Effect of altering local protein fluctuations using artificial intelligence

Abstract

The fluctuations in Arg111, a significantly fluctuating residue in cathepsin K, were locally regulated by modifying Arg111 to Gly111. The binding properties of 15 dipeptides in the modified protein were analyzed by molecular simulations, and modeled as decision trees using artificial intelligence. The decision tree of the modified protein significantly differed from that of unmodified cathepsin K, and the Arg-to-Gly modification exerted a remarkable effect on the peptide binding properties. By locally regulating the fluctuations of a protein, we may greatly alter the original functions of the protein, enabling novel applications in several fields.

A computational model for controlling conformational cooperativity and function in proteins

“…This knowledge is important for drug design, prediction of allosteric activity, mutation analysis, evolutionary processes, and design of nanosensors and nanoreactors. Recently, the role of fluctuations on ligand binding was computationally demonstrated by Nishiyama, 3 where local fluctuations of a residue were altered by inducing a single residue mutation that led to changes in binding propensities of a large number of ligands at a distant site from the mutation. Several other examples from known biological systems may be cited.…”