CONTRIBUTIONS OF GENETIC VARIATION AND PHENOTYPIC PLASTICITY TO VARIATION IN HIGH pH TOLERANCE IN RAINBOW TROUT by Sara Lynn Northrup B.Sc. The University of Guelph, 2003 M.Sc. The University of British Columbia, 2008 A THESIS SUBMITTED IN PARTIAL FULFULLMENT 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) October 2017 © Sara Lynn Northrup, 2017 ii Abstract High pH is physiologically stressful for Rainbow Trout, causing poor survival when fish are stocked into high pH lakes. To assess the relative contributions of genetic variation and phenotypic plasticity in high pH tolerance in Rainbow Trout, I examined high pH (pH 9.5) tolerance in three Rainbow Trout strains (Blackwater River, Eagle Lake and Fraser Valley Domestics) under four different rearing conditions: 1) near-neutral hatchery conditions (pH 7.2) from fertilization; 2) pH 8.5 from fertilization; 3) pH 8.8 from fertilization; and 4) near-neutral hatchery conditions from fertilization followed by acclimation to pH 8.8 for one month prior to testing (at fry and yearling). In general, I found that either rearing or acclimating fish to elevated pH improved high pH tolerance. Variation among strains was observed only at the fry life stage. I performed a genome wide association study to identify genetic variation that may be associated with differences in pH tolerance among strains. The results suggest that pH tolerance is likely controlled polygenically. To assess mechanisms underlying phenotypic plasticity in high pH tolerance, gill gene expression of fish reared under control conditions and those acclimated to pH 8.8 were compared using RNA-Seq. There were 140 genes that were significantly differentially expressed in response to high pH, but the most dramatic results were the strong interaction effects between pH and strain suggesting that each strain compensates for high pH conditions in different ways. Finally, the variation among strains and rearing treatments observed within the laboratory was tested in natural lakes. In general, short-term net pen trials were consistent with laboratory results showing higher pH tolerance in fish reared at or acclimated to elevated pH levels. Long-term survival trials indicate that the large differences in survival in natural lakes between strains mask subtler effects of prior exposure to high pH and require further investigation. My data suggest that it is the remarkable plasticity of Rainbow Trout rather than a specific strain or genotype which has the greatest effect on high pH tolerance, and that modifications of hatchery practices could be used to improve survival of stocked Rainbow Trout in high pH lakes. iii Lay Summary Many lakes in British Columbia are increasing in pH, which could lead to lower Rainbow Trout survival. The key goal of this research was to determine whether it would be more practical to breed trout with better pH tolerance or to use changes in rearing methods to improve pH tolerance for stocking fish in high pH lakes. I showed that although there is variation among Rainbow Trout strains in high pH tolerance, prior exposure to moderately high pH improved survival in both the laboratory and in natural high pH lakes in the short-term for all strains. However, I also showed that there were differences among the strains in the mechanisms involved in becoming tolerant to higher pH following exposure to moderately high pH. These results have management implications and I propose several recommendations that may help to maintain and restore fisheries in lakes affected by increasing pH. iv Preface Chapter 2 of this thesis is co-authored by Sara Northrup and Patricia M. Schulte. Data analysis and write up were performed by Sara Northrup. Chapter 3 of this thesis is co-authored by Sara Northrup, Timothy M. Healy and Patricia M. Schulte. All laboratory procedures were performed by Sara Northrup. Genome-by-sequencing and SNP calling was performed by the University of Cornell Genomic Diversity Facility. Timothy Healy performed paralog and missing data filtering. Sara Northrup ran all statistical analysis and wrote the chapter. Chapter 4 of this thesis is co-authored by Sara Northrup, David C.H. Metzger, Jason Blank and Patricia M. Schulte. All laboratory procedures were performed by Sara Northrup. Sequencing was performed by Genome Quebec. David C. H. Metzger performed bioinformatic and statistical RNA-Seq analysis. Jason Blank assisted with the functional annotation of genes that were significantly differentially regulated by high pH. Sara Northrup interpreted the results and wrote the chapter. Chapter 5 of this thesis is co-authored by Sara Northrup, Brett van Poorten and Patricia M. Schulte. All data were collected by Sara Northrup. Bayesian model creation was completed by Brett van Poorten and Sara Northrup. Sara Northrup ran all statistical analysis and wrote the chapter. v Table of Contents Abstract ............................................................................................................................ii Lay Summary .................................................................................................................. iii Preface ............................................................................................................................iv Table of Contents ............................................................................................................ v List of Tables ...................................................................................................................ix List of Figures ................................................................................................................. xii Acknowledgements ...................................................................................................... xvii Chapter 1: General introduction ...................................................................................... 1 1.1 Recreational fishing ............................................................................................... 1 1.2 High pH lakes in British Columbia .......................................................................... 2 1.3 Physiological response to high pH ......................................................................... 4 1.4 Effects of high pH on fish ....................................................................................... 6 1.5 Rainbow Trout ........................................................................................................ 8 1.5.1 Rainbow Trout strains ...................................................................................... 9 1.6 Fisheries response to high pH ............................................................................. 12 1.6.1 Acclimation .................................................................................................... 12 1.6.2 Strain ............................................................................................................. 15 1.7 Thesis objectives ................................................................................................. 18 1.8 Tables .................................................................................................................. 20 1.9 Figures ................................................................................................................. 21 Chapter 2: Variation in acute pH tolerance among and within strains of Rainbow Trout and between rearing environments ............................................................................... 25 2.1 Introduction .......................................................................................................... 25 2.2 Methods ............................................................................................................... 27 2.2.1 Experiment 1.1: Effects of strain .................................................................... 27 2.2.2 Experiment 1.2: Effects of high pH rearing .................................................... 29 2.2.3 Experiment 1.3: Effects of acclimation ........................................................... 30 2.2.4 Statistical analyses ........................................................................................ 31 2.3 Results ................................................................................................................. 31 2.3.1 Experiment 1.1: Variation in acute high pH tolerance among strains ............. 31 vi 2.3.2 Experiment 1.2: Variation in acute high pH tolerance between rearing environments .......................................................................................................... 32 2.3.3 Experiment 1.3: Variation in acute high pH tolerance between rearing environments and acclimation ................................................................................ 32 2.4 Discussion ............................................................................................................ 33 2.4.1 Variation in acute high pH tolerance between rearing environments and acclimation .............................................................................................................. 33 2.4.2 Variation in acute high pH tolerance among strains ....................................... 34 2.5 Conclusions ......................................................................................................... 37 2.6 Tables .................................................................................................................. 38 2.7 Figures ................................................................................................................. 42 Chapter 3: A GWAS analysis of the genetic basis of high pH tolerance across three strains of Rainbow Trout ............................................................................................... 53 3.1 Introduction .......................................................................................................... 53 3.2 Methods ............................................................................................................... 55 3.3 Results and discussion ........................................................................................ 57 3.4 Tables .................................................................................................................. 62 3.5 Figures ................................................................................................................. 63 Chapter 4: Gene expression plasticity in response to acclimation to high pH in three strains of Rainbow Trout ............................................................................................... 67 4.1 Introduction .......................................................................................................... 67 4.2 Methods ............................................................................................................... 69 4.3 Results ................................................................................................................. 71 4.3.1 Differential gene expression in response to high pH ...................................... 72 4.3.2 Differential gene expression in response to high pH between strains ............ 72 4.3.3 Differential gene expression in response to high pH within strain .................. 72 4.4 Discussion ............................................................................................................ 73 4.4.1 Effects of pH that are common to all strains .................................................. 74 4.4.2 Effects of pH within each strain ...................................................................... 76 4.4.3 Future analysis .............................................................................................. 78 4.4.4 Conclusion ..................................................................................................... 79 4.5 Tables .................................................................................................................. 80 4.6 Figures ................................................................................................................. 92 vii Chapter 5: Natural lake evaluation of high pH tolerance in Rainbow Trout from different rearing environments ..................................................................................................... 96 5.1 Introduction .......................................................................................................... 96 5.2 Methods ............................................................................................................... 97 5.3 Results ............................................................................................................... 101 5.3.1 Descriptive statistics .................................................................................... 101 5.3.2 Bayesian model ........................................................................................... 102 5.4 Discussion .......................................................................................................... 103 5.4.1 Effects of rearing and acclimation treatments .............................................. 103 5.4.2 Effects of strain ............................................................................................ 104 5.4.3 Short-term vs long-term survival .................................................................. 106 5.4.4 Size/ development stage ............................................................................. 107 5.4.5 Laboratory vs lake ....................................................................................... 107 5.4.6 Model vs ANOVA ......................................................................................... 108 5.4.7 Conclusions ................................................................................................. 109 5.5 Tables ................................................................................................................ 111 5.6 Figures ............................................................................................................... 123 Chapter 6: Conclusion ................................................................................................. 127 6.1 Variation in acute pH tolerance among and within strains of Rainbow Trout (Chapter 2) ............................................................................................................... 127 6.2 Variation in pH tolerance among Rainbow Trout between rearing environments (Chapter 2) ............................................................................................................... 128 6.3 A GWAS analysis of the genetic basis of high pH tolerance across three strains of Rainbow Trout (Chapter 3)....................................................................................... 129 6.4 Gene expression plasticity in response to acclimation to high pH in three strains of Rainbow Trout (Chapter 4)....................................................................................... 129 6.5 Natural lake evaluation of high pH tolerance in Rainbow Trout from different rearing environments (Chapter 5) ............................................................................ 130 6.6 Implications of research ..................................................................................... 130 6.7 Strengths and limitations .................................................................................... 132 6.8 Future research directions ................................................................................. 133 6.9 Conclusion and recommendations ..................................................................... 136 References .................................................................................................................. 138 Appendix Tables ......................................................................................................... 150 viii Appendix Figures ........................................................................................................ 172 ix List of Tables Table 1.1: Extent of occurrence of high pH conditions in the lakes stocked in British Columbia (FFSBC 2017), expressed as number of lakes and lake surface area .......... 20 Table 1.2: Average water pH in lakes by British Columbia management region (FFSBC 2017) ............................................................................................................................. 20 Table 2.1: Water chemistry parameters from Fraser Valley Hatchery Wells (FFSBC 2017) ...................................................................................................................................... 38 Table 2.2: pH settings for each head tank during rearing and acute high pH tolerance assay ............................................................................................................................. 39 Table 2.3: Controller pH settings for increasing pH during acclimation ......................... 39 Table 2.4: Mean and standard deviation of weight and length of each strain of fish for each experiment. BW = Blackwater, FV = Fraser Valley and EL = Eagle Lake. n=30 for each strain in each experiment. .................................................................................... 40 Table 2.5: Average testing pH during loss of equilibrium experiments. BW = Blackwater, FV = Fraser Valley and EL = Eagle Lake. Experiment 1.1 (Yearling) had 10 fish per strain per tank. Experiment 1.1 (Fry) had 6 fish per strain per tank. Experiment 1.2 and 1.3 had 30 fish for each strain. ................................................................................................... 41 Table 3.1: SNPs identified as significantly associated with high pH (pH 9.5) tolerance for Blackwater, Eagle Lake and Fraser Valley Domestics reared in near neutral well water (pH ~7.2) ....................................................................................................................... 62 Table 4.1: Mapping statistics for RNA-Seq gill libraries. Sample name indicates strain; Blackwater (BW), Eagle Lake (EL), or Fraser Valley (FV); treatment (fish acclimated to pH 8.8) or control (fish reared under standard hatchery conditions) and individual sample number. n=6 per strain per treatment ............................................................................ 80 Table 4.2: GO analysis of differentially expressed genes within combined strains between control and acclimated treatments in the gill ................................................................. 82 Table 4.3: GO analysis of upregulated differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments in the gill ........................... 82 Table 4.4: GO analysis of downregulated differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments in the gill ........................... 83 x Table 4.5: GO analysis of upregulated differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments in the gill ........................... 85 Table 4.6: GO analysis of downregulated differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments in the gill ........................... 86 Table 4.7: GO analysis of upregulated differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments in the gill ........................... 87 Table 4.8: GO analysis of downregulated differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments in the gill ................ 89 Table 5.1: Experimental lake parameters .................................................................... 111 Table 5.2: Experimental lake basic water chemistry measures ................................... 112 Table 5.3: Short-term experimental lake survival. Value recorded is the % of fish living at time of monitoring. ....................................................................................................... 113 Table 5.4: Experimental lake fish stocking for high pH tolerance: short (net pen) and long-term (into Lake) assessment. ...................................................................................... 114 Table 5.5: Gill net mesh set per experimental lake and day for long-term survival assessment. ................................................................................................................ 116 Table 5.6: Total gill net catch for long-term survival assessment. M&R indicates mark and recapture fish stocked one week prior to assessment. ......................................... 117 Table 5.7: Three-way ANOVA results from lake by strain by treatment comparison for short and long-term comparisons for each life stage stocked (fry and yearling). Significant results are in bold. ....................................................................................................... 120 Table 5.8: Long term lake analysis DIC model selection criteria for a set non-hierarchical (fixed effects) models (best fit models are in bold) ...................................................... 121 Table A.1: Breakdown of pH trends for 148 lakes in the Thompson-Nicola Region (1970-2010) which has the highest concentration of stocked lakes in BC (FFSBC 2017) ..... 150 Table A.2: pH tolerance and growth phenotypes for fish included in GBS……...….............................................................................................................…153 xi Table A.3: Genes associated with high pH tolerance from RNAseq analysis (sorted by FDR). Genes associated with ammonium transporter and carbonic anhydrase are highlighted in red and blue respectively……………………………………………………154 Table A.4: GO analysis of differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments…………………………………….….168 Table A.5: GO analysis of differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments…………………………………...……….….168 Table A.6: GO analysis of differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments………………………………….…….167 Table A.7: Water chemistry measures from experimental lakes …………..…………...171 xii List of Figures Figure 1.1: Provincial freshwater fisheries management regions in British Columbia (FFSBC 2017) ............................................................................................................... 21 Figure 1.2: All lakes stocked by FFSBC. Lakes with a pH less than 8.5 (black), lakes with a pH greater than 8.5 (green), lakes with a pH greater than 9.0 (purple). ..................... 22 Figure 1.3: pH measures in high pH lakes in British Columbia over time. Each point represents a single pH measurement. Data collection was halted in the 2000s until 2012. (FFSBC 2017) ............................................................................................................... 23 Figure 2.1: Summary of experimental design for a) Experiment 1.1: 6 fish were transferred from each strain’s individual rearing tank to the test tank for a total of 24 fish in each tank and 30 fish of each strain across the five testing tanks, b) Experiment 1.2: 30 fish were transferred of each strain from the three different rearing containers into a testing tank, with one testing tank per strain for Blackwater, Fraser Valley 2n and 3n. 10 Eagle Lake fish that had been reared at pH 7.2 were also transferred into each of the testing tanks. The Blackwater strain is shown as an example for the rearing tanks, and the other strains were transferred in the same fashion, c) Experiment 1.3: 30 fish of each were transferred from the three different rearing containers (with the exception of Eagle Lake which did not have a pH 8.8 reared group) into a testing tank such that there was one testing tank per strain for Blackwater, Eagle Lake, Fraser Valley 2n and 3n. The Blackwater strain is shown as an example for the rearing tanks, and the other strains were transferred in the same fashion. BW = Blackwater, FV = Fraser Valley, and EL= Eagle Lake .................. 43 Figure 2.2: Effect of strain on tolerance of acute transfer to pH 9.5. The data shown are percent of fish remaining as a function of time exposed to pH 9.5 (Experiment 1.1 Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW = black square), Eagle Lake (EL = red circles), Fraser Valley (FV2n = green triangles, and FV3n = blue diamonds) (5 tanks with 6 individuals each resulting in a total n=30 for each). Error bars indicate standard error. Significance is indicated by the letter to the right of each strain in the graph. ...................................................................................................................... 44 Figure 2.3: Size distribution of all fish separated by strain and state (exhibiting loss of equilibrium (LOE) or not) at the end of the acute tolerance assay from experiment 1.1 fall fry 2014 brood (White – fish that LOE; Red - fish that did not LOE). BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley. Error bars indicate standard error. n=30 for each strain. There were no significant differences in length between fish that exhibited loss of equilibrium (LOE) and those that did not for any strain. ................................................ 45 Figure 2.4: Percent of fish remaining as a function of time exposed to pH 9.5 Data are for Experiment 1.1 (Spring Yearlings 2013 brood) comparing Rainbow Trout strains xiii Blackwater (BW), Fraser Valley (FV2n and FV3n) (3 tanks with 10 individuals in each resulting in a total n=30 for each) a) replicate 1, b) replicate 2 (BW = black squares, FV2N = green triangles, FV3N = blue diamonds). Error bars indicate standard error. There was no significant difference found between the strains in either replicate. ......................... 46 Figure 2.5: Average daily rearing pH for each rearing group (pH 8.8 = green; pH 8.5 = red; control = black) during the length of 2014 brood growth cycle. .............................. 47 Figure 2.6: Effect of rearing water pH on Rainbow Trout tolerance to acute transfer to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5. Results are from Experiment 1.2 (Fall fry 2014 brood) comparing fish reared at pH 8.8 (green triangles), pH 8.5 (red squares) and control conditions (~pH 7.2) (black circles). Rainbow Trout strains a) Blackwater (BW), b) Fraser Valley (FV2n), and c) FV3n (n=30 for each). Significance is indicated by the letter to the right of each strain in the graph. 48 Figure 2.7: Length of fish tat the end of rearing by for treatment groups in three strains (white = control well water pH 7.2; red = reared at pH 8.5; green = reared at pH 8.8). BW = Blackwater, FV = Fraser Valley. Error bars indicate standard error. n=30. There were no significant differences in length among treatment groups for any strain ................... 49 Figure 2.8: Effect of acclimation to pH 8.8 on high pH tolerance following acute transfer to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5 for Experiment 1.3 (Fall fry 2014 brood) comparing fish reared at pH 8.8 (except for EL) (green), acclimation to pH 8.8 (blue) and control conditions (~pH 7.2) (black). Rainbow Trout strains a) Blackwater (BW), b) Eagle Lake (EL), c and d) Fraser Valley (FV2n and FV3n, respectively) (n=30 for each strain and treatment). Significance is indicated by the letter to the right of each strain in the graph. Note: The data for acclimated fish and fish reared at pH 8.8 often overlap which obscures one of the two treatments. .................................................................................................................... 50 Figure 2.9: Effect of strain on high pH tolerance following acute exposure to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5 Results from Experiment 1.2 (Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW), Eagle Lake (EL), Fraser Valley (FV2n and FV3n) (n=30 for each) all reared in control conditions. .......................................................................................... 51 Figure 2.10: Effect of strain on high pH tolerance following acute exposure to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5. results from Experiment 1.3 (Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW), Eagle Lake (EL), Fraser Valley (FV2n and FV3n) (n=30 for each) all reared in control conditions. ......................................................................................................... 52 xiv Figure 3.1: Number of SNPs per chromosome. The current version of the Rainbow Trout genome sequence (Berthelot et al. 2014) has not been fully assembled into chromosomes so there are an additional19,965 SNPs that were found on the large scaffold containing incompletely localized sequences that are not shown here a) Illustrates the SNPs found with specific areas on a chromosome. b) Illustrates the number of SNPs found to be on specific chromosomes but unknown areas on those chromosomes. ............................ 63 Figure 3.2: Plot of non-FDR corrected P-values for SNPs associated with high pH (9.5) tolerance in fish reared in near neutral well water (pH ~7.2) n=143. No SNPs were identified that were significantly associated with high pH tolerance with whole-genome level FDR correction. Red circles indicate SNPs that were significantly associated with high pH tolerance at the chromosome level, and grey open circles indicate SNPs with no significant association when raw P-values are FDR corrected. a) Illustrates the SNPs found with specific areas on a chromosome. b) Illustrates the number of SNPs found to be on specific chromosomes but unknown areas on those chromosomes. c) Illustrates the SNPs on the large scaffold containing incompletely localized sequences. .............. 65 Figure 3.3: qqplot demonstrating fit of the mixed linear model utilized to analyze the association of SNPs with high pH tolerance (pH 9.5). ................................................... 66 Figure 4.1: Principal component analysis of RNA-Seq data from the gills of three strains of Rainbow Trout under control hatchery conditions (pH 7.2) or acclimated to pH 8.8. Each point represents an individual fish. Squares are Blackwater, circles are Eagle Lake and triangles are Fraser Valley Rainbow Trout. Black represents control conditions and blue represents fish acclimated to high pH conditions. n=6 per strain per treatment .... 92 Figure 4.2: Heat map displaying expression patterns of genes (140) with significant effects of pH when all strains are considered. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per strain per treatment) and each row represents one gene. The colour coded expression values (log2 counts per million) for each gene have been normalized to mean expression values of that gene. BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley; control = fish acclimated to standard hatchery conditions (pH 7.2); treatment = fish acclimated to pH 8.8 for 4 weeks. The number indicates the replicate number. ....................................... 93 Figure 4.3: Heat map displaying expression patterns of genes (18,036) demonstrating a significant interaction between pH and strain. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per strain per treatment) and each row represents one gene. The colour coded expression values (log2 counts per million) for each gene have been normalized to mean expression values of that gene. BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley; control = fish xv acclimated to standard hatchery conditions (pH 7.2); treatment = fish acclimated to pH 8.8 for 4 weeks. The number indicates the replicate number. ....................................... 94 Figure 4.4: Venn diagram illustrating the total number of differentially expressed genes for which a main effect of pH treatment was detected within each strain; red is Blackwater, green is Eagle Lake and blue is Fraser Valley. n=6 per strain per treatment ................ 95 Figure 5.1: Variation in short-term proportional fry survival across in-lake experimental treatments. Points represent predicted proportional survival from the DIC-selected model relative to actual survival. ............................................................................................ 123 Figure 5.2: Variation in short-term proportional yearling survival across in-lake experimental treatments. Points represent predicted proportional survival from the DIC-selected model relative to actual survival. ................................................................... 124 Figure 5.3: Variation in long-term fry recaptures across in-lake experimental treatments. Points represent predicted recaptures from the DIC-selected model relative to actual survival. ....................................................................................................................... 125 Figure 5.4: Variation in long-term yearling recaptures across in-lake experimental treatments. Points represent predicted recaptures from the DIC-selected model relative to actual survival. ........................................................................................................ 126 Figure A.1: Heat map displaying expression patterns of genes (8,053) with significant effects of pH in the Blackwater (BW) strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene. Control = fish acclimated to pH 7.2; treatment = fish acclimated to pH 8.8 for 1 month………………………………………………………………………………………….172 Figure A.2: Heat map displaying expression patterns of genes (10,509) with significant effects interaction between pH in the Eagle Lake strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene……………………………………………………………………………………….…..173 Figure A.3: Heat map displaying expression patterns of genes (3445) with significant effects of pH in the Fraser Valley (FV) strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with xvi expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene. Control = fish acclimated to pH 7.2; treatment = fish acclimated to pH 8.8 for 1 month……………………………………………………………………………………….....174 xvii Acknowledgements I would like to extend my profound appreciation and thanks to Trish Schulte, my supervisor, for taking a chance on me and providing amazing guidance and advice. It has been wonderful being mentored by you. I would also like to thank my lab mates Heather Bryant, Dillon Chung, Marina Giacomin, Taylor Gibbons, Tim Healy, Xiang Lin Dave Metzger, and Tara McBryan for their help with sampling, analysis and being amazing soundboards. My committee members, Bob Devlin, Jeff Richards and Colin Brauner have been a great source of advice and support. I appreciate the Freshwater Fisheries Society of B.C. willingness to support and partner with me on this project. In particular, I would like to thank Charlotte Lawson and Dusty Waite who worked tirelessly to assist me with coordinating, rearing and sampling all of the fish required for this study with the support of all of the staff at Fraser Valley Trout Hatchery. I want to extend my wholehearted appreciation to Chris Johnsen, without whom there would not have been a pH rearing system not to mention his continuous encouragement, you are missed and I will be forever grateful that I knew you. I would like to further acknowledge the FFSBC Science Division. I am very grateful for Adrian Clarke’s encouragement, flexibility and unwavering support in allowing me to pursue this degree, as well as, for my Research and Development Section colleagues support and picking up the slack while I was occupied. Although I have not mentioned all the other members of FFSBC by name it is not because their contributions went unnoticed. Thank you to each and everyone of you who helped me with sampling, fish rearing, fish transport, paperwork, late night snacks and encouragement. And last but by no means least I would like to thank my family and friends. My beloved husband Kyle for his patience, understanding, humour and love during the long hours and many frustrations. My mom for always encouraging me to learn and supporting me through everything. My siblings for being brilliant so I feel the need to challenge myself to keep up the family name. My friends for often having no idea why I’m ‘still’ in school but respecting and encouraging me none the less. 1 Chapter 1: General introduction 1.1 Recreational fishing The recreational fishery is a valuable contributor to the socio-economic development of Canada with >$8 billion attributable to the direct (e.g. license sales) and indirect activities (e.g. tourism) of anglers. This is almost five times the economic benefit of the commercial fishing industry in Canada (at $1.7 billion in 2010). In British Columbia alone, ~400,000 freshwater anglers are drawn to BC lakes annually contributing $546 million in direct benefits to the province’s economy and $411 million in indirect and induced economic impacts. The sports fishery in BC also generates $144 million in provincial and federal tax revenues (Bailey and Sumaila, 2013). In British Columbia, the freshwater fishing industry is supported largely through stocking of over 800 lakes by the Freshwater Fisheries Society of British Columbia (FFSBC). The FFSBC is a private non-profit organization dedicated to the enhancement and conservation of BC's freshwater fish resources. The province of British Columbia is divided into eight fisheries management regions for freshwater fisheries (Figure 1.1), and the FFSBC stocks lakes in each of these regions with ~8 million fish a year, of which ~5 million are various strains of Rainbow Trout (Oncorhynchus mykiss). The FFSBC Rainbow Trout stocking program is built on collecting and fertilizing eggs from adult trout reared in several brood lakes. The resulting progeny are then reared in a hatchery environment until their release into the designated lakes. The age of release varies from fry (1g) to catchable size (250g). These fish are released either as diploids (2n) or sterile triploids (3n). The success of lake stocking varies depending on the strain of trout, the size of the fish when they are stocked, and on the environmental conditions of the lake. One important environmental variable that affects stocking success is lake pH. Trout experience substantial mortality when exposed to water at pH 9.0 or above (Witschi and Ziebell 1979, Murray and Ziebell 1984, and Wright and Wood 1985). High pH exposure in fish interferes with the excretion of nitrogenous wastes, resulting in increased plasma ammonia levels, increased blood pH, and a reduced ability to regulate plasma ion levels 2 (Yesaki 1990, Wilkie and Wood 1991). Fish transferred from near-neutral hatchery water to high pH lake water thus face physiological challenges that can lead to death (Witschi and Ziebell 1979, Murray and Ziebell 1984, Wright and Wood 1985). At present, more than 17% (by lake number and 31% by surface area) of the lakes stocked by the FFSBC have a mean pH greater than 8.5 (Figure 1.2, Table 1.1), with pH increasing to 9.0 and greater for short periods of time. Many of these “high pH” lakes are heavily used by anglers, thus increasing pH is having a detrimental effect on the sports fishery in British Columbia because lower fish survival results in a reduction in the quality of the angler experience and a reduction in angler days, with associated economic impacts. In addition to these present-day challenges, the pH of British Columbia lakes has been rising over the last few decades (Appendix Table A.1), which has the potential to cause even greater impacts on the sports fishery across British Columbia. 1.2 High pH lakes in British Columbia To understand the chemistry of high pH lakes in British Columbia, it is important to understand the related concepts of pH and alkalinity. pH is defined as the negative logarithm of the molar concentration of hydrogen ions (H+) in a solution. The pH of a solution can range from acidic (at pH <7) to basic (at pH>7) with acidic solutions having a higher concentration of H+ than basic. Alkalinity is the buffering capacity of a waterbody and describes the ability of the solution to neutralize any acid inputs such as acid rain; in other words, it is the ability of a solution to resist changes in pH. Most buffering systems in natural waters (e.g. in lakes, streams, and the ocean) are based on carbonate and bicarbonate (CO3-2 and HCO3-), which is generally measured as the concentration of calcium carbonate (CaCO3). Alkalinity and pH are intertwined. If the alkalinity of water is high then any acid added will be buffered and pH will remain unchanged, whereas if alkalinity of water is low then there is nothing to counteract acid introduced by rain, snow melt, agricultural runoff etc. and the pH of the water will decrease. Conversely, pH can affect alkalinity, because H+ interacts with the buffer, reducing the available buffering. Basic solutions are often referred to as alkaline; however, basic (alkaline) solutions may or may not have high alkalinity. Anything over pH 7 is considered basic whereas what is defined as high pH depends upon the specific context. In the case of lakes and Rainbow 3 Trout, and for the purposes of this study, moderate pH is defined as 8.0-8.5; moderately high pH as 8.5-8.9, and high pH as anything over 9.0. High pH in lakes can be caused by several different phenomena such as the geological composition of the lake (basalt or limestone) which alters the pH of groundwater, climate (dry climate, low flushing rates, high evaporation rates) or agricultural disturbances (Kruger and Waters 1983; Jones et al. 1998). Even photosynthesis can alter the pH of a lake as photosynthesis utilizes H+ thus decreasing the H+ concentration in a solution and increasing the pH. Normally, respiration and decomposition will lower the pH and balance the system. Together, photosynthesis, respiration and decomposition are a cause of fluctuations in pH throughout the day; however, if algae blooms occur, rates of photosynthesis can be faster than the rate of respiration and decomposition and the pH level within a water body will increase (Talling et al. 1973). Most high pH lakes have low species diversity, but there are several species that have adapted to and thrived in these conditions such as Alcolapia grahmai in Lake Magadi (Wood et al. 1989, Wilson et al. 2014) and Lahontan Cutthroat Trout (Oncorhynchus clarki henshawi) in Pyramid Lake (Galat et al.1985). However, in other systems high pH has caused fish populations to decline. The occurrence of rising pH impacting trout fisheries is well-documented in Western North America in areas where native trout reside, rain shadow climates to the East of coastal mountains, basalt or limestone geology and human activity all occur together. Many regions in British Columbia have high pH lakes; the Cariboo, Thompson-Nicola, Okanagan and the Kootenays are the most affected (Table 1.2, Figure 1.1). In more recent years there have been changes in weather conditions, such as decreasing precipitation and higher summer temperatures, as well as agricultural water extraction that have exacerbated the natural high pH conditions of the lakes. In many of these lakes, the pH is now higher than what is tolerated by most salmonids (pH=10; Yesaki and Iwama 1992, Wilkie et al. 1993). Rising pH and TDS (total dissolved solids) levels have rendered environments challenging for indigenous salmonid stocks. This leaves productive lakes 4 at greater risk of illegal stocking of invasive fish that can tolerate higher pH (e.g centrarchids). 1.3 Physiological responses to high pH The primary physiological systems thought to be affected by altered water pH are those involved in ionoregulation, ammonia excretion, and acid-base regulation. Below I briefly summarize these systems and how they are affected at high pH. Fish must maintain the osmolarity and ion composition of their internal fluids (blood, extracellular fluids, intracellular fluids) within a narrow range to survive. In freshwater, fish body fluids have a higher osmolarity than the surrounding water so the fish passively lose ions and gain water from the environment. To survive, they must compensate by actively transporting ions from the environment into the body. Maintaining osmolarity and ion balance is energetically costly and requires specific intracellular machinery (Evans et al. 2014). To minimize this cost, fish have many different strategies to deal with changing ionic gradients in their environment. These mechanisms are largely accomplished at the gill. In Rainbow Trout, the gill is a multipurpose organ that is involved in gas exchange and nitrogenous waste excretion. It is also the primary organ for ionoregulation (Evans et al. 1999). The structure of the gill consists of four paired gill arches each of which consists of numerous gill filaments which contain thousands of lamellae. The gill epithelium covers both the filaments and lamellae and is composed of several different cell types including pavement cells (PVCs) which typically comprises 90% of the cell population, ionocytes (previously referred to as chloride cells or mitochondrion rich cells), mucous cells, and stem cells. The ionocytes are the main cell type involved in ion transport (Evans et al. 2014). Ionoregulation and acid-base regulation in fish are tightly linked as the intake and production of hydrogen ions (H+) and bicarbonate (HCO3-) are coupled with the exchange of sodium ions (Na+) and chloride ions (Cl-), respectively (Perry and Gilmour 2006, Gilmour and Perry 2009). The current model of the Rainbow Trout gill outlines two different pathways for apical Na+ uptake. The first begins with a Na+ ion entering the gill 5 in exchange for a H+ ion via an electroneutral Na+/H+ exchanger (NHE). The second has Na+ ions enter the gill through an epithelial Na+ channel (ENaC) in exchange for a H+ ion removed by H+-ATPase (Dumowska et al. 2012, Evans et al. 2014). However, this process can be constrained by environmental conditions such as low Na+ or low H+ conditions. Rhesus glycoprotein (Rh) is co-expressed with H+-ATPase and / or NHE and works as a functional metabolon (Wright and Wood 2009) to overcome environmental constraints. The specific mechanism involved in Cl- update and HCO3- secretion is still unclear; however, it is strongly suggested that the two are linked. Carbonic anhydrase, Cl- cotransporters, and H+-ATPase are required for the function of Cl-/HCO3- exchange and likely assist with overcoming similar environmental constraints as were placed on Na+/H+, but additional study is required to understand the specifics (reviewed in Evans et al. 2014) (Figure1.4). The two major metabolic acid products excreted across the gills of fish are carbon dioxide (CO2) and ammonia (NH3) with CO2 being excreted at a rate ten times higher than that of ammonia (Randall and Wright 1989). The gill is permeable to CO2 and NH3 but not to the ionic forms HCO3- and NH4+. As carbon dioxide is excreted across the gill it is hydrated within the cell by the enzyme carbonic anhydrase (CA) to produce H+ and HCO3- which is then transported (Gilmour and Perry 2009). The resulting H+ ions acidify the boundary layer adjacent to the gill (Wright et al. 1986, Randall and Wright 1989) which assists with ammonia excretion (Figure 1.4). Ammonia is toxic and, therefore, must be either excreted or converted into less toxic compounds such as urea or glutamine. Rainbow Trout, like most teleosts, are ammonotelic meaning they generally excrete nitrogen waste as ammonia and this occurs primarily in the gills. The understanding of how the teleost gill functions to excrete ammonia is continually evolving. Currently however, ammonia is thought to be removed by passive diffusion of NH3 across brachial epithelia through Rh proteins (Wright and Wood 2009). This method is dependent on protonation of acid boundary layer of gill epithelia by the H+ protons produced by the CO2 hydration, catalyzed by carbonic anhydrase reaction as described above or H+-ATPase (Weihrauch et al. 2009). In neutral waters, ammonia excretion may be enhanced by active exchange of external Na+ and 6 NH4+. This reaction is dependent on availability of H+ ions for conversion of NH3 to NH4+ (Yesaki 1990). Rhesus glycoproteins work with the NHE (Na+/NH4+) and mediate ammonia transport (Evans 2010, Figure 1.4). 1.4 Effects of high pH on fish Excretion of CO2 and NH3 is influenced by the composition of the water near the gill. Under 'normal' conditions the described mechanisms facilitate ammonia excretion as H+ in the gill boundary layer trap NH3 as NH4+ maintaining a low NH3 partial pressure. When fish are introduced to high pH conditions the gill acid boundary layer becomes stripped of protons (H+) impairing NH3 diffusion and ammonia levels rise in the body (Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1994, Wilkie et al.1996). The pK’ of the NH3/ NH4+ equilibrium is 9.58 and when water pH exceeds this value, NH3 becomes the dominant form of ammonia present in solution reducing the NH3 partial pressure gradient resulting in lower diffusion gradients further impairing ammonia excretion (Cameron and Heisler 1983). As body ammonia levels increase, it eventually becomes lethal as elevated NH4+ depolarizes neurons causing activation of N-methyl-D-aspartate (NMDA) type glutamate receptor. This leads to an influx of excessive Ca2+ and subsequent cell death in the central nervous system, causing convulsions, seizures and eventual death (Ip and Chew 2010, Wilkie et al. 2011). High pH also alters ion balance as reductions in H+ and HCO3- lower the rate of Na+/H+ exchange and Cl-/HCO3- exchange leading to reduced ion levels in the blood (Wilkie et al. 1999, Laurent et al. 2000). As mentioned there is a balance between CO2 and NH3 and their reactions. The midpoint between the pK’s of the two reactions (CO2/ HCO3- pK’ is 6.08) provides an adequate blood to water gradient to exist for CO2 and NH3 to be excreted (Randall and Wright 1989). However, high pH levels alters the gradient and creates a CO2 vacuum that decreases the blood CO2 tension and increases blood pH and HCO3- as a result (Johansen et al.1975, Wilkie and Wood 1991). To compensate for the increase in blood pH, the gill exchanges Na+ in the blood for H+ in the water, and the Cl- in the water for HCO3- in the blood, reducing blood Na+ and increasing Cl- perturbation of blood osmolarity (Yesaki and Iwama 1992, Wilkie and Wood 1996). 7 In general, as pH in water increases there is a high ion efflux across the gills and inhibition of ion influx, and, therefore, loss of ions in plasma (Wilkie et al. 1999). Gill damage results in increases in mucous production, tight junction loosening and degradation of the secondary lamellae (Daye and Gardside 1976, Tang and Goodenough 2003) leading to increased gill permeability and ion efflux across the gill (Wilkie and Wood 1994, Wilkie et al. 1996). To balance ions, Rainbow Trout alter gill permeability by increasing ionocyte density (Wilkie and Wood 1994, Laurent et al. 2000), but this compensation is not always sufficient and inhibition of ion uptake persists. It has been suggested that Rainbow Trout cope with short-term acute high pH exposure through their ability to counteract high pH-induced disturbances to ammonia excretion, acid-base homeostasis and electrolyte balance (Wilkie et al. 1996). Fish acclimated to high pH conditions tend to be able to regulate plasma ions more effectively in high pH conditions with significant differences in plasma Na+ and Cl- when compared to non-acclimated fish. However, acclimation does not provide total protection against high pH conditions (Yesaki 1990). Long-term survival of Rainbow Trout (Wilkie et al. 1996) in high pH environments is facilitated by higher steady state internal ammonia concentration, the development of a sustained compensatory metabolic acidosis which offsets decreased plasma CO2 and effective electrolyte balance. Not all fish species are affected by high pH in the same way. Lahontan Cutthroat Trout (Oncorhynchus clarki henshawi) thrive when stocked from pH 8.4 well-water into Pyramid Lake (pH 9.4). It was concluded that Lahontan Cutthroat Trout can avoid chronically elevated levels of plasma ammonia by reducing ammonia production and maintaining a higher internal concentration (Wilkie et al. 1994). In contrast, Tilapia (Oreochromis alcalicus grahami) excrete their nitrogenous waste as urea rather than ammonia when placed in high pH conditions (Randall et al. 1989b). This is the first known instance of complete ureotelism in an aquatic teleost fish. This mechanism allows this species to avoid the build-up of ammonia when exposed to high pH conditions. It is the animal’s ureotelism along with modification to gill functional morphology that allows it to survive in environments (pH 10) that most fish cannot (Laurent et al. 2001). 8 In other species, there have been circumstances where fish appear to choose high pH conditions despite other options being available. Perch (Perca fluviatilis) (Scott et al. 2005) released into a lake where both high pH (pH 9.9) and moderate pH (pH 7.56) conditions were available were found in both areas. However, fish from the high pH section of the lake had three times the ammonia levels in their blood but there were no disturbances to Na+ or Cl-. The radio tracking information did not indicate any changes in behaviour such as fish seeking refuge from high pH conditions indicating there may be some advantages (better foraging, less predation) that outweighed benefits of moving to more neutral parts of lake. The survival of fish in high pH waters is not exclusively related to pH but also dependent on other environmental conditions such as temperature and ionic content of the water (Emerson et al. 1975, Yesaki and Iwama 1992, Toth and Tsumura 1993, Mathias et al. 1995). Yesaki and Iwama (1992) exposed Rainbow Trout to chronic levels of pH 10 in hard water (CaCO3 320mg/L) and soft water. They found that higher CaCO3 resulted in fewer ionic disturbances by improving ammonia excretion through a Na+ related exchange mechanism; however, Thompson et al. (2016) did not see this possibly due to insufficient observation time (8 days vs. 24h, respectively). In summary, exposure to high pH has dramatic effects on Rainbow Trout. At pH levels above 9.5, ionoregulation can be impaired, resulting in accumulation of ammonia in blood, increased blood pH and a reduced ability to regulate plasma ion levels (Cameron and Heisler 1983, Thurston et al.1984, Wright and Wood 1985, Randall and Wright 1988, Wright, Heming and Randall 1988, Wright and Wood 1988, Yesaki 1990, Wilkie and Wood 1991) resulting in difficulties surviving owing to a change in the balance of H+ and HCO3- ions. Although much is known about the physiological response to high pH, the genetic and phenotypic basis of the variation in tolerance between Rainbow Trout strains remains poorly understood. 1.5 Rainbow Trout In this thesis, I focus on Rainbow Trout because of their importance to the recreational fisheries stocking program in British Columbia. Rainbow Trout, 9 Oncorhynchus mykiss, are a member of the family Salmonidae and are closely related to Pacific Salmon. As the glaciers receded, Rainbow Trout and other fish moved up into river and lake systems from glacial refugia south of BC, and from Haida Gwaii (formerly the Queen Charlotte Islands) to colonize previously unoccupied waterbodies (Taylor et al. 2011). Since that time, populations have evolved to adapt to a suite of conditions associated with their local physical habitat, community and available resources (e.g. larger heads and mouths in piscivorous Rainbow Trout in lake populations (Keely et al. 2005)). Many of these adaptations are at least partly genetically-based and heritable (Pollard and Yesaki 2004). The natural distribution of Rainbow Trout in North America is mainly west of the Rocky Mountains, from Mexico to the Kuskokwim River, Alaska (Scott and Crossman 1973). They have also been stocked in many countries around the world, including those in Europe, Asia, Australia, Africa, South America and in many regions of North America outside their native range. In British Columbia, Rainbow Trout are the most prevalently stocked species for the recreational fishery. Beyond recreation they are also a major food source across the world resulting in many countries developing hatchery systems to support the demand for these fish (Behnke 1972). The temperature optimum of Rainbow Trout ranges between 13oC to 18oC but they are often found at higher temperatures when there are refugia offering cooler, well oxygenated water nearby (Scott and Crossman 1973). Beyond temperature, Rainbow Trout can tolerate a wide variety of environments and water qualities; however, there has been increasing concern in recent years over their tolerance to high pH conditions. Many studies have concluded that there is a detrimental effect on survival rates of fish in environments where pH exceeds 9.0 (Witschi and Ziebell 1979, Murray and Ziebell 1984, Wilkie and Wood 1991, Yesaki and Iwama 1992). 1.5.1 Rainbow Trout strains Below I briefly summarize key characteristics of the Rainbow Trout strains used in this thesis, and discuss other strains relevant to British Columbia for which information on high pH tolerance has been collected. 10 Blackwater River Blackwater River is a 280-km river located northwest of Quesnel in central British Columbia. The Blackwater River Rainbow Trout strain is the one most widely used in BC for stocking in multispecies competitive lake environments. The strain is relatively fast growing and can be highly piscivorous (Pollard and Yesaki 2004); however, there is also a high degree of variability in its performance which has been repeatedly observed between lake types and brood years (Askey 2007, Northrup and Godin 2009). The tolerance of this strain to high pH in the laboratory has been previously investigated and it represents one of the most tolerant wild strains among those currently stocked by the FFSBC (Thompson et al. 2015 and 2016). Fraser Valley Domestic Fraser Valley Domestic Rainbow Trout were first developed in the 1940s in Tacoma, Washington and this strain was purchased by the Provincial Fisheries Program in the 1960s and moved to Abbotsford, BC. The strain was relocated to the Vancouver Island Trout Hatchery in Duncan, BC in 2011. These domesticated fish have a very fast growth rate and are often utilized in urban, high use fisheries, as well as lakes prone to winter kill (Pollard and Yesaki 2004). As they are domesticated, only sterile Fraser Valley Rainbow Trout are stocked to protect against interbreeding with wild fish. Fraser Valley Rainbow Trout are extremely aggressive and are not generally wary of predators (Biro et al. 2004); therefore, they are not stocked into lakes with high predation or mixed species competition. The tolerance of this strain to high pH in the laboratory has been previously investigated and has been reported to be high compared to that of other strains (Thompson et al. 2015 and 2016). Eagle Lake Eagle Lake Rainbow Trout (Onchorhynchus mykiss aquilarum) is a subspecies of Rainbow Trout that originates from Eagle Lake in California, a naturally high pH lake (pH 8.6-9.7, average 9.1). It is no longer a self-sustaining population, depending on hatchery 11 propagation. Deriving from a high pH lake suggests that this strain is likely to have a good ability to tolerate high pH conditions. Other strains and populations Pennask Lake is a 964-ha monoculture lake, located approximately 50 km east of Merritt, BC which has a near neutral pH of 6.7. Rainbow Trout from this lake are traditionally utilized in productive monoculture lakes (Pollard and Yesaki 2004). Pennask is one of the oldest and most prominently stocked strains and is, therefore, often used as a baseline for comparing wild Rainbow Trout strains in British Columbia. The pH tolerance of this strain has been investigated previously in the laboratory (Thompson et al. 2015 and 2016) and lake experiments and has been found to be low except when naturalized to high pH conditions (Yesaki and Tsumura 1992, Toth and Tsumura 1993, Mathias et al. 1995). Green Lake is a large (2307-ha) eutrophic lake in British Columbia about 25km north east of Clinton. The lake has been stocked with a variety of Rainbow Trout strains; however, mortality is high following stocking likely due to its historically high pH of 9.3 (Figure 1.3). Those fish that do survive and grow in this high pH environment have been thought to possess a genotype that allows them to adapt to the normally adverse conditions (Godin et al.1995). It is for this reason that Rainbow Trout from this lake have been utilized in many high pH strain comparisons. When they were last genotyped they most closely resembled Pennask origin fish (Nelson 2004). Stump Lake is a 666-ha lake located midway between Kamloops and Merritt, BC with an average pH of 8.8 (Figure 1.3). Stump Lake does not support a natural Rainbow Trout population but rather has been, and continues to be, stocked with a variety of Rainbow Trout strains. There is an artificial spawning channel on the lake which allows for natural recruitment in the lake. Due to its periodic levels of high pH, this population of Rainbow Trout has been evaluated similarly to Green Lake Rainbow Trout, for utilization in high pH lake stocking. Due to its continued stocking, natural recruitment and varying levels of pH, the population is a mix of first generation hatchery fish and potentially high pH adapted individuals. 12 1.6 Fisheries response to high pH Recreational fisheries play a prominent role in our economy and thus many efforts have been made in the past to understand and combat these rising pH conditions. Initially, many strains of Rainbow Trout (Yesaki and Tsumura 1992, Toth and Tsumura 1993) and sizes (catchable size trout have slightly higher survival rates than fry when stocked) were tested to increase stocking success. Different native species such as kokanee (Toth and Tsumura 1993, Godin et al.1994, Malange et al.1997, Godin and Tsumura 2001) were also stocked. As well, the stocking of non-native high pH tolerant species (Lahontan Cutthroat) was also attempted. Water conservation projects have been utilized to flush lakes whenever possible. Many years ago, the Province of British Columbia initiated a collaborative study to investigate the lack of substantial fish populations in high pH lakes. This study and several follow up studies examined the effects of decreasing transport stress and acclimation of Rainbow Trout pre- and post-stocking (Yesaki 1990, Yesaki and Toth 1992, Toth and Tsumura 1993). Additional studies investigated the progeny of parents that have naturalized in high pH lakes (Toth and Tsumura 1993, Mathias et al.1995, FFSBC unpublished), as well as comparing stocking sizes and times (Malange et al.1997). However, despite these initiatives, a long-term management solution for increasing survival in high pH lakes has not been found. 1.6.1 Acclimation It has been hypothesized that one of the greatest hindrances to survival in high pH lakes is the stress caused by transferring fish from nearly neutral hatchery environments directly into high pH lakes (Witschi and Ziebell 1979, Murray and Ziebell 1984). Yesaki and Tsumura (1992) directly transferred yearling Rainbow Trout from near-neutral hatchery conditions to Green Lake (pH 9.3) and observed a 10% mortality after 24h. Witschi and Ziebell (1979) observed a 32% mortality rate when fish were exposed to pH 9.5 in the first 24h. Thus, various methods have been utilized to decrease this initial shock: 1) acclimating the fish prior to exposure, 2) acclimation within the release waters, and 3) reducing transport stress. 13 Methods of acclimating fish to high pH conditions prior to exposure vary, as do their success in improving high pH tolerance. Murray and Ziebell (1984) performed two different types of acclimation: gradual and rapid. When Rainbow Trout were gradually acclimated from 8.0 to 10.4 over a five day period fish could withstand pH levels of 9.8. It is unknown how long fish could have survived at pH 9.8 given that pH levels were increased until mortalities were observed. These authors also performed a rapid acclimation from pH 8.0 to 9.5 over 6h where stress was observed at 12h and within 24h mortalities were observed, but no mortality was observed in five days when fish were acclimated rapidly to pH 9.3. This study demonstrated that higher pH levels can be tolerated if acclimation is performed gradually. Yesaki (1990) tested slow acclimation in the field by acclimating Rainbow Trout over six days until the pH in the rearing tanks matched the target (pH 9.5) and fish were then transported and stocked into the lake the following day. Differences in total mortality over 120h were large; survival of acclimated fish was 82.5% while survival of control fish was 20.4%. Acclimated fish could regulate plasma ions more effectively in high pH water; however, even though acclimated fish regulated plasma ions more effectively they still experienced a decrease in plasma ion concentration, meaning they did not have total protection against high pH conditions (Yesaki 1990). As mentioned previously, survival in high pH lakes is not likely exclusively related to pH. Wagner et al. (1997) evaluated survival and stress response of multiple Rainbow Trout strains after exposure to different temperatures along with high pH. The general conclusion was that it was a combination of high pH and extreme temperatures that resulted in the highest mortality. Acclimating fish to high pH in the laboratory or hatchery prior to exposure does not allow acclimation to other environmental conditions; therefore, several studies have investigated acclimation within the release waters. Once again, several methods of acclimation within lake have been used. The first, acidifying the lake, was undertaken using carbon dioxide which decreased the pH of an area of the lake. Once acidification was stopped, the region of the lake gradually increased to match the original conditions of the lake environment (Yesaki 1990). This method acclimates the fish to the lake within the lake after stocking rather than before stocking. Stocking into net pens that had been acidified decreased mortality in the first 144h. A second method, 14 in-stream acclimation, was performed by Yesaki and Tsumura (1992). Yesaki and Tsumura (1992) acclimated fish by placing fish into a net pen and moving the pen gradually (over 6h) through an inlet stream (pH 8.02) towards a high pH lake (pH 9.3). After 31h, fish mortality was 1% compared to 9% observed when fish were directly transferred from near-neutral conditions. Toth and Tsumura (1993) compared acclimation methods of direct transfer, in-transport and in-stream acclimation using Green and Pennask strains of Rainbow Trout. In-transport acclimation was done by placing fish in low-density transport tanks altering the initial transport water from pH 8.15 to 8.5 and then gradually increasing the water using NaOH to pH 9.15 over 8 hours while on route to Green Lake. The fish were then placed directly into the net pens within the lake. In-stream acclimation followed the Yesaki and Tsumura (1992) method mentioned above. Fish were checked for mortalities after 24h and 7 days. Only one mortality was observed in the first 24h. The non-acclimated fish experienced 8.6% mortality, while the in-transport treatment experienced 4.0% mortality, and the in-stream acclimated treatment experienced 4.9% mortality. Transport and handling stress are well-known challenges in fish culture. Wedemeyer et al. (1985) sought to mitigate these stresses and improved survival of fish by adding MS222 during transport. Wedemeyer et al. (1985) further suggested including mineral salts (CaCl2) in transport media to add protection against osmoregulatory imbalances during stressful transports. Yesaki (1990) tested this idea in high pH conditions by combining mineral salts with lake acidification, and found that by decreasing transport stress (by adding a mild anesthetic to the transport tank) and acidifying the lake one could obtain better survival rates in high pH conditions; 87.2% mortality occurred in the group using standard transport methods, 76.1% mortality in the group that underwent in-lake acidification and 58% mortality in the groups exposed to mild anesthetic and acidification (Yesaki 1990). In general, acclimation to high pH conditions does appear to provide an advantage to fish over those directly transferred into high pH lakes. Further to this, acclimation prior to release is far more practical than acclimation within release waters. However, it is not 15 known if rearing fish at high pH conditions during their entire life span would improve survival to an equivalent degree or better than observed with acclimation. 1.6.2 Strain Beyond modifications to rearing techniques for stocked fish it has been hypothesized that some strains of Rainbow Trout might have evolved mechanisms that confer higher tolerance to high pH environments (Dwyer 1990). Thompson et al. (2015) compared multiple strains of fish within the laboratory environment evaluating physiological mechanisms of acute high pH tolerance. Several strains of Rainbow Trout currently utilized by FFSBC were compared, including Fraser Valley Domestic, Blackwater River and Pennask, as well as several that have been utilized previously and are being evaluated for use in the BC stocking program: Tzenziacut and Carp Lake, respectively. None of these strains are from high pH sources. Time to loss of equilibrium (LOE) was compared as a proxy for high pH tolerance. During the first brood year fish were observed for 2 days and 3 days in the second brood year. In both years, the Fraser Valley Domestics demonstrated the best tolerance compared to the wild strains. Despite the variation in time to LOE all strains showed considerable ammonia accumulation in plasma, brain and white muscle. Wagner et al. (1997) found no physiological (glucose or chloride) or survival differences between two strains of Rainbow Trout when comparing them across a variety of pH and temperature treatments. Studying natural environments is difficult due to their ever-changing conditions, often remote locations, and limited control. Meanwhile, the advantages of laboratory studies are many, not the least of which is the ability to control environments and factors to be tested one at a time to determine the effects on organisms. That is why many of the studies on high pH tolerance in Rainbow Trout have taken place in a laboratory setting. However, natural environmental complexities cannot be fully duplicated in the laboratory and as such important factors may be overlooked when determining methods to increase survival in natural environments. As pH increases in lakes it is more than interest in the factors that allows some fish to survive while others perish that drives investigation into this phenomenon, but also a need to improve survival of fish within the 16 lake. For these practical purposes lake experiments must be conducted to ensure that results determined from laboratory study are representative of the natural environment. Although fewer in number, natural lake studies on high pH survival of Rainbow Trout have been very informative. Yesaki and Tsumura (1992) released Bootjack, Tzenzaicut and Pennask (all near-neutral lake strains) Rainbow Trout as yearlings (18.6-23.9g) into Lake 5567 (pH 9.2). Short-term net pen trials showed 0% mortality after 24h. Using gillnets, longer term survival was assessed. Twenty five percent of fish released into the lake were recovered across three sampling periods (4 months, 1 year, 16 months post release). The Bootjack strain was recovered at double the rate of the other strains at 4 months post stocking. By 16 months post release there was no significant difference between relative survival, but Tzenzaicut fish were significantly larger. This recapture rate in the high pH lakes was relatively low compared to recaptures of up to 70% in lower pH lakes suggesting higher mortality. Two naturalized populations were identified from high pH lakes (Green and Stump, pH 9.3 and pH 9.2 respectively) as a potential source for broodstock to develop strains for stocking into high pH lakes. Their progeny were released into high pH lakes alongside a strain from a near-neutral pH lake (Pennask) to determine if growth and survival differed in high pH lakes between fish whose parents were from high pH and near-neutral lakes (Toth and Tsumura 1993). At 16 months of age (4 months post release) a significantly higher number of Green Lake fish were recovered compared to Stump and Pennask. In addition, it was found that, Stump fish were significantly larger. Mathias et al. (1995) summarized the results of short and long-term survival across multiple natural high pH lake environments in comparing two naturalized populations of Rainbow Trout from high pH lakes, Stump and Green, to that of the near-neutral strain Pennask. Short-term net pen trials were quite variable across strains, but both Stump and Green fish survived better in all cases than did Pennask fish. After three years of stocking the cumulative catches in Pigeon Lake (Lake 5567) were higher (>32%) for Stump and Green Lake fish compared to Pennask fish (19%). The same pattern was observed in Till Lake; however, recovery rates were less than 3% for all strains. Overall the study 17 demonstrated that Rainbow Trout strains that have naturalized to high pH environments have a survival advantage over near-neutral strains with the greatest difference in survival occurring over the long-term. The naturalized Green Lake population was utilized for a short time to stock high pH lakes; however, with the continued development of Rainbow Trout strains for other fisheries management and environments new strains were available for comparison. Shortly after the introduction of Blackwater River wild source strain the FFSBC (unpublished) compared the strain to the naturalized Green Lake population. Two years of stocking into Pigeon Lake (pH 9.0) showed Blackwater River Rainbow Trout to have similar relative survival and weights to age to those of Green Lake fish. Short-term net pen trials in four other high pH lakes showed Blackwater River fish had higher relative survival rates in all lakes after 20 days. Mortality was observed in fish by day 4 in Johnson Lake (pH 9.07) with >80% and 60% survival for Blackwater fish and Green Lake fish, respectively, and no further mortality was observed throughout remainder of the experiment. Fraser Valley Domestic trout were found not to tolerate high pH in Kootenay region lakes (Oliver 2004), and additional lake assessments by other regional biologists concurred with these assessments (FFSBC unpublished). These studies conflict with the laboratory comparisons completed by Thompson et al. (2015) that indicated Fraser Valley domestics had the highest tolerance when exposed in short-term (48 and 72h) trials. Many of these studies indicate that there is substantial variation in pH tolerance among the strains of Rainbow Trout currently or historically utilized by the FFSBC. The variation in protocols and strains makes it difficult to compare across studies and even within studies as year to year variation exists (Thompson et al. 2015). Further to that it has been noted that short-term survival of Rainbow Trout can be lake specific (Mathias et al. 1995) and that size of the fish, changes in pH and/ or other water chemistry characteristics affects survival of strains differently. Direct comparisons between controlled laboratory and natural lake systems within a single study are required for a full understanding on what is leading to decreased survival in high pH lakes. 18 1.7 Thesis objectives The overall goal of my thesis was to examine the natural variation in pH tolerance within and between multiple strains of Rainbow Trout reared in different pH environments. My first objective was to follow up on previously observed differences among strains in Rainbow Trout high pH tolerance (Thompson et al. 2015) utilizing three strains of Rainbow Trout. I wanted to expand on the previous work by not only testing an additional Rainbow Trout strain but by assessing the degree of plasticity of high pH tolerance in Rainbow Trout. I sought to investigate if exposure to increased pH during development or short-term acclimation to high pH improves tolerance of high pH. To test this, I bred and reared three strains of Rainbow Trout (Blackwater River, Eagle Lake and Fraser Valley Domestics) in four different rearing conditions: 1) in near-neutral hatchery conditions; 2) at pH 8.5 from fertilization; 3) at pH 8.8 from fertilization; and 4) at near-neutral hatchery conditions from fertilization then acclimated over one week to pH 8.8 and held for one month at pH 8.8 prior to testing. Tolerance was assessed in response to acute exposure to high pH (9.5) and my results are outlined in chapter 2 of my thesis. My subsequent objectives focused on determining the mechanism behind the observed differences between strains and rearing treatments. I used whole-genome genotyping to identify genetic variation that may be associated with differences in high pH tolerance among strains (Chapter 3). Analyses of changes in gene expression in response to pH acclimation (Chapter 4) were used to identify candidate genes and processes that may be associated with phenotypic plasticity in tolerance. I also investigated if differences in genetic variation and/or phenotypic plasticity due to prior experience of high pH that were observed in the lab were also observed in natural lakes of varying pH. I did this by rearing two strains of Rainbow Trout (Fraser Valley and Blackwater) to two different life stages (fry and yearling) at three different rearing conditions: 1) in near-neutral hatchery conditions; 2) at pH 8.8 from fertilization; and 3) at near-neutral hatchery conditions from fertilization then acclimated over one week to pH 8.8 and held for one month at pH 8.8 prior to testing. These fish were stocked into six different natural lakes and survival was assessed over short time periods (18h, 72h and 1 week) using fish released into net pens and over longer time periods (4 months 19 and a year) by releasing fish into the lakes and then recapturing the fish utilizing gill nets (Chapter 5). The overall goal of my thesis research was to provide insight into the mechanisms underlying variation in pH tolerance in trout, and disentangle the effects of genetic variation and rearing environment on pH tolerance. With this information FFSBC and other agencies that stock Rainbow Trout, can make informed decisions about the best approaches to increase survival of stocked Rainbow Trout in high pH lakes. In summary, the primary hypotheses tested in my theses are: 1. That Rainbow Trout populations vary in pH tolerance (Chapter 2) 2. That Rainbow Trout have the capacity to improve tolerance of high pH exposure both through short-term plasticity (acclimation) and as a result of long-term acclimation at moderately high pH (Chapter 2) 3. That there is a genetic basis of observed variation in tolerance among strains (Chapter 3) 4. That gene expression in the gill will change in response to high pH acclimation in ways that are consistent with observed plasticity in pH tolerance (Chapter 4) 5. That laboratory assessments of high pH tolerance are accurate approximations of levels of tolerance under field conditions in natural lakes(Chapter 5) 20 1.8 Tables Table 1.1: Extent of occurrence of high pH conditions in the lakes stocked in British Columbia (FFSBC 2017), expressed as number of lakes and lake surface area # of Lakes Surface Area (ha) Lakes Stocked 894 60,516 Lakes Above pH 8.5 153 19,237 Lakes Above pH 9.0 60 11,415 Table 1.2: Average water pH in lakes by British Columbia management region (FFSBC 2017) Region Region Name Average pH 1 Vancouver Island 7.27 2 Lower Mainland 7.43 3 Thompson-Nicola 8.12 4E East Kootenays 8.32 4W West Kootenays 7.63 5 Cariboo 8.5 6 Skeena 7.37 7A Omineca 7.86 7B Peace 7.93 8 Okanagan 7.8 21 1.9 Figures Figure 1.1: Provincial freshwater fisheries management regions in British Columbia (FFSBC 2017) 22 Figure 1.2: All lakes stocked by FFSBC. Lakes with a pH less than 8.5 (black), lakes with a pH greater than 8.5 (green), lakes with a pH greater than 9.0 (purple). 23 Figure 1.3: pH measures in high pH lakes in British Columbia over time. Each point represents a single pH measurement. Data collection was halted in the 2000s until 2012. (FFSBC 2017) 24 Figure 1.4: Ammonia excretion model for freshwater fish (adapted from Weihrauch et al. 2009)25 Chapter 2: Variation in acute pH tolerance among and within strains of Rainbow Trout and between rearing environments 2.1 Introduction Elevated pH is physiologically challenging for teleost fish such as Rainbow Trout. Increases in water pH levels are known to have negative physiological effects on fish, including increases in ammonia levels, blood pH and reduced ability to regulate plasma ion levels (Cameron and Heisler 1983, Randall and Wright 1989, Thurston et al. 1984, Wright et al. 1988, Wilkie and Wood 1991, Wright and Wood 1985, Wright and Wood 1988, Yesaki 1990), and when left uncorrected these effects can lead to death. Over the last several decades the pH of many lakes in British Columbia has been increasing to the point that they are at or near the tolerance limits for Rainbow Trout, as mentioned in chapter 1, which has the potential to decrease the success of stocking programs in these lakes. This project aims to provide information on variation of acute high pH tolerance among Rainbow Trout strains and determine if there is phenotypic plasticity in tolerance that could provide opportunities to alter hatchery practices that could improve survival. Based on previous work by Thompson et al. (2015) it is known that there is substantial variation in acute high pH tolerance among the Rainbow Trout strains which are currently part of BC’s stocking program. Thompson et al. (2015) demonstrated that Fraser Valley Domestic Rainbow Trout had higher tolerance than did four other wild strains of Rainbow Trout; however, they also observed variation in tolerance between brood years. The Thompson et al. (2015) study differs from lake assessment information which showed higher survival rates of wild strains of Rainbow Trout, particularly the Blackwater River strain, in natural lake environments compared to the domestic Fraser Valley strain (FFSBC unpublished data). Natural lake assessments tend to be confounded by other factors so it is not surprising to observe incongruent results; however, this lack of consensus across the literature, coupled with the observed year-to-year variation, and the limited number of families tested in Thompson et al. (2015), suggests additional comparisons between the strains is warranted. 26 Although there is evidence that genetic variation may play a role in determining variation in pH tolerance, it is unlikely to be the only cause for the phenotypic variation in acute high pH tolerance. One cause of phenotypic variation can be environmental exposures early in development because of developmental plasticity (Pigliucci 1996, 2001). Fishes exhibit extensive developmental plasticity in response to a variety of environmental factors (Pittman et al. 2013), but little is known about the potential for developmental plasticity in response to environmental pH (Reddon et al. 2013). Developmental plasticity may be of particular interest in the context of obtaining high pH tolerant fish for lake stocking, because developmental plasticity often results in irreversible changes to the phenotype (Pigliucci 1996, 2001). However, it is also possible for developmental plasticity to cause deleterious changes to the phenotype, which would not be of benefit when stocking fish into high pH lakes. If adaptive developmental plasticity for pH tolerance exists in Rainbow Trout, rearing larvae in elevated pH conditions could provide a cost-effective solution to increasing survival in high pH lakes compared to selective breeding. Previous work on trout reared in high pH lakes suggests the possible existence of adaptive developmental plasticity for high pH tolerance in trout, since trout reared in high pH lakes survive better than other strains when stocked into high pH lakes (Mathias et al. 1995). Short-term acclimation to moderately altered environmental conditions also has the potential to cause phenotypic adjustments that could improve survival at environmental extremes. Indeed, it has long been known that fish acclimated to high pH conditions tend to be able to regulate plasma ions more effectively in high pH conditions with significant differences in plasma Na+ and Cl- concentrations when compared to non-acclimated fish (Yesaki 1990). These acclimations have been performed within the lake (Yesaki 1990), as well as within the hatchery prior to stocking (Murray and Ziebell 1984, Yesaki 1990) with successful results. To assess the relative extents of phenotypic variation among strains, the effects of developmental plasticity, and the effects of acclimation, we assessed acute tolerance of high pH (9.5) in fish from three strains of Rainbow Trout reared under normal hatchery conditions (pH 7.2), elevated pH conditions (pH 8.5 or 8.8) and fish reared at 7.2 from 27 incubation and then acclimated to pH 8.8 for 1 month. Of the strains tested by Thompson et al. (2015), Fraser Valley Domestics demonstrated the greatest tolerance; therefore, the Fraser Valley strain was chosen for comparison to investigate variation in acute high pH tolerance amongst Rainbow Trout strains and across multiple rearing environments. However, as this strain has been domesticated it is not suitable for lake stocking in diploid form due to the potential for interbreeding with wild fish. In addition, domestication has led to poor competitiveness in lakes containing other fish (Pollard and Yesaki 2004). Thus, the Blackwater strain was also chosen for analysis. Although this strain was found to be less tolerant to high pH than the Fraser Valley strain, it has good acute high pH tolerance when compared to most other wild strains in the lab (Thompson et al. 2015, 2016), and it may perform relatively well in high pH lakes (Godin et al. 1994). It is also the strain most widely used in BC for lake stocking in multispecies competitive environments. Finally, the Eagle Lake strain is the only known Rainbow Trout strain to have evolved within a naturally high pH lake: Eagle Lake, California. Deriving from a high pH lake suggests that this strain is likely to have a good ability to tolerate high pH; however, it is not approved for stocking in British Columbia waters. The Eagle Lake strain was utilized in the laboratory experiments to compare fish that have evolved in a high pH environment with those that have not. 2.2 Methods 2.2.1 Experiment 1.1: Effects of strain Fish spawning and rearing: The Blackwater 2n, Fraser Valley 2n and Fraser Valley 3n fish for use in this experiment were spawned in two groups. The first group was spawned in April/May of 2013 and were reared under standard hatchery conditions at Fraser Valley Trout Hatchery. Fraser Valley fish were generated by batch spawning of five males and five females, resulting in a mixed group of up to 25 families. Because the Blackwater strain contains more segregating genetic variation, five males and 30 females were batch spawned to produce a mixed group of up to 150 families. These fish were ready for use as yearlings in April/May 2014. The second group included Blackwater 2n, Fraser Valley 2n, Fraser Valley 3n and Eagle Lake 2n spawned in April/ May 2014. The second group was created to allow direct comparison with Eagle Lake fish that were 28 spawned at Crystal Lake Hatchery in California and only available in 2014. Fertilized eggs that had reached the eyed stage were transported to Fraser Valley Trout Hatchery and reared in identical hatchery neutral water as the other strains in April 2014.For this second group of fish, all strains were created from ten males and ten females. Acute high pH tolerance assay: Acute high pH Tolerance testing took place at two different life stages (fry and yearling) as these are the two stages used in the BC stocking program. The 2013 brood were tested at the yearling stage and the 2014 brood were tested at the fry stage. Tolerance testing occurred by abruptly transferring fish from standard hatchery water (pH 7.2) to hatchery water adjusted to pH 9.5 using NaOH. We did not attempt to replicate the ionic profiles of natural lake water, which are variable among lakes, because prior experiments have yielded conflicting results with respect to the effects of water hardness on high pH tolerance in Rainbow Trout. For example, Yesaki (1990) observed improved tolerance of high pH in hard water, whereas Thompson et al. (2016) found that hardness had little effect on tolerance. Fraser Valley Trout Hatchery well water is moderately hard (Table 2.1). Prior to tolerance testing all fish were fasted for 48h in their rearing tanks. Fish were transferred in groups of 30 individuals into 200L oval tanks and monitored visually until loss of equilibrium (LOE) was observed (Figure 2.1a). Each strain was differentially clipped and distributed across acute tolerance testing tanks. Acute tolerance testing conditions were achieved using the same method utilized for altering rearing conditions with the pH controller set to a higher pH (9.5). This pH level is close to the NH3/ NH4+ equilibrium constant (9.58) resulting in NH3 becoming the dominant form of ammonia in the water which would significantly impair passive diffusion of NH3 and suggesting that this pH level should be challenging for Rainbow Trout. The pH probe was calibrated every 12h and a pH reading was measured within each tank every hour by a handheld Oakton pH meter to ensure pH remained at pH 9.5. Following transfer, fish were observed every hour for LOE. LOE was determined as in Thompson et al. (2015) and was defined to have occurred when fish no longer had the ability to maintain normal dorso-ventral orientation. Once a fish was not able to maintain orientation it was removed and the time of LOE was recorded. Following tolerance assessment, the fish were 29 sacrificed and tissue samples taken for DNA extraction and genotyping (Chapter 3). The 2013 yearlings were observed for 48h and the 2014 fry were observed for 72h. 2.2.2 Experiment 1.2: Effects of high pH rearing Fish spawning and rearing: The 2014 brood of Blackwater 2n, Fraser Valley 2n and Fraser Valley 3n fish, described above, were divided and reared in multiple environments from incubation to tolerance testing; standard hatchery control conditions (~pH 7.2), and elevated pH conditions (pH 8.5 and 8.8). As pH > 9.0 is stressful to Rainbow Trout (Witschi and Ziebell 1979, Murray and Ziebell 1984, Wright and Wood 1985), fish were reared at lower pH levels (pH 8.5 and pH 8.8) to reduce additional stress, to determine whether this pH is sufficient to confer a degree of tolerance when fish are exposed to even higher pH (>9.0) levels at a later time. Eagle Lake fish could not be reared in altered conditions as they were not obtained until the eyed egg stage and were quarantined upon arrival at Fraser Valley Trout Hatchery. Two hundred individuals of each strain were reared in two 200L oval tanks for each condition. Control of rearing pH: The pH of the water was controlled using a probe (American Marine Inc. Pinpoint) placed into the center of each head tank. The probe was then connected to an American Marine Inc. Pinpoint pH controller which was set to the designated pH ranges for each header tank (Table 2.2). The controller activated and deactivated a Schlobster 1.1ml/min dosing pump that was connected to a barrel of 50% sodium hydroxide (NaOH) and an output into the header tank. The dosing pump was turned on when the probe indicated that the head tank levels fell below the pH criteria set for that tank. All tanks were flow through; therefore, an acid pump was not required to lower the pH as the NaOH was flushed quickly from the system when pH levels rose above the required range. Dissolved gasses were maintained by utilizing a modified airlift system to strip carbon dioxide and improve dissolved oxygen. To compensate for changing fish biomass and delayed pump reaction, the pH levels for each head tank were set higher than the pH required in the rearing tanks as the metabolic acids produced from the fish lowered the pH to the desired rearing pH. The rearing tanks were monitored daily to ensure pH, temperature and oxygen were in appropriate ranges with nitrogenous 30 wastes measured monthly. Fish were hand fed daily on a prescribed growth curve set to ensure healthy development. Acute high pH tolerance assay: Tolerance testing comparing fish that were reared at control pH (7.2), pH 8.5 and pH 8.8 was performed on the 2014 brood year of fry of Blackwater 2n, Fraser Valley 2n and Fraser Valley 3n. Thirty fish from each strain and rearing group were differentially marked by fin clip into a near-neutral water tank. Once recovered all were directly transferred to pH 9.5 testing tanks. Each testing tank contained all treatments from a single strain with each strain tested in a separate tank except for Eagle Lake 2n fish reared under control conditions that were placed in all tanks. (Figure 2.1b). Tolerance testing occurred for 72h as described in Experiment 1.1. 2.2.3 Experiment 1.3: Effects of acclimation Fish spawning and rearing: A subset of the 2014 Blackwater 2n, Fraser Valley 2n, Fraser Valley 3n and Eagle Lake 2n fish reared under standard hatchery conditions (pH 7.2) were used to test the hypothesis that short-term acclimation to moderately elevated pH could confer improved tolerance of extreme high pH conditions. Thus far, all acclimation work has involved short-term acclimation (3-6 days) to high pH (9.5); however, this pH level is stressful on fish and more difficult to maintain in a hatchery system. Therefore, fish were reared under neutral conditions, then acclimated to a moderately high pH (8.8) gradually over the course of a week (Table 2.3) and held at that pH for one month prior to tolerance testing at pH (9.5) to determine if acclimation to intermediate pH improves tolerance of exposure to extreme pH. Sixty fish of each strain were transferred from control tanks (pH) to acclimation tanks. Acute high pH tolerance assay: Tolerance testing was performed to compare fish that were reared at pH 7.2 (control), pH 8.8 and fish acclimated to pH 8.8 for one month prior to testing. Thirty fish from each strain and treatment group were differentially marked by fin clip. Once recovered, all were directly transferred from their rearing tanks to pH 9.5 testing tanks. Each testing tank contained all treatments from a single strain with each strain tested in a separate tank (Figure 2.1c). Tolerance testing occurred for 72h as described in Experiment 1.1. 31 All experiments were conducted according to approved UBC animal care protocol number A14-0103. 2.2.4 Statistical analyses Kruskal-Wallis one-way ANOVA was performed on all LOE data sets with time as the dependent variable and strain the independent variable. When tests were significant, a Dunn post hoc analysis with Bonferroni corrections was performed. Cox proportional hazard tests were performed on LOE data sets which included time, strain and treatment. This test is a survival model that relates the time of the event to one or more covariates. Size variation was compared using Mann-Whitney U test. All data were analysed using R Studio version 1.0.136. 2.3 Results 2.3.1 Experiment 1.1: Variation in acute high pH tolerance among strains There were no significant differences in mean time to loss of equilibrium (LOE) across replicate tanks within a strain, and therefore all data were pooled for analysis using a Cox proportional hazard test. This test showed a significant effect of strain for loss of equilibrium (LOE) in fry (2014 brood) when fish were exposed to pH 9.5. Similar results were obtained analyzing the data using Kruskal-Wallis one way ANOVA (p=4.93x10-15). Trout began to lose equilibrium at 12h with Fraser Valley having the greatest loss of equilibrium (Figure 2.2). Blackwater had the fewest fish that lost equilibrium followed closely by Eagle Lake with over 50% of the fish remaining in both strains. Due to the varying spawning dates and growth rates of the fish all strains varied in size at testing (Table 2.4). Mann-Whitney U tests illustrated that there was no significant effect of size between tolerant (non-LOE) and intolerant (LOE) fish (p=2.2x10-16) of each strain (Figure 2.3). When fish were tested as yearlings (2013 brood), a Kruskal-Wallis one way ANOVA showed no significant difference between strains in either of the two replicate experiments that were performed (p=0.6429 and p=0.126) (Figure 2.4). 32 2.3.2 Experiment 1.2: Variation in acute high pH tolerance between rearing environments Although we made rigorous attempts to maintain constant pH during the months of rearing, rearing pH did fluctuate during the growth cycle (Figure 2.5); however, there was a difference in pH between each treatment (control, pH 8.5 and pH 8.8). A significant effect of rearing environment was observed for acute high pH tolerance (LOE) in fry (Figure 2.6) and verified by cox proportional hazard (p=0). The greatest differences were found to be between pH 8.8 and control across all strains with fish reared at pH 8.8 having the highest percentage remaining at the end of each experiment and fish reared under control conditions having the lowest. Fish reared at pH 8.5 were found to experience LOE between the other two treatments. Rearing environment did not affect growth (p=0.7080), but there was a significant difference in growth between the strains (p=0.0322). Differences in growth rates between strains and ploidy have been previously observed for these strains by the Fraser Valley Trout Hatchery under neutral hatchery conditions (unpublished). There was no significant interaction between rearing treatment and strain for growth (Figure 2.7). 2.3.3 Experiment 1.3: Variation in acute high pH tolerance between rearing environments and acclimation A significant effect of acclimation on LOE was observed in fry (Figure 2.8) and verified by cox proportional hazard analysis (p=1.11x10-16). There was no difference detected in tolerance between pH 8.8 reared fish and acclimated fish. Variation in the percentage of fish remaining of each strain of Rainbow Trout as fry was observed across each experiment. The percentage of fish remaining in the Blackwater 2n group was 67%, 4% and 52% across experiments 1.1, 1.2 and 1.3 (Figure 2.2, 2.9 and 2.10, Table 2.4). Although the pH range was set the same for each experiment the actual pH for each ranged from 9.45 to 9.68 (Table 2.5). Across all experiments it was noted that LOE began shortly after 12h of exposure and generally leveled off by 40h, often earlier in fry. 33 2.4 Discussion 2.4.1 Variation in acute high pH tolerance between rearing environments and acclimation The most notable results of these experiments were that acclimation and rearing at pH 8.8 both resulted in increased acute high pH tolerance during acute exposure. It is not surprising that acclimation to higher pH levels would improve acute high pH tolerance as acclimation of fish to high pH conditions prior to exposure has been performed previously in both laboratory environments (Murray and Ziebell 1984, Yesaki 1990, Yesaki and Iwama 1992) and in lake environments (Yesaki 1990). Murray and Ziebell (1984) acclimated Rainbow Trout to pH 9.7-9.9 over a 5-day period and reported trout successfully acclimated to pH levels over 9.0, depending on the rate of acclimation. When acclimated gradually, an increase in acute high pH tolerance was achieved (pH 9.8) with fish showing no alterations to fish behaviour or feeding activity. Yesaki (1990) acclimated fish utilizing two different methods: 1) by acidifying the lake area with CO2 prior to release and 2) by adjusting the pH of rearing conditions at the hatchery by adding NaOH over six days until it matched the target pH (pH 9.5). Acclimated fish regulated plasma ions more effectively in alkaline water. However, even though acclimated fish regulated plasma ions more effectively they still experienced decreases in plasma ion concentration. Although both methods succeeded in improving survival, acclimation prior to release is far more practical than acclimation within a lake (by acidifying the lake). Although various methods of acclimating fish have been performed in previous experiments (Murrary and Ziebell 1984, Yesaki 1990, Toth and Tsumura 1993, Yesaki and Tsumura 1993), this study is the first time that the effects of rearing fish under high pH conditions have been examined. Furthermore, acclimating and rearing large numbers of fish to pH conditions such as 9.5 (as used in previous experiments Wright and Wood 1985, Wilkie and Wood 1991, Wilkie et al.1996, Thompson et al. 2015) can be difficult. Therefore, we reared and acclimated fish at a lower pH level (8.5 and 8.8) to determine if improved tolerance could be achieved without the stress and technical challenges of high pH levels. Our data show that acute high pH tolerance increased as rearing pH increased (Figure 2.6). Rearing and acclimating at pH 8.8 demonstrated equivalent results (Figure 34 2.8). It could be argued that rearing at pH 8.8 is simply a longer acclimation period; therefore, it is not surprising that tolerance levels are similar between the two treatments. Although rearing in elevated pH conditions is a more cost-effective solution to increasing survival in high-pH lakes compared to selective breeding, acclimating fish just prior to release is even more practical. Our results demonstrate improved survival in pH 9.5 experimental conditions when fish are acclimated and reared at high pH levels fish were only held at high pH for 72h so questions remain about long-term effects of acclimation. Epigenetics may be playing a role in tolerance differences between control fish, fish reared at pH 8.5, 8.8 and fish acclimated to pH 8.8. Mathias et al. (1995) demonstrated that Rainbow Trout strains that have naturalized to high pH environments have a survival advantage over near-neutral strains with the greatest difference in survival occurring over the long-term. External stresses have been shown to have long-lasting effects on development and vary depending on when during development the stresses occur (Feil and Fraga 2012). Further research has demonstrated that epigenetic variation can be independent of genetic variation and be inherited (Bossdorf et al. 2008) indicating that epigenetics could explain why tolerance increased when generations were reared in high pH environment. 2.4.2 Variation in acute high pH tolerance among strains The other notable results from these experiments were the mixed effect of strain at various life stages. There have been only a few studies comparing high pH tolerance between strains of Rainbow Trout (Godin et al. 1994, Mathias et al. 1995, Godin unpublished, Thompson et al. 2015). Mathias et al. (1995) found that progeny of fish that had been stocked and naturalized to high pH lakes had a survival advantage over fish from lower pH lakes. Godin (unpublished 2000) found that Blackwater Rainbow Trout had similar survival rates to fish from naturalized high pH lakes when compared in short-term lake trials. The most comparable study (Thompson et al. 2015) suggested that there is substantial inter-strain variation in tolerance, and that this phenotype is repeatable at an individual level. Thompson et al. (2015) found Fraser Valley Rainbow Trout to have the best tolerance of all strains tested during two separate trials which included 35 Blackwater Rainbow Trout. In contrast, we found that Blackwater River and Eagle Lake Rainbow Trout had greater tolerance to high pH conditions than did the Fraser Valley strain, when tested as fry. Eagle Lake Rainbow Trout have evolved in high pH conditions so it is not surprising that their tolerance was high, but we did not expect the Blackwater Rainbow Trout to have better tolerance than the Fraser Valley strain, given the results of the previous work on these strains. However, there were many differences between the two studies that could account for the variation in tolerance that was observed. 1) The tank set up in Thompson et al. (2015) was such that each strain was placed into small buckets with holes drilled in the sides. All the small buckets were then placed within a larger tank that was controlled for pH. The pH within the smaller buckets was not individually controlled or measured and as the density of each changed as fish were removed it is unknown if each had the same pH throughout the entire experiment. 2) Thompson et al. (2015) used a gradual increase in pH over the first 6h of the experiment rather than using abrupt transfer as we did this study. 3) The fish were size matched in Thompson et al. (2015) which likely caused selection between the strains as each strain’s growth rate is quite different, whereas we randomly sampled fish from each strain such that they were not size matched. 4) Finally, only four families were utilized by Thompson et al. (2015) which could mean that the results might not be representative of the tolerance of the entire strain. When testing tolerance at the high pH values that are on the edge of Rainbow Trout’s tolerance, small variations in environmental pH could have large effects on the phenotype (LOE). In the experiments presented in this chapter, variation was observed in both the number of fish that lost equilibrium and rankings between strains when comparing control reared fish across experiments (Figure 2.2, 2.9 and 2.10). Fish for all fry experiments were the same brood year, reared in the same conditions and tanks, and the same number was utilized for each trial. The fish were randomly selected for size, therefore, size varied between experiments (Table 1.4), although not to a statistically significant degree. Experiment 1.1 was designed to specifically test if there were differences in tolerance between the strains, therefore, each strain was mixed across multiple tanks. Experiment 1.2 and 1.3, however, were designed to test difference in rearing methods and acclimation, therefore, each strain was tested individually. Due to 36 fluctuations in pH (Table 2.5) it is difficult to compare between strains unless they are tested within the same tanks. Therefore, experiment 1.1 is the only experiment in which we can make direct comparisons between strains. The only replicate experiment was with yearlings, although there were differences in the number of fish that LOE and rankings among the strains both replicates demonstrated no significant difference between strains (Figure 2.4). Beyond the differences observed between strains there were also differences within strains in the fact that some fish of each strain were able to tolerate high pH conditions and some were not. Genetic differences within and between the strains could account for these observations. Previous information (FFSBC unpublished) has shown that the Fraser Valley strain is less genetically variable than Blackwater River strain at neutral loci. Thus, it is possible that Blackwater’s diversity provided more opportunities for tolerance resulting in a greater tolerance overall and also accounts for differences between this study and Thompson et al. (2015). Genetic comparisons between strains and between LOE and non-LOE will be discussed in the next chapter. Size variation could have been another possible reason that some fish lost equilibrium throughout the experiment while others remained unaffected. It has been speculated that larger fish may be able to regulate physiological disturbances better than smaller (Malange et al. 1997). Other studies that utilized larger fish (200-400g). (Wilkie and Wood, 1991, Wilkie et al. 1996, Wilson et al. 1998) have shown greater survival in high pH conditions than did this study and Thompson et al. (2015) and at higher pH conditions. Rainbow Trout in high pH environments have been shown to have altered ion exchange (Yesaki and Iwama 1992, Wilkie and Wood 1996) thus smaller fish could loss ions at a faster rate due to their higher metabolic and oxygen consumption rates. Size variation within the same experiment in this study was compared and found to have no effect when testing the same life stages (Figure 2.3); however, there were no direct comparisons between life stages. Brood year, size and developmental stage may explain the differences observed between this study’s fry and yearling experiments. A threshold size difference may have to be met for size to show an effect, and fish across a larger 37 size range and across developmental stages would have to be directly compared for this to be tested. 2.5 Conclusions The results of this study illustrate that rearing environment and acclimation condition play a stronger and more consistent role in setting acute high H tolerance than does strain. By rearing fish at pH 8.8 or acclimating to pH 8.8 prior to exposure all strains were better able to tolerant acute high pH (9.5) conditions. As rearing fish from incubation at alternate pHs is difficult, acclimation is the most practical method for increasing tolerance in Rainbow Trout. Comparing the genotypes of the different strains of Rainbow Trout which lost equilibrium to those that remained upright at the end of each trial may assist in determining what factors within the genome are responsible for the differences observed within and between the strains and is the focus of the next chapter. However, since acclimation can improve the performance of all Rainbow Trout strains tested thus far it would be advantageous to use this method in the future. Future experiments should investigate varying lengths of acclimation and pH levels to determine the optimal conditions for improved survival when fish are stocked into high pH lakes. Investigating what is occurring at the tissue level during acclimation might assist in determining what mechanisms are being utilized to counteract the increased pH in environment during acclimation and acute exposure. Further to this, although acclimation has proven to be effective in increasing survival when fish are exposed to high pH for a short duration in laboratory conditions, it is still unknown if it would have protective effects in natural lake conditions and for longer exposures. 38 2.6 Tables Table 2.1: Water chemistry parameters from Fraser Valley Hatchery Wells (FFSBC 2017) Parameter Measured Units Well 1 Well 1 Well 4 Well 4 Sampling Date 22-Jun-15 18-Jun-14 22-Jun-15 25-Nov-14 pH 7.68 7.65 7.97 Alkalinity Total as CaCO3 mg/L 73 71.2 97 Total Hardness CaCO3 mg/L 137 137 153 152 Dissolved Hardness CaCO3 mg/L 140 131 168 163 Conductivity uS/cm 329 331 359 Total Dissolved Solids mg/L 216 217 220 Orthophosphate mg/L 0.0078 0.0089 0.0052 Dissolved Sulphate SO4 mg/L 28 28.3 37.1 Bromide Br mg/L 0.032 0.027 0.037 Dissolved Boron mg/L 21 <50 12 <50 Dissolved Chloride mg/L 13 12 17 Nitrate mg/L 8.64 11.6 1.43 9.12 Nitrite mg/L 0.0028 <0.0020 0.0821 0.0733 Total Organic Nitrogen mg/L 1.16 <0.20 2.04 Total Nitrogen mg/L 9.8 11.5 3.56 Ammonia mg/L <0.0050 0.02 0.0075 Total Phosphorus mg/L 0.0083 0.0092 0.0049 0.0048 Dissolved Phosphorus mg/L 0.0092 0.0089 0.0056 <0.010 Dissolved Calcium mg/L 38.2 34.6 49.9 48.7 Dissolved Magnesium mg/L 11 10.8 10 10.1 Total Aluminum ug/L <0.50 1.87 1.39 <3.0 39 Table 2.2: pH settings for each head tank during rearing and acute high pH tolerance assay Header Tank pH Controller Range pH 8.5 8.38-8.62 pH 8.8 8.68-8.92 pH 9.5 9.38-9.62 Table 2.3: Controller pH settings for increasing pH during acclimation Day Controller Day 1 8.06-8.28 Day 2 8.24-8.46 Day 3 8.48-8.71 Day 4 8.68-8.91 Day 5 8.90-9.13 Day 6 9.00-9.24 Day 7 9.29-9.53 40 Table 2.4: Mean and standard deviation of weight and length of each strain of fish for each experiment. BW = Blackwater, FV = Fraser Valley and EL = Eagle Lake. n=30 for each strain in each experiment. Strain Mean Length (mm) SD of Length Mean Weight (g) SD of Weight % Remaining Experiment 1.1 (Yearling - 2013 Brood - Replicate 1) BW 2n 114.58 20.15 15.12 7.51 41 FV 2n 143.1 21.24 34.99 15.72 54 FV 3n 144.17 22.47 35.2 18.3 52 Experiment 1.1 (Yearling - 2013 Brood - Replicate 2) BW 2n 123.8 18.34 27.76 30.85 40 FV 2n 144.64 25.37 49.00 51.36 18 FV 3n 146.04 18.96 36.82 13.84 29 Experiment 1.1 (Fry - 2014 Brood) BW 2n 62.9 5.9 2.78 0.91 67 FV 2n 69.93 5.375 4.35 1.11 30 FV 3n 80.37 4.69 6.23 1.4 37 EL 2n 88.37 12.02 7.95 3.45 63 Experiment 1.2 (Fry - 2014 Brood) BW 2n 63.49 5.14 2.91 0.76 4 FV 2n 69.17 4.27 4.21 1.37 17 FV 3n 79.52 3.88 8.25 12.42 0 EL 2n 86.35 10.62 7.38 2.55 28 Experiment 1.3 (Fry - 2014 Brood) BW 2n 68.17 5.57 3.49 1.5 52 FV 2n 79.33 9.35 7.21 2.82 87 FV 3n 87.83 7.69 13.89 23.51 77 EL 2n 104.43 9.52 12.13 3.5 57 41 Table 2.5: Average testing pH during loss of equilibrium experiments. BW = Blackwater, FV = Fraser Valley and EL = Eagle Lake. Experiment 1.1 (Yearling) had 10 fish per strain per tank. Experiment 1.1 (Fry) had 6 fish per strain per tank. Experiment 1.2 and 1.3 had 30 fish for each strain. Strain pH SD of pH Experiment 1.1 (Yearling - 2013 Brood - Replicate 1) Tank 1 9.47 0.08 Tank 2 9.37 0.07 Tank 3 9.24 0.11 Experiment 1.1 (Yearling - 2013 Brood - Replicate 2) Tank 1 9.49 0.08 Tank 2 9.52 0.07 Tank 3 9.5 0.06 Experiment 1.1 (Fry - 2014 Brood) Tank 1 9.55 0.1 Tank 2 9.56 0.11 Tank 3 9.58 0.11 Tank 4 9.68 0.14 Tank 5 9.68 0.13 Experiment 1.2 (Fry - 2014 Brood) BW 2n 9.51 0.07 FV 2n 9.53 0.07 FV 3n 9.67 0.1 EL 2n Spread between above three tanks Experiment 1.3 (Fry - 2014 Brood) BW 2n 9.56 0.16 FV 2n 9.45 0.15 FV 3n 9.47 0.13 EL 2n 9.54 0.27 42 2.7 Figures 43 Figure 2.1: Summary of experimental design for a) Experiment 1.1: 6 fish were transferred from each strain’s individual rearing tank to the test tank for a total of 24 fish in each tank and 30 fish of each strain across the five testing tanks, b) Experiment 1.2: 30 fish were transferred of each strain from the three different rearing containers into a testing tank, with one testing tank per strain for Blackwater, Fraser Valley 2n and 3n. 10 Eagle Lake fish that had been reared at pH 7.2 were also transferred into each of the testing tanks. The Blackwater strain is shown as an example for the rearing tanks, and the other strains were transferred in the same fashion, c) Experiment 1.3: 30 fish of each were transferred from the three different rearing containers (with the exception of Eagle Lake which did not have a pH 8.8 reared group) into a testing tank such that there was one testing tank per strain for Blackwater, Eagle Lake, Fraser Valley 2n and 3n. The Blackwater strain is shown as an example for the rearing tanks, and the other strains were transferred in the same fashion. BW = Blackwater, FV = Fraser Valley, and EL= Eagle Lake 44 Figure 2.2: Effect of strain on tolerance of acute transfer to pH 9.5. The data shown are percent of fish remaining as a function of time exposed to pH 9.5 (Experiment 1.1 Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW = black square), Eagle Lake (EL = red circles), Fraser Valley (FV2n = green triangles, and FV3n = blue diamonds) (5 tanks with 6 individuals each resulting in a total n=30 for each). Error bars indicate standard error. Significance is indicated by the letter to the right of each strain in the graph. (h) a a b b 45 Figure 2.3: Size distribution of all fish separated by strain and state (exhibiting loss of equilibrium (LOE) or not) at the end of the acute tolerance assay from experiment 1.1 fall fry 2014 brood (White – fish that LOE; Red - fish that did not LOE). BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley. Error bars indicate standard error. n=30 for each strain. There were no significant differences in length between fish that exhibited loss of equilibrium (LOE) and those that did not for any strain. 46 Figure 2.4: Percent of fish remaining as a function of time exposed to pH 9.5 Data are for Experiment 1.1 (Spring Yearlings 2013 brood) comparing Rainbow Trout strains Blackwater (BW), Fraser Valley (FV2n and FV3n) (3 tanks with 10 individuals in each resulting in a total n=30 for each) a) replicate 1, b) replicate 2 (BW = black squares, FV2N = green triangles, FV3N = blue diamonds). Error bars indicate standard error. There was no significant difference found between the strains in either replicate. (h) (h) 47 Figure 2.5: Average daily rearing pH for each rearing group (pH 8.8 = green; pH 8.5 = red; control = black) during the length of 2014 brood growth cycle. 48 Figure 2.6: Effect of rearing water pH on Rainbow Trout tolerance to acute transfer to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5. Results are from Experiment 1.2 (Fall fry 2014 brood) comparing fish reared at pH 8.8 (green triangles), pH 8.5 (red squares) and control conditions (~pH 7.2) (black circles). Rainbow Trout strains a) Blackwater (BW), b) Fraser Valley (FV2n), and c) FV3n (n=30 for each). Significance is indicated by the letter to the right of each strain in the graph.(h) (h) (h) a a b b c a b c c 49 Figure 2.7: Length of fish tat the end of rearing by for treatment groups in three strains (white = control well water pH 7.2; red = reared at pH 8.5; green = reared at pH 8.8). BW = Blackwater, FV = Fraser Valley. Error bars indicate standard error. n=30. There were no significant differences in length among treatment groups for any strain 50 Figure 2.8: Effect of acclimation to pH 8.8 on high pH tolerance following acute transfer to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5 for Experiment 1.3 (Fall fry 2014 brood) comparing fish reared at pH 8.8 (except for EL) (green), acclimation to pH 8.8 (blue) and control conditions (~pH 7.2) (black). Rainbow Trout strains a) Blackwater (BW), b) Eagle Lake (EL), c and d) Fraser Valley (FV2n and FV3n, respectively) (n=30 for each strain and treatment). Significance is indicated by the letter to the right of each strain in the graph. Note: The data for acclimated fish and fish reared at pH 8.8 often overlap which obscures one of the two treatments. (h) (h) (h) (h) a,a b a b a,a b a,a b 51 Figure 2.9: Effect of strain on high pH tolerance following acute exposure to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5 Results from Experiment 1.2 (Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW), Eagle Lake (EL), Fraser Valley (FV2n and FV3n) (n=30 for each) all reared in control conditions. (h) 52 Figure 2.10: Effect of strain on high pH tolerance following acute exposure to pH 9.5. Data are expressed as percent of fish remaining as a function of time exposed to pH 9.5. results from Experiment 1.3 (Fall fry 2014 brood) comparing Rainbow Trout strains Blackwater (BW), Eagle Lake (EL), Fraser Valley (FV2n and FV3n) (n=30 for each) all reared in control conditions. (h) 53 Chapter 3: A GWAS analysis of the genetic basis of high pH tolerance across three strains of Rainbow Trout 3.1 Introduction Over the last several decades, a number of lakes in British Columbia have undergone increases in pH, which has the potential to reduce the success of Rainbow Trout stocking programs (Tredger 1990, FFSBC 2017). One possible solution to this problem would be to breed a strain of Rainbow Trout with increased tolerance of high pH for stocking into these lakes. Previous studies (Yesaki and Tsumura 1992, Toth and Tsumura 1993, Mathias et al. 1995, Thompson et al. 2015) including my own (Chapter 2) have demonstrated that there is substantial variation both within and among strains of Rainbow Trout for tolerance of high pH, suggesting that developing a high pH tolerant broodstock could be possible. Marker-assisted selection (MAS) is a practice by which genetic markers linked to desirable traits are used to guide breeding decisions for agricultural or aquacultural species. For MAS to be useful, genetic markers that are closely associated with the desired phenotypes must first be identified within the genome. Recent advances in high throughput sequencing have made it possible to perform genotype-phenotype association studies without the requirement of a priori markers (e.g. Davey et al. 2011, Keller et al. 2013, Narum et al. 2013), and the sequencing of the Atlantic Salmon (Davidson et al. 2010) and Rainbow Trout (Berthelot et al. 2014) genomes have made the investigation of genotype-phenotype associations at a whole-genome scale possible in salmonids. For example, Atlantic Salmon studies have identified genetic markers associated with growth rate and age of sexual maturation (Gutierrez et al. 2015, Tsai et al. 2015) and resistance to sea lice (Tsai et al. 2016). While in Rainbow Trout the genetic basis for migration (Hecht et al. 2013), body weight traits (Gonzalez-Pena et al. 2016) and disease resistance (Campbell et al. 2014, Liu et al. 2015) have been examined, many of which are key aquacultural traits. Currently, whether there is genetically-based variation in tolerance of high pH within and among Rainbow Trout strains is unknown. 54 The goal of this study is to identify genetic markers associated with this variation that could potentially be used for marker assisted selection. Various methods have been developed over the last few decades that have advanced our ability to determine the genetic factors that cause various observed phenotypes. One such method is Quantitative Trait Locus (QTL) mapping and a second is Genome-Wide Association Studies (GWAS). Each method has advantages and limitations and can be used as complementary methods. The primary difference between QTL mapping and GWAS is that QTL mapping investigates the association between genotype and phenotype in families generated from crosses between parents with differing phenotypes. In contrast, GWAS examines the association between genotype and phenotype in a population generally of unrelated individuals (Mackay et al. 2009). GWAS aims to genotype a broad range of markers across the genome so that functional alleles will be in linkage disequilibrium with at least one of the genotyped markers (Myles et al. 2009). It provides the capability to integrate genomic data with phenotypic differences among individuals and populations to identify genetic variation associated with key phenotypes (Funk et al. 2012). The benefit of GWAS is that both neutral and adaptive loci can be investigated simultaneously, each of which provide different insights so when combined the most effective for management decisions can be reached (Funk et al. 2012). However, GWAS works most effectively when a trait is caused by a small number of loci with large effect sizes (Korte and Farlow 2013). Rare phenotypic variants and small effect size can limit how informative GWAS can be. In contrast, QTL mapping can allow the detection of genes associated with rare phenotypic variants, and has the potential to detect variants of relatively small effect size. However, QTL mapping requires large families (and ideally multiple families and multiple generations) and this can be a limiting factor for many studies, particularly when dealing with wild populations and species with long generation times (Korte and Farlow 2013). In addition, QTL mapping typically detects larger regions of the genome associated with a particular phenotypic variant, because of limited recombination, and subsequent fine-55 scale mapping studies in advanced generation backcrosses are required to detect causal genetic variants. Mapping studies in multiparental populations occupy an intermediate niche between QTL mapping of biparental populations and association mapping, and may be particularly useful in cultured fish species, such as salmonids, where generation times are long. Here we apply this approach to determine whether there are genetic markers associated with the variation in pH tolerance I observed in chapter 2. Genomics tools are found to be most beneficial in cases, such as Rainbow Trout, where differentiation among populations is expected (Keeley et al. 2005). Therefore, to ensure that I had a greater chance of detecting genes associated with variation in pH tolerance, I have chosen to genotype fish from multiple strains. Eagle Lake Rainbow Trout have evolved in a high pH lake; therefore, Eagle Lake Trout are less likely to have variation for pH tolerance within the strain. However, Blackwater and Fraser Valley strains have not evolved in high pH environments, but they have been shown to vary in pH tolerance in previous studies (Thompson et al. 2015), thus variation for pH tolerance should be observed when comparing between populations. The ability to detect this variation at a genetic level will depend on many factors including the effect size and the frequency of the alleles in the populations I am examining. 3.2 Methods Spawning, rearing and high pH tolerance testing: Genetic (fin clips) samples were collected following the high pH tolerance testing in chapter 2. All experiments were conducted according to UBC approved animal care protocol number A14-0103. Of the fish tested I selected a total of 48 fish from each strain across the three experiments (Appendix Table A.2) choosing individuals that had lost equilibrium as well as those that remained at the end of the 72 hours. Each fin clip was weighed and approximately 10mg were removed for DNA extraction using the Qiagen DNAeasy Kit (Qiagen Sciences Inc., Germantown, MD). Modifications to the protocol included gentle inversions in place of the vortexing steps to prevent DNA shearing, as well as, extending the proteinase K digestion step overnight to 56 ensure full digestion, and the addition of 5 µL of RNAse Cocktail (Thermo Fisher Scientific) to each sample prior to extraction to prevent RNA contamination. To ensure that DNA was of sufficient quality two additional steps were taken before genotyping by sequencing. All samples were quantified using a Quant-it Pico Green kit with fluorometer and aliquots of 10% percent of the samples were digested with restriction enzyme HindIII for 3 hours at 37oC to confirm the isolated DNA was intact and capable of effective digestion. Samples were diluted to 100ng/µL prior to submission to the Genomic Diversity Facility at Cornell University for genotyping by sequencing (GBS). Genotyping-by sequencing: GBS (Elshire et al. 2011, Davey et al. 2011) was used to genotype hundreds of thousands of genetic markers (single nucleotide polymorphisms, or SNPs) throughout the genome. GBS has been successfully utilized to identify genotype-phenotype associations in trout (Hecht et al. 2013), and is thus ideally suited for the detection of genetic variants associated with differences in high pH tolerance. Genotyping and SNP discovery was performed at the Institute for Genomic Diversity facility, Cornell University using a genotyping by sequencing (GBS) approach as described in Elshire et al. (2011). Genotype by sequencing (GBS) is a simple, highly multiplexed system for constructing libraries for next generation sequencing which include both GWAS and QTL mapping. Briefly, each sample was cut using HindIII restriction enzyme followed by the ligating an individual barcode to one end and a common adaptor to the other fragment end before pooling all the samples. Primers which were complimentary to adaptors and contained a region which was complimentary to the flow cell were used to amplify the fragments through PCR. A HiSeq 2500 next generation sequencer was used to sequence the fragments. Bioinformatics and statistical analyses: Analysis of the raw sequence data was performed using the TASSEL-GBS discovery pipeline defined in Glaubitz et al. (2014). In general, this pipeline utilizes the raw sequencing data from next generation sequencing and filters out the ‘good barcoded reads’: those that contain 64bp and have no unknown base pair reads in the sequence. It then sorts all of the reads into unique sequence ‘tags’ that may then be mapped to a reference genome. In this study, all data were filtered so that a minimum of 3 reads were available for each tag. SNPs were called utilizing the 57 Rainbow Trout genome assembly available at https://www.genoscope.cns.fr/trout/ (Berthelot et al. 2014). To ensure quality control all individuals were removed which had >60% missing data. Following this all SNP sites with >80% missing, >2 alleles and with minor allele frequencies less than 5% were removed. Rainbow Trout have undergone a recent genome duplication event which has resulted in numerous genetic paralogs within their genomes, only half of which have been lost, resulting in many remaining as almost identical duplicate copies (Berthelot et al. 2014). These near duplicates can cause errors in site mapping leading to inaccurate SNP calls. The final filtering step utilized a method developed by McKinney et al. (2016) for removing these errors. It identifies paralogs by expected heterozygosity frequency and deviations from expected singleton loci allele frequencies. Following filtering, the data were uploaded into TASSEL (Trait Analysis by aSSociation, Evolution, and Linkage) version 5.2.37 along with a trait file containing time to loss of equilibrium and weight for each fish. A general linear model (GLM) was run which included a principal components analysis (using 5 components) to account for population structure, DNA sequences and phenotypic data and a mixed linear model (MLM) which included the latter plus a kinship matrix to account for relatedness. Each model’s fit to the data was analyzed using a qq-plot. The qq-plots indicated a similar fit for both the GLM and MLM, with a MLM having a slightly better fit. Corrections for multiple comparisons were done at both the genomic and chromosomal level using the raw p-values produced from the MLM and FDR (false discovery rate) tests (Benjamin and Hochberg 1995) in R Studio version 1.0.136. Those SNPs found to be significant following FDR corrections were identified by using BLAST to compare the sequence of genes close to (within 100kbp) or encompassing that SNP against known genes in other species of fish. 3.3 Results and discussion Following data filtration 42,476 SNPs were identified among these individuals. The SNPs were broadly distributed across the Rainbow Trout genome sequence (Figure 3.1). 58 Note that the currently available version of the Rainbow Trout genome sequence (Berthelot et al. 2014) has not been fully assembled into chromosomes, and there is a large scaffold containing incompletely localized sequences, complicating the analysis of genotype-phenotype associations in this region. Using this genome sequence, we detected several SNPs that are putatively associated with variation in tolerance of high pH in Rainbow Trout. When a genome-wide FDR correction was applied, there were no SNPs which had a significant association with variation in high pH tolerance (Table 3.1, Figure 3.2). As an exploratory approach, we also conducted a chromosome-by-chromosome analysis, performing FDR-correction at the chromosome level. This analysis revealed two SNPs as significantly associated with variation in tolerance of high pH on chromosomes Un_29 and Un_26 (Table 3.2, Figure 3.2b). Genomic BLAST indicated that one of the significant SNPs was within a gene closely matched to Oncorhynchus kisutch polycomb group RING finger protein 3. How this gene affects high pH tolerance is unclear. The other SNP that we identified was not near a gene. There are two possible reasons for the limited detection of associations between genetic variants and phenotypic variation in this study: 1) inter-individual and inter-strain variation in high pH tolerance in Rainbow Trout is not genetically based, or 2) our study design did not have sufficient power to detect the relevant genetic variation. As can be seen from the results presented in chapter 2, high pH tolerance displays substantial phenotypic plasticity in Rainbow Trout, and thus it is possible that differences among strains and among individuals in pH tolerance could be the result of various forms of plasticity, including epigenetic variation or maternal effects. If this is the case, it is unsurprising that we detected limited genetic associations with variation in high pH tolerance. Alternatively, it is possible that variation in pH tolerance has a genetic component that we were unable to detect. As described in the general introduction of this thesis, high pH exposure is known to cause physiological response in multiple biochemical pathways 59 throughout the body, primarily those involved in ion and acid-base regulation (Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1994, Wilkie et al. 1996). This suggests that many biochemical processes may be involved in specifying tolerance of high pH, and thus many genes may be associated with this trait each having a small phenotypic effect. The genetic basis of polygenic traits is notoriously difficult to identify. In a typical association or QTL study using genotyping-by-sequencing, many thousands of SNPs are tested for associations with the phenotypic trait. As a result, the threshold for statistical significance is high and is typically chosen on the basis of ensuring a low ‘false discovery rate’. This means that a much higher threshold for significance is applied than would be used in a case where an association is tested between genetic variation in a single candidate locus and the phenotype of interest. Thus, for sample sizes that are practical in an aquaculture setting, only SNPs with large effects can be identified with statistical confidence. Thus, it is possible that many SNPs have effects on the trait, but that these effects are too small to be demonstrated. This issue has been thoroughly explored in the case of variation in human height. Classical genetics studies have suggested that approximately 80% of the variation in human height has a genetic basis (Brookfield 2013), but GWAS for human height have only identified SNPs that could account for a small fraction of this variation, in what has been called the “missing heritability problem” (Vineis and Pearce 2010). Recently, a number of studies have suggested that if variation at all of the SNP loci (whether detected as significantly associated or not) is considered together, a much greater proportion of the variation in the trait can be explained (Yang et al. 2010). These data suggest that there are many loci with additive effects on the trait in addition to those that reach a threshold for significance. In addition, these human GWAS typically use much larger sample sizes (in the thousands) than I used here (143), and thus my power to detect associations is even lower than in the studies discussed above. The cost and design of this experiment minimized the number of samples that were available to be genotyped. Although thirty fish from each strain for each experiment were run (Chapter 2) and samples were utilized across experiments not all fish were susceptible to acute high pH tolerance conditions which limited the number of fish 60 available to genotype. That along with the cost resulted in only 143 fish genotyped across three different strains and experiments. Three different strains of Rainbow Trout were assessed in this study. This population structure could act as a confounding factor; however, the qqplot (Figure 3.3) showed a very clear line, suggesting that our statistical model adequately controlled for population structure. The physiological acute tolerance trials showed variation within strains so it is likely that pH tolerance is at the individual level rather than the strain level. If any strain were to have fixed effects for high pH tolerance it should be Eagle Lake due to its history of exposure and evolution within a high pH environment. However, variation was observed in the Eagle Lake strain meaning not all fish were able to handle high pH conditions and the tolerance level was shown to be equivalent to the Blackwater strain when directly compared (Chapter 2 experiment 1.1). It becomes difficult to detect associations as the number of markers examined increases as the p-value required to show significance becomes very low. The number of SNPs (42,476) that were discovered was much higher than most Rainbow Trout GWAS studies (~4500 SNPs Campbell et al. 2014; ~3500 SNPs Hecht et al. 2014; and ~7800 Liu et al. 2015) resulting in little power to detect associations which might be present. Finally, rare alleles variants which were found to have a frequency of less than 5% were removed prior to analysis. If rare variants were associated with high pH tolerance then they might not have been present given our sample size or would have been removed prior to analysis. If sample sizes were increased dramatically another GWAS analysis could be run but given the current knowledge it would likely be more advantageous to focus on other aspects of high pH tolerance. In conclusion, a specific genetic association to high pH tolerance in Rainbow Trout was not found. Although two SNPs were found when exploring the data by chromosome their location within the genome did not provide any insight into the mechanisms associated with high pH tolerance. It is likely that high pH tolerance is a polygenic trait which would make it a poor candidate for marker assisted selection; however, a larger 61 sample size and more stringent filtering might aid in determining the specific mechanisms involved. 62 3.4 Tables Table 3.1: SNPs identified as significantly associated with high pH (pH 9.5) tolerance for Blackwater, Eagle Lake and Fraser Valley Domestics reared in near neutral well water (pH ~7.2) Marker Chromosome Raw P Value Genome FDR Chromosome FDR Distance to SNP (bp) BLAST Sequence Match Description S1_1717917424 chrUN_29 2.20E-06 0.09343 0.0015 Within gene Oncorhynchus kisutch polycomb group RING finger protein 3 (LOC109906020), mRNA S1_1663560086 chrUN_26 0.000162 0.80691 0.0405 not near gene 63 3.5 Figures Figure 3.1: Number of SNPs per chromosome. The current version of the Rainbow Trout genome sequence (Berthelot et al. 2014) has not been fully assembled into chromosomes so there are an additional19,965 SNPs that were found on the large scaffold containing incompletely localized sequences that are not shown here a) Illustrates the SNPs found with specific areas on a chromosome. b) Illustrates the number of SNPs found to be on specific chromosomes but unknown areas on those chromosomes. 64 a) b) 65 c) Figure 3.2: Plot of non-FDR corrected P-values for SNPs associated with high pH (9.5) tolerance in fish reared in near neutral well water (pH ~7.2) n=143. No SNPs were identified that were significantly associated with high pH tolerance with whole-genome level FDR correction. Red circles indicate SNPs that were significantly associated with high pH tolerance at the chromosome level, and grey open circles indicate SNPs with no significant association when raw P-values are FDR corrected. a) Illustrates the SNPs found with specific areas on a chromosome. b) Illustrates the number of SNPs found to be on specific chromosomes but unknown areas on those chromosomes. c) Illustrates the SNPs on the large scaffold containing incompletely localized sequences. 66 Figure 3.3: qqplot demonstrating fit of the mixed linear model utilized to analyze the association of SNPs with high pH tolerance (pH 9.5). 67 Chapter 4: Gene expression plasticity in response to acclimation to high pH in three strains of Rainbow Trout 4.1 Introduction Abrupt exposure to high pH is a very unusual circumstance in Rainbow Trout under natural conditions so it is understandable that transfer from neutral or near-neutral hatchery water to high pH lake water (e.g. during stocking) can cause stress which can lead to death (Witschi and Ziebell 1979, Murray and Ziebell 1984). Even the fish that survive acute exposure can be more susceptible to mortality from other causes (Wilkie and Wood 1991). Acclimation to environmental changes has long been known to lessen these reactions and has been performed within the lake (Yesaki 1990, Yesaki and Iwama 1992, Yesaki and Tsumura 1992, Toth and Tsumura 1993), as well as within the hatchery prior to stocking (Murray and Ziebell 1984, Yesaki 1990, Chapter 2). Although various methods have been used all concluded that short-term acclimation (from a few hours to a week) prior to high pH exposure improves survivorship, but none have looked at the underlying mechanisms which caused this improvement. Rainbow Trout are thought to compensate for acute high pH exposure by counteracting disturbances to ammonia exertion, acid-base homeostasis and electrolyte balance (Wilkie et al. 1996). As described in chapter 1 of this thesis, there are several different mechanisms which work together to regulate ions and acid-base balance in the gill. When exposed to high pH there is a general loss of H+ ions from the acid boundary layer resulting in the inability to convert NH3 to NH4+ (Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1994, Wilkie et al. 1996). One of the possible methods of re-establishing the ‘normal’ H+ ion concentration along the boundary layer would be to transfer H+ resulting from the hydration of CO2 (CO2 + H2O ↔HCO3- + H+). As the CO2 is excreted from the gills at a rate approximately ten times more than ammonia this may be a sufficient supply to compensate for the effects of high pH (Wilkie et al. 1994, Gilmour and Perry 2009). Carbonic anhydrase has been shown to catalyze this reaction (Gilmour and Perry 2009) and its mRNA and enzyme activity have been shown to decrease during ammonia loading (Nawata et al. 2007) suggesting its possible utility to compensate for changes when fish are placed in high pH environments. 68 Other methods of compensation to changing ammonia concentrations in the body as a result of high pH would be alterations to ammonia transport. One of the more recently discovered mechanisms in Rainbow Trout for ammonia transport are the Rhesus glycoproteins (Wright and Wood 2009). Alterations to mRNA expression patterns of any of the above or related are expected when comparing Rainbow Trout gill mRNA expression levels between control and acclimated to high pH samples. Glutamine synthetase (GSase) is a detoxification enzyme that catalyzes conversion of glutamate and ammonia to glutamine. Essex-Fraser et al. (2005) speculated that early induction of GSase genes in Rainbow Trout embryos may prevented excessive accumulation of ammonia. Perhaps alterations to GSase might be beneficial to Rainbow Trout adjustment to ammonia build up at times of high pH exposure. Acute high pH conditions have also been known to cause damage to the gill from alterations to tight junctions and secondary lamellae (Daye and Gardside 1976, Tang and Goodenough 2003). Exposure to high pH alters ion transport by reducing the amount of HCO3- available to exchange with Cl- and by a reduction in the number of Cl- transport sites with recovery resulting from increasing the number of Cl- transport sites (Wilkie et al. 1999, Laurent et al. 2000). Depending on how the fish are acclimated and the fish respond to acclimation to high pH mechanisms involved with cell structure may show variations in expression. Although these are a few of the suspected mechanisms involved in high pH compensation there are likely others that have yet to be identified. In this chapter, RNA-Seq was used to characterize differences in gene expression across pH acclimation groups to identify genes that may be important in conferring pH tolerance because of phenotypic plasticity in Blackwater, Eagle Lake and Fraser Valley Rainbow Trout strains. In these experiments, we acclimated fish at a lower pH level (pH 8.8), which we previously showed confers a degree of tolerance to acute exposure to higher pH (9.5) in Rainbow Trout (see Chapter 2). As previously discussed in this thesis, plasticity was observed across all strains with fish that were reared in pH 8.8 and acclimated to pH 8.8 having a lower percentage of individuals demonstrating loss of equilibrium when exposed to high pH conditions than those reared under near-neutral hatchery conditions. A subset of fish from previous acclimation to pH 8.8 experiments 69 were sampled at day 0, 2 weeks, and 1 month after transfer and varying expression levels in the gill were compared. Combining data from genotype-phenotype association studies with gene expression studies is a particularly powerful way to identify candidate genes underlying differences in tolerance (Jia et al. 2012). 4.2 Methods Fish spawning and rearing: The Blackwater 2n, Eagle Lake 2n, and Fraser Valley 2n fish used in these experiments spawned in April/May of 2014 and reared under standard hatchery conditions (pH 7.2). Details were described in chapter 2. All experiments were conducted according to UBC approved animal care protocol number A14-0103. Acclimation to moderate pH: Fish were acclimated following the methodology described in chapter 2. For RNA-Seq, we focused on fish that had been acclimated as fry to moderately high pH (8.8). These fish were reared at pH 7.2 and then at 6 months of age were gradually transferred over the course of a week to pH 8.8. Fish were then acclimated to this pH for 4 weeks. Fish were randomly selected from the acclimated and control rearing groups and sampled prior to acclimation, at 2 weeks’ post acclimation and 4 weeks’ post acclimation. Right gill arch samples were excised and flash frozen in liquid nitrogen. The 4 weeks’ post acclimation samples were used for RNA-Seq as expression levels were suspected to be the most divergent at this time period. RNA extraction: Total RNA was extracted from frozen gill tissue from six samples for each strain/ treatment group. The tissue was homogenized using a Next Advance Bullet Blender 24 with ten 1.0mm diameter Deria Stabilized Zirconium Oxide beads per sample at an instrument speed setting of 9 for 3 minutes (Next Advance Inc., Averill Park, NY, USA) and using TRIzol® Reagent (Life Technologies Inc., Burlington, ON, Canada) per the manufacturer’s protocol. RNA was then treated using the RNAeasy mini-kit and RNase-Free DNase Set (Qiagen Sciences Inc., Germantown, MD to remove genomic DNA contamination. Samples were diluted to 100ng/µL prior to submission to the McGill University and Genome Quebec Innovation Centre for sequencing. 70 All individuals used for gene expression analyses were genotyped to determine sex using the protocol outlined in Yano et al. (2012), the resulting products were visualized on 2% agarose gel. Strain, treatment, length, weight and sex for all samples are listed in table 4.1. Bioinformatics and Statistical analyses: Libraries were paired-end sequenced using Illumina HiSeq2000 (Illumina, Inc., San Diego, CA) at the Genome Quebec Facility. Samples were multiplexed in six groups of six and were evenly distributed across one flow cell with one individual from each strain/ treatment sequenced per lane. Sequencing reads were aligned to the Rainbow Trout genome (Bertholet et al. 2014) using CLC Genomics Workbench v 10.0. Average library size was 69,769,452 ± 12,701,743 SD. Average mapping efficiency was 81.39% (± 1.63% SD; see Table 4.1 for complete list of mapping statistics). Total gene reads were analyzed for differential expression using the package edgeR v3.12.0 following recommendations outlined in Lin et al. (2016). The data set was filtered to remove genes with no reads and then normalized utilizing the relative log expression (RLE) method (Anders and Huber 2010). Genes with low expression were then removed from the dataset. The criterion for retention was at least 0.5 counts per million (approximately ten reads within the sample with the smallest library). Robust tagwise dispersions were estimated for each gene (Zhou et al. 2014) and a negative binomial general linear model (GLM) was used to identify differentially expressed genes. Two separate approaches were used to assess the effects of strain and pH on gene expression levels. The first method utilized a contrast matrix that included strain, pH, and the interaction between them. The number of genes used for differential expression analysis following normalization and filtering was 33,585. The second method utilized a contrast matrix in which each strain was analyzed independently for the effects of pH on gene expression levels for each gene. The number of genes used for differential expression analysis for each strain following normalization and filtering were 32,016, 32,189 and 31,876 for Blackwater, Eagle Lake and Fraser Valley, respectively. Genes that were identified as differentially expressed between pH treatments for each strain were then compared to identify genes that were differentially expressed in two or more strains and genes that were uniquely differentially expressed in 71 each strain. False discovery rate (FDR) corrections for multiple comparisons (Benjamin-Hochberg 1995) were performed and a q-value threshold of 0.05 was established for significance. Variation among samples was assessed using principal component analysis (PCA). A two-way ANOVA was performed on the PCA results with strain and treatment as fix effects. Enrichment of gene ontology (GO) pathway annotations of the differentially expressed genes were conducted using goseq (v 1.22.0). GO IDs for Rainbow Trout transcripts were obtained from http://www.agbase.msstate.edu/. All analysis were conducted in R v3.12.0 (Robinson et al. 2010) unless otherwise stated. 4.3 Results Table 4.1 presents all mapping statistics for the RNA-Seq libraries. Principle component analysis (PCA) of differentially expressed genes identified several components that explained a high proportion of the variation, but did not clearly separate out samples based on either strain or treatment (Figure 4.1). PC1 explained approximately 31% of variation in the RNA-Seq data and roughly separated treatment groups within each strain, however, a two-way ANOVA showed this not to be significant (p=0.988003). The direction of this effect differed between strains with Eagle Lake strain treatment groups separating opposite to the other strains and Fraser Valley having the broadest distribution, demonstrated by a significant interaction of strain and treatment (P=0.000216). PC2 explained approximately 12% of the variation and separate the strains (p=5.59x10-5) but also illustrated an interaction effect between strain and treatment (p=0.00646). Of the 33,585 genes that met or exceeded the minimum expression threshold in this dataset, a total of 140 genes were identified as having a main effect of high pH across the three Rainbow Trout strains tested (Appendix Table A.3, Figure 4.2). There were also 18,036 genes which exhibited a significant interaction between pH and strain (Figure 4.3). The results suggest that acclimation to high pH has effects on Rainbow Trout and that that those effects differ among the strains. 72 4.3.1 Differential gene expression in response to high pH The expression patterns of the 140 genes that were differentially expressed when Rainbow Trout were acclimated to high pH are illustrated in Figure 4.2. Of those, 40 genes were found to be downregulated, including genes coding for glutamine synthetase (Appendix Table A.3). The remaining genes were upregulated when fish were acclimated to higher pH conditions and included genes coding for carbonic anhydrase and ammonium transporters. Functional GO enrichment analysis revealed two pathways that were significantly enriched in all strains between control and acclimated fish: female gamete generation and actin cytoskeleton organization (Table 4.2). 4.3.2 Differential gene expression in response to high pH between strains Of the genes that showed differential gene expression 18,036 demonstrated an interaction between pH and strain. In general, a more similar pattern was observed between Blackwater and Fraser Valley while Eagle Lake often showed the opposite pattern (Figure 4.3). The PCA supports this with the control and acclimated Eagle Lake fish separating in the opposite direction in PC1 compared to the Blackwater and Fraser Valley fish (Figure 4.1). The patterns observed within the control samples are more consistent among fish than the acclimated fish across all strains. T 4.3.3 Differential gene expression in response to high pH within strain Because of the large number of genes exhibiting significant interactions between strain and pH treatment, each strain of Rainbow Trout was analyzed separately to investigate the effect of high pH acclimation within each strain. There were 8,053 genes differentially expressed in the Blackwater strain, 10,509 in the Eagle Lake strain and 3,445 in the Fraser Valley strain demonstrating a different number of genes being effected by acclimation among the strains. Each strain showed an almost even number of upregulated and downregulated genes. Only 608 of the genes detected were differentially expressed in all strains, but when any two strains were compared a much higher number of genes was observed to overlap (Figure 4.4). For example, more than 50% of the differentially expressed genes overlapped between the Blackwater and Eagle 73 Lake strains, and a similar fraction overlapped between the Fraser Valley and Eagle Lake strains. Of the three strains, Blackwater showed the most consistent pattern of expression within the control samples and acclimated samples (Appendix Figure A.1). GO enrichment analysis of the genes upregulated by high pH detected enrichment in mitochondria related processes (Table 4.3) in Blackwater Rainbow Trout, whereas analysis of the down-regulated genes detected enrichment in biological processes related to membrane function such as cell adhesion and actin cytoskeleton which was also identified as significant in response to high pH when all strains were compared (Table 4.4). Eagle Lake had the largest number of genes that showed differential expression in response to acclimation to high pH (10,509) and the most variation within the control and acclimated samples (Appendix Figure A.2). GO analysis detected different biological processes being affected compared to those detected in the Blackwater strains with overrepresentation of processes corresponding to translation, phosphorylation, cell-matrix adhesion and more among the upregulated genes. Similar processes were also overrepresented within the downregulated genes (Table 4.5 and 4.6). The Fraser Valley strain had a relatively small number of differentially expressed genes compared to the other two strains (Figure 4.4, Appendix Figure A.3). GO enrichment analysis identified different functions being differentially regulated in Fraser Valley strain than in Blackwater or Eagle Lake strains, with oxygen transport and calcium ion transport showing the most significant overrepresentation in the upregulated genes and immune response showing the most significant overrepresentation among the down regulated genes (Table 4.7 and 4.8). 4.4 Discussion When comparing gill tissue from three strains of Rainbow Trout that had been reared under near-neutral hatchery conditions (pH 7.2) to those acclimated to elevated pH (pH 8.8), the most numerous effects (18,036 genes) were the interactions between strain and pH treatments. In contrast, only a small number of genes (140) showed a 74 consistent effect of pH acclimation across all strains, and when each strain was analyzed separately, only 608 overlapping genes were detected. This strong interaction effect illustrates that the response to high pH acclimation differs among all strains of Rainbow Trout suggesting that there is among-strain genetic variation for phenotypic plasticity for high pH tolerance in Rainbow Trout. 4.4.1 Effects of pH that are common to all strains GO enrichment analysis of the set of differentially expressed genes that overlap between all three strains revealed actin cytoskeleton organization as a significantly enriched biological processes. This GO-category consists of many different genes which contribute to assembly, arrangement or disassembly of cytoskeletal structures required for maintenance of epithelial barriers. As discussed in chapter 1, high pH conditions can cause gill damage which results in mucous production, tight junction loosening and degradation to the secondary lamellae (Daye and Gardside 1976, Tang and Goodenough 2003) leading to alterations in ionoregulation and acid-base balance (Wright and Wood 1985, Wilkie and Wood 1991, Yesaki and Iwama 1992, Thompson et al. 2015) as gill permeability alters. Rainbow Trout have been shown to compensate for these alterations by altering ionocyte density and cell surface area (Wilkie and Wood 1994, Wilkie et al. 1999, Laurent et al. 2000) and density of pavement cell microvilli (Laurent et al. 2000). Rodgers and Fanning (2011) reviewed and showed connections between the actin cytoskeleton and tight junctions and epithelial permeability thus alterations in action cytoskeleton biological processes are likely occurring to compensate for alterations to the ionoregulation disturbances caused by exposure to high pH. The significance of alterations to female gamete generation, which was another GO-term identified as having significant effect between control and acclimated fish (Table 4.3), and high pH exposure is not clear, but this category contains many genes coding for regulatory proteins, suggesting that aspects of cell signaling could differ between control and acclimated fish allowing for alterations to cell activities between the two environments. Because the current version of the Rainbow Trout genome is not well annotated, we manually curated the list of 140 genes that demonstrated a significant effect of pH acclimation to identify genes that may represent part of the common core response to 75 high pH in Rainbow Trout (Appendix Table A.3). When fish are exposed to high pH conditions the acid boundary layer of the gill becomes stripped of protons causing NH3 diffusion gradients to reverse and ammonia to build up in the body (Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1991, Wilkie et al. 1996). Acclimation to high pH resulted in the upregulation of ammonium transporter Rhesus glycoprotein type C and carbonic anhydrase genes both of which assist with regulation of ammonia during high pH conditions. The ammonia transporter Rhesus glycoprotein has been shown to act as an ‘ammonia pump’ from the gill by transporting NH3 out of the cell where it can then combine with H+ and the NH3 gradient is maintained (Wright and Wood 2009). One of the sources for H+ ions is when expired CO2 is converted to HCO3- and H+ by carbonic anhydrase in the gill mucous and the resulting acidified gill boundary layer then traps NH3 and NH4+ (Gilmour and Perry 2009). The increased mRNA levels observed of both Rhesus glycoproteins and carbonic anhydrase in the fish acclimated to high pH is consistent with their role in removing ammonia from the gill. Elevated mRNA levels of Rhesus glycoproteins is also consistent with what Sashew et al. (2010) observed in Rainbow Trout embryos when exposed to pH of 9.7. As mentioned, exposure to high pH causes ammonia to accumulate within the body of Rainbow Trout. Glutamine has been shown to be present in high levels in vertebrate brain tissues when exposed to external ammonia (Wicks and Randall 2002, Sanderson et al. 2010) or high pH (Thompson et al. 2015), and this is thought to protect the brain against the toxic effects of ammonia. GSase catalyzes the conversion of glutamate and ammonia to glutamine, and thus could play a role in protecting the brain. However, Sanderson et al. (2010) showed that GSase is not the only mechanism utilized to regulate ammonia and glutamate concentrations in the brain by inhibiting GSase in Rainbow Trout brains exposed to high ammonia. The downregulation of glutamine synthetase that was observed in the gill transcriptome of Rainbow Trout acclimated to high pH conditions suggests GSase, not surprisingly, is not the primary mechanism within the gill to regulate ammonia in response to high pH. 76 4.4.2 Effects of pH within each strain The largest number of genes which showed differential expression demonstrated an interaction between pH and strain. Figure 4.3 showed a similar pattern for control and acclimation samples in both Blackwater and Fraser Valley and may reflect the fact that both these strains are from moderate pH environments compared to the high pH environment that Eagle Lake Rainbow Trout evolved in. In addition, the GO enrichment analysis did not show similar enrichment patterns between the strains (Table 4.3 thru 4.8). The large number of interactions and few similarities among GO enrichment terms between the strains suggests that each strain is being affected by pH in a different manner. The most significant GO-terms identified in Blackwater Rainbow Trout between control and high pH acclimated fish were upregulation of H+ transmembrane transport and ATP synthesis coupled proton transport both of which assist with H+ movement across the cell membrane. When fish are exposed to high pH their acid boundary layer becomes stripped of H+ resulting in the inability to convert NH3 to NH4+ and a build up of ammonia in the body (Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1994, Wilkie et al.1996). The upregulation of genes related to processes which assist with H+ movement across the cell suggests a reacidification of the acid boundary layer to prevent ammonia within the body. GO enrichment analysis for the Blackwater strain of Rainbow Trout also shows many GO-terms related to mitochondria (Table 4.3) suggesting a possible switch to more mitochondria rich cells. As previously mentioned, an increase in the number of ionocytes, which are rich in mitochondria, has been reported (Wilkie and Wood 1994) likely to assist with the chloride ion imbalance resulting from high pH exposure. The downregulation of genes associated with cell adhesion in Blackwater suggests modification to epithelial permeability in response to compensate for ionoregulation disturbances as was described with alterations in expression across all strains in response to high pH. GO enrichment analysis indicated that upregulation of calcium ion transport genes were observed in Fraser Valley fish exposed to high pH. Calcium has been shown to strengthen the tight junctions in the gill which reduces ion loss from plasma (Perry and 77 Wood 1985). Previous studies have speculated about the protective effects of calcium and the possibility of its presence improving the effects of high pH ion loss (Yesaki and Iwama 1992, Wilkie and Wood 1994, Wilkie and Wood 1996). The benefits of increased water hardness on Rainbow Trout in high pH conditions has been explored (Yesaki and Iwama 1992, Thompson et al. 2016) and found to be conflicting. Yesaki and Iwama (1992) found increased calcium in the water to improve tolerance of high pH while Thompson et al. (2016) found it to have little effect. The two studies did not investigate the same sizes or strains of Rainbow Trout nor were their trials for equivalent lengths of time which may be the reason for the differing conclusions. The numerous differences between the strains observed in my study show that not all Rainbow Trout are affected equivalently. The fact that upregulation for calcium transport was observed in the Fraser Valley strain supports Yesaki and Iwama (1992) conclusions that increased water hardness would act as protective effect against elevated pH, at least for this strain. The physiological trials in chapter 2 showed that Fraser Valley fish had fewer individuals which could tolerate high pH. This could explain why there were two Fraser Valley fish acclimated to high pH (samples 3 and 6, Appendix Figure A.3) that showed an intermediate pattern of expression between the control and other acclimated fish. Phenotypic data was not collected from the same individuals that were used for RNA-Seq, so making direct comparisons between gene expression patterns and tolerance levels is not possible in this study. The Blackwater and Eagle Lake strains were shown to have equivalent high pH tolerance in physiological trials when tested as fry (Chapter 2), but despite this their gene expression patterns are quite different with Eagle Lake showing an opposite effect between treatments (Figure 4.3) and having very different biological processes involved in acclimation to high pH (compare Table 4.3 and 4.5). This is likely due to Eagle Lake strain having evolved in a high pH environment compared with the moderate pH origins of the Blackwater strain. Previous studies (Akman et al. 2005) have shown a relationship between differential expression patterns and source population environments. Blackwater and Fraser Valley are both from moderate pH origins and have differing levels of pH tolerance 78 when reared under near-neutral hatchery well conditions (Chapter 2) and although both showed similar phenotypes when acclimated to elevated pH differential gene expression patterns indicate that the different mechanisms are employed to produce this phenotype. As their genetic backgrounds differ the responses to elevated pH likely also differ due to various trade-offs and linkages as both strains diverged from their common ancestor and adapted to their environments. The ability of each of the Rainbow Trout strains to modify their pH tolerance through acclimation illustrates the role plasticity plays in shaping this trait. Although there is not always a strong correlation between abundance of mRNA and related proteins, due to post transcriptional gene regulation events (Vogel and Marcotte 2012, Payne 2015), measuring mRNA concentration levels is still a useful tool determining how transcriptional machinery of cell is affected in presence of external signals (different environments) (Sun et al. 2014). 4.4.3 Future analysis Future analysis should include evaluating the gene expression patterns of fish reared at elevated pH levels since fertilization, preferably those that have gone through a physiological trial to determine if the same processes are involved in acclimation as rearing from fertilization. Chapter 2 indicated that both acclimation and rearing from fertilization at elevated high pH in fry and yearling produced similar phenotypes, but the underlying mechanisms associated with different forms of plasticity have not been explored in this study. In early stages of development, Rainbow Trout can synthesize urea due to the expression of ornithine-urea cycle enzymes but these enzymes are not highly expressed in adults (Wright et al. 1995, Pilley and Wright 2000). Sashew et al. (2010) not only observed increased expression levels of Rhesus glycoproteins when Rainbow Trout embryos are exposed to high pH (pH 9.7), as previously mentioned, but also saw urea transporter mRNA concentrations increase 3.5-6.5 times. Perhaps early rearing at elevated pH levels would result in urea transporter remaining active resulting in a higher pH tolerance for Rainbow Trout. The data in chapter 2 indicated that rearing at high pH and acclimation had the same general effect on tolerance, however, the mechanism(s) that provided tolerance to fish when reared at elevated pH were not 79 investigated and could differ from that of acclimation and are worth exploring. It has been shown that both developmental plasticity and acclimation can contribute to adaptive responses to variation in environments (Brakefield et al. 2007) and additional investigations (qPCR or immunohistochemistry) into when acclimation takes place could provide insights into the role of developmental plasticity for increasing tolerance to high pH environments. Initial experiments focused on the gill due to its important role in acid-base and ionoregulation in fish; however, future studies should investigate the expression levels found in the brain as it is highly sensitive to disturbances in pH homeostasis and elevated ammonia levels (Ip and Chew 2010, Wilkie et al. 2011). It might also be worth exploring the role of the kidney in adapting to high pH environments. The kidney has also been shown to play a role in acid-base regulation, although to a lesser extent than the gill, and has been investigated in low pH conditions (Wood et al. 1999). 4.4.4 Conclusion In general, gene expression within the gill differed between control fish and fish acclimated to high pH. However, the results of this study show that the responses of the gill transcriptome to environmental pH are very different between strains of Rainbow Trout. Although not all the mechanisms identified in response to elevated pH could be explained, this study provided initial insights into some of the mechanisms involved. These results lend further support to the conclusion from chapter 3 that selective breeding for high pH tolerance amongst Rainbow Trout strains would be complicated by the need to undertake selection separately for each strain since the genes to target are numerous and differ between strains. 80 4.5 Tables Table 4.1: Mapping statistics for RNA-Seq gill libraries. Sample name indicates strain; Blackwater (BW), Eagle Lake (EL), or Fraser Valley (FV); treatment (fish acclimated to pH 8.8) or control (fish reared under standard hatchery conditions) and individual sample number. n=6 per strain per treatment Sample Name Length (mm) Weight (g) Sex # Reads Mapped as Pairs Reads Mapped as Broken Reads Total # of Reads Mapped % of Reads Mapped BW-Control-1 70 3.8 male 84,863,480 10,995,229 114,875,084 83.45 BW-Control-2 66 2.9 female 68,574,194 9,970,491 95,460,810 82.28 BW-Control-3 62 2.3 male 81,948,740 12,218,413 113,462,458 82.99 BW-Control-4 66 2.2 female 65,936,526 8,726,450 89,297,932 83.61 BW-Control-5 66 2.9 female 41,710,710 5,377,056 57,610,194 81.74 BW-Control-6 62 2.3 female 75,280,412 12,180,974 107,094,638 81.67 EL-Control-1 82 5.7 male 73,681,584 12,293,653 104,479,628 82.29 EL-Control-4 103 10.8 female 69,687,674 12,195,868 101,037,058 81.04 EL-Control-5 114 15 male 69,730,148 12,751,524 101,732,140 81.08 EL-Control-6 110 14 male 70,791,152 12,537,525 102,515,028 81.28 EL-Control-7 91 7.6 male 78,096,154 13,387,125 112,358,802 81.42 EL-Control 8 130 11 female 69333002 12571245 101007364 81.09 FV-Control-1 71 3.9 male 37,468,852 4,998,272 51,668,330 82.19 FV-Control-2 77 5.1 female 67,791,652 9,616,818 95,192,796 81.32 FV-Control-3 71 4.3 male 74,315,846 14,763,402 112,479,136 79.20 FV-Control-4 65 2.8 female 62,847,564 8,615,683 89,076,770 80.23 FV-Control-5 97 10.8 female 38,356,832 6,083,845 55,206,522 80.50 FV-Control-6 87 7.7 female 67,292,744 12,798,130 98,601,038 81.23 81 Sample Name Length (mm) Weight (g) Sex # Reads Mapped as Pairs Reads Mapped as Broken Reads Total # of Reads Mapped % of Reads Mapped BW-Treatment-1 79 4 male 98,379,968 15,861,400 137,298,802 83.21 BW-Treatment-2 81 5.5 female 78,168,848 13,664,026 112,932,762 81.32 BW-Treatment-3 71 3.8 male 74,940,268 13,486,111 108,035,010 81.85 BW-Treatment-4 58 2.1 female 81,083,106 13,515,466 115,430,844 81.95 BW-Treatment-5 65 2.9 male 81,416,336 13,881,252 117,684,836 80.98 BW-Treatment-6 69 3.3 female 73,826,966 11,666,741 104,502,390 81.81 EL-Treatment-1 91 7.6 male 67,481,162 11,127,746 96,907,272 81.12 EL-Treatment-2 91 8.2 female 69,910,354 8,870,709 97,555,206 80.76 EL-Treatment-3 112 13.8 male 79,035,022 11,266,735 109,407,074 82.54 EL-Treatment-4 116 14.5 female 72,908,296 10,362,262 101,463,634 82.07 EL-Treatment-5 96 10.5 male 46,447,468 6,455,123 72,097,782 73.38 EL-Treatment-6 94 8.5 female 55,902,186 8,081,910 79,007,274 80.99 FV-Treatment-1 96 10.9 male 77,569,876 13,863,780 111,374,076 82.10 FV-Treatment-2 78 5.7 female 75,610,424 12,013,776 107,132,830 81.79 FV-Treatment-3 80 6.2 male 77,889,158 13,357,869 111,740,264 81.66 FV-Treatment-4 83 6.7 female 63,435,106 10,930,431 90,835,748 81.87 FV-Treatment-5 87 8.2 male 71,982,412 13,534,106 105,658,404 80.94 FV-Treatment-6 79 5.9 male 68,006,054 12,853,713 99,269,912 81.45 82 Table 4.2: GO analysis of differentially expressed genes within combined strains between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0007292 female gamete generation BP 2.52x10-7 0.999999997 5 9 GO:0030036 actin cytoskeleton organization BP 1.63x10-5 0.999997585 10 94 Table 4.3: GO analysis of upregulated differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:1902600 hydrogen ion transmembrane transport BP 5.47x10-10 1 24 56 GO:0015986 ATP synthesis coupled proton transport BP 2.12x10-5 0.999996736 12 29 GO:0006122 mitochondrial electron transport, ubiquinol to cytochrome c BP 6.73x10-5 0.999998723 5 6 GO:0005743 mitochondrial inner membrane CC 2.63x10-7 0.999999931 28 92 GO:0070469 respiratory chain CC 1.36x10-6 0.99999988 11 20 GO:0005739 mitochondrion CC 9.74x10-6 0.999996078 39 175 GO:0004129 cytochrome-c oxidase activity MF 1.50x10-8 0.999999998 17 35 83 Table 4.4: GO analysis of downregulated differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0007155 cell adhesion BP 3.33x10-22 1 107 347 GO:0007156 homophilic cell adhesion via plasma membrane adhesion molecules BP 4.35x10-10 1 41 127 GO:0010951 negative regulation of endopeptidase activity BP 1.35x10-6 0.999999626 27 83 GO:0007275 multicellular organism development BP 2.63x10-6 0.999998905 48 191 GO:0030036 actin cytoskeleton organization BP 9.52x10-6 0.999996991 27 94 GO:0001558 regulation of cell growth BP 1.59E-5 0.999996568 17 44 GO:0001525 angiogenesis BP 5.58E-5 0.999989583 13 32 GO:0048731 system development BP 4.27x10-4 0.999990893 4 5 GO:0070831 basement membrane assembly BP 4.28x10-4 0.999990893 4 5 GO:0005576 extracellular region CC 2.44x10-16 1 120 422 GO:0005886 plasma membrane CC 6.72X10-13 1 92 354 GO:0005578 proteinaceous extracellular matrix CC 1.47X10-8 0.999999996 35 108 GO:0031012 extracellular matrix CC 1.12X10-6 0.999999723 24 72 GO:0005615 extracellular space CC 1.79X10-6 0.99999938 36 127 GO:0005882 intermediate filament CC 2.32x10-6 0.999999382 25 73 GO:0005581 collagen trimer CC 4.00x10-5 0.999992868 13 31 GO:0016020 membrane CC 8.64x10-5 0.999927378 943 6901 84 GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0005509 calcium ion binding MF 2.06x10-15 1 160 728 GO:0043565 sequence-specific DNA binding MF 1.75x10-7 1 101 479 GO:0003779 actin binding MF 2.33x10-7 1 70 336 GO:0005198 structural molecule activity MF 8.80x10-6 0.999996351 43 167 GO:0005520 insulin-like growth factor binding MF 1.64x10-5 0.999996446 17 44 GO:0008237 metallopeptidase activity MF 5.02x10-5 0.999980633 31 125 GO:0008201 heparin binding MF 5.27x10-5 0.999991159 12 28 GO:0050840 extracellular matrix binding MF 5.41x10-5 0.999997025 7 10 GO:0004867 serine-type endopeptidase inhibitor activity MF 5.82x10-5 0.999986239 16 46 GO:0004222 metalloendopeptidase activity MF 2.58x10-4 0.99988627 33 146 GO:0043395 heparan sulfate proteoglycan binding MF 3.08x10-4 0.999976017 6 11 GO:0051015 actin filament binding MF 3.59x10-4 0.999840602 32 142 GO:0005201 extracellular matrix structural constituent MF 4.16x10-4 0.999872681 17 57 85 Table 4.5: GO analysis of upregulated differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0006412 translation BP 3.13x10-23 1 169 284 GO:0019885 antigen processing and presentation of endogenous peptide antigen via MHC class I BP 2.33x10-4 0.999989659 9 10 GO:0046854 phosphatidylinositol phosphorylation BP 3.77x10-4 0.999853176 33 65 GO:0007160 cell-matrix adhesion BP 4.03x10-4 0.999905104 17 28 GO:0005840 ribosome CC 3.31x10-32 1 154 219 GO:0030529 intracellular ribonucleoprotein complex CC 2.03x10-11 1 83 142 GO:0005622 intracellular CC 4.62x10-5 0.999963787 503 1453 GO:0015934 large ribosomal subunit CC 6.34x10-5 0.999993956 13 16 GO:0003735 structural constituent of ribosome MF 8.30x10-28 1 150 224 GO:0004623 phospholipase A2 activity MF 8.06x10-4 0.999907242 10 13 GO:0004198 calcium-dependent cysteine-type endopeptidase activity MF 9.65x10-4 0.999721328 19 33 86 Table 4.6: GO analysis of downregulated differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0006412 translation BP 3.13x10-23 1 169 284 GO:0019885 antigen processing and presentation of endogenous peptide antigen via MHC class I BP 3.22x10-4 0.999989659 9 10 GO:0046854 phosphatidylinositol phosphorylation BP 3.77x10-4 0.999853176 33 65 GO:0007160 cell-matrix adhesion BP 4.03x10-4 0.999905104 17 28 GO:0005840 ribosome CC 3.31x10-32 1 154 219 GO:0030529 intracellular ribonucleoprotein complex CC 2.03x10-11 1 83 142 GO:0005622 intracellular CC 4.62x10-5 0.999963787 503 1453 GO:0015934 large ribosomal subunit CC 6.34x10-5 0.999993956 13 16 GO:0003735 structural constituent of ribosome MF 8.30x10-28 1 150 224 87 Table 4.7: GO analysis of upregulated differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0015671 oxygen transport BP 3.02x10-18 1 16 24 GO:0006816 calcium ion transport BP 5.92x10-12 1 25 71 GO:0007292 female gamete generation BP 5.36x10-8 0.999999999 8 9 GO:0006937 regulation of muscle contraction BP 2.02x10-6 0.99999995 6 9 GO:0048016 inositol phosphate-mediated signaling BP 2.69x10-5 0.99999879 6 10 GO:0070588 calcium ion transmembrane transport BP 3.39x10-5 0.999991328 17 78 GO:0051694 pointed-end actin filament capping BP 4.29x10-5 0.999998299 5 10 GO:0040029 regulation of gene expression, epigenetic BP 1.04x10-4 0.999998503 4 5 GO:0006096 glycolytic process BP 2.07x10-4 0.99997715 12 62 GO:0045446 endothelial cell differentiation BP 2.21x10-4 1 3 3 GO:0006942 regulation of striated muscle contraction BP 2.36x10-4 1 3 3 GO:0005833 hemoglobin complex CC 6.75x10-18 1 14 18 GO:0005861 troponin complex CC 1.39x10-8 0.999999999 10 22 GO:0016459 myosin complex CC 4.32x10-6 0.999998696 25 146 GO:0036379 myofilament CC 3.11x10-5 0.999998125 6 15 GO:0019825 oxygen binding MF 5.23x10-19 1 17 26 GO:0005344 oxygen transporter activity MF 3.02x10-18 1 16 24 GO:0020037 heme binding MF 1.32x10-8 0.999999997 22 118 88 GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0005262 calcium channel activity MF 4.39x10-8 0.999999995 14 34 GO:0003774 motor activity MF 2.98x10-6 0.99999898 31 197 GO:0005506 iron ion binding MF 1.88x10-5 0.999994583 20 146 GO:0005220 inositol 1,4,5-trisphosphate-sensitive calcium-release channel activity MF 2.69x10-5 0.99999879 6 10 GO:0005523 tropomyosin binding MF 4.29x10-5 0.999998299 5 10 GO:0003779 actin binding MF 1.62x10-4 0.999921894 40 336 89 Table 4.8: GO analysis of downregulated differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments in the gill GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0006955 immune response BP 1.43x10-13 1 33 120 GO:0051603 proteolysis involved in cellular protein catabolic process BP 1.47x10-11 1 19 47 GO:0060326 cell chemotaxis BP 1.80x10-11 1 17 36 GO:0006508 proteolysis BP 6.20x10-8 1 66 627 GO:0006935 chemotaxis BP 1.28x10-6 0.999999843 12 36 GO:0050829 defense response to Gram-negative bacterium BP 3.46x10-6 0.999999906 6 9 GO:0015991 ATP hydrolysis coupled proton transport BP 6.60x10-6 0.999998987 12 47 GO:0007034 vacuolar transport BP 1.17x10-5 0.999998751 9 27 GO:0006954 inflammatory response BP 3.42x10-5 0.999993171 13 55 GO:0006260 DNA replication BP 4.05x10-5 0.999990894 14 76 GO:0006621 protein retention in ER lumen BP 8.35x10-5 0.999998861 4 5 GO:0006098 pentose-phosphate shunt BP 1.99x10-4 0.99998266 6 16 GO:0007017 microtubule-based process BP 2.26x10-4 0.999947601 12 62 GO:0070098 chemokine-mediated signaling pathway BP 3.85x10-4 0.999950383 7 22 GO:0042254 ribosome biogenesis BP 3.86x10-4 0.999938249 8 32 GO:0000502 proteasome complex CC 2.14x10-12 1 20 49 GO:0005839 proteasome core complex CC 1.47x10-11 1 19 47 GO:0005783 endoplasmic reticulum CC 6.05x10-11 1 35 188 GO:0016020 membrane CC 2.01x10-9 1 449 6901 90 GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0005737 cytoplasm CC 2.33x10-9 1 78 749 GO:0016021 integral component of membrane CC 8.78x10-9 1 428 6578 GO:0033180 proton-transporting V-type ATPase, V1 domain CC 9.85x10-7 0.999999948 8 16 GO:0008250 oligosaccharyltransferase complex CC 1.30x10-5 1 4 4 GO:0019773 proteasome core complex, alpha-subunit complex CC 4.41x10-5 0.999995279 8 23 GO:0005623 cell CC 1.07x10-4 0.999963722 21 154 GO:0005789 endoplasmic reticulum membrane CC 1.38x10-4 0.999955396 19 132 GO:0008009 chemokine activity MF 8.23x10-14 1 20 40 GO:0004175 endopeptidase activity MF 3.72x10-12 1 19 44 GO:0004298 threonine-type endopeptidase activity MF 1.47x10-11 1 19 47 GO:0008233 peptidase activity MF 4.63x10-10 1 42 271 GO:0016787 hydrolase activity MF 4.67x10-9 1 83 865 GO:0046961 proton-transporting ATPase activity, rotational mechanism MF 7.93x10-6 0.999999551 7 14 GO:0000975 regulatory region DNA binding MF 1.08x10-5 0.99999889 9 25 GO:0005525 GTP binding MF 1.56x10-5 0.999992044 62 647 GO:0001530 lipopolysaccharide binding MF 6.54x10-5 0.999999165 4 5 GO:0046923 ER retention sequence binding MF 8.35x10-5 0.999998861 4 5 91 GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0004950 chemokine receptor activity MF 2.74x10-4 0.999967114 7 21 GO:0008234 cysteine-type peptidase activity MF 2.76x10-4 0.999915974 16 108 GO:0005200 structural constituent of cytoskeleton MF 3.47x10-4 0.999928821 10 49 GO:0000166 nucleotide binding MF 4.43x10-4 0.999694296 124 1886 GO:0003678 DNA helicase activity MF 5.09x10-4 0.999959075 5 14 92 4.6 Figures Figure 4.1: Principal component analysis of RNA-Seq data from the gills of three strains of Rainbow Trout under control hatchery conditions (pH 7.2) or acclimated to pH 8.8. Each point represents an individual fish. Squares are Blackwater, circles are Eagle Lake and triangles are Fraser Valley Rainbow Trout. Black represents control conditions and blue represents fish acclimated to high pH conditions. n=6 per strain per treatment 93 Figure 4.2: Heat map displaying expression patterns of genes (140) with significant effects of pH when all strains are considered. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per strain per treatment) and each row represents one gene. The colour coded expression values (log2 counts per million) for each gene have been normalized to mean expression values of that gene. BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley; control = fish acclimated to standard hatchery conditions (pH 7.2); treatment = fish acclimated to pH 8.8 for 4 weeks. The number indicates the replicate number. 94 Figure 4.3: Heat map displaying expression patterns of genes (18,036) demonstrating a significant interaction between pH and strain. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per strain per treatment) and each row represents one gene. The colour coded expression values (log2 counts per million) for each gene have been normalized to mean expression values of that gene. BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley; control = fish acclimated to standard hatchery conditions (pH 7.2); treatment = fish acclimated to pH 8.8 for 4 weeks. The number indicates the replicate number. 95 Figure 4.4: Venn diagram illustrating the total number of differentially expressed genes for which a main effect of pH treatment was detected within each strain; red is Blackwater, green is Eagle Lake and blue is Fraser Valley. n=6 per strain per treatment BW (8,053) FV (3,445) EL (10,509) 976 4414 1121 608 2055 740 4366 96 Chapter 5: Natural lake evaluation of high pH tolerance in Rainbow Trout from different rearing environments 5.1 Introduction The pH of British Columbia lakes has been rising over the last few decades and is reaching critical levels for fish survival in an increasing number of lakes (FFSBC 2017). Stocked Rainbow Trout, a popular target species in BC’s freshwater recreational fishery, are affected by the rising pH (Mathias et al. 1995). Therefore, there is a need to improve our understanding of the factors that influence survival of Rainbow Trout stocked in lakes with high pH, and to develop methods and practices that will improve survivability of stocked trout in high pH lakes. Most of our knowledge regarding the mechanisms involved in high pH tolerance in Rainbow Trout has been deduced from laboratory studies (e.g. Wright and Wood 1985, Yesaki 1990, Yesaki and Iwama 1992, Wilkie and Wood 1991, Wilkie and Wood 1994, Wilkie et al. 1996, Wilkie et al 1997), which are not necessarily accurate predictors of success in natural environments. There appears to be substantial variation among trout strains in tolerance of high pH (e.g. Mathias et al. 1995, Thompson et al. 2015, Chapter 2), but the relevance of these laboratory results to performance in natural lakes remains unclear. For example, the survival of fish in high pH waters is not exclusively related to pH but also dependent on other environmental conditions such as temperature and the ionic content of the water (Emerson et al. 1975, Yesaki and Iwama 1992, Iwama et al. 1997, Mathias et al. 1995). Further, it has been noted that short term survival of Rainbow Trout can be lake-specific (Mathias et al. 1995) and that size of the fish, changes in pH and/ or other water chemistry characteristics affects survival of strains differently. Although natural lake experiments are notoriously difficult to conduct due to the inability to control all factors that could affect stocking success, they are needed to develop successful fisheries management practices. There have been a few studies within natural environments which have demonstrated that, although survival is low when hatchery-reared fish are stocked into high pH lakes, a small number of individuals survive and can be naturalized (i.e. begin to 97 reproduce naturally in these lakes). The offspring of these naturalized trout have substantially higher survival in the high pH environment than do their parent populations (Mathias et al. 1995). This increase in survival is likely due to either 1) natural selection acting on existing genetic variation in the parent population, or 2) phenotypic plasticity in tolerance induced by rearing offspring in a high pH environment, or 3) a combination of these two factors. Alternatively, prior acclimation to high pH has been shown to result in an increase in survival when fish are stocked in high pH lakes (Yesaki 1990), suggesting that there may be phenotypic plasticity for pH tolerance in Rainbow Trout. However, there is little to no information on acclimation and in-lake performance utilizing current BC stocking Rainbow Trout strains. Nor is there information on the survival of trout stocked into naturally high pH lakes, following acclimation or rearing at high pH in the hatchery. The balance between genetic and environmental influences on pH tolerance remains to be determined. In the current study, I examined the effects of strain, rearing conditions, and acclimation treatment on the in-lake survival of Rainbow Trout. Different strains of Rainbow Trout differentially reared at high pH and released into natural lakes of different pH levels. Short- and long-term survival were measured in each strain-treatment group to begin teasing apart genetic and environmental influences on pH tolerance. Together with the other chapters of this thesis, this study provides critically important data to help choose the optimal strategy for improving survival and performance of stocked trout in high pH lakes, which will improve the cost-effectiveness and/or expand the number of lakes that can be stocked with Rainbow Trout. 5.2 Methods Rearing: Fish were spawned and incubated as described in chapter 2 with the exception of the Fraser Valley fish reared under control and acclimated conditions utilized for the fry stocking into Rexford and Morgan Lakes. Those fish were initially reared at Vancouver Island Hatchery and then transferred to Fraser Valley Hatchery in late July 2014. All experiments were performed according to a UBC approved animal care protocol A14-0103. This experiment was performed with Fraser Valley 3n and Blackwater 2n fish only because Fraser Valley 2n fish cannot be released into lakes due to the potential for 98 interbreeding with wild fish. Four thousand fish from each strain were exposed to each of the following pH treatments: 1) Fish reared at pH 7.2 (control) 2) Fish reared at pH 8.8 3) Fish reared at pH 7.2 and acclimated to pH 8.8 for one month prior to release At ponding, the fish designated for fry stocking were placed in separate 550L oval tanks (by strain and treatment) and reared until approximately 3g in size. Feed was measured daily and distributed using belt feeders for each tank. Daily water temperature, oxygen and pH measures were taken. Mortalities were removed each day and recorded prior to tank cleaning. At ponding, fish designated for yearling stocking were placed into 2400L circular tanks. Fish designated for acclimation were transferred from control tanks to 2400L acclimation tanks 6 weeks prior to stocking. The fish were allowed to recover from transfer for 1 week prior to starting the acclimation procedure which consisted of incremental increases in pH (0.2 units per day) to a final pH of 8.8 over one week, followed by holding at pH 8.8 for 4 weeks prior to stocking into the lakes. The pH was controlled using the system described in chapter 2. Stocking: Fish were fin clipped for identification and released into four high pH lakes (Morgan, Hosli, Whale, and Pigeon 2) and two neutral pH lakes (Rexford and Lower) at the fry (3g) and yearling (15-20g) stages, at a density of 200 fish per hectare. Table 5.1 summarizes the characteristics of the control and experimental lakes. Two different releases were chosen because survival and growth may differ depending on the life stage at which fish are stocked (Godin and Tsumura 1999, Northrup and Godin 2009). Fish were transported in oxygenated water at the pH that they had been most recently reared at (near-neutral pH and pH 8.8 for those reared and acclimated to pH 8.8). Fish designated for the short-term survival experiment were hand counted into individual mesh containers and directly placed into net pens upon arrival at the lake. The remaining fish were hand counted into the main transport tanks and released directly into the lake. Water chemistry: Complete water chemistry characteristics of each lake, including pH, total dissolved solids, and conductivity were measured with handheld instruments at 99 a minimum of two time points prior to stocking, at the time of stocking, and at the time of sampling (Table 5.2). Water samples were also collected prior to stocking, at each stocking event and during gill net sampling and submitted to Maxxam Analytics for a complete water chemistry profile (Table 5A.1). Short-term survival: For the purposes of this thesis ‘short-term’ refers to survival within the first week post stocking. To assess short-term survival, approximately 50 fish per lake (see Table 5.3) of each group of stocked fish were placed in net pens to assess relative short-term survival post-stocking. Net pens were checked at 18h, 24h, 72h, and 1 week after stocking as in Mathias et al. (1995) and Malange et al. (1997). Mortalities were removed and recorded at each interval. At the end of 1 week the surviving fish were released into the lake. Long-term survival: For the purposes of this thesis ‘long-term’ survival will refer to survival to age 1.5. A population assessment was carried out in the fall of 2015 (one year after the fry and four months after the yearlings were stocked into the lakes; i.e. at the age of 1.5 years in both cases) using gillnets as in Post et al. (1999) and Askey et al. (2007). Blackwater and Fraser Valley fish reared in the hatchery at pH 8.8 their entire lives were marked and released into each of the lakes one week prior to the full population assessment at ~10% of the original stocking number to provide control for capture success (Table 5.4). Nets were set for two nights with fish removed from the nets each morning and clip, length, and weight were recorded. The number of nets utilized was based on the surface area of the lake with adjustments made based on net set locations and day 1 catch rates (Table 5.5). Statistics: All data are reported as percent total survival. Separate analyses of variance (ANOVA) were performed on the fry and yearling data for both the short-term and long-term data. Each analysis included lake, strain and treatment as fixed factors. For the short-term analyses, only the 1 week data were included, and for long-term assessments data with multiple day catch rates were collapsed to focus on effects of treatment and strain. The effect of lake was used as an indication of pH effect, however, there were many factors which differs between the lakes beyond pH so pH effect alone 100 could not be determined. Chi-square tests were performed on individual lakes with high mortality to analyze at the data at a finer scale. Model development: Bayesian analysis allowed us to incorporate all the data and determine which factors played the most substantial roles in determining in-lake survival. Two models were built, one for short-term and one for long-term survival assessments. Bayesian approaches have been utilized for stock assessment (reviewed in Punt and Hilborn 1997) and mark-recapture analysis (Askey et al. 2007). The major advantage of Bayesian analysis is that prior information can be incorporated to improve individual estimates. In the present analysis, posterior distributions from the short-term analyses were utilized as prior distributions for the long-term analysis. The analysis was run using JAGS (Just Another Gibbs Sampler) which used Markov Chain Monte Carlo (MCMC) simulation. Vague priors were set for the short-term model and the posteriors from the best fit short-term model were utilized as the priors for the long-term model (Table 5.8). For each model, I ran the MCMC for 2,000,000 iterations and a burn-in period of 150,000. Convergence of the chains was determined using Gelman-Rubin Diagnostics and by visually assessing trace plots. Candidate models were compared in a Bayesian framework using the deviance information criterion (DIC) statistic, which evaluates model parsimony (Spiegelhalter et al. 2002). DIC combines goodness of fit (the posterior of mean deviance) with measure of model complexity (pD) which is the effective number of parameters. The data from the lakes stocked with fry and yearlings were run independently as developmental stage may have altered which parameters were affecting the overall survival. The parameters considered for model development to estimate survival were time (intercept for initial survival and time-dependent survival), lake, treatment, strain, treatment by lake interactions, strain by lake interactions and strain by treatment interactions. Models were developed for both short and long-term survival testing each of these factors singly and in combination. The models that were tested were based on the observations, ANOVA and previous lake experiments. 101 5.3 Results 5.3.1 Descriptive statistics A 3-way ANOVA detected no significant effect of lake, difference between strains, or effect of treatment, or a significant interaction between these factors in the short-term net pen trial for fry (Table 5.3). However, power to detect effects was low due to lack of replication, as only one lake of each pH type was used for each life stage, and all lakes have unique characteristics. In addition, in this experiment, some mortality was observed in the control lake (Rexford), which was almost certainly due to transport stress caused by a twist in the net bag which was separating the treatment groups as fish were observed to be in poor condition when placed into net pens. If only the data from Hosli Lake (moderate pH) and Morgan Lake (high pH) are considered, no fish died in Hosli Lake, while in Morgan Lake only 24% of the Fraser Valley fish and 84% of the Blackwater fish that were reared under control conditions remained at the end of one week, whereas survival was 100% for all fish reared or acclimated at pH 8.8 (Table 5.3). Survival was significantly different across groups in the high pH lake (Morgan Lake), as assessed using a Chi-square test, suggesting that the Fraser Valley strain reared under control conditions was less tolerant of high pH than the Blackwater strain, and that high pH rearing or acclimation can improve tolerance under natural lake conditions. Short-term survival of yearlings differed significantly among lakes, and treatments (3-way ANOVA detected significant effects of lake (p=0.00927), treatment (p=0.01856) and a lake/treatment interaction (p=0.0437), but no significant effect of strain (p=0.85716)) (Table 5.7). Of the lakes stocked with yearlings, considerable differences in short-term survival were observed across lakes, with Pigeon 2 and Whale Lake, two of the highest pH lakes, having the highest mortality (74% and 26-38% of fish reared under control conditions surviving in each of these lakes; Table 5.3). In contrast, there was only a single fish lost across all groups in Lower Lake (control pH). In both Pigeon 2 Lake (moderately high pH) and Whale Lake (high pH), high pH acclimated and high pH reared fish had better survival than did fish reared under control conditions. There was no significant effect of strain in the overall ANOVA for short-term survival of fish stocked as yearlings, and examining the data for Whale Lake (high pH) only, there was no difference 102 in survival between the Fraser Valley and Blackwater strains reared under control conditions, as analyzed using a Chi-square test. The data for short-term survival suggest that rearing at or acclimation to high pH can improve survival in high pH lakes regardless of size-at-release. The evidence for differences between the strains in pH tolerance is less clear. The data for fry suggest that the Blackwater strain may be more tolerant, but without replication, these data can only be regarded as suggestive. In contrast, there is no evidence for differences in tolerance among the strains for fish stocked as yearlings. Taken together, these data suggest that high pH rearing or acclimation have the strongest effect on short-term tolerance of high pH in natural lakes. The long-term survival trial yielded variable gill net recapture rates; ranging from 1-8% across lakes (Table 5.6). ANOVA detected a significant effect of strain in both the long-term fry (p=0.00528) and yearling (p=0.0239) experiments with Blackwater having a higher survival rate. ANOVA showed no significant effect of treatment in the long-term fry (p=0.48333) and yearling (p=0.9287) experiments (Table 5.7). 5.3.2 Bayesian model Due to the many factors involved, the limited replication among lakes, and the mixed sample sizes used in this experiment, it is difficult to decipher which parameters are having the greatest effect on survival across the sample lakes using ANOVA. Therefore, I also examined the data using Bayesian analysis, which offers a more flexible approach to data analyses that allows for imbalanced designs. The best fit model for short-term fry survival included lake, rearing/acclimation treatment and their interaction with a DIC of 495.7 (Table 5.8). Figure 5.1 illustrates how aligned the actual survival is to the estimate provided by the model for the short-term fry trial. The primary difference is that the model predicts mortality to occur more gradually. The same parameters, lake, rearing/acclimation treatment and their interaction, were also the best fit for the short-term yearling model with a DIC of 157.1 (Table 4.8). Figure 5.2 illustrates the alignment between the actual survival and the estimated survival for the short-term yearling trial. This predicted value fit better for the yearling than the fry which is reflected in figure 5.3 and by the lower DIC, 157.1. The next best model for the short- 103 term trial is lake, treatment and strain with a DIC of 54.6 and 47.7, for fry and yearling trials respectively. Bayesian analysis on the long-term survival data detected no effect of lake or rearing/acclimation treatment in either the fry or yearling data sets. The best fit model included an effect of strain with a DIC of 509 and 806.1 for fry and yearling, respectively (Table 5.8). This best fit model indicates that treatment did not noticeably improve long term survival, but the high DIC’s, similarity in the DIC’s between the models (Table 5.8) and deviation between predicted and actual catch rates (Figure 5.3 and 5.4) illustrate that there is minimal confidence that strain is the only parameter involved. 5.4 Discussion My data indicate that the greatest predictor of short-term survival in natural lakes is lake and rearing/acclimation treatment, but that these effects were not clearly detectable in the long-term survival trials, in which strain was the dominant effect, independent of lake and the pH variation between the lakes. Natural lakes are complex and fish survival is affected by many factors such as elevation, fish assemblages, avian predation, water chemistry, human impacts (angling and lake usage) (Mathias et al. 1995, Beckman et al. 2006). Capture rates are affected by lake bathymetry (Ward et al. 2012) which can alter net setting effectiveness. There is also net selectivity as fish will grow at different rates based on food resources, competition, predation and elevation (Biro et al. 2003, Biro et al. 2004, Mogensen et al. 2012) and catchability and movement of fish is different between strains within and between lakes (Askey 2007). In general, only a small percentage of fish were recaptured in some of the lakes despite high levels of netting effort, and this could indicate overall low long-term survival and/or growth. 5.4.1 Effects of rearing and acclimation treatments Many studies (Jordan and Lloyd 1964, Murray and Ziebell 1984, Yesaki 1990, Toth and Tsumura 1993) have illustrated that short-term acclimation to various high pH levels can improve survival when fish are exposed to high pH conditions when observed for 24h to 31days. However, previous research also found that even though acclimated fish were found to survive at a higher rate and regulate plasma ions more effectively they still 104 experienced decrease in plasma ion concentration so acclimation does not offer complete protection against high pH conditions (Yesaki 1990). Thus, questions remain regarding acclimation as a long-term solution to high pH lakes. Further, whether fish reared at a moderately high pH for the entirety of their lives prior to exposure to a natural high pH lake environment would have a greater tolerance to those acclimated to high pH has never been documented. This experiment sought to address these questions. Short-term net pen data for both fry and yearlings showed an effect of lake, rearing treatment and their interaction. As pH increased in lakes, fish that had been reared or acclimated to pH 8.8 prior to stocking had higher survival than those reared under standard hatchery conditions. Whale Lake, the highest pH lake in the yearling trial, was the only lake that showed a difference in performance between acclimation and rearing at high pH. These data along with the laboratory trials performed in chapter 2 demonstrate that rearing fish their entire lives at pH 8.8 are comparable to fish acclimated to pH 8.8 for one month prior to release as there was no significant difference in performance in the short-term. However, given the limited replication used in this study, these data must be considered to be preliminary. Whether acclimation is a long-term solution remains uncertain. There is no evidence from long-term lake sampling that acclimation improves long-term survival but sample size and confounding lake parameters could be masking the effects of acclimation meaning it should not be ruled out as a possible solution. 5.4.2 Effects of strain It has been shown that different strains have altered high pH tolerance and survival in both short and long-term lake experiments (Toth and Tsumura 1993, Mathias et al. 1995) and in laboratory environments (Thompson et al. 2015 and Chapter 2). There have also been several experiments comparing strains where no performance difference was observed (Yesaki and Tsumura 1993, Wagner et al. 1997). Bayesian analysis of the short-term survival data indicates that pH, treatment and their interaction influence short-term high pH tolerance but not strain for both fry and yearling release sizes. 105 There is some reason to doubt if the model is fully supported as DIC values are high and observational data indicates that strain may influence the short-term fry survival. In Morgan Lake 84% of Blackwater fish survived whereas only 24% of Fraser Valley fish survived to one week post-stocking. This aligns with the laboratory trials in chapter 2 of this thesis that indicated strain to have an effect in fry but not yearling. The second best fit Bayesian model indicates that strain along with lake, treatment and all their interactions is affecting survival in high pH lakes. The number of lakes utilized may have limited our ability to differentiate among parameters with smaller effects. Thus, although strain might influence high pH tolerance in fry the effect is likely not as large as that of treatment and lake. Thompson et al. (2015) and the study performed in chapter 2 detected differences in high pH tolerance between Blackwater and Fraser Valley Rainbow Trout in the laboratory with different strains having the best performance in those two studies. A variety of reasons are explored in chapter 2 to explain this discrepancy between studies; however, it is possible that differences observed are at the individual fish level rather than the strain. Godin et al. (1994) and Mathias et al. (1995) found that progeny of fish that had been stocked and naturalized to high pH lakes had a survival advantage over fish from lower pH lakes in both short and long-term trials. Godin (FFSBC unpublished 2000) found that Blackwater Rainbow Trout had similar survival rates to fish from naturalized high pH lakes when compared in short-term lake trials while yet another experiment looking at Blackwater and Fraser Valley show that Blackwater performed better than naturalized Green Lake fish (Pennask origin) (FFSBC unpublished 2001). Mathias et al. (1995) found inconsistency in strain performance from lake to lake and year to year in short-term trials. They hypothesized that water chemistry parameters and/or age of fish being tested was responsible for these differences. These studies demonstrate the difficulty in comparing strains in natural lake environments and the lack of repeatability across experiments and brood years. The short-term trial model conclusions that strain is not having a major effect suggests that the lack of repeatability in previous studies may have been due to the relatively modest effect of strain along with variations in lake water chemistry and fish ages. 106 Of the original groups stocked in this study, Blackwater were recaptured at a higher rate across all lakes in the long-term gill net recaptures and although not significant, Blackwater had a higher survival rate in the short-term fry experiment in high pH lakes as well. This study showed an effect of lake in the short-term. However, the long-term model, although not as strongly supported, indicates the Blackwater strain of fish fared better in all lakes, at all sizes of release regardless of the lake pH or rearing treatment. Fraser Valley fish were found not to tolerate high pH in Kootenay region lakes (Oliver 2004). The data from this and others studies mentioned suggests in natural lakes strain does influence performance but it appears to be independent of high pH. In other words, the strain of Rainbow Trout that is selected for stocking is of great importance for overall survival but does not differ between moderate and high pH lakes. 5.4.3 Short-term vs long-term survival The short-term model for fry and yearling indicates that lake, treatment and their interactions are having an effect while the long-term model shows that strain, regardless of the pH of the lake, is the primary factor affecting the survival. Although the long-term model is not as strongly supported as the short-term model there are reasons why different parameters could be related to improved performance across differing lengths of time. As mentioned, acclimation has been reported to improve regulation of ions in high pH environments, but ion levels are still altered compared to those in lower pH waters (Yesaki 1990). These altered ion levels are likely reflective of imbalance in other systems as well that could affect survival. The mechanisms that allow for survival in the short-term might not be sufficient for long-term survival. Further, natural lake environments have many factors that are in flux and affect Rainbow Trout survival. For example, Wagner et al. (1997) showed that pH and temperature both can alter fishes’ ability to survive. Therefore, alterations to the environment or additional stressors may occur during the fishes’ lifetime and alter their ability to compensate for high pH conditions. 107 5.4.4 Size/ development stage One of the questions that has arisen from previous studies is whether developmental stage and/or size of fish plays a role in high pH tolerance. In the short-term trials fry began to be affected at 18h while yearlings were not affected until 72h in the lake. In laboratory trials (see Chapter 2) both fry and yearling began to be affected at ~20h, but fry stabilized earlier (at ~30h) whereas yearlings were still being affected at 48h. Size was compared within each developmental stage and found not to be a significant predictor of tolerance. Both life stages were affected but data suggests that fry are affected more immediately which makes sense physiologically as it is primarily the build-up of ammonia and ion imbalance that causes pH intolerance (Cameron and Heisler 1983, Randall and Wright 1989, Thurston et al 1984, Wright et al. 1988, Wilkie and Wood 1991, Wright and Wood 1985, Wright and Wood 1988, Yesaki 1990) and this imbalance would be more pronounced in smaller fish. Mathias et al. (1995) suggested that size at stocking may have an effect as they found that fingerling Rainbow Trout generally had a higher initial survival rate than yearlings when stocked into high pH lakes. Kokanee (Oncorhynchus nerka) demonstrated a possible effect of developmental stage and/or body size with their ability to cope with high pH stress factors when stocked (Mathias et al. 1995, Malange et al. 1997). In that study, Kokanee were stocked at multiple times and sizes within high pH lakes and showed an increase in survival the later (and larger) they were stocked. Unfortunately, as my study does not have any direct comparisons between fish stocked as fry and yearling there is no ability to determine if fry or yearling stocked fish have a better survival rate in high pH lakes. 5.4.5 Laboratory vs lake In an earlier chapter of this thesis I described tolerance trials which occurred in the laboratory (Chapter 2). These trials tested acute pH tolerance at pH 9.5 in both fry and yearling utilizing the same strains and treatments compared within the lake. These trials demonstrated a difference in tolerance at high pH amongst strains. Blackwater performed consistently better than Fraser Valley across all trials; however, significant differences were only observed in fry (Chapter 2). This corresponds with the strain differences 108 observed in the short-term net pen lake trials, although strain was not detected as an important parameter in the Bayesian analysis. Laboratory trials have also clearly demonstrated an increase in survival when rearing conditions are altered. Rearing at pH 8.8 or acclimating fish to 8.8 prior to acute exposure to pH 9.5 increased survival compared to those fish reared in near-neutral hatchery well-water conditions when fish were tested under laboratory conditions (Chapter 2). This held true for the short-term net pen lake assessments as well. Laboratory trials were shorter, 72h and 48h for fry and yearling respectively, than the in-lake short-term net pen trials (18h to 1 week) but still allow for direct comparison. There are no laboratory trials that extended long enough to be able to compare them to long-term in-lake trials. Long-term survival within the lakes and between lakes was quite variable. Although the use of these standardized net assessments provides the ability to determine survival of the different treatment groups within and between lakes, and to compare population survival to other years and lake assessments, the overall recapture numbers were not sufficient to overcome the overwhelming strain effect observed across all lakes regardless of pH. The alignment of the laboratory and in-lake short-term trials provides further evidence that pH is the driving force for the higher acute mortality that has been observed in high pH lakes. If another of the numerous water chemistry factors were confounding the results the laboratory and lake results should not align as only pH was altered in the laboratory. This agreement suggests that laboratory environments can act as an appropriate proxy for investigating Rainbow Trout reactions to high pH, at least for short-term survival. As there was no comparable long-term laboratory trial and the long-term lake experiment is plagued by low sample size, whether laboratory measures can act as a proxy for long-term assessments is still uncertain. 5.4.6 Model vs ANOVA Standard ANOVAs were performed to allow for comparison to historical data and for the most part agreed with the best fit models. The one model and ANOVA analysis that does not agree is the short-term fry experiment. The model shows significant effects 109 for lake, treatment and their interaction while the ANOVA sees no difference. However, the ANOVA required collapsing and combining data. These data were also affected by the transport stress to one treatment of fish in Rexford lake. For these reasons, it is understandable that it would not always agree with the more comprehensive Bayesian model and why it is not the preferred method of analysis for in-lake assessment. The advantage of the Bayesian model is the ability to incorporate different pH levels, differences in catchability, effects of short-term (using short-term model priors for long-term) and small sample sizes to look at complex system. A model can be created using all parameters of experiment and their interactions, or findings from previous studies can be used to decide which parameters to focus on. The model predicts posterior probabilities which in the case of the short-term experiment were then incorporated into the long-term model to increase the predictive power. The best fit Bayesian model indicates that lake, treatment, and their interaction are the major factors affecting short-term survival in the lake. The results of the long-term trials indicate that the underlying differences in survival by strain mask subtler effects of pH on survival. Additional lakes will need to be assessed before a conclusive statement can be made that acclimation and / or rearing at pH 8.8 are not affecting long-term survival in natural lakes. Although the sample size in the long-term lake experiment was not sufficient to provide confident information on treatment, the use of Bayesian model means that information from other sampling can be incorporated to improve individual estimates. 5.4.7 Conclusions My data demonstrate that in natural lake environments, there is strong evidence that rearing environment plays a vital role in short-term survival with both acclimation to pH 8.8 and rearing at pH 8.8 resulting in better short-term survival in high pH lakes compared to fish reared under control hatchery conditions. Long-term assessment of survival shows a very clear difference in survival between the strains regardless of the pH of the lake, but the small sample sizes did not allow for conclusive statements about the long-term effects of treatment and strain on survival of Rainbow Trout in high pH lakes. This study has provided clearer direction for stocking programs as they seek to improve survival in high pH tolerant lakes. A follow-up study focussing on one specific 110 strain and acclimation should assist with addressing if long-term survival can be improved through alterations to rearing. 111 5.5 Tables Table 5.1: Experimental lake parameters Lake Name Location (UTM) Surface Area (ha) Elevation (m) Mean Depth (m) Maximum Depth (m) Average pH Average Conductivity (uS) Experiment Rexford 10 U 696888 5670910 5 871 6.4 11.6 8.15 298 Fry Stocking Hosli 10 U 702360 5600278 21 1179 5.1 14 8.66 563 Fry Stocking Morgan 10 U 658180 5623478 8 606 5.8 12.4 9.09 1276 Fry Stocking Lower 10 U 626453 5734910 15 921 6 12 8.47 544 Yearling Stocking Pigeon 2 10 U 57862 5710986 23 1110 4 9 8.76 662 Yearling Stocking Whale 10 U 592224 5742461 22 1158 4 11 8.86 828 Yearling Stocking 112 Table 5.2: Experimental lake basic water chemistry measures Lake Season Surface Temperature (oC) pH TDS Conductivity (µS) # Measures Rexford Fall 2014 15.6 8.24 146 293 4 Rexford Spring 2015 20.1 7.93 153 305 1 Hosli Fall 2014 14.7 8.9 267 538 3 Hosli Spring 2015 16.3 8.95 284 568 1 Morgan Fall 2014 15.2 9.17 566 1130 5 Morgan Spring 2015 17.4 9.31 602 1209 1 Lower Spring 2015 19.3 8.24 567 284 5 Pigeon 2 Spring 2015 16.1 8.64 332 664 5 Whale Spring 2015 16.8 8.94 441 883 5 113 Table 5.3: Short-term experimental lake survival. Value recorded is the % of fish living at time of monitoring. Lake Species, Strain, Ploidy Rearing Treatment Stocking # 18 Hour Survival 72 Hour Survival 1 Week Survival Fry Stockings Rexford (control pH) Blackwater 2n Control 50 100% 100% 100% Fraser Valley 3n Control 50 100% 100% 100% Blackwater 2n Acclimation 50 100% 100% 100% Fraser Valley 3n Acclimation 50 100% 100% 100% Blackwater 2n pH 8.8 39 69% 69% 69% Fraser Valley 3n pH 8.8 49 86% 86% 86% Hosli (moderately high pH) Blackwater 2n Control 50 100% 100% 100% Fraser Valley 3n Control 50 100% 100% 100% Blackwater 2n Acclimation 49 100% 100% 100% Fraser Valley 3n Acclimation 50 100% 100% 100% Blackwater 2n pH 8.8 53 100% 100% 100% Fraser Valley 3n pH 8.8 50 100% 100% 100% Morgan (high pH) Blackwater 2n Control 50 88% 84% 84% Fraser Valley 3n Control 50 28% 24% 24% Blackwater 2n Acclimation 50 100% 100% 100% Fraser Valley 3n Acclimation 50 100% 100% 100% Blackwater 2n pH 8.8 49 100% 100% 100% Fraser Valley 3n pH 8.8 50 100% 100% 100% Yearling Stockings Lower (control pH) Blackwater 2n Control 50 100% 100% 100% Fraser Valley 3n Control 50 100% 100% 100% Blackwater 2n Acclimation 50 100% 100% 100% Fraser Valley 3n Acclimation 50 100% 100% 100% Blackwater 2n pH 8.8 50 100% 100% 100% Fraser Valley 3n pH 8.8 50 100% 100% 98% Pigeon 2 (moderately high pH) Blackwater 2n Control 50 98% 92% 74% Fraser Valley 3n Control 50 98% 96% 74% Blackwater 2n Acclimation 50 100% 100% 98% Fraser Valley 3n Acclimation 50 100% 100% 100% Blackwater 2n pH 8.8 50 98% 98% 98% Fraser Valley 3n pH 8.8 50 100% 100% 100% Whale (high pH) Blackwater 2n Control 50 98% 42% 26% Fraser Valley 3n Control 50 96% 62% 38% Blackwater 2n Acclimation 50 92% 90% 84% Fraser Valley 3n Acclimation 50 98% 92% 82% Blackwater 2n pH 8.8 50 100% 98% 98% Fraser Valley 3n pH 8.8 50 100% 100% 96% 114 Table 5.4: Experimental lake fish stocking for high pH tolerance: short (net pen) and long-term (into Lake) assessment. Strain, Ploidy Stocking Date Life Stage Rearing Treatment Clip Average stocking size (g) # Fish into Net Pens #Fish into Lake Stocking Quantity (Total) Rexford Blackwater 2n 25-Sep-14 Fry Control ARV 2.10 ± 0.41 50 117 167 Fraser Valley 3n 25-Sep-14 Fry Control RV 3.56 ± 0.64 50 117 167 Blackwater 2n 25-Sep-14 Fry Acclimation ALM 1.85 ± 0.36 50 117 167 Fraser Valley 3n 25-Sep-14 Fry Acclimation LM 3.07 ± 0.51 49 117 166 Blackwater 2n 25-Sep-14 Fry pH 8.8 ALV 2.43 ± 0.48 39 117 156 Fraser Valley 3n 25-Sep-14 Fry pH 8.8 LV 3.20 ± 0.57 50 117 167 Blackwater 2n 08-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 150 150 Fraser Valley 3n 08-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 150 150 Hosli Blackwater 2n 22-Sep-14 Fry Control ARV 2.10 ± 0.41 50 650 700 Fraser Valley 3n 22-Sep-14 Fry Control RV 3.56 ± 0.64 50 650 700 Blackwater 2n 22-Sep-14 Fry Acclimation ALM 1.85 ± 0.36 49 650 699 Fraser Valley 3n 22-Sep-14 Fry Acclimation LM 3.07 ± 0.51 50 650 700 Blackwater 2n 22-Sep-14 Fry pH 8.8 ALV 2.43 ± 0.48 53 650 703 Fraser Valley 3n 22-Sep-14 Fry pH 8.8 LV 3.20 ± 0.57 50 650 700 Blackwater 2n 01-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 210 210 Fraser Valley 3n 01-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 210 210 Morgan Blackwater 2n 29-Sep-14 Fry Control ARV 2.10 ± 0.41 50 217 267 Fraser Valley 3n 29-Sep-14 Fry Control RV 3.56 ± 0.64 50 217 267 Blackwater 2n 29-Sep-14 Fry Acclimation ALM 1.85 ± 0.36 50 217 267 Fraser Valley 3n 29-Sep-14 Fry Acclimation LM 3.07 ± 0.51 50 217 267 Blackwater 2n 29-Sep-14 Fry pH 8.8 ALV 2.43 ± 0.48 49 517 566 Fraser Valley 3n 29-Sep-14 Fry pH 8.8 LV 3.20 ± 0.57 50 217 267 Blackwater 2n 21-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 80 80 Fraser Valley 3n 21-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 80 80 115 Strain, Ploidy Stocking Date Life Stage Rearing Treatment Clip Average stocking size (g) # Fish into Net Pens #Fish into Lake Stocking Quantity (Total) Lower Blackwater 2n 26-May-15 Yearling Control ARV 14.49 ± 6.10 50 450 500 Fraser Valley 3n 26-May-15 Yearling Control RV 12.27 ± 4.91 50 450 500 Blackwater 2n 26-May-15 Yearling Acclimation ALM 16.67 ± 6.64 50 450 500 Fraser Valley 3n 26-May-15 Yearling Acclimation LM 14.52 ± 5.47 50 450 500 Blackwater 2n 26-May-15 Yearling pH 8.8 ALV 16.89 ± 8.08 50 450 500 Fraser Valley 3n 26-May-15 Yearling pH 8.8 LV 11.70 ± 5.07 50 450 500 Blackwater 2n 14-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 427 427 Fraser Valley 3n 14-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 119 119 Pigeon 2 Blackwater 2n 25-May-15 Yearling Control ARV 14.49 ± 6.10 50 717 767 Fraser Valley 3n 25-May-15 Yearling Control RV 12.27 ± 4.91 50 717 767 Blackwater 2n 25-May-15 Yearling Acclimation ALM 16.67 ± 6.64 50 717 767 Fraser Valley 3n 25-May-15 Yearling Acclimation LM 14.52 ± 5.47 50 717 767 Blackwater 2n 25-May-15 Yearling pH 8.8 ALV 16.89 ± 8.08 50 717 767 Fraser Valley 3n 25-May-15 Yearling pH 8.8 LV 11.70 ± 5.07 50 717 767 Blackwater 2n 14-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 230 230 Fraser Valley 3n 14-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 230 230 Whale Blackwater 2n 25-May-15 Yearling Control ARV 14.49 ± 6.10 50 683 733 Fraser Valley 3n 25-May-15 Yearling Control RV 12.27 ± 4.91 50 683 733 Blackwater 2n 25-May-15 Yearling Acclimation ALM 16.67 ± 6.64 50 683 733 Fraser Valley 3n 25-May-15 Yearling Acclimation LM 14.52 ± 5.47 50 683 733 Blackwater 2n 25-May-15 Yearling pH 8.8 ALV 16.89 ± 8.08 50 683 733 Fraser Valley 3n 25-May-15 Yearling pH 8.8 LV 11.70 ± 5.07 50 683 733 Blackwater 2n 14-Sep-15 1.5 years pH 8.8 ALVRM 37.22 ± 14.79 0 220 220 Fraser Valley 3n 14-Sep-15 1.5 years pH 8.8 LVRM 37.90 ± 12.96 0 220 220 116 Table 5.5: Gill net mesh set per experimental lake and day for long-term survival assessment. Lake Total Mesh # of Nets Total Soak Time Mesh/ha Rexford 1264.92 13 278.57 252.98 Day 1 487.68 5 100.20 97.54 Day 2 777.24 8 178.37 155.45 Hosli 2225.04 22 472.25 105.95 Day 1 1005.84 10 202.77 47.90 Day 2 1219.2 12 269.48 58.06 Morgan 1691.64 17 352.97 211.46 Day 1 807.72 8 150.10 100.97 Day 2 883.92 9 202.87 110.49 Lower 5897.88 19 1124.75 393.19 Day 1 2971.8 10 573.60 198.12 Day 2 2926.08 9 551.15 195.07 Pigeon 2 2438.4 24 514.88 106.02 Day 1 1219.2 12 253.08 53.01 Day 2 1219.2 12 261.80 53.01 Whale 2240.28 22 451.67 101.83 Day 1 1036.32 10 186.77 47.11 Day 2 1203.96 12 264.90 54.73 117 Table 5.6: Total gill net catch for long-term survival assessment. M&R indicates mark and recapture fish stocked one week prior to assessment. Day of Capture Strain, ploidy Treatment Total Catch Average Size (g) Rexford Day1 Blackwater 2n Control 7 182.59 Day1 Fraser Valley 3n Control 2 240.45 Day1 Blackwater 2n Acclimation 4 177.78 Day1 Fraser Valley 3n Acclimation 0 0.00 Day1 Blackwater 2n pH 8.8 11 189.52 Day1 Fraser Valley 3n pH 8.8 1 321.40 Day1 Blackwater 2n pH 8.8 - M&R 4 36.75 Day1 Fraser Valley 3n pH 8.8 - M&R 25 45.31 Day 2 Blackwater 2n Control 8 184.00 Day 2 Fraser Valley 3n Control 1 182.00 Day 2 Blackwater 2n Acclimation 4 182.53 Day 2 Fraser Valley 3n Acclimation 1 301.7 Day 2 Blackwater 2n pH 8.8 2 162.10 Day 2 Fraser Valley 3n pH 8.8 0 0.00 Day 2 Blackwater 2n pH 8.8 - M&R 5 48.62 Day 2 Fraser Valley 3n pH 8.8 - M&R 17 41.25 Hosli Day1 Blackwater 2n Control 36 175.76 Day1 Fraser Valley 3n Control 7 252.31 Day1 Blackwater 2n Acclimation 27 733.98 Day1 Fraser Valley 3n Acclimation 12 206.52 Day1 Blackwater 2n pH 8.8 39 181.36 Day1 Fraser Valley 3n pH 8.8 2 228.85 Day1 Blackwater 2n pH 8.8 - M&R 7 67.66 Day1 Fraser Valley 3n pH 8.8 - M&R 15 45.01 Day 2 Blackwater 2n Control 17 168.01 Day 2 Fraser Valley 3n Control 4 239.45 Day 2 Blackwater 2n Acclimation 8 162.26 Day 2 Fraser Valley 3n Acclimation 8 204.89 Day 2 Blackwater 2n pH 8.8 18 200.88 Day 2 Fraser Valley 3n pH 8.8 1 214.00 Day 2 Blackwater 2n pH 8.8 - M&R 18 45.12 Day 2 Fraser Valley 3n pH 8.8 - M&R 35 51.2 118 Day of Capture Strain, ploidy Treatment Total Catch Average Size (g) Morgan Day1 Blackwater 2n Control 1 214.9 Day1 Fraser Valley 3n Control 0 0 Day1 Blackwater 2n Acclimation 2 261.2 Day1 Fraser Valley 3n Acclimation 1 340.3 Day1 Blackwater 2n pH 8.8 3 265.4 Day1 Fraser Valley 3n pH 8.8 0 0 Day1 Blackwater 2n pH 8.8 - M&R 5 57.58 Day1 Fraser Valley 3n pH 8.8 - M&R 5 45.24 Day 2 Blackwater 2n Control 4 291.4 Day 2 Fraser Valley 3n Control 0 0 Day 2 Blackwater 2n Acclimation 2 269.4 Day 2 Fraser Valley 3n Acclimation 0 0 Day 2 Blackwater 2n pH 8.8 2 325.3 Day 2 Fraser Valley 3n pH 8.8 0 0 Day 2 Blackwater 2n pH 8.8 - M&R 5 51 Day 2 Fraser Valley 3n pH 8.8 - M&R 6 46.83 Lower Day1 Blackwater 2n Control 0 0 Day1 Fraser Valley 3n Control 0 0 Day1 Blackwater 2n Acclimation 1 54.7 Day1 Fraser Valley 3n Acclimation 0 0 Day1 Blackwater 2n pH 8.8 3 86.43 Day1 Fraser Valley 3n pH 8.8 0 0 Day1 Blackwater 2n pH 8.8 - M&R 5 54.52 Day1 Fraser Valley 3n pH 8.8 - M&R 12 45.19 Day 2 Blackwater 2n Control 1 80.7 Day 2 Fraser Valley 3n Control 0 0 Day 2 Blackwater 2n Acclimation 1 79.4 Day 2 Fraser Valley 3n Acclimation 1 19.1 Day 2 Blackwater 2n pH 8.8 0 0 Day 2 Fraser Valley 3n pH 8.8 1 77.8 Day 2 Blackwater 2n pH 8.8 - M&R 5 36.6 Day 2 Fraser Valley 3n pH 8.8 - M&R 4 39.3 119 Day of Capture Strain, ploidy Treatment Total Catch Average Size (g) Pigeon 2 Day1 Blackwater 2n Control 42 201.87 Day1 Fraser Valley 3n Control 6 123.52 Day1 Blackwater 2n Acclimation 40 226.52 Day1 Fraser Valley 3n Acclimation 8 180.96 Day1 Blackwater 2n pH 8.8 49 266.71 Day1 Fraser Valley 3n pH 8.8 8 203.29 Day1 Blackwater 2n pH 8.8 - M&R 49 41.41 Day1 Fraser Valley 3n pH 8.8 - M&R 46 39.92 Day 2 Blackwater 2n Control 28 196.23 Day 2 Fraser Valley 3n Control 0 0 Day 2 Blackwater 2n Acclimation 21 164.59 Day 2 Fraser Valley 3n Acclimation 4 214.18 Day 2 Blackwater 2n pH 8.8 24 270.45 Day 2 Fraser Valley 3n pH 8.8 2 162.2 Day 2 Blackwater 2n pH 8.8 - M&R 33 36.15 Day 2 Fraser Valley 3n pH 8.8 - M&R 47 45.84 Whale Day1 Blackwater 2n Control 11 164.55 Day1 Fraser Valley 3n Control 1 116.7 Day1 Blackwater 2n Acclimation 14 191.6 Day1 Fraser Valley 3n Acclimation 1 150.7 Day1 Blackwater 2n pH 8.8 23 181.73 Day1 Fraser Valley 3n pH 8.8 2 136.05 Day1 Blackwater 2n pH 8.8 - M&R 14 37.83 Day1 Fraser Valley 3n pH 8.8 - M&R 0 0 Day 2 Blackwater 2n Control 7 125.87 Day 2 Fraser Valley 3n Control 1 106.8 Day 2 Blackwater 2n Acclimation 16 137.78 Day 2 Fraser Valley 3n Acclimation 0 0 Day 2 Blackwater 2n pH 8.8 14 188.29 Day 2 Fraser Valley 3n pH 8.8 1 94 Day 2 Blackwater 2n pH 8.8 - M&R 25 43.66 Day 2 Fraser Valley 3n pH 8.8 - M&R 4 50.98 120 Table 5.7: Three-way ANOVA results from lake by strain by treatment comparison for short and long-term comparisons for each life stage stocked (fry and yearling). Significant results are in bold. Comparison p-value Short-Term Fry Lake 0.60200 Strain 0.68300 Treatment 0.85200 Lake: Strain 0.84300 Lake: Treatment 1.00000 Strain: Treatment 0.85200 Lake: Strain: Treatment 1.00000 Short-Term Yearling Lake 0.00927 Strain 0.88313 Treatment 0.01856 Lake: Strain 0.85716 Lake: Treatment 0.04370 Strain: Treatment 0.82905 Lake: Strain: Treatment 0.75799 Long-Term Fry Lake 0.95931 Strain 0.00528 Treatment 0.48333 Lake: Strain 0.73121 Lake: Treatment 0.59564 Strain: Treatment 0.42651 Lake: Strain: Treatment 0.91050 Long-term Yearling Lake 0.33860 Strain 0.02390 Treatment 0.92870 Lake: Strain 0.37560 Lake: Treatment 0.90280 Strain: Treatment 0.97550 Lake: Strain: Treatment 0.84410 121 Table 5.8: Long term lake analysis DIC model selection criteria for a set non-hierarchical (fixed effects) models (best fit models are in bold) Model pD DIC DIC Short Term- Fry Intercept 1.1 990.9 495.2 Lake 6 801.7 306 Strain 5.7 1371.7 876 Treatment 5.9 3267 2771.3 Lake & strain 10.5 810.3 314.6 Treatment & strain 9.9 691.2 195.5 Lake & treatment 11.9 653.1 157.4 Lake, treatment & strain 16.7 661.9 166.2 Lake, treatment & their interaction 29 495.7 0 Lake, treatment, strain, all interactions 55.3 550.3 54.6 Short Term - Yearling Intercept 1 623.2 466.1 Lake 4.7 386.3 229.2 Strain 1.8 621.9 464.8 Treatment 3.4 417.1 260 Lake & strain 8.2 392.7 235.6 Treatment & strain 5.3 419.3 262.2 Lake & treatment 8.2 325.4 168.3 Lake, treatment & strain 11.1 330.4 173.3 Lake, treatment & their interaction 20.5 157.1 0 Lake, treatment, strain, all interactions 48.5 204.8 47.7 Long Term - Fry Intercept 4 546.4 37.4 Lake 6.7 557.5 48.5 Strain 8.1 509 0 Treatment 6.7 568.4 59.4 Lake & strain 12 513 4 Treatment & strain 11.7 571.2 62.2 Lake & treatment 11 565.3 56.3 Lake, treatment & strain 15 520.4 11.4 Lake, treatment & their interaction 18.7 575 66 Lake, treatment, strain, all interactions 37.2 553.7 44.7 122 Model pD DIC DIC Long Term - Yearling Intercept 3.7 914.7 108.6 Lake 5.7 916.9 110.8 Strain 8.3 806.1 0 Treatment 7 913.5 107.4 Lake & strain 10.9 812 5.9 Treatment & strain 11 812.6 6.5 Lake & treatment 8.9 916.6 110.5 Lake, treatment & strain 15.6 820.7 14.6 Lake, treatment & their interaction 16.3 924.9 118.8 Lake, treatment, strain, all interactions 37.7 857.4 51.3 123 5.6 Figures Figure 5.1: Variation in short-term proportional fry survival across in-lake experimental treatments. Points represent predicted proportional survival from the DIC-selected model relative to actual survival. 00.20.40.60.810 0.2 0.4 0.6 0.8 1Predicted Proportional Survival Actual Proportional Survival 124 Figure 5.2: Variation in short-term proportional yearling survival across in-lake experimental treatments. Points represent predicted proportional survival from the DIC-selected model relative to actual survival. 00.20.40.60.810 0.2 0.4 0.6 0.8 1Predicted proportional survivalActual proportional survival 125 Figure 5.3: Variation in long-term fry recaptures across in-lake experimental treatments. Points represent predicted recaptures from the DIC-selected model relative to actual survival. 010203040500 10 20 30 40 50Predicted RecapturesActual Recaptures 126 Figure 5.4: Variation in long-term yearling recaptures across in-lake experimental treatments. Points represent predicted recaptures from the DIC-selected model relative to actual survival. 010203040500 10 20 30 40 50Predicted RecapturesActual Recaptures 127 Chapter 6: Conclusion The overall goal of the research presented in this thesis was to provide insights into the mechanisms underlying natural variation in pH tolerance in Rainbow Trout and to disentangle the effects of genetic variation and rearing environment on tolerance of high pH. The results of this work also provide clear direction on the best approach for improving the survival of Rainbow Trout stocked in high pH lakes. To achieve this overall goal, I addressed five primary objectives: 1) to characterize variation in tolerance of high pH among and within strains of Rainbow Trout, 2) to characterize the effects of rearing environment and acclimation conditions on high pH tolerance, 3) to begin to identify the genetic basis for variation in pH tolerance within and among several strains of Rainbow Trout, 4) to examine the effects of high pH acclimation on gill gene expression to provide insight into the mechanisms involved in high pH acclimation, and 5) to determine whether the effects observed in laboratory trials also improved survival in natural high pH lakes. In this chapter, I summarize the major findings of my experiments, discuss some potential limitations of my work, outline potential future directions, and provide recommendations to managers regarding the most efficient and effective approaches to improving the survival of fish stocked into high pH lakes. 6.1 Variation in acute pH tolerance among and within strains of Rainbow Trout (Chapter 2) To address the first objective of my thesis I reared three different strains of Rainbow Trout in near-neutral hatchery conditions (pH 7.2) and evaluated their acute high pH tolerance to determine the level of variation present within and among the strains. I observed clear differences between Rainbow Trout strains as fry for high pH tolerance with Blackwater and Eagle Lake fish having the highest tolerance, and Fraser Valley fish having the lowest tolerance. Similar trends of variation between the strains were also observed when Rainbow Trout were yearlings, although the differences between strains were not statistically significant at this life stage. There was also substantial variation in pH tolerance within strains at both the fry and yearling stages. The rank order of tolerance among these strains was not consistent with the results of previous 128 studies (Thompson et al. 2015) in which the Fraser Valley strain was observed to be superior to that of the Blackwater strain. At present, it is not possible to determine whether these differences among studies are due to differences in the methods used to assess tolerance, to differences among ages or year classes of fish, or to other unknown factors. However, these data point to the complexity of assessing levels of tolerance of high pH, and determining whether this variation may be genetically based. 6.2 Variation in pH tolerance among Rainbow Trout between rearing environments (Chapter 2) My second objective involved rearing Rainbow Trout in different hatchery conditions to determine if exposure to higher pH during development or short-term acclimation to high pH improved tolerance to acute high pH exposure. In general, I found that rearing fish at an elevated pH increased pH tolerance as assessed under laboratory conditions, and that rearing at pH 8.8 was more effective at improving tolerance than was rearing fish at pH 8.5. Short-term acclimation to pH 8.8 was also increased tolerance of acute exposure to even higher pH (9.5) as assessed under laboratory conditions. In fact, high pH tolerance was equivalent between fish that were reared at pH 8.8 for the entirety of their life and those acclimated to pH 8.8 for one month prior to exposure to acute high pH conditions, across all strains of Rainbow Trout tested at fry life stages. When comparing the effects of strain and rearing on high pH tolerance the results presented in this thesis chapter indicate that rearing environment and acclimation play a stronger and more consistent role in setting acute high pH tolerance than does strain in the three strains evaluated. Although acclimation and rearing at the same elevated pH are both effective at increasing tolerance of high pH, I recommend that short-term acclimation be pursued in the future as a method of increasing the tolerance of fish to high pH prior to stocking into high pH lakes, as rearing fish from incubation to release at an elevated pH is technically challenging to achieve in a production setting. 129 6.3 A GWAS analysis of the genetic basis of high pH tolerance across three strains of Rainbow Trout (Chapter 3) My third objective was to compare the genotypes of multiple individuals from the different strains of Rainbow Trout to determine whether there were any consistent genetic differences between fish that lost equilibrium and those that remained upright at the end of each physiological trial GWAS was employed to investigate if the variation observed could be traced to specific gene(s) that could help explain the mechanisms and/or be utilized as a marker if selective breeding for a high pH tolerant strain were desired. I found little evidence of genetic associations with high pH tolerance in Rainbow Trout, and as such, it is likely that high pH tolerance is a polygenic trait which would make it a poor candidate for marker assisted selection. Although a relatively limited sample size affected my ability to conclusively evaluate genetic associations, I do not recommend pursuing additional GWAS with an increased sample size nor genomic selection, as the combined results of the experiments within this thesis support the suggestion that strain does not have as great of an effect as rearing treatment on overall pH tolerance in Rainbow Trout. 6.4 Gene expression plasticity in response to acclimation to high pH in three strains of Rainbow Trout (Chapter 4) The fourth objective of my thesis was to investigate changes in gene expression at the tissue level during acclimation to aid in determining what mechanisms are associated with the observed plasticity in pH tolerance that I observed in chapter 2. My results showed that many genes altered in expression during acclimation to high pH. Only a relatively small portion of these changes were consistent across all strains. The differential expression levels found in ammonia transport, carbonic anhydrase, glutatmine synthetase support the current physiological understanding of high pH; however, expression levels also altered in other genes whose relation to high pH tolerance has yet to be determined. 130 In contrast, a very large number of genes with significant interactions between pH and strain were detected. This dramatic pattern indicates that the changes occurring during acclimation are different in each strain, demonstrating that a single phenotype (increases in high pH tolerance) can be associated with different physiological compensations. 6.5 Natural lake evaluation of high pH tolerance in Rainbow Trout from different rearing environments (Chapter 5) In chapter 2, I showed that rearing and acclimation at pH 8.8 are both effective in increasing survival when fish are exposed to high pH for a short duration in laboratory conditions, but the applicability of this observation to the stocking program in unclear, because increases in tolerance as assessed in the laboratory do not necessarily result in tolerance in natural lake conditions and for longer exposures. The final objective of my thesis evaluated fish from multiple strains and differential rearing methods in natural lakes to determine if genetic variation and/or phenotypic plasticity due to prior experience of high pH affects survival in natural lakes. Consistent with the results of the laboratory tolerance assessments, the short-term lake trials showed strong evidence that rearing environment plays a key role in survival with both acclimation to high pH and rearing during the entire life cycle at pH 8.8 having high acute survival in high pH lakes compared to those reared under control conditions. Although the long-term dataset had a low sample size it did demonstrate that strain plays a key role in survival in natural environments regardless of pH. However, the data were not sufficient to determine if alterations to rearing would improve long-term survival of Rainbow Trout in high pH lakes. 6.6 Implications of research Taken together, the results of my thesis demonstrate that there is substantial variation for high pH tolerance in Rainbow Trout that likely has a genetic or epigenetic basis. My thesis provides two independent lines of evidence supporting variation among the strains. First, the strains tested differed in tolerance under common conditions in the laboratory (Chapter 2). Second, although all strains improved in tolerance as a result of 131 high pH acclimation, the underlying mechanism likely differ among strains because the strains had very different gene expression patterns in the gill following acclimation (Chapter 4). This work contributes to and extends previous observations of variation among strains of Rainbow Trout (Yesaki and Tsumura 1992, Toth and Tsumura 1993, Mathias et al. 1995, Thompson et al. 2015), and suggests that this system could be an excellent one in which to further explore the mechanistic basis of high pH tolerance. Another key finding of my research is that my data suggest that the differences among strains are the result of changes in multiple underlying processes. For example, the GWAS performed in chapter 3 did not detect any single gene of major effect associated with high pH tolerance, which is consistent with a polygenic trait caused by multiple genes of small effect. Similarly, acclimation to high pH resulted in changes in the expression of many genes, again suggesting a complex underlying mechanism associated with tolerance. Previous physiological work also suggests that high pH tolerance is likely to be a complex trait because many physiological mechanisms are known to be associated with the response to high pH exposure, including changes in ammonia excretion, ion balance, and blood pH (Cameron and Heisler 1983, Randall and Wright 1989, Thurston et al. 1984, Wright et al. 1988, Wilkie and Wood 1991, Wright and Wood 1985, Wright and Wood 1988, Yesaki 1990) as well as effects on the brain, (Wilkie et al. 2011, Thompson et al. 2015), which suggests that high pH tolerance is likely to be a complex trait. My RNA-Seq data provide the first whole-transcriptome analysis of the effects of high pH acclimation, and have the potential to provide new insight into the mechanisms involved in this process. My initial analysis of the genes that were differentially expressed in response to high pH acclimation across all the strains supports the current physiological understanding of this process (Wilkie and Wood 1994, Wilkie et al. 1999, Laurent et al. 2000, Wright and Wood 2009). Once a more completely annotated version of the Rainbow Trout genome is available, it will be possible to more fully analyze the functions of the many genes whose expression differed among the strains in response to high pH acclimation. Thus, in the long-term, my data will provide a resource for future studies that will provide new insights into the physiological mechanisms of high pH tolerance. 132 6.7 Strengths and limitations One of the major strengths of my thesis is that acute and long-term high pH tolerance was investigated in Rainbow Trout looking at physiological performance, genetic variation, expression levels, and lake survival across the same groups of fish. I studied three strains of Rainbow Trout (two utilized in the lake studies), two development stages, and four rearing treatments (three utilized in the lake studies) as well as investigating short and long-term exposure in the lake. This is the first study in which differing rearing methods were investigated, as previous studies have only examined short-term acclimation prior to exposure (Murray and Ziebell 1984, Yesaki 1990, Yesaki and Tsumura 1992, Toth and Tsumura 1993). Prior to these experiments no one had ever reported rearing fish at elevated high pH conditions for extended lengths of time and there was no knowledge on how rearing fish at high pH levels would affect high pH tolerance in natural lakes. Despite these strengths, and in some cases because of them, there were limitations to these experiments. With so many strains and rearing treatments sample sizes were limited. The relatively low sample sizes most directly affected the GWAS analysis and the long-term lake analysis. The GWAS analysis sampled across experiments lowered the ability to compare within strain as the different experiments varied slightly in test pHs. In this thesis, differential gene expression patterns were observed between control and higher pH acclimated treatments to assist it determining the mechanisms which lead to high pH tolerance. However, gene expression patterns do not necessarily correspond to levels of proteins or in the activities of proteins. As a result, gene expression patterns can only be used to generate hypotheses that can be investigated further with other approaches. My experimental design also did not allow direct comparison between developmental stages (as fish tested were from different brood years and both developmental stages were not in the same lake) which prevented my ability to address if/how developmental stage affected pH tolerance. 133 Finally, these experiments oversimplified the chemical profiles of high pH lakes because I altered pH in the laboratory trials by adding only NaOH to increase pH in the laboratory pH tolerance experiments and the altered rearing environments. Within natural lakes many different chemical reactions occur which can alter the pH or alter the response to pH. Studies have been conducted that have compared differences in water hardness (Lloyd and Jordan 1964, Yesaki 1990, McGeer and Eddy 1998, Thompson et al. 2016), ammonia (Eshchar et al. 2006), temperature (Wagner et al. 1997), oxygen (Serafy and Harrell 1993) and their effect on pH tolerance along with many other water chemistry parameters that may influence tolerance. Although these features were monitored within my experiments the adjustment of NaOH alone as a proxy for all high pH environments may limit our ability to determine what mechanisms are involved in high pH tolerance in natural lake systems. Although the mechanisms for high pH tolerance have not been fully explained the knowledge obtained in this study will greatly benefit those who continue in this area of study and does provide a clearer direction to follow to improve survival in high pH lakes in the future, as well as an analysis model which can be utilized when additional data becomes available. The overall goal of this thesis was to provide data that could be mused to inform how stocking programs could be modified to improve survival in high pH lakes. Continued research is required; however, a focus for future explorations has been determined. 6.8 Future research directions A few of the questions that remain to be addressed on high pH tolerance following these evaluations are: 1. What (if any) is the genetic basis for variation in high pH tolerance? 2. Do high pH tolerance mechanisms differ between fish that have been reared at elevated pH levels and those that acclimated to elevated pH levels? 3. What is the optimum pH level to acclimate to long-term pH tolerance? 4. What are the long-term effects of acclimation in a natural lake environment? 5. Is there a difference in survival between different life stages? 134 To assist with the current need to improve survival of Rainbow Trout stocked in high pH lakes, future research should focus on optimizing rearing methods and determining which developmental stage to stock to improve the stocking programs. In addition, much remains to be learned about the mechanisms underlying variation in pH tolerance. To address the applied needs of the stocking program to cost-effectively improve high pH tolerance of Rainbow Trout, future experiments should investigate varying lengths of acclimation and pH levels to determine the optimal conditions for improved survival when fish are stocked into high pH lakes. As rearing treatment showed the greatest overall improvement in the laboratory and were found to also have an interaction effect with pH in the short-term lake experiment it is an area which warrants further investigation. As mentioned above, rearing at pH 8.8 showed equivalent advantages to acclimating fish to pH 8.8 and as acclimating fish is far more practical for stocking programs it may provide more realistic options. It is recommended that the acclimation procedure be optimized, following the suggestions outlined in chapter 2. This includes determining the optimum pH level at which to acclimate the fish. My thesis as well as previous studies (Murray and Ziebell 1984, Yesaki 1990, Yesaki and Iwama 1992, Yesaki and Tsumura 1992, Toth and Tsumura 1992) evaluated pH at varying levels and found the process to improve high pH tolerance. If acclimation is the main objective, then it may be advantageous to do so at a higher pH. This experiment acclimated to pH 8.8 in order to match elevated rearing environments; however previous experiments have demonstrated that higher pH levels (9.5 – 10) are also effective (Murray and Zieball 1984, Yesaki 1990). My thesis showed that rearing at pH 8.8 was better than rearing at pH 8.5 indicating that tolerance improves as levels increase, this is likely true for acclimation as well. However, is there an optimum pH level where fish will be tolerant and yet not overly stressed prior to transport and stocking? This would be interesting from an expression point of view as well as an applied point of view as rearing fish even for short periods at high pH conditions can be challenging so choosing the lowest yet most effective level would be preferred. It would also be advantageous to refine the length of time to hold fish at the elevated pH. The current study increased pH over one week and held fish for one month prior to exposure to high pH conditions while previous experiments acclimated fish over 6 hours to 6 days (Murray and Zieball 1984, Toth and Tsumura 1993, Yesaki 1990). 135 It is recommended that comparisons be made on tolerance to acute pH over a range of acclimation time periods, pH levels and rates at which pH is increased to determine which optimal acclimation method while monitoring expression to determine what the key factor (s) is in determining high pH tolerance and if it can be achieved through alterations to environment. It would be interesting to increase the length of the acute tolerance trials to determine if and when the benefits of acclimation cease. This would be of particular interest if done prior to additional lake experiments. If acclimation ceases to confer an advantage after a specific period of time in the laboratory, where all other variables are controlled then it is unlikely it would be beneficial for long-term survival in natural lakes. Depending on the results of this extended acute tolerance trial, additional lake experiments could be conducted. I was unable to assess the long-term effects of acclimation on lake survival due low recapture rates. A focus on one strain (Blackwater) and two treatments (acclimation and controlled) as well as utilizing more high pH lakes for stocking would increase the sample size to see if acclimation is advantageous to long-term survival in natural lakes. The results of my study of differential expression in gill tissues between control and acclimated conditions demonstrate that although the phenotypes (high pH tolerance) of each of the strains tested were similar following acclimation, the physiological process by which they achieved tolerance differed. Therefore, if acclimation proves successful for long-term survival in Blackwater Rainbow Trout it is not guaranteed to be applicable to other strains of Rainbow Trout and thus should be verified before universal application. The ability to acclimate any strain to high pH conditions would be ideal for stocking as various wild strains of Rainbow Trout are stocked into different lakes based on environment and fisheries management goals. One of the remaining questions to be answered is whether exposure to high pH is affected by the developmental stage of the fish (i.e. would fry or yearling fish survive better long-term when stocked into high pH lakes?). This could not be addressed in this thesis as there were no comparisons between fry and yearling survival in the same lake. Both 136 fry and yearling could be placed in the same lakes for both short-term and long-term to determine if one developmental stage has increased survival. As natural environments do not have static pH levels it would be interesting to expose fish to high pH levels and then transfer to lower pH and re-expose to high pH conditions again to compare physiological responses. Finally, as mentioned additional gene expression levels could be compared at different acclimation pH levels and time periods in gill as well as other tissues. This thesis focused initially on gill; however, kidney, brain, and other organs contribute to ion and acid-base regulation and may provide a clearer understanding regarding the mechanisms involved in high pH tolerance. 6.9 Conclusion and recommendations This thesis has demonstrated that rearing environment has an effect on high pH tolerance. The physiological trials demonstrated that all strains of Rainbow Trout tested improved performance when they were reared or acclimated to elevated pH levels. This was supported by the short-term net pen lake trials. Although the long-term trials were inconclusive regarding the role of rearing environments, the general conclusion of this thesis is that it is worth further exploration. The physiological trials are supported by the gene expression, GWAS analysis and lake trials indicating that strain is not the primarily factor affecting high pH tolerance. The physiological and short-term lake trials demonstrated a consistent effect of altered rearing across strains and developmental stages to improved high pH tolerance. However, variation in pH tolerance among strain differed between developmental stages and was found to be difficult to replicate between experiments (in this thesis and previous research Thompson et al. 2015). The strong interaction between pH and strain that was found in gene expression analysis demonstrates that each strain is responding to altered pH conditions in a different way. The GWAS analysis was not able to identify a specific genetic association to high pH tolerance in Rainbow Trout. The likely polygenic nature of high pH tolerance makes it a poor candidate for marker assisted selection. The fact that Eagle Lake Rainbow Trout did not all have higher pH tolerance than fish from the other 137 strains suggests that evolution in a high pH environment has not reduced variation in the strain to only individuals with high pH tolerance. This could be due to the likely polygenic source of high pH tolerance and/or to the fact that pH levels are not static in a high pH environment resulting in varying levels of selection. The similar tolerance levels observed in Blackwater and Eagle Lake Rainbow Trout, two strains of Rainbow Trout that have evolved in very different pH environments, lends additional support to the fact that differences between strains is not the predominant factor determining high pH tolerance. Overall, the studies presented in this thesis have shown that it is plasticity of the Rainbow Trout species rather than a specific strain or genotype which has the greatest effect on high pH tolerance. Prior to this thesis biologists interested in providing a solution to increased mortality when stocking high pH lakes did not know where to focus attention. Various research had been completed investigating alternate strains of Rainbow Trout or acclimating fish to high pH conditions but whether developing a high pH tolerant strain, altering rearing practices or a combination of the two would be the most beneficial was unknown. By addressing these questions in a single study, I am able to provide a clear recommendation for managers moving forward. I recommend that focus be placed on optimizing acclimation techniques and confirming its application in natural lake environments in the future. 138 References Anders S., Huber W. 2010. Differential expression analysis for sequence count data. Genome Biology. 11: R106. Askey, P. 2007. Towards optimal management of spacially-stucture, recreation fisheries: Linking ecology and angler dynamics in British Columbia rainbow trout (Oncohynchus mykiss) lakes. PhD Thesis. University of Calgary. National Library of Canada. Pp 193. Askey, P.J., Post, J.R., Parkinson, E.A, Rivot, E., Paul, A.J., Biro, P.A. 2007. Estimation of gillnet efficiency and selectivity across multiple sampling units: A hierarchical Bayesian analysis using mark-recapture data. Fisheries Research. 83: 167-174. Bailey, M. and Sumaila, U.R. 2013. Freshwater Angling and the BC Economy. Freshwater Fisheries Society of British Columbia. Technical Report. 60pp. Behnke, R.J.1972. The rationale for preserving genetic diversity: examples of the utilization of intraspecific races of salmonid fishes in fisheries management. Proceedings of the Annual Conference Western Association of State Game and Fish Commissioners. 52:559-561. Benjamin, 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 (Methodological). 57: 289–300. Berthelot C., Brunet F., Chalopin D., Juanchich A., Bernard M., Noel B., et al. 2014. The Rainbow Trout genome provides novel insights into evolution after whole-genome duplication in vertebrates. Nature communications. 5:3657. Biro, P.A., Abrahams, M.V., Post, J.R. and Parkinson, E.A. 2004. Predators select against high growth rates and risk taking behaviour in domestic trout populations. Proceedings of the Royal Society of London Series B – Biological Sciences. 271: 2233-2237. Biro, P.A., Post, J.R., and Parkinson, E.A. 2003. Population consequences of a predator-induced habitat shift by trout in whole-lake experiments. Ecology. 84(3):691-700. Bossdorf, O., Richards, C.L., and Pigliucci, M. 2008. Epigenetics for ecologists. Ecology Letters. 11:106-115 Bradley, R.W. and Sprague, J.B. 1985. The influence f pH, water hardness, and alkalinity on the acute lethality of zinc in rainbow trout (Salmo gairdneri). Canadian Journal of Fisheries and Aquatic Science. 42: 731-736. Brakefield, P.M., Pijpe, J., and Zwaan, B.J. 2006. Developmental plasticity and acclimation both contribute to adaptive responses to alternating seasons to plenty and of stress in Bicyclus butterflues. Journal of Biosciences. 32(3): 465-475. 139 Brauner, C.J., Gonzalez, R.J. and Wilson, J.M. 2013. Extreme environments: Hypersaline, alkaline, and ion-poor waters. Fish Physiology. 31: 435-476. Brookfield, J.F.Y. 2013. Quantitative Genetics: Heritability is not always missing. Current Biology. 23(7): R276-278. Cameron, J.N., Heisler, N. 1983. Studies of ammonia in the rainbow trout: physio-chemical parameters, acid-base behaviour and respiratory clearance. Journal of Experimental Biology. 105 (1): 107-125. Campbell, N.R., LaPatra, S.E., Overturf, K., Towner, R. and Narum, S.R. 2014. Association mapping of disease resistance traits in rainbow trout using restriction site associated DNA sequencing. G3: Genes, Genomes, Genetics. 4(12): 2473-2481. Davidson, W.S., Koop, B.F., Jones, S.J.M., Iturra, P., Vidal, R., Maass, A. et al.2010. Sequencing the genome of the Atlantic salmon (Salmo salar). Genome Biology. 11:403. Daye, P.G. and Gardside, E.T. 1976. Histopathologic changes in surficial tissues of brook trout, Salvelinus fontainalis (Mitchel), exposed to acute and chronic levels of pH. Canadian Journal of Fisheries and Aquatic Science. 54: 2140-2155. Dwyer, R.J. 1990. Catchability of three strains of cutthroat trout. North American Journal of Fisheries Management. 10:458-461. Dymowska, A.K., Hwang, P.P. and Goss, G.G. 2012. Structure and function of ionocytes in the freshwater fish gill. Respiratory Physiology and Neurobiology. 184: 282-292. Elshire, R. J., Glaubitz, J. C., Sun, Q., Poland, J. A., Kawamoto, K., Buckler, E. S., and Mitchell, S. E. 2011. A robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. PloS one. 6(5), e19379. Emerson, K., Russo, R.E., Lund, R.E., Thurston, R.V. 1975. Aqueous ammonia equilibrium calculations: effects of pH and temperature. Journal of Fisheries Research Board of Canada. 32: 2379-2383. Eshchar, M., Lahav, O., Mozes, N., Peduel, A., and Ron, B. 2006. Intensive fish culture at high ammonium and low pH. Aquaculture. 255:301-313. Essex‐Fraser, P. A., Steele, S. L., Bernier, N. J., Murray, B. W., Stevens, E. D., and Wright, P. A. 2005. Expression of four glutamine synthetase genes in the early stages of development of rainbow trout (Oncorhynchus mykiss) in relationship to nitrogen excretion. Journal of Biological Chemistry. 280, 20268–20273. Evans, D.E. 2010. A brief history of the study of fish osmoregulation: the central role of the Mt. Desert Island Biological Laboratory. Frontiers in Physiology. 1(13): 1-10. 140 Evans, D.H., Claiborne, J.B., Currie, S. 2014. The Physiology of Fishes Fourth Edition. CRC Press. Pg 205-233. Feil, R. and Fraga, M.F. 2012. Epigenetics and the environment: emerging patterns and implications. Nature Reviews: Genetics. 13:97-109 FFSBC. 2017. Freshwater Fisheries Society of BC Small Lakes Database [WQ_General Water Chemistry; Table 1b – TDS; Table 4 - Effort]. Retrieved from SLDB [April 14, 2017]. Funk, W.C., McKay, J.K., Hohenlohe, P.A., Alllendorf, F.W. 2012. Harnessing genomics for delineating conservation units. Trends in Ecology and Evolution. 27 (9): 489-496. Galet, D.L., Lider, E.L., Vigg, S., and Robertson, S.R. 1985. Histological changes in the gill, kidney and liver of Lahontan cutthroat trout, Salmo clarki henshawi, living in lakes of different salinity-alkalinity. Journal of Fish Biology. 27(5):533-552. Gilmour, K.M. and Perry, S.F. 2009. Carbonic anhydrase and acid–base regulation in fish. Journal of Experimental Biology. 212(11): 1647-1661. Glaubitz, J.C., Casstevens, T.M., Lu, F., Harriman, J., Elshire, R.J., Sun, Q. and Buckler, E.S. (2014). TASSEL-GBS: a high capacity genotyping by sequencing analysis pipeline. PloS ONE. 9(2): e90346. Godin, T. and Tsumura, K. 1999. Size at release comparisons of Blackwater strain rainbow trout in monoculture and coarsefish lakes (1995-1998). British Columbia Ministry of Fisheries Management Report No. 111. 17pp. Godin, T.I. and Tsumura, K. 2001. Kokanee for High pH Lakes. Habitat Conservation Fund Final Report 1997-2001. British Columbia Ministry of Fisheries Project Report No. RD86. 1-15pp + Appendices. Godin, T., Tsumura, K., Mathias, K. 1994. Alkaline lakes enhancement. British Columbia Ministry of Environment, Lands and Parks Fisheries Project Report No. RD42. 17 pp. Gonzalez-Pena, D., Gao, G., Baranski, M., Moen, T., Cleveland, B.M., Kenney, P.B., Vallejo, R.L., Palti, Y. and Leeds, T.D. 2016. Genome-Wide Association Study for Identifying Loci that Affect Fillet Yield, Carcass, and Body Weight Traits in Rainbow Trout (Oncorhynchus mykiss). Frontiers in Genetics. 7: 1-14. Gutierrez, A.P., Yáñez, J.M., Fukui, S., Swift, B., Davidson, W. S. 2015. Genome-Wide Association Study (GWAS) for Growth Rate and Age at Sexual Maturation in Atlantic Salmon (Salmo salar). PLoS ONE. 10(3): e0119730. 141 Hargis, J.R. 1976. Ventilation and metabolic rate of young rainbow trout (Salmno gairdneri) exposed to sublethal environmental pH. Journal of Experimental Zoology. 196(1): 39-44. Hecht, B.C., Campbell, N.R., Holecek, D.E. and Narum, S.R. 2013. Genome‐wide association reveals genetic basis for the propensity to migrate in wild populations of rainbow and steelhead trout. Molecular Ecology. 22(11): 3061-3076. Ip, Y.K. and Chew, S.F. 2010. Ammonia Production, Excretion, Toxicity, and Defense in Fish: A Review. Front Physiology. 1:134. Jia, P, Liu, Y, Zhao, Z. 2012. Integrative pathway analysis of genome-wide association studies and gene expression data in prostate cancer. BMC Systems Biology. 6 Suppl 3:S13. Johansen, K., Maloiy, G., and Lykkeboe, G. 1975. A fish in extreme alkalinity. Respiration Physiology. 24(2): 159-162. Jones, B.E., Grant, W.D., Duckworth, A.W., Owenson, G.G. 1998. Microbial diversity of soda lakes. Extremophiles. 2:191-200. Jordan, D.R.H. and Lloyd, R. 1964. The resistance of rainbow trout (Salmo gairdnerii Richardson) and roach (Rutius rutilus (L.) to alkaline solutions. International Journal of Air and Water Pollution. 8: 405-409 Keeley, E.R., Parkinson, E.A, Taylor, E.B. 2005. Ecotypic differentiation of native rainbow trout (Oncorhynchus mykiss) populations from British Columbia. Canadian Journal of Fisheries Aquatic Sciences. 62: 1523-1539. Keller I., Wagner C. E., Greuter L., Mwaiko S., Selz O. M., et al., 2013 Population genomic signatures of divergent adaptation, gene flow and hybrid speciation in the rapid radiation of Lake Victoria cichlid fishes. Molecular Ecology. 22: 2848–2863. Korte, A., and Farlow, A. 2013. The advantages and limitations of trait analysis with GWAS: a review. Plant Methods. 9(1), 29. Krueger, C.C., Gharrett, A.J., Dehring, T.R., Allendorf, F.W. 1981. Genetic aspects of fisheries rehabilitation programs. Canadian Journal of Fisheries and Aquatic Science. 38:1877-1881. Krueger, C.C. and Waters, T.F. 1983. Annual production of macroinvertebrates in three streams of different water quality. Ecology. 64(4):840-850. Laurén, D.J. and McDonald, D.G. 1986. Influence of water hardness, pH and alkalinity on the mechanisms of copper toxicity in juvenile rainbow trout, Salmo gairdneri. Canadian Journal of Fisheries and Aquatic Science. 43 (8): 1488-1496. 142 Laurent, P., Wilkie, M.P., Chevalier, C. and Wood, C.M. 2000. The effect of highly alkaline water (pH 9.5) on the morphology and morphometry of chloride cells and pavement cells in the gills of the freshwater rainbow trout: relationship to ionic transport and ammonia excretion. Canadian Journal of Zoology. 78 (2): 307-319. Lin, Y., Golovnina, K., Chen, Z.-X., Lee, H.N., Negron, Y.L.S., Sultana, H., et al. 2016. Comparison of normalization and differential expression analyses using RNA-Seq data from 726 individual Drosophila melanogaster. BMC Genomics. 17: 28. Liu, S., Vallejo, R.L., Palti, Y., Gao, G., Marancik, D.P., Hernandez, A.G. and Wiens, G.D. 2015. Identification of single nucleotide polymorphism markers associated with bacterial cold water disease resistance and spleen size in rainbow trout. Frontiers in Genetics. 6: 298. Lloyd, R. and Jordan, D.H.M. 1964. Some factors affecting the resistance of rainbow trout (Salmo gairdnerh Richardson) to acid waters. International Journal of Air and Water Pollution. 8:393-403. Mackay, T.F.C., Stone, E.A., and Ayroles, J.F. 2009. The genetics of quantitative traits: challenges and prospects. Nature Reviews: Genetics. 10: 565-577. Malange, K, Tsumura, K., Godin, T.I. 1997. Kokanee for high pH lakes. Habitat Conservation Trust Fund Report (1997-1998). British Columbia Ministry of Fisheries Project Report No. RD 73. 15pp. Mathias, K.L. Tsumura, K., Godin, T.I. 1995. Alkaline Lakes Enhancement. Habitat Conservation Fund Final Report. British Columbia Ministry of Environment, Lands, and Parks, Fisheries Branch. Fisheries Project Report No. RD43. 22pp. McDonald, D.G., Hobe, H., Wood, C.M. 1980. The influence of calcium on the physiological responses of the rainbow trout, Salmo gairdneri, to low environmental pH. Journal of Experimental Biology. 88:109-131. McGreer, J.C. and Eddy F.B. 1998. Ionic Regulation and Nitrogenous Excretion in Rainbow Trout Exposed to Buffered and Unbuffered Freshwater of pH 10.5. Physiological Zoology. 71(2): 179-190. McKinney, G.J., Waples, R.K., Seeb, L.W. and Seeb, J.E. 2016. Paralogs are revealed by proportion of heterozygotes and deviations in read ratios in genotyping‐by‐sequencing data from natural populations. Molecular Ecology Resources. Doi: 10.1111/1755-0998/12613 Milligan, C.I. and Wood, C.M. 1982. Disturbances in haematology, fluid volume distribution and circulatory function associated with low environment pH in rainbow trout, Salmo gairdneri. Journal of Experimental Biology. 99: 397-415. 143 Mogensen, S. and Post, J. 2012. Energy allocation strategy modifies growth-survival trade-offs in juvenile fish across ecological and environmental gradients. Oecologia. 168: 923-933 Murray, C. A., and Ziebell, C. D. 1984. Acclimation of rainbow trout to high pH to prevent stocking mortality in summer. The Progressive Fish-Culturist. 46(3): 176-179. Nawata, C.M., Hung, C.C.Y., Tsui, T.K.M., Wilson,J.M., Wright, P.A., and Wood, C.M. 2007. Ammonia excretion in rainbow trout (Oncorhynchus mykiss) evidence for Rh glycoproteins and H+-ATPase involvement. Physiol. Genomics. 31:463-474. Narum S. R., Campbell N. R., Meyer K. A., Miller M. R., Hardy R. W. 2013. Thermal adaptation and acclimation of ectotherms from differing aquatic climates. Molecular Ecology. 11: 3090–3097. Nelson, J. and Yzabeck, G. 2004. Microsatellite DNA analysis of Rainbow Trout Broodstock. Prepared for FFSBC by SeaStar Biotech Inc., Victoria, BC. Pp 11. Northrup, S. and Godin, T.I. 2009. A Review of Rainbow Trout Strain Performance: Historical and Contemporary Data from British Columbia Lakes. Freshwater Fisheries Society of British Columbia Report No 6. Pp 31. Oliver, G.G. 2004. Kootenay region small lakes stocking assessment: 2003. Prepared for Ministry of Water Land and Air Protection, Cranbrook, BC. Pp 112. Payne, S.H. 2015. The utility of protein and mRNA correlation. Trends in Biochemical Sciences. 40(1):1-3 Pigliucci, M. 1996. How organisms respond to environmental changes: from phenotypes to molecules (and vice versa). Trends in Ecology and Evolution. 11:168-173. Pigliucci, M. 2001. Phenotypic Plasticity: Beyond Nature and Nurture. Johns Hopkins University Press. Pilley, C.M. and Wright, P.A. 2000. The mechanisms of urea transport by early life stages of rainbow trout (Oncorhynchus mykiss). The Journal of Experimental Biology. 203: 3199-3207. Pittman, K., Yufera, M., Pavlidis, M., Geffen, A.J., Koven, W., Ribeiro, L., Zambonino-Infante, JL., Tandler, A. 2013. Fantastically plastic: fish larvae equipped for a new world. Reviews in Aquaculture. 5 (S1): S224-S267. Pollard, S. and Yesaki, T. 2004. Rainbow trout strains currently stocked in BC waters. Freshwater Fisheries Society of British Columbia Report No 1. Pp 61. 144 Post, J. R., Parkinson, E.A., Johnston, N.T. 1999. Density-dependent processes in structured fish populations: interaction strengths in whole-lake experiments. Ecological Monographs. 69:155–175. Punt, A. and Hilborn, R. 1997. Fisheries stock assessment and decision analysis: the Bayesian approach. Reviews in Fish Biology and Fisheries. 7(1): 35-63. R Core Team. 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. Randall, D.J., Wright, P.A. 1989. The interaction between carbon dioxide and ammonia excretion and water pH in fish. Canadian Journal of Zoology. 67(12): 2936-2942. Randall, D.J., Wood, C.M., Perry, S.F., Bergman, H., Maloiy, et al. 1989b. Urea excretion as a strategy for survival in a fish living in a very alkaline environment. Nature. 337(6203):165-166. Randall, D.J. and Tsui, T.K.N. 2002. Ammonia toxicity in fish. Marine pollution bulletin. 45(1): 17-23. Reddon, A.R., Hurd, P.L. 2013. Water pH during early development influences sex ratio and male morph in a West African cichlid fish, Pelvicachromis pulcher. Zoology. 116: 139-143. Ricker, W.E. 1975. Computation and interpretation of biological statistics offish populations. Bull. Fish. Res. Board Can. 191: 382 pp. Rodgers, L.S. and Fanning, A.S. 2011. Regulation of epithelial permeability by the actin cytoskeleton. Cytoskeleton. 68: 653-660. Robinson, M.D., McCarthy, D.J., Smyth, G.K. 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 26, 139–140. Sanderson, L.A., Wright, P.A., Robinson, J.W., Ballantyne, J.S., and Bernier, N.J. 2010. Inhibition of glutamine synthetase during ammonia exposure in rainbow trout indicates a high reserve capacity to prevent brain ammonia toxicity. The Journal of Experimental Biology. 213: 2343-2353. Sashaw, J., Nawata, M., Thompson, S., Wood, C.M., and Wright, P.A. 2010. Rhesus glycoprotein and urea transporter genes in rainbow trout embryos are upregulated in response to alkaline water (pH 9.7) but not elevated water ammonia. Aquatic Toxicology. 96: 308-313. Scott, W.B., Crossman, E.J. 1973. Freshwater Fishes of Canada. Bull. Fish. Res. Board Can. pp 184-191. 145 Scott, D.M., Lucas, M.C. and Wilson, R.W. 2005. The effect of high pH on ion balance, nitrogen excretion and behaviour in freshwater fish from an eutrophic lake: a laboratory and field study. Aquatic Toxicology. 73(1): 31-43. Serafy, J.E. and Harrell, R.M. 1993. Behavioural response of fishes to increasing pH and dissolved oxygen: field and laboratory observations. Freshwater Biology. 30: 53-61 Shafer, A.B.A, Wolf, J.B.W, Alves, P.C., Bergstrӧm, L., Bruford, M.W., Brännstrӧm,I. et al. 2015. Genomics and the challenging translation into conservation practice. Trends in Ecology and Evolution. 30(2): 78-87. Spiegelhalter, S.C., Best, N.G., Carlin, B.P., van der Linde, A. 2002. Bayesian measures of model complexity and fit. Journal of the Royal Statistical Society: Series B (Statistical Methodology). 64:583-639. Sun, J., Zhang, G.L., Li, S., Ivanov, A.R., Fenyo, D., Lisacek, F., et al. 2014. Pathway analysis and transciptomics improve protein identification by shotgun proteomics from samples comprising small number of cells – a benchmarking study. BMC Genomics. 15 (Suppl 9): S1. Taling, J. F., Wood, R. B., Prosser, M. V. and Baxter, R. M. 1973. The upper limit of photosynthetic productivity by phytoplankton: evidence from Ethiopian soda lakes. Freshwater Biology. 3: 53–76 Tang, V.W. and Goodenough, D.A. 2003. Paracellular ion channel at the tight junction. Biophysical Journal. 84: 1660-1673. Taylor, E. B., Tamkee, P., Keeley, E. R., and Parkinson, E. A. 2011. Conservation prioritization in widespread species: the use of genetic and morphological data to assess population distinctiveness in rainbow trout (Oncorhynchus mykiss) from British Columbia, Canada. Evolutionary Applications. 4(1): 100-115. Thompson, W. A., Rodela, T. M., and Richards, J. G. 2015. The effects of strain and ploidy on the physiological responses of rainbow trout (Oncorhynchus mykiss) to pH 9.5 exposure. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology. 183: 22-29. Thompson, W. A., Rodela, T. M., and Richards, J. G. 2016. Hardness does not affect the physiological responses of wild and domestic strains of diploid and triploid rainbow trout Oncorhynchus mykiss to short‐term exposure to pH 9.5. Journal of Fish Biology. 89(2): 1345-1358. Thurston, R.V., Russo, R.C., Luedtke, R.J., Smith, C.E., Meyn, E.L., Chakoumakos, C., Wang, K.C., Brown, C.J.D. 1984. Chronic toxicity of ammonia in rainbow trout. Transactions of American Fisheries Society. 113:56-73. 146 Toth, B., and Tsumura, K. 1993. Alkaline Lakes Enhancement. Habitat Conservation Fund Progress Report British Columbia Ministry of Environment, Lands and Parks. Fisheries Report No. RD33, 29pp. Toth, B., and Tsumura, K. 1994. Alkaline Lakes Enhancement. Habitat Conservation Fund Progress Report (1993-94). British Columbia Ministry of Environment, Lands and Parks. Fisheries Report No. RD38, 16pp. Tredger, D. 1990. Small lakes index management: analysis of 1989 results. British Columbia Ministry of Environment. Fisheries Project Report No. FAIU 17. 18pp. Tsai, H.Y., Hamilton, A., Tinch, A.E., Guy, D.R., Gharbi, K., Stear, M., Matika, O., Bishop, S.C. and Houston, R.D. 2015. Genome wide association and genomic prediction for growth traits in juvenile farmed Altantic salmon using a high density SNP array. BMC Genomics. 16(1): 969. Tsai, H.Y., Hamilton, A., Tinch, A.E., Guy, D.R., Bron, J.E., Taggart, J.B., Gharbi, K., Stear, M., Matika, O., Pong-Wong, R. and Bishop, S.C. 2016. Genomic prediction of host resistance to sea lice in farmed Atlantic salmon populations. Genetics Selection Evolution. 48(1): 47. VanPoorten, B.T., Walters, C.J. 2010. Estimation of bioenergetics parameters for Rainbow Trout (Oncorhynchus mykiss) using capture-recapture data with comparison to estimates from a laboratory-based model. The Open Fish Science Journal. 3: 69-79. Vineis, P. and Pearce, N. 2010. Missing heritability in genome-wide association study research. Nature Reviews Genetics. 11: 589. Vogel, C. and Marcotte, E.M. 2012. Insights into the regulation of protein abundance from proteomics and transcriptomic analyses. Nature Reviews: Genetics. 12: 227- 232. Wagner, E.J., Bosakowski, T., and Intelmann, S. 1997. Combined effects of temperature and high pH on mortality and the stress response of Rainbow Trout after stocking. Transactions of the American Fisheries Society. 126(6): 985-998. Ward, H.G.M., Askey, P.J., Post, J.R., Varkey, D.A., and McAllister, M.K. 2012. Basic characteristics and temperature improve abundance estimates from standard index netting of rainbow trout (Oncorhynchus mykiss) in small lakes. Fisheries Research. 131-133:52-59. Wedemeyer, G.A., Palmisano, A.N., and Salsbury, L.E. 1985. Development of an effective transport media for juvenile spring chinook salmon to mitigate and improve smolt survival during Columbia River fish hauling operations. Bonneville Power Administration, Fish and Wildlife Division. 147 Weihrauch, D., Wilkie, M.P. and Walsh, P.J. 2009. Ammonia and urea transporters in gills of fish and aquatic crustaceans. Journal of Experimental Biology. 212: 1716-1730. Wicks, B. J., and Randall, D. J. 2002. The effect of sub‐lethal ammonia exposure on fed and unfed rainbow trout: The role of glutamine in regulation of ammonia. Comparative Biochemistry and Physiology Part A: Molecular and Integrative Physiology. 132(2): 275–285. Wilkening, S., Tekkedil, M.M., Lin, G., Fritsch, E.S., Wei, W., Gagneu,r J., Lazinski, D.W., Camilli, A., Steinmetz, L.M. 2013. Genotyping 1000 yeast strains by next-generation sequencing. BMC Genomics 14:90. Wilkie, M.P. 1994. The physiological adaptations of two salmonids, the rainbow trout (Oncorhynchus mykiss) and the high pH tolerant Lahontan cutthroat trout (Oncorhynchus clarki henshawi), to highly alkaline environments. PhD thesis, McMaster University, Hamilton, Ontario. Wilkie, M.P. 2002. Ammonia excretion and urea handling by fish gills: present understanding and future research challenges. Journal of Experimental Zoology. 293(3): 284-301. Wilkie, M. P., Simmons, H. E., and Wood, C. M. 1996. Physiological adaptations of rainbow trout to chronically elevated water pH (pH= 9.5). Journal of Experimental Zoology. 274(1): 1-14. Wilkie, M.P., Pamenter, M.E., Duquette, S., Dhiyebi, H., Sangha, N., Skelton, G., Smith, M.D. and Buck, L.T. 2011. The relationship between NMDA receptor function and the high ammonia tolerance of anoxia-tolerant goldfish. Journal of Experimental Biology. 214(24): 4107-4120. Wilkie, M.P., and Wood, C.M. 1991. Nitrogenous waste excretion, acid-base regulation, and ionoregulation in rainbow trout (Oncorhynchus mykiss) exposed to extremely alkaline water. Physiological Zoology 64: 1069-1086. Wilkie, M.P. and Wood, C.M. 1994. The effects of extremely alkaline water (pH 9· 5) on rainbow trout gill function and morphology. Journal of fish biology, 45(1): 87-98. Wilkie, M.P., and Wood, C.M. 1995. Recovery from high pH exposure in the rainbow trout: white muscle ammonia storage, ammonia washout, and the restoration of blood chemistry. Physiological Zoology, 68(3): 379-401. Wilkie, M.P. and Wood, C.M. 1996. The adaptations of fish to extremely alkaline environments. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology, 113(4): 665-673. 148 Wilkie, M.P., Wright, P. A., Iwama, G.K., and Wood, C. M. 1994. The physiological adaptations of the Lahontan cutthroat trout (Oncorhynchus clarki henshawi) following transfer from well water to the highly alkaline waters of Pyramid Lake, Nevada (pH 9.4). Physiological Zoology, 67(2): 355-380. Witschi, W.A., and Ziebell, C.D. 1979. Evaluation of pH shock on hatchery-reared rainbow trout. The Progressive Fish Culturist. Vol. 41, No. 1:3-5. Wood, C.M., Perry, S.F., Wright, P.A., Bergman, H.L., and Randall, D.J. 1989. Ammonia and urea dynamics in the Lake Magadi tilapia, a ureotelic teleost fish adapted to an extremely alkaline environment. Respiration Physiology. 77:1-20. Wood, C.M., Bergman, H.L., Bianchini, A., Laurent, P., Maina, J., Johannsson, O.E., Bianchini, L.F., Chevalier, C., Kavembe, G.D., Papah, M.B. and Ojoo, R.O. 2012. Transepithelial potential in the Magadi tilapia, a fish living in extreme alkalinity. Journal of Comparative Physiology B. 182 (2): 247-258. Wood, C.M., Milligan, C.L., and Walsh, P.J. 1999. Renal responses of trout to chronic respiratory and metabolic acidosis and metabolic alkalosis. American Journal of Physiology. 277: R482-492 Wright, P.A., Felskie, A. and Anderson, P.M. 1995. Induction of ornithine-urea cycle enzymes and nitrogen metabolism and excretion in rainbow trout (Oncorhynchus mykiss) during early life stages. Journal of Experimental Biology. 198: 127-135 Wright, P.A., Heming, T.A., Randall, D.J. 1986. Downstream pH changes in water flowing over the gills of rainbow trout. Journal of Experimental Biology. 126: 499-512. Wright, P.A., and Wood, C.M. 1985. An analysis of branchial ammonia excretion in freshwater rainbow trout: effects of environmental pH change and sodium uptake blockade. Journal of Experimental Biology. 114: 329-353. Wright, P.A., and Wood, C.M. 2009. A new paradigm for ammonia excretion in aquatic animals: role of Rhesus (Rh) glycoproteins. Journal of Experimental Biology. 212(15): 2303-2312. Yang, J., Benyamin, B., McEvoy, B.P., Gordon, S., Henders, A.K., Nyholt, D.R. et al. 2010. Common SNPs explain a large portion of the heritability for human height. Nature Genetics. 42(7):565-569. Yano, A., Nicol, B., Jouanno, E., Quillet, E., Fostier, A., Guyomard, E., and Guiguen, Y. 2012. The sexually dimorphic on the Y-chromosome gene (sDY) is a conserved male-specific Y-chromosome sequence in many salmonids. Evolutionary Applications. 6: 486-496. 149 Yesaki, T.Y. 1990. The survival and physiology of rainbow trout (Oncorhynchus mykiss) in hard alkaline waters. Masters thesis. University of British Columbia. Yesaki, T.Y., and Iwama, G.K. 1992. Survival, acid-base regulation, ion regulation, and ammonia excretion in rainbow trout in highly alkaline hard water. Physiological Zoology. 65 (4): 763-787. Yesaki, T.Y., and Tsumura, K. 1992. Alkaline lakes enhancement. Habitat Conservation Fund Progress Report (1991-92). British Columbia Ministry of Environment, Lands and Parks. Fisheries Project Report No. RD31. 14 pp. Zhou, X., Lindsay, H., Robinson M.D. 2014. Robustly detecting differential expression in RNA sequencing data using observation weights. Nucleic Acids Research. 42, e91. 150 Appendix Tables Table A.1: Breakdown of pH trends for 148 lakes in the Thompson-Nicola Region (1970-2010) which has the highest concentration of stocked lakes in BC (FFSBC 2017) # of Lakes % of Lakes Increasing pH 92 62 Static pH 5 3 Decreasing pH 51 35 151 Table A.2: pH tolerance and growth phenotypes for fish included in GBS. BW = Blackwater, EL = Eagle Lake, FV = Fraser Valley Sample ID Experiment Strain Rearing Treatment Exposure Time (h) Length (mm) Weight (g) Condition Factor 2014-1555 Exp 1.1 BW Control 14 55 2 1.20 2014-1560 Exp 1.1 BW Control 16 57 2.4 1.30 2014-1590 Exp 1.1 BW Control 22 65 3.1 1.13 2014-1591 Exp 1.1 BW Control 22 50 1.4 1.12 2014-1607 Exp 1.1 BW Control 24 66 3.5 1.22 2014-1609 Exp 1.1 BW Control 24 62 2.8 1.17 2014-1615 Exp 1.1 BW Control 25 66 3.3 1.15 2014-1627 Exp 1.1 BW Control 36 63 2.7 1.08 2014-1638 Exp 1.1 BW Control 64 62 2.6 1.09 2014-1644 Exp 1.1 BW Control Survived 66 2.9 1.01 2014-1651 Exp 1.1 BW Control Survived 60 2.5 1.16 2014-1653 Exp 1.1 BW Control Survived 63 2.6 1.04 2014-1663 Exp 1.1 BW Control Survived 67 3.1 1.03 2014-1691 Exp 1.1 BW Control Survived 69 3.5 1.07 2014-1699 Exp 1.1 BW Control Survived 57 2 1.08 2014-1693 Exp 1.1 BW Control Survived 69 3.3 1.00 2014-1558 Exp 1.1 EL Control 16 75 4.5 1.07 2014-1563 Exp 1.1 EL Control 17 75 4.5 1.07 2014-1566 Exp 1.1 EL Control 18 91 8.4 1.11 2014-1571 Exp 1.1 EL Control 19 73 4.4 1.13 2014-1586 Exp 1.1 EL Control 21 74 4.8 1.18 2014-1587 Exp 1.1 EL Control 21 99 11.5 1.19 2014-1600 Exp 1.1 EL Control 23 79 5.5 1.12 2014-1614 Exp 1.1 EL Control 25 68 3.3 1.05 2014-1621 Exp 1.1 EL Control 29 86 7.4 1.16 2014-1624 Exp 1.1 EL Control 32 79 5.6 1.14 2014-1625 Exp 1.1 EL Control 36 82 6.6 1.20 2014-1631 Exp 1.1 EL Control 40 83 7.3 1.28 2014-1630 Exp 1.1 EL Control 40 79 6.1 1.24 2014-1641 Exp 1.1 EL Control Survived 103 10.8 0.99 2014-1646 Exp 1.1 EL Control Survived 117 17.1 1.07 2014-1648 Exp 1.1 EL Control Survived 101 10.2 0.99 2014-1649 Exp 1.1 EL Control Survived 87 7.3 1.11 2014-1551 Exp 1.1 FV Control 12 68 3.7 1.18 2014-1552 Exp 1.1 FV Control 13 74 5.9 1.46 2014-1554 Exp 1.1 FV Control 14 76 5.3 1.21 2014-1559 Exp 1.1 FV Control 16 71 4.5 1.26 2014-1562 Exp 1.1 FV Control 17 68 3.8 1.21 2014-1564 Exp 1.1 FV Control 17 63 2.9 1.16 2014-1568 Exp 1.1 FV Control 18 64 3.6 1.37 2014-1577 Exp 1.1 FV Control 19 65 2.9 1.06 152 Sample ID Experiment Strain Rearing Treatment Exposure Time (h) Length (mm) Weight (g) Condition Factor 2014-1569 Exp 1.1 FV Control 19 66 3.5 1.22 2014-1575 Exp 1.1 FV Control 19 76 5 1.14 2014-1581 Exp 1.1 FV Control 20 64 3.4 1.30 2014-1580 Exp 1.1 FV Control 20 83 7.2 1.26 2014-1582 Exp 1.1 FV Control 20 77 5.9 1.29 2014-1584 Exp 1.1 FV Control 20 68 5.6 1.78 2014-1588 Exp 1.1 FV Control 21 67 4.3 1.43 2014-1594 Exp 1.1 FV Control 23 74 5.1 1.26 2014-1595 Exp 1.1 FV Control 23 77 5.9 1.29 2014-1597 Exp 1.1 FV Control 23 65 3.6 1.31 2014-1610 Exp 1.1 FV Control 24 72 5.1 1.37 2014-1619 Exp 1.1 FV Control 27 65 3.2 1.17 2014-1637 Exp 1.1 FV Control 60 75 5.2 1.23 2014-1677 Exp 1.1 FV Control 72 74 4.9 1.21 2014-1660 Exp 1.1 FV Control 72 70 4.1 1.20 2014-1667 Exp 1.1 FV Control 72 61 3.3 1.45 2014-1909 Exp 1.2 BW Control 17 61 2.4 1.06 2014-1928 Exp 1.2 BW Control 23 73 3.8 0.98 2014-1940 Exp 1.2 BW Control 30 70 3.5 1.02 2014-1944 Exp 1.2 BW Control 31 70 3.7 1.08 2014-2155 Exp 1.2 BW Control Survived 73 3.9 1.00 2014-2162 Exp 1.2 BW Control Survived 67 3.2 1.06 2014-1270 Exp 2 BW Control 22 66 3.3 1.15 2014-1169 Exp 2 BW Control 17 69 3.7 1.13 2014-1170 Exp 2 BW Control 17 58 1.9 0.97 2014-1171 Exp 2 BW Control 17 58 2.2 1.13 2014-1185 Exp 2 BW Control 18 63 2.8 1.12 2014-1186 Exp 2 BW Control 18 66 2.9 1.01 2014-1201 Exp 2 BW Control 19 54 1.6 1.02 2014-1202 Exp 2 BW Control 19 53 1.7 1.14 2014-1227 Exp 2 BW Control 20 64 2.9 1.11 2014-1245 Exp 2 BW Control 21 55 2.1 1.26 2014-1243 Exp 2 BW Control 21 58 2 1.03 2014-1284 Exp 2 BW Control 23 66 3.4 1.18 2014-1285 Exp 2 BW Control 23 67 3.4 1.13 2014-1286 Exp 2 BW Control 23 62 2.6 1.09 2014-1295 Exp 2 BW Control 24 64 2.9 1.11 2014-1314 Exp 2 BW Control 25 54 1.9 1.21 2014-1315 Exp 2 BW Control 25 64 2.5 0.95 2014-1313 Exp 2 BW Control 25 71 4.2 1.17 2014-1325 Exp 2 BW Control 26 67 3.6 1.20 2014-1349 Exp 2 BW Control 28 64 3.3 1.26 2014-1374 Exp 2 BW Control 32 71 4.3 1.20 153 Sample ID Experiment Strain Rearing Treatment Exposure Time (h) Length (mm) Weight (g) Condition Factor 2014-1378 Exp 2 BW Control 34 69 3.5 1.07 2014-1389 Exp 2 BW Control 38 69 3.7 1.13 2014-1393 Exp 2 BW Control 39 67 4 1.33 2014-1397 Exp 2 BW Control 41 63 2.8 1.12 2014-1404 Exp 2 BW Control 48 68 3.6 1.14 2014-1548 Exp 2 BW Control Survived 64 2.4 0.92 2014-1157 Exp 2 EL Control 14 68 3.1 0.99 2014-1160 Exp 2 EL Control 15 66 2.6 0.90 2014-1168 Exp 2 EL Control 17 79 5.6 1.14 2014-1198 Exp 2 EL Control 18 80 5.7 1.11 2014-1192 Exp 2 EL Control 18 75 4.4 1.04 2014-1207 Exp 2 EL Control 19 104 12.3 1.09 2014-1208 Exp 2 EL Control 19 84 6.3 1.06 2014-1228 Exp 2 EL Control 20 67 3.7 1.23 2014-1233 Exp 2 EL Control 20 71 4.1 1.15 2014-1264 Exp 2 EL Control 21 81 5.9 1.11 2014-1247 Exp 2 EL Control 21 94 9.1 1.10 2014-1273 Exp 2 EL Control 22 78 5.3 1.12 2014-1278 Exp 2 EL Control 22 96 9.4 1.06 2014-1301 Exp 2 EL Control 24 97 9.7 1.58 2014-1309 Exp 2 EL Control 24 99 10.5 1.08 2014-1330 Exp 2 EL Control 26 102 12.1 1.14 2014-1329 Exp 2 EL Control 26 90 8.2 1.12 2014-1332 Exp 2 EL Control 27 71 4 1.12 2014-1337 Exp 2 EL Control 27 88 7.7 1.13 2014-1359 Exp 2 EL Control 28 88 8.1 1.19 2014-1363 Exp 2 EL Control 29 85 11.9 1.94 2014-1368 Exp 2 EL Control 30 98 11.4 1.21 2014-1372 Exp 2 EL Control 31 76 5.1 1.16 2014-1383 Exp 2 EL Control 35 92 8.6 1.10 2014-1392 Exp 2 EL Control 38 83 5.9 1.03 2014-1396 Exp 2 EL Control 41 80 6.1 1.19 2014-1399 Exp 2 EL Control 44 91 8.3 1.10 2014-1415 Exp 2 EL Control Survived 84 6.6 1.11 2014-1442 Exp 2 EL Control Survived 92 7 0.90 2014-1460 Exp 2 EL Control Survived 84 6.2 1.05 2014-1488 Exp 2 EL Control Survived 99 9.6 0.99 2014-1161 Exp 2 FV Control 15 67 3.6 1.20 2014-1174 Exp 2 FV Control 17 69 3.6 1.10 2014-1175 Exp 2 FV Control 17 73 4.1 1.05 2014-1176 Exp 2 FV Control 17 75 4.7 1.11 2014-1187 Exp 2 FV Control 18 69 3.7 1.13 2014-1188 Exp 2 FV Control 18 67 3.5 1.16 154 Sample ID Experiment Strain Rearing Treatment Exposure Time (h) Length (mm) Weight (g) Condition Factor 2014-1191 Exp 2 FV Control 18 68 3.9 1.24 2014-1189 Exp 2 FV Control 18 57 2.4 1.30 2014-1190 Exp 2 FV Control 18 72 4.8 1.29 2014-1203 Exp 2 FV Control 19 62 3 1.26 2014-1204 Exp 2 FV Control 19 68 3.8 1.21 2014-1206 Exp 2 FV Control 19 73 4.7 1.21 2014-1246 Exp 2 FV Control 21 69 3.9 1.19 2014-1272 Exp 2 FV Control 22 69 3.9 1.19 2014-1287 Exp 2 FV Control 23 67 3.7 1.23 2014-1296 Exp 2 FV Control 24 68 3.8 1.21 2014-1297 Exp 2 FV Control 24 64 3.1 1.18 2014-1299 Exp 2 FV Control 24 66 3.4 1.18 2014-1308 Exp 2 FV Control 24 68 3.8 1.21 2014-1333 Exp 2 FV Control 27 70 4.4 1.28 2014-1362 Exp 2 FV Control 29 75 5.4 1.28 2014-1371 Exp 2 FV Control 31 80 6.7 1.31 2014-1418 Exp 2 FV Control Survived 67 4.8 1.60 2014-1462 Exp 2 FV Control Survived 70 4.9 1.43 155 Table A.3: Genes associated with high pH tolerance from RNAseq analysis in the gills (sorted by FDR). Genes associated with ammonium transporter and carbonic anhydrase are highlighted in red and blue respectively. Name Log Fold Change Raw P Value FDR P Value Description Gene_42029 1.586925 1.57x10-12 6.12E-08 PREDICTED: Salmo salar protein FAM131A-like (LOC106568863), transcript variant X3, mRNA Gene_8113 1.4096039 2.64x10-11 5.16E-07 PREDICTED: Salmo salar ammonium transporter Rh type C (LOC106587392), mRNA Gene_35631 -0.849326 1.36x10-10 1.77E-06 Oncorhynchus mykiss glutamine synthetase (gs04), mRNA Gene_42513 1.9383789 3.66x10-10 3.57E-06 Salmo salar Carbonic anhydrase 12 (cah12), mRNA Gene_42512 2.5100271 2.46x10-9 1.92E-05 Salmo salar Carbonic anhydrase 12 (cah12), mRNA Gene_26121 1.2698303 4.37x10-9 2.84x10-05 PREDICTED: Salmo salar synaptogyrin-1-like (LOC106607452), transcript variant X1, mRNA Gene_34959 1.2106505 5.88x10-9 3.28x10-05 PREDICTED: Salmo salar ammonium transporter Rh type C-like (LOC106562250), mRNA Gene_7693 1.1999758 8.15x10-9 3.98x10-05 PREDICTED: Salmo salar signal-transducing adaptor protein 1-like (LOC106604676), mRNA Gene_12782 1.2455496 1.29x10-8 5.61x10-05 Salmo salar clone ssal-eve-574-243 Cytochrome c oxidase subunit 4 isoform 2, mitochondrial precursor putative mRNA, complete cds Gene_42514 1.7473016 2.05x10-8 8.01x10-05 Salmo salar Carbonic anhydrase 12 (cah12), mRNA Gene_34435 -3.323748 4.15x10-8 0.0001472 Oncorhynchus mykiss mRNA for inducible nitric oxide synthase (iNOS gene) Gene_13989 1.1480679 7.51x10-8 0.0002444 PREDICTED: Salmo salar transient receptor potential cation channel subfamily M member 5-like (LOC106596144), partial mRNA Gene_12591 1.0205748 1.45x10-7 0.0004326 PREDICTED: Salmo salar transient receptor potential cation channel, subfamily M, member 5 (trpm5), partial mRNA 156 Name Log Fold Change Raw P Value FDR P Value Description Gene_3468 0.8436541 1.55x10-7 0.0004326 PREDICTED: Salmo salar Neural cell adhesion molecule 1-A (nca11), transcript variant X3, mRNA Gene_3214 -8.013175 1.72x10-7 0.0004484 PREDICTED: Salmo salar CUB and zona pellucida-like domain-containing protein 1 (LOC106587759), mRNA Gene_13536 -3.423262 2.19x10-7 0.0005348 PREDICTED: Salmo salar cis-aconitate decarboxylase-like (LOC106591804), mRNA Gene_23595 -0.704754 2.54x10-7 0.000583 PREDICTED: Salmo salar insulin-induced gene 1 protein-like (LOC106578605), mRNA Gene_21972 -1.047889 3.04x10-7 0.0006603 Oncorhynchus mykiss cytochrome P450 family 1 subfamily B polypeptide 1 (cyp1b1), mRNA Gene_26526 -0.915091 3.33x10-7 0.0006845 PREDICTED: Salmo salar N-acetyllactosaminide beta-1,3-N-acetylglucosaminyltransferase 2-like (LOC106604396), mRNA Gene_40146 -1.320238 7.42x10-7 0.0014494 Rainbow trout (S.gairdneri) cytochrome P450IA1 mRNA, complete cds Gene_41162 -1.177387 7.86x10-7 0.0014619 PREDICTED: Salmo salar cell death activator CIDE-3-like (LOC106596534), mRNA Gene_14766 1.0093415 1.25x10-6 0.0020748 PREDICTED: Salmo salar acetyl-CoA carboxylase 2-like (LOC106585658), transcript variant X4, mRNA Gene_28755 1.146295 1.32x10-6 0.0020748 PREDICTED: Salmo salar neural cell adhesion molecule 1-B-like (LOC106605634), transcript variant X2, mRNA Gene_4518 0.7905335 1.33x10-6 0.0020748 PREDICTED: Salmo salar receptor-type tyrosine-protein phosphatase delta-like (LOC106608606), partial mRNA Gene_21511 1.3101553 1.33x10-6 0.0020748 PREDICTED: Salmo salar uncharacterized LOC106570022 (LOC106570022), transcript variant X4, mRNA Gene_42515 1.9449047 1.61x10-6 0.0023754 PREDICTED: Salmo salar carbonic anhydrase 4-like (LOC106569487), transcript variant X2, mRNA 157 Name Log Fold Change Raw P Value FDR P Value Description Gene_23407 0.6228236 1.64x10-6 0.0023754 PREDICTED: Salmo salar folliculin-interacting protein 1-like (LOC106603820), mRNA Gene_41766 -2.192908 1.78x10-6 0.0024892 PREDICTED: Salmo salar apolipoprotein F (apof), transcript variant X1, mRNA Gene_35876 0.7537355 2.35x10-6 0.0031626 PREDICTED: Salmo salar regulator of nonsense transcripts 1-like (LOC106599024), transcript variant X3, mRNA Gene_42511 1.9779746 2.50x10-6 0.0031626 Salmo salar Carbonic anhydrase 12 (cah12), mRNA Gene_6047 0.9458802 2.51x10-6 0.0031626 PREDICTED: Salmo salar uridine phosphorylase 2-like (LOC106586173), mRNA Gene_10618 0.5141628 3.05x10-6 0.0036753 PREDICTED: Salmo salar pleckstrin homology-like domain family A member 1 (LOC106597692), mRNA Gene_9578 -4.05673 3.16x10-6 0.0036753 PREDICTED: Salmo salar STIP1 homology and U-box containing protein 1, E3 ubiquitin protein ligase (stub1), mRNA Gene_5550 0.9941901 3.20x10-6 0.0036753 PREDICTED: Salmo salar acetyl-CoA carboxylase 2-like (LOC106585658), transcript variant X9, mRNA Gene_23982 1.3527246 4.35x10-6 0.0048587 PREDICTED: Salmo salar centromere protein E, 312kDa (cenpe), transcript variant X10, mRNA Gene_3219 1.2560623 4.62x10-6 0.0049168 PREDICTED: Salmo salar uncharacterized LOC106589496 (LOC106589496), mRNA Gene_10026 -1.120278 4.66x10-6 0.0049168 PREDICTED: Salmo salar beta-2-glycoprotein 1-like (LOC106605775), transcript variant X2, mRNA Gene_38851 -1.025219 4.98x10-6 0.0051149 PREDICTED: Salmo salar C-C motif chemokine 19-like (LOC106565855), mRNA Gene_17279 -2.394893 5.46x10-6 0.0054708 Oncorhynchus mykiss mRNA for interleukin-22 precursor (il22 gene), allele a Gene_30512 1.0005499 6.24x10-6 0.0060914 PREDICTED: Salmo salar spectrin, beta, non-erythrocytic 5 (sptbn5), transcript variant X3, mRNA 158 Name Log Fold Change Raw P Value FDR P Value Description Gene_19559 1.1708019 7.43x10-6 0.0070825 Salmo salar Probable phosphatase phospho1 (phop1), mRNA Gene_37192 0.880408 8.05x10-6 0.0074923 PREDICTED: Salmo salar neural-cadherin-like (LOC106580611), mRNA Gene_20929 0.7765549 8.72x10-6 0.0079191 PREDICTED: Salmo salar bone morphogenetic protein 8B-like (LOC106570141), mRNA Gene_15159 1.0561885 9.48x10-6 0.0084128 PREDICTED: Salmo salar neural cell adhesion molecule 1-B-like (LOC106605634), transcript variant X1, mRNA Gene_30511 0.8800129 9.92x10-6 0.0086128 PREDICTED: Salmo salar spectrin, beta, non-erythrocytic 5 (sptbn5), transcript variant X3, mRNA Gene_13928 0.5805282 1.05x10-5 0.0089081 PREDICTED: Salmo salar Krueppel-like factor 4 (LOC106561364), transcript variant X2, mRNA Gene_30192 0.5364345 1.10x10-5 0.0091644 PREDICTED: Salmo salar UPF0392 protein F13G3.3-like (LOC106574488), transcript variant X4, mRNA Gene_6318 -0.490333 1.29x10-5 0.0105154 PREDICTED: Salmo salar regulator of G-protein signaling 1-like (LOC106584150), transcript variant X1, mRNA Gene_9522 -1.038095 1.38x10-5 0.011012 Salmo salar Duodenase-1 (ddn1), mRNA Gene_18510 0.4727638 1.55x10-5 0.0121118 PREDICTED: Salmo salar transcriptional activator GLI3-like (LOC106578726), transcript variant X2, mRNA Gene_9617 0.7369162 1.63x10-5 0.0124947 PREDICTED: Salmo salar amyloid beta (A4) precursor protein-binding, family B, member 2 (apbb2), transcript variant X7, mRNA Gene_37721 0.6595437 1.95x10-5 0.0143857 PREDICTED: Salmo salar zinc finger protein Gfi-1-like (LOC106584164), transcript variant X2, mRNA Gene_29457 0.8947942 1.95x10-5 0.0143857 PREDICTED: Salmo salar chromodomain-helicase-DNA-binding protein 6-like (LOC106565042), transcript variant X8, mRNA 159 Name Log Fold Change Raw P Value FDR P Value Description Gene_16011 1.3833927 2.01x10-5 0.0145448 PREDICTED: Salmo salar F-box only protein 40-like (LOC106602853), transcript variant X2, mRNA Gene_40498 0.7460666 2.12x10-5 0.0150797 PREDICTED: Salmo salar protein MRVI1-like (LOC106587202), transcript variant X2, mRNA Gene_23063 -0.890581 2.29x10-5 0.0157577 PREDICTED: Salmo salar scrapie-responsive protein 1-like (LOC106610171), transcript variant X2, misc_RNA Gene_11997 -1.322909 2.31x10-5 0.0157577 PREDICTED: Salmo salar phospholipase B1, membrane-associated-like (LOC106571734), mRNA Gene_26120 0.9617085 2.34x10-5 0.0157577 PREDICTED: Salmo salar synaptogyrin-1-like (LOC106607452), transcript variant X2, mRNA Gene_14783 0.4840515 2.54x10-5 0.0166186 PREDICTED: Salmo salar ethanolamine kinase 1-like (LOC106609326), mRNA Gene_40461 1.2346993 2.55x10-5 0.0166186 PREDICTED: Salmo salar ammonium transporter Rh type C (LOC106587392), mRNA Gene_12693 -0.666851 2.63x10-5 0.0166398 PREDICTED: Salmo salar proto-oncogene c-Fos-like (LOC106571028), transcript variant X2, mRNA Gene_3469 0.6924091 2.64x10-5 0.0166398 PREDICTED: Salmo salar Neural cell adhesion molecule 1-A (nca11), transcript variant X2, mRNA Gene_32443 0.910759 2.79x10-5 0.0170399 PREDICTED: Salmo salar dexamethasone-induced Ras-related protein 1-like (LOC106589334), mRNA Gene_45564 0.8224554 2.79x10-5 0.0170399 PREDICTED: Salmo salar myosin light chain kinase, smooth muscle-like (LOC106581830), transcript variant X4, mRNA Gene_2414 -1.180632 3.10x10-5 0.0186553 PREDICTED: Salmo salar basic leucine zipper transcriptional factor ATF-like (LOC106571033), mRNA Gene_29758 0.5844184 3.28x10-5 0.0193937 PREDICTED: Salmo salar ubiquitin carboxyl-terminal hydrolase CYLD-like (LOC106565781), transcript variant X3, mRNA 160 Name Log Fold Change Raw P Value FDR P Value Description Gene_37210 0.6981533 3.33x10-5 0.0193995 PREDICTED: Salmo salar annexin A2-like (LOC106580392), mRNA Gene_19474 0.4142777 3.54x10-5 0.0203412 PREDICTED: Salmo salar kinesin-like protein KIF13B (LOC106567792), transcript variant X7, mRNA Gene_15918 -1.595564 3.95x10-5 0.0223772 PREDICTED: Salmo salar uncharacterized LOC106587829 (LOC106587829), mRNA Gene_32699 -1.031189 4.04x10-5 0.0223772 PREDICTED: Salmo salar parvalbumin, thymic CPV3-like (LOC106589582), transcript variant X2, mRNAr Gene_3559 0.7214169 4.07x10-5 0.0223772 PREDICTED: Salmo salar ral guanine nucleotide dissociation stimulator-like 1 (LOC106607078), transcript variant X2, mRNA Gene_35955 -1.286028 4.12x10-5 0.0223772 PREDICTED: Salmo salar regulator of G-protein signaling 2-like (LOC106598560), mRNA Gene_11751 0.5315441 4.28x10-5 0.0228822 Salmo salar potassium voltage-gated channel, Isk-related family, member 4 (kcne4), mRNA Gene_11751 0.5315441 4.28x10-5 0.0228822 PREDICTED: Salmo salar long-chain-fatty-acid--CoA ligase 3-like (LOC106581298), mRNA Gene_22709 0.4167207 4.67x10-5 0.024285 PREDICTED: Salmo salar amyloid beta A4 precursor protein-binding family B member 1-like (LOC106603266), transcript variant X5, mRNA Gene_2686 0.8712749 4.71x10-5 0.024285 PREDICTED: Salmo salar beta-1,4-galactosyltransferase 6-like (LOC106560385), transcript variant X2, mRNA Gene_5706 -0.578067 4.72x10-5 0.024285 PREDICTED: Salmo salar fat storage-inducing transmembrane protein 2-like (LOC106572608), mRNA Gene_10818 1.8194202 4.85x10-5 0.0245997 PREDICTED: Salmo salar E3 ubiquitin-protein ligase RNF213-like (LOC106583653), mRNA Gene_2222 -3.702477 4.99x10-5 0.0250025 PREDICTED: Salmo salar CUB and zona pellucida-like domain-containing protein 1 (LOC106587759), mRNA 161 Name Log Fold Change Raw P Value FDR P Value Description Gene_35542 0.4709678 5.29x10-5 0.0258217 PREDICTED: Salmo salar autophagy-related protein 9A-like (LOC106599848), transcript variant X2, mRNA Gene_24330 0.5487228 5.33x10-5 0.0258217 PREDICTED: Salmo salar zinc finger protein 438-like (LOC106578558), transcript variant X2, mRNA Gene_34337 0.8472925 5.35x10-5 0.0258217 PREDICTED: Salmo salar homeobox protein six1a-like (LOC106612510), transcript variant X1, mRNA Gene_19156 1.1047811 5.42x10-5 0.0258217 PREDICTED: Salmo salar G protein-coupled receptor 160 (gpr160), transcript variant X2, mRNA Gene_46088 0.8558402 5.50x10-5 0.0258992 PREDICTED: Salmo salar transmembrane protein 44 (tmem44), transcript variant X1, mRNA Gene_33739 0.9488313 5.79x10-5 0.0268615 PREDICTED: Salmo salar POU class 2 homeobox 1 (pou2f1), transcript variant X17, mRNA Gene_7235 0.8425595 5.84x10-5 0.0268615 PREDICTED: Salmo salar neuroligin-2-like (LOC106608803), mRNA Gene_12402 1.4273241 6.31x10-5 0.0285807 PREDICTED: Salmo salar zinc finger protein GLIS2-like (LOC106606494), mRNA Gene_20591 0.9898313 6.36x10-5 0.0285807 PREDICTED: Salmo salar catenin delta-2-like (LOC106578422), transcript variant X2, mRNA Gene_19599 0.6262146 6.44x10-5 0.0286079 PREDICTED: Salmo salar ensconsin-like (LOC106611719), transcript variant X1, mRNA Gene_8860 0.4452057 6.66x10-5 0.0289782 PREDICTED: Salmo salar AF4/FMR2 family member 4-like (LOC106603851), transcript variant X2, mRNA Gene_6339 1.0532778 6.68x10-5 0.0289782 PREDICTED: Salmo salar WD repeat-containing protein on Y chromosome-like (LOC106580697), mRNA Gene_26169 -0.64555 6.89x10-5 0.0295918 PREDICTED: Salmo salar E3 ubiquitin/ISG15 ligase TRIM25-like (LOC106590355), transcript variant X2, mRNA Gene_17427 1.2465099 7.16x10-5 0.030384 PREDICTED: Salmo salar potassium channel subfamily K member 10-like (LOC106608174), mRNA 162 Name Log Fold Change Raw P Value FDR P Value Description Gene_9311 0.4833946 7.35x10-5 0.0306973 PREDICTED: Salmo salar LON peptidase N-terminal domain and RING finger protein 1-like (LOC106609964), mRNA Gene_29280 0.4511452 7.39x10-5 0.0306973 PREDICTED: Salmo salar inactive rhomboid protein 2-like (LOC106564910), transcript variant X2, mRNA Gene_25016 0.77895 7.64x10-5 0.031315 PREDICTED: Salmo salar asialoglycoprotein receptor 1-like (LOC106567873), mRNA Gene_39802 0.4176234 7.70x10-5 0.031315 PREDICTED: Salmo salar zinc finger protein 462-like (LOC106585746), transcript variant X2, mRNA Gene_29136 -0.662815 7.92x10-5 0.0318103 PREDICTED: Salmo salar solute carrier family 25 member 34-like (LOC106565132), transcript variant X5, mRNA Gene_42748 0.6396924 8.06x10-5 0.0318103 PREDICTED: Salmo salar integrin alpha-1-like (LOC106561820), transcript variant X1, mRNA Gene_10093 0.7062323 8.06x10-5 0.0318103 PREDICTED: Salmo salar NACHT, LRR and PYD domains-containing protein 12-like (LOC106591987), mRNA Gene_6824 0.6183224 8.25x10-5 0.0322371 PREDICTED: Salmo salar peroxisome proliferator-activated receptor gamma coactivator 1-beta (LOC100136477), transcript variant X2, mRNA Gene_32576 1.5010049 8.61x10-5 0.0332532 PREDICTED: Salmo salar dynein, axonemal, heavy chain 17 (dnah17), transcript variant X1, mRNA Gene_29038 -1.701983 8.68x10-5 0.0332532 PREDICTED: Salmo salar peptidyl-prolyl cis-trans isomerase FKBP5-like (LOC106565346), mRNA Gene_25119 -0.996334 9.71x10-5 0.0368237 PREDICTED: Salmo salar relaxin receptor 1-like (LOC106568031), mRNA Gene_43647 0.4714384 9.94x10-5 0.0373548 PREDICTED: Salmo salar semaphorin-3ab-like (LOC106572304), mRNA Gene_38528 1.0080946 1.03x10-4 0.0378264 PREDICTED: Salmo salar DNA-binding protein RFX7-like (LOC106584737), mRNA 163 Name Log Fold Change Raw P Value FDR P Value Description Gene_185 -3.024819 1.32x10-4 0.0378264 PREDICTED: Salmo salar CUB and zona pellucida-like domain-containing protein 1 (LOC106562604), mRNA Gene_39807 1.4665135 1.04x10-4 0.0378264 PREDICTED: Salmo salar lipocalin-like (LOC106585399), mRNA Gene_22466 0.6478082 1.06x10-4 0.0380219 PREDICTED: Salmo salar inositol 1,4,5-trisphosphate receptor type 2-like (LOC106573309), transcript variant X4, mRNA Gene_12174 0.6306884 1.06x10-4 0.0380219 PREDICTED: Salmo salar serine/threonine-protein kinase/endoribonuclease IRE1a-like (LOC106562560), transcript variant X2, mRNA Gene_25037 0.7473985 1.07x10-4 0.0380219 PREDICTED: Salmo salar transmembrane protease serine 13-like (LOC106603367), mRNA Gene_14772 0.93567 1.09x10-4 0.0382963 Oncorhynchus mykiss toxin-1 (LOC100135970), mRNA Gene_29556 -0.486721 1.12x10-4 0.0389612 PREDICTED: Salmo salar neuroblast differentiation-associated protein AHNAK-like (LOC106587602), mRNA Gene_19199 -1.869064 1.14x10-4 0.0393831 Oncorhynchus mykiss cytochrome P450 1A2 (CYP1A2) mRNA, complete cds Gene_25498 0.8208758 1.17x10-4 0.0398733 PREDICTED: Salmo salar nuclear pore complex protein Nup155-like (LOC106563918), transcript variant X4, mRNA Gene_25803 -1.172093 1.17x10-4 0.0398733 PREDICTED: Salmo salar hemicentin-1-like (LOC106601185), misc_RNA Gene_28837 1.3457363 1.22 x10-4 0.0411505 PREDICTED: Salmo salar nck-associated protein 5-like (LOC106565432), transcript variant X10, mRNA Gene_40423 0.5822357 1.28 x10-4 0.0426843 PREDICTED: Salmo salar DNA-binding protein RFX7-like (LOC106587434), transcript variant X1, mRNA Gene_31702 1.4755893 1.29 x10-4 0.0426843 Salmo salar IgH locus B genomic sequence Gene_1787 0.9302684 1.32 x10-4 0.0431896 PREDICTED: Salmo salar chromodomain-helicase-DNA-binding protein 6-like (LOC106565042), transcript variant X2, mRNA 164 Name Log Fold Change Raw P Value FDR P Value Description Gene_3215 -3.759777 1.36 x10-4 0.0443438 PREDICTED: Salmo salar integrin beta-like protein A (LOC106587907), partial mRNA Gene_30397 -0.929679 1.38 x10-4 0.0443438 PREDICTED: Salmo salar phospholipase A2 inhibitor 31 kDa subunit-like (LOC106575194), transcript variant X2, mRNA Gene_34211 0.9580678 1.40 x10-4 0.0443438 PREDICTED: Salmo salar sideroflexin-1-like (LOC106577244), mRNA Gene_23123 0.7305482 1.40 x10-4 0.0443438 PREDICTED: Salmo salar translation initiation factor IF-2-like (LOC106602737), mRNA Gene_33153 0.5564065 1.44 x10-4 0.0453767 PREDICTED: Salmo salar galactose-3-O-sulfotransferase 4 (gal3st4), mRNA Gene_37855 0.6815812 1.46 x10-4 0.0455444 PREDICTED: Salmo salar SH3-containing GRB2-like protein 3-interacting protein 1 (LOC106583991), mRNA Gene_23408 0.4753649 1.52 x10-4 0.0471022 PREDICTED: Salmo salar folliculin-interacting protein 1-like (LOC106603820), mRNA Gene_5694 0.7915528 1.57 x10-4 0.0478471 PREDICTED: Salmo salar WD repeat-containing protein on Y chromosome-like (LOC106580697), mRNA Gene_19368 -0.946509 1.57 x10-4 0.0478471 PREDICTED: Salmo salar Krueppel-like factor 9 (LOC106585715), transcript variant X2, mRNA Gene_18378 -0.554233 1.62 x10-4 0.0489874 Salmo salar Acetyl-CoA acetyltransferase, cytosolic (thic), mRNA Gene_33207 -1.389056 1.63 x10-4 0.049033 PREDICTED: Salmo salar cadherin-related family member 5-like (LOC106609137), transcript variant X3, mRNA Gene_1251 0.6292917 1.65 x10-4 0.0493313 PREDICTED: Salmo salar receptor-type tyrosine-protein phosphatase S-like (LOC106574153), transcript variant X16, mRNA Gene_12938 0.6913653 1.69 x10-4 0.0493313 PREDICTED: Salmo salar histone deacetylase 5-like (LOC106607176), transcript variant X6, mRNA 165 Name Log Fold Change Raw P Value FDR P Value Description Gene_32954 -0.952747 1.71 x10-4 0.0493313 PREDICTED: Salmo salar S-arrestin-like (LOC106579793), mRNA Gene_9356 0.4646784 1.72 x10-4 0.0493313 PREDICTED: Salmo salar oxysterol-binding protein-related protein 3-like (LOC106570528), transcript variant X3, mRNA Gene_31865 1.1599133 1.73 x10-4 0.0493313 PREDICTED: Salmo salar nestin (nes), transcript variant X1, mRNA Gene_28244 0.6372017 1.73 x10-4 0.0493313 PREDICTED: Salmo salar synapse differentiation-inducing gene protein 1-like (LOC106605956), transcript variant X3, mRNA Gene_22337 0.5451894 1.74 x10-4 0.0493313 PREDICTED: Salmo salar mitogen-activated protein kinase kinase kinase kinase 3-like (LOC106577155), transcript variant X7, mRNA Gene_18607 1.7659721 1.75 x10-4 0.0493313 Salmo salar homeobox protein HoxB13ab (hoxb13ab), mRNA Gene_45912 -1.595224 1.77 x10-4 0.0493313 Salmo salar homeobox protein HoxD10aa (hoxd10aa), mRNA Gene_34927 0.6512817 1.77 x10-4 0.0493313 Salmo salar NFAT5b1 mRNA, complete cds, alternatively spliced 166 Table A.4: GO analysis of differentially expressed genes within Blackwater Rainbow Trout between control and acclimated treatments GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0007155 cell adhesion BP 2.64x10-8 1 122 347 GO:0005576 extracellular region CC 6.35x10-10 1 157 422 GO:0003779 actin binding MF 4.63x10-8 1 117 336 GO:0005509 calcium ion binding MF 6.43x10-6 0.999995846 216 728 Table A.5: GO analysis of differentially expressed genes within Eagle Lake Rainbow Trout between control and acclimated treatments GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0006412 translation BP 3.13x10-23 1 169 284 GO:0005840 ribosome CC 3.31x10-32 1 154 219 GO:0030529 intracellular ribonucleoprotein complex CC 2.03x10-11 1 83 142 GO:0005622 intracellular CC 4.62x10-5 0.999963787 503 1453 GO:0015934 large ribosomal subunit CC 6.34x10-5 0.999993956 13 16 GO:0003735 structural constituent of ribosome MF 8.30x10-28 1 150 224 167 Table A.6: GO analysis of differentially expressed genes within Fraser Valley Rainbow Trout between control and acclimated treatments GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0015671 oxygen transport BP 2.93x10-11 1 16 24 GO:0060326 cell chemotaxis BP 1.50x10-8 0.999999998 17 36 GO:0006816 calcium ion transport BP 2.47x10-8 0.999999995 26 71 GO:0051603 proteolysis involved in cellular protein catabolic process BP 5.67x10-8 0.999999991 19 47 GO:0006955 immune response BP 9.25x10-8 0.999999974 33 120 GO:0007292 female gamete generation BP 4.04x10-7 0.999999994 8 9 GO:0050829 defense response to Gram-negative bacterium BP 7.40x10-5 0.999996456 6 9 GO:0006937 regulation of muscle contraction BP 9.12x10-5 0.999995457 6 9 GO:0007015 actin filament organization BP 1.33x10-4 0.999963132 17 59 GO:0006935 chemotaxis BP 1.50x10-4 0.999969525 12 36 GO:0048016 inositol phosphate-mediated signaling BP 3.34x10-4 0.999975653 6 10 GO:0005833 hemoglobin complex CC 1.73x10-11 1 14 18 GO:0000502 proteasome complex CC 2.29x10-8 0.999999996 20 49 GO:0005839 proteasome core complex CC 5.67x10-8 0.999999991 19 47 GO:0005783 endoplasmic reticulum CC 3.41x10-7 0.999999874 44 188 GO:0005861 troponin complex CC 2.75x10-5 0.999996633 10 22 GO:0033180 proton-transporting V-type ATPase, V1 domain CC 8.18x10-5 0.999991766 8 16 GO:0005737 cytoplasm CC 8.20x10-5 0.999947578 113 749 168 GO-ID Term Ontology p-value (over-represented) p-value (under-represented) Number of Differentially Expressed Genes Total Number of Genes GO:0008250 oligosaccharyltransferase complex CC 1.10x10-4 1 4 4 GO:0005623 cell CC 3.12x10-4 0.999863352 31 154 GO:0019825 oxygen binding MF 1.15x10-11 1 17 26 GO:0005344 oxygen transporter activity MF 2.93x10-11 1 16 24 GO:0008009 chemokine activity MF 2.10x10-10 1 20 40 GO:0004175 endopeptidase activity MF 1.54x10-8 0.999999998 19 44 GO:0004298 threonine-type endopeptidase activity MF 5.67x10-8 0.999999991 19 47 GO:0005262 calcium channel activity MF 9.15x10-6 0.99999847 14 34 GO:0020037 heme binding MF 2.02x10-5 0.999992991 28 118 GO:0046961 proton-transporting ATPase activity, rotational mechanism MF 2.13x10-4 0.999979286 7 14 GO:0008233 peptidase activity MF 2.36x10-4 0.999877489 48 271 GO:0005220 inositol 1,4,5-trisphosphate-sensitive calcium-release channel activity MF 3.34x10-4 0.999975653 6 10 169 Table A.7: Water chemistry parameters for experimental lakes Name Units Rexford Hosli Sampling Date 23-Jun-14 29-Sep-14 31-May-15 14-Sep-15 20-May-14 22-Sep-14 20-May-15 08-Sep-15 pH 8.2 8.2 8.06 8.34 8.64 8.74 8.52 8.71 Alkalinity Total as CaCO3 mg/L 159 163 302 319 Total Hardness CaCO3 mg/L 149 166 152 151 223 291 239 266 Dissolved Hardness CaCO3 mg/L 142 167 150 155 216 278 236 271 Conductivity uS/cm 291 303 301 313 513 558 573 581 Total Dissolved Solids mg/L 152 198 190 202 318 332 358 324 Orthophosphate P mg/L <0.0010 0.0013 0.0043 0.0016 0.0018 Dissolved Sulphate SO4 mg/L 0.99 <0.50 0.54 0.65 0.74 0.53 0.87 <0.50 Bromide Br mg/L 0.015 0.015 0.016 0.015 <0.10 0.06 0.076 0.062 Dissolved Boron B mg/L <50 <20 <10 <10 <50 <20 12 12 Fluoride F mg/L 0.22 0.23 0.4 0.46 Dissolved Chloride Cl mg/L 1.7 1.7 1.4 1.6 4.4 4.9 5 5.2 Nitrate N mg/L 0.0044 <0.0020 <0.0020 <0.0020 0.0066 <0.0020 0.002 0.0022 Nitrite N mg/L <0.0020 <0.0020 <0.0020 <0.0020 0.0025 <0.0020 <0.0020 <0.0020 Total Organic Nitrogen N mg/L 0.377 0.808 0.343 0.351 1.22 1.26 1.18 1.31 Total Nitrogen N mg/L 0.396 0.844 0.35 0.373 1.3 1.26 1.24 1.35 Ammonia N mg/L 0.015 0.035 0.011 0.022 0.073 <0.0050 0.059 0.03 Total Phosphorus P mg/L 0.0045 0.0055 0.0124 0.0057 0.0065 Dissolved Phosphorus P mg/L 0.0033 0.006 0.0078 0.0038 0.0064 Dissolved Calcium Ca mg/L 36.3 44.5 38.4 39.8 19.9 25.3 22.4 23.7 Dissolved Magnesium Mg mg/L 12.4 13.7 13 13.4 40.4 52.2 44 51.3 Total Aluminum Al ug/L 3.28 2.77 1.58 2.89 7.3 6.37 3.9 6.79 170 Name Units Morgan Lower Sampling Date 20-May-14 29-Sep-14 20-May-15 30-Sep-15 19-May-14 26-May-15 28-Sep-15 pH 9.03 9.12 9.01 8.99 8.47 8.48 8.73 Alkalinity Total as CaCO3 mg/L 590 663 289 319 330 Total Hardness CaCO3 mg/L 365 434 384 422 246 253 277 Dissolved Hardness CaCO3 mg/L 364 424 406 421 233 256 271 Conductivity uS/cm 1150 1170 1170 1290 519 570 588 Total Dissolved Solids mg/L 766 816 790 870 316 234 312 Orthophosphate P mg/L 0.0052 0.005 0.0053 0.0024 0.0013 <0.0010 Dissolved Sulphate SO4 mg/L 91.9 106 99.7 102 0.87 0.98 1.05 Bromide Br mg/L 0.11 0.126 0.1 0.29 <0.10 0.036 0.036 Dissolved Boron B mg/L 1140 1440 1410 1360 <50 <10 10 Fluoride F mg/L 0.21 0.22 0.18 Dissolved Chloride Cl mg/L 7 7.6 7.4 7.9 2.8 4.1 3.8 Nitrate N mg/L 0.0049 0.0482 0.007 0.037 <0.0020 0.0025 <0.0020 Nitrite N mg/L <0.0020 0.0211 <0.0020 0.013 0.0031 <0.0020 <0.0020 Total Organic Nitrogen N mg/L 1.15 1.47 1.25 1.31 0.729 0.811 0.693 Total Nitrogen N mg/L 1.19 1.62 1.3 1.51 0.747 0.86 0.76 Ammonia N mg/L 0.028 0.085 0.049 0.15 0.013 0.05 0.068 Total Phosphorus P mg/L 0.0224 0.0285 0.0368 0.0156 0.0142 0.0102 Dissolved Phosphorus P mg L mg/L 0.0222 0.0198 0.0358 0.0125 0.0121 0.0102 Dissolved Calcium Ca mg/L 12.3 12.2 14.2 10 35 36.6 38.1 Dissolved Magnesium Mg mg/L 80.9 95.4 90 96.1 35.4 40 42.8 Total Aluminum Al ug/L 8.28 9 11.4 9.89 11.2 2.85 2.05 171 Name Units Pigeon 2 Whale Sampling Date 19-May-14 25-May-15 21-Sep-15 19-May-14 25-May-15 23-Sep-15 pH 8.75 8.76 8.76 8.96 9.05 9.06 Alkalinity Total as CaCO3 mg/L 390 400 393 552 574 590 Total Hardness CaCO3 mg/L 318 319 341 471 543 Dissolved Hardness CaCO3 mg/L 306 313 345 454 529 Conductivity uS/cm 640 673 666 872 888 928 Total Dissolved Solids mg/L 372 388 349 544 562 574 Orthophosphate P mg/L 0.009 0.0013 <0.0010 <0.0010 <0.0010 0.0015 Dissolved Sulphate SO4 mg/L <0.50 <0.50 <0.50 <0.50 0.58 <0.50 Bromide Br mg/L <0.10 <0.10 0.024 <0.10 <0.10 <0.10 Dissolved Boron B mg/L <50 <10 <10 <50 <10 Fluoride F mg/L 0.24 0.18 Dissolved Chloride Cl mg/L 1.8 1.6 2.1 3.9 4.4 4.7 Nitrate N mg/L <0.0020 <0.0020 <0.0020 0.0026 0.0024 <0.0020 Nitrite N mg/L <0.0020 <0.0020 <0.0020 0.0044 <0.0020 <0.0020 Total Organic Nitrogen N mg/L 0.77 0.724 0.775 1 0.991 1.05 Total Nitrogen N mg/L 0.783 0.77 0.818 1.06 1.02 1.1 Ammonia N mg/L 0.013 0.042 0.044 0.051 0.031 0.05 Total Phosphorus P mg/L 0.0082 0.0163 0.0102 0.0099 0.0077 0.0159 Dissolved Phosphorus P mg L mg/L 0.0092 0.0107 0.0096 0.0079 0.0073 0.0134 Dissolved Calcium Ca mg/L 14.2 13.3 13.8 7.63 7.39 Dissolved Magnesium Mg mg/L 65.7 68 74.4 106 124 Total Aluminum Al ug/L 2.61 0.98 0.27 10.9 1.78 172 Appendix Figures Figure A.1: Heat map displaying expression patterns of genes (8,053) with significant effects of pH in the Blackwater (BW) strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene. Control = fish acclimated to pH 7.2; treatment = fish acclimated to pH 8.8 for 1 month. 173 Figure A.2: Heat map displaying expression patterns of genes (10,509) with significant effects interaction between pH in the Eagle Lake strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene. 174 Figure A.3: Heat map displaying expression patterns of genes (3445) with significant effects of pH in the Fraser Valley (FV) strain of Rainbow Trout. Blue indicates genes with expression lower than mean expression across all samples. Yellow indicates genes with expression higher than mean across all samples. Each column represents one individual (n=6 per group) and each row represents expression values (log2 counts per million) for one gene which have been normalized to mean expression values of that gene. Control = fish acclimated to pH 7.2; treatment = fish acclimated to pH 8.8 for 1 month.