Robust and predictive 2D/3D-QSAR models for predicting the activities of novel pleuromutilin derivatives as antibacterial agents

Abstract

In recent years, antibiotics have been widely used and even caused abuse, many kinds of bacteria are becoming more and more resistant to existing antibiotics. Drug resistance has become a major problem to be solved, therefore, development of novel antibiotics is imperative. Studies have proven that novel pleuromutilin derivatives would provide powerful means to solve the problem of medical infection of drug-resistant pathogens, thus, it is noteworthy to in-depth study this series of derivatives. Therefore, the antibacterial activity of a series of novel pleuromutilin derivatives were investigated by 2D-QSAR and 3D-QSAR modeling. The CoMFA and CoMSIA methods were conducted to build up the 3D-QSAR models. Satisfactory statistical parameters (R2cv=0.608, R2pred=0.8596 for S. aureus-25923/CoMFA, R2cv=0.780, R2pred=0.8030 for S. aureus-25923/CoMSIA; R2cv=0.588, R2pred=0.6062 for MRSE-51625/CoMFA, R2cv=0.571, R2pred=0.5998 for MRSE-51625/CoMSIA; R2cv=0.604, R2pred=0.8061 for MRSA-337371/CoMFA, R2cv=0.737, R2pred=0.8241 for MRSA-3373715/CoMSIA; R2cv=0.704, R2pred=0.7504 for S. dysgalactiae/CoMFA, R2cv=0.683, R2pred=0.7852 for S. dysgalactiae/CoMSIA; R2cv=0.708, R2pred=0.7663 for S. agalactiae/CoMFA, R2cv=0.734, R2pred=0.6692 for S. agalactiae/CoMSIA) were obtained. To validate the results of 3D-QSAR models, MLR method was used to develop the 2D-QSAR models, which would provide structural information for the ligand modification. All the results would be of great help for the design of new antibacterial agents.