Structural and Functional Differentiation of Two ... - Semantic Scholar

Structural and Functional Differentiation of Two Clinally Distributed Glucosephosphate Isomerase Allelic Isozymes from the Teleost Fundulus keferocZitusl Rebecca J. Van Beneden and Dennis A. Powers3 Department

of Biology, The Johns Hopkins University

The teleost Fundulus heteroclitus(L.) possesses two loci, Gpi-A and Gpi-B, for the glycolytic enzyme, glucose-phosphate isomerase (GPI; D-glucose-6-phosphate ketolisomerase; E.C. 5.3.1.9). The Gpi-B locus is polymorphic in Fundulus, with two common alleles, Gpi-Bb and Gpi-B’, distributed in a clinal manner in populations along the east coast of North America. Since this clinal distribution is strongly correlated with a temperature gradient, we asked whether the GPI-B2 allozymes were functionally adapted to the thermal environment in which a given phenotype predominated. The two major GPI-B2 allozymes were purified to homogeneity and were characterized as to molecular weight, isoelectric pH, thermal denaturation, and kinetic parameters. Both GPI-Bi and GPI-B; allozymes have molecular masses of 110 kD, and they have isoelectric pHs of 6.4 and 6.6, respectively. The GPI-B; allozyme was more stable to thermal denaturation than was the GPI-BS enzyme. Kinetic properties of the allelic isozymes were investigated both as a function of pH and as a function of temperature. At 25”C, over the pH range considered, there were no significant differences between allozymes, either in K, for fi-uctose6-phosphate or in Ki for 6-phosphogluconate, but apparent V,,, values differed between pH 7.5 and pH 8.5. All steady-state kinetic parameters showed strong temperature dependence, but the allozymes differed only in the Ki for 6-phosphogluconate at temperatures >3O”C. On the basis of the observed structural and functional differences alluded to above, the hypothesis that the major allelic isozymes of the Gpi-B locus were functionally equivalent was rejected. However, it is not yet known whether these structural and functional differences have any significance at higher levels of biological organization.

Introduction Aquatic poikilotherms inhabit a wide range of habitats encompassing extremes of temperature, salinity, and pressure. Aquatic organisms that live in temperate estuarine environments are subjected to additional daily and seasonal perturbations. Complex interactions and modifications at behavioral, physiological, metabolic, and

biochemical levels enable these organisms to accommodate some environmental changes. Since the development of electrophoretic techniques to examine genetic vari1. Key words: allelic isozymes,cline, GPI. Abbreviations: GPI-A2 = muscle glucose-phosphate isomerase enzyme (Gpi-A = GPI-A2 locus); GPI-B2 = liver glucose-phosphate isomerase enzyme (Gpi-B = GPI-B2 locus); Fru-6-P = fructose-6-phosphate; BSA = bovine serum albumin; E, = energy of activation; Glc-6-P = glucosed-phosphate; e.u. = enzyme units. 2. Current address: Duke University Marine Laboratory, Beaufort, North Carolina 285 16.

3. Current address and address for correspondence and reprints: Dennis A. Powers, Hopkins Marine Station, Department of Biological Sciences, Stanford University, Pacific Grove, California 93950. Mol. Biol. Evol. 6(2): 155-170. 1989. 0 1989 by The University of Chicago. All rights reserved.

0737-4038/89/0602-0004$02.00

155

156 Van Beneden and Powers

ation in natural populations (Markert and Moller 1959)) attention has focused on the role of multiple enzyme forms in adaptation to environmental variations. Somero and Hochachka (1976) postulated several ways in which genetic adaptation at the protein level might effect evolutionary changes: (1) alteration of the in vivo concentration of an enzyme or its substrate, (2) modification of the activity of a preexisting protein, and / or ( 3 ) development of genetic variants with different catalytic efficiencies. Two classes of genetic variants have been described. The first, the multilocus isozymes, have analogous functions but are encoded by different loci. The second class are the allelic isozymes (allozymes), the multiple allelic products of a single locus. The role of allelic isozymes in adaptation of organisms has been a topic of active investigation among population biologists for many years. Still debatable is what fraction of these electrophoretic variants are maintained by natural selection and what fraction are selectively neutral. One step toward addressing this question is to test the hypothesis that allelic isozymes are functionally equivalent. To that end, we have examined the glucose phosphate isomerase (GPI; E.C. 5.3.1.9) system of Fun&&s heteroclitus.[In keeping with the IUPAC-IUB recommendations for multiple-enzyme nomenclature (J. Biol. Chem. 252:5939-594 1,1979) and previous reports on glucosephosphate isomerase (Place and Powers 1978; Palumbi et al. 198 l), the muscle and liver glucose-phosphate isomerase isozymes are designated as GPI-A2 and GPI-B2, respectively.] Fundulus heteroclituspossessestwotissue-specific Gpi loci (Avise and Kitto 1973). The GPI-A2 isozyme predominates in anaerobic tissue such as white muscle, while GPI-B2 is present in aerobic tissues such as liver, heart, and other internal organs (Place and Powers 1978). The physiological roles of the GPI multilocus isozymes (GPI-AZ and GPI-B2) have been described recently in a detailed kinetic analysis of purified enzymes (Van Beneden and Powers 1985). Electrophoretic variation at one of these loci ( Gpi-B) has been reported elsewhere (Place and Powers 1978). The two common alleles ( Gpi-Bband Gpi-Be) at this locus are inherited in a Mendelian manner (Place and Powers 1978). Herein we present an analysis of the GPI-B2 allozymes that includes their clinal distribution, their purification and physical characterization, and the kinetic analysis of these allelic isozymes as a function of pH and temperature. Material and Methods

Fish were caught by using Gee-type minnow traps or minnow seines, when conditions permitted. Sampling for cline studies was done both along the eastern U.S. coast from Maine to Florida and within the Chesapeake Bay and its major tributaries. Fish at each location were pooled, with no distinction made for sex or age. Livers from adult fish were dissected in the field, immediately frozen in liquid nitrogen, and kept frozen in liquid nitrogen until needed. Livers used in purification of GPI-B$ for kinetic analysis were primarily taken from fish collected in Wiscasset, Me. Tissue from Wiscasset showed Gpi-BCallele at frequencies of BO.94. Liver tissue used for the isolation of GPI-Bt was obtained from fish from either Sapelo Island, Ga., or Stone Harbor, N.J. Since the Gpi-Bb allele does not approach fixation at either of these locations, individual fish were typed either by fin clipping live fish (Place and Powers 1978 ) or by using liver chips from individual frozen livers which were homogenized and resolved via starch gel electrophoresis to enable the phenotypes to be identified by histological staining (Powers and Place

7o”

w

5o” w

60°W

50°N

‘45ON

40°N

30°N

Scale

of Miles

Glucosephosphate Isomerase Allozymes

159

of further fractionation procedures. Purity of 95%-98% was obtained for most prep arations; however, occasionally 92%-95% purity was obtained. Molecular mass determined by gel filtration was 110 + 5.0 kD for each allozyme. Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate yielded a single band of between 54 and 55 kD for GPI-B! and GPI-B$ . These data are typical of values obtained for GPI from other organisms (Takeda et al. 1967; Carter and Dando, 1973; Hill and Carter 1973; Kempe et al. 1974; Gracy and Tilley 1975; MO et al. 1975; Phillips et al. 1976; Charles and Lee 1980; Gee et al. 1980; Schreyer and Bock 1980; Gottlieb and Greve 198 1; Weeden and Gottlieb 1982) and suggest that the allozymes are dimeric proteins. The isoelectric pH of each allozyme, obtained from the isoelectric focusing column used as the last step in purification, was 6.4 +_ 0.2 and 6.6 +, 0.1 for GPI-Bt and GPI-B 5, respectively. Relative Specific Activities GPI-B2 activities were measured for fish from different populations. In all cases, there were no significant differences. For example, GPI activities from fish from Maine with GPI-B$ were identical to those of fish from Georgia with GPI-B! ( 1.9 f 0.3 mol/ min / mg protein ) . Heat Denaturation Results of the heat denaturation experiments are shown in figures 5 and 6. The TsO’s-i.e., temperatures at which half the original enzyme activity remains-are 44.6”C f 0.5”C and 42.3”C f 0.5”C for GPI-Bt and GPI-B$, respectively. These values were not significantly altered in the presence of substrate [fructose-6-phosphate (Fru-6-P)] or phosphate buffer (data not shown). Thus, neither the inclusion of substrate nor incubation in higher-ionic-strength buffer appeared to affect the thermal stability. The differences in thermal stability were most apparent when the enzymes were incubated at 42°C and when activity was monitored as a function of time (fig. 6). The kinetics of denaturation are consistent with a simple first-order process. Kinetic Characterization:

Effect of pH and Temperature

Figures 7-9 summarize the effect of pH on the kinetic parameters of the GPI-B2 allozymes at a constant temperature (25 “C) . In general, over the pH range examined at 25”C, there are no significant differences between the allozymes, either in K,,, for Fru-6-P or in the inhibition constant (&) for 6-phosphogluconate. The K, for Fru6-P changes dramatically between pH 7.5 and pH 8.5 for both allelic isozymes (fig. 7). This change in K, suggests that at least one amino acid side chain affecting the binding of Fru-6-P is being titrated. On the other hand, the Ki for 6-phosphogluconate shows no such effect ( fig. 8 ) , a finding suggesting either that the pH-dependent amino acid residue that affects the K, for Fru-6-P is not involved in the binding of 6-phosphogluconate or that the real inhibitor is 6-phosphogluconolactone, whose concentration in the mixture would decrease with increasing pH. Significant differences are apparent between V,,, values over the pH range of 7.5-8.5, but no significant differences in this parameter exist at either pH 7.0 or pH 9.0 (fig. 9). Figures 10 and 11 summarize the effect of temperature on K,,, for Fru-6-P and

FIG. 2.-Location

of sampling sites for Fundulus heteroclitus along the eastern U.S. coastline

76’

38”

7O

FIG. 3.-Location tributaries.

‘6”

of sampling sites for Fundulus

160

heteroclitus

in the Chesapeake Bay and its major

42

40 38 36 Latitude to North )

34

32

variation in Gpi-Bf gene frequency along the eastern U.S. coast and within the FIG. 4-Geographical Chesapeake Bay. Open symbols (0) represent data obtained from fish collected in the Chesapeake Bay in 1979. Blackened symbols represent data calculated from fish collected along the coast over a 9-year time span: 0 = fish collected 1972-75 (Powers and Place 1978); w = fish collected in 1977-78; A = collections made in 1981.

100

-$ .f B

75

% .?z > ._ 5 a

50

8

25

0

45

40 Temperature FIG. 5.-Thermal

denaturation

of Fund&s

50

i” C)

heteroclitusGPI-B2 allelic isozymes. Cl = GPI-BP; A

= GPI-B; .

161

162 Van Beneden and Powers

IO

20

30 TIME

FlG. 6.-Time

( MINUTES

50

40

60

)

dependence of thermal stability of GPI-B2 allelic isozymes at 42°C. 0 = GPI-B!; A

= GPI-B;.

on Ki for 6-phosphogluconate, respectively. All parameters show strong temperature dependence. No differences are apparent between K,,, (Fru-6-P) values. At temperatures >3O”C, the Ki values for 6-phosphogluconate are significantly different. At these temperatures, the GPI-Bg allozyme has a larger Ki than does GPI-B$ (fig. 11). The Arrhenius plots for the GPI-B2 allelic isozymes were linear (e.g., see Van Beneden and Powers 1985), indicating that the rate-limiting constant did not change over the temperature range studied. Small differences in the thermodynamic activation parameters E, and AH* were found between the allelic isozymes. For example, the AH* values for the GPI-B!j and GPI-B$ allozymes were 14.1 + 0.3 and 13.5 + 0.2 IScal/mol, respectively. However, these small differences are primarily the result of the IL,, value for GPI-Bs at 25°C. If that value is eliminated from the analysis, the thermodynamic parameters are identical. Over the temperature range studied, the pseudo-first-order rate constant at very low substrate concentrations (V&J K,,,) did not differ significantly between the allozymes (fig. 12). Discussion

Several clines have been reported for a variety of polymorphic enzyme systems in Fundulus heteroclitus (Powers and Place 1978; Cashon et al. 198 1; Powers et al. 1986; Ropson et al., submitted). Three points from the Gpi-B cline data should be emphasized: (1) The Gpi-B”allele appears to be fixed in the northern part of the coast and also predominates in the northern areas of the Chesapeake Bay. (2) A cline present in the Chesapeake Bay parallels that of the east coast but is shifted in latitude, to the

Glucosephosphate Isomerase Allozymes

I

7.0

I

7.5

I

8.0

I

8.5

163

I

9.0

PH FIG. 7.--K,,, as a function of pH. Cl = GPI-Bt; A = GPI-B;

south of its east coast counterpart, by 100-200 miles. ( 3) An apparent temporal shift in gene frequency (toward the c allele) is observed in the middle ranges of the species’ distribution (34”N-37’N). This cline in Gpi-B gene frequency could have arisen by either primary or secondary integradation (see Powers et al. 1986). Gonzalez-Villasenor and Powers (submitted) have presented mtDNA restriction-fragment-length polymorphism data that support the secondary-integradation hypothesis. This is an important finding that provides insight concerning the general nature of gene diversity in this species, but it does not provide insight concerning the relative contributions of chance and adaptive forces that drive the required genetic differentiation for both the primary- and the secondary-integradation models. Yet, it is the relative role of chance and adaptive forces that strikes at the very heart of the “neutralist/selectionist” controversy that has gone unresolved for >2 decades. The neutralist hypothesis implies that most genetic variation is functionally equivalent; that is, different allelic variants behave in an identical fashion in terms of cellular and organismal function. Therefore, this hypothesis is weakened for specific loci whenever functional nonequivalence between the genetic alternatives can be established. Of course, establishment of in vitro functional nonequivalence of genetic alternatives is just the first step in the long process of determining whether an enzyme polymorphism is subject to natural selection.

164 Van Beneden and Powers

loo 70

50

30 a z j 20 iz

IO

5 I

7.0

I

7.5

I

8.0

I

8.5

I

9.0

PH FIG. 8.--6Phosphogluconate

Ki as a function of pH. El = GPI-Bq; A = GPI-B;

Clarke ( 1975 ) has established a four-step strategy toward addressing the problem of genetic variation at enzyme-synthesizing loci. The first step is to “make a detailed biochemical and physiological study of enzymatic products of the alleles, noting any differences between them.” The present paper and the accompanying paper of Ropson and Powers (1989) address this point by analyzing the structural and functional characteristics of allelic isozymes. Heat Denaturation

of GPI-B Allozymes

The results of heat denaturation studies showed that GPI-Bt , the allozyme most predominant in southern latitudes, was more resistant to thermal denaturation than was GPI-Bs (see figs. 5, 6). Differences in heat stabilities between allelic variants of Gpi-B have been reported in rabbit (Welch et al. 1970), mouse (Charles and Lee

Glucosephosphate Isomerase Allozymes

165

300 -250 -i r TE200 E iii r" -2 150 is E > 100

7.0

7.5

0.0

8.5

9.0

PH FIG. 9.-pH

dependence of apparent V,,,,, GPI-B! ( Cl) and V,,,,, GPI-BC,(A)

1980)) chinchilla and cuis (Hill and Carter 1973)) and humans (Tilly et al. 1974), just to name a few species. In colias butterflies, heat-stability differences in GPI allozymes are also correlated with environmental temperature (Watt 1977). While heat-stability studies are useful as physical probes for enzyme flexibility and structure, differences in heat stability in vitro do not necessarily mean that there is an in vivo adaptive advantage for the organism. The TSO,the temperature at which 50% of the enzyme is denatured, reported for most enzymes is often much higher than the upper lethal temperature of the organism. Although the T&s for GPI-B$ (44.6”C) and GPI-Bt (42.3”C) are close to the range of short-term temperatures experienced by F. heteroclitus fry during the summer months (D. A. Powers, unpublished data), the normal upper lethal temperature of adult F. heteroclitus is 36°C (Garside and Jordan 1968). On the other hand, at 35”C, the GPI-Bt allozyme is fully active, while the GPI-B? allozyme only has 85% of its activity remaining after 10 min of incubation (fig. 5). If these enzymes are incubated for 2 h at 35°C GPI-Bt will be essentially fully active, while GPI-B$ will still have no detectible activity. If in vitro denaturation data are representative of enzyme stability in vivo, then, at higher temperatures (e.g., 35 “C), fish with the GPI-B! allozyme would have a significant advantage over fish with GPI-Bs .

166 Van Beneden and Powers

80 60 z ‘;'40 x

20

I

I

IO

I5

FIG. lO.-Temperature

I

I

I

I

I

20

25

30

35

40

dependence of K,,, for Frud-P. Cl = GPI-B!; A = GPI-B’,

Kinetic Parameters Catalytic-parameter differences between the GPI-B2 allelic isozymes are more subtle than those seen for the multilocus isozymes (GPI-B2 and GPI-A2 ; Van Beneden and Powers 1985). Under all conditions of temperature and pH examined, K,,, and I&,,/&, values were statistically the same for the ahelic isozymes. While small differences were observed in apparent V,,, values at 25°C over the physiological pH range ( pH 7.5-8.5 ) , no significant differences were observed at other temperatures. Since l&,x = kat [ &I, where kat is the catalytic rate constant and [ I.$,] is the active-

I

I

IO

I5

I

I

I

I

I

20

25

30

35

40

80 60 h

40

5 Y iz

20

IO

TEMPERATURE (“Cl FIG. 11.-Temperature = GPI-B;; A = GPI-B;.

I

I2 IO

-

dependence

I

of Ki for the competitive

I

I

inhibitor 6-phosphogluconate.

I

I

Cl

I

00

8

a 2

I

I

IO

15

I

I

20

25

I

TEMPERATURE FIG. 12.-Temperature

dependence of V,,,/K,.

30 ("C)

I

I

35

40

Cl = GPI-B:; A = GPI-BC,

-

168 Van Beneden and Powers

enzyme concentration, it would take a very small difference in active enzyme to account for the apparent difference illustrated in figure 9. Therefore, the most conservative approach would be to discount these apparent differences in V,,, . On the other hand, there were consistent statistically significant differences in the Ki for 6-phosphogluconate at temperatures >3O”C. Moreover, this difference increased directly with temperature. The GPI-Bg allozyme had a higher Ki for 6-phosphogluconate than did GPI-B$. Thus, the allelic isozyme that has less sensitivity to 6-phosphogluconate inhibition at high temperatures predominates in F. heteroclitus populations in southern climates. Determinations of metabolite levels in livers from Fundulus acclimated to different temperatures (D. A. Powers, unpublished data) show that fish acclimated to 30°C have significantly higher 6-phosphogluconateglc-6-P ratios than do fish acclimated at 10°C. Liver [6-phosphogluconate]:[Glc-6-P] ratios were 0.138:0.020 and 0.76:0.219 for 10°C and 30”Cacclimated fish, respectively. The higher Ki for 6-phosphogluconate exhibited by the GPI-B$ allozyme may provide an adaptative advantage to compensate for the increased concentrations of 6-phosphogluconate present at higher temperatures. However, metabolic-flux studies will be necessary to test this hypothesis. In summary, data obtained from heat denaturation studies and from some of the kinetic analysis of purified enzymes indicate that the GPI-B2 allelic isozymes in F. heteroclitus are functionally nonequivalent. In vitro, GPI-Bt is more stable with respect to denaturation at higher temperatures and is less sensitive to inhibition by 6phosphogluconate at temperatures >3O”C. Fish with the GPi-B? enzyme predominate in the warmer regions of the cline. While we can reject the hypothesis that the GPIB2 allelic isozymes are functionally equivalent, we have no evidence at the present time that these in vitro differences are reflected at those higher levels of biological organization that would affect the differential fitness. However, the biochemical differences between GPI-B2 allelic isozymes, differences documented in the present paper, can now be used to predict differences in cell physiology, organismic responses, and, ultimately, fitness. Those predictions can be tested by appropriate experimental design in a manner similar to that which we have used for the LDH-B4 allelic isozymes of F. heteroclitus (reviewed in Powers 1987). Acknowledgments

We thank Dave Morgan, Michelle Queen, and Jonathan Swift for their invaluable help. We thank Streamson Chua and Charles Montague for their assistance with the computer analysis. We appreciate the efforts of all those who helped with field collections, including Drew Brown, Robert Cashon, Mitchell Hobish, Julie Linker, Lewis Linker, Ira Ropson, Henrik Steinberg, Dennis Wrightson, Captain Bill Harris and the crew of the R. V. RidgeZy Warfield, the staff at the University of Georgia Sapelo Island Marine Institute, and the staff at Lehigh University Wetlands Institute, Stone Harbor, N.J. We thank Dianne Powers for her technical and artistic assistance. LITERATURE

CITED

AVISE, J. C., and G. B. KITTO. 1973. Phosphoglucose isomerase gene duplication in the bony fishes: an evolutionary history. Biochem. Genet. 8: 113- 132. CARTER, N. D., and P. R. DANDO. 1973. Phosphoglucose isomerase in teleostean fish. B&hem. Sot. Trans. 1:1263- 1264. CASHON,R. E., R. J. VAN BENEDEN,and D. A. POWERS.198 1. Biochemical genetics of Funduh

Glucosephosphate Isomer-ax Allozymes

169

heteroclitus (L). IV. Spatial variation in gene frequencies of Idh-A, Idh-B, 6-Pgdh-A, and E&S. Biochem. Genet. 19:7 15-727. CHARLES, D. J., and C.-Y. LEE. 1980. Biochemical and immunological characterization of genetic variants of phosphoglucose isomerase from mouse. Biochem. Genet. 18: 153- 169. CLARKE, B. 1975. The contribution of ecological genetics to evolutionary theory: detecting the direct effects of natural selection on particular polymorphic loci. Genetics 79: 10 1- 108. GARSIDE, E. T., and C. M. JORDAN. 1968. Upper lethal temperatures at various levels of salinity in the euryhaline cyprinodontids Fundulus heteroclitus and F. diuphanus after isosmotic acclimation. J. Fisheries Res. Board Can. 24:27 17-2720. GEE, D. M., G. M. HATHAWAY, R. H. PALMIERI, and E. A. NOLTMANN. 1980. Molecular characterization of pig muscle phosphoglucose isomerase. J. Mol. Biol. 142:29-42. GONZALEZ-VILLASENOR,L. I., and D. A. POWERS. Mitochondrial DNA restriction site polymorphisms in the teleost Fundulus heteroclitus supports secondary integradation. Evolution (submitted). GOTTLIEB, L. D., and L. C. GREVE. 198 1. Biochemical properties of duplicated isozymes of phosphoglucose isomerase in the plant Clurkia xuntiana. Biochem. Biophys. 19: 155- 172. GRACY, R. W., and B. E. TILLEY. 1975. Phosphoglucose isomerase of human erythrocytes and cardiac tissue. Methods Enzymol. 4 1 [B] : 392-400. HILL, M. R., and N. D. CARTER. 1973. Comparative properties of rodent phosphoglucose isomerases. Biochem. Sot. Trans. 1: 1264- 1266. KEMPE, T. D., Y. NAK AGAWA, and E. A. NOLTMANN . 1974. Physical and chemical properties of yeast phosphoglucose isomerase isoenzymes. J. Biol. Chem. 249:46 17-4624. MARKERT, C. L., and F. MOLLER. 1959. Multiple forms of enzymes: tissue, ontogenetic and species-specific patterns. Proc. Natl. Acad. Sci. USA 45:753-763. MO, Y., C. D. YOUNG, and R. W. GRACY. 1975. Isolation and characterization of tissuespecific isozymes of glucosephosphate isomerase from catfish and conger. J. Biol. Chem. 250:6747-6755. PALUMBI,S. R., B. D. SIDELL,R. J. VAN BENEDEN,and D. A. POWERS. 1980. Glucosephosphate isomerase (GPI) of the teleost Fundulus heteroclitus (Linneaus): isozymes, allozymes and their physiological roles. J. Comp. Physiol. 138:49-57. PHILLIPS, T. L., J. M. TALENT, and R. W. GRACY. 1976. Isolation of rabbit muscle glucosephosphate isomerase by a single-step substrate elution. B&him. Biophys. Acta 429:624628. PLACE,A. R., and D. A. POWERS. 1978. Genetic bases for protein polymorphism in Fundulus heteroclitus (L.) . I. Lactate dehydrogenase (Ldh-B), malate dehydrogenase (Mdh-A ), glucosephosphate isomerase ( Gpi-B), and phospho-glucomutase (Pgm-A ) . Biochem. Genet. 16:577-59 1. POWERS, D. A. 1987. A multidisciplinary approach to the study of genetic variation in species. Pp. 102-l 34 in M. E. FEDER, A. F. BENNET, W. W. BURGGREN, and R. B. HUEY, eds. New directions in physiological ecology. Cambridge University Press, New York. POWERS, D. A., and A. R. PLACE.1978. Biochemical genetics of Fundulus heteroclitus (L.) I. Temporal and spatial variation in gene frequencies of Ldh-B, Mdh-A, Gpi-B, and Pgm-A. Biochem. Genet. 16:593-607. POWERS,D. A., I. ROBON, D. C. BROWN, R. J. VAN BENEDEN,R. E. CASHON,L. I. GONZALEZVILLASENOR,and J. A. DIMICHELE. 1986. Genetic variation in Fundulus heteroclitus: geographic distribution. Am. Zool. 26: 13 1- 144. ROPSON , I. J., D. C. BROWN, and D. A. POWERS.Biochemical genetics of Fundulus heteroclitus (L.) VI. Spatial variation in the gene frequencies of 15 loci. Evolution (submitted). ROPSON, I. J., and D. A. POWERS. 1989. The allelic isozymes of hexosed-phosphate dehydrogenase isolated from Fundulus heteroclitus: physical characteristics and kinetic properties. Mol. Biol. Evol. 6:171-185. SCHREYER,R., and A. BOCK. 1980. Phosphoglucose isomerase from Escherischia coli KlO:

170

Van Beneden and Powers

purification, properties and formation under aerobic and anaerobic conditions. Arch. Microbiol. 127:289-296. SOMERO, G. N., and P. W. HOCHACHKA. 1976. Biochemical adaptations to temperature. Pp. 125-190 in R. C. NEWELL, ed. Adaptation to environment: essays on the physiology of marine animals. Butterworths, Boston. TAKEDA ,Y., S. HISUKURI, and Z. NIKUNI . 1967.Crystallization and properties of pea glucosephosphate isomerase. Biochim. Biophys. Acta 146:568-575. TILLEY, B. E., R. W. GRACY, and S. G. WELCH. 1974. A point mutation increasing the stability of human phosphoglucose isomerase. J. Biol. Chem. 249:457 l-4579. VAN BENEDEN,R. J., and D. A. POWERS. 1985. The isozymes of glucose-phosphate isomerase (GPI-A2 and GPI-B2) from the teleost fish Fund&s heteroclitus(L.) J. Biol. Chem. 260: 14596-14603. WATT, W. B. 1977. Adaption at specific loci. I. Natural selection on phosphoglucose isomerase of Colias butte&es: biochemical and population aspects. Genetics 87: 177- 194. WEEDEN, N. F., and L. D. GOTTLIEB. 1982. Dissociation, reassociation, and purification of plastid and cytosolic phosphoglucose isomerase isozymes. Plant Physiol. 69:7 17-723. WELCH, S., L. FITCH, and C. PARR. 1970. A variant of rabbit phosphoglucose isomerase. Biochem. J. 117:525-53 1. WALTER M. FITCH, reviewing Received

June 5, 1987; revision

editor received

November

15, 1988