Nature 535, 376–381 (2016)

The composition and metabolic activities of the human gut microbiome can be correlated with metabolic conditions such as insulin resistance (IR). Serum levels of amino acids (particularly branched-chain amino acids, BCAAs), lipopolysaccharides, and certain triacylglycerols and membrane phospholipids have also been linked to IR, raising the possibility that the gut microbiome is a source of these metabolites. To test this idea, Pedersen et al. first determined the fasting serum metabolomes of non-diabetic human patients whose potential for IR or metabolic syndrome was assessed by clinical measures. The metabolome analyses revealed that 19 clusters of metabolites, including BCAAs, lipopolysaccharides and triacylglycerols, were significantly associated with both clinically determined IR and metabolic syndrome (defined as the IR metabotype). The authors then found functional alignments between the metabotype data and microbiome modules in the KEGG database and determined that these correlations are primarily driven by six bacterial species, including Prevotella corpi. Consistent with this finding, individuals with detectable levels of P. corpi showed a positive correlation with one BCAA-containing metabolite cluster, and challenging mice fed a high-fat diet with P. corpi aggravated insulin resistance and glucose intolerance. These results define an increase in potential BCAA biosynthesis in the gut microbiome of individuals with IR, possibly mediated directly by P. corpi, and a potential causal role of microbially synthesized BCAAs in metabolic disorders.