The larger data-set of COI sequences generated in this study confirms the previous results for female Perna perna, but interestingly, no genetic structure was found in the males. These sex-specific differences are unlikely to be an artifact of two mitochondrial genomes being subject to differential evolutionary constraints sensu 10], because this should have resulted in two copies of COI amplifying with the universal primers used here. Maternally- and paternally-inherited mtDNA genomes are usually highly distinct [10, 16, 17], but the fact that we did not even find genetic structure between males and females (which would be expected if there was a recent masculinisation of female-transmitted mtDNA) suggests that P. perna does not exhibit DUI (see also ) and so can serve as a model for other marine invertebrates in which the sexes are separate and mtDNA is inherited only in the female line.
Although more ITS-2 alleles were recovered for males than for females, no genetic structure was found with this genetic marker for either sex. Possible reasons for this include less genetic variation and the larger effective population size of the nuclear genome .
Our results have important implications for interpreting genetic structure and highlight the value of analysing genetic data from males and females separately. Surprisingly, in many studies on species with breeding behaviours that differ between the sexes, genetic data from males and females have been combined [2–4]. Gender-specific differences in genetic structure based on maternally-inherited mtDNA sequences have so far only been reported for social mammals and have been attributed to the sexes exhibiting different dispersal patterns [20–22]. As males cannot pass their mitochondrial genome to the next generation, their lack of genetic structure implies considerably greater contemporary gene flow than in the females.
Male-biased gene flow in P. perna (which lacks both social structure and external sexual dimorphism) could be explained by sex-specific differences in larval behaviour, including differences in the way this behaviour influences their position in the water column , and differences in larval development time. Although larval behaviour in P. perna has not been investigated, this seems unlikely as larvae of this species disperse as passive particles , and we are aware of no studies reporting sex-specific differences in the larval behaviour of mussels. Alternatively, as the magnitude of genetic structure depends on effective population sizes , it is possible that the differences between genders in mtDNA structure are due to lower female population sizes. Preliminary data indicate that the sex ratio in P. perna is male-biased (coast: 2.8 males : 1 female; n = 103; bay: 1.2 males : 1 female, n = 326), so the stronger male bias on the open coast, where wave action is stronger, could reflect the negative consequences of weakened attachment strength due to greater reproductive effort by the female mussels . We nonetheless consider it unlikely that the resulting reduction in female effective population size is sufficient to explain the observed genetic structure. The mtDNA diversity of males reflects that of females from the previous generation. Lack of structure suggests not only that there is a large amount of gene flow between bays and coastal habitats, but also that the pool of female larvae from the present generation is unlikely to have lower mtDNA diversity than that of the males. Instead, it is possible that in every generation, large numbers of females are eliminated because they are less likely to survive in habitats from which they did not originate, thus reinforcing genetic differentiation between habitats.
Females of bay populations tend to have a larger number of private haplotypes than those of coastal populations . Hence, even though there are no distinct habitat-linked mtDNA lineages of P. perna, bay individuals having certain haplotypes may be particularly vulnerable to strong wave action on the open coast. We hypothesise that females having these haplotypes expend more energy on reproduction rather than attachment, which results in an overall greater gamete output in mussels that reside in bays . Sex-specific genetic structure in the mtDNA of the mussel P. perna may therefore stem not from differential dispersal of the sexes, but from sex- and habitat-specific differences in reproductive effort and the effects of differential selection pressures between bays and the open coast.