The ecological context of enzymatic variation

Among the challenges in understanding microbial ecosystems is the presence of physics at vastly different scales. Reactions, catalyzed by enzymes but regulated at the level of the cell, propel the flux of carbon and nitrogen through our atmosphere. Compounding this is the presence of pervasive enzyme sequence variation; this variation has been shown to contain coevolving modes of amino acids which encode the enzyme's evolutionary history. In this work, we take steps towards bridging the gap between the enzymatic variation present in an ecosystem and its subsequent activity. We employ the reduction of nitrate by NarG as a model system, which acts as an essential step in the nitrogen cycle by mediating both the return of di-nitrogen to the atmosphere and the assimilation of nitrate into biomass. Considering both metagenomic reconstructions as well as functional data from soil nitrate reducers, we find that sequence variants of enzymes obey predictable responses to environmental fluctuations. That is, while prior community-level metagenomic studies have characterized the response of bacterial strains to the environment, our study provides an enzyme variant level sequence-to-response map. Further, we demonstrate that a simple statistical model can predict the organismal phenotype from variant sequence; in soil samples not originally seen by that model, the prediction of a variant's reduction rate correlates with how much cells with that variant grow in abundance.