Enzyme ‘Social’ Networks: Tracking Sequence / Function Relationships to Elucidate Determinants of Enzyme Specificity and Mechanism

Abstract

Integrated analysis of sequence, structure and functional properties of enzymes is an established approach for determining their mechanisms and specificities. As sequence data continue to mount, analyzing the natural variation of homologs offers a powerful context for guiding experiments, while turning studies of single enzymes or large superfamilies into bioinformatic 'big data' problems. Many current tools designed to tease out potential functional residues are not sensitive enough to accommodate the specific intricacies of enzyme superfamilies, yielding biased predictions and misannotation. Here, we present the Structure‐Function Linkage Database (http://sfld.rbvi.ucsf.edu), a hierarchical classification resource and bioinformatic infrastructure that enables research of structure‐function relationships in functionally diverse enzyme superfamilies. It offers manually curated and automatic classification of enzymes, informed by the mechanistic blueprint of their corresponding superfamilies. Importantly, annotation and data exploration in the database takes advantage of sequence similarity networks to visualize functional trends mapped to the context of sequence similarity. We will illustrate how networks are used for enzyme classification in the glutathione transferase superfamily and show how mapping active site residue variation across a large number of proteins aids in elucidating specificity determinants in the Sirtuin superfamily of deacetylases. In summary, our approach allows a unique way to effectively utilize massive multidimensional data for generating hypotheses about enzyme mechanism and specificity.