Convergent Functional Constraints Make Biochemical Reaction Diversity Predictable

Abstract

The predictability of biochemical evolution remains highly debated, as it is unknown if biological function can be predicted a priori. Here, we explore predictability of function not in individual reactions, but in ensembles. Sampling from 13,777 bacterial, 371 archaeal, and 203 eukaryotic genomes, we use Enzyme Commission (EC) classification to hierarchically group enzyme-catalyzed reactions that perform similar transformations into functional equivalence classes. We show that organisms partition their reaction diversity among these functional equivalence classes in a predictable way, identifying over 120 new, system-size dependent functional scaling relationships. We find distantly related lineages, acting on distinct molecular substrates, can display similar functional scaling, demonstrating that convergence is not driven by phylogeny or common reactions. We demonstrate how transitions in functional scaling can be identified with physiological shifts, using as case studies O2-utilizing oxidoreductases and hexosyltransferases. Taken together, our results open novel avenues for predicting global features of evolving enzyme populations, independent of protein structures. These functional constraints may have broad implications, including for predicting biochemical diversity, designing synthetic organisms, and modeling the evolution of reaction diversity in cases where the exact identity of catalysts is not known, such as during the emergence of life and for potential alternative forms of biochemistry.