Distinct N and C Cross-Feeding Networks in a Synthetic Mouse Gut Consortium

Abstract

The complex interactions between the gut microbiome and host or pathogen colonization resistance cannot be understood solely from community composition. Missing are causal relationships, such as metabolic interactions among species, to better understand what shapes the microbiome. Here, we focused on metabolic niches generated and occupied by the Oligo-Mouse-Microbiota (OMM) consortium, a synthetic community composed of 12 members that is increasingly used as a model for the mouse gut microbiome. Combining monocultures and spent medium experiments with untargeted metabolomics revealed broad metabolic diversity in the consortium, constituting a dense cross-feeding network with more than 100 pairwise interactions. Quantitative analysis of the cross-feeding network revealed distinct C and N food webs, highlighting the two Bacteroidetes members Bacteroides caecimuris and Muribaculum intestinale as primary suppliers of carbon and a more diverse group as nitrogen providers. Cross-fed metabolites were mainly carboxylic acids, amino acids, and the so far not reported nucleobases. In particular, the dicarboxylic acids malate and fumarate provided a strong physiological benefit to consumers, presumably used in anaerobic respiration. Isotopic tracer experiments validated the fate of a subset of cross-fed metabolites, such as the conversion of the most abundant cross-fed compound succinate to butyrate. Thus, we show that this consortium is tailored to produce the anti-inflammatory metabolite butyrate. Overall, we provide evidence for metabolic niches generated and occupied by OMM members that lays a metabolic foundation to facilitate an understanding of the more complex in vivo behavior of this consortium in the mouse gut. IMPORTANCE This article maps out the cross-feeding network among 10 members of a synthetic consortium that is increasingly used as the model mouse gut microbiota. Combining metabolomics with in vitro cultivations, two dense networks of carbon and nitrogen exchange are described. The vast majority of the ∼100 interactions are synergistic in nature, in several cases providing distinct physiological benefits to the recipient species. These networks lay the groundwork toward understanding gut community dynamics and host-gut microbe interactions.

Keywords: food web; metabolic interactions; metabolism; metabolomics; microbial communities.

FIG 1

Exometabolome dynamics of OMM species in mBHI medium. (A) Growth curves of 10 OMM species in mBHI medium. Shaded areas indicate the standard deviations from the means (n = 3 to 4 replicates). (B) Metabolic footprint heat map of all 10 OMM species during growth in mBHI medium. Secretion is indicated in red, and consumption is in blue. Intensities are scaled to ±1 by dividing each metabolite by the maximum observed change in abundance in all species. Hierarchical clustering was performed for metabolites, using Euclidian distances and centroid linkage. (C) Heat map describing the different types of potential interactions extrapolated from the consumption and secretion profiles in panel B.

FIG 2

Metabolite cross-feeding among OMM members in mBHI spent medium. (A) Maximum OD of OMM species during growth in a spent medium mixture of 50% culture supernatant and 50% mBHI medium. Data points are the means from duplicate measurements (see Table S1B in the supplemental material). The species origin of the culture supernatant is indicated by the color of the points. The relative maximum OD achieved by each of the other species is given on the y axis, relative to the maximum OD achieved in fresh mBHI medium. Relative maximum ODs of 1.1, 1, 0.5, and 0.4 are indicated by the dotted lines. (B) Metabolite interaction network of OMM species inferred from spent medium experiments. Metabolites secreted by a producer in fresh mBHI medium and consumed in spent media by a second species were classified as cross-fed. Consumed metabolites in spent medium experiments were identified by filtering all decreasing annotated ions, based on either a significant correlation with the culture OD over time (Pearson correlation coefficient of less than −0.7; P value of <0.05) or a significant goodness of linear or exponential fit (R² of >0.7; P value of <0.05). Bacteria that represent more than 10% of the community are shown in a larger size in a study by Yilmaz et al. (21). (C) Relative metabolite class distributions of public and private cross-fed compounds. Percentages were calculated from the total numbers of metabolites within a class divided by the total number of metabolites, including the ones without a specific class associated.

FIG 3

Carbon and nitrogen interaction networks of the OMM consortium. Interactions were inferred from growth experiments in a mix of 50% complex mBHI medium and 50% spent medium of each OMM species. (A) Compound-specific OMM carbon interaction network. (B) Compound-specific OMM nitrogen interaction network. For both networks, the amount of cross-fed compounds containing carbon or nitrogen was quantified and multiplied by the number of carbon or nitrogen atoms per molecule, respectively. Only compounds above 0.1 C-mmol or N-mmol exchange are displayed. Bacteria that represent more than 10% of the community are shown in a larger size in a study by Yilmaz et al. (21).

FIG 4

Supplementation and ¹³C-tracing experiments reveal the impact of cross-feeding interactions. (A) Impact of cross-fed nutrient supplementation on the maximum OD and growth rate in mBHI medium supplemented with one metabolite of the indicated compound class (n = 3 replicates per experiment). Values are shown relative to the growth rate and maximum OD obtained without supplementation. Color indicates metabolite class, and the dot size is proportional to the significance of the P value (by Student’s t test). The dotted horizontal lines at 1.1 and 0.9 are shown for reference. (B) Extracellular time course of fully ¹³C-labeled succinate and malate in L. reuteri mBHI medium cultures supplemented with ¹³C-malate. Shaded areas represent the standard deviations from the experiments (n = 3 replicates per experiment). (C) Normalized area under the OD curve (AUC) of C. clostridioforme grown in mBHI medium supplemented with different concentrations of cysteine. Shaded areas represent the standard deviations from the means (n = 3 replicates per experiment). (D) Levels of ¹³C-succinate and ¹³C-butyrate over time when C. clostridioforme was grown in mBHI medium supplemented with ¹³C-succinate. Shaded areas represent the standard deviations from the means (n = 3 replicates per experiment). (E) Levels of succinate and butyrate in monocultures and coculture of C. clostridioforme and B. caecimuris in GMM normalized to the maximum value across all experiments (n = 3 replicates per experiment).