Quantitative Structure-Activity Relationship and Virtual Screening of ω-Conotoxins

Abstract

ω-Conotoxins are active peptides composed of 24-31 amino acids isolated from venomous marine predatory cone snails. ω-Conotoxins selectively inhibit voltage-gated calcium channels (VGCCs) in nociceptors, so are considered attractive molecules for drug design. In this study, based on a set of new amino acid structure descriptors (c-scales) and genetic partial least squares (G/PLS) regression method, quantitative structure- activity relationship (QSAR) models for N-type and P/Q-type VGCC antagonists of ω-conotoxins were developed. Two virtual polypeptide libraries with 2244 peptides were designed and established for N-type and P/Q-type VGCC antagonists, respectively. Then, based on the biological activities predicted from the constructed QSAR models and chemical similarities to the probes MVIIAand MVIIC, the polypeptide libraries were virtually screened. As a result, the established QSAR models had good predictability (cross-validated correlation coefficient CV-r 2 >0.89). The structural diversity of the libraries was validated using principal component analysis (PCA) and hierarchical cluster analysis (HCA) approaches. Six N-type and nineteen P/Q-type VGCC antagonists with high selectivity and activity were identified by virtual screening. The results of this study will be valuable for finding highly active polypeptide and non-peptide mimetics. Furthermore, the established polypeptide QSAR models and virtual screening strategy can also be applied to other peptide systems.