Introduction
Introduced and invading populations generally show low genetic variability, and a different genetic structure than in their native range, due to small founder populations, and demographic bottlenecks (Hufbauer, Bogdanowicz, & Harrison, 2004). Low genetic variability may in turn influence population level dynamics, persistence, and evolutionary potential of introduced populations (Fauvergue, Vercken, Malausa, & Hufbauer, 2012; Szucs, Melbourne, Tuff, & Hufbauer, 2014). The individuals founding new populations usually bring along various symbiotic passengers. A common example of such a symbiont is the α-Proteobacterium Wolbachia pipientis - a maternally inherited endosymbiotic bacterium that infects over 40% of insect species (Detcharoen, Arthofer, Schlick-Steiner, & Steiner, 2019).Wolbachia can be intimately involved in the biology of their hosts, through either manipulating the host reproductive system (O’Neill, Hoffman, & Werren, 1997), susceptibility to predators, parasites or pathogens (Fytrou, Schofield, Kraaijeveld, & Hubbard, 2006; Hedges, Brownlie, O’Neill, & Johnson, 2008; Osborne, Iturbe-Ormaetxe, Brownlie, O’Neill, & Johnson, 2012; van Nouhuys, Kohonen, & Duplouy, 2016), metabolism (Gruntenko et al., 2017; Gruntenko et al., 2019), or dispersal capacities (Evans et al., 2009).Wolbachia -mediated costs and benefits have been shown to affect host population dynamics (Charlat et al., 2009; Duplouy, Hurst, O’Neill, & Charlat, 2010; Verne, Johnson, Bouchon, & Grandjean, 2012), select for particular host genotypes (Signor, 2017), or even hamper the evolution of host traits in infected populations (Martinez et al., 2016). Consequently, studying spatio-temporal patterns in the penetrance and prevalence of symbionts in host populations along with the genetic structure of introduced and original host populations, can provide crucial insights into how both intentionally and accidentally introduced species may successfully establish, persist and further disperse across habitats (Lu, Hulcr, & Sun, 2016).
The Glanville fritillary butterfly, Melitaea cinxia (L.) (Lepidoptera: Nymphalidae) lives as a classical metapopulation in Åland, Finland (I. Hanski, Pakkala, Kuussaari, & Lei, 1995). The (meta)population ecology and dynamics of the butterfly and associated community of parasitoid species has been under study since the early 90’s (van Nouhuys & Hanski, 2005), revealing, for example, that the butterfly population dynamics dictates the population sizes of its associated parasitoids (I. Hanski et al., 2017). In August 1991, seventy-two families of gregarious M. cinxia caterpillars were intentionally introduced on the island of Sottunga, on the East side of the Åland archipelago (Fountain et al., 2018; I. Hanski et al., 2004; I. Hanski et al., 2017). Through natural parasitism of about a third of the introduced butterfly larvae, the specialist parasitoid waspHyposoter horticola (Gravenhorst) (Hymenoptera: Ichneumonidae: Campoplaginae), and its own specialist hyperparasitoid Mesochoruscf. stigmaticus (Hymenoptera: Ichneumonidae: Mesochorinae) (I. Hanski et al., 2004; G. C. Lei, Vikberg, Nieminen, & Kuussaari, 1997; Montovan, Couchoux, Jones, Reeve, & van Nouhuys, 2015; Shaw, Stefanescu, & Van Nouhuys, 2009; van Nouhuys & Ehrnsten, 2004) were simultaneously accidentally introduced in Sottunga. As they make up the highest trophic levels of insect communities, parasitoids are extremely sensitive to the spatio-temporal dynamics and structure of their host resources (Cronin & Reeve, 2005; Gagic et al., 2012; Gagic et al., 2011; Kaartinen & Roslin, 2011; Nair, Fountain, Ikonen, Ojanen, & van Nouhuys, 2016; van Nouhuys, 2005). Nonetheless, despite occasional strong bottlenecks through local butterfly population crashes (Fountain et al., 2016; I. Hanski et al., 2004; van Bergen et al., 2020), the three species have persisted on Sottunga, more than 30km away from the mainland population, and more than 12km away from any other island population. In this system, Wolbachia is only known to infect the parasitoid wasp, H. horticola (Duplouy, Couchoux, Hanski, & van Nouhuys, 2015). In Åland, the host population infection occurs at an intermediate and stable rate of about ≈50% across the archipelago (Duplouy et al., 2015), but the local prevalence of the bacterium differs between the mainland and neighbouring isolated islands (Duplouy et al., 2015), and the infection is more often associated to one of the two previously described mitochondrial haplotypes of the host (Duplouy et al., 2015). Finally, the infection is costly and increases, by almost two folds, the susceptibility of infected individuals to the hyperparasitoid (van Nouhuys et al., 2016), which itself varies in prevalence and rate of hyperparasitism across local populations in Åland (Montovan et al., 2015; Nair et al., 2016).
We analysed spatio-temporal variations in both the genetic structure of the parasitoid host, H. horticola, and the infection rate of the parasitoid by the endosymbiont Wolbachia on the island of Sottunga, and four other regions in the Åland islands. We used 14 nuclear microsatellite markers and one mitochondrial marker to genotype 324 wasps, to infer history and outcome of the accidentally introduced small population, over a 22year period (1992-2013), and screened the wasps for infection with Wolbachia over the same period. We investigated (1) whether gene flow occurred after the accidental introduction of the parasitoid species on the island of Sottunga, and migrations supported persistence of the neighbouring island populations despite occasional population crashes, and (2) whether variations in the local levels of hyperparasitism selected for Wolbachia -infected or uninfected host genotypes in isolated local populations.