From species divergence to population structure: A multimarker approach on the most basal lineage of Salamandridae, the spectacled salamanders (genus Salamandrina) from Italy

Abstract

The Apennine Peninsula is one of Europe’s main glacial refugial areas and harbors a large number of lineages and species. Here, a pattern of higher genetic diversity in the south compared to that of the north is characteristic of most vertebrates; however, most studies that have produced these results have relied only on inferences based on mitochondrial DNA. The spectacled salamanders (genus Salamandrina) are endemic to the Apennine Peninsula and have diverged into two sibling species: S. terdigitata (in the south) and S. perspicillata (in the north), presumably in the late Miocene or early Pliocene. By sequencing one mitochondrial (cytb) and two nuclear genes (RAG1 and POMC) and genotyping 10 microsatellite loci, we traced the evolution of these sibling species from their divergence to their contemporary population structure at a fine scale. Using a multilocus coalescent-based approach, we estimated the temporal divergence of both species at approximately 2.25mya (million years ago), which, hence, is much younger than previous estimates. The classical pattern of high genetic diversity in the south and lower diversity in the north was confirmed only for some markers, and the demographic histories of the two species differed substantially. Whereas S. perspicillata (north) expanded from a single major refugium in the center of the Apennine Peninsula, populations of S. terdigitata (south) persisted through cooler periods in multiple refugia. Further, the fine-scale population genetic structure of 16 S. perspicillata populations revealed significant genetic differentiation, even across short geographic distances. The results of our study stress that for a better understanding of phylogeographic patterns and past demographic processes, both mitochondrial and multiple nuclear loci should be analyzed to avoid gene-specific, and possibly biased results.