Protein Flexibility in Structure-Based Drug Design.

Abstract

Structure-based drug design (SBDD) is defined as the use of three-dimensional structural data to advance lead development and optimization studies. Many SBDD projects have used a rigid protein structure to represent the receptor target in order to gain greater throughput with minimal computational time. However, numerous studies have illustrated the significant influence protein flexibility exerts upon binding predictions. Inclusion of protein flexibility has become essential due to the need for ligands with novel scaffolds and unique modes of action that combat increasing rates of drug resistance and decreasing approval of clinical candidates. Additionally, accurate modeling of protein flexibility may reveal unknown allosteric sites and increase the number of viable lead compounds for a given target. Previously, Carlson et al. incorporated structural flexibility into pharmacophore modeling through the development of the multiple protein structure (MPS) method. This technique was the first computational-mapping algorithm to identify experimentally-validated lead compounds. Probe mapping is a common computational technique for identifying potential binding pockets along a protein surface. However, the efficacy of most methods has been limited by neglecting desolvation penalties. To broaden the impact of our studies, we have developed an improved technique for probe mapping, Mixed Solvent Molecular Dynamics (MixMD), which extends our MPS approach by simultaneously incorporating flexibility and solvent competition. This technique has been validated on the canonical hen egg-white lysozyme system and has been generalized across a series of pharmaceutically-relevant targets. MixMD can be used to develop accurate pharmacophores of druggable hot spots through the incorporation of several different probe types. As a complement to our methodology development, we have specifically targeted protein flexibility in another canonical protein system. HIV-1 Protease (HIVp) is an exceptional test case due to the abundance of structural data available, its importance as a pharmaceutical target, and its potential for allosteric regulation. Three allosteric sites have been hypothesized for HIVp: the elbow site, the eye site, and the dimer interface. We have used MD simulations to probe the allosteric control possible at the elbow and eye sites by small molecules. Our studies have identified important features for designing effective allosteric inhibitors of HIVp.