3D descriptors calculation and conformational search to investigate potential bioactive conformations, with application in 3D-QSAR and virtual screening in drug design. Issue 13 (3rd October 2017)

• Record Type: Journal Article
• Authors: da Silva, Carlos Henrique Tomich de Paula; Taft, Carlton Anthony
• Is Part Of: Journal of biomolecular structure & dynamics. Volume 35:Issue 13(2017)
• Page Start: 2966
• Page End: 2974
• Publication Date: 2017-10-03
• Subjects: bioactive conformations -- 3D descriptors -- GRIND -- medicinal chemistry -- drug design; Biomolecules -- Periodicals; Molecular structure -- Periodicals; Molecular Biology -- Periodicals; Biomechanics -- Periodicals; 572
• DOI: 10.1080/07391102.2016.1237382
• Languages: English
• ISSNs: 0739-1102

Abstract

Abstract : The knowledge of the bioactive conformation for an active hit is relevant because of the easier interpretation and the general quality of the recognition models of protein and ligand. With the aim of investigating potential bioactive conformations without previous structural knowledge of the molecular target, we present herewith a 'protocol' that could be used which includes generation of low-energy conformations, calculations of tridimensional descriptors and investigation of structural similarity via principal component analysis. The protocol was used in the search for potential bioactive conformations. An initial selection of targets was made from a set of protein–ligand complexes with structure deposited in the Protein Data Bank, which was systematically filtered by lead-like rules, resulting in 45 ligands of 8 important therapeutic targets. After extensive optimization of the protocol and parameters of both OMEGA and Pentacle softwares, the best results were obtained for series of compounds such as the beta-trypsin and urokinase inhibitors, which are more structurally related among each other, inside the respective therapeutic class. Future improvements of the protocol, including a suitable choice and combination of robust 3D descriptors, could yield more reliable and less restrictive results, with general and diverse applications in drug design, in particular for improving the 3D-QSAR methodologies as well as virtual screening experiments for a more reliable selection of new lead compounds for different molecular targets.