Ligand-guided receptor optimization.
ABSTRACT Receptor models generated by homology or even obtained by crystallography often have their binding pockets suboptimal for ligand docking and virtual screening applications due to insufficient accuracy or induced fit bias. Knowledge of previously discovered receptor ligands provides key information that can be used for improving docking and screening performance of the receptor. Here, we present a comprehensive ligand-guided receptor optimization (LiBERO) algorithm that exploits ligand information for selecting the best performing protein models from an ensemble. The energetically feasible protein conformers are generated through normal mode analysis and Monte Carlo conformational sampling. The algorithm allows iteration of the conformer generation and selection steps until convergence of a specially developed fitness function which quantifies the conformer's ability to select known ligands from decoys in a small-scale virtual screening test. Because of the requirement for a large number of computationally intensive docking calculations, the automated algorithm has been implemented to use Linux clusters allowing easy parallel scaling. Here, we will discuss the setup of LiBERO calculations, selection of parameters, and a range of possible uses of the algorithm which has already proven itself in several practical applications to binding pocket optimization and prospective virtual ligand screening.
SourceAvailable from: Xiang Simon WangStructure 01/2014; 22:1120-1139.
[Show abstract] [Hide abstract]
ABSTRACT: Retrospective small-scale virtual screening (VS) based on benchmarking data sets has been widely used to estimate ligand enrichments of VS approaches in the prospective (i.e. real-world) efforts. However, the intrinsic differences of benchmarking sets to the real screening chemical libraries can cause biased assessment. Herein, we summarize the history of benchmarking methods as well as data sets and highlight three main types of biases found in benchmarking sets, i.e. "analogue bias", "artificial enrichment" and "false negative". In addition, we introduced our recent algorithm to build maximum-unbiased benchmarking sets applicable to both ligand-based and structure-based VS approaches, and its implementations to three important human histone deacetylase (HDAC) isoforms, i.e. HDAC1, HDAC6 and HDAC8. The Leave-One-Out Cross-Validation (LOO CV) demonstrates that the benchmarking sets built by our algorithm are maximum-unbiased in terms of property matching, ROC curves and AUCs. Copyright © 2014. Published by Elsevier Inc.Methods 12/2014; 71. DOI:10.1016/j.ymeth.2014.11.015
[Show abstract] [Hide abstract]
ABSTRACT: The aryl hydrocarbon receptor (AHR) is a ligand-activated transcription factor that regulates the expression of a diverse group of genes. Exogenous AHR ligands include the environmental contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), which is a potent agonist, and the synthetic AHR antagonist N-2-(1H-indol-3yl)ethyl)-9-isopropyl-2- (5-methylpyridin-3-yl)-9H-purin-6-amine (GNF351). As no experimentally determined structure of the ligand binding domain exists, homology models have been utilized for virtual ligand screening (VLS) to search for novel ligands. Here, we have developed an "agonist-optimized" homology model of the human AHR ligand binding domain, and this model aided in the discovery of two human AHR agonists by VLS. In addition, we performed molecular dynamics simulations of an agonist TCDD-bound and antagonist GNF351-bound version of this model in order to gain insights into the mechanics of the AHR ligand-binding pocket. These simulations identified residues 307-329 as a flexible segment of the AHR ligand pocket that adopts discrete conformations upon agonist or antagonist binding. This flexible segment of the AHR may act as a structural switch that determines the agonist or antagonist activity of a given AHR ligand.Biology 12/2014; 3(4):645-669. DOI:10.3390/biology3040645