Distribution of Subspecies of the House Mouse, Mus musculus (Rodentia: Muridae) in East China as Inferred from Mitochondrial D-loop Sequences

Tianzhu Chao, Huiqiang Cai, Yuxun Zhou, Kai Li* and Junhua Xiao

College of Chemistry, Chemical Engineering and Biotechnology, Donghua University, Shanghai, China

ABSTRACT

The patterns of subspecies distribution, diversity and differentiation of the house mouse Mus musculus provide insights into past dispersal events in natural populations of this species. We investigated the molecular phylogenetics of Chinese house mice based on 466 DNA sequences from mitochondrial D-loop fragments. Our analyses revealed that Mus musculus musculus and M. m. castaneus are older colonized subspecies with broad distributions in East China, while M. m. domestics is a newly arrived subspecies currently restricted to Shanghai, indicating the elevated risk in this cosmopolitan cities. Phylogenetic analysis indicates that M. m. musculus migrated from northern China to eastern China and that M. m. castaneus invaded from southern China. The two subspecies form a hybrid zone in eastern China. There were three major well-differentiated clades within M. m. castaneus, which were designated as CAS1, CAS2 and CAS3 and confirmed previous internal sublineage differentiation of this subspecies. Of the three clades, only CAS1 was detected in all the samples from mainland China. M. m. musculus exhibited no broadly distributed clade, which might be explained by frequent human-assisted movements. Our study offers a newly detailed perspective on the subspecies distribution of M. musculus and its history of migration to eastern China.

Article Information

Received 22 November 2016

Revised 13 December 2016

Accepted 30 December 2016

Available online 12 June 2017

Authors’ Contribution

TC, KL and JX conceived and designed the study. KL, YZ and JX collected samples. TC, HC and KL extracted mouse DNA, performed mtDNA sequencing and analyzed the data. TC, YZ and KL wrote the article.

Key words

House mouse, Mus musculus, Phylogenetics, Mitochondrial DNA, Control region.

DOI: http://dx.doi.org/10.17582/journal.pjz/2017.49.4.1175.1184

* Corresponding author: likai@dhu.edu.cn

0030-9923/2017/0004-1175 $ 9.00/0

INTRODUCTION

Although the house mouse Mus musculus has been used as an important model for mammalian speciation (Geraldes et al., 2008), migration (Gabriel et al., 2010), introgression (Staubach et al., 2012), laboratory biology (Montero et al., 2013), and medicine (Guenet and Bonhomme, 2003; Flint and Eskin, 2012), its natural population structure within China is still incompletely described. The species as a whole probably originated in South or Southwest Asia approximately half a million years ago (Geraldes et al., 2008) and it is usually understood to contain at least three subspecies Mus musculus domesticus, M. m. musculus, and M. m. castaneus on account of their ability to cross yet maintain stable hybrid zones (Teeter et al., 2008).

As commensals of people, the global distribution of M. musculus shows many features suggestive of migration associated with human migration and trade (Gabriel et al., 2010). Dispersal of M. m. domesticus into Western Europe is relatively well dated to approximately 3000 years ago due to a well-studied archaeological record and intensive genetic surveys (Cucchi et al., 2006). By contrast, the timing and dispersal routes of M. m. musculus and M. m. castaneus are more complicated. It is likely that M. m. musculus reached Eastern Europe and East Asia via an initial northern immigration followed by eastward and westward dispersals, respectively (Suzuki et al., 2013; Kodama et al., 2015). One particular M. m. castaneus sublineage arrived to eastern Asia via a Southeast Asia road (Suzuki et al., 2013).

Currently, M. m. domesticus is indigenous to western Europe, while M. m. musculus is found in northern and eastern Europe and across northern Asia to the Pacific coast in China, Korea (Boursot et al., 1993) and Japan (Terashima et al., 2006). M. m. castaneus is present in western, southern, southeastern, and eastern Asia, including China and Japan (Terashima et al., 2006). Among the three subpopulations, M. m. castaneus has the highest genetic diversity and a larger effective population size (Phifer-Rixey et al., 2012). M. m. domesticus and M. m. musculus meet in a well-studied hybrid zone ranging across Denmark, the Netherlands, Germany, Czech Republic and Bulgaria (Jones et al., 2010; Wang et al., 2011), and M. m. musculus and M. m. castaneus meet in a poorly studied hybrid region in mainland China and Japan (Kawashima et al., 1991, 1995).

Previous studies of Chinese mice reported the general distribution pattern of the two subspecies with M. m. musculus distributed in northern China and M. m. castaneus in southern China (Bao et al., 1995, 1999). In 2002, high levels of genetic variation as determined by microsatellite loci analysis were found in the Taiwan and Fujian populations in southeast China (Yu and Peng, 2002). Subsequently, samples from Taiwan were also found to belong to M. m. castaneus based on mitochondrial DNA (Geraldes et al., 2008). More recently, the wider geographic relationships of the East Asian populations of M. m. musculus and M. m. castaneus were established (Suzuki et al., 2013). An interesting finding for M. m. castaneus was the presence of two sublineages on Taiwan compared with only one on mainland China. In another recent study, the presence of an extensive hybrid zone between M. m. castaneus and M. m. musculus that corresponds approximately with Yangtze River was confirmed to (Jing et al., 2014). Therefore, a high-density population of wild mice would provide more detailed information about this hybrid zone.

In our previous study in Shanghai (eastern China) (Fan et al., 2008), fluorescence-based conformation sensitive gel electrophoresis (F-CSGE) and DNA sequence analyses were performed to analyze the coding regions of mitochondrial DNA. Different types of novel mutations were recorded from 64 wild house mice. We undertook further geographic sampling of wild house mice in East China with a view to generating a distribution map of their mitochondrial diversity and a better understanding of their history of immigration and contact.

MATERIALS AND METHODS

Sampling locations

We trapped 191 samples of house mice from eastern and other parts of China. The sampling localities are in five main districts: southwest China (SWC), southeast China (SEC), north China (NC), central China (CC) and east China (EC) (Fig. 1, Table I). All sampling sites were at least 10 km apart. Our field sampling was carried out in accordance with the recommendations of and approved by the Laboratory Animal Committee of Donghua University.

mtDNA sequencing

DNA was extracted from mouse tails using the phenol-chloroform method. The mtDNA control region and tRNA sequences were amplified using 5’-ATTACTCTGGTCTTGTAAACC-3’ and 5’-TATAAGGCCAGGACCAAACCT-3’ primers, which were synthesized by Sangon (Shanghai Sangon Biological Engineering and Technology and Service S. Ltd.). The optimized PCR mixture contained 1×PCR Buffer, 2.5 mM MgCl2, 0.25 mM each dNTP, 0.1 µM forward and reverse primers, and 0.5 U hotstarTaq polymerase (Qiagen, Germany). The PCR cycling conditions were as follows: 15 min of denaturation at 94ºC, 35 cycles of 30 s at 94ºC, 60 s at 59ºC, and 1 min at 72ºC, and a final 5 min extension step at 72ºC. For each analysis, a negative control was included with the samples. The PCR products were detected by agarose gel electrophoresis. All sequences were determined in both directions on an ABI 3730 sequencer (ABI, USA) by Sangon (Shanghai Sangon Biological Engineering and Technology and Service S. Ltd.).

A total of 191 new mtDNA sequences of house mouse origin are deposited in GenBank under accession numbers (KJ681118 - KJ681308). A further 275 sequences were retrieved from NCBI and M. gentilulus was used as the out-group. The combination of new sequences and those downloaded from GenBank produced a final data alignment of 466 sequences of 926 bp length.

Sequence analysis

Sequences were aligned with MAFFT 7. Fifty percent majority-rule consensus trees were generated using MrBayes 3.2 (Ronquist and Huelsenbeck, 2003), and the run conditions were 10 million generations of two independent runs of five chains.

Table I.- Locations and numbers of samples used in this study.

Province	Locality	Latitude N	Longitude E	CAS1	Mus	Dom
Helongjiang (NC)	Mohe	52.25	121.07		1
	Hulin	45.75	133.97		2
Liaoning (NC)	Shenyang	41.48	123.23		1
Henan (NC)	Sanmengxia	34.47	111.12		1
Shandong (EC)	Zaozhuang	34.52	117.33	1	2
	Zaozhuang	35.11	116.48		1
	Zaozhuang	34.7	117.23		2
	Linyi	35.1	118.01		2
	Zaozhuang	34.69	117.55		2
	Zaozhuang	34.77	117.58		1
	Linyi	34.86	118.1		1
	Linyi	35	118.37	1
	Yancheng	34.87	118.41		1
Jiangsu (EC)	Xuzhou	34.15	117.16		2
	Xuzhou	34.37	117.44	1
	Xuzhou	34.42	117.45		1
	Taizhou	32.45	119.92		2
	Taizhou	32.41	119.9	1	1
	Taizhou	32.34	119.88	1
	Nanjing	31.56	119.1		4
	Yixing	31.07	119.31	2
Shanghai (EC)	Qingpu	31.1	121.1		20	2
	Pudong	31.12	121.31	2	12	3
	Yangpu	31.27	121.52	1	1
	Songjiang	31	121.24		11	5
	Chongming	31.73	121.4	16	10
	Fengxian	30.92	121.46		4
	Fengxian	30.97	121.55		2
	Jinshan	30.89	121.16		28
	Nanhui	31.05	121.76		3	1
	Jiading	31.4	121.24	2	5
	Jiading	31.23	121.15		5
	Huangpu	31.23	121.48	1
Zhejiang (EC)	Jiaxing	30.77	120.76		1
	Huzhou	31.02	119.9	4	1
	Ningbo	30.15	120.18	1
Anhui (EC)	Hefei	31.51	117.17		1
Hubei (CC)	Wuchan	30.57	114.3	2
	Xiaogan	31.92	113.9	2
Hunan (CC)	Yongzhou	26.22	111.63	1
Guangxi (SWC)	Daxin	22.85	107.21	1
Yunnan (SWC)	Lijiang	26.83	100.47	1
	Lijiang	26.86	100.25	2
	Kunming	25.02	102.43	2
	Dali	25.43	100.1	4
Total				49	131	11

CAS1 are the sub-clade of Mus musculus castaneus identified in the phylogenetic tree. Mus and Dom represent M. m. musculus and M. m. domestics, respectively.

The burn-in was determined from the convergence of the two runs using the potential scale reduction factor (PSRF) statistics. The analysis was performed using gamma-distributed rates across sites and variable transition and transversion rates as determined by JMODELTEST 0.1.1 (Posada, 2008). Polymorphic and divergence parameters was calculated with DnaSP version 5 (Librado and Rozas, 2009). Tajima’s D and Fu and L’s D neutrality tests, average number of nucleotide substitutions per site between populations (Dxy), and the net number of nucleotide substitutions per site between populations (Da) were also calculated with DnaSP. FST between subspecies and between populations of a given subspecies were calculated using Arlequin 3.5 (Excoffier and Lischer, 2010). Neighbor-Net networks of M. m. musculus haplotypes and M. m. castaneus haplotypes of the CAS1 clade were generated with SplitsTree 4.14 software (Huson and Bryant, 2006).

RESULTS

Subspecies attribution of mtDNA haplotypes

A Bayesian phylogenetic tree constructed from the D-loop alignment shows a clear division into three main lineages, with ample placement consistent with subspecies annotations in NCBI (Fig. 2). The newly sequenced samples included 49 with M. m. castaneus haplotypes, 131 with M. m. musculus haplotypes, and 11 with M. m. domesticus haplotypes (Table I).

mtDNA lineage distributions

Two main subspecies, Mus m. musculus and M. m. castaneus, coexist in EC, while M. m. musculus or M. m. castaneus alone were found in SWC, SEC, NC, and CC. All individuals from the four sites in NC to the north of the Yellow River were found to possess M. m. musculus mtDNA haplotype. By contrast, mice from CC, SWC and SEC all possessed the M. m. castaneus haplotypes. M. m. musculus and M. m. castaneus haplotypes co-occur over a large area in EC from Shandong Province south to Zhejiang Province (Fig. 1). All individuals with M. m. domesticus haplotypes were collected in Shanghai where they co-occur with M. m. musculus and M. m. castaneus haplotypes (Fig. 1, Table I).

mtDNA phylogeography

There were 76 haplotypes detected within M. m. musculus (Fig. 3); however, no coincident geographic pattern was found. The M. m. musculus haplotypes cluster into a few small geographically based sub-clades; these are labeled on Figure 2 as Mus1, Mus2, and Mus3 with posterior probabilities of 1, 0.97, and 0.97, respectively. A series of 16 individuals from EC were grouped together within Mus1. The other two sub-clades (Mus2 and Mus3) are not represented in China but are reported from mice from Europe, central Asian, and Oceania. Mus2 consisted of the individuals from Germany, Czech, Austria, and New Zealand. Mus3 was found in two sequences from Georgia in central Europe.

The M. m. castaneus lineages encompassed three main clades, which are named CAS1, CAS2 and CAS3; each clade is well supported, with posterior probabilities of 0.99, 1 and 1, respectively. These three clades are almost equivalent to the haplogroups CAS-1, CAS-2, and CAS-3 that were described by Suzuki’s group (Suzuki et al., 2013) or HG3, HG1, and HG2 described by Bonhomme’s group (Rajabi-Maham et al., 2012). All samples from mainland China fall into CAS1.

CAS1 is the most widespread clade within M. m. castaneus. It is represented on at least three continents: Asia, Africa and Oceania, and there were 61 mtDNA haplotypes detected (Fig. 4). These haplotypes formed four sub-clades, CAS1A, CAS1B, CAS1C, and CAS1D. Indian samples were found in CAS1A, CAS1B, and CAS1C, and CAS1A contains samples only from India.

Table II.- Polymorphisms within subspecies and the divergence between these subspecies and Mus gentilulus.

Populations	N	Polym orphic nucleo tide sites	Haplo types	Hd	л(%)	θ(%)	Tajima's D	Fu’s Fs	Dxy	Da(%)
musChina	131	66	52	0.9	0.614	1.717	-2.0332 (p<0.05)	-3.58578 (p<0.02)	2.906	2.599
musWorld	176	83	77	0.94	0.626	2.079	-2.1805 (p<0.01)	-5.02493 (p<0.02)	2.875	2.562
casChina	104	63	42	0.93	0.808	1.303	-1.205 (p>0.1)	-1.51694 (p>0.05)	2.919	2.515
cas Mainland China	50	36	30	0.90	0.589	1.119	-1.5932 (p>0.1)	-2.17622 (p>0.05)	3.23	2.936
CAS1	131	52	61	0.93	0.516	1.368	-1.9425 (p<0.05)	-2.93925 (p<0.05)	2.904	2.646
CAS2	64	50	42	0.98	0.813	1.437	-1.4588 (p>0.1)	-1.85853 (p>0.1)	2.915	2.508
CAS3	75	55	59	0.99	0.91	1.52	-1.3384 (p>0.1)	-3.29877 (p<0.05)	3.14	2.687
casWorld	270	109	161	0.983	1.25	2.7	-1.6527 (p>0.05)	-4.87136 (P<0.02)	2.976	2.356

musChina and musworld, represent the sample sequences of Mus musculus musculus from China and elsewhere in the world, respectively; casChina, casMainlandChina, and casWorld represent the sequences of M. m. castaneus from China, mainland China, and elsewhere in the world, respectively. CAS1, CAS2, and CAS3 indicate the three main clades of M. m. castaneus shown in Figure 2.

CAS1B consisted of samples from India, Taiwan Island, and Jinmen Island. CAS1C was composed of mice in Africa from Kenya and in Oceania from New Zealand (Figs. 2, 4) together with those from Taiwan Island (China), Japan, Jinmen Island (China), and India (Fig. 4). Nearly all samples of M. m. castaneus from mainland China fall into CAS1D. Samples from the SEC region form a relatively independent subgroup, while nearly all the remaining samples of CAS1D came from the EC region.

CAS3 consists exclusively of haplotypes found in mice from Indian, Pakistan and Iran. CAS2 is best represented in mice from Iran and Pakistan but a few samples from Taiwan Island belong to this clade (Fig. 2). Suzuki’s group (Suzuki et al., 2013) showed that house mice of this clade were also observed in Vietnam.

Within the M. m. castaneus subspecies worldwide, the CAS3 sub-clade is the most variable, followed by CAS2 and CAS1 (Table II). Although Chinese populations are restricted to only one of these clades (CAS3), these show more nucleotide diversity (π=0.808±0.082%) than Chinese M. m. musculus (π=0.614±0.054%) (Table II).

Tajima’s D and Fu’s Fs values for various subspecies and populations are shown in Table II. For both worldwide and Chinese samples M. m. musculus there are significantly negative for Tajima’s D-values and for Fu’s Fs. This indicates both global and regional population increases for this subspecies, either following period of positive selection or possibly the occurrence of bottlenecks. By contrast, Chinese populations of M. m. castaneus do not show significant negative values, either taken as a whole or on for mainland China only.

Pattern of genetic differentiation are compared between and within subspecies in Table III. In general, FST values are higher for comparisons between subspecies than within subspecies. There was more differentiation between M. m. musculus and M. m. castaneus in China than between the two subspecies in the other locations worldwide. The highest levels of differentiation within subspecies were between the CAS1 and CAS3 sub-clades of M. m. castaneus, followed by CAS2 and CAS3, and CAS1 and CAS2.

Table III.- Patterns of differentiation between different population or subspecies pairs.

	cas- musa	CAS1-CAS2	CAS3-CAS1	CAS3-CAS2	cas-mus (China)b
poly-morphism	147	85	91	80	92
Fixed differences	0	0	2	1	4
Shared mutation	33	16	11	25	15
Fst	0.551	0.492	0.515	0.392	0.654
Dxy (%)	2.2	1.61	1.148	1.641	2.185
Da (%)	1.303	0.929	0.754	0.78	1.535

CAS1, CAS2, and CAS3 indicate the three main clades of M. m. castaneus shown in Figure 2. acas-mus indicates all sequences from China and other countries. bcas-mus (China) indicates all sequences of M. m. castaneus and M. m. musculus from China.

Diversity, differentiation, and divergence time

As reported previously by Suzuki group (Suzuki et al., 2013), there is considerable variations in the level of polymorphisms among the various subspecies and populations of M. musculus. With our new samples included, M. m. castaneus contains nearly twice the nucleotide diversity (π=1.252±0.34%) found in M. m. musculus (π=0.626±0.041%). In M. m. castaneus, more than 100 polymorphic sites contribute to 161 haplotypes (Hd=0.983±0.0034). By comparison, M. m. musculus has only 83 variable sites contributing to 77 haplotypes (Hd=0.940±0.013).

Within the M. m. castaneus subspecies worldwide, the CAS3 sub-clade is the most variable, followed by CAS2 and CAS1 (Table II). Although Chinese populations are restricted to only one of these clades (CAS3), these show more nucleotide diversity (π=0.808±0.082%) than Chinese M. m. musculus (π=0.614±0.054%) (Table II).

Tajima’s D and Fu’s Fs values for various subspecies and populations are shown in Table II. For both worldwide and Chinese samples M. m. musculus there are significantly negative for Tajima’s D-values and for Fu’s Fs. This indicates both global and regional population increases for this subspecies, either following period of positive selection or possibly the occurrence of bottlenecks. By contrast, Chinese populations of M. m. castaneus do not show significant negative values, either taken as a whole or on for mainland China only.

Pattern of genetic differentiation are compared between and within subspecies in Table III. In general, FST values are higher for comparisons between subspecies than within subspecies. There was more differentiation between M. m. musculus and M. m. castaneus in China than between the two subspecies in the other locations worldwide. The highest levels of differentiation within subspecies were between the CAS1 and CAS3 sub-clades of M. m. castaneus, followed by CAS2 and CAS3, and CAS1 and CAS2.

DISCUSSION

Molecular markers for subspecies identification

Our phylogenetic tree suggests that the control region of mitochondrial DNA is suitable for subspecies or taxa of lower rank (i.e., population) for the availability of sequences in public database, although complete mitochondrial DNA is more desirable for distinguishing house mouse subspecies (Katouzian and Rajabi-Maham, 2013). The mutation rate per generation per site in the control region of M. musculus is approximately 0.41 million (Suzuki et al., 2004; Geraldes et al., 2008), which indicates a significant number of polymorphic sites suitable for identification. For instance, 147 variable sites were found between M. m. musculus and M. m. castaneus, although a fixed SNP has not been identified (Table III). All of our samples and downloaded sequences were precisely classified at the subspecies level. In terms of the descriptions of clades within individual subspecies, such as CAS1, CAS2, and CAS3, our results were remarkably consistent with those of previous studies (Rajabi-Maham et al., 2012; Suzuki et al., 2013).

Nuclear genes in autosomes are useful for studying mixed populations of wild mice from hybrid zones and could reveal additional details about gene flow (Yu and Peng, 2002). However, this method has produced complex and confusing results within M. musculus, as phylogenetic trees based on one gene may not be in agreement with trees based on other genes (Geraldes et al., 2008), even if those genes are nearby in the chromosome (Nunome et al., 2010). Therefore, the use of nuclear genes is important. Fine-scale phylogenetic analysis throughout the whole mouse genome might solve this problem and provide a large amount of phylogenetic history at the genomic scale of three closely related subspecies: M. m. musculus, M. m. castaneus, and M. m. domesticus (White et al., 2009).

Biogeography and evolution

Our study of mtDNA of Chinese house mice has shed new light on the regional mtDNA diversity and distributions of each of M. m. musculus, M. m. castaneus, and M. m. domesticus. For each subspecies, mtDNA diversity within China represents a restricted subset of that observed globally. This is consistent with current belief that each of the subspecies originated within a relatively small geographic area in southwest Asia, followed by dramatic episodes of range expansion, much of it achieved through human commensalism (Geraldes et al., 2008; Suzuki et al., 2013).

Although it seems reasonable to postulate that colonization of China by house mice occurred largely through their commensalism with humans, little is known regarding how and when this occurred for each of the subspecies. Though, there is still debate about precise timing of modern human settlement in mainland China, one point of general agreement is that Han populations were first established in South China, followed by a northward expansion (Shi et al., 2008). On these grounds, we might expect that the earliest introduction of house mice was into South China, and that these mice belonged to the M. m. castaneus lineage. A similar conclusion was reached in the previous mtDNA RFLP study (Yonekawa et al., 1988).

All individuals of M. m. castaneus from mainland China belong to CAS1 clade (Fig. 2). Within this clade, all individuals from mainland China are in CAS1D subgroups, except one individual in CAS1C (Fig. 4). Based on this evidence, we hypothesize that CAS1D house mice were the main M. m. castaneus mice introduced into China and dispersed to other part of this country. Moreover, samples from Southwest Asia (Iran), South Asia (India and Pakistan), Southeast Asia (Thailand), East Asia (Japan, Taiwan), Africa (Kenya), and Oceania (New Zealand) were included within CAS1. A previous study also confirmed that southwestern Asia is the ancestral homeland of M. m. castaneus (Suzuki et al., 2013). Because this high regional diversity was not observed for CAS2 and CAS3, our data support the hypothesis that CAS1 M. m. castaneus is the primary clade that has expanded within Asia, and to Africa and Australia.

Within M. m. musculus, there is less differentiation and no strong phylogeographic patterning. This contrast with M. m. castaneus may reflect the contrasting topographies and histories of the main area of geographic distribution of the two subspecies. In the south of Eurasian continent, the area occupied by M. m. castaneus, mountain ranges and large river basins create a naturally fragmented landscape and this has been reflected in ancient division of civilization into multiple regional centers. In contrast, northern Eurasia is a landscape dominated by vast plateau and human history has been more strongly influenced by large scale migration at a population scale. At the global scale, M. m. castaneus harbors more mtDNA variation than M. m. musculus in this study and previous studies (Baines and Harr, 2007; Geraldes et al., 2008), both in terms of variable sites and haplotype number (Table II). In contrast, nucleotide diversity of mainland Chinese populations of M. m. castaneus is lee than that of Chinese M. m. musculus. We believe that this is explained by their contrasting opportunities for invasion and ongoing interchange with populations of the corresponding subspecies outside of China. Essentially, Chinese populations of M. m. musculus are more likely part of a larger regional population of this subspecies that stretches west into central Eurasia and beyond, whereas Chinese populations of M. m. castaneus, seem more isolated from other populations of this subspecies. In particular the last 1500 years has seen the repeated expansion of Proto-Mongal and Mongol peoples creating a cultural and biological realm stretching at times from the Caspian Sea to the Pacific coastline. The occurrence of M. m. domesticus in China is a more restricted and modern and modern phenomenon. To date, this classic marker of western European interaction is recorded exclusively from Shanghai, China’s most cosmopolitan city with the most extensive network of global connections.

Human migration and geographic isolation have also determined the local populations of genetic diversity in each of M. m. musculus and M. m. castaneus. Jinmen Island is 4.3 km off the coast of Fujian Province of China (Fig. 1) and is the site of Jinmen City constructed in 1385. Today it is administered by Taiwan. Although we might expect to find commonality in M. m. castaneus sequences from Jinmen Island with the population in Fuzhou (capital city of Fujian Province) belonging to CAS1D, instead the Jinmen samples share CAS1B and CAS2C haplotypes with individuals from Taiwan (Figs. 2, 4). Thus the genetic pattern appears to mirror the political and human isolation between Jinmen (administered by Taiwan) and mainland China over the last 60years, since The Third Revolutionary Civil War in modern China. Similar instances of house mouse populations mirroring recent political or human immigration events are known from other parts of the world. For example, a single M. m. musculus sequence from New Zealand is identical to records of central Europe, consistent with a significant German immigration into New Zealand (Searle et al., 2009); and Japanese M. m. musculus sequences are very similar to those from Korean mice, again reflecting the recent cultural exchange between these regions (Nunome et al., 2010).

Our expanded geographic sampling of Chinese house mice has more precisely defined the distributions in China of M. m. musculus and M. m. castaneus. Importantly, our sampling has demonstrated a relatively large area of geographic overlap between the two mtDNA lineages in eastern China, including parts of Shangdong, Jiangsu and Zhejiang Provinces. No population of M. m. castaneus occurs north of the Yellow River. However, in the south several populations of M. m. musculus occur south of the Yangtze River. While these long and wide rivers are likely to pose significant barriers to the natural dispersal of a small terrestrial mammal such as M. musculus, a commensal mammal has the advantage of human-assisted dispersal across such barriers.

The interdigitating geographic distribution of the M. m. musculus and M. m. castaneus mtDNA types raises numerous questions regarding the biological interaction between the subspecies. Do the two subspecies overlap in range with limited exchange of genetic material? Alternatively, is the mosaic of mtDNA types indicative of a hybrid zone? If so, what is the nature of the genetic interaction that is occurring? Is it limited to introgression of mtDNA and/or other restricted genetic component? Or is it characterized by broad-scale exchange of genetic component? Further studies of the musculus/castaneus interaction in East China would represent a fascinating counterpoint to the detailed analyses available for the Musculus/domesticus hybrid zone in Europe (Teeter et al., 2008).

ACKNOWLEDGMENTS

We thank Shimin Bao of Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences for his kind help in sample collection. This work was supported by the Key Project of Science and Technology Commission of Shanghai Municipality (No. 13140900300), the National Science Foundation of China (No. 31171199), the Fundamental Research Funds for the Central Universities (No. 2232013A3-06), and the DHU Distinguished Young Professor Program (B201308).

Statement of conflict of interest

Authors have declared no conflict of interest.

REFERENCE

Bao, S., Jin, M. and Shi, M., 1995. Taxonomical characteristics of geneus Mus in Hainan, China. China Forestry Publishing House, China, pp. 203-210 (in Chinese).

Bao, S., Zhao, G., Zhang, R., Gu, Z. and Xu, P., 1999. Study on genetic differentiation of Mus musculus of China. Zoology Society of China. Forestry Publishing House, China, pp. 987-994 (in Chinese).

Baines, J.F. and Harr, B., 2007. Reduced X-linked diversity in derived populations of house mice. Genetics, 175: 1911-1921. https://doi.org/10.1534/genetics.106.069419

Boursot, P., Auffray, J.C., Brittondavidian, J. and Bonhomme, F., 1993. The evolution of house mice. Annu. Rev. Ecol. System., 24: 119-152. https://doi.org/10.1146/annurev.es.24.110193.001003

Cucchi, T., Orth, A., Auffray, J.C., Renaud, S., Fabre, L., Catalan, J., Hadjisterkotis, E., Bonhomme, F. and Vigne, J.D., 2006. A new endemic species of the subgenus Mus (Rodentia, Mammalia) on the Island of Cyprus. Zootaxa, 1241: 1-36.

Excoffier, L. and Lischer, H.E.L., 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol. Ecol. Resour., 10: 564-567. https://doi.org/10.1111/j.1755-0998.2010.02847.x

Fan, Z.P., Zhu, W.S., Zhang, C., Zhou, Y.X., Li, K., Liang, Y.M., Xing, Z.H., Chen, G.Q., Bai, X. and Xiao, J.H., 2008. SNP discovery by F-CSGE in the coding region of mitochondrial DNA in wild house mice from Shanghai suburb. Yi Chuan, 30: 475-482 (in Chinese). https://doi.org/10.3724/SP.J.1005.2008.00475

Flint, J. and Eskin, E., 2012. Genome-wide association studies in mice. Nature Rev. Genet., 13: 807-817. https://doi.org/10.1038/nrg3335

Gabriel, S.I., Johannesdottir, F., Jones, E.P. and Searle, J.B., 2010. Colonization, mouse-style. BMC Biol., 8: 131. https://doi.org/10.1186/1741-7007-8-131