In order to elucidate the nucleotide sequence divergence of mitochondrial DNAs between two Japanese pond frog species Rana nigromaculata and R. brevipoda, and the mode of inheritance of cytoplasmic genomes in the female hybrid lineages of the two species, the cleavage patterns of mtDNAs digested with 10 restriction endonucleases were examined by agarose gel electrophoresis using a total of 52 frogs including Rana nigromaculata and R. brevipoda, their reciprocal hybrids and the backcross offspring (B1 and B2) derived from female hybrids by crossing with the paternal species. The cleavage patterns for mtDNA of Rana nigromaculata were different from those of R. brevipoda digested with all the restriction endonucleases used except EcoRV. The nucleotide sequence divergence of mtDNAs between these two species was roughly estimated to be 8.5%. The cleavage patterns for mtDNAs of the reciprocal hybrids and the B1 and B2 offspring were clearly similar to those of the maternal species, and paternal mtDNAs could not be detected. On the other hand, the proportions of original maternal nuclear genes at the 22 allozyme loci were 50% in the reciprocal hybrids, 21.9% or 25.6% in the B1 offspring, and 7.5% in the B2 offspring. These results demonstrate that nuclear genomes decrease the original maternal constitution in the female hybrid lineages generationally, whereas the mtDNAs are inherited maternally during repeated backcrossing.
INTRODUCTION
The two Japanese pond frog species, Rana nigromaculata and R. brevipoda (= R. porosa brevipoda), are nearly completely isolated by male hybrid sterility (Moriya, 1951, 1960; Kawamura, 1962; Kawamura and Nishioka, 1977). By contrast, the females of reciprocal hybrids between these two species are fertile to a large extent. Kawamura and Nishioka (1978) examined the change of reproductive capacity in the descendants of their reciprocal hybrids, and clarified that female hybrids were more or less inferior to their parents in reproductive capacity. There was scarcely any improvement in females of the B1, B2 or B3 generation on an average in reproductive capacity as compared with their female parents. In each of four lineages derived from the reciprocal female hybrids by backcrossing with the paternal and maternal species, males were distinctly inferior to their sisters in reproductive capacity.
Mitochondrial DNA (mtDNA) has been shown to be predominantly maternally inherited in Xenopus (Dawid and Blackler, 1972), Drosophila (Reilly and Thomas, 1980) and Heliothis (Lansman et al., 1983). Maternal inheritance of mtDNA has also been observed in horse-donkey hybrids (Hutchison et al., 1974), the rat Rattus norvegicus (Buzzo et al., 1978; Francisco et al., 1979; Hayashi et al., 1978; Kroon et al., 1978), the white-footed mouse Peromyscus polionotus (Avise et al., 1979), Mus (Gyllensten et al., 1985), human (Giles et al., 1980), Onchorhynchus (Ginatulina and Maksimovich, 1994), chicken-quail hybrids (Watanabe et al., 1985) and ducks (Lin et al., 1990). Strictly maternal inheritance of mtDNA has been demonstrated during extensive backcrossing in both directions between Mus domesticus and Mus spretus (Gyllensten et al., 1985) and between Heliothis virescens and H. subflexa (Lansman et al., 1983).
As to the nuclear genomes, the preferential expression of a maternal or paternal allele has been reported frequently in the somatic cells of hybrids (Castro-Sierra and Ohno, 1968; Hitzeroth et al., 1968; Whitt et al., 1972, 1973; Honjo and Reeder, 1973; Yamaguchi and Goldberg, 1974; Schmidtke et al., 1976; Etkin, 1977; Durica and Krider, 1977), but rarely in the germ cells of hybrids (Brown and Blackler, 1972; Vogel and Chen, 1976; Graf et al., 1977) (Elinson, 1981).
In the present study, the cleavage patterns of mtDNAs were examined using two Japanese pond frog species Rana nigromaculata and R. brevipoda, their reciprocal hybrids and the B1 and B2 offspring derived from fertile female hybrids backcrossed with the paternal species, in order to elucidate the nucleotide sequence divergence of mtDNAs between two species and the mode of inheritance of cytoplasmic genomes during repeated backcrossing of Japanese pond frogs. The enzymes and blood proteins encoded by nuclear genes were also examined to confirm the expression of parental genes and the chromosomal genome constitutions in the reciprocal hybrids and the backcross offspring.
MATERIALS AND METHODS
A total of 52 mature frogs was used in the present study (Table 1). These included eight Rana nigromaculata, eight R. brevipoda (= R. p. brevipoda), 24 reciprocal hybrids, and 12 B1 and B2 offspring derived from reciprocal female hybrids. All the crossing experiments were carried out by artificial fertilization method during the breeding seasons of 1990~1994. Ovulation was accelerated by injection of suspension of bullfrog pituitaries into the body cavity. Tadpoles were fed on boiled spinach. Metamorphosed frogs were fed on two-spotted crickets. The 14 enzymes extracted from livers and skeletal muscles and the three blood proteins of the above 52 frogs were analyzed by means of starch gel electrophoresis according to the method of Nishioka et al. (1980, 1992) with a slight modification, whereas the amylase of blood sera was examined by polyacrylamide gel electrophoresis (Table 2). Each enzyme was detected by the agar-overlay method outlined by Harris and Hopkinson (1976). The detection of blood proteins was made with the amido-black staining method.
Table 1
Number of specimens used in the present study

Table 2
Enzymes and blood proteins analyzed in the present study

Mitochondria were isolated from livers or ovaries. Tissues were homogenized in a decuple volume of STE buffer cooled with ice (0.25 M sucrose, 0.03 M Tris-HCl, 0.01 M EDTA, 0.11 M NaCl, pH 7.6), and centrifuged at 800 × g for 10 min at 2°C. The supernatant was centrifuged at 10,000 × g for 10 min at 2°C. MtDNAs were purified by CsCl ethidium bromide density gradient centrifugation described by Yonekawa et al. (1980) as follows.
Mitochondria obtained from each frog were lysed completely by suspending the pellet in 3.6 ml (final volume) of 0.6% sarcosyl, 10 mM EDTA and 10 mM Tris-HCl (pH 8.0). The lysate was then dialyzed against the above buffer for 4 hr at room temperature. The volume of lysate was adjusted to 3.6 ml by the same buffer, and 3.6 grams of solid CsCl and 0.24 ml of 4.6 mg/ml ethidium bromide were added to the lysate and mixed well. Then the lysate was centrifuged at 36,000 rpm (Hitachi RPS-40 rotor) for 40 hr at 20°C. The fraction containing closed circular mtDNA was collected, extracted three times with CsCl- saturated isopropanol to remove the dye and dialyzed against 0.1 mM EDTA (pH 8.0) for 20 hr with two changes of the medium. The mtDNA solutions were stored at −20°C until used.
Ten kinds of six-base recognizing restriction endonucleases, Bam HI, Eco RI, Eco RV, Hae II, HindIII, Hpa I, PstI, Pvu II, SacI and Xba I, were purchased from TaKaRa. The mtDNAs were digested by incubating at 37°C for 2 hr with appropriate amounts of the enzymes under conditions described by the suppliers. Agarose slab gel (1% Sigma agarose) electrophoresis was carried out in the standard TAE buffer (0.04 M Tris, 0.001 M EDTA, 0.02 M sodium acetate, pH 8.3). After electrophoresis, the gels were stained with 0.1 μg/ml ethidium bromide and photographed under ultraviolet light. λ-DNA digested with HindIII was used as molecular weight standards. Nucleotide sequence divergence was calculated by the method of Gotoh et al. (1979).
The abbreviations of N and B refer to a genome of Rana nigromaculata and a genome of R. brevipoda, respectively. The letters in parentheses indicate sources of cytoplasm. The first and second generation backcrosses are abbreviated to B1 and B2, respectively.
RESULTS
Mitochondrial DNAs
Figure 1 shows a comparison of the cleavage patterns for mtDNAs of two Japanese pond frog species, Rana nigromaculata and R. brevipoda, digested with eight restriction endonucleases, Bam HI, Eco RV, Hae II, HindIII, Hpa I, Pst I, Pvu II and Xba I. The cleavage patterns for mtDNA of R. nigromaculata were different from those of R. brevipoda digested with all the enzymes used except EcoRV. The molecular weight of each restriction fragment of these mtDNAs was estimated by comparing relative mobility to those of molecular weight standards (Table 3). The molecular weight of total mtDNA genome was estimated to be 19.1 ± 0.19 kb in Rana nigromaculata and 18.4 ± 0.18 kb in R. brevipoda. The numbers of common and different cleavage sites between R. nigromaculata and R. brevipoda were inferred from those of the fragments. The nucleotide sequence divergence between these two species was roughly estimated to be 8.5%.
Fig. 1
The cleavage patterns for mtDNAs of Rana nigromaculata and R. brevipoda digested with eight restriction endonucleases. M = λ-DNA digested with HindIII as molecular weight standards. The cleavage patterns for mtDNAs of R. nigromaculata were different from those of R. brevipoda digested with all the enzymes used except Eco RV.

Table 3
Restriction fragments of mtDNAs from Rana nigromaculata and R. brevipoda

Figures 2 and 3 show the cleavage patterns for mtDNAs of the reciprocal hybrids between these two species and the B1 and B2 offspring digested with three restriction endonucleases, BamHI, HindIII and PstI. The cleavage patterns for mtDNAs of 13 hybrids (N)NB between female Rana nigromaculata and male R. brevipoda, four B1 offspring (N)NB × BB obtained from matings between female hybrids and male R. brevipoda, and four B2 offspring (N)NB × BB × BB obtained from matings between female B1 offspring and male R. brevipoda were clearly similar to those of R. nigromaculata. Likewise the cleavage patterns for mtDNAs of 11 hybrids (B)BN between female R. brevipoda and male R. nigromaculata, and four B1 offspring (B)BN × NN obtained from matings between female hybrids and male R. nigromaculata were clearly similar to those of R. brevipoda. The mtDNA cleavage patterns observed in the hybrids and the B1 and B2 offspring are summarized in Table 4. The paternal mtDNA cleavage patterns could not be detected at all in these hybrids nor backcrosses. These results demonstrate that mtDNAs were primarily maternally inherited in these hybrids and backcrosses.
Fig. 2
The cleavage patterns for mtDNAs of the reciprocal hybrids between two species and the B1 and B2 offspring digested with Bam HI and HindIII. M = λ-DNA digested with HindIII as molecular weight standards. The cleavage patterns for mtDNAs of the reciprocal hybrids and the B1 and B2 offspring were clearly similar to those of the maternal parent.

Fig. 3
The cleavage patterns for mtDNAs of the reciprocal hybrids between two species and the B1 and B2 offspring digested with Pst I. M = λ- DNA digested with HindIII as molecular weight standards. The cleavage patterns for mtDNAs of the reciprocal hybrids and the B1 and B2 offspring were clearly similar to those of the maternal parent.

Table 4
MtDNA cleavage patterns of the reciprocal hybrids and the backcross offspring

Nuclear genome constitutions
Figures 4 and 5 show the electrophoretic patterns of enzymes and blood proteins in R. nigromaculata, R. brevipoda, their reciprocal hybrids and the B1 and B2 offspring. The electrophoretic patterns of two species were clearly different from each other at all the 22 diagnostic loci (Table 5). They were NN and BB in genotype, respectively. The electrophoretic patterns of each hybrid consisted of the sum of those of the two parental species at all the 22 loci. They were NB or BN in genotype. In the four B1 offspring (N)NB × BB, 7~10 of 20 loci showed hybrid pattern, NB in genotype, and the remaining 10~13 loci revealed the brevipoda pattern, BB in genotype (Table 5). In the four B1 offspring (B)BN × NN, 7~14 of 22 loci showed hybrid pattern, BN in genotype, and the other 8~15 loci revealed the nigromaculata pattern, NN in genotype (Table 5). Of the four B2 offspring (N)NB × BB × BB, one showed the brevipoda pattern, BB in genotype, at all the 20 loci examined. The remaining three showed the hybrid pattern, NB in genotype, at 3~5 loci and the brevipoda pattern, BB in genotype, at the other 15~17 loci (Table 5).
Fig. 4
The electrophoretic patterns of ADA, GPD, LDH and PEP-C in R. nigromaculata, R. brevipoda, the reciprocal hybrids and the B1 and B2 offspring. In the PEP-C patterns, the substrate L-leucyl-alanine also detected the PEP-A.

Fig. 5
The electrophoretic patterns of PEP-D, SOD, serum proteins and Hb in R. nigromaculata, R. brevipoda, the reciprocal hybrids and the B1 and B2 offspring.

Table 5
Genotypes at 22 loci of enzymes and blood proteins in the reciprocal hybrids and the backcross offspring

The proportions of original maternal genes were 50% in the reciprocal hybrids, 21.9% or 25.6% in the B1 offspring, and 7.5% in the B2 offspring (Table 5). Thus it was found that the nuclear genomes decreased the original maternal constitution in the female hybrid lineages generationally.
DISCUSSION
The present results demonstrate that the inheritance of mtDNAs is not governed by the same rules that apply to chromosome genes. As shown in Fig. 6, mtDNAs are maternally inherited; the mtDNAs of the original maternal parent are transmitted in the fertile female hybrid lineage during extensive backcrossing. On the other hand, both maternal and paternal nuclear genes are expressed in the reciprocal hybrids, and chromosomal genomes decrease the constitution of the original maternal parent in the fertile female hybrid lineage generationally (Fig. 6). Nishioka and Ohtani (1986) examined the lampbrush chromosome constitution of oocytes in 100 female backcrosses produced from female hybrids between female brevipoda and male nigromaculata by backcrossing with male nigromaculata. Of 1300 bivalents in total, 656 were BN in chromosome constitution and the other 644 were NN in chromosome constitution. These results showed that 656 (25.2%) of all the lampbrush chromosomes were derived from maternal species of the hybrids. Theoretically, in the course of extensive backcrossing of the fertile female hybrids with paternal species, the original maternal chromosomes or genes decrease to 50% in reciprocal hybrids, to 25.0% in the B1 offspring and to 12.5% in the B2 offspring. In the present study, the proportions of original maternal genes were 50.0% in reciprocal hybrids, 21.9% or 25.6% in the B1 offspring and 7.5% in the B2 offspring. These values almost coincide with the expected values stated above (χ2 = 0~2.29, P > 0.13).
Fig. 6
Scheme illustrating the inheritance of the nuclear and cytoplasmic genomes in the female hybrid lineage of two Japanese pond frog species. The chromosomal genomes decrease the original maternal constitution generationally, whereas the mitochondrial genomes are inherited maternally during repeated backcrossing. Circular, female; square, male; black, nigromaculata genome; dot, brevipoda genome.

Nishioka, Ohtani and Sumida (1980, 1987), Nishioka and Ohtani (1986), Nishioka and Sumida (1994a, b), Sumida and Nishioka (1994a, b) and Sumida (1996) reported several linkage groups of enzyme and blood protein loci, color mutant genes and sex-linked genes in the Rana nigromaculata group and the Japanese brown frog Rana japonica. According to these studies, 10 linkage groups were established in the Rana nigromaculata group; Alb, ADH-2 and albino gene b on chromosome No. 1; GPD, SOD-1, PEP-C, ME-2 and albino genes a and c on chromosome No. 2; MDH-1, ME-1, albino gene e and olive mutant gene on chromosome No. 3; LDH-B, PEP-B, HK, MPI, SORDH and ENO on chromosome No. 4; PEP-A on chromosome No. 5; Hb and IDH-1 on chromosome No. 6; blue mutant gene on chromosome No. 8; Prot-C, ALD and albino gene d on chromosome No. 9; PEP-D, EST-1, EST- 2, EST-4 and EST-5 on chromosome No. 10; ADA on chromosome No. 11. The present results gave definite evidence of linkages between the SOD-1 and PEP-C loci, between LDH-B and HK loci, between SORDH and ENO loci and between Hb and IDH-1 loci (Table 5), where the recombination rates were 16.7% (χ2 = 5.33, P < 0.03), 8.3% (χ2 = 8.33, P < 0.004), 8.3% and 16.7%, respectively.
The evidence presented here demonstrates that the mtDNAs were primarily maternally inherited during repeated backcrossing. The earliest evidence for maternal inheritance of mtDNA in animals came from the crossing experiments performed by Dawid and Blackler (1972) on two species of Xenopus. The studies of maternal inheritance using intra- and interspecific or intergeneric hybrids and repeated backcrosses also produced results concordant with the present ones (Avise et al., 1979; Buzzo et al., 1978; Francisco et al., 1979; Giles et al., 1980; Ginatulina and Maksimovich, 1994; Gyllensten et al., 1985; Hayashi et al., 1978; Kroon et al., 1978; Lansman et al., 1983; Lin et al., 1990; Reilly and Thomas, 1980; Watanabe et al., 1985).
According to Hutchison et al. (1974), two mechanisms for maternal inheritance are considered: (1) paternal mitochondria may be incapable of replication during development of the fertilized eggs, (2) maternal inheritance may result from a quantitative preponderance of maternal mitochondria in the zygotes. Gillham (1978) has raised the possibility that maternal inheritance in a single generation cross could be determined by interaction between nuclear and mitochondrial genes. It is believed that as little as 5% paternal mtDNAs will have been detectable in this experiment, however the presence of smaller amounts of paternal mtDNAs cannot be excluded. In Xenopus laevis, an egg contains more than 108 mtDNA molecules, whereas each sperm is estimated to contain about 100 molecules (Dawid, 1966; Dawid and Blackler, 1972). 10−6 of the mtDNA paternally derived would be below the limit of detection in this experiment. Nevertheless, some papers have shown paternal leakage of mtDNAs in the interspecific backcrosses of Drosophila (Kondo et al., 1990) or Mus (Gyllensten et al., 1991) using sufficiently sensitive techniques such as the Southern hybridization or the polymerase chain reaction. Kaneda et al. (1995) demonstrated that in intraspecific hybrids of Mus musculus, the paternal mtDNA genome got into the egg but was rapidly destroyed, whereas in interspecific hybrids between Mus musculus and M. spretus, the paternal mtDNA contribution persisted at least up until the neonate stage of development. It would be of interest to know whether leaky paternal inheritance of mtDNAs occurs during early development in the female hybrid lineages of the two Japanese pond frog species. These female hybrid lineages are considered to be very useful and attractive systems for examining the paternal mtDNA contribution during early development. Subsequent examination using polymerase chain reaction will clarify this point.
Acknowledgments
The author wishes to thank Dr. M. Nishioka for her helpful comments and advice, Dr. H. Yonekawa for his kind guidance regarding the method for mtDNA preparation, and Dr. O. Gotoh for supplying the program software for calculating the nucleotide sequence divergence. This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.