Japanese Journal of Herpetology 17(3): 91-97., June 1998 (C)1998 by The HerpetologicalSociety of Japan Local Population Differentiation and Phylogenetic Relationships of Russian Brown Frog, Rana amurensis Inferred by Mitochon- drial Cytochrome b Gene Sequences (Amphibia, Ranidae) TOMOKO TANAKA-UENO, MASAFUMI MATSUI, TAKANORI SATO, SEN TAKENAKA AND OSAMU TAKENAKA Abstract: In order to assess local population differentiation and phylogenetic relation- ships of a Russian brown frog, Rana amurensis, the sequences of 587 base pairs of the mitochondrial cytochrome b genes are compared with seven species of Japanese and one species complex of Taiwanese brown frogs (R. pirica, R. ornativentris, R. dybowskii, R. japonica, R. okinavana, R. tagoi, R. tsushimensis, and the R. sauteri complex). Genetic differentiation between populations of R. amurensis from Sakhalin and the Maritime Territory was found to be minimal. The resultant phylogenetic tree indicates monophyly of brown frogs and earliest divergence of R. amurensis among all the brown frogs studied. For this reason, separation of R. amurensis from the R. japonica group is suggested, but separation of the R. sauteri complex as a distinct ge- nus or subgenus Pseudorana is not supported. Key words: Brown frog; Cytochrome b; Phylogeny; Pseudorana; Rana amurensis; Russia Rana amurensis Boulenger, 1886 occupies a ships of Japanese brown frogs have been report- very wide range, including western Siberia east ed by Nishioka et al. (1992) and Tanaka et al. to Sakhalin, northern and eastern Mongolia, (1996). In the present study, we first investigat- northeastern China and Korea, and regions ed the degree of genetic differentiation between north to bevond the Arctic Circle to 71°N populations of R, amurensis in Sakhalin and (Frost, 1985). In several previous studies, some continental Russia using the nucleotide sequence brown frog species were proven to be highly of mitochondrial DNA cytochrome b gene. differentiated genetically among conspecific Rana amurensis has 2n=26 chromosomes (Orlo- populations without showing marked morpho- va et al., 1977; Wu and Yin, 1983), but its logical variations (e. g., Sumida, 1996; Tanaka et phylogenetic relationships with other brown al., 1994, 1996, Tanaka-Ueno, Matsui, Sato et frogs are still poorly understood, and only two al., 1998). Because R. amurensis is noted for its studies based on the isozyme data mentioned immense distribution as mentioned above, the above are available (Nishioka et al., 1992; Green presence of considerable genetic differentiation and Borkin, 1993). In both of those studies, R. is expected in this species. However, the amurensis is inferred to be involved in very early populational genetic differentiation in R. divergence among the brown frogs with 2n=26 amurensis has been studied only from the elec- chromosomes. Thus, our second purpose was trophoretic analyses of allozymes and proteins to clarify the phylogenetic relationships of R. between the populations of the Amur Basin and amurensis with respect to seven species of brown Sakhalin Island (Green and Borkin, 1993) or frogs in Japan on the basis of sequence variation among the populations of north Mongolia, of cytochrome b. From our previous study northeastern China, and the Maritime Territory (Tanaka-Ueno, Matsui, Chen et al., 1998), the (Nishioka et al., 1992). The results reported by R. sauteri complex from Taiwan (Chou and Lin, these authors indicate absence of notable genetic 1997), which is occasionally assigned to a dis- differentiations among the populations studied. tinct genus (Fei et al., 1990) or subgenus Analyses of allozymes and DNA nucleotide (Dubois, 1992), has been suggested to form a sequence do not always give congruent results. sister group with some Japanese brown frogs. For example, discordant phylogenetic relation- Our third purpose was to examine how the relationships of these frogs change when R. Accepted 27 Dec. 1997 amurensis is added to them. 92 Jpn. J. Herpetol. 17 (3) 1998 FIG. 1. Aligned sequences of a 572-bp segment of the cytochrome b gene. Dots indicate identity with the sequence of Xen opus laevis. TANAKA-UENO ET AL.-RELATIONSHIPS OF BROWN FROGS 93 FIG. 2. A neighbor-joining tree rooted at the mid- point of the longest path (A), a maximum-likelihood tree (B), and a bootstrapped parsimony tree (C). Nodal values in A and C indicate percent support for branches in 1,000 bootstrap replicates. method (Saitou and Nei, 1987), using the NJ MATERIALS AND METHODS procedure included in the PHYLIP package DNA was extracted from small amounts (less (Felsenstein, 1993), the maximum-likelihood than 50mg) of frozen liver, muscle, heart, or method using the DNAML procedure of eggs of a total of 12 specimens of R. amurensis PHYLIP, and the parsimony method using the from Russia and referential R. dybowskii from heuristic search algorithm in PAUP (Swofford, Tsushima (see appendix). 1993) with sequences added in a random order The methods used to extract, amplify, and se- for 1,000 repetitions. quence are as described previously (Tanaka et For comparisons, the published sequences of al., 1994, 1996). A part of the mtDNA cyto- seven species of Japanese and one species com- chrome b gene was amplified using the poly- plex of Taiwanese brown frogs (R. pirica, R. or- merase chain reaction (PCR) and approximately nativentris, R, japonica, R. okinavana, R. tagoi, 600 bP of this gene were sequenced. Primers R. tsushimensis, and the R. sauteri complex), newly designated for the amplification and se- and an outgroup taxon, R. catesbeiana were quencing are: L14731 (5'-GAAAAACTATCG- used (Tanaka-Ueno, Matsui, Chen et al., 1998). TTGTTATTCAACTA-3'), L14850 (5'-TCTCA- The published sequence of Xenopus laevis TCCTGATGAAACTTTGGCTC-3'), H15502 (Dunon-Bluteau et al., 1985) was also incorpo- (5'-GGATTAGCTGGTGTGAAATTGTCTGGG rated into the analysis as that for another out- -3'), and H15584 (5'-CCTAGTTTATTAGGGA- group. When a taxon showed some variation in TGGAGCGGAG-3'). The sequence numbering the nucleotide sequence, the most frequently ap- system followed that of the human sequence pearing sequence was considered as the (Anderson et al., 1981). representative sequence for the taxon. Three different methods were employed to in- fer relationships among taxa on the basis of se- RESULTS quence data obtained-the neighbor-joining Nucleotide sequence data of 587 by were con- 94 Jpn. J. Herpetol. 17(3) 1998 stantly obtained for all samples (Fig. 1). Of the (72.3% and 89.6% iterations, respectively). R. amurensis from Sakhalin, two samples from Slepikovskogo had an identical sequence, and DISCUSSION they exhibited a high nucleotide sequence Sequence similarities found between samples similarity of 99.7% with the sample from Aniva. of R. amurensis from Lazo and Sakhalin (98.5- Similarly, two samples from continental Lazo 99.0%) are higher than the values reported for had a high similarity of 99.7%. Between samples populations of Japanese R. tagoi (93.2-96.8%: from Sakhalin and Lazo, similarities varied Tanaka et al., 1994), and interpopulation genetic from 98.5 to 99.0%. variation in this species is judged to be very low. Brown frogs showed a monophyletic relation- Similar low genetic differentiation between ship with high bootstrap iterations (99.2% in populations was reported from the analyses of neighbor-joining and 97.6% in maximum-par- allozymes and proteins by Nishioka et al.(1992). simony analyses: Fig. 2A, C). Maximum-likeli- According to their report, Nei's (1975) genetic hood analysis also resulted in identical grouping distances among populations of R. amurensis (Fig. 2B). from northern Mongolia, the Maritime Territory Samples of R. amurensis from the continent of Russia, and northeastern China range from and Sakhalin were monophyletic with bootstrap 0.005 to 0.008, and are smaller than interpopu- iterationsof 100%in both the neighbor-joining lation distances among the Japanese R, japonica and maximum-parsimony trees. They exhibited that they studied. Green and Borkin (1993) also the earliestdivergence both in neighbor-joining analyzed allozymes and showed low genetic and maximum-likelihood trees,but this diver- differentiation between populations of R. gence was supported only in very low iterations amurensis from the Amur Basin (Khabarovsk) of 38.5% in the maximum-parsimony analysis. and Sakhalin. From these previous results of The order of divergence among the remaining allozyme analyses and our DNA analyses, genet- brown frogs was the same as that reported by ic variation appears to be low at least among Tanaka-Ueno, Matsui, Chen et al. (1998). In populations of R. amurensis occurring from the neighbor-joining tree (Fig. 2A), trichoto- Mongolia and northeastern China to the Far mous relationships were observed; Rana tagoi East Asia. Therefore, R. amurensis currently and the R. sauteri complex from Sanyi formed found in this region is considered to have the first cluster with relatively low iterations diverged in a short time rather recently. (59.9%). The second cluster was supported by However, the distribution of R. amurensis is high iterations of 92.1% and consisted of R. very wide, and several morphologically distinct okinavana and the subcluster containing R. subspecies, such as R. amurensis coreana from tsushimensis and the R. sauteri complex from Korea (Shannon, 1956), have been recognized in Wulai. The sister relationship of the latter sub- some parts of its species range, but these forms cluster was completely supported (itera- have been poorly studied. tions=100%). In the third cluster which was The Strait of Mamiya (Tatal) separating Sak- supported by 88.4% iterations, R. japonica was halin and the Maritime Territory of the conti- separated from the three species with 2n=24 nent with a depth of 25m, is estimated to have chromosomes. The latter group, supported by been about -30±5m deep about 8,500 years B. iterations of 92.5%, consisted of R. ornativen- P.,before the separation of the two areas started tris and the group of R. pirica and R. dybowskii (Ohshima,1990).