Rossmann-Like Proteins Function and Evolution Analysis of a Fifth of the Protein World

Abstract

Rossmann folds are ancient, frequently diverged domains found in many biological reaction pathways where they have adapted different functions and constitute about 20% of all known protein structures. The classical Rossmann fold, also known as a doubly-wound three layer α/β sandwich, consists of two β-α-β-α-β units (with 321456 topology) that form a single parallel sheet flanked by α-helices on both sides and contain a characteristic crossover between strands 3 and 4. We defined its core minimal Rossmann-like motif (RLM) unit of three β-strands flanked by two α-helices and found all known protein structures containing the RLM. We showed that RLM enzymes utilize ligands from 20 chemical superclasses of organic and inorganic compounds. The ability of these proteins to bind many types of ligands is provided by the incorporation of an RLM into a broad array of structural contexts. Homologous RLM domains can exhibit diverging active sites that accommodate alternate ligands, but with similar binding modes. We show that RLM enzymes function predominantly in metabolism, covering 38% of reference metabolic pathways. Closely related, homologous RLM enzyme families can catalyze different reaction chemistries using a similar topology. Conversely, non-homologous RLM domains can converge to catalyze the same reactions or to bind the same ligand with alternate binding modes. The Rossmann fold is considered one of the most ancient folds, utilizing iron-sulfur clusters as cofactors and being the part of ancient energy metabolism, the Wood-Ljungdahl pathway, used by LUCA. We found that strong inhibiting agents, such as methotrexate, can inhibit multiple non-homologous RLM enzymes, which is particularly relevant in the light of polypharmacology theory. Our data suggests that the top three disease categories with mutations in RLM proteins are diseases of endocrine system, nervous system and developmental anomalies.