Reproductive isolation between sympatric Anguilla japonica and Anguilla marmorata

The European eel Anguilla anguilla (L. 1758) and American eel Anguilla rostrata (LeSueur 1817) are two species of the genus Anguilla Schrank 1798 that spawn in the Sargasso Sea (Schmidt, 1922). The occurrence of natural hybrids (not only F1 but also admixed) between the A. anguilla and A. rostrata has been confirmed by numerous genetic analyses (Pujolar et al., 2014; Ulrik et al., 2014; Wielgoss et al., 2014; Jacobsen et al., 2016 and references therein). Natural hybrids have been observed in other species, such as Anguilla marmorata Quoi & Gairnard 1824 and Anguilla megastoma Kaup 1856 (Schabetsberger et al., 2015). Hybrids between Anguilla australis Richardson and A. anguilla, Anguilla japonica Temminck & Schlegel 1846 and A. anguilla, as well as those between Anguilla dieffenbachii Grey 1842 and A. australis have been produced under laboratory conditions (Lokman & Young, 2000; Okamura et al., 2004; Burgerhout et al., 2011), suggesting that reproductive isolation among the species of the genus Anguilla is not strict. Spawning areas of two Pacific species A. japonica and A. marmorata have been located at the west of the southern West Mariana Ridge in the North Equatorial Current (NEC) by finding fertilized eggs, pre-leptocephali and matured adult eels (Tsukamoto, 1992, 2006; Ishikawa et al., 2001; Miller et al., 2002; Chow et al., 2009; Kurogi et al., 2011; Tsukamoto et al., 2011). Extensive collections of Anguilla leptocephali performed in the NEC indicated that A. japonica and A. marmorata have similar spawning areas at west of the southern West Mariana Ridge (Kuroki et al., 2009). Furthermore, the sympatric reproduction of A. japonica and A. marmorata was corroborated by capturing matured adults of both species in the same tow (Chow et al., 2009; Tsukamoto et al., 2011), raising the question of whether these two species are reproductively well isolated (Kuroki et al., 2009). To address the above mentioned issue, a nuclear DNA marker was developed specific to A. japonica and A. marmorata. This nuclear marker together with mitochondrial (mt)DNA sequence analysis was applied to small leptocephali to silver-stage eels to check possible hybridization between A. japonica and A. marmorata.

Information on the glass, yellow and silver-stage individuals of A. japonica and A. marmorata used for the present study are represented in Table I and 10 areas, designated as A to J, where individuals were caught are shown in Fig. 1. Phenotypic differences in eels from the glass to adult-stages between these two species have been well described (Ege, 1939; Tabeta et al., 1976; Tzeng & Tabeta, 1983). These individuals were anaesthetised before fixation or before performing any surgical procedure. Eel leptocephali were collected by plankton surveys in the western North Pacific Ocean (Fig. 1 and Table II) from 2008 to 2010 and from 2013 to 2016 by Fisheries Agency of Japan R. V. Kaiyo-Maru and R. V. Shoyo-Maru. Anguilla spp. leptocephali were sorted out on board, their total length (L_T) was measured and the total myomere (N_M) numbers were counted under a binocular microscope. Subsequently, these leptocephali were frozen at −40 °C or preserved in 70% ethanol and transferred to the laboratory.

Crude DNA was extracted from muscle tissues using a DNA extraction kit (Genomic Prep Cells and Tissue DNA Isolation Kit, Amersham Bioscience; www.gelifesciences.com). The primer sequences to amplify the entire 5th intron of the androgen receptor gene (ar5) were 5′-TGTGAAGTCCAGGATGCTTT-3′ (ARex5F) and 5′-GGATGCAGTGTTCATACATG-3′ (ARex6R). To amplify the partial mitochondrial 16S rRNA gene (16 s), a pair of primers from Kurogi et al. (2012) or 16Sar-L and 16Sbr-H from Palumbi et al. (1991) were used. PCR amplification was performed in a 12 µl final volume sample containing 1 µl of template DNA (1–10 ng µl⁻¹), 1·2 µl of x10 reaction buffer, 1 mM of each deoxynucleotide triphosphate, 0·4 µM of each primer and 0·5 U of EX Taq polymerase (TaKaRa Bio, Inc.; www.takara-bio.com). The reaction mixtures were preheated to 94 °C for 4 min, followed by 35 amplification cycles [denaturation at 94 °C for 0·5 min, annealing at 55 °C (for 16s) or 58 °C (for ar5) for 0·5 min and extension at 72 °C for 1 min], with a final extension at 72 °C for 7 min. The PCR products of ar5 of standard A. japonica and A. marmorata specimens were cloned using a pGEM-T Easy Vector System I (Promega; www.promega.com) and sequenced using M13 and PCR primers.

The ar5 sequences determined for A. japonica and A. marmorata are available in the DNA database of Japan (DDBJ), European Molecular Biology Laboratory (EMBL) and GenBank database (accession numbers LC190928 and LC190929). The ar5 amplicons of both species consisted of a 28 bp partial exon 5 at the 5′ end and a 21 bp partial exon 6 at the 3′ end, whereas the intron 5 was 1295 bp for A. japonica and 1286 bp for A. marmorata.

These sequences were aligned using the CLUSTAL W algorithm implemented in MEGA 6 (Tamura et al., 2013) and was manually edited to minimize the number of gaps. Among nucleotide substitutions between species, those associated with palindrome sequences were investigated. Among several candidate restriction enzymes, Dpn II that recognizes the GATC palindrome was selected according to the number of restriction sites and deduced restriction profiles. Five of the six restriction sites of Dpn II were homologous between these species and the remaining sites may produce diagnostic restriction fragment length polymorphism (RFLP) between species (Fig. 2). According to the nucleotide sequence analysis, A. japonica was expected to have 592 and 503 bp fragments and A. marmorata was expected to have 565 and 505 bp fragments. The actual RFLP profiles of the ar5 amplicons of A. japonica and A. marmorata are shown in Fig. 3, which were in line with what was expected from nucleotide sequencing (Fig. 2). No RFLP was observed between individuals within species (51 A. japonica and 21 A. marmorata), The A. japonica-type and A. marmorata-type RFLP profiles were determined to be aa and bb homozygotes, respectively. The 16s PCR products of leptocephali were directly sequenced using PCR primers and basic local-alignment search tool (BLAST) homology search of the GenBank database was performed for the 16s sequences. Because Minegishi et al. (2009) reported 16s sequence differences among all known Anguilla spp. to be greater than 1·11%, the top BLAST match with greater than 99% homology was deemed to be an identified species.

The BLAST search for 16s sequences (413–598 bp) determined for 328 leptocephali indicated 194 specimens to be A. japonica, 128 to be A. marmorata, 5 to be A. bicolor pacifica (McClelland, 1844) and 1 to be A. luzonensis (Watanabe et al., 2009). All sequence data are available in the DDBJ–EMBL–GenBank database (accession numbers LC199026 to LC199244, LC213576 to LC213608, LC218744 to LC218816 and LC229308 to LC229310). Nucleotide sequence divergences [per cent Kimura 2 parameter (K2P) distance] within and between species are summarized in Table III. The mean percentages of K2P distance within species ranged from 0 to 0·22% and those between species ranged from 1·34 to 5·00%, in line with the results of Minegishi et al. (2009).

An N_M number of leptocephali larger than 10 mm was adopted, in which the N_M number ranged from 109 to 120 (114·0 ± 2·4 S.D.) in A. japonica (n = 164), 99 to 109 (103·5 ± 2·2) in A. marmorata (n = 126) and 104 to 107 (105·6 ± 1·1) in A. bicolor pacifica (n = 5). The N_M of A. luzonensis (n = 1) was 105. These results are in agreement with the N_M or vertebral counts of corresponding species reported previously (Ege, 1939; Matsui et al., 1968; Matsui, 1971; Miller et al., 2002; Watanabe et al., 2009). The results from mitochondrial 16s sequence and ar5 RFLP analyses were concordant; A. japonica leptocephali identified by 16s sequence analysis were aa homozygotes at the ar5 locus and all A. marmorata leptocephali were bb homozygotes. Moreover, the five A. bicolor pacifica and one A. luzonensis leptocephali had different RFLP profiles from those of A. japonica and A. marmorata (Fig. 4).

In contrast to the highly polymorphic microsatellite loci (Als et al., 2011; Wielgoss et al., 2014) and the large number of amplified fragment length polymorphism (AFLP) and single nucleotide polymorphism (SNP) markers (Albert et al., 2006; Pujolar et al., 2014; Ulrik et al., 2014; Jacobsen et al., 2016) used for addressing complicated hybridization issues in A. anguilla, the present nuclear DNA-RFLP marker is simple and is diagnostic for A. japonica and A. marmorata. No heterozygote was observed in the entire eel life cycle and in the entire distribution range, particularly with respect to A. japonica, indicating a strong reproductive isolation between A. japonica and A. marmorata.