Assessing pre- and post-zygotic barriers between North Atlantic eels (Anguilla anguilla and A. rostrata)

Abstract

Elucidating barriers to gene flow is important for understanding the dynamics of speciation. Here we investigate pre- and post-zygotic mechanisms acting between the two hybridizing species of Atlantic eels: Anguilla anguilla and A. rostrata. Temporally varying hybridization was examined by analyzing 85 species-diagnostic single-nucleotide polymorphisms (SNPs; F_ST ⩾0.95) in eel larvae sampled in the spawning region in the Sargasso Sea in 2007 (N=92) and 2014 (N=460). We further investigated whether genotypes at these SNPs were nonrandomly distributed in post-F1 hybrids, indicating selection. Finally, we sequenced the mitochondrial ATP6 and nuclear ATP5c1 genes in 19 hybrids, identified using SNP and restriction site associated DNA (RAD) sequencing data, to test a previously proposed hypothesis of cytonuclear incompatibility leading to adenosine triphosphate (ATP) synthase dysfunction and selection against hybrids. No F1 hybrids but only later backcrosses were observed in the Sargasso Sea in 2007 and 2014. This suggests that interbreeding between the two species only occurs in some years, possibly controlled by environmental conditions at the spawning grounds, or that interbreeding has diminished through time as a result of a declining number of spawners. Moreover, potential selection was found at the nuclear and the cytonuclear levels. Nonetheless, one glass eel individual showed a mismatch, involving an American ATP6 haplotype and European ATP5c1 alleles. This contradicted the presence of cytonuclear incompatibility but may be explained by that (1) cytonuclear incompatibility is incomplete, (2) selection acts at a later life stage or (3) other genes are important for protein function. In total, the study demonstrates the utility of genomic data when examining pre- and post-zyotic barriers in natural hybrids.

Figure 1

Map showing all localities of new and previously published data analyzed in this study. Black circles denote localities with new data; gray circles denote localities with both new and already published data and white squares show localities with only data already published. See Table 1 for information on exact sampling location and sample size.

Figure 2

Maps showing (a) sampling localities of the leptocephali larvae caught during the 2014 expedition and genotyped in this study, and (b) the frequencies of the two species at each of the localities sampled during the 2014 expedition. The two red circles denote the localities from which hybrids (bAR × AA) were detected in 2014 and red asterisks show the positions of two hybrids found among larvae sampled in 2007. The numbers in brackets denote sample sizes of analyzed larvae for each locality.

Figure 3

Expected and observed genotype frequencies at ATP5c1 of the examined hybrids, given the estimated hybrid classes. The estimated expected frequencies are based on assumed Hardy–Weinberg proportions and previously reported allele frequencies of the ATP5c1 gene in American eel (Gagnaire et al., 2012). The probabilities of the observed genotype combinations are shown to the right and the individual showing ATP6–ATP5c1 mismatch is denoted by an asterisk. Sample sizes (N) are shown in the respective histograms. AA, European eel (A. Anguilla); AR, American eel (A. rostrata).