infection
Endosymbionts such as Wolbachia, are selfish entities that have evolved to promote their own prevalence in their host populations through increasing the fitness of the infected individuals over their uninfected counterparts (O’Neill et al., 1997). The success of symbioses is however often context-dependent (Ferrari & Vavre, 2011; Hajek, Morris, & Hendry, 2019). While the Wolbachia strain w Hho occurs at an intermediate prevalence across the Åland Islands (Duplouy et al., 2015), on the islands of Seglinge-Kumlinge, it has been almost at fixation since the year 2000, and potentially earlier. As shown by Duplouy et al. (2015), this Wolbachia strain appears not to affect dispersal capacity of H. horticola , thus the highWolbachia prevalence in Seglinge-Kumlinge is unlikely to be due to differential dispersal of infected wasps to these islands. However,Wolbachia increases the susceptibility of the wasp H. horticola to hyperparasitism by the wasp Mesochorus cf.stigmaticus (van Nouhuys et al., 2016). The hyperparasitoid is common across the Åland mainland (Nair et al., 2016), and is also present in Sottunga and Föglö, but is absent from Seglinge-Kumlinge islands (van Nouhuys & Hanski, 2005). The hyperparasitoid wasp restrains the spread of w Hho in H. horticola by keeping the infection at lower prevalence when highly abundant (van Nouhuys et al., 2016), while the absence of M. cf. stigmaticus in Seglinge-Kumlinge (Nair et al., 2016; van Nouhuys & Hanski, 2005; van Nouhuys et al., 2016) releases the selection pressure onw Hho-infected H. horticola wasps, and allows the spread of the symbiont in this isolated wasp population (van Nouhuys et al., 2016). Although primarily vertically transmitted, Wolbachia has been suggested to also occasionally transfer horizontal between hosts, with parasitoids being one of the suggested ecological routes supporting such transfer (Duplouy, Pranter, Warren-Gash, Tropek, & Wahlberg, 2020; Vavre, Fleury, Lepetit, Fouillet, & Boulétreau, 1999).Mesochorus cf. stigmaticus wasps are however not known to carry Wolbachia , nor to vector the symbiont between hosts in the Åland population of their food web insect community (Duplouy. Pers. Obs.).
The spread of a maternally inherited symbiont, such as Wolbachia , leads to a simultaneous increase in prevalence of the host mitochondrial haplotypes associated with the symbiont (Charlat et al., 2009; Duplouy et al., 2010; Schuler et al., 2016). In H. horticola in Åland, C- and T-mitotypes associate with the w Hho infection. The C-mitotypes are however less associated with the infection, potentially because the trans-generation transmission of this Wolbachiastrain is less efficient in females carrying the C- over the T-mitotype (Duplouy et al., 2015). Consequently, we expected that the T-mitotype would be found at low frequency in Seglinge-Kumlinge, as it was in the original population of Finström (16%), as well as in Sottunga (29%). Instead, we found that the T-mitotype is prevalent in Seglinge-Kumlinge (72%). The spread of the wasps carrying the T-mitotype in Seglinge-Kumlinge could result from: (I) the selective sweep of the T-matriline during the spread of Wolbachia in Seglinge-Kumlinge in the absence of pressures from the hyperparasitoid, (II) strong bottlenecks randomly selecting for individuals from the T-matriline over the C-matriline in Seglinge-Kumlinge during migration events from Sottunga; and (III) uncharacterized fitness benefits associated with the T- mitotype. According to our data, Seglinge-Kumlinge is colonized by several genotypes also found in the rest of Åland, which suggests that several migration events have occurred between the two populations over the 22 years period of our study. This rate of migration considerably reduces the probability of a high frequency of the T- mitotype in Seglinge-Kumlinge due to bottlenecks, especially because the C-mitotype is significantly more prevalent in the potential source population of Sottunga. It is thus more likely that selection acts on the wasps after migration in Seglinge-Kumlinge. Additionally, there is currently no evidence that the T-mitotype provides any benefit to its host that the C-mitotype would not, and it remains generally rare across Åland (Duplouy et al., 2015). Consequently, the high prevalence of the T-mitotype in Seglinge-Kumlinge is most likely due to the spread ofw Hho under relaxed pressures of the hyperparasitoid.