Evolutionary and dispersal history of Eurasian house mice Mus musculus clarified by more extensive geographic sampling of mitochondrial DNA

Abstract

We examined the sequence variation of mitochondrial DNA control region and cytochrome b gene of the house mouse (Mus musculus sensu lato) drawn from ca. 200 localities, with 286 new samples drawn primarily from previously unsampled portions of their Eurasian distribution and with the objective of further clarifying evolutionary episodes of this species before and after the onset of human-mediated long-distance dispersals. Phylogenetic analysis of the expanded data detected five equally distinct clades, with geographic ranges of northern Eurasia (musculus, MUS), India and Southeast Asia (castaneus, CAS), Nepal (unspecified, NEP), western Europe (domesticus, DOM) and Yemen (gentilulus). Our results confirm previous suggestions of Southwestern Asia as the likely place of origin of M. musculus and the region of Iran, Afghanistan, Pakistan, and northern India, specifically as the ancestral homeland of CAS. The divergence of the subspecies lineages and of internal sublineage differentiation within CAS were estimated to be 0.37-0.47 and 0.14-0.23 million years ago (mya), respectively, assuming a split of M. musculus and Mus spretus at 1.7 mya. Of the four CAS sublineages detected, only one extends to eastern parts of India, Southeast Asia, Indonesia, Philippines, South China, Northeast China, Primorye, Sakhalin and Japan, implying a dramatic range expansion of CAS out of its homeland during an evolutionary short time, perhaps associated with the spread of agricultural practices. Multiple and non-coincident eastward dispersal events of MUS sublineages to distant geographic areas, such as northern China, Russia and Korea, are inferred, with the possibility of several different routes.

Figure 1

Collection of localities and mitochondrial genotypes in Eurasia of M. musculus samples examined in this study (a). New samples genotyped for this study are shown. Detailed locality names and sample codes are listed in Supplementary Table 1. Five major mitochondrial groups representing five subspecies groups, M. m. musculus (blue: MUS), M. m. domesticus (red: DOM), and M. m. castaneus (yellow: CAS), M. m. gentilulus (white: GEN) and the divergent lineage occurring in Nepal (orange: NEP) are differentially shown. The specific haplotype group of DOM that broadly dispersed to a variety of countries (Australia, Canada, China, Germany, Indonesia, Senegal, Somalia) are marked with arrowheads. Together with those from Prager et al. (1998), spatial patterns for the mitochondrial genotypes are shown for mice from Central Asia based on combination of new and previously published sequences (sources) (b), where further sub-division of the CAS lineage into four (CAS-1, CAS-2, CAS-3, CAS-4) are detected. The types of the four sub-groups of CAS are shown in circle with numerical numbers (black, Prager et al., 1998; red, in this study). Further sub-division of the MUS lineages into two, MUS-1 (light blue) and MUS-2 (dark blue), and the MUS-1 sublineage into three (MUS-1a, MUS-1b, MUS-1c) is suggested in this study (a and c).

Figure 2

NN networks tree based on the cytochrome b gene (Cytb; a, f and h) and control region (CR; b–d, e and g) of the mitochondrial DNA, with tip labels for the three major subspecies groups, M. m. musculus (MUS), M. m. castaneus (CAS) and M. m. domesticus (DOM) and two rather geographically confined groups of M. m. gentilulus (GEN) and Nepalease mice (NEP). The portion of the CR network was enlarged to show the details of the branching patterns for CAS-2, in which most of members possess a 75-bp repeat (d). The codes for the HGs in the CR (g) and Cytb (h) network for DOM were taken from those used in Bonhomme et al. (2011).

Figure 3

ML trees for mitochondrial DNA sequences of the cytochrome b gene (a) and control region (b). The PhylML algorithm (Guindon and Gascuel, 2003) was used for the tree reconstruction and bootstrap analysis (100 replications). Bootstrap values (>50%) are shown under basal branches.

Figure 4

ML tree for concatenated mitochondrial DNA haplotypes (control region and cytochrome b gene) using representatives for the four major HGs of M. musculus and M. macedonicus as outgroup. Bootstrap values (>50%) are shown under basal branches (ML/MP/NJ).

Figure 5

NN networks of concatenate sequences of control region and cytochrome b gene (ca. 2020 bp) from individuals representing the sublineage of CAS, CAS-1 (a) and MUS (b); individuals from Korea and Japan are marked with * and **, respectively. Prominent sub-groups appeared in the networks are indicated.

Figure 6

Divergence time estimates (million years ago, mya) of M. musculus phylogroups and its closely related species, based on a Bayesian-relaxed molecular clock applied to the mitochondrial cytochrome b sequences (1140 bp). The posterior probability and 95% highest posterior density intervals of node ages in mya (gray bars) are shown in particular nodes with ancient divergent. The time estimates of 1.7 mya for the root node of the divergence of M. spretus and the other species of M. musculus Species Group (Suzuki et al., 2004) was used as calibration point. Sequences obtained from the databases are marked with their accession numbers and asterisks.