ABSTRACT
We compared the deduced amino acid sequences, heat stabilities and thermal sensitivities of a kinetic property, the apparent Michaelis–Menten constant (Km) of pyruvate, of A4-lactate dehydrogenase (A4-LDH) in four species of goby fishes (Family Gobiidae), adapted to different temperatures, to examine how changes in primary structure influence the adaptation of enzymes. The effect of temperature on Km of pyruvate reflected each species’ environmental temperature. For the most eurythermal species, Gillichthys seta, which is endemic to shallow intertidal regions of the upper Gulf of California and encounters temperatures between approximately 9 and 40 °C, Km of pyruvate was minimally affected by temperature, compared with the A4-LDH orthologues from a less eurythermal congener, G. mirabilis (9–30 °C), a cold temperate goby, Coryphopterus nicholsi (10–18 °C) and a tropical species, C. personatus (25–32 °C). Heat denaturation profiles failed to correlate with habitat temperature; G. mirabilis A4-LDH was most thermally stable, followed by the orthologues of C. nicholsi and G. seta. Complementary DNAs (cDNAs) encoding LDH-As of G. seta, Gulf of California and Pacific coast populations of G. mirabilis and C. nicholsi were isolated and sequenced, and the corresponding amino acid sequences deduced. The nucleotide sequences of LDH-A of the two populations of G. mirabilis were identical. Five nucleotide differences in the coding region and one amino acid substitution (at position 78) distinguished LDH-As of G. mirabilis and C. nicholsi. The substitution of a glycyl residue (C. nicholsi) for an alanyl residue (G. mirabilis) may account for the difference in thermal stability between these two orthologues. Comparisons of the LDH-A cDNAs of G. mirabilis and G. seta revealed four differences in nucleotide sequence in the coding region, but all nucleotide substitutions were synonymous. The identical deduced primary structures of the two enzymes suggested the possibility of different protein conformational variants (‘conformers’) in the two species. This hypothesis is supported by electrospray ionization mass spectrometry, which indicates that the masses of the A4-LDH orthologues of the two species are the same within the resolution of the technique. To explore the possibility that the two enzymes were different conformers of the same primary structure, we treated purified G. seta and G. mirabilis A4-LDHs with 3.0 mol l−1 urea or 6 mol l−1 guanidine–HCl and, after removing the denaturant, compared their kinetic properties and heat stabilities. Neither treatment had an effect on the A4-LDH of G. mirabilis, but both converted the Kmversus temperature profile of the G. seta enzyme to that of the G. mirabilis A4-LDH. The thermal stability of neither enzyme was affected. We propose, as has been suggested in several previous studies of A4-LDH, that this enzyme can fold into a number of conformers with different stabilities and functional properties. The A4-LDH of G. seta furnishes evidence that such conformers may provide an important mechanism for adaptation of proteins to temperature.
Introduction
Temperature is a major factor limiting the distribution of ectothermic marine species (Hubbs, 1948; Fields et al. 1993), in part because it has such powerful effects on the structures and activities of enzymes and other proteins (Jaenicke, 1991; Somero, 1995). The role of temperature in affecting the evolution of proteins is indicated by studies that have compared interspecific homologues of enzymes (orthologous homologues, or orthologues) from species adapted to different temperatures. Consistent interspecific differences in kinetic properties, such as Michaelis–Menten constants (Km) and catalytic rate constants (kcat), which appear to be temperature-adaptive, have been identified (Johnston and Walesby, 1977; Yancey and Somero, 1978; Graves and Somero, 1982; Yancey and Siebenaller, 1987; Dahlhoff and Somero, 1993; Somero, 1995; Holland et al. 1997). Thermal stability of structure also typically reflects adaptation temperature (Jaenicke, 1991; McFall-Ngai and Horwitz, 1990; Dahlhoff and Somero, 1993), although exceptions to this pattern have been found in closely related organisms (Place and Powers, 1984; Holland et al. 1997).
Elucidation of the changes in protein primary structure (amino acid sequence) that underlie adaptive differences in kinetic properties and thermal stabilities can provide important new insights in several contexts, including molecular biology, environmental physiology and evolutionary biology. These studies should reveal structure–function relationships that link specific amino acid substitutions to adaptive alterations in traits such as Km and structural stability. Studies of orthologues from differently adapted species complement investigations that use laboratory-generated protein variants to deduce the effects of specific changes in primary structure on stability and function (e.g. Matthews et al. 1987). By revealing the amounts and types of sequence change that are needed to adapt a protein to a new thermal regime, insights into the mechanisms and rates of evolutionary change may also be possible.
To facilitate these types of comparative studies, both the species and enzyme to be studied must be carefully selected. It is essential to choose study species that are adapted to different temperature conditions, yet are close enough, phylogenetically, to ensure that the number of differences in amino acid sequence among orthologues is small enough to allow relatively straightforward interpretations of structure–function relationships. The ideal study system would appear to be a set of congeneric or confamilial species that have evolved in different thermal conditions. The enzyme to be studied should be one for which there is substantial information on primary and three-dimensional structure, to allow amino acid substitutions to be related to alterations in function and thermal stability.
With these considerations in mind, we initiated a study of A4-lactate dehydrogenase (A4-LDH; EC 1.1.1.27, NAD+:lactate oxidoreductase) orthologues from fishes of the Family Gobiidae adapted to widely different thermal conditions. Gillichthys mirabilis is a eurythermal estuarine and mudflat species that occurs along the Pacific coast of North America from Tomales Bay, California, into northern Baja California and also has what is thought to be a disjunct population within the Gulf of California (Miller and Lea, 1972; Thomson et al. 1979). G. mirabilis encounters temperatures from approximately 9 to 30 °C and appears to avoid high temperatures by burrowing into the sediment (Barlow, 1961). Its congener, G. seta, is found exclusively in the shallow rocky intertidal zone of the northern Gulf of California (Thomson et al. 1979), where it encounters temperatures from approximately 9 to 40 °C and possibly higher. Gillichthys seta may be derived from G. mirabilis through progenesis (Barlow, 1961), and morphological similarities indicate that it is closely related to G. mirabilis.
The second species pair examined is from the genus Coryphopterus, which was divided by the development of the Isthmus of Panama between 3 and 4 million years ago (Coates et al. 1993; Knowlton et al. 1993). Coryphopterus nicholsi occurs from British Columbia to Baja California (Miller and Lea, 1972) and is found deeper at the southern extent of its range. Off the coast of southern California, C. nicholsi generally occurs below the thermocline at 20–30 m, where temperature varies annually from 10 to 18 °C (R. McConnaughey, Scripps Institution of Oceanography, personal communication). Coryphopterus personatus occurs in Bermuda, southern Florida and throughout the Caribbean (Zaneveld, 1983) to depths of 30 m. The sampled population from the Florida Keys encounters an annual temperature range of approximately 25–32 °C (Dr M. E. Clarke, University of Miami, personal communication). These four species provided an excellent study system for examining the evolution of A4-LDH in different thermal conditions over varying periods of evolutionary time.
To compare the A4-LDH orthologues of each species and of the Pacific coast and Gulf of California populations of G. mirabilis, we measured the effects of temperature on the Km of pyruvate, a highly temperature-sensitive kinetic trait (Yancey and Somero, 1978). For all but C. personatus, we also had sufficient material to conduct thermal stability measurements and complementary DNA (cDNA) sequencing. We report that some of the differences in thermal sensitivity of the LDH-A orthologues can be explained by observed differences in sequence, but that differences in kinetics and thermal stabilities between the orthologues of G. mirabilis and G. seta may be due to alternative conformational variants (conformers) of an identical primary structure.
Materials and methods
Collection and care of specimens
Gillichthys seta
Ginsburg were trapped by hand in the intertidal habitat approximately 3 km north of San Felipe, Baja California, Mexico. Gulf of California G. mirabilis Cooper were collected from a slough approximately 3 km south of San Felipe, and Pacific G. mirabilis were collected from a small slough approximately 10 km north of San Diego, CA. G. mirabilis were caught in minnow traps baited with chopped mackerel. All fish were transported to the laboratory at the Scripps Institution of Oceanography (SIO) in oxygenated sea water at ambient temperature. Fish were held at ambient temperature (approximately 15 °C) for over 3 months prior to use, and were fed fish pellets and minced mackerel twice per week.
Coryphopterus nicholsi
Bean were collected using SCUBA in 20–30 m of water off the coast of La Jolla, CA. There was heavy mortality among the fish collected, so specimens were frozen at −80 °C immediately upon their arrival at SIO.Coryphopterus personatus Jordan and Marsh were collected from the Florida Keys by Dr M. Elizabeth Clarke, of the University of Miami, and her associates. Fish were frozen at −80 °C upon capture, and were shipped to San Diego on solid carbon dioxide. Upon arrival, they were placed at −80 °C until used.
Determination of kinetics of A4-LDH
Fish were decapitated after anesthetization in ice-cold sea water. Approximately 2 g of white muscle was dissected away from the body, with care being taken not to contaminate the sample with red muscle, skin or viscera. If 2 g could not be obtained from one fish, tissue from two or more individuals was pooled.
The A4-LDH of each species was purified using agarose–oxamate column chromatography and dialyzed against 50 mmol l−1 potassium phosphate buffer, pH 6.8, as described in Yancey and Somero (1978). Complete separation of LDH from other proteins was shown by the presence of a single protein band on silver-stained sodium dodecyl sulfate polyacrylamide gels. Non-denaturing polyacrylamide gels stained for LDH activity also revealed a single band, indicating that the only isoenzyme present was A4-LDH. Apparent Michaelis–Menten constants for pyruvate of A4-LDH were determined using a Perkin Elmer Lambda 3B spectrophotometer equipped with a temperature-controlled cell attached to a Lauda RM6 recirculating water bath. The temperature of the cuvette sample was measured with a Digisense model 8528-20 thermocouple thermometer and maintained to within ±0.2 °C.
To ensure that pH levels in the experimental system corresponded to intracellular values over a wide range of temperatures, an imidazole chloride buffer (80 mmol l−1 imidazole chloride, 150 μmol l−1 NADH; pH 7.0 at 20 °C) was used during enzyme activity determinations (Yancey and Somero, 1978; Somero, 1981; Hochachka and Somero, 1984). For each Km determination, eight concentrations of pyruvate were used: 0.083, 0.100, 0.125, 0.143, 0.167, 0.200, 0.250 and 0.333 mmol l−1 for Km determinations at 10, 15, 20 and 25 °C, and 0.100, 0.125, 0.167, 0.200, 0.250, 0.333, 0.400 and 0.500 mmol l−1 for determinations at 30, 35 and 40 °C.
Each assay was initiated by addition of 10 μl of A4-LDH to the imidazole/NADH/pyruvate solution, and the rate of decrease in absorbance was measured at 340 nm. Velocity determinations for the eight pyruvate concentrations at each experimental temperature were entered into a computer program (Wilman Km-fitting; Brooks and Suelter, 1986) to calculate Km using weighted linear regression as outlined by Wilkinson (1961).
Determination of thermal stabilities of A4-LDH
The thermal stabilities of A4-LDH from G. mirabilis, G. seta and C. nicholsi were determined at 45, 47.5, 50, 52.5 and 55 °C using enzyme purified as described above. 400 μl of enzyme in 50 mmol l−1 potassium phosphate buffer, pH 6.8, was added to 1.60 ml of thermal denaturation solution [1.4 g of KCl, 125 mg of bovine serum albumin (BSA), 100 ml of H2O; Place and Powers, 1984] in a 13 mm×100 mm glass tube, and the tube was placed in a temperature-controlled Lauda RM6 recirculating water bath (±0.1 °C). Three 75 μl samples were removed at 0 (before incubation), 5, 10, 20, 40 and 60 min. The activity of the samples was measured as in the kinetics experiment, using 150 μmol l−1 NADH and 2 mmol l−1 pyruvate. No decrease in activity was found in the 0 min samples when held on ice for over 1 h. The three values for each time point were averaged and divided by the mean value for 0 min to allow comparison of residual activities.
Sequencing of LDH-A cDNA
Total RNA was purified from white muscle tissue of G. mirabilis, G. seta and C. nicholsi using a guanidinium thiocyanate–phenol extraction method adapted from the method of Chomczynski and Sacchi (1987). To separate messenger RNA (mRNA) from total RNA, microscopic (2.8 mm) magnetized beads to which 25-mers of poly-deoxythymidylate had been covalently attached (Dynabeads; Dynal, Inc.) were used to bind reversibly the poly-adenylate tails of the mRNA molecules (Hornes and Korsnes, 1990). Complementary DNA was synthesized from the purified mRNA using the RACE procedure (rapid amplification of cDNA ends; Frohman et al. 1988). Amplification of LDH-A cDNA from the synthesized cDNA thus obtained was accomplished using the polymerase chain reaction (PCR; Mullis and Faloona, 1987). LDH-A-specific primers were designed on the basis of the LDH-A sequence of barracuda (genus Sphyraena; Holland et al. 1997). These primers were adequate to amplify the coding region of ldh-a, but were not able to provide amplification of the 5′-noncoding region. This region was amplified by single-strand ligation of cDNA (Tessier et al. 1986; Edwards et al. 1991) using reagents provided with the 5′-amplifinder race kit from Clontech Laboratories, Inc.
The complementary DNA sequence was obtained through the dye-labeled ddNTP terminator protocol (Smith et al. 1987a,b) using a prism ready reaction dyedeoxy terminator cycle sequencing kit from Applied Biosystems and the Applied Biosystems model 373A automated sequencer.
Treatment of A4-LDH with urea and guanidine–HCl
The effects of incubation in 3.0 mol l−1 urea on A4-LDH kinetics and thermal stabilities were determined for G. mirabilis, G. seta and C. nicholsi. In addition, the effects of denaturation in 6 mol l−1 guanidine–HCl (Gdn–HCl) on A4-LDH thermal stability were examined for G. mirabilis and G. seta. For urea treatment, 1.13 ml of A4-LDH in 50 mmol l−1 potassium phosphate buffer, pH 6.8, was mixed with 0.20 ml of 10 mg ml−1 BSA, to which 0.667 ml of 9 mol l−1 urea was added, for a final urea concentration of 3 mol l−1. The mixture was allowed to incubate for 15 min at room temperature, and then urea was removed by centrifugal filtration in a Centricon 30 concentrator (Amicon). The BSA served to prevent loss of enzyme activity during the filtration step due to adherence of enzyme to the Centricon filter. After concentration to approximately 100 μl, the retentate was diluted with 1 ml of 50 mmol l−1 potassium phosphate buffer (pH 6.8 at 20 °C) and re-filtered; this process was repeated three times, and the retentate was returned to a final volume of 1.13 ml. Kinetics and thermal stabilities were then determined for the urea-treated A4-LDH as described above.
To induce denaturation by Gdn–HCl, 0.833 ml of A4-LDH in 50 mmol l−1 potassium phosphate buffer, pH 6.8, was mixed with 1.67 ml of 9 mol l−1 Gdn–HCl and 10 mmol l−1 2-mercaptoethanol (2-ME), to give a final Gdn–HCl concentration of 6 mol l−1. The sample was held at room temperature for 1 h. It was then diluted to a final volume of 150 ml (i.e. 60-fold dilution) with 50 mmol l−1 Tris–HCl, 5 mmol l−1 2-ME and 350 μmol l−1 NADH, pH 7.5. This was held at room temperature for 48 h to permit renaturation. The sample was then concentrated using an Amicon model 52 ultrafiltration cell with a BSA-pretreated (1 mg ml−1) YM30 filter. After concentration to 5 ml, the sample was washed twice with 50 mmol l−1 potassium phosphate buffer, pH 6.8, concentrated to 5 ml, and further concentrated with a BSA-pretreated Centricon-30 to approximately 1 ml. Thermal stabilities were then determined for the A4-LDH samples as described above.
Mass determination by electrospray ionization mass spectrometry
Molecular masses were determined for the LDH-A monomers of G. mirabilis and G. seta in order to determine whether the enzyme orthologues contain different post-translational modifications. To prepare enzyme samples for electrospray ionization (ESI) mass spectrometry, approximately 72 μg (2 nmol) of each was dialyzed against 20 mmol l−1 of the volatile buffer triethanolamine HCl. The samples were then concentrated under vacuum to a final volume of 25 μl.
These samples were sent to Peptidogenic Research, Inc. (Livermore, CA), where mass determinations were performed on a Perkin Elmer API-Plus mass spectrometer.
Results
Kinetics of A4-LDH
The results of the kinetics studies for the Gillichthys congeners are shown in Fig. 1A, and the results for the Coryphopterus congeners are given in Fig. 1B. As can be seen in the inset of Fig. 1A, there was no significant difference in the A4-LDH kinetics of the two populations of G. mirabilis within the temperature range examined (log-transform analysis of covariance, F=0.751, P>0.50); thus, kinetic data from these populations were pooled. When compared with the values for G. mirabilis, the affinity of G. seta A4-LDH for pyruvate remains significantly higher (i.e. the Km remains lower) at temperatures greater than 20 °C (see legend to Fig. 1). However, when the Km values are compared within the environmental temperature range of each species (solid lines, Fig. 1A), it is apparent that the orthologues of A4-LDH maintain similar pyruvate affinities. These differences in Km of pyruvate values are not due to the effects of acclimatization, because specimens were held for over 3 months at a common temperature (approximately 15 °C) prior to being killed. Further, we obtained identical Km values for enzymes purified from specimens collected and immediately frozen in the field at different seasons.
Fig. 1B shows that the A4-LDH of C. personatus has a greater affinity for pyruvate than does that of C. nicholsi at all temperatures tested except 40 °C. Although small sample sizes preclude determination of statistical significance at individual temperatures, a paired-sample t-test indicates that C. nicholsi A4-LDH has significantly higher Km values (t=6.534; P<0.01). Again, however, an examination of Km values within the normal thermal range of each species (solid lines, Fig. 1B) reveals similar pyruvate affinities.
Thermal stabilities of A4-LDH
Residual activities of A4-LDH from G. mirabilis, G. seta and C. nicholsi after incubation at 50 °C are shown in Fig. 2. G. mirabilis A4-LDH retained greater activity than the other two orthologues at this temperature, possessing 75.3 % of its original activity after 60 min, compared with 23.3 % for C. nicholsi and 8.4 % for G. seta. Thus, it appears that the mechanisms responsible for producing relatively temperature-insensitive kinetics in the G. seta enzyme, when compared with the G. mirabilis orthologue (Fig. 1A), do not lead to a higher resistance to heat denaturation. The absence of a linkage between thermal effects on kinetic properties and structural stability is also indicated by the comparison between the orthologues of G. mirabilis and C. nicholsi, in which identical effects of temperature on Km of pyruvate (Fig. 1A,B) are associated with different thermal stabilities.
cDNA sequences and deduced amino acid sequences of LDH-A
Nucleotide sequence differences of ldh-a among the species of goby studied are given in Fig. 3. There was no difference in nucleotide sequence between the two populations of G. mirabilis. Between G. seta and G. mirabilis, there were a total of four base differences in the coding region, three of which occur in the first 100 bases (positions 15, 30 and 69). The final substitution occurred at position 648. In addition, there was one base difference in the 31 bases sequenced in the 5′-noncoding region (not shown). The nucleotide sequence of C. nicholsi did not show much greater disparity in comparison with G. mirabilis than did the G. seta sequence. In the coding region there were five base differences, which are concentrated in the first few hundred base pairs (at positions 126, 195, 231, 233 and 234); there were also three substitutions in the 5′-noncoding region.
The differences in amino acid sequences of the LDH-A protein deduced from the nucleotide sequences described above are also given in Fig. 3. Because there are no base differences in the LDH-A cDNA sequence of the G. mirabilis populations, there are no amino acid differences. Remarkably, however, there are also no differences in the deduced amino acid sequences of the G. mirabilis and G. seta orthologues because all of the nucleotide substitutions are for synonymous codons. This result was unexpected, given the differences in kinetics and thermal stability between the G. mirabilis and the G. seta A4-LDHs. To ensure that no contamination had occurred in the handling of the G. seta mRNA or cDNA, the full sequencing process was repeated, using new fish and new sequencing reagents. The nucleotide sequence derived from the second G. seta preparation was identical to that derived from the first.
The ldh-a nucleotide sequence of C. nicholsi does code for one difference in amino acid sequence when compared with the Gillichthys orthologue. Although four of the five base differences are silent, the substitution of deoxyguanosine for deoxycytidine at position 233 in the nucleotide sequence results in the replacement of alanine in the Gillichthys protein by glycine in that of C. nicholsi at position 78 (Fig. 3). The difference in amino acid sequence between the Gillichthys spp. and C. nicholsi may contribute to the differences in electrophoretic mobility (data not shown). On a non-denaturing polyacrylamide gel stained for LDH activity, the A4-LDH bands of G. mirabilis and G. seta were identical, but differed from that of C. nicholsi.
Effects of urea and Gdn–HCl on the kinetics and thermal stabilities of A4-LDH
Although treatment with 3 mol l−1 urea did not denature any of the orthologues tested, as shown by retention of essentially full activity at the end of the incubation period, we observed significant effects on the kinetics of the enzyme of G. seta (Fig. 4). In contrast to the absence of any effect of urea on the Kmversus temperature profile for the enzymes of G. mirabilis (Fig. 4) and C. nicholsi (data not shown), the affinity of G. seta A4-LDH for pyruvate was reduced above 20 °C. At 40 °C, this reduction in affinity was substantial, with the pyruvate Km of G. seta A4-LDH increasing from 0.261 mmol l−1 to 0.568 mmol l−1. Thus, after treatment with 3 mol l−1 urea, the kinetics of G. seta A4-LDH became indistinguishable from that of the G. mirabilis orthologue. The same effect was observed after treatment of the two orthologues with 6.0 mol l−1 Gdn–HCl, a much stronger denaturing regimen than 3.0 mol l−1 urea (data not shown).
The effects of treatment with 3 mol l−1 urea on the thermal stabilities of G. mirabilis, G. seta and C. nicholsi A4-LDH from 45 to 55 °C are shown in Fig. 5. These effects were slight and non-uniform, leading in some cases to small increases in retention of activity and, in other cases, to minor decreases. However, despite the ability of urea to convert the kinetics of the G. seta A4-LDH to the pattern seen for G. mirabilis, after the same treatment the thermal stabilities of the enzymes remained distinct.
The effect of treatment with 6 mol l−1 Gdn–HCl on the stabilities at 50 °C of G. mirabilis and G. seta A4-LDH are also shown in Fig. 5 (inset). Despite the strong denaturing properties of Gdn–HCl, there is no apparent alteration in the thermostability of either enzyme.
Molecular mass determinations
Determinations of the subunit mass of LDH-A were made using ESI mass spectrometry on two separate enzyme preparations for each species, yielding values of 36 162.10±2.73 and 36 168.38±2.70 Da for G. mirabilis LDH-A and 36 151.03±2.44 and 36 156.82±2.77 Da for G. seta LDH-A; each mass value represents the mean ± S.D. of 11–16 peaks in an individual mass-to-charge (m/z) spectrum. The small difference in the mean masses of the G. mirabilis and G. seta LDH-A subunits cannot be regarded as significant in view of the resolving power of this technique (Edmonds and Smith, 1990).
Discussion
Kinetics of A4-LDH
The effects of temperature on Km of pyruvate of A4-LDH in each species reflect the temperatures found in the species’ habitats. Although Km values tend to rise with increasing temperature, numerous studies have shown that A4-LDH orthologues conserve Km values within a narrow range when measured at normal body temperatures (Yancey and Somero, 1978; Yancey and Siebenaller, 1987; Somero, 1995; Holland et al. 1997). Furthermore, these Km values consistently fall within the range of pyruvate concentrations found in skeletal muscle (Walsh and Somero, 1982). Trends found in these earlier studies are clearly reflected in data from the four goby species. For the two Coryphopterus species, the environmental temperature ranges are offset by approximately 15 °C, but have similar breadth (C. nicholsi 10–18 °C; C. personatus 25–32 °C). However, when Km values are compared within the normal temperature range of each species, they are found to be similar (C. nicholsi, 0.130–0.165 mmol l−1 pyruvate; C. personatus, 0.148–0.229 mmol l−1; see Fig. 1B).
The two G. mirabilis populations show no significant differences in A4-LDH kinetics across the temperature range tested (inset, Fig. 1A). Although the Gulf of California population probably experiences higher temperatures than the Pacific coast population, their A4-LDH orthologues are equally sensitive to temperature. However, the differences between the thermal habitats of the two G. mirabilis populations are not as extreme as those between the Coryphopterus species, and there is evidence that G. mirabilis behaviorally thermoregulates by burrowing (Barlow, 1961). Thus, the high surface-water temperatures of habitats in the Gulf of California may not be an adequate measure of the actual thermal conditions encountered by G. mirabilis. The lack of divergence in the A4-LDH kinetics of the two G. mirabilis populations may therefore be due to a number of factors, including the inherent eurythermy of the species and an ability to ameliorate environmental temperature fluctuations through behavioral thermoregulation. In addition, the lack of differences in A4-LDH sequence and function between the populations could reflect either ongoing gene flow between the populations sampled or relatively recent geographical separation (not more than 10 000–15 000 years bp, the approximate time at which waters warmed after the last glaciation and coastal marine populations migrated northward).
In comparison with the other gobies examined, the kinetic parameters of G. seta A4-LDH are remarkable in their insensitivity to temperature. G. seta A4-LDH maintains a relatively constant affinity for pyruvate at temperatures where the Km values of its confamilials rise steeply and reach values far above physiological pyruvate concentrations. In agreement with the extreme eurythermy of G. seta, its A4-LDH has the most temperature-insensitive Km of any A4-LDH yet examined (Yancey and Somero, 1978; Somero, 1995).
Although the time at which G. seta and G. mirabilis last shared a common ancestor cannot be accurately determined, certain factors suggest that divergence was relatively recent. These congeners are nearly indistinguishable morphologically as juveniles and occur in the same habitats (Barlow, 1961). Barlow suggests that G. seta is a paedomorphic form of G. mirabilis, adapted to life in an extreme and unstable environment. If this hypothesis is correct, it supports the view of Gould (1977, 1988) that paedomorphosis through progenesis can be the result of rapid adaptation to a variable environment.
Thermal stability of A4-LDH
Although no difference was noted in A4-LDH kinetics between the G. mirabilis populations and C. nicholsi, the A4-LDH of the latter species loses activity more rapidly than G. mirabilis A4-LDH when incubated at temperatures between 47.5 and 55 °C (Figs 2, 5). The only difference in the amino acid sequences of these orthologues is at position 78, where the alanine found in the orthologue of G. mirabilis is replaced by a glycine in the A4-LDH of C. nicholsi. This location is contained within a six-stranded parallel sheet that contributes to the stability of the LDH-A subunit (Abad-Zapatero et al. 1987). Although replacement of an alanyl residue by a glycyl residue represents only a minor change in the overall structure of a protein, this type of substitution has been shown to effect shifts in thermal stability of other enzymes. Work by Matthews et al. (1987) involving site-directed mutations in bacteriophage T4 lysozyme showed that a glycine-to-alanine shift increased the thermal stability of the enzyme. These authors argued that the replacement of a glycyl residue by an alanyl residue should enhance protein thermal stability by reducing the configurational entropy of the unfolded protein. Imanaka et al. (1986) likewise found that glycine-to-alanine substitutions enhanced thermal stability, and attributed this effect in part to an enhanced potential for hydrophobic interactions in the alanine-containing protein. Thus, there is a strong basis for concluding that the change in amino acid sequence at position 78 accounts for the difference in thermal stability between the A4-LDH orthologues of G. mirabilis and C. nicholsi.
The large difference in thermal stability between G. mirabilis and G. seta A4-LDHs is more enigmatic, because it cannot be explained by an underlying difference in amino acid sequence. Further, the thermal stabilities of the enzymes of these species are reversed with respect to the thermal sensitivity of their kinetics. Relative to the A4-LDH of G. mirabilis, the orthologue of G. seta denatures at a lower temperature despite having a more thermally stable Km of pyruvate.
These findings, notably, the difference in sequence between the orthologues of C. nicholsi and G. mirabilis, suggest that the kinetics and thermal stability of an enzyme can evolve independently and that different regions of a protein may be involved in effecting these two types of change. This conclusion is supported by the findings of Powers et al. (1993), who showed that the thermal stability of B4-LDH of the killifish (Fundulus heteroclitus) could be modified without any effect on the kinetic properties of the enzyme. Recently, Holland et al. (1997) showed that some single amino acid substitutions in A4-LDHs of barracuda congeners modify only Km, others modify only thermal stability and still other substitutions lead to changes in both characteristics.
Potential mechanisms accounting for differences in kinetics and thermal stability between A4-LDHs of G. mirabilis and G. seta
The discovery that orthologues of a protein with identical deduced amino acid sequences are characterized by different heat stabilities and temperature effects on Km suggests that the maturation processes that lead to the physiologically active form of A4-LDH may differ between the Gillichthys congeners. Potential differences between the mature A4-LDHs of G. mirabilis and G. seta include (1) covalent modification of the proteins, (2) non-covalent binding of a modulator of enzyme function and stability, and (3) the assumption of different conformational states in the two proteins, as a result of different pathways of folding, subunit assembly, or both.
In order to differentiate among these possibilities, we began a series of experiments that involved the treatment of A4-LDH with the denaturants urea and Gdn–HCl. Our initial method for incubating enzymes in 3 mol l−1 urea and subsequently removing the urea by dialysis was based on the discovery by Ozernyuk et al. (1994) that such urea treatment could modify the kinetic properties of A4-LDH of the loach (Misgurnus fossilis). In this fish, A4-LDH from specimens acclimated to different temperatures had different kinetic properties, with the Kmversus temperature curve for the cold-acclimated form being left-shifted with respect to that of the warm-acclimated form. This acclimation-induced difference in kinetic properties disappeared after treatment of the enzyme with 3 mol l−1 urea, such that the kinetics of the two forms became identical and intermediate. Although the work of Ozernyuk et al. (1994) focused on thermal acclimation in one species, as opposed to adaptation in two related species, the lack of difference between the Gillichthys LDH-A amino acid sequences suggested that a similar approach might provide evidence regarding the mechanism underlying the kinetic and thermal stability differences noted in the present study.
The treatments with urea and Gdn–HCl did not affect thermal stability, possibly because the conditions used were insufficient either to change the conformation or to remove a non-covalently bound factor. Attempts to denature the enzymes more fully with combinations of Gdn–HCl and high temperature failed to yield an active enzyme upon renaturation (data not shown). Consequently, data from the thermal stability experiments do not allow distinction among the hypotheses listed above. However, the results of kinetics studies (Fig. 4) do provide evidence in support of a mechanism involving either conformational differences or a non-covalently bound modulator. We have shown that the Km of pyruvate of G. seta A4-LDH can be modified to become indistinguishable from that of the G. mirabilis enzyme, and we believe that the ability of urea to alter the Km of pyruvate of G. seta A4-LDH is evidence against differential covalent modification as a mechanism for the interspecific differences observed, because washing with 3 mol l−1 urea is unlikely to affect covalent bonds.
The finding that the molecular masses of the A4-LDHs of the Gillichthys congeners cannot be discriminated by the sensitive technique of electrospray ionization mass spectrometry is further evidence that the differences between the enzymes are due either to alternative conformations or to binding of a modulator to the enzyme of G. seta. While the molecular mass data do not allow us to exclude the possibility that identical covalent modifications at different sites on the molecules may be responsible for the variations in Km and thermal stability, we feel that such an explanation for the identical masses is unlikely.
Our data do not allow us to distinguish whether the differences in Kmversus temperature relationships and thermal stability between the A4-LDHs of the Gillichthys congeners are due to different conformational states or to non-covalent binding of a modulator to the LDH-A from G. seta. However, a potential role for the former mechanism is supported by results from several other studies of LDHs and additional enzymes, in which evidence for multiple conformations of a single protein is presented. In addition to the work of Ozernyuk et al. (1994), there are several other reports of conformers of A4-LDH that differ in kinetic or structural properties. Xue and Yeung (1995) used a laser-based methodology that allowed replicate measurements of the activities of single A4-LDH molecules. They found that the activities of individual enzyme molecules from an electrophoretically pure preparation differed by up to fourfold, a result that they attribute to the occurrence of several stable conformers of the enzyme. Levi and Kaplan (1971), through experiments involving denaturation and renaturation of A4-LDH of dogfish, showed that LDH can exist in multiple catalytically active conformations. The conformation of the renatured, active enzyme in this study was different from that of the native form, as measured by protein fluorescence, optical rotary properties and the reactivity of sulfhydryl groups. King and Weber (1986a,b) used high pressure and low temperature to dissociate LDH, and subsequently followed the time course of reassociation. They provided evidence that LDH subunits can exist in multiple conformations that differ both in their ability to assemble into tetramers and in their enzymatic activity. They further showed that the conformational changes responsible for the loss of activity were independent of those that affected tetramer reassembly. Several others enzymes have been shown to exist in multiple conformational states, including alcohol dehydrogenase (Jacobson et al. 1970), glyceraldehyde-3-phosphate dehydrogenase (Marangos and Constantinides, 1974) and pyruvate kinase (Susor and Rutter, 1968; Somero, 1969). Interestingly, the temperature-dependent conformers of pyruvate kinase of the king crab (Paralithodes camtschatica) differed in their kinetic properties in an apparently adaptive manner (Somero, 1969). Nickerson (1973) has presented a theoretical analysis of what he terms ‘multistable’ proteins, emphasizing that a single primary structure may have multiple folded states with different free energies, and that the conformation assumed by a protein under physiological conditions need not be the one that is thermodynamically most stable. He argues that multiple conformers of enzymes can be generated under different environmental conditions and that these may have adaptive significance.
A further example of conformational variation in polypeptides sharing a single amino acid sequence is provided by the prion phenomenon. A protein normally found in the brain cells of humans and other mammals, prion protein (PrPC) has been shown to undergo a conformational shift to a form termed PrPSc for the disease scrapie (Prusiner, 1991). This form shows extensive β-sheets in place of the α-helices of PrPC (Pan et al. 1993) and differs from the native form in important biochemical properties, including a tendency to aggregate and resistance to proteolysis (Prusiner, 1994). Interestingly, PrPC can convert irreversibly to PrPSc in the presence of other PrPSc molecules, resulting in intracellular protein aggregations that cause a number of diseases including scrapie, bovine spongiform encephalopathy and Creutzfeldt–Jakob disease (Telling et al. 1996). Each of these is infectious through contact with diseased tissue, but has no apparent nucleic acid component. These findings imply that PrPSc acts as a template that converts PrPC molecules to the PrPSc form (Prusiner, 1994).
A fascinating prion analog in yeast is the [psi+] trait, involving the aggregation of the Sup35 protein, normally involved in translation termination (Ter-Avanesyan et al. 1994). Sup35 polypeptides in a particular conformation, [psi+], appear to act as templates that cause a shift in conformation of [psi−] proteins to the [psi+] form. The heat shock protein Hsp104 appears to mediate this conversion, such that when Hsp104 levels are transiently increased, Sup35 in [psi+] cells begins to disaggregate and revert to the native [psi−] conformation. If Hsp104 remains at high levels, the yeast cell can be ‘cured’ of the [psi+] condition and shows no more aggregates (Patino et al. 1996). These results indicate that intracellular proteins can exhibit alternative conformers depending on the chaperone proteins present; furthermore, such conformational shifts can be due to changes in Hsp levels, as opposed to changes in the Hsp structures themselves (Chernoff et al. 1995; Lindquist and Kim, 1996).
The diverse studies described above show that conformers of proteins can exhibit differences in functional characteristics and thermal stability, despite the absence of differences in sequence and covalent modification. These results thus provide support for our hypothesis that conformational variants can account for the differences in Kmversus temperature relationships of the A4-LDHs of the Gillichthys congeners. However, the present study is unprecedented in suggesting that conformers may be responsible for adaptive differences between species.
The mechanisms that could generate these hypothesized conformers of A4-LDH in the Gillichthys congeners remain unknown. However, at least one potential mechanism can be excluded: temperature-dependent folding. Because specimens of G. seta and G. mirabilis were acclimated to a common temperature (approximately 15 °C), the alternative conformations proposed for their A4-LDHs could not simply be a direct effect of temperature on the processes of folding and subunit assembly. Rather, unidentified cellular factors are hypothesized to direct folding along different maturation pathways in the two species. These factors might include low-molecular-mass constituents of the cell that bind to the nascent LDH-A polypeptides and affect folding or assembly. Possibly, these low-molecular-mass factors remain bound non-covalently to the protein and affect its stability and kinetics. Because the kinetic properties of LDHs are influenced by the energy changes that result from conformational alterations during binding and release of substrates and cofactors (see Holland et al. 1997), bound solutes that affect conformational mobility could have significant effects on kinetic constants such as Km. Thus, non-covalent binding of low-molecular-mass solutes to A4-LDH, while seemingly an alternative explanation for the differences between the enzymes of G. seta and G. mirabilis, may also exert its effects by influencing the conformation of the enzyme or, more precisely, the energy changes that occur during functionally important shifts in conformation.
The involvement of molecular chaperones is hypothesized to be another potential mechanism for generating conformers. Although chaperones are generally regarded as facilitating folding without determining the final folded state of the protein (Hartl, 1996), the discovery that Hsp104 can affect the conformation of Sup35 protein suggests that differences in chaperoning activity could potentially lead to alternative folding of A4-LDH in the two Gillichthys species.
Whatever the mechanism is that underlies the differences in kinetics and stability between the mature A4-LDHs of the Gillichthys congeners, our results suggest that differences in amino acid sequence are not always necessary for adapting proteins to temperature. If the kinetics and stabilities of enzymes can be modified through the production of stable alternative conformers during the maturation of a newly synthesized protein, then another mechanism for the adaptation of enzymes, in addition to changes in their amino acid sequence, may exist.
ACKNOWLEDGEMENTS
We thank Ms L. Z. Holland for her help in preparation and sequencing of the goby cDNA. We thank Mr R. McConnaughey and Dr M. E. Clarke for their assistance in collecting specimens. P.A.F. would also like to express his gratitude to Dr J. B. Graham for helpful advice and guidance during much of this work. The work described in this paper was funded in part by NSF grant IBN92-06660, in part by a grant from the National Sea Grant College Program, National Oceanic and Atmospheric Administration, US Department of Commerce, under grant number NA89AA-D-SG138, project number 72-C-N, through the California Sea Grant College, and in part by the G. Unger Vetlesen Foundation. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies. The US Government is authorized to reproduce and distribute for governmental purposes.