| Edited by Toshihiko Shiroishi. Hitoshi Suzuki: Corresponding author. E-mail: htsuzuki@ees.hokudai.ac.jp |
Black rats (Rattus rattus sensu lato, also commonly known as ‘house rats’) nowadays occur in almost all regions of the world, largely as a consequence of commensalism and recent human migration and trade. In many places, their close association with people results in great human cost, whether measured through damage to field and stored crops, or through the impact of zoonotic diseases transmitted to people and livestock. Remarkably, given the undoubted economic and medical importance of the group (e.g. Lokugamage et al., 2004), black rats have received little attention from taxonomists and geneticists and there is no firm agreement as to exactly how many species and subspecies should be recognized among the ‘black’ or ‘house’ rats (Yosida, 1980; Baverstock et al., 1983; Musser and Carleton, 1993, 2005; Aplin et al., 2003, 2004) and only a partial understanding of the potential for hybridization and genetic introgression between different populations (Yosida, 1980).
Black rats display a high level of chromosomal diversity, as investigated most extensively by Yosida and coworkers (Yosida, 1980 and refs therein). To date, four major chromosomal groups of black rats have been identified: an ‘Asian’ group with 2n = 42, an ‘Oceanian’ group with 2n = 38, a ‘Ceylonese’ group with 2n = 40, and a ‘Mauritius’ type with a distinct 2n = 42. A recent study (Cavagna et al., 2002) using the chromosome painting technique suggested a higher level of karyotypic diversity than that detected by conventional G- and Q-banding studies (Yosida, 1980).
Fewer studies have investigated the patterns of genetic diversity within the R. rattus group. Moriwaki et al. (1975) and Yosida (summarized in Yosida, 1980) examined the variation in serum transferrin polymorphism in the R. rattus group by starch gel electrophoresis; they found two major ‘series’ of alleles: the R-series in ‘Asian’ rats and the C-series in ‘Oceanian’ rats–with marked geographic variation in allele frequencies within the former group. Brown and Simpson (1981) estimated the overall DNA sequence divergence between various populations of the R. rattus group as 0.4–9.6%, based on the fraction of conserved enzyme restriction sites. Their sampling included animals from Japan, Hong Kong, Sri Lanka and various localities in North America. The highest divergences (5.3–9.6%) were observed between the Sri Lankan population and all others; however, consistently high divergences (3.4–4.5%) were also observed between East Asian and North American populations. By contrast, the Norway rat, Rattus norvegicus, showed only limited divergence between populations, ranging from 0.4–1.8% across a similar geographic range. Baverstock et al. (1983) measured genetic differentiation among the karyotypic forms of black rat using the methods of isozyme electrophoresis and microcomplement fixation. They concluded that the ‘Oceanian’ 2n = 38 and ‘Asian’ 2n = 42 were sufficiently distinct to be treated as ‘incipient species’. Musser and Carleton (1993, 2005) cited the karyotypic and genetic evidence in support of their taxonomic separation of the R. rattus groups into two species: R. tanezumi for the ‘Asian’ house rats; and R. rattus for the ‘Oceanian’, ‘Ceylonese’ and ‘Mauritius’ populations.
Japanese black rat populations have been subject to more intensive sampling and analysis than those of any other region. Yosida and his co-workers found that most Japanese populations possessed 2n = 42 ‘Asian’ karyotypes, albeit with geographic variation in autosomal morphology (Yosida, 1980 and refs. therein). However, one population on Chichijima Island in the Ogasawara group was identified as showing evidence of hybrid interaction between the ‘Asian’ and ‘Oceanian’ types as indicated by their intermediate karyotypes and serum transferrin profiles (Yosida, 1980). More recently, a population of the 2n = 38 ‘Oceanian’ karyotypic group was identified at Otaru on Hokkaido (Suzuki et al., 2001). The ‘Asian’ black rat types probably invaded the Japanese island several thousands of years ago, most likely during Neolithic times (Kowalski and Hasegawa, 1976; Kawamura, 1989); in contrast, the Otaru ‘Oceanian’ population is believed to be a recent historic arrival (Suzuki et al., 2001).
We conducted a pilot survey of Japanese black rat populations using molecular sequencing methods. Our main objectives were: 1) to document the level of genetic divergence between Japanese populations of the 2n = 38 and 2n = 42 karyotypic groups; and 2) to investigate the nature, extent and consequences of interaction between these populations. We used the cytochrome b gene (cyt b; 1140 bp) as a mitochondrial marker and the gene encoding interphotoreceptor retinoid binding protein (IRBP; 1152 bp) as a nuclear marker. IRBP is a single-copy gene (Borst and Nickerson, 1988; Stanhope et al., 1996; Springer et al., 1997) that has proven useful for analyses at various phylogenetic levels, including analyses of higher-level relationships among mammals (Stanhope et al., 1996; Springer et al., 1997, 1999, 2001; Jansa and Voss, 2000), interrelationships among genera (Suzuki et al., 2000, 2004; Sato et al., 2004), and interrelationships among congeneric species (Serizawa et al., 2000) and even within species (Suzuki et al., 2004).
Samples were obtained from Hokkaido (Otaru), Honshu (Tokyo), and Kyushu (Kagoshima, Miyazaki and Shibushi) islands, and from two islands (Chichijima and Hahajima) in the Ogasawara group (Table 1, Fig. 1A). Individuals were trapped mostly around harbors and ports in order to maximize the chances of finding evidence of genetic interaction between the two previously documented lineages. Notes were made on the external appearance of the captured animals. Tissue samples (liver and kidney) were stored in 95% ethanol solution in the collection of the Laboratory of Ecology and Genetics, Graduate School of Environmental Earth Science, Hokkaido University. Chromosome preparations were made from bone marrow cells extracted from the femur and stained by conventional Giemsa staining, using methods described by Yosida (1980) and Suzuki et al. (2001).
 View Details | Table 1. List of samples of black rats from Japan used in this study |
 View Details | Fig. 1. Collection localities of black rats Rattus rattus from the Japanese Islands (A), and phylogenetic reconstruction based on cyt b (B) and nuclear IRBP (C) gene sequences. Locality numbers are as in Table 1, associated with the cyt b haplotypes; closed circles = ‘Oceanian’ haplotypes and open circles = ‘Asian’ haplotypes. The illustrated tree topology was obtained from analysis of the Kimura-2-parameter metric by the neighbor joining method. In all analyses, all codon positions and all substitutions were considered. Bootstrap values are expressed as percentages of 1000 replications for the neighbor joining, maximum parsimony and maximum likelihood analyses; these are represented as NJ/MP/ML. In the IRBP tree, haplotypes include those in the homozygotic condition (Homo-1, Homo-2, and Homo-3) and those reconstructed from heterozygous diplotypes (Hetero-1, Hetero-2, and Hetreo-3). Roman numerals represent the three major IRBP lineages in the black rat. Sequences of Rattus argentiventer (AB033701 for cyt b), Rattus norvegicus (AB033713 for cyt b; AB033714 for IRBP), Diplothrix legata (AB033696 for cyt b; AB033712 for IRBP) and Mus musculus (AB125774 for cyt b; AB125808 for IRBP) were obtained from the DDBJ/EMBL/GenBank DNA database. |
DNA extraction and the polymerase chain reaction and direct sequencing of the cyt b (1140 bp) and IRBP (1152 bp) genes were performed according to previously described methods (Suzuki et al., 2003). We sequenced both strands with a Dye Terminator Kit (ABI) and ABI3100 DNA autosequencer. The obtained sequences were deposited in the DDBJ/EMBL/GenBank DNA database with access numbers from AB211039 to AB211053. Subcloning of PCR fragments was performed according to previously described methods (Arakawa et al., 2002). IRBP fragments from a Kagoshima specimen (HS2433; Table 1) with 10 heterozygotic sites were cloned into pGEM-T vector (Promega). We examined multiple clones and obtained a ‘consensus’ sequence for each haplotype, neglecting clone-specific substitutions that were probably due to replication errors of the polymerase used (AmpliTaq polymerase; Perkin-Elmer) for the PCR reaction. Sequences of Rattus argentiventer (access numbers: AB033701 for cyt b), Rattus norvegicus (AB033713 for cyt b; AB033714 for IRBP), Diplothrix legata (AB033696 for cyt b; AB033712 for IRBP) and Mus musculus (AB125774 for cyt b; AB125808 for IRBP) were obtained from the DNA databases for use as outgroups in phylogenetic analyses.
Phylogenetic analyses were conducted using MEGA version 2.1 (Kumar et al., 2001) for the neighbor-joining method (NJ; Saitou and Nei, 1987) and PAUP* (4.0b8) (Swofford, 2001) for the maximum parsimony (MP; Swofford and Olsen, 1990) and maximum likelihood (ML; Felsenstein, 1981) methods. For NJ tree construction, the substitution model chosen was the Kimura-2 parameter model (Kimura, 1980). The appropriate model of nucleotide substitution for maximum-likelihood (ML) analysis was determined using the MODELTEST 3.06 program (Posada and Crandall, 1998). We used GTR+G and TrN+G models for the cyt b (-lnL = 3376.7378; A = 0.3087, C = 0.3012, G = 0.1246, T = 0.2655; gamma shape parameter = 0.1849) and IRBP (-lnL = 2235.2483; A = 0.2073, C = 0.2951, G =0.2972, T =0.2004; gamma shape parameter = 0.1149) data, respectively. For both markers, all substitutions were considered, since saturation does not affect phylogenies at this level. Bootstrap analysis was carried out with 1000, 1000, and 100 replications in the NJ, MP, and ML analyses, respectively.
A total of 37 black rats were karyotyped. All individuals from Honshu, Kyushu and the Ogasawara Islands were found to have typical ‘Asian’ karyotypes with 2n = 42 (Table 1). In contrast, all individuals from Otaru on Hokkaido had 2n = 38 ‘Oceanian’ karyotypes, consistent with previous results for this locality (Suzuki et al., 2001).
Specimens from Honshu and Kyushu had the typical appearance of R. rattus ‘tanezumi’, as described by Imaizumi (1960). Those from Otaru were larger, darker and longer-tailed, as described previously by Suzuki et al. (2001) for this population of ‘Oceanian’ black rats. Specimens from Chichijima Island were somewhat mixed in appearance, with some individuals having an intermediate appearance between ‘Asian’ and ‘Oceanian’ types.
We found five distinct cyt b haplotypes in the 37 individuals (haplotypes a-e; Table 1). Haplotype b was the most commonly encountered, being present in all individuals from Tokyo (n = 4), Shibushi (n = 6) and the Ogasawara Islands (n = 14), and in one of seven individuals from Kagoshima. The other haplotypes were each found at single localities: haplotype a at Otaru (n = 5); haplotype c at Miyazaki (n = 1); and haplotypes d and e at Kagoshima (n = 5 and 1, respectively). The populations from each of Honshu, Kyushu and Ogasawara Islands thus were not distinguished by their cyt b sequences. Phylogenetic analyses of the cyt b haplotypes by NJ, MP and ML methods (Fig. 1B) suggest a deep division between two haplotype lineages. One lineage contained haplotypes b and c, which differed from each other by a single nucleotide substitution; the other lineage contained haplotypes a, d and e, which differed from each other by one or two substitutions. As indicated in Fig. 1B, the two haplotype lineages crosscut the ‘Oceanian’ and ‘Asian’ groups as defined by karyotypes.
The two haplotype lineages had an average between-group sequence divergence of 3.8%, compared with only 0.1–0.2% divergence within each lineage. Pairwise differences between R. rattus as a whole and the outgroup taxa ranged from 10% (Rattus argentiventer) to 19% (Mus musculus).
DNA from 35 individuals was sequenced for the IRBP gene (Table 1). Across all sequences, 15 (1.3%) were variable. Homozygous sequences were recovered from 23 individuals and heterozygous sequences from 12 individuals, each of which possessed from 1 to 11 single nucleotide polymorphisms (SNPs; Table 1).
Three unique IRBP sequences (Homo-1, Homo-2, and Homo-3) were identified among the homozygous individuals. These differed from each other by seven to 11 nucleotide substitutions or by sequence divergences of 0.6–1.0%. Among the homozygous individuals, Homo-1 was represented exclusively in the Otaru sample, associated with ‘Oceanian’ 2n = 38 karyotypes and haplotype a for cyt b. In contrast, Homo-2 was widely represented among homozygous individuals with the ‘Asian’ 2n = 42 karyotype from Tokyo, Miyazaki, Shibushi, Kagoshima and from the Ogasawara Islands. Homo-1 and Homo-2 sequences differ by 7 nucleotide substitutions. Homo-3 was found in the homozygous state in a single individual from Shibushi (HS2395); this animal had the ‘Asian’ 2n = 42 karyotype and haplotype b for cyt b. Homo-3 differs from Homo-1 by 10 nucleotide substitutions and from Homo-2 by 11 nucleotide substitutions.
A total of seven unique diplotypes were identified among the heterozygotic sequences (Table 2). We employed a parsimony criterion to separate the diplotypes into component haplotypes, on the principle that hypotheses involving known haplotypes (i.e. those found in the homozygous condition) should take precedence over hypotheses that required completely novel haplotypes. Some of the diplotypes can be explained as straightforward combinations of Homo-1, Homo-2, and Homo-3. For example, diplotypes in individuals from Tokyo (HS2609) and Chichijima (HS3154, HS3156) can be explained as combinations of Homo-2 and Homo-3, while other individuals from Chichijima either had diplotypes combining Homo-1 and Homo-2 (HS3162, HS3169) or Homo-1 and Homo-3 (HS3159, HS3173). However, for five of the heterozygotic individuals it was necessary to postulate novel sequences to account for the diplotypes. Three such haplotypes were identified, each differing from one or other of the homozygous sequences by a single nucleotide substitution, and are herein designated as Hetero-1, Hetero-2, and Hetero-3 (Table 2). Specifically, Hetero-1 is linked to Homo-1; Hetero-2 to Homo-2; and Hetero-3 to Homo-3. Combinations identified under this method are shown in Table 1 and Table 2. For one Kagoshima individual (HS2433), the identity of the predicted diplotype was tested by subcloning analysis. Two types of clones were detected: as predicted, these conformed with Homo-2 and Hetero-3, respectively.
 View Details | Table 2. Polymorphic sites of IRBP in Japanese black rats, and haplotype analysis |
As noted above, the Homo-1 sequence was found in the homozygous condition only at Otaru. However, the Homo-1 and related Hetero-1 sequences were also found in heterozygous combinations (variously paired with Homo-2, Hetero-2 and Homo-3) in six of 11 individuals from Chichijima Island in the Ogasawara Islands (Table 1).
The Homo-2 and related Hetero-2 IRBP sequences were found in the homozygous condition at various places on Honshu and Kyushu, and on both of the Ogasawara Islands. These sequences were also found in the same localities in heterozygous combinations. On Kyushu, all heterozygous individuals possessed combinations of Homo-2 with either Homo-3 or the related Hetero-3 haplotype. In contrast, the sample from Chichijima Island in the Ogasawara group showed greater diplotype diversity (Table 1), with three individuals pairing Homo-2 with Homo-1, and one individual pairing Homo-2 with each of Homo-3, Hetero-1 and Hetero-2. Two other diplotype combinations found in the Chichijima Island sample involved Hetero-2 paired with Homo-1 and Homo-1 paired with Homo-3.
The rare haplotype Homo-3 was recovered in a homozygous state only in one individual (HS2395) from Shibushi, Kyushu. However, it was detected in heterozygous combinations in the Tokyo sample (one individual, paired with Homo-2) and on Chichijima Island (four individuals; two paired with Homo-2, two paired with Homo-1). As noted above, the related Hetero-3 sequence was found in two individuals from Kagoshima, Kyushu (HS2433 and HS2434), in both cases paired with Homo-2.
Phylogenetic analysis of the IRBP sequence data (Fig. 1C) clearly showed the association of the three ‘Homo’ and ‘Hetero’ haplotype pairs with moderate or high bootstrap values (62–100%), as well as the relatively deep separation of the three lineages (I, II, and III). Black rat IRBP lineages I and II formed a monophyletic cluster on the NJ MP, and ML trees, but with relatively low supporting bootstrap values of 53–76%. These trees generally did not feature a monophyletic clade including all of the black rat IRBP lineages, but rather a poorly resolved polytomy involving these lineages, other Rattus species and Diplothrix legata. This presumably reflects the overall low level of divergence of IRBP sequences among these taxa, rather than genuine polyphyly of the black rat group. As in the cyt b gene tree, the IRBP lineages crosscut the karyotypic groups.
The results of the present study indicate that the genetic composition of Japanese black rats may be more complex than previously understood, perhaps for several reasons. As mentioned earlier, previous studies of karyotypes and serum transferrins identified populations as belonging either to an ‘Asian’ group, characterised by a 2n = 42 karyotype and R-series serum transferrins, or an ‘Oceanian’ group with 2n = 38 and C-series serum transferrins. To date, ‘Oceanian’ black rats have been detected in two areas only—Otaru on Hokkaido (Suzuki et al., 2001; this study) and Chichijima Island in the Ogasawara group (Yosida, 1980), in the latter case based on probable hybrid individuals.
Each of the mitochondrial and nuclear genes examined displayed significant sequence diversity among the various Japanese black rat populations. In the case of cyt b, the five distinct haplotypes fall into two major lineages, with an average between-lineage divergence of 3.8%. In contrast, the six IRBP haplotypes sequences fall into three lineages, with between-lineage divergences of 0.6–1.0%. For both genes, the level of sequence divergence is similar to that usually observed between subspecies or sibling species of murine rodents (cyt b: Suzuki et al., 2004; IRBP: Serizawa et al., 2000; Suzuki et al., 2004).
Our findings are consistent with the notion that Japanese black rat populations have been derived from several different source areas (Yosida, 1980; Suzuki et al., 2001). However, the disparity in the number of lineages identified by each genetic marker begs explanation, as does the lack of clear association between the major lineages within each gene and the karyotypic groups. It is tempting to interpret the two major cyt b lineages as representing separate invasions of Japan by ‘Asian’ and ‘Oceanian’ black rat populations. This would be consistent with previous interpretations based on the karyotypic and serum transferrin characteristics (Yosida, 1980) and with the historical evidence for the recent arrival of the ‘Oceanian’ black rats at Otaru (Suzuki et al., 2001). It is further tempting to regard Homo-1 as an ‘Oceanian’ IRBP lineage, based on its restricted homozygotic occurrence at Otaru in association with the ‘Oceanian’ karyotype and cyt b haplotype a. Conversely, Homo-2 and Homo-3 might be identified as ‘Asian’ IRBP lineages, based on their geographically widespread association (within Japan) with the 2n = 42 karyotype and cyt b haplotypes b and c. However, Homo-3 is the most divergent of the three IRBP haplotype lineages, with 10–11 nucleotide substitutions separating it from the other two. In contrast, Homo-1 and Homo-2 differ by only seven substitutions. If the ‘Asian’ black rat group in Japan represents a unified historical lineage, it would follow that its IRBP gene preserves ancient lineages that have not been eliminated by recombination. Alternatively, it is conceivable that three distinct groups of black rats have contributed to the genetic diversity observed today in Japan, but with loss of lineage diversity in the mitochondrial gene. To resolve this issue clearly will require extensive sampling in potential source areas outside of Japan.
The discordance between the genetic variation and the karyotypic groups raises the interesting possibilities of hybridization and introgression in Japanese black rats. The critical evidence comes from two of the sampling areas: Kagoshima on Kyushu and Chichijima in the Ogasawara Islands.
F1 hybrids between the ‘Oceanian’ and ‘Asian’ black rats are readily obtained under laboratory conditions and have 40 chromosomes (Yosida, 1980). However, the yield of F2 offspring is typically poor, suggesting that the F1 animals are semi-sterile. Among five F2 animals reported by Yosida (1980: Table 20), three had 39 chromosomes, while one each had 40 and 42 chromosomes. Fertility appears to be higher in backcrosses between F1 hybrids and ‘Asian’ individuals; the majority (71%) of these have 41 chromosomes, with equal minorities having 40 and 42 chromosomes.
Although we did not detect any animals with obvious hybrid karyotypes among the Japanese black rats, the pattern of the distribution of the mitochondrial and nuclear gene haplotypes is strongly suggestive of a history of hybridization and introgression. In the case of the Kagoshima sample, the key evidence is the presence of two cyt b haplotypes (d and e) that both are closely related to the haplotype (a) found in the Otaru population. This might be construed as evidence for introgression of ‘Oceanian’ mitochondrial haplotypes into the Kagoshima population, albeit not from Otaru but from an as yet undiscovered source. Interestingly enough, no such genetic admixture is reflected in the Kagoshima population by gross alteration of the karyotype or by introgression of the putative ‘Oceanian’ Homo-1 marker for IRBP. In the Chichijima population, the evidence for hybrid interaction comes mainly from the IRBP sequence data, specifically, the presence of Homo-1 lineage haplotypes in six out of the 11 individuals analysed for this gene, although the variability in external appearance of this population also points to possible hybridity. In contrast, the karyotypes and mitochondrial cyt b of all Chichijima individuals were of the ‘Asian’ type.
In our view, the apparent admixture in some Japanese black rat populations of mitochondrial and nuclear lineages of likely ‘Asian’ and ‘Oceanian’ types is most likely a consequence of genetic introgression. Interestingly enough, this appears to have had variable genetic outcomes in different localities. At Kagoshima on Kyushu, the result appears to be introgression of one or more ‘Oceanian’ lineages (haplotypes d, e) of the mitochondrial cyt b gene, and the karyotype and external morphology have apparently stabilized back to the original ‘Asian’ type. In contrast, on Chichijima Island, introgression of the ‘Oceanian’ (Homo-1) lineage of the nuclear IRBP gene apparently occurred without any admixture of mitochondrial lineages. As in the Kagoshima population, the karyotype at Chichijima apparently has stabilized to the ‘Asian’ type but the external morphology has become more variable.
Few other cases of natural hybridization are known between karyomorphs of the Rattus rattus group. Yosida et al. (1971) reported natural hybrids between 2n = 38 and 2n = 42 karyomorphs on Eniwetok Island, as evidenced by individuals heterozygous for serum transferrins and with intermediate karyotypes resembling those obtained during laboratory crosses. Later, Yosida (1980) reported further apparent hybrids between these karyomorphs from Karachi in Pakistan, as well as probable hybrids between the 2n = 38 and the 2n = 40 ‘Ceylonese’ karyomorphs from Anuradhapura in the central lowlands of Sri Lanka. Lakhotia et al. (1973) and Yosida (1980) both reported a zone of possible natural sympatry between the 2n = 38 and 2n = 42 karyomorphs in southwestern India, with little if any evidence of hybridization between them. However, more recent investigations indicate that the animals with 2n = 42 from this region belong to a closely related but distinct species of Rattus, identified tentatively as R. satarae (Verneau et al., 1997; Musser and Carleton, 2005). Further karyotypic and genetic studies in areas of established and novel sympatry will be needed to determine the circumstances that variably inhibit or facilitate hybridization and effective gene flow among members of the Rattus rattus complex.
We thank Bisho Sasamoto and Hiroshi Shimamura for help with collecting samples. This study was supported in part by grants-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology, of Japan.
|
Index