HIGH GENE-FLOW BETWEEN DIANTHUS SYLVESTRIS POPULATIONS FROM SERPENTINE AND LIMESTONE SUGGESTS AN OLIGOGENIC NATURE OF LOCAL ADAPTATION

Local adaptation and gene flow are the two sides of the same coin: while gene flow tends to homogenize the allele frequencies, decreasing the variability between populations, local adaptation acts selecting and promoting those alleles better performing in each environment, thus increasing differences among populations. In the ecological speciation, local adaptation has a central role in initiating population divergence as natural selection on traits that confer ecological divergence could also produce, as a by-product, divergent selection on the traits that confer reproductive isolation, such as, for instance, flowering time differences (1). However, although divergent selection can favour reproductive isolation at a local scale, incipient divergence can be prevented by the homogenising effect of gene flow between adjacent populations, maintaining species cohesion (2). Investigations into transient stages in the continuum between local adaptation and incipient speciation are integral to speciation research. Serpentine soils represent an excellent system for investigating the contribution of local adaptation to plant speciation as indirectly testified by the striking levels of endemism and the distinctive flora they possess relative to surrounding areas. The main characteristics of serpentine soils, defined “serpentine syndrome”, are low amounts of calcium opposed to high amounts of magnesium, relatively high concentrations of nickel, chromium and other heavy metals, and low water retention (3). Thus, adaptation to serpentine soils can contribute indirectly to pre- or post-zygotic reproductive barriers that genetically isolate serpentine populations from non-serpentine relatives. Additionally, the patchy distribution of serpentine soils can contribute to the geographic isolation of populations by limiting the homogenizing effect of inter-population gene flow. Here we investigated the amount of gene-flow and genetic divergence among populations of Dianthus sylvestris (Caryophyllaceae) occurring both on serpentine outcrops and adjacent limestone sites in order to highlight if local adaptation enhanced the genetic divergence among populations. Plants from serpentine outcrops and calcareous soils were collected from populations of north-centre of Italy and levels of population divergence and genetic connection were determined with molecular markers. Alleles of 10 microsatellites loci (SSR) derived from an EST library were characterized. All populations showed an high percentage of polymorphic loci (ranging from 71% to 100%) and the distribution of allele frequencies showed no significant differences among populations. Similarly allele richness was comparable among populations of serpentine and limestone. The M-ratio approach ruled out that the populations, even if small in size, have been subject to recent bottleneck, although the positive values of the inbreeding index (0.172<Fis<0.2894) suggest the occurrence of crosses among related individuals. Both ANOVA (86% of the variability within populations, 14% between populations) and the low values of differentiation among populations (mean Fst= 0.119, among populations) confirmed the low overall genetic differentiation with the notable exception of a single population from Val di Susa, which showed higher values of Fst (mean Fst = 0.3194, between Val di Susa and other populations). In particular the locus DSL16 was found more strongly differentiated than expected by chance (Fst = 0.713, p≤0.05), consistent with a scenario of divergent selection. Nevertheless this locus was not found as outlier in any Fst estimation between serpentine and limestone populations. Bayesian (STRUCTURE) and multivariate approach (PCoA) ruled out that populations of limestone and serpentine soils cluster in two genetically differentiated groups and Mantel test did not identify a significant correlation with geographic distance. Thus, in contrast with several previous studies, D. sylvestris does not show any evident genetic differentiation among populations from serpentine and limestone. Instead, these likely behave as a meta-population, with the intensity of gene flow apparently masking any potential effect of local adaptation. This suggests, for D. sylvestris, that adaptation to serpentine soil may have an oligogenic basis, such as that local divergence in the few genes conferring adaptation can be maintained in spite of extensive gene-flow among demes.