Systematic dissection of a complex gut bacterial community

Abstract

Efforts to model the gut microbiome have yielded important insights into the mechanisms of interspecies interactions, the impact of priority effects on ecosystem dynamics, and the role of diet and nutrient availability in determining community composition. However, the model communities studied to date have been defined or complex but not both, limiting their utility. Here, we construct a defined community of 104 bacterial strains composed of the most common taxa from the human gut microbiota. By propagating this community in growth media missing one amino acid at a time, we show that branched-chain amino acids have an outsize impact on community structure and identify a pathway in Clostridium sporogenes for generating ATP from arginine. We constructed and propagated the complete set of single-strain dropout communities, revealing a sparse network of strain-strain interactions including a novel interaction between C. sporogenes and Lactococcus lactis driven by metabolism. This work forms a foundation for studying strain-strain and strain-nutrient interactions in highly complex defined communities, and it provides a starting point for interrogating the rules of synthetic ecology at the 100+ strain scale.

Microcosm cultures of a complex synthetic community reveal ecology and genetics of gut microbial organization

“…predictions by altering the expression of identified gene families using gene knockout, knockdown or knockin experiments [62][63][64]. Second, while our in vitro results generally parallel those from earlier in vivo work [15,26], we do find limited discrepancies (e.g., microcosm depletion of Actinobacteriota), meaning our current platform provides a close but still imperfect replica of the in vivo gut environment. More realistic culture conditions can be explored, potentially through modification of media conditions (e.g., addition of bile acids, different carbon sources).…”

“…By using metagenomic sequencing, we are able to profile microbes with high taxonomic resolution, enabling strain- and gene-level analysis. We use a synthetic 123 strain community modeled closely after the recently published hCom2 community [26, 27], cultured in vitro with added mucin microcosms to provide a mucosal-like substrate for bacterial attachment distinct from the surrounding liquid supernatant [28, 29]. To identify genetic correlates of microcosm colonization, we develop a computational workflow that uses a comprehensive search across KEGG Orthology (KO) gene families [30] to identify associations between gut spatial organization and underlying microbial genotypes, using phylogenetic regression to account for evolutionary relationships between taxa [31–33].…”

Model-guided design of the diversity of a synthetic human gut community

“…This ability to rationally inform community experiments with high-throughput monoculture data should make our approach useful for larger communities, potentially even up to 100 members 59 . As species richness increases, the degree of metabolic similarity among species would increase (i.e., metabolic redundancy), leading to potential challenges in identifying specific nutrients that can tune the growth of individual species in the community.…”

Pitfalls of genotyping microbial communities with rapidly growing genome collections

“…We used 327 UHGG species with at least two high-quality genomes and at least one CRS (interspecies ANI ≥ 92%) to further explore reference bias and to evaluate cross-mapping, alignment uniqueness, and alignment sequence identity ( Table S4 ). We used 86 NCBI genomes from species commonly found in the human gut (Cheng et al, 2021) and represented in UHGG with at least one related species (inter-species ANI ≥ 80%) to evaluate performance differences between reference and alternative alleles, as well as the effects of post-alignment filtering and database customization ( Table S5 ). All genomes used as simulation templates were high quality (Completeness ≥ 90, Contamination ≤ 5), and only species with at least two high-quality genomes were used.…”