The lesser Japanese mole (Mogera imaizumii) is mainly distributed in eastern Honshu, the largest island of the Japanese archipelago. From central Honshu to the western islands, on the other hand, is almost exclusively inhabited by their closely related species, the large Japanese mole (M. wogura), with several allopatric fragmented populations of M. imaizumii. Our study aims to analyze the phylogenetic relationships and historical dynamics of M. imaizumii populations in western Japan using mitochondrial cytochrome b gene (Cytb) sequences (1140 bp). The findings indicate that the populations in western Honshu (Mt. Hiba), Shikoku (Mts. Ishizuchi and Tsurugi), and two offshore islands (Dogo and Shodoshima) belonged to a distinct lineage (Mim-IV) that diverged about 700 000 years ago from the three major lineages previously identified in eastern Japan and the Kii Peninsula. Mim-IV harbored closely related Cytb sequences within the lineage, with genetic divergences occurring at least 200 000 years ago. We hypothesize that the genetic structures of M. imaizumii and M. wogura in western Japan may have been formed by the repeated expansion and contraction in response to the recurrent environmental fluctuations during the 100 000-year Quaternary glacial cycle.
Published online 25 December, 2024; Print publication 31 January, 2025
Understanding the generation and persistence of genetic diversity among geographically isolated populations is of great importance in evolutionary biology (Allmon 1992). Physical barriers such as mountains and oceans are abiotic factors that contribute to genetic diversity, whereas competitive exclusion is an important biotic factor that can maintain historical phylogeographic division (Waters 2011). The 100 000-year glacial cycle of the middle to late Quaternary influenced the phylogeographic/genetic differentiation of various organisms (Hewitt 2000; Shafer et al. 2010), with genetic differentiation occurring during glacial periods (McCormack et al. 2008; He and Jiang 2014). For example, interglacial warming during the Quaternary allowed species that had adapted to a warmer climate to expand, whereas competing non-adapted species retreated to higher elevations to form populations fragmented in sky islands, which can serve as generators of genetic diversity (McCormack et al. 2008, 2009).
The keys to understanding the evolutionary history of terrestrial animals in the Japanese archipelago are the dramatic changes in population size in response to climate changes during the Quaternary, including the glacial and interglacial cycles, and the repeated formation of land bridges between Japan's four main islands and the Asian continent, as well as the smaller surrounding islands (Millien-Parra and Jaeger 1999). For example, these evolutionary drivers have shown to have had a large impact on the large Japanese wood mouse, Apodemus speciosus (Suzuki et al. 2015), and large mole species of genus Mogera (Iwasa et al. 2006; Kirihara et al. 2013; Nakamoto et al. 2021). Another important aspect of evolution in this region is that distributions of species or groups of closely related species expand and contract reciprocally between populations at the northern and southern extremes of the Japanese archipelago in a manner consistent with Quaternary climate changes. This trend has been reported in large mammals as shown in a study on Asian black bears in Japan (Wu et al. 2015). In small mammals such as voles and moles, northern (or eastern) species can have fragmented populations within the ranges of southern (or western) species (Kirihara et al. 2013; Honda et al. 2019). For example, the northern species Myodes andersoni has a fragmented population in the Kii Peninsula within the ranges of the southern species M. smithii (Honda et al. 2019). This fragmented population is thought to reflect the repeated expansion and contraction of the northern species, highlighting the impact of the Quaternary climate changes.
The temperate zone of the Japanese archipelago (i.e., Honshu, Shikoku, and Kyushu Islands) contains several mole and shrew-mole species (family Talpidae), most of which are endemic, following repeated ancient migrations from the continent to Japan via the Tsushima Strait (Shinohara et al. 2005; Kirihara et al. 2013; Sato 2017). Large mole species of genus Mogera have a subterranean lifestyle, which allows them to expand into both plains and mountains; they feed mainly on earthworms or other invertebrates (Abe 1968; Yokohata 2005) and evolved under the influence of Quaternary environmental changes (Iwasa et al. 2006; Kirihara et al. 2013; Nakamoto et al. 2021). The lesser Japanese mole (Mogera imaizumii) inhabits eastern Japan (northern Honshu and enclaves in western Honshu and Shikoku; Fig. 1), and the large Japanese mole (M. wogura) inhabits western Japan (southern Honshu, Shikoku, Kyushu, and other islands including the Oki Islands, Shodoshima Island, Yakushima Island, Tanegashima Island, and the Tsushima Islands). In addition, M. tokudae is endemic to Sado Island, and M. etigo is endemic to the Echigo Plain (Abe et al. 2008; Ohdachi et al. 2009). Mogera wogura is also found on the Asian continent (Korean Peninsula, northeastern China, and Russian Primorsky Territory), and these continental populations are considered as a separate species, M. robusta (Zemlemerova et al. 2019). Mogera wogura is generally larger than M. imaizumii and commonly excludes M. imaizumii through competition under soft and fertile soil conditions (Abe 1974, 2001, 2010; Moribe and Yokohata 2011). A previous study suggested that M. imaizumii, which has traits that are likely cold-adapted, moved south during glacial periods of the Quaternary cycle, and M. wogura, which has a larger body size and therefore dominant over M. imaizumii, moved north during interglacial periods, in response to environmental fluctuation (Kirihara et al. 2013). In addition to its Kyoto and Kii Peninsula populations, M. imaizumii has enclave populations on Mt. Hiba in Chugoku, Mts. Ishizuchi and Tsurugi in Shikoku Island, Shodoshima Island, the Nobi Plain in Tokai, and the Suzuka Mountains in Kansai (Kushihashi 1982; Abe et al. 2008; Ohdachi et al. 2009). A new population of M. imaizumii was recently discovered on Dogo Island (Tsunoi et al. 2023) among the Oki Islands in Chugoku (Fig. 1). The five isolated relict populations of M. imaizumii in western Japan, including the two offshore islands inhabited by M. wogura, provide opportunities for investigating the factors contributing to habitat segregation among the relict populations.
Phylogeographic variation in M. imaizumii and M. wogura has been investigated with sequences of the cytochrome b gene (Cytb) in mitochondrial DNA (mtDNA; Tsuchiya et al. 2000; Iwasa et al. 2006; Kirihara et al. 2013; Nakamoto et al. 2021). Mogera imaizumii has three distinct phylogroups generally corresponding to the administrative units of Japan: Mim-I (Tohoku), Mim-II (Kanto), and Mim-III (Hokuriku and Kansai) (Fig. 1A; see Fig. 1C for the names of the administrative units of Japan). Mogera wogura has three phylogroups in the Japanese archipelago (Fig. 1A)—Mwo-I (Tokai and Kansai), Mwo-II (Hokuriku, eastern Chugoku, and Shikoku), and Mwo-III (western Chugoku and Kyushu)—as well as one phylogroup (Mwo-IV) in the neighboring continental area (Korean Peninsula and Russian Far East). In Kansai, M. imaizumii has enclave populations in Kyoto, Nara, and Wakayama Prefectures (Mim-III), which show substantial genetic differentiation from Mim-I and Mim-II (Kirihara et al. 2013; Nakamoto et al. 2021). Okamoto (1999) examined mitochondrial cytochrome c oxidase subunit 1 gene (CO1) sequences in enclave M. imaizumii populations of Mt. Hiba on western Honshu, Mts. Ishizuchi and Tsurugi on Shikoku Island, and an offshore island (Shodoshima). The enclave mtDNA haplotypes formed a distinct cluster separated from other M. imaizumii clusters, although no sequences were deposited in the nucleotide database. Recently, Tsunoi et al. (2023) discovered a fragmented M. imaizumii population on Dogo Island and demonstrated that its lineage (Mim-IV) was distinct from the three known lineages. Previous studies based on Cytb sequences did not examine all enclave populations (Tsuchiya et al. 2000, Kirihara et al. 2013, Nakamoto et al. 2021), and the phylogeographical history of the enclave populations was not discussed in Okamoto (1999). Thus, the phylogenetic relationships of all known M. imaizumii enclave populations remain poorly understood, and the factors leading to the generation and persistence of these fragmented populations have not been studied in detail.
Fig. 1.
(A) Map of the Japanese archipelago indicating sampling locations for Mogera imaizumii and M. wogura used in this study. Collection localities, described in Supplementary Table S1 (09tsunoi_2023-0069_suppl.pdf), are indicated for intraspecies phylogroups based on cytochrome b gene (Cytb) sequences in M. imaizumii and M. wogura. (B) Maximum likelihood (ML) tree for M. imaizumii, M. wogura, and M. tokudae based on Cytb sequences (1140 bp). The evolutionary history was inferred with Tamura-Nei model (Tamura and Nei 1993). Bootstrap scores for 500 replicates are shown in percentages at clade nodes. Branch lengths were measured in the number of substitutions per site. (C) Administrative units of Japan (Tohoku, Koshinetsu, Kanto, Hokuriku, Tokai, Kansai, Chugoku, Shikoku, and Kyushu) are used to describe regionally cohesive sampling localities and phylogroups.
In this study, we examined M. imaizumii samples collected from Shodoshima, and Mts. Hiba, Mts. Ishizuchi and Tsurugi and determined their Cytb sequences to analyze their phylogenetic relationships and historical dynamics. We performed molecular phylogenetic analyses of these sequences together with those reported from Dogo Island (Tsunoi et al. 2023) to elucidate the evolutionary history of isolated relict populations of M. imaizumii in western Japan. We further investigated how the evolutionary history of M. imaizumii was influenced by late Quaternary environmental changes and interactions with the closely related species M. wogura, which has been possibly competitive at the boundary between two species distributions.
Materials and methods
Materials
We collected 43 specimens of M. imaizumii from 20 localities, with a special emphasis on isolated populations in western Honshu (Mt. Hiba, Shodoshima Island, and Mts. Ishizuchi and Tsurugi; Fig. 1A, Supplementary Table S1 (09tsunoi_2023-0069_suppl.pdf)). Thirteen specimens were newly collected for M. wogura. Genomic DNA was extracted with a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) from tissue samples (liver, muscle, or bone) preserved in 99.5% ethanol solution. Nucleotide sequence data newly obtained in this study are available in the DDBJ/ENA/GenBank databases under the accession numbers LC777529–LC777584, LC807168, and LC833857.
Determination of mitochondrial Cytb sequences
Mitochondrial Cytb sequences (1140 bp) were determined with polymerase chain reaction (PCR) and Sanger sequencing. PCR was performed with the universal primer pair L14724 and H15915 (Kocher et al. 1989). The thermocycling parameters for the PCR were 94°C for 2 min; 30 cycles at 98°C for 10 s, 50°C for 30 s, and 60°C for 2 min; followed by 72°C for 7 min. The PCR mixtures (20 µL) consisted of 1 µL DNA solution, 10 µL 2x buffer, 2 µL 2mM dNTPs, 0.6 µL of each primer (concentration, 10 pmol), 5.4 µL deionized water, and 0.4 µL KOD FX Neo polymerase (Toyobo, Osaka, Japan).
Sequencing was performed with a BigDye Terminator Cycle Sequencing Kit v. 3.1 (Applied Biosystems, Waltham, MA, USA) and the sequence primers L14724 and H15915. The labeled PCR products were sequenced with an ABI3130 automated sequencer (Applied Biosystems). The sequences were aligned in MEGA v10.0 (Kumar et al. 2018).
Phylogenetic analyses
Including the newly determined sequences (n = 56) and those obtained from the GenBank, ENA, and DDBJ nucleotidedatabases,weused122and157Cytbsequences (1140 bp) of M. imaizumii and M. wogura, respectively, in our phylogenetic analyses. Sequences of M. tokudae and M. etigo were obtained from the databases and used as outgroups for phylogenetic inference.
A maximum likelihood (ML) phylogenetic tree was constructed in MEGA v10.0, with the substitution models TN93 + G + I. The best-fitting model was determined with Akaike's information criterion in MEGA. Bootstrap scores were calculated using 500 replicates. Divergence time estimation was performed with BEAST v1.6.1 (Drummond and Rambaut 2007). We used an evolutionary rate of 0.03 substitutions/site/million years for the TN93 + G + I substitution model based on rates for mole species determined in the previous studies (Kirihara et al. 2013; Nakamoto et al. 2021).
Results
Sample collection and Cytb sequence determination were conducted mainly for M. imaizumii in the Chugoku and Shikoku regions to analyze their phylogenetic relationships and historical dynamics. The ML phylogenetic tree constructed from sequences of geographically isolated populations of M. imaizumii from western Japan (Dogo Island, Mt. Hiba, Shodoshima Island, and Mts. Ishizuchi and Tsurugi), together with those obtained from the databases, indicated four major monophyletic clades (Mim-I–IV; Fig. 1B). All haplotypes from Chugoku and Shikoku were integrated into a single clade, Mim-IV, that was previously assigned to the lineage of Dogo Island haplotypes (Tsunoi et al. 2023). The remaining three lineages represented specific geographic regions, as reported previously (Kirihara et al. 2013; Nakamoto et al. 2021; Fig. 1B).
The Mim-IV clade exhibited three distinct sublineages: Shodoshima–Shikoku Island (Mim-IVa), Mt. Hiba (Mim-IVb), and Oki–Dogo (Mim-IVc) (Fig. 1B).
Fig. 2.
Chronogram of Mogera imaizumii, M. wogura, M. tokudae, and M. etigo based on Cytb sequences obtained using the BEAST software. Estimated divergence times and 95% highest posterior density (HPD) intervals are indicated at clade nodes. The horizontal bar on each node indicates 95% HPD intervals.
The ML tree placed the Mim-II lineage (Kanto) in the basal position. The divergence patterns of Mim-I (Tohoku), -III (Kansai), and -IV were not clearly resolved, which suggests that they diverged within a short period of evolutionary time. In M. wogura, the divergence of Mwo-III (Kyushu) and Mwo-IV (continental) was followed by that of Mwo-I (Tokai and Kansai) and Mwo-II (Chugoku and Shikoku) (Fig. 1B).
The BEAST analyses produced a different topology for the phylogenetic relationships from that estimated by the ML analysis. The analyses estimated the basal divergence of Mim-II at approximately 0.96 million years ago (mya, 95% highest posterior density [HPD], 0.75–1.18), that of Mim-IV vs. Mim-I and Mim-III at 0.71 mya (95% HPD, 0.55–0.88), and that of Mim-I vs. Mim-III at 0.71 mya (95% HPD, 0.55–0.89; Fig. 2). In Mim-IV, sublineage divergence times were estimated to be 0.22 mya (95% HPD, 0.14–0.31) and 0.21 mya (95% HPD, 0.12–0.29) for the early divergence of Mim-IVa (Dogo Island) and that of Mim-IVc (Shodoshima and Shikoku Island) and Mim-IVb (Mt. Hiba), respectively. Divergence of the Mim-IIIb (Kyoto) and Mim-IIIc (Kii Peninsula) lineages from Mim-IIIa lineage (Hokuriku) was estimated to have occurred at approximately 0.30 mya (95% HPD, 0.20–0.41).
The major phylogroups of M. wogura were estimated to have diverged 1.23 mya (95% HPD, 0.98–1.49) for Mwo-IV, 1.16 mya (95% HPD, 0.90–1.40) for Mwo-I, and 1.1 mya (95% HPD, 0.81–1.39) for Mwo-II and -III. The haplotypes of M. wogura from Shodoshima appear to have been integrated into the Kansai and Tokai Mwo-I phylogroup rather than Mwo-II.
Discussion
Intraspecific phylogenetic history of moles of the Japanese archipelago
Our previous studies of mole evolution using mtDNA sequences identified three major phylogroups (Mim-I–III), including the Kii Peninsula population (Tsuchiya et al. 2000; Kirihara et al. 2013; Nakamoto et al. 2021). In this study, we found that all M. imaizumii haplotypes from Chugoku and Shikoku were members of a fourth lineage (Mim-IV) of M. imaizumii from Dogo in the Oki Islands, where this species was only recently discovered (Tsunoi et al. 2023; Fig. 1). By contrast, Kansai populations (Kii Peninsula and Kyoto) had the third major clade (Mim-III), which is represented by haplotypes from Hokuriku.
Phylogenetic analyses showed that among Chugoku and Shikoku populations, Mim-IV diverged from the northwestern Honshu lineages (Mim-I and -III) approximately 0.71 mya (Fig. 2). This finding suggests that the Chugoku–Shikoku lineage has persisted in western Japan as early as the mid-Pleistocene.
Phylogenetic relationships among distinct populations of western Japan showed three sublineages: the Oki Islands (Mim-IVa), Mt. Hiba (Mim-IVb), and Shodoshima and Shikoku Island (Mts. Ishizuchi and Tsurugi; Mim-IVc, Fig. 1B). These sublineages are estimated to have diverged 205 000–219 000 years ago. The ancient divergence of isolated M. imaizumii populations in western Japan appears to have not been associated with the last glacial maximum (LGM) but rather with the post-penultimate (PGM) and post-antepenultimate glacial maxima, which has not been discussed in previous studies on climatic effects on granivorous and herbivorous small mammals in Japan (Oshida et al. 2009; Suzuki et al. 2015; Hanazaki et al. 2017; Honda et al. 2019). Our results clearly indicate a similarity between haplotypes from Shodoshima and Shikoku Island, which suggests their post-LGM divergence. In summary, our results imply that the observed mtDNA diversity among M. imaizumii populations in western Japan could be influenced by a series of expansion and contraction events, possibly in response to glacial and interglacial periods during the 100 000-year cycle of Quaternary environmental changes.
The Cytb sequences between Ootoyo-cho (Fig. 1A, locality 56) and Kumakogen-cho (locality 55) in Shikoku differed by a single nucleotide. This suggests that the genetic differentiation may not be significant between the two populations of these localities, despite previous beliefs that M. imaizumii have two isolated habitats on Mts. Ishizuchi and Tsurugi in Shikoku (Ohdachi et al. 2009; Fig. 1A). Recent distribution surveys suggest that M. imaizumii is found in the mountains between Mts. Ishizuchi and Tsurugi (Tanioka 2020), which is in line with our current findings.
Impact of Quaternary climate changes and interspecies competitive exclusion
Our results are consistent with the established view that interaction between M. imaizumii and M. wogura influenced their evolution. Mogera imaizumii may have adapted to cold environments and expanded its territory during glacial periods (Kirihara et al. 2013; Nakamoto et al. 2021), whereas M. wogura is thought to have experienced rapid demographic expansion during periods of abrupt warming. For example, Nakamoto et al. (2021) reported expansion signals of post-LGM and post-PGM interglacial periods in Cytb sequence data for M. wogura from western Japan. Thus, it is reasonable to hypothesize that both species underwent changes in their territory and ecological dominance throughout the 100 000-year Quaternary glacial cycle (Kirihara et al. 2013; Nakamoto et al. 2021).
Our BEAST divergence time estimation showed that western Japanese M. wogura diverged as early as 1 mya (Fig. 2). The Mim-IV lineage was present in western Japan for at least 700 000 years; therefore, M. imaizumii may have played a role in shaping the three main lineages observed in M. wogura distribution in western Japan. The mechanisms underlying the creation and maintenance of the distribution boundary between Mwo-II and Mwo-III mtDNA lineages of M. wogura in Chugoku, at the westernmost tip of Honshu, have not been elucidated to date. Mim-IV lineage may have existed near the boundary between Mwo-II and Mwo-III (e.g., Fig.1, locality 50). It can be inferred that the population of Mim-IV lineage had intermittently interrupted the range expansions and gene flows of Mwo-II and Mwo-III during middle Quaternary period, which would have led to the current phylogeographic patterns of Mwo-II and Mwo-III. Likely, Kirihara et al. (2013) have speculated that the range expansions of M. imaizumii may have contributed to the genetic distinctiveness of Mwo-I. We also consider the presence of the Mim-IV populations of M. imaizumii (e.g., localities 39–43) during glacial periods as a potential factor in the segregation of Mwo-I and Mwo-II. Previous studies have also speculated that the presence of an inland sea (Kirihara et al. 2013) and a difference in ecological preference for mountains and plains (Nakamoto et al. 2021) may have formed a boundary between Mwo-I and Mwo-II, and these multiple factors may have worked complexly.
Conclusions
This study is the first to provide the genetic structure of M. imaizumii, with an emphasis on isolated populations in western Japan. Our phylogenetic analyses showed that mtDNA sequences from enclave populations were integrated into a fourth major lineage of M. imaizumii (Mim-IV). Divergence times within Mim-IV were estimated at 200 000 years ago, which suggests that in the Japanese archipelago, cold-adapted M. imaizumii underwent genetic exchange between sites during glaciation two and three epochs ago and that the origin of these western Japanese lineages dates to 700 000 years ago. Our results suggest that the 100 000-year Quaternary glacial cycle caused mole populations to shrink and expand repeatedly. The factors that determine the distributions of M. imaizumii and M. wogura are only partly uncovered, but we speculate that they had been in a state of mutual competitive exclusion during the long period of the Quaternary, and M. imaizumii may have been involved in the generation of the three main lineages of M. wogura in western Japan. Mole species in the Japanese archipelago appear to have been affected by a variety of factors, including Quaternary environmental changes, geographic barriers, interspecific competition, and regional habitat diversity. In conclusion, both biotic and abiotic factors are likely responsible for creating and maintaining the within-species genetic diversity in M. imaizumii and M. wogura in the Japanese archipelago.
Supplementary data
Supplementary data are available at Mammal Study online.
Supplementary Table S1 (09tsunoi_2023-0069_suppl.pdf). List of Mogera samples used in this study.
Acknowledgments:
We thank Kimiyuki Tsuchiya, Masahiro Iijima, Akio Shinohara, and Gohta Kinoshita for providing valuable comments on an earlier draft of this manuscript. We also thank Shin-ichiro Kawada for providing an M. imaizumii sample for this study. We also thank two anonymous reviewers for their comments, which helped us to improve the manuscript. This study was conducted with the support of a Grant-in-Aid for Scientific Research on Innovative Areas to HS (no. JS18H05508).
© The Mammal Society of Japan
