Engineered Interspecies Amino Acid Cross-Feeding Increases Population Evenness in a Synthetic Bacterial Consortium

Abstract

In nature, microbes interact antagonistically, neutrally, or beneficially. To shed light on the effects of positive interactions in microbial consortia, we introduced metabolic dependencies and metabolite overproduction into four bacterial species. While antagonistic interactions govern the wild-type consortium behavior, the genetic modifications alleviated antagonistic interactions and resulted in beneficial interactions. Engineered cross-feeding increased population evenness, a component of ecological diversity, in different environments, including in a more complex gnotobiotic mouse gut environment. Our findings suggest that metabolite cross-feeding could be used as a tool for intentionally shaping microbial consortia in complex environments.IMPORTANCE Microbial communities are ubiquitous in nature. Bacterial consortia live in and on our body and in our environment, and more recently, biotechnology is applying microbial consortia for bioproduction. As part of our body, bacterial consortia influence us in health and disease. Microbial consortium function is determined by its composition, which in turn is driven by the interactions between species. Further understanding of microbial interactions will help us in deciphering how consortia function in complex environments and may enable us to modify microbial consortia for health and environmental benefits.

Keywords: metabolite cross-feeding; microbial consortia; synthetic biology.

FIG 1

Engineered metabolite overproduction and growth requirements. (a) Consortium design. Each strain is auxotrophic for three amino acids and overproduces one. (b) Quantification of overproduction and growth requirement (EC₅₀). Overproduction in each engineered strain was measured via LC-MS with an appropriate amino acid standard. Metabolite growth requirement EC₅₀ was measured in medium with full supplementation of two amino acids and various amounts of one other. (c) Cross-feeding capabilities of each strain were assessed by testing for rescue of auxotroph growth in supernatants obtained from overproducers. (d) Growth assay of auxotrophic strains. Auxotroph strains were grown in medium without and with addition of all cross-fed amino acids (left side of graphs, −/+), which served as a reference point. Each strain was also grown in fresh medium without supplementation and supernatant of engineered overproducers or WT counterpart at a 1:1 ratio. Shown are three biological replicates with medians indicated as horizontal lines.

FIG 2

Engineered consortium exhibits added beneficial interactions and reduced antagonistic interactions. (a and b) Growth trajectories in anaerobic batch cocultures for WT (a) and engineered (b) consortia. All strains were inoculated at equal ratios and grown anaerobically at 37°C for 27 h in M9 minimal medium with specific modifications as described in Materials and Methods, without supplementation of any of the cross-fed amino acids, and with 0.5% starch and 0.5% glucose as carbon sources. The engineered consortium grows to an order-of-magnitude-lower density than the WT counterpart. (c and d) Growth trajectories of consortia with each strain inoculum reduced by 10-fold from WT (c) and engineered (d) consortia. These data combined were used to infer a network interaction using the MDSINE algorithm. (e and f) Inferred interaction networks for WT (e) and engineered (f) consortia. WT consortia show mainly negative interactions with large Bayes factors. Engineered consortia’s negative interactions have reduced Bayes factors, and positive interactions with strong Bayes factors occurred between B. fragilis and E. coli and S. Typhimurium.

FIG 3

Engineered consortia exhibit increased population evenness in vitro. WT and engineered consortia were grown anaerobically in batch coculture for 24 h anaerobically in modified M9 medium as described in Materials and Methods with and without supplementation of 1 mM cross-fed amino acids and 0.2% starch and 0.5% glucose as carbon sources. (Upper panel) Average population composition after 24 h. (Lower panel) Quantified Pielou evenness. The engineered consortium has higher population evenness with or without amino acid supplementation than the WT consortium.

FIG 4

Consortium engineering increases population evenness in the mammalian gut in a diet-dependent manner. Four groups of germfree mice (n = 5, except second group from the left, which is n = 4) were fed either a low-protein diet or standard diet and inoculated with either the engineered or WT bacterial consortium. Fecal samples 10 days postinoculation were analyzed via strain-specific qPCR to assess concentrations of each consortium species. Population evenness of consortia was calculated. Bar indicates median. The Mann-Whitney test showed significantly increased population of the engineered consortia in mice that were fed the low-protein diet compared to the consortia in the three other groups (P values, 0.024, 0.032, and 0.015). Yellow, E. coli; purple, S. Typhimurium; green, B. thetaiotaomicron; blue, B. fragilis.