DNA preparation

Permission from the Environmental Agency and the Agency for Cultural Affairs of Japan was given to S. Minato to examine specimens of the Japanese dormouse. A total of seven individual Japanese dormice, one forest dormouse and one common dormouse (Table 1, Fig. 1) were phylogenetically compared. Skin samples were cultured by a standard method without sacrificing the animals, then DNA was isolated as described by Maniatis et al. (1982). Nuclear DNA was also prepared from the liver tissue of a Syrian hamster, which was a gift from Dr. A. Matsuda (Institute of Medical Science, University of Tokyo).

Construction of restriction maps of the nuclear rDNA spacers

Southern blots were performed as described by Suzuki et al. (1994b). To construct restriction maps for the various types of repeating unit of rDNA (repetypes), genomic DNA was digested with the restriction enzymes, Aatl, BamHl, Bcll, BglII, Dral, EcoRl, HindIII, Kpnl, Pstl, Pvull, Sacl, and Xbal and resolved by electrophoresis. The fragments were transferred to nylon filters and sequentially hybridized with the ³²P-labeled rDNA probes, 18SB, 28S, and INT (see Fig. 2). We constructed restriction maps for the coding and spacer regions of genes for rRNA based on hybridization profiles after single digestions (Suzuki et al., 1994b).

Fig. 2

Restriction maps of the major rDNA repetypes of the Japanese dormice from Nagano, Wakayama, and Kochi Prefectures, and of continental forest and common dormice. With respect to the restriction sites on the flanking spacers, only those nearest the distal end of the genes for 18S or 28S rRNA are shown. The top diagram shows the conserved restriction sites in the coding and internal spacer regions of the genes for 18S and 28S RNA (see text), which are not represented in the lower maps. Positions of probes are also shown by arrows. Asterisks and brackets indicate polymorphic sites that show substantial heterogeneity within genomes and among localities, respectively. Small characters represent types of sites identified after comparing restriction maps. A, Aatl; B, BamHl; D, Dral; E, EcoRl; G, Bglll, H, HindIII; K, Kpnl; L, Bcll; P, Pstl; S, Sacl; V, Pvull; X, Xbal.

Direct sequencing

“Semi-nested” polymerase chain reactions (PCRs) and direct sequencing were performed according to instructions supplied with the automated sequencer (ABI, Division of Applied Biosystems, Perkin Elmer). DNA was amplified in 20-μl reaction mixtures containing 10 mM Tris (pH 8.3), 50 mM KCl, 0.01% gelatin, 0.1% Triton X-100, 2.5 mM MgCl₂, 0.2 mM each dNTP, 0.05 μM each primer (1 pmol of each primer per reaction), 0.5 units of Amplitaq DNA polymerase (Perkin Elmer) and 0.1–0.5 μg of template total genomic DNA using an automated thermal cycler (model PJ2000; Perkin Elmer). The thermal cycling parameters were as follows: initial denaturation at 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min and extension at 60°C for 1 min. A 0.5-μl aliquot of each reaction mixture after PCR, was used as the template for the second PCR in a 20-μl reaction mixture with the same reagents and under the same conditions as the first PCR, apart from the primer pairs and the concentration of MgCl₂ (1.25 mM). To construct the primer pair for the second PCR, the sequence of the 18-meric dye-labeled primer, M13RP1 (R) or −21 M13 (U), from ABI was attached to the 5′ end of each of the gene-specific primers. Both strands of the product of the second PCR, were directly sequenced by an automated method using the Dye Primer Cycle Sequencing Kit (ABI) and an automated sequencer (model 373A; ABI).

Sequencing the mitochondrial gene for cytochrome b

A 1.2-kb fragment of the gene for cytochrome b was amplified as described above for the first PCR using the primers [universal primers of Kocher et al. (1989)] L14724 and H15915 (Fig. 2b). L and H refer to light and heavy strands. Numbers refer to the position of the 3′ base of the primer in the complete sequence of human mtDNA (Anderson et al., 1981). For the second PCR, the 3′ part of the forward primer, R-L14724 (5′-CAGGAAACAGCTATGACCGATATGAAAAA-CCATCGTTG-3′), was a shortened version of Kocher's universal primer (Kocher et al., 1989) and the reverse primer, U-H15155 (5′-TGTAAAACGACGGCCAGTTGCCCCTCAAAAGGATATTTG-3′), was a sequence that should anneal to all mammalian mtDNAs around nucleotide positions 15155–15176 in view of the known sequences of mammalian genes for cytochrome b (Irwin et al., 1991).

Sequencing of the mitochondrial gene for 12S rRNA

A 0.9-kb fragment of the 12S rRNA gene was amplified by PCR (Fig. 2c) with the primer pair L613 and H1478 (Kocher et al., 1989). Then first half of the fragment was amplified with the primer pair R-L613 (5′-CAGGAAACAGCTATGACCACACAAAGCATGGCACTG-AA-3′) and U-H1066 (5′-TGTAAAACGACGGCCAGTGGGGTATCTA-ATCCCAGTTTG-3′). The second half was amplified with the primer pair R-L1091 (5′-CAGGAAACAGCTATGACCACTGGGATTAGATAC-CCCACTAT-3′) and U-H1478 (5′-TGTAAAACGACGGCCAGTCAGA-GGGTGACGGGCGGTGTG-3′). To sequence the region around nucleotide position 1091, a fragment of 0.9 kb was amplified with the primer pair R-L613 and U-H1478, and the product was purified by agarose gel electrophoresis using a Geneclean Kit (BIO 101 Inc.). The amplified fragments were subsequently sequenced described above.

Sequencing of flanking portion of the Sry gene

The Sry gene is located on the Y chromosome and is the most likely sex-determining gene in mammalian genomes (Gubbay et al., 1990; Sinclair et al., 1990). The Sry gene has a conserved DNA-binding motif, known as the HMG box, while its flanking regions have rapidly evolved and are male-specific phylogenetic markers for related species and subspecies (Lundrigan and Tucker, 1994; Nagamine et al., 1994). The HMG box provided one side of the primer-anchoring regions for amplification by PCR. To choose other primer-anchoring sites, we compared the sequences of the human and murine Sry genes (Gubbay et al., 1990; Sinclair et al., 1990) by analyzing a homology plot and found an appropriate region about 300 bp upstream from the 3′ end of the HMG box. PCR was performed first with the primer pair SRY286 (5′-GTTGGGCGGGGTTGAGGGGGTGTTGAGGGCGGAG-3′) and HMG777 (5′-GGCCGATACTTATAGTTTCGGGTATTT-3′). Fragments were then amplified with R-SRY286 (5′-CAGGAAACAGCTATGACCTGAGGGGGTGTTGAGGGCGGAG-3′) and U-HMG597 (5′-TGTAAAACGACGGCCAGTCACACGATAAATG-CGTTCATGGG-3′) as primers. The number of each primer refers to the position of the 3′ base of the primer in the sequence of the human Sry gene (Sinclair et al., 1990).

Construction of phylogenetic trees

To estimate sequence divergence among major rDNA repetypes, we compared the arrangement of restriction sites between pairs of rDNA repetypes. We counted common and different sites for a total of about 27 sites, namely 24 sites from both sides of the external spacer region and a few from the internal spacer regions, as described criteria (Suzuki et al., 1994b). We used the method developed by Gotoh et al. (1979), in which backward mutations and parallel mutations are taken into account (Jukes and Cantor, 1969), to produce a matrix of sequence divergence among all possible combinations of repetypes. Sequences of the gene for 12S rRNA were aligned manually, with introduction of gaps to maximize homology. Excluding gap regions, we selected 804 bp with which to consider sequence divergence. We produced a matrix of sequence divergence for all possible combinations of sequences of genes for cytochrome b (402 bp) and 12S rRNA (830 bp) using a computer program (DNADIST), with a “Kimura's 2-parameter” option, in PHYLIP 3.5 developed by Felsenstein (1993). We constructed phylogenetic trees by the neighbor-joining (NJ) method (Saitou and Nei, 1987) and performed a bootstrap analysis (500 replications) to check the topology for robustness using software (SEQBOOT and NEIGHBOR) in PHYLIP 3.5.

Intraspecies variation in the Japanese dormouse

We analyzed restriction length polymorphism (RFLP) in the spacers of the genes for rRNA, which provides genetic relationships between closely related species or populations (Suzuki et al., 1990, 1994a; Wakana et al., 1996). DNA from a total of nine individuals (Table 1), including seven Japanese dormice from four localities (Fig. 1) and two individuals of continental species, the common dormouse and the forest dormouse, was Southern hybridized after digestion with 12 restriction enzymes. Five restriction maps of rDNA (Fig. 2) constructed from the hybridization profiles, represented the major repetypes of these individuals. We compared the repetypes and estimated the sequence divergence (Table 1) according to the criteria described in Materials and Methods. We then constructed a phylogenetic tree by the NJ method (Fig. 3a). In agreement with the morphological and behavioral findings, we found considerable intraspecies differences among the Japanese dormouse. As shown in Fig. 2 and Table 1, individuals from Nagano and Yamanashi Prefectures, had the same repetype, Nagano. Individuals from Fukui and Wakayama Prefectures, also had the same repetype, Wakayama. The Wakayama repetype differed slightly from that of the individual from Kochi Prefecture, with an estimated SD of 0.6% (Table 1). However, the Nagano repetype was quite distinct from the those of southwestern Japan, namely, Fukui, Wakayama and Kochi. The average SD between them was 2.8% (7-7.5 sites were different among 28-29 sites examined), which is relatively high for populations that belong to the same species: the SD value was higher than between two subspecies of mouse, Mus musculus domesticus and M. m. musculus (1.5% SD, four sites differed among 24 examined; Suzuki et al., unpublished data), or between two related species of field mice, Apodemus flavicollis and A. sylvaticus (1% SD; Suzuki et al., 1990). Assuming that the rate of evolution of the spacer of rDNA is 1-2% per million years (Myr; Suzuki et al., 1994b), the estimated time of the split between the Nagano-Yamanashi and Wakayama-Fukui-Kochi groups of the Japanese dormouse was 1.4-2.8 million years ago (Mya). The Wakayama and Kochi repetypes differed slightly, with estimated an SD of 0.6% (Table 1).

Fig. 3

Neighbor-joining trees showing the phylogenetic position and geographic variation of the Japanese dormouse inferred from (a) the nuclear rDNA-RFLP, (b) mitochondrial cytochrome b gene sequences, and (c) mitochondrial 12S ribosomal RNA gene sequences. In trees b and c, bootstrap scores expressed in percentages out of 500 replicates are given above each branch. The human sequence (Anderson et al., 1981) was used to root tree c. Japanese dormice were collected from Nagano, Yamanashi, Fukui, Wakayama, and Kochi Prefectures.

An apparent intraspecies divergence in sequences of mitochondrial genes for cytochrome b was also observed in populations of Japanese dormice. An examination of 402 bp fragments of the cytochrome b gene from ten dormice from six localities (Table 1) revealed six sequences (haplotypes) that were specific to each locality. We constructed a phylogenetic tree by the NJ method based on these sequences (Fig. 3b). The sequences from individuals from the central Japan, namely, Nagano and Yamanashi Prefectures were similar to each other but distinct from those from southwestern Japan, namely, Fukui, Wakayama and Kochi Prefectures (Fig. 3b). The SD was over 7%, which corresponds to divergence 2.8 Mya if the rate of evolution of the gene for cytochrome b is 2.5% per Myr (Irwin et al., 1991; Masuda and Yoshida, 1994). In contrast to the results obtained for rDNA, the mtDNA haplotype from Wakayama was distinct from those from Fukui and Kochi with SD of 9.2 and 9.1%, respectively. The inconsistency between the results for rDNA and mtDNA might reflect the specific nature of mtDNA inheritance. The mtDNA phylogeny reflects only the movement of female individuals. Genetic interchanges among various geographic populations might have occurred during the ice age (1–2 Mya), when Honshu and Shikoku (Fig. 1) were attached by land bridges on several occasions. Such discontinuous geographic communication might have influenced the close relationship between the rDNA repetypes from Wakayama and Kochi. Since mtDNA is maternally inherited and the choice of certain haplotypes after populational fusion and subsequent isolation must be determined in a stochastic manner, ancestrally diverged haplotypes might have been preserved in some localities.

While the geographic distribution of mtDNA haplotypes depends on the movement of female individuals, intraspecific differentiation of the Sry-associated region must reflect the movement of male individuals. To examine the extent of the geographic differentiation in the Y chromosome-specific region in the Japanese dormouse, we sequenced approximately 300 bp of the flanking region of the HMG box in three male individuals, one from Yamanashi, one from Wakayama, and one from Kochi Prefecture (Fig. 4). The extent of sequence differences was not very high but one base change and one gap differed between the sequences from Yamanashi and Wakayama, while those from Wakayama and Kochi differed by only a single base change. In a 515-bp region from Sry in M. musculus (Lundrigan and Tucker, 1994), the intraspecific sequence difference ranges from 0.39% (2/515, between European M. m. musculus and M. m. domesticus) to 0.78% (4/512, between Asian M. m. musculus and M. m. domesticus). Our data indicate that phylogenetic differences between the two groups of Japanese dormice from central and southwestern Japan, are also evident in the Sry-associated region. In addition, our data indicate that the Sry variation will be useful in analyzing populational differentiation.

Fig. 4

Sequences of the region of the Y-linked, sex-determining locus, Sry, of Japanese dormice from Yamanashi, Wakayama, and Kochi Prefectures.

Interspecies relationships of the Japanese dormouse

To estimate the extent of interspecies differences between Japanese and continental common and forest dormice, we examined sequences of the mitochondrial DNA as well as the RFLP of the nuclear rDNA. The estimated SD among the genes for cytochrome b from the three species was over 20%. The extent of substitution is likely to be saturated and thus, these data may not reflect the true interspecies phylogenetic relationship. The estimated SD among the three species on the basis of the RFLP of nuclear rDNA was over 13% (Table 1, Fig. 3a). Although the data suggest that these species diverged at quite an ancient time, for example 10 Mya, the reliability of the estimation might be not very high because the extent of changes in restriction sites tended to be saturated. Sequence analysis of the mitochondrial gene for 12S rRNA should provide useful information over a longer evolutionary period (Mindel et al., 1991). We sequenced about 830 bp of the gene for 12S rRNA from two individuals of the Japanese dormouse from Wakayama and Nagano, one common and one forest dormouse. To compare the extent of divergence with those of other relatively well-examined rodents, we also sequenced this region from the Syrian hamster (Mesocricetus auratus) and obtained the analogous human, mouse and rat sequences from a database (Anderson et al., 1981; Bibb et al., 1981; Kobayashi et al., 1981). As shown in Fig. 3c, the extent of divergence between the Japanese and forest dormice was similar to that between the Japanese and common dormice, suggesting that these three species diverged at about the same time. The extent was similar to that of divergence between the rat and the hamster.