Seasonal patterns of population structure in a colonial marine invertebrate (Bugula stolonifera, bryozoa).
Larvae are the primary means of dispersal for sessile invertebrates, although other processes such as fragmentation and rafting can result in the dispersal of post-metamorphic individuals (e.g., Bruno, 1998; Watts et at, 1998). The degree to which these mechanisms transport individuals away from their natal grounds can have significant ecological implications, including affecting levels of inbreeding within populations and connectivity between populations. It has been widely accepted that larval dispersal distance can be directly affected by the length of the larval planktonic period (e.g., Crisp, 1978; Bohonak, 1999; Shanks et at, 2003), which in turn can be influenced by such factors as reproductive and developmental mode (see Strathmann, 1985), egg size (see Thorson, 1950), amount of endogenous reserves (Wendt, 2000), presence of appropriate settlement cues (see Pawlik, 1992), and ability to withstand a protracted larval period (see Pechenik, 2006). Indeed, results from studies examining larval dispersal in either sympatric or closely related species with contrasting reproductive strategies have shown that increased larval duration can result in increased dispersal (reviewed by Bohonak, 1999). Conversely, a growing number of studies have documented that larvae of some species recruit back to their natal populations, regardless of the length of the larval dispersal period (e.g., Palumbi, 2001; Swearer et al., 2002; Warner and Cowen, 2002).
Knowlton and Keller (1986) were among the first to empirically demonstrate that localized recruitment in long-lived larvae ([approximately equal to] 1 week for alpheid shrimp) could occur within meters of the birth site. Subsequent studies have demonstrated that this type of settlement by planktotrophic larvae might be widespread (e.g., Ayre and Hughes, 2000; Jones et at, 2005). Studies such as these have not only shifted conventional wisdom away from equating dispersal potential with realized dispersal, but have also led to investigations of how environmental factors (e.g., Dupont et at, 2007; Underwood et al., 2009) or larval behavior can result in lower than expected realized dispersal (see Levin, 2006). Todd et at (1998) investigated genetic structure in a nudibranch with a pelagic lecithotrophic larval stage that, although competent to metamorphose within 2 days of release, could remain planktonic for several weeks. The significant genetic differentiation that these authors documented in populations countered expectations of dispersal potential and suggested that larval behavior was responsible for the observed local recruitment. Hence, it has become apparent that in evaluating the realized dispersal achieved by long-lived larvae one must consider variables other than simply dispersal potential. Although the paradox of long-lived larvae demonstrating local recruitment appears to be driving the idea that dispersal potential is a poor indicator of dispersal distance, some studies have reported higher than expected larval dispersal in species with low dispersal potential (e.g., Maier et at, 2009).
Miller and Ayre (2008) examined genetic structure in Goniastrea favulus, a broadcast-spawning coral with negatively buoyant eggs that adhere to the parent colony through fertilization, and whose larvae are also negatively buoyant through development. Despite expectations of reduced dispersal and localized recruitment, the authors found similar levels of genetic subdivision in comparisons to another species with buoyant eggs and larvae, suggesting that the two species had similar dispersal patterns. The authors mention that for G. favulus, the reduced mobility found in the larvae may last only a few days, after which larvae swim actively. It could be that the ability to withstand an extended larval swimming duration and cope with the potential deleterious effects of this protracted phase is the primary factor allowing for greater than expected dispersal by short-lived nonfeeding larvae. Conversely, an inability to cope with a protracted larval phase could result in very low dispersal by larvae and surprisingly small-scale genetic structure. Yund and O'Neil (2000) found genetic structure in a brooding colonial ascidian over distances of 8 m, while Calderon et al. (2007) found structure in a brooding sponge on a scale of tens of centimeters. The effects of extended larval swimming in these species is not known; however, Marshall et al. (2003) found that extending the swimming duration of larvae by 2-3 h in the colonial ascidian Diplosoma listerianum resulted in reduced fitness due to a smaller postmetamorphic size.
Our study investigated genetic structure in the marine bryozoan Bugula stolonifera. As with most species in the phylum Bryozoa, B. stolonifera is characterized as having low dispersal potential. Metamorphosis is irreversible, and adult colonies are sessile. This species releases short-lived nonfeeding larvae that are competent to metamorphose within about 1 h and will usually commence metamorphosis within about 4 h (Woollacott et at, 1989; Wendt and Woollacott, 1999). Previous work (Woollacott et at, 1989) has documented significant deleterious effects associated with a protracted larval phase, as extended larval swimming results in significant decreases in juvenile survival and growth. Despite this, Johnson and Woollacott (2010) found high amounts of mixing within conspecific aggregations. For the present study, a suite of 10 microsatellite loci was used to determine if high mixing within an aggregation also resulted in high levels of mixing between aggregations. For B. stolonifera, however, multiple generations occur throughout the reproductive season; the time to reach reproductive maturity is about 12 days (Johnson, 2010), while the embryonic brooding period has been observed in the laboratory to last about 7 days (unpubl. obs.). Thus, additional analyses were performed to determine (1) if genetic differentiation between sites was consistent throughout the reproductive season, resulting in stable micro-geographic structure for these animals, (2) if there was any evidence for genetic mixing that might result in increased homogeneity between sites, and (3) if inter-annual changes in genetic variability existed within sites. For some erect bryozoans, the erect portions of colonies are not perennial. Rather, these portions experience seasonal die-backs during unfavorable conditions (e.g., Keough and Chernoff, 1987). It is thought, however, that root-like projections emanating from these colonies can survive and bud new zooids when favorable conditions return (Numakunai, 1960, 1967). By examining inter-annual genotypic variation in Eel Pond, it is possible to determine whether this annual cycle of die-off and regrowth can result in a population bottleneck, or shift in genetic composition.
Materials and Methods
About 30 adult colonies of Bugula stolonifera Ryland 1960 were collected from each of five sites within Eel Pond, Woods Hole, Massachusetts, in August 2009 (Fig. 1). A more distant site (Hadley Harbor) was also sampled for comparison. To ensure adequate sampling, animals were sampled from the entire area ([approximately equal to] 5-15 [m.sup.2]) within each site. Because these animals brood embryos within the colony, only portions of colonies lacking brood chambers were selected for genotyping. Results from genetic analyses conducted in 2009 showed significant genotypic differentiation between most sites. These analyses were extended in 2010 and 2011 to investigate potential inter-annual changes in genetic structure within each site in Eel Pond, as well as intra-annual changes in genetic structure within and between the Eel Pond sites. In Eel. Pond, B. stolonifera colonies are reproductively active from about June through November. The Eel Pond sites used in 2009 were sampled early, mid-, and late summer in 2010, and then in early summer 2011. As previously mentioned, multiple generations can occur throughout the reproductive season, so sampling discrete cohorts of individuals is not possible. However, by preferentially sampling small colonies over large, established individuals at each sampling time, it was possible to investigate a change in genetic structure over time. Additionally, under-sampling individuals during each collection could result in an inadequate representation of the genetic diversity within each site. To ensure that potential significant differences within and between sites were not due to our sampling protocol, an additional collection of animals was conducted at EP 5 during the mid-summer 2010 collection.
Animals were fixed in 95% EtOH, and DNA was extracted using the EZ-10 96-well-plate genomic DNA isolation kit (Bio Basic Inc.). DNA was extracted only from portions of colonies lacking brood chambers. Also, because colonies can asexually reproduce via root-like projections, only individuals with unique genotypes were used for analyses. Ten microsatellite loci were amplified using published primer pairs (Johnson and Woollacott, 2010). These pairs were divided into three groups on the basis of annealing temperature. The 5' end of the forward primer in each group was fluorescently labeled with either 6-FAM, VIC, NED, or PET dye and utilized in a Qiagen Multiplex PCR reaction. Reactions were performed in 10-[micro]1, volumes containing 20 ng of DNA, 1 x Multiplex PCR Master Mix, 1 x primer mix (0.3 [micro]mol[1.sup.1] each), and 2.75 [micro]1 of d[H.sub.2]2O. The following thermal cycler program was used: 94 [degrees]C for 15 min, 30 cycles of 94 [degrees]C for 30 s, annealing temperature for 90 s, and 72 [degrees]C for 60 s, followed by a final extension of 72 [degrees]C for 10 min. PCR products were run on an AB1 3730 xl DNA sequencer using GeneScan-500 LIZ as the size standard. Results were analyzed using PeakScanner Software ver. 1.0 (Applied Biosystems). For loci that failed to amplify during the Multiplex reaction, single-locus PCR reactions were performed using the protocol in Johnson and Woollacott (2010).
Descriptive statistics including the number of alleles per locus, percentage of polymorphic loci, and observed and expected heterozygosity were calculated with GDA ver. 1.1 (Lewis and Zaykin, 2002). The presence of null alleles at each locus was investigated using Micro-Checker ver. 2.2.3 (van Oosterhout et al., 2004). Exact tests examining linkage disequilibrium between all pairs of loci within each year and deviations from Hardy-Weinberg Equilibrium (HWE) within each site were conducted in Genepop (Raymond and Rousset, 1995) ver. 4.0, with Markov Chain parameters set to 10,000 dememorizations, 500 batches, and 10,000 iterations per batch. Estimations of Wright's inbreeding coefficient (F15) were calculated and tested for significance using Genetix ver. 4.05 (Belkhir et al., 1996-2004) set to 10,000 permutations. The sequential Bonferroni correction was utilized to adjust significance levels compensating for multiple comparisons within the same test (Rice, 1989).
To estimate the spatial and temporal components of population genetic structure over the duration of the study, a hierarchical analysis of molecular variance (AMOVA) was performed in Arlequin ver. 18.104.22.168 (Excoffier and Lischer, 2010). Genotypic data from each sampled individual were compiled by sampling year and then grouped according to site. As the Hadley Harbor and EP 2 sites were sampled only in 2009, these data were excluded from this analysis. Additionally, although multiple collections were conducted in 2010, data were compiled into a single sampling period within each site. The locus-by-locus AMOVA for genotypic data in Arlequin partitioned the molecular variance among spatial groups, among collection years within spatial groups, and among samples. Estimated fixation indices were averaged across all loci, and their significance was determined using 10,000 permutations. After this analysis, data were examined for genetic differentiation between samples and to investigate the potential effects of the seasonal colony die-back within and between sites.
For the 2009 data, genetic differences between the Eel Pond and Hadley Harbor collection sites were initially investigated using the exact test for genotypic differentiation in Genepop. This test analyzes the distribution of diploid genotypes in all pairs and assumes that genotypes are distributed equally (Raymond and Rousset, 1995). Settings for the Markov Chain reaction were as previously described. Genotypic differentiation between all pairs of sites was also analyzed by calculating pairwise [F.sub.ST]values in Arlequin. These values were tested for significance using 10,000 permutations. The relationship between genetic and geographic distance was analyzed by a Mantel test performed in Genetic Analysis in Excel (GenAlEx) ver. 6.0 (Peakall and Smouse, 2006). The genetic distance matrix used in this analysis was based on mean pairwise population differences, while the geographic distance matrix was based on decimal latitude and longitude coordinates for each collection site. Significance was determined on the basis of 9999 permutations.
To investigate potential genetic differentiation resulting from the annual cycle of colony die-back and regrowth, the 2009 data and the early summer 2010 data were subjected to the exact test for genotypic differentiation in Genepop and the estimation of pairwise [F.sub.ST] values in Arlequin. To determine if the observed pattern was consistent across multiple years, comparisons were also conducted between the late summer 2010 and early summer 2011 data. Settings for each test were as previously described. Results from these analyses provided evidence for significant differentiation within sites. The early summer 2010 and 2011 data were then investigated for a potential population bottleneck by using the program Bottleneck ver. 1.2.02 (Cornuet and Luikart, 1996). Bottleneck utilizes allele frequency data to detect recent reductions in effective population size by comparing observed genetic heterozygosity to the expected equilibrium heterozygosity. Because the allelic diversity decreases faster than genetic heterozygosity during a bottleneck event, populations that have experienced a recent bottleneck should be more heterozygous than expected on the basis of the observed allelic diversity. Data were analyzed using the two-phase model of mutation (TPM). The TPM is recommended for microsatellite data as it is a compromise between the stepwise mutation model and the infinite allele model, allowing for mostly stepwise mutations but incorporating a small percentage of multistep mutations. Estimations from the program were based on 10,000 iterations. In addition to the use of Bottleneck, potential changes in allelic diversity after the annual die-back were examined. If numerous individuals were unable to survive the overwintering process, it might be expected that rare alleles would drop out from each population, resulting in an overall decrease of allelic diversity. The allelic diversity averaged over all loci was calculated for each site at each sampling time in Arlequin.
Prior to examining intra-annual changes in genotypic differentiation within and between sites for the 2010 data, we conducted an initial comparison between the two mid-summer collections at EP 5 to ensure that significant differentiation was not due to some artifact of our sampling protocol. The data were subjected to the tests for genotypic differentiation in Genepop and Arlequin as previously described. Because results documented no significant difference, these data were combined for all subsequent analyses, and the differences between sites within each sampling time were determined. Additionally, intra-annual changes within each site were investigated by comparing the early summer to the mid-summer collection period, and then comparing the mid-summer to the late summer collections. These data were analyzed in Genepop and Arlequin as previously described. Because Eel Pond B. swlonifera colonies experienced widespread die-back between the mid- and late summer collections, the late summer data were also analyzed for a potential population bottleneck resulting from this colony die-back. These data were subjected to the tests in Bottle-neck and examined for changes in allelic diversity as previously described.
Descriptive statistics and results from AMOVA
Throughout the study, 648 individuals were genotyped from five sites within and one site outside of Eel Pond, Woods Hole, Massachusetts (Fig. 1). The average number of alleles per locus ranged from 6.6 to 8.7 for each site, and all loci were found to be polymorphic (Table 1). All sites in each year were found to deviate significantly from Hardy-Weinberg equilibrium (HWE) (P < 0.0001). Expected heterozygosity ranged from 0.641 to 0.707, while observed heterozygosity ranged from 0.354 to 0.557. Results from Micro-Checker analyses suggested that null alleles were present at 6 out of 10 loci due to heterozygote deficiencies at these loci. The extremely low occurrence of null allele homozygosity across all loci ([approximately equal to]0.17%), however, suggests that deviations from HWE stemmed from other causes. Inbreeding coefficients ranged from 0.205 to 0.478, and all were significantly different than zero (Table 1). There was no evidence for linkage disequilibrium for any pairs of loci after sequential Bonferroni correction, with the exception of Bug.stol4 and Bug.stol6 in the EP 1 site in 2009 and Bug. stol3 and Bug.stol7 in the EP 4 site in 2011 (see Johnson and Woollacott, 2010, for locus descriptions).
The hierarchical AMOVA investigating overall genetic structure documented both a highly significant spatial component and a highly significant temporal component to the observed genetic variability found throughout the duration of the study (Table 2).
Genotypic differentiation 2009
Pairwise comparisons examining genetic differentiation between sampling sites found significant differentiation between most sites, even those separated by as few as 100 m (e.g., EP 4 and EP 5) (Table 3). For the site outside of Eel Pond, significant differentiation was found between the Hadley Harbor collection site and all sites in Eel Pond (P S 0.0001 for all tests). Within Eel Pond, results from the exact test for genotypic differentiation conducted in Genepop showed significant differences in all comparisons with the exception of EP 2 and EP 3 (P = 0.070) and EP 1 and EP 5 (P = 0.127) (Table 3). Likewise, most pairwise FsT values calculated in Arlequin were significantly different from zero, with only the comparison between EP 2 and EP 3 not significant (P = 0.133). No significant relationship was found among comparisons of genetic and geographic distances across all sites (R2 = 0.138, P = 0.324).
Animal collection 2010-2011
Collection sites within Eel Pond were monitored biweekly beginning in April 2010 for initial colony regrowth. The site designated EP 2 (Fig. 1), a floating dock, was no longer located within Eel Pond and could not be used for the 2010 analyses. Colonies of Bugula stolonifera appeared in Eel Pond in late May, and by mid-June each site contained sufficient biomass to allow for collection. For the early summer time period, the site designated EP 5 was sampled on 4 June 2010, and the sites designated as EP 1, EP 3, and EP 4 were sampled on 17 June 2010. These sites were sampled again on 27 July 2010 for the mid-summer time period (40 d from end of previous collection). By this time, B. stolonifera abundance was such that colonies carpeted the sides of floating docks and piers where they occurred. For the late summer time period, sampling sites were inspected on 6 September 2010 (41 d from previous collection). At this time, however, all sites lacked sufficient biomass for collection, and it was evident that there had been widespread colony die-back within Eel Pond. Sites were monitored over the next several weeks for colony regrowth, and multiple sampling days were required for each site. Collections were conducted on 13 and 17 September and 4 October 2010 to garner enough genetically distinct individuals for genetic analyses.
Table 1 Summary of genetic diversity ftr Bugula stolonifera colonies sampled at Eel Pond (EP) and Hadley Harbor (HH) collection sites in 2009 and early, mid-, and late summer 2010 Year Site N [P.sub.L] A [H.sub.E] [H.sub.O] [F.sub.IS] * 2009 EP 1 29 1.0 7.6 0.670 0.483 0.283 EP 2 30 1.0 8.0 0.698 0.545 0.222 EP 3 19 1.0 6.6 0.700 0.432 0.390 EP 4 30 1.0 7.7 0.701 0.492 0.302 EP 5 17 1.0 7.8 0.682 0.378 0.451 HH 30 1.0 7.9 0.707 0.480 0.324 2010 EP 1 30 1.0 7.7 0.690 0.363 0.478 early EP 1 29 1.0 7.8 0.685 0.383 0.446 mid EP 1 28 1.0 7.6 0.670 0.482 0.284 late EP 3 28 1.0 7.9 0.701 0.439 0.378 early EP3 30 1.0 8.7 0.687 0.470 0.320 mid EP 3 28 1.0 8.0 0.700 0.527 0.250 late EP 4 30 1.0 8.0 0.683 0.422 0.387 early EP4 30 1.0 8.1 0.699 0.383 0.456 mid EP 4 30 1.0 8.3 0.686 0.531 0.230 late EP 5 30 1.0 7.3 0.641 0.450 0.301 early EP 5 30 1.0 7.3 0.665 0.421 0.371 mid-1 EP 5 29 1.0 7.6 0.654 0.354 0.464 mid-2 EP 5 28 1.0 7.6 0.676 0.468 0.311 late 2011 EP 1 23 1.0 7.0 0.697 0.557 0.205 EP 3 25 1.0 8.0 0.703 0.540 0.235 EP 4 26 1.0 8.2 0.696 0.542 0.226 EP 5 30 1.0 7.8 0.675 0.510 0.245 The site labeled EP 5 mid-2 was intended for comparison to EP 5 mid-1 to ensure that our sampling technique adequately represented the genetic variability of each site at the time of collection. Number of individuals genotyped (N), percentage of polymorphic loci ([P.sub.L]). average number of alleles perlocus CA), expected ([H.sub.E]) and observed (He) heterozygosity calculated in GDA. Wright's inbreeding coefficient (F15) calculated in Genetix ver. 4.05. * A1l values of [F.sub.IS] differed significantly from zero (P < 0.0001). Table 2 Results from hierarchical AMOVA investigating the spatial and temporal components of genetic struture over the duration. of the study Source of Sum of Variance Percentage Fixation P variation squares components of variation indices Among sites 75.593 0.05538 Va 1.57 [F.sub.CT] <0.0001 = 0.0157 Among years 51.808 0.03981 Vb 1.13 [F.sub.SC] <0.0001 within = 0.0115 sites Among 3993.322 3.43665 Vc 97.31 [F.sub.ST] <0 0001 samples = 0.0270 Because the Hadley Harbor and EP 2 sites were sampled only in 2009. these sites were not included in this analysis. Data were compiled by sampling year and then grouped according to sampling site. Degrees of freedom were calculated for each locus and are thus not included. Variance components. percentage of variation, and fixation indices were estimated for each locus and then averaged across all loci. Significance of fixation indices is based on 10,000 permutations. Table 3 Matrix of pairwise [F.sub.ST] values (population genetic differentiation, Arlequin ver. 3.5) below diagonal and results from exact tests for genotypic diffrrentiation (Genepop 4.0) above diagonal for Eel Pond (EP) and Hadley Harbor (HH) sites collected in 2009 Site EP 1 EP 2 EP 3 EP 4 EP 5 HH EP 1 -- 40.46 42.50 79.51 27.30 71.22 EP 2 0.020 -- 30.02 104.10 40.64 78.55 EP 3 0.035 0.013 -- 69.86 40.98 69.75 EP 4 0.040 0.046 0.034 -- 106.19 [varies] EP 5 0.026 0.017 0.024 0.046 -- 74.05 HH 0.043 0.040 0.047 0.052 0.051 -- Bolded values indicate significance ([alpha] = 0.05).
Collection sites within e1 Fond were again monitored beginning in April 2011 for colony regrowth. B. stolonifiera colonies appeared in Eel Pond in mid-June, but growth was delayed relative to the previous year. The sites designated EP 4 and EP 5 were sampled on 29 June 2011; however, EP I and EP 3 did not have sufficient biomass to allow for collecting until mid-July. These sites were sampled on 12 July 2011.
Genotypic differentiation 2010-2011
An initial comparison of genetic differentiation was conducted on the EP 5 mid-summer 1 and mid-summer 2 collections to test whether potential genetic differences were due to some artifact of our sampling protocol. No significant difference in genotypic differentiation was found in either Genepop (P = 0.773) or Arlequin (P = 0.666), suggesting that our sampling protocol adequately represented the genetic variability found at that site (Table 4). The genotypic data for these two mid-summer collections were then combined and used for all other comparisons.
Investigations examining inter-annual genotypic differentiation within sites found that significant differentiation can exist between years for some sites, and that this pattern was consistent across multiple years (Table 4). Results from the program Bottleneck suggest that these significant changes were not due to a population bottleneck. All loci for all sampling sites were shown to fit the TPM model in 2010 (P [greater than or equal to] 0.348) and in 2011(P [greater than or equal to] 0.246), documenting that none of the sampled sites possessed the excess in genetic heterozygosity expected from a recent bottleneck event. Additionally, no substantial decrease in allelic diversity was observed in any site (Fig. 2).
For potential intra-annual genotypic differentiation within each site, comparisons found no significant difference between the early and mid-summer collection times for the majority of comparisons (Table 4). In contrast, the mid-to late summer tests showed evidence of differentiation within sites. Results from Genepop showed significant genotypic differentiation for each site, while the [F.sub.ST] value for EP 1 calculated in Arlequin was significantly different from zero (P = 0.025) (Table 4). Similar to the inter-annual genotypic differentiation within each site, the genetic differences between the mid- and late summer collection periods do not appear to be a result of a population bottleneck, as all loci for all sampling sites were shown to fit the TPM model (Bottleneck: P [greater than or equal to] 0.461) and there was no observable decrease in the allelic diversity for any site (Fig. 2).
As in the 2009 comparison, differentiation was found between all sites for the early 2010 collection. An overall calculation of [F.sub.ST] for Eel Pond was 0.034, which was significantly different from zero (P <0.0001). Results from Genepop and Arlequin showed significant differences in all pairwise comparisons (Table 5). To examine a change in genetic structure between collection sites throughout the reproductive season, newly settled individuals were preferentially selected over established colonies during the mid-summer collection period. The overall [F.sub.ST] for Eel Pond was found to decrease to a value of 0.020 (P = 0.001), documenting continued heterogeneity within Eel Pond. Results from pairwise comparisons did show decreased genotypic differentiation between some sites, but results from the majority of comparisons continued to be significant (Table 5). The collection sites designated EP 1 and EP 3 were no longer genetically differentiated, with nonsignificant results found from analyses conducted in both Genepop (P = 0.258) and Arlequin (P = 0.141). No trend toward homogeneity persisted, however, through the late summer collection period. The overall [F.sub.ST] for Eel Pond at this time was 0.024 (P < 0.0001), and results from Genepop and Arlequin showed continued significant genotypic differentiation in most pairwise comparisons (Table 5). For this collection time, the only nonsignificant comparison was the pairwise [F.sub.ST] calculated between EP 1 and EP 3 (P = 0.115).
Table 4 Results from pair-wise F.sub.ST] calculations and exact tests for genotypic differentiation fbr changes in genetic structure within collection sites in Eel Pond, Woods Hole, Massachusetts Site [F.sub.ST] Genotypic Differentiation Negative control EP5 mid-1 &- 2 0.005 15.07 Inter-annual comparisons 2009-2010 EP 1 0.015 36.22 EP 3 0.030 48.62 EP 4 0.011 43.85 EP 5 0.017 45.75 2010-2011 EP 1 0.032 46.72 EP 3 0.013 36.90 EP 4 0.021 44.05 EP 5 0.006 28.21 Intra-annual comparisons Early-Mid EP 1 0.005 20.32 EP 3 0.012 29.97 EP 4 0.009 26.58 EP 5 0.011 32.04 Mid-Late EP 1 0.019 40.45 EP 3 0.011 44.18 EP 4 0.015 38.29 EP 5 0.011 37.12 The negative control was conducted to ensure that an artifact of sampling of each site was not responsible for significant differences between sampling periods within sites. Inter-annual comparisons were conducted between groups of animals collected from each site in August 2009 and early summer 2010, and between late summer 2010 and early summer 2011. The late summer collection period in 2010 spanned 3 weeks (9/13/2010-10/4/2010) due to widespread Bugula stolonsfera colony die-off after the mid-summer collection period, possibly resulting from hurricane activity and increased freshwater influx to Eel Pond. Bolded values indicate significance ([alpha] = 0.05).
Previous work examining fine-scale genetic structure in the marine bryozoan Bugula stolonifera found high levels of mixing within conspecific aggregations (Johnson and Woollacott, 2010). A group of individuals occupying an area of up to 21 [cm.sup.2] possessed as much genetic variability as was found for an entire sampling site (area [approximately equal to] 15 [m.sup.2]). In the present study, we examined whether this greater than expected mixing within an aggregation also resulted in increased levels of genetic mixing among aggregations within Eel Pond. Results from collections conducted in 2009 found significant genotypic differentiation between most comparisons, suggesting that minimum genetic exchange existed between sites. This investigation was extended to 2010 and 2011, whereby potential changes in genetic heterogeneity within and between sampling sites throughout the reproductive season were examined. Results from these analyses established that some mixing could occur between sites, but that this mixing was not ubiquitous within Eel Pond. Indeed, sites separated by as little as 100 m showed no evidence of interbreeding, suggesting that significant barriers to genetic exchange can exist on small spatial scales. Analyses of temporal genetic structure documented that significant differentiation can occur inter-annually within sites, most likely due to the presumably random survival or differential growth of individuals during the annual die-back experienced by these animals. Taken together these results suggest that regardless of any low probability dispersal events that might lead to increased homogeneity between sites in Eel Pond, the overwintering strategy by these animals will likely lead to increased differentiation at the beginning of the next reproductive season.
Genetic variability within sampling sites
Modes of dispersal have long been known to affect population distribution and connectivity (Jackson, 1986; Gros-berg and Cunningham, 2001; Palumbi, 2004). For sessile organisms, the absence of a long dispersal phase could result in the accumulation of closely related individuals on a small spatial scale. Although a few bryozoans have planktonic larval development, the vast majority of species brood embryos within the colony and release short-lived larvae with limited dispersal potential (see Zimmer and Woollacott, 1977). As it has also been shown that extended swimming by nonfeeding larvae can result in decreased survival and postmetamorphic fitness (Woollacott et al., 1989; Wendt, 1996, 1998), it might be expected that these larvae will settle soon after release. The genetic signature deriving from this settlement pattern would be expected to deviate significantly from HWE, as was found in this study. All sampling sites deviated significantly from HWE in 2009, 2010, and 2011 due to a deficiency in heterozygotes (Table 1). Further, all sites had high inbreeding coefficients that were significantly different from zero. These data suggest that limited larval dispersal can result in the clumping of closely related kin, increasing the potential for inbreeding in these conspecific aggregations. Alternatively, similar deviations from HWE could result in populations that were isolated from receiving novel genes or recruits for long periods of time. Eel Pond is a relatively closed body of water with a single opening to Vineyard Sound (Fig. 1), and it might be expected that over time the population within Eel Pond could become increasingly inbred. It is unclear which mechanism is responsible for the observed deviations from HWE in our sampled sites, or if the two processes worked in concert. That the population sampled in neighboring Hadley Harbor, an open body of water that is free to mix with surrounding areas, was also outside of HWE suggests that limited larval dispersal alone could result in a population deviating significantly from HWE (Table 1).
Genotypic differentiation between sites
In addition to resulting in the formation of conspecific aggregations, limited larval dispersal has also been tightly coupled with genetic structure (e.g., Todd et al., 1998; Watts and Thorpe, 2006). Indeed, studies have shown that this structure can occur on small spatial scales for species with short-lived larvae (e.g., Yund and O'Neil, 2000; Calderon et al., 2007). In our study, results from analyses conducted in 2009 appear to support this assertion. Genotypic differentiation was found in most pairwise comparisons, even between sampling sites separated by only 100 m (e.g., EP 4 and EP 5) (Table 3). By examining intra-annual changes in genetic structure among sites in 2010, we were able to show that although some mixing was possible, sites within Eel Pond were not freely interbreeding. For instance, the site designated EP 4 showed no evidence of mixing with any other site during the study (Table 5), and when coupled with the high levels of genotypic differentiation, suggests that this site was isolated from other sites in Eel Pond. Importantly, the negative control conducted during the mid-summer 2010 collection documents that the observed pattern was not due to an artifact of sampling (Table 4). Rather, these results suggest that meaningful barriers to genetic exchange can exist on small spatial scales for these animals.
Table 5 Matrix of pairwise [F.sub.ST] values below diagonal arid results from exact tests for genotypic differentiation above diagonal for samples conducted in Eel Pond in early, mid-, and late summer 2010 EP 1 EP 3 EP 4 EP 5 Early summer 2010 EP 1 -- 38.43 70.60 41.01 EP 3 0.019 -- 88.92 76.64 EP 4 0.036 0.049 [varies] [varies] EP 5 0.024 0.055 0.060 -- Early summer 2010 EP 1 -- 23.62 [varies] 32.83 EP 3 0.012 -- 61.65 47.84 EP 4 0.043 0.030 -- [varies] EP 5 0.013 0.014 0.044 -- Late summer 2010 EP 1 -- 49.62 81.35 50.01 EP 3 0.012 -- 65.10 61.95 EP 4 0.038 0.030 -- 82.95 EP 1 0.026 0.017 0.045 -- Early summer animals were collected in June 2010. while mid-summer animals were collected in July 2010. The duration of the late summer collection was from 13 September to 4 October 2010 due to widespread colony die-off after the mid-summer collection period. Bolded values indicate significance ([alpha] = 0.05).
Although these Eel Pond sites do not appear to be freely interbreeding, there was some evidence that low levels of mixing were occurring between some sites. By the mid-summer collection period, EP 1 and EP 3 were no longer significantly differentiated, and comparisons of these two sites with EP 5 showed decreased values in results from analyses conducted in Arlequin and Genepop (Table 5). How this mixing occurred remains unclear. As the reproductive season progressed, an increasing number of colonies would become reproductively mature, resulting in an increase in larval output. It seems likely that this increase in larvae would also increase the likelihood of low-probability long-distance larval exchange between sites. Alternatively, this pattern of increased mixing could also have resulted from two other possibilities: long-distance sperm transfer or anthropogenically mediated dispersal.
Although fertilization is internal in bryozoans, sperm are released to the water column, a reproductive strategy that is relatively widespread among invertebrate taxa (Bishop and Pemberton, 2006). Animals that use this strategy are not only thought to have long-lived sperm, but also to be able to utilize sperm at very dilute concentrations. The longevity of sperm released by Bugula spp. has not been investigated, but sperm longevity has been reported for other bryozoans. Mandquez et al. (2001) reported a half-life of 1.2 h for sperm from the bryozoan Celleporella hyalina. Also, Yund and McCartney (1994) provided evidence of long-distance fertilization while conducting field-based mating assays with C. hyalina, as well as the colonial ascidian Botryllus schlosseri. Finally, Temkin (1994) reported that spermatozeugmata released from the bryozoan Membranipora membranacea experienced periods of quiescence while in the water column and only commenced strong undulating movements when contacting tentacles of a lophophore. It remains possible that the observed mixing documented in our study resulted not only from larval dispersal, but also from long-distance sperm dispersal, whereby sperm were released at one site and passively transported among other sites.
The second possibility is that the increased homogeneity observed between these sites occurred as a consequence of anthropogenically mediated dispersal. Eel Pond is a heavily trafficked area that houses many watercraft. Colonies of B. stolonifera routinely grow on boat hulls, and the transport of reproductively mature colonies among the sampling sites is certainly possible. If this were the case, however, it is unclear why the site designated EP 4 did not show the signs of increased mixing that were evident between EP 1 and EP 3 (Table 5). Additionally, the site designated EP 3, a permanent floating dock anchored within Eel Pond, is a nonfunctioning dock. Despite these factors, the role of anthropogenically mediated dispersal cannot be discounted completely, and the mechanism for the increased homogeneity between EP 1 and EP 3 remains uncertain.
Any increase in homogeneity among sites observed in the mid-summer collection period of 2010 was not continued through to the late summer collection period (Table 5), most likely due to the widespread colony die-back and regrowth observed after the mid-summer collection. For many invertebrates, fluctuations in environmental conditions have been shown to adversely affect adult and offspring survival. For instance, adult mortality due to decreased salinity from heavy rainfall was reported for the ascidians Ascidia nigra (Goodbody, 1962) and Corella willmeriana (Lambert, 1968). In fact, decreased salinity from heavy rain, flooding, or ice melt has been implicated in the mass mortality of numerous shallow marine invertebrate taxa (e.g., Goodbody, 1961, and references therein). Eel Pond did receive elevated levels of rainfall in late August and early September that could have resulted in the decline in adult biomass first observed on 6 September. The Woods Hole region received about 50 mm of rain from 23 through 26 August and an additional 70 mm of rain associated with Hurricane Earl on 4 September (http://water.weather.gov/precip/). Similar to the seasonal die-back experienced in the winter by B. stolonifera (see below), results from analyses seem to suggest that this mid-season die-back did not culminate in a population bottleneck. It could be, however, that because these animals experience a perpetual cycle of bottleneck and population expansion due to the annual die-back, the number of individuals surviving a mass mortality event would have to be minuscule for a bottleneck to be detectable. However, there was no observable decrease in the index of allelic diversity for any site following a die-back event (Fig.2), further suggesting that these events do not culminate in the massive die-off of individuals. Regardless, the presumably random survival of individuals during the mid-season die-back or the differential growth of individuals after the event resulted not only in significant genotypic differentiation within each site between the mid- and late summer collection periods (Table 4), but also in increased genotypic differentiation among sampling sites in the late summer collection compared to the mid-summer collection (Table 5).
The patterns of genetic structure found in this study also illustrate the problem of conducting population genetic surveys using a single sampling period, as we did in 2009. For species that have short generation times (e.g., Dybern, 1965; Wendt, 1996; Burton, 1997; Massaro and Rocha, 2008), intra-annual changes in population genetic structure are possible. Hence, multiple collections over time are needed to fully understand the population dynamics of these types of species.
Implications of seasonal colony die-back
Although the reproductive season for B. stolonifera (roughly June--Oct.) in Eel Pond results in high densities of adults and thick mats of colonies where the animal occurs, the winter months see a decline in animal abundance, culminating in the absence of the erect portions of colonies at or just below the water line in most areas. How these animals overwinter to repopulate in the summer months remains unclear. One possibility is that a standing stock of animals survives at depth, presumably below each of our collection sites. Even though Eel Pond routinely freezes over in winter, colonies found at depth might be able to survive. The sandy bottom of Eel Pond is, however, not appropriate for attachment by B. stolonifera. In addition, neither an adult nor any suitable attachment substrate was found on visual inspection of the bottom under the MBL Pier (EP 5) in 2009 (unpubl. obs.). It seems unlikely, therefore, that the summer population arises from a standing stock of animals beneath the surtace. Alternatively, it has been proposed that bryozoans might overwinter via root-like projections (Numakunai, 1967) that emanate from various parts of the adult colony and attach to the substrate to assist in anchoring. These projections can also bud additional zooids that, through asexual reproduction, can develop into a genet. Hence, even if the erect portion of the colony dies off each winter as the temperature of Eel Pond decreases, the root-like projection could remain attached to the substrate. When favorable conditions return, a zooid could form that eventually develops into an adult. Results from analyses suggest that this overwintering strategy does not result in a population bottleneck, nor does it result in a decrease in allelic diversity (Fig. 2). Rather, the random survival of individuals after this die-back can produce significant inter-annual genotypic differentiation within collection sites (Table 4). Further, it could also lead to the high levels of genotypic differentiation observed among sites (Tables 3 and 5). Hence, any homogeneity between sites resulting from low-probability long-distance genetic mixing during a reproductive season will likely be lost at the beginning of the next season due to the annual cycle of die-back and regrowth.
We are indebted to Edward Enos and Rhys Probyn (both of the Marine Biological Laboratory, Woods Hole) for their assistance in animal procurement. We also thank Peter Girguis and Gonzalo Giribet (both of Harvard University) for technical advice, and we are grateful to Matthew Klooster (Center College of Kentucky) and Sonia Andrade (Harvard University) for their assistance in analyses. Scott Edwards, Helene Ferranti, James McCarthy (all of Harvard University). Per Palsboll (University of Groningen), and two anonymous reviewers provided helpful comments improving this manuscript. This research was supported by funds from Harvard University.
Received 27 September 2011; accepted 9 April 2012.
* To whom correspondence should be addressed. E-mail: email@example.com
Abbreviations: AMOVA, analysis of molecular variance; HWE, Hardy-Weinberg equilibrium; TPM, two-phase model of mutation.
Ayre, D. J., and T. P. Hughes. 2000. Genotypic diversity and gene flow in brooding and spawning corals along the Great Barrier Reef, Australia. Evolution 54: 1590-1605.
Belkhir, K., P. Borsa, L. Chikhi, N. Raufaste, and F. Bonhomme. 1996-2004. GENETIX 4.05, logiciel sous Windows TM pour la genetique des populations. Laboratoire Genome, Populations, Interactions, CNRS UMR 5000, Universite de Montpellier II, Montpellier, France.
Bishop, J. D. D., and A. J. Pemberton. 2006. The third way: spermcast mating in sessile marine invertebrates. Integr. Comp. Biol. 46: 398-406.
Bohonak, A. J. 1999. Dispersal, gene flow, and population structure. Q. Rev. Biol. 74: 21-45.
Bruno, J. F. 1998. Fragmentation in Madracis mirabilis (Duchassaing and Michelotti): How common is size-specific fragment survivorship in corals? J. Exp. Mar. Biol. Ecol. 230: 169-181.
Burton, R. S. 1997. Genetic evidence for long term persistence of marine invertebrate populations in an ephemeral environment. Evolution 51: 993-998.
Calderon, I., N. Ortega, S. Duran, M. Becerro, M. Pascual, and X. Turon. 2007. Finding the relevant scale: clonality and genetic structure in a marine invertebrate (Crambe crambe, Porifera). Mot Ecol. 16: 1799-1810.
Cornuet, J. M., and G. Luikart. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014.
Crisp, D. J. 1978. Genetic consequences of different reproductive strategies in marine invertebrates. Pp. 257-273 in Marine Organisms: Genetics, Ecology, and Evolution, B. Battaglia and J. A. Beardmore, eds. Plenum Press, New York.
Dupont, L., F. Viard, and J. D. D. Bishop. 2007. Combined effects of bottlenecks and selfing in populations of Corella eumyota, a recently introduced sea squirt in the English Channel. Divers. Distrib. 13: 808-817.
Dybern, B. I. 1965. The life cycle of Ciona intestinalis (L.) 1. typica in relation to the environmental temperature. Oikos 16: 109-131.
Excoffier, L., and H. E. Lischer. 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics under Linux and Windows. Mot Ecol. Resour. 10: 564-567.
Goodbody, I. 1961. Mass mortality of a marine fauna following tropical rains. Ecology 42: 150-155.
Goodbody, I. 1962. The biology of Ascidia nigra (Savigny). 1. Survival and mortality in an adult population. Biol. Bull. 122: 40-51.
Grosberg, R., and C. W. Cunningham. 2001. Genetic structure in the sea: from populations to communities. Pp. 61-84 in Marine Community Ecology, M. D. Bertness, S. D. Gaines, and M. E. Hay, eds. Sinauer Associates, Sunderland, MA.
Jackson, J. B. C. 1986. Modes of dispersal of clonal benthic invertebrates: consequences for species' distributions and genetic structure of local populations. Bull. Mar. Sci. 39: 588-606.
Johnson, C. H. 2010. Effects of selling on offspring survival and reproduction in a colonial simultaneous hermaphrodite (Bugula stolonifera, Bryozoa). Biol. Bull. 219: 27-37.
Johnson, C. H., and R. M. Woollacott. 2010. Larval settlement preference maximizes genetic mixing in an inbreeding population of a simultaneous hermaphrodite (Bugula stolonifera, Bryozoa). Mot Ecol. 19: 5511-5520.
Jones, G. P., S. Planes, and S. R. Thorrold. 2005. Coral reef fish larvae settle close to home. Cult Biol. 15: 1314-1318.
Keough, M. J., and H. Chernoff. 1987. Dispersal and population variation in the bryozoan Bugula neritina. Ecology 68: 199-210.
Knowlton, N., and B. D. Keller. 1986. Larvae which fall far short of their potential: highly localized recruitment in an alpheid shrimp with extended larval development. Bull. Mar. Sci. 39: 213-223.
Lambert, G. 1968. The general ecology and growth of a solitary ascidian, Corella willmeriana. Biol. Bull. 135: 296-307.
Levin, L. A. 2006. Recent progress in understanding larval dispersal: new directions and digressions. Integr. Comp. Biol. 46: 282-297.
Lewis, P. 0., and D. Zaykin. 2002. Genetic data analysis: a computer program for the analysis of allelic data. [Online]. Available: http://hydrodictyon.eeb.uconn.edu/people/plewis/software.php [2012, May 7].
Maier, E., R. To'Irian, and B. Niirnberger. 2009. Fine-scale analysis of genetic structure in the brooding coral Seriatopora hystrix from the Red Sea. Coral Reefs 28: 751-756.
Manriquez, P. H., R. N. Hughes, and J. D. D. Bishop. 2001. Age-dependent loss of fertility in water-borne sperm of the bryozoan Celleporella hyalina. Mar. Ecol. Prog. Ser. 224: 87-92.
Marshall, U. J., J. A. Pechenik, ana M. J. .Keough. Larval activity levels and delayed metamorphosis affect post-larval performance in the colonial ascidian Diplosoma listerianum. Mar. Ecol. Pros. Ser. 246: 153-162.
Massaro, F. C., and 0. Rocha. 2008. Development and population growth of Hydra viridissima Pallas, 1766 (Cnidaria, Hydrozoa) in the laboratory. Braz. J. Biol. 68: 379-383.
Miller, K. J., and D. J. Ayre. 2008. Population structure is not a simple function of reproductive mode and larval type: insights from tropical corals. J. Anim. Ecol. 77: 713-724.
Numakunai, T. 1960. An observation on the budding of the stolon of a bryozoan, Bugula neritina Linne Bull. Mar. Biol. Stn. Asamushi 10: 99-101.
Numakunai, T. 1967. Histological observations on the budding of the over-wintering stolon of a bryozoan, Bugula neritina L. Sci. Rep. Tohoku Univ. 4 Ser. (Biol.) 33: 499-508.
Palumbi, S. it 2001. The ecology of marine protected areas. Pp. 509-530 in Marine Community Ecology, M. D. Bertness, S. D. Gaines, and M. E. Hay, eds. Sinauer Associates, Sunderland, MA.
Palumbi, S. It 2004. Marine reserves and ocean neighborhoods: the spatial scale of marine populations and their management. Annu. Rev. Environ. Resour. 29: 31-68.
Pawlik, J. it 1992. Chemical ecology of the settlement of benthic marine invertebrates. Pp. 273-335 in Oceanography and Marine Biology: An Annual Review, Vol. 30, M. Barnes, A. D. Ansell, and R. N. Gibson, eds. University College London Press, London.
Peakall, R., and P. E. Smouse. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6: 288-295.
Pechenik, J. A. 2006. Larval experience and latent effects--metamorphosis is not a new beginning. Integr. Comp. Biol. 46: 323-333.
Raymond, M., and F. Rousset. 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J. Hered. 86: 248 -249.
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225.
Shanks, A. L., B. A. Grantham, and M. H. Carr. 2003. Propagule distance and the size and spacing of marine reserves. Ecol. Appl. 13: 5159-S169.
Strathmann, R. R. 1985. Feeding and nonfeeding larval development and life-history evolution in marine invertebrates. Annu. Rev. Ecol. Syst. 16: 339-361.
Swearer, S. E., J. S. Shima, M. E. Hellberg, S. R. Thorrold, G. P. Jones, D. R. Robertson, S. G. Morgan, K. A. Selkoe, G. M. Ruiz, and R. R. Warner. 2002. Evidence of self-recruitment in demersal marine populations. Bull. Mar. Sci. 70: 251-271.
Temkin, M. H. 1994. Gamete spawning and fertilization in the gymnolaemate bryozoan Membranipora membranacea. Biol. Bull. 187: 143-155.
Thorson, G. 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol. Rev. 25: 1-45.
Todd, C. D., W. J. Lambert, and J. P. Thorpe. 1998. The genetic structure of intertidal populations of two species of nudibranch molluscs with planktotrophic and pelagic lecithotrophic larval stages: Are pelagic larvae "for" dispersal? J. Exp. Mar. Biol. Ecol. 228: 1-28.
Underwood, J. IN., L. D. Smith, J. H. Madeleine, H. van Oppen, and J. P. Gilmour. 2009. Ecologically relevant dispersal of corals on isolated reefs: implications for managing resilience. Ecol. Appl. 19: 18-29.
van Oosterhout, C., W. F. Hutchinsons, D. P. M. Wills, and P. Shipley. 2004. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol. Ecol. Notes 4: 535-538.
Warner, R. R., and R. K. Cowen. 2002. Local retention of production in marine populations: evidence, mechanisms, and consequences. Bull. Mar. Sci. 70: S245--S249.
Watts, P. C., and J. P. Thorpe. 2006. Influence of contrasting larval developmental types upon the population-genetic structure of cheilostome bryozoans. Mar. Biol. 149: 1093-1101.
Watts, P. C., J. P. Thorpe, and P. D. Taylor. 1998. Natural and anthropogenic dispersal mechanisms in the marine environment: a study using cheilostome Bryozoa. Philos. Trans. R. Soc. B 353: 453464.
Wendt, D. E. 1996. Effect of larval swimming duration on success of metamorphosis and size of the ancestrular lophophore in Bugula neritina (Bryozoa). Biol. Bull. 191: 224-233.
Wendt, D. E 1998. Effect of larval swimming duration on growth and reproduction of Bugula neritina (Bryozoa) under field conditions. Biol. Bull. 195: 126-135.
Wendt, D. E. 2000. Energetics of larval swimming and metamorphosis in four species of Bugula (Bryozoa). Biol. Bull. 198: 346-356.
Wendt, D. E., and R. M. Woollacott. 1999. Ontogenies of phototactic behavior and metamorphic competence in larvae of three species of Bugula (Bryozoa). Invertebr. Biol. 118: 75-84.
Woollacott, R. M., J. A. Pechenik, and K. M. Imbalzano. 1989. Effects of duration of larval swimming period on early colony development in Bugula stolonifera (Bryozoa: Cheilostomata). Mar. Biol. 102: 57-63.
Yund, P. 0., and M. A. McCartney. 1994. Male reproductive success in sessile invertebrates: competition for fertilizations. Ecology 75: 2151-2167.
Yund, P. 0., and P. G. O'Neil. 2000. Microgeographic genetic differentiation in a colonial ascidian (Botryllus schlosseri) population. Mar. Biol. 137: 583-588.
Zimmer, R. L., and R. M. Woollacott. 1977. Structure and classification of gymnolamaete larvae. Pp. 57-89 in Biology of Bryozoans, R. M. Woollacott and R. L. Zimmer, eds. Academic Press, New York.
COLLIN H. JOHNSON AND ROBERT M. WOOLLACOTT
Department of Organismic and Evolutionary Biology and the Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138
|Printer friendly Cite/link Email Feedback|
|Author:||Johnson, Collin H.; Woollacott, Robert M.|
|Publication:||The Biological Bulletin|
|Date:||Jun 1, 2012|
|Previous Article:||Annual variation in the timing of coral spawning in a high-latitude environment: influence of temperature.|
|Next Article:||Comparative analysis of movement characteristics of lancelet and fish spermatozoa having different morphologies.|