Contrasting genetic structure of rear edge and continuous range populations of a parasitic butterfly infected by <it>Wolbachia</it>

Abstract Background Climatic oscillations are among the long-term factors shaping the molecular features of animals and plants and it is generally supposed that the rear edges (i.e., the low-latitude limits of distribution of any given specialised species) situated closer to glacial refugia are vital long-term stores of genetic diversity. In the present study, we compared the genetic structure of several populations of an endangered and obligate myrmecophilous butterfly (Maculinea arion) from two distinct and geographically distant parts of its European distribution (i.e., Italy and Poland), which fully represent the ecological and morphological variation occurring across the continent. Results We sequenced the COI mitochondrial DNA gene (the ‘barcoding gene’) and the EF-1α nuclear gene and found substantial genetic differentiation among M. arion Italian populations in both markers. Eleven mtDNA haplotypes were present in Italy. In contrast, almost no mtDNA polymorphisms was found in the Polish M. arion populations, where genetic differentiation at the nuclear gene was low to moderate. Interestingly, the within-population diversity levels in the EF-1α gene observed in Italy and in Poland were comparable. The genetic data did not support any subspecies divisions or any ecological specialisations. All of the populations studied were infected with a single strain of Wolbachia and our screening suggested 100% prevalence of the bacterium. Conclusions Differences in the genetic structure of M. arion observed in Italy and in Poland may be explained by the rear edge theory. Although we were not able to pinpoint any specific evolutionarily significant units, we suggest that the Italian peninsula should be considered as a region of special conservation concern and one that is important for maintaining the genetic diversity of M. arion in Europe. The observed pattern of mtDNA differentiation among the populations could not be explained by an endosymbiotic infection.