Abstract
Quantitative structure-activity relationships for quinolone antibacterials have been previously examined and a steric parameter L for the N1-substituents found to be important in QSAR equations. But some compounds for which previous QSAR equations could not predict the activity have appeared recently. In this study, conformations of a variety of N1-substituents of quinolone antibacterials were analyzed by a molecular modeling method. An active conformation of each of the compounds was estimated with information of the energy profile calculated by molecular orbital methods and of its biological activity. A model of a receptor corresponding to the N1-substituents was constructed by superposing van der Waals volumes of active conformer of highly active compounds. As a result of these conformational analyses and receptor mapping, it is proposed that there are two different optimum volumes for increasing the activity of quinolone antibacterials and two unfavorable regions for reducing the activity. It is suggested that the steric parameter L which appeared in previous QSAR equations corresponds to one of the optimum volumes of the proposed receptor model. With this receptor model, a relation between structure and activity of the compounds, including those mispredicted compounds in previous QSAR equations is able to be rationalized qualitatively and elegantly. We believe that this receptor model is useful for a prediction of the activity of compounds not yet synthesized as well as for designing new quinolone antibacterials.