Computational design and discovery of nanomolar inhibitors of IκB kinase β

Abstract

IκB kinase β (IKKβ) is a useful target for the discovery of new medicines for cancer and inflammatory diseases. In this study, we aimed to identify new classes of potent IKKβ inhibitors based on structure-based virtual screening, de novo design, and chemical synthesis. To increase the probability of finding actual inhibitors, we improved the scoring function for the estimation of the IKKβ-inhibitor binding affinity by introducing proper solvation free energy and conformational destabilization energy terms for putative inhibitors. Using this modified scoring function, we have been able to identify 15 submicromolar-level IKKβ inhibitors that possess the phenyl-(4- phenyl-pyrimidin-2-yl)-amine moiety as the molecular core. Decomposition analysis of the calculated binding free energies showed that a high biochemical potency could be achieved by lowering the desolvation cost and the conformational destabilization for the inhibitor required for binding to IKKβ as well as by strengthening the interactions in the ATP-binding site. The formation of two hydrogen bonds with backbone amide groups of Cys99 in the hinge region was found to be necessary for tight binding of the inhibitors in the ATP-binding site. From molecular dynamics simulations of IKKβ-inhibitor complexes, we also found that complete dynamic stability of the bidentate hydrogen bond with Cys99 was required for low nanomolar-level inhibitory activity. This implies that the scoring function for virtual screening and de novo design would be further optimized by introducing an additional energy term to measure the dynamic stability of the key interactions in enzyme−inhibitor complexes.