The Evolution of Thermal Compensation in Antarctic Fish Parvalbumins Arthur Carl Whittington

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2011 The Evolution of Thermal Compensation in Antarctic Fish Parvalbumins Arthur Carl Whittington

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] THE FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

THE EVOLUTION OF THERMAL COMPENSATION IN ANTARCTIC FISH

PARVALBUMINS

ARTHUR CARL WHITTINGTON

A dissertation submitted to the Department of Biological Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Summer Semester, 2011

The members of the committee approve the dissertation of Arthur Carl Whittington defended on June 14th, 2011.

______W. Ross Ellington Professor Directing Dissertation

______Timothy S. Moerland Professor CoDirecting Dissertation

______Timothy M. Logan University Representative

______P. Bryant Chase Committee Member

______Gavin J.P. Naylor Committee Member

Approved:

______P. Bryant Chase, Chairperson, Department of Biological Science

The Graduate School has verified and approved the abovenamed committee members.

ACKNOWLEDGEMENTS

I have been very fortunate during my graduate studies to have such a supportive and knowledgeable doctoral committee. I would like to thank my comajor professors, Tim Moerland and Ross Ellington. Tim Moerland accepted me into his lab, introduced me to research and comparative biochemistry, and provided key direction during my dissertation work. He has been an excellent teacher, mentor and friend. Ross Ellington took me on as his last graduate student and has provided me with sound guidance and has made himself available at all times despite his busy schedule. The completion and quality of this dissertation would not have been possible without his participation. Tim Logan and Gavin Naylor have given time and valuable input throughout my doctoral work. I would like to thank Bryant Chase for allowing me to be a foster student in his lab group. He has provided advice on research, writing and on science in general. Florida State University provides an excellent level of support for research through its core facilities. The Analytical lab, Sequencing lab, and Molecular Cloning facility and their personnel have provided valuable assistance throughout this project. All calcium binding constant measurements were made in the Physical Biochemistry Facility of the Institute of Molecular Biophysics. All stoppedflow measurements were performed in the laboratory of Dr. Jonathan Davis at Ohio State Medical Center. I would also like to thank former lab members. Brian Storz is a great friend, family man, and excellent scientist. Neil Sanscrainte and Jennifer Benbow provided friendship and assistance during the early stages of my graduate career. Gregg Hoffman has taught me a great deal about protein expression and purification. Theresa Grove taught me how to sequence my first cDNA and has been an excellent lab mate, collaborator and friend. Danielle SandozOsmus provided assistance during the sequencing phase of the project. The late Bruce Sidell made possible the collection of Antarctic fish specimens that made this project possible as well as provided key input during the lab’s initial work on parvalbumin. I want to give special thanks to George Somero, Donal Manahan and the NSF International Antarctic Biology Training Course for taking me to McMurdo Station, Antarctica. George is an excellent mentor and provided me with one of my most cherished learning and life experiences.

iii

I would like to thank my good friends in Tallahassee and Jacksonville for all their support and providing me with an outlet for fun and relaxation. I am thankful for the special friendships I have made with fellow graduate students here at FSU. My sister Christi Whittington has been a collaborator and one of my best friends. I would like to thank Sarah Patton for her love and friendship over the years. Finally, I want to thank and recognize my parents Cathy and Curtis Whittington for their never ending and unconditional support.

TABLE OF CONTENTS

List of Tables ...... vi

List of Figures...... vii

Abstract...... ix

1. INTRODUCTION...... 1 2. ANCESTRAL SEQUENCE RECONSTRUCTION AND HOMOLOGY MODELING REVEAL POTENTIAL FUNCTIONALLY ADAPTIVE SUBSTITUTIONS IN ANTARCTIC FISH PARVALBUMINS ...... 12

Introduction ...... 12

Materials and methods...... 13

Results ...... 17

Discussion...... 28

3. EXPRESSION, PURIFICATION AND FUNCTIONAL CHARACTERIZATION OF RECOMBINANT PARVALBUMINS...... 32 Introduction ...... 32

Materials and methods...... 34

Results ...... 42

Discussion...... 56

4. CONCLUSIONS...... 60

APPENDIX...... 64

REFERENCES ...... 77

BIOGRAPHICAL SKETCH ...... 84

LIST OF TABLES

Table 2.1 GenBank accession numbers for teleost parvalbumin cDNA sequences...... 15

Table 2.2. Predicted free energy of folding values for homology models of extant and reconstructed PV sequences and predicted effects of virtual mutagenesis...... 26

Table 3.1. Calcium dissociation constants for GBPV WT plus mutants at a range of temperatures...... 51

Table 3.2. Calcium dissociation rates measured by stoppedflow spectrometry using terbium fluorescence as a reporter...... 53

Table 3.3. Calcium onrates for GBPV WT and mutants...... 55

LIST OF FIGURES

Figure 1.1 Corresponding states theory ...... 2

Figure 1.2. Evolutionary compensation of protein function...... 4

Figure 2.1 Antarctic fish parvalbumin protein sequence alignment...... 18

Figure 2.2. Ancestral sequence reconstruction guide tree and method outline...... 19

Figure 2.3. Reconstructed ancestral sequences shows unique substitution pattern between Perciformes and notothenioid ancestral PV sequences...... 21

Figure 2.4. Ambiguity in sequence reconstruction...... 22

Figure 2.5. Sequence alignment of PAPV and NAPV with the extant GBPV...... 23

Figure 2.6. Location of nonconservative substitutions identified by ASR...... 25

Figure 2.7. Homology models showing reintroduction of hydrogen bonds by virtual mutagenesis...... 27

Figure 2.8. Functional tuning by substitutions away from the active site in GBPV WT ...... 29

Figure 3.1. Representative example of fluorescence data collected during a titration of 1.25 M fluo3 and 1.25 M PV ...... 38

Figure 3.2. Representative fluo3 calcium Kd determination curve...... 40

Figure 3.3. Representative PV calcium binding curve ...... 41

Figure 3.4. Representative gel showing analytical scale expression results of GBPV WT...... 45

Figure 3.5. Representative AKTA chromatograph of GBPV WT DEAE ionexchange chromatography ...... 46

Figure 3.6. Representative SDSPAGE gels from ionexchange chromatography ...... 47

Figure 3.7. Representative AKTA results for GBPV WT S200 sizeexclusion chromatography ...... 48

Figure 3.8. Representative SDSPAGE gels from GBPV WT size exclusion chromatography ...49

Figure 3.9. Temperature affects steepness of the PV titration curve...... 50

vii

Figure 3.10. Calcium dissociation constants for GBPV WT and mutants...... 52

Figure 3.11. Calcium dissociation rates, koff, measured as a function of temperature...... 54

Figure 3.12. Proposed mechanism for thermal compensation in PV...... 59

viii

ABSTRACT

Protein function is acutely sensitive to temperature. Investigations of enzymes have demonstrated that the thermal adaptation of proteins is achieved through adjustments in conformational flexibility that lead to changes in protein function. Subtle changes in primary structure drive these adjustments in tertiary structure and function. Antarctic fish of the Perciformes suborder Notothenioidei survive and thrive in subzero temperatures. Key to the evolutionary success of this group is the adaptation of proteins to the cold environment. Previously, it has been shown that parvalbumin (PV), a nonenzymatic, sarcoplasmic calcium buffer, from Antarctic fish shows a characteristic pattern of thermal sensitivity of calcium binding. At common measurement temperatures, PV from Antarctic fish displays a weaker calcium binding affinity than PVs from temperate counterparts, but at physiological temperatures function is highly conserved. Ancestral sequence reconstruction (ASR) was used to probe the evolutionary trajectory of Antarctic fish PV. Homology modeling was used to view the results of ASR in a threedimensional context. The evolutionary and modeling results revealed two substitutions at positions 8 and 26 that are most likely to have shifted the function of PV from a temperate Perciformes ancestor to the present thermal sensitivity pattern of extant notothenioids. We hypothesized that these substitutions caused the evolutionary loss of two hydrogen bonds leading to increased conformational flexibility, which, in turn, compensates for the cold environment of the Southern Ocean. Further, we predicted that these single mutations would cause an intermediate shift in calcium binding affinity while the double mutant would show a full conversion of thermal sensitivity pattern to that of a temperateadapted PV. Functional characterization of the recombinant wildtype protein, two single mutants and double mutant confirmed our hypothesis. Calcium dissociation thermal sensitivity patterns showed an intermediate phenotype for the single mutants and a full right shift to a temperate profile for the double mutant. Furthermore, measurements of calcium rate constants allowed for the development of a structural model, based on the binding energy funnel model, for the shift in calcium binding affinity seen in Antarctic fish PV. Subtle adjustments in the bound and apostate of PV may lead to the displayed shifts in phenotype. This study revealed the underlying evolutionary steps taken to achieve coldadaptation of PV found in Antarctic fish.

CHAPTER ONE

INTRODUCTION

Temperature Sensitivity of Protein Function Temperature influences physiological processes at all levels of biological organization. At the fundamental level these processes are mediated by biomolecular interactions and reactions. All such interactions and reactions (e.g. diffusion, membrane fluidity and protein catalysis/ligand binding) are acutely sensitive to temperature (Hazel and Prosser, 1974). The

effect of temperature on protein function can be described by the temperature coefficient, Q10. A

Q10 of 2, a typical value in enzyme systems measured near physiological temperatures, describes

a doubling of catalytic rate (kcat) or binding (KM) constant (binding constants being equivalent to o the ratio of ligand offrate, koff, to onrate, kon) for every 10 C rise or fall, respectively, in temperature. For enzyme activity, the decrease in kcat with decreasing temperatures can be attributed to a decrease in enthalpy in the system. As the kinetic energy of the system decreases, a smaller fraction of the population of enzymesubstrate complexes will have sufficient energy to overcome the activation energy barrier.

For binding constants, the decreased KM, or in the case of this study Kd, can be explained by the following functional rationale. Proteins exist as a series or ensemble of conformations. At a given temperature only a subset of these conformations will have a conformational structure that is bindingcompetent. As temperature decreases, the protein will sample less conformational space resulting in a higher proportion of molecules in a binding competent state. It is thought that at physiological temperature enzymes from organisms adapted to different thermal habitats sample a similar subset of conformational space. This “corresponding states” model predicts that differentially adapted orthologs (homologous protein isoforms separated by speciation) will sample a similar subset of conformational states at their physiological temperature, giving proteins the appropriate amount of flexibility to provide the optimal level of function (Zavodszky et al., 1998; Hochachka and Somero, 2002). This is illustrated in Figure 1.1. Two hypothetical orthologous proteins “adapted” to different temperatures are shown occupying a roughly Gaussian distribution of conformational microstates. At physiological temperatures the populations of protein molecules inhabit a similar subset of conformational space leading to similar Km values.

Figure 1.1. Corresponding states theory (Somero, 1978, 1995). Proteins exist in a series of conformational microstates with a subset of the population in a binding competent state (grey boxes). At physiological temperatures orthologs from different thermally adapted organisms will sample a similar subset of conformational space and display a similar level of function.

In general, proteins are marginally stable. The difference in free energy between the folded and unfolded forms of a protein is equivalent to just a few weak, noncovalent interactions (25 kcal/mol) which dominate the folding, stability and function of proteins. There are two main categories of these weak interactions: electrostatic interactions (including ionpairs, hydrogen bonds, and van der Waals interactions) and hydrophobic interactions. These interactions exert their effects through interresidue contacts between amino acids within the protein, through solvent interactions and also between active site residues and substrates during ligand binding. Each type of noncovalent interaction is sensitive to temperature. In free energy terms, thermal sensitivity changes in equilibrium binding constants are driven by changes in net enthalpy change. A binding event with a negative enthalpy change will be promoted by decreased temperature while an event with a positive enthalpy change will be inhibited. Similarly, weak interactions with a negative enthalpy of formation (i.e. van der Waals, hydrogen bond, and ionic interactions) are strengthened by decreasing temperature while those with a positive enthalpy of formation (i.e. hydrophobic interactions) are disrupted with decreasing temperature (Somero and Low, 1977; Hochachka and Somero, 2002). The above explanation gives a framework for how organisms can evolve to fill a new thermal habitat while maintaining an optimal level of function/metabolism. Figure 1.2 graphically shows the competing effects of temperature and evolutionary adjustments on conformational flexibility. As a group of organisms evolves to fill a new thermal niche, adjustments in protein flexibility compensate for the function altering effects of temperature. Over evolutionary time scales thermal compensation is accomplished through adjustments in protein primary structure. It appears that these compensating amino acid substitutions are generally excluded from the active site residues which tend to be conserved to sustain substrate specificity (Wilks et al., 1988; Golding and Dean, 1998). A recent example of this principle is provided by the work of Dong and Somero (2009) on malate dehydrogenase (MDH) in a series of congeneric, intertidal marine molluscs. MDH substrate binding ability,

measured as KM for NADH, correlated tightly with the latitudinal position of a given species as well as vertical position in the intertidal zone. At common measurement temperatures, the KM values were different, but at physiological temperatures they fell within a narrow range indicating that function is conserved in the normal range of habitat temperatures for the species.

Figure 1.2. Evolutionary compensation of protein function. Temperature and adaptation have competing effects on protein flexibility and function. Evolutionary adjustments in conformational flexibility compensate for the impact of temperature change on protein function.

The differences in thermal sensitivity between protein variants from the more northern and southern species is attributable to one amino acid difference found away from the active site residues. The southern species, which experiences warmer temperatures, has more hydrogen bonding potential at a position distant from the substrate binding site, presumably leading to less conformational flexibility. This general trend of subtle changes in primary structure correlating with conservation of function at physiological temperature is mirrored in work on other enzymes (Holland et al., 1997; Fields and Somero, 1998; Johns and Somero, 2004). It is important to note that conformational flexibility is a phenomenon distinct from thermal stability. Thermal stability, or macrostability, is a protein’s resistance to denaturation in the face of temperature or chemical denaturants. Thermal stability maintains the integrity of protein tertiary structure. Conformational flexibility, or microstability, is somewhat of a catch all term describing a variety of protein movements including but not limited to breathing motions, loop movements, and transiently unfolded local regions all of which affect the rigidity of the folded protein structure (Privalov and Tsalkova 1979; Vihinen 1987; Zavodszky et al., 1998). While both correlate with an organisms physiological temperature, it has been shown experimentally that these two parameters can in fact be decoupled (Miyazaki et al., 2000; LeMaster et al., 2005). For proteins from coldadapted organisms, the characteristic thermolability seems to be a result of a lack of selection for thermal stability as these proteins never experience temperatures that would cause denaturation. Changes in conformational flexibility, however, are selected for to maintain optimal protein function (Fields and Somero, 1998; Fields et al., 2001). Flexibility has been probed experimentally by a variety of techniques including acrylamide quenching of tryptophan fluorescence (Varley and Pain, 1991) and hydrogen deuterium exchange as measured by Fouriertransformed infrared resonance spectroscopy (Zavodszky et al., 1998; Svingor et al., 2001; Hadju et al., 2008). These techniques investigate the solvent interaction of buried residues that may be exposed by transient unfolding of local regions of the protein. This technique limits the definition of flexibility, however, as it measures only one aspect of flexibility. Caution is required in interpreting these data as a proxy for whole molecule flexibility. Thermal adaptation through optimization of conformational flexibility can be achieved through a variety of structural adjustments such as changes in hydrophobic packing density,

solvent interaction of surface residues, and optimization of electrostatics. Comparative molecular dynamics studies provide a method to monitor the motions of all the atoms of a protein which allow the discovery of the exact structural mechanisms required for adaptation (Brandsdal et al., 1999; Papaleo et al., 2006, 2007, 2008; Xie et al., 2009; Tiberti and Papaleo, 2011). Work on enzymes by Somero and colleagues has elucidated much of what is known about the thermal sensitivity of protein function. Much less is known about the origin and nature of thermal adaptation in noncatalytic proteins. Recently, the calcium (Ca2+) binding protein parvalbumin (PV) was shown to display thermal sensitivity patterns similar to what has been found in enzymes (Erickson et al., 2005; Erickson and Moerland, 2006). PV isoforms from 2+ Antarctic species show a weaker binding affinity, measured as Ca Kd, than temperate counterparts at a common measurement temperature, but show similar binding abilities at physiological temperature (Erickson et al., 2005). Additionally, a fish from the polar region of the Northern Atlantic Ocean, Boreogadus saida (of the teleost order Gadiformes) expresses a PV with an equivalent thermal sensitivity pattern to that of PV found in Antarctic fish (Erickson and Moerland, 2006). The finding of convergent evolution of PV function in disparately related fish suggests that the thermal sensitivity pattern found in Antarctic fish PVs (Erickson et al., 2005) is not a phylogenetic effect, but is an evolutionary response to environmental temperature. The polar and temperate species studied by Erickson et al. (2005) do not fall into the same style of study system as the enzyme work mentioned above. Instead of being congeneric or confamilial species, these polar (Gobionotothen gibberifrons) and temperate (Cyprinus carpio) species are highly divergent. Their last common teleost ancestor probably lived during the Triassic 200250 million years ago (Patterson, 1993; Palmer, 1999; Nelson, 2006). Over this evolutionary distance neutral substitutions build up and can confound the typical simple sequence alignment analysis searching for functionally adaptive substitutions (Bae and Phillips, 2004). A major goal of this work is to find a generally applicable way to search for these adaptive amino acid changes against the background of ‘phylogenetic noise.’

Parvalbumin Parvalbumins are small (MW ~ 1012 kD), acidic (pI ~ 35), vertebratespecific proteins found abundantly in the cytosol of fast twitch skeletal muscle (Rall, 1996). It is especially abundant in the white muscle of fish. PV is a member of the EFhand family of calcium binding

6 proteins and contains two of the eponymous domains that are functional and able to bind calcium, and, with a lesser affinity, magnesium. A third domain is thought to be an ancestral EF hand that has lost function (Kawasaki et al., 1998). As calcium is a master regulator of cellular function, EFhand proteins play a critical role in physiology (McPhalen et al., 1991). The diverse array of processes regulated by EFhand proteins is facilitated by the wide range of Ca2+ binding abilities displayed by the various members of this family which span six

orders of magnitude of Kd values (Falke et al., 1994). This group of proteins has been extensively studied, and while knowledge of thermal sensitivity is relatively lacking there is a large body of literature concerning structurefunction relationships in EFhand proteins laying the groundwork for structure/function studies of PV. Proteins of the EFhand family consist of one or more pairs of the helixloophelix binding motif. The calcium coordinating residues are contained in a conserved twelve residue loop at relative positions 1, 3, 5, 7, 9, and 12 that correspond to a coordinate system: X, X, Y, Y, Z, Z binding in a pentagonal bipyramidal arrangement (LewittBentley and Rety, 2000). PV contains three EFhand motifs named after the helices they contain: AB, CD, and EF. The CD and EF domains are the functional ionbinding domains (Pauls et al., 1996). The AB domain is nonfunctional, and is considered to be the remnants of an ancestral binding site that has lost function due to the loss of its paired EFhand and two residues in its loop region (Cox et al., 1999). The AB domain forms a hydrophobic patch that covers the back of the CD and EF domains, and it appears to play a role in modulating ionbinding affinity as it provides key stabilization to the protein (Permyakov et al., 1991). Substitutions of active site residues appear to be the primary mechanism for changing binding affinity and kinetics among different subfamilies of EFhand proteins, (e.g. calmodulin, troponin C [TnC], PV, etc). This agrees with the work on enzymes mentioned above which demonstrated that substitutions outside of protein active sites are generally responsible for environmental adaptation and functional tuning of enzymes (Fields and Somero 1998; Hochachka and Somero 2002; Dong and Somero 2009). It reinforces the principle that small changes in weak interactions can change the net enthalpy change of a protein binding event, thus adjusting the function and thermal sensitivity. Parvalbumin coordinating residues are almost completely conserved from fish to humans with most variability occurring at the Y position, which coordinates with its backbone oxygen

and not its side chain. Analysis of an alignment of PVs from a variety of vertebrates reveals that most substitutions are found in the AB domain away from the binding loops. Evidence points to this region being important for modulating PV affinity. Remodeling of the AB domain in an attempt to restore its ancestral binding ability was not successful, but it did show that the AB domain is very important in stabilizing the structure of PV. In the engineered mutant of PV the αhelical content was reduced to 2/3 of the wild type which also reduces the ion binding affinity (Cox et al., 1999). Henzl et al. (2004) used mixed complexes of fragments from two different PV lineages (α and β) known to have different ionbinding abilities, and found that heterologous complexes (e.g. αPV AB complexed with βPV CD/EF) produced different binding affinities, which were higher than homologous complexes in some cases. This suggests that differences in the AB domain sequence can modulate affinity in the functional binding sites (Henzl et al., 2004). Recent work by Henzl and colleagues has also shown that for rat PV the ionbinding ability is quite sensitive to small sequence changes away from the active site. By altering the interactions and movements of the hydrophobic core, these substitutions affect the function of the highly conserved active sites (Henzl et al., 2008). Physiological PV function reflects its ability to reversibly bind ionic calcium and magnesium. Evidence has accumulated showing a strong correlation between muscle relaxation rate and PV concentration (Heizmann et al., 1982). Significantly longer contractionrelaxation cycles were found for fast twitch muscles of PV knockout mice (Schwaller et al., 1999), while relaxation speed is enhanced in rat slow twitch muscle by direct injection of PV cDNA (Muntener et al., 1995). This function is mimicked by injection of EDTA, a reversible calcium buffer, into muscle cells that normally do not express PV (Johnson et al., 1999). The consensus view is that PV, due to its ability to bind Ca2+, acts to promote relaxation from the active state in rapidly contracting muscles by allowing faster dissocation of Ca2+ from TnC during the relaxation phase of the contraction/relaxation cycle (Rall, 2005). From the above evidence for PV function we can make a general speculation of the importance of PV in the ecology of fish: Faster contraction/relaxation cycles due to the presence of PV would lead to faster burst swimming in fish. A fish with a higher speed of burst swimming is better able to catch prey and evade predators, providing a selective advantage to the fish with PV in its fasttwitch muscle, and certainly this particular aspect of its physiology would need to be tuned to its thermal habitat.

Parvalbumin provides an excellent model protein for studies of structure/function relationships, and adaptation to environmental variables. It is small, monomeric, highly soluble and relatively easily purified. Other EFhands play roles as regulatory proteins, but PV function seems to be solely dependent on its ability to reversibly bind divalent cations. Additionally, it has a minimum amount of posttranslational modification; at the most the first methionine is removed and the protein is acetylated. Thus, folding of PV and tertiary structure are determined by its primary structure.

Antarctic Fish Poikilotherms (organisms whose body temperature conforms to ambient temperature) are susceptible to environmental temperature fluctuations which may vary on a variety of timescales. With only a few exceptions the body temperatures of teleost fish are the same as the water in which they live, making this group of organisms an excellent study system for comparative studies on the effects of temperature on biological function (Hazel and Prosser, 1974). This gives the opportunity to study temperature effects in organisms living between 1.86°C (for fish on the Antarctic shelf) (Eastman, 1993) to nearly 40°C (for fish living in lakes in East Africa) (Lowe McConnell, 1987). This study focuses on the diverse Perciformes suborder Notothenioidei. This group of Antarctic fish, which dominates the fish fauna in the Southern Ocean, has evolved from a benthic ancestor to fill a variety of niches in a unique environment. Antarctica and its surrounding waters have been isolated for around 25 million years (my) since the opening of the Drake Passage and the formation of the Antarctic Circumpolar Front. Water temperatures in this region have been below 5°C for the last 14 my, and are now extremely stable. On the Antarctic Peninsula, temperatures range from 1.8°C in winter to only 1.5°C in the summer while on the Ross Ice Shelf the sea temperature is a constant 1.86°C year round (Eastman, 1993). Notothenioid fishes are extreme stenotherms that have developed a suite of adaptations for living in the frigid waters of their habitat, including antifreeze glycoproteins and metabolic proteins tuned to function optimally in these extremely cold waters (Sidell, 2000; Petricorena and Somero, 2007). The diverse and endemic notothenioids provide an excellent study system to investigate thermal adaptation at the protein level.

Antarctic Fish Parvalbumins and Cold-adaptation Based on the previous work on PV discussed above (Erickson et al., 2005, Erickson and Moerland, 2006), we have sought to identify the amino acid substitutions that have occurred during the evolution of Antarctic fish that lead to the characteristic polar thermal sensitivity pattern of notothenioid PV function. The polar phenotype is represented by Gobionotothen gibberifrons PV (GBPV) and temperate by Cyprinus carpio PV (CCPV). G. gibberifrons expresses one isoform of PV which was used in the work of Erickson et al. (2005). C. carpio expresses multiple isoforms and the major isoform was used. Simple sequence comparison between these two PV orthologs reveals 23 substitutions between them. In the present effort we used ancestral sequence reconstruction (ASR) to narrow this set of substitutions down to functionally significant substitutions. Using this approach, we mapped orthologous PV sequences onto an evolutionary guide tree reflecting currently understood teleost systematics. Ancestral PV sequences were then estimated for the branch points of the tree. Comparison of the notothenioid ancestral PV sequence with the PV from the Perciformes ancestor, a fish thought to have lived in warm/temperate seas, showed 16 substitutions in the notothenioid ancestral PV that would have occurred during the invasion and subsequent cooling of the Southern Ocean. Next, the structural effects of the 8 nonconservative substitutions revealed by ASR were investigated through use of homology models of the notothenioid and Perciformes ancestral PVs. The free energy of folding was calculated for extant and ancestral PV structures. This metric correlated well with habitat temperature, with warm/temperate PVs having a more stable structure than forms from coldadapted fish. Utilizing virtual site directed mutagenesis, the substitutions identified by ASR were introduced into the wildtype recombinant G. gibberifrons PV (GBPV) in order to reverse engineer the structure of the Perciformes ancestral PV. Two substitutions reintroduced hydrogen bonds into the GBPV scaffold and increased the calculated stability approximately half way to that of the Perciformes ancestral PV and other extant warm/temperate isoforms. The mutant GBPV containing both hydrogen bonds displayed a full conversion to the ancestral warm/temperate structure. Based upon these results, we hypothesized that the two individual substitutions identified by ASR and homology modeling would show a phenotype intermediate between that of the Antarctic fish PV and the temperate fish PV. Further, the double mutant would show a full conversion of phenotype from polar to temperate. Our model was then used to guide sitedirected

mutagenesis and subsequent expression and purification of PV from G. gibberifrons, a notothenioid common on the Antarctic Peninsula. The calcium binding function of the expressed wild type PV and mutant constructs was measured using a fluorescence based competitive binding assay. The binding assay revealed that individually these two substitutions gave an intermediate phenotype. Moreover, the double mutant yielded the full temperate phenotype. Additionally, Ca2+ offrates were determined using stoppedflow spectrometry with terbium fluorescence used as a reporter of ligand binding. With these steadystate and kinetics data we developed a statistical thermodynamic model using a binding energy funnel to describe the nature of the structural changes that alter the thermal sensitivity pattern of the PV constructs. The data suggest that the loss of one hydrogen bond from the temperate Perciformes ancestral PV caused a shift in Ca2+ offrate destabilizing the bound form of PV. Then the loss of a second hydrogen bond caused a shift in Ca2+ onrate destabilizing the apostate. These substitutions identified in this work are sufficient to explain the current polar thermal sensitivity pattern of Antarctic notothenioid PV function.

CHAPTER TWO Ancestral Sequence Reconstruction and Homology Modeling Reveal Potential Functionally Adaptive Substitutions in Antarctic Fish Parvalbumins

Introduction

As discussed in Chapter 1, a body of evidence has accumulated demonstrating that the ability of organisms to adapt to a wide variety of thermal habitats relies in part on evolutionary adjustments to protein structure (Moerland, 1995; Hochachka and Somero, 2002). Subtle adjustments in protein primary structure functionally compensate for the effects of temperature. For proteins involved in exothermic processes, such as parvalbumin (PV) Ca2+ binding, decreased temperature promotes ligand binding (Somero and Low, 1977). In the case of Antarctic fish PV functioning at 1.86°C, without compensatory sequence changes the dissociation constant (Kd) would be lower than PV from the temperate ancestor of notothenioids. This could potentially result in a deleterious decrease in burst swimming speed as PV with too high of a Ca2+ affinity could overbuffer Ca2+ causing decreased force production. Antarctic fish

PVs, however, do not have a lower Kd than temperate counterparts when measured at physiological temperature. Conservation of function in PVs from Antarctic and temperate fish has been demonstrated previously (Erickson et al., 2005). At physiological temperatures PVs from Antarctic and temperate fish show a similar level of function (Erickson et al., 2005). Thus, we would expect to find a small subset of compensatory sequence differences to be responsible for the observed thermal sensitivity of Antarctic fish PV. It is difficult to discern the evolutionary steps taken in the process of thermal adaptation when comparing disparately related organisms. Neutral substitutions accumulate over evolutionary time masking the few functionally adaptive substitutions (Kimura, 1983). A sequence comparison of a representative Antarctic PV from the notothenioid Gobionotothen gibberifrons (GBPV) and PV from a temperate counterpart, Cyprinus carpio (CCPV), whose last common teleost ancestor lived 200250 million years ago (mya) (Patterson, 1993; Palmer, 1999;

Nelson, 2006) reveals 23 amino acid differences. Ancestral sequence reconstruction (ASR) provides a method to identify the functionally adaptive substitutions above the background of neutral substitutions that accumulate in orthologous protein sequences from disparately related organisms (Zuckerkandl and Pauling, 1965; Thornton, 2004). Indeed, ASR has been used successfully to investigate environmental adaptation (Thomson et al., 2005; Gaucher et al., 2008; Yokoyama et al., 2008). By mapping orthologous sequences to an evolutionary tree and estimating the ancestral sequences at the nodes of the tree, the course that evolution has taken to achieve adaptation by the extant Antarctic notothenioids to the frigid Southern Ocean can be visualized. While ASR can provide a subset of residues to target, the subset may still be too large to effectively consider during mutagenesis studies. In this case 16 substitutions were found between the reconstructed Perciformes ancestral PV (PAPV) and the notothenioid ancestral PV (NAPV). Critical to determining potential effects of these substitutions is viewing them in their three dimensional context rather than just scanning the primary structure. Homology modeling allows us to construct 3D models of PV orthologs and pinpoint which of the substitutions identified by ASR are most likely to affect function. We identified two substitutions, K8N and K26N, that we hypothesize are sufficient to explain the adaptation of GBPV and, by proxy, all Antarctic notothenioids, to cold. Homology modeling indicated that these two substitutions constitute an evolutionary loss of two hydrogen bonds. Increased conformational flexibility of the Antarctic fish PVs missing the two hydrogen bonds is suggested by folding energy estimates. As discussed previously, a decrease in net

enthalpy change due to the loss of two hydrogen bonds would lead to a higher apparent Kd in the Antarctic fish PV which corresponds to observations in the literature (Erickson et al., 2005). ASR and homology modeling provided a guide for sitedirected mutagenesis of the representative notothenioid PV GBPV. We expected that reintroduction of the hydrogen bonds would convert the thermal sensitivity profile of GBPV to that of its temperate counterpart CCPV.

Materials and Methods

Animals

Samples of Antarctic fish Chaenodraco wilsoni, Champsocephalus gunnari, Chionodraco rastrospinosus, Dissostichus mawsoni, Gobionotothen gibberifrons, Lepidonotothen nudifrons, Notothenia coriiceps, Notothenia rossi, Parachaenicthys charcoti, Patagonotothen ramsayi, Pseudochaenichthys georgianus, and Trematomus hansoni were collected by trawl or trap from the ARSV Laurence M. Gould in Dallman Bay, Antarctica in 2003. Dissostichus eleginoides samples were purchased from a local fish market. All fish were collected according to a protocol approved by the Animal Care and Use Committee of Florida State University (FSU) (Protocol #9304).

Sequencing In order to perform ASR for Antarctic fish PV, it was necessary to enrich the PV sequence database. G. gibberifrons skeletal muscle total RNA was isolated using the Trizol reagent protocol (Invitrogen). RACE ready cDNA was generated using the BD SMART RACE kit (Clontech). cDNA was amplified using gene specific primers (For: 5’ TTGCACAGCTGCTGAGTCTTTCAAGC3’, Rev: 5’ CAGGAAAGCCTTGGTCTCAGCATCAG3’) (PCR protocol: 94°C for 2:00 min. 35 cycles of 94°C for 45 sec, 66°C for 1:00 min, 72°C for 45 sec. 72°C for 5:00 min). These primers were designed from the Chaenocephalus aceratus PV cDNA sequence (Hendrickson, 2005). The RACE fragments were gel purified (Qiagen) and subcloned according to the TOPO TA Cloning protocol (Invitrogen). Positive clones were isolated by miniprep (Qiagen) and submitted for sequencing by the DNA Sequencing Laboratory housed in the Department of Biological Science, FSU. Primers were then designed for the 5’ and 3’ untranslated regions (For: 5’ CAACTGAACGAATCCACTCTAGTCT3’, Rev: 5’CAGTCAGGGGAGTGAGAGGTC3’) (PCR protocol: 94°C for 2:00 min. 35 cycles of 94°C for 45 sec, 65°C for 1:00 min, 72°C for 45 sec. 72°C for 5:00 min). PCR was used to amplify the full coding region of the PV gene which was then cloned and sequenced as above. These primers were used successfully to obtain PV cDNA sequences for all Antarctic fish. These sequences have been deposited in GenBank (See Table 2.1 for accession numbers). An undergraduate student researcher Danielle SandozOsmus assisted in the protocol for obtaining PV cDNA sequences from N. coriiceps, C. wilsoni, C. rastrospinosus, L. nudifrons, T. hansoni, C. gunnari, and P. georgianus.

Table 2.1. GenBank accession numbers for teleost parvalbumin cDNA sequences. Animal Name Accession Number Parachaenicthys charcoti FJ696942 Dissostichus eleginoides FJ696943 Gobionotothen gibberifrons FJ696944 Dissostichus mawsoni FJ696945 Patagonotothen ramsayi FJ696946 Notothenia rossii FJ696947 Notothenia coriiceps FJ696948 Pseudochaenicthys georgianus FJ696949 Champsocephalus gunnari FJ696950 Trematomus hansoni FJ696951 Lepidonotothen nudifrons FJ696952 Chionodraco rastrospinosus FJ696953 Chaenodraco wilsoni FJ696954 Chaenocephalus aceratus FJ696955 Boreogadus saida FJ696956 Micropterus salmoides FJ696957 Fundulus heteroclitus FJ696958 Fundulus similis FJ696959 Fundulus grandis FJ696960 Anguilla_japonica AB375263 Danio_rerio BC164913 Cyprinus_carpio AJ292212 Aristichthys_nobilis FJ013047 Ictalurus_punctatus AF227795 Sardinops_melanostictus AB375262 Sardinops_sagax FM177701 Salmo_salar X97824 Salvelinus_alpinus AF538283 Theragra_chalcogramma AY035587 Gadus_morhua AM497927 Orysias_latipes AU176885 Kryptolebias_marmoratus AY682950 Sebastes_inermis DQ374441 Paralicthys_olivaceus AB375266 Hippoglossus_hippoglossus EU412911 Tetraodon_nigroviridis CAAE01014738 Takifugu_rubripes CK829797 Sparus_aurata AY550962 Evynnis_japonicus AB375264 Lates_calcarifer AY626067 Lutjanus_argentimaculatus EF591789 Trachurus_japonicus AB211364 Scomber_japonicus AB091470 Oreochromis_mossambicus DQ124253 Gasterosteus aculeatus BT026859 Katsuwonus pelamis AB375265

Phylogenetic Analysis and Sequence Reconstruction DNA and protein sequence alignments were done using Clustal W2 (Chenna et al., 2003) on the EBI webserver. An empirical Bayesian approach was used for ASR (Huelsenbeck and Bollback, 2001). A phylogenetic guide tree of teleost fish corresponding to those species with available PV sequences (Table 2.1) was constructed with the topology conforming to known and wellsupported evolutionary relationships (Briolay et al., 1998; Miya et al., 2003; Teletchea et al., 2006; Nelson, 2006; Li et al., 2007; Near and Cheng, 2008; Hertwig, 2008). Branch lengths were allowed to vary and were estimated from the data during the analysis. The ModelTest (Posada and Crandall, 1998; Posada, 2006) and ProtTest (Abascal et al., 2005) servers were used to identify the appropriate models of nucleotide and amino acid sequence evolution, respectively. MrBayes v 3.2 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003) was used for ASR of PV nucleotide sequences. Reconstruction of cDNA sequences was performed using the general time reversible model with a gamma distribution of rates with a shape parameter, α, of 0.52311, and a proportion of invariable sites estimated by the program. Four chains were run for 1,000,000 generations with a sample frequency of 1000. The first 25% were discarded as the sampling during this phase of the analysis can be biased by the starting point of the chain. This portion of the sample is known as “burnin” and after this period the Markov Chain Monte Carlo (MCMC) sampling is more thorough and robust. The ancestral sequences were then annotated from the stat files. Ambiguity in the reconstructions due to model choice and algorithm was investigated by using the ASR program FASTML v. 2.02 (Pupko et al., 2000; 2002). Nucleotide sequences were reconstructed using the JukesCantor model, the Yang codon model, Goldman Yang codon model and an empirical codon model. ProtTest identified the Jones, Taylor, Thornton amino acid substitution model based on the PV alignment and this was also used in FASTML.

Homology Modeling All PV homology models were constructed using the SwissModel webserver (Arnold et al., 2006). The crystal structure of C. carpio PV, pdb 4CPV, was used as the template for all models. Model validity was evaluated using the built in methods of the SwissModel server. This includes the atomic empirical mean force potential ANOLEA which assesses packing quality of the model, and the GROMOS force field to assess local quality of residues. Models were

visualized with PyMol (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC.) and VMD (Humphrey et al., 1996). Hydrogen atoms were added to the models in PyMol. Protein energies, hydrogen bonds, and distances were calculated in DeepView.

Results

Sequencing In order to effectively estimate the PV sequence of the NAPV, the sequence database of PV from extant notothenioids needed to be enriched. Initial primers for PV sequencing were designed based on the cDNA sequence obtained by Hendrickson (2005). These primers were used to amplify RACEready cDNA from total RNA obtained from the white muscle of G. gibberifrons. Once the 5’ and 3’ untranslated regions of the PV cDNA had been sequenced, primers located in these regions were developed. Due to the high homology of the notothenioid PV genes these primers were successfully used to amplify PV cDNAs for all other notothenioid fish. These sequences have been deposited in GenBank (Accession numbers found in Table 2.1). Notothenioid PV cDNA sequences can be found in Appendix 1 while deduced amino acid sequences are found in Figure 2.1.

Ancestral Sequence Reconstruction Figure 2.2 shows the guide tree and general steps involved in ASR. A guide tree for ASR was constructed using the PV sequences of 51 teleost species (accession numbers can be found in Table 2.1). This phylogeny shown in Figure 2.2 is a composite tree and represents well established teleost relationships (Briolay et al., 1998; Miya et al., 2003; Teletchea et al., 2006; Nelson, 2006; Li et al., 2007; Near and Cheng, 2008; Hertwig, 2008). It should be noted that this tree only reflects topology or order of speciation. Evolutionary distances or branch lengths were estimated from the PV sequence data during the analysis. For visualization purposes the branch lengths in Figure 2.2 were set to the arbitrary value of 1. Ancestral sequences were estimated in MrBayes using a general time reversible (GTR) model of sequence evolution as identified by the program ModelTest. Sequences were estimated

10 20 30 40 50 ....|....| ....|....| ....|....| ....|....| ....|....| Dissostichus mawsoni MALAGNLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFLV Dissostichus eleginoides MALAGNLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFLV Patagonotothen ramsayi MALAGTLKEA DITAALAACT AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Trematomus hansoni MAPAGTLKEA DITAALAACT AAESFKHKEF FAKVGLSAKS ADDIKKAFLV Lepidonotothen nudifrons MALAGTLKEA DITAALAACT AAESFKHKEF FAKVGLSAKS ADDIKKAFLV Gobionotothen gibberifrons MALAGTLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Notothenia rossii MALAGTLKEA DITAALAACT AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Notothenia coriiceps MALAGTLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Parachaenichthys charcoti MALAGTLKEA DITAALAACS AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Pseudochaenichthys georgianus MALAGTLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Chionodraco rastrospinosus MALAGTLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Chaenodraco wilsoni MALAGTLKEA DITAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Chaenocephalus aceratus MALAGTLKEA DITAALAACT AAESFKHKEF FAKVGLSAKS ADDIKKAFGV Champsocephalus gunnari MALAGTLKAA DIAAALAACK AAESFKHKEF FAKVGLSAKS ADDIKKAFGV

60 70 80 90 100 ....|....| ....|....| ....|....| ....|....| ....|....| Dissostichus mawsoni IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI Dissostichus eleginoides IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI Patagonotothen ramsayi IDQDQSGFIE EEELKLFLQN FSAGARALTV GETKAFLKAG DIDGDGMIGI Trematomus hansoni IDQDQSGFIE EEELKLFLQN FAAGARALTV AETKAFLKAG DIDGDGMIGI Lepidonotothen nudifrons IDQDQSGFIE EEELKLFLQN FAAGARALTD AETKAFLKAG DIDGDGMIGI Gobionotothen gibberifrons IDQDQSGFIE EEELKLFLQN FSAGARALTD GETKAFLKAG DIDGDGMIGI Notothenia rossii IDQDQSGFIE EEELKLFLQN FSAGARALTV GETKAFLKAG DIDGDGMIGI Notothenia coriiceps IDQDQSGFIE EEELKLFLQN FSAGARALTD GETKAFLKAG DIDGDGMIGI Parachaenichthys charcoti IDQDQSGFIE EDELKLFLQN FSASARALTV AETKAFLKAG DIDGDGMIGI Pseudochaenichthys georgianus IDQDQSGFIE EEELKLFLQN FSASARALTD AETKAFLKAG DIDGDGMIGI Chionodraco rastrospinosus IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI Chaenodraco wilsoni IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI Chaenocephalus aceratus IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI Champsocephalus gunnari IDQDQSGFIE EEELKLFLQN FSAGARALTD AETKAFLKAG DIDGDGMIGI

....|.... Dissostichus mawsoni DEFATMVKA Dissostichus eleginoides DEFATMVKA Patagonotothen ramsayi DEFATMVKA Trematomus hansoni DEFATMVKA Lepidonotothen nudifrons DEFATMVKA Gobionotothen gibberifrons DEFATMVKA Notothenia rossii DEFATMVKA Notothenia coriiceps DEFATMVKA Parachaenichthys charcoti DEFATMVKA Pseudochaenichthys georgianus DEFASMVKA Chionodraco rastrospinosus DEFASMVKA Chaenodraco wilsoni DEFASMVKA Chaenocephalus aceratus DEFASMVKA Champsocephalus gunnari DEFASMVKA

Figure 2.1 Antarctic fish parvalbumin protein sequence alignment. Identical residues are shown in black, similar residues in red and nonconservative differences in blue. All accession numbers can be found in Table 2.1.

Figure 2.2. Ancestral sequence reconstruction guide tree and method outline. 51 teleost PV sequences were mapped to this composite tree for reconstruction. Of main interest is the branch leading from the Perciformes ancestor to the notothenioid ancestor. This represents the invasion of the Southern Ocean and subsequent cooling experienced by notothenioid ancestral fish.

for five ancestors of modern, extant notothenioids (Fig. 2.3). This particular reconstruction scenario places the eel Anguilla japonica as the outgroup with Acanthomorpha containing the rest of the teleost orders represented in the tree. In order to discern the robustness of ASR, ambiguity in the reconstructions (defined as different possible reconstructions) was investigated by estimating sequences using different models of sequence evolution and a different program, FastML, which uses a different algorithm than MrBayes. There was a small amount of sequence ambiguity found among the different models and programs as compared to the GTR model used in MrBayes (Fig. 2.4). Comparing the ancestral PV sequences reveals a unique set of substitutions between PAPV and NAPV. This set of 16 substitutions correlates with paleontological and geological events highlighted in Figure 2.2 including the cooling of the Southern Ocean to the current sub zero temperatures. Of these 16 substitutions (Fig. 2.5) eight are conservative and eight are non conservative. The extant GBPV provides the template for sitedirected mutagenesis studies.

Homology Modeling In order to determine the subset of residues likely responsible for the observed thermal sensitivity profiles of extant notothenioid PVs and the potential structural and functional differences between the two ancestral PVs, homology models of both ancestral PVs were made using the major PV isoform from C. carpio, CCPV, (PDB: 4CPV; Kumar et al., 1990) as the template structure. This allows for the substitutions to be viewed in a threedimensional context. The sequence identities between the template and target structures are at least 78%. Models built with this level of sequence identity have been shown to be accurate and have resolutions comparable to that of crystal structures (Nayeem et al., 2006). Root mean square deviation (rmsd) of all the models as compared to the template structured, 4CPV, was within 0.05Å which is well within the resolution of the template structure, 1.5Å. Additionally, evaluation by ANOLEA and GROMOS force fields within the SwissModel server showed a high level of quality of all the models.

Figure 2.3. Reconstructed ancestral sequences shows unique substitution pattern between Perciformes and notothenioid ancestral PV sequences. A. Deduced amino acid sequences from cDNA reconstruction in MrBayes. Identities are shown as dots. B. Example phylogram showing evolutionary relationships of ancestral fish whose PV sequences were estimated by ASR.

MrBayes GTR DNA to AA: Acanthomorpha MAFAGILNDADITAALEACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAALVKA Acanthopterygii MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAAMVKA Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAAMVKA Notothenioid MALAGTLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGMIGIDEFATMVKA

FastML JC DNA to AA Acanthomorpha MAFAGILNDADITAALEACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Acanthopterygii MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSAKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Notothenioid MALAGNLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGMIGIDEFATMVKA

FastML Yang Codon DNA to AA Acanthomorpha MAFAGILNDADITAALEACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Acanthopterygii MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSAKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Notothenioid MALAGNLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGMIGIDEFATMVKA

FastML Goldman-Yang Codon DNA to AA Acanthomorpha MAFAGILNDADITAALEACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Acanthopterygii MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSAKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Notothenioid MALAGNLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGMIGIDEFATMVKA

FastML Empirical Codon DNA to AA Acanthomorpha MAFAGILNDADITAALEACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Acanthopterygii MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSAKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAALVKA Notothenioid MALAGNLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGTIGIDEFATMVKA

FastML JTT AA Acanthomorpha MAFAGVLNDADITAALEACKAADSFNHKAFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAALVKA Acanthopterygii MAFAGVLNDADITAALAACKAADSFNHKAFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Percomorpha MAFAGVLNDADITAALAACKAADSFNHKAFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Perciformes MAFAGVLNDADITAALAACKAADSFNHKAFFAKVGLSGKSADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDSDGDGKIGVDEFAAMVKA Notothenioid MALAGTLKEADITAALAACKAAESFKHKEFFAKVGLSAKSADDIKKAFLVIDQDQSGFIEEEELKLFLQNFSAGARALTDAETKAFLKAGDIDGDGMIGIDEFATMVKA

Figure 2.4. Ambiguity in sequence reconstruction. Comparing the sequences reconstructed using different models of sequence evolution (General time reversible (GTR), JukesCantor (JC), three different codon models, and JonesTaylorThornton model) and different reconstruction programs (MrBayes and FastML). “DNA to AA” denotes reconstructed cDNA sequences translated to amino acid sequences while AA denotes reconstructed amino acid sequences. Nonconservative substitutions are shown in yellow and conservative in blue. Ambiguous positions are in red. Note that there is no ambiguity at positions 8 and 26.

AB † † † † Perciformes MAFAGILNDADITAALAACKAADSFNHKDFFAKVGLSGKS Notothenioid --L--T-KE------E--K--E------A-- G_gibberifrons --L--T-KE------E--K--E------A-- 1 * * * *

CD EF † † † † Perciformes ADDIKKAFAIIDQDKSGFIEEDELKLFLQNFSAGARALTDAETKAFLKAGDTDGDGKIGVDEFAAMVKA Notothenioid ------LV----Q------E------I----M--I----T---- G_gibberifrons ------GV----Q------E------I----M--I----T---- 41 * * * * * *

Figure 2.5. Sequence alignment of PAPV and NAPV with the extant GBPV. Calcium binding loops are shown in grey. Identical residues are indicated by . Nonconservative substitutions indicated by †.

In this study we focused on the eight nonconservative substitutions as these are most likely to affect function. Figure 2.6 shows models of PAPV and NAPV and the position of the eight nonconservative substitutions identified by ASR. Viewed in the tertiary structure, most of the substitutions reside outside of the active sites. One substitution at position 97 affects a Ca2+ coordinating residue. This residue, however, coordinates with its backbone oxygen and can accept a higher degree of variability than the other coordinating residues which are almost completely conserved (Falke et al., 1994). Virtual sitedirected mutagenesis of GBPV, an Antarctic fish PV which has been characterized functionally (Erickson et al., 2005), was used to determine the effect that each of the PAPV to NAPV substitutions had on extant PV structure and potentially function. In this process each substitution between PAPV and NAPV was substituted in reverse into the GBPV primary structure and then modeled to confirm that the virtual mutations would recapitulate the interactions present in PAPV (e.g. Fig. 2.7). Each of these models was evaluated as above. Energies of unfolding of the homology models were calculated in DeepView using the GROMOS force field as a representative measure of relative conformational flexibility/thermal stability (Table 2.2). Our results showed a correlation between this measure of relative stability and thermal habitat. The NAPV and GBPV had similar energies which were both higher than the warm/temperate adapted Perciformes ancestral PV and the PV from the temperate representatives CCPV and PV from Micropterus salmoides (MSPV). Most of the substitutions did not substantially affect free energy of folding The substitutions at position 8 and 26 individually both reduced the energy of GBPV to approximately half way towards the temperate/warm PVs. The double mutant, K8N K26N, had an energy value similar to the Perciformes PV structure indicating that the PAPV structure, and potentially function, is recapitulated.

Figure 2.6. Location of nonconservative substitutions identified by ASR. The yellow to blue gradient (A.) corresponds to the branch leading from the Perciformes ancestor to the notothenioid ancestor in Figure 2.2. Panels B. and C. show the Perciformes ancestral PV and the notothenioid ancestral PV, respectively, with calcium ions (red) oriented to the rear to show the position of the substitutions found in the AB domain. Panels D. and E. show the calcium binding loops and the substitutions found near the active sites.

Table 2.2. Predicted free energy of folding values for homology models of extant and reconstructed PV sequences and predicted effects of virtual mutatgenesis.

Exant and Reconstructed Predicted Free Energy of Folding Proteins (kJ/mol) (Gromos Force Field) Thermal Environment Perciformes Ancestral PV -2605 Temperate/Warm Notothenioid Ancestral PV -2214 Polar Wild-type G.gibberifrons PV -2147 Polar C. carpio PV -2705 Temperate M. salmoides PV -2589 Temperate

Predicted Free Energy of Folding Mutant Proteins (kJ/mol) (Gromos Force Field) Predicted Effect of Substitution L3F -2192 Network of Aromatic Interactions T6I -2039 Alter solvent accessibility K8N -2343 Introduce H-bond K26N -2341 Introduce H-bond Q55K -2100 Both polar aliphatic residues I92T -2180 Alter solvent accessibility M97K -2149 Introduce cation-pi bond T105A -2131 Alter solvent accessibility/H-bond K8N K26N -2537 Introduce two H-bonds

Figure 2.7. Homology models showing reintroduction of hydrogen bonds by virtual mutagenesis. GBPV WT (A.) showing the residues that are completely conserved in Antarctic notothenioids and found in the notothenioid ancestral PV. The double mutant (B.) showing the reintroduction of the hydrogen bonds found in the Perciformes ancestral PV.

To determine if these findings support the idea that functiontuning substitutions are located away from the active site, the distance from the active site was measured in PyMol (Fig. 2.8). Distances were measured from the alphacarbon of D11 and E29. These two residues are the hydrogen bonding partners of the putative functionally adaptive substitution sites, 8 and 26. Further support for the hypothesis that K8N and K26N will convert GBPV to a temperate, ancestral state is given by the complete conservation among Antarctic notothenioids (Fig. 2.1) of K8 and its hydrogen bonding partner D11 as well as K26 and its partner E29. The importance of the evolutionary loss of the two hydrogen bonds as compensatory substitutions is supported by their complete conservation.

Discussion

After invasion of the Southern Ocean by the notothenioid temperate ancestor, the group faced the extreme cooling that has led to the current subzero temperatures found around Antarctica (Eastman, 1993). For Antarctic fish PVs to function optimally we would expect to find evolutionary substitutions of amino acid residues compensating for the decreased ambient temperature. Such changes would have occurred sometime after the Antarctic notothenioids diverged from their temperate ancestor. Due to the exothermic nature of PV Ca2+ binding, decreased temperature would cause a decrease in 2+ 2+ Ca Kd. In white muscle this increase in PV Ca affinity would disrupt the correct timing of muscle contraction by diverting Ca2+ from binding to troponin C (TnC). The resulting decreased force production presumably would put a fish at a disadvantage when trying to feed on swimming prey or evading predators. Most comparative studies of enzyme systems seeking to determine the molecular basis of thermal adaptation have focused on proteins from congeneric or confamilial species (Hochachka and Somero, 2002). This allows the identification of the few residues responsible for functional tuning as they are not swamped by the neutral substitutions that accrue in more disparately related groups. This type of study system, however, is not always attainable. In the present effort we have sought to discover the molecular basis for

Figure 2.8. Functional tuning by substitutions away from the active site in GBPV WT. As seen in Figure 2.7 Asp11 and Glu29 have the potential to hydrogen bond with residues at position 8 and 26, respectively. Distances from the alphacarbon of position 11 and 29 (shown in CPK coloring) to the active site calcium ions (yellow) are given in angstroms. The rest of the protein is shown in light blue ribbons. Of note is that even though the potential hydrogen bonds identified by ASR and homology modeling are far removed from the active sites, they may cause important changes in binding ability.

29 coldadaptation found in PV from Antarctic notothenioids as compared to a temperate counterpart, CCPV (Erickson et al., 2005). Notothenioids and C. carpio last shared a common ancestor at least 200 million years ago. Over this long evolutionary time 23 amino acid differences have built up between CCPV and GBPV. The majority of these are likely to be functionally neutral. In Figures 2.2 and 2.6 the yellow to blue gradient represents the several events that would have occurred during the evolution of Antarctic notothenioids leading to the current thermal sensitivity profile of GBPV and the other notothenioid PVs. Evidence has accumulated that supports the corresponding states theory (Somero, 1978; 1995) where thermal adaptation of proteins is accomplished through adjustments in conformational flexibility. As protein structure is only marginally stable, we would expect only small, subtle changes in the network of noncovalent interactions that dominate protein folding and structure to elicit measurable changes in protein function. In thermodynamic terms, adjustments in noncovalent interactions alter the net enthalpy change of ligand binding. These changes provide compensation for the effect of changing temperature on protein function. In the case of notothenioid PV evolution, as the temperature of the Southern Ocean decreased we would expect to find evolutionary adjustments in protein sequence that compensate for the subzero temperatures found in the waters surrounding Antarctica. We have identified through ASR and homology modeling two substitutions in Antarctic notothenioid PV, at positions 8 and 26, which could be sufficient to explain the observed thermal sensitivity pattern of notothenioid PV (Erickson et al., 2005). By comparing the reconstructed PV sequences for the coldadapted notothenioid ancestor and several of its teleost ancestors (Fig. 2.2 and Fig. 2.3) it was possible to track the potential changes that occurred during notothenioid evolution. By viewing the PAPV and NAPV (Fig. 2.5) the set of potentially function altering substitutions was narrowed from 23 to 16, of these eight were nonconservative. In order to view the structural effects that the nonconservative substitutions between PAPV and NAPV would have on an extant notothenioid PV, the PAPV residues were reintroduced into GBPV. Table 2.2 gives the predicted effect of each of these substitutions. The substitutions at positions 8 and 26 stand out as being important as they introduce hydrogen bonds into the GBPV model. As

30 discussed in the Introduction, hydrogen bonds have a negative enthalpy of formation and are stabilized by decreased temperature. The evolutionary removal of two hydrogen bonds through mutation at positions 8 and 26 would serve to decrease the net enthalpy of Ca2+ binding in Antarctic fish PV. These compensatory substitutions would presumably lead to higher conformational flexibility in Antarctic fish PV as compared to temperate counterparts which has been shown to be a key feature of coldadaptation (Fields and Somero, 1998; Zavodsky et al., 1998). Analysis of relative stability through estimates of the free energy of folding of our homology models showed that estimated stability was correlated with environmental temperature. Based on previous studies that show a correlation of PV Ca2+ binding function and environmental temperature, the folding energy estimates suggest that PV function will correlate with free energy of folding values. This gives confidence to the virtual mutagenesis results which leads to our hypothesis that reintroduction of the hydrogen bonds individually will provide an intermediate, but not necessarily additive decrease in conformational flexibility, and intermediate thermal sensitivity pattern. Moreover, we propose that the double mutant will show a full conversion from the Antarctic wildtype thermal sensitivity pattern and that of the temperate counterparts. These results guided empirical sitedirected mutagenesis studies with GBPV serving as a representative template to test our hypothesis concerning coldadaptation of Antarctic fish PV as described in the next chapter.

CHAPTER THREE Expression, Purification and Functional Characterization of Recombinant Parvalbumins

Introduction

Most of the knowledge concerning the adaptation of proteins to varied thermal habitats comes from research on enzymes (Hochachka and Somero, 2002). According to the corresponding states theory, differences in thermal sensitivity patterns of substrate binding constants, Km, of orthologous enzymes from poikilotherms adapted to different thermal regimes are driven by subtle changes in protein primary structure leading to changes in conformational flexibility. These changes in conformational flexibility offset the function altering effects of changing temperatures, thus conserving optimal function at physiological temperature. In thermodynamic terms, at low temperatures there is less heat available to drive physiochemical reactions. Coldadapted orthologs respond with a smaller net enthalpy change and a larger change in entropy than warm/temperate adapted orthologs. This leads to a similar change in free energy and conserved binding ability at physiological temperatures. In prior work, this knowledge gained from enzymes was extended to the non catalytic calcium binding protein parvalbumin (PV) (Erickson et al., 2005; Erickson and Moerland, 2006). PV purified from white muscle of coldadapted Antarctic fish of the Perciformes suborder Notothenioidei and PV purified from temperate counterparts displayed the characteristic pattern of thermal sensitivity of species adapted to different temperatures. Specifically, at a common measurement temperature Antarctic fish PV

showed a weaker binding affinity, as evidenced by a higher Kd, than temperate counterparts, but at physiological temperatures function was highly conserved. In this study we sought to determine the specific changes in PV primary structure that occurred during notothenioid evolution that led to thermal compensation and conservation of function. We outlined our selection of mutant PVs (K8N, K26N and the double mutant,

DM) in Chapter 2. This chapter discusses studies wherein mutations were introduced into the wildtype Gobionotothen gibberifrons PV (GBPV WT) through sitedirected mutagenesis. Lysine to asparagine mutations were engineered at positions 8 (GBPV K8N) and 26 (GBPV K26N). A double mutant construct was also produced (GBPV DM). We were also successful with expression and purification of the mutant versions of GBPV. Here we use a competitive binding assay based on the fluorescent indicator fluo3 originally developed by Eberhard and Erne (1993) to estimate the Ca2+ dissociation constant for recombinant and mutants. This technique has been used previously in the lab to characterize PV purified from white muscle of polar and temperate fish including Gobionotothen gibberfrons (GBPV) (Erickson et al., 2005; Erickson and Moerland, 2006), the species providing the template PV (wildtype, WT) for this study, and from a species of stingray known to inhabit habitats with varying salinity (Heffron and 2+ Moerland, 2008). Additionally, Ca dissociation rates, koff, were measured using stoppedflow spectrometry with terbium fluorescence as a reporter ligand (Hou et al., 1992). The amino acid substitutions identified by ancestral sequence reconstruction (ASR) and homology modeling (Chapter 2) and introduced into GBPV WT via site directed mutagenesis produced significant increases in Ca2+ binding affinity confirming our hypothesis that the single mutants, K8N and K26N, would have an intermediate thermal sensitivity pattern and the double mutant, DM, would have a fully rightshifted thermal sensitivity pattern characteristic of PV from a temperate fish recapitulating the evolutionary trajectory of PV coldadaptation in Antarctic fish. These results confirm that these substitutions are sufficient to explain the characteristically polar pattern of thermal sensitivity found for Antarctic notothenioid PVs (Erickson et al., 2005).

Interestingly, the koff estimates did not correspond fully to the steadystate measurements. While there was a right shift in thermal sensitivity of koff for the mutant PVs as compared to the WT isoform, the three mutant constructs had very similar off rates. The combined steadystate and kinetics data suggest a structural mechanism for the coldadaptation of Antarctic fish PVs. As the Southern Ocean began to cool, the Perciformes ancestral PV lost one of two hydrogen bonds destabilizing the apostate

33 leading to slower Ca2+ onrates, but relatively similar offrates. With continued cooling the loss of the second hydrogen bond in the notothenioid ancestral PV led to destabilization of the bound state and faster offrates and the current thermal sensitivity pattern displayed by GBPV WT and other Antarctic fish PV isoforms (Erickson et al., 2005). Based on the functional data presented here, we propose a thermodynamic/structural mechanism in the context of known structural data of PV in the bound and apostates. We discuss this model in the context of the corresponding states theory outlined in Chapter 1.

Materials and Methods

Preparation of GBPV WT cDNA Isolation of total RNA was performed as described in Chapter 2, according to the Trizol reagent protocol (Invitrogen). First strand cDNA synthesis was accomplished using the Superscript III Reverse TranscriptionPCR protocol (Invitrogen). Expression primers (For: 5’ATGGCACTTGCAGGAAC3’; Rev: 5’TTAAGCCTTGACCATGGA 3’) were designed based on the cDNA sequence of Gobionotothen gibberifrons (GenBank accession number: FJ696944) (Sequencing described in Chapter 2. Sequence found in Appendix 1). PV cDNA was amplified with the following PCR protocol: 94°C for 2:00 min. 35 cycles of 94°C for 45 sec, 56°C for 1:00 min, 72°C for 45 sec. 72°C for 5:00 min. PV cDNA was gel purified from a 1% agarose gel using the QIAquick Gel Extraction Kit Protocol (Qiagen).

Cloning in Novablue Cells GBPV WT cDNA was ligated into the pETBlue1 plasmid vector (Novagen) using the Clonables Ligation Premix Kit (Novagen) protocol which contains ligase and appropriate buffers and salts and then transformed into NovaBlue competent cells (Novagen). Positive clones were identified by blue/white screening. Vector was isolated by QIAPrep Spin Miniprep Kit (Qiagen) and subjected to restriction enzyme digest to confirm correct insertion orientation of the gene. Plasmids were then sent for sequencing

34 at the Florida State University Department of Biological Science DNA Sequencing Facility.

Expression in DE3(BL21) Cells Once the cDNA insert sequence and orientation was confirmed, the plasmid was transformed into DE3 expression cells (Novagen). The cells were plated and grown over night on LuriaBertani (LB) broth agar plates. To determine appropriate expression conditions, an analytical scale expression was done. 1 mL of LB broth with antibiotics and glucose was inoculated with a colony of DE3 cells and grown overnight. Then 5 mL of LB broth of inoculated with 50 L of overnight culture. Each tube was grown at 37°C until it reached an optical density of 0.6 1.0 measured at 600 nm. Expression was induced with various concentrations of isopropyl βD1thiogalactopyranoside (IPTG) ranging from 0.1 to 2 mM. 1 mL samples were taken and processed at 2, 4 and 6 h postinduction. Once optimum expression conditions were determined, a large scale expression was done. 50 mL of LB broth containing 50 g/mL ampicillin and 1% glucose was inoculated with GBPV WT and grown overnight at 37°C. 1 L of LB broth was inoculated with the overnight culture. After 2 h of growth the culture was induced with 1 mM IPTG. After 4 h of growth the cells were harvested by centrifugation.

Site-directed Mutagenesis The QuikChange mutagenesis protocol (Stratagene) was used to introduce substitutions at positions 8 and 26 of GBPV WT. Primers were designed using the PrimerX program available online (http://www.bioinformatics.org/primerx/) (K8N: For: 5’GCAGGAACCCTGAATGAGGCTGACATCAC3’; Rev: 5’ GTGATGTCAGCCTCATTCAGGGTTCCTGC3’; K26N: For: 5’ CTGCTGAGTCTTTCAATCACAAGGAATTCTTCG3’; Rev: 5’ CGAAGAATTCCTTGTGATTGAAAGACTCAGCAG3’). Mutagenesis steps were performed by the Molecular Cloning Facility housed in the Department of Biological Science at Florida State University. Plasmids were sequenced by the Sequencing Facility

35 to confirm that sitedirected mutagenesis was successful. Once mutagenesis was complete all recombinant proteins were expressed as described above.

Protein Purification All recombinant proteins were purified according to a protocol adapted from Cates et al (1999) and Erickson et al (2005). Cells from 1 L of culture were pelleted using a Beckman JA14 centrifuge rotor set at 10,500 rpm for 30 min. Cells were resuspended in ice cold 25 MES buffer at pH 5.7. Cells were lysed by sonication with 60 sec burst with 90 sec rest periods for 6 cycles. Cellular debris was removed from the sonicate by centrifugation in a Beckman JA20 rotor set at 18,250 rpm for 30 min. The supernatant was collected and concentrated to 50 mL or less using Amicon Centrifugal Filter Devices or a Millipore Stirred Cell Device, depending on supernatant volume. The supernatant was then dialyzed against 2 x 2 L of 25 mM MES (pH 5.7) in preparation for ionexchange chromatography. Chromatography steps were performed on an AKTApurifier chromatography system (GE Healthcare). The supernatant was applied to a HiPrep 16/10 DEAE column equilibrated with dialysis buffer. Proteins were eluted with a 0120 mM NaCl gradient applied over 10 column volumes. The PV containing fractions were identified by SDSPAGE analysis. These fractions were then pooled and concentrated to 5 mL or less as above. The resultant PV sample was dialyzed against 2 x

2 L of S200 column buffer containing 50 mM Na2HPO4 and 150 mM KCl (pH 7.0). The sample was applied to the column and then eluted without collecting fractions for 0.43 column volumes. Fractions were collected for 0.27 column volumes. PV containing fractions were collected and pooled. PV was judged to be 95% pure based on SDSPAGE and Western Blot analysis.

Decalcification of Buffers and Proteins Assay buffer and proteins were stripped of divalent cations using procedures modified from previously published protocols (Erickson et al., 2005, 2006; Heffron and Moerland 2008) To decalcify assay buffer (20 mM HEPES, 150 mM KCl) a one step procedure was used. 50 mL of assay buffer was mixed with 5% chelex100 resin (Sigma) in a 50 mL plastic conical tube and covered with parafilm. The tubes were swirled on a

36 rotary shaker overnight to allow removal of divalent cations. Next, the tubes were centrifuged in a table top hangingbucket centrifuge for 20 min at 3000 rpm to concentrate the chelex at the bottom of the tube. The buffer was then decanted into acid washed 50 ml conical tubes and covered with parafilm. To adjust the pH of the assay buffer while avoiding contamination, small aliquots were removed and the pH measured. Then small aliquots of molecular biology grade, concentrated HCl (Sigma) were added to the stock buffer container to adjust the pH to 7.2. Purified proteins were stripped of divalent cations using a four step dialysis procedure. First, proteins were dialyzed against 1 L of assay buffer containing 5% chelex, 40 M EDTA and 40 M EGTA to scavenge cations, and 4 M urea to provide a denaturing environment. Second, the proteins were refolded by dilution in assay buffer with no urea containing 5% chelex, 40 M EDTA and 40 M EGTA. The third and fourth steps were identical containing assay buffer and 5% chelex. After dialysis, proteins were transferred to acid washed plastic 15 ml conical tubes. Protein concentrations were then determined by the BCA assay (Pierce).

Determination of Calcium Dissociation Constants, Kd A competition assay using the fluorescent Ca2+ indicator fluo3 (Molecular Probes) based on previously published methods (Eberhard and Erne, 1991; Erickson et 2+ al., 2005; Heffron and Moerland, 2008) was used to determine the Ca Kd for all proteins. Two mL of assay buffer containing 1.25 M fluo3 was titrated with 10 aliquots

of 100 M CaCl2 (Orion) at 5 L/aliquot. One 5 L aliquot of 1 mM CaCl2 was then

added followed by one 5 L aliquot of 100 mM CaCl2 to ensure saturation of fluo3 yielding the maximum fluorescence. Figure 3.1 displays the change in fluorescence intensity with increasing [Ca2+]. Fluorescence intensity measurements were converted to the concentration of fluo3 bound with Ca2+ using the equation:

2+ F [fluo− 3* Ca ] = ⋅ [ fluo − 3]total (1) Fmax

Figure 3.1. Representative example of fluorescence data collected during a titration of 1.25 M fluo3 and 1.25 M PV. Fluorescence has been converted to relative fluorescence with maximum fluorescence set to 1.

where F is the normalized fluorescence intensity and Fmax is the maximum fluorescence. Free calcium is determined by:

2+ 2 + 2 + [Ca ]free= [ Ca ] total − [ fluo − 3* Ca ] (2)

Plotting fluo3 bound with Ca2+ vs. free Ca2+ provides binding curves for fluo3. Hyperbolic nonlinear leastsquares fits of binding curves provide estimation of fluo3

Kd. Figure 3.2 shows a representative fluo3 titration curve fit with a hyperbolic function.

Titration of 1.25 M fluo3 in the presence of PV allows estimation of PV Kd. The concentration of fluo3 bound with Ca2+ is determined as in equation 1. Then using the fluo3 Kd allows estimation of free calcium:

K⋅[ fluo − 3* Ca2+ ] []Ca2+ = d (3) free [fluo− 3] − [ fluo − 3* Ca2+ ]

The concentration of Ca2+ bound to PV is found by:

2+ 2 + 2 + 2 + [*PV Ca ][= Ca ]total − [ fluo − 3* Ca ][ − Ca ] free (4)

2+ 2+ PV binding curves ([PV*Ca ] vs. [Ca ]free) were fit as described above to provide estimates of PV Kd. Figure 3.3 shows a representative PV titration fit with a hyperbolic function. All titrations were performed using a Varian Cary Eclipse fluorescence spectrometer with internal temperature control housed in the Physical Biochemistry Facility at Florida State University. Excitation wavelength was 505 nm and emission wavelength was 530 nm. Titrations were performed from 5 to 25°C at 5 degree increments.

2+ Figure 3.2. Representative fluo3 Ca Kd determination curve. 1.25 M fluo3 titrated with ten 5 L aliquots of 100 M CaCl2 (inset) and one 5 L aliquot of 1mM CaCl2 (far right point on top curve). Maximum fluorescence values were obtained by adding one aliquot of 100 mM CaCl2. Nonlinear regression analysis yielded Kd estimates.

Figure 3.3. Representative PV Ca2+ binding curve. 1.25 M fluo3 was titrated in the presence of 1.25 M PV. Nonlinear regression analysis gives estimates of Kd.

Preliminary fluo3 experiments showed inconsistent and at times a complete lack of Ca2+ binding by GBPV WT and mutant constructs. Initially, it was assumed that the proteins were not fully decalcified. One test of decalcification showed that GBPV WT had Ca2+ binding ability immediately after decalcification treatment, but then lost this capacity incrementally over 3 d until no binding ability was measured. Addition of β mercaptoethanol (BME) to 14 mM in protein stocks restored binding ability while addition of BME to 7 mM in the assay buffer immediately before titrations maintained a reducing environment throughout the assay.

Calcium unidirectional rate constants, koff and kon Calcium offrates were measured using terbium fluorescence as a reporter ligand based on established methods (Hou et al, 1991, 1992). Parvalbumin has a higher affinity for terbium and the offrate of Ca2+ can be measured as it is replaced by terbium and terbium fluorescence increases. Excitation of intrinsic phenylalanine residues in PV allows excitation of terbium through resonance energy transfer. Offrates were measured using an Applied Photophysics Ltd. model SX.18 MV stoppedflow instrument housed in the laboratory of Dr. Jonathan Davis at the Ohio State University Medical Center. The instrument has a mixing time of 1.4 ms at 15°C. PV samples were prepared as described above including adding 1 mM DTT to the protein stocks. For offrate measurements, 20

M CaCl2 was added to 10 M PV. This solution was rapidly mixed with 500 M TbCl2 in the instrument, excited at 250 nm and emission monitored through a 547 nm emission filter. Terbium fluorescence versus time curves were fitted with single and double exponential curves. A double exponential provided a better fit and these results are 2+ reported. Ca association rate, kon, was determined using the relationship kon = koff/Kd.

Results

Production of Recombinant Constructs Figure 3.4 shows the time course of expression of GBPV WT in pilot experiments. There was evidence of basal expression of soluble PV which was dramatically enhanced by addition of IPTG. After establishing conditions for expression

conditions, PV GBPV WT and the single and double mutant PV constructs were successfully expressed and purified. Typically, a 2 L culture yielded ~12 mL of pooled PV containing fractions at a concentration of 110 M as judged by BCA assay (Pierce) giving a total yield of ~15 mg of protein. Figures 3.5 through 3.8 show the purification stages for GBPV WT; similar chromatographic behavior was observed for the mutant constructs. This purification scheme represents an optimization of previous protocols (Cates et al., 1999; Erickson et al., 2005) as there is no ammonium sulfate cut which can decrease the yield of protein. Also, all chromatography steps were performed on an AKTApurifier system which decreases the amount of time needed for each purification step (e.g. AKTA S200 takes 7 h versus traditional low pressure gel filtration which takes ~18 h to elute PV.) and gives greater reproducibility as all flow rates and gradients are computer controlled.

Calcium Kd Measurements PV, which has two EFhand ion binding sites, has been previously shown to bind two Ca2+ molecules with no difference in affinity between sites (Pauls et al., 1993; Eberhard and Erne, 1994; Agah et al., 2003). This has also been shown to be the case for Antarctic fish PV (Erickson et al., 2005). In the present effort, all PV titration curves were fit with a hyperbolic function accounting for single site binding (for instance see Fig. 3.9). All four proteins showed the same general trend of decreased binding ability, as

evidenced by a larger Kd, with increasing temperature, indicative of an exothermic process (Fig. 3.9; Table 3.1). At each measurement temperature the WT protein gave the highest Kd. All three mutant PVs showed a right shift in thermal sensitivity (Fig. 3.10) 2+ indicating increased Ca binding ability. GBPV WT showed a sharper increase in Kd above 20 °C than the three mutants indicating less thermal sensitivity for the three mutant constructs. Twoway analysis of variance (ANOVA) was used to test for significance between all four proteins with Kd as the measurement variable and temperature and PV construct as the nominal variables. All values were found to be significantly different for the four proteins (p < 0.0001) except for the comparison of the K8N and K26N constructs for which no significant difference was found (p = 0.8647) indicating that K8N and K26N PVs show equivalent Ca2+ binding ability.

Calcium koff Measurements While the two EF hand ionbinding sites of PV were assumed to be equivalent in 2+ analysis of Kd data, Ca dissociation rates using terbium fluorescence as a reporter displayed better fit using a double exponential regression. Table 3.2 shows values for the fast rate (Rate 1) and slow rate (Rate 2). All three mutants showed a right shift in koff (Rate 1) vs temperature curves as compared to the WT construct (Fig. 3.11). The three mutants, however, showed a nearly identical shift in offrates with the two single mutants showing slightly slower rates than the double mutant, which does not correspond to the results of the steady state experiments, but does hint at a structural mechanism for shifts in thermal sensitivity in response to environmental temperature which will be explored below. A twoway ANOVA of Rate 1 values indicated that all recombinant constructs were significantly different from each other with respect to the fast rate (p < 0.0001). Ca2+ onrates which were calculated as described above are shown in Table 3.3. Because kon is derived from the ratio of two means of different size n values, there is no variance for the estimates of onrates.

Figure 3.4. Representative SDSPAGE gel showing analytical scale expression results of GBPV WT. Shown is the total protein and soluble fraction at 0, 2, 4, and 6 h post induction. Results show that while there is some basal expression after induction (identified by arrow), GBPV WT overexpresses and is found in the soluble fraction. L: Kaleidoscope protein ladder (BioRad).

Figure 3.5. Representative AKTA chromatograph of GBPV WT DEAE ionexchange chromatography. The large peak after sample insertion corresponds to the flowthrough of bacterial proteins. Blue numbers refer to time of peak. PV is found in fractions 2733 (indicated by arrows). See Materials and Methods for chromatography conditions.

Figure 3.6. Representative SDSPAGE gels from ionexchange chromatography. L: Precision Plus Protein ladder (BioRad); D: Dialyzate applied to column; Numbers correspond to fractions in Figure 3.2. Fractions 2731 were pooled and prepared for size exclusion chromatography.

Figure 3.7. Representative AKTA results for GBPV WT S200 sizeexclusion chromatography. PV is found in fraction 1113 (indicated by arrows).

Figure 3.8. Representative SDSPAGE gels from GBPV WT size exclusion chromatography. L: Precision Plus Protein ladder (BioRad); C: Concentrated PV containing fractions from ionexchange chromatography; Numbers correspond to fractions in Figure 3.4.

Figure 3.9. Temperature affects steepness of the PV titration curve. At lower temperatures PV binds Ca2+ with higher affinity leading to a left shift in the binding curve and a lower Kd estimate.

Table 3.1. Calcium dissociation constants, Kd, for GBPV WT plus mutants at a range of temperatures. All values are means ± 1 s.d. (n = 3). The units for all Kd values are nM.

Temperature (°C) WT K8N K26N DM

5 14.39 ± 1.15 8.45 ± 0.10 9.22 ± 0.96 6.46 ± 0.37 10 16.57 ± 1.09 11.67 ± 0.39 10.90 ± 0.69 5.39 ± 0.33 15 18.34 ± 0.54 16.08 ± 0.45 14.46 ± 1.30 7.71 ± 0.51 20 22.66 ± 1.24 19.43 ± 0.81 20.00 ± 0.37 11.40 ± 0.68 25 39.46 ± 1.17 27.06 ± 1.03 29.20 ± 0.90 18.28 ± 0.25

Figure 3.10. Calcium dissociation constants, Kd, for GBPV WT and mutants. All points represent the mean ± 1 s.d. (n = 3). Values for all PV constructs were found to be significantly different by twoway ANOVA (p < 0.0001) except for K8N and K26N which were found to not be significantly different (p = 0.8647).

Table 3.2. Calcium dissociation rates, koff, measured by stoppedflow spectrometry using terbium fluorescence as a reporter. Values 1 represent the mean ± 1 s. d. (n = 45). The units for all koff are s .

______WT______K8N______K26N______DM

Temperature (°C) Rate 1 Rate 2 Rate 1 Rate 2 Rate 1 Rate 2 Rate 1 Rate 2

5 1.22 ± 0.10 0.183 ± 0.049 0.69 ± 0.06 0.094 ± 0.022 0.65 ± 0.02 0.054 ± 0.004 0.76 ± 0.03 0.086 ± 0.014

15 1.99 ± 0.12 0.149 ± 0.048 1.15 ± 0.07 0.082 ± 0.011 0.99 ± 0.06 0.169 ± 0.234 1.32 ± 0.03 0.071 ± 0.016

25 2.45 ± 0.09 0.099 ± 0.038 1.77 ± 0.41 0.066 ± 0.009 1.80 ± 0.20 0.048 ± 0.006 1.92 ± 0.10 0.155 ± 0.253

Figure 3.11. Calcium dissociation rates, koff, measured as a function of temperature. Values represent the mean ± 1 s. d. (n = 45).

Table 3.3. Calcium onrates for GBPV WT and mutants. Onrates were calculated from the 1 1 7 simple relationship kon = koff/Kd. The units for all kon values are M s x 10 .

Temperature (°C) WT K8N K26N DM

5 8.48 8.19 6.70 25.49 15 10.83 7.13 9.04 17.17 25 6.20 6.54 6.17 10.51

Discussion

The steadystate Ca2+ binding measurements reported here confirm our hypothesis that the single mutants, K8N and K26N, would show a rightshifted thermal sensitivity pattern intermediate between the polar and warm/temperate patterns observed previously for PV (Erickson., et al 2005). The double mutant, DM, showed a full conversion of phenotype to that of a warm/temperate adapted PV, indicating the successful recapitulation of the evolutionary steps toward thermal compensation in Antarctic fish PV. Additionally, these results show that the WT

PV is more thermolabile than the mutant forms showing a sharp increase in Kd at the highest measurement temperature. Erickson et al (2005) found that at physiological temperatures the cold and temperateadapted PVs had a similar free energy change (G), but the coldadapted PV had a smaller enthalpy and larger entropy change upon Ca2+ binding as determined by isothermal titration calorimetry. Thus, it appears that changes in flexibility/stability drive the differences seen in binding ability among PV orthologs adapted to different thermal environments and that these changes in flexibility can be explained by small changes in primary structure (K8N and K26N) at positions distant from the ligand binding sites. In order to further characterize the evolution of thermal compensation in Ca2+ binding ability of Antarctic fish PV we determined the unidirectional rate constants of Ca2+ binding. Our results show an interesting pattern. K8N and K26N have a slower offrate than WT (Table 3.2;

Figure 3.11), but very similar onrates (Table 3.3). This suggests that the differences in Kd between the single mutants and the WT are mediated by a stabilization of the Ca2+ bound state. DM, alternatively, does not show a further decrease in offrate indicating that the further increase in binding affinity seen with both substitutions is mediated through a faster onrate suggesting a stabilized apostate. The results reported here allow for the development of a statistical thermodynamic model of PV Ca2+ binding based on the idea of folding/binding energy funnels (Leopold et al., 1992; Miller and Dill, 1997; Dill and Chan, 1997). The folding of proteins can be described by an energy landscape where linear strings of amino acids with high entropy fold through multiple energy pathways leading to the same end state: a folded protein. This model captures the multiple conformations and folding intermediates found experimentally (Dill and Chan, 1997). This idea has been expanded to capture ligand binding (Miller and Dill, 1997; Tsai et al., 1999;

Ma et al; 1999). In this formulation of the model different types of ligand binding are described by different energy landscapes. An energy funnel with a rugged bottom corresponds to “promiscuous” proteins that bind multiple substrates. At the opposite end of the spectrum would be an energy funnel with a single, smooth minimum that describes highly selective, rigid binding (Ma et al, 1999). Applying these ideas to PV Ca2+ binding allows an explanation that accounts for experimental results measuring binding ability described here, thermodynamic data presented elsewhere (Erickson et al., 2005) and the currently held view of PV function in muscle physiology as well as confirming the corresponding states theory (Hochachka and Somero, 2002). In the sarcoplasm of resting fasttwitch skeletal muscle fibers, PV is loaded with Mg2+ due to the prevailing high [Mg2+] and low [Ca2+]. In our model (Fig. 3.12) this corresponds to one of two smooth energy minima. The other minimum describes the Ca2+ bound state (PV Ca2+ affinity is higher than Mg2+, thus, the lower minimum for Ca2+). In order for PV to function as an effective Ca2+ buffer and increase the speed of muscle relaxation during prolonged stimulations (e.g. burstswimming in fish), it must exchange Mg2+ for Ca2+ during the high [Ca2+] present in contracting muscle. Once Mg2+ moves off of PV, PV goes through a conformational change to a relatively disordered apostate (Henzl and Tanner, 2007; 2008). Ca2+ binding affinity is affected by the structural rearrangements necessary to move from the apostate to the Ca2+ bound state. By shifting the stability of the apostate, the onrate of Ca2+ can be shifted consequently shifting Ca2+ affinity. A difference in apostate stability is proposed as a mechanism for the observed higher binding affinity of rat αPV as compared to rat βPV (Henzl and Tanner, 2008). Additionally, affinity could be shifted by adjustments in bound state stability. The same idea holds for Mg2+ affinity which could also be shifted by adjusting the stability of the apo or bound state. Mg2+ binding by PV has been shown to vary with temperature in a similar manner to Ca2+ binding (Erickson and Moerland, 2005). In each of these cases shifting stability involves changes in conformational flexibility. As a molecule moves higher in the funnel it has higher entropy, higher energy and a higher conformational flexibility sampling more microstates. This provides us with a model of how to adjust PV function. By adjusting the stability and amount of conformational space sampled by the apo and bound states, the binding affinity can be changed. In the case of the PV studied in the present effort, flexibility is being adjusted in response to temperature. Merging this model with

57 the theory of corresponding states to describe thermal adaptation in orthologs yields the following explanation: As temperature is lowered a protein moves further down the funnel to the lower energy states where it samples a smaller number of conformations with a higher percentage that are binding competent. This leads to a tighter binding affinity that could be more than the physiologically optimal level. To compensate for decreased temperature amino acid substitutions adjust the inherent flexibility of the protein shifting it back up the energy funnel allowing optimal binding affinity at the colder physiological temperature. Our results can be explained using this model as applied to thermal compensation in Antarctic fish PV (Fig. 3.12). In order to compensate for the cooling temperatures of the Southern Ocean, as notothenioid PV evolved from the Perciformes ancestral state (corresponding to the DM isoform), first the single substitutions at positions 8 or 26 removed hydrogen bonds leading to a more flexible apostate and slower Ca2+ onrates. While Mg2+ and Ca2+ are exchanging, the single mutants’ apostate has higher entropy than the Perciformes ancestral PV and DM meaning it is sampling more conformational microstates. Most of these states will not 2+ be binding competent leading to the observed increased Ca Kd. These single mutants correspond to a transistionary period of a coldtemperate thermal environment. Then as the environment continued to cool, the removal of both hydrogen bonds destabilized the boundstate leading to the faster Ca2+ offrate seen in the WT. These subtle adjustments in primary structure led to adjustments in entropy/conformational flexibility in either the apo or bound states shifting the thermal sensitivity pattern of PV allowing functional compensation as the thermal environment changed to the current subzero state of the Southern Ocean.

Figure 3.12. Proposed mechanism for thermal compensation in PV. We propose that the two single mutants stabilize the bound state leading to slower offrates and similar onrates as compared to WT while the two substitutions together stabilize the apostate leading to lower entropy (smaller number of microstates) and faster onrates. This provides a structural mechanism to explain the evolution of thermal compensation in PV from Antarctic notothenioid fish.

CHAPTER FOUR

CONCLUSIONS

Antarctic fish of the suborder Notothenioidei inhabit what would seem to humans to be a very harsh environment. These organisms, however, have evolved a suite of adaptations that have allowed them to become the dominant group of fish in the freezing waters of the Southern Ocean. Notothenioids are thought to have evolved from a temperate, benthic ancestor and now fill a variety of niches (Eastman 1993). Key to the evolutionary success of this group of fish is the functional “tuning” of proteins to compensate for changes in habitat temperature. Indeed, this phenomenon plays a role in the fitness of all organisms. A characteristic pattern has been found for protein function in response to changing temperature. At common measurement temperatures orthologs from coldadapted organisms show a weaker ligand binding affinity than warm or temperateadapted counterparts, but at physiological temperature function is highly conserved. This characteristic trend in thermal sensitivity patterns from polar and temperate fish was demonstrated in PVs from Antarctic fish and fish from warmer habitats (Erickson et al., 2005). At a common measurement temperature PV from Antarctic fish had a higher Ca2+ affinity than carp PV, a temperate counterpart, but at physiological temperatures binding affinity was conserved. Additionally, this study found that at physiological temperatures the cold and temperate PVs had a similar free energy change (G), but the cold PV had a smaller enthalpy and larger entropy change upon Ca2+ binding as determined by isothermal titration calorimetry (Erickson et al., 2005). This leads to the explanation using the corresponding states theory outlined in Chapter 1 that the coldadapted PV has a more flexible structure, allowing it to sample more conformational microstates (larger entropy) compensating for the decreased temperature. Without an increase in flexibility mediated through changes in primary structure the cold temperature would cause an increase in binding

affinity (lower Kd) beyond optimal levels. The study presented here sought to determine the underlying structural correlates and physicochemical mechanisms responsible for the displayed thermal sensitivity pattern in Antarctic notothenioid PV. A simple sequence alignment of notothenioid PV sequences with those from temperate fish counterparts showed that the comparison was confounded by the large number of sequence differences that have accumulated since these groups shared a common

60 ancestor (approximately 200250 million years ago). To counter this “phylogenetic noise” we took an evolutionary approach, namely ancestral sequence reconstruction. By mapping PV sequences to an evolutionary tree of teleost fish we were able to reconstruct the Antarctic notothenioid ancestral PV sequence and the sequence of its most recent temperate ancestor corresponding to an ancestral fish of the Perciformes order. Comparison of the notothenioid and Perciformes ancestral PVs narrowed the set of candidate substitutions from 23 to 16. We further narrowed the subset by viewing the nonconservative substitutions in a three dimensional context through homology modeling. Using the homology model of PV from Gobionotothen gibberifrons (GBPV WT) as a template structure, we performed virtual mutagenesis to determine the effects of the nonconservative substitutions. The subset was finally narrowed to two substitutions, at position 8 and 26, that appear most likely to affect function. Either of these substitutions reintroduces a single hydrogen bond into the WT notothenioid structure to match the temperate Perciformes ancestral isoform. Interestingly, these same two hydrogen bonds are in the structure of PV from the temperate fish, Cyprinus carpio (CCPV). In addition to viewing the structural effects of the substitutions, we estimated the free energy of folding for all PV structures in our search set. Each of the two substitutions, K8N and K26N, shifted the free energy of folding of the template GBPV WT to an intermediate point between similar values found for the GBPV WT and notothenioid ancestral PV and similar values found for the Perciformes ancestral PV and CCPV. Adding both substitutions to the template structure completed the shift in free energy of folding. This led to the following hypothesis: thermal compensation in Antarctic notothenioid PV was accomplished by the evolutionary loss of two hydrogen bonds from a warm/temperateadapted ancestral PV. Reintroduction of the hydrogen bonds individually will cause an intermediate increase in Ca2+ binding affinity and adding both hydrogen bonds will recapitulate the Perciformes ancestral PV phenotype. Additionally, we predicted that function would be conserved at what would be the ancestral PV’s estimated habitat temperature. That is, when compared to the binding affinity of GBPV WT at low temperatures, the single mutants would show an equivalent binding affinity at intermediate temperatures and the double mutant at high temperatures. To test this hypothesis we introduced the identified substitutions via sitedirected mutagenesis and then expressed the constructs in bacteria. Once purified, the proteins were stripped of contaminating cations to allow accurate determination of Ca2+ dissociation constants

61 using a competition assay based on the fluorescent indicator fluo3. The results of these assays confirmed our hypothesis. Reintroduction of the temperate ancestor associated hydrogen bonds 2+ into GBPV WT showed a rightshifted phenotype of Ca Kd. Additionally, the double mutant showed a full conversion of phenotype to that characteristic of a temperate ortholog.

Interestingly, the WT construct showed a sharp increase in Kd at high temperature as compared to the single and double mutants. This indicates a lower thermal stability for the WT and hints at a mechanism for the increased binding affinity seen in the mutant constructs. The reintroduced hydrogen bonds increased the stability of the protein matching the prediction from the corresponding states theory that the single and double mutants would be sampling a smaller subset of conformational states. In order to determine a more specific mechanism for the observed shift in Ca2+ binding affinity, unidirectional rate constants were measured using a stoppedflow spectrometry technique with terbium fluorescence as a reporter. The stoppedflow results did not show a one toone correspondence with the Kd data, but they revealed a possible mechanism for the structural changes that lead to the functional “tuning” as identified by the fluo3 competition assay. We proposed a model based on the binding energy funnel model (Leopold et al., 1992; Tsai et al., 1999) in which the two single mutants, K8N and K26N, had slower offrates than the WT isoform, but had very similar onrates, indicating a stabilization of the Ca2+ bound state and a lower level of entropy in the energy funnel model. The double mutant, in contrast, displayed a very similar pattern of offrates as compared to the single mutants, but with increased binding

affinity (lower Kd) indicating a pattern of faster onrates than the single mutants and the WT isoform. This suggests that having both of the reintroduced hydrogen bonds provides additional stabilization, this time in the apostate, leading to faster Ca2+ on rates and a higher apparent affinity. Due to the high sequence similarity among notothenioid PVs and the demonstrated equivalence in thermal sensitivity of Ca2+ binding function of two notothenioid PVs, our results provide a general model for the evolution of thermal compensation in PVs from all Antarctic notothenioid fish. In summary, the present results demonstrate the apparent evolutionary steps taken to achieve thermal compensation in notothenioid PV. The loss of two hydrogen bonds is sufficient to explain the observed thermal sensitivity patterns in extant notothenioid PVs as compared to temperate counterparts through adjustments of the contributions in entropy

(conformational flexibility) and enthalpy of Ca2+ binding. These results are explained through a binding energy funnel model that is consistent with the corresponding states theory originally developed to explain observed thermal compensation in enzymatic proteins.

APPENDIX A

AMINO ACID AND CODING DNA SEQUENCES FOR ANTARCTIC NOTOTHENIOID PARVALBUMINS

Dissostichus mawsoni

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAATCCCAAGTAAAAAACAA 50

------|------|------|------|------| 51 AAatggcacttgcaggaaacctgaaggaggctgacatcactgcagccctc 100 1 M A L A G N L K E A D I T A A L 16

------|------|------|------|------| 101 gcagcatgcaaagctgctgagtctttcaagcacaaggaattcttcgccaa 150 17 A A C K A A E S F K H K E F F A K 33

------|------|------|------|------| 151 ggtcggcctgtccgccaagagcgccgatgacatcaagaaagctttcttgg 200 34 V G L S A K S A D D I K K A F L V 50

------|------|------|------|------| 201 tcattgaccaggaccagagtggcttcattgaggaggaggagctgaagctg 250 51 I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 ttcctgcagaacttctctgccggtgccagagctctgactgatgctgagac 300 67 F L Q N F S A G A R A L T D A E T 83

------|------|------|------|------| 301 caaggctttcctgaaggccggtgacatcgatggtgatggcatgatcggaa 350 84 K A F L K A G D I D G D G M I G I 100

------|------|------|------|------| 351 tcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAACA 400 101 D E F A T M V K A * 109

------|------|------|------|------| 401 TCCTACATCCCTTCTCTTAATTTAGGAACAACTTGAGCCAAGACATAGTC 450

------|------|------|------|------| 451 CTCTCTTTTCCCTTTCACTCTCTCTCTCCCATTCACTACACCAGTTTTAT 500

------|------|------501 ACAATTGACCTCTCACTCCCCTGACTG 527

Dissostichus eleginoides

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaacctgaaggaggctgacatcactgcagccct 100 1 M A L A G N L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcatgcaaagctgctgagtctttcaagcacaaggaattcttcgcca 150 17 A A C K A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgccgatgacatcaagaaagctttcttg 200 34 V G L S A K S A D D I K K A F L 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagctgaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccggtgccagagctctgactgatgctgaga 300 67 F L Q N F S A G A R A L T D A E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgatcgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTTAGGAACAACTTGAGCCAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCTCCCATTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Patagonotothen ramsayi

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcatgcacagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C T A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctggccgccaagagcgccgatgacatcaagaaagctttcggg 200 34 V G L A A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagctgaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccggtgccagagctctgactgacggtgaga 300 67 F L Q N F S A G A R A L T D G E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgatcgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagtccgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E S A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACGCTCTCTCTCCCACTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Trematomus hansoni

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacctgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A P A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcatgcacagctgctgagtctttcaaacacaaggaattcttcgcta 150 17 A A C T A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcttg 200 34 V G L S A K S A D D I K K A F L 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagctgaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttcgctgccggtgccagagctctgactgttgctgaga 300 67 F L Q N F A A G A R A L T V A E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgatcgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCAAGACAGTCC 450

------|------|------|------|------| 451 TCTCTTTTCCATTTCACTCTCTCTCTCCCACTCACTACACCAGTTTTATA 500

------|------|------501 CAATTGACCTCTCACTCCCCTGACTG 526

Lepidonotothen nudifrons

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAATCA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcatgcacagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C T A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcttg 200 34 V G L S A K S A D D I K K A F L 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagctgaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttcgctgccggtgccagagctctgactgatgctgaga 300 67 F L Q N F A A G A R A L T D A E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgatcgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCAATCTCTCTCTCCCACTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Gobionotothen gibberifrons

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcttgcaaagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C K A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcggg 200 34 V G L S A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagcttaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccggtgccagggctctgactgacggtgaga 300 67 F L Q N F S A G A R A L T D G E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggctggtgacatcgatggtgatggcatgattgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGATAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCAAGGCATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCTCCCACTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Notothenia rossii

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAATCCCAAGTAAAAAACC 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcttgcacagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C T A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcggg 200 34 V G L S A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagcttaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccggtgccagagcactgaccgtaggtgaga 300 67 F L Q N F S A G A R A L T V G E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgattgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCACGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCTCCCACTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Notothenia coriiceps

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcttgcaaagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C K A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcggg 200 34 V G L S A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggaggagcttaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccggtgccagagctctgaccgatggtgaga 300 67 F L Q N F S A G A R A L T D G E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgattgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCATTCTCTTAATTAAGGAACAACTTCAGCCAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCTCCCACTCACTACACCAGTTTTA 500

------|------|------501 TACAATTGACCTCTCACTCCCCTGACTG 528

Parachaenichthys charcoti

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAAACCCCAAGTAAAAACCA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcttgctcagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C S A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcggg 200 34 V G L S A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgaccaggaccagagtggcttcattgaggaggacgagctgaagct 250 50 V I D Q D Q S G F I E E D E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccagtgccagagctctgactgttgctgaga 300 67 F L Q N F S A S A R A L T V A E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggccggtgacatcgatggtgatggcatgattgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgccaccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A T M V K A * 109

------|------|------|------|------| 401 ATCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCAAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCCCACTCACTACACCAGTTTTATA 500

------|------|------501 CAATTGACCTCTCACTCCCCTGACTG 526

Pseudochaenichthys georgianus

------|------|------|------|------| 1 CAACTGAACGAATCCACTCTAGTCTCTAGATAATCCCCAAGTAAAAAACA 50

------|------|------|------|------| 51 AAAatggcacttgcaggaaccctgaaggaggctgacatcactgcagccct 100 1 M A L A G T L K E A D I T A A L 16

------|------|------|------|------| 101 cgcagcttgcaaagctgctgagtctttcaagcacaaggaattcttcgcta 150 17 A A C K A A E S F K H K E F F A K 33

------|------|------|------|------| 151 aggtcggcctgtccgccaagagcgctgatgacatcaagaaagctttcggg 200 34 V G L S A K S A D D I K K A F G 49

------|------|------|------|------| 201 gtcattgatcaggaccagagtggcttcattgaggaggaggagctgaagct 250 50 V I D Q D Q S G F I E E E E L K L 66

------|------|------|------|------| 251 gttcctgcagaacttctctgccagtgccagagctctgactgatgctgaga 300 67 F L Q N F S A S A R A L T D A E T 83

------|------|------|------|------| 301 ccaaggctttcctgaaggctggtgacatcgatggtgatggcatgattgga 350 84 K A F L K A G D I D G D G M I G 99

------|------|------|------|------| 351 atcgatgagttcgcctccatggtcaaggcttaaATGGAGAATCAGACAAC 400 100 I D E F A S M V K A * 109

------|------|------|------|------| 401 TTCCTACATCCCTTCTCTTAATTAAGGAACAACTTGAGCCAAGACATAGT 450

------|------|------|------|------| 451 CCTCTCTTTTCCCTTTCACTCTCTCTCTCACTCACTACACCAGTTTTATA 500

------|------|------501 CAAATGACCTCTCACTCCCCTGACTG 526

Chionodraco rastrospinosus