Maternally inherited symbionts are common in arthropods and many have important roles in host adaptation. The observation that specific symbiont lineages infect distantly related host species implies new interactions are commonly established by lateral transfer events. However, studies have shown that symbionts often perform poorly in novel hosts. We hypothesized selection on the symbiont may be sufficiently rapid that poor performance in a novel host environment is rapidly ameliorated, permitting symbiont maintenance. Here, we test this prediction for a Spiroplasma strain transinfected into the novel host Drosophila melanogaster. In the generations immediately following transinfection, the symbiont had low transmission efficiency to offspring and imposed severe fitness costs on its host. We observed that effects on host fitness evolved rapidly, being undetectable after 17 generations in the novel host, whereas vertical transmission efficiency was poorly responsive over this period. Our results suggest that long-term symbiosis may more readily be established in cases where symbionts perform poorly in just one aspect of symbiosis.

Maternally inherited endosymbionts are found in numerous insect species and have various effects on host ecology. New symbioses are most commonly established following lateral transfer of an existing symbiont from one host species to another. Laboratory study has demonstrated that symbionts commonly perform poorly in novel hosts, with weak vertical transmission and maladaptive pathogenicity being observed in the generations following transfer. This poor performance probably limits symbiont occurrence. We here use microarray technology to test whether poor symbiont performance observed following 1 year of vertical transmission through a new host is associated with alteration in host gene expression or whether it occurs independently of this. We utilize the Drosophila melanogaster--Spiroplasma interaction and test the response of the host in the presence of both natural Spiroplasma infections and novel Spiroplasma infections transinfected previously from other host species. None of the Spiroplasma infections investigated produced upregulation in host haemolymph/fat body-based immune responses, and we therefore rejected the hypothesis that failure to thrive was associated with immune upregulation. One infection was associated with a downregulation of genes associated with egg production compared to uninfected controls, indicative of damage to the host. The Spiroplasma infection showed that the weakest vertical transmission showed no significant disturbance to host gene expression compared to uninfected controls. We conclude that the failure of Spiroplasma in novel host species is associated either with causing harm to their new hosts or through a failure to thrive in the new host that occurs independently of host responses to infection.

Insects deploy cellular and humoral defences to defend themselves against pathogens and parasites. The recent discovery that fly defences are commonly supplemented by inherited protective microbes suggests that symbiont-mediated protection is common. Resistance evolution may have a more complex dynamic than previously described.

No abstract