Microsatellites have proven their use as powerful markers for population and clonality studies in the sponge Scopalina lophyropoda
. The differentiation statistics used revealed that the genetic diversity was spatially structured in this rare species at the three spatial scales analyzed (as revealed by the FST
values and the AMOVA). Most diversity was found within the populations, but differentiation among populations and among geographic regions was also significant. The Bayesian clustering method (STRUCTURE), however, turned out to be more conservative and only detected three genetically independent groups, which correspond to the three geographical regions. However, when the polymorphism shown by the selected markers is limited, as it occurs in sponge microsatellites, or there is a strong isolation by distance pattern, the STRUCTURE algorithm is limited to resolve population clusters [33
], and has been reported to provide a lower resolution than differentiation statistics [34
]. Thus, considering all other results (Mantel test, AMOVA, FST
values, and the presence of private alleles), we can conclude that the sampled populations are genetically well structured, though some gene flow may take place within regions.
The positive regression between genetic differentiation and geographic distance between populations of this species revealed distance isolation and short-range effective dispersal. This pattern agrees with the philopatric behavior reported for S. lophyropoda
], which, in combination with the strict habitat requirement of the species [13
], could contribute to its extremely fragmented distribution. Environmental stress is likely to be stronger in small fragmented populations than in large populations [37
], and it has been reported to lead to rapid genetic divergence [38
]. The later studies have suggested that the intensity and direction of the selection could differ among small populations [37
] and lead to speciation. The presence of several cryptic Scopalina
species in the distribution area of S. lophyropoda
] could be interpreted in that way.
Gene flow can be maintained by dispersal of sexual and/or asexual propagules. In S. lophyropoda
, the contribution of clonality to the populations was minor. Only 4% (9 out of 222) of the ramets analyzed can be assumed to be the result of asexual reproduction. A plausible explanation is that fissions are balanced with fusions between clone-mates [30
], since clones remain attached to the rocky substrate and can contact each other when growing.
Since asexual dispersal appears to be unimportant in the species and the few clones identified use to be found at smaller distances than 0.5 m [32
], gene flow between the populations depends on the dispersal of sexual products: either larvae, or sperm. Unfortunately, little is known about the dispersal of sponge sperm in general, and nothing is known about S. lophyropoda
sperm in particular but the reported short-range dispersal [22
] of S. lophyropoda
larvae, fully agrees with the strong genetic structure in the species. The strong genetic structuring (AMOVA, FST
values and Mantel test results) indicates that there is a current absence of connectivity among the regions but STRUCTURE results point to some, although restricted, gene flow among the populations within a region, which may occur through sporadic stochastic dispersal events. A combined sexual-asexual dispersal mechanism consisting in asexual fragments of ripe individuals containing larvae has been speculated [31
], which, if it occurs in the field, could enhance gene flow over longer distances than those suggested by the observed larval behavior in the field.
Bottleneck tests under IAM and TPM did not reveal any recent reduction in size of the populations, which is in agreement with the growth detected during the last 20 years in the several populations studied (authors' obs.). Only under the less probable mutation model for the microsatellites (SMM) [39
], the bottleneck analysis indicated recent reductions in size in some populations. Although we do not have information on the mutation model of S. lophyropoda
microsatellites, some studies in other taxa (e.g. humans, fish, insects) reported that microsatellites fit the mixed TPM or the IAM, but not the SMM [39
]. Thus, field monitoring, the best models for microsatellite mutation, and the presence of private alleles in all populations (rare alleles are the first that disappear after a bottleneck), indicate that if any size reduction has occurred, it was not recently.
Most marine invertebrates with non-feeding and short dispersed larvae present genetically structured populations and inbreeding, whatever the Phylum they belong to [2
]. As for sponges, only one study of population genetics using microsatellites is available up to now (Crambe crambe
]). Some population traits of S. lophyropoda
are shared with those reported for C. crambe
, whereas other aspects are contrasting. The populations are genetically structured in both species, despite their respective patchy vs
. continuous distributions and the differences in larval behavior [28
]. Restricted dispersal seems to be the main factor responsible for the isolation by distance that is found in the two sponges. However, the contribution of clonality to the population makeup was rather insignificant in S. lophyropoda
(4% on average) while it reached 23% in C
]. There are two possible explanations for the differences found between both species. First, in the study on C. crambe
, several individuals did not amplify for each locus, which may contribute to a failure in differentiation of some individuals that were not real clones. On the other hand, fusion rates between ramets (also much more frequent in S. lophyropoda
than in C. crambe
) are more balanced in the former with fission rates and thus could account for the low extent of clonality in our "snapshot" sampling.
Moreover, all C. crambe
populations showed inbreeding [2
], as expected in long-lived, low-dispersing organisms due to temporary Wahlund effects [46
]. Furthermore, individuals in small populations usually present inbreeding, which can increase their extinction probability by decreasing their potential for adaptation [39
], especially when fragmented [14
]. However, contrary to most predictions, we show here that S. lophyropoda
has no inbreeding depression and exhibits heterozygote excess for most of the loci studied. In fact, our results contrast with theory and experimental evidences, which suggest that the effective size of small populations, and its response to selection, decreases with time [48
]. The unexpected genetic results in the S. lophyropoda
populations, however, support studies that predict that inbreeding should not be that common in marine invertebrates, despite their restricted dispersal, strong population structure, and consanguinity (e.g.,[49
]). Our study provides indirect evidence on the existence of genetic strategies for maintaining the high genetic diversity even when the populations are isolated by distance and there is restricted gene flow among the populations. Selection against homozygote and/or for outcrossing, extreme genet longevity combined with stochastic recruitment events from other populations [51
] are among the strategies that may allow S. lophyropoda
and other rare, marine, benthic invertebrates to persist in a fragmented scenario.