Mitochondrial phylogeography of the Holarctic Parnassius phoebus complex supports a recent refugial model for alpine butterflies

Aim Our study provides a description of the mitogenetic structure of alpine butterflies of the Parnassius phoebus complex throughout their Holarctic distribution. Our analyses extend and reassess population history models for alpine butterflies under an explicit calibration of their mitochondrial DNA (mtDNA) substitution rate. Location Mountain ranges of the Holarctic region. Methods A fragment (824 bp) of the mitochondrial cytochrome c oxidase subunit I (COI) gene was sequenced in 203 samples (72 locations), and combined with previously available COI sequences (499 samples), to obtain full coverage of the Holarctic distribution of the P. phoebus complex. A global species distribution model (SDM) was calculated by the maximum entropy (Maxent) approach, allowing assignment of samples into geographically consistent operational units. Phylogenetic and coalescent methods were applied to describe the global mitogenetic structure and estimate population genetics parameters. Geological and palaeoecological evidence was used for internal calibration and validation of a COI substitution rate. Results Eurasian (including Alaskan) and North American populations form two distinct mitochondrial clades. The mitochondrial time to most recent common ancestor (TMRCA) of the North American clade was estimated at less than 125 ka, and the TMRCA of the EurasianAlaskan clade at less than 80 ka, except for a single divergent sequence from Mongolia. Pairwise divergence times between all geographical units within each continent date well within the last 100 ka, and most likely, the last 5010 ka. Main conclusions In contrast with its currently scattered distribution within each of Eurasia and North America, the mitogenetic structure of the P. phoebus complex in both continents is shallow and weak, and shows no evidence of geographical structure dating back earlier than the last glacial cycle. We argue that mtDNA data are consistent with recent (Wurm/Wisconsin) range expansion across each of the two continents and with persistent glacial long-range gene flow which ceased during the Holocene. We propose that P. phoebus may represent a model for Holarctic alpine invertebrates with moderate dispersal abilities in that its genetic structure at a continental scale reflects extensive connectivity during the most recent glacial phases.