Population genetics of bidentatus (: ): the effect of planktonic development on gene flow

S. W. Schaeffer 1, E. C. Kellerl, Jr. & N. E. Buroker2 i Dept. of Biology, West Virginia University, Morgantown, WV26506, USA 2 Center for Environmental and Estuarine Studies, University of Maryland, Crisfield, Md, USA

Abstract

Melampus bidentatus has restricted gene flow of gametes and the potential for extensive gene flow of their planktonic larvae. We used genetic data to examine the influence of these opposing forces on the population structure of six Delmarva Peninsular populations. We analyzed the population structure of these six populations using twenty-three electrophoretic loci. Two ocean populations gave the appearance of large panmictic populations, while the other four populations appeared to be highly subdivided due to either restricted gene flow or age-structuring of the population. Two distinct groups of populations are apparent, which accounts for the extreme interpopulation heterogeneity. These data suggest that the direction of tides and currents limit the effectiveness of the planktonic dispersal stage of M. bidentatus in overcoming restricted gene flow of gametes.

Introduction out into the offshore plankton layer with the ebbing tide. The larvae spend two weeks developing into Successful gene flow requires dispersal of gam- adults in the plankton layer. Following planktonic etes or newly formed zygotes to a new popu- development, the adults settle into the salt marsh lation and subsequent establishment of these gam- where they reside for their remaining two to three etes or zygotes (Endler, 1977; Scheltema, 1975). years of life (Russell-Hunter et al., 1972). McCracken & Bussard (1981) showed that re- The terrestrial existence of M. bidentatus should stricted gene flow in terrestrial snails increased the increase genetic differences between populations, genetic differences between populations. Other but a planktonic larval stage has the potential to sedentary organisms show similar patterns (Levin & reduce these genetic differences. In this paper we Kerster, 1974; Schaal, 1975; Workman & Nis- examine the influence of these two opposing forces wander, 1970). On the other hand, extensive gene on the population structure of natural populations flow in marine invertebrates reduces genetic differ- of M. bidentatus. ences between populations (Buroker et al., 1979; Koehn et al.. 1976). These experimental results are Material and methods consistent with Wright's (1943) isolation-by-dis- tance model. We sampled six natural populations ofM. biden- Melampus bidentatus (Say) is a pulmonate snail tatus located on the peninsula formed by Delaware, which lives above the average high-tide mark of Maryland and Virginia (Delmarva Peninsula). coastal and estuarine salt marshes along the Atlan- Three Atlantic Ocean localities (Fenwick Island, tic Ocean and Gulf of Mexico (Holle & Dineen, Delaware; Wallops Island, Virginia; and Cape 1957). Hatching M. bidentatus larvae are washed Charles, Virginia) and three Chesapeake Bay sites

Genetica 66, 223-229 (1985). © Dr W. Junk Publishers, Dordrecht. Printed in the Netherlands. 224

phoglucose isomerase (pgi), phosphoglucomutase ~,. Surfacel~ .--P7" Current I'-k,~ .... -~ y" (pgm), and superoxide dismutase (sod). The no- menclature of Ayala et aL (1972) was used to de- scribe our electrophoretic loci. "'Current I I k N We calculated allele frequencies for each poly- morphic locus in each population. A locus was considered polymorphic if the frequency of the most common allele in any of the six populations sampled was less than or equal to 0.95. An analysis of population structure compares the observed gene and genotypic frequencies with those expected under the assumption of random mating. Depar- tures from random mating suggest that selection, inbreeding or other evolutionary forces are acting I ..... on the individuals of the population. We tested departures of genotypic frequencies from the ex- pected proportions under random mating by means of chi-squared tests for goodness-of-fit. A chi- Chesapeake - f squared test for homogeneity or gene frequency was B j, c v , ,/. Y. ' . used to detect departures from random mating be- \ tween populations (Schaal, 1975; Workman & Nis- wander, 1970). Genetic similarities and genetic dis- tances for all pairwise comparisons were calculated using Nei's (1972) formulas. We clustered Nei's ccV# / genetic distances using the UPGMA algorithm of Sheath & Sokal (1973).

Results The six populations were monomorphic for the Fig. 1. A map of the Delmarva Peninsula showing the six Me- same electrophoretic allele at thirteen enzyme loci lampus bidentatus sampling sites. The map also shows the cur- of the twenty-three loci scored for M. bidentatus. rents along the Atlantic coast and in the Chesapeake Bay (after Scheletema, 1975). The six populations are: Fenwick Island Ten loci examined were polymorphic and in some (FI), Wallops Island (Wl), Cape Charles (CC), Eastville (EV), cases highly so with six electromorphs or more. The Cashville (CV), and Crisfield (CF). polymorphic loci were: acp-2, ald, ap, est-2, idh, !ap-1, lap-2, lap-3, 6-pgd, and sod. Our estimate of (Eastville, Virginia; Cashville, Virginia; and Cris- percent polymorphic loci, P = 0.435 + 0.073, was field, Maryland) were selected (Fig. 1). We col- within the range observed for other marine and lected 125 snails from each population and stored land snails (Nevo, 1978; Selander, 1976). The aver- them at -20 o C for electrophoresis. Salinity meas- age individual heterozygosity, H = 0.196 + 0.005, urements were taken at each site. Table 1. Average salinity ± standard error for six Delmarva Fourteen enzyme systems involving 23 loci were Peninsular sampling sites. Salinity is given in parts per thousand. examined using horizontal starch gel electrophore- sis (see Schaeffer, 1980). We stained for the follow- Population Salinity ± SE ing enzymes: acid phosphatase (acp), aldolase (aid), Fenwick Island 26.0 + 2.5 aminopeptidase (ap), aspartate aminotransferase Wallops Island 36.9 + 4.3 (aat), esterase (est), isocitrate dehydrogenase (idh), Cape Charles 31.7 + 6.0 leucine aminopeptidase (lap), malate dehydrogen- Eastville 20.6 ± 1.3 Cashville 19.4 + 1.5 ase (mdh), malic enzyme (me), muscle protein (rap), Crisfield 18.3 ± 3.5 6-phosphogluconate dehydrogenase (6-pgd), phos- 225 was greater than that reported in land or marine ville, Cashville, and Crisfield populations. These snails. Salinities for each site are given in Table !. populations had a deficiency of heterozygotes at We observed significant departures of our ob- most polymorphic loci. The other two populations, served genotypic frequencies from those expected Wallops Island and Cape Charles, exhibited few under random mating in the Fenwick Island, East- significant departures (Table 2).

Table 2. Gene frequencies, number of alleles (N), relative mobilities (R M), observed heterozygosity (HO), and expected heterozygosity (HE) for six Delmarva Peninsular populations of Melampus bidentatus.

Locus RM FI WI CC EV CV CF a~H-2 N 250*** 250 248 246*** 248 250 100 0.120 0.976 0.952 0.439 0.040 0.028 85 0.880 0.024 0.048 0.561 0.960 0.972 HO 0.016 0.048 0.097 0.016 0.048 0.056 HE 0.211 0.047 0.092 0.493 0.077 0.054 aM N 248 250 246 248 248 250*** 100 0.956 1.000 1.000 0.130 0.024 0.120 98 0.044 0.870 0.976 0.876 96 0.004 HO 0.056 0.195 0.056 0.136 HE 0.085 0.226 0.047 0.218 ap N 250*** 248 250 246*** 248 250*** 108 0,020 0.004 0.008 104 0.056 0.081 0.084 0.020 0.008 I00 0.132 0.746 0.680 0.183 0.004 96 0.772 0.161 0.200 0.138 0.032 0.016 92 0.020 0.008 0.024 0.114 0.109 0.084 88 0.004 0.455 0.802 0.856 84 0.073 0.044 0.028 80 0.004 0.016 78 0.012 HO 0.240 0.403 0.440 0.480 0.315 0.184 HE 0.383 0.411 0.490 0.721 0.341 0.259 est-2 N 248*** 248 244 246*** 248*** 250 109 0.020 0.024 0.073 0.053 107 0.012 0.125 0.115 0.053 0.016 0.004 105 0.024 0.246 0.139 0.053 103 0.036 0.218 0.152 0.061 0.065 0.044 102 0.036 0.069 0.119 0.049 100 0.032 0.238 0.217 0.045 0.028 98 0.024 0.048 0.078 0.118 0.085 0.076 97 0.060 0.016 0.045 0.045 96 0.016 95 0.141 0.016 0.020 0.045 0.351 0.088 93 0.185 0.025 0.159 0.073 0.244 91 0.153 0.008 0.102 0.113 0.228 89 0.052 0.077 0.016 0.024 87 0.125 0.004 0.098 0.129 0.212 85 0.057 0.024 0.008 0.044 83 0.028 0.004 0.016 0.089 0.028 79 0.012 0.004 0.012 0.008 HO 0.806 0.669 0.828 0.846 0.661 0.816 HE 0.891 0.812 0.868 0.914 0.821 0.825 idh N 248*** 250 250 244*** 244* 250 102 0.004 0.008 0.004 0.028 100 0.855 1.000 1.000 0.873 0.902 0.772 98 0.137 0.115 0.094 0.200 226

Table 2. Continued.

Locus RM FI Wl CC EV CV CF

96 0,004 0.004 HO 0.153 0.057 0.148 0.280 HE 0.025 0.225 0.178 0.363 lap-I N 240 - 230 232 214"** 244*** 102 0.146 0.135 0.177 0.098 0.123 100 0.604 0.683 0.603 0.505 0,594 98 0.250 0.182 0.220 0,397 0.283 HO 0.525 0.461 0.526 0,449 0.533 HE 0.551 0.481 0.556 0.578 0.552

lap-2 N 248 224 246* 246*** 248*** 250 112 0.044 0.037 0.028 0.032 I10 0.331 0.138 0.254 0.192 108 0.423 0.289 0.544 0.644 106 0.109 0.085 0,149 0.112 104 0.032 0.054 0.069 0.130 0.020 0,020 102 0.028 0.295 0.256 0,110 0,004 100 0.016 0.527 0.463 0.126 98 0.016 0.103 0.167 0.073 96 0.022 0,045 0.012 HO 0,766 0.580 0,545 0.797 0.694 0.488 HE 0.695 0.621 0.685 0.838 0.616 0.534 lap-3 N 246*** 200 246 240*** 246*** 248 104 0.025 0.057 0.029 102 0.160 0,203 0.150 100 0.705 0.480 0.254 98 0.090 0.195 0.071 96 0.069 0.015 0.065 0.133 0.089 0,065 94 0.191 0,005 0.300 0.211 0.161 92 0,268 0.050 0.394 0.657 90 0.362 0,012 0.293 0.113 88 0.110 0.012 0.004 HO 0.675 0.510 0.691 0.592 0,675 0.565 HE 0.744 0.468 0.683 0.797 0.706 0.525

6-pgd N 216 188 78 218" 168 232 104 0.011 102 0.009 0.032 0.013 0.014 0.012 0.022 100 0.069 0.505 0.090 0.069 0.071 0.060 98 0.829 0.394 0.782 0,743 0,744 0.763 96 0.088 0.058 0,115 0.147 0.167 0,151 94 0.005 0.028 0.006 0.004 HO 0,343 0.617 0.359 0.358 0.440 0.405 HE 0.301 0.585 0.367 0.421 0.413 0.391 sod N 248 196" 192"** 242 246*** 248*** 102 0.036 0.526 0.328 0.021 0.195 0.048 100 0,282 0.464 0.312 0.178 0.398 0.335 98 0,516 0,010 0.297 0,566 0.301 0,460 96 0,125 0.057 0.219 0.093 0.109 94 0,040 0.005 0.017 0.012 0,048 HO 0,645 0.653 0.458 0.686 0,691 0.653 HE 0,635 0.703 0.703 0.599 0.704 0.660

Symbols: Fenwick Island (FI), Wallops Island (Wl), Cape Charles (CC), Eastville (EV), Cashville (CV), Crisfield (CF). Significant departures from genotypic proportions expected under rando m mating are denoted: * P < 0.05; * * - P < 0.01 ; *** - P < 0.00 I. N o data were available for the lap-I locus at Wallops Island which is denoted by (). 227

Table3. Chi-squared test for homogeneity of allele frequencies Table 4. Nei's genetic similarity/genetic distance matrix for all among six Delmarva Peninsular populations of Melampus bi- pairwise comparisons of six Delmarva Peninsular populations dentatus for ten polymorphic enzyme loci. of Melampus bidentatus. Genetic similarities are found above the diagonal; genetic distances are located below the diagonal. Locus Chi-squared* DF Site FI Wl CC EV CV CF aqJ-2 988.427 5 aid 1207.805 5 FI 0.872 0.876 0.897 0.844 0.887 ap 1665.981 25 WI 0.137 0.984 0.875 0.800 0.789 est-2 1178.989 75 CC 0.132 0.016 0.909 0.829 0.825 idh 98.675 5 EV 0.109 0.134 0.095 0.965 0.963 lap-1 33.759 8 CV 0.123 0.233 0.187 0.036 - 0.990 lap-2 1226.581 40 CF 0.120 0.237 0.192 0.038 0.010 lap-3 1558.837 40 6-pgd 264.021 15 Symbols: Fenwick Island (FI), Wallops Island (WI), Cape sod 474.368 20 Charles (CC), Eastville (EV), Cashville (CV), Crisfield (CF).

* All Chi-squared values have a P < 0.0001. in the Crisfield, Cashville, Eastville, and Fenwick Significant genic heterogeneity among popula- Island populations (Table 2). The deficiency of tions was observed at all loci (Table 3). The heterozygotes in these populations suggests that UPGMA cluster of Nei's genetic distances shows planktonic larval dispersal is insufficient to over- two groups of populations, Wallops Island and come the restricted gamete flow of adult snails be- Cape Charles, and Fenwick Island, Eastville, Cash- tween demes. When larvae are washed out of these ville, and Crisfield (Fig. 2). Notice that the Wallops populations the currents may be unidirectional so Island and Cape Charles populations have a genetic that few larvae settle into the marsh where they similarity expected for neighboring demes (Ayala, originated to increase gene flow between local 1975), while the Wallops Island and Crisfield popu- demes. Thus, the Crisfield, Cashville, Eastville, and lations have a genetic similarity corresponding to Fenwick Island populations give the appearance of subspecies. One curious feature is that the Fenwick highly subdivided populations (Wright, 1943). If, Island population shows a greater genetic similarity on the other hand, larvae do return to their origin, to the Crisfield population than the nearby Wallops then what we may be observing is an age-structured Island population (Table 4) (Fig. 2). population with several generations of snails each having different gene frequencies (Charlesworth, 1980). At this time we have no data to reject this Discussion hypothesis but it could be tested by aging snails according to shell size and electrophoresing indi- We found departures of genotypic frequencies viduals in the different size classes. Wallops Island from expected proportions under random mating and Cape Charles appear to be large panmictic populations since few departures from random mating were observed. In these populations, larval CAPE !- CHARLES dispersal appears to be sufficient to overcome the WALLOPS ISLAND effects of restricted gene flow of gametes. CRISFIELD An alternative explanation for the severe hetero- zygote deficiency might be natural selection acting CASHVILLE in a heterogeneous environment (Levene, 1953). EASTVILLE Each salt marsh may be divided into small micro- FENWICK ISLAND geographic environments. Sampling these micro- = t I L t I I / 0.16 014 0.12 010 0.08 0.06 004 0.02 geographic environments as one population will GENETIC DISTANCE (D) give the appearance of a subdivided population if selection is acting differentially on the same geno- Fig. 2. The dendrogram formed by the UPGMA cluster of Nei's types. At this time we have no evidence to support genetic distances for six populations of Melampus bidentatus. or reject this hypothesis. 228

The genic heterogeneity observed among the six Acknowledgements populations suggests that planktonic larvae are not exchanged between ocean and bay populations. We would like to thank the Center for En- The flushing of the Chesapeake Bay may be so vironmental and Estuarine Studies at Crisfield, strong that planktonic larvae settle into marshes Maryland and the Department of Biology at West downstream from their origin site. Also, larvae Virginia University for supplying support for this from oceanic populations rarely penetrate the bay, research. Also, we would like to thank Wyatt W. although some oceanic genotypes are found in the Anderson for useful comments on this manuscript. Eastville population (Table 2) (Schaeffer, 1980). The net effect is that the bay populations are highly subdivided and distinct from the ocean popula- R eferences tions. If currents and tides are responsible for the Ayala, F. J., 1975. Genetic differentiation during the speciation observed pattern in M. bidentatus, then the Fen- process. In: Evolutionary biology 8: 1-78. T. Dobzhansky, wick Island population should be more closely re- M. K. Hecht & W. C. Steere, eds, Plenum Press, New York. lated to the oceanic populations than the bay popu- Ayala, F. L., Powell, J. R., Mourao, G. A. & Perez-Salas, S., 1972. Enzyme variability in the Drosophila willistonigroup. lations. This contradiction suggests that natural IV. Genic variations in populations of Drosophila willistoni. selection or some other evolutionary force is acting Genetics 70:113-139. on these snails. Buroker, N. E., Hershberger, W. K. & Chew, K. K., 1979. Popu- The obvious environmental variable that could lation genetics of the family Ostreidae 1. lntraspecific studies be responsible for the observed cline in gene fre- of Crassostrea gigas and Saccostrea commercialis. Mar. Biol. 54: 157-169. quency in M. bidentatus around the Delmarva Charlesworth, B., 1980. Evolution in age-structured popula- Peninsula is salinity. The two genetic groups corre- tions. Cambridge University Press, New York. late well with the salinity of the environment in Endler, J. A., 1977. Geographic variation, speciation and clines. which they are found (Table 1) (Schaeffer, 1980) Princeton University Press, Princeton, New Jersey. providing indirect evidence that selection is acting Holle, P. A. & Dineen, C. F., 1957. Life history of the salt marsh snail, Melampus bidentatus Say. Nautilus 70: 90-95. on allozyme genotypes. A direct proof would show Koehn, R. K., Bayne, B. L., Moore, M. N. & Siebenaller, J. F., that certain allozymic genotypes are favored in a 1980. Salinity related physiological and genetic differences more brackish environment (Koehn et al., 1980). between populations of Mytilis edulis. Biol. J. Linnean So- The genetic similarity observed between the bay ciety 14: 319-334. group and the ocean group suggests that these two Koehn, R. K., Milkman, R. & Mitton, J. B., 1976. Population genetics of marine pelecypods. IV. Selection, migration, and groups could be subspecies. The geographic isola- genetic differentiation in the blue mussel Mytilus edulis. tion of these snails could have occurred just by the Evolution 30: 2-32. nature of currents and tides near large estuaries. Levene, H., 1953. Genetic equilibrium when more than one The currents around the peninsula may prevent ecological niche is available. Am. Nat. 87: 331-333. mixing of the two groups of snails. Scheletema Levin, D. A. & Kerster, H. W., 1974. Gene flow in seed plants. In: Evolution Biology 7:139 220. T. Dobzhansky & M. K. (1975) found that surface dwelling larvae rarely Hecht, eds, Plenum Press, New York. enter major estuaries along the Atlantic coast. Our McCracken, G. F. & Brussard, P. F., 1980. The population biol- data are consistent with a surface dwelling larval ogy of the white-lipped snail, Triodopsis albolabris: Genetic stage for M. bidentatus since little mixing occurs variability. Evolution 34:92 104. between the ocean populations and bay popula- Nei, M., Genetic distance between populations. Am. Nat. 106: 283 292. tions. Thus, bay snails became geographically iso- Nevo, E., 1978. Genetic variation in natural populations: Pat- lated and differentiated from the ocean snails. terns and theory. Theor. Pop. Biol. 13: 121-177. Our study indicates that the extensive gene flow Russell-Hunter, W. D., Apley, M. L. & Hunter, R. D., 1972. is dependent on the strength of good dispersal vec- Early life history of Melampus and the significanceof semi- tors, in this case the ocean currents. Gene flow lunar synchrony. Biol. Bull. 143: 623-656. Schaeffer, S. W., 1980. Genetic variation of the salt marsh snail appeared to be high when the ocean currents trav- (Melampus bidentatus) around the southern Delmarva elled between populations. On the other hand, gene peninsula. M.S. Thesis, West Virginia University, Morgan- flow was low when the currents were blocked. town, W.V. Schaal, B., 1975. Population structure and local differentiation in Liatris cylindracea. Am. Nat. 109:511-528. 229

Scheltema, R.S., 1975. Relationship of larval dispersal, gene Sheath, P. H. A. & Sokal, R. R., 1973. Numerical . flow and natural selection to geographic variation of benthic W. H. Freeman, San Francisco. invertebrates in estuaries and along coastal regions. In: Es- Workman, P. L. & Niswander, J. D., 1970. Population studies tuarine research: Chemistry, biology and the estuarine sys- on southwestern Indian tribes. 11. Local genetic differentia- tem. L. E. Cronin, ed., Academic Press, New York. tion in the Papago. Am. J. Human Genet. 22: 24-49. Selander, R. K., 1976. Genic variation in natural populations. Wright, S., 1943. Isolation by distance. Genetics 28:114 138. In: Molecular evolution. F. J. Ayala ed., Sinauer Associates, Inc., Sunderland, Massachusetts. Received 10.4.1984 Accepted 15.10.1984.