Genome-scale metabolic models reveal phylogenetic conservation of amino acid fermentation-linked metabolic traits in soil bacteria

Abstract

Soil microorganisms drive biogeochemical cycling of carbon, nitrogen, and other elements, but the partitioning of metabolic fluxes among taxa remains poorly understood, limiting our ability to link community composition to ecosystem function. Genome-scale metabolic models (GEMs) offer a mechanistic framework to connect taxa and their metabolic pathways to these fluxes. However, heterogeneity of soil communities has hampered recovery of high-quality metagenome-assembled genomes (MAGs), constraining application of flux balance analysis (FBA) to soil metagenomes and skewing available GEMs towards cultured model organisms. This gap limits our understanding of soil microbial physiology and responses to disturbance. Here, we present a collection of GEMs for 550 MAGs (497 bacterial, 53 archaeal) recovered from 94 soil samples across the continental United States. As an example application, we used FBA to estimate uptake and secretion of 44 commonly exchanged metabolites and tested whether these fluxes are phylogenetically conserved. While the overall bacterial phylogenetic signal was weak (median r = 0.029), eight metabolites showed significant conservation: amino acids (L-methionine, L-leucine, L-isoleucine), fermentation intermediates (pyruvate, acetaldehyde, ethanol), and amino acid degradation by-products NH and CO2. Most bacterial MAGs were predicted to import branched-chain amino acids, and a subset also secreted pyruvate and ethanol, identifying candidate cross-feeding interactions. Thus, fermentation-linked traits may be phylogenetically conserved and predictable from community composition, whereas most other metabolites showed no detectable signal. Ultimately, this resource enables genome-resolved prediction of soil microbial metabolisms, from cross-feeding and carbon-use efficiency to community responses to disturbance, providing a mechanistic bridge between microbial taxonomy and ecosystem-scale biogeochemistry.