C D Klein1, A J Hopfinger. 1. Laboratory of Molecular Modeling and Design, M/C 781, College of Pharmacy, University of Illinois at Chicago 60612-7231, USA.
Abstract
PURPOSE: This study was done to explore the relationships of both macroscopic and molecular level physicochemical properties to in-vivo antiarrhythmic activity and interactions with phospholipid membranes for a set of cationic-amphiphilic analogs. METHODS: The 4D-QSAR method, recently developed by Hopfinger and co-workers (1), was employed to establish 3D-QSAR/QSPR models. Molecular dynamics simulations provided the set of conformational ensembles which were analyzed using partial least squares regression in combination with the Genetic Function Approximation algorithm to construct QSAR and QSPR models. RESULTS: Significant QSAR models for in-vivo antiarrhythmic activity were constructed in which logP (the partition coefficient), and specific grid cell occupancy (spatial) descriptors are the main activity correlates. LogP is the most significant QSAR descriptor. 4D-QSPR models were also developed for two analog-membrane interaction properties, the change in a membrane transition temperature and the ability of the analogs to displace adsorbed Ca(2+)-ions from phosphatidylserine monolayers. CONCLUSIONS: Spatial features, represented by grid cell occupancy descriptors, supplement partition coefficient, which is the most important determinant of in-vivo antiarrhythmic activity, to provide a comprehensive model for drug action. The QSPR models are less significant in statistical measures, and limited to interpretation of possible molecular mechanisms of action.
PURPOSE: This study was done to explore the relationships of both macroscopic and molecular level physicochemical properties to in-vivo antiarrhythmic activity and interactions with phospholipid membranes for a set of cationic-amphiphilic analogs. METHODS: The 4D-QSAR method, recently developed by Hopfinger and co-workers (1), was employed to establish 3D-QSAR/QSPR models. Molecular dynamics simulations provided the set of conformational ensembles which were analyzed using partial least squares regression in combination with the Genetic Function Approximation algorithm to construct QSAR and QSPR models. RESULTS: Significant QSAR models for in-vivo antiarrhythmic activity were constructed in which logP (the partition coefficient), and specific grid cell occupancy (spatial) descriptors are the main activity correlates. LogP is the most significant QSAR descriptor. 4D-QSPR models were also developed for two analog-membrane interaction properties, the change in a membrane transition temperature and the ability of the analogs to displace adsorbed Ca(2+)-ions from phosphatidylserine monolayers. CONCLUSIONS: Spatial features, represented by grid cell occupancy descriptors, supplement partition coefficient, which is the most important determinant of in-vivo antiarrhythmic activity, to provide a comprehensive model for drug action. The QSPR models are less significant in statistical measures, and limited to interpretation of possible molecular mechanisms of action.