RNA binding proteins play vital roles in RNA metabolism and function, including splicing, translation, localization, stability and degradation. In addition to canonical RNA binding proteins where RNA binding is an aspect of their main function, dozens of moonlighting proteins have been found that combine an enzymatic function in sugar, lipid, and amino acid metabolism with RNA binding function. The wide occurrence and conservation of RNA binding ability across distant branches of the evolutionary tree suggest that these moonlighting enzymes/RNA binding proteins are involved in connections between intermediary metabolism and gene expression that comprise far more extensive regulatory networks than previously thought. Combining catalytic and RNA binding functions in one protein can be a mechanism to coordinate cellular activities, for example, by sensing the cell’s metabolic state through availability of the enzyme’s ligands and responding by regulating translation of specific transcripts. Conversely, RNA binding could regulate the enzyme’s catalytic activity, through binding in the active site, allosteric effects, acting as a scaffold, or sequestering enzymes. How these moonlighting enzymes bind to RNA is largely unknown, and most do not contain canonical RNA binding domains. We are using a combination of biochemical assays, structure determination, and computational methods to identify the RNA binding sites and investigate the molecular mechanisms of RNA binding, including the conformation of the protein and the structure of the RNA in each complex. We are developing a new version of partial order optimum likelihood (POOL) to identify RNA binding sites. POOL is a machine-learning method originally developed for predicting the locations of catalytic and ligand binding sites in protein structures. It utilizes computed features of the protein structure as input: electrostatic and chemical properties of the individual amino acids as well as surface topological information.
Tu et al. (Sun,) studied this question.