Stress is significant to biological communities as it affects species diversity, community composition, and frequency/kind of interactions that occur. Many ecological theories have been applied to and constructed from studies of stress on animal and plant communities; however, macro-organismal communities represent a small proportion of the species and functional diversity found globally relative to microbial taxa. These microbial communities are integral to ecosystems and have effects that span from pairwise interactions (e.g. plant-fungal associations) to larger global processes (e.g. carbon cycling). While we understand the significance of microbes in ecosystems, there remains a dearth of information on microbiome responses to stress and whether principles derived from macro-organisms scale down to the microbial world. To study the effects of stress gradients on microbiomes, we used the Florida rosemary scrub, which provides a gradient along three distinct persistent and highly replicated habitats: rosemary scrub (high stress), scrubby flatwoods (intermediate stress), and flatwoods (low stress). We sampled soil from 71 sites at two depths within each of these habitat types. We then sequenced the 16S_V4 and ITS1 amplicon regions to quantify the diversity, composition, and complexity of prokaryotic and fungal communities, respectively. We used established QIIME methods and co-occurrence networks.
Results/Conclusions
When comparing richness (observed operational taxonomic units) along stress gradients, microbial richness was higher in flatwoods (low stress) at both depths relative to less stressful habitats although evenness was unaffected by the gradient. Crust and subterranean communities were distinct in phylogeny and abundance (Weighted UniFrac distances: Prokaryotic PCoA1: 36.7% of variance, Fungal PCoA1: 24.7%) while composition across the stress gradient for both communities was more stochastic at both sampling depths. However, much of the variance in community composition (43%, p < 0.001) could be explained by composition of the other microbial community. This suggests that a primary determinant of microbiome composition is dependent on species interactions; if this was not the case, we would expect little variation to be explained by composition of the counterpart microbial community. Facilitative interactions also followed the stress-gradient hypothesis as more stressful habitats had higher ratios of putative positive-to-negative interactions as inferred from co-occurrence networks. Our work agrees with the growing evidence for the importance of microbial interactions in defining these communities and suggests that ecological theories derived from macro-organisms cannot be directly applied to microbes without understanding interspecific interactions.