Obligate parasites represent a convenient system within which to test potential drivers of microbiome variation for several reasons. The number of factors influencing variation may be more limited in the microbiomes of obligate parasites than in environmental microbiomes (e.g., soil or water) or microbiomes of free-living host species. Unlike free-living host species, obligate ectoparasites have extremely specialized diets, and their movements in the broader environment beyond their host are constrained by their dependence on a host to survive. Ectoparasitic arthropods also have characteristically depauperate microbiome communities compared to arthropods with diverse diets [19]. We can leverage the hierarchical nature of the host-parasite-microbiome system to clearly delimit the microbiome community and restrict the sources of colonizing bacteria that may invade the parasite microbiome, hence providing a manageable system for testing hypotheses about community composition and factors governing assembly of the microbiome.
In this study, we used bat flies (Diptera: Streblidae and Nycteribiidae), which are obligate blood-feeding ectoparasites of bats, to assess community composition of parasite-associated microorganisms across a fragmented landscape in the Atlantic Forest. The microbiome of bat flies may be influenced by the parasitic bat fly, the host bat, and the environment in several ways (Figure 1A). First, bat fly microbiomes may be vertically inherited or horizontally acquired, as in other insect-microbiome associations (Figure 1A, yellow line; [20]. Second, bat flies are host-specific and depend on their host for dispersal [21, 22], meaning that the host bat may also influence the bat fly microbiome by altering the community of bat flies from which bacteria may be horizontally acquired (Figure 1A, dark purple line). For example, bat maternity colonies likely support more abundant parasite communities than bachelor colonies, because bat flies preferentially parasitize female and young bats [23]. By supporting a smaller community of bat flies, bachelor colonies may decrease the sources from which bat flies horizontally acquire bacteria, leading to host-sex-based variation in the microbiome. Other aspects of bat phylogeny, ecology, and behavior may similarly influence bat fly microbiomes, including roost preference, feeding guild, and host bat species identity. Third, habitat patches support a specific bat community based on the availability of roosting and foraging sites, which subsequently alters the diversity and abundance of bat flies [24]. These changes in the local bat fly community may be reflected in associated microbiomes. Lastly, the environment may directly impact bat fly microbiome composition (Figure 1A, dark green line). Bat flies and all other members of the Hippoboscoidea are adenotrophically viviparous, a condition in which a single egg hatches inside the female fly and the larva feeds from milk glands until it is ready to pupate [21]. In the case of bat flies, the female fly leaves the host bat to deposit the larva on the roost substrate [21], providing opportunities for the environment to act as a source of bacteria for the microbiome of bat flies (Figure 1A). Beyond acting as a source of bacteria for the microbiome, deforestation may increase the local temperature of small habitat patches and directly impact arthropod-associated microbiomes due to thermal constraints of some bacteria [25, 26]. Using the mosaic landscape of the Atlantic Forest, we can examine whether bat fly-associated microbiomes respond to environmental change following island biogeography theory or whether the host bat and parasitic bat fly more strongly determine bat fly microbiome composition.
Methods
Sample collection and landscape metrics
Bat flies were collected from bats in 11 habitat patches of the Atlantic forest of Brazil, State of Rio de Janeiro from 18 December 2015 to 19 January 2017 (Table 1, Figure 1B, Table S1, Figure S1; [27], including a large protected area of pristine and secondary forest belonging to the Reserva Ecológica de Guapiaçu (REGUA). REGUA is the third largest remaining expanse of Atlantic Forest and was sampled in three separate locations. Samples were additionally collected from three geographically distant habitat fragments (southern sites; Figure 1B). Each site was sampled for 6 nights, 6 hours per night or at least 2 hours if there was heavy rain, and between 7 and 10 ground-level mist nets were used to capture bats each night (approximately 60m of nets were set per night; [27]. Bats were removed from mist nets and placed into freshly washed cloth bags for holding and to minimize cross-contamination of ectoparasites. Each bat was searched for approximately 45s for ectoparasites, which were captured using featherweight forceps and immediately transferred to tubes containing 92% ethanol, stored at room temperature overnight, and subsequently transferred to -20°C. Bats were identified in the field following [28, 29]. All capture and handling methods followed recommendations in [30]. Because many bat species were only captured in a subset of sampled sites, we selected bat flies from the six most well-represented bat species for microbiome analysis.