No abstract

Predicting antibiotic efficacy within microbial communities remains highly challenging. Interspecies interactions can impact antibiotic activity through many mechanisms, including alterations to bacterial physiology. Here, we studied synthetic communities constructed from the core members of the fruit fly gut microbiota. Co-culturing of Lactobacillus plantarum with Acetobacter species altered its tolerance to the transcriptional inhibitor rifampin. By measuring key metabolites and environmental pH, we determined that Acetobacter species counter the acidification driven by L. plantarum production of lactate. Shifts in pH were sufficient to modulate L. plantarum tolerance to rifampin and the translational inhibitor erythromycin. A reduction in lag time exiting stationary phase was linked to L. plantarum tolerance to rifampicin, opposite to a previously identified mode of tolerance to ampicillin in E. coli. This mechanistic understanding of the coupling among interspecies interactions, environmental pH, and antibiotic tolerance enables future predictions of growth and the effects of antibiotics in more complex communities.

SummaryMechanistic understanding of the impacts of the gut microbiota on human health has been hampered by limited throughput in animal models. To enable systematic interrogation of gut-relevant microbial communities, here we generated hundreds of in vitro communities cultured from diverse stool samples in various media. Species composition revealed stool-derived communities that are phylogenetically complex, diverse, stable, and highly reproducible. Community membership depended on both medium and initial inoculum, with certain media preserving inoculum compositions. Different inocula yielded different community compositions, indicating their potential for personalized therapeutics. Communities were robust to freezing and large-volume culturing, enabling future translational applications. Defined communities were generated from isolates and reconstituted growth and composition similar to those of communities derived from stool inocula. Finally, in vitro experiments probing the response to ciprofloxacin successfully predicted many changes observed in vivo, including the resilience and sensitivity of each Bacteroides species. Thus, stool-derived in vitro communities constitute a powerful resource for microbiota research.

Diet can impact host health through changes to the gut microbiota, yet we lack mechanistic understanding linking nutrient availability and microbiota composition. Here, we use thousands of microbial communities cultured in vitro from human feces to uncover simple assembly rules and develop a predictive model of community composition upon addition of single nutrients from central carbon metabolism to a complex medium. Community membership was largely determined by the donor feces, whereas relative abundances were determined by the supplemental carbon source. The absolute abundance of most taxa was independent of the supplementing nutrient, due to the ability of fast-growing organisms to quickly exhaust their niche in the complex medium and then exploit and monopolize the supplemental carbon source. Relative abundances of dominant taxa could be predicted from the nutritional preferences and growth dynamics of species in isolation, and exceptions were consistent with strain-level variation in growth capabilities. Our study reveals that community assembly follows simple rules of nutrient utilization dynamics and provides a predictive framework for manipulating gut commensal communities through nutritional perturbations.

No abstract