QSAR and proteo-chemometric analysis of the interaction of a series of organic compounds with melanocortin receptor subtypes.

We have created quantitative structure-activity relationship (QSAR) models describing the interaction of a series of 54 organic compounds with four melanocortin (MC) receptor subtypes, MC(1), MC(3), MC(4), and MC(5). In addition to traditional QSAR analysis, we applied our recently developed proteo-chemometrics approach. Proteo-chemometrics is based on the combined analysis of series of receptors and ligands, wherein descriptions of ligands, proteins, and so-called ligand-protein cross-terms are correlated with interaction activities. The compounds were characterized by structural descriptors, including three-dimensional grid-independent descriptors (GRINDs), topological descriptors, and geometrical descriptors. Description of receptors was obtained by computing the receptors' amino acid sequence identities. Both the QSAR and proteo-chemometrics approaches resulted in models with essentially the same statistical significance: the cross-validated correlation coefficient q(2) for the proteo-chemometric model being 0.71, while for the QSAR models the q(2)s were 0.75, 0.68, 0.63, and 0.71 for the MC(1), MC(3)(-)(5) receptor, respectively. However, the proteo-chemometrics modeling provided more detailed information about receptor-ligand interactions and determinants for receptor subtype selectivity than did QSAR.