Abstract: The relationship between structures and antibacterial activities of 15 quinoline derivatives were studied by using both density functional theory (DFT) at the B3LYP/6-31G (d, p) basis set and stepwise multiple linear regression analysis (SMLR) method. The main independent factors which affected the activity of the compounds were determined by stepwise regression analysis, then a QSAR model was established. The stability and predictive ability of the model were examined by “leave-one-out”(LOO) cross-validation method. The result suggested that the nucleophilic frontier electron density of C(5) atom ( ), the bond order of C(9)-N(1) ( ) were the predominant factors affecting antibacterial activity against S. aureus. The fitting correlation coefficient (R2) and the cross-validation coefficient (q2) values were 0.814 and 0.734 respectively, showing this model with a good predictability. Based on this QSAR study, 4 new compounds with high antibacterial activity had been theoretically designed.

Key words: Quinoline, Antibacterial activity, S. aureus, Density functional theory, QSAR