Exploiting receptor flexibility to improve the accuracy of structure-based virtual screening

Structure-based virtual screening is usually challenged by numerous false positives in the candidate drugs predicted by molecular docking tools. Here, on the basis of the binding energy landscape theory, a hypothesis that a true binder can bind to different conformations of the binding site favorabl...

Full description

Saved in:
Bibliographic Details
Main Author: Yaw, Awuni
Other Authors: Mu Yuguang
Format: Theses and Dissertations
Language:English
Published: 2017
Subjects:
Online Access:http://hdl.handle.net/10356/69603
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Structure-based virtual screening is usually challenged by numerous false positives in the candidate drugs predicted by molecular docking tools. Here, on the basis of the binding energy landscape theory, a hypothesis that a true binder can bind to different conformations of the binding site favorably was put forth, and related strategies to defeat this challenge were devised; reducing false positives when receptor plasticity is considered. The receptors in the study are the influenza A nucleoprotein (NP) and the bacterial actin-like MreB. The structural flexibility of the receptors was explored by molecular dynamics (MD) simulations. The resultant distinctive structures and the respective crystal structures were used as receptor models in docking exercises in which two binding sites of NP, the tail-loop binding pocket and the RNA binding site, and the A22 binding site of MreB were targeted with molecule libraries using the GOLD software. In each case, the intersection ligands that were listed in the top-ranked molecules from all receptor models were selected. Such selection strategy successfully distinguished high-affinity and low-affinity control molecules added to the molecule libraries. This study provides an applicable approach for reducing false positives and selecting true ligands from molecule libraries. Furthermore, X-ray crystallographic studies involving the molecules selected for the A22 binding site and MreB were carried to validate the selection criterion. Eleven hits, representing 0.037% of the total 30036 molecules making up the Maybridge fragment library were selected for crystallographic studies. Three of the hits were detected in the active site of MreB. The binding modes of these molecules are close to A22. The data obtained support the hypothesis that a true binder can bind to varying conformations of the binding site. In addition, it was sought to understand the dynamics of the nucleotide-A22 binding site of the bacterial actin MreB. A22 is an antibiotic-like small molecule that perturbs the rod cell shape of bacteria and has been suggested to inhibit MreB by targeting ATP hydrolysis. However, without the elucidation of the structure of the ATP-A22-bound state of MreB, the mechanism of A22 inhibition is still not clear. Here conventional MD simulation was applied to explore the dynamics of the active site of MreB in complex with A22 and different nucleotides. The data obtained suggest that A22 may not function as an inhibitor of ATP hydrolysis. Analysis of the dynamics of water molecules in the MreB active site reveals that in the presence of A22 water molecules are able to occupy positions suitable for ATP hydrolysis. Thus these observations are consistent with a mechanism in which A22 affects MreB polymerization/depolymerization dynamics in part through slowing phosphate release rather than by inhibiting ATP hydrolysis. This is partially supported by the fact that the crystallographic structure of the ATP-bound state has not been solved in the presence of A22. These data can be incorporated in the design/development of the next generation of inhibitors of MreB.