Y, Ishikawa Y (2007) Relationship between rapid cold-hardening and cold acclimation in the eggs of the yellow-spotted longicorn beetle, Psacothea hilaris. Journal of Insect Physiology, 53, 1055–1062. Si Y, Ding Y, He F et al. (2016) DNA methylation level of cyp19a1a and Foxl2 gene related to their expression patterns and reproduction traits during ovary development stages of Japanese flounder (Paralichthys olivaceus). Gene, 575, 321–330. Smith G, Smith C, Kenny JG, Chaudhuri RR, Ritchie MG (2015) Genome-wide DNA methylation patterns in wild samples of two morphotypes of threespine stickleback (Gasterosteus aculeatus). Molecular Biology and Evolution, 32, 888–895. Sollid J (2005) Temperature alters the respiratory surface area of crucian carp Carassius carassius and goldfish Carassius auratus. Journal of Experimental Biology, 208, 1109–1116. Somero GN (2005) Linking biogeography to physiology: evolutionary and acclimatory adjustments of thermal limits. Frontiers in zoology, 2, 1–9. Somero GN (2010) The physiology of climate change: how potentials for acclimatization and genetic adaptation will determine “winners” and “losers.” Journal of Experimental Biology, 213, 912–920. Somero GN (2012) The physiology of global change: linking patterns to mechanisms. Annual Review of Marine Science, 4, 39–61. Song Y, Liu L, Feng Y et al. (2015) Chilling- and freezing- induced alterations in cytosine methylation and its association with the cold tolerance of an Alpine aubnival plant, Chorispora bungeana. PLoS ONE, 10, e0135485. Sopinka NM, Jeffrey JD, Burnett NJ et al. (2017) Maternal programming of offspring hypothalamic–pituitary–interrenal axis in wild sockeye salmon (Oncorhynchus nerka). General and Comparative Endocrinology, 242, 30–37. Sreenivasan R, Cai M, Bartfai R et al. (2008) Transcriptomic analyses reveal novel genes with sexually dimorphic expression in the zebrafish gonad and brain. PLoS ONE, 3, e1791. Stein LR, Bell AM (2014) Paternal programming in sticklebacks. Animal Behaviour, 95, 165–171. Stillman JH (2002) Causes and consequences of thermal tolerance limits in rockey intertidal porcelain crabs, genus Petrolisthes. Integrative and Comparative Biology, 42, 790–796. Stockwell PA, Chatterjee A, Rodger EJ, Morison IM (2014) DMAP: differential methylation analysis package for RRBS and WGBS data. Bioinformatics, 30, 1814–1822. Storm JJ, Lima SL (2010) Mothers forewarn offspring about predators: a transgenerational maternal effect on behavior. The American naturalist, 175, 382–90. 158  Streisinger G, Walker C, Dower N, Knauber D, Singer F (1981) Production of clones of homozygous diploid zebra fish (Brachydania rerio). Nature, 291, 293–296. Suetake I, Shinozaki F, Miyagawa J, Takeshima H, Tajima S (2004) DNMT3L stimulates the DNA methylation activity of Dnmt3a and Dnmt3b through a direct Interaction. Journal of Biological Chemistry, 279, 27816–27823. Sunday JM, Bates AE, Dulvy NK (2011) Global analysis of thermal tolerance and latitude in ectotherms. Proceedings of the Royal Society B: Biological Sciences, 278, 1823–1830. Sunday JM, Bates AE, Dulvy NK (2012) Thermal tolerance and the global redistribution of animals. Nature Climate Change, 2, 686–690. Sved J, Bird A (1990) The expected equilibrium of the CpG dinucleotide in vertebrate genomes under a mutation model. Proceedings of the National Academy of Sciences, 87, 4692–4696. Szyf M, Weaver ICG, Champagne FA, Diorio J, Meaney MJ (2005) Maternal programming of steroid receptor expression and phenotype through DNA methylation in the rat. Frontiers in Neuroendocrinology, 26, 139–162. Takai D, Jones PA (2002) Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proceedings of the National Academy of Sciences of the United States of America, 99, 3740–5. Talkowski ME, Mullegama S V., Rosenfeld JA et al. (2011) Assessment of 2q23.1 microdeletion syndrome implicates MBD5 as a single causal locus of intellectual disability, epilepsy, and autism spectrum disorder. American Journal of Human Genetics, 89, 551–563. Talkowski ME, Rosenfeld JA, Blumenthal I et al. (2012) Sequencing chromosomal abnormalities reveals neurodevelopmental loci that confer risk across diagnostic boundaries. Cell, 149, 525–537. Taugbøl A, Arntsen T, Østbye K, Vøllestad LA (2014) Small changes in gene expression of targeted osmoregulatory genes when exposing marine and freshwater threespine stickleback (Gasterosteus aculeatus) to abrupt salinity transfers. PLoS ONE, 9, e106894. Taylor EB, McPhail JD (1999) Evolutionary history of an adaptive radiation in species pairs of threespine sticklebacks (Gasterosteus): insights from mitochrondrial DNA. Biological Journal of the Linnean Society, 66, 271–291. Teets NM, Denlinger DL (2013) Physiological mechanisms of seasonal and rapid cold-hardening in insects. Physiological Entomology, 38, 105–116. Teigen LE, Orczewska JI, McLaughlin J, O’Brien KM (2015) Cold acclimation increases levels of some heat shock protein and sirtuin isoforms in threespine stickleback. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 188, 139–147. Thornton PK, Ericksen PJ, Herrero M, Challinor AJ (2014) Climate variability and 159  vulnerability to climate change: a review. Global Change Biology, 20, 3313–3328. Timme-Laragy AR, Meyer JN, Waterland RA, Di Giulio RT (2005) Analysis of CpG methylation in the killifish CYP1A promoter. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, 141, 406–411. Toli E-A, Calboli FCF, Shikano T, Merilä J (2016) A universal and reliable assay for molecular sex identification of three-spined sticklebacks (Gasterosteus aculeatus). Molecular Ecology Resources, 16, 1389–1400. Traut W, Winking H (2001) Meiotic chromosomes and stages of sex chromosome evolution in fish: Zebrafish, platyfish and guppy. Chromosome Research, 9, 659–672. Trucchi E, Mazzarella AB, Gilfillan GD et al. (2016) BsRADseq: screening DNA methylation in natural populations of non-model species. Molecular Ecology, 25, 1697–1713. Tyler CR, Sumpter JP (1996) Oocyte growth and development in teleosts. Reviews in Fish Biology and Fisheries, 6, 287–318. Umer M, Herceg Z (2013) Deciphering the epigenetic code: an overview of DNA methylation analysis methods. Antioxidants & Redox Signaling, 18, 1972–1986. Varriale A (2014) DNA methylation, epigenetics, and evolution in vertebrates: facts and challenges. International Journal of Evolutionary Biology, 2014, 1–7. Varriale A, Bernardi G (2006) DNA methylation and body temperature in fishes. Gene, 385, 111–21. Vaughn MW, Tanurdžić M, Lippman Z et al. (2007) Epigenetic natural variation in Arabidopsis thaliana. PLoS Biology, 5, e174. Veilleux HD, Ryu T, Donelson JM et al. (2015) Molecular processes of transgenerational acclimation to a warming ocean. , 5. Verhoeven KJF, Jansen JJ, van Dijk PJ, Biere A (2010) Stress-induced DNA methylation changes and their heritability in asexual dandelions. New Phytologist, 185, 1108–1118. Verhoeven KJF, VonHoldt BM, Sork VL (2016) Epigenetics in ecology and evolution: what we know and what we need to know. Molecular Ecology, 25, 1631–1638. Véron N, Peters AHFM (2011) Epigenetics: Tet proteins in the limelight. Nature, 473, 293–294. Vidalis A, Živković D, Wardenaar R et al. (2016) Methylome evolution in plants. Genome Biology, 17, 264. Viggiano L, de Pinto MC (2017) Dynamic DNA methylation patterns in stress response. In: Plant Epigenetics (eds Rajewsky N, Jurga S, Barciszewski J), pp. 281–302. Springer International Publishing, Cham. Waddington CH (1942) The epigenotype. Endeavour, 18–20. Wade J, Peabody C, Coussens P et al. (2004) A cDNA microarray from the 160  telencephalon of juvenile male and female zebra finches. Journal of Neuroscience Methods, 138, 199–206. Wan ZY, Xia JH, Lin G et al. (2016) Genome-wide methylation analysis identified sexually dimorphic methylated regions in hybrid tilapia. Scientific Reports, 6, 35903. Wang HQ, Tuominen LK, Tsai CJ (2011) SLIM: A sliding linear model for estimating the proportion of true null hypotheses in datasets with dependence structures. Bioinformatics, 27, 225–231. Wang Y, Wang C, Zhang J, Chen Y, Zuo Z (2009) DNA hypomethylation induced by tributyltin, triphenyltin, and a mixture of these in Sebastiscus marmoratus liver. Aquatic Toxicology, 95, 93–98. Wang G, Yang E, Smith KJ et al. (2014) Gene expression responses of threespine stickleback to salinity: implications for salt-sensitive hypertension. Frontiers in genetics, 5, 312. Warton DI, Hui FKC (2011) The arcsine is asinine: the analysis of proportions in ecology. Ecology, 92, 3–10. Weaver ICG, Cervoni N, Champagne FA et al. (2004) Epigenetic programming by maternal behavior. Nature neuroscience, 7, 847–54. Weaver ICG, D’Alessio AC, Brown SE et al. (2007) The transcription factor nerve growth factor-inducible protein a mediates epigenetic programming: altering epigenetic marks by immediate-early genes. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 1756–68. Weinstock M (2007) Gender differences in the effects of prenatal stress on brain development and behaviour. Neurochemical Research, 32, 1730–1740. Wen AY, You F, Sun P et al. (2014) CpG methylation of dmrt1 and cyp19a promoters in relation to their sexual dimorphic expression in the Japanese flounder Paralichthys olivaceus. Journal of Fish Biology, 84, 193–205. West-Eberhard MJ (2005) Developmental plasticity and the origin of species differences. Proceedings of the National Academy of Sciences, 102, 6543–6549. White MA, Kitano J, Peichel CL (2015) Purifying selection maintains dosage-sensitive genes during degeneration of the threespine stickleback Y chromosome. Molecular Biology and Evolution, 32, 1981–1995. Windahl SH, Treuter E, Ford J et al. (1999) The nuclear-receptor interacting protein (RIP) 140 binds to the human glucocorticoid receptor and modulates hormone-dependent transactivation. The Journal of steroid biochemistry and molecular biology, 71, 93–102. Wissink S, van Heerde EC, Schmitz ML et al. (1997) Distinct domains of the RelA NF-kappaB subunit are required for negative cross-talk and direct interaction with the glucocorticoid receptor. The Journal of biological chemistry, 272, 22278–84. Wolff GL, Kodell RL, Moore SR, Cooney CA (1998) Maternal epigenetics and methyl 161  supplements affect agouti gene expression in Avy/a mice. The FASEB Journal, 12, 949–57. Wootton RJ (1984) A functional biology of sticklebacks. University of California Press. Wreczycka K, Gosdschan A, Yusuf D et al. (2017) Strategies for analyzing bisulfite sequencing data. bioRxiv. Wright AE, Dean R, Zimmer F, Mank JE (2016) How to make a sex chromosome. Nature Communications, 7, 1–8. Wu C -t., Morris J (2001) Genes, genetics, and epigenetics: a correspondence. Science, 293, 1103–1105. Wu H, Wang C, Wu Z (2015) PROPER: Comprehensive power evaluation for differential expression using RNA-seq. Bioinformatics, 31, 233–241. Wu S-F, Zhang H, Hammoud SS et al. (2011) DNA methylation profiling in zebrafish. In: Methods in cell biology, pp. 327–339. Elsevier Inc. Xie W, Schultz MD, Lister R et al. (2013) Epigenomic analysis of multilineage differentiation of human embryonic stem cells. Cell, 153, 1134–1148. Yamada M, Higuchi M, Goto A (2001) Extensive introgression of mitochondrial DNA found between two genetically divergent forms of threespine stickleback, Gasterosteus aculeatus, around Japan. Environmental Biology of Fishes, 61, 269–284. Yelina NE, Lambing C, Hardcastle TJ et al. (2015) DNA methylation epigenetically silences crossover hot spots and controls chromosomal domains of meiotic recombination in Arabidopsis. Genes and Development, 29, 2183–2202. Yokomine T, Hata K, Tsudzuki M, Sasaki H (2006) Evolution of the vertebrate DNMT3 gene family: a possible link between existence of DNMT3L and genomic imprinting. Cytogenetic and Genome Research, 113, 75–80. Young MD, Wakefield MJ, Smyth GK, Oshlack A (2010) Gene ontology analysis for RNA-seq: accounting for selection bias. Genome biology, 11, R14. Zemach A, McDaniel IE, Silva P, Zilberman D (2010) Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science, 328, 916–919. Zerbino DR, Achuthan P, Akanni W et al. (2018) Ensembl 2018. Nucleic Acids Research, 46, D754–D761. Zhang YY, Fischer M, Colot V, Bossdorf O (2013) Epigenetic variation creates potential for evolution of plant phenotypic plasticity. New Phytologist, 197, 314–322. Zhang Y, Liu H, Lv J et al. (2011) QDMR: a quantitative method for identification of differentially methylated regions by entropy. Nucleic acids research, 39, e58. Zhang W, Wang X, Yu Q, Ming R, Jiang J (2008) DNA methylation and heterochromatinization in the male-specific region of the primitive Y chromosome of papaya. Genome Research, 18, 1938–1943. Zhang X, Wen H, Wang H et al. (2017) RNA-Seq analysis of salinity stress–responsive 162  transcriptome in the liver of spotted sea bass (Lateolabrax maculatus). PLoS ONE, 12, e0173238. Zhou X, Lindsay H, Robinson MD (2014) Robustly detecting differential expression in RNA sequencing data using observation weights. Nucleic Acids Research, 42, 1–10. Zhu LJ, Gazin C, Lawson ND et al. (2010) ChIPpeakAnno: a Bioconductor package to annotate ChIP-seq and ChIP-chip data. BMC Bioinformatics, 11, 237. Ziller MJ, Hansen KD, Meissner A, Aryee MJ (2015) Coverage recommendations for methylation analysis by whole-genome bisulfite sequencing. Nature Methods, 12, 230–232. Zucchi FCR, Yao Y, Ward ID et al. (2013) Maternal stress induces epigenetic signatures of psychiatric and neurological diseases in the offspring. PLoS ONE, 8, e56967.  163  Appendices  Appendix A  Supporting information for chapter two  Table A.1: Sample sizes from each family in each developmental treatment   Family 1 Family 2 Family 3 Family 4 Family 5 Family 6 12 °C Development 2 0 1 2 1 0 18 °C Development 1 1 0 3 0 1 24 °C Development 2 1 1 2 0 0   Table A.2: Genomic mean and median percent methylation Treatment Mean genomic percent DNA methylation Standard deviation                                   Median genomic percent DNA methylation Standard deviation 12 °C development, 18 °C acclimation 73.66% 2.21% 82.62% 3.68% 24 °C development, 18 °C acclimation 72.31% 0.89% 80.30% 0.74% 18 °C development, 18 °C acclimation 72.03% 2.20% 80.98% 3.32% 18 °C development, 5 °C acclimation 73.62% 1.24% 83.38% 2.07% 18 °C development, 25 °C acclimation 73.96% 1.48% 84.12% 3.56% n=6 	 	 	 	  Table A.3: Kolmogorov-Smirnov test statistics    p-value D statistic 12 °C Development, 18 °C acclimation 2.2 E-16 0.02336 24 °C Development, 18 °C acclimation 2.2 E-16 0.042084 18 °C Development, 5 °C acclimation 2.2 E-16 0.05126 18 °C Development, 25 °C acclimation 2.2 E-16 0.081229              164  Table A.4: Difference in percent DNA methylation level for DMRs common to all treatments                                         DMR position 12 °C Development 24 °C Devlopment 5 °C Acclimation 25 °C Acclimation chr start end              qvalue DiffMeth qvalue DiffMeth qvalue DiffMeth qvalue DiffMeth groupI 9551801 9551900 3.52E-02 21 1.29E-02 23 9.78E-04 26 1.03E-02 24 groupI 24585101 24585200 1.23E-05 -16 6.34E-05 -14 3.24E-05 -15 9.61E-04 -13 groupI 27994101 27994200 7.90E-05 20 4.49E-03 17 2.48E-08 28 6.72E-03 17 groupII 18989201 18989300 8.13E-06 15 4.91E-08 18 3.00E-16 24 3.31E-04 16 groupII 21905101 21905200 1.44E-02 10 5.17E-03 11 4.30E-06 15 5.37E-03 11 groupIII 879601 879700 5.51E-09 23 3.10E-06 19 4.03E-03 12 1.15E-15 31 groupIII 9626701 9626800 2.23E-67 32 3.89E-07 12 1.18E-18 19 2.31E-30 25 groupIV 132901 133000 1.57E-03 14 3.36E-05 16 7.14E-03 11 9.56E-03 13 groupIV 2314301 2314400 2.43E-15 -10 8.36E-17 -10 9.82E-30 12 3.58E-34 -15 groupIV 2758901 2759000 3.15E-02 14 3.08E-02 15 3.91E-03 18 2.05E-03 19 groupIV 11319801 11319900 4.00E-04 -14 3.95E-03 -11 4.07E-04 -13 4.39E-06 -17 groupIV 12200201 12200300 1.07E-03 17 4.32E-03 16 6.87E-06 22 8.46E-04 19 groupIV 29800001 29800100 1.03E-06 15 5.09E-08 18 7.18E-19 26 6.79E-13 28 groupIX 353601 353700 9.20E-06 -13 4.96E-04 -11 4.24E-14 -22 4.86E-11 -18 groupIX 8238301 8238400 7.43E-03 -15 2.42E-04 -19 2.11E-03 -15 6.51E-04 -17 groupIX 17934101 17934200 4.70E-02 13 6.68E-03 16 5.86E-03 15 2.00E-02 16 groupV 9288901 9289000 3.64E-04 -14 1.34E-02 -11 6.73E-10 -25 1.27E-02 -10 groupVI 1530101 1530200 4.31E-08 26 5.66E-03 16 1.97E-11 27 8.73E-03 14 groupVI 4132101 4132200 1.01E-37 13 1.80E-12 -12 1.11E-52 -27 3.86E-37 13 groupVII 5901801 5901900 4.96E-11 -20 3.75E-18 -27 3.17E-51 -55 9.13E-06 -12 groupVII 15623001 15623100 4.60E-06 21 9.36E-03 14 3.56E-04 17 6.62E-04 17 groupVIII 351301 351400 3.04E-02 12 1.23E-03 18 1.08E-08 27 1.33E-05 27 groupX 97101 97200 9.08E-10 36 9.73E-05 22 2.46E-15 37 9.17E-04 23 groupX 2872701 2872800 4.20E-06 -16 2.29E-04 -12 6.29E-05 -13 8.74E-04 -11 groupXI 2274501 2274600 8.61E-05 20 1.22E-02 15 6.30E-07 24 2.93E-06 24 groupXII 7504001 7504100 7.04E-04 14 7.12E-04 14 3.86E-06 18 1.42E-02 11 groupXIII 8471801 8471900 9.75E-06 32 1.38E-04 30 4.85E-09 37 1.28E-03 24 groupXIV 3410601 3410700 2.65E-02 -14 4.41E-02 -13 3.38E-02 -13 2.62E-02 -13 groupXIV 3943001 3943100 3.40E-02 -14 1.56E-02 -14 4.76E-05 -23 8.94E-03 -12 groupXIV 10441601 10441700 2.17E-02 -14 1.60E-04 -20 7.08E-07 -25 7.12E-03 -16 groupXIX 8793401 8793500 8.41E-08 22 2.15E-06 17 1.08E-09 23 1.44E-03 14 groupXIX 10454801 10454900 3.29E-03 -13 1.49E-07 -21 2.58E-10 -23 5.99E-03 -12 groupXIX 15351001 15351100 2.35E-03 15 6.51E-03 13 1.44E-04 18 2.20E-03 15 groupXV 1970801 1970900 1.34E-04 -21 5.07E-06 -25 9.26E-04 -17 1.54E-03 -16 groupXV 4546301 4546400 4.46E-187 36 1.83E-100 27 1.56E-52 19 4.13E-109 29 groupXVI 8645601 8645700 5.14E-03 23 2.09E-02 19 5.51E-03 20 3.47E-02 17 groupXVI 17985101 17985200 2.93E-06 32 4.05E-04 26 6.06E-08 38 1.18E-07 37 groupXVI 17985201 17985300 8.29E-05 23 1.01E-02 16 4.03E-06 26 1.54E-08 31 groupXVII 5773201 5773300 8.78E-03 -14 1.04E-02 -12 9.66E-11 -27 4.55E-02 -12 groupXVII 11445901 11446000 8.57E-04 16 1.13E-03 16 2.00E-05 20 7.49E-03 14 groupXVIII 7284901 7285000 3.63E-02 16 4.98E-02 13 3.09E-04 23 3.65E-05 24 groupXVIII 15625401 15625500 1.20E-12 26 3.85E-06 18 2.37E-04 17 1.06E-04 18 groupXXI 5934501 5934600 5.07E-06 -28 1.44E-05 -30 5.01E-06 -27 1.62E-04 -24 groupXXI 8811701 8811800 7.07E-04 12 2.27E-03 10 1.32E-04 13 1.05E-12 30 scaffold_1331 2901 3000 3.73E-05 -20 9.70E-05 -19 2.79E-07 -22 3.76E-07 -25 scaffold_1498 5201 5300 1.62E-07 11 4.63E-06 10 1.50E-13 16 7.16E-11 15 scaffold_167 70001 70100 4.98E-12 -25 2.32E-03 -15 3.08E-13 -24 3.31E-10 -24 scaffold_235 12101 12200 4.14E-03 17 8.66E-03 17 1.11E-05 23 1.07E-03 19 scaffold_515 5601 5700 5.90E-16 16 2.25E-13 13 5.99E-45 22 6.37E-35 18 scaffold_990 401 500 2.99E-03 22 4.79E-02 13 6.34E-08 32 2.42E-03 19   165   Figure A.1: The effects of developmental temperature on time to hatch of threespine stickleback embryos. Histograms represent the number of threespine stickleback embryos that hatched on a particular day post fertilization when reared at 12 °C (A; white), 18 °C (B; grey), or 24 °C (C; black) . Day 0 represents the day of fertilizaiton.  0 5 10 15 20 25020406080Number of hatched eggs0 5 10 15 20 25020406080Days post fertilization0 5 10 15 20 25020406080Days post fertilizationNumber of eggs hatchedA	B	C	166   Figure A.2: Genomic distributions of (A) significantly hypermethylated and (B) significantly hypomethylated cytosines for each treatment. Different colors correspond to the different treatments with 12 °C development as light blue, 5 °C acclimation as dark blue, 25 °C acclimation as red, 24 °C development as pink. The background distribution of the sequenced CpG loci is in grey and is replicated in panels A and B.              A	 B	%	hyper	methylated	DMCs	%	hypo	methylated	DMCs	167  Appendix B  Supporting information for chapter three  Table B.1: RNA-sequencing library sizes                                    Sample ID Developmental Temperature (°C) Acclimation Temperature  (°C) Sequencing library size (# of reads) TSS1 12 18 31366410 TSS2 12 18 36206008 TSS3 12 18 41908012 TSS4 12 18 32977256 TSS5 12 18 35373834 TSS6 12 18 44243698 TSS7 18 18 32459910 TSS8 18 18 41973418 TSS9 18 18 37587956 TSS10 18 18 40668456 TSS11 18 18 45922122 TSS12 18 18 36567744 TSS13 24 18 39881784 TSS14 24 18 40313224 TSS15 24 18 41374452 TSS16 24 18 37452680 TSS17 24 18 36030292 TSS18 24 18 31786428 TSS19 18 5 46315050 TSS20 18 5 41645694 TSS21 18 5 28716160 TSS22 18 5 36478166 TSS23 18 5 35062444 TSS24 18 5 42850618 TSS25 18 25 49687968 TSS26 18 25 43084226 TSS27 18 25 35147542 TSS28 18 25 33675164 TSS29 18 25 34663342 TSS30 18 25 40017684 TSS31 18 18 57013628 TSS32 18 18 63133950 168    Figure B.1: Hatching success of threespine stickleback reared at 12, 18, and 24 °C. Hatching success is presented as the proportion of fertilized eggs that hatched per family (n=6 families/temperature). The lines in each box plot indicates the median proportion of hatched individuals for each temperature, the box defines the interquartile range (IQR), and the whiskers represent the maximum and minimum values, excluding values greater than 1.5x IQR (which are shown as individual points). There were no significant differences in hatching success among the temperature treatments (one-way ANOVA p>0.05).  12 18 240.00.20.40.60.81.0Developmental temperature (°C)Proportion of eggs hatched169   Figure B.2: Length (A/B) and weight (C/D) of nine-month-old threespine stickleback exposed to different temperatures during development until hatch and then kept at 18 °C until nine months old (A/B) or developed at 18 °C to eight months old and then acclimated to different temperatures for four weeks as adults (B/D). The lines in each box plot indicate the median value for each group, the box defines the interquartile range (IQR), and the whiskers represent the maximum and minimum values, excluding values greater than 1.5x IQR (which are shown as individual points). A significant effect of developmental temperature on wet weight was detected by a one-way ANOVA analysis (p <0.05). An asterisk (*) indicates a significant difference in mean value from the other two treatments based on a Tukey’s post hoc analysis (p < 0.05).        12 18 24303540455055Length (mm)Developmental Temperature (°C)A5 18 25303540455055Length (mm)Acclimation Temperature (°C)B12 18 240.00.51.01.5Weight (g)Developmental Temperature (°C)C*5 18 250.00.51.01.5Weight (g)Acclimation Temperature (°C)D170  Appendix C  Supporting information for chapter four C.1 Effects of environmental salinity on fertilization and hatching success  Salinity had a significant effect on both fertilization (p < 2.2 x 10-16) and hatching success (p < 2.2 x 10-16). Percent fertilization was highest at 2 ppt, intermediate at salinities from 7 ppt to 21 ppt, and lowest at 28 and 35 ppt (Figure S1). Similarly, the percent of fertilized embryos that hatched was highest at salinities from 2 to 14 ppt, intermediate at 21 ppt, and lowest at 28 and 35 ppt. These data indicate that fertilization and hatching success the “fully plated” marine stickleback ecotype used in this study are optimal at low and brackish salinities and that higher salinities closer to full strength sea water represent a stressful reproductive environment and early developmental environment. Therefore, for subsequent studies of the epigenome we selected the salinities of 2 ppt and 21 ppt as representative of salinity extremes that do not impose excessive mortality in this population.   C.2 Characterizing family effects on DNA methylation patterns in stickleback  Principle component analysis of genome-wide DNA methylation patterns in stickleback clearly separated the samples into three distinct groups based on family, and the first three principle components accounted for 31.73 % of the variation in the dataset (PC1 = 11.03 %, PC2 = 10.65 %, PC3 = 10.05 %; Figure S2). Hierarchical cluster analysis also revealed three major groupings corresponding to the three separate families used in this study. Within each stickleback family individuals clustered either by sex (two of three families) or by environmental salinity (one family, Figure S3). Together, these 171  data indicate that family has a larger effect on inter-individual variation in DNA methylation levels than either sex or environmental salinity.   C.3 Supporting figures and tables for chapter four  Figure C.1: Percent fertilization and hatching at different salinities. A) Percent of embryos that were successfully fertilized and B) percent of fertilized embryos that successfully hatched at salinities 2, 7, 14, 21, 28, or 35 ppt. Different letters indicate a significant difference between two groups detected using a Tukey test (q ≤ 0.05).  172   Figure C.2: Principle component analysis of DNA methylation levels. 3D plot of the first three principal components (PC) of the variation in DNA methylation levels among all twelve individuals. PC1 = 11.03 %, PC2 = 10.65 %, and PC3 = 10.05 % of the variation in the data. Shapes represent families. Colors represent rearing salinity and sex. Dark red = females reared at 21 ppt, light red = females reared at 2 ppt, dark blue = males reared at 21 ppt, light blue = males reared at 2 ppt.   173   Figure C.3: Cluster analysis of whole genome bisulfite sequencing data of twelve individual threespine stickleback. Each family consists of two females (F) and two males (M) reared at an environmental salinity of either 2 ppt or 21 ppt.                   0.00.10.20.30.40.50.6CpG methylation clusteringDistance method: "correlation"; Clustering method: "ward"SamplesHeightF3 M3M6 F6 M5M2 F2 F5 F4 F1 M1M4Family	3	 Family	2	 Family	1	F21	F2	 M2	 M21	 F21	F2	M2	M21	 F21	 F2	 M2	 M21	0.6	0.5	0.4	0.3	0.2	0.1	0.0	Height	174    Figure C.4: Chromosomal DNA methylation levels for each individual. Each point represents the mean DNA methylation level in a 10 kb window for an individual stickleback that were either male stickleback reared at 2 ppt (light blue) or 21 ppt (dark blue) or female stickleback reared at 2 ppt (light red) or 21 ppt (dark red). Solid lines represent the smooth spline fit for each individual.                     175   Table C.1: DNA hypomethylated canyon gene annotations    Table C.2: Differentially methylated cytosine distributions Ensembl Gene ID Chromosome Gene Name Ensembl Gene ID Chromosome Gene NameENSGACG00000017891 chr4 trpc7 ENSGACG00000005622 chr11 si:dkeyp-101e12.1ENSGACG00000017895 chr4 not annotated ENSGACG00000005623 chr11 not annotatedENSGACG00000017896 chr4 diaph7 ENSGACG00000005626 chr11 not annotatedENSGACG00000017901 chr4 pcdhgc5 (1 of many) ENSGACG00000005628 chr11 hoxb4aENSGACG00000017905 chr4 pcdhgc5 (1 of many) ENSGACG00000005631 chr11 hoxa3ENSGACG00000017907 chr4 not annotated ENSGACG00000005633 chr11 hoxb2aENSGACG00000017910 chr4 not annotated ENSGACG00000005635 chr11 not annotatedENSGACG00000017911 chr4 not annotated ENSGACG00000021289 chr11 not annotatedENSGACG00000017912 chr4 not annotated ENSGACG00000022266 chr11 not annotatedENSGACG00000017913 chr4 pcdh2g17 ENSGACG00000009377 chr12 calcoco1aENSGACG00000017914 chr4 pcdh2g12 ENSGACG00000009389 chr12 hoxc13aENSGACG00000017915 chr4 pcdh2ac ENSGACG00000009391 chr12 hoxc12aENSGACG00000017917 chr4 not annotated ENSGACG00000009392 chr12 hoxc11aENSGACG00000017919 chr4 not annotated ENSGACG00000009394 chr12 hoxc10aENSGACG00000017920 chr4 si:ch73-233f7.1 ENSGACG00000009396 chr12 hoxc9aENSGACG00000017921 chr4 fgf18a ENSGACG00000009401 chr12 hoxc8aENSGACG00000017925 chr4 fbxw11a ENSGACG00000009405 chr12 hoxc6aENSGACG00000017928 chr4 etf1a ENSGACG00000009416 chr12 hoxc5aENSGACG00000022802 chr4 vault ENSGACG00000009421 chr12 hoxc4aENSGACG00000007085 chr10 not annotated ENSGACG00000009429 chr12 not annotatedENSGACG00000007090 chr10 not annotated ENSGACG00000009430 chr12 not annotatedENSGACG00000007094 chr10 hoxb3a ENSGACG00000009431 chr12 not annotatedENSGACG00000007100 chr10 hoxa4 ENSGACG00000009433 chr12 cbx5ENSGACG00000007108 chr10 hoxa5a ENSGACG00000009435 chr12 hnrnpa1bENSGACG00000007112 chr10 hoxa7 ENSGACG00000009442 chr12 nfe2ENSGACG00000007123 chr10 hoxa9a ENSGACG00000021244 chr12 not annotatedENSGACG00000007128 chr10 not annotated ENSGACG00000021427 chr12 not annotatedENSGACG00000007132 chr10 not annotated ENSGACG00000004548 chr16 hoxd3aENSGACG00000007134 chr10 hoxa13a ENSGACG00000004551 chr16 hoxd4aENSGACG00000007148 chr10 evx1 ENSGACG00000004556 chr16 hoxd9aENSGACG00000007155 chr10 hibadha ENSGACG00000004564 chr16 hoxd10aENSGACG00000022170 chr10 not annotated ENSGACG00000004569 chr16 hoxd11aENSGACG00000004574 chr16 hoxd12aENSGACG00000004579 chr16 not annotatedENSGACG00000004584 chr16 lnpaENSGACG00000021403 chr16 not annotatedDMC	Analysis Promoter Exon Intron Intergenic CpGi CpGi	shores CpGi	otherBackground	Distrubtions	for	All	CpG	Loci 12.33 5.75 26.36 55.56 3.34 26.43 70.23Sex 13.69 5.12 23.94 57.25 2.41 23.57 74.01Salinity 16.92 6.75 25.81 50.52 3.18 29.94 66.88Hypomethylated	in	Females 18.27 6.04 25.44 50.26 3.87 31.11 65.02Hypermethylated	in	Females 13.15 5.01 23.77 58.07 2.24 22.69 75.06Hypomethylated	at	High	salinity 16.94 6.57 25.40 51.09 3.43 28.16 68.41Hypermethylatd	at	High	Salinity 16.83 7.69 27.88 47.60 1.92 38.94 59.13       176 Table C.3: Top ten enriched biological process gene ontology categories  *p-value < 0.05 after FDR correction                  DMC	analysis category over	represented	pvalue under	represented	pvalue numDEInCat numInCat term ontologySex GO:0007411 1.20E-06* 0.999999299 123 610 axon	guidance BPSex GO:0030036 2.97E-05 0.999987721 38 146 actin	cytoskeleton	organization BPSex GO:0030574 3.97E-05 0.999987447 24 77 collagen	catabolic	process BPSex GO:0061337 4.24281E-05 0.999983448 33 122 cardiac	conduction BPSex GO:0030198 8.31041E-05 0.99995392 69 330 extracellular	matrix	organization BPSex GO:0022617 9.54257E-05 0.999960232 34 132 extracellular	matrix	disassembly BPSex GO:0034765 0.000211102 0.999911756 31 121 regulation	of	ion	transmembrane	transport BPSex GO:0051056 0.000313878 0.999851784 38 162 regulation	of	small	GTPase	mediated	signal	transduction BPSex GO:0030199 0.000536729 0.999840788 16 50 collagen	fibril	organization BPSex GO:0007268 0.00069141 0.999553526 90 486 chemical	synaptic	transmission BPSalinity GO:0051924 1.98E-05 0.999997917 8 37 regulation	of	calcium	ion	transport BPSalinity GO:0007156 7.17E-05 0.999982454 14 128 homophilic	cell	adhesion	via	plasma	membrane	adhesion	molecules BPSalinity GO:0033280 0.000214507 0.999984769 5 18 response	to	vitamin	D BPSalinity GO:0072006 0.000325122 0.999994625 3 5 nephron	development BPSalinity GO:0061444 0.001054532 1 2 2 endocardial	cushion	cell	development BPSalinity GO:0006812 0.001587505 0.999810676 5 27 cation	transport BPSalinity GO:0010976 0.001640774 0.999564522 10 101 positive	regulation	of	neuron	projection	development BPSalinity GO:0007096 0.001692083 0.999930411 3 8 regulation	of	exit	from	mitosis BPSalinity GO:0034765 0.001958714 0.999425251 11 121 regulation	of	ion	transmembrane	transport BPSalinity GO:0051056 0.002457297 0.999165564 13 162 regulation	of	small	GTPase	mediated	signal	transduction BP       177 Table C.4: Top ten enriched cellular component gene ontology categories  *p-value < 0.05 after FDR correction                 DMC	analysis category over	represented	pvalue under	represented	pvalue numDEInCat numInCat term ontologySex GO:0016020 1.07E-06* 0.999999226 345 2056 membrane CCSex GO:0045211 6.89E-06* 0.999996657 62 267 postsynaptic	membrane CCSex GO:0043195 8.1882E-05 0.999971094 26 90 terminal	bouton CCSex GO:0031095 0.000151272 0.99998868 7 11 platelet	dense	tubular	network	membrane CCSex GO:0042734 0.00032697 0.999877656 24 87 presynaptic	membrane CCSex GO:0016529 0.000370163 0.999900351 15 44 sarcoplasmic	reticulum CCSex GO:0008021 0.000442366 0.999811842 29 115 synaptic	vesicle CCSex GO:0030054 0.000533742 0.999650229 100 546 cell	junction CCSex GO:0030424 0.000552069 0.999686132 58 284 axon CCSex GO:0005903 0.000741204 0.999751522 18 61 brush	border CCSalinity GO:0005911 0.000646235 0.999782529 16 195 cell-cell	junction CCSalinity GO:0005886 0.00243756 0.998240739 141 3505 plasma	membrane CCSalinity GO:0005891 0.002512918 0.999570773 6 43 voltage-gated	calcium	channel	complex CCSalinity GO:0005958 0.003095305 0.999965854 2 3 DNA-dependent	protein	kinase-DNA	ligase	4	complex CCSalinity GO:0060187 0.003095305 0.999965854 2 3 cell	pole CCSalinity GO:1990423 0.003095305 0.999965854 2 3 RZZ	complex CCSalinity GO:0043235 0.003447861 0.998841329 12 149 receptor	complex CCSalinity GO:1902711 0.004251369 0.999530043 4 21 GABA-A	receptor	complex CCSalinity GO:0030424 0.005489377 0.997564702 18 284 axon CCSalinity GO:0070852 0.006032377 0.99955649 3 12 cell	body	fiber CC       178 Table C.5: Top ten enriched molecular function gene ontology categories         DMC	analysis category over	represented	pvalue under	represented	pvalue numDEInCat numInCat term ontologySex GO:0005509 4.58E-06* 0.999997239 124 632 calcium	ion	binding MFSex GO:0004252 5.50E-05 0.999979802 29 103 serine-type	endopeptidase	activity MFSex GO:0000146 7.74151E-05 0.999986659 12 27 microfilament	motor	activity MFSex GO:0005089 8.1882E-05 0.999971094 26 90 Rho	guanyl-nucleotide	exchange	factor	activity MFSex GO:0005516 0.000103789 0.999948231 50 221 calmodulin	binding MFSex GO:0003779 0.000106332 0.999937966 79 393 actin	binding MFSex GO:0030020 0.000123928 0.999994691 6 8 extracellular	matrix	structural	constituent	conferring	tensile	strength MFSex GO:0005220 0.00031733 1 4 4 inositol	1,4,5-trisphosphate-sensitive	calcium-release	channel	activity MFSex GO:0038191 0.000346476 0.999948396 9 19 neuropilin	binding MFSex GO:0005524 0.000359602 0.99972823 243 1492 ATP	binding MFSalinity GO:0005261 0.000367817 0.999969914 5 20 cation	channel	activity MFSalinity GO:0005509 0.000800742 0.999592781 36 632 calcium	ion	binding MFSalinity GO:0005096 0.001254622 0.999502823 19 268 GTPase	activator	activity MFSalinity GO:0004222 0.002704885 0.999228699 10 108 metalloendopeptidase	activity MFSalinity GO:0008381 0.003095305 0.999965854 2 3 mechanically-gated	ion	channel	activity MFSalinity GO:0030594 0.003095305 0.999965854 2 3 neurotransmitter	receptor	activity MFSalinity GO:0005230 0.0035311 0.999632183 4 20 extracellular	ligand-gated	ion	channel	activity MFSalinity GO:0046872 0.004488283 0.997071819 64 1420 metal	ion	binding MFSalinity GO:0004890 0.005064548 0.999407937 4 22 GABA-A	receptor	activity MFSalinity GO:0019992 0.006032377 0.99955649 3 12 diacylglycerol	binding MF       179 Appendix D  Supporting information for chapter five  D.1 Fish collection and acclimation  Adult threespine stickleback (G. aculeatus) used in this study were collected in May 2013 from Oyster lagoon (British Columbia, Canada, GPS: 49.6121,-124.0314) and transported to UBC. This population contains almost entirely fish of the fully plated “marine” ecotype (Barrett et al. 2008), and all fish used in this experiment were fully plated. Captured fish were transported to UBC where they were separated into six 110-litre glass tanks (20 fish/tank) and acclimated to 20 ppt salt water (instant ocean), 18°C and 14:10h light:dark photoperiod. Fish were fed daily to satiation with Hakari Bio-Pure frozen Mysis Shrimp. Fish were acclimated to these conditions for three weeks prior to the initiation of the experimental stress treatment.  D.2 Experimental stress treatment, crossing design and rearing  Following the three-week acclimation period to laboratory conditions, three tanks were designated as the unstressed treatment (tank #’s 1-3) and three tanks were designated as the stressed treatment (tank #’s 4-6). The four sides of the tanks containing the unstressed group were wrapped in black plastic to reduce visual disturbances. The top of the tanks as well as a small window on the front of each tank were left uncovered to allow for continued health monitoring of the fish during the treatment. Fish in the stressed treatment were chased once daily with an 8 in x 6 in blue aquarium fish net for 30sec before being captured and held out of water for a further 30 sec and then returned to their tank.  After two weeks, gravid females were harvested from both the stressed and unstressed treatments. As not all females reached reproductive readiness at the same time, eggs were        180 harvested over the course of 4 days. The daily stress treatment continued throughout. Six unstressed males were crossed to six unstressed females and six stressed females to generate a total of twelve families (6 half-sib pairs). Gravid females were removed from the experimental tanks and eggs were collected by gently applying pressure to the abdomen with a thumb and forefinger. Testes were dissected from unstressed males and macerated in a 1.75 mL microcentrifuge tube containing 100 uL Ginzberg’s fish ringers solution. Eggs collected from females were arranged as a monolayer in petri dishes containing 5mL of 20ppt seawater. 50 µL of the sperm solution was applied directly on the egg mass and left for 30 min to allow fertilization to occur. Following fertilization, an additional 10mL of 20 ppt saltwater was added to each petri dish. Petri dishes were partially covered to prevent water loss from evaporation and allow for surface gas exchange. Eggs were monitored twice daily during which time any unfertilized eggs were removed and 10 mL of water was changed to prevent mold growth. After all fish in a petri dish had hatched and the yolks had been absorbed (~15 days post fertilization), larvae were transferred to 110 L glass aquaria equipped with hanging box filters (Aquaclear) and sponge filters for filtration and aeration. Larval stickleback were fed live brine shrimp nauplii twice daily ad libitum until they were large enough to feed on frozen Mysis shrimp. Each family was reared in a separate tank. Although density varied because family size varied, as the juvenile stickleback grew towards adulthood families were culled to a maximum final density of ~1 fish per 5 L. After one year post hatch, and prior to reaching sexual maturity, whole brain tissue was dissected and immediately frozen in liquid nitrogen and stored at -80°C until further use. Family information for the fish used for RNA-seq is provided in Table D.1 (Appendix D).         181 D.3 DNA Isolation and Sex Identification  Genomic DNA was isolated from fin clips using the Qiagen DNeasy Blood & Tissue mini spin columns (Product # 69504). The sex of each fish was determined by PCR (Toli et al. 2016). PCR amplification was conducted using PTC-200 a Peltier Thermal Cycler (MJ Research) in 10 µl reactions containing 0.5 µM of each primer, 5 ng DNA, 0.2 mM dNTPs, 2 mM MgCl2, and 0.25 units Taq polymerase (Fermentas) with the following thermal cycling conditions: 95 °C for 5 min followed by 40 cycles of 95 °C for 30 s, 50 °C for 30 s, and 72 °C for 1 min with a final extension of 72 °C for 10 min. PCR products were separated by gel electrophoresis on a 2% agarose gel containing SYBR safe DNA gel Stain (product # S33102; ThermoFisher Scientific).  D.4 Plasma cortisol  To verify that the maternal stress treatment induced an increase in the circulating cortisol levels, two gravid female fish were randomly selected from each of the three control and three experimental tanks (n = 6 for each treatment) and rapidly sacrificed with a cranial blow. Blood was collected into heparinized capillary tubes by caudal severance. Plasma was separated by centrifugation, frozen in liquid nitrogen and stored at -80°C until use. Steroids were extracted with diethyl ether, and cortisol levels were assessed using an enzyme-linked immunosorbent assay (ELISA) (product# 402710; Neogen) according to the manufacturer’s instructions.  D.5 RNA isolation  Total RNA was isolated from stickleback brain tissue using TRIzol Reagent (Invitrogen Life Technologies) following the manufacturer’s recommended protocol. Briefly, tissue was        182 homogenized in 1 mL of TRIzol containing approximately 10, 1.0 mm ceria stabilized zirconium oxide beads (Next Advance, NY, USA) using a Bullet Blender24 (Next Advance, NY, USA). Total RNA was DNase treated using the Qiagen RNeasy (product # 74104) DNase I (product # 79254) on column DNA digestion protocol. Total RNA was quantified using the QBit® RNA broad range assay kit (product # Q10210; ThermoFisher Scientific) and an Invitrogen™ Qubit® 2.0 Flurometer. RNA quality was assessed using the Agilent RNA 6000 Pico Kit (product # 5067-1514) and an Agilent 2100 Bioanalyzer (Agilent Technologies).   D.6 Gene Ontology and pathway analysis  Orthologous Ensembl gene IDs for stickleback and human and the corresponding human UniProt accession numbers were obtained from the Ensembl biomart database and identifiers for Gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) IDs were obtained using the UniProt online mapping tool (http://www.uniprot.org/uploadlists/) to generate a database of stickleback Ensembl gene IDs linked to orthologous human GO and KEGG gene IDs. The keggLink() function in the KEGGREST v1.6.4 R package was then used to obtain KEGG pathway identifiers for the KEGG gene IDs obtained from the UniProt database. A total of 15,646 genes were assigned GO identifiers and 3,213 genes were assigned KEGG pathway IDs using this method. GO and KEGG pathway enrichment analyses was conducted using the goseq (v1.22.0) R package (Young et al. 2010). The goseq package contains gene length information for stickleback ("gasAcu1) based on Ensembl gene IDs (“ensGene"). Results from the enrichment analysis were corrected for the false discovery rate (FDR) of multiple comparisons (Benjamini & Hochberg 1995) using a q value cutoff of <0.05.         183 D.7 Differential expression between males and females  1,255 genes exhibited a significant effect of sex, independent of maternal stress (at an FDR q<0.05). Figure D.2 (Appendix D) provides a heatmap summarizing the expression patterns of these genes.  D.8 Effects of increased stringency thresholds on detection of sexually-dimorphic effects of maternal stress  To determine the robustness of the detected sexually dimorphic response to maternal stress we analyzed the RNA-seq data using two different approaches to reduce the false discovery cost. When we applied cutoff of q<0.01 (increasing the stringency of the false discovery rate criterion), we detected 821 genes with a significant interaction between the effects of sex and maternal stress (compared with 1650 with a cutoff of q<0.05). Figure D.4A (Appendix D) provides a heatmap summarizing the patterns of expression for these genes. When we applied a fold-change cutoff of >2.0 we detected 675 genes with a significant interaction between the effects of sex and maternal stress. Figure D.4B provides a heatmap summarizing the patterns of expression for these genes. If both criteria are applied simultaneously 472 genes are detected with a significant interaction between the effects of sex and maternal stress.   GO-enrichment analysis of these datasets detected different enriched processes in males and females, as was observed in the analyses presented in the main text. Enrichment results for genes with a fold-change greater than two are summarized in Table D.3.         184 D.9 Candidate genes involved in the stress response and epigenetic regulation  Genes with known roles in the regulation of the glucocorticoid receptor expression and activity exhibited significant interactions between maternal stress and offspring sex are listed in Table D.4. In addition, the early growth response gene egr, which is known to regulate the transcription of the glucocorticoid receptor (Szyf et al. 2005) demonstrated a significant effect of sex, with higher expression in males compared to females and there was a trend for an interaction between maternal stress and sex (p = 0.053)   With respect to genes with known roles in epigenetic processes, we identified several members of the ten-eleven translocation (TET) family of enzymes, which are involved in active DNA demethylation (Li & Zhang 2014). A significant interaction between maternal stress and offspring stress was identified for two members of the TET family of DNA-demethylases, tet2 and tet3, which were up-regulated in male offspring of stressed mothers, but down-regulated in female offspring. These sex-specific effects of maternal stress on the expression of DNA-demethylases suggest a potential role for epigenetic mechanisms influencing the dimorphic expression of genes in males and females from stressed mothers.  In contrast, we did not identify a significant interaction between maternal stress and offspring sex for any of the DNA (Cytosine-5-)-Methyltransferase 3 Alpha (dmnt3a ) isoforms that are involved in regulating de novo methylation patterns in the genome (Li & Zhang 2014). However, dnmt3aa and dmnt3ab were both differentially expressed between the sexes and for dnmt3aa there was a trend towards an interaction between offspring sex and maternal stress (p = 0.083; FDR corrected).   A variety of other genes whose products are known to associate with methylated DNA also exhibited sex-specific effects of maternal stress, including zinc finger and BTB domain        185 containing 7A (zbtb7a), methyl-CpG binding domain protein 5 (mbd5), and myocyte enhancer factor 2C (mef2c), which are also up-regulated in male, but not female, offspring of stressed mothers. Zbtb7a (also referred to as FBI-1) has been shown to interact with the methyl binding domain containing protein mbd3 to recruit DNA methylation machinery to silence the promoters of target genes (Choi et al. 2013). mbd5 is another member of the MBD protein family that plays an important role regulating expression of genes involved in neural development (Talkowski et al. 2011, 2012) and growth (Du et al. 2012) but cannot bind methylated DNA (Laget et al. 2010). mbd5 is thought bind heterochromatin by interacting with mef2c (Gigek et al. 2016).  Genes involved in a variety of other epigenetic mechanisms also exhibited sex-specific effects of maternal stress including bromodomain adjacent to zinc finger domain 2A (baz2a), which is involved in gene silencing by recruiting enhancer of zeste 2 polycomb repressive complex 2 subunit (ezh2) which has known histone methyltransferase activity (Gu et al. 2015).  The GO category “translational gene silencing by RNA” (GO: 0035194) was also significantly enriched among genes that demonstrated a significant interaction between offspring sex and maternal stress. The genes associated with this term include several members of the argonaute (ago) and trinucleotide repeat containing 6 (tnrc) gene family including ago2, ago3, ago4, tnrc6a, tnrc6b, tnrc6c, and tnrc6c2. These genes have been shown to interact with each other and regulate micro RNA (miRNA) silencing activity (Liu et al. 2004; Rehwinkel et al. 2005; Baillat & Shiekhattar 2009).            186 D.10 Supporting tables and figures for chapter five Table D.1: Fish used for RNA-seq Offspring Fish # Offspring Sex Maternal Stress Treatment Mother # Mother Tank # Father # Father Tank #  UM1 Male Unstressed 1 1 1 1 UM2 Male Unstressed 2 1 2 1 UM3 Male Unstressed 1 1 1 1 UF1 Female Unstressed 3 2 3 2 UF2 Female Unstressed 4 3 4 3 UF3 Female Unstressed 5 1 5 1 SM1 Male Stressed 6 4 1 1 SM2 Male Stressed 7 5 4 3 SM3 Male Stressed 8 4 5 1 SF1 Female Stressed 9 5 2 1 SF2 Female Stressed 10 6 6 2 SF3 Female Stressed 9 5 2 1   Table D.2: RNA quality and library sizes Sample Name RNA Integrity Number (RIN) RNA-seq Library Size UM1 8.5 11383120 UM2 7.9 11891132 UM3 8.9 22702505 UF1 8.9 21626954 UF2 7.9 18841249 UF3 9.1 31112715 SM1 8.1 18149183 SM2 8.2 16236161 SM3 8.7 18104469 SF1 8.5 15446367 SF2 9.0 24948538 SF3 9.0 27018611                          187 Table D.3: GO enrichment analysis of genes with fold-change >2.0 with a significant interaction between maternal stress and offspring sex Enriched GO terms of genes expressed at higher levels in males from stressed mothers Enriched GO terms of genes expressed at higher levels in females from stressed mothers Biological Process GO:0007268   Synaptic transmission GO:0044281 Small molecule metabolic process GO:0043401   Steroid hormone mediate signaling pathway GO:0044237  Cellular metabolic process GO:0043484  Regulation of RNA splicing GO:0022904  Respiratory electron transport chain   GO:0006414  Translational elongation    GO:0006415  Translational termination    GO 0016259  Selenocysteine metabolic process   GO:0019083  Viral transcription    GO:0001887  Selenium compound metabolic process Cellular Component GO:0005887  Integral component of plasma membrane GO:0005743 Mitochondrial inner membrane GO:0030425  Dendrite    GO:0045211 Post-synaptic membrane    Molecular function GO:0003700  Transcription factor activity, sequence-specific DNA binding   GO:0003707  Steroid hormone receptor activity     Table D.4: Candidate genes with known roles in the regulation of glucocorticoid receptor expression and activity Gene symbol Gene name Protein function crebbpb CREB binding protein b  histone acetyltransferase (Weaver et al. 2007; McGowan et al. 2009; Hellstrom et al. 2012) ncoa1 Nuclear receptor coactivator 1  steroid receptor coactivator protein; (Heitzer et al. 2007) ncoa3  Nuclear receptor coactivator 3 steroid receptor coactivator protein (Heitzer et al. 2007) nrip1b Nuclear receptor interacting protein 1b nuclear receptor interacting protein (Windahl et al. 1999) rela v-rel avian reticuloendotheliosis viral oncogene homolog A NF-ϰB subunit (Wissink et al. 1997) pou2f1 POU class 2 homeobox 1  POU domain factor with roles in neuroendocrine function (Préfontaine et al. 1998; Andersen & Rosenfeld 2001) pou2f2 POU class 2 homeobox 2 POU domain factor with roles in neuroendocrine function (Préfontaine et al. 1998; Andersen & Rosenfeld 2001)         188  Table D.5: Significantly enriched KEGG pathways     Enriched	KEGG	Pathway	ID	(FDR	q	<	0.05) KEGG	ID pvalueNumber	of	DE	Genes	With	AnnotationNumber	of	Genes	With	AnnotationRibosome hsa03010 1.81E-08 22 67Huntington's	disease hsa05016 2.72E-08 30 90Oxidative	phosphorylation hsa00190 3.00E-08 22 65Parkinson's	disease hsa05012 2.13E-07 22 71Alzheimer's	disease hsa05010 4.37E-07 29 89Non-alcoholic	fatty	liver	disease hsa04932 1.25E-06 19 61Nicotine	addiction hsa05033 2.17E-05 14 33Glutamatergic	synapse hsa04724 0.00012917 23 72Morphine	addiction hsa05032 0.00038931 14 48Retrograde	endocannabinoid	signaling hsa04723 0.00080708 19 65Oxidative	phosphorylation hsa00190 5.66E-08 22 65Parkinson's	disease hsa05012 1.23E-06 21 71Huntington's	disease hsa05016 6.70E-06 21 90Ribosome hsa03010 1.24E-05 22 67Alzheimer's	disease hsa05010 3.39E-05 19 89Non-alcoholic	fatty	liver	disease hsa04932 6.18E-05 16 61Nicotine	addiction hsa05033 1.94E-07 14 33Morphine	addiction hsa05032 6.79E-07 14 48Glutamatergic	synapse hsa04724 2.28E-06 22 72Retrograde	endocannabinoid	signaling hsa04723 3.86E-06 19 65Neuroactive	ligand-receptor	interaction hsa04080 2.24E-05 19 120Cocaine	addiction hsa05030 0.0001104 10 34Cell	adhesion	molecules hsa04514 0.00015629 11 41GABAergic	synapse hsa04727 0.00017636 12 50Rap1	signaling	pathway hsa04015 0.00027316 18 94cAMP	signaling	pathway hsa04024 0.00094925 19 105Transcriptional	misregulation	in	cancer hsa05202 0.00166575 10 52Long-term	potentiation hsa04720 0.00171489 12 43MAPK	signaling	pathway hsa04010 0.00173098 18 127Long-term	depression hsa04730 0.00238347 10 38Calcium	signaling	pathway hsa04020 0.00246922 18 89All	DE	GenesDE	genes	up-regulated	in	FemalesDE	genes	up-regulated	in	Males       189  Figure D.1: Plasma cortisol levels of adult stickleback mothers from the unstressed and stressed treatments. Data are presented as box and whisker plots generated using default settings in R. Asterisk indicates significant difference in mean plasma cortisol levels between stressed and unstressed mothers detected by one-way ANOVA in R (p = 0.0014). n = 6 fish per treatment group.   Figure D.2: Heat map of all differentially expressed (DE) genes for which a significant effect of offspring sex was identified. Expression results are displayed as log2 counts per million normalized to the mean expression for all samples. Blue indicates lower expression and yellow indicates higher expression.  Males Females10.50-0.5-1       190   Figure D.3: Summary of gene ontology (GO) categories for genes that exhibited sexually dimorphic expression patterns independent of maternal stress. GO categories represent those that contained at least ten genes that were differentially expressed between males and females and where 80% of the genes in that GO category exhibit a bias of higher expression in males (A) or higher expression in females (B). Blue bars represent the number of genes expressed at higher levels in males while red bars represent genes expressed at higher levels in females. The percent of genes within a category that exhibit a given bias in expression is also shown.           191  Figure D.4. Cluster analysis of the significantly enriched cellular component gene ontology (GO) terms over-represented among genes that increased in males (blue circles) or increased in females (red circles) from stressed mothers compared to unstressed mothers. Size of the circle is representative of the total number of DE genes in that GO category. Note that some names have been omitted from the figure due to semantic similarity.    Figure D.5: Heatmap displaying the expression patterns of genes with a significant interaction between sex and maternal stress with A) FDR q<0.01 or B) fold change >2.0. Expression results are displayed as log2 counts per million normalized to the mean expression for all samples. Blue indicates lower expression and yellow indicates higher expression.  Fig S4.Extracellular vesicular exosomeSynaptic vesicleMitochondrionMitochondrial inner membraneNucleusRibosomeAxonDendriteSynapseCell junctionPlasma membranePostsynaptic membraneSemantic space x0-5 50-44Semantic space yFig S4.Unstressed Mothers Stressed MothersMale MaleFemale Female10.50-0.5-1Unstressed Mothers Stressed MothersMale MaleFemale FemaleA	 B	

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0365265/manifest

Comment

Related Items