To elucidate the phylogenetic relationships among four species belonging to the genus Petaurista (P. alborufus castaneus, P. alborufus lena, P. leucogenys leucogenys, P. leucogenys nikkonis, P. petaurista melanotus, and P. philippensis grandis), we investigated the partial sequences (1,068 bp) of the mitochondrial cytochrome b gene for these giant flying squirrels. Phylogenetic trees (NJ, MP, and ML trees) constructed from cytochrome b sequences indicated that P. leucogenys was grouped independently with other species, and that P. philippensis was most closely related to P. petaurista with 99–100% bootstrap values. In addition, two subspecies of P. alborufus did not form a single clade: P. alborufus castaneus from China was most distantly related to the other species, whereas P. alborufus lena from Taiwan was closely related to P. petaurista and P. philippensis with 82–90% bootstrap values. This result suggests that it is reasonable to regard P. alborufus lena as a distinct species from P. alborufus castaneus.
INTRODUCTION
Flying squirrels belonging to the genus Petaurista had been classified traditionally into five species: P. alborufus, P. elegans, P. leucogenys, P. magnificus, and P. petaurista (Corbet and Hill, 1980), each of which was intricately divided into various subspecies (Lekagul and McNeely, 1988). Recently, Corbet and Hill (1991, 1992) renewed the classification and recognized five additional species: P. caniceps, P. nobilis, P. philippensis, P. sybilla, and P. xanthotis, from five species classified previously (Corbet and Hill, 1980). Such classification disrupts phylogenetic study of the giant flying squirrels. Oshida et al. (1992) investigated the karyotaxonomy of Petaurista and concluded that P. petaurista melanotus was more closely related to P. alborufus lena than to P. petaurista grandis (P. philippensis grandis). In addition, Oshida et al. (1996) examined the mitochondrial 12S ribosomal RNA sequences on P. leucogenys from Japan, P. petaurista from Laos (P. petaurista melanotus), and P. petaurista from Taiwan (P. philippensis grandis), and reported that two subspecies of P. petaurista were closely related to each other and that P. leucogenys could have early diverged from P. petaurista.
Mitochondrial DNA (mtDNA) is a valuable molecule in investigating the phylogenetic relationships among populations, subspecies, and species. Features of mtDNA such as the maternal inheritance and rapid evolutionary rate advance the rapid geographic sorting of haplotypes in the absence of gene flow (Avise et al., 1984). Accordingly, using the information of mtDNA, we are able to infer the interspecific relationships, the intraspecific situations of population subdivision, and the genetic differentiation beyond the resolving ability of non-molecular approaches. In the present study, we determined partial sequences (1,068 base pairs: bp) of the mitochondrial cytochrome b gene for four species: P. alborufus, P. leucogenys, P. petaurista, and P. philippensis, and discuss the phylogenetic relationships among them as well as the taxonomic status of P. alborufus from continental China and Taiwan.
MATERIALS
Flying squirrels examined in the present study are shown in Table 1. Classification of species and subspecies followed the description of Corbet and Hill (1991, 1992), Imaizumi (1960), and Lekagul and McNeely (1988). Two samples of P. alborufus lena and one individual of P. philippensis grandis were captured in central Taiwan. Three individuals of P. alborufus castaneus imported from Hong-Kong to Japan in 1996 were commercially obtained. Muscle tissues of two individuals of P. leucogenys leucogenys were provided from Mr. Koichi Ikeda of the Fukuoka Prefecture Forest Research and Extension Center, Fukuoka, Japan, and Mr. Takehito Okayama of the Omogo Mountain Museum, Ehime, Japan. The other Petaurista specimens were commercially obtained. The outgroup taxon, Pteromys volans was provided from the Noboribetsu Bear Park, Hokkaido, Japan.
Table 1
Species of the genera Petaurista and Pteromys examined in this study
METHODS
From homogenated muscle tissues, genomic DNAs were extracted according to the phenol/proteinase K/sodium dodecyl sulfate method of Sambrook et al. (1989). The whole region of the mtDNA cytochrome b gene was amplified with polymerase chain reaction (PCR) using the following two primers: L14724 5′-GATATGAAAAACCATCGTTG-3′ and H15910 5′-GATTTTTGGTTTACAAGACCGAG-3′. Primer names correspond to the light (L) or heavy (H) strand and the 3′end-position of the primers in the human mtDNA sequence (Anderson et al., 1981). The former was made from the sequence described by Kocher et al. (1989) and the latter was newly designed from the conservative sequences between human (Anderson et al., 1981) and Rattus norvegicus (Gadaleta et al., 1989). The 50 μl of reaction mixture contained 100 ng of genomic DNA, 25 picomoles of each primer, 200 μM dNTPs, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, and 2.5 units of rTaq DNA polymerase (Takara). Amplification was carried out for 35 cycles using the following cycle program: 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min. The extension reaction was completed by incubation at 72°C for 10 min. PCR products were purified with the Qia-quick PCR purification kit (QIAGEN) and directly sequenced using an automated DNA sequencer (SQ5500, Hitachi).
Sequence alignment was carried out using GeneWorks (Intelligenetics). The phylogenetic trees were constructed using the neighbor-joining (NJ) method (Saitou and Nei, 1987) in Clustal W (Thompson et al., 1994), using the maximum parsimony (MP) method with the branch and bound search algorithm (Hendy and Penny, 1982) in PAUP (Swofford, 1993), and using the maximum likelihood (ML) method with DNAML in PHYLIP package program (Felsenstein, 1993). In NJ and ML methods, numbers of nucleotide substitutions per site were estimated for multiple substitutions using the Kimura's two-parameter method (Kimura, 1980). To assess the confidence of branching, the bootstrap analyses (Felsenstein, 1985) were performed with 1,000 replications in NJ and MP methods and at 100 replications in ML method.
Based on 0.5%/million years (Myr) reported as the transversional divergence rate at the third codon positions of mammalian cytochrome b gene (Irwin et al., 1991), divergence time was estimated.
RESULTS AND DISCUSSION
Phylogenetic relationships within the genus Petaurista
In this study, the partial sequences (1,068 bp) of cytochrome b gene were determined from 12 individuals of the genus Petaurista and one individual of the genus Pteromys as an out-group. Table 2 shows percentage sequence differences and numbers of transversions and transitions obtained from pairwise comparison. In addition, to estimate divergence between species, the transversional substitutions at the third codon positions were obtained by pairwise comparison (Table 3). Phylogenetic trees reconstructed using the NJ and the MP methods indicated essentially the same branching patterns (Figs. 1a and b): the first dichotomy isolated P. alborufus castaneus from the other flying squirrels and then P. leucogenys split from a clade formed by P. alborufus lena, P. philippensis, and P. petaurista. Consequently, the giant flying squirrels analyzed in the present study formed three groups: P. alborufus castaneus, P. leucogenys, and the other species or subspecies (82–90% bootstrap values). Although the branching order in the phylogenetic tree constructed with the ML method (Fig. 1c) was different from those of NJ and MP trees, three groups recognized in NJ and MP trees were obviously observed in ML tree (82-100% bootstrap values). However, the bootstrap values to support the branching orders of three groups were not high: 34% in NJ tree (Fig. 1a), 75% in MP tree (Fig. 1b), and 43% in ML tree (Fig. 1c). Owing to these low bootstrap values and the branching order differences between NJ and MP trees and ML tree, their phylogenetic relationships were not obvious in the present study.
Table 2
Pairwise comparisons of cytochrome b nucleotide sequences (1,068 bp) between 13 flying squirrel specimens
Table 3
Pairwise comparisons of the transversional substitution at the third codon positions of cytochrome b gene between 12 specimens of Petaurista
Fig. 1
Phylogenetic trees constructed by (a) the neighbor-joining (NJ), (b) the maximum parsimony (MP), and (c) the maximum likelihood (ML) methods based on the cytochrome b nucleotide sequences. Scale bars for the NJ and the ML trees represent branch length in terms of nucleotide substitutions per site. Numbers above branches indicate bootstrap values (%) derived from 1,000 replications for NJ and MP trees and 100 replications for ML tree.
Phylogeny of Petaurista alborufus
P. alborufus from southern China, Taiwan, Burma, and Thailand had been divided into seven subspecies: barroni (Kloss, 1916; Ellerman and Morrison-Scott, 1951), castaneus (Thomas, 1923), candidula (Wroughton, 1911; Ellerman and Morrison-Scott, 1951), lena (Thomas, 1907), leucocephalus (Hilzheimer, 1905), ochraspis (Thomas, 1923), and taylori (Thomas, 1914; Ellerman and Morrison-Scott, 1951). However, Corbet and Hill (1992) have recognized only four variations as subspecies of P. alborufus: castaneus, lena, leucocephalus, and ochraspis. In the present study, it is noteworthy that P. alborufus castaneus was distantly related to P. alborufus lena (Fig. 1). Based on pelage characteristics, P. alborufus lena was once treated as a distinct species P. pectoralis (Swinhoe, 1870). In addition, Corbet and Hill (1992) also suggested that P. alborufus lena is distinct enough to merit specific rank. Phylogenetic trees (Fig. 1) obtained in the present study did not conflict with the phylogenetic position of the form lena proposed by Swinhoe (1870) and Corbet and Hill (1992). Assuming that castaneus destributed in the mainland is a representative subspecies of P. alborufus, lena living only in Taiwan may be regarded to be distinct from P. alborufus. Moreover, P. alborufus lena was closely related to the clade of P. petaurista and P. philippensis with high bootstrap values (82% in NJ tree of Fig. 1a, 90% in MP tree of Fig. 1b, and 82% in ML tree of Fig. 1c). Based on the morphological characteristics such as externals and dental forms, Corbet and Hill (1992) regarded three subspecies (barroni, candidula, and taylori) of P. alborufus as P. petaurista. P. alborufus has often been confused with P. petaurista owing to the complicated morphological variation of these species. Based on the chromosomal characteristics, Oshida et al. (1992) have reported that P. alborufus lena was more closely related to P. petaurista melanotus than to P. petaurista grandis (P. philippensis grandis). The present molecular findings support their view.
Hsu (1990) reported that the Taiwan island rose from the sea floor on the Eurasian Continent approximately 4.0 Myr ago. Moreover, based on the faunistic and geological analyses, Kano (1940) and Liu and Ding (1984) concluded that the Taiwan island had been connected with the Eurasian Continent at least twice due to the glacial eustacy, initially during the Pliocene and subsequently during the Pleistocene. It is likely that the multiple faunistic exchanges between the Taiwan island and the Eurasian Continent had occurred through these connections. Lin and Lin (1983) explained the complicated zoogeography of Taiwanese mammals as follows: from a paleoenvironmental point of view, the first mammal group which immigrated from the Eurasian Continent to the Taiwan island during the glacial period of the Pliocene had adapted themselves to the cold environment. However, after the glacial period, to avoid the environment being warm, they had to move to the high elevation areas of Taiwan. Subsequently, the mammal group which immigrated to the island during the glacial periods of the Pleistocene had succeeded in expanding their ranges throughout the low elevation areas of Taiwan.
P. alborufus lena inhabits areas with elevation of 1,200 to 3,750 meters above sea level (Chang, 1985). In contrast, the distribution area of P. philippensis grandis living in Taiwan widely ranges from 700 to 2,600 meters above sea level (Chang, 1985). Accepting the hypothesis of Lin and Lin (1983), it is likely that the incursion of P. alborufus lena was earlier than that of P. philippensis grandis. In the present study, by using the available divergence rate estimated from the mammalian cytochrome b genes (Irwin et al., 1991) and our data (Table 3), the divergence between P. alborufus lena and P. petaurista melanotus and that between P. alborufus lena and P. philippensis grandis were estimated to have occurred approximately 6.2–6.8 Myr ago and 7.4–7.8 Myr ago, respectively. It seems resonable to suppose that, after the deviation from the lineages of P. petaurista and P. philippensis in the Eurasian Continent during the late Miocene, P. alborufus lena immigrated to Taiwan island and adapted itself to the alpine region in Taiwan. P. alborufus lena might have evolved independently from other Petaurista species due to the geographic isolation. Thus, our results support that this giant flying squirrel should be treated as a distinct species, as described originally by Swinhoe (1870) and Thomas (1907).
Phylogeny of Petaurista leucogenys
P. leucogenys is indigenous to the Japanese main islands except for Hokkaido (Corbet and Hill, 1991). Although Corbet and Hill (1980) described that P. leucogenys is distributed on the Japanese islands and central China, recently they changed the classification and newly treated the leucogenys population of central China as a distinct species P. xanthotis (Corbet and Hill, 1991, 1992). Based on the poor information about the pelage, Imaizumi (1960) classified P. leucogenys into three subspecies: leucogenys, nikkonis, and oreas, although this classification of subspecies has not been generelly accepted. In the present study, P. leucogenys clearly formed a single clade with high bootstrap values (100% in NJ, MP, and ML trees). Meanwhile, the divergence time estimated from our data (Table 3) was approximately 10.2–10.6 Myr ago between P. leucogenys and P. petaurista, 8.4–9.0 Myr ago between P. leucogenys and P. philippensis, 9.0–10.2 Myr ago between P. leucogenys and P. alborufus lena, and 9.0–9.6 Myr ago between P. leucogenys and P. alborufus castaneus. Although it is hard to determine which species is most closely related to P. leucogenys, this species could be an independent lineage in the genus Petaurista at least in the late Miocene. Based on the fossil records, Kawamura (1988, 1990) and Kawamura et al. (1989) showed that P. leucogenys immigrated from the Eurasian Continent to the Japanese islands through the land bridges around the early to the middle Pleistocene, and that this species had been isolated due to the separation of the Japanese islands from the Eurasian Continent in the Pleistocene. Accepting Kawamura's hypothesis, our results suggest that an ancestral stock of P. leucogenys had diverged from the other Petaurista species within the Eurasian Continent prior to its immigration to Japan.
Phylogeny of Petaurista petaurista and P. philippensis
P. petaurista melanotus and P. philippensis grandis formed a single clade with high bootstrap values (100% in NJ and MP trees and 99% in ML tree, Fig. 1). P. petaurista, which is one of the most dominant species in the genus Petaurista, is distributed throughout southern parts of the Eurasian Continent and Southeast Asia (Corbet and Hill, 1980; Lekagul and McNeely, 1988). On the other hand, P. philippensis had been treated as a subspecies of P. petaurista until Corbet and Hill (1991, 1992) established it as a distinct species. A Taiwanese form (P. philippensis grandis) examined here was previously considered as P. petaurista grandis by Swinhoe (1870). Based on sequence data of the 12S rRNA gene, Oshida et al. (1996) reported that the genetic distance between P. philippensis grandis (P. petaurista from Taiwan) and P. petaurista melanotus (P. petaurista from Laos) was almost parallel to intraspecific differences within P. leucogeneys. Cytochrome b data in the present study supported that, although the genetic distance between P. philippensis grandis and P. petaurista melanotus corresponded to approximately twice of intraspecific differences within P. leucogenys (Fig. 1a), P. philippensis is most closely related to P. petaurista.
Acknowledgments
We would like to thank Mr. L. S. Tzen, Mr. K. Ikeda (Fukuoka Prefecture Forest Research and Extension Center), Mr. T. Okayama (Omogo Mountain Museum), and Dr. K. Gouda and Dr. M. Satoh (Noboribetsu Bear Park) for supplying specimens. We thank Dr. T. Tanaka-Ueno (Kyoto University) for invaluable suggestion on the phylogenetic analysis. This study was partly supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture.
