Genes Genet. Syst. (1997) 72, p. 79–90 Mitochondrial DNA differentiation in the Japanese brown japonica as revealed by restriction endonuclease analysis

Masayuki Sumida Laboratory for Biology, Faculty of Science, Hiroshima University, Higashihiroshima, Hiroshima 739, Japan

(Received 17 February 1997, accepted 28 April 1997)

To elucidate mtDNA differentiation in the Japanese brown frog Rana japonica, and compare it with results from allozyme analysis and crossing experiments, RFLP analysis was conducted on 78 from 16 populations in Honshu. Purified mtDNA was digested with eight six-base recognizing restriction enzymes and analyzed by 1% agarose-slab gel electrophoresis. Cleavage patterns of the mtDNA showed three distinct genome size classes: small (18.5 kb), middle (20.0 kb) and large (21.5 kb). Ten haplotypes (I~X) were observed among the 16 populations. The expected nucleotide divergences within populations ranged from 0 to 0.47% with a mean of 0.08%. The net nucleotide divergences among 16 populations ranged from 0 to 7.74% with a mean of 3.49%. The UPGMA dendrogram and NJ tree, which were constructed based on the net nucleotide divergences, showed that R. japonica diverged first into the eastern and western groups. The eastern group subsequently differentiated into a subgroup containing six populations and the Akita population, and the western group divided into several subgroups. These results, as well as the results of allozyme analysis and crossing experiments, suggest that the eastern and western groups have experienced secondary contact, and introgression has occurred in the Akita population.

eastern, western and northwestern groups, were repro- INTRODUCTION ductively isolated from one another by male hybrid steril- The Japanese brown frog, Rana japonica, is widely dis- ity (Sumida, 1996). tributed through Japan, from Honshu (except for the Mitochondrial genomes of higher are mater- northern end) to Kyushu and into China (except for the nally inherited, closed circular duplex DNA molecules. northeastern regions) (Okada, 1931, 1966; Maeda and They are compact in organization, conserved in gene con- Matsui, 1989). In Japan, the Ichinoseki population (east- tent, and range in size, among species, from 15,000 to ern Japan) of R. japonica is reproductively isolated from 23,000 base pairs (Brown, 1983; Kessler and Avise, 1985; the Hiroshima population (western Japan) by sterility in Moritz and Brown, 1986). The evolutionary rate at the male hybrids (Sumida, 1981). These male hybrids show nucleotide sequence level for mtDNA is rapid, per- varying degrees of testicular abnormality, and in the first haps 1-10 times faster than a typical single copy nuclear spermatogenetic meiosis, homologous chromosomes fail to DNA (Brown et al., 1979; Vawter and Brown, 1986). pair, resulting in the formation of univalents and varying Variations in fragment patterns revealed following diges- degrees of degeneration of the spermatids (Sumida, 1994). tion with restriction enzymes are referred to as restriction Starch-gel electrophoretic analyses of 25 loci encoding 15 fragment length polymorphisms (RFLPs). Base substitu- enzymes and three blood proteins of 505 individuals be- tion can create or eliminate cleavage sites for a particular longing to 25 populations showed distinct differentiation enzyme, thereby altering the number and size of frag- between the eastern and western groups. The introgres- ments detected by that enzyme alone (Dowling et al., sion of eastern alleles was noted at several loci in the 1996). A great deal of attention has been given to RFLP northwestern populations such as the Akita and Joetsu analyses of animal mtDNA within and among populations populations (Sumida and Nishioka, 1994a). The locus (Harrison, 1989; Avise, 1994; Moritz, 1994). Mitochon- linked with the sex-determining gene differed with the lo- drial DNA markers, originally used mainly for the analy- cal populations (Sumida and Nishioka, 1994b). Extensive sis of population structure and matriarchal phylogeny, crossing experiments among 20 populations from Honshu have also been used to elucidate the phenomena of hybrid- and Kyushu showed that three population groups, the ization and introgression (Streit et al., 1994). In the con- 80 M. SUMIDA

text of hybridization, the most frequently used approach was centrifuged at 36,000 rpm (Hitachi RPS-40 rotor) for has been to apply nuclear markers (mostly allozymes and 40 h at 20°C. The fraction containing closed circular ribosomal DNA sequences) and an RFLP analysis of mtDNA was collected, extracted three times with CsCl- mtDNA in concert to reveal the process and directionality saturated isopropanol to remove the dye and dialyzed of hybridization. against 0.1 mM EDTA (pH 8.0) for 20 h with two changes In the present study, the restriction fragment analysis of the medium. The mtDNA solutions were stored at of mtDNA was used to examine intraspecific variation in -20°C until used. Rana japonica and to compare the results with those obtained from crossing hybridization experiments and Restriction endonuclease digestion. Eight six-base allozyme analysis. recognizing restriction endonucleases, BglII, EcoRI, EcoRV, HpaI, PstI, SacI, SmaI and XhoI, were purchased from TaKaRa. The mtDNA was digested by incubating MATERIALS AND METHODS at 37°C for 2 h with appropriate amounts of the enzymes Animals and mtDNA preparation. A total of 78 adult under conditions described by the suppliers. Agarose- Rana japonica (66 females and 12 males) was collected slab gel (1% Sigma agarose) electrophoresis was carried from 16 localities throughout Honshu (Table 1). Most of out in the standard TAE buffer (0.04 M Tris, 0.001 M them were also used for allozyme analysis (Sumida and EDTA, 0.02 M sodium acetate, pH 8.3). After electro- Nishioka, 1994a) and crossing experiments (Sumida, phoresis, the gels were stained with 0.1 µg/ml ethidium 1996). Mitochondria were isolated from fresh livers or bromide and photographed under ultraviolet light. λ- ovaries. Tissues were homogenized in a decuple volume DNA digested with HindIII was used as molecular weight of STE buffer (0.25 M sucrose, 0.03 M Tris-HCl, 0.01 M standards. The cleavage patterns with each enzyme EDTA, 0.11 M NaCl, pH 7.6) cooled with ice, and centri- were designated by A, B, C, etc., in the order of discovery. fuged at 800 ×g for 10 min at 2°C. The supernatant was centrifuged at 10,000 ×g for 10 min at 2°C. MtDNA was Nucleotide sequence divergence and dendrogram. purified by CsCl ethidium bromide density gradient cen- Nucleotide sequence divergence between mtDNA haplo- trifugation described by Yonekawa et al. (1980) as follows types was calculated by the method of Gotoh et al. (1979). (Sumida, 1997). Mitochondria obtained from each frog The expected value for the nucleotide sequence divergence were lysed completely by suspending the pellet in 3.6 ml within a population (nucleotide diversity; π) was calcu- (final volume) of 0.6% sarcosyl, 10 mM EDTA and 10 mM lated from the estimated nucleotide divergence and fre- Tris-HCl (pH 8.0). The lysate was then dialyzed against quencies (Nei and Tajima, 1981). The nucleotide diver-

the same buffer for 4 h at room temperature. The volume gence (πxy) and the net nucleotide divergence (δ) between of lysate was adjusted to 3.6 ml by the same buffer, and 3.6 populations were estimated according to the method of Nei g of solid CsCl and 0.24 ml of 4.6 mg/ml ethidium bromide and Tajima (1981). Based on the net nucleotide diver- were added to the lysate and mixed well. Then the lysate gences, dendrograms were constructed by two different

Table 1. Specimens of Rana japonica used in the present study No. of specimens Prefecture Locality Population Total Female Male Eastern Iwate Ichinoseki-shi, Sannoseki 11 10 1 Ichinoseki Fukushima Onuma-gun, Aizutakada-machi 2 2 0 Aizutakada Tochigi Yaita-shi 3 2 1 Yaita Chiba Sawara-shi 4 4 0 Sawara Chiba Mobara-shi 6 6 0 Mobara Kanagawa Isehara-shi, Sannomiya 5 4 1 Isehara Northwestern Akita Akita-shi, Toyoiwaishidazaka 7 6 1 Akita Niigata Kitakanbara-gun, Nakajo-machi 5 5 0 Nakajo Niigata Santo-gun, Izumozaki-machi 1 1 0 Izumozaki Niigata Joetsu-shi 2 2 0 Joetsu Western Fukui Sakai-gun, Mikuni-cho 2 2 0 Mikuni Shizuoka Fujieda-shi 3 1 2 Fujieda Aichi Ama-gun, Tatsuta-mura 2 1 1 Tatsuta Shimane Ochi-gun, Ochi-cho 3 3 0 Ochi Hiroshima Higashihiroshima-shi, Saijo-cho 12 11 1 Saijo Hiroshima Saiki-gun, Saiki-cho, Iinoyama 10 6 4 Saiki Total 78 66 12 Mitochondrial DNA differentiation in Rana japonica 81 methods: the unweighted pair-group arithmetic average cleases. The cleavage patterns of each enzyme except clustering method (Sneath and Sokal, 1973) and the BglII, SacI, and SmaI, with which no site variations were neighbor-joining method (Saitou and Nei, 1987). found among individuals, are shown in Fig. 1. Hetero- plasmy was not observed in any individuals. The num- bers of cleavage patterns produced by site variation were RESULTS three (A~C) in HpaI and XhoI, four (A~D) in EcoRI and Cleavage pattern and genome size of mtDNA. PstI, and eight (A~H) in EcoRV (Fig. 1). Fragment Restriction enzyme cleavage patterns of R. japonica lengths of each cleavage pattern were determined by their mtDNA were analyzed with eight restriction endonu- mobilities relative to those of λ-DNA fragments digested

Fig. 1. Cleavage patterns of Rana japonica mtDNA digested with five restriction enzymes. A~H = types of cleavage patterns with each enzyme; M = λ-DNA digested with HindIII as molecular weight standards. l, m, s = three genome size (large, middle, and small) classes. 82 M. SUMIDA

with HindIII (Table 2). kb, and the large type (l) having about 21.5 ± 0.4 kb (Table Cleavage patterns of mtDNA showed three genome size 2, Fig. 2). These size variations were clearly observed in classes, with the small type (designated by s) having about one fragment of digests treated with BglII, EcoRI, EcoRV, 18.5 ± 0.4 kb, the middle type (m) having about 20.0 ± 0.3 HpaI, PstI and XhoI (Fig. 3).

Table 2. Lengths of restriction fragments of mtDNA in 16 populations of Rana japonica

Enzyme Lengtha Enzyme Lengtha

BglII l m s HpaIAl Bl Cl 11.5 11.5 11.5 14.7 14.5 9.5 10.0 8.5 7.0 3.5 3.3 4.4 Sum 21.5 20.0 18.5 3.3 3.1 3.5 – 0.6 3.3

EcoRI Am Bl Cm Ds – – 0.8 17.8 15.5 14.0 14.7 Sum 21.5 21.5 21.5 1.2 4.8 3.8 3.8

1.0 1.2 1.2 – PstIAs Bs Cm Ds – – 1.0 – 18.5 16.7 13.8 12.3 – 1.8 4.4 6.2 Sum 20.0 21.5 20.0 18.5 – – 1.8 – EcoRV As Bs Cs Ds Sum 18.5 18.5 20.0 18.5 18.5 18.0 15.5 13.5

– 0.5 2.5 4.5 XhoIAl Bl Cm – – 0.5 0.5 21.5 13.5 8.0 – 8.0 7.0 Sum 18.5 18.5 18.5 18.5 – – 5.0 Es Fl Gl Hm Sum 21.5 21.5 20.0 15.5 17.9 15.4 12.2 3.0 3.6 3.6 4.4 – – 2.5 2.5 –––0.9 Sum 18.5 21.5 21.5 20.0 a The lengths of fragments are approximately given in kilobase pairs (kb).

Fig. 2. BglII and EcoRV cleavage patterns of Rana japonica mtDNA showing three genome size classes. s = small type (18.5 kb); m = middle type (20.0 kb); l = large type (21.5 kb); M = λ-DNA digested with HindIII as molecular weight standards. Mitochondrial DNA differentiation in Rana japonica 83

Mapping of cleavage sites was carried out to identify tions (Table 4, Fig. 4). Haplotype I was predominant in common and different restriction sites between the cleav- six eastern populations, and haplotype II was also found in age patterns (Fig. 3). In these molecules, the fragment the Yaita and Mobara populations. Haplotype III was containing the size variation was represented by dashed observed only in the Akita population. Haplotype IV was lines. Restriction sites ranging in number from one to predominant in three northwestern populations, and five per cleavage pattern (2.5 sites on average) were found haplotype V was also found in the Nakajo population. on these maps. As shown in Table 3, 10 haplotypes were Haplotype VI was predominant in the Fujieda and Tatsuta observed in 16 populations of R. japonica. They were des- populations, and haplotype VII was only predominant in ignated by Roman numerals I~X. Each haplotype had the Mikuni population. Haplotype IX was observed in 14~22 cleavage sites. three western populations, and haplotype VIII and X were also found in the Ochi and Saijo populations, respectively Geographic variation of mtDNA. The geographic dis- (Table 4, Fig.4). tributions of three genome sizes and 10 haplotypes are shown in Table 4 and Fig. 4. The frequencies of small Nucleotide sequence divergence among mtDNA size mtDNA, s, were high in five eastern and three west- haplotypes. The numbers of common and different sites ern populations. The frequencies of middle size mtDNA, between pairs of mtDNA haplotypes were totaled from the m, were high in the Nakajo, Izumozaki, Fujieda, and cleavage patterns, and the nucleotide sequence diver- Tatsuta populations. The large size mtDNA, l, was pre- gences were estimated by the method of Gotoh et al. (1979) dominant in the Isehara, Akita, Joetsu, and Saiki popula- (Table 5). The nucleotide sequence divergences between

Fig. 3. Cleavage maps of Rana japonica mtDNA digested with eight restriction enzymes. O = common sites of all cleavage types; A~H = the sites of each cleavage type; l, m, s = three genome size (large, middle, and small) classes; numbers = lengths (kb) of restriction frag- ments. The linear maps are arranged by setting one of common sites of all cleavage pat- terns on the left tip. On these maps, the common and different sites between cleavage types are distinctly showed although the orders of sites are not represented.

Table 3. Haplotypes of mtDNA in 16 populations of Rana japonica

Restriction enzyme Total number of Haplotype BglII EcoRI EcoRV HpaI PstI SacI SmaI XhoI cleavage sites I ABDAAAAA 15 IIABBAAAAA 14 IIIABDBAAAA 16 IVAAGCBAAB 19 V AAFCBAAB 18 VIACHCBAAC 22 VIIADCABAAB 16 VIII AB AAB AAB 15 IXABEACAAB 17 X ABEADAAB 16 84 M. SUMIDA

Table 4. Frequencies of genome sizes and haplotypes of mtDNA in 16 populations of Rana japonica

Genome sizes Haplotypes PopulationNo.of (Frequency) (Frequency) frogs l m s I II III IV V VI VII VIII IX X Ichinoseki 11 2 9 11 (0.18) (0.82) (1.00) Aizutakada 2 2 2 (1.00) (1.00) Yaita 3 3 2 1 (1.00) (0.67) (0.33) Sawara 4 4 4 (1.00) (1.00) Mobara 6 6 5 1 (1.00) (0.83) (0.17) Isehara 5 3 2 5 (0.60) (0.40) (1.00) Akita 7 5 2 7 (0.71) (0.29) (1.00) Nakajo 5 2 3 3 2 (0.40) (0.60) (0.60) (0.40) Izumozaki 1 1 1 (1.00) (1.00) Joetsu 2 2 2 (1.00) (1.00) Mikuni 2 2 2 (1.00) (1.00) Fujieda 3 1 2 3 (0.33) (0.67) (1.00) Tatsuta 2 2 2 (1.00) (1.00) Ochi 3 1 2 2 1 (0.33) (0.67) (0.67) (0.33) Saijo 12 3 2 7 84 (0.25) (0.17) (0.58) (0.67) (0.33) Saiki 10 6 4 10 (0.60) (0.40) (1.00)

10 haplotypes ranged from 0.46% between IV and V to nucleotide divergences among the six eastern populations 8.46% between III and V. The nucleotide sequence diver- ranged from 0 to 0.06% with a mean of 0.02%, whereas gences between haplotypes belonging to the same popula- those among the six western populations ranged from 0 to tions were 0.58% between I and II, 0.46% between IV and 4.39% with a mean of 2.85%. The net nucleotide diver- V, 1.07% between VIII and IX, and 0.51% between IX and gences among three northwestern populations except the X. The nucleotide sequence divergences between haplo- Akita population ranged from 0 to 0.07% with a mean of types belonging to different populations ranged from 0.05%, whereas those between these three populations 1.68% between I and III and between VIII and X to 8.46% and the Akita population ranged from 7.45% to 7.74% with between III and V with a mean of 4.19% (Table 5). a mean of 7.55%. The net nucleotide divergences be- tween the six eastern populations and the Akita popula- Intra- and inter-populational variability. The nucle- tion ranged from 1.68% to 1.77% with a mean of 1.70%, otide diversities (π) within the Yaita, Mobara, Nakajo, whereas those between the six western populations and Ochi, and Saijo populations were estimated to be 0.26%, the Akita population ranged from 4.28% to 7.37% with a 0.16%, 0.22%, 0.47%, and 0.23%, respectively, whereas mean of 5.86% (Table 6). those within the other 11 populations were zero. The av- erage nucleotide divergences between 16 populations of Phenetic relationships among populations. Based

Rana japonica (πxy) ranged from 0 to 7.85% with a mean of on the net nucleotide divergences (Table 6), two dendro- 3.57% (Table 6; above diagonal). Considering intrapopu- grams were constructed by using the UPGMA and NJ lational variability, the net nucleotide divergences be- methods (Figs. 5 and 6). The UPGMA dendrogram tween populations (δ) were calculated as suggested by Nei showed that R. japonica distributed in Japan was divided and Tajima (1981) (Table 6; below diagonal). The net into the eastern group containing seven populations and Mitochondrial DNA differentiation in Rana japonica 85 the western group containing nine populations. The group was divided into two subgroups, one of which con- eastern group diverged into the Akita population and a tained the Mikuni, Ochi, Saijo, and Saiki populations and subgroup containing six populations, whereas the western another which contained the Tatsuta, Fujieda, Izumozaki, Joetsu, and Nakajo populations (Fig. 5). The NJ tree also showed that Rana japonica distributed in Honshu di- verged into the eastern and western groups. The western group differentiated into a subgroup containing the Ochi, Saijo, and Saiki populations, which was followed by the Mikuni population, and the group divided into two sub- groups containing five populations. On the other hand, the eastern group differentiated into a subgroup contain- ing six populations, which was followed by the Akita popu- lation (Fig. 6).

DISCUSSION Inter- and intra-populational variabilities. Mito- chondrial DNA variation has been extensively surveyed in a number of animal species at the intraspecific level (Avise et al., 1987; Avise, 1994). According to these studies, in- traspecific variation of mtDNA was 0.1~8.7%, 2.8% on av- erage, in 21 species including invertebrates, fishes, am- phibians, reptiles, birds, and mammals. The present study revealed that interpopulational variation of mtDNA among 16 populations of R. japonica was 0~7.7%, 3.5% on average. These values are not significantly different from those stated above. In , sequence diver- gences within species were reported in several species (Table 7), all of which are also within the values reported by Avise et al. (1987). On the other hand, interspecific and intersubspecific se- quence divergences were estimated among several am- phibian species or subspecies (Table 8). The present study revealed that sequence divergences between the eastern and western groups of R. japonica were 2.3~7.7%, 4.8% on average. These values are somewhat smaller Fig. 4. Geographic distributions of three genome sizes and 10 than most of the interspecific sequence divergences shown haplotypes of Rana japonica mtDNA. in Table 8, and are similar to those for four European

Table 5. Numbers of common (ci) and different (di) sites, and nucleotide sequence divergence (%) between pairs of mtDNA haplotypes

Haplo- I II III IV V VI VII VIII IX X type ci di ci di ci di ci di ci di ci di ci di ci di ci di ci di I – 14 1 14 3 12 10 11 11 13 11 12 7 13 4 13 6 13 5 II 0.58 – 13 4 12 9 11 10 13 10 12 6 13 3 13 5 13 4 III 1.68 2.36 – 11 13 10 14 12 14 11 10 12 7 12 9 12 8 IV 5.64 5.17 7.45 – 18 1 18 5 14 7 13 8 14 8 13 9 V 6.53 6.05 8.46 0.46 – 17 6 13 8 13 7 14 7 13 8 VI 5.71 5.28 7.37 2.14 2.67 – 15 8 14 9 15 9 14 10 VII 4.18 3.65 6.05 3.65 4.37 3.86 – 13 5 14 5 13 6 VIII 2.36 1.80 4.18 4.37 3.90 4.54 2.89 – 15 2 14 3 IX 3.40 2.89 5.17 4.10 3.65 4.28 2.70 1.07 – 16 1 X 2.89 2.36 4.68 4.83 4.37 4.96 3.40 1.68 0.51 –

Numbers of common and different sites are given above the diagonal. Nucleotide sequence divergence is given below. 86 M. SUMIDA

Table 6. Average nucleotide divergence (%) and net nucleotide divergence (%) of mtDNA among 16 populations of Rana japonica

Population No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Ichinoseki 1 – 0 0.19 0 0.10 0 1.68 6.00 5.64 5.64 4.18 5.71 5.71 2.70 3.23 3.40 Aizutakada 2 0 – 0.19 0 0.10 0 1.68 6.00 5.64 5.64 4.18 5.71 5.71 2.70 3.23 3.40 Yaita 3 0.06 0.06 – 0.19 0.22 0.19 1.90 5.84 5.49 5.49 4.01 5.57 5.57 2.52 3.06 3.23 Sawara 4 0 0 0.06 – 0.10 0 1.68 6.00 5.64 5.64 4.18 5.71 5.71 2.70 3.23 3.40 Mobara 5 0.02 0.02 0.01 0.02 – 0.10 1.80 5.92 5.56 5.56 4.09 5.64 5.64 2.61 3.14 3.31 Isehara 6 0 0 0.06 0 0.02 – 1.68 6.00 5.64 5.64 4.18 5.71 5.71 2.70 3.23 3.40 Akita 7 1.68 1.68 1.77 1.68 1.72 1.68 – 7.85 7.45 7.45 6.05 7.37 7.37 4.51 5.01 5.17 Nakajo 8 5.89 5.89 5.60 5.89 5.73 5.89 7.74 – 0.18 0.18 3.94 2.35 2.35 4.10 4.16 3.92 Izumozaki 9 5.64 5.64 5.36 5.64 5.48 5.64 7.45 0.07 – 0 3.65 2.14 2.14 4.28 4.34 4.10 Joetsu 10 5.64 5.64 5.36 5.64 5.48 5.64 7.45 0.07 0 – 3.65 2.14 2.14 4.28 4.34 4.10 Mikuni 11 4.18 4.18 3.88 4.18 4.01 4.18 6.05 3.83 3.65 3.65 – 3.86 3.86 2.83 2.93 2.70 Fujieda 12 5.71 5.71 5.44 5.71 5.56 5.71 7.37 2.24 2.14 2.14 3.86 – 0 4.45 4.50 4.28 Tatsuta 13 5.71 5.71 5.44 5.71 5.56 5.71 7.37 2.24 2.14 2.14 3.86 0 – 4.45 4.50 4.28 Ochi 14 2.47 2.47 2.39 2.47 2.30 2.47 4.28 3.76 4.05 4.05 2.60 4.22 4.22 – 0.91 0.72 Saijo 15 3.12 3.12 2.93 3.12 2.95 3.12 4.90 3.94 4.23 4.23 2.82 4.39 4.39 0.56 – 0.17 Saiki 16 3.40 3.40 3.10 3.40 3.23 3.40 5.17 3.81 4.10 4.10 2.70 4.28 4.28 0.49 0.06 –

Average nucleotide divergence (πxy) is given above the diagonal and the net nucleotide divergence (δ) is given below.

Fig. 5. UPGMA dendrogram for 16 populations of Rana japonica based on the net nucleotide diver- gences.

Triturus species or three Xenopus laevis subspecies. each of R. japonica, R. ornativentris and R. tagoi. Thus it is supposed that the differentiation between the Shaffer and Mcknight (1996) revealed that within-locality eastern and western groups of R. japonica is at a subspe- variation was 0.1% in Ambystoma velasci and A. tigrinum cific level on the basis of mtDNA sequence divergences. nebulosum. Hanzawa et al. (1987) estimated that the ex- The present study showed that the expected nucleotide pected intrapopulational variability was 0.07~0.23% in divergences within populations ranged from 0 to 0.47% Japanese dace, Tribolodon hakonensis. Avise and with a mean of 0.08%. There is little data for the level of Lansman (1983) reported that the intrapopulational vari- variability within a population derived from a single loca- ability of pocket gophers was less than 1%. The values tion in amphibians. Tanaka et al. (1994) reported that for R. japonica of the present study correspond to those for the intrapopulational sequence divergences were 0.8% in amphibians, fish and mammals stated above. Mitochondrial DNA differentiation in Rana japonica 87

Fig. 6. Neighbor-joining tree for 16 populations of Rana japonica based on the net nucleotide divergences.

Table 7. Nucleotide sequence divergence of mtDNA within species estimated from RFLP analysis No. of No. of No. of Sequence Species populations animals restriction divergence (%) Reference surveyed surveyed enzymes (average) Bombina orientaris 3 28 11 0.6 – 1.6 (1.1) Lee and Park (1991b) Bufo bufo 3 8 11 0.1 – 1.8 (1.1) Lee and Park (1992a) B. stejnegeri 2 12 11 0.2 Lee and Park (1992a) B. terrestris 24 117 – 1.0 Avise et al. (1987)* Hyla japonica 4 38 11 0.1 – 2.6 (1.7) Lee and Park (1992b) Rana amurensis 2 8 11 2.2 Lee and Park (1991a) R. dybowskii 3 17 11 0.4 – 0.9 (0.7) Lee and Park (1991a) R. nigromaculata 3 13 11 0.7 – 1.1 (0.9) Lee and Park (1991a) R. rugosa 3 19 11 0 – 1.1 (0.7) Lee and Park (1991a) R. japonica 16 78 8 0 – 7.7 (3.5) Present study Rhacophorus arboreus 4 22 12 0.3 – 1.5 (0.9) Wilkinson et al. (1996)** Rh. schlegelii 4 4 12 1.0 – 2.0 (1.3) Wilkinson et al. (1996)** Rh. viridis 3 5 12 0.6 – 1.4 (1.1) Wilkinson et al. (1996)** Ambystoma texanum 9 11 18 0.2 – 2.3 (1.2) Spolsky et al. (1992) A. jeffersonianum 3 3 18 1.0 – 1.6 (1.3) Spolsky et al. (1992) A. laterale 3 4 18 0.3 – 0.6 (0.4) Spolsky et al. (1992) * cited from unpublished data by Bermingham and Avise. ** A 2-kb fragment of mtDNA, which includes part of the 12S rRNA, 16S rRNA, and the intermittent tRNAVal genes, was amplified and used for RFLP analysis.

Geographic population divergence. The present (range, 1.68% to 1.77%), whereas that between the former study on mtDNA variation suggests that R. japonica in and the nine northwestern and western populations was Honshu was first divided into the eastern and western 6.42% (range, 4.28% to 7.74%). This suggests that the groups. The eastern group thereafter diverged into the mtDNA of the Akita population was derived from that of Akita population and a subgroup containing six popula- the eastern group. tions, and the western group differentiated into several Conditions suitable for interspecific transfer of mtDNA subgroups. The mean divergence between the Akita may occasionally occur either when sympatric but ecologi- population and the six eastern populations was 1.70% cally separated species interbreed locally or when species 88 M. SUMIDA

Table 8. Nucleotide sequence divergence of mtDNA among species or subspecies estimated from RFLP analysis No. of No. of Sequence Genus or species or restriction divergence (%) Reference species subspecies enzymes (average) surveyed Bombina 2 species 17 9.4 Szymura et al. (1985) Bufo 2 species 11 6.0 – 8.1 ( 7.3) Lee and Park (1992a) Hyla 5 species 16 15.6 – 60.0 (26.0) Kessler and Avise (1985) Hyla 2 species 11 13.1 – 14.6 (14.2) Lee and Park (1992b) Rana 2 species 19 8.1 Spolsky and Uzzell (1984) Rana 5 species 11 9.2 – 21.9 (19.2) Lee and Park (1991a) Rhacophorus 3 species 12 1.6 – 4.7 ( 2.8) Wilkinson et al. (1996)* Xenopus 7 species 11 11.0 – 40.0 (22.8) Carr et al. (1987) Ambystoma 3 species 18 7.8 – 9.2 ( 8.5) Spolsky et al. (1992) Triturus 4 species 11 3.9 – 7.1 Wallis and Arntzen (1989) Xenopus laevis 3 subspecies 11 3.0 – 7.0 ( 5.0) Carr et al. (1987) Rh. viridis 2 subspecies 12 1.4 Wilkinson et al. (1996)* R. japonica 2 groups 8 2.3 – 7.7 ( 4.8) Present study

* A 2-kb fragment of mtDNA, which includes part of the 12S rRNA, 16S rRNA, and the intermittent tRNAVal genes, was amplified and used for RFLP analysis.

with parapatric distributions hybridize along their zone of population was formed by an introgression between the contact (Szymura et al., 1985). Several studies have pro- eastern and western groups followed by random drift at vided evidence for interspecific mtDNA transfer in the ab- each locus. sence of nuclear gene introgression (Ferris et al., 1983; Sumida (1996) showed that although there was neither Powell, 1983; Spolsky and Uzzell, 1984). On the other gametic isolation nor hybrid inviability among popula- hand, nuclear and mitochondrial genes were found to tions of R. japonica, a remarkable preponderance of male change concordantly within hybrid populations or in a reciprocal hybrids was observed among three groups of transect across hybrid zones (Avise et al., 1984; Szymura populations, the eastern, western, and northwestern et al., 1985; Harrison et al., 1987; Wallis and Arntzen, groups. In crosses within each of the three groups, 1987; Rand and Harrison, 1989; Sumida and Ishihara, 52.7~55.7% of frogs were males, whereas 93.0 and 88.0% 1997). The present study suggests that mitochondrial of frogs were males in the reciprocal hybrids between the gene flow may have occurred into the Akita population eastern and western groups and between the northwest- from the eastern populations of R. japonica. The pat- ern and western groups, respectively. These three terns for mtDNA variation of R. japonica were largely groups were reproductively isolated from one another by concordant with allozyme data. It is probable that mito- male hybrid sterility. The degree of male hybrid sterility chondrial transfer may have occurred in the Akita popula- was largest in reciprocal hybrids between the western and tion in addition to some nuclear gene flow. eastern groups, and smallest in reciprocal hybrids be- Sumida and Nishioka (1994a) analyzed 25 loci control- tween the northwestern and eastern groups. Also, the ling 15 enzymes and three blood proteins of 505 individu- number of abnormal testes in crosses within the north- als belonging to 25 populations of R. japonica using western group was slightly higher than that in crosses starch-gel electrophoresis. Distinct differentiation was within eastern or western groups. observed between the eastern and western groups. The The data on mitochondrial DNA variations as well as genetic distances between the western and eastern groups those on allozyme analysis and crossing experiments sug- ranged from 0.104 to 0.239, with a mean of 0.156. At the gest that after R. japonica diverged into the eastern and Hb-II locus, the eastern and western groups were charac- western groups by geographical isolation, they might have terized by a fixed allelic difference, whereas at the Pep-A come into secondary contact with each other, and hybrid- and Alb loci, introgressions of the ‘eastern’ alleles were ization occurred in the northwestern region such as the noted in the northwestern populations. A steep cline in Akita population, leading to substantial introgression of allelic frequencies was observed in the northwestern popu- mtDNA and nuclear genes. They were probably isolated lations at the Pep-A locus. In the Akita population, the geographically thereafter by marine transgression during Hb-II locus is fixed for the ‘western’ allele, whereas the Alb the interglacial period or other biogeographical events and Pep-A loci are fixed for the ‘eastern’ allele. Unique (Minato et al., 1965). alleles at the LDH-B and MDH-B loci are found in the Based on a conventional mtDNA clock calibration for Akita population. These results suggest that the Akita mammals of 2% nucleotide sequence divergence per mil- Mitochondrial DNA differentiation in Rana japonica 89 lion years (Brown et al., 1979; Brown, 1985), it is assumed (1984) Characterization of mitochondrial DNA variability in that R. japonica diverged into the eastern and western a hybrid swarm between subspecies of bluegill sunfish (Lepomis macrochirus). Evolution 38, 931–941. groups over two million years ago, and then the Akita Avise, J. C., Arnold, J., Ball, R. M., Bermingham, E., Lamb, T., population diverged from the former about 0.85 million Neigel, J. E., Reeb, C. A., and Saunders, N. C. (1987) In- years ago. When divergence time is estimated from Nei’s traspecific phylogeography: The mitochondrial DNA bridge genetic distances of allozyme data by the equation of T = between population genetics and systematics. Annu. Rev. 5 × 106D (Nei, 1975, 1987), it was speculated that R. Ecol. Syst. 18, 489–522. Benzen, P., Leggett, W. C., and Brown, G. G. (1988) Length and japonica diverged into the eastern and western groups restriction site heteroplasmy in the mitochondrial DNA of about 0.8 million years ago and then the Akita population American shad (Alosa sapidissima). Genetics 118, 509–518. diverged from the former about 0.5 million years ago Bermingham, E., Lamb, T., and Avise, J. C. (1986) Size polymor- (Sumida and Nishioka, 1994a). The differences in diver- phism and heteroplasmy in the mitochondrial DNA of lower gence time between mtDNA and allozyme data are prob- vertebrates. J. Hered. 77, 249–252. Brown, W. M. (1983) Evolution of animal mitochondrial DNA. ably related to independent calibration of divergence time. In: Evolution of Genes and Proteins (eds.: M. Nei and R. Koehn), pp. 62–88. Sinauer, Sunderland, MA. Size variation. The present study revealed that there Brown, W. M. (1985) The mitochondrial genome of animals. In: were three size classes (s, l, and m) of mtDNA in R. Molecular Evolutionary Genetics (ed.: R. Maclntyre), pp. 95– japonica, corresponding to total genome sizes of approxi- 130. Plenum, New York. Brown, W. M., George, M. Jr., and Wilson, A. C. (1979) Rapid evo- mately 18.5, 20.0 and 21.5 kb, respectively. The small lution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. type (s) was found in most of the eastern populations, USA 76, 1967–1971. whereas the middle (m) and large (l) types were found in Carr, S. M., Brothers, A. J., and Wilson, A. C. (1987) Evolutionary the northwestern and western populations. Large-scale inferences from restriction maps of mitochondrial DNA from size variations of the mtDNA molecules have been ob- nine taxa of Xenopus frogs. Evolution 41, 176–188. Cook, D. I. and Zouros, E. (1994) The highly variable and highly served in many species of invertebrates, fish, amphibians, mutable mitochondrial DNA molecule of the deep sea scallop and reptiles (Reilly and Thomas, 1980; Densmore et al., Placopecten megellanicus. Nautilus (Suppl.) 2, 85–90. 1985; Harrison et al., 1985; Bermingham et al., 1986; Densmore, L. D., Wright, J. W., and Brown, W. M. (1985) Length Moritz and Brown, 1986, 1987; Benzen et al., 1988; Lee variation and heteroplasmy are frequent in mitochondrial and Park, 1991a, 1992b; Lee et al., 1992; Lee and Kim, DNA from parthenogenetic and bisexual lizards (genus Cnemidophorus). Genetics 110, 689–707. 1993; Cook and Zouros, 1994). The present results sug- Dowling, T. E., Moritz, C., Palmer, J. D., and Rieseberg, L. H. gest that mtDNA length variations were generated by ei- (1996) Nucleic acids III: Analysis of fragments and restriction ther accumulations of individually small additions/dele- sites. In: Molecular Systematics (eds.: D. M. Hillis, C. tions or saltational large-size changes, but it does not al- Moritz and B. K. Mable), pp. 249–320. Sinauer, Sunderland, low for a conclusion about the mechanisms of mtDNA ge- MA. Ferris, S. D., Sage, R. D., Huang, C.-M., Nielsen, J. T., Ritte, U., nome size changes. Further sequencing or fine-scale and Wilson, A. C. (1983) Flow of mitochondrial DNA across a mapping will be required to decide the mechanistic basis of species boundary. Proc. Natl. Acad. Sci. USA 80, 2290– changes in the mtDNA genome size of R. japonica. 2294. Gotoh, O., Hayashi, J.-I., Yonekawa, H., and Tagashira, Y. (1979) The author thanks M. Nishioka for her advice and encourage- An improved method for estimating sequence divergence be- ment during the research, H. Yonekawa for his kind guidance re- tween related DNAs from changes in restriction endonu- garding the method for mtDNA preparation, and O. Gotoh for clease cleavage sites. J. Mol. Evol. 14, 301–310. supplying the program software for calculating the nucleotide se- Hanzawa, N., Yonekawa, H., and Numachi, K.-I. (1987) Variabil- quence divergences among haplotypes. Thanks are also ex- ity of mitochondrial DNA in Japanese dace, Tribolodon tended to H. Ueda, M. Ryuzaki, K. Sekiya, K. Kinebuchi, T. hakonensis (Cyprinidae). Jpn. J. Genet. 62, 27–38. Ishihara, H. Ohtani, Y. Hasegawa, and J. Marunouchi for collect- Harrison, R. G. (1989) Animal mitochondrial DNA as a genetic ing and providing valuable specimens, and to C. Katagiri, T. Fujii, marker in population and evolutionary biology. Trends T. Hongo, S. Nakagawa, H. Ikeda, and Y. Kokuryo for their kind Ecol. Evol. 4, 6–11. aid in collecting specimens. This study was supported by a Harrison, R. G., Rand, D. M., and Wheeler, W. C. (1985) Mito- Grant-in-Aid for Scientific Research from the Ministry of Educa- chondrial DNA size variation within individual crickets. tion, Science, Sports, and Culture of Japan. Science 228, 1446–1448. Harrison, R. G., Rand, D. M., and Wheeler, W. C. (1987) Mito- chondrial DNA variation in field crickets across a narrow hy- REFERENCES brid zone. Mol. Biol. Evol. 4, 144–158. Kessler, L. G. and Avise, J. C. (1985) A comparative description of Avise, J. C. (1994) Molecular markers, natural history and evolu- mitochondrial DNA differentiation in selected avian and tion. Chapman Hall, New York. other vertebrate genera. Mol. Biol. Evol. 2, 109–125. Avise, J. C. and Lansman, R. A. (1983) Polymorphism of mito- Lee, H. Y. and Kim, I. S. (1993) Analysis of mitochondrial DNA chondrial DNA in populations of higher animals. In: Evolu- size difference in Korean frog, Rana amurensis. Korean J. tion of Genes and Proteins. (eds.: M. Nei and R. Koehn), pp. Genet. 15, 17–22. 147–164. Sinauer, Sunderland, MA. Lee, H. Y. and Park, C. S. (1991a) Genetic studies on Korean Avise, J. C., Bermingham, E., Kessler, L. G., and Saunders, N. C. anurans: On the mitochondrial DNA differentiation in frogs 90 M. SUMIDA

of the genus Rana. Korean J. Genet. 13, 1–16. revisited: Genetic differentiation and molecular phylogene- Lee, H. Y. and Park, C. S. (1991b) On the fragment patterns and tics of the tiger salamander Ambystoma tigrinum (Amphibia: variation of mitochondrial DNA in the red bellied frog, Caudata) complex. Evolution 50, 417–433. Bombina orientalis and small round, Kaloula borealis. Sneath, P. H. A. and Sokal, R. R. (1973) Numerical . Bull. Inst. Basic Sci. Inha Univ. 12, 83–87. W. H. Freeman and Co., San Francisco. Lee, H. Y. and Park, C. S. (1992a) Genetic studies on Korean Spolsky, C. and Uzzell, T. (1984) Natural interspecies transfer of anurans: Mitochondrial DNA variation of asian toad (Bufo mitochondrial DNA in amphibians. Proc. Natl. Acad. Sci. bufo) and water toad (B. stejnegeri). Korean J. Genet. 14, USA 81, 5802–5805. 247–255. Spolsky, C., Phillips, C. A., and Uzzell, T. (1992) Antiquity of Lee, H. Y. and Park, C. S. (1992b) Genetic studies on Korean clonal salamander lineage revealed by mitochondrial DNA. anurans: Length and restriction site variation in the mito- Nature 356, 706–708. chondrial DNA of tree frogs, Hyla japonica and H. Streit, B., Stadler, T., Schwenk, K., Ender, A., Kuhn, K., and suweonensis. Korean J. Zool. 35, 219–225. Schierwater, B. (1994) Natural hybridization in freshwater Lee, H. Y., Park, C. S., Jung, E. K., and Kim, I. S. (1992) The study animals. Naturwissenschaften 81, 65–73. on the macro-mitochondrial DNA using restriction enzymes Sumida, M. (1981) Studies on the Ichinoseki population of Rana in Korean anura, Rana rugosa. Korean J. Genet. 14, 303– japonica. Sci. Rep. Lab. Amphibian Biol. Hiroshima Univ. 314. 5, 1–46. Maeda, N. and Matsui, M. (1989) Frogs and toads of Japan. Sumida, M. (1994) Abnormalities of meioses in male reciprocal Bun-ichi Sogo Shuppan, Tokyo, Japan. (in Japanese). hybrids between the Hiroshima and Ichinoseki populations of Minato, M., Gorai, M., and Hunahashi, M. (1965) Geologic devel- Rana japonica. Experientia 50, 860–866. opment of the Japanese Islands. Tsukiji Shokan, Tokyo, Ja- Sumida, M. (1996) Incipient intraspecific isolating mechanisms in pan. (in Japanese). the Japanese brown frog Rana japonica. J. Herpetol. 30, Moritz, C. (1994) Applications of mitochondrial DNA analysis on 333–346. conservation: A critical review. Mol. Ecol. 3, 401–411. Sumida, M. (1997) Inheritance of mitochondrial DNAs and Moritz, C. and Brown, W. M. (1986) Tandem duplication of D-loop allozymes in the female hybrid lineages of two Japanese pond and ribosomal RNA sequences in lizard mitochondrial DNA. frog species. Zool. Sci. 14, 277–286. Science 233, 1425–1427. Sumida, M. and Ishihara, T. (1997) Natural hybridization and in- Moritz, C. and Brown, W. M. (1987) Tandem duplication in ani- trogression between Rana nigromaculata and Rana porosa mal mitochondrial DNAs: Variation in incidence and gene porosa in central Japan. Amphib.-Reptilia 18, 121–129. content among lizards. Proc. Natl. Acad. Sci. USA 84, 7183– Sumida, M. and Nishioka, M. (1994a) Genetic differentiation of 7187. the Japanese brown frog, Rana japonica, elucidated by elec- Moritz, C., Schneider, C. J., and Wake, D. B. (1992) Evolutionary trophoretic analyses of enzymes and blood proteins. Sci. relationships within the Ensatina eschscholtzii complex con- Rep. Lab. Amphibian Biol. Hiroshima Univ. 13, 137–171. firm the ring species interpretation. Syst. Biol. 41, 273–291. Sumida, M. and Nishioka, M. (1994b) Geographic variability of Nei, M. (1975) Molecular Population Genetics and Evolution. sex-linked loci in the Japanese brown frog, Rana japonica. North-Holland Publ. Co., Amsterdam. Sci. Rep. Lab. Amphibian Biol. Hiroshima Univ. 13, 173–195. Nei, M. (1987) Molecular evolutionary genetics. Columbia Univ. Szymura, J. M., Spolsky, C., and Uzzell, T. (1985) Concordant Press, New York. change in mitochondrial and nuclear genes in a hybrid zone Nei, M. and Tajima, F. (1981) DNA polymorphism detectable by between two frog species (genus Bombina). Experientia 41, restriction endonucleases. Genetics 97, 146–163. 1469–1470. Okada, Y. (1931) The tailless batrachians of Japanese Empire. Tanaka, T., Matsui, M., and Takenaka, O. (1994) Estimation of Imp. Agric. Exp. Station, Nishigahara, Tokyo. phylogenetic relationships among Japanese brown frogs from Okada, Y. (1966) Fauna Japonica, anura (amphibia). Biogeogr. mitochondrial cytochrome b gene (Amphibia: Anura). Zool. Soc. Japan, Tokyo. Sci. 11, 753–757. Powell, J. R. (1983) Interspecific cytoplasmic gene flow in the ab- Vawter, L. and Brown, W. M. (1986) Nuclear and mitochondrial sence of nuclear gene flow: Evidence from Drosophila. Proc. DNA comparisons reveal extreme rate variation in the mo- Natl. Acad. Sci. USA 80, 492–495. lecular clock. Science 234, 194–196. Rand, D. M. and Harrison, R. G. (1989) Ecological genetics of a Wallis, G. P. and Arntzen, J. W. (1987) Mitochondrial-DNA in the mosaic hybrid zone: Mitochondrial, nuclear, and reproductive crested newt superspecies: Limited cytoplasmic gene flow differentiation of crickets by soil type. Evolution 43, 432– among species. Evolution 43, 88–104. 449. Wilkinson, J. A., Matsui, M., and Terachi, T. (1996) Geographic Reilly, J. G. and Thomas, C. A. (1980) Length polymorphisms, re- variation in a Japanese tree frog (Rhacophorus arboreus) re- striction site variation, and maternal inheritance of mito- vealed by PCR-aided restriction site analysis of mtDNA. J. chondrial DNA of Drosophila melanogaster. Plasmid 3, Herpetol. 30, 418–423. 109–115. Yonekawa, H., Moriwaki, K., Gotoh, O., Watanabe, J., Hayashi, Saitou, N. and Nei, M. (1987) The neighbor-joining method: A new J.-I., Miyashita, N., Petras, M. L., and Tagashira, Y. (1980) method for reconstructing phylogenetic trees. Mol. Biol. Relationship between laboratory mice and the subspecies Evol. 4, 406–425. Mus musculus domesticus based on restriction endonuclease Shaffer, H. B. and Mcknight, M. L. (1996) The polytypic species cleavage patterns of mitochondrial DNA. Jpn. J. Genet. 55, 289–296.