Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Adaptive variation of mitochondrial function in response to oxygen variability in intertidal sculpins… Lau, Gigi Yik Chee 2017

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

Item Metadata

Download

Media
24-ubc_2018_february_lau_gigi.pdf [ 3.38MB ]
Metadata
JSON: 24-1.0357971.json
JSON-LD: 24-1.0357971-ld.json
RDF/XML (Pretty): 24-1.0357971-rdf.xml
RDF/JSON: 24-1.0357971-rdf.json
Turtle: 24-1.0357971-turtle.txt
N-Triples: 24-1.0357971-rdf-ntriples.txt
Original Record: 24-1.0357971-source.json
Full Text
24-1.0357971-fulltext.txt
Citation
24-1.0357971.ris

Full Text

ADAPTIVE VARIATION OF MITOCHONDRIAL FUNCTION IN RESPONSE TO OXYGEN VARIABILITY IN INTERTIDAL SCULPINS (COTTIDAE, ACTINOPTERYGII)  by  Gigi Yik Chee Lau  B.Sc., The University of British Columbia, 2007 M.Sc., The University of British Columbia, 2010   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) November 2017  © Gigi Yik Chee Lau, 2017  ii  Abstract Variation in environmental oxygen (O2) poses a significant physiological challenge to animals, not only because of the impact on aerobic metabolism, but also because it can lead to generation of potentially harmful reactive oxygen species (ROS). In this thesis, I aimed to investigate the interplay between two aspects of O2 use at the mitochondria, aerobic respiration and ROS metabolism, using species of intertidal sculpins (Cottidae, Actinopterygii) which are distributed along the marine intertidal zone, exposed to varying O2 conditions and vary in their tolerance to low O2 (hypoxia).  I first hypothesized that there would be a relationship between whole animal hypoxia tolerance and mitochondrial and cytochrome c oxidase (COX) O2-binding affinity, whereby hypoxia tolerant sculpins would have higher mitochondrial and COX O2-binding affinity than less hypoxia tolerant sculpins. This hypothesis was supported with functional analysis. In silico modelling of the COX catalytic core revealed that the variation in O2 binding was related to interspecific differences in the interaction between COX3 and membrane phospholipid, cardiolipin, which could impact O2 diffusion to its binding site.  I then investigated whether intact mitochondria from hypoxia tolerant sculpins were able to use O2 more efficiently such that phosphorylation efficiency was improved and ROS generation was reduced compared to mitochondria from less hypoxia tolerant sculpins. Although there were relationships between hypoxia tolerance and complex I and II dependencies, there were no interspecies differences in phosphorylation or mitochondrial coupling that would indicate differences in aerobic metabolism. Moreover, mitochondria from hypoxia tolerant sculpins generated more ROS under resting conditions and were more perturbed by in vitro redox and anoxia-recovery challenges.  Finally, I confirmed consistent responses of mitochondria to in vivo responses with a whole animal study comparing ROS metabolism (redox status, mitochondrial H2O2, oxidative damage and scavenging capacity) between two sculpin species with different hypoxia tolerance to hypoxia, hyperoxia, with normoxia-recovery exposures.  Taken together, my thesis demonstrates that hypoxia tolerance is associated with improved O2 binding at the mitochondria and COX. Further, hypoxia tolerance in sculpins is associated with iii  higher ROS generation compared to less tolerant species, suggesting a potentially important role of ROS in mediating hypoxia tolerance.   iv  Lay Summary  Animals are highly dependent on oxygen (O2). The use of O2, however, comes with both advantages and disadvantages. On one hand, O2 is key to the process that produces chemical energy within mitochondria inside the cell, on the other hand the use of O2 generates reactive oxygen species (ROS), which are harmful byproducts. In this thesis, I investigated these two aspects of O2 use at mitochondria from a group of sculpin fishes that are distributed along the marine intertidal zone and are naturally exposed to daily fluctuations of O2. I found that more hypoxia tolerant sculpins improved O2 binding at the level of mitochondria. However, this increased O2 binding was not associated with increased aerobic energy production, and counterintuitively, there was higher ROS generation in more hypoxia tolerant sculpins. It is possible that higher ROS generation in hypoxia tolerant sculpins is part of the strategy in surviving the O2 variable intertidal.  v  Preface A version of Chapter 2 has been published as “Lau, G.Y., Mandic, M. and Richards, J.G. (2017). Evolution of cytochrome c oxidase in hypoxia tolerant sculpins (Cottidae, Actinopterygii). Molecular Biology and Evolution. 34, 2153-2162”. I designed and carried out the protocols for the high-resolution respirometry experiments with input from Dr. Jeff Richards. I also carried out the in silico protein analyses. Dr. Milica Mandic isolated mitochondrial samples that were used for the COX Km,app O2 determination and also provided advice on statistical analyses. I wrote the manuscript with editorial input from Drs. Mandic and Richards.  I designed and carried out the protocols for the high-resolution respirometry and fluorometry in Chapter 3. I wrote the manuscript with editorial input from Dr. Richards.  I designed and carried out the whole animal exposure in Chapter 4 with input from Drs. Sabine Arndt, Michael Murphy and Richards. Dr. Arndt also performed mass spectrometry analysis on the MitoB samples. I performed the biochemical analyses on all tissue samples. I wrote the final manuscript with editorial input from Drs. Arndt, Murphy, and Richards.   All experiments performed for the various sculpin species were approved by the UBC Animal Care Committee (Protocol A13-0309).  vi  Table of Contents  Abstract ..................................................................................................................................... ii Lay Summary ............................................................................................................................ iv Preface ....................................................................................................................................... v Table of Contents ..................................................................................................................... vi List of Tables .............................................................................................................................x List of Figures .......................................................................................................................... xi Acknowledgements ................................................................................................................. xv Chapter One: General Introduction .......................................................................................... 1 1.1 OXYGEN AND THE EVOLUTION OF AEROBIC METABOLISM .............................................................. 1 1.2 THE MITOCHONDRION .......................................................................................................................... 3 1.3 FACTORS AFFECTING MITOCHONDRIAL FUNCTION ........................................................................... 3 1.3.1 Proton leak ....................................................................................................................................... 3 1.3.2 Electron leak .................................................................................................................................... 4 1.3.3 O2 delivery to mitochondria .......................................................................................................... 6 1.4 MAMMALIAN MITOCHONDRIAL RESPONSES TO O2 VARIABILITY ..................................................... 7 1.4.1 Mammalian mitochondrial responses to acute changes in O2 .................................................. 7 1.4.2 Mammalian mitochondrial responses to chronic changes in O2 .............................................. 9 1.5 O2 VARIABILITY IN THE NATURAL ENVIRONMENT ........................................................................... 10 1.6 DEFINING HYPOXIA TOLERANCE ....................................................................................................... 11 1.7 ADAPTATION AND PLASTICITY OF MITOCHONDRIA TO O2 VARIABILITY ...................................... 13 1.7.1 O2 binding at mitochondria and COX ....................................................................................... 13 1.7.2 Mitochondrial ETS function ....................................................................................................... 15 1.7.3. ROS metabolism .......................................................................................................................... 16 1.8 THESIS OBJECTIVES AND CHAPTER HYPOTHESES ............................................................................. 17 1.8.1 Using intertidal sculpins (Cottidae, Actinopterygii) as model ................................................. 17 1.8.2 Chapter hypotheses ....................................................................................................................... 18 Chapter Two: Evolution of cytochrome c oxidase in hypoxia tolerant sculpins (Cottidae, Actinopterygii) ......................................................................................................................... 24 2.1 INTRODUCTION ..................................................................................................................................... 24 2.2 MATERIALS AND METHODS .................................................................................................................. 26 2.2.1 Chemicals ........................................................................................................................................ 26 2.2.2 Species collection and holding .................................................................................................... 26 2.2.3 Mitochondrial isolation ................................................................................................................. 26 vii  2.2.4 COX Km,app O2 ............................................................................................................................... 27 2.2.5 Mitochondrial P50 .......................................................................................................................... 27 2.2.6 COX respiration rate (ascorbate-TMPD/FCCP respiration rate) ......................................... 28 2.2.7 COX voltage recovery rate .......................................................................................................... 28 2.2.8 cox1 and cox3 sequencing .............................................................................................................. 29 2.2.9 Protein in silico analyses ................................................................................................................. 29 2.2.10 Statistical analyses ........................................................................................................................ 30 2.3 RESULTS AND DISCUSSION ................................................................................................................... 30 2.3.1 Interspecific variation in COX function and mitochondrial P50 ............................................ 30 2.3.2 COX protein in silico analyses ...................................................................................................... 34 2.3.3 COX1 protein structure ............................................................................................................... 35 2.3.4 COX3 protein structure and stability ......................................................................................... 35 2.3.5 Interspecific variation in mitochondrial kinetics does not affect proton pumping ............. 36 2.3.6 COX3 protein stability may affect cardiolipin interactions and mitochondrial function ... 36 2.4 SUMMARY ................................................................................................................................................ 37 Chapter Three: Hypoxia tolerance is associated with higher mitochondrial ROS emission in intertidal sculpins (Cottidae, Actinopterygii) ......................................................................... 46 3.1 INTRODUCTION ..................................................................................................................................... 46 3.2 METHODS ............................................................................................................................................... 48 3.2.1 Species collection and holding .................................................................................................... 48 3.2.2 Isolation of brain mitochondria .................................................................................................. 48 3.2.3 Part I: Mitochondrial respiration ................................................................................................. 49 3.2.4 Part II: Mitochondrial respiration and simultaneous ROS measurements ........................... 49 3.2.5 Part III: In vitro redox challenge .................................................................................................. 50 3.2.6 Part IV: Recovery from in vitro anoxia........................................................................................ 51 3.2.7 Biochemical analyses ..................................................................................................................... 51 3.2.8 Calculations and statistical analyses ............................................................................................ 52 3.3 RESULTS .................................................................................................................................................. 53 3.3.1 Part I: Interspecific relationship between hypoxia tolerance and complex I and II respiratory flux capacity ......................................................................................................................... 53 3.3.2 Part II: Interspecific differences in ROS/O2 in state IV ......................................................... 53 3.3.3 Part III: Effects of manipulating the redox environment on ROS/O2 ................................. 54 3.3.4 Part IV: In vitro anoxia-reoxygenation exposure ....................................................................... 54 3.4 DISCUSSION ............................................................................................................................................ 55 3.4.1 Hypoxia tolerant sculpins utilize less of complex I in overall ETS flux ............................... 55 3.4.2 Hypoxia tolerant sculpins emit higher ROS/O2 with lower state IV respiration rate ........ 57 3.4.3 Hypoxia tolerant sculpins buffer mitochondrial redox changes better but generate higher ROS/O2.................................................................................................................................................... 58 3.4.4 More hypoxia tolerant sculpins do not recover better from in vitro anoxia .......................... 60 3.4.5 A potential role for ROS in response to environmental O2 variability ................................. 60 viii  3.5 SUMMARY ................................................................................................................................................ 61 Chapter Four: Whole animal responses of ROS metabolism to hypoxia- and hyperoxia- recovery .................................................................................................................................... 70 4.1 INTRODUCTION ..................................................................................................................................... 70 4.2 METHODS ............................................................................................................................................... 71 4.2.1 Equipment ...................................................................................................................................... 71 4.2.2 Animals ........................................................................................................................................... 72 4.2.3 Validation of MitoB use in sculpins ........................................................................................... 72 4.2.4 Experimental protocol and sampling ......................................................................................... 72 4.2.5 Purification of tissue samples for MitoP/MitoB ...................................................................... 73 4.2.6 Tissue glutathione redox status (GSH:GSSG) .......................................................................... 74 4.2.7 Thiobarbuturic acid reactive substances (TBARS) ................................................................... 74 4.2.8 Total oxidative scavenging capacity (TOSC)............................................................................. 75 4.2.9 Statistical analyses .......................................................................................................................... 75 4.3 RESULTS .................................................................................................................................................. 76 4.3.1 Validation of MitoB and MitoP in marine sculpins ................................................................. 76 4.3.2 Hypoxia and hypoxia-recovery at a common PO2 (3.5kPa) .................................................... 76 4.3.3 Effects of hypoxic PO2 on responses in O. maculosus (2.3 and 3.5kPa) ................................. 77 4.3.4 Hyperoxia and hyperoxia-recovery at a common PO2 (64kPa) .............................................. 79 4.4. DISCUSSION ........................................................................................................................................... 81 4.4.1 MitoB for the detection of in vivo mitochondrial ROS generation in marine sculpins ........ 82 4.4.2 Hypoxia and hyperoxia exposure had greater effects on ROS metabolism in O. maculosus than S. marmoratus .................................................................................................................................... 83 4.4.3 Tissue specific responses of ROS metabolism to O2 stress .................................................... 85 4.4.4 Effect of PO2 on ROS metabolism in O. maculosus .................................................................. 86 4.4.5 Greater response in ROS metabolism as an adaptive response to O2 variability ................ 87 4.4.6 Summary ......................................................................................................................................... 88 Chapter Five:  General Discussion and Conclusion ............................................................. 102 5.1 MAJOR FINDINGS ............................................................................................................................... 102 5.1.1 Adaptive variation in oxygen binding of mitochondria and COX (Chapter 2) ................ 102 5.1.2 Hypoxia tolerance in sculpins is not associated with efficient mitochondrial O2 use (Chapter 3) ............................................................................................................................................ 105 5.1.3 Hypoxia tolerant sculpins generate more ROS (Chapters 3&4) ......................................... 106 5.1.4 Summary model: hypoxia tolerant vs. hypoxia intolerant sculpin ....................................... 108 5.2 STUDY CONSIDERATIONS ................................................................................................................. 108 5.2.1 Interacting abiotic factors in the intertidal.............................................................................. 108 5.2.2 Technical challenges of assessing ROS metabolism in vivo and in vitro ............................... 109 5.2.3 Multiple factors affect ROS metabolism ................................................................................ 110 5.3 FUTURE STUDIES ................................................................................................................................ 111 ix  5.3.1 Are there other underlying mechanisms contributing to interspecies variation in COX function? ................................................................................................................................................ 111 5.3.2 What are the functional consequences of the difference in ETS complex flux capacities when comparing sculpins of varying hypoxia tolerance? ............................................................... 112 5.3.3 Is there a role of ROS in the adaptive response to O2 stress? ............................................. 113 5.4 CONCLUSION ....................................................................................................................................... 113 Bibliography ........................................................................................................................... 118 Appendix ................................................................................................................................ 135  x  List of Tables  TABLE 2.1 Regressions tested under ordinary least squares (OLS) and phylogenetic generalized least squares (PGLS) with Pagel’s model of evolution (Pagel 1999)………………………………..45 TABLE 3.1 ADP/O, respiratory control ratios, and mitochondrial complex maximal activities from five species of sculpins. …………………………………………………………………………...68 TABLE 3.2 Regressions were tested with ordinary least squares (OLS) and phylogenetic generalized least squares (PGLS) with Pagel’s model of evolution (Pagel 1999) ……………………………….69  xi  LIST OF FIGURES  FIGURE 1.1 Vertebrate electron transport system………………………………………………..21 FIGURE 1.2 Balance of ROS scavenging rate and generation rate that results in accumulation of ROS………………………………………………………………………………………………22 FIGURE 1.3. Hypothesis for an O2 efficient electron transport system…..………………………23 FIGURE 2.1 (A) Relationship between whole animal hypoxia tolerance (Pcrit) and brain mitochondrial P50 and COX Km,app O2; and, relationship between brain mitochondrial P50 and (B) COX Vmax enzyme activities among sculpins, and (C) ascorbate-TMPD stimulated mitochondrial respiration (COX respiration rate)………………………………………………………………...39 FIGURE 2.2 (A) Structure of whole COX enzyme (bovine heart 3ABM PDB structure) with COX1, COX2, COX3, and COX7a highlighted; (B) COX1 and COX3 structures that were investigated for interspecific differences between sculpin species; (C) Relationship between brain COX Km,app O2 and COX3 subunit protein stability (estimated as free energy of unfolding, in kcal/mol); (D) COX3 structure showing sculpin interspecific differences in amino acid residues…………………………………………………………………………………………...41 FIGURE 2.3 Phylogenetic transitions of two amino acid residues 55 and 224 on COX3………...43 FIGURE 3.1. Relationship between whole animal hypoxia tolerance as assessed by critical oxygen tensions (Pcrit) and mitochondrial respiration rates………………………………………………...62 FIGURE 3.2. Substrate-inhibitor titration protocol with simultaneous measurement of oxygen consumption rate and reactive oxygen species generated (expressed as ROS/O2) in isolated brain mitochondria in three sculpin species……………………………………………………………...64 FIGURE 3.3. Relationship between mitochondrial pellet GSH:GSSG and ROS/O2 in O. maculosus and S. marmoratus …………………………………………………………………………………65 FIGURE 3.4. The effect of 20min in vitro anoxia-recovery on O2 consumption and ROS emission in isolated brain mitochondria from three sculpin species……...………………………………….66 FIGURE 4.1. The effect of hypoxia (3.5kPa)-recovery and hyperoxia (64.0kPa)-recovery in brain of Oligocottus maculosus and Scorpaenichthys marmoratus on ROS metabolism…………………………….89 FIGURE 4.2. The effect of hypoxia (3.5kPa)-recovery and hyperoxia (64.0kPa)-recovery in liver of O. maculosus and S. marmoratus on ROS metabolism………………….…………………………….91 xii  FIGURE 4.3. The effect of hypoxia (3.5kPa)-recovery and hyperoxia (64.0kPa)-recovery in gill of O. maculosus and S. marmoratus on ROS metabolism…………………….………………………….93 FIGURE 4.4. The effect of 2.3kPa and 3.5kPa hypoxia-recovery in brain of O. maculosus on ROS metabolism………………………………………………………………………………………..95 FIGURE 4.5. The effect of 2.3kPa and 3.5kPa hypoxia-recover in liver of O. maculosus on ROS metabolism………………………………………………………………………………………..97 FIGURE 4.6. The effect of 2.3kPa and 3.5kPa hypoxia-recovery in gill of O. maculosus on ROS metabolism………………………………………………………………………………………..99 FIGURE 4.7. Effect of relative hypoxia exposure and recovery in liver TBARS and TOSC of O. maculosus and S. marmoratus (2.3kPa for O. maculosus and 3.5kPa for S. marmoratus…………………101 FIGURE 5.1. Relationship of the various levels of the sculpin oxygen transport cascade………..115 FIGURE 5.2. Revised model of brain mitochondria ETS associated with hypoxia tolerance in sculpins…………………………………………………………………………………………..116 FIGURE 5.3. Revised model comparing the oxygen transport cascade between hypoxia tolerant and intolerant sculpins…………………………………………………………………………...117  xiii  LIST OF ABBREVIATIONS [E] enzyme concentration  ADP adenosine diphosphate  AIC Akaike's Information Criterion  ANOVA analysis of variance ANT adenine nucleotide transporter  ATP adenosine triphosphate BSA bovine serum albumin  CO2 carbon dioxide  COX cytochrome c oxidase  Cu copper  DNA  deoxyribonucleic acid  DTNB 5,5'-dithio-bis(2-nitrobenzoic acid) EDTA ethylenediaminetetraacetic acid  EGTA  ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid  ETS electron transport system  FADH2 flavin adenine dinucleotide, reduced  FCCP carbonyl cyanide 4-(trifluoromethyoxy)phenylhydrazone  GPDH glycerol-3-phosphate dehydrogenase  GR glutathione reductase  GSH reduced glutathione GSSG glutathione disulfide; oxidized glutathione  H2O2 hydrogen peroxide  Hb-O2 P50 hemoglobin-O2 binding affinity  HCl hydrochloric acid  HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HIF-1α hypoxia inducible factor 1, alpha  hr hour  HRP horseradish peroxidase IMM inner mitochondrial membrane  IMS intermitochondrial space K2HPO3 potassium phosphate, dibasic  kcat catalytic rate  KCl potassium chloride  KH2PO4 potassium phosphate, monobasic Km Michaelis Menten constant  Km,app apparent Michaelis-Menten constant  kPa kilopascal LOE loss of equilibrium  LOE50 time to loss of equilibrium  MgCl2 magnesium chloride xiv  min  minute  Mitochondrial P50 mitochondrial oxygen binding affinity  mRNA  messenger ribonucleic acid NAD+ nicotinamide adenine dinucleotide NADH nicotinamide adenine dinucleotide, reduced NADP+ nicotinamide adenine dinucleotide phosphate NADPH nicotinamide adenine dinucleotide phosphate, reduced  O2 oxygen  O2·- superoxide radical oC degrees celsius  OH·- hydroxyl radicals  OLS ordinary least squares PaO2 arterial partial pressure of oxygen  Pcrit  critical oxygen tension of metabolic rate  PGLS phylogenetically generalized least squares  PO2 partial pressure of oxygen  RCR respiratory control ratio  RNAi RNA interference  ROS reactive oxygen species  Rot rotenone  s.e.m. standard error of mean  sec second  SUIT substrate utilization inhibitor titration  TBA thiobarbituric acid  TBARS thiobarbituric acid reactive substances TMPD N,N,N',N'-tetramethyl-p-phenylenediamine  TOSC total oxidative scavenging capacity  TPP+ tetraphenylphosphonium ion  UCP uncoupling proteins  Vmax maximum enzyme activity  ΔG change in Gibbs free energy  Δp protonmotive force  ΔΔG change in ΔG       xv  Acknowledgements  Having been at UBC for awhile, it is not an exaggeration to say I grew up a lot here. I have the deepest gratitude for everyone that has made graduate school such a great experience. I will miss being a part of UBC Zoology.  My supervisor Dr. Jeff Richards, thankfully, had a first-come-first-serve policy when I was looking for an undergraduate project. He has been an incredible mentor through the years, has let me run with ideas and projects, trusted me to get things done, and always made time for me. I will miss sitting in front of his computer together to go over data while sharing an orange and catching up over chocolate peanut-butter cups.  I consider myself very fortunate to have overlapped with so many talented individuals in the Richards lab (roughly in the order of appearance in the lab): Milica Mandic, Lindsay Jibb, Ben Speers-Roesch, Patrik Henriksson, Travis Van Leeuwen, Gina Galli, Dave Allen, Matt Regan, Rush Dhillon, Lili Yao, Mark Scott, Derrick Groom, Andrew Thompson, Tammy Rodela, Joshua Emerman, Crisostomo Gomez, Victor Chan, Yuanchang Fang, Derek Somo, and Shuang Liu. Special thanks to: Ben and Milica who were there from the beginning and made me feel included as an undergraduate student. I have fond memories of sitting on the floor of their Wesbrook office watching the hamster ball roll around as I sat in on conversations; Rush who always made time for brainstorming protocols and still giggles at Oroboros emails with me via text; Josh and Andrew who both dropped everything to help me get fish (while I stood on the shore pointing to spots where I wanted fish from), eat bacon, and catch up on South Park and Ancient Aliens in Bamfield.  A big thank you to Tammy Rodela and the Hannah’s! I would have lost a lot of hair from pulling out of frustration and a poor diet during this PhD if not for her. Also, a big shout out to Patrick Tamkee and the rest of the Tamkee’s! Without Pat and his truly endless patience, I would know nothing about keeping fish alive or the precise temperature to cook fish at. Also, thank you to Stella Lee and Georgina Cox, who started in UBC Zoology with me. Even though we were in three different places near the end of my PhD, they were there for me with pom-pom waving emojis, Skype calls, and offered to format my thesis for me. Also, thank you to Charissa Fung for being a xvi  wonderful friend, and to fellow mitochondriacs Dillon Chung and Heather Bryant for letting me bend their ears when I needed help troubleshooting or ran out of substrate!  Dr. Agnes Lacombe encouraged me to pursue an undergraduate research project, then my MSc, and again when I was thinking about a PhD. She always made time for me. The image of her sitting in the back row of my first guest lecture for her BIOL363 class, waving wildly and giving me the thumbs-up makes me happy.   My committee members Drs. Patricia Schulte, Bill Milsom, and Bill Sheel have challenged me to think broadly throughout my degree and have made this thesis much better, thank you; especially to Trish who was also on my MSc committee and has been so generous with her time and advice.  A big thanks to Dr. Mike Murphy who was supportive of the multiple comparative projects that stemmed from my visit to the MRC Mitochondrial Biology Unit. Also, thanks to the rest of the Murphy lab for making me feel welcomed for my 5-month stay even though practically the whole lab was against me in the great cookie vs. biscuit debate.  To the friends outside of academia who understood why I would show up late or reschedule meals entirely because “I have to test one more thing…”; especially to Hedy Li who camped out with me at cafes to work on weekends, and to Jane Lee who has patiently endured just about every rant about graduate school over sushi.  Finally, thank you to my family, especially my parents, aunt Dara and brother Andy. If graduate school was a rollercoaster ride for me, it was an even bigger one for my parents. None of this would have been possible without them.  1  Chapter One: General Introduction 1.1 Oxygen and the evolution of aerobic metabolism The rise of O2 in the atmosphere 2.4-2.1 billion years ago coincided with the rise of eukaryotes and complex life on earth. The success of eukaryotes in the newly oxic environment is thought to be due to the endosymbiotic event whereby an α-protobacterium was engulfed by a methanogen (Lane, 2006; Martin and Müller, 1998). This merger not only provided the methanogen with an intracellular mechanism to metabolize toxic O2, but it eventually provided the methanogen with the protein toolkit to harness energy from carbon substrates using O2 in what is now referred to as the mitochondrion. Compared to O2-independent energy metabolism, aerobic metabolism in eukaryotes allows for a more complete oxidation of carbon substrates, thus transferring a greater proportion of their chemical energy to ATP. This greater capacity to generate ATP supported the greater energy requirements of life on earth as organisms evolved to grow in complexity, in size, and to explore more diverse environmental niches (Lane, 2006). Modern eukaryotes are highly dependent upon O2 to sustain energy metabolism and in most eukaryotes even short periods of O2 lack can threaten the maintenance of cellular energy balance (Boutilier and St-Pierre, 2000). Indeed, over 90% of O2 delivered to tissues was estimated to be used by mitochondria for aerobic respiration (Rolfe and Brown, 1997). Under O2 limiting conditions, cellular ATP levels can fall when ATP supply no longer matches ATP demand, resulting in the failure of ion-motive ATPases (among other ATPases), membrane potential depolarization, calcium ion influx, and the initiation of necrotic cell death (Boutilier and St-Pierre, 2000). As such, even short disruptions of O2 supply to tissues can have dire consequences. In mammals, which are generally sensitive to low O2 exposure (hypoxia), seconds of disrupted blood supply to critical tissues such as the brain can cause ATP levels to fall over the course of 5-10sec resulting in irreparable cellular damage (Hansen, 1985; Lutz et al., 2003).  Though modern eukaryotes have evolved a heavy dependence on O2 for energy metabolism, the threat of O2 toxicity still looms. Oxygen can react with electrons to form reactive oxygen species (ROS) and this occurs primarily at sites within the mitochondrial electron transport system (ETS). Low levels of ROS are constantly produced by mitochondria through electron slip, which are believed to play an important role in cell signaling (D’Autréaux and Toledano, 2007), and are scavenged by mitochondrial and cellular antioxidant mechanisms (Turrens 2003). When ROS are 2  produced at high levels, which can occur under highly reducing redox conditions, they can overwhelm the cell’s antioxidant capacity and cause cellular damage by reacting with lipids, proteins, and DNA (Cadet, 2003; Gutteridge, 1995; Reznick and Packer, 1994). Therefore, O2 use by the cell is a proverbial “double-edged sword”, where both the good and bad of O2 use must be kept in careful balance. A steady substrate and O2 supply to mitochondrial complexes must be maintained to sustain ATP production, meanwhile electron slip must be avoided in order to prevent increases in ROS generation. Indeed, mitochondrial aerobic and ROS metabolism have been the focus of numerous studies because of the pathological conditions that can arise from or cause variable O2 supply to tissues (e.g. heart attack, stroke, sleep apnea; Chouchani et al., 2014; Piantadosi and Zhang, 1996; Troncoso Brindeiro et al., 2007) The vast majority of work examining the interaction between aerobic metabolism and ROS generation has been conducted in mammalian models that have evolved to function under fairly constant O2 conditions. Animals in nature, however, can be found in environments that vary immensely in O2 from hypoxia or anoxia, to hyperoxia, over various time periods. For example, animals are able to survive months in cold anoxia during the winter months (Nilsson, 2001), and animals that inhabit intertidal environments can experience wide variability in O2 due to tidal cycle (Richards, 2011). Compared with studies focused on hypoxia-sensitive mammalian systems, few studies have taken advantage of the organisms that inhabit different O2 environments to examine putative adaptations of mitochondrial O2 use for aerobic and ROS metabolism. There are a few exceptions with a study on intertidal elasmobranchs (Hickey et al. 2012) and laboratory bred deer mice of high altitude ancestry (Peromyscus maniculatus; Mahalingham et al. 2017). A comparative approach investigating mitochondria from multiple species that naturally experience O2 variability would help to illuminate the underlying mechanisms essential to inhabiting these challenging environments. Further, the impacts of natural O2 oscillations on aerobic and ROS metabolism are unclear, and hypoxia and hyperoxia both can pose very different challenges on mitochondrial function. Thus, the goal of this thesis is to elucidate the balance and trade-offs between the two aspects of mitochondrial O2 use, aerobic respiration and ROS metabolism, in animals that inhabit different O2 environments and vary in hypoxia tolerance. In this introduction, I will first briefly describe how mitochondria use O2 and what affects aerobic and ROS metabolism, followed by what is currently known from well-studied mammalian models and the lesser-studied comparative models of adaptation to hypoxia.   3  1.2 The Mitochondrion Mitochondria are often described as the “powerhouses of the cell” because of their critically important role in energy metabolism. Briefly, the mitochondrial electron transport proteins capture free energy from substrate oxidation in the form of a proton gradient that is used to facilitate the phosphorylation of ADP to ATP in a process termed oxidative phosphorylation (Mitchell 1961, 1966). Processes involved in substrate oxidation (namely glycolysis, the Kreb’s cycle and ß-oxidation) generate reducing equivalents (NADH, succinate, FADH2) which donate electrons to the protein components of the mitochondrial ETS (complexes I to IV; Fig.1.1). Complexes I (electrons from NADH), II (electrons from succinate), glycerol-3-phosphate dehydrogenase (electrons from glycerophosphate oxidation), electron-transferring flavin protein (electrons from fatty acid beta-oxidation) all donate electrons to the Q (ubiquinone) cycle at complex III. Another electron carrier, cytochrome c, receives electrons at complex III. Reduced cytochrome c then docks and donates electrons to complex IV (cytochrome c oxidase or COX). The electrons transferred to COX are eventually used to reduce O2 to form water. As electrons move along the ETS, the resulting difference in free energy from the redox reactions is used by complexes I, III and IV to remove and/or transport protons from the mitochondrial matrix into the intermembrane space (IMS), to form a proton electroconcentration gradient or protonmotive force (Δp) across the inner mitochondrial membrane (IMM). In the final step of oxidative phosphorylation, complex V (F1FO-ATP synthase) couples the movement of protons back into the matrix with ADP phosphorylation to ATP (Fig.1.1). ATP is then transported out of the mitochondria into the cytoplasm via the adenine nucleotide transporter (ANT) to be used by the cell. 1.3 Factors affecting mitochondrial function The proton gradient generated by the mitochondrial ETS complexes drives complex V activity to phosphorylate ADP into ATP. There are several factors that can influence ETS function and ultimately the rate and efficiency of phosphorylation of mitochondria, which include proton leak, electron leak, and O2 availability.  1.3.1 Proton leak  Changes to proton movement from the matrix to the IMS could affect mitochondrial phosphorylation efficiency, which is empirically assessed by the ADP/O or P/O. Proton leak from the IMS into the matrix can occur via ANT (which exchanges ATP out of the matrix for ADP and a 4  proton into the matrix), uncoupling proteins (UCPs; Jastroch et al., 2011), and due to general mitochondrial membrane leakiness (Brand et al., 1994). This is commonly referred to as mitochondrial proton leak and results in ETS electron flux, proton pumping, and O2 use without ATP generation. This futile cycling of protons has a role in heat generation in endotherms and can account for up to 15% of basal metabolic rate (Rolfe et al., 1999). It is unlikely, however, that proton leak functions solely for thermogenesis as it also occurs in ectotherms and in some cases can account for 20-30% of routine metabolic rate in lizards and frogs hepatocytes (Brand et al. 2000). As this proton leak incurs higher metabolic costs, it might be assumed that a lower proton leak would be beneficial and result in a higher phosphorylation efficiency and in a high protonmotive force. This would, however, also require maintaining the ETS complexes in a more reduced redox state which could stimulate ROS production at sites on the ETS complexes (discussed below). Thus, it has been suggested that the dissipation of the proton gradient would reduce mitochondrial ROS generation and relieve oxidative stress, which has been coined the ‘uncouple to survive’ hypothesis (Brand, 2000; Brand and Esteves, 2005).  1.3.2 Electron leak  Some of the electrons that enter the ETS leak out and react with O2 to form ROS at a rate that accounts for 1-2% of total O2 used by mitochondria (Boveris and Chance, 1973). Mitochondrial ROS have been observed to be generated from eleven sites along the ETS where one leaked electron reduces O2 to form superoxide radicals, or two leaked electrons form the more stable H2O2 (Murphy, 2009; Quinlan et al., 2013; Brand 2016). Electrons that become substrates for ROS generation do not participate in the ETS redox reactions, which reduces the protons pumped to contribute to the protonmotive force. As a result, changes in ETS flux that increase ROS generation would impact phosphorylation efficiency. Additionally, changes in ETS flux that occur due to dysfunction or inhibition of ETS complexes that lead to increases in ROS generation can overwhelm cellular antioxidant mechanisms and lead to ROS accumulation (Ott et al., 2007; Sies, 1997). ROS accumulation can cause oxidative damage to the ETS and the phospholipid membrane, negatively impacting ETS flux (Paradies et al., 2004; Petrosillo et al., 2003).  ROS generation from the various ETS sites differ depending on the protonmotive force, redox environment, and energy status of the mitochondrion (Aon et al., 2010; Barja, 2002; Quinlan et al., 2013), which has made it challenging to assess ROS emission under physiologically relevant 5  conditions. ROS scavenging processes are also highly dependent upon the cellular and mitochondrial redox environment (which is the sum of various redox couples in the different cellular compartments)1. In order to better understand and predict the relationship between ROS generation/scavenging and redox environment, Aon et al. (2012) developed a hypothesis that ROS accumulates under both highly reduced and oxidized redox states. Under normal physiological conditions, ROS is generated at low rates, plays a role in cell signaling (D’Autréaux and Toledano, 2007) and is scavenged by cellular antioxidant defences (via a combination of small molecular weight antioxidants and antioxidant enzymes; Point 1 in Fig.1.2; Pamplona and Costantini, 2011). Under conditions of low ETS flux and high protonmotive force (point 2 in Fig1.2), such as in hypoxia, electrons can build up at ETS complex redox centres which result in more reduced ETS complexes and lead to a high rate of ROS generation. This high rate of ROS generation can overwhelm the scavenging capacity of mitochondria and increase net ROS emission. Under conditions of high ETS flux and lower protonmotive force (point 3 in Fig.1.2), lower electron availability can cause the mitochondrial environment and ETS to become relatively more oxidized which can lower overall redox buffering capacity (e.g. by lowering NADPH/NADP+ redox couple and reducing recycling of other antioxidant redox couples) and compromise cellular scavenging capacity, also culminating in elevated ROS emission (Aon et al., 2010; Munro and Treberg, 2017).   While better coupled mitochondria with low proton leak results in efficient ATP production, this also leads to an increased rate of ROS generation by promoting a more reduced state in the ETS. As such, there is a clear trade-off with aerobic metabolism whereby improved phosphorylation efficiency can result in a mitochondrial environment conducive to ROS generation. Thus, ETS function must be carefully regulated to maintain ATP production to support cellular activities under                                                  1 The main redox couples within the cell are NADH/NAD+, NADPH/NADP+, and GSH/GSSG. NADH/NAD+ is an important redox couple functioning in electron transport, for example, donating to ETS complex I. The NADPH/NADP+ pair is involved in maintaining redox balance, for example, NADPH is important for the regulation of the GSH/GSSG redox couple as it donates electrons to the reaction of glutathione reductase to recycle GSSG into GSH. GSH/GSSG is a dominant redox couple as well and is associated with maintaining important redox-linked reactions in different cellular compartments, such as maintaining critical protein sulphydryls in the nucleus for DNA repair and maintaining an oxidizing environment in the endoplasmic reticulum for protein disulfide bond formation (reviewed in Mari et al., 2013).   6  a range of physiological and environmental conditions while also preventing excessive ROS generation that would lead to oxidative stress. 1.3.3 O2 delivery to mitochondria  The supply of terminal electron acceptor, O2, to the ETS is crucial in maintaining ETS flux. As the terminal electron acceptor, O2 limitation to COX could lead to slowing of overall ETS flux, lower the maintenance of the proton gradient and ultimately affect complex V activity to phosphorylate ADP. The process of transporting O2 from the environment to mitochondria involves both diffusive and convective movement along the O2 transport cascade. Central to the efficient movement of O2 across the steps of the O2 transport cascade is the maintenance of partial pressure gradients so that O2 readily diffuses from higher to lower PO2. When O2 comes into contact with the respiratory surface via ventilation, it diffuses across the respiratory surface into blood (typical mammalian arterial PO2 is 10 to 13kPa; Malatesha et al., 2007), where convective movement of blood with O2-bound hemoglobin is circulated to tissues. Oxygen then diffuses across the capillary walls into the interstitial fluid (mammalian PO2 estimates of 1.3 to 2.3kPa; reviewed in Mik et al. 2009) and across the cell membrane into the cytoplasm (PO2 of 0.4 to 0.9kPa; reviewed in Mik et al. 2009), and finally diffusing to mitochondria (PO2 of 0.3 to 0.7kPa; Jones 1986, Whalen and Nair 1967) where binding to COX facilitates ATP generation.  It is estimated that 90% of O2 is used by mitochondria and COX, while the remainder is attributed to O2 use at peroxisomes and endoplasmic reticulum that can contribute to ROS generation (Brown and Borutaite, 2012). Recent estimates of mitochondrial PO2 in rat hearts reported a much higher mean mitochondrial PO2 4.7kPa (~46µM O2)2 than previously reported values, although there appears to be great variation in mitochondrial PO2 with values as low as 0 to 1.3kPa (0 to 14µM O2) reported in the same tissue (Mik et al. 2009). This variation is likely dependent upon the location of mitochondria within the cell. In comparison, mitochondrial P50 (i.e. O2 concentration at half-maximal O2 consumption rate) is many fold lower than most of the estimated mitochondrial PO2 values, and range between 0.05-0.1kPa (0.1-0.2µM) at rest and 0.15-0.24kPa (0.3-0.5µM) in phosphorylating mitochondria (Gnaiger et al. 1998) and the Michaelis-Menten constant (Km; an indicator of substrate affinity) of mammalian COX for O2 is estimated to be between 0.5-1µM in bovine COX (reviewed in Brunori et al. 1987 and Nicholls and                                                  2 At 37oC, barometric pressure of 100.7kPa, O2 concentration at air saturation is 208.87µM.  7  Chance 1974). Thus, normal intracellular physiological PO2 are many fold above what is required for ETS to function maximally, suggesting that mitochondria and COX do not lack O2 under normal physiological conditions. However, reductions in environmental PO2 or disruption of blood supply to tissues can compress diffusion gradients such that O2 does become limiting to mitochondria which causes ETS flux to slow as it is without a terminal electron acceptor. This slowing of electron flux subsequently slows the redox-linked proton pumping reactions at complexes I, III, and IV, which lowers the proton gradient and ultimately reduces the drive for complex V to phosphorylate ADP. Thus, complex feedback mechanisms involving O2 levels, ATP turnover, and substrate oxidation ensure that cellular energy demands are matched with aerobic ATP supply (Balaban 1990; Brown 1992), such that increases in cellular activity that reduce intracellular O2 levels will stimulate tissue blood flow to maintain PO2 levels (Hogan 2001; LaManna et al. 2004). This feedback regulation results in tight coupling that attempts to maintain tissue PO2 levels in order to sustain mitochondrial ETS function. However, if environmental hypoxia worsens, O2 will eventually become limiting for mitochondria and threaten energy and redox balance.    1.4 Mammalian mitochondrial responses to O2 variability  Much of what is currently known of how mitochondria respond to O2 variability is from work conducted on hypoxia-sensitive mammals to investigate the implications of various O2 related pathologies. The pathologies and tissue damage associated with ischemia-reperfusion injury (e.g. heart attack and stroke) are thought to be due to the ROS surge that occurs when O2 supply to the tissue recovers after hypoxia (Murphy and Steenbergen, 2008). As a result, tremendous research efforts have focused on how to mitigate ROS generation and also to alleviate downstream effects to improve prognosis (e.g. Chouchani et al., 2014).  1.4.1 Mammalian mitochondrial responses to acute changes in O2  Hypoxia  The main response of mammalian mitochondria to acute hypoxia exposure is a reduction in respiration rate via a decrease in ETS capacity (Kuroda et al., 1996; Magalhães et al., 2005; Schumacker et al., 1993; Sims and Pulsinelli, 1987). The reduction in mitochondrial aerobic metabolism then necessitates an increase in O2-independent glycolytic flux in attempts to maintain cell function, which has a lower ATP yield. The hypoxia-induced reduction in state III ADP-stimulated respiration rate appears to be reversible and mediated through direct modifications of the 8  ETS complexes (Schumacker et al. 1993) causing them to adopt a more reduced redox status (Chandel et al., 1996; Vollmar et al., 1997). This lowered ETS capacity has been shown to be, in part, mediated through COX. Oxygen can allosterically modify COX in bovine mitochondria and lower its activity under hypoxic conditions (Chandel et al., 1996). Other mechanisms of modifying COX activity have been observed in hypoxia (6-12hrs at 0.13kPa), such as a reduction in COXI, IV, and Vb subunit mRNA content that is associated with an overall loss of COX enzyme in mice macrophages, and decreases in COX activity with a slight increase in catalytic turnover observed in rat PC12 cells (Vijayasarathy et al., 2003). The other ETS complexes also play a role in coordinating the mitochondrial response to hypoxia. Complex I-linked respiration was inhibited to a greater extent (32-46%) than complex II-linked respiration (25%) in rat parietal cortex mitochondria exposed to 7% O2 ischemic conditions (Gilland et al., 1998). These rapid modifications to ETS complexes can occur via post-translational modifications. For instance, S-nitrosation (R-SNO) of mitochondrial protein thiols at sites on complexes II, III, IV, and electron-transferring flavoproteins (ETF), have been observed to exert protective effects on mouse hearts during ischemia (Chouchani et al., 2017). Also, multiple phosphorylation sites have been identified on COX (Helling et al., 2012) and are believed to play a role in modifying COX function in response to hypoxia (Prabu et al. 2005). Thus, mammalian mitochondria inhibit ETS flux in response to low O2 supply.  Normoxic-recovery from hypoxia   As cellular PO2 ranges between 0.4 to 0.9kPa (Mik et al. 2009), studies that utilize normoxic-recovery following hypoxia exposure are in fact exposing mitochondria to non-physiological oscillation in O2 levels. Indeed, the vast majority of mitochondrial studies, including many performed in this thesis, undertake mitochondrial analysis under air-saturated conditions which does not replicate in vivo conditions. From studies of this nature, however, it is possible to understand the effects of large changes in O2 on overall mitochondrial function. For instance, ischemic-tissue that is reoxygenated experience a large increase in O2 levels. Upon normoxia-recovery, hypoxia-induced reduction in ETS flux needs to be reversed in order to restore aerobic ATP production and full cellular function. However, a quick return to normoxic and relatively high O2 levels can cause a surge in ROS generation. High electron flux along with high phosphorylation rate lowers the protonmotive force leading to a relatively oxidized redox environment, which with high O2 availability favors an increase in ROS generation from ETS sites (Brand 2000). It appears that an 9  inhibition of ETS flux upon hypoxia-reoxygenation may be part of a slow controlled recovery to reduce ROS generation. In fact, pharmacological inhibitions of both complex I (via S-nitrosation modification; Chouchani et al., 2013) and II (via malonate inhibition; Chouchani et al., 2014) have been shown to have cardioprotective effects and alleviated ischemic-reperfusion injury in murine models. Also, there may be temporal differences of in vivo regulation of the different ETS complexes. In gerbils recovering from cerebral ischemia, complex II-linked state III respiration rate recovered more quickly (5min) compared to complex I-linked state III respiration rate (30min; Almeida et al., 1995). It is also possible that the reduction in ETS complex activities is caused by a secondary effect of oxidative damage to the lipid bilayer in mammals. After ischemia exposure, an increase in H2O2 generation was accompanied by a loss of cardiolipin (Chen and Lesnefsky, 2006; Paradies et al., 2004), which is a mitochondrial phospholipid important for ETS complex assembly and for stabilizing respiratory supercomplexes (Mileykovskaya and Dowhan, 2014; Pfeiffer et al., 2003).  1.4.2 Mammalian mitochondrial responses to chronic changes in O2  Hypoxia Many mammals have the ability to acclimate to longer term exposure to moderate hypoxia. Acclimation or acclimatization to prolonged exposure to new O2 environments are mediated by large-scale changes in gene and protein expression, epigenetic factors which have roles in stabilizing hypoxia inducible factor (HIF) and activating HIF directly (Watson et al., 2010), and via post-translational modifications of proteins (Kumar and Klein, 2004), primarily to reduce O2-dependent pathways and to increase O2-independent energy production. The transcription factor HIF-1α plays a key role in coordinating cellular responses to hypoxia. HIF-1α is targeted for degradation in normoxia, but is stabilized when O2 levels become limiting and acts on nuclear targets. HIF-1α has also been shown to be stabilized by H2O2 released by complex III, pointing to a role of H2O2 in hypoxic signaling (Chandel et al., 2000; Guzy et al., 2005). Stabilized HIF-1α then targets a number of genes that generally coordinate an increase in O2-independent ATP producing pathways and reduce substrate supply to the mitochondria, including upregulation of glycolytic genes (e.g. glucose transporter GLUT1), lactate dehydrogenase-A to increase glycolytic flux (Kim et al., 2006; Semenza, 2007), and pyruvate dehydrogenase kinase that phosphorylates pyruvate dehydrogenase to inhibit pyruvate entry into the Kreb’s cycle and mitochondrial respiration (Papandreou et al., 2006). HIF-1α 10  also targets COX nuclear-encoded subunit 4 and initiates the switching of COX4-1 to COX4-2 during hypoxia. While COX protein with subunit 4-1 under normoxic conditions is inhibited by high ATP levels which would lower ETS flux, COX protein with 4-2 is not sensitive to high ATP levels and allows ETS flux to be maintained (Fukuda et al., 2007; Horvat et al., 2006). This COX4 subunit switch is thought to optimize efficiency of electron flux and minimizes ROS generation during hypoxia. Acclimation to moderate hypoxia has also been shown to result in morphological changes such as increased capillary density, increased mitochondrial volume density, and altered tissue mitochondrial distribution (in liver and heart mitochondria; Costa et al., 1997; Hoppeler et al., 2008; Howald et al., 2008), which would presumably reduce O2 diffusion distances and improve O2 delivery to mitochondria.  Hyperoxia  When PO2 levels increase beyond normoxic levels, the increase in substrate (O2) causes the rate of mitochondrial ROS generation to also increase (Jamieson et al., 1986). Although there are no natural exposures to hyperoxia in mammals, O2 therapy is often used in clinical settings when the benefits of treatment outweigh the damaging effects of hyperoxic exposure (Thomson et al., 2002). The typical mammalian mitochondrial response to high O2 tensions is an overall reduction in ETS flux that reduces ROS generation and accumulation. Chinese hamster ovarian cells exposed to hyperoxia (98% O2, 2% CO2) showed an 80% decrease in respiration rate within 3 days of exposure which was related to the selective inhibition of complexes I and II, although the cell line eventually died due to ATP depletion (Schoonen et al., 1990). ROS generated from hyperoxia exposure also caused cardiolipin loss (mice exposed to 100% O2 for 72hrs; Tyurina et al., 2010), which may also negatively impact ETS complex activities similar to the response to hypoxia-recovery as discussed above. An increase in COX activity, however, may be important under high O2 levels. HeLa cells maintained at 80% O2 increased COX capacity by two-fold which was associated with two-fold lower ROS levels (Campian et al., 2007) indicating that COX may have a role in helping to mediate ETS flux to mitigate oxidative damage experienced in hyperoxia.  1.5 O2 variability in the natural environment   Most mammals rarely experience hypoxic environments and are generally hypoxia sensitive as they have not evolved mechanisms to survive O2 limiting conditions. In fact, many organisms, especially those in aquatic environments, are exposed to natural patterns of environmental O2 variability that 11  can differ in spatial and temporal patterns and vary from hypoxia to hyperoxia (Diaz and Breitburg, 2009).   Timescale of O2 changes can vary from minutes to hours, to seasonal changes. Environments such as the marine intertidal zone, estuaries, swamps, and marshes (Diaz and Breitburg, 2009; Nikinmaa, 2002; Richards 2011), show fluctuations in the span of hours due to the tidal cycle. These quick changes are due to periods of high mixing at high tide that would bring about well-mixed oxygenated water, and periods of low tide that may create isolated areas with little to no mixing. In tropical freshwater systems, the rainy season can bring about increased water flow and oxygenation compared to during the dry season. There are also environments that are chronically hypoxic. Animals found living at high altitudes, in poorly ventilated burrows, and in the oceans’ O2 minimum zone are constantly under O2 limiting conditions (Storz et al., 2007; Larson et al., 2004; Seibel, 2011; Diaz and Breitburg, 2009).  Spatial variability of O2 can be driven by biotic and abiotic factors. Stagnant waters with high amount of organic matter and high biomass can become severely hypoxic and even anoxic due to high O2 consumption. Also, eutrophic environments that are abundant in green plants are prone to periods of hyperoxia as high photosynthetic rates generate high O2 levels (e.g. up to 300% air saturation was measured in high tidepool in the intertidal zone; Richards, 2011). Hypoxia or anoxia can occur easily in aquatic environments as O2 solubility is lower in water and diffusion rate is also slower in water than in air (Graham 1990). Both O2 solubility and diffusion rate vary with abiotic factors such as temperature and salinity (Diaz and Breitburg, 2009). Additional variation in O2 can be due to differences in mixing that could depend on density, current, wind, and also depth of the different layers (Diaz and Breitburg 2009).  As O2 variability occurs frequently in aquatic environments, many organisms have been shown to display diverse responses and mechanisms, behavioural, morphological, physiological, and biochemical, that enable them to inhabit and thrive in these O2 environments. Thus, aquatic organisms show wide variation in hypoxia tolerance. 1.6 Defining hypoxia tolerance  In defining hypoxia tolerance, we should first consider what happens when environmental PO2 level is reduced. As PO2 is reduced from air saturation, O2 content in the blood is maintained with high 12  blood hemoglobin concentration and increased O2 binding in circulation. Many aquatic animals will exhibit behaviors such as avoidance and aquatic surface respiration to seek oxygenated water to respire, and also reduced spontaneous swimming to lower metabolic demand (although the thresholds for these behaviors are species and even individual dependent; reviewed in Chapman and Mckenzie, 2009). The hypoxic ventilatory response is observed in many fish species where there are changes in ventilation (volume, frequency, amplitude, and/or stroke volume depending on species under investigation), ultimately to increase oxygenated water flowing pass gill lamella to increase O2 extraction (reviewed in Perry et al. 2009). Eventually when PaO2 has been impacted, aerobic scope is lowered and with it a number of physiological processes (e.g. growth, reproduction) are suppressed but standard metabolic rate is maintained. At the point where aerobic scope is reduced to zero, O2 supply is no longer able to support standard metabolic rate and animals switch from being oxy-regulating to oxy-conforming at the critical O2 tension of O2 consumption rate (Pcrit). Below Pcrit, animals increase their reliance on anaerobic respiration in attempts to maintain cellular energy balance and extend survival time in hypoxia (Farrell and Richards, 2009).  Species with a lower Pcrit are presumed to be better able to extract O2 from their environment so that standard metabolic rate is maintained to a lower environmental PO2 and the animal is thus more hypoxia tolerant. Thus, Pcrit has been frequently used in literature as an indicator of hypoxia tolerance (Deustch et al. 2015; Regan and Richards, 2017; Speers-Roesch et al., 2013). Hypoxia tolerant organisms typically have larger respiratory surface area (Mandic et al., 2009; Nilsson, 2007), higher hemoglobin-O2 binding affinity (Jensen and Weber, 1982), and lower tissue O2 demands, all of which facilitate the maintenance of aerobic function even during hypoxia and would lower Pcrit. There are a couple other metrics commonly used to assess hypoxia tolerance. Animals that have a longer time to loss of equilibrium (LOE50) at a hypoxic PO2 are deemed more hypoxia tolerant, as they are able to survive long periods of O2 limitation (Chapman et al., 1995; Mandic et al., 2013). The 50% effective/lethal concentration (EC50/LC50), which is the O2 concentration that causes 50% reduction in hypoxic survival has been used to determine hypoxia tolerance as well (Andrade et al., 2017; Irving et al., 2004; Wu et al., 2002). These two metrics assess the species’ overall ability to maintain activity using both aerobic and anaerobic processes. A strong relationship between LOE50 and Pcrit has been shown in species of intertidal sculpin fishes, where species with a longer time to LOE50 also exhibit a lower Pcrit (Mandic et al., 2013), supporting that the maintenance of aerobic respiration is an important strategy to survive hypoxia exposure.  13  Further, the ability of organisms to recover from hypoxia should also be considered as a part of hypoxia tolerance. The reintroduction of O2 signifies that the processes which have redirected substrates away from mitochondria towards O2-independent energy pathways can now be reversed. As I mentioned above, the reoxygenation of mammalian ischemic tissue causes a burst of ROS generation that can lead to extensive tissue damage. The extent of recovery of state III ADP-stimulated mitochondrial respiration after ischemia, the amount of ROS generated, the oxidative damage sustained upon normoxic recovery, and the ability to repair oxidative damage are indicators that have been used to assess the ability to recovery from O2 limitation (Almeida et al., 1995; Chouchani et al., 2014; Shiva et al., 2007). For example, mouse heart mitochondria exposed to 30min in vitro anoxia typically recovers only 50% of state III respiration rate in normoxia, which is thought to be due to ROS damage to the ETS (Shiva et al., 2007). Hypoxia tolerant animals would presumably have strategies to deal with the challenges of exiting the hypoxia bout as well (Bickler and Buck, 2007), but it is unclear what adaptive traits at the mitochondrial level are associated with coordinating recovery from hypoxia. 1.7 Adaptation and plasticity of mitochondria to O2 variability  Hypoxia tolerance in animals is often associated with modifications to the O2 transport cascade that improve O2 extraction from the environment and increase supply to the mitochondria. There are several detailed studies investigating mitochondrial characteristics from laboratory-selected lines of flies maintained at 4% O2 (herein referred to as hypoxic flies; Ali et al., 2012; Yin et al., 2013), laboratory-bred lines of deer mice with highland ancestry (herein referred to as highland deer mice; Mahalingam et al., 2017), and elasmobranchs that inhabit the marine intertidal environment (Hickey et al., 2012). However, the results from these studies do not suggest a unified evolutionary strategy of mitochondria to O2 variability. In this section of the Introduction, I will summarize what is currently known of three aspects of mitochondrial function as they relate to aerobic and ROS metabolism from organisms that are characterized as hypoxia tolerant or show variation in hypoxia tolerance.  1.7.1 O2 binding at mitochondria and COX  Although few studies have attempted to directly determine whether there are differences in mitochondrial and COX O2 binding in animals that vary in hypoxia tolerance, the few that do exist suggest that adaptation and acclimation to hypoxia exposure can result in modifications to 14  mitochondrial P50 and COX Km in order to improve mitochondrial function at a lower PO2. For example, highland deer mice have lower mitochondrial P50 under resting respiration state conditions compared with a lowland population (Mahalingham et al., 2017). This difference of mitochondrial P50 between populations were not modified by hypoxia acclimation. Similarly, killifish (Fundulus heteroclitus) showed no changes in mitochondrial P50 when acclimated to hypoxia  (Du et al., 2016). In overwintering frogs (Rana temporaria), however, hypoxia acclimation after 1-month reduced mitochondrial P50 compared to normoxic and 4-month acclimated frogs. These studies suggest species-specific responses in plasticity of mitochondrial O2 binding to O2 limitation.  These shifts in mitochondrial P50 are likely due to differences in COX activity. Higher COX enzyme content within mitochondria is thought to increase mitochondrial O2 binding affinity (Gnaiger et al., 1998). Hypoxic flies and highland deer mice show higher COX activity or respiration rates compared to normoxic flies and lowland deer mice population (Ali et al., 2012; Mahalingam et al., 2017), which would be consistent with increased mitochondrial O2 binding (although this was only empirically measured in deer mice and not in flies). In elasmobranchs, however, the opposite trend was observed where less hypoxia tolerant shovelnose rays (Aptychotrema rostrata) had higher COX respiration rate compared to more hypoxia tolerant epaulette sharks (Hemiscyllum ocellatum; Hickey et al. 2012), suggesting higher COX enzyme content in the less tolerant elasmobranch. Although there have not been direct measurements of COX O2 binding across species or populations, there has been strong evidence for adaptive changes in COX substrate affinity in hypoxia tolerant organisms. In both high-altitude bar-headed geese (Anser indicus) and Tibetan locusts (Locusta migratoria), COX affinity for its electron donor, reduced cytochrome c, was observed to be higher than in their low-altitude counterparts (Scott et al., 2011; Zhang et al., 2013). The differences in cytochrome c binding, at least in bar-headed geese, appeared to be due to a single amino acid residue difference on the COX3 subunit (Scott et al., 2011). While these studies allude to differences in COX function that may be associated with hypoxia tolerance, in fact, COX modeling studies did not show a clear mechanistic link between cytochrome c and O2 binding (Krab et al., 2011), so variation in cytochrome c binding may not reflect differences in O2 binding. Thus, whether there are functional differences in O2 binding of COX in animals varying in hypoxia tolerance needs to be empirically determined.  15  1.7.2 Mitochondrial ETS function Animals that inhabit O2 variable environments appear to differ in mitochondrial ETS function. In hypoxic flies, there was lower state III ADP-stimulated respiration rate compared to normoxic flies, which was associated with 30% lower complex II activity, and 20% higher complex III and IV activities (Ali et al., 2012). This result suggests a shift to an increased dependence on complex I, which is a proton pump contributing to the proton gradient, rather than on complex II, which does not contribute to the proton gradient. An increased complex I dependency would presumably increase ADP/O leading to more efficient use of O2. There was also an increase in leak respiration in hypoxic flies which indicates higher uncoupling and possibly associated with lower ROS generation (see discussion in 1.6.3). These observations, however, contrast what was observed in a comparison of elasmobranchs that inhabit different O2 environments. There were no interspecies differences in state III respiration rates (both complex I and/or complex II fuelled) of permeabilized ventricular fibres, but the more hypoxia tolerant epaulette shark had lower leak respiration rates, resulting in more coupled mitochondria (higher respiratory control ratio; RCR) compared to the less hypoxia tolerant shovelnose ray (Hickey et al. 2012). Comparisons between highland and lowland populations of deer mice, however, showed no differences in leak respiration rate, and state III respiration rate was higher in the highland population compared to the lowland population indicating higher respiration capacity (Mahalingham et al. 2017). These results reveal divergent characteristics of mitochondrial ETS from different animal models that inhabit hypoxic environments.  Hypoxia tolerant species generally respond to hypoxia exposure by lowering of ETS activity and increasing their dependence on anaerobic respiration. Overwintering frogs submerged for 4 months in hypoxia lowered state III ADP-stimulated and state IV resting respiration rates by 60% in skeletal muscle which was mediated by a lowering of the protonmotive force (St-Pierre et al. 2000a, b). A reduction in mitochondrial capacity was accompanied by an increase in glycolytic capacity as evidenced by an increase in transcription of lactate dehydrogenase and phosphofructokinase in liver and muscle of longjaw mudsucker (Gillichthys mirabilis; Gracey et al., 2001) and increased in activities of a number of glycolytic enzymes in tench (Tinca tinca; Johnston and Bernard, 1982). However, hypoxia-induced plasticity of mitochondria ETS appears to be species-specific as some species are also able to maintain ETS function even in the face of severe hypoxia. In killifish Fundulus heteroclitus, hypoxia, whether intermittent or chronic exposures for 28-33 days at 5kPa, did not affect liver mitochondrial 16  respiratory capacities or mitochondrial O2 kinetics indicating that ETS flux was maintained (Du et al., 2016). Similarly, ventricular fibres from epaulette sharks showed no significant changes in mitochondrial function after whole animal hypoxia exposure, whereas fibres from the shovelnose ray exposed to hypoxia had 50% lower complex I and II fluxes rates and 33% lower COX flux rates compared with fibres from normoxia-acclimated rays. These results suggest that while there are animals that reduce ETS function in response to O2 limitation, certain species of hypoxia tolerant fish were able to maintain mitochondrial function in response to acute hypoxia. 1.7.3. ROS metabolism  While mammals sustain extensive ROS damage after ischemia-reperfusion, it has been proposed that hypoxia tolerant animals would be able minimize the oxidative damage by either lowering ROS generation or increasing in ROS scavenging during normoxia recovery from hypoxia. Despite this intuitive expectation, neither whole animal studies looking at responses to O2 challenges or mitochondrial studies in hypoxia tolerant or adapted animals have yielded consistent trends. The hypoxia tolerant epaulette shark generated 60-70% less ROS/O2 at the mitochondria when examined under both working or resting conditions than the less hypoxia tolerant shovelnose ray (Hickey et al., 2012). Mitochondria from hypoxic flies showed reduced superoxide generation (with complex I substrates only), likely due to the increased uncoupling (measured higher leak respiration) that would decrease the protonmotive force and reduce the drive for ROS generation (Ali et al. 2012). In contrast, mitochondria from highland and lowland deer mice populations did not show differences in ROS emission under both state III and IV conditions (Mahalingham et al. 2017). Further, no study has shown consistent ROS responses at both the mitochondria and the tissue level during whole animal hypoxia exposure. The difference between diverse species is possibly due to variation in antioxidant defenses (Leveelahti et al., 2014). For instance, elasmobranchs are well known to have lower antioxidant activities than teleosts (Filho and Boveris, 1993; Gorbi et al., 2004) and thus could result in species differences in ROS accumulation kinetics.  Our current understanding of mitochondrial function in animals tolerant to natural variations in O2 is scattered. First of all, previous studies have focused primarily on hypoxia exposure and characterizing hypoxia tolerance, and far less is known of hyperoxia tolerance. While it has been suggested that mitochondria and COX from hypoxia tolerant animals may be better able to function under low PO2, this has not been explicitly measured. Further, while there is strong evidence that 17  mitochondria have a role in determining whole animal hypoxia tolerance, it has only been addressed in a few species that live in hypoxia environments (where hypoxia tolerance has not been determined), that are distantly related, exposed to different O2 levels, under substrate and air-saturating, supraphysiological O2 conditions, and also studied at various levels of biological organization (whole animal tissue responses vs permeabilized tissue vs isolated mitochondria). Therefore, characterizing mitochondrial traits of multiple closely-related species with known whole animal hypoxia tolerance would be insightful.   1.8 Thesis objectives and chapter hypotheses The overall goal of this thesis is to determine the mitochondrial traits that have shaped whole animal tolerance to environmental O2 variability. In particular, I will address the three aspects of mitochondrial function that I have discussed above: O2 use at mitochondria and COX (Chapter 2), and mitochondrial ETS flux and ROS metabolism (Chapter 3 & 4) in a group of fish species commonly called sculpins that live along the highly variable marine near-shore environment. 1.8.1 Using intertidal sculpins (Cottidae, Actinopterygii) as model  The near-shore marine intertidal environment is heavily influenced by the daily tidal cycle, resulting in an environmental gradient of variability in a number of abiotic factors. In the rocky intertidal, ebb tide causes pools of water to become isolated from bulk seawater forming tidepools, which can occur at different levels within the intertidal zone. The longer the tidepool is emerged from the bulk ocean, the greater the fluctuations in environmental characteristics (0 to 400% O2 air saturation, pH 7.0 to 9.5, 12 to 24oC ; Richards, 2011). In contrast, the lower intertidal may still be submerged during low tide and as a result experience less dramatic fluctuation in environmental variables (0 to 200% O2 air saturation, pH 7 to 8.5, and 12 to 18oC; Richards 2011). The subtidal and offshore environments stay relatively homogenous with the bulk ocean.  The variation in environmental gradient has a strong effect on species distribution along the marine near-shore environment, including multiple species of sculpins (Cottidae, Actinopterygii; Knope, 2013). Intertidal sculpins display high site fidelity and therefore they experience the environmental fluctuations particular to their habitat for prolonged periods of time, including drastic fluctuations in O2 levels (Green, 1971; Knope et al., 2017). As such, species that inhabit the higher intertidal which are more prone to both hypoxic and hyperoxic exposures have been shown to be more hypoxia tolerant with longer time to LOE than species found in the lower intertidal and subtidal 18  environments (Mandic et al., 2013). This greater hypoxia tolerance in higher intertidal sculpins was significantly correlated with a lower Pcrit which was further correlated with a higher gill surface area and higher hemoglobin O2-binding affinity (Hb-P50); both traits increase O2 movement from the environment into the animal’s circulation (Mandic et al., 2009). At the biochemical level, sculpins from the intertidal environment had higher brain LDH activity, which increases their anaerobic capacity and also their ability to buffer redox imbalances in the tissue (Mandic et al. 2013).  The sculpin system is ideal for the studies in this thesis for multiple reasons. First, evidence of adaptive variation in O2 extraction and delivery has been observed at multiple steps of the sculpin O2 transport cascade (Mandic et al., 2009; Mandic et al., 2013); however, whether there is similar interspecific variation in O2 use at the mitochondrial level has not been studied. Second, while traits to enhance O2 extraction and delivery would be beneficial during periods of O2 lack, the daily tidal cycle also brings about periods of hyperoxia, and tolerance to hyperoxia has yet to be investigated in sculpins. Hypoxia and hyperoxia place very different demands on processes involved in aerobic respiration and ROS metabolism. As the natural variation in O2 in the different intertidal zones have shaped sculpin hypoxia tolerance, these species provide an ideal system to study the interplay and potential evolutionary trade-offs between mitochondrial aerobic and ROS metabolism. Finally, a cross-lineage approach with closely-related species allows for the correlation of measured traits to the corresponding environmental factor while considering phylogenetic relationships. Where possible, I have chosen to characterize mitochondrial traits in multiple species (vs two species comparison) so as to draw conclusions and to form additional hypothesis about evolutionary processes that have shaped physiological traits (Garland and Adolph 1994). This has proven to be a valuable approach to account for the effects of phylogeny and identifying relationships of traits potentially adaptive in challenging environments. As such, with the sculpin model I will be able to assess the relationship between mitochondrial traits and whole animal hypoxia tolerance while incorporating phylogenetic relationships.   1.8.2 Chapter hypotheses  In the following three chapters, I assessed aspects of mitochondrial function using the intertidal sculpin model with the hypothesis that mitochondria from hypoxia tolerant species would show putatively adaptive traits that improve O2 binding, maximize phosphorylation efficiency, and 19  minimize ROS emission compared to less tolerant species, resulting in an overall more O2-efficient ETS (Fig.1.3). Hypothesis: Hypoxia tolerant sculpins exhibit higher O2 binding affinity at the mitochondria and COX compared with less hypoxia tolerant species. In Chapter 2, I studied the O2 kinetics of intact brain mitochondria and semi-purified COX. As mentioned earlier in the Introduction, few studies have attempted to examine whether there is variation in mitochondrial function as it relates to hypoxia tolerance and those that have, have generally focused on differences in how COX interacts with its other substrate, electron donor cytochrome c (Scott et al., 2011; Zhang et al., 2013), and have not directly studied COX O2 kinetics. In fact, it appears that there is little evidence for a mechanistic link between the binding affinities of cytochrome c and O2 (Krab et al., 2011). As the maintenance of O2 binding to COX is likely paramount in sustaining aerobic metabolism in hypoxia, I developed a protocol to investigate O2 kinetics of isolated brain mitochondria and semi-purified COX in 12 species of sculpins. Further, I used in silico protein modeling techniques to investigate possible underlying explanations for the interspecific variation in COX O2 binding affinity, potentially via modifications to COX3 subunit interaction with membrane phospholipid, cardiolipin (Lau et al., 2017).  Hypothesis: Mitochondria from hypoxia tolerant sculpins would have more efficient O2 use showing higher phosphorylation efficiency and lower ROS emission compared to less tolerant species.  In Chapter 3, I explored how the variation of mitochondrial and COX O2 binding affinities in Chapter 2 is manifested in intact mitochondrial function in regards to phosphorylation efficiency and ROS emission. Aerobic and ROS metabolism in mitochondria are tightly linked. While O2 usage powers ATP production and supports cellular work, at the same time it leads to generation of ROS which when accumulated can be damaging. With hypoxia tolerant animals having adapted mechanisms to improve O2 delivery and enhance O2 use, the question remains whether there are evolutionary trade-offs between aerobic and ROS metabolism. In this chapter, I investigated differences in mitochondrial respiration and ROS emission in isolated brain mitochondria from multiple sculpin species in order to identify adaptive traits that relate to whole animal ability to tolerate O2 variability. I hypothesized that mitochondria from hypoxia tolerant sculpins would have 20  more efficient O2 use which would be reflected in higher phosphorylation efficiency (P/O) as a result of increased complex I dependency (point (2) in Fig.1.3), better mitochondrial coupling (higher RCR due to reduced proton leak; point (3) in Fig.1.3), and lower overall ROS emission (lower ROS/O2) compared to less tolerant species (point (4) in Fig.1.3). Hypothesis: Hypoxia tolerant sculpins would reduce ROS accumulation and show less effects of ROS in response to hypoxia, hyperoxia, and normoxia recovery from both exposures compared to less tolerant species.  Previous studies have found species- and tissue-specific responses of ROS generation and scavenging capacities during exposure to O2 variability (Leveelahti et al., 2014). Building upon the characterization of isolated mitochondria (Chapter 3), in Chapter 4 I investigated whether there are consistent responses of ROS metabolism observed at the whole animal level. I exposed two sculpin species to 6hrs of hypoxia or hyperoxia followed by a quick normoxic recovery, the timing of which mimics the duration of a typical tidal cycle. I then assessed different aspects of ROS metabolism, including redox status, mitochondrial ROS, oxidative damage, and scavenging capacity. In order to assess mitochondrial H2O2 levels, I used a mitochondria-targeted mass spectrometry probe, MitoB, to measure in vivo changes in ROS (Logan et al., 2014).  Finally, in the General Discussion (Chapter 5), I summarize the major findings from my thesis research, discuss how my findings extend our understanding of mitochondrial physiology, and also pose future research questions generated from my work.  21   Figure 1.1 Vertebrate electron transport system (ETS). The ETS protein complexes are imbedded in the inner mitochondrial membrane (IMM), where NADH donates electrons to complex I (NADH oxidoreductase; CI) and succinate donates electrons to complex II (succinate dehydrogenase; CII), and along with glycerol-3-phosphate dehydrogenase (GPDH) and electron-transferring flavin protein (ETF) donate electrons to ubiquinone/ubiquinol (UbQ). UbQ subsequently donates electrons to complex III (cytochrome bc1 complex/ cytochrome c reductase; CIII) which donates its electrons to cytochrome c (Cyt c). Reduced cytochrome c binds to complex IV (cytochrome c oxidase; CIV) which is also where oxygen receives electrons and is reduced to water. Complexes I, III, and IV removes and/or pump protons (H+) from the mitochondrial matrix to the intermembrane space (IMS) to form the electrochemical gradient. The movement of protons back into the matrix via complex V (F1Fo-ATP synthase; CV) is used to drive the phosphorylation of ADP into ATP. More details of this process of oxidative phosphorylation is in section 1.2.  22    Figure 1.2. Balance of ROS scavenging rate and generation rate that results in accumulation of ROS. Under a normal redox environment (1), a high ROS scavenging rate with a low ROS generation rate results in a low net ROS level. Under highly reduced redox environment (2), high ROS generation rate exceeds ROS scavenging rate, resulting in an increase in ROS levels. Under highly oxidized redox environment (3) which lowers the cell’s redox buffering capacity and thus lowers ROS scavenging rate, results in an accumulation of ROS. Detailed description of figure in section 1.3.2. This figure is modified from Aon et al. (2010).   23   Figure 1.3. A hypothesis for an O2 efficient electron transport system. I hypothesized that the increased O2 delivery and binding upstream in the O2 transport cascade observed in the intertidal sculpin model (Mandic et al., 2009), would extend to the level of mitochondria such that there would be higher mitochondrial and COX O2 binding affinity (1). A higher dependency on complex I, which is a proton pump, would result in higher phosphorylation efficiency (assessed as P/O; 2). Further, lower proton leak (3) would result in higher mitochondrial coupling (assessed with respiratory control ratio; RCR) and lower futile proton cycling. Finally, lower ROS generation (4) would lower the amount of O2 used to generate a potentially harmful byproduct, ROS.    24  Chapter Two: Evolution of cytochrome c oxidase in hypoxia tolerant sculpins (Cottidae, Actinopterygii)3  2.1 Introduction Environmental hypoxia in marine ecosystems is increasing in both severity and duration due to climatic shifts and wide-spread eutrophication (Diaz and Breitburg, 2009). The increasing prevalence of hypoxia threatens to compress viable marine habitats, but whether a species will be impacted is dependent upon their hypoxia tolerance, which among fish is known to vary to a great degree (Chapman and Mckenzie, 2009). Indeed, naturally occurring hypoxia has been an important evolutionary driving force in fish, resulting in both convergent and divergent selection of physiological traits that enhance hypoxia tolerance (Richards, 2009; Richards, 2011). Understanding the determinants of variation in physiological function as it pertains to hypoxia tolerance is increasingly important in order to parametrize predictive models that seek to define how organisms will respond to the greater prevalence of hypoxia in the marine environment.  Adaptive modifications of the vertebrate oxygen (O2) transport cascade are well described in various groups of organisms that encounter hypoxia including fish (Chapman and Mckenzie, 2009), high altitude geese (Scott et al., 2009) and deer mice (Lui et al., 2015; Natarajan et al., 2015; Storz et al., 2009). Hypoxia tolerant organisms typically have larger respiratory surface areas (gills, Nilsson, 2007; lungs in birds, Scott et al., 2011), higher haemoglobin-O2 binding affinity (Jensen and Weber, 1982), and lower tissue O2 demands (Hopkins and Powell, 2001) all of which facilitate the maintenance of aerobic function even in the presence of sometimes severe environmental hypoxia. Indeed, among sculpins (diverse species of fish from the family Cottidae; Actinopterygii) that live along the marine near-shore environment, an environment typified by strong spatial and temporal variation in hypoxia exposure, Mandic et al.(2009) demonstrated a phylogenetically-independent relationship between the critical PO2 for O2 consumption rate (Pcrit; the environmental PO2 below which animal O2 consumption rate conforms to decreasing environmental PO2), and several traits along the O2 transport cascade. Hypoxia tolerant sculpins have larger mass specific gill surface area, higher                                                  3A version of this chapter has been published: Lau GY, Mandic, M, and Richards JG. Evolution of cytochrome c oxidase in hypoxia tolerant sculpins (Cottidae, Actinopterygii). Molecular Biology and Evolution. 34, 2153-2162.  25  hemoglobin-O2 binding affinity (low Hb-O2 P50), and generally lower metabolic rates compared with less hypoxia tolerant species. Although it is well established that modifications to the O2 transport cascade are important determinants of hypoxia survival in fishes and other organisms, much less is known about how variation in hypoxia tolerance is related to mitochondrial function and the actions of cytochrome c oxidase (COX), which is the protein responsible for the majority of whole-animal O2 use during respiration.   Cytochrome c oxidase is an ancient, multi-subunit enzyme that is thought to be older than the surge in atmospheric O2 that occurred 2.4-2.1 billion years ago and led to the explosion of eukaryotic biodiversity (Castresana et al., 1994; Lyons et al., 2014). Although the original function of COX remains unknown, COX in oxic mitochondria catalyze the final transfer of electrons from the electron transport system (ETS) to O2, reducing it to water while simultaneously pumping protons to generate a proton electrochemical gradient for ATP synthesis via the F1FO ATP-synthase. The subunits that compose the COX catalytic core, COX1, 2, and 3 are all coded for by the mitochondrial genome which is typified by a high mutation rate that can be ~10 times greater than that of the nuclear genome (Brown et al., 1979; Pierron et al., 2012). These COX subunits could thus be hotspots of genetic and hence functional variation upon which natural selection can act. Indeed, non-synonymous substitutions in COX 1 and 2 have been identified between low and high-altitude pika (Ochotona curzoniae) (Luo et al., 2008) and functional analysis of COX orthologues among geese (Scott et al., 2011) and locusts (Zhang et al., 2013) has revealed putatively-adaptive variation whereby COX from the more hypoxia tolerant, high-altitude species or populations have a higher binding affinity for cytochrome c than the lower altitude organisms. Although these studies suggest that aspects of COX function are under selection in organisms inhabiting hypoxic environments, no study has yet to examine whether there is adaptive variation in the kinetics of COX interactions with O2, which is the critical element limiting survival in the hypoxic environments.  The goal of this study is to determine if there is functional variation in the kinetics of mitochondrial and COX interactions with O2 among species of marine sculpins that vary in their ability to withstand O2 deprivation. We chose to focus on mitochondria and COX from the brain because of its importance to the maintenance of whole-animal function in hypoxia. Further, to provide a mechanistic understanding of COX functional differences in the absence of being able to generate 26  recombinant protein of the complex multi-subunit COX enzyme, in silico analyses of deduced protein sequences were performed to highlight the putatively important amino acid sites on COX under selection by hypoxia.  2.2 Materials and methods 2.2.1 Chemicals Cytochrome c (from equine heart; Sigma-Aldrich) was reduced via dialysis with ascorbate in 50mM Tris-HCl buffer, pH8.0. The concentration of reduced to oxidized cytochrome c was determined by spectrophotometry as A550/A280 (between 1.1 to 1.3) and was stored in aliquots at -80oC until use. Unless otherwise specified, all chemicals were purchased from Sigma-Aldrich (Canada).  2.2.2 Species collection and holding  Sculpins were collected near Bamfield Marine Sciences Centre (British Columbia, Canada) at Ross Islets (48o52.4’N, 125o9.7’W) and Wizard’s Rock (48o51.5’N, 125o9.4’W) using either handheld nets or pole seines at the lowest tidal cycle. Animals were either sampled after one week of housing in flow-through seawater (12°C) or transported to The University of British Columbia (UBC) and housed in a recirculation system with artificial seawater (12°C) and maintained on a diet of shrimp, Atlantic krill, and bloodworms for at least 3 weeks. All experimental procedures were reviewed and approved by the UBC Animal Care Committee (A13-0309).  2.2.3 Mitochondrial isolation  Mitochondria were isolated from whole brain of 8 species of sculpins including (Oligocottus maculosus, Myoxocephalus polyacanthocephalus, Clinocottus globiceps, Hemilepidotus hemilepidotus, Leptocottus armatus, Artedius fenestralis, Artedius lateralis, Blepsias cirrhosus). Briefly, fish were stunned via concussion, euthanized by spinal severance and the brain was dissected and minced with a razor blade on ice. The minced tissues were transferred to a glass homogenizer containing isolation media (in mM: 25 KH2PO4, 50 KCl, 10 HEPES, 0.5 EGTA, 250 sucrose, 0.5% bovine serum albumin (BSA), pH 7.4) and homogenized with 3 full passes.  The resulting homogenate was transferred to centrifuge tubes and centrifuged at 600g for 10min at 4oC. The supernatant was collected, filter though glass-wool, and centrifuged at 9000g for 10min at 4oC. The pellet was suspended in isolation media and centrifuged again at 9000g for 10min at 4°C.  The final pellet was suspended in isolation media 27  without BSA and either used immediately (for whole mitochondrial measurements) or frozen in aliquots in liquid nitrogen (for COX Km,app O2) and stored at -80oC for later analysis.  2.2.4 COX Km,app O2 Frozen mitochondria were freeze-thawed three times and assayed after the third thaw in the Oroboros oxygraph (Innsbruck, Austria) that was calibrated daily to 100% air saturation and anoxia at 12oC in assay buffer (in mM at pH 7.2: 25 K2HPO4, 5 MgCl2, 100 KCl, 2.5mg/mL BSA). An assay temperature of 12°C was chosen for this analysis because it is the typical marine water temperature off the coast of British Columbia and the temperature to which these animals were acclimated. Buffer pH was kept consistent across species to enable comparisons of COX Km,app O2 under constant conditions for all species. Once the air calibration signal stabilized, chamber O2 was reduced to ~100µM by passing a stream of nitrogen gas over the surface of an open oxygraph chamber. The chamber was then sealed and 4mM ascorbate, 0.5mM N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD), and 100µM reduced cytochrome c were added sequentially, followed by the addition of the freeze-thawed mitochondria. The progression of O2 consumption was then monitored until anoxia.  Background auto-oxidation was determined in separate trials (without freeze-thawed mitochondria), fit with a two-phase decay non-linear curve, and used to correct experimental O2 consumption rate. The corrected O2 consumption curves were then fit with a two-phase decay and the fast phase half-life was taken as the COX Km,app O2.   2.2.5 Mitochondrial P50 Mitochondrial respiration was measured with the Oroboros oxygraphy, which was calibrated daily to 100% air saturation and anoxia at 18oC in assay buffer (MiR05; in mM at pH 7.1: 0.5 EGTA, 3 MgCl2·6H2O, 60 lactobionic acid, 20 taurine, 10 KH2PO4, 20 HEPES, 110 D-sucrose, 1g/L BSA). We chose to perform the analysis of mitochondrial P50 at 18°C instead of at their acclimation temperature of 12°C because of the higher signal-to-noise-ratio at the warmer temperature.  Once the air calibration signal stabilized, mitochondria (~0.1mg protein) was introduced into the chamber and the chamber was then sealed. To fuel complex I, 5mM pyruvate, 2mM malate, and 10mM glutamate were added, following which 10mM succinate was added to fuel complex II. 1mM ADP was then added to stimulate maximum state III respiration rate. The progression of O2 consumption 28  was then monitored until anoxia. The analyses of mitochondrial P50 was performed with DatLab2 (Oroboros; Innsbruck, Austria).  2.2.6 COX respiration rate (ascorbate-TMPD/FCCP respiration rate) Samples of whole brain mitochondria were introduced into the calibrated Oroboros chamber set to 18oC. After complex I (5mM pyruvate, 2mM malate, and 10mM glutamate) and II (10mM succinate)-fuelled state III respiration was established (data not shown), the ETS was uncoupled with titration of two to three 0.05µM steps of carbonyl cyanide 4-(trifluoromethyoxy)phenylhydrazone (FCCP). After mitochondria were fully uncoupled (when respiration rate no longer increased with FCCP titration), 0.5µM rotenone and 2.5µM antimycin A were added to the chamber to inhibit complexes I and III respectively, followed by 2mM ascorbate and 0.5mM TMPD that were added to donate electrons to cytochrome c and maximally stimulate COX respiration rate. The COX respiration rate was corrected for the empirically determined chemical background of ascorbate-TMPD autoxidation at various PO2 values. COX respiration rate was normalized to the uncoupled respiration rate determined in the presence of FCCP and complex I and II substrates (an estimate of ETS capacity). The fact that the normalized values are close to or slightly below 1 (but not significantly; one sample t-test with Bonferonni correction for multiple comparisons) does not affect the relationship between COX respiration between species as the same trend is observed when COX respiration is expressed to complex I and II-fuelled state III ADP-stimulated respiration rate (data not shown). 2.2.7 COX voltage recovery rate Membrane potential was monitored in freshly isolated mitochondria using tetraphenylphosphonium (TPP+) ion selective electrodes connected to the oxygraph unit. Four additions of 0.5mM TPP were used for calibration.  After TPP+ calibration, mitochondria were injected into the chamber. Complex I (5mM pyruvate, 2mM malate, and 10mM glutamate) and II (10mM succinate) substrates were used to induce state II respiration. 4mM ADP was added to stimulate state III respiration. The O2 inside the chamber was then allowed to deplete and during the resulting anoxic period (not exceeding 10min), the following were introduced into the 2mL oxygraph chamber: 560U catalase, 0.35µM ascorbate, 0.1µM TMPD, 1µM rotenone, 5µM antimycin A, 5µM oligomycin to ensure the other ETS complexes were inhibited, and that COX had saturating levels of electron donors to achieve maximal activity when 29  O2 was reintroduced to the chamber. To reoxygenate, 0.08µM (in 3µL) H2O2 was injected into the chamber and in the presence of catalase it quickly reintroduced O2 into the chamber.  To estimate membrane potential recovery rate, the initial velocity of TPP signal recovery was calculated per trial, and normalized to the maximum TPP signal to which the sample recovered to after reoxygenation. In order to account for varying concentrations of COX enzyme between samples, this recovery rate was then expressed relative to the ascorbate-TMPD stimulated respiration rate (expressed as sec-1 max COX respiration rate-1).  2.2.8 cox1 and cox3 sequencing cox1 and cox3 genes which make up the catalytic core were sequenced from six sculpin species including Oligocottus maculosus, Artedius lateralis, Artedius fenestralis, Myoxocephalus polyacanthocephalus, and Blepsias cirrhosus. These species were chosen because they are found on different clades of the sculpin phylogeny (Knope, 2013), and they show a range in whole animal hypoxia tolerance (Pcrit) and COX Km,app O2 (Figure 1). Mitochondrial DNA was extracted from muscle of 2-3 individuals per species using DNeasy Tissue Kit (Qiagen). Degenerate and specific primers (Supplementary Table 3) for PCR amplification were designed for cox1 and cox3 using GeneTool Lite 1.0 (BioTools). PCR products were purified with QIAquick PCR purification kit (Qiagen), and sequenced using an Applied Systems 3730 DNA Analyzer. A consensus sequence for each species was determined using Geneious (Drummond et al. 2012). Sequences were submitted to Genbank (accession numbers KY356329-KY356352). 2.2.9 Protein in silico analyses  The deduced consensus cox1 and cox3 sequences for each species were translated in Geneious (with BLOSUM matrix) and modeled separately using bovine heart COX structure (3ABM PDB) as template. Swiss-PdbViewer was used to view and manipulate protein data bank (PDB) files. The 3D structures were created using PyMOL. Comparisons of functional domains were determined after aligning to annotated bovine heart COX (3ABM) Protein stability analyses FoldX (Schymkowitz et al., 2005) (with YASARA view (Krieger and Vriend, 2014)) was used to calculate protein stability (ΔG). To prepare data for FoldX analyses, modeling results from Swissmodel were exported as PDB files, which were first repaired using FoldX (using the 30  REPAIRPDB function to undergo energy minimization of overall protein structure). The FoldX algorithm was then applied to calculate protein stability (ΔG). Subunit and phospholipid interactions were investigated using PDBePISA (EMBL-EBI) after modeling all three subunits to the 3ABM catalytic core. ABS-Scan (Anand et al., 2014) online platform was used to carry out in silico alanine scanning mutagenesis on COX3 and cardiolipin (CDL270 in 3ABM PDB file) interaction. The larger the ΔΔG value indicates larger 𝛥G between wild-type and mutated protein, and the more the alanine mutation disrupted protein stability.   2.2.10 Statistical analyses Correlative analysis was performed using both ordinary least squares (OLS) and phylogenetically generalized least squares (PGLS) using ape (Paradis et al., 2004), Geiger (Pennell et al., 2014) and nlme (Pinheiro et al., 2014) packages in R (Team, 2016). For the phylogenetic analyses we used the most updated phylogenetic tree of the marine species of Superfamily Cottoidei (Knope, 2013) and dropped the tips of the tree for species with no available data. Each set of regressions from our data set was tested under OLS and PGLS (Pagel, 1999). Pagel’s λ of 0 indicated that the correlation was independent of phylogeny, whereas λ value of 1 is consistent with the constant-variance model (or Brownian motion model). The model with the lower Akaike’s Information Criterion (AIC) value represented the better fitting model (Table 1).  2.3 Results and Discussion 2.3.1 Interspecific variation in COX function and mitochondrial P50 In order to assess whether there is adaptive variation in how COX interacts with O2, we assessed the kinetic properties of COX orthologues from the brain of eight species of sculpins, previously shown to vary in hypoxia tolerance (assessed as time to loss of equilibrium) and Pcrit (Mandic et al., 2009; Mandic et al., 2013). Among the eight species of sculpins, there was large variation in the apparent Michaelis-Menten constant for O2 binding to COX (Km,app O2) and this variation was significantly correlated with previously determined Pcrit values (Fig.2.1A; Table2.1). COX from hypoxia tolerant sculpins (those with lower Pcrit values) had a lower Km,app O2 than the less hypoxia tolerant species. These empirically determined values for COX Km,app O2 are generally lower than the values reported in mammals (ranging between 0.25 to 0.66 µM in this study compared to 0.5-1µM Km of O2 in bovine COX (pH 7.4 at 25oC; reviewed in Brunori et al., 1987, and Nicholls and Chance 1974), which may not be surprising considering the general sensitivity of most mammals to O2 deprivation. 31  Notwithstanding the generally lower COX Km,app O2 in fish relative to mammals, this is the first study to show that the evolution of hypoxia tolerance is directly associated with functional modifications to COX that increase O2 binding affinity and facilitate the maintenance of aerobic metabolism in hypoxia.   In addition to the adaptive variation in COX Km,app O2 observed herein, previous comparative studies have shown a lower apparent binding affinity of COX for cytochrome c in high altitude species or populations compared with lower altitude species and populations (Scott et al., 2011; Zhang et al., 2013). Although it is tempting to think that the apparent binding affinity of COX for both cytochrome c and O2 would co-vary, there appears to be no clear mechanistic link of steps in the COX redox cycle that involve cytochrome c electron transfer to CuA on COX2 exerting control over Km for O2 (Krab et al., 2011). As such, the mechanistic origin of the adaptive variation in COX Km,app O2 does not appear to be linked to cytochrome c binding.   COX Km,app O2 can be modified by the energy and redox states of the ETS (Krab et al., 2011). In order to investigate whether respiratory and redox states affect the interspecific relationship between Pcrit and COX Km,app O2 we assessed brain mitochondrial P50 in five species of sculpins under state III phosphorylating conditions with complex I and II fuels. Under these strongly reducing conditions, we observe a significant relationship between Pcrit and mitochondrial P50, with the most hypoxia tolerant sculpin having the lowest mitochondrial P50 (Fig.2.1A, Table2.1).  The interspecific relationships between Pcrit and both mitochondrial P50 and COX Km,app O2 were roughly parallel (Fig.2.1A), but the mitochondrial P50 was 1.6 to 2-fold higher than the associated COX Km,app O2, which may be due to greater diffusion distances for O2 in the intact mitochondria compared with the semi-purified COX protein. Cellular O2 levels are estimated to be 2-5µM O2 (~0.14-0.35kPa), and below 2µM O2 (~0.14kPa) closer to mitochondrial cluster (Jones, 1986). These O2 concentrations are many fold higher than what the mitochondrial ETS normally requires to maintain function, suggesting that mitochondria and COX do not lack O2 under normal physiological conditions but when environment O2 is diminished, cellular PO2 will decrease necessitating adaptations at the mitochondrial and COX level to improve O2 kinetics in species inhabiting these environments. The similar interspecific relationships between Pcrit and both COX Km,app O2 and mitochondrial P50 lead us to conclude that the variation in COX Km,app O2 among 32  sculpins is due to intrinsic properties of the COX enzyme and not due to modifying effects of mitochondrial respiration or redox state.  In combination with previous studies (Mandic et al., 2009; Mandic et al., 2013), we have now characterized five steps of the O2 transport cascade for multiple species of sculpins that differ in whole animal hypoxia tolerance (represented as time to loss of equilibrium (LOE) in Supplementary Fig.2.1), from the extraction of environmental O2 to the level of O2 interactions with COX in mitochondria. Our interspecific analysis reveals a strong relationship between Pcrit, whole red blood cell P50, and stripped hemoglobin-O2 P50, and also between mitochondrial P50 and COX Km,app O2, whereby hypoxia tolerance sculpins show coordinated changes at multiple steps in the O2 transport cascade which would serve to not only improve O2 extraction from the hypoxic environment, but also improve or sustain O2 delivery to mitochondria. In addition to COX Km,app O2, COX Vmax and respiration rate also shows interspecific differences related to hypoxia tolerance. Hypoxia tolerant sculpins with a lower mitochondrial P50 have a lower COX Vmax (determined on semi-purified COX; Fig.2.1B, Table2.1), but higher COX respiration rate relative to total ETS capacity (ascorbate-N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD)/ carbonyl cyanide 4-(trifluoromethyoxy)phenylhydrazone (FCCP) respiration rate of whole brain mitochondria; Fig.2.1C; Table2.1; note that no value in Fig.2.1C is lower than 1 according to one-sample t-test with Bonferroni correction for multiple comparisons). The lower COX Vmax (Fig.2.1B) in hypoxia tolerant sculpins with lower mitochondrial P50 suggests that these species have a less powerful COX or lower COX protein content (Vmax = kcat x [E]), but the higher COX respiration rate determined in uncoupled mitochondria (Fig.2.1C) suggests that the COX protein operates at a higher level relative to total ETS flux in hypoxia tolerant sculpins compared with hypoxia intolerant sculpins. These divergent relationships between COX Vmax and COX respiration rate suggest that other mitochondrial components, perhaps the phosphorylation system or allosteric regulation by ATP/ADP (Ludwig et al., 2001), may contribute to regulating COX activity in the intact mitochondria and that hypoxia tolerant sculpins appear to have COX proteins that can function to a greater extent of mitochondrial maximal capacity than do hypoxia intolerant sculpins. Excess COX capacity is widely acknowledged in the mammalian literature and several hypotheses have been put forward to explain the potential benefit of higher, or even excess COX capacity, toward aerobic metabolism under O2 limiting conditions. First, as PO2 is reduced, the flux control coefficient of 33  COX over ETS flux increases such that the total ETS flux is increasingly controlled by COX. As such, in organisms that frequently encounter hypoxic conditions, higher COX enzyme capacity could lower the flux control coefficient and reduce COX control over ETS flux, which would support relatively higher ETS flux even when PO2 is decreasing (Gnaiger et al., 1998; Suarez et al., 1996). Second, it has been hypothesized that a greater mitochondrial COX capacity could serve to store excess electrons and reduce reactive O2 species generation, which could potentially alleviate oxidative stress during recovery from hypoxia, although this hypothesis has not been empirically tested (Campian et al., 2007). A similar relationship in COX Vmax as the one in sculpins was observed when comparing the hypoxia tolerant epaulette shark (Hemiscyllum ocellatum) and to the less tolerant shovelnose ray (Aptychoterma rostrata) where the less tolerant species exhibited higher COX activity (Hickey et al., 2012). However, the opposite trend was observed relating hypoxia tolerance and COX Vmax among triplefin fishes (familes Bellapiscis and Forsterygion; Hilton et al., 2010), another group of fish that inhabit the marine near-shore environment, and also in highland deer mice that have higher COX Vmax than their lowland counterparts (Cheviron et al., 2014; Lui et al., 2015). These conflicting results suggest that a lower COX Vmax is not a universal evolutionary strategy for surviving environments prone to fluctuating PO2.  Our interspecific comparison of sculpins inhabiting the near-shore marine environment clearly suggests that a low mitochondrial P50 and low COX Km,app O2 are putative adaptations that enhance hypoxic survival in the more hypoxia tolerant sculpin species found in the O2 variable intertidal zone, compared with species from the more O2 stable subtidal environment. However, O2 is not the only abiotic factor that varies in the intertidal environment. Tidepool temperatures can vary from ~12°C during high tide (bulk ocean temperature near collection sites) or low tide at night to upwards of 24oC or higher during daytime emergence (Richards, 2011). In ectotherms, increases in temperature necessitate increases in O2 uptake in order to support a higher metabolic rate, but for many enzymes, increases in temperature are known to increase Km for substrate binding (Holland et al., 1997). Furthermore, temperature-induced changes in cellular pH (alpha-stat) may also impact protein function and Km (reviewed in Burton, 2002). Thus, it is also tempting to speculate that temperature variation may also serve as a selective pressure underlying the interspecific variation in COX Km,app O2 whereby species that experience warm temperatures have evolved a COX protein with lower Km,app O2 to offset temperature/pH dependent effects. However, the natural patterns of O2 and temperature fluctuations experienced by our most hypoxia tolerance species do not support 34  temperature as an important variable underlying the low COX Km,app O2. For example, daytime emergence of tidepools is associated with both warm temperatures and hyperoxia (up to ~400% air saturation; Richards 2011), thus any warm temperature induced increases in COX Km,app O2 would be offset by higher O2 availability. In contrast, nighttime emergence is associated with often severe hypoxia, but at temperatures that are near the bulk ocean temperatures of 12°C. Thus, based on the natural fluctuations in O2 and temperature experienced by fish living in tidepools, it seems likely that the interspecific variation in COX Km,app O2 is due to variability in O2, not temperature, but this assertion requires validation. 2.3.2 COX protein in silico analyses  In order to gain insight into the mechanistic underpinnings of the adaptive variation in COX Km,app O2 among sculpins, we adopted an in silico approach based on protein modelling of deduced amino acid sequences of the mitochondrial-encoded COX core. The catalytic core of the COX is comprised of three mitochondrial-encoded subunits: subunit 1 (COX1) contains the binuclear O2 binding site (containing heme a3 and CuB), subunit 2 (COX2) contains the docking site for reduced cytochrome c and CuA, and subunit 3 (COX3) which is essential for maintaining protein function (Fig.2A&B; Bratton et al., 1999; Wikström et al., 2015), serves as the putative entry point for O2, and protects the entrance of the D proton transfer pathway at the COX1/COX3 interface (Hosler, 2004; Sharma et al., 2015). Given the lack of a clear mechanistic link between cytochrome c binding to COX2 and O2 binding (Krab et al., 2011), we elected to focus our analysis on COX1 and COX3. Variation in the apparent binding affinity of O2 to COX could be due to differences in the fast trapping of O2 by COX1 through the rapid electron transfer from the heme a and a3 (Verkhovsky et al., 1996) or due to differences in the rate of O2 diffusion from outside of this multimeric protein to COX1 heme a3 (Riistama et al., 1996). As all vertebrates contain the same aa3-type COX1 (with heme a and heme a3) it is unlikely that the variation that we measure in COX Km,app O2 is due to variation in O2 trapping, thus we focus on the role of variation in the O2 diffusion pathways through the COX protein. A putative COX O2 diffusion pathway was elucidated in the crystal structure of bovine heart COX and also via site-directed mutagenesis studies of aa3-COX from Paracoccus denitrificans, which involves a hydrophobic channel that extends from the COX surface through a COX3 v-cleft structure to the COX1 binuclear O2 binding site (Riistama et al., 1996; Tsukihara et al., 1996). We therefore hypothesized that amino acid variation in COX1 and/or COX3 subunits could 35  affect the pathway of O2 diffusion and thus sculpin Km,app O2. To investigate this possibility, we evaluated sequence-based differences of COX1 and 3 subunits by aligning and modeling the two COX subunits from six sculpin species that varied in COX Km,app O2 using bovine heart COX protein crystal structure as template.  2.3.3 COX1 protein structure  In COX1, all residues important for heme a, a3 binding were conserved as were all known residues participating in the proton pumping D-pathway, through which chemical and pumped protons are transferred (Wikström et al., 2015). Indeed, our sequence analysis demonstrated a high degree of conservation in COX1 sequence among the sculpins investigated and two species that differed in COX Km,app O2 (Myoxocephalus polyacanthocephalus and Blepsias cirrhosus; species 2 and 8 from Fig.2.1) had identical deduced amino acid sequences. At the four positions where we observe variation in amino acid sequence, the residues extend into both the intermembrane space and matrix and not towards the binuclear catalytic site (Supplementary Table1), suggesting that their role in defining COX Km,app O2 is minimal.  2.3.4 COX3 protein structure and stability  It was previously thought that the only role of COX3 was to impart structural integrity to the catalytic core, but it is now recognized that COX3 plays an essential role in maintaining COX activity at high pH by protecting the microenvironment of the proton acceptor of the D-pathway, as well as preventing suicide inactivation during the catalytic cycle (Bratton et al., 1999; Hosler, 2004). In addition, COX3 is the putative site for O2 entry into the large COX complex en route to the COX1 catalytic site via a distinct v-cleft structure formed from two bundles of seven transmembrane helices of COX3. Three histidine residues near the N-terminus of COX3 have been shown to play a critical role in accepting and donating protons for possible use by D-pathway (Alnajjar et al., 2014). Combined, these findings suggest that COX3 not only has an important role in maintaining O2 diffusion into the protein core but also proton uptake.  Cumulative point mutations on a protein surface can affect its interactions with the surroundings and alter protein stability/protein function (Eijsink et al., 2004; Strickler et al., 2006). As COX3 does not contain any catalytic residues, we first estimated in silico protein stability of the COX3 orthologues to determine whether interspecific amino acid substitutions could potentially affect COX3 protein surface and be related to COX Km,app O2. This analysis revealed a significant 36  relationship between COX3 protein stability with brain COX Km,app O2 (estimated as free energy of unfolding; Fig2.2B; Table2.1), suggesting that COX3 protein may play a role in determining Km,app O2. A more stable COX3 protein in more hypoxia tolerant sculpin suggests that they either have a more compact and rigid protein, and/or they differ in how COX3 interacts with its surroundings due to surface mutations (e.g. with other protein subunits or phospholipids). The COX3 models reveal that its interface with COX1 is well conserved across sculpin species, whereas the opposing interface with the phospholipid bilayer showed interspecific amino acid residue differences (Fig.2.2D). In particular, there were three residues showing interspecies variation in amino acid functional groups on helix 2 which forms one arm of the previously thought highly conserved COX3 v-cleft structure (Fig.2.2B&D, Supplementary Table 2.2). Sequence analysis further suggests that interspecific differences in COX3 residues could also affect interactions with a high affinity bound cardiolipin adjacent to the v-cleft, and also with nuclear subunits 5b, 6a, 6b, and 7a (Supplementary Table 2.2).  2.3.5 Interspecific variation in mitochondrial kinetics does not affect proton pumping Since differences in COX3 could potentially affect proton uptake to the D-pathway in COX1, we assessed whether the interspecific variation in O2 kinetics was associated with differences in COX proton pumping, estimated as the speed with which anoxic brain mitochondria recovered the proton gradient upon re-oxygenation. However, this analysis did not reveal a significant relationship between COX Km,app O2 and the rate of proton gradient recovery (Supplementary Fig.2.2). There may, however, be interspecific variation in the rate of recovery of post-anoxia membrane potential suggesting possible modifications to proton transfer unrelated to D-pathway residues in COX1 and these modifications deserve more in-depth study.  2.3.6 COX3 protein stability may affect cardiolipin interactions and mitochondrial function Cardiolipin is a membrane phospholipid found exclusively in mitochondria and plays important roles in the function of COX (Alnajjar et al., 2015; Hofacker and Schulten, 1998). Four cardiolipin molecules are directly associated with COX, and one in particular, situated within the COX3 v-cleft with its head group in contact with COX7a (Arnarez et al., 2013), is thought to be critical for O2 uptake by the protein due to the higher partition coefficient of O2 in phospholipids than aqueous media (Hofacker and Schulten, 1998). Indeed, Sedlak and Robinson (2015) demonstrated that the cardiolipin associated with COX7a (and COX3) is critical for the function of COX. Given the 37  interspecific variation in COX O2 kinetics (Fig.2.1A), COX3 protein stability (Fig.2.2C) and COX3 residues that interact with COX7a (Fig.2.2A; Supplementary Table 2.2), we hypothesized that residues important for cardiolipin interaction would vary between sculpin species. We carried out in silico alanine scanning mutagenesis to mutate residues between COX3 and cardiolipin on the bovine COX crystal structure to alanine to identify important sites involved in the protein-ligand (COX3-cardiolipin) interaction. These analyses yield a ΔΔG, representing the difference in ΔG between the native protein and a mutated protein where the native amino acid is replaced with an alanine residue (with a small methyl side chain) to investigate protein-ligand interactions. The results of this analyses predicted that alterations in both positions 55 and 224 would have the greatest impact on COX3-cardiolipin interactions, pointing to both positions having key roles in cardiolipin recognition by forming hydrogen bonds with the structure (Fig.2.2E). Mapping of residues 55 and 224 onto a simplified sculpin phylogeny indicates a pattern that suggests that variation in this cardiolipin interaction may have been important in sculpins invading the upper intertidal environment (Fig.2.3, Supplementary Fig.2.3), which is typified by daily bouts of hypoxia at night. Two exceptions to this pattern can be observed in M. polyacanthocephalus and O. maculosus, both of which have similarly low COX Km,app O2 but differ at residues 55 and 224. Additional comparison between M. polyacanthocephalus and B. cirrhosus, our most hypoxia intolerant species, reveals two of the four amino acid differences (positions 41 and 47) are also found on helix 2 (same as position 55) which forms an arm of the COX3 v-cleft (Fig.2.3, Supplementary Table 2.2). This provides further evidence for a role of the COX3 v-cleft in determining COX Km,app O2 in sculpins and that the low COX Km,app O2 seen in hypoxia tolerant sculpins can be achieved via genetic mechanisms. As such, we provide the first mechanistic hypothesis for a previously unrecognized, but potentially critical adaptation that ensures hypoxia tolerant organisms maintain mitochondrial function and aerobic metabolism to a lower PO2 than in hypoxia intolerant species.  2.4 Summary The present study provides novel evidence of adaptive variation in the function of COX, arguably the most important protein in aerobic respiration, where organisms that have evolved a higher degree of hypoxia tolerance possess a high O2 affinity COX that functions to a greater extent of its maximal activity. These adaptive modifications to COX and mitochondria in hypoxia tolerant sculpins translates into these species being better able to maintain mitochondrial function to a lower cellular PO2 than in hypoxia intolerant sculpins. As such, we provide strong interspecific evidence 38  that the mitochondrion and O2-binding COX protein are under selection for improved function in organisms that experience hypoxia more frequently in their natural environment. Furthermore, we have identified several amino acid residues on the sculpin COX3 structure that are strong candidates for explaining the adaptive variation in COX Km,app O2 through modulation of the high-affinity interaction between COX3 and cardiolipin.  39  Figure 2.1 (A) Relationship between whole animal hypoxia tolerance (Pcrit) and brain mitochondrial P50 (Data are mean ± s.e.m..; phylogenetic generalized least squares (PGLS), p=0.03, y= 40.07x + 2.71; hollow squares) and COX Km,app O2 (PGLS, p< 0.0001, y= 37.12x + 3.75; black circles) and relationship between brain mitochondrial P50 and (B) COX Vmax enzyme activities among sculpins (PGLS, p<0.0001, y= 4607.04x + 173.74), and (C) ascorbate-TMPD stimulated mitochondrial respiration (COX respiration rate; ordinary least squares (OLS), p= 0.015, y= -17.80x + 1.90). Species indicated in the figure are as follows: (1) Oligocottus maculosus, (2) Myoxocephalus polyacanthocephalus, (3) Clinocottus globiceps, (4) Hemilepidotus hemilepidotus, (5) Leptocottus armatus, (6) Artedius fenestralis, (7) Artedius lateralis, (8) Blepsias cirrhosus. Pcrit values used in panel A are taken from Mandic et al. 2009.  40   Figure 2.1  41  Figure 2.2 (A) Structure of whole COX enzyme (bovine heart 3ABM PDB structure) with COX1 (in orange), COX2 (in green), COX3 (in blue) and COX7a (in pink) highlighted; (B) COX1 (in orange) and COX3 (in blue) structures that were investigated for interspecific differences between sculpin species (heme a3, a part of the binuclear site, is shown in pink in the COX1 structure); (C) Relationship between brain COX Km,app O2 and COX3 subunit protein stability (estimated as free energy of unfolding, in kcal/mol; data are means ± s.e.m.; OLS, p= 0.029, y= 0.0052x +0.40); (D) COX3 structure showing sculpin interspecific differences in amino acid residues. Cardiolipin (CDL270) is shown in green. Blue in panel D identifies residues where the least hypoxia tolerant species in our study Blepsias cirrhosus is different from the other species, and orange highlights positions where more hypoxia tolerant species Oligocottus maculosus, Artedius fenestralis, and Artedius lateralis are different from the others (refer to Supplementary Table 2.2); (E) ΔΔG of mutants from alanine mutagenesis analyses of COX3 interacting residues with cardiolipin (CDL270) in the bovine structure, showing particular importance of residues 55 and 224 for ligand recognition.  42   Figure 2.2  43  Figure 2.3 Phylogenetic transitions of two amino acid residues 55 and 224 on COX3 (shown here on bovine 3ABM chain C structure; UnitProtKB accession number P00396) identified from the alanine mutagenesis analyses show that upper intertidal species have functionally different amino acid residues when compared to lower intertidal species. Each species is indicated with a number corresponding to those in Figure 2.1.   44   Figure 2.3     45  Regression (x, y) Test (OLS/PGLS) Slope estimate slope p value AIC Pagel’s λ Brain COX Km,app O2, Pcrit OLS 50.95 0.12 21.87  PGLS 37.12 <0.001 9.80* 1.21 Brain mitochondrial P50, Pcrit OLS 41.41 0.081 11.90  PGLS 40.07 0.030 6.20* 1.08 COX Vmax, brain mitochondrial P50 OLS 5432.10 0.066 59.74  PGLS 4607.04 <0.001 31.51* 1.21 Ascorbate-TMPD/FCCP, brain mitochondrial P50 OLS -17.80 0.014 -3.41*  PGLS -21.64 0.0033 -2.83 -0.64 COXIII protein stability, brain COX Km,app O2 OLS 0.0052 0.029 -39.79*  PGLS 0.0046 0.014 -16.87 1.16  Table 2.1 Regressions tested under ordinary least squares (OLS) and phylogenetic generalized least squares (PGLS) with Pagel’s model of evolution (Pagel 1999). Pagel’s λ is an indication of whether the phylogeny correctly predicts the patterns of covariance among species on a given trait. λ value of 0 indicates that the correlation is independent of phylogeny, whereas value of 1 is consistent with the constant-variance model (or Brownian motion model) being a good representation of the data. Model that best represented the data with the lowest Akaike’s Information Criterion (AIC) value is indicated with an asterisk.   46  Chapter Three: Hypoxia tolerance is associated with higher mitochondrial ROS emission in intertidal sculpins (Cottidae, Actinopterygii)  3.1 Introduction  Oxygen (O2) use in eukaryotic cells can be a “double-edged sword”. On one side, O2 binding to cytochrome c oxidase (COX) serves as a critical step in the function of the mitochondrial electron transport system (ETS) to support respiration and aerobic ATP production. On the other side, stray electrons from ETS redox centres can bind O2 to form reactive oxygen species (ROS; Hoffman and Brookes, 2009; Quinlan et al., 2013), which at low levels play a signaling role and are scavenged by cellular and mitochondrial antioxidant systems (Jones, 2006; Turrens, 2003). When ROS generation exceeds scavenging capacity however, ROS can accumulate and lead to oxidative damage (Ott et al., 2007; Sies, 1997). The relative rates of mitochondrial respiration and ROS generation by the ETS are dependent upon the redox environment (Aon et al., 2010), whereby mitochondrial ROS production is minimal in respiring mitochondria with a more oxidized redox environment. Mitochondrial ROS production increases as the ETS complexes become more reduced and the protonmotive force increases, as occurs in low O2 (hypoxia) when respiration is inhibited. As such, mitochondrial redox environment serves as the mechanistic link explaining the inverse relationship between mitochondrial respiration and ROS generation and underlies the redox imbalance that occurs in mammalian systems during pathological states such as those experienced during ischemia-reperfusion (Aon et al., 2010). However, much less is known about the natural variation in mitochondrial respiration and ROS metabolism across organisms that have evolved to thrive in environments that experience fluctuations in O2. The evolution of hypoxia tolerance necessitates numerous modifications to the O2 transport cascade that improves O2 extraction from the environment and delivery to mitochondria (Mandic et al., 2009; Mandic et al., 2013; Scott et al., 2011). Hypoxia tolerance is often associated with a higher mitochondrial respiratory capacity as observed in intraspecific comparisons among high and low altitude native deer mice (Peromyscus maniculatus; Mahalingam et al., 2017) or higher ETS capacity as observed in interspecific comparisons in elasmobranchs inhabiting different intertidal environments (Hemiscyllum ocellatum vs. Aptychoterma rostrate; Hickey et al., 2012). Related to greater respiratory and ETS capacity in hypoxia tolerant organisms is a generally lower ROS/O2 (Hickey et al., 2012; Mahalingam et al., 2017), which, if consistent across hypoxia tolerant species, may contribute to 47  explaining why mitochondria from hypoxia tolerant organisms tend to be more resistant to in vitro acute anoxia-reoxygenation stress and recover ADP-stimulated respiration to a greater extent after anoxia exposure than do hypoxia intolerant organisms (Hickey et al., 2012; Ivanina and Sokolova, 2016). The relative use of complexes I and II in supporting mitochondrial respiration can impact the efficiency of oxidative phosphorylation due to the higher ADP/O associated with complex I-fuelled respiration. Complexes I and II (along with other ETS sites) also contain important and well-studied sites of ROS generation but the magnitude of ROS generation from complex I can exceed that of complex II (Quinlan et al., 2013). Indeed, a laboratory-selected line of Drosophila melanogaster maintained at 4% O2 had approximately 30% lower complex II activity relative to a normoxic strain of flies, indicating a higher dependency on complex I concomitant with a reduction in superoxide generation (with complex I substrates; Yin et al. 2013; Ali et al. 2012). As such, there still is a general lack of understanding of the relationship between respiratory capacity and ROS/O2 across organisms that vary in hypoxia tolerance, also, the role of the redox environment in defining mitochondrial ROS generation as it relates to respiration has not been examined in a comparative context.  The goal of the present study was to assess the relationship between mitochondrial substrate preference, respiration, ROS/O2 and redox environment across species that vary in hypoxia tolerance. To examine this relationship, we conducted four sets of analyses: first, we determined whether there was a relationship between respiratory capacity and hypoxia tolerance; second, we investigated whether differences in mitochondrial capacity were consistent with variation in ROS emission; third, in order to investigate ROS emission kinetics under redox-controlled conditions, we manipulated extramitochondrial redox environment to compare ROS/O2 between species; and fourth, we compared the responses of mitochondria to in vitro anoxia-recovery. These experiments were conducted in mitochondria isolated from the brain of a well-characterized group of fish species, commonly called sculpins (Cottidae, Actinopterygii) that are distributed along the marine intertidal zone (Richards, 2011) and show interspecific variation in whole animal hypoxia tolerance (Mandic et al., 2013). Their ability to extract O2 from the environment as assessed by Pcrit, which is the PO2 at which O2 consumption rate transitions from oxy-regulating to oxy-conforming, which in turn is due to modifications to several steps in the O2 transport cascade that enhances O2 uptake and delivery to tissues (Mandic et al., 2009). Further, variation in hypoxia tolerance is also associated with modifications to the mitochondrion where cytochrome c oxidase (COX) of hypoxia tolerant 48  sculpins has an increased O2 binding affinity than COX from hypoxia intolerant sculpins (Lau et al., 2017). The clear functional variation in mitochondrial O2 binding among sculpins presents us with an opportunity to assess the relationship between mitochondrial respiration and ROS generation among multiple species that vary in hypoxia tolerance.   3.2 Methods 3.2.1 Species collection and holding  All species of sculpins used in this study were collected near Bamfield Marine Sciences Centre (British Columbia, Canada) at Ross Islets (48o52.4’N, 125o9.7’W) or Wizard’s Rock (48o51.5’N, 125o9.4’W) using either handheld dipnets or pole seines at lowest point in the daily tidal cycle. Animals were transported to The University of British Columbia (UBC) and housed in a recirculation system with artificial seawater (12°C) and maintained on a diet of shrimp, Atlantic krill, and bloodworms for at least 1 month before experimentation. All experimental procedures were reviewed and approved by the UBC Animal Care Committee (A13-0309).  3.2.2 Isolation of brain mitochondria Individual fish from each of the following species, Oligocottus maculosus (6.6 ± 0.3g), Artedius fenestralis (38.2 ± 5.3g), Artedius lateralis (34.7 ± 7.7g), Leptocottus armatus (91.3 ± 18.4g), Scorpaenichthys marmoratus (160.5 ± 40.0g), and Blepsias cirrhosus (30.0 ± 3.6g) were netted from their stock tank, weighed, stunned via concussion, euthanized by spinal severance, and the entire brain was quickly dissected for mitochondrial isolation. For fish <10g (i.e. O. maculosus), tissue from four animals were pooled to get sufficient yield for experiments, and for fish between 10 and 40g (e.g. A. lateralis), tissue from two animals were pooled. Each pooled sample is considered a single replicate for statistical analysis.  Briefly, brains were roughly chopped on an ice cooled surface with a razor blade in isolation buffer (250mM sucrose, 5mM Tris, 1mM EGTA, pH 7.4) and transferred into a Potter-Elvehjem tissue grinder and homogenized manually with 6-8 passes of polytetrafluoroethylene pestle. The tissue homogenate was then centrifuged at 1000g for 3min at 4oC and the supernatant was filtered through four layers of cheesecloth. The filtrate was then centrifuged at 10,000g for 10min at 4oC and the resulting supernatant was removed. The mitochondrial pellet was resuspended in cold isolation buffer and centrifuged again at 10,000g for 10min at 4oC to generate the final mitochondrial pellet, 49  which was resuspended in isolation buffer and kept on ice until analysis. Protein content of the mitochondrial suspension was determined with the Bradford’s assay (Sigma-Aldrich). An aliquot of the isolated mitochondria suspension was frozen in liquid N2 and stored at -80oC for the determination of mitochondrial ETS complex maximal activity.   3.2.3 Part I: Mitochondrial respiration Mitochondrial respiration was measured with an Oroboros oxygraph high-resolution respirometer (Innsbruch, Austria). The polarographic oxygen sensors were calibrated daily with air saturated and anoxic MiR05 buffer (in mM: 0.5 EGTA, 3 MgCl2·6H2O, 60 lactobionic acid, 20 taurine, 10 KH2PO4, 20 HEPES, 110 sucrose, and 1 g/L bovine serum albumin; pH 7.1; Gnaiger et al. 2000) at 18oC. We chose an assay temperature of 18°C to maximize signal to noise ratio in our respiration data, which is within the range of temperatures that sculpins are exposed to in the intertidal (Richards 2011).   To assess mitochondrial respiration, we employed a substrate utilization inhibitor titration (SUIT) protocol. Briefly, isolated mitochondria (~0.2 to 0.5mg mitochondrial protein) were first introduced into the respirometry chamber containing MiR05 buffer at 18°C. Following the introduction of mitochondria, complex I fuels were added (10mM pyruvate, 1mM malate, and 10mM glutamate; PMG) to yield state II (PMG) respiration rate. This was followed by addition of 0.75mM ADP to stimulate complex I-fueled state III respiration rate, then 10mM succinate was added to assess complex I and II fueled state III respiration rate (PMGS). State IV respiration was established with 2.5µM oligomycin. This was followed by titration of 2-3, 0.5µM steps of carbonyl cyanide 4-(trifluoromethyoxy)phenylhydrazone (FCCP) to fully uncouple the mitochondria. Once fully uncoupled, 0.5µM rotenone (Rot) was added to inhibit complex I, followed by addition of 2.5µM antimycin A to inhibit complex III.  Respiratory control ratio (RCR) was calculated as state III (both complex I and II fuelled) to state IV (oligomycin-induced) respiration rate. ADP/O was determined as the ratio of ADP phosphorylated divided by the O2 consumed during the transition from state III to IV.  3.2.4 Part II: Mitochondrial respiration and simultaneous ROS measurements To assess mitochondrial respiration simultaneously with ROS emission, mitochondrial respiration was assessed with steps described above and changes in ROS emission were assessed 50  fluorometrically using Amplex Ultrared (ex 568nm/em 581nm; Invitrogen) with the Oroboros respirometer and fluorometer. Isolated mitochondria (~0.5-0.6mg mitochondrial protein) were introduced into the chamber calibrated with MiR05 maintained at 18oC, followed by the addition of 5µM Amplex Ultrared, 0.004U horseradish peroxidase (HRP), and 13U superoxide dismutase. To calibrate the ROS detection methods, four additions of H2O2 were sequentially added to the chamber to a total of 0.56µM H2O2 while monitoring changes in the product of the reaction between H2O2 and Amplex Ultrared. In order to confirm that ROS was the primary reactive species being detected (Amplex Ultrared is known to be sensitive to several reactive species, including reactive sulfur species (DeLeon et al., 2016)), we added catalase to catalyse the conversion of H2O2 to O2 at the end of several experimental runs (after GSH titrations in Experimental Part III) which resulted in a 75 to 80% reduction in ROS/mg protein in all species examined, suggesting that the major reactive species detected in our experiments was H2O2. Thus, we have elected to refer to the reactive species measured in this study as ROS.  For this part of the study, isolated brain mitochondria were first incubated with 10mM pyruvate, 5mM malate (excess malate was used to ensure oxaloacetate removal to avoid inhibition of complex II; Hickey et al., 2012), and 10mM glutamate to establish complex I-fuelled state II respiration rate. This was followed by 10mM succinate to assess complex I and II fuelled state II, after which 16µM ADP was added to stimulate respiration to assess ADP/O. Once state IV respiration was obtained following the sub-maximal ADP addition, 0.75mM ADP was added to stimulate maximal state III respiration rate, followed by 2.5µM oligomycin to establish state IV respiration, and 0.5µM rotenone to inhibit complex I in state IV conditions.  3.2.5 Part III: In vitro redox challenge In order to assess the effects of redox environment on mitochondrial ROS emission, we aimed to manipulate the redox environment of the mitochondrial matrix by adjusting the extra-mitochondrial GSH:GSSG while measuring mitochondrial GSH:GSSG and ROS emission under state II conditions. Briefly, the oxygraph was set up as described in Part II with Amplex Ultrared. Once the mitochondrial sample was introduced into the chamber and the H2O2 calibration was complete, complexes I and II substrates (PMGS) were added to establish stable state II respiration and ROS emission rates, after which the contents of the oxygraph chamber were removed and centrifuged at 10528g for 5min. The resulting mitochondrial pellet was washed twice with MiR05, and the 51  mitochondrial pellet was frozen in liquid nitrogen and stored at -80oC until analyses of GSH and GSSG (see below). To adjust the extra-mitochondrial GSH:GSSG, the oxygraph was set up as described above, but after state II (PMGS) respiration was established, four injections of 0.6mM GSH were added to the chamber to a final concentration of 2.4mM GSH. State II respiration rate and ROS emission were monitored simultaneously with the oxygraphy as GSH was titrated. At the end of the titrations, the entire mitochondrial suspension was sampled and processed as described previously for analysis of GSH and GSSG. GSH was below the level of detection in the supernatant from the last MiR05 wash.  3.2.6 Part IV: Recovery from in vitro anoxia To assess potential interspecific differences in recovery from in vitro anoxia, the oxygraph was set up as described in Part II with Amplex Ultrared. Once the mitochondrial sample was introduced and H2O2 calibration was complete, complex I and II (PMGS) substrates were added followed by 0.75mM ADP to stimulate state III respiration. At this point, chamber O2 concentration was adjusted to 100µM by passing a stream of nitrogen gas over the surface of the mitochondrial suspension. The chamber was then sealed and mitochondria allowed to deplete chamber O2 to anoxia, at which point mitochondria were maintained in anoxia for 20min. The chamber was opened to reoxygenate with air, after which 0.5-1mM ADP was added to stimulate state III respiration rate. State IV was then induced with 2.5µM oligomycin, and finally 0.5µM rotenone was added to inhibit complex I.  3.2.7 Biochemical analyses  Mitochondrial GSH:GSSG  Mitochondrial GSH:GSSG was determined spectrophotometrically with the recycling method modified from Rahman et al. (2006) for isolated mitochondrial pellets. Briefly, frozen mitochondrial pellets were resuspended in 80µL buffer with 0.1M KH2PO4, 5mM EDTA disodium salt at pH7.5 (KPE buffer), with 0.1% triton X-100 and 0.6% sulfosalicylic acid and thawed in a room temperature water bath. The suspension was freeze-thawed twice and 50µL of the resulting suspension was used to determine glutathione disulfide (GSSG; oxidized glutathione) and the remainder of the sample for determining total glutathione (both reduced and oxidized forms). To determine GSSG, the GSSG samples and standards were incubated at room temperature with 1µL of vinylpyridine (diluted 1:10 v/v KPE) to derivatize endogenous GSH in the sample. After 1hr, 52  3µL of triethanolamine (diluted 1:6 v/v KPE) was added to the samples and standards and incubated for 10min at room temperature. This was followed by the addition of 3µL of 1M HCl to neutralize the sample. The GSSG and total glutathione samples were then assayed with the following protocol in which glutathione reductase (GR) converts GSSG into GSH. Briefly, equal volumes of 1.7mM [5,5’-dithio-bis(2-nitrobenzoic acid)] (DTNB) and glutathione reductase (3.33U/mL KPE) were mixed as the assay buffer, of which 125µL was added to 20µL of sample/standard. After 30sec to allow GR activity to convert GSSG into GSH, 60µL of 0.8mM ß-NADPH was used to start the reaction and the rate of TNB formation was monitored at 412nm for 5min. GSH was calculated as the difference between total glutathione and GSSG. The redox status in (mV) was calculated using a simplified Nernst equation for GSH:GSSG: Ehc = Eo + 30 log([GSSG]/[GSH]2) where Eo is -264mV at pH 7.4 (Garcia et al. 2010; Jones 2002).   Mitochondrial complex maximal activities (Vmax) on isolated mitochondria To determine the maximal activity of complexes I, II, and V, frozen samples of isolated mitochondrial were thawed on ice and centrifuged at 15000g for 10min at 4oC. The pellet was resuspended in hypotonic medium (25mM K2HPO4, 5mM MgCl2·6H2O at pH7.2) and freeze-thawed thrice before proceeding with assay protocol described in Galli et al. (2013). To determine the maximal activity of complexes III, the frozen mitochondrial sample was thawed on ice and freeze-thawed twice before proceeding with assay described in Galli et al. (2013).  3.2.8 Calculations and statistical analyses  Due to the different body weights of sculpin species used in this study, we first confirmed using regression analysis that there were no relationships between body weight and measurements made in Part I.  To determine if there were interspecific differences in respiration rates and flux capacities (Part I and II), we performed one-way ANOVA followed by Tukey’s multiple comparisons tests. Correlative analysis in Part I was performed using both ordinary least squares (OLS) and phylogenetically generalized least squares (PGLS) using ape (Paradis et al., 2004), Geiger (Pennell et al., 2014) and nlme (Pinheiro et al., 2014) packages in R (Team, 2016). We used the most updated phylogenetic tree of the marine species of Superfamily Cottoidei (Knope, 2013) for the phylogenetic analyses and dropped the tree tips for species without available data. Each set of regressions from our data set 53  was tested under OLS and PGLS (Pagel, 1999). Pagel’s λ of 0 indicated that the correlation was independent of phylogeny, whereas λ value of 1 is consistent with the constant-variance model (or Brownian motion model). The model with the lower Akaike Information Criterion (AIC) value represented the better fitting model (Table 3.2).  The slopes and intercepts of the simple linear regressions in Part III were compared between species using the student’s t-test. The pre- and post- anoxia differences in respiration states for Part IV were tested using student t-test.  3.3 Results 3.3.1 Part I: Interspecific relationship between hypoxia tolerance and complex I and II respiratory flux capacity The relative rate of complex I-fuelled ADP-stimulated respiration to ETS capacity differed between sculpins (one-way ANOVA; p=0.005) and was related to interspecific variation in Pcrit (OLS; p=0.05; Fig.3.1A, Table 3.2). Hypoxia tolerant sculpins with lower Pcrit had lower complex I-fuelled ADP stimulated respiration rates than less tolerant species (Fig.3.1A). The relative rate of complex I and II-fuelled ADP-stimulated respiration to ETS capacity did not differ between species (p=0.12) and was not associated with Pcrit (data not shown; p=0.45). In uncoupled mitochondria, there were differences in complex I flux capacity between species (p=0.0035), which were related to Pcrit (OLS; p=0.03; Fig.3.1B, Table3.2). There were no differences in complex II flux capacity between species (p=0.074), but there was a strong inverse interspecific relationship with Pcrit (OLS; p=0.002; Fig.3.1C, Table 3.2). Overall, species with lowest Pcrit (O. maculosus) had lower complex I flux capacity (65% of ETS capacity) and higher complex II flux capacity (29%) than species with highest Pcrit (B. cirrhosus), which had higher complex I flux capacity (86%) and lower complex II flux capacity (8%; c.f. Fig3.1B&C). These differences in complex flux capacities were not associated with differences in ADP/O (per mg mitochondrial protein), respiratory control ratio (RCR), or the maximal activity (Vmax) of complex I or II, all of which did not vary between species (Table 3.1). 3.3.2 Part II: Interspecific differences in ROS/O2 in state IV We propose that the interspecific differences in complex dependency could potentially drive differences in ROS accumulation from mitochondria. In light of this, we compared ROS metabolism in three species of sculpin: O. maculosus, which has lowest Pcrit among the species 54  investigated and the more hypoxia tolerant, with two other species A. lateralis and S. marmoratus both with higher Pcrit values and which are less hypoxia tolerant (though not different from one another).  Consistent with our previous findings in Part I (Supplementary Fig.3.1), there were no differences in state II respiration rates (PMG or PMGS) between the three sculpin species tested in this part of the study (Fig.3.2A). O. maculosus had lower state III respiration rate (Fig.3.2A; one-way ANOVA, p=0.018) and lower oligomycin-induced state IV respiration rate (p=0.042) than S. marmoratus. There were no differences in rotenone-inhibited state IV respiration rates between the three species. Simultaneous measurement of ROS emission rate expressed relative to O2 consumption rate (ROS/O2) showed no differences between species in all states of respiration, except under state IV conditions where O. maculosus showed higher ROS/O2 than the other two species (p=0.016; Fig.3.2B). There were no differences in ROS generated per mg mitochondrial protein in all states of respiration (Supplementary Table 3.1).  3.3.3 Part III: Effects of manipulating the redox environment on ROS/O2    To investigate the effects of redox environment on ROS generation, the glutathione redox environment was manipulated by titrating GSH (to reduce the assay buffer environment) to relate ROS/O2 to the measured mitochondrial GSH:GSSG. This part of the study was carried out with two species, O. maculosus and S. marmoratus, which differed in state IV ROS/O2 in Part II (Fig.3.2B). As the redox environment shifted towards a more reduced state with GSH titrations (a more negative GSH:GSSG in mV), both species increased ROS/O2 (Fig.3.3A) and ROS/mg protein (Supplementary Fig.3.1). The slopes of the relationship between mitochondrial GSH:GSSG and ROS/O2 were -0.0030±-0.00033 and -0.0016±0.00020 for O. maculosus and S. marmoratus, respectively, which were significantly different (Analysis of covariance; p=0.011), where the same concentration of extramitochondrial GSH reduced S. marmoratus mitochondrial GSH:GSSG more (more negative mV value) than in O. maculosus (by 11mV). Overall, these results indicate that O. maculosus produce more ROS/O2 than S. marmoratus across various mitochondrial GSH:GSSG.  3.3.4 Part IV: In vitro anoxia-reoxygenation exposure  We compared the responses of brain mitochondria from O. maculosus, A. lateralis, and S. marmoratus to 20min of in vitro anoxia followed by reoxygenation. Anoxia-reoxygenation significantly reduced state III respiration rate in O. maculosus and A. lateralis, but not in S. marmoratus (Fig.3.4A).  There was no effect of anoxia-reoxygenation on state IV or rotenone-inhibited state IV respiration rate in any 55  species examined (Fig.3.4A). The ROS/O2 under state III conditions was not affected by anoxia-reoxygenation and did not differ between species, whereas ROS/O2 in state IV was significantly reduced following anoxia-reoxygenation in S. marmoratus. Further, there were no differences in rotenone-inhibited ROS/O2 following anoxia-reoxygenation (Fig.3.4B). ROS/mg protein under state III conditions was not affected by anoxia-reoxygenation, but it was reduced in in O. maculosus and S. marmoratus under state IV conditions. Further, S. marmoratus reduced rotenone-inhibited ROS/mg protein (Fig.3.4C).  3.4 Discussion  The present study clearly demonstrates that interspecific variation in hypoxia tolerance among sculpins is related to how mitochondria from the brain utilized carbon substrates to support respiration. This variation, however, was not related to improving mitochondrial phosphorylation efficiency or reducing ROS emission. In fact, we show that the hypoxia tolerant O. maculosus generated more ROS/O2 under reducing state IV conditions than the less tolerant A. lateralis and S. marmoratus. O. maculosus was more capable of buffering mitochondrial redox environment under conditions where extra-mitochondrial redox was reduced, but when ROS/O2 was examined under similar mitochondrial GSH:GSSG conditions, brain mitochondria from O. maculosus had a higher ROS/O2 than S. marmoratus. Finally, unlike that previously shown in elasmobranchs (Hickey et al., 2012) and clams (Ivanina and Sokolova, 2016), the hypoxia tolerant O. maculosus experienced a reduction in state III respiration rate after in vivo anoxia-recovery, which could be a sign of mitochondrial damage, following anoxia-reoxygenation compare to S. marmoratus which showed no changes in respiration rate. Both O. maculosus and S. marmoratus showed significant reduction in ROS/mg protein in state IV, which in S. marmoratus was due to a reduction in ROS/O2. These outcomes suggest that hypoxia tolerance in sculpins is associated with a reduced reliance on complex I and higher mitochondrial ROS emission.    3.4.1 Hypoxia tolerant sculpins utilize less of complex I in overall ETS flux Among six species of sculpin, there were strong but opposite relationships between complex I and II flux capacities and Pcrit, where hypoxia tolerant species had reduced complex I and increased complex II dependency and intolerant species increased complex I and reduced complex II dependency. This variation suggests that these differences are associated with whole animal hypoxia tolerance (Fig.3.1). Complex I is a proton pump fuelled by NAD+-linked substrates and is composed 56  of both nuclear and mitochondrial-encoded subunits, whereas complex II is nuclear-encoded, fuelled by succinate and does not directly contribute to proton pumping. It is therefore not surprising that the selective pressures of routine hypoxia exposure in sculpins has resulted in variation in how these two complexes are used. However, it was unexpected that the hypoxia tolerant species would show reduced complex I and increased complex II dependency, as opposed to relying more on complex I which would result in a higher ADP/O and be consistent with increased efficiency of O2 use in generating the electrochemical gradient. In the present study, however, the interspecific shifts in complex flux capacity did not affect the measured phosphorylation efficiency (ADP/O) or mitochondrial coupling as assessed by RCR (Table 3.1). There are several possible explanations for why the reduced complex I flux capacity in the hypoxia tolerant sculpins does not result in lower phosphorylation efficiency.  First, the phosphorylation efficiency was determined in coupled mitochondria whereas complex I and II flux capacities are determined in uncoupled mitochondria, thus there is a possibility that the complexes are differential regulated in the coupled and uncoupled states, which may minimize the effects of differences in flux capacity on ADP/O between the species. This is supported by the fact that state III respiration rates (PMG or PMGS; Supplemental Fig. 3.1) were not significantly different among the species. Second, it is also possible that other ETS proton pumps (complex III and IV) compensate for the reduced complex I flux capacity in the hypoxia tolerant sculpins.  Complex III Vmax (Table 3.1 and antimycin-A inhibited respiration rate; Supplementary Fig. 3.1) did not vary among sculpins, but in our previous work, we demonstrated that complex IV (cytochrome c oxidase; COX)-specific respiration rate was higher, with a lower Km,app for O2 in hypoxia tolerant sculpins than intolerant sculpins (Lau et al., 2017), which may point to an increased role of the O2-binding ETS complex in the maintenance of ADP/O across the species of sculpins. The lack of an interspecific effect of variation in complex I flux capacity on ADP/O strongly suggests that the variation in complex I has little to do with its role as a proton pump, thus suggesting that there is another explanation for why hypoxia tolerant sculpins have a reduced complex I flux capacity compared with hypoxia intolerant sculpins.   While there are a number of sites along the ETS that generate ROS, complexes I sites are considered to be the major ROS generation sites (Quinlan et al., 2011; Quinlan et al., 2013; Treberg et al., 2011). It is possible that the lower dependency on complex I reduces ROS generation. Additionally, even though the present study is the first to show putatively adaptive variation in ETS substrate 57  dependency across species that vary in hypoxia tolerance, previous studies have shown that hypoxia exposure differentially affects complex I and II flux rates. For example, in murine brain mitochondria, complex I and II were inhibited to different degrees in response to hypoxia (complex II-fuelled state III was lowered by 25% compared to complex I-fuelled state III lowered more by 32-46%), and upon reoxygenation complex II-linked state III respiration recovered faster than complex I-linked state III respiration (Almeida et al., 1995; Gilland et al., 1998). Similarly, permeabilized heart fibres from cold anoxia acclimated turtles lowered complex I-IV flux, but not complex II-IV flux compared with normoxia-acclimated turtles (Galli et al., 2013). Thus, complexes I and II may be regulated differently in response to hypoxia exposure with complex I flux being inhibited to a greater extent and slower to recover than complex II. These acute responses to hypoxia are consistent with the innate interspecific differences observed in sculpins, where hypoxia tolerance is associated with reduced complex I and increased complex II dependency.  3.4.2 Hypoxia tolerant sculpins emit higher ROS/O2 with lower state IV respiration rate The marine intertidal zone is heavily influenced by the daily ebb and flow of the tide, which results in a strong gradient of spatial and temporal oscillations in abiotic factors including O2. Isolated tidepools from the upper intertidal zone experience greater fluctuations in O2 (from hypoxic to hyperoxic levels), more so than the homogenous subtidal environments (Richards 2011). In hypoxia tolerant sculpins, enhanced steps of the sculpin O2 transport cascade (lower whole animal O2 consumption rate, higher gill surface area, lower Hb-P50, lower mitochondrial and COX Km,app O2; Lau et al., 2017; Mandic et al., 2009; Mandic et al., 2013) would presumably increase O2 extraction and delivery to tissues, and illustrates the importance of efficient O2 use in determining whole animal hypoxia tolerance. Mechanisms to increase the efficiency of O2 use would presumably also minimize ROS generation in order to restrict the use of already limited O2 towards the generation of a potentially harmful by-product. We thus investigated whether hypoxia tolerant sculpins would have lower ROS/O2, indicating that less O2 is reacting with electrons to form ROS.  We chose three species that varied in Pcrit and complex dependency in Part I to investigate the relationship between respiration and mitochondrial ROS emission. Among the three species, O. maculosus had a significantly lower state IV respiration rate, which is consistent with lower futile proton cycling and thus better mitochondrial coupling (although our six species comparison in part I showed no relationship between Pcrit and RCR). Increased mitochondrial coupling brings about a 58  higher proton gradient which is typically associated with higher ROS generation (Brand, 2000; Brand and Esteves, 2005), and indeed, O. maculosus had higher ROS/O2 in state IV compared to S. marmoratus. This pattern of higher ROS/O2 with lower leak respiration only emerged under state IV conditions (with high protonmotive force) rather than in state II, which indicates intrinsic differences in the capacity to generate ROS between sculpin species.  Although we hypothesized that variation in complex I flux capacity would be associated with reduced ROS emission, the interspecies differences in ROS/O2 was not observed in rotenone-inhibited state IV (Fig.3.2B). Under state IV conditions (high protonmotive force, high reduced redox environment) and in the presence of rotenone and complex I and II substrates, the ROS generated is likely from the quinone-binding IQ site on complex I (Quinlan et al., 2013). A more detailed study would be required to also assess the ROS generation capacity from the complex I NADH-binding flavin IF site.  Overall, ROS levels detected in sculpins are comparable to other studies. In state II (PMGS), O. maculosus and S. marmoratus emitted 2pmol H2O2/s/mg protein (Supplementary Fig.3.2), and similar levels were measured in isolated killifish Fundulus heteroclitus liver mitochondria in leak state (Du et al., 2016). Similar ranges of values were also observed in mammalian mitochondria (rat skeletal muscle and guinea pig heart mitochondria; Aon et al., 2010; Munro et al., 2016). Relative to respiration rate (2pmol H2O2/s/mg protein and simultaneous 0.09nmol O2/s/mg protein), sculpin H2O2 emission rates are consistent with the estimation that about 1-2% of O2 used by mitochondria is emitted as ROS (Turrens, 2003).  3.4.3 Hypoxia tolerant sculpins buffer mitochondrial redox changes better but generate higher ROS/O2    Mitochondrial ROS generation is highly affected by mitochondrial energy status and the redox environment (Aon et al., 2010; Munro and Treberg, 2017), both of which can be profoundly affected by the availability of O2. Hypoxia results in a reduction in ETS flux and leads to a highly reduced redox environment, whereby excess electrons in the ETS complex redox centres favor ROS generation even though O2 levels are low. Hyperoxia can lead to an oxidized mitochondrial redox environment due to high O2 availability, also favoring ROS generation. At either extreme of the redox environment, the combined effects of ROS generation and scavenging capacity can result in higher net ROS emission (Aon et al., 2010). We hypothesized that the more hypoxia tolerant sculpin 59  would lower ROS emission under a stressful redox environment when compared to a less tolerant species.  To manipulate the mitochondrial redox environment, we altered the GSH pool as it is a major redox couple within the cell and is also easily assessed by monitoring concentrations of GSH and GSSG (Rahman et al., 2006; Schafer and Buettner, 2001). Mitochondria contains two GSH pools, the intermembrane space (IMS) and matrix pools and although the communication between these two pools is not understood (e.g. Kojer et al., 2012; Mari et al., 2013), it is generally recognized that mitochondria are not capable of synthesizing GSH, thus there must be a transport mechanism between the IMS and the matrix. Furthermore, we were effective at changing the combined mitochondrial GSH:GSSG by altering the extra-mitochondrial GSH:GSSG pool. Under conditions where we were able to adjust extra-mitochondrial GSH:GSSG, we predicted that the more hypoxia tolerant sculpin would (1) be able to maintain mitochondrial redox status with changes in extra-mitochondrial GSH:GSSG potentially due to better scavenging capacities, and (2) accumulate lower ROS/O2 as GSH:GSSG became more reduced.  Indeed, mitochondrial redox environment was better buffered in the hypoxia tolerant O. maculosus (Fig.3.3) compared with S. marmoratus during titration of extra-mitochondrial GSH, but after each titration the same relative increase in ROS/O2 were observed in both species. Further, when ROS/O2 was examined at a single mitochondrial GSH:GSSG, it was consistently higher in O. maculosus than in S. marmoratus. O. maculosus experience more variable PO2 in their environment (anoxia to 400% air saturation), which may translate to their mitochondria also experiencing large fluctuations in PO2 and possibly variable redox conditions through the daily tidal cycle. Resistance to changes in the mitochondrial redox environment would ensure that the various redox-linked processes, such as thioredoxin/peroxiredoxin which are major antioxidant mechanisms (Drechsel and Patel, 2010) within the mitochondrial matrix are left unperturbed. One mechanism that would result in increased redox buffering capacity could be a higher NADPH/NADP+, another redox couple that donates electrons to aid the recycling of GSSG back to GSH to maintain glutathione redox status (Rahman et al., 2006). This ability to maintain matrix GSH:GSSG, however, did not result in minimizing ROS emission, as indicated by the higher ROS/O2 in O. maculosus compared to S. marmoratus at the same matrix GSH:GSSG. This difference suggests possible differences in ROS generation capacity from the ETS sites and/or differences in scavenging capacities in the matrix.  60  3.4.4 More hypoxia tolerant sculpins do not recover better from in vitro anoxia Injury is sustained from ischemia-reperfusion due to surge of ROS generated in normoxic recovery following a period of O2 deprivation. The oxidative damage caused by the transition from anoxia/hypoxia to normoxia results in incomplete recovery of respiration in mammalian mitochondria (Almeida et al., 1995; Du et al., 1998; Shiva et al., 2007). In fact, the interspecies difference in recovery from 20min of in vitro anoxia showed that hypoxia tolerant O. maculosus brain mitochondria aligned more with typical mammalian response in O2 sensitive tissues, and a complete recovery of state III respiration was observed in less tolerant S. marmoratus (Fig.3.4). While both O. maculosus and S. marmoratus significantly reduced state IV ROS/mg protein, only S. marmoratus significantly reduced state IV ROS/O2. This indicates that both tolerant and intolerant species reduced ROS generation, but the tolerant O. maculosus achieved this via an overall reduction in respiration capacity (as indicated by the reduction in state III respiration rate), and the less tolerant S. marmoratus achieved this by altering ROS generation capacity (ROS/O2). Further, this lower ROS generation capacity in S. marmoratus appears to be in part mediated by changes in ROS generated from the IQ site as evident by the reduced rotenone-induced ROS/mg protein.  3.4.5 A potential role for ROS in response to environmental O2 variability Our study in sculpin mitochondria paints a different picture from previous studies showing that hypoxia tolerant animals generally lower mitochondrial ROS emission. Our interspecies comparison shows that, in fact, hypoxia tolerance may be associated with higher ROS generation under state IV conditions (Fig.3.2B), higher sensitivity of ROS generation to changes in mitochondrial GSH:GSSG but better ability to maintain matrix redox status (Fig.3.3), and reduced ability to recover from in vitro anoxia (Fig.3.4). While more ROS accumulation is often associated with oxidative damage, perhaps an increase in ROS levels may potentially be an important response to O2 variability. ROS has been shown to act as an agonists to covalently modify proteins (such as iron-sulfur clusters) which can be part of specific signaling processes mediating cellular responses to oxidative damage (D’Autréaux and Toledano, 2007). For instance, the Kreb’s cycle enzyme aconitase is targeted by reactive superoxide to reduce Kreb’s cycle flux (Powell and Jackson, 2003). ROS generated from complex III has also been shown to play a role in stabilizing the transcription factor hypoxia inducible factor 1 (HIF-1) during hypoxia, which subsequently acts on downstream targets to coordinate cellular responses to hypoxia (Chandel et al., 2000; Guzy et al., 2005; Klimova and Chandel, 2008). There is also an emerging view that mitochondria are not only generators of ROS, but also act as regulators 61  to maintain steady state ROS levels for the purpose of cell signaling (Munro and Treberg, 2017; Starkov, 2008). It is possible that an increase in ROS generation with relatively small changes in mitochondrial redox status is a part of the strategy of hypoxia tolerant sculpins to deal with variations in O2 in the intertidal zone.  3.5 Summary  In animals that frequently encounter hypoxia in their natural environment, it has been observed that there are putative adaptations in the O2 transport cascade presumably to maintain O2 supply to mitochondria to maintain energy output from oxidative phosphorylation even at low environmental PO2. Changes in ETS flux can cause a rise in ROS emission from mitochondria, which when accumulated can result in oxidative damage. Although it was previously proposed that hypoxia tolerant animals or animals that are frequently exposed to O2 limitations would be able to reduce ROS generation, in fact, in the sculpin model this was not what we observed. In this study, we first provided evidence that there was a difference in reliance of carbon substrates between species of varying hypoxia tolerance, with more hypoxia tolerant species showing lower complex I and higher complex II dependency compared to less tolerant species. We then demonstrated that brain mitochondria from hypoxia tolerant species of sculpin in fact generated more ROS under state IV conditions and with changes in extramitochondrial redox environment. We also observed that species which vary in hypoxia tolerance have different strategies of reducing ROS/mg protein after in vitro anoxia exposure. It is possible that higher mitochondrial ROS emission has a role in determining hypoxia tolerance in intertidal sculpins, but this will require further investigation.  62  Figure 3.1. Relationship between whole animal hypoxia tolerance as assessed by critical oxygen tensions (Pcrit; Mandic et al. 2009) and mitochondrial respiration rates (A) under ADP-stimulated respiration state with complex I substrates (10mM pyruvate, 1mM malate, and 10mM glutamate; expressed to FCCP-fully uncoupled respiration rate), (B) complex I flux capacity (with pyruvate, malate, and glutamate) determined as the rotenone-sensitive respiration rate in uncoupled mitochondria (expressed to FCCP-uncoupled respiration rate), and (C) complex II flux capacity (PMG with 10mM succinate; determined as the antimycin A-sensitive respiration rate in the presence of rotenone in uncoupled mitochondria; expressed to FCCP-uncoupled respiration rate). Data are means ± standard error. Species indicated in the figure are as follows: (1) Oligocottus maculosus, (2) Artedius fenestralis, (3) Artedius lateralis, (4) Leptocottus armatus, (5) Scorpaenichthys marmoratus, and (6) Blepsias cirrhosus.  63    Figure 3.1 64   Figure 3.2. Substrate-inhibitor titration protocol with simultaneous measurement of (A) oxygen consumption rate and (B) reactive oxygen species generated (expressed as ROS/O2) in isolated brain mitochondria in three sculpin species (from left to right: O. maculosus (species 1 in Fig.3.1) in squares, A. lateralis (species 3) in triangles, and S. marmoratus (species 5) in circles). State II PMG (with complex I fuels pyruvate, malate, and glutamate), state II PMGS (addition of complex II fuel succinate), state III (ADP maximally stimulated), state IV (oligomycin-induced), Rotenone (inhibition of complex I). Within each respiration state, symbols with different letters are significantly different (P<0.05 one-way ANOVA with Bonferroni multiple comparison corrections). 65    Figure 3.3. (A) Relationship between mitochondrial pellet GSH:GSSG and ROS/O2 in O. maculosus (species 1; squares and solid lines; y=-0.0030x-0.95) and S. marmoratus (species 5; circles and dotted lines; y=-0.0016x-0.48). Mitochondrial GSH:GSSG was manipulated by varying extramitochondrial GSH:GSSG with four steps to a total of 2.4mM GSH (to make more reduced, GSH:GSSG more negative). All four GSH doses are shown but only first and last doses (solid symbols) had mitochondrial pellet GSH:GSSG measurements, whereas the middle doses (hollow symbols) are estimates.   66  Figure 3.4. The effect of 20min in vitro anoxia-recovery on O2 consumption and ROS emission in isolated brain mitochondria from three sculpin species (from left to right: O. maculosus in squares, A. lateralis in triangles, and S. marmoratus in circles) with measurements of (A) O2 consumption, (B) ROS/O2 measurement, and (C) ROS/mg protein of state III, IV, and rotenone addition to state IV (expressed relative to values in normoxic respiration states). Asterisks indicate significance from multiple student’s t-tests with Holm-Sidak method for multiple comparison corrections.    67    Figure 3.4  68   Species  ADP/O  RCR (State III/IV) Mitochondrial complex maximal activities (Vmax; nmol/min/mg mitochondrial protein) I II III V Oligocottus maculosus 5.0 ± 0.7ab 5.9 ± 0.3d 34.2 ± 3.1ef 14.2 ± 1.1g 25.5 ± 5.3h 135.2 ± 25.1j Artedius fenetralis 4.3 ± 0.8 ab 4.6 ± 0.2d 42.2 ± 3.9e 20.1 ± 2.5g 17.6 ± 2.7h 174.0 ± 29.6j Artedius lateralis 5.2 ± 0.5 ab 5.7 ± 0.3d 23.6 ± 2.1f 16.6 ± 1.3g 23.0 ± 1.9h 72.7 ± 31.2j Leptocottus armatus 5.4 ± 0.9 a  5.5 ± 0.4d 42.9 ± 3.2ef 17.6 ± 2.3g 25.4 ± 3.1h 165.7 ± 14.2j Scorpaenichthys marmoratus 2.4 ± 0.1 b 6.3 ± 0.3d     Blepsias cirrhosus 4.6 ± 0.8 ab 5.2 ± 0.3d 39.9 ± 2.3ef 16.5 ± 1.9g 48.9 ± 7.3i 140.4 ± 2.4j  Table 3.1. ADP/O, respiratory control ratios (RCR; state III/state IV respiration rates), and mitochondrial complex maximal activities (Vmax of complexes I, II, III, and V; complex IV Vmax is published in Lau et al. 2017; nmole/min/mg mitochondrial protein) from five species of sculpins. Data are means ± standard error. Letters indicate the results from Tukey’s multiple comparisons test following one-way ANOVA. Mitochondrial complex Vmax for S. marmoratus is not available.    69  Regression (x,y) Test (OLS/PGLS) Slope estimate Slope p value AIC Pagel’s λ Complex I fueled ADP-stimulated respiration rate, Pcrit OLS PGLS 0.058 0.057 0.046 0.055 12.91* 22.50  0.31 FCCP-Rot/FCCP (Complex I flux capacity), Pcrit OLS PGLS 0.074 0.080 0.032 0.017 11.85* 21.33  -0.60 Rot-Ama/FCCP (Complex II flux capacity), Pcrit OLS PGLS -0.11 -0.11 0.0016 0.0007 3.12* 14.97  -0.78 Table 3.2. Regressions were tested with ordinary least squares (OLS) and phylogenetic generalized least squares (PGLS) with Pagel’s model of evolution (Pagel 1999). Pagel’s λ is an indication of whether the phylogeny correctly predicts patterns of covariance among species for a given trait. λ value of 0 indicates that correlation is independent of phylogeny, whereas value of 1 indicates that correlation is consistent with constant-variance model (or Brownian motion model) being a good representation of the data. Model with the lowest AIC value represents the best fitting model (indicated with an asterisk).   70  Chapter Four: Whole animal responses of ROS metabolism to hypoxia- and hyperoxia- recovery   4.1 Introduction  Changes in environmental O2 elicit not only changes in aerobic metabolism, but also in reactive oxygen species (ROS) metabolism. ROS can be generated from multiple sites along the mitochondrial electron transport system (ETS; Quinlan et al., 2013) under conditions of high protonmotive force where ETS complexes are in a reduced redox state and capable of donating electrons directly to O2.  Mitochondria continuously produces ROS at low rates, which have cellular signaling functions (D’Autréaux and Toledano, 2007) and are scavenged by antioxidant mechanisms. At high rates of production, ROS can overwhelm cell or mitochondria scavenging capacities,  and then accumulate and cause oxidative damage to DNA, proteins, and lipids (Cadet, 2003; Gutteridge, 1995; Reznick and Packer, 1994). When O2 becomes limiting, the disruption of electron flux through the ETS can result in an increase in ROS generation. Animals that are sensitive to changes in environmental O2 (i.e. the majority of mammalian species), are sensitive because of devastating effects of low O2 (hypoxia) on cellular energy status, but also because rapid changes in O2 such as those that occur during reoxygenation after hypoxia can induce significant tissue injury due to the effects of ROS (i.e. ischemia-reperfusion injury; Chouchani et al., 2014).  There are animals that inhabit and thrive in environments that are prone to changes in O2, which have potentially evolved strategies to minimize ROS production or mitigate the damaging effects of ROS. From the literature in this area, however, there does not appear to be a consistent response within ROS metabolism to changes in O2, particularly hypoxia, and there are considerable tissue- and species-specific differences (Leveelahti et al., 2014). For example, several studies looking at mitochondrial ROS metabolism suggest that the evolution of hypoxia tolerance was associated with  lower mitochondrial ROS generation (elasmobranchs, molluscs, and hypoxia-acclimated killifish Fundulus heterolitus; Du et al., 2016; Hickey et al., 2012; Ivanina and Sokolova, 2016) as typically expected; however, our recent study using isolated mitochondria from multiple species of intertidal sculpins (Cottidae, Actinpterygii) demonstrated the opposite where the more hypoxia tolerant species had higher ROS/O2 under state IV conditions and also under highly reduced glutathione redox status when compared to less tolerant species (Chapter 3). These studies suggest that even at the mitochondrial level the relationship between hypoxia tolerance and ROS metabolism is nuanced 71  and not straightforward, possibly due to variation in the role of ROS in cell signaling and the role of mitochondria as regulators of ROS metabolism in vivo (Munro and Treberg, 2017). In addition, it is unclear whether studies of ROS metabolism conducted in vitro at the level of mitochondria apply to in vivo responses during whole animal exposure to O2 stress.  The goal of this present study was to investigate whether there are interspecific differences in aspects of in vivo ROS metabolism in response to O2 variation as it relates to whole animal hypoxia tolerance. In order to assess in vivo relationships between redox environment and ROS emission, and to extend our interspecific investigation to in vivo cellular damage and ROS scavenging, we exposed two species of sculpin, Oligocottus maculosus and Scorpaenichthys marmoratus, that differ in hypoxia tolerance and mitochondrial function (Chapter 3) to hypoxia (2.3kPa and 3.5kPa), hypoxia followed by reoxygenation, and hyperoxia (64kPa). Sculpin are distributed along the marine intertidal zones where depending on location, the species experience different fluctuation in O2 due to the influence of the tidal cycle, with sculpins (e.g. O. maculosus) located higher in the intertidal experiencing greater O2 fluctuations than those in the more homogenous subtidal (e.g. S. marmoratus). We chose a 6hr hypoxia and hyperoxia exposure duration followed by a quick normoxia recovery to mimic exposures patterns that occur in the higher intertidal environment (Richards, 2011). Following exposure to hypoxia, hyperoxia, and normoxic recovery, aspects of in vivo ROS metabolism were assessed in liver, gill, and brain. Tissue redox status was assessed as the relative ratio of reduced and oxidized glutathione, which is a dominant redox couple within the cell (Rahman et al., 2006). For this study, we used the ratiometric mitochondrial-targeted mass spectrometry probe MitoB to measure changes in H2O2, which is boronic acid conjugated to a tetraphenylphosphonium ion (TPP+; Logan et al., 2014). Once collected within the mitochondrial matrix, it reacts with H2O2 to form a stable product, MitoP. We quantified oxidative damage as lipid peroxidation, which was measured as thiobarbuturic acid reactive substances (TBARS) levels (Gutteridge, 1995). We also assessed tissue total oxidative scavenging capacity (TOSC).  4.2 Methods  4.2.1 Equipment MitoB compounds used for this study were synthesized by Dr. Richard Hartley (University of Glasgow). Chemicals were purchased from Sigma-Aldrich unless otherwise specified.  72  4.2.2 Animals O. muculosus (average 4g) and S. marmoratus (average 16g) were collected near Bamfield Marine Sciences Centre (British Columbia, Canada) at Ross Islets (48o52.4’N, 125o9.7’W) and Wizard’s Rock (48o51.5’N, 125o9.4’W) using either handheld nets or pole seines at the lowest tidal cycle. Animals were transported to the University of British Columbia (UBC) and housed in a recirculation system with artificial seawater at 12°C and maintained on a diet of shrimp, Atlantic krill, and bloodworms for at least 3 weeks before experimentation. All experimental procedures were reviewed and approved by the UBC Animal Care Committee under animal use protocol number A13-0309.  4.2.3 Validation of MitoB use in sculpins Individual O. maculosus were retrieved from the stock tank and injected intraperitoneally with 50µL of MitoB (2nmol/g fish in phosphate-buffered saline) using a BD Ultra-Fine™ syringe (6mm needle). Following injection, fish were transferred to separate plastic mesh baskets (with three fish each) and held in a wet table with recirculating seawater at 12oC. Three fish at each timepoint (0.5, 1, 4, 8, 24, and 72hrs post-injection) were sampled. Animals were euthanized with 0.5g/L benzocaine. Gills, brain, muscle, and liver tissues were sampled at each of the timepoints, frozen in liquid nitrogen and stored at -80oC until processing (see below).  4.2.4 Experimental protocol and sampling  Both species of sculpin were exposed to normoxia, hypoxia and hypoxia-recovery, and hyperoxia and hyperoxia-recovery to assess the effects of different PO2’s on ROS metabolism. For hypoxia treatment, the less hypoxia tolerant species S. marmoratus was exposed to 3.5kPa, which is equivalent to 65% of the species’ Pcrit value (Mandic et al., 2009), whereas the more hypoxia tolerant species was exposed to two different hypoxia PO2 treatments of 2.3 and 3.5kPa, corresponding to 65% Pcrit and at Pcrit respectively. These hypoxic PO2 levels were chosen to facilitate the interspecies comparison at a common PO2 exposure (3.5kPa) and also at a hypoxic exposure that is relative to each species’ Pcrit (65% of Pcrit), which has been shown in other species to control for the level of hypoxemia (Speers-Roesch et al., 2013). For hyperoxia treatment, both species were exposed to 64kPa, which corresponds to a level of hyperoxia that is frequently encountered by O. maculosus in the higher intertidal (Richards 2011). Both species were also held in normoxia (21kPa) and normoxic control treatments which were run in parallel with each of the hypoxia and hyperoxia treatments. The treatments were set up by placing two covered aquaria (one for normoxic control animals and 73  the other for O2 treatments) on a wet table with recirculating water chilled to 12oC. Small circulating water pumps were placed inside each aquarium to ensure adequate mixing throughout the experiment.  To initiate the experiment, sculpins of each species were taken from their stock tank and injected intraperitoneally with MitoB (2nmol/g fish) and placed into 1.1L plastic mesh baskets with gravel on the bottom (4-6 of O. maculosus and 2-3 of S. marmoratus in each basket) held within each of the treatment aquaria. The PO2 was then adjusted to the desired level with either nitrogen gas (for hypoxia), 100% O2 (for hyperoxia), or aerated with compressed air (normoxia) and maintained at these levels for 6hrs. Oxygen levels were monitored with a hand-held O2 probe (Oakton; for hypoxia) or a FOXY Ocean Optics fluorescent O2 probe (for hyperoxia). At 6hr, a total of 12 individuals of O. maculosus and 8 individuals of S. marmoratus were sampled as described below. Following sampling at 6hrs, the hypoxic aquaria at 2.3kPa or 3.5kPa were then quickly returned to normoxia (within 30min) by aeration and a total of 12 individuals of O. maculosus and 8 individuals of S. marmoratus were sampled at 1hr recovery. In order to account for excretion rates of MitoB and MitoP, paired normoxic controls were also sampled at the same time to compare to O2 treatment samples. To sample fish, the mesh baskets were removed from the treatment aquaria and placed in a 3.7L container with anaesthetic (0.5g/L benzocaine). Once the fish were unresponsive to touch, they were removed from the anaesthetic bath and brain, gill, and liver were dissected and frozen in liquid N2 and stored at -80oC until further analyses.   4.2.5 Purification of tissue samples for MitoP/MitoB Briefly, tissues were homogenized in 200µL ice-cold 60% acetonitrile/0.1% formic acid in the bullet blender bead homogenizer (Next Advance). Homogenates were then centrifuged at 16000g for 10min at 4oC, after which after which the supernatant was transferred to a new vial.  The tissue was resuspended in 200µL 60% acetonitrile/0.1% formic acid and homogenized and centrifuged as described above. The resulting supernatant was combined with the previous supernatant and 10µL of internal standard (10µM d15-MitoB/5µM d15-MitoP) was added to each sample, vortexed for 10sec, and incubated for 30min at 4oC. Samples were then centrifuged at 16000g for 10min, and the supernatant was filtered with the Millipore centrifugal filter plate (0.45µm hydrophilic, low protein binding Durapore membrane; centrifuged at 3000g for 10min). The filtrate was collected and the samples were dried in a vacuum speed centrifuge (Labconco Centrivap Concentrator). The dried 74  sample was then resuspended in 250µL 20% acetonitrile/0.1% formic acid, vortexed for 5min to resuspend and centrifuged at 16000g for 10min. 200µL was used for LC-MS/MS analysis. Individual sample variation was normalized to deuterated internal standards. Standard curves for MitoB (0 to 1000pmol) and MitoP (0 to 1000pmol) were prepared with salmon tissue. Salmon tissue were homogenized in 60% acetonitrile/0.1% formic acid as described above. After two centrifugation steps, 10µL of internal standard and 10µL of standard was added to the supernatant. The standards were then purified with the same procedure as described above (standard curves in Supplementary Fig.4.1).  4.2.6 Tissue glutathione redox status (GSH:GSSG) Reduced (GSH) and oxidized (GSSG) glutathione were assayed with the enzymatic recycling method described in Rahman et al. (2006). Briefly, tissues were homogenized in 80µL buffer with 0.1M KH2PO4 and 5mM EDTA at pH7.5 (KPE buffer) with 0.1% triton X-100 and 0.6% sulfosalicylic acid. The homogenized sample was then centrifuged at 8000g for 10min at 4oC. The supernatant was then divided for assessment of total GSH and GSSG. To determine GSSG, 50µL of the samples or GSSG standards were incubated for 1hr at room temperature with 1µL vinylpyridine (1:10 v/v in KPE buffer) to derivatize GSH. After 1 hr, 3µL triethanolamine (1:6 in KPE buffer) was added to the samples/standards and incubated for 10min at room temperature, which was followed by the addition of 3µL 1M HCl to neutralize the sample. The GSSG and total GSH samples were then assayed using the same protocol in which glutathione reductase (GR) converts GSSG into GSH. For the assay, the buffer was prepared with equal volumes of 1.7mM [5,5’-dithio-bis(2-nitrobenzoic acid)] (DTNB) and glutathione reductase (3.33U/mL KPE), of which 125µL is added to 20µL of standard/sample. 60µL of 0.8mM ß-NADPH was added to start the reaction, and the rate of TNB formation was monitored at 412nm for 5min. GSH was calculated as the difference between total glutathione and GSSG. The redox status in (mV) was calculated using a simplified Nernst equation for GSH:GSSG: Ehc = Eo + 30 log([GSSG]/[GSH]2) where Eo is -264mV at pH 7.4 (Garcia et al., 2010; Jones, 2002).   4.2.7 Thiobarbuturic acid reactive substances (TBARS) TBARS was assessed using a commercially available kit (TBARS ParameterTM kit; R&D systems). Tissues were homogenized in homogenization buffer (in mM at pH7.75: 100 TrisHCl, 2 EDTA, and 5 MgCl2·6H2O) after which the suspension was centrifuged at 1600g for 10min at 4oC. The 75  supernatant was divided into aliquots for the total scavenging oxidative capacity assay (TOSC; described below), for the determination of protein content using the Bradford’s method (Bradford, 1976), and the remaining supernatant was used for the determination of TBARS. Briefly, equal volumes of sample and acid reagent were combined and incubated for 15min at room temperature. The sample was then centrifuged twice at 12000g for 4min at 4oC, and the supernatant was used for the assay. TBARS standards were prepared according to manufacturer instructions. The samples and standards were then incubated with TBA reagent (2:1 v/v) for 3hrs at 50oC, and absorbance was measured at 532nm. TBARS levels were normalized to tissue protein concentration.  4.2.8 Total oxidative scavenging capacity (TOSC) TOSC was determined as the overall activity of H2O2 removal, where H2O2 in the sample was monitored with Amplex Ultrared (Invitrogen). 100µL of assay buffer (with 0.1mM Amplex Ultrared, 1U/mL horseradish peroxidase in 50mM sodium citrate at pH6.0) was added to 50µL catalase standard or sample (homogenate prepared for the TBARS assay) in a spectrophotometer plate and pre-read at 565nm. 50µL of 160µM H2O2 was then added to each well and the endpoint absorbance was measured after 5min. The final TOSC was expressed as units of catalase activity (catalase standard curve was generated (from 0 to 45 units of activity) with an exponential fit). TOSC was normalized to tissue protein concentration.  4.2.9 Statistical analyses In order to account for MitoP and MitoB decay in tissues over time due to excretion, each timepoint had to be compared to a control sampled at the same time. The normoxic controls sampled at 6hrs and after an additional 1hr of recovery were not significantly different, and thus were grouped into a single normoxic sample to compare with O2 treatment and recovery samples.  In order to compare interspecies response to the different O2 treatments, two-way ANOVA with Holm-Sidak’s multiple comparison correction was performed to compare the effect of species (O. maculosus and S. marmoratus), treatment (normoxia, 6hrs hypoxia/hyperoxia treatment, or recovery from hypoxia/hyperoxia) for each measure (MitoP/MitoB, GSH:GSSG, TBARS, and TOSC). TBARS and TOSC values were normalized to normoxic control values.  76  4.3 Results  4.3.1 Validation of MitoB and MitoP in marine sculpins  As this is the first study to use MitoB in marine fish, we first validated its use in O. maculosus by examining tissue distribution and changes in concentrations of MitoB and the product MitoP over 72hr following injection of MitoB, which is similar to the validation approach used by Salin et al. (2017) for work in brown trout Salmo trutta. MitoB was reliably detected within 30min of injection in liver, brain, and gill, but the extent of uptake varied between tissues (in pmol/mg tissue: gill 1.42 ± 0.23, brain 0.38 ± 0.08, liver 13.71 ± 3.78; Supplementary Fig.4.2). MitoP was also detected after 30 min in the liver, brain and gill, but at lower concentrations (in pmol/mg tissue: gill 0.51 ± 0.07, brain 0.13 ± 0.06, liver 4.64 ± 1.75). In white muscle, MitoB and MitoP levels peaked at 1hr after injection, but tissue concentrations were much lower compared with the other tissues sampled (highest levels at 1hr of 0.09 MitoB and 0.06 MitoP pmol/mg tissue), so white muscle was not analysed for this study. Both MitoB and MitoP decreased exponentially over the 72hr time course in liver, brain, and gill, but we were able to accurately determine the MitoP/MitoB for the full 8hr required for treatment and normoxia exposures for our study.  4.3.2 Hypoxia and hypoxia-recovery at a common PO2 (3.5kPa) Brain Two-way ANOVA revealed no significant effect of normoxia, 6hr hypoxia (3.5kPa), or hypoxia-recovery (treatment effect, F= 3.26, p= 0.058), no significant effect of species (F= 0.091, p= 0.77), and no significant treatment x species interaction (F=0.47, p=0.63) on brain GSH:GSSG (Fig.4.1A). Similarly, there was no effect of treatment (F=2.43, p=0.10), species (f=1.74, p=0.20), or treatment x species interaction (F=1.34, p=0.27) on MitoP/MitoB (Fig.4.1B). There were, however, significant effects of treatment (F=5.34, p= 0.012), species (F=34.66, p<0.0001), and a significant treatment x species interaction (F=10.12, p=0.0006) on brain TBARS. Post-hoc analysis revealed that O. maculosus increased TBARS after 6hrs exposure to hypoxia compared with both normoxia and hypoxia-recovery treatments, whereas S. marmoratus showed no changes (Fig.4.1C). Although there were no effects of treatment (F=2.56, p=0.094), there were significant effects of species (F=7.60, p=0.01), and treatment x species interaction (F=4.28, p=0.024) on brain TOSC. Post-hoc analysis shows that O. maculosus increased TOSC after 6hrs exposure to hypoxia compared with both normoxia and hypoxia-recovery treatments, whereas S. marmoratus showed no changes (Fig.4.1D).  77  Liver  Two-way ANOVA revealed a significant species effect (F=18.39, p=0.0002), but no treatment effect (F=1.72, p=0.20), and of treatment x species interaction (F=0.17, p=0.85) on liver GSH:GSSG. Post-hoc analysis showed that O. maculosus had a more oxidized GSH:GSSG compared to S. marmoratus during normoxia (Fig.4.2A). Similarly, there was a significant species effect (F=13.45, p=0.0007), but no effect of treatment (F=1.303, p=0.28) or treatment x species interaction (F=0.31, p=0.74) on MitoP/MitoB. Post-hoc analysis revealed a significantly higher MitoP/MitoB in O. maculosus than in S. marmoratus in hypoxia-recovery (Fig.4.2B). There was a significant effect of species (F=8.77, p=0.0062), but no effect of treatment (F=3.27, p=0.053) or treatment x species interaction (F=2.17, p=0.13) on TBARS. Post-hoc analysis showed that O. maculosus had significantly higher TBARS level with 3.5kPa hypoxia exposure compared to levels during normoxia and hypoxia-recovery, whereas S. marmoratus showed no changes (Fig.4.2C). There was a significant effect of treatment (F=4.19, p=0.023) and species (F=6.39, p=0.016), but not treatment x species interaction (F=2.18, p=0.13) on liver TOSC levels. Post-hoc analysis revealed that S. marmoratus had significantly higher TOSC levels during hypoxia-recovery compared to levels during normoxia and hypoxia exposure, whereas O. maculosus showed no changes (Fig.4.2D).  Gill  There was a significant effect of treatment (F=3.84, p=0.04), but no effect of species (F=2.35, p=0.14) or treatment x species interaction (F=2.88, p=0.073) on gill GSH:GSSG. From the post-hoc analysis, hypoxia exposure resulted in an oxidized GSH:GSSG in S. marmoratus compared with normoxia and hypoxia-recovery, whereas O. maculosus showed no changes in GSH:GSSG (Fig.4.3A). There were no significant effects of treatment (F=1.90, p=0.16), species (F=0.036, p=0.85), or treatment x species interaction (F=0.44, p=0.64) on MitoP/MitoB. There were also no effects of treatment (F=0.13, p=0.88), species (F=3.46, p=0.074), or treatment x species interaction (F=0.86, p=0.43) on TBARS (Fig.4.3C). Similarly, there were no effects of treatment (F=0.76, p=0.47), species (F=1.59, p=0.21), or treatment x species (F=0.40, p=0.67) interaction on TOSC (Fig.4.3D). 4.3.3 Effects of hypoxic PO2 on responses in O. maculosus (2.3 and 3.5kPa) Brain 78  In order to account for the fact that the O. maculosus and S. marmoratus differ in hypoxia tolerance and Pcrit, and exposure of both species to a common hypoxia PO2 of 3.5kPa represents a more severe stress for S. marmoratus (3.5kPa represents 65% Pcrit) than for O. maculosus (3.5kPa is at Pcrit), we assessed the effects of PO2 on the responses of O. maculosus. In O. maculosus exposed to normoxia, two levels of hypoxia (2.3kPa which represents 65% Pcrit for this species and 3.5kPa which is at Pcrit for this species), and hypoxia with 1hr normoxia recovery, two-way ANOVA revealed a significant treatment effect (normoxia, hypoxia, hypoxia/recovery; F=6.41, p=0.0059), but no effect of PO2 (2.3 and 3.5 kPa; F=0.40, p=0.53) and treatment x PO2 interaction (F=1.77, p=0.19) on GSH:GSSG. Post-hoc analysis revealed that O. maculosus exposed to hypoxia-recovery had significantly reduced GSH:GSSG compared with 3.5 kPa hypoxia, whereas O. maculosus from the 2.3kPa hypoxia treatments showed no changes (Fig.4A). There were no significant effects of treatment (F=1.51, p=0.23), PO2 (F=3.12, p=0.086), or treatment x PO2 interaction (F=0.79, p=0.46) on brain MitoP/MitoB (Fig.4.4B). There was a significant effect of treatment (F=15.45, p<0.0001), but no effect of PO2 (F=3.977, p=0.056) or treatment x PO2 interaction (F=1.244, p=0.30) on brain TBARS. Post-hoc analysis showed a significant increase in TBARS in O. maculosus exposed to both 3.5kPa hypoxia and 2.3kPa hypoxia, which was significantly reduced in hypoxia-recovery after exposure to 3.5 kPa, but not 2.3kPa (Fig.4.4C). There was a significant effect of treatment (F=4.30, p=0.023), but no effect of PO2 (F=0.016, p=0.90) or treatment x PO2 interaction (F=2.16, p=0.13) on TOSC. Post-hoc analysis showed that O. maculosus after exposure to 3.5kPa hypoxia had significantly higher TOSC levels than during normoxia and hypoxia-recovery, whereas there were no changes in O. maculosus exposed to 2.3kPa hypoxia and recovered (Fig.4.4D).  Liver Comparing O. maculosus exposed to 2.3kPa and 3.5kPa hypoxia, there were no significant effects of treatment (F=0.35, p=0.71), PO2 (F=1.93, p=0.16), or treatment x PO2 interaction (F=0.47, p=0.63; Fig.4.5A) on liver GSH:GSSG. There were also no significant effects of treatment (F=1.93, p=0.16), PO2 (F=0.086, p=0.77), or treatment x PO2 interaction (F=0.38, p=0.69; Fig.4.5B) on MitoP/MitoB. There were significant effects of PO2 (F=7.79, p=0.0094), but no effects of treatment (F=2.068, p=0.15) or treatment x PO2 interaction (F=2.58, p=0.093) on TBARS. Post-hoc analysis revealed that O. maculosus exposed to 3.5kPa hypoxia had significantly higher TBARS levels compared to in normoxia and hypoxia/recovery, whereas those exposed to 2.3kPa showed no 79  changes (Fig.4.5C). There were significant effects of PO2 (F=7.55, p=0.0092), but no effects of treatment (F=2.55, p=0.092) and treatment x PO2 interaction (F=4.89, p=0.013) on TOSC levels. Post-hoc analysis revealed that O. maculosus exposed to 3.5kPa hypoxia had significantly lower TOSC levels compared to normoxia and hypoxia-recovery, whereas O. maculosus exposed to 2.3kPa showed no changes (Fig.4.5D).  In light of the significant effects of PO2 in liver TBARS and TOSC in O. maculosus (Fig.4.5C&D), we compared O. maculosus and S. marmoratus at their respective 65% Pcrit exposures, i.e. O. maculosus at 2.3kPa and S. marmoratus at 3.5kPa. There were no significant effects of treatment (F=0.019, p=0.39), species (F=0.16, p=0.86), and treatment x species interaction (F=0.23, p=0.79) on TBARS (Fig.4.7A). There were no effects of treatment (F=0.59, p=0.56) or species (F=0.46, p=0.50), but a significant treatment x species interaction (F=3.65, p=0.036) for TOSC levels. Specifically, S. marmoratus had significantly higher TOSC levels with hypoxia-recovery compared to levels during hypoxia, but not different from levels during normoxia, whereas O. maculosus showed no changes (Fig.4.7B).  Gill  There were no significant effects of treatment (F=1.71, p=0.20), PO2 (F=2.95, p=0.095), or treatment x species interactions (F=0.74, p=0.48) on gill GSH:GSSG (Fig.4.6A). There were also no significant effects of treatment (F=3.06, p=0.058), PO2 (F=1.87, p=0.18), and treatment x PO2 interaction (F=1.04, p=0.36) on MitoP/MitoB. However, post-hoc analysis revealed that O.maculosus in hypoxia-recovery from 3.5kPa exposure had significantly lower MitoP/MitoB than in normoxia, whereas O. maculosus exposed to 2.3kPa and recovered show no changes (Fig.4.6B). There were no significant effects of treatment (F=0.32, p=0.73), PO2 (F=0.0056, p=0.94), and treatment x PO2 interaction (F=0.036, p=0.96) on TBARS (Fig.4.6C). Also, there were no significant effects of treatment (F=0.14, p=0.87), PO2 (F=0.0026, p=0.96), and treatment x PO2 interaction (F=0.080, p=0.92) on TOSC (Fig.4.6D).  4.3.4 Hyperoxia and hyperoxia-recovery at a common PO2 (64kPa) Brain  Two-way ANOVA revealed no significant effect of 6hr hyperoxia (64kPa) or hyperoxia-recovery (treatment effect, F= 0.27, p= 0.76) on brain GSH:GSSG. However, there was a significant effect of 80  species on brain GSH:GSSG (F= 6.86, p= 0.0.16) with no treatment x species interaction (F=3.05, p=0.068). Post-hoc analysis showed that O. maculosus had a significantly more oxidized GSH:GSSG with hyperoxia-recovery compared to S. marmoratus (Fig.4.1E). Similarly, there was a significant species effect (F=10.07, p=0.0031), but no effect of treatment (F=0.27, p=0.76) and treatment x species interaction (F=0.3.05, p=0.068) of MitoP/MitoB. Specifically, O. maculosus had a significantly higher MitoP/MitoB during 6hr hyperoxia compared to S. marmoratus (Fig.4.1F). There was no effect of treatment (F=0.21, p=0.81), but significant effects of species (F=20.78, p<0.0001), and treatment x species interaction (F=5.25, p=0.012) on TBARS. Post-hoc analysis showed significantly higher TBARS level in O. maculosus with hyperoxia-recovery compared to normoxia and during hyperoxia, whereas S. marmoratus showed no changes (Fig.4.1G). There was no effect of treatment (F=0.98, p=0.39), but there were significant effects of species (F=23.65, p<0.0001), and treatment x species interaction (F=7.10, p=0.0031) on TOSC. Specifically, O. maculosus had significantly higher TOSC levels during hyperoxia and hyperoxia-recovery compared to normoxic levels, whereas S. marmoratus showed no changes (Fig.4.1H).  Liver  There was a significant effect of treatment (F=3.48, p=0.045), and species (F=11.35, p=0.0022), but no effect of treatment x species interaction (F=1.89, p=0.17) on liver GSH:GSSG. Post-hoc analyses showed that O. maculosus had significantly more oxidized GSH:GSSG in hyperoxia-recovery compared to during hyperoxia, which was not different from normoxic levels. S. marmoratus showed no changes in GSH:GSSG in hyperoxia-recovery (Fig.4.2E). There were significant effects of treatment (F=3.43, p=0.042), and species (F=20.55, p<0.0001), but no effect of treatment x species interaction (F=0.013, p=0.99) on MitoP/MitoB. Post-hoc analyses revealed that O. maculosus had significantly higher MitoP/MitoB than S. marmoratus in normoxia, hyperoxia, and hyperoxia-recovery (Fig.4.2F). There was no effect of treatment (F=1.69, p=0.20), but there was a significant effect of species (F=7.93, p=0.0091) and treatment x species interaction (F=3.54, p=0.044) on TBARS. Post-hoc analyses revealed significantly higher TBARS during hyperoxia in O. maculosus relative to normoxia and hyperoxia-recovery levels, whereas S. marmoratus showed no changes (Fig.4.1G). There was significant effect of treatment (F=4.30, p=0.022), species (F=26.51, p<0.0001) and treatment x species interaction (F=6.78, p=0.0035) on TOSC. Post-hoc analyses showed 81  significantly higher TOSC levels in S. marmoratus during treatment and hyperoxia-recovery compared to normoxic levels, whereas O. maculosus showed no changes (Fig.4.2H).  Gill  There was a significant species effect (F=6.95, p=0.013), but no effects of treatment (F=0.14, p=0.87) and treatment x species interaction (F=0.14, p=0.87) on GSH:GSSG. There was a significant effect of treatment (F=3.44, p=0.042), but no effects of species (F=1.42, p=0.24) and treatment x species interaction (F=0.10, p=0.91) of MitoP/MitoB (Fig.4.3F). There were no effects of treatment (F=0.26, p=0.77), species (F=0.40, p=0.53), and treatment x species interaction (F=0.12, p=0.89) on TBARS (Fig.4.3G). There were also no effects of treatment (F=0.40, p=0.67), species (F=0.35, p=0.56), or treatment x species interaction (F=0.23, p=0.79) on TOSC (Fig.4.3H).  4.4. Discussion Our previous in vitro work on sculpins has shown that interspecific variation in hypoxia tolerance was associated with variation in mitochondrial ROS generation, with mitochondria from more hypoxia tolerant species generating more ROS/O2 under reducing conditions than those from hypoxia intolerant species. In addition, mitochondria from hypoxia tolerant sculpins were more sensitive to changes in the mitochondrial redox environment, and did not recover state III respiration after an in vitro anoxia-reoxygenation exposure to the same extent as the less hypoxia tolerant sculpins (Chapter 3). These counterintuitive results, however, were based on experiments using mitochondria isolated from the brain which may not accurately reflect in vivo conditions and responses. Thus, the primary goal of the present study was to determine if the observed results from in vitro studies (Chapter 3) were also observed in vivo and whether there were tissue specific differences in ROS metabolism in response to hypoxia, hypoxia-recovery, hyperoxia, hyperoxia-recovery exposures. Indeed, the more hypoxia tolerant O. maculosus generally exhibited greater responses in ROS metabolism, i.e. changes in glutathione redox status, TBARS, and TOSC, to both hypoxia and hyperoxia exposures followed by normoxia recovery, and there were tissue specific differences. However, the changes in oxidative damage and scavenging capacity in response to hypoxia and hyperoxia do not appear to be caused by ROS accumulated in the mitochondrial matrix as indicated by no changes in MitoP/MitoB, pointing to other possible sources of ROS.  82  4.4.1 MitoB for the detection of in vivo mitochondrial ROS generation in marine sculpins The recent development of the mitochondrially-targeted MitoB has proven to be a valuable tool in the assessment of in vivo mitochondrial ROS production and has contributed greatly to our understanding of mammalian ROS production in response to cardiac ischemic-reperfusion (Chouchani et al., 2014). Although MitoB is well used in the study of human-related pathologies in murine models, far fewer studies have employed MitoB in the comparative or environmentally-oriented context. Salin et al. (2017) used MitoB and showed that brown trout with high metabolic rates had lower liver ROS generation. MitoB was also used in drosophila to show that ageing was related to increase in H2O2 accumulation (Cochemé et al., 2011), and also that hyperoxia led to an increase in H2O2 (Cochemé et al., 2012). MitoB has not yet been used in an interspecies comparison, in animals that inhabit different O2 environments. Even though MitoB has already been used in teleost fish (Salin et al., 2015), the use of the probe must first be validated in any study organism to account for species-specific excretion rates and tissue-specific uptake. Our validation study in sculpins shows that for the most part MitoB accumulated in various tissues to a similar extent as in other animal systems (Cochemé et al., 2012; Logan et al., 2014; Salin et al., 2017). MitoB quickly accumulated in brain, gill and liver after injection (within 30min), after which there was an exponential decay with reliable detection for up to 8hrs post-injection (Supplementary Fig.4.2). MitoB was also quickly converted to MitoP in brain, gill, and liver tissues, which similarly showed an exponential decay over the 72hrs. The concentrations of MitoP and MitoB result in a stable MitoP/MitoB over the first 8hrs post-injection (Supplementary Fig.4.2). The timeline of accumulation, however, differed from that of the freshwater brown trout where MitoB was still at high levels in liver and white muscle at 72hrs post-injection (Salin et al. 2017). There were also differences in tissue MitoP/MitoB between sculpins and brown trout, where brown trout white muscle (0.27) showed a much higher ratio than in liver (0.073), whereas in sculpins we were unable to reliably detect MitoP/MitoB in white muscle due to the slow uptake of MitoB (Supplementary Fig.4.2; Salin et al. 2017). These species-specific differences in tissue MitoB uptake and excretion could be due to comparison between freshwater and seawater fish and due to the difference in osmoregulatory strategies. It is also important to note that we observed MitoB uptake in the sculpin brain, which has not been previously observed in mammals (personal observation), likely due to the presence of a tight blood brain barrier. Teleost fish are also believed to have a tight endothelial-based blood brain barrier analogous to that of other vertebrates, but the 83  uptake of MitoB in sculpins brain suggest that there are functional differences in the blood brain barrier between teleosts and mammals that allowed for uptake of this charged molecule (Kniesel and Wolburg, 2000; Wolburg et al., 1983). These differences in MitoB uptake across species emphasize the importance of validating its use in a new species before undertaking experiments.   4.4.2 Hypoxia and hyperoxia exposure had greater effects on ROS metabolism in O. maculosus than S. marmoratus In addition to using MitoB to measure mitochondrial H2O2, we chose to measure representative indices of tissue redox status, oxidative damage, and scavenging capacity to study different aspects of ROS metabolism. Tissue GSH:GSSG represent the dominant redox pair within the cell and is easily assessed by measuring the concentrations of GSH and GSSG (Rahman et al., 2006). Changes in glutathione redox status have been used to describe changes in the cellular redox environment, which can drive changes in ROS generation rate (Aon et al., 2010; Jones, 2002; Schafer and Buettner, 2001). To assess oxidative damage, we quantified TBARS levels as an indicator of lipid peroxidation, although proteins, lipids, and DNA are also susceptible to oxidative damage, which may occur over different time scales (Cadet, 2003; Reznick and Packer, 1994). Redox imbalance and oxidative damage could signal an increase in ROS scavenging capacity. In lieu of measuring only a few select antioxidant enzymes, we assessed the overall ability of the tissue to metabolize H2O2 with the TOSC assay. However, there are different types of antioxidants, enzymatic and non-enzymatic, each with different chemical properties and reactivity to different ROS species that would not be encompassed by this TOSC measure. Overall, the indices that we chose provide us with an overview of how the two sculpin species respond to changes in environmental PO2.  Hypoxia- and hyperoxia-recovery generally had a greater impact on ROS metabolism in O. maculosus than S. marmoratus. O. maculosus showed signs of increased oxidative damage with a simultaneous increase in TOSC levels during hypoxia, hyperoxia, and hyperoxia-recovery, whereas S. marmoratus showed fewer responses.   In O. maculosus, hypoxia at a common PO2 and recovery yielded no effects on redox balance and H2O2 in the brain, liver, and gill. Although there were no signs of redox imbalance and H2O2 accumulation in this hypoxia tolerant species, there were significant effects of ROS in brain and 84  liver, but not the gills. In the brain, the increase in TBARS was concomitant with an increase in TOSC, suggesting that the increased scavenging capacity was not enough to prevent oxidative damage. These increased levels of brain TBARS and TOSC capacity, for the most part, recovered after 1hr at normoxia, which is within the timeframe of a typical tidal cycle. There was a similar increase of TBARS in the liver, however, it was not accompanied by an increase in TOSC. In contrast, S. marmoratus showed relatively few changes in hypoxia-recovery. There was no indication of redox imbalance, changes in H2O2, oxidative damage, or scavenging capacity in the brain to hypoxia-recovery. The liver also did not show indications of redox imbalance, changes in H2O2, or oxidative damage. However, there was an increase in TOSC levels in recovery from hypoxia, suggesting that the upregulation of scavenging capacity may have prevented any oxidative damage (i.e. increase in TBARS). In the gills, there was oxidation of glutathione redox status to hypoxia that was recovered in normoxia but this sign of redox imbalance did not yield changes in H2O2 and there was also no evidence of oxidative damage.  In O. maculosus, hyperoxia did not lead to redox imbalance or changes in H2O2 in the brain. However, there were increases in TBARS and TOSC that remained high in recovery. In the liver, GSH:GSSG was significantly reduced in hyperoxia but was recovered in normoxia. This did not cause any changes in H2O2. Similar to the hypoxia response, increase in liver TBARS was not mirrored by TOSC levels. Finally in the gills, GSH:GSSG was reduced in hyperoxia and recovered in normoxia, but no changes in H2O2 and no signs of oxidative damage were observed. S. marmoratus did not show any response to hyperoxia-recovery, except for a significant increase in liver TOSC in hyperoxia that remained high in recovery.  Our results showed few changes in MitoP/MitoB in response to hypoxia and hyperoxia in either species, suggesting that very little mitochondrial H2O2 was generated even though there was evidence of ROS effects with changes in TBARS and TOSC. There are several possible explanations that may contribute to this discrepancy. First, the oxidative damage could have been caused by ROS released by ETS complexes sites orientated toward the cytoplasmic side (complex III and glycerol-3-phosphate dehydrogenase (GPDH)) and not detected in the matrix by MitoB. Second, ROS other than H2O2 (e.g. superoxide and hydroxyl radicals which are normally quickly scavenged and converted to less reactive H2O2), and also reactive species other than ROS (e.g. reactive sulfur and nitrogen species; DeLeon et al., 2016) may be causing oxidative damage in response to O2 stress 85  rather than H2O2. Third, non-mitochondrial sites have been shown to have the capacity to generate a significant amount of ROS (Brown and Borutaite, 2012). It was estimated in the rat liver that mitochondria contribute 18% of total ROS, but microsomes and peroxisomes can generate 45% and 35% respectively of total ROS (Chance et al., 1979). Lastly, although less likely, is that the MitoB probe is not sensitive to mitochondrial ROS production in marine teleosts. Although it is unlikely that the chemistry of the probe itself would differ in a different animal model (and our validation shows that MitoB works similarly in sculpins), it is conceivable that there are details that require further validation with a positive control using a teleost study model (e.g. using MitoB to detect superoxide generation stimulated by mitochondrial-targeted paraquat, a redox cycler). 4.4.3 Tissue specific responses of ROS metabolism to O2 stress The gills in both species were generally more resistant to both hypoxia and hyperoxia, showing no changes in mitochondrial H2O2, TBARS or TOSC, whereas the brain and liver tissues showed greater responses in ROS metabolism to O2 variability. No changes in TBARS level would suggest that no ROS was generated in response to hypoxia and hyperoxia, and thus it was not necessary to elicit any TOSC responses. Given its functional role as the gas exchange surface, exposed gills would presumably be more susceptible to oxidative stress. Thus, it was surprising that we did not observe large responses in ROS metabolism in the gills that would act to buffer changes in O2 levels to minimize the impact on internal organs. Gills in fish represent a unique respiratory surface compared to other organisms in that they not only act as the site of gas exchange, but also for ion exchange. It is possible that there are adaptive mechanisms to protect the gills in order to sustain both of these essential functions. Curiously, the resistance of gills to oxidative damage was not due to constitutively higher levels of TOSC that were at similar levels to the brain (Supplementary Table 4.1). However, our assessment of TOSC (i.e. ability to metabolize H2O2) does not encompass all of the antioxidant mechanisms within the cell, and thus there could be other ROS scavenging processes that give the gills higher buffering capacity to any changes in ROS metabolism. Also, gills in anoxia and hypoxia tolerant animals have been shown to have incredible morphological plasticity and can be rapidly modified in response to changes to environmental O2 (Sollid and Nilsson 2006).  Tissue redox status of liver from O. maculosus was generally more oxidized (concomitant with a higher MitoP/MitoB) than liver from S. marmoratus. Under hypoxic conditions, lowered ETS flux would increase protonmotive force and drive ETS complexes into a more reduced state, resulting in 86  an overall reduced redox environment. Under hyperoxic conditions, high O2 availability leads to a relatively more oxidized redox environment. In fact, liver and gill of hyperoxia exposed O. maculosus had more reduced GSH:GSSG which was recovered in normoxia (Fig.4.2E & 4.3E), and S. marmoratus hypoxic gill had more oxidized GSH:GSSG which was recovered in normoxia (Fig.4.3A). These changes in GSH:GSSG are contrary to what would be predicted (as O2 limitation would typically lead to more reduced redox status), potentially indicating regulation of tissue GSH:GSSG in response to the O2 exposure in both species. In fact, tissues and cells have been observed to export GSSG in attempts to preserve GSH:GSSG, and it appears that this ability to regulate GSH:GSSG is correlated with protection of the tissue from oxidative stress (Ishikawa and Sies, 1984; Sies and Akerboom, 1984). Thus, O. maculosus may be actively regulating GSH:GSSG in liver and gills in response to hyperoxia exposure. Of note, the pattern of changes in GSH:GSSG in gills and liver in both species exposed to hyperoxia-recovery were similar, but the response in ROS, TBARS and TOSC differed in both of these tissues.  The TOSC response to hypoxia and hyperoxia differed between tissues in both species. Although the brain and liver had similar patterns of change in TBARS (i.e. where S. marmoratus showed no changes, and O. maculosus showed an increase in TBARS during O2 treatment), the TOSC response differed. In the brain, the increase in TBARS in O. maculosus was concomitant with an increase in TOSC, which indicates that the increased scavenging capacity was likely overwhelmed and excess ROS likely caused TBARS to increase. This is in contrast to what was observed in the liver, where TOSC did not increase with the increase in TBARS in O. maculosus in hypoxia and hyperoxia, and S. marmoratus showed a significant increase in TOSC that appeared to mitigate oxidative damage.  4.4.4 Effect of PO2 on ROS metabolism in O. maculosus  As we are comparing responses in two species that differ in hypoxia tolerance, exposure to a common PO2 represents a more severe hypoxic stress for the less hypoxia tolerant species than the more tolerant species (e.g. 3.5kPa is 65% Pcrit for S. marmoratus and at Pcrit for O. maculosus). In order to address this issue, we also exposed O. maculosus to a PO2 equivalent to 65% Pcrit (2.3kPa) and compared the responses to those of O. maculosus exposed to a PO2 at Pcrit (3.5kPa). These differences in PO2 had little effects on ROS metabolism across tissues and in fact, there were only significant effects of this lower PO2 on liver TBARS and TOSC. O. maculosus at 3.5kPa had significantly higher TBARS than in fish exposed to 2.3kPa, which was recovered in normoxia, but at the same time 87  TOSC levels were significantly reduced during exposure to 3.5kPa whereas those under 2.3kPa showed no changes in TOSC (Fig.4.5C & D). The response of O. maculosus to 65% Pcrit (2.3kPa) hypoxia is not unlike S. marmoratus, which would indicate similar responses at similar relative levels of PO2, although S. marmoratus showed an increase in TOSC to hypoxia-recovery (Fig.4.2D).  It is interesting to consider what occurs at Pcrit to explain how hypoxia exposure at Pcrit (3.5kPa for O. maculosus) induced more oxidative damage than exposure to the lower 65% Pcrit (2.3kPa for O. maculosus) exposure in O. maculosus. Pcrit is the PO2 at which animals switch from being oxy-regulating to oxy-conforming as environmental PO2 continues to decline, and it reflects the ability of an animal to extract sufficient O2 from the environment to sustain a constant rate of metabolism. Below Pcrit, the dependency upon anaerobic metabolism increases, and ultimately results in a significant accumulation of lactate in anoxia-tolerant goldfish and tilapia (Regan et al., 2017; Speers-Roesch et al., 2010). At Pcrit (3.5kPa), reducing equivalents may still be available to mitochondria to support ETS flux, but the change in ETS flux caused by changes in PO2 still led to an increase in oxidative damage, whereas below Pcrit a downregulation of ETS flux (regulated by pyruvate dehydrogenase; Papandreou et al., 2006b; Richards et al., 2008) may reduce ROS generation which would explain the lack of TBARS accumulation observed with 2.3kPa hypoxia exposure.   As there were significant effects of PO2 on O. maculosus liver TBARS and TOSC (Fig.4.5C & D), we also compared O. maculosus and S. marmoratus both at 65% Pcrit hypoxia exposure (Fig.4.7). While there were no interspecies differences in TBARS (Fig.4.7A), S. marmoratus significantly increased TOSC levels during hypoxia-recovery whereas O. maculosus showed no changes in TOSC (Fig.4.7B). Assuming that the relative hypoxia exposure led to the same arterial PO2 (Speers-Roesch et al., 2013), this indicates that the only difference in response to the same level of hypoxemia is in liver scavenging capacity with normoxia-recovery. This signifies inherent differences in the O2 sensitivity of tissue scavenging capacity in liver. 4.4.5 Greater response in ROS metabolism as an adaptive response to O2 variability  Our results show that ROS metabolism in the more hypoxia tolerant O. maculosus were more perturbed by O2 variability when compared to the less tolerant S. marmoratus in the timeframe that mimics a typical tidal cycle. Similarly, O. maculosus that were exposed to temperature changes exhibited a more sensitive response in heat shock protein expression (Todgham et al. 2006; Chapter 3). These studies show that quick responses mirroring changes in abiotic factors in the environment 88  may confer survival advantage in the higher intertidal. However, it appears that the timing of responses depends on which level of biological organization is in question. At the transcript level,  the less hypoxia tolerant sculpin Blepsias cirrhosus showed changes in mRNA expression between 3-24hrs in response to hypoxia whereas O. maculosus showed changes later between 24-72hrs, indicating that upper intertidal O. maculosus adopted a more generalist approach to more prolonged exposure to hypoxia (Mandic et al., 2014). Perhaps the preferred strategy for higher intertidal living is dependent upon sensitive detection of environmental changes and initiating plastic responses to quickly modify existing proteins such as antioxidant enzymes (e.g. via post-translational modifications) and slower changes in mRNA transcription. The coordination of a global transcriptional response would be energetically expensive and is likely reserved for coordinating acclimation response to chronic O2 stress. 4.4.6 Summary Overall, we found that the more hypoxia tolerant sculpin O. maculosus generally showed more effects of ROS metabolism (changes in levels of TBARS and TOSC) compared to the less hypoxia tolerant sculpin S. marmoratus, indicating a more sensitive response to hypoxia and hyperoxia under O2 conditions frequently encountered in the intertidal environment. These results from whole animal exposures are generally consistent with what has been observed in in vitro work on isolated brain mitochondria (Chapter 3), whereby more hypoxia tolerant sculpins generated more ROS under resting conditions and were more sensitive to changes in the extramitochondrial redox environment. The overall greater responses in more hypoxia tolerant sculpins may be part of a sensitive detection of changes in environmental O2 to coordinate quick cellular responses. 89  Figure 4.1. The effect of hypoxia (3.5kPa)-recovery (A-D) and hyperoxia (64.0kPa)-recovery (E-H) in brain of Oligocottus maculosus (hollow squares) and Scorpaenichthys marmoratus (solid squares) on ROS metabolism as assessed by (A, E) tissue GSH:GSSG redox status, (B, F) MitoP/MitoB, (C, G) TBARS (normalized to normoxia control value), (D, H) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia or hyperoxia, recovery) within a species are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between species within an O2 treatment. No significant differences in datasets are indicated as ‘n.s.’ in the figure.  90   Figure 4.1.  91  Figure 4.2. The effect of hypoxia (3.5kPa)-recovery (A-D) and hyperoxia (64.0kPa)-recovery (E-H) in liver of Oligocottus maculosus (hollow squares) and Scorpaenichthys marmoratus (solid squares) on ROS metabolism as assessed by (A, E) tissue GSH:GSSG redox status, (B, F) MitoP/MitoB, (C, G) TBARS (normalized to normoxia control value), (D, H) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia or hyperoxia, recovery) within a species are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between species within an O2 treatment. No significant differences in datasets are indicated as ‘n.s.’ in the figure. 92  Figure 4.2.  93  Figure 4.3. The effect of hypoxia (3.5kPa)-recovery (A-D) and hyperoxia (64.0kPa)-recovery (E-H) in gill of Oligocottus maculosus (hollow squares) and Scorpaenichthys marmoratus (solid squares) on ROS metabolism as assessed by (A, E) tissue GSH:GSSG redox status, (B, F) MitoP/MitoB, (C, G) TBARS (normalized to normoxia control value), (D, H) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia or hyperoxia, recovery) within a species are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between species within an O2 treatment. No significant differences in datasets are indicated as ‘n.s.’ in the figure. 94  Figure 4.3 95  Figure 4.4. The effect of 2.3kPa (solid triangle) and 3.5kPa (clear square) hypoxia-recovery in brain of O. maculosus on ROS metabolism assessed by (A) tissue GSH:GSSG redox state, (B) MitoP/MitoB, (C) TBARS (normalized to normoxia control value), (D) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia, and recovery) are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between 2.3 and 3.5kPa hypoxia exposures. No significant differences in datasets are indicated as ‘n.s.’ in the figure.  96   Figure 4.4  97  Figure 4.5. The effect of 2.3kPa (solid triangle) and 3.5kPa (clear square) hypoxia-recover in liver of O. maculosus on ROS metabolism assessed by (A) tissue GSH:GSSG redox state, (B) MitoP/MitoB, (C) TBARS (normalized to normoxia control value), (D) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia, and recovery) are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between 2.3 and 3.5kPa hypoxia exposures. No significant differences in datasets are indicated as ‘n.s.’ in the figure. 98   Figure 4.5 99  Figure 4.6. The effect of 2.3kPa (solid triangle) and 3.5kPa (clear square) hypoxia-recovery in gill of O. maculosus on ROS metabolism assessed by (A) tissue GSH:GSSG redox state, (B) MitoP/MitoB, (C) TBARS (normalized to normoxia control value), (D) TOSC (normalized to normoxia control value). Data are means ± standard error of mean. The effects of O2 treatments (normoxia, hypoxia, and recovery) are shown with letters, where data points with different letters are significantly different. Asterisks indicate significant differences between 2.3 and 3.5kPa hypoxia exposures. No significant differences in datasets are indicated as ‘n.s.’ in the figure. 100   Figure 4.6 101   Figure 4.7. Effect of relative hypoxia exposure and recovery in liver (A) TBARS and (B) TOSC of O. maculosus and S. marmoratus (2.3kPa for O. maculosus (hollow triangles) and 3.5kPa for S. marmoratus (solid squares)). The effects of O2 treatments (normoxia, hypoxia, and recovery) are shown with letters, where data points with different letters are significantly different. There were no significant differences between species within an O2 treatment. No significant differences in datasets are indicated as ‘n.s.’ in the figure.   102  CHAPTER FIVE:  GENERAL DISCUSSION AND CONCLUSION  The overall objective of my PhD thesis was to determine if there is evidence of adaptive variation in mitochondrial function across species that inhabit O2 variable environments. Specifically, I aimed to investigate the interplay between aerobic metabolism and mitochondrial ROS emission, which are both processes that environmental O2 variability could exert strong selective pressure on. To do so, I used multiple species of sculpin that live along the marine intertidal zone as my study model since these species have previously been shown to vary in their hypoxia tolerance (Mandic et al., 2009; Mandic et al., 2013). In chapter 2, I demonstrated that the adaptive variation in the O2 transport cascade previously described in sculpins (Mandic et al. 2009) extends to the level of mitochondria and COX. I then explored the in vitro relationship between oxidative phosphorylation and ROS metabolism in isolated mitochondria from sculpins in chapter 3. Finally, I examined ROS metabolism in vivo by investigating whether there were species differences to whole animal exposure to hypoxia, hyperoxia and normoxic recovery in Chapter 4.  In this discussion, I will first summarize the major findings from my thesis and develop a revised model of the O2 transport cascade that distinguishes between hypoxia tolerant and intolerant sculpins. I will then describe how my research fits with and extends what is currently known about the responses of organisms to O2 fluctuations, and finish with some considerations for current and futures studies on ROS metabolism.  5.1 Major Findings 5.1.1 Adaptive variation in oxygen binding of mitochondria and COX (Chapter 2) Using intertidal sculpins, I provide the first evidence of adaptive variation in mitochondrial and COX O2 binding across species that vary in hypoxia tolerance. In order to investigate the mechanism underlying the variation in COX Km,app O2, I performed in silico protein analyses on the catalytic core and showed a strong relationship between COX Km,app O2 and COX3 subunit stability, which suggests a significant role of COX3 in determining the O2 kinetics of the overall protein. Even though the COX3 subunit is not thought to directly interact with O2, my results from comparative protein modeling revealed residue differences in the COX3 v-cleft structure which 103  could influence interactions with the mitochondrial membrane lipid, cardiolipin, and affect O2 diffusion into the COX protein en route to the COX1 catalytic site.  Previous studies in animals that have evolved to inhabit low O2 environments have shown evidence of adaptive variation at multiple steps of the O2 transport cascade down to the level of O2 diffusion at the cellular level (Mahalingam et al., 2017; Scott et al., 2011; Mandic et al., 2009; Mandic et al., 2013). My thesis extends this general phenomenon to the site of O2 use at the subcellular level in intertidal sculpins. Indeed, sculpins show parallel interspecific relationships between Pcrit, whole red blood cell P50, stripped hemoglobin-O2 P50, and also at the subcellular level between mitochondrial P50 and COX Km,app O2 (Fig.5.1) such that hypoxia tolerance is associated with improved or sustained O2 diffusion and extraction from the environment to mitochondria. The parallel relationships between the different traits indicate that the maintenance of the gradients of O2 diffusion is important for environments that are or frequently become hypoxic, presumably to improve O2 supply to mitochondria.  Additionally, my results demonstrate that there is selective pressure on both COX Vmax and COX respiration as there were significant relationships of both to mitochondrial P50. The opposite relationships of mitochondrial O2 kinetics, and COX Vmax and respiration are intriguing; hypoxia tolerant sculpins have lower COX enzyme (Vmax= kcat x [E]), but higher COX respiration rate in uncoupled mitochondria altogether indicating that hypoxia tolerant sculpins have a higher functioning COX enzyme relative to that in hypoxia intolerant sculpins.   Influences of protein stability on enzyme function  Variation in protein stability and net protein surface charge can have a number of effects, direct or indirect, on protein function. For instance, variation in myoglobin surface charge was demonstrated to be associated with the maximum concentration of myoglobin and used to infer mammalian dive times (Mirceta et al., 2013). Mechanistically, increased electrostatic repulsion between myoglobin molecules is thought to prevent aggregation at high myoglobin concentrations, which would maintain high O2 binding affinity and support longer dive times (Mirceta et al. 2013). In the case of COX in the sculpins, the variation in COX3 protein stability may reflect differences in how COX3 interacts with its immediate environment, including interactions with other COX subunits or the mitochondrial lipid bilayer. Changes in protein stability can also exert a more indirect effect; a more 104  stable protein is an indication that the protein is more compact, which can affect catalytic rate by bringing catalytically important residues within the active site closer together. Additionally, an increase in overall protein stability has been hypothesized to increase the evolvability of a protein by stabilizing mutations that would otherwise destabilize the protein and cause unfolding (Bloom et al., 2006). With increasingly more powerful protein modeling methods that are easier to use, functional analyses can continue to be combined with in silico analyses to investigate how the environment has shaped protein adaptation in both direct and indirect ways.  The power (and limitations) of comparative protein modeling  Online tools that are available to assess protein function rely on a well-resolved crystal structure (which include relevant ligands, such as the high-affinity bound cardiolipin in the case of my study on sculpin COX). Using the bovine protein structure, I was able to isolate the mitochondrially-encoded catalytic core of COX for further in silico investigation of possible interspecific differences in the sculpin proteins that would underlie variation in COX O2 kinetics. Using modeling programs, I generated predicted sculpin COX protein structures to estimate protein stability of the three COX subunits in isolation, study protein subunit interaction, and perform in silico alanine-mutagenesis to investigate differences in cardiolipin and COX3 interactions. The available tools to perform in silico protein analyses will provide increasingly powerful methods to illuminate underlying mechanisms of protein function for empirical testing. Following the discovery of the relationship between COX Km,app O2 and COX3 protein stability, I hypothesized that there would be variation in the second functional role of COX3 which is to protect the microenvironment of the proton acceptor of the proton-pumping D-pathway. I thus designed a protocol to test whether there were differences in proton pumping rate between sculpin species (Fig.5.1). Although there was no evidence of a relationship between COX Km,app O2 and proton pumping rate, it does appear to be interspecies differences in proton pumping rate that would warrant further detailed study (Supplementary Fig.2.2).  It is important, however, to keep in mind the limitations of comparative protein modeling. Although COX is a highly-conserved protein, it is possible that there are minute differences that are lost when using a mammalian protein template to estimate a protein structure. Also, there could be influences of the nuclear-encoded COX subunits on overall COX function that needs to be further 105  investigated4. Thus, while I have isolated a potential mechanism that may underlie interspecies variation in COX Km,app O2, there is still a lot more to learn about adaptive variation in COX function (discussed in Section 5.3.1 under Future Directions).     5.1.2 Hypoxia tolerance in sculpins is not associated with efficient mitochondrial O2 use (Chapter 3) The finding that hypoxia tolerance in sculpins is associated with a higher mitochondrial and COX O2 binding affinity led me to hypothesize that the intact ETS would be more O2 efficient, which would be reflected in increased mitochondrial coupling and phosphorylation efficiency. My in vitro analysis using mitochondria isolated from the brain of various species of sculpin however, did not yield evidence to support my hypothesis. In fact, there was no evidence for an association between interspecific variation in hypoxia tolerance and more efficient O2 use in sculpin brain mitochondria including no relationship between Pcrit and mitochondrial coupling (respiratory control ratio; RCR), leak respiration (state IV respiration), or phosphorylation efficiency (ADP/O; Fig.5.2). Intriguingly, there were significant correlations between complex I and II flux capacities and Pcrit, an indicator of hypoxia tolerance, where species with lower Pcrit had low complex I flux capacity and high complex II flux capacity and the opposite was observed in hypoxia intolerant sculpins. Complex I is a proton pump that contributes directly to the generation of the proton motive force whereas complex II is not a proton pump, thus reduced dependence on complex I and increased dependence on complex II in hypoxia tolerant sculpins, would presumably reduce the magnitude of the protonmotive force and reduce phosphorylation efficiency (Fig.5.2). However, this did not appear to be associated with any differences in phosphorylation efficiency (ADP/O) or mitochondrial coupling (respiratory control ratio; RCR). Complex I is also a major site of mitochondrial ROS generation, thus it is possible that the lower dependency on complex I in hypoxia tolerance sculpins compared with intolerance sculpins could be associated with the need to reduce ROS generation. To investigate whether the variation in complex I was associated with mitochondrial ROS emission, I continued                                                  4 Certain nuclear subunits show tissue-specific expression, such as COX4 in Nile tilapia (Porplycia et al. 2017), COX6a in rats and cattle (Schlerf et al. 1988; Schillace et al. 1994). Also, different isoforms of the COX subunits can be expressed under certain situations, such as the COX4-1 to 4-2 switch observed in mammals under hypoxic conditions (Semenza 2007), although this switch was not observed in hypoxia-exposed Nile tilapia, goldfish, or turtles (Porplycia et al. 2017). The exact functions of the nuclear-encoded COX subunits are still relatively unclear in the literature. 106  with in vitro analysis and also combined it with in vivo experiments to characterize whole animal responses to O2 variability.  5.1.3 Hypoxia tolerant sculpins generate more ROS (Chapters 3&4) Contrary to my expectation that more hypoxia tolerant sculpin that inhabit the more O2 variable environment would generate less ROS, my in vitro and in vivo analyses demonstrate that more hypoxia tolerant sculpins, in fact, generate more ROS compared to less tolerant species. Lower ROS generation would reduce oxidative damage, especially in sculpins that live in O2 variable upper intertidal environments. However, mitochondria from the brain of hypoxia tolerant sculpins generated more ROS/O2 (under state IV conditions) than those from less tolerant sculpins. As the redox environment has been shown to have a profound effect on ROS emission which also likely varies with O2 variability, I also assessed the responses of mitochondria to changes in the extramitochondrial redox environment. While hypoxia tolerant sculpins appeared to buffer matrix glutathione redox status (GSH:GSSG) better, they were generally more sensitive to changes in extramitochondrial GSH:GSSG and had higher ROS/O2 than less tolerant species (Fig.5.2). Also, when challenged with 20min in vitro anoxia-recovery, brain mitochondria from hypoxia tolerant sculpins showed a significant 25% decrease in state III respiration rate, whereas that from the less tolerant species did not show any reduction in state III respiration upon recovery. Interestingly, the post-anoxic reduction in state III respiration rate observed in the hypoxia tolerant sculpin is similar to that observed in typical hypoxia-sensitive mammalian systems. Isolated rat heart mitochondria recovered only 50% of state III respiration after 30min of in vitro anoxia exposure (Shiva et al., 2007) and liver mitochondria recovered 75% after 1 or 10min anoxia exposure (Du et al., 1998). While the anoxia-induced reduction in state III respiration rate in mammals is often attributed to oxidative damage, it may not be the case in sculpin brain mitochondria. It would appear that all three sculpin species reduced ROS/mg protein (in state IV; although not significant in the mid-tolerance species Artedius lateralis), but there was interspecific variation in how this was achieved. In hypoxia tolerant Oligocottus maculosus this may have been due to a general reduction in ETS capacity, as indicated by the reduction in state III respiration rate and state IV ROS/mg protein, whereas in the less tolerant Scorpaenichthys marmoratus there was a change in the ‘efficiency’ of ROS emission, as there was no change in respiration rate but a reduction in ROS/O2. Whether there are advantages to either strategy in response to anoxia-recovery remains to be determined.  107  Although the results from my in vitro work are counterintuitive, this is the first study to compare these responses in a system with closely-related species with the goal of understanding adaptive variation in mitochondrial ROS metabolism. Thus, it sets new directions for further investigation to determine how these variations would give certain species an evolutionary advantage to inhabit O2 variable environments.  Building on results from isolated brain mitochondria in chapter 3, I hypothesized that there would be consistent responses in vivo with whole animal exposures to hypoxia and hyperoxia, and hypoxia tolerant sculpins would show more oxidative damage (estimated as lipid peroxidation levels with TBARS measurements) and generally greater responses in ROS metabolism (GSH redox status, H2O2, and TOSC). I also predicted that there would be tissue-specific responses as tissue vary in metabolic activity and O2 requirements. As ROS generation rates are heavily dependent on mitochondrial energy status, tissues with high metabolic activity, more O2-sensitive, and generally function closer to state III conditions would presumably generate ROS at lower rates than tissues that function closer to state IV resting conditions. To address these hypotheses, I focused on two species that differed in hypoxia tolerance and showed differences in ROS emission (difference in ROS/O2 and response with GSH manipulations; Chapter 3) and exposed both species to hypoxia (at a common PO2 and also relative PO2 exposure to each species’ Pcrit), hyperoxia, and normoxia-recovery and monitored the response in ROS metabolism. The interspecies differences to whole animal exposure to hypoxia and hyperoxia were consistent with what was observed at the mitochondrial level (Chapter 3). In general, O. maculosus were more perturbed by O2 variability than S. marmoratus. Although both species showed changes in GSH:GSSG (though the directionality would suggest regulation of GSH:GSSG), there were no changes in MitoP/MitoB which indicates no accumulation of matrix H2O2 with hypoxia or hyperoxia with normoxic recovery. However, the more hypoxia tolerant O. maculosus showed more ROS effects, showing increases in TBARS and generally similar patterns of TOSC increase in both the brain and liver. Further, there were tissue specific differences in ROS metabolism, where the gills were more resistant to O2 variability, where there were only significant changes in GSH:GSSG but no changes in H2O2, TBARS, and TOSC. Similar patterns of GSH:GSSG change were associated with similar changes in TBARS and TOSC in the brain, but not the liver. This difference in TOSC response to O2 variability suggests tissue difference in regulation of scavenging capacity.   108  Collectively, the results from my in vitro and in vivo studies revealed that hypoxia tolerance in sculpins is associated with more sensitive responses of ROS metabolism. Though the response patterns were tissue specific, changes in GSH:GSSG, TBARS, and TOSC were for the most part quickly reversed with 1hr normoxic recovery. These quick responses were possibly coordinated by modifications of existing proteins in order to scavenge ROS emitted and repair lipid peroxidation.  5.1.4 Summary model: hypoxia tolerant vs. hypoxia intolerant sculpin I have summarized the major findings of this thesis in Figure 5.3 to compare the responses of hypoxia tolerant vs. hypoxia intolerant sculpins. There was evidence of adaptive variation in mitochondrial function in sculpins that inhabit environments that vary in O2 levels. Previous studies determined that hypoxia tolerance was associated with a number of changes in various levels of the O2 transport cascade (at the respiratory surface, in circulation, and capillary diffusion distance; Lui et al., 2015; Mandic et al., 2009; Scott et al., 2011), but we now know that this variation in O2 transport and binding extends to the mitochondrial and COX levels. The differences in COX O2 kinetics were likely associated with interspecific variation in COX3 subunit interaction with cardiolipin, which may subsequently affect the path of O2 travel to the catalytic site. These variations in O2 kinetics however were not related to differences in mitochondrial phosphorylation efficiency and coupling, in fact, the relationship between Pcrit and complex I (lower in hypoxia tolerant sculpin) and complex II (higher in hypoxia tolerant sculpin) was not consistent with increased contribution of protons to the proton gradient. It appears that more O2 in hypoxia tolerant sculpins is used for higher generation of ROS compared to less tolerant sculpins, and overall more sensitive responses to in vitro redox challenge (GSH:GSSG manipulation) and anoxia-recovery challenge. These in vitro responses were consistent with in vivo responses where more hypoxia tolerant species showed higher levels of lipid peroxidation (increase TBARS levels) and increased scavenging capacity (increase in TOSC) compared to the less tolerant species. Given the emerging role of ROS in cell signaling, it is possible that the sensitive responses observed in hypoxia tolerant sculpins may be part of the adaptive strategy of living in the more O2 variable higher intertidal.  5.2 Study Considerations 5.2.1 Interacting abiotic factors in the intertidal  Although I have focused my thesis on O2 variability and the intertidal sculpin model was chosen because they show variability in hypoxia tolerance, sculpins experience changes in other abiotic 109  factors in their environment that may also impose powerful selective pressures and have interacting effects with the putative adaptations to hypoxia. For instance, the fluctuations in temperature on top of variation in PO2 may have an additional impact on the level of available O2. Daytime emergence of tidepools occurs with both warm temperatures and hyperoxia (up to 400% air saturation; Richards, 2011). As discussed in chapter 2, warmer temperatures and the alpha-stat effect may cause a general increase in enzyme Km values and reduce substrate binding, including COX, but hyperoxic conditions may be able to compensate by increasing O2 supply to tissues in order to maintain aerobic metabolism. In order to fully understand the selective pressures of life in the intertidal, the physiological effects of these multiple stressors that mimic the naturally fluctuating patterns in the natural environment need to be characterized.  5.2.2 Technical challenges of assessing ROS metabolism in vivo and in vitro The assessment of ROS formation and accumulation remains challenging. For my in vitro analysis, I chose to use Amplex Ultrared, which is a commonly used fluorophore that reacts with H2O2 in a reaction catalyzed by horseradish peroxidase. Amplex Ultrared, and another commonly used ROS detection method, 2’,7’-dihydrodichlorofluorescein, however, have also been shown to be sensitive to reactive sulfur species and possibly reactive nitrogen species (DeLeon et al., 2016). The non-specificity of the current in vitro tools we have to investigate ROS metabolism makes it challenging to assign causal relationships to ROS specifically. Thus, it is important with existing reactive species detection tools to be aware of the limitations of the technology and to make efforts to confirm the actual role of ROS. For instance, to confirm that H2O2 is the major contributor to the Amplex Ultrared signal by using catalase (which catalyzes the reaction of H2O2 to O2) to remove H2O2 from solution, or by assessing known specific ROS targets, e.g. aconitase activity is reversibly inhibited by superoxide and often also measured to demonstrate the presence of ROS (Armstrong et al., 2004). With their limitations in mind, fluorophores like Amplex Ultrared are useful in vitro tools to monitor ROS emission from isolated mitochondria and cells in real-time. However, they are not ideal for whole animal studies as tissue level responses would be impossible to view in real-time and ROS generation from sampled tissue is hard to assess due to the reactive nature of ROS.  Chemical probes like MitoB offers an effective solution to some of the limitations of the fluorometic ROS detection probes and enable the study of ROS metabolism in vivo. This probe is targeted to the mitochondrial matrix (as it is conjugated to a tetraphenylphosphonium ion; TPP+) 110  and it forms a stable product upon reaction with H2O2. The product MitoP is stable throughout the sampling and processing procedures and thus one can gain an accurate estimation of in vivo ROS accumulation. As previously mentioned, in both brain and liver there were no changes in MitoP/MitoB indicating no accumulation of matrix H2O2, but there were changes in both GSH:GSSG, TBARS and TOSC in response to hypoxia and hyperoxia exposure and normoxic recovery. In other words, there was evidence of ROS effects, but no evidence of the presence of mitochondrial ROS. There are three possible explanations for this observation: (1) that the ROS effects were not caused by ROS generated by ETS sites within the mitochondrial matrix, but instead by sites that face the cytoplasmic side (complex III and GPDH), (2) that the ROS effects were caused by non-mitochondrial sources (e.g. ER, peroxisomes), and/or (3) that MitoB does not work in this teleost study model. The second explanation would be exciting since mitochondria have been presumed to be the major site of ROS generation in response to O2 variability. This possibility is also supported by recent studies that have identified non-mitochondrial sites of ROS generation that can function at high rates (Brown and Borutiate 2012). Whether these non-mitochondrial sites of ROS generation are at play in the sculpins requires further study with careful analysis to isolate which cellular compartments contribute to ROS generation in vivo. The less likely option is the latter one, that MitoB did not work in the sculpin model. As with any new technique, it is important to carefully validate its use in a new animal model. From our validation of MitoB, it appears that MitoB works similarly in sculpins as in murine and drosophila studies (Cochemé et al., 2012; Logan et al., 2014). Although MitoB has been previously used in brown trout, there has yet to be a positive control of MitoB generated in a teleost study model. This positive control can be generated in sculpins using both MitoB and MitoParaquat (a mitochondrial-targeted redox cycler; Mulvey et al., 2017) simultaneously, which would stimulate superoxide generation that is quickly converted to H2O2 in the matrix via superoxide dismutase activity and thus confirm that MitoB can indeed detect H2O2 in the matrix in sculpins.  5.2.3 Multiple factors affect ROS metabolism As illustrated by Aon et al. (2012), multiple factors must be considered when studying ROS emission, including energy status, redox environment, and O2 levels. Depending on the species and tissue type, the relationship of mitochondrial ROS generation and each of these factors may differ (Chapter 3&4; Aon et al., 2010; Munro and Treberg, 2017). Rather than only assessing amounts of ROS emitted under a single condition, I think a more insightful approach for future interspecies 111  comparison of ROS metabolism would be to compare ROS emission kinetics with manipulation of a factor of interest. For instance, I manipulated extramitochondrial GSH:GSSG to investigate the potential differences in the response to an in vitro redox challenge (Chapter 3). Similarly, monitoring ROS emission kinetics simultaneously with O2 kinetics may illuminate potential O2 dependency of ROS emission (Hoffman and Brookes, 2009), and ROS generation under different energy states of mitochondria monitored with state III to state IV transitions (Munro and Treberg 2017).  Moreover, we should also aim to combine observations at multiple levels of biological organization as we continue to understand the subtleties of the interaction between aerobic and ROS metabolism. For example, elasmobranchs generally show reduced antioxidant levels when compared to teleosts (Leveelahti et al., 2014). Thus, by comparing observations from in vitro and in vivo studies in elasmobranchs and teleosts, we can potentially gain further insight into the regulation of antioxidant responses to O2 variability.  5.3 Future Studies 5.3.1 Are there other underlying mechanisms contributing to interspecies variation in COX function?  Other mechanisms beyond those examined in this thesis could potentially contribute to interspecies variation in COX function, one of which is the variation in membrane phospholipids which are intimately linked to membrane proteins and affect protein function. Particularly, cardiolipin differences can have direct effects on mitochondrial enzyme function (such as the effects I described on COX in Chapter 2) and also in the formation of supercomplexes (Claypool, 2009; Zhang et al., 2005). The composition of phospholipids can vary with diet, changing the acyl chain composition of mitochondrial phospholipids, and influence mitochondrial oxidative capacities (shown in rainbow trout Onchorhynchus mykiss; Guderley et al., 2008). Thus, it is important to compare the phospholipid composition between sculpins that vary in hypoxia tolerance. Alternatively, COX from different species can be purified and reconstituted into phospholipid vesicles such that the enzymes would be within a common membrane background and any differences in function (e.g. Km,app O2) can be attributed to the COX enzyme itself.  It would also be interesting to perform more detailed studies on the cardiolipin that binds to the COX3 v-cleft discussed in Chapter 2. The variation in COX3 amino acid residues between species 112  that interact with the cardiolipin would presumably affect the affinity of the cardiolipin to its ligand binding site. However, this is challenging as it only takes one cardiolipin molecule to fill the binding site, and due to its high affinity would be difficult to chemically remove from the protein structure without causing damage to the rest of the enzyme. It would also be interesting to investigate whether the chemical species of that particular cardiolipin molecule differ between sculpins and has functional consequences on COX.  A number of other regulatory factors on COX function require further investigation. The nuclear-encoded subunits of COX are thought to have regulatory roles on overall COX function, so that COX can be regulated under different physiological conditions (e.g. changes in ADP/ATP, or O2) but the specific effects on COX function are unclear (Kocha et al., 2014; Little et al., 2010). Thus, it is important to determine whether sculpins differ in the composition of nuclear-encoded subunits in the multisubunit COX. Also, there are gases that regulate COX function which could play an important role when O2 levels vary. For instance, nitric oxide generated by nitric oxide synthase has been shown to inhibit COX in vivo, which could be a quick method of reversible modification of ETS flux when O2 is scarce (Cooper, 2002). Hydrogen sulfide has been shown to act as a substrate and inhibitor to COX, and accumulate in vivo during myocardial ischemia (Arndt et al. 2017). Investigation in hypoxia tolerant animals is necessary to establish the role of these modulators in response to O2 variability.  5.3.2 What are the functional consequences of the difference in ETS complex flux capacities when comparing sculpins of varying hypoxia tolerance?  Hypoxia tolerant sculpins had higher complex I flux capacity and lower complex II flux capacity compared with less hypoxia tolerant species (Chapter 3). I also observed a relationship between Pcrit and COX activity where more hypoxia tolerant sculpins had more powerful COX, i.e. COX Vmax (which is the product of kcat and enzyme concentration) was lower but COX respiration was higher, compared to less tolerant species (Chapter 2). None of these variations in complex flux capacities appeared to be related to more efficient O2 use in isolated brain mitochondria, which would have been reflected in increased ADP/O and/or reduced ROS generation. To further investigate the role of varying the ETS complex flux capacities on mitochondrial function, it would be interesting to inhibit each of these complexes (I, II, and IV) by titrating low doses of specific inhibitors and assess the subsequent impact on phosphorylation efficiency, membrane potential, O2 kinetics and ROS 113  generation. In particular, I would be interested in the role of COX, given it is well known to be present in excess capacity relative to the other ETS capacities. This excess COX capacity has been thought to be beneficial when O2 supply varies to enhance mitochondrial P50, lower the COX flux control ratio to maintain mitochondrial function under O2 limiting conditions, and potentially alleviate oxidative damage to hyperoxic conditions (discussed in Chapter 2; Campian et al., 2007; Gnaiger et al., 1998; Suarez et al., 1996). The functional consequence of more powerful COX in more hypoxia tolerant sculpins can be investigated with titration of potassium cyanide in order to relate variation to its possible role in whole animal tolerance to environmental O2 variability.  5.3.3 Is there a role of ROS in the adaptive response to environmental O2 variability?  It seems counterintuitive that while hypoxia tolerant sculpins have more efficient O2 transport to mitochondria, and also mitochondria that are more prone to higher ROS emission compared to less tolerant sculpins. Perhaps this is because maintaining aerobic metabolism during O2 variability is more important than ROS metabolism, and having mechanisms to quickly repair oxidative damage (Chapter 4) is the strategy to life in the higher intertidal. It is possible that the increase in ROS emission plays a large part in coordinating cellular responses to O2 variability. ROS has emerged as an important signaling molecule with specific targets (D’Autréaux and Toledano, 2007). For instance, superoxide generated from complex III is thought to stabilize transcription factor HIF-1α (Klimova and Chandel, 2008) and initiate transcription of hypoxia-protective genes. Increases in ROS can also serve as a mechanism to signal changes in cellular O2 levels (Guzy and Schumacker, 2006). Thus, further work is needed to determine whether this increase in ROS emission plays a role in this increased hypoxia tolerance in sculpins. This could potentially be investigated by increasing dietary antioxidants (e.g. vitamin C) in sculpins and observing in vivo responses in ROS metabolism with whole animal exposure to hypoxia and hyperoxia. In vitro studies can also be performed by removing H2O2 that is generated by mitochondria (e.g. by providing an antioxidant in the mitochondrial suspension buffer) and observing performance following O2 insults such as anoxia-recovery exposure (e.g. recovery of state III respiration rate and also changes in ROS emission rate).  5.4 Conclusion  The relationship between mitochondria and O2 is not straightforward. Animals that are frequently exposed to O2 variation face constant pressure to balance both sides of a “double-edged sword”, where on one side O2 is essential to aerobic metabolism and maintenance of cellular energy balance 114  and on the other O2 generates potentially harmful ROS. Although we know that hypoxia tolerant animals or animals that live in O2 variable conditions show adaptive traits to increase O2 movement in the O2 transport cascade and increase O2 delivery to tissues, I have demonstrated in this thesis that this O2 may not be used any more efficiently at mitochondria in hypoxia tolerant sculpins compared to less tolerant sculpins (from the perspective of ADP phosphorylation). In fact, this thesis shows that hypoxia tolerance in intertidal fish is associated with higher ROS generation and generally more sensitive response of ROS metabolism at both the mitochondrial and tissue level.  115    Figure 5.1. Relationship of the various levels of the sculpin oxygen transport cascade (Pcrit, Hemoglobin (Hb) P50, stripped Hb P50, mitochondrial P50 and COX Km,app O2 (Mandic et al. 2009 and 2013) to whole-organism time to loss of equilibrium (LOE). The equations for the OLS regressions are y= -0.0044x +6.03 for Pcrit (circle), y= -0.0078x +7.77 for Hb P50 (square), y=-0.002x +1.34 for stripped Hb P50 (triangle), y=-0.000075x +0.079 for mitochondrial P50 (inverted triangle), and y= -0.000051x +0.047 for COX Km,app O2 (diamond). This figure is also Supplementary Fig 2.1.     116   Figure 5.2. Revised model of brain mitochondria electron transport system associated with hypoxia tolerance in sculpins. More hypoxia tolerant sculpins compared to less hypoxia tolerant sculpins showed (1) increased mitochondrial and COX O2 binding, (2) reduced complex I and increased complex II dependency, which was not associated with any differences in phosphorylation efficiency (3) or proton leak (4). Finally, more hypoxia tolerant sculpins showed increased ROS generation and redox sensitivity compared to less hypoxia tolerant sculpins (5).  117   Figure 5.3. Revised model comparing the oxygen transport cascade between hypoxia tolerant and intolerant sculpins. Detailed description in Section 5.1.4.         118  Bibliography  Ali, S. S., Hsiao, M., Zhao, H. W., Dugan, L. L., Haddad, G. G. and Zhou, D. (2012). Hypoxia-adaptation involves mitochondrial metabolic depression and decreased ROS leakage. PLoS One 7, e36801. Almeida, A., Allen, K. L., Bates, T. E. and Clark, J. B. (1995). Effect of reperfusion following cerebral ischaemia on the activity of the mitochondrial respiratory chain in the gerbil brain. J. Neurochem. 65, 1698–1703. Alnajjar, K. S., Hosler, J. and Prochaska, L. (2014). Role of the N-terminus of subunit III in proton uptake in cytochrome c oxidase of Rhodobacter sphaeroides. Biochemistry 53, 496–504. Alnajjar, K. S., Cvetkov, T. and Prochaska, L. (2015). Role of phospholipids of subunit III in the regulation of structural rearrangements in cytochrome c oxidase of Rhodobacter sphaeroides. Biochemistry 54, 1053–63. Anand, P., Nagarajan, D., Mukherjee, S. and Chandra, N. (2014). ABS–Scan: In silico alanine scanning mutagenesis for binding site residues in protein–ligand complex. F1000Research. doi: 10.12688/f1000research.5165.2 Andrade, T. S., Henriques, J. F., Almeida, A. R., Soares, A. M. V. M., Scholz, S. and Domingues, I. (2017). Zebrafish embryo tolerance to environmental stress factors- Concentration-dose response analysis of oxygen limitation, pH, and UV-light irradiation. Environ. Toxicol. Chem. 36, 682–690. Aon, M. A., Cortassa, S. and O’Rourke, B. (2010). Redox-optimized ROS balance: A unifying hypothesis. Biochim. Biophys. Acta - Bioenerg. 1797, 865–877. Armstrong, J. S., Whiteman, M., Yang, H. and Jones, D. P. (2004). The redox regulation of intermediary metabolism by a superoxide-aconitase rheostat. BioEssays 26, 894–900. Arnarez, C., Marrink, S. J. and Periole, X. (2013). Identification of cardiolipin binding sites on cytochrome c oxidase at the entrance of proton channels. Sci. Rep. 3, 1263. Arndt, S., Baeza-Garza, C.D., Logan, A., Rosa, T., Wedmann, R., Prime, T.A., Martin, J.L., Saeb-Parsy, K., Krieg, T., Filipovic, M.R. and Hartley, R.C., 2017. Assessment of H2S in vivo using the newly developed mitochondria-targeted mass spectrometry probe MitoA. Journal of Biological Chemistry, 292, 7761-7773. Barja, G. (2002). The quantitative measurement of H2O2 generation in isolated mitochondria. J. Bioenerg. Biomembr. 34, 227–33. Bickler, P. E. and Buck, L. T. (2007). Hypoxia tolerance in reptiles, amphibians, and fishes: Life with variable oxygen availability. Annu. Rev. Physiol. 69, 145–170. 119  Bloom, J. D., Labthavikul, S. T., Otey, C. R. and Arnold, F. H. (2006). Protein stability promotes evolvability. Proc. Natl. Acad. Sci. 103, 5869–5874. Boutilier, R. G. and St-Pierre, J. (2000). Surviving hypoxia without really dying. Comp. Biochem. Physiol. A 126, 481–490. Boveris, A. and Chance, B. (1973). The mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen. Biochem. J. 134, 707–716. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. Brand, M. (2000). Uncoupling to survive? The role of mitochondrial inefficiency in ageing. Exp. Gerontol. 35, 811–820. Brand, M.D., 2016. Mitochondrial generation of superoxide and hydrogen peroxide as the source of mitochondrial redox signaling. Free Radical Biology and Medicine, 100, 14-31. Brand, M. D. and Esteves, T. C. (2005). Physiological functions of the mitochondrial uncoupling proteins UCP2 and UCP3. Cell Metab. 2, 85–93. Brand, M. D., Chien, L. F., Ainscow, E. K., Rolfe, D. F. S. and Porter, R. K. (1994). The causes and functions of mitochondrial proton leak. Biochim. Biophys. Acta - Bioenerg. 1187, 132–139. Bratton, M. R., Pressler, M. A. and Hosler, J. P. (1999). Suicide inactivation of cytochrome c oxidase: catalytic turnover in the absence of subunit III alters the active site. Biochemistry 38, 16236–16245. Brown, G. C. and Borutaite, V. (2012). There is no evidence that mitochondria are the main source of reactive oxygen species in mammalian cells. Mitochondrion 12, 1–4. Brown, W. M., George, M. and Wilson, A. C. (1979). Rapid evolution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. 76, 1967–1971. Brunori, M., Antonini, G., Malatesta, F., Sarti, P. and Wilson, M. (1987). Structure and function of cytochrome oxidase: a second look. Adv. Inorg. Biochem. 7, 93–153. Burton, R. F. (2002). Temperature and acid-base balance in ectothermic vertebrates: the imidazole alphastat hypotheses and beyond. J. Exp. Biol. 205, 3587–600. Cadet, J. (2003). Oxidative damage to DNA: formation, measurement and biochemical features. Mutat. Res. Mol. Mech. Mutagen. 531, 5–23. Campian, J. L., Gao, X., Qian, M. and Eaton, J. W. (2007). Cytochrome c oxidase activity and oxygen tolerance. J. Biol. Chem. 282, 12430–8. 120  Castresana, J., Lübben, M., Saraste, M. and Higgins, D. G. (1994). Evolution of cytochrome oxidase, an enzyme older than atmospheric oxygen. EMBO J. 13, 2516. Chance, B., Sies, H. and Boveris, A. (1979). Hydroperoxide metabolism in mammalian organs. Physiol. Rev. 59, 527–605. Chandel, N. S., Budinger, G. R. S. and Schumacker, P. T. (1996). Molecular oxygen modulates cytochrome c oxidase function. J. Biol. Chem.  271, 18672–18677. Chandel, N. S., McClintock, D. S., Feliciano, C. E., Wood, T. M., Melendez, J. A., Rodriguez, A. M. and Schumacker, P. T. (2000). Reactive oxygen species generated at mitochondrial complex III stabilize hypoxia-inducible factor-1α during hypoxia. J. Biol. Chem. 275, 25130–25138. Chapman, L. J. and Mckenzie, D. J. (2009). Behavioral responses and ecological consequences. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P.25-77.  Chapman, L. J., Kaufman, L. S., Chapman, C. A. and McKenzie, F. E. (1995). Hypoxia tolerance in twelve species of east African cichlids: Potential for low oxygen refugia in Lake Victoria. Conserv. Biol. 9, 1274–1288. Chen, Q. and Lesnefsky, E. J. (2006). Depletion of cardiolipin and cytochrome c during ischemia increases hydrogen peroxide production from the electron transport chain. Free Radic. Biol. Med. 40, 976–82. Cheviron, Z. A., Bachman, G. C., Connaty, A. D., McClelland, G. B. and Storz, J. F. (2012). Regulatory changes contribute to the adaptive enhancement of thermogenic capacity in high-altitude deer mice. Proc. Natl. Acad. Sci. 109, 8635–8640. Cheviron, Z. A., Connaty, A. D., Mcclelland, G. B. and Storz, J. F. (2014). Functional genomics of adaptation to hypoxic cold-stress in high-altitude deer mice: Transcriptomic plasticity and thermogenic performance. Evolution 68, 48–62. Chouchani, E. T., Methner, C., Nadtochiy, S. M., Logan, A., Pell, V. R., Ding, S., James, A. M., Cochemé, H. M., Reinhold, J., Lilley, K. S., et al. (2013). Cardioprotection by S-nitrosation of a cysteine switch on mitochondrial complex I. Nat. Med. 19, 753–759. Chouchani, E. T., Pell, V. R., Gaude, E., Aksentijević, D., Sundier, S. Y., Robb, E. L., Logan, A., Nadtochiy, S. M., Ord, E. N. J., Smith, A. C., et al. (2014). Ischaemic accumulation of succinate controls reperfusion injury through mitochondrial ROS. Nature 515, 431–435. Chouchani, E. T., James, A. M., Methner, C., Pell, V. R., Prime, T. A., Erikson, B. K., Forkink, M., Lau, G. Y., Bright, T. P., Menger, K. E., et al. (2017). Identification and quantification of 121  protein S-nitrosation by nitrite in the mouse heart during ischemia. J. Biol. Chem. 292, 14486-14495.  Claypool, S. M. (2009). Cardiolipin, a critical determinant of mitochondrial carrier protein assembly and function. Biochim. Biophys. Acta 1788, 2059–68. Cochemé, H. M., Quin, C., McQuaker, S. J., Cabreiro, F., Logan, A., Prime, T. A., Abakumova, I., Patel, J. V., Fearnley, I. M., James, A. M., et al. (2011). Measurement of H2O2 within living Drosophila during aging using a ratiometric mass spectrometry probe targeted to the mitochondrial matrix. Cell Metab. 13, 340–350. Cochemé, H. M., Logan, A., Prime, T. A., Abakumova, I., Quin, C., McQuaker, S. J., Patel, J., Fearnley, I., James, A., Porteous, C., et al. (2012). Using the mitochondria-targeted ratiometric mass spectrometry probe MitoB to measure H2O2 in living Drosophila. Nat. Protoc. 7, 946–958. Cooper, C. E. (2002). Nitric oxide and cytochrome oxidase: substrate, inhibitor or effector? Trends Biochem. Sci. 27, 33–39. Costa, L. E., Mendez, G. and Boveris, A. (1997). Oxygen dependence of mitochondrial function measured by high-resolution respirometry in long-term hypoxic rats. Am. J. Physiol.-  Cell Physiol. 273, C852–C858. D’Autréaux, B. and Toledano, M. B. (2007). ROS as signalling molecules: mechanisms that generate specificity in ROS homeostasis. Nat. Rev. Mol. Cell Biol. 8, 813–824. DeLeon, E. R., Gao, Y., Huang, E., Arif, M., Arora, N., Divietro, A., Patel, S. and Olson, K. R. (2016). A case of mistaken identity: are reactive oxygen species actually reactive sulfide species? Am. J. Physiol. - Regul. Integr. Comp. Physiol. 310, R549-60. Deutsch, C., Ferrel, A., Seibel, B., Portner, H., and Huey, R.B. (2015). Climate change tightens a metabolic constraint on marine habitats. Science 348, 1132-1135.  Diaz, R. J. and Breitburg, D. L. (2009). The Hypoxic Environment. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P.1–23.  Drechsel, D. A. and Patel, M. (2010). Respiration-dependent H2O2 removal in brain mitochondria via the thioredoxin/peroxiredoxin system. J. Biol. Chem. 285, 27850–8. Du, G., Mouithys-Mickalad, A. and Sluse, F. E. (1998). Generation of superoxide anion by mitochondria and impairment of their functions during anoxia and reoxygenation in vitro. Free Radic. Biol. Med. 25, 1066–1074. 122  Du, S. N. N., Mahalingam, S., Borowiec, B. G. and Scott, G. R. (2016). Mitochondrial physiology and reactive oxygen species production are altered by hypoxia acclimation in killifish (Fundulus heteroclitus). J. Exp. Biol. 219, 1130–1138. Eijsink, V. G. H., Bjørk, A., Gåseidnes, S., Sirevåg, R., Synstad, B., Burg, B. van den and Vriend, G. (2004). Rational engineering of enzyme stability. J. Biotechnol. 113, 105–120. Farrell, A. P. and Richards, J. G. (2009). Defining hypoxia: an integrative synthesis of the responses of fish to hypoxia. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P.487–503. Filho, D. W. and Boveris, A. (1993). Antioxidant defences in marine fish. 2. Elasmobranchs. Comp. Biochem. Physiol. C. 106C, 415–418. Fukuda, R., Zhang, H., Kim, J., Shimoda, L., Dang, C. V and Semenza, G. L. (2007). HIF-1 regulates cytochrome oxidase subunits to optimize efficiency of respiration in hypoxic cells. Cell 129, 111–22. Galli, G. L. J., Lau, G. Y. and Richards, J. G. (2013). Beating oxygen: chronic anoxia exposure reduces mitochondrial F1FO-ATPase activity in turtle (Trachemys scripta) heart. J. Exp. Biol. 216, 3283–3293. Garcia, J., Han, D., Sancheti, H., Yap, L.-P., Kaplowitz, N. and Cadenas, E. (2010). Regulation of mitochondrial glutathione redox status and protein glutathionylation by respiratory substrates. J. Biol. Chem. 285, 39646–39654. Gilland, E., Puka-Sundvall, M., Hillered, L. and Hagberg, H. (1998). Mitochondrial function and energy metabolism after hypoxia-ischemia in the immature rat brain: involvement of NMDA-receptors. J. Cereb. Blood Flow Metab. 18, 297–304. Gnaiger, E., Kuznetsov, A. V., Schneeberger, S., Seiler, R., Brandacher, G., Steurer, W., & Margreiter, R. (2000). Mitochondria in the cold. In: Heldmaier, G. Klingenspor, M. editors. Life in the Cold. Berlin, Heidelberg: Springer. P.431-442.  Gnaiger, E., Lassnig, B., Kuznetsov, A., Rieger, G., Margreiter, R. (1998). Mitochondrial oxygen affinity, respiratory flux control and excess capacity of cytochrome c oxidase. J. Exp. Biol.  201, 1129–1139. Gorbi, S., Pellegrini, D., Tedesco, S. and Regoli, F. (2004). Antioxidant efficiency and detoxification enzymes in spotted dogfish Scyliorhinus canicula. Mar. Environ. Res. 58, 293–297. Gracey, A. Y., Troll, J. V. and Somero, G. N. (2001). Hypoxia-induced gene expression profiling in the euryoxic fish Gillichthys mirabilis. Proc. Natl. Acad. Sci. 98, 1993–1998. 123  Graham, J. (1990). Ecological, evolutionary, and physical factors influencing aquatic animal respiration. Integr. Comp. Physiol. 30, 137-146.  Green, J. M. (1971). High tide movements and homing behaviour of the tidepool sculpin Oligocottus maculosus. J. Fish. Board Canada 28, 383–389. Guderley, H., Kraffe, E., Bureau, W. and Bureau, D. P. (2008). Dietary fatty acid composition changes mitochondrial phospholipids and oxidative capacities in rainbow trout red muscle. J. Comp. Physiol. B 178, 385–399. Gutteridge, J. M. (1995). Lipid peroxidation and antioxidants as biomarkers of tissue damage. Clin. Chem. 41, 1819–1829. Guzy, R. D. and Schumacker, P. T. (2006). Oxygen sensing by mitochondria at complex III: the paradox of increased reactive oxygen species during hypoxia. Exp. Physiol. 91, 807–819. Guzy, R. D., Hoyos, B., Robin, E., Chen, H., Liu, L., Mansfield, K. D., Simon, M. C., Hammerling, U. and Schumacker, P. T. (2005). Mitochondrial complex III is required for hypoxia-induced ROS production and cellular oxygen sensing. Cell Metab. 1, 401–408. Hansen, A. J. (1985). Effect of anoxia on ion distribution in the brain. Physiol. Rev. 65, 101–148. Helling, S., Hüttemann, M., Ramzan, R., Kim, S. H., Lee, I., Müller, T., Langenfeld, E., Meyer, H. E., Kadenbach, B., Vogt, S., et al. (2012). Multiple phosphorylations of cytochrome c oxidase and their functions. Proteomics 12, 950–959. Hickey, A. J. R., Renshaw, G. M. C., Speers-Roesch, B., Richards, J. G., Wang, Y., Farrell, A. P. and Brauner, C. J. (2012). A radical approach to beating hypoxia: depressed free radical release from heart fibres of the hypoxia-tolerant epaulette shark (Hemiscyllum ocellatum). J. Comp. Physiol. B. 182, 91–100. Hilton, Z., Clements, K. and Hickey, A. (2010). Temperature sensitivity of cardiac mitochondria in intertidal and subtidal triplefin fishes. J. Comp. Physiol. B. 180, 979–990. Hofacker, I. and Schulten, K. (1998). Oxygen and proton pathways in cytochrome c oxidase. Proteins Struct. Funct. Genet. 30, 100–107. Hoffman, D. L. and Brookes, P. S. (2009). Oxygen sensitivity of mitochondrial reactive oxygen species generation depends on metabolic conditions. J. Biol. Chem. 284, 16236–16245. Holland, L. Z., McFall-Ngai, M. and Somero, G. N. (1997). Evolution of lactate dehydrogenase-A homologs of barracuda fishes (Genus Sphyraena) from different thermal environments: Differences in kinetic properties and thermal stability are due to amino acid substitutions outside the active site. Biochemistry 36, 3207–3215. 124  Hopkins, S. R. and Powell, F. L. (2001). Common themes of adaptation to hypoxia. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P.153–167.  Hoppeler, H., Kleinert, E., Schlegel, C., Claassen, H., Howald, H., Kayar, S. R. and Cerretelli, P. (2008). II. Morphological adaptations of human skeletal muscle to chronic hypoxia. Int. J. Sports Med. 11, S3–S9. Horvat, S., Beyer, C. and Arnold, S. (2006). Effect of hypoxia on the transcription pattern of subunit isoforms and the kinetics of cytochrome c oxidase in cortical astrocytes and cerebellar neurons. J. Neurochem. 99, 937–951. Hosler, J. P. (2004). The influence of subunit III of cytochrome c oxidase on the D pathway, the proton exit pathway and mechanism-based inactivation in subunit I. Biochim. Biophys. Acta - Bioenerg. 1655, 332–339. Howald, H., Pette, D., Simoneau, J.-A., Uber, A., Hoppeler, H. and Cerretelli, P. (2008). III. Effects of chronic hypoxia on muscle enzyme activities. Int. J. Sports Med. 11, S10–S14. Hylland, P., Milton, S., Pek, M., Nilsson, G.E. and Lutz, P.L., 1997. Brain Na+/K+-ATPase activity in two anoxia tolerant vertebrates: crucian carp and freshwater turtle. Neuroscience letters, 235, 89-92. Irving, E. C., Liber, K. and Culp, J. M. (2004). Lethal and sublethal effects of low dissolved oxygen condition on two aquatic invertebrates, Chironomus tentans and Hyalella Azteca. Environ. Toxicol. Chem. 23, 1561. Ishikawa, T. and Sies, H. (1984). Cardiac transport of glutathione disulfide and S-conjugate. J. Biol. Chem. 259, 3838–3843. Ivanina, A. V. and Sokolova, I. M. (2016). Effects of intermittent hypoxia on oxidative stress and protein degradation in molluscan mitochondria. J. Exp. Biol. 219, 3794–3802. Jamieson, D., Chance, B., Cadenas, E. and Alberto, B. (1986). The Relation of Free Radical Production to Hyperoxia. Annu. Rev. Physiol. 48, 703–719. Jastroch, M., Divakaruni, A. S., Mookerjee, S., Treberg, J. R. and Martin, D. (2011). Mitochondrial proton and electron leak. Essays Biochem. 47, 53–67. Jensen, F. B. and Weber, R. E. (1982). Respiratory properties of tench blood and hemoglobin. Adaptation to hypoxic-hypercapnic water. Mol. Physiol 2, 235–250. Johnston, I. and Bernard, L. (1982). Ultrastructure and metabolism of skeletal muscle fibres in the tench: Effects of long-term acclimation to hypoxia. Cell Tissue Res. 227, 179–199. 125  Jones, D. P. (1986). Intracellular diffusion gradients of O2 and ATP. Am. J. Physiol. - Cell Physiol. 250, C663–C675. Jones, D. P. (2002). Redox potential of GSH/GSSG couple: Assay and biological significance. Methods Enzymol. 348, 93–112. Jones, D. P. (2006). Disruption of mitochondrial redox circuitry in oxidative stress. Chem. Biol. Interact. 163, 38–53. Kim, J., Tchernyshyov, I., Semenza, G. L. and Dang, C. V (2006). HIF-1-mediated expression of pyruvate dehydrogenase kinase: A metabolic switch required for cellular adaptation to hypoxia. Cell Metab. 3, 177–185. Klimova, T. and Chandel, N. S. (2008). Mitochondrial complex III regulates hypoxic activation of HIF. Cell Death Differ. 15, 660–666. Kniesel, U. and Wolburg, H. (2000). Tight Junctions of the Blood – Brain Barrier. Cell. Mol. Neurobiol. 20, 57–76. Knope, M. L. (2013). Phylogenetics of the marine sculpins (Teleostei: Cottidae) of the North American Pacific Coast. Mol. Phylogenet. Evol. 66, 341–349. Knope, M. L., Tice, K. A. and Rypkema, D. C. (2017). Site fidelity and homing behaviour of intertidal sculpins revisited. J. Fish Biol. 90, 341–355. Kocha, K. M., Reilly, K., Porplycia, D. S., McDonald, J., Snider, T. and Moyes, C. D. (2014). Evolution of the oxygen sensitivity of cytochrome c oxidase subunit 4. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 308, R305–R320. Kojer, K., Bien, M., Gangel, H., Morgan, B., Dick, T. P. and Riemer, J. (2012). Glutathione redox potential in the mitochondrial intermembrane space is linked to the cytosol and impacts the Mia40 redox state. EMBO J. 31, 3169–3182. Krab, K., Kempe, H. and Wikström, M. (2011). Explaining the enigmatic KM for oxygen in cytochrome c oxidase: A kinetic model. Biochim. Biophys. Acta - Bioenerg. 1807, 348–358. Krieger, E. and Vriend, G. (2014). YASARA View--molecular graphics for all devices--from smartphones to workstations. Bioinformatics 30, 2981–2982. Kumar, G. K. and Klein, J. B. (2004). Analysis of expression and posttranslational modification of proteins during hypoxia. J. Appl. Physiol.  96, 1178–1186. Kuroda, S., Katsura, K., Hillered, L., Bates, T. E. and Siesjö, B. K. (1996). Delayed treatment with alpha-phenyl-N-tert-butyl nitrone (PBN) attenuates secondary mitochondrial dysfunction after transient focal cerebral ischemia in the rat. Neurobiol. Dis. 3, 149–57. 126  Lane, N. (2006). Power, sex, suicide: mitochondria and the meaning of life. New York: Oxford University Press. Lau, G. Y., Mandic, M. and Richards, J. G. (2017). Evolution of cytochrome c oxidase in hypoxia tolerant sculpins (Cottidae, Actinopterygii). Mol. Biol. Evol. 34, 2153-2162.  Leveelahti, L., Rytkönen, K. T., Renshaw, G. M. C. and Nikinmaa, M. (2014). Revisiting redox-active antioxidant defenses in response to hypoxic challenge in both hypoxia-tolerant and hypoxia-sensitive fish species. Fish Physiol. Biochem. 40, 183–191. Little, A. G., Kocha, K. M., Lougheed, S. C. and Moyes, C. D. (2010). Evolution of the nuclear-encoded cytochrome oxidase subunits in vertebrates. Physiol. Genomics 42, 76–84. Logan, A., Cochemé, H. M., Li Pun, P. B., Apostolova, N., Smith, R. A. J., Larsen, L., Larsen, D. S., James, A. M., Fearnley, I. M., Rogatti, S., et al. (2014). Using exomarkers to assess mitochondrial reactive species in vivo. Biochim. Biophys. Acta - Gen. Subj. 1840, 923–930. Ludwig, B., Bender, E., Arnold, S., Hüttemann, M., Lee, I. and Kadenbach, B. (2001). Cytochrome c oxidase and the regulation of oxidative phosphorylation. ChemBioChem 2, 392–403. Lui, M.A., Mahalingam, S., Patel, P., Connaty, A. D., Ivy, C. M., Cheviron, Z.A., Storz, J. F., McClelland, G. B. and Scott, G. R. (2015). High-altitude ancestry and hypoxia acclimation have distinct effects on exercise capacity and muscle phenotype in deer mice. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 308, R779–R791. Luo, Y., Gao, W., Gao, Y., Tang, S., Huang, Q., Tan, X., Chen, J. and Huang, T. (2008). Mitochondrial genome analysis of Ochotona curzoniae and implication of cytochrome c oxidase in hypoxic adaptation. Mitochondrion 8, 352–357. Lutz, P. L., Nilsson, G. E. and Prentice, H. (2003). The brain without oxygen: Causes of failure molecular and physiological mechanisms for survival. Dordrecht: Kluwer Academic Publishers.  Lyons, T. W., Reinhard, C. T. and Planavsky, N. J. (2014). The rise of oxygen in Earth’s early ocean and atmosphere. Nature 506, 307–315. Magalhães, J., Ascensão, A., Soares, J. M. C., Ferreira, R., Neuparth, M. J., Marques, F. and Duarte, J. A. (2005). Acute and severe hypobaric hypoxia increases oxidative stress and impairs mitochondrial function in mouse skeletal muscle. J. Appl. Physiol.  99, 1247–1253. Mahalingam, S., McClelland, G. B. and Scott, G. R. (2017). Evolved changes in the intracellular distribution and physiology of muscle mitochondria in high-altitude native deer mice. J. Physiol. 595, 4785–4801. 127  Malatesha, G., Singh, N. K., Bharija, A., Rehani, B. and Goel, A. (2007). Comparison of arterial and venous pH, bicarbonate, PCO2 and PO2 in initial emergency department assessment. Emerg. Med. J. 24, 569–571. Mandic, M., Todgham, A. E. and Richards, J. G. (2009). Mechanisms and evolution of hypoxia tolerance in fish. Proc. R. Soc. B Biol. Sci. 276, 735–44. Mandic, M., Speers-Roesch, B. and Richards, J. G. (2013). Hypoxia tolerance in sculpins is associated with high anaerobic enzyme activity in brain but not in liver or muscle. Physiol. Biochem. Zool. 86, 92–105. Mandic, M., Ramon, M. L., Gracey, A. Y. and Richards, J. G. (2014). Divergent transcriptional patterns are related to differences in hypoxia tolerance between the intertidal and the subtidal sculpins. Mol. Ecol. 23, 6091–6103. Mari, M., Morales, A., Colell, A., Garcia-Ruiz, C., Kaplowitz, N. and Fernandez-Checa, J. C. (2013). Mitochondrial glutathione: Features, regulation and role in disease. Biochim. Biophys. Acta - Gen. Subj. 1830, 3317–3328. Martin, W. and Müller, M. (1998). The hydrogen hypothesis for the first eukaryote. Nature 392, 37–41. Mileykovskaya, E. and Dowhan, W. (2014). Cardiolipin-dependent formation of mitochondrial respiratory supercomplexes. Chem. Phys. Lipids 179, 42–48. Mirceta, S., Signore, A. V., Burns, J. M., Cossins, A. R., Campbell, K. L. and Berenbrink, M. (2013). Evolution of mammalian diving capacity traced by myoglobin net surface charge. Science 340, 1234192–1234192. Mitchell, P. (1966). Chemiosmotic coupling in oxidative and photosynthetic phosphorylation. Biol. Rev. 41, 445–501. Mulvey, J., Logan, A., Arndt, S., Pell, V., Caldwell, S., Hartley, R., Murphy, M. and Krieg, T. (2017). 208 Cardioprotection by the mitochondria-targeted superoxide generator MitoParaquat in a murine model of acute myocardial ischaemia reperfusion injury. Heart 103, A138.3-A139. Munro, D. and Treberg, J. R. (2017). A radical shift in perspective: mitochondria as regulators of reactive oxygen species. J. Exp. Biol. 220, 1170–1180. Munro, D., Banh, S., Sotiri, E., Tamanna, N. and Treberg, J. R. (2016). The thioredoxin and glutathione-dependent H2O2 consumption pathways in muscle mitochondria: Involvement in H2O2 metabolism and consequence to H2O2 efflux assays. Free Radic. Biol. Med. 96, 334–346. Murphy, M. P. (2009). How mitochondria produce reactive oxygen species. Biochem. J. 417, 1–13. 128  Murphy, E. and Steenbergen, C. (2008). Mechanisms underlying acute protection from cardiac ischemia-reperfusion injury. Physiol. Rev. 88, 581–609. Natarajan, C., Hoffmann, F. G., Lanier, H. C., Wolf, C. J., Cheviron, Z. A., Spangler, M. L., Weber, R. E., Fago, A. and Storz, J. F. (2015). Intraspecific polymorphism, interspecific divergence, and the origins of function-altering mutations in deer mouse hemoglobin. Mol. Biol. Evol. 32, 978–997. Nikinmaa, M. (2002). Oxygen-dependent cellular functions—why fishes and their aquatic environment are a prime choice of study. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 133(1), 1-16. Nilsson, G.E. (2007). Gill remodeling in fish–a new fashion or an ancient secret? J. Exp. Biol. 210, 2403–2409. Nilsson, G.E. (2001). Surviving anoxia with the brain turned on. Physiology, 16(5), 217-221.Ott, M., Gogvadze, V., Orrenius, S. and Zhivotovsky, B. (2007). Mitochondria, oxidative stress and cell death. Apoptosis 12, 913–922. Pagel, M. (1999). Inferring the historical patterns of biological evolution. Nature 401, 877–884. Pamplona, R. and Costantini, D. (2011). Molecular and structural antioxidant defenses against oxidative stress in animals. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 301, R843–R863. Papandreou, I., Cairns, R. A., Fontana, L., Lim, A. L. and Denko, N. C. (2006). HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab. 3, 187–197. Paradies, G., Petrosillo, G., Pistolese, M., Di Venosa, N., Federici, A. and Ruggiero, F. M. (2004). Decrease in mitochondrial complex I activity in ischemic/reperfused rat heart: Involvement of reactive oxygen species and cardiolipin. Circ. Res. 94, 53–59. Paradis, E., Claude, J. and Strimmer, K. (2004). APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20, 289–290. Pennell, M. W., Eastman, J. M., Slater, G. J., Brown, J. W., Uyeda, J. C., FitzJohn, R. G., Alfaro, M. E. and Harmon, L. J. (2014). geiger v2.0: an expanded suite of methods for fitting macroevolutionary models to phylogenetic trees. Bioinformatics 30, 2216–2218. Perry, S.F., Jonz, M.G., and Gilmour, K.M. (2009). Oxygen sensing and the hypoxic ventilatory response. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P.193-253.  Peter Mitchell (1961). Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nature 191, 144. 129  Petrosillo, G., Ruggiero, F. M., Di Venosa, N. and Paradies, G. (2003). Decreased complex III activity in mitochondria isolated from rat heart subjected to ischemia and reperfusion: role of reactive oxygen species and cardiolipin. FASEB J. 17, 714–6. Pfeiffer, K., Gohil, V., Stuart, R. A., Hunte, C., Brandt, U., Greenberg, M. L. and Schagger, H. (2003). Cardiolipin stabilizes respiratory chain supercomplexes. J. Biol. Chem. 278, 52873–52880. Piantadosi, C. A. and Zhang, J. (1996). Mitochondrial generation of reactive oxygen species after brain ischemia in the rat. Stroke 27, 327–332. Pierron, D., Wildman, D. E., Hüttemann, M., Markondapatnaikuni, G. C., Aras, S. and Grossman, L. I. (2012). Cytochrome c oxidase: Evolution of control via nuclear subunit addition. Biochim. Biophys. Acta - Bioenerg. 1817, 590–597. Pinheiro, J., Bates, D., DebRoy, S. and Sarkar, D. (2014). nlme: linear and nonlinear mixed effects models. R package version 3.1-117. Retrieved March 23rd, 2017, from https://CRAN.R-project.org/package=nlme. Porplycia, D., Lau, G.Y., McDonald, J., Chen, Z., Richards, J.G., Moyes, C.D. (2016). Subfunctionalization of COX4 paralogs in fish. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 312, R671-R680.  Powell, C. S. and Jackson, R. M. (2003). Mitochondrial complex I, aconitase, and succinate dehydrogenase during hypoxia-reoxygenation: modulation of enzyme activities by MnSOD. Am. J. Physiol. - Lung Cell. Mol. Physiol. 285, L189–L198. Prabu, S.K., Anandatheerthavarada, H.K., Raza, H., Srinivasan, S., Spear, J.F., and Avadhani, N.G., (2006). Protein kinase A-mediated phosphorylation modulates cytochrome c oxidase function and augments hypoxia and myocardial ischemia-related injury. J. Biol. Chem. 281, 2061-2070.  Quinlan, C. L., Gerencser, A. A., Treberg, J. R. and Brand, M. D. (2011). The mechanism of superoxide production by the antimycin-inhibited mitochondrial Q-cycle. J. Biol. Chem.  286, 31361–31372. Quinlan, C. L., Perevoschikova, I. V., Goncalves, R. L. S., Hey-Mogensen, M. and Brand, M. D. (2013). The determination and analysis of site-specific rates of mitochondrial reactive oxygen species production. Methods Enzymol. 526, 189-217.  R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Retrieved on March 23rd, 2017, from http://www.R-project.org/  130  Rahman, I., Kode, A. and Biswas, S. (2006). Assay for quantitative determination of glutathione and glutathione disulfide levels using enzymatic recycling method. Nat. Protoc. 1, 3159–65. Reeves, R. B. (1972). An imidazole alphastat hypothesis for vertebrate acid-base regulation: Tissue carbon dioxide content and body temperature in bullfrogs. Respir. Physiol. 14, 219–236. Regan, M. D. and Richards, J. G. (2017). Rates of hypoxia induction alter mechanisms of O2 uptake and the critical O2 tension of goldfish. J. Exp. Biol. 220, 2536–2544. Regan, M. D., Gill, I. S. and Richards, J. G. (2017). Calorespirometry reveals that goldfish prioritize aerobic metabolism over metabolic rate depression in all but near-anoxic environments. J. Exp. Biol. 220, 564-572. Reznick, A. Z. and Packer, L. (1994). Oxidative damage to proteins: Spectrophotometric method for carbonyl assay. Methods Enzymol. 233, 357–363. Richards, J. G. (2009). Metabolic and molecular responses of fish to hypoxia. In: Richards, J.G. Farrell, A.P., Brauner, C.J. editors. Hypoxia. Burlington (MA): Academic Press. P. 443-485.  Richards, J. G. (2011). Physiological, behavioral and biochemical adaptations of intertidal fishes to hypoxia. J. Exp. Biol. 214, 191–9. Richards, J. G., Sardella, B. A. and Schulte, P. M. (2008). Regulation of pyruvate dehydrogenase in the common killifish, Fundulus heteroclitus, during hypoxia exposure. Am. J. Physiol. – Regul. Integr. Comp. Physiol. 295, R979-90. Riistama, S., Puustinen, A., García-Horsman, A., Iwata, S., Michel, H. and Wikström, M. (1996). Channelling of dioxygen into the respiratory enzyme. Biochim. Biophys. Acta - Bioenerg. 1275, 1–4. Rolfe, D. F. and Brown, G. C. (1997). Cellular energy utilization and molecular origin of standard metabolic rate in mammals. Physiol. Rev. 77, 731-758.  Rolfe, D. F. S., Newman, J. M. B., Buckingham, J. A., Clark, M. G. and Brand, M. D. (1999). Contribution of mitochondrial proton leak to respiration rate in working skeletal muscle and liver and to SMR. Am. J. Physiol. - Cell Physiol.  276, C692–C699. Salin, K., Auer, S. K., Rudolf, A. M., Anderson, G. J., Cairns, A. G., Mullen, W., Hartley, R. C., Selman, C., Metcalfe, N. B. (2015). Individuals with higher metabolic rates have lower levels of reactive oxygen species in vivo. Biol. Lett. 11, 20150538.  Salin, K., Auer, S. K., Villasevil, E. M., Anderson, G. J., Cairns, A. G., Mullen, W., Hartley, R. C. and Metcalfe, N. B. (2017). Using the MitoB method to assess levels of reactive oxygen species in ecological studies of oxidative stress. Sci. Rep. 7, 41228. 131  Schafer, F. Q. and Buettner, G. R. (2001). Redox environment of the cell as viewed through the redox state of the glutathione disulfide/glutathione couple. Free Radic. Biol. Med. 30, 1191–1212. Schillace, R, Preiss, T., Lightowlers, R.N., Capaldi, R.A. (1994). Developmental regulation of tissue-specific isoforms of subunit VIa of beef cytochrome c oxidase. Biochim. Biophys. Acta 1188, 391-7.  Schlerf, A., Droste, M., Winter, M. and Kadenbach, B. (1988). Characterization of two different genes (cDNA) for cytochrome c oxidase subunit VIa from heart and liver of the rat. EMBO J. 7, 2387–91. Schoonen, W. G., Wanamarta, A. H., van der Klei-van Moorsel, J. M., Jakobs, C. and Joenje, H. (1990). Respiratory failure and stimulation of glycolysis in Chinese hamster ovary cells exposed to normobaric hyperoxia. J. Biol. Chem. 265, 11118–11124. Schumacker, P. T., Chandel, N. and Agusti, A. G. (1993). Oxygen conformance of cellular respiration in hepatocytes. Am. J. Physiol. - Lung Cell. Mol. Physiol. 265, L395–L402. Schymkowitz, J., Borg, J., Stricher, F., Nys, R., Rousseau, F. and Serrano, L. (2005). The FoldX web server: an online force field. Nucleic Acids Res. 33, W382–W388. Scott, G. R., Egginton, S., Richards, J. G. and Milsom, W. K. (2009). Evolution of muscle phenotype for extreme high altitude flight in the bar-headed goose. Proc. Biol. Sci. 276, 3645–3653. Scott, G. R., Schulte, P. M., Egginton, S., Scott, A. L. M., Richards, J. G. and Milsom, W. K. (2011). Molecular evolution of cytochrome c oxidase underlies high-altitude adaptation in the bar-headed goose. Mol. Biol. Evol. 28, 351–63. Sedlák, E. and Robinson, N. C. (2015). Destabilization of the quaternary structure of bovine heart cytochrome c oxidase upon removal of tightly bound cardiolipin. Biochemistry 54, 5569–5577. Semenza, G. L. (2007). Oxygen-dependent regulation of mitochondrial respiration by hypoxia-inducible factor 1. Biochem. J. 405, 1–9. Seibel, B.A. (2011). Critical oxygen levels and metabolic suppression in oceanic oxygen minimum zones. J. Exp. Biol. 214, 326-336.  Sharma, V., Ala-Vannesluoma, P., Vattulainen, I., Wikström, M. and Róg, T. (2015). Role of subunit III and its lipids in the molecular mechanism of cytochrome c oxidase. Biochim. Biophys. Acta - Bioenerg. 1847, 690–697. Shiva, S., Sack, M. N., Greer, J. J., Duranski, M., Ringwood, L. A., Burwell, L., Wang, X., MacArthur, P. H., Shoja, A., Raghavachari, N., et al. (2007). Nitrite augments tolerance to 132  ischemia/reperfusion injury via the modulation of mitochondrial electron transfer. J. Exp. Med. 204, 2089–2102. Sies, H. (1997). Oxidative stress: oxidants and antioxidants. Exp. Physiol.  82, 291–295. Sies, H. and Akerboom, T. P. M. (1984). [59] Glutathione disulfide (GSSG) efflux from cells and tissues. Methods Enzymol. 105, 445–451. Sims, N. R. and Pulsinelli, W. A. (1987). Altered mitochondrial respiration in selectively vulnerable brain subregions following transient forebrain ischemia in the rat. J. Neurochem. 49, 1367–1374. Skulachev, V. P. (1996). Role of uncoupled and non-coupled oxidations in maintenance of safely low levels of oxygen and its one-electron reductants. Q. Rev. Biophys. 29, 169. Sollid, J., and Nilsson, G.E. (2006). Plasticity of respiratory structures- adaptive remodeling of fish gills induced by ambient oxygen and temperature. Respir. Physiol. Neurobiol. 154, 241-251.  Speers-Roesch, B., Sandblom, E., Lau, G. Y., Farrell, A. P. and Richards, J. G. (2010). Effects of environmental hypoxia on cardiac energy metabolism and performance in tilapia. Am. J. Physiol. – Regul. Integr. Comp. Physiol. 298, R104-19. Speers-Roesch, B., Richards, J. G., Brauner, C. J., Farrell, A. P., Hickey, A. J. R., Wang, Y. S. and Renshaw, G. M. C. (2011). Hypoxia tolerance in elasmobranchs. I. Critical oxygen tension as a measure of blood oxygen transport during hypoxia exposure. J. Exp. Biol. 215, 93-102.  Speers-Roesch, B., Mandic, M., Groom, D. J. E. and Richards, J. G. (2013). Critical oxygen tensions as predictors of hypoxia tolerance and tissue metabolic responses during hypoxia exposure in fishes. J. Exp. Mar. Bio. Ecol. 449, 239–249. St-Pierre, J., Tattersall, G. J. and Boutilier, R. G. (2000a). Metabolic depression and enhanced O2 affinity of mitochondria in hypoxic hypometabolism. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 279, R1205-14. St-Pierre, J., Brand, M. D. and Boutilier, R. G. (2000b). Mitochondria as ATP consumers: Cellular treason in anoxia. Proc. Natl. Acad. Sci.  97, 8670–8674. Starkov, A. A. (2008). The role of mitochondria in reactive oxygen species metabolism and signaling. Ann. N. Y. Acad. Sci. 52, 37–52. Storz, J.F., Sabatino, S.J., Hoffmann, F.G., Gering, E.J., Moriyama, H., Ferrand, N., Monteiro, B. and Nachman, M.W. (2007). The molecular basis of high-altitude adaptation in deer mice. PLoS genetics, 3, e45. 133  Storz, J. F., Runck, A. M., Sabatino, S. J., Kelly, J. K., Ferrand, N., Moriyama, H., Weber, R. E. and Fago, A. (2009). Evolutionary and functional insights into the mechanism underlying high-altitude adaptation of deer mouse hemoglobin. Proc. Natl. Acad. Sci. 106, 14450–14455. Strickler, S. S., Gribenko, A. V, Gribenko, A. V, Keiffer, T. R., Tomlinson, J., Reihle, T., Loladze, V. V and Makhatadze, G. I. (2006). Protein stability and surface electrostatics: a charged relationship. Biochemistry 45, 2761–6. Suarez, R. K., Lighton, J. R., Joos, B., Roberts, S. P. and Harrison, J. F. (1996). Energy metabolism, enzymatic flux capacities, and metabolic flux rates in flying honeybees. Proc. Natl. Acad. Sci. U. S. A. 93, 12616–12620. Todgham, A. E., Iwama, G. K. and Schulte, P. M. (2006). Effects of the natural tidal cycle and artificial temperature cycling on Hsp levels in the tidepool sculpin Oligocottus maculosus. Physiol. Biol. Zool. 79, 1033–1045. Thomson, A. J., Webb, D. J., Maxwell, S. R. J. and Grant, I. S. (2002). Oxygen therapy in acute medical care. BMJ 324, 1406–7. Treberg, J. R., Quinlan, C. L. and Brand, M. D. (2011). Evidence for two sites of superoxide production by mitochondrial NADH-ubiquinone oxidoreductase (complex I). J. Biol. Chem. 286, 27103–27110. Troncoso Brindeiro, C. M., da Silva, A. Q., Allahdadi, K. J., Youngblood, V. and Kanagy, N. L. (2007). Reactive oxygen species contribute to sleep apnea-induced hypertension in rats. Am. J. Physiol. Hear. Circ. Physiol. 293, H2971–H2976. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. and Yoshikawa, S. (1996). The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. Science 272, 1136–1144. Turrens, J. F. (2003). Mitochondrial formation of reactive oxygen species. J. Physiol. 552, 335–344. Tyurina, Y. Y., Tyurin, V. A., Kaynar, A. M., Kapralova, V. I., Wasserloos, K., Li, J., Mosher, M., Wright, L., Wipf, P., Watkins, S., et al. (2010). Oxidative lipidomics of hyperoxic acute lung injury: mass spectrometric characterization of cardiolipin and phosphatidylserine peroxidation. Am. J. Physiol. - Lung Cell. Mol. Physiol. 299, L73-L85.  Verkhovsky, M. I., Morgan, J. E., Puustinen, A. and Wikström, M. (1996). Kinetic trapping of oxygen in cell respiration. Nature 380, 268. Vijayasarathy, C., Damle, S., Prabu, S. K., Otto, C. M. and Avadhani, N. G. (2003). Adaptive changes in the expression of nuclear and mitochondrial encoded subunits of cytochrome c oxidase and the catalytic activity during hypoxia. Eur. J. Biochem. 270, 871–879. 134  Vollmar, B., Burkhardt, M., Minor, T., Klauke, H. and Menger, M. D. (1997). High-resolution microscopic determination of hepatic NADH fluorescence for in vivo monitoring of tissue oxygenation during hemorrhagic shock and resuscitation. Microvasc. Res. 54, 164–73. Watson, J. A., Watson, C. J., Mccann, A. and Baugh, J. (2010). Epigenetics, the epicenter of the hypoxic response. Epigenetics 5, 293–296. Wikström, M., Sharma, V., Kaila, V. R. I., Hosler, J. P. and Hummer, G. (2015). New perspectives on proton pumping in cellular respiration. Chem. Rev. 115, 2196–2221. Wolburg, H., Kastner, R. and Kurz-Isler, G. (1983). Lack of orthogonal particle assemblies and presence of tight junctions in astrocytes of the goldfish (Carassius auratus). A freeze-fracture study. Cell Tissue Res. 234, 389–402. Wu, R. S., Lam, P. K. and Wan, K. (2002). Tolerance to, and avoidance of, hypoxia by the penaeid shrimp (Metapenaeus ensis). Environ. Pollut. 118, 351–355. Yin, S., Xue, J., Sun, H., Wen, B., Wang, Q., Perkins, G., Zhao, H. W., Ellisman, M. H., Hsiao, Y., Yin, L., et al. (2013). Quantitative evaluation of the mitochondrial proteomes of Drosophila melanogaster adapted to extreme oxygen conditions. PLoS One 8, e74011. Zhang, M., Mileykovskaya, E. and Dowhan, W. (2005). Cardiolipin is essential for organization of complexes III and IV into a supercomplex in intact yeast mitochondria. J. Biol. Chem. 280, 29403–8. Zhang, Z.-Y., Chen, B., Zhao, D.-J. and Kang, L. (2013). Functional modulation of mitochondrial cytochrome c oxidase underlies adaptation to high-altitude hypoxia in a Tibetan migratory locust. Proc. R. Soc. B Biol. Sci. 280, 20122758–20122758.      135  Appendix   Supplementary Figure 2.1. Relationship of the various levels of the sculpin oxygen transport cascade (Pcrit, Hemoglobin (Hb) P50, stripped Hb P50, mitochondrial P50 and COX Km,app O2 (Mandic et al. 2009 and 2013). The equations for the OLS regressions are y= -0.0044x +6.03 for Pcrit (circle), y= -0.0078x +7.77 for Hb P50 (square), y=-0.002x +1.34 for stripped Hb P50 (triangle), y=-0.000075x +0.079 for mitochondrial P50 (inverted triangle), and y= -0.000051x +0.047 for COX Km,app O2 (diamond).   136    Supplementary Figure 2.2. Measurement of post-anoxic recovery rate of membrane potential in isolated brain mitochondria of five species of sculpins (data are means ± s.e.m.).  No significant differences between means with one-way ANOVA.137    Supplementary Figure 2.3. Phylogenetic transitions of the important amino acid residues 55 and 224 on COX3 in sculpins (shown with Pcrit (Mandic et al., 2009), mitochondrial P50 (mito P50), and COX Km,app O2 values that show that upper intertidal species have functionally different amino acid residues when compared to lower intertidal species (data are means ± s.e.m.); expanded from Figure 2.3 in the main manuscript.     138    COX1 OLMA ARFE ARLA MYPO BLCI LEAR Interacting residues on SAME CHAIN Interacting residues on DIFFERENT CHAIN position Amino acyl residue (on cottid) Amino acyl residue (on bovine) chain residue chain residue 133 A A S A A A THR124, VAL128, TYR129, PRO130, PRO131, LEU132, ALA133, GLY134, ASN135, HIS138, VAL143, LEU209, ARG213         270 Y H H Y Y Y PRO267, PHE268, GLY269, TYR270, MET271, GLY272, MET273, VAL274, GLN178, TYR179, GLN180, THR181, PRO182, LEU183, TRP186 F SER67     484 A Q N A A A VAL482, LEU283, THR484, VAL485, ASP486 D VAL6, SER15, TYR16, VAL17 M ILE1, THR2 489 T A T T T T LEU487, THR488, THR489, THR490, ASN491 F VAL69, ILE70, TRP71, PHE72      Supplementary Table 2.1. COX1 interspecies residue differences that show changes in amino acid functional groups (chain F= subunit 5b, chain D= subunit 4-1, chain M= subunit 8b); Species abbreviated as follows: O. maculosus= OLMA, A. fenestralis= ARFE, A. lateralis= ARLA, M. polyacanthocephalus= MYPO, B. cirrhosus= BLCI, L. armatus= LEAR   139  COX3 OLMA ARFE ARLA MYPO BLCI LEAR Interacting residues on SAME CHAIN Interacting residues on DIFFERENT CHAIN position Amino acyl residue (on cottid) Amino acyl residue (on bovine) chain residue chain residue 41 T T T T I T SER39, MET40, THR41, LEU42, LEU43, MET44, ILE45 J TYR45     47 T T T T M T MET27, LEU42, LEU43, MET44, ILE45, GLY46, LEU47, THR48, THR49, ASN 50, MET51 G DMU272     55 L L L F F Y ASN50, MET51, LEU52, THR53, MET54, TYR55, GLN56, TRP57, TRP58, ARG59, ASP60, CDL270 J LEU31     115 Y H H C C C PRO108, THR109, LEU112, GLY113, GLY114, CYS115, TRP116, PRO117, PRO118, THR119         119 A T S S A A LEU112, GLY113, CYS115, PRO117, PRO118, THR119, GLY120, ILE121 G TYR50, HIS52, LEU53 H PRO82, GLY83 147 A T T A A A SER143, ILE144, THR145, TRP146, ALA147, HIS148, HIS149, SER150, LEU151, HIS158, MET159, ALA162, LEU163, PHE235         153 A A A E E E HIS149, SER150, LEU151, MET152, GLU153, GLY154, ASP155 G GLY12, ALA13, ARG13     164 A A A A T T MET159, LEU160, GLN161, ALA162, LEU163, PHE164, ILE165, THR166, ILE167, THR168, LEU169, PHE219, CHD271         224 F L L L L L PHE220, ARG221, GLN222, LEU223, LYS224, PHE225, HIS226, ARG156, CDL270 J ASN3, VAL5     225 Y Y Y H H H ARG221, GLN222, LEU223, LYS224, PHE225, HIS226, PHE227, ARG156, CHD271 F GLY5, PRO7, GLN12 J PHE1, GLU2, ASN3, VAL5 230 Q E E E E E THR228, SER229, ASN230, HIS231, HIS232, GLN76 F ASP9     Supplementary Table 2.2. COX3 interspecies residue differences that show changes in amino acid functional groups (chain J= subunit 7a, chain G= subunit 6a, chain H= subunit 6b, chain F= subunit 5b); Species abbreviated as follows: O. maculosus= OLMA, A. fenestralis= ARFE, A. lateralis= ARLA, M. polyacanthocephalus= MYPO, B. cirrhosus= BLCI, L. armatus= LEAR   140  Species Subunit Forward primer Reverse primer Annealing ToC Oligocottus maculosus cox1 5'- CAG CTA AGC GCN CAA ACC AG -3' 5'- GGG TGT AGG CAT CCG GGT AG -3' 60 5'- CAG CTA AGC GCY YWA ACC AG -3' 5'- CTG CRA YRA TRG CRA AKA C -3' 56 5'- CTY CAY GGV GGC TCH ATC AA TG -3' 5'- GCG GWY ATG TGG TTG GCT TGA AA -3' 59 cox3 5'- GCC CAY ACC TTC CCR AAY GA -3' 5'- ATC CGA AGT GGT GTT SSG AKG TA -3' 57 5'- GGC CCA AGC ACA CGC ATA C -3' 5'- GGT CKG GGG YTA TTT GRG GRA G -3' 60      Myoxocephalus polyacantheus cox1 5'- CGC YCA AAC CAG AGA GCW TCC A -3' 5'- GCG CGG GTG TCY ACR TCC ATG -3' 61 5'- GGG GCA GGA ACC GGG TGA AC -3' 5'- GGG GA GGG CAG CCG TGT AGT -3' 64 5'- CTY CAY GGV GGC TCH ATC AA TG -3' 5'- GCG GWY ATG TGG TTG GCT TGA AA -3' 59 cox3 5'- ACG AA YCA RCC AAC VCA YG -3' 5'- GGT CKG GGG YTA TTT GRG GRA G -3' 58 5'- GGC CCA AGC ACA CGC ATA C -3' 5'- GGT CKG GGG YTA TTT GRG GRA G -3' 60      Leptocottus armatus cox1 5'- CCT GWT AAG ACT TGC RGG RKA T - 3' 5'- CTG CRA YRA TRG CRA AKA C -3' 56 5'- CGT GCH TAC TTT ACA TCY GCY ACW A -3' 5'- GGG CGG GAT GCR ATR AAR GC -3' 59 cox3 5'- CMG CCG CCT TYG YYC TTC TMT C -3' 5'- GGA GGC AAG AAG RAC KGC TGT GT -3' 57 5'- GCT TTT TAC CAC GCA AGC CT -3' 5'- ACC GGG TGA TTG GAA GTC AC -3' 56      Artedius lateralis cox1 5'- CGC YCA AAC CAG AGA GCW TCC A -3' 5'- GCG CGG GTG TCY ACR TCC ATG -3' 61 5'- CAC CCY GAA GTY TAY ATY CTY A -3' 5'- CTG CRA YRA TRG CRA AKA C -3' 55 5'- CTY CAY GGV GGC TCH ATC AA TG -3' 5'- GCG GWY ATG TGG TTG GCT TGA AA -3' 59 cox3 5'- CGC CTA CAT AAT CCC CAC ACA AG -3' 5'- CGC TGT GTT GAG GAG GGG GAC TT -3' 61 5'- GGG GCC ATC GCT GCC CTT TT -3' 5'- TGG GGT TTA ACC AAG ACC GGG TG -3' 61           141   Artedius fenestralis cox1 5'- CGC YCA AAC CAG AGA GCW TCC A -3' 5'- GCG CGG GTG TCY ACR TCC ATG -3' 61 5'- GCC GGG GCA TCT GTG GAC TTA AC -3' 5'- GGG GGA GGG CAG CCG TGT AGT- 3' 63 5'- CTY CAY GGV GGC TCH ATC AA TG -3' 5'- GCG GWY ATG TGG TTG GCT TGA AA -3' 59 5'- GGG GCC ATC GCT GCC CTT TT -3' 5'- TGG GGT TTA ACC AAG ACC GGG TG -3' 61 cox3 5'- CTY CAY GGV GGC TCH ATC AA TG -3' 5'- GCG GWY ATG TGG TTG GCT TGA AA -3' 59 5'- CGC TGT GTT GAG GAG GGG GAC TT -3' 5'- CCC GCC CAG GAA TTT TCT AC -3' 64      Blepsias cirrhosus cox1 5'- CAG CTA AGC GCY YWA ACC AG -3' 5'- CTG CRA YRA TRG CRA AKA C -3' 56 5'- CAC GTG CCT ACT TTA CAT CTG -3' 5'- GGG CCA TTA GGG TAA GAC AC -3' 56 cox3 5'- ACG AA YCA RCC AAC VCA YG -3' 5'- TTG GAA GTG GAA YCA GGT TGC -3' 55 5'- CCC CTT TGA AGT CCC YCT BC -3' 5'- CGK AGG GAR AAR GGH ARC C -3' 59 5'- GRC CCA TCA AGC ACA CSC MTA C -3' 5'- CSC CRT CWG CRA TTG TAA AKG G -3' 55  Supplementary Table 2.3. Degenerate and specific primers designed (using GeneTool) to sequence cox1 and cox3 genes from six cottid species.     142   Supplementary Material Figure 3.1. Brain mitochondrial respiration rate (expressed to mg mitochondrial protein) at various steps of substrate-utilization inhibitor titration (SUIT) protocol for Part I in six species of sculpins.  143   Part II ROS/mg protein Species St II (PMG) St II (PMGS) St III St IV Rot O. maculosus 0.42 ± 0.063 2.03 ± 0.31 0.33 ± 0.045 2.11 ± 0.32 1.52 ± 0.19 A. lateralis 0.39 ± 0.068 1.88 ± 0.48 0.40 ± 0.083 1.58 ± 0.12 1.17 ± 0.11 S. marmoratus 0.47 ± 0.069 2.29 ± 0.34 0.39 ± 0.052 1.69 ± 0.19 1.51 ± 0.31  Supplementary Table 3.1. ROS/mg protein values from Experiment Part II in Figure 3.2.    144  Part IV ROS/mg protein  (expressed to normoxic value) Species State III State IV Rot O. maculosus 0.99 ± 0.29 0.56 ± 0.08 0.77 ± 0.12 A. lateralis 0.89 ± 0.18 0.67 ± 0.12 0.64 ± 0.13 S. marmoratus 1.02 ± 0.18 0.89 ± 0.34 0.67 ± 0.16  Supplementary Table 3.2. ROS/mg protein values from Experiment Part IV in Figure 3.4.  145   Supplementary Figure 3.2. Relationship between mitochondrial pellet GSH:GSSG and ROS/mg protein data in O. maculosus (species 1; squares and solid line) and S. marmoratus (species 5; circles and dotted line) from Experiment Part III in Figure 3.3A.  146   Supplementary Figure 4.1. MitoB (top; R2= 1.00) and MitoP (bottom; R2= 0.99) standard curves     147   Supplementary Figure 4.2. Normoxic timecourse over 72hrs for MitoP (top), MitoB (middle), and MitoP/MitoB (bottom) to determine excretion rates; Gill in circle, brain in square, liver in triangle, and muscle in hollow square.  148   Oligocottus maculosus Scorpaenichthys marmoratus Variable Normoxia Hypoxia Hypoxia-Recovery Hyperoxia Hyperoxia-Recovery Normoxia Hypoxia Hypoxia-Recovery Hyperoxia Hyperoxia-Recovery TBARS           Brain 0.31 ± 0.020 0.68 ± 0.11 0.47 ± 0.039 0.42 ± 0.097 0.46 ± 0.026 0.51 ± 0.073 0.39 ± 0.058 0.32 ± 0.021 0.31 ± 0.031 0.33 ± 0.032 Liver 0.24 ± 0.019 0.38 ± 0.045 0.31 ± 0.026 0.35 ± 0.023 0.30 ± 0.025 0.22 ± 0.020 0.23 ± 0.0068 0.21 ± 0.0099 0.20 ± 0.021 0.23 ± 0.013 Gill 1.017 ± 0.26 0.80 ± 0.081 0.87 ± 0.035 0.82 ± 0.26 0.89 ± 0.16 0.39 ± 0.027 0.51 ± 0.099 0.53 ± 0.091 0.40 ± 0.067 0.35 ± 0.043 TOSC           Brain 1.21 ± 0.046 2.02 ± 0.57 1.23 ± 0.060 1.91 ± 0.21 1.65 ± 0.19 1.66 ± 0.18 1.38 ± 0.044 1.39 ± 0.072 1.21 ± 0.041 1.43 ± 0.15 Liver 8.83 ± 0.94 3.12 ± 0.94 5.99 ± 1.27 7.81 ± 1.79 5.37 ± 0.72 5.66 ± 0.52 4.28 ± 0.77 8.55 ± 1.93 12.92 ± 2.79 10.00 ± 1.88 Gill 2.23 ± 0.23 2.10 ± 0.16 2.34 ± 0.058 2.19 ± 0.13 2.59 ± 0.36 1.85 ± 0.18 2.11 ± 0.29 2.32 ± 0.36 1.78 ± 0.25 1.86 ± 0.20  Supplementary Table 4.1. TBARS (in µM) and TOSC levels (in units of catalase activity) to 3.5kPa hypoxia-recovery and 64.0kPa hyperoxia-recovery     

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0357971/manifest

Comment

Related Items