The scleractinian coral Goniastrea aspera (Verrill) undergoes both broadcast spawning and planulae brooding in the Ryukyu Archipelago of southern Japan. Genetic variation and gene flow in G. aspera were studied using allozyme electrophoresis. We tested the hypothesis that gene flow is determined by the competency period of the planulae. We also assessed the relative contributions of sexual and asexual reproduction to recruitment. For the five staining systems surveyed, G. aspera encoded five polymorphic loci and one monomorphic locus. The genotype frequencies in each population significantly differed from the expected Hardy-Weinberg equilibrium (HWE), indicating that the local populations of G. aspera are not fully panmictic. The high ratio of the observed number of genotypes to the number of individuals (0.90±0.07, mean NG:N±SD) and the observed to expected genotypic diversity (0.84±0.11, mean GO:GE±SD) suggested that each population is likely maintained by sexual reproduction. The genetic differentiation (FST) and value of average number of migrants per generation (Nem) among and within regions ranged from 0.025 to 0.104 and 2.2 to 9.6, respectively. Comparisons with other species demonstrated that larva survival rates also influence gene flow. In addition, gene flow on distant reefs by planulae originating from spawning might prevent divergence by planulae originating from brooding for short-distant dispersal among and within populations of G. aspera in the Ryukyu Archipelago.
The effects of differences in reproductive modes on genetic variability and gene flow have been studied in some marine invertebrates, i. e., gamete spawners or planulae brooders (Benzie and Williams, 1997; Ayre and Dufty, 1994; Ayre et al., 1997a, b; Yu et al., 1999; Ayre and Hughes, 2000; Bastidas et al., 2001, 2002; Nishikawa et al., in press). Several population genetic analyses using allozyme electrophoresis have shown that brooding species have restricted gene flow (Nem: number of migrants per generation) compared with spawning species in ascidians, scleractinian corals, and alcyonacean corals (Hellberg, 1996; Ayre et al., 1997a; Bastidas et al., 2001, 2002; Nishikawa et al., in press). These reports are in good agreement with the characteristics of the reproductive mode; larvae produced by brooding generally settle quickly after their release from the parental colony, while larvae produce by spawning delay settling for at least a few days (e. g., Svane and Young, 1989; Harrison and Wallace, 1990). By contrast, an allozyme study of five brooding and four spawning scleractinian coral species on the Great Barrier Reef (GBR) did not fully support this simple prediction (Ayre and Hughes, 2000), as it demonstrated that larval dispersal was sufficient to maintain moderate to high levels of gene flow along the entire GBR in three of five brooding species.
The scleractinian coral Goniastrea aspera (Verrill) is widely distributed in Indo-Pacific reefs. Abe (1937) and Motoda (1939) reported that G. aspera in Palau released brooded planulae. By contrast, G. aspera on the GBR is a hermaphroditic spawner that spawns once per year (Babcock, 1984; Harrison et al., 1984; Willis et al., 1985; Babcock et al., 1986). In the Ryukyus, Sakai (1997) reported that individual colonies of G. aspera brooded planulae after spawning gametes. Goniastrea aspera broods planula, which is probably only capable of short distance dispersal since laboratory experiments have shown that brooded planulae settle within one hour of release. Nishikawa et al. (in press) compared larval settlement rates and gene flow in scleractinian corals, using electrophoresis to examine a spawner (Acropora tenuis) and a brooder (Stylophora pistillata) in the Ryukyu Archipelago. They estimated that the settlement-competency period in A. tenuis and S. pistillata was 69 and 51 days, respectively. Moreover, they also estimated that Nem in A. tenuis and S. pistillata was 16.4 and 1.5, respectively. They concluded that the longer competency period of A. tenuis larvae versus those of S. pistillata was one of the major cause of the difference in Nem. The estimated competency period of planulae originating from spawning in G. aspera in the Ryukyus was 63–70 days, similar to that of A. tenuis (Nozawa, 2000). These results suggest that planulae originating from spawning in G. aspera have the potential to disperse long distances. Due to the unusual reproductive mode of G. aspera in the Ryukyus, we hypothesized that gene flow in G. aspera would exceed that in the brooding coral S. pistillata and not differ greatly from that in the spawning coral A. tenuis. Although brooded larvae may display restricted gene flow among and within populations due to their short-distance dispersal, the recruitment of planulae originating from spawning on distant reefs might prevent the fixation of genetic differences.
In Western Australia, the brooded planulae of the viviparous coral Pocillopora damicornis are thought to be derived from asexual reproduction (Stoddart, 1983, 1984a, b). He reported that larvae of P. damicornis are genotypically identical to their brood parents; local populations are dominated by a small number of clones (highly replicated multi-locus genotypes), which are generated asexually. By contrast, there was no evidence of asexual replication of genotypes in P. damicornis on the GBR (Benzie et al., 1995; Ayre et al., 1997b). The relative importance of sexual and asexual reproduction has profound effects on the genetic structure of populations (Jackson, 1986). The occurrence of asexual reproduction can lead to gene frequencies that are skewed from those expected under random mating. In addition, clonal life should also accelerate evolutionary change within, and divergence among, local populations (Wright, 1978). For G. aspera in the Ryukyus, Sakai (1997) suggested that brooded planulae are very likely the products of sexual reproduction, because eggs remained in the polyps after spawning and there was no sign of asexual reproduction. However, population genetic data on G. aspera in the Ryukyu has been limited.
It should be noted that indirect estimates of genetic differentiation (FST) may low precision of model-based estimates (Bossart and Prowell, 1998) and raise mathematical and statistical problems (Whitlock and McCauley, 1999). Small value of FST cannot be precisely estimated because they correspond to differences in population allele frequencies that are small relative to differences that arise by chance in samples of populations. Because of the non-linearity of the relationship between FST and Nem, when FST is small the variance in its estimator is transformed into a much larger variance in the estimator of Nem (Waples, 1998). Neigel (2002) argue that direct estimation based on mark-and recapture or genetic identification of migrant sources is the preferred (e.g., Palsboll et al., 1997). Moreover, he mentioned that, although gene flow should be estimated by more powerful approaches whenever practical, FST remains a useful measure of the average effects of gene flow and will continue to be used for comparative purposes (Neigel, 2002). For avoid such problems, we showed not only Nem but also FST in this study.
In this paper, we measured local (two or three stations in a region) and regional (Okinawa, Kerama and Ishigaki) patterns of genetic variation and gene flow in G. aspera in the Ryukyu Archipelago, southern Japan, using allozyme electrophoresis. We assessed the relative contribution of sexual and asexual reproduction to recruitment. We also tested the hypothesis that gene flow is determined by their planulae competency in a comparison with other species. The reproductive mode of G. aspera in the Ryukyus is unique, since other scleractinian corals reproduce only by planula brooding or gamete spawning (reviewed by Harrison and Wallace, 1990). Therefore, a comparison of population genetic data between G. aspera and other corals in the Ryukyus provides further information to the relationship between reproductive mode and gene flow in corals.
MATERIALS AND METHODS
Collection and storage of specimens
We established three geographic sampling regions (Okinawa Islands, Kerama Islands and Ishigaki Island), and collected coral samples from two or three reefs that were more than 5 km apart within each region (Fig. 1). Fragments were taken from 38 to 50 colonies at least 3 m apart, which were haphazardly selected at each reef. We kept the collected fragments alive in a small plastic tank filled with seawater, and brought them to the laboratory at the Tropical Biosphere Research Center (TBRC) at the University of the Ryukyus. The fragments were alive after transportation and were reared in an outdoor holding tank (2.0×1.0×3.5 m) supplied with running seawater. The coral fragments were stored at −80°C for at least 1 hr before electrophoresis.
Electrophoresis was used to detect variation at polymorphic loci following the methods of Selander et al. (1971) and Hillis et al. (1996), using horizontal starch gels (11–12%; at 100–180 V and 36–48 mA for 6–9 hr). Five enzyme systems were used: malate dehydrogenase (Mdh; E.C.#188.8.131.52), phosphoglucomutase (Pgm; E.C.#184.108.40.206), hexokinase (Hk; E.C.#220.127.116.11), and peptidase using leucyl-tyrosine (Lt; E.C.#3.4.11) and leucyl-glycyl-glycine (Lgg; E.C.#3.4.11) as the substrate. Tris EDTA citrate (pH 7.5) buffer was used for Mdh, Pgm, and Hk, and Tris citrate (pH 8.0) buffer was used for Lgg and Lt.
The allele frequencies, measures of genetic variability, and genetic differentiation among and within regions were analyzed using the computer program TFPGA (Miller, 1997). We also assessed the magnitude and direction of departures from Hardy-Weinberg equilibrium (HWE). These departures were expressed as Wright's (1978) fixation index ƒ, where positive and negative values represent deficits or excesses of heterozygotes, respectively. A total of five tests were conducted and Bonferroni's adjustment was used to test the significance of departures from HWE (significant: P<0.01).
Two measures were used to assess the possible effects of asexually derived recruits on the genotypic diversity of the colonies. First, each colony was assigned a multilocus (clonal) genotype. The number of multilocus genotypes detected (NG) is an estimate of the minimum number of clones present within a population; and the ratio NG:N (N is the number of individual colonies collected) provides a simple index of the effect of asexual reproduction on genotypic diversity. Second, the ratio of observed multilocus genotypic diversity (GO) to that expected under conditions of sexual reproduction (GE) was calculated following Stoddart and Taylor (1988). Departure of GO:GE from unity is an index of the combined effect of departures from single-locus HWE and multilocus linkage disequilibrium. A genetically variable population with high levels of asexual recruitment should have a low ratio of observed to expected genotypic diversity. We also tested for significant differences between GO and GE using the t-test (Stoddart and Taylor, 1988).
We used Wright's (1969) standardized genetic variance (FST) to quantify the levels of allelic variation among regions or reefs, thereby inferring the degree of population subdivision. We calculated this parameter as Weir & Cockerham's (1984) θ, using the program TFPGA (Miller, 1997), which executes numerical resampling to estimate the variance at each locus (jackknifing) and the variance across loci (bootstrapping). We also estimated gene flow using Wright's (1969) island model (Nem=[(1/θ)−1]/4, where Ne is the effective population size and m is the proportion of migrants per generation).
For the five staining systems surveyed, five polymorphic loci and one monomorphic locus were encoded for Goniastrea aspera (Table 1). The mean number of alleles per locus (Na) in each reef ranged from 2.7 to 3.5, and the percentage of polymorphic loci (P95) ranged from 66.7 to 83.3. The observed mean heterozygosity in each reef was lower than the expected heterozygosity assuming HWE (t-test, P<0.01, Table 1).
Goniastrea aspera. Genetic variability in eight reefs, respectively. Locality abbreviation as in Fig. 1.
The allele frequencies of G. aspera differed among the three study regions (Table 2). Although Pgm*53 and Mdh*80 were observed only in the Yaeyama region, Pgm*67 was not observed there. In addition, the frequencies of Pgm*100 in the Yaeyama region was higher than 0.5, but it ranged from 0.235 to 0.410 in other populations.
Goniastrea aspera. Allele frequency of eight reefs. Collection sites as in Fig. 1. N, number of individuals.
The ratio of the number of observed genotypes (NG) to the number of individuals sampled (N) ranged from 0.82 to 1.00, and the ratio of the observed multilocus genotypic diversity (GO) to that expected under conditions of sexual reproduction (GE) among collection sites ranged from 0.68 to 1.00 (Table 3). The ratios NG:N and GO:GE were both high (0.90±0.11 and 0.84±0.07, respectively, mean ± SD). However, the differences between GO and GE were significant only for SHI, TOK, and GIT (unpaired t-test).
Goniastrea aspera. Estimates of contribution of asexual reproduction in each eight reefs (N: number of individuals, NG: number of multi-locus genotypes, Go: observed genotypic diversity, G E: expected genotypic diversity). Locality abbreviations as in Table 1. (unpaired t-test, *p<0.05, **p<0.01)
Wright's (1969) fixation index (ƒ) of all loci was positive for all reefs except for AKA in the Kerama region. If the significant deviation from HWE were due to subdivision within populations, one would expect parallel patterns for different loci within each population. However, there is strong pattern of differences among loci, with 5 of 9 significant departures within Lt-1. This significant deviation from HWE is likely due to the locus where the electromorphs differed in mobilites least (based on numbering in Table 2). This tendency is in agreement with significant deviations from HWE that were not found at great mobilite locus of Pgm (Table 4). In excess and deficits number of heterozygotes, there were almost all deficits (31 out of 38 cases), and significant deviations from HWE were seen in 9 cases.
Goniastrea aspera. Wright's fixation index (F) indicating heterozygote excess (negative number) or deficit (positive number) for each locus in reefs. Asterisks show significant differences among reef after Bonfferoni adjustment for multiple comparisons. (siginificant level: * p<0.05, ** p<0.01, *** p<0.001), (n: no data available).
FST (genetic differentiation) differed significantly from zero in all pairs of regions, except for all regions between Yaeyama and Kerama, ranged from 0.025 to 0.104 (P<0.05, Table 5). The estimated gene flow (the number of migrants per generation: Nem) calculated from FST using Wright's island model (1969) ranged from 2.2 to 9.6 among and within regions. In addition, Nem showed similar patterns among each pair of regions (all regions > all reefs > reefs-region, Table 5).
Goniastrea aspera. Genetic differentiation (FST ±} SE) and gene flow (Nem, average number of migration pergeneration) in three hierachical level are estimated in each pair of region. Statistical significance are calculated by 95 and 99% CI. (Significant level: *P<0.05, **P<0.01). (O: Okinawa, K: Kerama, and Y: Yaeyama)
Localized asexual recruitment via asexual reproduction or fragmentation had little effect on maintaining populations of G. aspera. In the previous study, Sakai (1997) suggested that brooded planulae are very likely the products of sexual reproduction, because eggs remained in the polyps after spawning and there was no sign of asexual reproduction. Asexual reproduction in many clonal marine invertebrates appear to produce (1) low ratios of both NG:N and GO:GE, and (2) a similar number of heterozygotes excesses and deficits (Stoddart, 1983, 1984a, b; Johnson and Threlfall, 1987; Ayre and Willis, 1988; Burnett et al., 1995; Adjeroud and Tsuchiya, 1999; Ayre and Hughes, 2000). Previous studies have demonstrated that both ratios are low in coral populations with asexual recruitment. For example, Stoddart (1984) showed that in Pocillopora damicornis, which has asexual planulae, NG:N and GO:GE were 0.40±0.17 (mean ± SD) and 0.27±0.15, respectively, and Ayre & Willis (1988) demonstrated that in Pavona cactus, which propagates sexually and asexually (fragmentation), the respective ratios were 0.35±0.23 and 0.35±0.31. We found that in G. aspera, both NG:N and GO:GE were relatively high: 0.90±0.07 and 0.84±0.11, respectively. In addition, the heterozygotes deficits from HWE were over four times more than the excess. Although this study paid little attention to time scales that may affect ecological and genetic estimates (see McFadden, 1997), we conclude that asexual recruitment likely contributes little to the maintenance of G. aspera populations by Sakai's histological observation and present population genetic data.
This study supported our hypothesis that genetic differentiation and gene flow are greater in G. aspera than in the brooding coral S. pistillata. The values of FST and Nem among and within regions of a spawning and brooding coral G. aspera were significantly greater than that of a brooding coral S. pistillata in the Ryukyu Archipelago (P<0.01 and P<0.01, respectively, Wilcoxon signed-ranks test, Table 5). Although there has been no report on brooded planulae of G. aspera, the competency period is higher in planulae originating from spawning of G. aspera than in brooded planulae of S. pistillata (63–70 days and 51 days, respectively, Nozawa, 2000; Nishikawa et al., in press). Generally, gene flow is a powerful cohesive force (Slatkin, 1987); one effective migrant exchanged between subpopulations per generation is sufficient to stop the subpopulations from drifting to fixation (alternative gene expression), although higher levels are required to maintain homogeneity (Allendorf and Phelps, 1981). The differences in the competency periods of the two corals indicate that restricted gene flow is likely in S. pistillata as compared to G. aspera. Moreover, the percentage of mean settlement rate 1 day after planulae release, based on the mean total settlement rates, is 41.5–60.0 and 71.4% in brooded planulae of G. aspera and S. pistillata, respectively. These results indicate that settlement near parental colonies is more likely in S. pistillata than in G. aspera. In order to discuss the influence of these initially settlement rates on gene flow, comparisons with other species are needed (i. e., with differential initial settlement rates but similar competency periods) in a future study.
This study did not support our hypothesis that gene flow in G. aspera is not greatly different from that in the spawning coral Acropora tenuis. Planulae originating from spawning in these two corals have similar competency periods (63–70 and 69 days, respectively, Nozawa, 2000; Nishikawa et al., in press). However, the genetic differentiation (FST) and number of migrants per generation (Nem) among and within regions of the spawning and brooding coral G. aspera was significantly lower than that of the spawning coral A. tenuis in the Ryukyu Archipelago (P<0.05 and P<0.05, respectively, Wilcoxon signed-ranks test, Table 5). Our results indicate that gene flow in G. aspera is relatively restricted as compared to that in the spawning coral A. tenuis. One possible explanation for the difference in gene flow between the two is a difference in survival rates of larvae; the mean survival rate was less than 10% of the initial larvae 35 and 59 days after gamete release in G. aspera and A. tenuis, respectively (Nozawa, 2000; Nishikawa et al., in press). These results indicate that both the competency period and survival rates are important for gene flow in these corals. This study demonstrated that relatively long distance dispersal is likely in A. tenuis as compared to G. aspera, owing to the different survival rates.
In conclusion, although there are two possibilities of asexual reproduction in scleractinian corals, i.e. asexual production of planula (Pocillopora damicornis; Stoddart, 1983) and the fragmentation (e.g., Pavona cactus; Ayre and Willis, 1988), asexual reproduction unlikely contributes to maintain local G. aspera populations in the Ryukyu Archipelago. Because there was no sign of asexual production of planula by histological study (Sakai, 1997) and the fragmentation unlikely occur in the small massive coral of G. aspera in the Ryukyu. Gene flow in G. aspera is intermediate between that of Acropora tenuis (spawning coral) and Stylophora pistillata (brooding coral). These results for G. aspera suggest that gene flow attribute to planulae originating from spawning on distant reefs prevents the accumulation of fixed genetic differences among and within regions in the Ryukyu Archipelago.
We are indebted to the staff of Sesoko Station, Tropical Biosphere Research Center, University of Ryukyus, staff of International Coral Reef Research and Monitoring Center, the staffs of Akajima Marine Science Laboratory, the staff of National Okinawa Youth Center, Yaeyama Marine Park Research Station. Thanks are due to many students in TBRC for their encouragement, help and advice. We would like to thank two anonymous reviewers for comments on the manuscript. This study was partly supported by grants from Japanese Ministry of Education, Science, Sports and Culture (Nos. 12640619 and 14205071-7), and from the Sumitomo Foundation. (No. 993212).