Connecting trait-based microbe- and substrate-explicit modeling: an example application in the rhizosphere

Carbon use efficiency (CUE)the amount of carbon incorporated into biomass versus respired to the atmospherereflects dynamic allocation of a microbes energy budget to growth under given thermodynamic constraints. CUE is an important parameter in ecosystem models and a frequently cited driver of soil carbon decomposition. In the rhizosphere, it has been hypothesized that organic matter formation is associated with high CUE. To better understand the variation in CUE in distinct soil microbial communities from rhizosphere and bulk soils, we used a dynamic energy budget model (DEBmicroTrait), which allows for a thermodynamically consistent treatment of the balance between structural maintenance, structural growth and extracellular enzyme production in microbial metabolism. We used a genomes-to-traits workflow (microTrait), together with allometric scaling theory and biophysical modeling to synthesize information in genome annotations and infer potential microbial phenotypes, including minimum generation times, cell size, cell stoichiometry, and to constrain model parameters for substrate uptake, assimilation efficiency, depolymerization rates and enzyme allocation, protein synthesis, and maintenance rates. We then conducted batch simulations for 39 bacterial isolates individually grown on 84 root exudate compounds from the wild oat grass Avena barbata. We found a significant association between experimentally observed rhizosphere response (isolates that increased in relative abundance in response to plant growth) and CUE; the CUE of rhizosphere-adapted bacteria was consistently higher across low molecular weight substrate classes. DEBmicroTrait predicts a substantial amount of variation in CUE, both at broad (class, ~20%) and fine (strain, ~40%) taxonomic levels. While exudate compound type was a weak predictor across species (~6%), it explained ~50% of variation in CUE within species. Our study suggests that genome-level information linked with dynamic energy budget trait-based and substrate-explicit modeling can resolve variations in CUE within and across microbial communities, and ecosystem scale models should consider such variations.