Zusammenfassung
Starting from Western Europe, the house mouse (Mus musculus domesticus) has spread across the globe in historic times. However, most of the southern oceanic islands were colonized by mice only within the past 300 years. This makes them an excellent model for studying the evolutionary processes during early stages of new colonization and for understanding mechanisms of early adaptation. Twenty-four microsatellite loci and the mitochondrial D-loop sequence were typed in a total of 534 mice mainly from the Kerguelen Archipelago but also from Falkland Islands, Marion Island, Amsterdam Island, Antipodes Island, Macquarie Island, Auckland Islands, and one sample from South Georgia. Although there was heavy ship traffic for over a hundred years to the Kerguelen Archipelago, it appears that only the mice that have arrived first have colonized the main island (Grande Terre) and most of the associated small islands indeed mice shared the same D-loop haplotype as well as the same Y chromosomal haplotype. The second mice invasion has occurred on islands that are detached from Grande Terre (Cimetière Island and Cochons Island) and were likely to have had no resident mice prior to their arrival. They displayed a different mitochondrial D-loop haplotype and were genetically distinct in the autosomes. However, the Y chromosome haplotype was related to the one found in Grande Terre, suggesting that they came from a related source population. These data suggest that an area colonized by mice is refractory to further introgression, possibly due to fast adaptations of the resident mice to local conditions. Interestingly, single step mutational derivatives of one of the major mitochondrial haplotypes were found several times in different southern hemisphere islands, suggesting an unusually high mutation rate or the putative presence of a selective sweep in the mitochondrial genome. In order to investigate further the genetic basis of adaptation on southern hemisphere islands a genome-wide microsatellite screen for selective sweeps was performed. 737 markers dispersed around the entire genome were typed in pooled samples from Kerguelen Archipelago populations and compared to European populations in order to pre-selected candidate loci for selective sweeps. A total of 38 pre-selected candidates’ loci were then individually typed from five different islands (Kerguelen Archipelago, Marion Island, Auckland Island, Marion Island, Antipodes Island) representing six mouse populations in order to identify genomic regions displaying similar patterns and to decrease the impact of the demographic history of the island on the hitchhiking mapping approach used. Five microsatellite candidate loci, all of them associated with a gene, were identified. Interestingly, one of the candidates has a known function during parasite infection. Another candidate gene is a sub-unit of a K+ channel and, surprisingly, the second sub-unit of this K+ channel was picked up during the less stringent pre-selection of the 38 loci, pointing to the importance of this channel for mouse adaptation to their new environment. Additional experiments are required in order to confirm these five candidate genes, but nevertheless this study demonstrates that a genome-wide microsatellite locus screen is a valuable approach to identify genes which are implicated in recent adaptations.
