Molecular Phylogeny of Japanese Leporidae, the Amami Rabbit

Genes Genet. Syst. (2002) 77, p. 107–116 Molecular phylogeny of Japanese Leporidae, the Amami rabbit Pentalagus furnessi, the Japanese hare Lepus brachyurus, and the mountain hare Lepus timidus, inferred from mitochondrial DNA sequences

Fumio Yamada1*, Mika Takaki2 and Hitoshi Suzuki2 1Wildlife Ecology Laboratory, Forestry and Forest Products Research Institute, PO Box 16, Tsukuba-Norin, Ibaraki 305-8687, Japan 2Graduate School of Environmental Earth Science, Hokkaido University, Kita-ku, Sapporo 060-0810, Japan

(Received 8 December 2001, accepted 12 February 2002)

We determined mitochondrial 12S ribosomal RNA (rRNA) and cytochrome b (cyt b) gene sequences in three leporid species of Japan, the Amami rabbit Pentalagus furnessi from the Ryukyu Islands, the Japanese hare Lepus brachyurus from Honshu, and a Japanese form of the mountain hare Lepus timidus ainu from Hokkaido. We compared the sequences with those of other taxa of leporids avail- able in databases. Phylogenetic trees of the 12S rRNA gene sequences indicated that the lineage of P. furnessi diversified during the generic radiation of the lep- orids at an ancient time, which was estimated to have been the middle Miocene. Cyt-b gene trees revealed that the lineage of L. brachyurus branched off at an early stage in the speciation of Lepus, probably at the beginning of the Pliocene. The cyt b sequences of L. t. ainu were somewhat distinct from those of continental conspecific populations; this lineage divergence is likely to have occurred during the middle or late Pleistocene. The results show that the three regions of the Japanese archipelago, Ryukyu, Honshu-Shikoku-Kyushu, and Hokkaido, now preserve their own leporid taxa, each with a different extent of genetic endemicity. It is possible that the zoogeographic traits of the Japanese leporids are a consequence of the evolutionary dynamics of leporids in East Asia, in that the radiation centers of leporids are likely to have shifted from tropical, through temperate, to arctic zones.

conservation biology. INTRODUCTION The family Leporidae (Lagomorpha, Mammalia) is Japanese mammals, especially small species, reflect a composed of 53 species of 11 extant genera (Chapman and considerable amount of endemicity from a taxonomic Flux, 1990). Nine of the genera, however, are mono- viewpoint. Many are endemic species, and some are typic, as in the case of Pentalagus (Amami rabbit) and classified as endemic genera. Meanwhile, through Romerolagus (volcano rabbit), or poorly speciated, as recent progress made in molecular phylogeny, high levels in the case of Pronolagus (rock hare), with three of endemicity have been demonstrated at the molecular species. These lineages are thought to retain primitive level in many taxonomic groups, such as rodents (e.g., morphological characters. In contrast, Lepus (jack- Serizawa et al., 2000; Suzuki et al., 2000), moles rabbits and hares) and Sylvilagus (cottontails) are spe- (Tsuchiya et al., 2000), and mustelids (Hosoda et al., cies-rich genera, comprising 29 and 13 species, 2000). The specific properties of the fauna of Japanese respectively, and are thought to be advanced genera on terrestrial mammals have begun to be understood in the basis of their morphology (Dawson, 1981; Angermann the context of the evolution of mammals in East et al., 1990; Chapman and Flux, 1990). Asia. Japanese taxa of leporids (rabbits and hares), how- The Japanese leporids comprise three taxa: the Amami ever, have not yet been explained, in spite of their contro- rabbit Pentalagus furnessi from the Amami Islands versial status in taxonomy and their importance in (Amamioshima and Tokunoshima islands, in the Ryukyu Islands in southernmost Japan), the Japanese hare L. Edited by Toshihiko Shiroishi brachyurus from the mainland (Honshu, Kyushu, and * Corresponding author. E-mail: [email protected] Shikoku), and the mountain hare L. timidus ainu from

108 F. YAMADA et al.

Hokkaido. Phylogenetic information will be useful for occurred during the early stages of speciation within the conservation of these taxa, especially for P. furnessi, Lepus (Halanych et al., 1999). In addition, Halanych which is categorized as endangered (IUCN, 2000). The et al. (1999) used cyt-b gene sequences to examine status of this species is very serious; its habitat has intraspecific relationships in the most variable species of undergone considerable fragmentation and the size of the Lepus, L. timidus. They found that the species L. timi- population has greatly declined (2700–6500 rabbits in dus, L. arcticus, and L. othus form a clade, and that L. 1995, Sugimura et al., 2000; Yamada et al., 2000). timidus displays clear geographic subdivisions. Pentalagus furnessi is considered one of the most prim- The objective of this study was to elucidate the phylo- itive Leporidae (Corbet, 1983; Matsuzaki et al., 1989). It genetic relationships of the three taxa of Japanese lep- has been stated that the Japanese taxon Pentalagus is orids in relation to other leporid lineages of the world, most similar to the African taxon Pronolagus in morpho- based on mitochondrial DNA (mtDNA) sequences. We logical characteristics, and that they evolved from the examined the sequence variation of the mitochondrial common ancestor Pliopentalagus (Fejfar, 1961; Hibbart, 12S rRNA and cyt b genes in the three Japanese species 1963; Daxner and Fejfar, 1967; Dawson, 1981; Tomida, of leporids and compared them with those of other lago- 1997). The mainland taxon L. brachyurus was once morph lineages available in databases. regarded as the same species as that of the continent, L. mandshuricus, but is now classified as a separate species (Angermann et al., 1990; Hoffmann 1993). The MATERIALS AND METHODS Hokkaido form of L. timidus is often treated as a subspe- cies (L. t. ainu) distinct from those of Sakhalin and the Materials All three species of Japanese leporids were continent (Ellerman and Morrison-Scott, 1951; Corbet, examined in this study (Table 1). Tissue samples of P. 1978; Flux and Angermann, 1990). furnessi, which is one of the Natural Monuments of Mitochondrial DNA has been used to examine the evo- Japan, were collected after obtaining permission from the lutionary relationships of leporid genera. The generic Environment Agency and the Agency for Cultural Affairs relationships among seven leporid taxa, Brachylagus, of Japan. Bunolagus, Lepus, Oryctolagus, Pronolagus, Romero- lagus, and Sylvilagus, have been investigated using the PCR and sequencing Polymerase chain reactions variation in the mitochondrial 12S ribosomal RNA (PCRs) and direct sequencing were performed according (rRNA) gene, which evolves at a relatively slow rate to the methods described by Suzuki et al. (1997, (Halanych and Robinson 1999; Surridge et al., 1999). It 2000). A DNA fragment covering the entire target gene has been suggested that a rapid radiation occurred in six region was first amplified in 20-µl reaction mixtures con- genera, and that Pronolagus branched off slightly earlier taining 10 mM Tris (pH 8.3), 50 mM KCl, 0.01% gelatin, from the main leporid current (Halanych and Robinson 0.1% Triton X-100, 2.5 mM MgCl2, each dNTP at 0.2 mM, 1999). The 12S sequences of Nesolagus have been each primer at 0.05 mM (1 pmol of each primer per reac- reported by Surridge et al. (1999) following their discov- tion), 0.5 units of Amplitaq DNA polymerase (Perkin ery of a new species (Nesolagus sp.) from Vietnam that is Elmer), and 0.1–0.5 µg of total genomic DNA as template, related to the Sumatran rabbit, Nesolagus netscheri. with an automated thermal cycler (model TP400; Takara, The interspecies relationships of Lepus have been Japan). The thermal cycling parameters were 35 cycles examined using mitochondrial cytochrome b (cyt b) gene of denaturation at 94°C for 0.5 min, annealing at 50°C for sequences (Halanych et al., 1999; Halanych and Robin- 0.5 min, and extension at 60°C for 0.5 min. A 0.5-µl ali- son, 1999), and revealed several major lineages of Lepus quot of each reaction mixture after PCR was then used as spread across North America, Europe, Asia and template for the second PCR in a 20-µl reaction mixture Africa. These analyses suggest that a rapid radiation with the same reagents and under the same conditions as

Table 1. Japanese lagomorphs analyzed in this study Taxon Common name N Individual code Collection locality Pentalagus furnessi Amami rabbit 3 Pfur_Amami1 Sumiyo, Kagoshima, Japan (28° 19´ N, 129° 22´ E) Pfur_Amami2 Sumiyo, Kagoshima, Japan (28° 19´ N, 129° 22´ E) Pfur_Amami3 Uken, Kagoshima, Japan (28° 14´ N, 129° 20´ E)

Lepus brachyurus Japanese hare 2 Lbra_Honshu1 Ushiku, Ibaraki, Japan (35° 58´ N, 140° 11´ E) Lbra_Honshu2 Ami, Ibaraki, Japan (35° 58´ N, 140° 11´ E)

Lepus timidus ainu mountain hare 2 Ltim_Hokkaido1 Sapporo, Hokkaido, Japan (42° 59´ N, 141° 25´ E) Ltim_Hokkaido2 Sapporo, Hokkaido, Japan (42° 59´ N, 141° 25´ E)

Molecular phylogeny of Japanese leporids 109 for the first PCR, apart from the primer pairs and the (MP) analyses using PAUP* 4.0b8 (Swofford, 2001). In concentration of MgCl2 (1.25 mM). To construct the the NJ tree, sequence divergences were calculated by primer pair for the second PCR, the sequence of an 18- Kimura’s (1980) two-parameter method. The MP analyses meric dye-labeled primer, M13RP1(R) or -21M13(U), from were performed using a priori weighting (transversions : Applied Biosystems, was attached to the 5’ end of each of transitions = 2:1). A heuristic search was performed with the gene-specific primers. Both strands of the second 10 random addition replicates and the tree bisection and PCR product were directly sequenced by an automated reconnection option (TBR). The statistical confidence method using a DyePrimer Cycle Sequencing Kit and an of the sequence clusters in the NJ and MP trees was automated sequencer (model 373A; Applied Biosystems). evaluated by bootstrap percentages derived from 500 A 0.9-kb fragment of the 12S rRNA gene was amplified replications. In addition, we assessed the level of confi- with the primer pair L613 (Mindell et al., 1991) and dence with decay indices (Bremer, 1988) for the MP trees H1478 (Kocher et al., 1989), in which the letters L and H using TreeRot v2 (Sorenson, 1999). refer to the light and heavy strands, respectively, and each number refers to the position of the 3’ base of the RESULTS primers in the complete sequence of human mtDNA (Anderson et al., 1981). Then, the first half of the frag- Generic relationships We determined sequences of ment was amplified with the primer pair R-L613 and the 12S rRNA gene (832-839 bp) and the cyt b gene (1140 U-H1066 (Suzuki et al., 1997). The second half was bp) in seven individuals of the three Japanese species of amplified with the primer pair R-L1063 (5’-CAGG- Lagomorpha (Table 1). Only slight differences within AAACAGCTATGACCTTGAACACACAATAGCTAAGA-3') species were seen in the 12S sequences. One polymorphic and U-H1478 (Suzuki et al., 1997). The amplified frag- site was found in the three individuals of P. furnessi from ments were subsequently sequenced in the manner Amamioshima Island, two sites in the two individuals of described above. L. brachyurus from the same locality of Honshu (Ibaraki A 1.2-kb fragment of the cyt b gene was first amplified Pref.) and one site in the two individuals of L. timidus with the universal primers L14724 and H15915 (Irwin et from Hokkaido. On the other hand, in the cyt-b gene al., 1991). In the second PCR, three segments were sequences, the two individuals of L. brachyurus differed amplified using the first PCR product as template with at seven sites (p = 0.6), while P. furnessi and L. timidus the following primer sets: R-L14724 and U-H15155 showed three (p = 0.3 on average) and two polymorphic (Suzuki et al., 1997), R-L15131 and U-H15598 (Suzuki et sites (p = 0.2), respectively, implying substantial cyt-b al., 2000), and R-L15525 (Hosoda et al., 2000) and U- variation within a local population of L. brachyurus. H15916 (Suzuki et al., 2000). The phylogenetic information content of the markers

DNA sequences obtained from databases Sequences of homologous regions of the 12S rRNA and cyt b genes were obtained from the DDBJ/EMBL/GenBank Inter- national Nucleotide Sequence Databases (Irwin and Arn- ason, 1994; Halanych et al., 1999; Halanych and Robinson, 1999; Surridge et al., 1999; Murphy et al., 2000; Takaki et al., unpublished): Brachylagus idahoensis (U58930, U58921), Bunolagus monticularis (U58931, U58922), Lepus alleni (AF010156), L. americanus (AF010152), L. arcticus (AF010153), L. californicus (U58925), L. euro- paeus (AF010161), L. othus (AF010154), L. timidus from Russia (AF010155) and the UK (AF009732), L. townsen- dii (AF009733), Nesolagus sp. (AF176588), Oryctolagus cuniculus (U59264, U07566), Pronolagus crassicaudatus (U31044, U58935), Romerolagus diazi (AB053205, AB0- 53206), Sylvilagus audubonii (AF038019, U58928) and S. floridianus (AY012126). Sequences of Ochotona hyper- Fig. 1. Transitions (Ti) and transversions (Tv) of the cyt b gene borea (AB053258, AB053257) were used as the outgroup. versus all substitutions of the 12S rRNA gene. Ochotona hyperborea and ten leporid taxa were subjected to a pairwise Construction of phylogenetic trees To infer molec- comparison (p distance). The three blocks seen in the transver- sion analysis correspond to comparisons between families (Lep- ular phylogenetic relationships among the nine leporid oridae vs. Ochotonidae) (a), between Pronolagus and the genera of our data set, we conducted neighbor-joining remaining genera of Leporidae (b), and between the remaining (NJ; Saitou and Nei, 1987) and maximum parsimony genera (c). 110 F. YAMADA et al.

Fig. 2. Phylogenetic trees inferred from the 12S rRNA sequences of nine genera of Leporidae. (a) Neighbor-joining tree based on genetic distances computed by Kimura’s (1980) two-parameter method; (b) 50% majority-rule consensus tree from six equally parsimo- nious trees (L= 597; CI = 0.55) recovered in weighted (2:1) maximum parsimony analysis. Sequences other than those of Japanese taxa were obtained from databases. The sequences of the 12S rRNA gene were aligned manually, introducing gaps to maximize homology. Excluding such regions, 710 selected sites were used. Only bootstrap values (based on 500 replicates) >50% are shown beside the relevant node. Numbers below nodes in the MP tree are Bremer support index values. Molecular phylogeny of Japanese leporids 111 was assessed by a scatterplot analysis with taxa for which As for the relationships among the seven remaining both 12S (> 720 bp) and cyt-b (671 bp) sequences were genera, each internal branch was short, bootstrap values available. In addition to the Japanese taxa, we analyzed were low, and no genera, including P. furnessi from the sequences obtained from databases: seven taxa for Lep- Ryukyu Islands, were clustered with strong support in oridae (B. idahoensis, B. monticularis, O. cuniculus, P. the NJ tree. The generic relationships are accordingly crassicaudatus, S. audubonii, S. floridanus, and R. diazi), consistent with those inferred from the MP tree con- and one for Ochotonidae (Ochotona hyperborea). In the structed with the same data set. These results agree 12S sequences, gaps were eliminated and 710 aligned with the conclusion of Halanych and Robinson (1999) that sites were used. Pairwise comparisons of transversions most of the leporid genera arose via a rapid divergence of the cyt b gene versus substitutions of the 12S gene did event. not appear to plateau, while those of all transitions of the cyt b gene showed apparent saturation in the compari- Lepus phylogeny The phylogenetic positions of the sons between genera (Fig. 1). Japanese lineages of hares was assessed by constructing an NJ tree with the cyt-b gene sequences (671 Generic phylogeny The 12S sequences for 12 taxa, bp). Halanych et al. (1999) reported that the majority of including Nesolagus sp. (Surridge et al., 1999), were used Lepus lineages were the result of a rapid radiation. Our to determine the phylogenetic position of the Amami rab- data clearly indicate that the lineage of the Japanese bit, P. furnessi. The average transversion : transition hare L. brachyurus is positioned at the base of the radia- ratio was 2.16 ± 0.09 SE. Among the 710 aligned sites tion (Fig. 3). In the data set, six distinct lineages of of the 12S region, 221 were variable, and 114 of these Lepus were evident. Each internal branch was short, were parsimony-informative. bootstrap values were low (<60%), and no lineages were We constructed phylogenetic trees with 12S gene clustered with strong support. This trend was consis- sequences, using Ochotona hyperborea as the outgroup tently observed in the NJ analysis with genetic distances (Fig. 2). In the NJ and MP trees, the two genera Neso- of transversions and in the MP analysis with various lagus and Pronolagus were shown to have separated weighting criteria (e.g. 8:1) with the same data set (data first and second, respectively, from the seven other not shown). In these analyses, Lepus brachyurus tended genera. The early splitting of Nesolagus was supported to be associated with a major lineage consisting of L. tim- by high supporting values (bootstrap = 78% in NJ and idus and related species (the timidus group) with moder- 91% in MP; Bremer support index = 11). The two lin- ate support values (bootstrap value 60–75%). However, eages of Nesolagus and Pronolagus had relatively long the genetic distance between the two major lineages was branches in the NJ tree. Sequence divergences (d) substantial (all substitutions: 0.093 on average), and at a between these two genera and the other seven genera, level similar to those among the six major lineages (0.097 Brachylagus, Bunolagus, Lepus, Oryctolagus, Pentalagus, on average; Table 3). Romerolagus, and Sylvilagus, were 0.129 on average, Lepus timidus ainu was shown to be positioned in while those among the seven genera were 0.084 on aver- the clade of the timidus group (Fig. 3). This lineage age (Table 2). appeared to have a close relationship with the following

Table 2. Pairwise comparisons of the 12S sequence differences for 11 leporid taxa plus an outgroup taxon Taxon 1 2 3456789101112 1 Ochotona hyperborea – 2 Brachylagus idahoensis 0.203 – 3 Bunolagus monticularis 0.210 0.102 – 4 Lepus brachyurus 0.210 0.104 0.083 – 5 Lepus timidus 0.197 0.100 0.081 0.026 – 6 Nesolagus sp. 0.205 0.139 0.137 0.121 0.118 – 7 Oryctolagus cuniculus 0.201 0.099 0.094 0.076 0.072 0.131 – 8 Pentalagus furnessi 0.186 0.087 0.084 0.078 0.073 0.130 0.076 – 9 Pronolagus crassicaudatus 0.224 0.148 0.141 0.108 0.098 0.144 0.117 0.126 – 10 Romerolagus diazi 0.195 0.099 0.075 0.078 0.070 0.123 0.078 0.064 0.119 – 11 Sylvilagus audubonii 0.212 0.095 0.090 0.065 0.065 0.143 0.074 0.069 0.117 0.077 – 12 Sylvilagus floridanus 0.182 0.078 0.091 0.065 0.067 0.130 0.065 0.057 0.110 0.065 0.045 – Note: Kimura two-parameter distances with all substitutions (710 bp). 112 F. YAMADA et al.

Fig. 3. Neighbor-joining tree of the cytochrome b gene (671 bp) with sequence divergences computed by Kimura’s (1980) two-param- eter method. Sequences other than those of Japanese taxa were obtained from databases. Only bootstrap values (based on 500 rep- licates) >50% are shown beside the relevant node. The evolution of leporids can be explained by three classes of rapid differentiation: emergence of the major generic lineages (I), of the major Lepus lineages (II), and of lineages of the L. timidus group (III).

taxa of Lepus from nearby regions of the Bering Strait: L. Hokkaido type differs substantially from those of the timidus from far eastern Russia, L. othus from Alaska, eastern tip of Eurasia (Chukotsk Peninsula) and central and L. arcticus from Greenland, at genetic distances of North America (L. townsendii), with genetic distances of 0.012–0.014 (Table 3). It is apparent, however, that the 0.029 (on average) and 0.041 (on average), respectively. Molecular phylogeny of Japanese leporids 113

Table 3. Pairwise comparisons of the cytochrome b sequence differences for 15 taxa of Lepus

Taxon 123456789101112131415 1 L. callotis – 2 L. californicus 0.035 – 3 L. alleni 0.039 0.024 – 4 L. americanus 0.094 0.090 0.092 – 5 L. arcticus 0.096 0.094 0.090 0.117 6 L. brachyurus (Honshu1) 0.092 0.099 0.094 0.107 0.092 – 7 L. brachyurus (Honshu2) 0.097 0.104 0.099 0.109 0.094 0.004 – 8 L. europaeus 0.110 0.117 0.108 0.122 0.099 0.106 0.112 – 9 L. othus 0.098 0.096 0.092 0.119 0.001 0.094 0.092 0.101 – 10 L. saxatilis 0.094 0.097 0.090 0.110 0.101 0.095 0.101 0.108 0.103 – 11 L. timidus (Hokkaido1) 0.099 0.097 0.094 0.115 0.011 0.098 0.103 0.101 0.012 0.099 – 12 L. timidus (Hokkaido2) 0.099 0.097 0.094 0.115 0.012 0.096 0.101 0.103 0.014 0.099 0.001 – 13 L. timidus (Russia) 0.094 0.096 0.089 0.119 0.001 0.094 0.096 0.101 0.003 0.103 0.012 0.014 – 14 L. timidus (UK) 0.096 0.097 0.087 0.110 0.018 0.085 0.084 0.089 0.017 0.083 0.029 0.031 0.020 – 15 L. townsendii 0.090 0.092 0.082 0.113 0.035 0.080 0.082 0.103 0.037 0.096 0.042 0.043 0.037 0.023 – Note: Kimura two-parameter distances with all substitutions (671 bp).

are considered ancestral or convergent on the basis of the DISCUSSION present phylogenetic study. Assuming a molecular clock Phylogenetic position of the Japanese leporids We and 30–40 Mya for the split between Ochotonidae and have elucidated the phylogenetic positions of the three Leporidae (branch lengths, 0.09; Fig. 2), the divergence Japanese leporid taxa, P. furnessi, L. brachyurus, and L. times for the majority of the genera (branch length 0.035 timidus ainu, endemic to the Ryukyu Islands, Honshu- on average) can be calculated to be 12–16 Mya. Shikoku-Kyushu, and Hokkaido, respectively. These Lepus brachyurus appears to have arisen in a rapid three taxa were proven to have independent mtDNA lin- radiation of Lepus species, because the lineage of this spe- eages that are separate from those of their continental cies differs distinctly from any other lineage of Lepus so counterparts, as discussed below. The present data pro- far examined (Fig. 3, Table 3). This indicates that L. vide useful clues for understanding the evolution of lep- brachyurus might display genetic endemism to a rela- orids in East Asia. tively large extent. However, further analysis is needed The lineages of all of the genera examined, including P. of geographically proximate species, including L. mand- furnessi, were shown to be distinct from one another (Fig. shuricus, which is distributed in northern Manchuria in 2). The early splitting of Nesolagus and Pronolagus is China and Korea (Jones and Johnson, 1965) and has been supported by considerably high bootstrap values, but the regarded as a subspecies of L. brachyurus (Radde, 1861; cladistic patterns of the remaining seven genera are Ellerman and Morrison-Scott, 1951; Imaizumi, 1960; rather ambiguous. The present results suggest a rapid Angermann, 1966, 1983; Corbet, 1978; Honacki et al., radiation of the nine generic lineages, as Halanych and 1982; Luo, 1988; Flux and Angermann, 1990; Hoffman, Robinson (1999) concluded based on a data set of 12S 1993). Recent studies of mtDNA sequences of continen- sequences of seven genera. Although we must await tal taxa have revealed that Lepus species from China, data for the two remaining extant taxa, an African (Poe- including L. mandshuricus, do not have a close relation- lagus) and a Himalayan (Caprolagus) endemic genus, to ship with L. brachyurus (Wu et al., 2000; Y. P. Zhang, obtain a complete picture of the phylogenetic relation- personal communication). The "second" radiation, speci- ships among the extant genera, it is worth emphasizing ation within Lepus, is estimated to have occurred 4-5 that the nine genera, including Pentalagus, have been Mya, i.e., in the early Pliocene, taking 12-16 Mya for the proven to have considerable evolutionary histories after split of genus Lepus (the genetic distance in the 12S the first radiation of the Leporidae. Pentalagus and between the taxa in the Lepus radiation is about 30% of Pronolagus, which share similar dental morphology that of the generic radiation; Table 2). (Hibbart, 1963; Dawson, 1981; Tomida, 1997), were once The tree of the cyt b gene (Fig. 3) suggests that Lepus considered to have evolved from a common ancestor, Plio- timidus ainu has a closer relationship with members of pentalagus, in Europe during the Pliocene (Fejfar, 1961; the Arctic clade (L. timidus from far eastern Russia, L. Daxner and Fejfar, 1967), but these morphological traits othus from Alaska and L. arcticus from Greenland; 114 F. YAMADA et al.

Halanych et al., 1999) than with L. timidus from the 2000), and even mustelids (Hosoda et al., 2000). UK. On the other hand, the mitochondrial lineage of the A similar substantial subdivision in the mitochondrial Hokkaido subspecies L. t. ainu is distinct from those of genes of a Hokkaido form from the continental and the related taxa, with a genetic distance (d) of >0.011 Sakhalin forms is evident in red-backed voles (Iwasa et (Table 3). Recent work on L. timidus from Italy (Pierpa- al., 2000, 2002b) and field mice (Serizawa et al., in oli et al., 1999) and China (Y. P. Zhang, personal commu- press). It is conceivable that the landmass of Hokkaido nication) supports our phylogenetic view of L. t. acted as a refugium for small mammals during several ainu. Since there are several subspecies of L. timidus periods of Quaternary glacial ages. geographically close to L. t. ainu (L. t. abei in the Kuril From the present study, it is apparent that the three Islands and L. t. orii in Sakhalin; Ellerman and Morrison- areas of the Japanese archipelago, Ryukyu, Honshu- Scott, 1951; Corbet, 1978; Flux and Angermann, 1990), Shikoku-Kyushu, and Hokkaido, now each preserve their further genetic study of the L. timidus populations in respective leporid taxon with different amounts of genetic northeastern Asia will be necessary to obtain a complete endemicity. It is possible that the zoogeographic traits view of the evolutionary relationships of the mtDNA of the Japanese leporids are a consequence of the evolu- lineages. From the genetic distance for cyt-b between L. tionary dynamics of leporids on the continents, in which t. ainu and Arctic-clade species (d = 0.012) and between three radiation events with intercontinental dispersal Lepus species (d = 0.1), L. t. ainu is thought to have sep- occurred at different geological times (I, II, and III in Fig. arated from the continental subspecies 0.5–0.6 Mya, i.e., 3). The radiation centers are likely to have shifted from in the middle or late Pleistocene, assuming the diver- low altitude to high altitude in the Northern Hemisphere, gence time of Lepus lineages to be 4–5 Mya. This esti- and from tropical (generic radiation), through temperate mate seems reasonable, because the island of Hokkaido (Lepus radiation), to Holarctic (radiation of the L. timidus has occasionally been connected with the Eurasian Con- group) zones. The present distribution of extant species tinent during the last 1 million years (Ono, 1990). of monotypic genera, Lepus species, and the timidus groups suggests that a wide geographic area consisting of Biogeography of the Japanese leporids The diver- North America, Africa, and Eurasia, has been involved in gence time of the Ryukyu taxon P. furnessi (12–16 Mya) the radiation of leporids. The distribution of three Jap- is congruent with the geological view that the Ryukyu anese leporids in Ryukyu, Honshu, and Hokkaido, is Archipelago, including the Amamioshima and Tokuno- apparently linked to the worldwide evolutionary history shima Islands on which P. furnessi now occurs, were con- of the Leporidae. nected to the Eurasian continent in the middle to late The geological ages of the three different classes of Miocene and isolated from the continent in the Pliocene genus, species (Lepus), and subspecies (L. timidus) can be (Kisaki and Ohshiro, 1981). It is also in accord with the estimated to be 12–16 Mya, 4–5 Mya, and 0.5–0.6 Mya, molecular information obtained for a Ryukyu endemic respectively, as mentioned above. This is more or less rodent, Tokudaia osimensis, in which divergence from concordant with geological changes in the climate, which related rodents is reported to have occurred in the middle became cool and dry through the last 15 Myr (Kennett, Miocene (ca. 12–14 Mya), assuming the rat-mouse split to 1995; Denys 1999), and may be related to three drastic be at 14 Mya (Suzuki et al., 2000). It is possible that the changes occurring 15 Mya, 5 Mya, and during the last 1- currently endangered endemic species Pentalagus fur- 2 Myr. Lepus is thought to have adapted to temperate nessi, as well as T. osimensis, have maintained their grasslands, and to have expanded its distribution by mak- populations in a small geographic area for a long evolu- ing use of such unique characteristics as acquisition of tionary time. precocity in the reproduction system and cursorial adap- Our result, 4–5 Mya, is earlier than the date suggested tation with enlargement of body size and strong hind by the paleontological evidence of Kawamura et al. legs suitable for an open habitat (Yamada et al., 1988, (1989), who concluded that L. brachyurus has existed on 1990). Thus, the development of Lepus may be coinci- the main island of Japan (Honshu) since the middle dent with a global development of grassland that would Pleistocene (0.5–0.6 Mya), based on fossil evidence. One have started 5 Mya. Accordingly, the environmental explanation might be that the genealogical splitting pre- changes of the middle Miocene (15 Mya) and the Quater- ceded the divergence of species, partly due to ancient nary ice ages, which would have provided new unoccupied polymorphism. However, it is more likely that the ecological niches, may have accelerated the global devel- ancestor of L. brachyurus immigrated to Japan during the opment of leporid genera and the timidus group, respec- Pliocene since the Honshu-Shikoku-Kyushu area is rich tively. in endemic mammals with such an ancient origin, as In conclusion, it is clear that the Japanese Islands have shown by molecular studies in dormice (Suzuki et al., played an important role in the differentiation and con- 1997), field mice (Serizawa et al., 2000), voles (Suzuki et servation of Leporidae lineages. These taxa, each inhab- al., 1999; Iwasa et al., 2002a), moles (Tsuchiya et al., iting a specific geographic area, have proven to be Molecular phylogeny of Japanese leporids 115 valuable for understanding not only leporid evolution but DICE, 1931 und Pliopentalagus GUREEV, 1964 (Lagomor- also environmental changes in East Asia from the late pha, Mammalia). Ann. Naturhistor. Mus. Wien 71, 37–55. Tertiary to the present. The leporids in East Asia, from Denys, C. (1999) Evolution of rodents in the New World and the Old World. In: Factors in the Emergence and Control of both the continent and the Japanese Islands, would pro- Rodent-borne Viral Diseases (Hantavirial and Arenal Dis- vide good points of reference for the evolution of other ter- eases) (eds.: J. F. Saluzzo, and B. Dodet), pp. 23–37. Elsevier, restrial animals that spread from the tropical to the Paris. northern temperate zone of the eastern periphery of Asia. Ellerman, J. R., and Morrison-Scott, T. C. S. (1951) Checklist of Palaearctic and Indian Mammals 1758 to 1946. Brit. Mus. (Nat. Hist.), London. We are most grateful to Dr. S. Higashi of Hokkaido University Fejfar, O. (1961) Die plio-pleistozänen Wirbeltierfaunen von ∨ for valuable suggestions and financial support. We thank Dr. S. Hajnácka∨ und Ivanovce (Slowakei), CSSR. III. Lagomorpha. Miura of the Forestry and Forest Products Research Institute, N. Jb. Geol. Päleont. Monatshefte 112, 267–282. Dr. S. Abe of Hokkaido University and Dr. Y. P. Zhang of Kun- Felsenstein, J. (1993) PHYLIP, version 3.5. Department of ming Institute of Zoology for helpful suggestions. We thank S. Genetics, University of Washington, Seattle. Wakana of the Central Institute for Experimental Animals, Dr. Flux, J. E. C., and Angermann, R. (1990) The hare and jack- H. Hirakawa of the Forestry and Forest Products Research Insti- rabbit. In: Rabbits, Hares and Pikas (eds.: A. Chapman et tute, and T. Nakaoka of the Erimo Town Museum Horoizumi, al.), pp. 61–94. IUCN, Gland. Hokkaido for providing the samples. We also wish to thank Halanych, K. M., Demboski, J. R., Van Vuuren, B. J., Klein, D. Drs. M. A. Iwasa and K. Serizawa of Hokkaido University and R., Robinson, T. J., and Cook, J. A. (1999) Cytochrome b S. Abe and Y. Handa of the Mammalogical Society of Amami, for phylogeny of North American hares and jackrabbits (Lepus, help and technical support. This study was supported in part Lagomorpha) and the effects of saturation in outgroup by a grant from the Japan Environment Agency (Global Environ- taxa. Mol. Phylogent. Evol. 11, 213–221. ment Research Program, F-1-4-2) given to F. Y. from 1996 to Halanych, K. M., and Robinson, T. J. (1999) Multiple substitu- 1998, and by the Joint Research Project under the Japan-China tions affect the phylogenetic utility of cytochrome b and 12S Scientific Cooperation Program by Japan Society for the Promo- rRNA data: Examining a rapid radiation in leporid (Lago- tion of Science (JSPS) and the Natural Science Foundation of morpha) evolution. J. Mol. Evol. 48, 369–379. China (NSFC). Hibbard, C. W. (1963) The origin of the P3 pattern of Sylvilagus, The nucleotide sequences reported in this paper appear in the Caprolagus, Oryctolagus and Lepus. J. Mammal. 44, 1–15. DDBJ, EMBL, and GenBank nucleotide sequence databases with Hoffman, R. S. (1993) Order Lagomorpha. In: Mammal Species the following accession numbers AB058604-AB058616, and of the World (eds.: D. E. Wilson, D. M. Reeder), pp. 807–827. AB059258. Smithsonian Inst. Press, Washington. Honacki, J. H., Kinman, K.E., and Koeppl, J. W. (1982) Mammal Species of the World. Allen Press, Lawrence. REFERENCES Hosoda, T., Suzuki, H., Harada, M., Tsuchiya, T., Han, S. H., Zhang, Y. P., Kryukov, A. P., and Lin, L. K. (2000) Evolu- Anderson, S., Bankier, A. T., Barrell, B. G., de Bruijn, M. H. L., tionary trends of the mitochondrial lineage differentiation Couslon, A. R., Drouin, J., Eperon, I. C., Nierlich, D. P., Roe, in species of genera Martes and Mustela. Genes Genet. B. A., Sanger, F., Schreier, P. H., Smith, A. J. H., Staden, Syst. 75, 259–267. R., Young, I. G. (1981) Sequence and organization of the Imaizumi, Y. (1960) An Illustrated Book of Japanese human mitochondrial genome. Nature 290, 457–465. Mammals. Hoikusha, Osaka. (in Japanese) Angermann, R. (1966) Der taxonomisch Status von Lepus Irwin, D. M., and Arnason, U. (1994) Cytochrome b gene of brachyurus und Lepus mandschuricus. Mitt. Zool. Mus. marine mammals: phylogeny and evolution. J. Mammal. Berl. 42, 321–335. Evol. 2, 37–55. Angermann, R. (1983) The taxonomy of Old World Lepus. Acta. Irwin, D. M., Kocher, T. D., and Wilson, A. C. (1991) Evolution Zool. Fenn. 174, 17–21. of the cytochrome b gene of mammals. J. Mol. Evol. 32, Angermann, R., Flux, J. E. C., Chapman, J. A., and Smith, A.T. 128–144. (1990) Lagomorph Classification. In Rabbits, Hares and Iwasa, M. A. and Suzuki, H. 2002a. Evolutionary networks of Pikas (eds.: A. Chapman et al.), pp. 7–13. IUCN, Gland. maternal and paternal gene lineages in Eothenomys voles Bremer, K. (1988) The limits of amino acid sequence data in endemic to Japan. J. Mammal., 83, in press. angiosperm phylogenetic reconstruction. Evolution 42, Iwasa, M. A., Kartavtseva, I. V., Dobrotvorsky, A. K., Panov, V. 795–803. V. and Suzuki, H. 2002b. Local differentiation of the north- Chapman, J. A., and Flux, J. E. C. (1990) Introduction and over- ern red-backed vole Clethrionomys rutilus (Rodentia, Arvi- view of the Lagomorphs. In: Rabbits, Hares and Pikas (eds.: colinae) in northeastern Asia inferred from mitochondrial A. Chapman et al.), pp. 1–6. IUCN, Gland. gene sequences. Z. Säugetierkd. (Mammalian Biology). 67, Corbet, G. B. (1978) The Mammals of the Palaearctic Region: a in press. Taxonomic Review. Brit. Mus. (Nat. Hist.), Cornell Univ. Iwasa, M. A., Utsumi, Y., Nakata, K., Kartavtseva, I. V., Neve- Press, London,. domskaya, I. A., Kondoh, N., and Suzuki, H. (2000). Geo- Corbet, G. B. (1983) A review of classification in the family Lep- graphic patterns of cytochrome b and Sry gene lineages in oridae. Acta. Zool. Fennica 174, 11–15. gray red-backed vole, Clethrionomys rufocanus (Mammalia, Dawson, M. R. (1981) Evolution of the modern lagomorphs. In: Rodentia) from Far East Asia including Sakhalin and Proceedings of the World Lagomorph Conference (eds.: K. Hokkaido. Zool. Sci. 17, 477–484. Myers, and C. D. MacInnes), pp. 1–8. University of Guelph IUCN (2000) 2000 IUCN Red List of Threatened Animals. Press, Guelph. IUCN, Gland. Daxner, G., and Fejfar, O. (1967) Über die Gattungen Alilepus Jones, J. K., and Johnson, D. H. (1965) Synopsis of the lagomor- 116 F. YAMADA et al.

phs and rodents of Korea. Univ. Kansas Publ. Mus. Nat. and Han, S.H. (2002) A spatial aspect on mitochondrial Hist. 16, 357–407. DNA genealogy in Apodemus peninsulae from East Kennett, J. .P. (1995) A review of polar climatic evolution during Asia. Biochem. Genet. 40. (in press) the Neogene, based on the marine sediment record. In: Sorenson, M. D. (1999) TreeRot, version 2. Boston University, Paleoclimate and Evolution with Emphasis on Human Ori- Boston, MA. gins. (eds.: E. S. Vrba, G. H Denton, T. C. Patrirdge, and L Sugimura, K., Sato, S., Yamada, F., and Hirakawa, H. (2000) . H. Burckle), pp. 49–64. Yale University Press, New Distribution and abundance of the Amami rabbit Penta- Haven. lagus furnessi in the Amami and Tokuno Islands, Kawamura, Y., Kamei, T., and Taruno, H. (1989) Middle and Japan. Oryx 34, 198–206. late Pleistocene mammalian faunas in Japan. Quat. Res. Surridge, A. K., Timmins, R. J., Hewitt, G. M., and Bell, D. J. 28, 317–326. (1999) Striped rabbits in Southeast Asia. Nature 400, 726. Kimura, M. (1980) A simple method for estimating evolutionary Suzuki, H., Iwasa, M., Harada, M., Wakana, S., Sakaizumi, M., rate of base substitutions through comparative studies of Han, S. H., Kitahara, E., Kimura, Y., Kartavtseva, I., and nucleotide sequences. J. Mol. Evol. 16, 111–120. Tsuchiya, K. (1999) Molecular phylogeny of red-backed voles Kisaki, K., and Ohshiro, I. (1981) A background of the Ryukyu in Far East Asia based on variation in ribosomal and mito- Islands. In: Natural History of the Ryukyu Islands (ed.: K. chondrial DNA. J. Mammal. 80, 512–521. Kisaki), pp. 8–37. Tsukiji-shoten, Tokyo. (in Japanese) Suzuki, H., Minato, S., Sakurai, S., Tsuchiya, K., and Fokin, I. Kocher, T. D., Thomas, W. K., Meyer, A., Edwards, S. V., Pääbo, M. (1997) Phylogenetic position and geographic differentia- S., Villablanca, F. X., and Wilson, A. C. (1989) Dynamics of tion of the Japanese dormouse, Glirulus japonicus, revealed mitochondrial DNA evolution in animals: amplification and by variations among rDNA, mtDNA and the Sry gene. Zool. sequencing with conserved primers. Proc. Natl. Acad. Sci. Sci. 14, 167–173. USA 86, 6196–6200. Suzuki, H., Tsuchiya, K, and Takezaki, N. (2000) A molecular Luo, Z. (1988) The Chinese Hare. China Forestry Publishing phylogenetic framework for the Ryukyu endemic rodents House, Peking. Tokudaia osimensis and Diplothrix legata. Mol. Pylogenet. Matsuzaki, T., Kamiya, M., and Suzuki, H. (1989) Laboratory Evol. 15, 15–24. rearing of the Amami rabbits (Pentalagus furnessi Stone, Swofford, D. L. (2001). PAUP*. Phylogenetic Analysis Using 1900) in captivity. Exp. Anim. 38, 65–69. Parsimony (*and Other Methods). Version 4.02b. Sinauer Mindell, D. P., Dick, C. W., and Baker, R. J. (1991) Phylo- Associates, Sunderland, MA. genetic relationships among megabats, microbats, and Tomida, Y. (1997) Why has Amami rabbit been classified in the primates. Proc. Natl. Acad. Sci. USA 88, 10322–10326. subfamily Palaeolaginae in Japan? FOSSILS 63, 20–28 (in Murphy,W. J., Eizirik, E., Johnson,W. E., Zhang, Y. P., Ryder, Japanese) O. A., and O'Brien, S. J. (2001) Molecular phylogenetics and Tsuchiya, K., Suzuki, H., Shinohara, A., Harada, M., Wakana, the origins of placental mammals. Nature 409, 614–618. S., Sakaizumi, M., Han, S. H., and Kryukov, A. P. (2000) Ono, Y. (1990) The northern landbridge of Japan. Quat. Res. Molecular phylogeny of East Asian moles inferred from the 29, 183–192. sequence variation of the mitochondrial cytochrome b gene. Pierpaoli, M., Riga, F., Trocchi, V., and Rndi, E. (1999) Species Genes Genet. Syst. 75, 17–24. distribution and evolutionary relationships of the Italian Wu, C. H., Li, H. P., Wang, Y. X., and Zhang, Y. P. (2000) Low hare (Lepus corsicanus) as described by mitochondrial DNA genetic variation of the Yunnan hare (Lepus comus G. Allen sequencing. Mol. Ecol. 8, 1805–1817. 1927) as revealed by mitochondrial cytochrome b gene Radde, G. (1861) Neue Säugetier-Arten aus Ost-Sibir. sequences. Biochem. Genet. 38, 149–155. Melanges. Biol. Bullet. de l’Acad. Imp. de St. Petersburg 3, Yamada, F., Shiraishi, S., Taniguchi, A., and Uchida, T. A. 684–686. (1990) Growth, development and age determination of the Saitou, N., and Nei, M. (1987) The neighbor-joining method: a Japanese hare, Lepus brachyurus. J. Mamm. Soc. Japan new method for reconstructing phylogenetic trees. Mol. 14, 65–77. Biol. Evol. 4, 406–425. Yamada, F., Shiraishi, S., and Uchida, T. A. (1988) Parturition Serizawa, K., Suzuki, H., and Tsuchiya, K. (2000) A phylo- and nursing behaviours of the Japanese hare, Lepus genetic view on species radiation in Apodemus inferred from brachyurus. J. Mamm. Soc. Japan 13, 59–68. variation of nuclear and mitochondrial genes. Biochem. Yamada, F., Sugimura, K., Abe, S., and Handa, Y. (2000) Genet. 38, 27–40. Present status and conservation of the endangered Amami Serizawa, K., Suzuki, H., Iwasa, A. I., Tsuchiya, K., Pavlenko, rabbit Petalagus furnessi. Tropics 10, 87–92 M. V., Kartavtseva, I. V., Chelomina, G. N., Dokuchaev, N.,