Introduction

Corbicula leana clams have been regarded as simultaneous hermaphrodites and can reproduce by self-fertilization. Hermaphroditic Corbicula clams reproduce by androgenesis (Komaru et al. 1998). At metaphase in the oocytes, all the meiotic chromosomes become extruded into two polar bodies, and a male pronucleus is transformed into the zygote nucleus. Consequently, only the nucleus of spermatozoa is transmitted to the progeny in androgenetic clams. C. leana specimens found in Japan are hermaphroditic and triploids (Komaru et al. 1997; Ishibashi and Komaru 2003). We proposed a hypothesis of ploidy elevation from diploid to triploid by formation of a female pronucleus in these clams due to the aberrant meiosis of eggs (Komaru et al. 2006). The hermaphrodite clam Corbicula fluminea recently introduced into Japan includes diploids and triploids (Ishibashi and Komaru 2003).

We found that 78.2 % of C. leana specimens found in Shiga Prefecture, Japan, were diploid males and 21.8 % were triploid hermaphrodites (Houki et al. 2011). The males produced nonreductional and biflagellate diploid spermatozoa. They produce more spermatozoa than the hermaphrodites. Analysis of mitochondrial cytochrome b revealed that 28 of 35 males had the H1 haplotype (DNA Data Base of Japan (DDBJ) accession no. AB551543) and seven had H2 (AB551535). Of ten hermaphrodites, nine had H2 and only one had H3 (AB598630). The similarity of mitochondrial haplotypes suggested that the males have been derived from hermaphrodites in C. leana (Houki et al. 2011).

In androgenetic C. leana, the oocytes and spermatozoa released from a single hermaphrodite can self-fertilize. Oocytes from hermaphrodites might be also fertilized by spermatozoa from males occasionally, leading to the replacement of nuclear DNA, as suggested by Houki et al. (2011). In this case, nuclear gene flow should continue between males and hermaphrodites if the progeny develop into males or hermaphrodites. On the other hand, male lineages would be genetically isolated from hermaphrodites if the progeny develops into only males. This question can be answered by testing their identity in nuclear genes.

In the present study, we sequenced partial nuclear 28S rDNA containing the domain that is divergent between other eukaryotes (Larson 1991) and mollusks (Park and Foighil 2000). A haplotype network for hermaphroditic and male C. leana and related dioecious (bisexual) species was constructed to elucidate the origin of the male C. leana genome.

Materials and methods

Specimens of C. leana were collected from an irrigation ditch flowing out from the Yasu River in Ritto, Shiga Prefecture, Japan, in July 2009. Six males (diploids) and six hermaphrodites (triploids) were examined. Six individuals of the dioecious Corbicula sandai were collected in August 2008 from Lake Biwa, Hikone, Shiga Prefecture, Japan, and seven dioecious Corbicula japonica specimens were collected at Ano River, Tsu, Mie Prefecture, Japan, in July 2009. These samples have been used previously for mtDNA sequence and ploidy analysis, and for histological observations of the gonads (Houki et al. 2011).

Extraction of template DNA from tissue

The foot muscle was dissected using scissors and put into test tubes with 700 μl buffer (10 mM Tris–HCl, pH 7.5; 125 mM NaCl; 10 mM EDTA; 0.5 % sodium dodecyl sulfate; 4 M urea) containing 20 μl proteinase K. Samples were processed for DNA extraction according to the procedure of Lansman et al. (1981).

Polymerase chain reaction amplification of nuclear 28S ribosomal genes

The primer set D23F (5′-GAGAGTTCAAGAGTACGTG-3′) and D4RB (5′-TGTTAGACTCCTTGGTCCGTGT-3′) was used to amplify the nuclear 28S rDNA region (Park and Foighil 2000) for constructing the Corbicula phylogenic tree. The expansion region (also called the divergent domain) of this sequence is polymorphic. Polymerase chain reaction (PCR) was conducted according to the manufacturer's protocol for Ex Taq polymerase (Takara Bio, Shiga, Japan). The PCR conditions for Ex Taq amplification were as follows: following initial denaturation for 2 min at 94°C, denaturation was performed at 94°C for 30 min, annealing at 54°C for 45 s, and extension at 72°C for 2 min (35 cycles), followed by a final extension for 5 min at 72°C.

Cloning and sequencing

Heterogeneity in the nucleotide sequences of the nuclear 28S rDNA PCR products from the same clams meant that the sequences could not be determined directly; therefore, all material was cloned using standard methods. Nucleotide sequences were analyzed using a 3100 Genetic Analyzer (Applied Biosystems, Tokyo, Japan). Sequence alignments were performed using DANASYS version 3.6.1 (Hitachi Software, Tokyo, Japan). The haplotype network was constructed using TCS 1.21 (http://darwin.uvigo.es/software/tcs.html). Molecular diversity indexes were estimated using Arlequin version 3.1.1 (http://cmpg.unibe.ch/software/arlequin3/). Statistical analyses of molecular diversity index were performed with R software, version 1.12.1 (http://www.r-project.org/). The Kruskal–Wallis test was used for comparing more than two groups. If the difference in means was significant, the Bonferroni/Dunn procedure was used as a post hoc test. Differences were considered significant when P < 0.05.

Results and discussion

Haplotypes from the nuclear 28S rDNA partial sequence

As shown in Table 1, analysis of the partial sequences of the 28S rRNA gene (which is 441–444 bp long) revealed 69 different haplotypes in the species studied: from C. sandai (S1–S21; DDBJ accession nos. AB661640–AB661660) hermaphrodite C. leana (H1–H17; AB661719–AB661735), male C. leana (M1–M18; AB661737–AB661754), and C. japonica (Y1–Y13; AB661755–AB661767). The sequences of haplotype S3 in C. sandai and H2 in C. leana were the same (S3 = H2 in Table 1). This was the predominant haplotype in both species. The other haplotypes in C. sandai, C. leana, and C. japonica were species specific. The haplotypes differed between hermaphroditic and male C. leana. In C. sandai, 21 haplotypes (S1–S21) were detected from six clams (Table 2), with S1 and S2 accounting for 52 %. The six hermaphrodite and six male C. leana showed 17 and 18 haplotypes, respectively (Tables 3, 4), with H2 and M1 accounting for 70 and 63 % in hermaphrodites and male clams, respectively. A total of 13 haplotypes were detected from seven C. japonica clams, with the predominant haplotypes Y1 and Y2 accounting for 79 % (Table 5). These predominant and less frequent haplotypes were found from a single clam in each species. The 28S rDNA region sequence in Corbicula species was polymorphic, showing intra-individual sequence variation.

Table 1 Haplotype of 28S rDNA partial sequence (441–444 bp) in Corbicula species
Full size table
Table 2 Frequency of haplotypes of 28 s rDNA in C. sandai
Full size table
Table 3 Frequency of haplotypes of 28S rDNA in hermaphrodite C. leana
Full size table
Table 4 Frequency of haplotypes of 28S rDNA in male C. leana
Full size table
Table 5 Frequency of haplotypes of 28 s rDNA in C. japonica
Full size table

Distinct haplotypes of males from those of hermaphrodites

Figure 1 shows the haplotype network of Corbicula species based on the 28S rDNA sequences. The predominant haplotype S3 (=H2) was shared by hermaphrodite C. leana and C. sandai. In contrast, hermaphroditic (H1–17) and male (M1–18) C. leana did not share any haplotypes. Males had relatively heterogeneous haplotypes, such as M15–18, which differed from the predominant male haplotype M1 by at least 13 nucleotide substitutions.

Fig. 1
figure 1

Haplotype network of partial 28S rDNA sequences in dioecious C. sandai, C. japonica, androgenetic hermaphrodite C. leana and male C. leana. Large circles indicate haplotypes that were detected at a frequency of more than 20 % in each species

Full size image

Table 6 shows the genetic diversity indexes of 28S rDNA in Corbicula species. The number of haplotypes per individual did not differ obviously between the species (Table 6). Male C. leana showed higher nucleotide diversity (π) and higher mean numbers of substitutions and indels than did the hermaphroditic C. leana or dioecious C. japonica and C. sandai species. In C. japonica, C. sandai, and hermaphrodite C. leana, the mean number of transitions was greater than the mean number of transversions, while the reverse was true in male C. leana.

Table 6 Genetic diversity indexes of 28S rDNA in C. japonica, C. sandai, and C. leana
Full size table

The mutation rates in the male germ line in mammals have been shown to be higher than in the female germ line; molecular evolution has also been suggested to be male driven in other animals, such as birds (see review by Li et al. 2002). Therefore, an interpretation of these results is that the nuclear genome in male germ cells of C. leana is subject to a high frequency of mitosis, and this might result in a high mutation rate because of the peculiar mode of reproduction involved in androgenetic clams, especially in the male.

Our previous study on mtDNA showed that the haplotypes of dioecious C. sandai and androgenetic C. leana were clearly distinct and appeared in different clades in a neighbor-joining tree with a 100 % bootstrap value (Houki et al. 2011). The branching patterns of the three domestic species, C. sandai, C. leana, and C. japonica, were in accordance with those shown in previous studies based on mtDNA (Lee et al. 2005; Yamada et al. 2010). Dioecious C. sandai and androgenetic C. leana are closely related and form clearly different clades in the mtDNA tree. In contrast, in the haplotype network based on nuclear 28S rDNA, C. sandai and C. leana shared the haplotype S3. From the common haplotype S3 (=H2), the species-specific predominant haplotypes S1 and H1were derived by one base substitution. These results suggest that both species might have had a very recent common ancestor, suggesting that the transition from dioecy to hermaphroditism and androgenesis could be a recent evolutionary event in Corbicula. The slowly evolving nuclear genes remain in common in both species, although analysis of mtDNA places them in distinct clades because of its rapid rate of molecular evolution.

28S rDNA and mtDNA divergence

Table 7 shows the pairwise sequence divergence among the 28S rDNA haplotypes in the present study and that reported for mtDNA cytochrome b (Houki et al. 2011). The mean pairwise 28S rDNA sequence divergences among C. japonica, C. sandai, and hermaphrodite C. leana were smaller than those for mtDNA. In contrast, the 28S rDNA sequence divergence between hermaphroditic and male C. leana was 1.54 %, even though the mtDNA haplotype was almost identical.

Table 7 The mean pairwise sequence divergence (in percent) of mitochondiral cytochrome b haplotypes (621 bp), the nuclear 28S rDNA haplotypes (447 bp), and ratio of mtDNA/nuclear 28S rDNA among dioecious and androgenetic Corbicula clams
Full size table

In most animal species, the substitution rates of nuclear genes are much higher than those of mtDNA (Brown 1981). For mammals, estimates of 0.1–0.16 % divergence/site/million years have been made for the divergent domain of nuclear 28S rDNA, while the rate was lower among salamander species, at approximately 0.02 % divergence/site/million years (Larson 1991). In contrast, mtDNA changes very rapidly in vertebrates compared with single nuclear genes, and the estimated substitution rate is 0.5–2 %/genome/million years (Brown et al. 1979). In the present study, the nuclear 28S divergent domain showed relatively low substitution rates (1.10–2.68 %) between the haplotypes of C. japonica, C. sandai, and hermaphrodite C. leana compared with those of mitochondrial cytochrome b.

The similarities in mtDNA haplotypes found in our previous study (Houki et al. 2011) suggested that male C. leana was derived from hermaphrodites. Unexpectedly, the present data on nuclear 28S rDNA genes showed that the haplotypes of male and hermaphrodite clams were clearly distinct. No haplotype was common between hermaphroditic and male C. leana.

Our results support a hypothesis of Houki et al. (2011) that spermatozoa from males fertilize oocytes from hermaphrodites, and the male nuclear genome replaces the oocyte genome. MtDNA in egg cytoplasm remains in the progeny produced by this cross breeding between a male and a hermaphrodite. The nucleotide divergence of 28S rDNA haplotypes between hermaphrodites and males was 1.54 %, while that of mtDNA (cytochrome b) was only 0.16 % (Table 7). This divergence of nuclear 28S rDNA supports a hypothesis that cross breeding between males and hermaphrodites generates only males and so no gene flow between them. Males are likely to be a genetically distinct lineage and recently invaded into the triploid hermaphrodite population.