(648c) Structure Guided Phylogenetic Tree for Novel Hydrolase Conceptualization

Hydrolases are an important class of enzyme for their environmental and therapeutic application such as, bioremediation, wastewater treatment, probiotics, drug metabolism, antibiotic formulations, and vaccine production. While a sequence-based alignment of all hydrolases would reveal similarity across the family of all hydrolases spanning different species, it does not necessarily translate to functional similarity of hydrolases. Overall structure and local geometry of how certain residues are packed dictate function of each hydrolase with higher fidelity. To this end, we have first constructed a database of 29,938 hydrolases from the Protein Data Bank spanning 3 domains of life – eukaryotes, bacteria and archaea. The median length of hydrolases in this set is 254 amino acids, with the longest ones being 2579 amino acids long and the shortest being 11 amino acids long (both from Homo Sapiens). We constructed the entire inter-residue distance map for each member of the hydrolase family and chartered their closest contacts within 6.5 Å. This provides us with insights on interacting amino acid types that are crucial for holding each hydrolase structure together. We compute the similarity along such intra-contact landscape to yield a function-aware phylogenetic tree. By inspecting and comparing various branches of the hydrolase family tree we glean rules for designing novel hydrolases with (a) altered substrate specificity, and (b) desired catalytic turnover (i.e., activity or kcat) – yet, avoiding specific amino acid transitions at select residues to keep structural stability intact. We are thus able to unravel a transition map across hydrolases by performing minimal number of mutations to switch their substrate specificities from one to another, and corroborate our findings experimentally.