QSAR and pharmacophore approach on substituted imidazole derivatives as angiotensin II receptor antagonists

Abstract

A QSAR analyses of 32 aminomethyl and acylaminomethyl substituents derivatives were carried out to interpret the relationship between structural properties and angiotensin II AT1 receptor activity. Two-dimensional (2D-QSAR), Group-based (G-QSAR), 3D-QSAR, and pharmacophore mapping studies were performed using partial least square and k-nearest neighbor methodology coupled with various feature selection methods, viz. stepwise, genetic algorithm, and simulated annealing (SA) to derive QSAR models which were further validated for statistical significance and predictive ability by internal and external validation. The activity contributions of whole compounds and their substituents were determined from regression equation. The statistically significant best 2D-QSAR model 1 having r 2 = 0.8754 and q 2 = 0.7231 with $$ {\text{pred}}\_r^{ 2} = \, 0. 8 3 8 9 $$ was developed by GA-PLS with the descriptors like SdssCcount, SssNHcount and SaaaCcount. The best Group-based QSAR model-5 having r 2 = 0.7516 and q 2 = 0.6714 with $$ {\text{pred}}\_r^{ 2} = \, 0. 7 3 0 9 $$ was developed by GA-PLS. The 2D-QSAR results agreed with those from the Group based QSAR model and suggested the structural features of SssOE-index, Mol.Wt and SssNHE-index functionality that would enhance the antihypertensive activity of the imidazole derivatives in this series. The results of two-dimensional, Group-based QSAR showed that a combination of topological indices, hydrophobic properties, and auto-correlation descriptors of different atomic properties could be explored to design potent antihypertensive agents. The k-nearest neighbor approach was used to generate three-dimensional quantitative structure–activity relationship (3D-QSAR) models for these sets of molecules. The present study is an attempt in this direction seeking for the development and comparison of QSAR models of substituted imidazoles by different feature selection methods, which ultimately establishes the superiority of the GA-based models. Statistically genetic algorithm (GA) selection k-nearest neighbor (GA-kNN-MFA) model with respect to both the internal (q 2 = 0.7856) as well as external (pred_r 2 = 0.8193) model validation and correctly predicts the activity of calculated 78.56 % and calculated 81.93 % for the training and test set, respectively. Continuing with the series of substituted imidazoles derivatives, chemical feature-based pharmacophore models with lowest RMSD value (0.1397 A), consisting of one aromatic carbon center, two hydrogen bond acceptor, and one hydrogen bond donor features, was developed. The information rendered by 2D-QSAR, Group-based QSAR, k-nearest neighbor, and pharmacophore identification models may lead to a better understanding of structural requirements of antihypertensive agents and can help in the design of novel potent molecules.