New QSAR prediction models derived from GPCR CB2-antagonistic triaryl bis-sulfone analogues by a combined molecular morphological and pharmacophoric approach

Abstract

In order to build quantitative structure–activity relationship (QSAR) models for virtual screening of novel cannabinoid CB2 ligands and hit ranking selections, a new QSAR algorithm has been developed for the cannabinoid ligands, triaryl bis-sulfones, using a combined molecular morphological and pharmacophoric search approach. Both pharmacophore features and shape complementarity were considered using a number of molecular descriptors, including Surflex–Sim similarity and Unity Query fit, in addition to other molecular properties such as molecular weight, ClogP, molecular volume, molecular area, molecular polar volume, molecular polar surface area and dipole moment. Subsequently, partial least squares regression analyses were carried out to derive QSAR models linking bioactivity and the descriptors mentioned, using a training set of 25 triaryl bis-sulfones. Good prediction capability was confirmed for the best QSAR model by evaluation against a test set of a further 20 triaryl bis-sulfones. The pharmacophore and molecular shape-based QSAR scoring function now established can be used to predict the biological properties of virtual hits or untested compounds obtained from ligand-based virtual screenings.

Keywords:

• cannabinoid (CB2) receptor
• triaryl bis-sulfone
• quantitative structure–activity relationship (QSAR)
• pharmacophore feature
• morphological descriptor
• virtual screening

Acknowledgements

This project was supported by grants from NIH NIDA R01DA025612. We are grateful to Tripos International for granting an evaluation license for Surflex–Sim for shape-based molecular similarity simulation. We also acknowledge the technical support kindly provided by Drs Tom Jones and Lei Wang at Tripos.