Protection of rainforest frogs in recolonised populations from Batrachochytrium dendrobatidis infection by symbiotic microbiota: Repeated, divergent evolution of the hologenome?

Background/Question/Methods

Epidemics of chytridiomycosis, caused by the amphibian chytrid fungus Batrachochyrtium dendrodatitis (Bd), caused frog declines in Australian Wet Tropics rainforests.  Declines occurred only at high elevations, and several species have subsequently reappeared at many upland sites.  Gene flow among these sites is unlikely, so these reappearances may represent independent acquisition of greater resistance to chytridiomycosis.  Mutualist skin microbes that inhibit Bd can increase resistance.  We set out to determine how the microbiota of vulnerable frogs at several sites in the Wet Tropics interacted with Bd, whether these interactions differed among sites, and how the microbiota at one site changed over time on individual frogs.  We collected isolates from the skins of three species of frogs, and screened them for anti-Bd activity.  Isolates that produced anti-Bd metabolites were sequenced (16S rDNA) and identified.  We used these results to compare the prevalence and assemblage composition of anti-Bdbacteria among sites. We also collected  swab samples from individual frogs captured multiple times at one high elevation site over one year, determined their infection status, and assessed the diversity of their microbial assemblages using DGGE.  This enabled us to examine the stability and composition of the microbiota over an extended period in the field.

Results/Conclusions

At one upland site (Kirrama uplands) we found an unusually high diversity of inhibitory bacteria inhabiting an unusually high proportion of frogs.  The modal number of inhibitory bacterial genera per frog at Kirrama Uplands was three, while at all other sites it was zero, and 85% of frogs at Kirrama Uplands had one or more genus of bacteria that inhibited the growth of Bd; this may be a high enough prevalence to confer herd immunity.  At another site (Wooronooran uplands) the inhibitory microbiota was less diverse and not as prevalent as at Kirrama, but uninfected frogs shared a common core assemblage of bacterial species, while infected frogs had more variable assemblages that did not overlap greatly with the core assemblage.   Our results suggest that the reappearance of declining species may have been facilitated by their microbiota, and that it is possible that different defenses have evolved at different sites.