A chemical basis for metabolic cooperation in microbial communities

What distinguishes autonomous and cooperating metabolic networks in microbes? How does the underlying universal chemical network constrain metabolic cooperation between individuals? Answering these questions remains a conceptual roadblock given the dearth of culturable, sequenced microorganisms: only about 1% of the expected diversity. Here, we attempt to sidestep this experimental limitation by algorithmically generating reaction networks from the repertoire of chemical reactions in KEGG. We generate a large set of reaction networks, both autonomous and cooperating (cross-feeding). We survey their size, energy (ATP) and biomass yields, as well as their response to a variety of perturbations. We find that while rare, cross-fed pairs can best even the most productive autonomous networks without an obvious compromise to stability. This central result is robust to changing environmental conditions, biomass compositions and the precise method to calculate yields. As a proof-of-concept that a “productivity boost” is possible by cross-feeding, our study provides a chemical basis for the prevalence of metabolic diversity and cooperation in naturally-occurring microbial communities.