Evolutionary Convergence in Adaptation of Proteins to Temperature: A4-Lactate Dehydrogenases of Pacific Damselfishes (Chromis spp.)
http://www.100md.com
分子生物学进展 2004年第2期
Hopkins Marine Station, Stanford University, Pacific Grove, California 1
E-mail: somero@stanford.edu.
Abstract
We have compared the kinetic properties (Michaelis-Menten constant [Km] and catalytic rate constant [kcat]) and amino acid sequences of orthologs of lactate dehydrogenase-A (A4-LDH) from congeners of Pacific damselfishes (genus Chromis) native to cold-temperate and tropical habitats to elucidate mechanisms of enzymatic adaptation to temperature. Specifically, we determined whether the sites of adaptive variation and the types of amino acids involved in substitutions at these sites were similar in the Chromis orthologs and other orthologs of warm-adapted and cold-adapted A4-LDH previously studied. We report striking evolutionary convergence in temperature adaptation of this protein and present further support for the hypothesis that enzyme adaptation to temperature involves subtle amino acid changes at a few sites that affect the mobility of the portions of the enzyme that are involved in rate-determining catalytic conformational changes. We tested the predicted effects of differences in sequence using site-directed mutagenesis. A single amino acid substitution in a key hinge region of the A4-LDH molecule is sufficient to change the kinetic characteristics of a temperate A4-LDH to that of a tropical ortholog. This substitution is at the same location that was identified in previous studies of adaptive variation in A4-LDH and was hypothesized to be important in adjusting Km and kcat. Our results suggest that certain sites within an enzyme, notably those that establish the energy changes associated with rate-limiting movements of protein structure during catalysis, are "hot spots" of adaptation and that common types of amino acid substitutions occur at these sites to adapt structural "flexibility" and kinetic properties. Thus, despite the wide array of options that proteins have to adjust their structural stabilities in the face of thermal stress, the adaptive changes that couple "flexibility" to alterations of function may be limited in their diversity.
Key Words: temperature adaptation ? A4-LDH ? enzyme kinetics ? ldh-a gene ? ortholog evolution ? Chromis
Introduction
Proteins are highly sensitive to temperature, and, therefore, their structural and functional characteristics mirror the thermal conditions encountered by species during their evolutionary histories. Two general types of conclusions have been reached about adaptation to temperature by proteins (Fields 2001; Hochachka and Somero 2002). First, a high degree of conservation in the values of temperature-sensitive functional properties (Km and kcat) is seen among orthologs from differently adapted species at their normal body temperatures. Second, differences in global structural stability exist among orthologs, and these differences are proposed to be important in adaptation of function. In particular, the lower stabilities characteristic of orthologs from cold-adapted species have been proposed to be responsible for the higher Km and kcat values of these enzymes, a conjecture based on the concept of "flexibility" of protein conformation (Fields 2001).
Several key questions remain about the structure-function relationships associated with protein adaptation to temperature. First, how much change in primary structure is needed to effect adaptive change in function and structural stability? Second, are particular sites within the sequence and three-dimensional structure of the protein "hot spots" of adaptive change; that is, is there evolutionary convergence in the regions in which adaptive change occurs? Third, is there evolutionary convergence in the types of amino acid substitutions that characterize adaptation to temperature? Fourth, what physical-chemical principles explain how differences in amino acid composition lead to shifts in kinetic properties; for example, how do putative differences in "flexibility," whether global or local, lead to alterations in Km and kcat?
In addressing these questions an initial experimental challenge involves selecting a study system that maximizes the likelihood that functionally important differences in sequence can be detected and their effects interpreted mechanistically. Appropriate study systems should be characterized by the following properties. The protein to be studied should be well understood in terms of mechanism, specifically the relationship between the conformational changes that accompany binding and catalysis and values for kinetic parameters such as Km and kcat. The orthologs to be compared should be from phylogenetically closely related organisms adapted to different temperatures to allow adaptively important differences in sequence to be discerned most unambiguously. Orthologous variants of the A-isoform of lactate dehydrogenase (A4-LDH, EC 1.1.1.27:lactate:NAD+ oxidoreductase) from congeneric fishes adapted to different temperatures satisfy these criteria (Holland, McFall-Ngai, and Somero 1997; Fields and Somero 1998; Fields 2001). A4-LDH is well understood in terms of three-dimensional structure and the roles played by changes in conformation during the catalytic cycle (Abad-Zapatero et al. 1987; Dunn et al. 1991). Previous studies of orthologs of A4-LDH have revealed consistent patterns of adaptive variation in Km and kcat and have provided at least initial indications of how these adaptive differences in kinetic properties are established through variation in sequence (Fields 2001).
The present study exploited congeners of temperate and tropical damselfishes (genus Chromis) to extend the analysis of adaptive structure-function variation in A4-LDH. Two tropical species from identical thermal niches were included in the comparisons to facilitate separating neutral mutations from those that underlie temperature-adaptive variation in function. We were particularly interested in determining whether sites of putative adaptive variation identified in previous studies were also sites of change in A4-LDHs of Chromis congeners, and, if so, whether the types of amino acid substitutions at these sites were similar to those observed in earlier studies. Thus, our focus was on the question of evolutionary convergence in location of adaptive change and in the mechanism used to alter functional properties. We tested empirically the hypothesized importance of observed differences in sequence among the orthologs using site-directed mutagenesis. Our results indicate that different evolutionary lineages may show markedly similar patterns of temperature-adaptive change in enzyme structure and function.
Materials and Methods
Species Choice and Specimen Collection
The temperate Blacksmith (Chromis punctipinnis) is one of the most cold-adapted damselfish species. Its biogeographic range extends from Monterey Bay, California (122°W, 37°N) to Punta San Pablo, Baja California, Mexico (115°W, 27°N) covering the San Diego and southern portion of the Oregon biogeographic provinces (Miller and Lea 1972). Over this latitudinal range, Blacksmith experience water temperatures from 10°C to 22°C on an annual basis. The Dusky Chromis, C. caudalis, and the Yellow-axil Chromis, C. xanthochira, have similar distributions on outer reef slopes of islands throughout the tropical western Pacific (Allen 1991; Myers 1999). These species experience a much warmer and more stable thermal environment of approximately 29°C, with seasonal variation typically less than 1°C (Levitus and Boyer 1994). All specimens were hand collected by SCUBA diving. Chromis punctipinnis individuals were collected from Monterey, California and La Jolla, California (ambient seawater temperature 12°C at both collecting occasions; no differences in enzyme kinetics or cDNA sequence were observed in fish from either location). Chromis caudalis and C. xanthochira were collected on outer reef slopes near Koror, Republic of Palau, in the tropical western Pacific. White muscle tissue samples were taken immediately after killing the fish, frozen in liquid nitrogen, and stored at –70°C.
A4-LDH Purification and Kinetic Characterization
A4-Lactate dehydrogenase was purified from white muscle by oxamate affinity chromatography as described previously (Yancey and Somero 1978). All purifications showed only a single band on both silver-stained SDS-PAGE (corresponding to the LDH monomer) and activity-stained native PAGE (corresponding to the A4 tetramer) gels.
All kinetic assays were performed spectrophotometrically in 80 mM imidazole·HCl buffer, pH 7.0 at 20°C, by measuring the conversion of NADH to NAD+ at 340 nm (Yancey and Somero 1978). Measurements were made on the purified and dialyzed A4-LDH within 24 h of purification using a Shimadzu BioSpec 1601 UV/Vis spectrophotometer with temperature control (±0.1°C) and UVPC Kinetics software (Shimadzu, Kyoto, Japan). Kinetic studies were conducted from 5.0°C to 35.0°C at 5-degree intervals. Assays were performed at eight pyruvate concentrations covering the range of one-third to three times the predicted Km, with three replicates at each concentration. Enzyme concentration was calculated spectrophotometrically using the Pierce Coomassie protein assay reagent. The Michaelis-Menten constant, Km, and maximal velocity, Vmax, were calculated by the weighted linear regression method (Wilkinson 1961). The catalytic constant, kcat, was calculated from Vmax and the enzyme concentration.
Nucleic Acid and Deduced Amino Acid Sequences
Total RNA was extracted from white muscle tissue using Trizol (Invitrogen, San Diego, Calif.). From this, cDNA was reverse transcribed using an oligo-dT primer. Single-stranded cDNA was used as template for PCR using two damselfish-specific primers in the 5' and 3' UTR regions (PomLDH-5'S 5'-GACACCCACACGGCAGCAGCACTTCT-3' and PomLDH-3'A 5'-ACGGGAGGTATGGATTGAGGGAGTC-3'; PCR conditions: 28 cycles of 94°C for 20 s, 55°C for 30 s, and 72°C for 90 s). PCR products were gel purified and cloned into the pGEM T-Easy vector (Promega, Madison, Wis.). The complete ldh-a inserts were sequenced in both directions from two independent clones for each of four individuals from each species, using an ABI 373 XL automated sequencer. Sequence analysis was performed with the DNAStar suite of software (DNAStar, Inc., Madison, Wis.), and the presumptive amino acid sequences were inferred from the ldh-a cDNA sequences.
To corroborate the phylogenetic relatedness of these congeners, a 526-bp fragment of the mitochondrial cytochrome b gene was PCR amplified and sequenced using primers designed from GenBank sequences AF119392 and AF119395 for a congener and closely related confamilial, respectively (Primers: Chromis-CBL 5'-CAACCCCCGCCAACATTTCAG-3' and Chromis-CBH 5'-GAGCTCAGGCCAGTTGGGTTATTG-3'; PCR conditions: 35 cycles of 94°C for 20 s, 48°C for 30 s, and 72°C for 90 s). The phylogenetic analysis of the aligned sequences was performed using the PAUP* computer program under various distance, parsimony, and maximum-likelihood models (Swofford 2002).
Site-Directed Mutagenesis and in Vitro Expression
To assess the importance of specific amino acid substitutions between tropical and temperate LDH-A sequences, we utilized site-directed mutagenesis of the cloned ldh-a gene coupled with the in vitro overexpression of the protein. To accomplish this, the ldh-a gene from the temperate C. punctipinnis was taken as the starting wild-type sequence and placed into an expression vector (pAlter-Ex1 [Promega]) under control of the tac promoter. A Histidine Affinity Tag (BD Biosciences Clontech, Palo Alto, Calif.) was incorporated onto the 5' end of the gene, producing a histidine-rich, 31–amino acid epitope on the N-terminal end of the LDH-A monomer. An enterokinase cleavage recognition site was incorporated with this immediately preceding the serine at position 1 in the native LDH-A sequence. This plasmid construct, representing the recombinant wild-type ldh-a gene (rLDH-wt), was then transformed into E. coli JM109 cells.
After growing the transformed cells to log-phase growth, rLDH expression was induced with 1 mM isopropyl ?-D-thiogalactopyranoside (IPTG) for 4 h. The cells were then lysed and centrifuged to collect the expressed protein. This resulting protein mixture was further purified by first applying the supernatant to an immobilized metal affinity chromatography column (IMAC [Clontech]), which retained only highly charged proteins, including the rLDH with the histidine tag. The rLDH was then further purified using oxamate affinity chromatography, resulting in pure rLDH as measured by a single band of predicted molecular weight on a silver-stained SDS-PAGE gel. The purified protein was then cleaved with enterokinase (Roche Applied Sciences, Indianapolis, Ind. [1:20 w/w, 24 h at 37°C]). The cleaved protein was once again purified by oxamate affinity chromatography. The resulting recombinant protein showed a single band of the same molecular weight as the C. punctipinnis LDH-A on a silver-stained SDS-PAGE gel.
After confirming that the recombinant protein had the same Michaelis-Menten kinetics as the C. punctipinnis A4-LDH, specific amino acid substitutions were incorporated by site-directed mutagenesis utilizing the Altered Sites II system (Promega). Each mutant was purified in the same manner as the recombinant wild-type LDH. Finally, the Michaelis-Menten kinetic constant, Km, was measured for each recombinant protein following the procedure listed above at an intermediate temperature of 20°C.
Results
Temperature Adaptation of Kinetic Properties
The values of Kmpyr and kcat for the three orthologs reflect the pattern found in earlier studies of A4-LDHs of vertebrates: a cold-adapted ortholog has significantly higher values for Km and kcat than warm-adapted orthologs, across the range of measurement temperatures, (fig. 1). However, at their respective physiological temperatures (10°C to 22°C for C. punctipinnis and 29°C for C. caudalis and C. xanthochira), the orthologs have similar kinetic properties. Kmpyr is essentially the same for orthologs from all species (fig. 1, upper panel) and kcat exhibits a high degree of compensation to temperature (fig. 1, lower panel).
FIG. 1. A4-lactate dehydrogenase enzyme kinetic constants for one temperate and two tropical species of Chromis across temperatures spanning the physiological range of all species. (A) The apparent Michaelis-Menten constant, Km, of pyruvate (±SE) for the temperate Chromis punctipinnis (open circle) as compared with two tropical species, C. caudalis (shaded square) and C. xanthochira (shaded triangle). (B) The catalytic constant, kcat, for the same species and temperatures. Arrows indicate the environmental temperature range experienced by temperate (open arrows) and tropical (solid arrows) Chromis species
Sequence Differences and Their Location in the Three-Dimensional Structure
The deduced amino acid sequences reveal 13 polymorphic sites among the three species (fig. 2). Sequencing of multiple individuals of a species revealed no amino acid polymorphisms. Given the observed differences in the kinetic properties of temperate-adapted and tropical-adapted orthologs, we reasoned that those sites that have amino acid residues that differ between orthologs from different environments, but are the same in orthologs from the same environment, might be the most likely candidates for causing the observed kinetic differences for orthologs from different environments. Only four such polymorphic sites matched these criteria (fig. 2).
FIG. 2. Alignment of the presumptive amino acid sequence of lactate dehydrogenase-A from three species of Chromis. The complete sequence of the temperate species, C. punctipinnis, is given, and identity to this sequence in the two tropical species, C. caudalis and C. xanthochira, is indicated by (.). Numbering reflects the removal of the amino-terminal methionine residue. Secondary structure motifs, as deduced from the crystal structure of the dogfish A4-LDH (Abad-Zapatero et al. 1987), are included as a guide. Amino acid substitutions from the temperate ortholog that are shared between both tropical orthologs are highlighted in bold
If these four sites are in fact important in the adaptation of A4-LDH to temperature, then the type of substitution, as well as its location in the three-dimensional structure, should be of great importance. The locations of the four sites in the native tetramer were determined using a crystal structure model of C. caudalis LDH-A that was derived from the known crystal structures of the dogfish, pig, and human LDH-A (PDB codes 1LDM [Abad-Zapatero et al. 1987], 9LDT [Dunn et al. 1991], and 1I10 [Read et al. 2001], respectively). The four sites were at the base of the 1G-2G helix, the base of the H helix, and within the ?J-1G loop. Indeed, this loop has the highest thermal vibrational B factor in the crystal structure, partially representing the disordered nature of this loop (Abad-Zapatero et al. 1987) but also because of its high exposure to solvent even in the tetrameric form (Read et al. 2001). These two helices and loop, along with the catalytic loop itself, are the regions of the enzyme that undergo the largest movements during the conformational changes that accompany substrate and cofactor binding (shaded regions of fig. 3 [Dunn et al. 1991; Gerstein and Chothia 1991]).
FIG. 3. Three-dimensional model of a single LDH-A monomer from the tropical damselfish, Chromis caudalis. The major moving portions of the molecule are darkly shaded and labeled. The position of the ThrAla (temperatetropical) mutation at position 219 is also indicated. This model was produced by the SWISS-MODEL homology modeling program using the crystal structure of LDH-A holoenzyme from dogfish, pig, and human (PDB codes 1LDM, 9LDT, and 1I10, respectively), and visualized with the Swiss-PdbViewer program (Guex and Peitsch 1997)
Site-Directed Mutagensis: Targets and Effects
To test the functional importance of different amino acid substitutions at the four sites hypothesized to be important in adaptation to temperature, we utilized site-directed mutagenesis and in vitro expression. Although the Km of the recombinant wild-type enzyme, rLDH, was not significantly different from the Km of the temperate ortholog, the different mutant substitutions had varying effects on Km (fig. 4). At one extreme, the GluAsp substitution at position 310 at the base of the H helix actually resulted in a slight increase in Km. The AspGlu substitution at position 220 in the ?J-1G loop appears to have no effect, and the LysAsn substitution at position 225 immediately preceding the 1G-2G helix appears to have only a moderate effect on lowering the Km from a temperate to a tropical-like value. Finally, the most radical substitution, ThrAla at position 219 in the ?J-1G loop, appears sufficient, by itself, to lower the Km from a value associated with a temperate ortholog to a value comparable with a tropical ortholog.
FIG. 4. Apparent Michaelis-Menten constant of pyruvate (±SE) at 20°C for different native and recombinant ldh-a genotypes. The purified recombinant A4-LDH (rLDH-wt) has the exact amino acid sequence as the temperate ortholog. Each of the other recombinant genotypes has the same amino acid sequence as the recombinant wild-type, with the exception of the amino acid substitution indicated. The recombinant wild-type, E310D, and D220E are not significantly different from the ortholog of the temperate C. punctipinnis, but do differ from the tropical C. caudalis or C. xanthochira orthologs. T219A is not significantly different from the tropical orthologs, but is different from the temperate ortholog. The intermediate K225N is not significantly different from orthologs of any species (95% confidence intervals; Dunn-idák adjustment for multiple means)
Discussion
These results corroborate a growing base of research demonstrating a conservation of biochemical properties among enzyme orthologs when examined under their respective physiological conditions (fig. 1). Foremost among these is the observation that, at any one temperature and for both the Michaelis-Menten constant, Km, and the catalytic constant, kcat, the tropical adapted species show more efficient substrate binding and a lower catalytic efficiency. Conversely, when all orthologs are compared at their respective physiological temperatures, both the substrate binding and the catalytic efficiency are roughly comparable.
These opposite trends in substrate binding and catalytic efficiency are likely the result of the balance of changes in enthalpy and entropy associated with a posited stability/flexibility compromise. The reaction catalyzed by orthologs of A4-LDH from species adapted to different temperatures exhibits a regular variation in thermodynamic activation parameters (Low, Bada, and Somero 1973), a trend observed for other enzymes as well (Lonhienne, Gerday, and Feller 2000). The reactions of cold-adapted species have lower enthalpies of activation (H) and more negative entropies of activation (S) than do the reactions catalyzed by warm-adapted orthologs. Thus, for cold-adapted orthologs relative to warm-adapted orthologs, the enthalpy barrier to the reaction is less, although the enzyme becomes relatively more ordered during the rate-limiting step in the reaction. The free energy of activation (G = H – T S), which determines kcat, is lower for cold-adapted orthologs because the enthalpy term has the stronger contribution to G.
Now that the role of catalytic conformational changes in the LDH reaction is understood (Dunn et al. 1991; Gerstein and Chothia 1991), it is possible to interpret this variation in activation parameters. The rate-limiting conformational changes that accompany binding of substrate and cofactor, leading to the establishment of the catalytic vacuole (Dunn et al. 1991), are endothermic and occur with a decrease in the entropy of the protein. More warm-adapted enzymes must increase their rigidity to maintain an adequate proportion of binding-competent microstates of the enzyme to allow effective substrate binding at the physiologically relevant substrate concentrations and temperature. The increased rigidity creates a higher enthalpy barrier for conformational changes and results in higher activation enthalpy (H). This increased rigidity also leads to a lower (less negative) activation entropy (S), as there is likely a narrower distribution of conformational microstates associated with a more rigid molecule such that the formation of the catalytic vacuole involves a smaller increase in order relative to the more flexible cold-adapted ortholog.
To examine the mechanisms that might explain the observed differences in kinetics between the orthologs of the tropical and temperate damselfishes, we first examined the sites where sequence differences exist that could potentially affect the energy changes that occur during catalysis. After this analysis, we exploited site-directed mutagenesis to alter residues that we hypothesized might be important in adaptation to temperature. In designing our experiments, we reasoned that the amino acid substitutions that would be most likely responsible for the differences in kinetics between tropical and temperate orthologs would be those residues that are the same in the tropical orthologs but different in the temperate ortholog. In this way, we were able to reduce the 13 polymorphic sites to just four that were the most likely candidates for providing functionally significant differences between orthologs (fig. 2).
Structural analysis yielded a number of correlations with previous results from studies of A4-LDH evolution. Most significantly, the location of the candidate substitutions for adaptive variation on the three-dimensional crystal structure model is restricted to a limited number of sites in the LDH-A monomer that previously have been identified as regions that undergo large movements during function and manifest substantial sequence variation among species (fig. 3 [Dunn et al. 1991; Gerstein and Chothia 1991]). It is the movement of these portions of the molecule, notably helices 1G-2G, that are the rate-limiting steps in the reaction. Other components of the enzyme move during catalysis as well; for example, the catalytic loop that comprises residues 95 to 112 in our residue numbering system. However, because the catalytic loop contains residues that interact with the substrate and cofactor, its sequence is nearly invariant across all known ldh-a sequences (see below). Thus, adaptive change that modulates the energetics of catalytic conformational changes may be restricted to only some of the "moving parts" of the enzyme. Consequently, the primary candidates for adaptive change are two helical regions, 1G-2G and H, plus the disordered loop region that acts as a hinge for the G-2G helix. The four amino acid polymorphisms implicated in differences between orthologs of tropical and temperate damselfishes were located in these regions. These are the same regions of the molecule implicated as "hot-spots" for cold-adaptation in Antarctic notothenioids (Fields and Somero 1998). Thus, in the context of adaptive mechanisms for adjusting kinetic properties, this represents an instance of enzymatic convergent evolution—separate evolutionary lineages evolving amino acid substitutions in the same locations of the molecule to effect the same sorts of kinetic changes associated with different adaptation temperatures.
One apparent exception to this paradigm is found in the comparison of A4-LDH orthologs from different temperate and cosmopolitan barracuda species (Holland, McFall-Ngai, and Somero 1997). Significant differences in Km and thermal stability were found among orthologs, but amino acid differences that were likely responsible were located in the N-terminal regions of the LDH-A molecule, away from the major moving portions of the LDH-A subunit. However, the substitutions they observed at position 8 and at positions 61 to 68 interact with the H helix and with residues adjacent to the 2G helix of other subunits in the quaternary structure of A4-LDH, respectively (Abad-Zapatero et al. 1987). Thus, the substitutions in the warm-adapted orthologs may serve to stabilize the quaternary structure and thereby attenuate the movements of the major moving portions of the molecule without altering the function of the catalytic loop. In particular, substitutions in the N-terminal helix could affect the ease with which the 1G-2G helices move, thereby affecting the energetics of the conformational change thought to be rate-limiting to catalysis (Dunn et al. 1991). Therefore, alterations in the energetics of catalytic conformational changes can arise from shifts in the intrinsic flexibility of portions of the individual LDH-A subunit or from changes in subunit-subunit interactions that alter the ease with which catalytic conformation changes can take place.
The question might arise as to whether the two different tropical species examined here in fact represent a sort of convergence in themselves in that they might have each evolved separately to the tropics via independent molecular mechanisms. However, our independent cytochrome b mitochondrial DNA phylogeny of the species examined clearly demonstrates that the two tropical Chromis species are much more closely related to each other (1.0% maximum-likelihood corrected sequence divergence) than either is to the temperate Chromis species (18.3% average maximum-likelihood corrected sequence divergence). Accepting all of the caveats of a molecular clock and assuming a standard rate calibration (e.g. Johns and Avise 1998), this tentatively suggests that temperate C. punctipinnis has diverged from the tropical C. caudalis and C. xanthochira over 9 MYA, whereas the two tropical species have diverged less than 0.5 MYA. This is consistent with a single adaptation to the tropical environment followed by a small amount of divergence between tropical species, but the alternative possibility of a single adaptation event of the temperate Chromis to colder waters cannot be excluded. Regardless, the majority of Chromis species, as well as the majority of damselfishes overall, represent broadly distributed species occupying tropical and subtropical environments (Allen 1991). This distribution pattern combined with the lower tropical ocean temperatures in the geologic past (e.g., Lea, Pak, and Spero 2000) suggests that the ancestral Chromis of the species examined here likely had an intermediate environmental regime, and these species represent divergence to two different extremes.
To ascertain the functional significance for changes in enzyme kinetics brought about by each of the putatively important amino acid substitutions, we utilized site-directed mutagenesis and an in vitro expression system. The results of this analysis suggest several points about the importance of these sites: (1) not all of the substitutions change the Km in the predicted manner, (2) some substitutions only have marginal effects on the change in Km, and (3) in some cases (e.g., Thr219Ala), a single mutation can account for the full difference in Km between a temperate and a tropical ortholog (fig. 4). In the case of Glu310Asp, the shorter side chain associated with the aspartic acid residue in the tropical ortholog potentially reduces by one the number of hydrogen bonds available to stabilize the H helix. This likely results in a more flexible molecule that is less often in a binding-competent microstate and hence has a higher Km. This finding, along with the finding that Asp220Glu and Lys225Asn mutants only marginally reduce the Km, suggest that temperature adaptation for overall Km may reflect selection on the net global effects of all substitutions combined. The effects of the interaction of multiple substitutions on the overall change in the Km await further investigation. However, that fact that the single substitution of Thr219Ala was able to reduce the Km from a level comparable with the temperate Chromis punctipinnis to a level comparable with the tropical C. caudalis and C. xanthochira demonstrates that a single substitution in the highly mobile ?J-G loop region is sufficient to cause temperature-adaptive change in the enzyme.
In this case, a single transition mutation at position 658 in the ldh-a gene, from an adenine to a guanine nucleotide in the tropical ortholog, results in the replacement of the hydrophilic threonine amino acid with a hydrophobic alanine residue. The importance of this change in hydrophobicity is indicated by that fact that the substitution is located in a region of the molecule that not only is a key "hinge" for the moving 1G-2G helix (Fields and Somero 1998) but also is one of the most solvent-exposed loops of the tetrameric enzyme (Read et al. 2001). Thus, the increased hydrophobicity of this loop could potentially force it to remain closer to the overall molecule and result in a less flexible "hinge" for the movement of the 1G-2G helix. Interestingly, this same transition mutation is evident in the comparisons of Antarctic notothenioids to the temperate teleost consensus sequence (658 here; 655 in Fields and Somero's [1998] numbering system), with the adenine nucleotide encoding alanine in the warm-adapted ortholog, and the guanine nucleotide encoding threonine in the cold-adapted ortholog in both cases. Consequently, this may represent a striking example of convergent evolution at this single base pair of the ldh-a gene. However, the kinetic difference associated with this specific substitution in the notothenioids has yet to be determined.
Indeed, the significance of this surface loop, along with its putatively greater accessibility to solvent in the more cold-adapted ortholog, agrees well with other studies that have shown that psychrophilic enzymes tend to have very hydrophilic surfaces with improved solvent interactions leading to a less compact protein shell as compared with their mesophilic counterparts (for review see Feller and Gerday [1997]). Specifically, in the case of subtilisin, the psychrophilic ortholog contains four amino acid insertions in surface loops that additionally increase the solvent interactions and hence reduce the compactness and stabilization of the psychrophilic ortholog (Davail et al. 1994).
In an effort to elucidate how these observations correlated with the broader patterns of vertebrate ldh-a evolution, we examined the patterns of nucleotide and amino acid substitution of all ldh-a sequences available in GenBank (including the sequences presented here) representing 52 vertebrate sequences. The analysis of amino acid polymorphisms in LDH-A when combined with the three-dimensional structure of A4-LDH supports three points: (1) the active site and catalytic loop are almost entirely conserved across all LDH-A sequences, (2) the areas of highest polymorphism typically occur on the most solvent-exposed portions of the monomer, and (3) in the tetramer, there are two classes of hypervariable regions—those that are on the 1G-2G and H helices or their associated hinge regions and those that interact with these moving helices in the quaternary structure. These same trends are even more clearly apparent when the analyses are restricted to the 32 euteleost fish ldh-a sequences available in GenBank.
These conclusions for A4-LDH suggest that, for other enzyme systems, adaptation of orthologs to differing physiological temperature is perhaps more likely to occur in regions of the molecule that will have the greatest influence on changing the flexibility of the moving portions of the molecule but do not affect the movement, shape, or specificity of the catalytic vacuole. Those mutations that occur with relatively greater frequency but nonetheless have significant effects on temperature-dependent enzyme kinetics, such as the adenine to guanine transition at position 658 in ldh-a, may show repeated instances of convergent evolution. The ubiquity of such hot spots of adaptation in other proteins remains to be empirically ascertained, but this burgeoning era of proteomics, particularly as it spreads to nonmodel species, holds with it the promise of at least the potential of gleaning common strategies or constraints to the mechanism of physiological adaptation of enzymes. This not only would prove invaluable for our understanding of the evolutionary history of extant enzyme orthologs but also would provide insights into their capacity for future adaptive change for both natural invasions of new species into new ecological niches and also in response to global environmental change.
Supplementary Material
The sequences reported in this paper have been deposited in the GenBank database under accession numbers AY289557 to AY289562.
Acknowledgements
We would like to thank Peter A. Fields, Andrew Y. Gracey, and Anthony W. DeTomaso for their assistance and expertise with molecular techniques. Lori and Patrick Colin, Larry Sharon, Ron McConnaghey, and Ed Kisfaludy were a tremendous assistance in the collection of specimens. This work was funded by NSF grant IBN-0133184 to G.N.S.
Literature Cited
Abad-Zapatero, C., J. P. Griffith, J. L. Sussman, and M. G. Rossman. 1987. Refined crystal structure of dogfish M4 apo-lactate dehydrogenase. J. Mol. Biol. 198:445-467.
Allen, G. R. 1991. Damselfishes of the world. Mergus, Melle, Germany.
Davail, S., G. Feller, E. Narinx, and C. Gerday. 1994. Cold adaptation of proteins: purification, characterization, and sequence of the heat-labile subtilisin from the Antarctic psychrophile bacillus TA41. J. Biol. Chem. 269:17448-17453.
Dunn, C. R., H. M. Wilks, D. J. Halsall, T. Atkinson, A. R. Clarke, H. Muirhead, and J. J. Holbrook. 1991. Design and synthesis of new enzymes based on the lactate dehydrogenase framework. Phil. Trans. R. Soc. Lond. B Biol. Sci. 332:177-184.
Feller, G., and C. Gerday. 1997. Psychrophilic enzymes: molecular basis of cold adaptation. Cell. Mol. Life Sci. 53:830-841.
Fields, P. A. 2001. Review: protein function at thermal extremes: balancing stability and flexibility. Comp. Biochem. Physiol. (part A) Mol. Integr. Physiol. 129:417-431.
Fields, P. A., and G. N. Somero. 1998. Hot spots in cold adaptation: localized increases in conformational flexibility in lactate dehydrogenase A4 orthologs of Antarctic notothenioid fishes. Proc. Natl. Acad. Sci. USA 95:11476-11481.
Gerstein, M., and C. Chothia. 1991. Analysis of protein loop closure two types of hinges produce one motion in lactate dehydrogenase. J. Mol. Biol. 220:133-149.
Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modelling. Electrophoresis 18:2714-2723.
Hochachka, P. W., and G. N. Somero. 2002. Biochemical adaptation: mechanism and process in physiological evolution. Oxford University Press, Oxford, UK.
Holland, L. Z., M. McFall-Ngai, and G. N. Somero. 1997. Evolution of lactate dehydrogenase-A homologs of barracuda fishes (genus Sphyraena) from different thermal environments: differences in kinetic properties and thermal stability are due to amino acid substitutions outside the active site. Biochemistry 36:3207-3215.
Johns, G. C., and J. C. Avise. 1998. A comparative summary of genetic distances in the vertebrates from the mitochondrial cytochrome b gene. Mol. Biol. Evol. 15:1481-1490.
Lea, D. W., D. K. Pak, and H. J. Spero. 2000. Climate impact of late Quaternary equatorial Pacific sea surface temperature variations. Science 289:1719-1724.
Levitus, S., and T. P. Boyer. 1994. World ocean atlas 1994, Vol. 4. Temperature. Department of Commerce, Washington, DC.
Lonhienne, T., C. Gerday, and G. Feller. 2000. Psychrophilic enzymes: revisiting the thermodynamic parameters of activation may explain local flexibility. Biochim. Biophys. Acta 1543:1-10.
Low, P. S., J. L. Bada, and G. N. Somero. 1973. Temperature adaptation of enzymes: roles of the free energy, the enthalpy, and the entropy of activation. Proc. Natl. Acad. Sci. USA 70:430-432.
Miller, D. J., and R. N. Lea. 1972. Guide to the coastal marine fishes of California, Fish Bulletin No. 157. California Department of Fish and Game, Sacramento.
Myers, R. F. 1999. Micronesian reef fishes. 3rd edition. Coral Graphics, Guam.
Read, J. A., V. J. Winter, C. M. Eszes, R. B. Sessions, and R. L. Brady. 2001. Structural basis for altered activity of M- and H-isozyme forms of human lactate dehydrogenase. Proteins 43:175-185.
Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4b10. Sinauer Associates, Sunderland, Mass.
Wilkinson, G. N. 1961. Statistical estimations in enzyme kinetics. Biochem. J. 80:324-332.
Yancey, P. H., and G. N. Somero. 1978. Temperature dependence of intracellular pH: its role in the conservation of pyruvate apparent Km values of vertebrate lactate dehydrogenases. J. Comp. Physiol. 125:129-134.(Glenn C. Johns and George)
E-mail: somero@stanford.edu.
Abstract
We have compared the kinetic properties (Michaelis-Menten constant [Km] and catalytic rate constant [kcat]) and amino acid sequences of orthologs of lactate dehydrogenase-A (A4-LDH) from congeners of Pacific damselfishes (genus Chromis) native to cold-temperate and tropical habitats to elucidate mechanisms of enzymatic adaptation to temperature. Specifically, we determined whether the sites of adaptive variation and the types of amino acids involved in substitutions at these sites were similar in the Chromis orthologs and other orthologs of warm-adapted and cold-adapted A4-LDH previously studied. We report striking evolutionary convergence in temperature adaptation of this protein and present further support for the hypothesis that enzyme adaptation to temperature involves subtle amino acid changes at a few sites that affect the mobility of the portions of the enzyme that are involved in rate-determining catalytic conformational changes. We tested the predicted effects of differences in sequence using site-directed mutagenesis. A single amino acid substitution in a key hinge region of the A4-LDH molecule is sufficient to change the kinetic characteristics of a temperate A4-LDH to that of a tropical ortholog. This substitution is at the same location that was identified in previous studies of adaptive variation in A4-LDH and was hypothesized to be important in adjusting Km and kcat. Our results suggest that certain sites within an enzyme, notably those that establish the energy changes associated with rate-limiting movements of protein structure during catalysis, are "hot spots" of adaptation and that common types of amino acid substitutions occur at these sites to adapt structural "flexibility" and kinetic properties. Thus, despite the wide array of options that proteins have to adjust their structural stabilities in the face of thermal stress, the adaptive changes that couple "flexibility" to alterations of function may be limited in their diversity.
Key Words: temperature adaptation ? A4-LDH ? enzyme kinetics ? ldh-a gene ? ortholog evolution ? Chromis
Introduction
Proteins are highly sensitive to temperature, and, therefore, their structural and functional characteristics mirror the thermal conditions encountered by species during their evolutionary histories. Two general types of conclusions have been reached about adaptation to temperature by proteins (Fields 2001; Hochachka and Somero 2002). First, a high degree of conservation in the values of temperature-sensitive functional properties (Km and kcat) is seen among orthologs from differently adapted species at their normal body temperatures. Second, differences in global structural stability exist among orthologs, and these differences are proposed to be important in adaptation of function. In particular, the lower stabilities characteristic of orthologs from cold-adapted species have been proposed to be responsible for the higher Km and kcat values of these enzymes, a conjecture based on the concept of "flexibility" of protein conformation (Fields 2001).
Several key questions remain about the structure-function relationships associated with protein adaptation to temperature. First, how much change in primary structure is needed to effect adaptive change in function and structural stability? Second, are particular sites within the sequence and three-dimensional structure of the protein "hot spots" of adaptive change; that is, is there evolutionary convergence in the regions in which adaptive change occurs? Third, is there evolutionary convergence in the types of amino acid substitutions that characterize adaptation to temperature? Fourth, what physical-chemical principles explain how differences in amino acid composition lead to shifts in kinetic properties; for example, how do putative differences in "flexibility," whether global or local, lead to alterations in Km and kcat?
In addressing these questions an initial experimental challenge involves selecting a study system that maximizes the likelihood that functionally important differences in sequence can be detected and their effects interpreted mechanistically. Appropriate study systems should be characterized by the following properties. The protein to be studied should be well understood in terms of mechanism, specifically the relationship between the conformational changes that accompany binding and catalysis and values for kinetic parameters such as Km and kcat. The orthologs to be compared should be from phylogenetically closely related organisms adapted to different temperatures to allow adaptively important differences in sequence to be discerned most unambiguously. Orthologous variants of the A-isoform of lactate dehydrogenase (A4-LDH, EC 1.1.1.27:lactate:NAD+ oxidoreductase) from congeneric fishes adapted to different temperatures satisfy these criteria (Holland, McFall-Ngai, and Somero 1997; Fields and Somero 1998; Fields 2001). A4-LDH is well understood in terms of three-dimensional structure and the roles played by changes in conformation during the catalytic cycle (Abad-Zapatero et al. 1987; Dunn et al. 1991). Previous studies of orthologs of A4-LDH have revealed consistent patterns of adaptive variation in Km and kcat and have provided at least initial indications of how these adaptive differences in kinetic properties are established through variation in sequence (Fields 2001).
The present study exploited congeners of temperate and tropical damselfishes (genus Chromis) to extend the analysis of adaptive structure-function variation in A4-LDH. Two tropical species from identical thermal niches were included in the comparisons to facilitate separating neutral mutations from those that underlie temperature-adaptive variation in function. We were particularly interested in determining whether sites of putative adaptive variation identified in previous studies were also sites of change in A4-LDHs of Chromis congeners, and, if so, whether the types of amino acid substitutions at these sites were similar to those observed in earlier studies. Thus, our focus was on the question of evolutionary convergence in location of adaptive change and in the mechanism used to alter functional properties. We tested empirically the hypothesized importance of observed differences in sequence among the orthologs using site-directed mutagenesis. Our results indicate that different evolutionary lineages may show markedly similar patterns of temperature-adaptive change in enzyme structure and function.
Materials and Methods
Species Choice and Specimen Collection
The temperate Blacksmith (Chromis punctipinnis) is one of the most cold-adapted damselfish species. Its biogeographic range extends from Monterey Bay, California (122°W, 37°N) to Punta San Pablo, Baja California, Mexico (115°W, 27°N) covering the San Diego and southern portion of the Oregon biogeographic provinces (Miller and Lea 1972). Over this latitudinal range, Blacksmith experience water temperatures from 10°C to 22°C on an annual basis. The Dusky Chromis, C. caudalis, and the Yellow-axil Chromis, C. xanthochira, have similar distributions on outer reef slopes of islands throughout the tropical western Pacific (Allen 1991; Myers 1999). These species experience a much warmer and more stable thermal environment of approximately 29°C, with seasonal variation typically less than 1°C (Levitus and Boyer 1994). All specimens were hand collected by SCUBA diving. Chromis punctipinnis individuals were collected from Monterey, California and La Jolla, California (ambient seawater temperature 12°C at both collecting occasions; no differences in enzyme kinetics or cDNA sequence were observed in fish from either location). Chromis caudalis and C. xanthochira were collected on outer reef slopes near Koror, Republic of Palau, in the tropical western Pacific. White muscle tissue samples were taken immediately after killing the fish, frozen in liquid nitrogen, and stored at –70°C.
A4-LDH Purification and Kinetic Characterization
A4-Lactate dehydrogenase was purified from white muscle by oxamate affinity chromatography as described previously (Yancey and Somero 1978). All purifications showed only a single band on both silver-stained SDS-PAGE (corresponding to the LDH monomer) and activity-stained native PAGE (corresponding to the A4 tetramer) gels.
All kinetic assays were performed spectrophotometrically in 80 mM imidazole·HCl buffer, pH 7.0 at 20°C, by measuring the conversion of NADH to NAD+ at 340 nm (Yancey and Somero 1978). Measurements were made on the purified and dialyzed A4-LDH within 24 h of purification using a Shimadzu BioSpec 1601 UV/Vis spectrophotometer with temperature control (±0.1°C) and UVPC Kinetics software (Shimadzu, Kyoto, Japan). Kinetic studies were conducted from 5.0°C to 35.0°C at 5-degree intervals. Assays were performed at eight pyruvate concentrations covering the range of one-third to three times the predicted Km, with three replicates at each concentration. Enzyme concentration was calculated spectrophotometrically using the Pierce Coomassie protein assay reagent. The Michaelis-Menten constant, Km, and maximal velocity, Vmax, were calculated by the weighted linear regression method (Wilkinson 1961). The catalytic constant, kcat, was calculated from Vmax and the enzyme concentration.
Nucleic Acid and Deduced Amino Acid Sequences
Total RNA was extracted from white muscle tissue using Trizol (Invitrogen, San Diego, Calif.). From this, cDNA was reverse transcribed using an oligo-dT primer. Single-stranded cDNA was used as template for PCR using two damselfish-specific primers in the 5' and 3' UTR regions (PomLDH-5'S 5'-GACACCCACACGGCAGCAGCACTTCT-3' and PomLDH-3'A 5'-ACGGGAGGTATGGATTGAGGGAGTC-3'; PCR conditions: 28 cycles of 94°C for 20 s, 55°C for 30 s, and 72°C for 90 s). PCR products were gel purified and cloned into the pGEM T-Easy vector (Promega, Madison, Wis.). The complete ldh-a inserts were sequenced in both directions from two independent clones for each of four individuals from each species, using an ABI 373 XL automated sequencer. Sequence analysis was performed with the DNAStar suite of software (DNAStar, Inc., Madison, Wis.), and the presumptive amino acid sequences were inferred from the ldh-a cDNA sequences.
To corroborate the phylogenetic relatedness of these congeners, a 526-bp fragment of the mitochondrial cytochrome b gene was PCR amplified and sequenced using primers designed from GenBank sequences AF119392 and AF119395 for a congener and closely related confamilial, respectively (Primers: Chromis-CBL 5'-CAACCCCCGCCAACATTTCAG-3' and Chromis-CBH 5'-GAGCTCAGGCCAGTTGGGTTATTG-3'; PCR conditions: 35 cycles of 94°C for 20 s, 48°C for 30 s, and 72°C for 90 s). The phylogenetic analysis of the aligned sequences was performed using the PAUP* computer program under various distance, parsimony, and maximum-likelihood models (Swofford 2002).
Site-Directed Mutagenesis and in Vitro Expression
To assess the importance of specific amino acid substitutions between tropical and temperate LDH-A sequences, we utilized site-directed mutagenesis of the cloned ldh-a gene coupled with the in vitro overexpression of the protein. To accomplish this, the ldh-a gene from the temperate C. punctipinnis was taken as the starting wild-type sequence and placed into an expression vector (pAlter-Ex1 [Promega]) under control of the tac promoter. A Histidine Affinity Tag (BD Biosciences Clontech, Palo Alto, Calif.) was incorporated onto the 5' end of the gene, producing a histidine-rich, 31–amino acid epitope on the N-terminal end of the LDH-A monomer. An enterokinase cleavage recognition site was incorporated with this immediately preceding the serine at position 1 in the native LDH-A sequence. This plasmid construct, representing the recombinant wild-type ldh-a gene (rLDH-wt), was then transformed into E. coli JM109 cells.
After growing the transformed cells to log-phase growth, rLDH expression was induced with 1 mM isopropyl ?-D-thiogalactopyranoside (IPTG) for 4 h. The cells were then lysed and centrifuged to collect the expressed protein. This resulting protein mixture was further purified by first applying the supernatant to an immobilized metal affinity chromatography column (IMAC [Clontech]), which retained only highly charged proteins, including the rLDH with the histidine tag. The rLDH was then further purified using oxamate affinity chromatography, resulting in pure rLDH as measured by a single band of predicted molecular weight on a silver-stained SDS-PAGE gel. The purified protein was then cleaved with enterokinase (Roche Applied Sciences, Indianapolis, Ind. [1:20 w/w, 24 h at 37°C]). The cleaved protein was once again purified by oxamate affinity chromatography. The resulting recombinant protein showed a single band of the same molecular weight as the C. punctipinnis LDH-A on a silver-stained SDS-PAGE gel.
After confirming that the recombinant protein had the same Michaelis-Menten kinetics as the C. punctipinnis A4-LDH, specific amino acid substitutions were incorporated by site-directed mutagenesis utilizing the Altered Sites II system (Promega). Each mutant was purified in the same manner as the recombinant wild-type LDH. Finally, the Michaelis-Menten kinetic constant, Km, was measured for each recombinant protein following the procedure listed above at an intermediate temperature of 20°C.
Results
Temperature Adaptation of Kinetic Properties
The values of Kmpyr and kcat for the three orthologs reflect the pattern found in earlier studies of A4-LDHs of vertebrates: a cold-adapted ortholog has significantly higher values for Km and kcat than warm-adapted orthologs, across the range of measurement temperatures, (fig. 1). However, at their respective physiological temperatures (10°C to 22°C for C. punctipinnis and 29°C for C. caudalis and C. xanthochira), the orthologs have similar kinetic properties. Kmpyr is essentially the same for orthologs from all species (fig. 1, upper panel) and kcat exhibits a high degree of compensation to temperature (fig. 1, lower panel).
FIG. 1. A4-lactate dehydrogenase enzyme kinetic constants for one temperate and two tropical species of Chromis across temperatures spanning the physiological range of all species. (A) The apparent Michaelis-Menten constant, Km, of pyruvate (±SE) for the temperate Chromis punctipinnis (open circle) as compared with two tropical species, C. caudalis (shaded square) and C. xanthochira (shaded triangle). (B) The catalytic constant, kcat, for the same species and temperatures. Arrows indicate the environmental temperature range experienced by temperate (open arrows) and tropical (solid arrows) Chromis species
Sequence Differences and Their Location in the Three-Dimensional Structure
The deduced amino acid sequences reveal 13 polymorphic sites among the three species (fig. 2). Sequencing of multiple individuals of a species revealed no amino acid polymorphisms. Given the observed differences in the kinetic properties of temperate-adapted and tropical-adapted orthologs, we reasoned that those sites that have amino acid residues that differ between orthologs from different environments, but are the same in orthologs from the same environment, might be the most likely candidates for causing the observed kinetic differences for orthologs from different environments. Only four such polymorphic sites matched these criteria (fig. 2).
FIG. 2. Alignment of the presumptive amino acid sequence of lactate dehydrogenase-A from three species of Chromis. The complete sequence of the temperate species, C. punctipinnis, is given, and identity to this sequence in the two tropical species, C. caudalis and C. xanthochira, is indicated by (.). Numbering reflects the removal of the amino-terminal methionine residue. Secondary structure motifs, as deduced from the crystal structure of the dogfish A4-LDH (Abad-Zapatero et al. 1987), are included as a guide. Amino acid substitutions from the temperate ortholog that are shared between both tropical orthologs are highlighted in bold
If these four sites are in fact important in the adaptation of A4-LDH to temperature, then the type of substitution, as well as its location in the three-dimensional structure, should be of great importance. The locations of the four sites in the native tetramer were determined using a crystal structure model of C. caudalis LDH-A that was derived from the known crystal structures of the dogfish, pig, and human LDH-A (PDB codes 1LDM [Abad-Zapatero et al. 1987], 9LDT [Dunn et al. 1991], and 1I10 [Read et al. 2001], respectively). The four sites were at the base of the 1G-2G helix, the base of the H helix, and within the ?J-1G loop. Indeed, this loop has the highest thermal vibrational B factor in the crystal structure, partially representing the disordered nature of this loop (Abad-Zapatero et al. 1987) but also because of its high exposure to solvent even in the tetrameric form (Read et al. 2001). These two helices and loop, along with the catalytic loop itself, are the regions of the enzyme that undergo the largest movements during the conformational changes that accompany substrate and cofactor binding (shaded regions of fig. 3 [Dunn et al. 1991; Gerstein and Chothia 1991]).
FIG. 3. Three-dimensional model of a single LDH-A monomer from the tropical damselfish, Chromis caudalis. The major moving portions of the molecule are darkly shaded and labeled. The position of the ThrAla (temperatetropical) mutation at position 219 is also indicated. This model was produced by the SWISS-MODEL homology modeling program using the crystal structure of LDH-A holoenzyme from dogfish, pig, and human (PDB codes 1LDM, 9LDT, and 1I10, respectively), and visualized with the Swiss-PdbViewer program (Guex and Peitsch 1997)
Site-Directed Mutagensis: Targets and Effects
To test the functional importance of different amino acid substitutions at the four sites hypothesized to be important in adaptation to temperature, we utilized site-directed mutagenesis and in vitro expression. Although the Km of the recombinant wild-type enzyme, rLDH, was not significantly different from the Km of the temperate ortholog, the different mutant substitutions had varying effects on Km (fig. 4). At one extreme, the GluAsp substitution at position 310 at the base of the H helix actually resulted in a slight increase in Km. The AspGlu substitution at position 220 in the ?J-1G loop appears to have no effect, and the LysAsn substitution at position 225 immediately preceding the 1G-2G helix appears to have only a moderate effect on lowering the Km from a temperate to a tropical-like value. Finally, the most radical substitution, ThrAla at position 219 in the ?J-1G loop, appears sufficient, by itself, to lower the Km from a value associated with a temperate ortholog to a value comparable with a tropical ortholog.
FIG. 4. Apparent Michaelis-Menten constant of pyruvate (±SE) at 20°C for different native and recombinant ldh-a genotypes. The purified recombinant A4-LDH (rLDH-wt) has the exact amino acid sequence as the temperate ortholog. Each of the other recombinant genotypes has the same amino acid sequence as the recombinant wild-type, with the exception of the amino acid substitution indicated. The recombinant wild-type, E310D, and D220E are not significantly different from the ortholog of the temperate C. punctipinnis, but do differ from the tropical C. caudalis or C. xanthochira orthologs. T219A is not significantly different from the tropical orthologs, but is different from the temperate ortholog. The intermediate K225N is not significantly different from orthologs of any species (95% confidence intervals; Dunn-idák adjustment for multiple means)
Discussion
These results corroborate a growing base of research demonstrating a conservation of biochemical properties among enzyme orthologs when examined under their respective physiological conditions (fig. 1). Foremost among these is the observation that, at any one temperature and for both the Michaelis-Menten constant, Km, and the catalytic constant, kcat, the tropical adapted species show more efficient substrate binding and a lower catalytic efficiency. Conversely, when all orthologs are compared at their respective physiological temperatures, both the substrate binding and the catalytic efficiency are roughly comparable.
These opposite trends in substrate binding and catalytic efficiency are likely the result of the balance of changes in enthalpy and entropy associated with a posited stability/flexibility compromise. The reaction catalyzed by orthologs of A4-LDH from species adapted to different temperatures exhibits a regular variation in thermodynamic activation parameters (Low, Bada, and Somero 1973), a trend observed for other enzymes as well (Lonhienne, Gerday, and Feller 2000). The reactions of cold-adapted species have lower enthalpies of activation (H) and more negative entropies of activation (S) than do the reactions catalyzed by warm-adapted orthologs. Thus, for cold-adapted orthologs relative to warm-adapted orthologs, the enthalpy barrier to the reaction is less, although the enzyme becomes relatively more ordered during the rate-limiting step in the reaction. The free energy of activation (G = H – T S), which determines kcat, is lower for cold-adapted orthologs because the enthalpy term has the stronger contribution to G.
Now that the role of catalytic conformational changes in the LDH reaction is understood (Dunn et al. 1991; Gerstein and Chothia 1991), it is possible to interpret this variation in activation parameters. The rate-limiting conformational changes that accompany binding of substrate and cofactor, leading to the establishment of the catalytic vacuole (Dunn et al. 1991), are endothermic and occur with a decrease in the entropy of the protein. More warm-adapted enzymes must increase their rigidity to maintain an adequate proportion of binding-competent microstates of the enzyme to allow effective substrate binding at the physiologically relevant substrate concentrations and temperature. The increased rigidity creates a higher enthalpy barrier for conformational changes and results in higher activation enthalpy (H). This increased rigidity also leads to a lower (less negative) activation entropy (S), as there is likely a narrower distribution of conformational microstates associated with a more rigid molecule such that the formation of the catalytic vacuole involves a smaller increase in order relative to the more flexible cold-adapted ortholog.
To examine the mechanisms that might explain the observed differences in kinetics between the orthologs of the tropical and temperate damselfishes, we first examined the sites where sequence differences exist that could potentially affect the energy changes that occur during catalysis. After this analysis, we exploited site-directed mutagenesis to alter residues that we hypothesized might be important in adaptation to temperature. In designing our experiments, we reasoned that the amino acid substitutions that would be most likely responsible for the differences in kinetics between tropical and temperate orthologs would be those residues that are the same in the tropical orthologs but different in the temperate ortholog. In this way, we were able to reduce the 13 polymorphic sites to just four that were the most likely candidates for providing functionally significant differences between orthologs (fig. 2).
Structural analysis yielded a number of correlations with previous results from studies of A4-LDH evolution. Most significantly, the location of the candidate substitutions for adaptive variation on the three-dimensional crystal structure model is restricted to a limited number of sites in the LDH-A monomer that previously have been identified as regions that undergo large movements during function and manifest substantial sequence variation among species (fig. 3 [Dunn et al. 1991; Gerstein and Chothia 1991]). It is the movement of these portions of the molecule, notably helices 1G-2G, that are the rate-limiting steps in the reaction. Other components of the enzyme move during catalysis as well; for example, the catalytic loop that comprises residues 95 to 112 in our residue numbering system. However, because the catalytic loop contains residues that interact with the substrate and cofactor, its sequence is nearly invariant across all known ldh-a sequences (see below). Thus, adaptive change that modulates the energetics of catalytic conformational changes may be restricted to only some of the "moving parts" of the enzyme. Consequently, the primary candidates for adaptive change are two helical regions, 1G-2G and H, plus the disordered loop region that acts as a hinge for the G-2G helix. The four amino acid polymorphisms implicated in differences between orthologs of tropical and temperate damselfishes were located in these regions. These are the same regions of the molecule implicated as "hot-spots" for cold-adaptation in Antarctic notothenioids (Fields and Somero 1998). Thus, in the context of adaptive mechanisms for adjusting kinetic properties, this represents an instance of enzymatic convergent evolution—separate evolutionary lineages evolving amino acid substitutions in the same locations of the molecule to effect the same sorts of kinetic changes associated with different adaptation temperatures.
One apparent exception to this paradigm is found in the comparison of A4-LDH orthologs from different temperate and cosmopolitan barracuda species (Holland, McFall-Ngai, and Somero 1997). Significant differences in Km and thermal stability were found among orthologs, but amino acid differences that were likely responsible were located in the N-terminal regions of the LDH-A molecule, away from the major moving portions of the LDH-A subunit. However, the substitutions they observed at position 8 and at positions 61 to 68 interact with the H helix and with residues adjacent to the 2G helix of other subunits in the quaternary structure of A4-LDH, respectively (Abad-Zapatero et al. 1987). Thus, the substitutions in the warm-adapted orthologs may serve to stabilize the quaternary structure and thereby attenuate the movements of the major moving portions of the molecule without altering the function of the catalytic loop. In particular, substitutions in the N-terminal helix could affect the ease with which the 1G-2G helices move, thereby affecting the energetics of the conformational change thought to be rate-limiting to catalysis (Dunn et al. 1991). Therefore, alterations in the energetics of catalytic conformational changes can arise from shifts in the intrinsic flexibility of portions of the individual LDH-A subunit or from changes in subunit-subunit interactions that alter the ease with which catalytic conformation changes can take place.
The question might arise as to whether the two different tropical species examined here in fact represent a sort of convergence in themselves in that they might have each evolved separately to the tropics via independent molecular mechanisms. However, our independent cytochrome b mitochondrial DNA phylogeny of the species examined clearly demonstrates that the two tropical Chromis species are much more closely related to each other (1.0% maximum-likelihood corrected sequence divergence) than either is to the temperate Chromis species (18.3% average maximum-likelihood corrected sequence divergence). Accepting all of the caveats of a molecular clock and assuming a standard rate calibration (e.g. Johns and Avise 1998), this tentatively suggests that temperate C. punctipinnis has diverged from the tropical C. caudalis and C. xanthochira over 9 MYA, whereas the two tropical species have diverged less than 0.5 MYA. This is consistent with a single adaptation to the tropical environment followed by a small amount of divergence between tropical species, but the alternative possibility of a single adaptation event of the temperate Chromis to colder waters cannot be excluded. Regardless, the majority of Chromis species, as well as the majority of damselfishes overall, represent broadly distributed species occupying tropical and subtropical environments (Allen 1991). This distribution pattern combined with the lower tropical ocean temperatures in the geologic past (e.g., Lea, Pak, and Spero 2000) suggests that the ancestral Chromis of the species examined here likely had an intermediate environmental regime, and these species represent divergence to two different extremes.
To ascertain the functional significance for changes in enzyme kinetics brought about by each of the putatively important amino acid substitutions, we utilized site-directed mutagenesis and an in vitro expression system. The results of this analysis suggest several points about the importance of these sites: (1) not all of the substitutions change the Km in the predicted manner, (2) some substitutions only have marginal effects on the change in Km, and (3) in some cases (e.g., Thr219Ala), a single mutation can account for the full difference in Km between a temperate and a tropical ortholog (fig. 4). In the case of Glu310Asp, the shorter side chain associated with the aspartic acid residue in the tropical ortholog potentially reduces by one the number of hydrogen bonds available to stabilize the H helix. This likely results in a more flexible molecule that is less often in a binding-competent microstate and hence has a higher Km. This finding, along with the finding that Asp220Glu and Lys225Asn mutants only marginally reduce the Km, suggest that temperature adaptation for overall Km may reflect selection on the net global effects of all substitutions combined. The effects of the interaction of multiple substitutions on the overall change in the Km await further investigation. However, that fact that the single substitution of Thr219Ala was able to reduce the Km from a level comparable with the temperate Chromis punctipinnis to a level comparable with the tropical C. caudalis and C. xanthochira demonstrates that a single substitution in the highly mobile ?J-G loop region is sufficient to cause temperature-adaptive change in the enzyme.
In this case, a single transition mutation at position 658 in the ldh-a gene, from an adenine to a guanine nucleotide in the tropical ortholog, results in the replacement of the hydrophilic threonine amino acid with a hydrophobic alanine residue. The importance of this change in hydrophobicity is indicated by that fact that the substitution is located in a region of the molecule that not only is a key "hinge" for the moving 1G-2G helix (Fields and Somero 1998) but also is one of the most solvent-exposed loops of the tetrameric enzyme (Read et al. 2001). Thus, the increased hydrophobicity of this loop could potentially force it to remain closer to the overall molecule and result in a less flexible "hinge" for the movement of the 1G-2G helix. Interestingly, this same transition mutation is evident in the comparisons of Antarctic notothenioids to the temperate teleost consensus sequence (658 here; 655 in Fields and Somero's [1998] numbering system), with the adenine nucleotide encoding alanine in the warm-adapted ortholog, and the guanine nucleotide encoding threonine in the cold-adapted ortholog in both cases. Consequently, this may represent a striking example of convergent evolution at this single base pair of the ldh-a gene. However, the kinetic difference associated with this specific substitution in the notothenioids has yet to be determined.
Indeed, the significance of this surface loop, along with its putatively greater accessibility to solvent in the more cold-adapted ortholog, agrees well with other studies that have shown that psychrophilic enzymes tend to have very hydrophilic surfaces with improved solvent interactions leading to a less compact protein shell as compared with their mesophilic counterparts (for review see Feller and Gerday [1997]). Specifically, in the case of subtilisin, the psychrophilic ortholog contains four amino acid insertions in surface loops that additionally increase the solvent interactions and hence reduce the compactness and stabilization of the psychrophilic ortholog (Davail et al. 1994).
In an effort to elucidate how these observations correlated with the broader patterns of vertebrate ldh-a evolution, we examined the patterns of nucleotide and amino acid substitution of all ldh-a sequences available in GenBank (including the sequences presented here) representing 52 vertebrate sequences. The analysis of amino acid polymorphisms in LDH-A when combined with the three-dimensional structure of A4-LDH supports three points: (1) the active site and catalytic loop are almost entirely conserved across all LDH-A sequences, (2) the areas of highest polymorphism typically occur on the most solvent-exposed portions of the monomer, and (3) in the tetramer, there are two classes of hypervariable regions—those that are on the 1G-2G and H helices or their associated hinge regions and those that interact with these moving helices in the quaternary structure. These same trends are even more clearly apparent when the analyses are restricted to the 32 euteleost fish ldh-a sequences available in GenBank.
These conclusions for A4-LDH suggest that, for other enzyme systems, adaptation of orthologs to differing physiological temperature is perhaps more likely to occur in regions of the molecule that will have the greatest influence on changing the flexibility of the moving portions of the molecule but do not affect the movement, shape, or specificity of the catalytic vacuole. Those mutations that occur with relatively greater frequency but nonetheless have significant effects on temperature-dependent enzyme kinetics, such as the adenine to guanine transition at position 658 in ldh-a, may show repeated instances of convergent evolution. The ubiquity of such hot spots of adaptation in other proteins remains to be empirically ascertained, but this burgeoning era of proteomics, particularly as it spreads to nonmodel species, holds with it the promise of at least the potential of gleaning common strategies or constraints to the mechanism of physiological adaptation of enzymes. This not only would prove invaluable for our understanding of the evolutionary history of extant enzyme orthologs but also would provide insights into their capacity for future adaptive change for both natural invasions of new species into new ecological niches and also in response to global environmental change.
Supplementary Material
The sequences reported in this paper have been deposited in the GenBank database under accession numbers AY289557 to AY289562.
Acknowledgements
We would like to thank Peter A. Fields, Andrew Y. Gracey, and Anthony W. DeTomaso for their assistance and expertise with molecular techniques. Lori and Patrick Colin, Larry Sharon, Ron McConnaghey, and Ed Kisfaludy were a tremendous assistance in the collection of specimens. This work was funded by NSF grant IBN-0133184 to G.N.S.
Literature Cited
Abad-Zapatero, C., J. P. Griffith, J. L. Sussman, and M. G. Rossman. 1987. Refined crystal structure of dogfish M4 apo-lactate dehydrogenase. J. Mol. Biol. 198:445-467.
Allen, G. R. 1991. Damselfishes of the world. Mergus, Melle, Germany.
Davail, S., G. Feller, E. Narinx, and C. Gerday. 1994. Cold adaptation of proteins: purification, characterization, and sequence of the heat-labile subtilisin from the Antarctic psychrophile bacillus TA41. J. Biol. Chem. 269:17448-17453.
Dunn, C. R., H. M. Wilks, D. J. Halsall, T. Atkinson, A. R. Clarke, H. Muirhead, and J. J. Holbrook. 1991. Design and synthesis of new enzymes based on the lactate dehydrogenase framework. Phil. Trans. R. Soc. Lond. B Biol. Sci. 332:177-184.
Feller, G., and C. Gerday. 1997. Psychrophilic enzymes: molecular basis of cold adaptation. Cell. Mol. Life Sci. 53:830-841.
Fields, P. A. 2001. Review: protein function at thermal extremes: balancing stability and flexibility. Comp. Biochem. Physiol. (part A) Mol. Integr. Physiol. 129:417-431.
Fields, P. A., and G. N. Somero. 1998. Hot spots in cold adaptation: localized increases in conformational flexibility in lactate dehydrogenase A4 orthologs of Antarctic notothenioid fishes. Proc. Natl. Acad. Sci. USA 95:11476-11481.
Gerstein, M., and C. Chothia. 1991. Analysis of protein loop closure two types of hinges produce one motion in lactate dehydrogenase. J. Mol. Biol. 220:133-149.
Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modelling. Electrophoresis 18:2714-2723.
Hochachka, P. W., and G. N. Somero. 2002. Biochemical adaptation: mechanism and process in physiological evolution. Oxford University Press, Oxford, UK.
Holland, L. Z., M. McFall-Ngai, and G. N. Somero. 1997. Evolution of lactate dehydrogenase-A homologs of barracuda fishes (genus Sphyraena) from different thermal environments: differences in kinetic properties and thermal stability are due to amino acid substitutions outside the active site. Biochemistry 36:3207-3215.
Johns, G. C., and J. C. Avise. 1998. A comparative summary of genetic distances in the vertebrates from the mitochondrial cytochrome b gene. Mol. Biol. Evol. 15:1481-1490.
Lea, D. W., D. K. Pak, and H. J. Spero. 2000. Climate impact of late Quaternary equatorial Pacific sea surface temperature variations. Science 289:1719-1724.
Levitus, S., and T. P. Boyer. 1994. World ocean atlas 1994, Vol. 4. Temperature. Department of Commerce, Washington, DC.
Lonhienne, T., C. Gerday, and G. Feller. 2000. Psychrophilic enzymes: revisiting the thermodynamic parameters of activation may explain local flexibility. Biochim. Biophys. Acta 1543:1-10.
Low, P. S., J. L. Bada, and G. N. Somero. 1973. Temperature adaptation of enzymes: roles of the free energy, the enthalpy, and the entropy of activation. Proc. Natl. Acad. Sci. USA 70:430-432.
Miller, D. J., and R. N. Lea. 1972. Guide to the coastal marine fishes of California, Fish Bulletin No. 157. California Department of Fish and Game, Sacramento.
Myers, R. F. 1999. Micronesian reef fishes. 3rd edition. Coral Graphics, Guam.
Read, J. A., V. J. Winter, C. M. Eszes, R. B. Sessions, and R. L. Brady. 2001. Structural basis for altered activity of M- and H-isozyme forms of human lactate dehydrogenase. Proteins 43:175-185.
Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4b10. Sinauer Associates, Sunderland, Mass.
Wilkinson, G. N. 1961. Statistical estimations in enzyme kinetics. Biochem. J. 80:324-332.
Yancey, P. H., and G. N. Somero. 1978. Temperature dependence of intracellular pH: its role in the conservation of pyruvate apparent Km values of vertebrate lactate dehydrogenases. J. Comp. Physiol. 125:129-134.(Glenn C. Johns and George)