- Email: [email protected]
Correlations between water quality and frequencies of allozyme genotypes in spotfin shiner (Notropis spilopteris) populations
Environmental Pollution 81 (1993) 147-150
CORRELATIONS BETWEEN WATER QUALITY A N D FREQUENCIES OF ALLOZYME GENOTYPES IN SPOTFIN SHINER (Notropis spilopteris) POPULATIONS R. B. Gillespie Department of Biological Sciences, Indiana University-Purdue University at Fort Wayne, Indiana 46805-1499, USA
&
S. I. Guttman Department of Zoology, Miami University, Oxford, Ohio 45056, USA (Received 12 February 1992; accepted 22 May 1992)
Genotype frequencies of glucose-6-phosphate isomerase (GPI) allozymes differed significantly among populations of the spotfin shiner Notropis spilopterus from sites with varying water quality. Frequencies of shiners with a GPI-2 BB genotype decreased significantly at sites with reduced water quality. Alternatively, the total frequencies of shiners with a genotype of GPI-2 AA and AB increased at sites with reduced water quality. Individuals with certain allozyme genotypes may be more sensitive to the toxic effects of polluted waters than those with other genotypes. The selection of individuals with sensitive genotypes may reduce genetic diversity in populations and thus increase the susceptibility of these populations to additional novel stresses. Because allele and genotype frequencies of GPI-2 were correlated with water quality, electrophoretic determination of genetic structure in fishes may be a useful tool for monitoring the health of aquatic populations.
genetic characteristics in wild populations. Differences in allozyme frequencies in populations of the central stoneroller were associated with different levels of exposure to contaminants in a stream receiving non-point sources of run-off from a uranium-enrichment facility (Gillespie & Guttman, 1989). The genetic diversity of a population may therefore moderate the toxic effects of pollutants and provide a mechanism of population survival. Assessments of allozyme composition in aquatic populations may provide a sensitive end-point for the effects of environmental pollutants. Analyses of allozyme frequencies may therefore act as a sensitive population-level biomarker for exposure to contaminants in aquatic populations. The objective of this study was to test the hypotheses that allozyme-genotype frequencies in populations of fishes differ significantly among sites with varying water quality and are correlated with water quality.
INTRODUCTION
MATERIALS A N D M E T H O D S
Allozymes, or alleles at the same enzyme-coding locus, have been used to quantify genetic variability in many populations of organisms. Recently, studies have shown that aquatic animals with different allozyme genotypes vary in their sensitivity to the toxic effects of contaminants (Diamond et al., 1989; Chagnon & Guttman, 1989). Differential tolerance to mercury was linked to varying genotypes of allozymes in shrimp (Nevo et al., 1981) and mosquitofish (Diamond et al., 1989). Similarly, enzyme polymorphism in marine invertebrates may allow for adaptations to the toxic effects of copper and zinc (Lavie & Nevo, 1982), as well as detergents and crude-oil-surfactant mixtures (Lavie et al., 1984). Differential tolerance to the toxic effects of contaminants may result in changes of population
Dicks Creek, near Middletown, Ohio, receives effluent from a steel-production plant at several locations (Fig. 1). Chemical and physical parameters at four sites on Dicks Creek were obtained from environmental staff at the steel-production facility (ARMCO, unpublished) and from the Ohio EPA (unpublished). These data indicated that water quality (thirteen parameters) varied among the four sites in Dicks Creek. The spotfin shiner Notropis spilopteris was chosen as a study species because of its abundance at all sites in Dicks Creek and because electrophoretic analyses showed evidence of allozyme variability for several enzymes. Shiners were collected with seine nets from four sites in Dicks Creek (n = 431; mean total length = 58-6 + 8-8 mm) during October 1987, March 1988, and April 1988. All fish were placed individually in plastics bags and stored at -70°C. Individuals were
Abstract
Environ. Pollut. 0269-7491/93/$06.00 © 1993 Elsevier Science Publishers Ltd, England. Printed in Great Britain 147
148
R. B. Gillespie, S. I. Guttman
total number of alleles in the population. Genotype frequencies (percentage of total) were calculated simply by dividing the number of individuals with a particular genotype by the total number of individuals collected from that site. Allele and genotype frequencies were analyzed by chi-square contingency tests by using the statistical package Statview for the Macintosh computer. Correlation coefficients were calculated (Cricket Graph for Macintosh) to determine the strength of the relationship between genotype frequencies and waterquality parameters in Dicks Creek.
.._ ~ i . ~ c k s Creek
RESULTS
,%
Fig. 1. Locations of sampling sites for spotfin shiners and water-quality parameters on Dicks Creek, Middletown, Ohio, USA. homogenized on ice with an extraction solution of 2% 2-phenoxyethanol and transferred to centrifuge tubes. Homogenates were centrifuged at about 50 000 g for 30 min at 0°C, and the resulting supernatant was stored at -70°C. Horizontal starch-gel (13-15%) electrophoretic techniques and histochemical analyses were used for the resolution of genetic loci (Harris & Hopkinson, 1976). An amino-morpholine (Clayton & Tretiak, 1972; pH 6-1) gel was used to resolve glucose-6-phosphate isomerase 5.3.1.9 (GPI-2) and malate dehydrogenase 1.1.1.37 (MDH-1 and MDH-2). Additionally, a Clayton-Tretiak gel and a tris-citric acid (pH 8.5) gel were used to resolve GPI-1. Banding patterns were interpreted by labelling alleles of fastest migration 'A' and slower-migrating alleles as 'B', etc., for each enzyme system (Fig. 2). Isozyme loci were labelled as 1 for the fastest-migrating locus and 2 for the slower-migrating locus. Allele frequencies (percentage of total) were calculated by dividing the number of specific alleles for each enzyme locus by the
Two enzymes coded by four loci were polymorphic on spotfin shiners: glucose-6-phosphate isomerase (GPI-1, GPI-2,) and malate dehydrogenase (MDH-1, MDH-2). Frequencies of alleles and genotypes of GPI-1 and MDH-1,2 did not differ significantly (P > 0.05) among populations of spotfins and will not be discussed further. Of four GPI-2 genotypes characterized, GPI-2 AA, AB, and BB comprised from 85 to 97% of all individuals. For statistical comparisons, the remaining genotypes of GPI-2 (3-15%) from each site were therefore grouped into an 'other' category and analyzed as one group. The frequencies of alleles and genotypes of GPI-2 did not differ significantly (P > 0.05) between 1987 and 1988 at sites DC2 and DC6, where collections were made in both years. Data from both years were 142
151
76
62
70
v
~
~r
50 GPI-2 A
30
Others
l0 Q
60
m, . , Q , l ~ , I R a
A
wl °.(.~iole0
BC
B
o
•
,.0 AC
50 ~v
40
o
30 LL
20
a
10
Allele A Allele B
1
I
I
l
l
/
/
/
/
/
DC2
DC6
DCIO
DC12
Site
Gen0type
AA
BB
AB
AA
AA
BB
AB
AA
Fig. 2. Diagrammatic representation of allele and genotype
characterizations for a 'typical' enzyme detected by electrophoresis. Each band represents an individual genotype at a single enzyme-coding locus. Frequencies are calculated by determining the percentage occurrence of a specific allele or enzyme genotype in the population.
Fig. 3. Allele and genotype frequencies of GPI-2 in spotfin shiners from Dicks Creek (DC2-DC12). Allele frequencies did not differ significantly among sites (P > 0.05). Numbers above allele frequencies are the sample sizes of fish collected from each site. For genotype frequencies, similar letters among sites indicate no significant difference in frequency (P > 0-05).
Water quality and frequencies of allozyme genotypes
149
Table 1. Mean water quality data for four sites in Dicks Creek a
Site
DC2 DC6 DC10 DC12
Water-quality parameterb T °C
pH
20 24 25 22
7.4 8.1 7.9 7.7
Hard NH 3 N-Tot Phos Zn Fe Pb mg/litre mg/litre mg/litre mg/litre mg/litre mg/litre mg/litre 714 680 507 542
0-001 0.05 0.27 0.02
0.47 1.15 1.90 1.65
0.06 0.27 0.27 0.23
10 64 70 66
222 522 390 407
4.0 7.2 7.0 6.7
0.96
0.99
0.99
Correlations with GPI-2 BB frequency r
0.69
0.80
0.32
0.16
0-64
0.97
a Means of two samples collected during September 1987 (Ohio EPA, unpublished). b T -- temperature, Hard = hardness as CaCO3, N-Tot = total nitrogen, Phos -- total phosphorus. r = correlation coefficient. therefore combined for these sites prior to statistical analyses. Allozyme frequencies
Although a trend is apparent, allele frequencies of GPI-2 B and GPI-2 A did not differ significantly (P > 0-05) among any of the sites in Dicks Creek (Fig. 3). Frequencies of individuals with the genotype GPI-2 BB decreased between DC2 and DC12 (Fig. 3). The frequency of individuals with the genotype GPI-2 AA and AB increased between DC2 and DC12. Genotype frequencies differed significantly (P < 0.05) between spotfins from DC2 and DC6, between those from DC2 and DC10, and between those from DC10 and DCI2. Water quality
With the exception of hardness, water quality was better at site DC2 in Dicks Creek than sites downstream of DC2 (Table 1). This was expected because DC2 receives only one effluent source, whereas sites downstream receive three or four additional effluent discharges (Fig. 1). Frequencies of spotfin shiners with the genotype GPI-2 BB were negatively correlated (r = 0.69-0.99) with levels of contaminants in Dicks Creek (Table 1). The strongest correlations were found for phosphorus, zinc, iron, and lead. DISCUSSION Allele and genotype frequencies of GPI-2 in spotfin shiners showed patterns that may have been associated with differences in water quality in Dicks Creek. Frequencies of allozyme genotypes of GPI-2 in spotfin shiner populations differed significantly among sites in Dicks Creek. Furthermore, allozyme frequencies of GPI-2 BB were correlated with water-quality parameters. These results suggest that contaminants may have selected against sensitive genotypes in spotfin shiners and that changes in proportions of allozyme genotypes in aquatic populations are sensitive biological indicators of changes in water quality. Differences in allozyme frequencies of populations of fishes occur owing to differences in variability of
habitat and geographic isolation (Hedrick et al., 1976; Smith et al., 1983). Because genotypic frequencies in spotfin shiners showed a pattern that was correlated with concentrations of contaminants, differences in genotypic frequencies among shiner populations in Dicks Creek may be due to water-quality-induced selection against sensitive genotypes rather than stochastic processes. We were unable to establish a cause-and-effect relationship between specific contaminants in Dicks Creek and differences in allozyme frequencies of spotfin shiners. Although concentrations of some metals (Zn, Fe, Pb) were higher at downstream sites than upstream sites, associated changes in genotype frequencies were only circumstantial. However, fishes with certain allozyme genotypes have been shown to be more sensitive to the toxic effects of metals than those with other genotypes (Chagnon & Guttman, 1989; Diamond et al., 1989; Gillespie & Guttman, 1989). Individuals with certain allozyme genotypes may therefore be more sensitive to the toxic effects of specific contaminants and complex effluents than fish with other genotypes. We do not know whether the mechanism of selection against individuals with sensitive enzyme genotypes is directed towards the specific enzyme or if the selection is associated with other biochemical processes. However, differential survivorship has been associated with different allozyme genotypes of GPI in marine snails exposed to Zn, Cu, Cd, and organics (Lavie & Nevo, 1982, 1986; Lavie et al., 1984). Additionally, the time to death in mosquitofish exposed to Hg was significantly affected by their allozyme genotype of GPI (Diamond et al., 1989). The sensitivity of GPI to the toxic effects of contaminants has been shown for a variety of fish and invertebrate taxa; this suggests that contaminant-induced selection is directed at the GPI locus. Other evidence indicates that contaminant-induced selection at the GPI locus may explain differences in senstivity of animals to toxic effects of contaminants. Inhibition of phosphoglucomutase by heavy metals has been attributed to competition with magnesium (Milstein, 1961). Because different aUozymes of GPI may vary in
150
R. B. Gillespie, S. I. Guttman
their catalytic activity (Gottlieb & Greve, 1981), contaminant-induced mortality of sensitive individuals could be mediated through these same mechanisms at the GPI-2 locus (Lavie & Nevo, 1982; Lavie et al., 1984). Current methods of biological monitoring for water quality involve observations of resident populations of aquatic biota. Because allele and genotype frequencies were correlated with water-quality parameters, assessments of allozyme frequencies in aquatic populations may be sensitive indicators and useful predictors of the impact of reduced water quality on aquatic populations. To determine its usefulness, more research is needed to correlate specific water-quality parameters with changes in allozyme frequencies of allozyme genotypes in aquatic populations. Furthermore, laboratory experiments are needed to determine the mechanisms of contaminant-induced mortality of sensitive allozyme genotypes. Finally, studies are needed to determine the impact of reduce allozyme variability on biological fitness in aquatic populations. Because frequencies of allozyme genotypes may change prior to more significant impacts (e.g. reduced reproduction and increased acute toxicity), monitoring these allozyme frequencies in natural populations may be a more sensitive biological indicator of reduced water quality than those currently in use. ACKNOWLEDGEMENTS We thank A R M C O Steel Inc. and Ohio EPA for providing us with water-quality data on Dicks Creek. REFERENCES
Chagnon, N. L. & Guttman, S. I. (1989). Differential survivorship of allozyme genotypes in mosquitofish popu-
lations exposed to copper or cadmium. Environ. Toxicol. Chem., 8, 319-26. Clayton, J. W. & Tretiak, D. N. (1972). Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Bd Can., 29, 1169-72. Diamond, S. A., Newman, M. C., Mulvey, M., Dixon, P. M. & Martinson, D. (1989). Allozyme genotypes and time to death of mosquitofish, Gasmbusia affinis (Baird and Giraard), during acute exposure to inorganic mercury. Environ. Toxicol. Chem., 8, 613-22. Gillespie, R. B. & Guttman, S. I. (1989). Effects of contaminants on the frequencies of allozymes in populations of the central stoneroller. Environ. Toxicol. Chem., 8, 309-17. Gottlieb, L. D. & Greve, L. C. (1981). Biochemical properties of duplicated isozymes of phosphoglucose isomerase in the plant Clarkia xantia. Biochem. Gen., 19, 155-72. Harris, H. & Hopkinson, D. A. (1976). Handbook of Enzyme Electrophoresis in Human Genetics. North Holland Publishing Company, Amsterdam, the Netherlands. Hedrick, P. W., Ginevan, M. E. & Ewing, E. P. (1976). Genetic polymorphism in heterogenous environments. Annu. Rev. Ecol. Syst., 7, 1-32. Lavie, B. & Nevo, E. (1982). Heavy metal selection of phosphoglucose isomerase allozymes in marine gastropods. Mar. BioL, 71, 17-22. Lavie, B. & Nevo, E. (1986). The interactive effects of cadmium and mercury pollution on allozyme polymorphisms in the marine gastropod Cerithium scabridium. Mar. Pollut. Bull., 17, 21-3. Lavie, B., Nevo, E. & Zoller, Y. (1984). Differential viability of phosphoglucose isomerase allozyme genotypes of marine snails in nonionic detergent and crude oil-surfactant mixtures. Environ. Res., 35, 270q5. Milstein, C. (1961). On the mechanism of activation of phosphoglucomutase by metal ions. Biochem. J., 79, 574-84. Nevo, E., Perl, T., Beiles, A. & Wood, D. (1981). Mercury selection of allozyme genotypes in shrimps. Experientia, 37, 1152-4. Smith, M. W., Smith, M. H. & Chesser, R. K. (1983). Biochemical genetics of mosquitofish. I. Environmental correlates, and temporal and spalial heterogeneity of allele frequencies within a river drainage. Copeia, 1983, 182-93.
CORRELATIONS BETWEEN WATER QUALITY A N D FREQUENCIES OF ALLOZYME GENOTYPES IN SPOTFIN SHINER (Notropis spilopteris) POPULATIONS R. B. Gillespie Department of Biological Sciences, Indiana University-Purdue University at Fort Wayne, Indiana 46805-1499, USA
&
S. I. Guttman Department of Zoology, Miami University, Oxford, Ohio 45056, USA (Received 12 February 1992; accepted 22 May 1992)
Genotype frequencies of glucose-6-phosphate isomerase (GPI) allozymes differed significantly among populations of the spotfin shiner Notropis spilopterus from sites with varying water quality. Frequencies of shiners with a GPI-2 BB genotype decreased significantly at sites with reduced water quality. Alternatively, the total frequencies of shiners with a genotype of GPI-2 AA and AB increased at sites with reduced water quality. Individuals with certain allozyme genotypes may be more sensitive to the toxic effects of polluted waters than those with other genotypes. The selection of individuals with sensitive genotypes may reduce genetic diversity in populations and thus increase the susceptibility of these populations to additional novel stresses. Because allele and genotype frequencies of GPI-2 were correlated with water quality, electrophoretic determination of genetic structure in fishes may be a useful tool for monitoring the health of aquatic populations.
genetic characteristics in wild populations. Differences in allozyme frequencies in populations of the central stoneroller were associated with different levels of exposure to contaminants in a stream receiving non-point sources of run-off from a uranium-enrichment facility (Gillespie & Guttman, 1989). The genetic diversity of a population may therefore moderate the toxic effects of pollutants and provide a mechanism of population survival. Assessments of allozyme composition in aquatic populations may provide a sensitive end-point for the effects of environmental pollutants. Analyses of allozyme frequencies may therefore act as a sensitive population-level biomarker for exposure to contaminants in aquatic populations. The objective of this study was to test the hypotheses that allozyme-genotype frequencies in populations of fishes differ significantly among sites with varying water quality and are correlated with water quality.
INTRODUCTION
MATERIALS A N D M E T H O D S
Allozymes, or alleles at the same enzyme-coding locus, have been used to quantify genetic variability in many populations of organisms. Recently, studies have shown that aquatic animals with different allozyme genotypes vary in their sensitivity to the toxic effects of contaminants (Diamond et al., 1989; Chagnon & Guttman, 1989). Differential tolerance to mercury was linked to varying genotypes of allozymes in shrimp (Nevo et al., 1981) and mosquitofish (Diamond et al., 1989). Similarly, enzyme polymorphism in marine invertebrates may allow for adaptations to the toxic effects of copper and zinc (Lavie & Nevo, 1982), as well as detergents and crude-oil-surfactant mixtures (Lavie et al., 1984). Differential tolerance to the toxic effects of contaminants may result in changes of population
Dicks Creek, near Middletown, Ohio, receives effluent from a steel-production plant at several locations (Fig. 1). Chemical and physical parameters at four sites on Dicks Creek were obtained from environmental staff at the steel-production facility (ARMCO, unpublished) and from the Ohio EPA (unpublished). These data indicated that water quality (thirteen parameters) varied among the four sites in Dicks Creek. The spotfin shiner Notropis spilopteris was chosen as a study species because of its abundance at all sites in Dicks Creek and because electrophoretic analyses showed evidence of allozyme variability for several enzymes. Shiners were collected with seine nets from four sites in Dicks Creek (n = 431; mean total length = 58-6 + 8-8 mm) during October 1987, March 1988, and April 1988. All fish were placed individually in plastics bags and stored at -70°C. Individuals were
Abstract
Environ. Pollut. 0269-7491/93/$06.00 © 1993 Elsevier Science Publishers Ltd, England. Printed in Great Britain 147
148
R. B. Gillespie, S. I. Guttman
total number of alleles in the population. Genotype frequencies (percentage of total) were calculated simply by dividing the number of individuals with a particular genotype by the total number of individuals collected from that site. Allele and genotype frequencies were analyzed by chi-square contingency tests by using the statistical package Statview for the Macintosh computer. Correlation coefficients were calculated (Cricket Graph for Macintosh) to determine the strength of the relationship between genotype frequencies and waterquality parameters in Dicks Creek.
.._ ~ i . ~ c k s Creek
RESULTS
,%
Fig. 1. Locations of sampling sites for spotfin shiners and water-quality parameters on Dicks Creek, Middletown, Ohio, USA. homogenized on ice with an extraction solution of 2% 2-phenoxyethanol and transferred to centrifuge tubes. Homogenates were centrifuged at about 50 000 g for 30 min at 0°C, and the resulting supernatant was stored at -70°C. Horizontal starch-gel (13-15%) electrophoretic techniques and histochemical analyses were used for the resolution of genetic loci (Harris & Hopkinson, 1976). An amino-morpholine (Clayton & Tretiak, 1972; pH 6-1) gel was used to resolve glucose-6-phosphate isomerase 5.3.1.9 (GPI-2) and malate dehydrogenase 1.1.1.37 (MDH-1 and MDH-2). Additionally, a Clayton-Tretiak gel and a tris-citric acid (pH 8.5) gel were used to resolve GPI-1. Banding patterns were interpreted by labelling alleles of fastest migration 'A' and slower-migrating alleles as 'B', etc., for each enzyme system (Fig. 2). Isozyme loci were labelled as 1 for the fastest-migrating locus and 2 for the slower-migrating locus. Allele frequencies (percentage of total) were calculated by dividing the number of specific alleles for each enzyme locus by the
Two enzymes coded by four loci were polymorphic on spotfin shiners: glucose-6-phosphate isomerase (GPI-1, GPI-2,) and malate dehydrogenase (MDH-1, MDH-2). Frequencies of alleles and genotypes of GPI-1 and MDH-1,2 did not differ significantly (P > 0.05) among populations of spotfins and will not be discussed further. Of four GPI-2 genotypes characterized, GPI-2 AA, AB, and BB comprised from 85 to 97% of all individuals. For statistical comparisons, the remaining genotypes of GPI-2 (3-15%) from each site were therefore grouped into an 'other' category and analyzed as one group. The frequencies of alleles and genotypes of GPI-2 did not differ significantly (P > 0.05) between 1987 and 1988 at sites DC2 and DC6, where collections were made in both years. Data from both years were 142
151
76
62
70
v
~
~r
50 GPI-2 A
30
Others
l0 Q
60
m, . , Q , l ~ , I R a
A
wl °.(.~iole0
BC
B
o
•
,.0 AC
50 ~v
40
o
30 LL
20
a
10
Allele A Allele B
1
I
I
l
l
/
/
/
/
/
DC2
DC6
DCIO
DC12
Site
Gen0type
AA
BB
AB
AA
AA
BB
AB
AA
Fig. 2. Diagrammatic representation of allele and genotype
characterizations for a 'typical' enzyme detected by electrophoresis. Each band represents an individual genotype at a single enzyme-coding locus. Frequencies are calculated by determining the percentage occurrence of a specific allele or enzyme genotype in the population.
Fig. 3. Allele and genotype frequencies of GPI-2 in spotfin shiners from Dicks Creek (DC2-DC12). Allele frequencies did not differ significantly among sites (P > 0.05). Numbers above allele frequencies are the sample sizes of fish collected from each site. For genotype frequencies, similar letters among sites indicate no significant difference in frequency (P > 0-05).
Water quality and frequencies of allozyme genotypes
149
Table 1. Mean water quality data for four sites in Dicks Creek a
Site
DC2 DC6 DC10 DC12
Water-quality parameterb T °C
pH
20 24 25 22
7.4 8.1 7.9 7.7
Hard NH 3 N-Tot Phos Zn Fe Pb mg/litre mg/litre mg/litre mg/litre mg/litre mg/litre mg/litre 714 680 507 542
0-001 0.05 0.27 0.02
0.47 1.15 1.90 1.65
0.06 0.27 0.27 0.23
10 64 70 66
222 522 390 407
4.0 7.2 7.0 6.7
0.96
0.99
0.99
Correlations with GPI-2 BB frequency r
0.69
0.80
0.32
0.16
0-64
0.97
a Means of two samples collected during September 1987 (Ohio EPA, unpublished). b T -- temperature, Hard = hardness as CaCO3, N-Tot = total nitrogen, Phos -- total phosphorus. r = correlation coefficient. therefore combined for these sites prior to statistical analyses. Allozyme frequencies
Although a trend is apparent, allele frequencies of GPI-2 B and GPI-2 A did not differ significantly (P > 0-05) among any of the sites in Dicks Creek (Fig. 3). Frequencies of individuals with the genotype GPI-2 BB decreased between DC2 and DC12 (Fig. 3). The frequency of individuals with the genotype GPI-2 AA and AB increased between DC2 and DC12. Genotype frequencies differed significantly (P < 0.05) between spotfins from DC2 and DC6, between those from DC2 and DC10, and between those from DC10 and DCI2. Water quality
With the exception of hardness, water quality was better at site DC2 in Dicks Creek than sites downstream of DC2 (Table 1). This was expected because DC2 receives only one effluent source, whereas sites downstream receive three or four additional effluent discharges (Fig. 1). Frequencies of spotfin shiners with the genotype GPI-2 BB were negatively correlated (r = 0.69-0.99) with levels of contaminants in Dicks Creek (Table 1). The strongest correlations were found for phosphorus, zinc, iron, and lead. DISCUSSION Allele and genotype frequencies of GPI-2 in spotfin shiners showed patterns that may have been associated with differences in water quality in Dicks Creek. Frequencies of allozyme genotypes of GPI-2 in spotfin shiner populations differed significantly among sites in Dicks Creek. Furthermore, allozyme frequencies of GPI-2 BB were correlated with water-quality parameters. These results suggest that contaminants may have selected against sensitive genotypes in spotfin shiners and that changes in proportions of allozyme genotypes in aquatic populations are sensitive biological indicators of changes in water quality. Differences in allozyme frequencies of populations of fishes occur owing to differences in variability of
habitat and geographic isolation (Hedrick et al., 1976; Smith et al., 1983). Because genotypic frequencies in spotfin shiners showed a pattern that was correlated with concentrations of contaminants, differences in genotypic frequencies among shiner populations in Dicks Creek may be due to water-quality-induced selection against sensitive genotypes rather than stochastic processes. We were unable to establish a cause-and-effect relationship between specific contaminants in Dicks Creek and differences in allozyme frequencies of spotfin shiners. Although concentrations of some metals (Zn, Fe, Pb) were higher at downstream sites than upstream sites, associated changes in genotype frequencies were only circumstantial. However, fishes with certain allozyme genotypes have been shown to be more sensitive to the toxic effects of metals than those with other genotypes (Chagnon & Guttman, 1989; Diamond et al., 1989; Gillespie & Guttman, 1989). Individuals with certain allozyme genotypes may therefore be more sensitive to the toxic effects of specific contaminants and complex effluents than fish with other genotypes. We do not know whether the mechanism of selection against individuals with sensitive enzyme genotypes is directed towards the specific enzyme or if the selection is associated with other biochemical processes. However, differential survivorship has been associated with different allozyme genotypes of GPI in marine snails exposed to Zn, Cu, Cd, and organics (Lavie & Nevo, 1982, 1986; Lavie et al., 1984). Additionally, the time to death in mosquitofish exposed to Hg was significantly affected by their allozyme genotype of GPI (Diamond et al., 1989). The sensitivity of GPI to the toxic effects of contaminants has been shown for a variety of fish and invertebrate taxa; this suggests that contaminant-induced selection is directed at the GPI locus. Other evidence indicates that contaminant-induced selection at the GPI locus may explain differences in senstivity of animals to toxic effects of contaminants. Inhibition of phosphoglucomutase by heavy metals has been attributed to competition with magnesium (Milstein, 1961). Because different aUozymes of GPI may vary in
150
R. B. Gillespie, S. I. Guttman
their catalytic activity (Gottlieb & Greve, 1981), contaminant-induced mortality of sensitive individuals could be mediated through these same mechanisms at the GPI-2 locus (Lavie & Nevo, 1982; Lavie et al., 1984). Current methods of biological monitoring for water quality involve observations of resident populations of aquatic biota. Because allele and genotype frequencies were correlated with water-quality parameters, assessments of allozyme frequencies in aquatic populations may be sensitive indicators and useful predictors of the impact of reduced water quality on aquatic populations. To determine its usefulness, more research is needed to correlate specific water-quality parameters with changes in allozyme frequencies of allozyme genotypes in aquatic populations. Furthermore, laboratory experiments are needed to determine the mechanisms of contaminant-induced mortality of sensitive allozyme genotypes. Finally, studies are needed to determine the impact of reduce allozyme variability on biological fitness in aquatic populations. Because frequencies of allozyme genotypes may change prior to more significant impacts (e.g. reduced reproduction and increased acute toxicity), monitoring these allozyme frequencies in natural populations may be a more sensitive biological indicator of reduced water quality than those currently in use. ACKNOWLEDGEMENTS We thank A R M C O Steel Inc. and Ohio EPA for providing us with water-quality data on Dicks Creek. REFERENCES
Chagnon, N. L. & Guttman, S. I. (1989). Differential survivorship of allozyme genotypes in mosquitofish popu-
lations exposed to copper or cadmium. Environ. Toxicol. Chem., 8, 319-26. Clayton, J. W. & Tretiak, D. N. (1972). Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Bd Can., 29, 1169-72. Diamond, S. A., Newman, M. C., Mulvey, M., Dixon, P. M. & Martinson, D. (1989). Allozyme genotypes and time to death of mosquitofish, Gasmbusia affinis (Baird and Giraard), during acute exposure to inorganic mercury. Environ. Toxicol. Chem., 8, 613-22. Gillespie, R. B. & Guttman, S. I. (1989). Effects of contaminants on the frequencies of allozymes in populations of the central stoneroller. Environ. Toxicol. Chem., 8, 309-17. Gottlieb, L. D. & Greve, L. C. (1981). Biochemical properties of duplicated isozymes of phosphoglucose isomerase in the plant Clarkia xantia. Biochem. Gen., 19, 155-72. Harris, H. & Hopkinson, D. A. (1976). Handbook of Enzyme Electrophoresis in Human Genetics. North Holland Publishing Company, Amsterdam, the Netherlands. Hedrick, P. W., Ginevan, M. E. & Ewing, E. P. (1976). Genetic polymorphism in heterogenous environments. Annu. Rev. Ecol. Syst., 7, 1-32. Lavie, B. & Nevo, E. (1982). Heavy metal selection of phosphoglucose isomerase allozymes in marine gastropods. Mar. BioL, 71, 17-22. Lavie, B. & Nevo, E. (1986). The interactive effects of cadmium and mercury pollution on allozyme polymorphisms in the marine gastropod Cerithium scabridium. Mar. Pollut. Bull., 17, 21-3. Lavie, B., Nevo, E. & Zoller, Y. (1984). Differential viability of phosphoglucose isomerase allozyme genotypes of marine snails in nonionic detergent and crude oil-surfactant mixtures. Environ. Res., 35, 270q5. Milstein, C. (1961). On the mechanism of activation of phosphoglucomutase by metal ions. Biochem. J., 79, 574-84. Nevo, E., Perl, T., Beiles, A. & Wood, D. (1981). Mercury selection of allozyme genotypes in shrimps. Experientia, 37, 1152-4. Smith, M. W., Smith, M. H. & Chesser, R. K. (1983). Biochemical genetics of mosquitofish. I. Environmental correlates, and temporal and spalial heterogeneity of allele frequencies within a river drainage. Copeia, 1983, 182-93.