Bacterial microbiota composition of Ixodes ricinus ticks: the role of environmental variation, tick characteristics and microbial interactions

Abstract

Ecological factors, host characteristics and/or interactions among microbes may all shape the occurrence of microbes and the structure of microbial communities within organisms. In the past, disentangling these factors and determining their relative importance in shaping within-host microbiota communities has been hampered by analytical limitations to account for (dis)similar environmental preferences ('environmental filtering'). Here we used a joint species distribution modelling (JSDM) approach to characterize the bacterial microbiota of one of the most important disease vectors in Europe, the sheep tick Ixodes ricinus, along ecological gradients in the Swiss Alps. Although our study captured extensive environmental variation along elevational clines, the explanatory power of such large-scale ecological factors was comparably weak, suggesting that tick-specific traits and behaviours, microhabitat and -climate experienced by ticks, and interactions among microbes play an important role in shaping tick microbial communities. Indeed, when accounting for shared environmental preferences, evidence for significant patterns of positive or negative co-occurrence among microbes was found, which is indicative of competition or facilitation processes. Signals of facilitation were observed primarily among human pathogens, leading to co-infection within ticks, whereas signals of competition were observed between the tick endosymbiont Spiroplasma and human pathogens. These findings highlight the important role of small-scale ecological variation and microbe-microbe interactions in shaping tick microbial communities and the dynamics of tick-borne disease.

Keywords: Borrelia burgdorferi; Community composition; Lyme disease; Species distribution modelling; Tick-borne pathogens.

Figure 1. Location of tick sampling sites in the Swiss Alps.

Different shapes (i.e., circle, square and triangle) represent the different locations and different colours represent elevation (white: low, grey: middle, black: high). Rivers and motorway are shown in black. Map data ©2019 Google, GeoBasis-DE/BKG.

Figure 2. Tick microbial community variance partitioning for different fixed and random effects.

The first three columns represent tick endosymbionts, the next three columns are OTUs which are both tick endosymbionts and human pathogens and the subsequent six columns represent human pathogens. The other columns represent the 88 most common OTUs found in I. ricinus, ordered by read frequency. Month, sampling site, location and tick ID were included in the model as random effects, whereas fixed effects were divided into environmental (elevation, temperature, precipitation, forest coverage, slope, aspect, vole abundance and vole-to-other-rodents ratio) and tick-specific variables (life stage or sex, individual heterozygosity, abundance, expected population heterozygosity). See raw data in Figshare for information on OTU labels (DOI: 10.6084/m9.figshare.7380767.v3).

Figure 3. Residual association patterns among endosymbionts and human pathogens within ticks on (A) individual tick-level and (B) on site-level after accounting for shared environmental preference.

Red lines represent positive associations and blue lines negative associations. Only associations with strong statistical support (i.e., based on the 90% central credible interval) are presented. Darker colors indicate stronger associations.