Ligand binding to a receptor is necessary but not sufficient for a functional response, and development of small molecule agonists in medicinal chemistry is often hampered by the fact that even minor modifications of a promising agonist scaffold lead to decrease or even loss of functional potency and intrinsic activity. Starting from the observation that in vitro binding affinities (pIC50) and functional potencies (pEC50) for a set of nicotinic α4β2 agonists were only weakly correlated, we developed a “Consensus 3D-QSAR Approach” to investigate if it is possible to combine the information from two GRID/GOLPE models, correlating molecular interaction fields (MIFs) with pIC50s and pEC50s respectively, to gain insight in the structural motifs which influence binding and functional behaviour at the α4β2 receptors. Using a 5-point fitting scheme,1 all possible conformations of the training set compounds within a 1-kcal/mol range from the global minimum were considered as templates. Two 3D-QSAR models were then built for each alignment, one based on binding affinities and the other on functional potencies; the models were subsequently scored according to their ability to predict both properties for an external validation set. The consistency of this scoring criterion can be appreciated by the fact that the highest ranking models are characterized by good, balanced performance with either set of dependent variables. The alignment which originated the best model (SDEPs of 0.30 and 0.68 for binding affinities and functional potencies, respectively) is supported by two predictive 3D-QSAR models whose dependent variables are only weakly correlated, and can thus be envisaged to be a reasonable guess of the bioactive conformation. The interpretation of the 3D PLS coefficient plots, with the aid of an in-house homology model of the α4β2 receptor, yielded valuable information with regard to the differences between a “good binder” and a “good agonist”. While confirming the importance of hydrogen bonding interactions between the pyridine nitrogen and a highly conserved water molecule, as well as between the protonated nitrogen of the alicyclic ring and the backbone carbonyl oxygen of Trp147 as far as binding affinity is concerned,1 an intriguing possible role emerged for the substituents at the pyridine ring, which seem to control the functional potency of the ligands at α4β2 receptors; also the size of the ring bearing the protonated nitrogen appears to modulate agonism.
Complementary 3D-QSAR modelling of binding affinity and functional potency: a study on α4β2 nicotinic ligands
TOSCO, Paolo;
2009-01-01
Abstract
Ligand binding to a receptor is necessary but not sufficient for a functional response, and development of small molecule agonists in medicinal chemistry is often hampered by the fact that even minor modifications of a promising agonist scaffold lead to decrease or even loss of functional potency and intrinsic activity. Starting from the observation that in vitro binding affinities (pIC50) and functional potencies (pEC50) for a set of nicotinic α4β2 agonists were only weakly correlated, we developed a “Consensus 3D-QSAR Approach” to investigate if it is possible to combine the information from two GRID/GOLPE models, correlating molecular interaction fields (MIFs) with pIC50s and pEC50s respectively, to gain insight in the structural motifs which influence binding and functional behaviour at the α4β2 receptors. Using a 5-point fitting scheme,1 all possible conformations of the training set compounds within a 1-kcal/mol range from the global minimum were considered as templates. Two 3D-QSAR models were then built for each alignment, one based on binding affinities and the other on functional potencies; the models were subsequently scored according to their ability to predict both properties for an external validation set. The consistency of this scoring criterion can be appreciated by the fact that the highest ranking models are characterized by good, balanced performance with either set of dependent variables. The alignment which originated the best model (SDEPs of 0.30 and 0.68 for binding affinities and functional potencies, respectively) is supported by two predictive 3D-QSAR models whose dependent variables are only weakly correlated, and can thus be envisaged to be a reasonable guess of the bioactive conformation. The interpretation of the 3D PLS coefficient plots, with the aid of an in-house homology model of the α4β2 receptor, yielded valuable information with regard to the differences between a “good binder” and a “good agonist”. While confirming the importance of hydrogen bonding interactions between the pyridine nitrogen and a highly conserved water molecule, as well as between the protonated nitrogen of the alicyclic ring and the backbone carbonyl oxygen of Trp147 as far as binding affinity is concerned,1 an intriguing possible role emerged for the substituents at the pyridine ring, which seem to control the functional potency of the ligands at α4β2 receptors; also the size of the ring bearing the protonated nitrogen appears to modulate agonism.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
