Accurate Activity Predictions of B-Raf Type Ii Inhibitors Via Molecular Docking and Qsar Methods
Liu Hai-Chun,Lu Shuai,Ran Ting,Zhang Yan-Min,Xu Jin-Xing,Xiong Xiao,Xu An-Yang,Lu Tao,Chen Ya-Dong
DOI: https://doi.org/10.3866/pku.whxb201510134
2015-01-01
Acta Physico-Chimica Sinica
Abstract:B-Raf kinase plays an important role in the mitogen-activated protein kinase (MAPK) signaling transmission pathway and has been identified as an attractive target for cancer therapy. The exploitation of novel and efficient B-Raf inhibitors has become a hot research topic. In this study, we investigated quantitative structure activity relationship (QSAR) to probe the origins of the inhibitory activities of B-Raf Type II inhibitors. We used structurally diverse B-Raf Type II, inhibitors and an integrated docking and QSAR extended method. We focused mainly on two themes: bioactive conformations and descriptors. First, various molecular docking methods (Glide, Gold, LigandFit, Cdocker, and Libdock) were evaluated, and then all molecules were docked into the B-Raf active site to obtain the bioactive conformations. Secondly, based on the docking results, 16 scoring functions and 21 docking-generated energy-based descriptors were calculated to construct regression models. The results gave highly accurate fitting and had strong predictive abilities (M1: r(2) = 0.852, r((CV))(2) = 0.790, r(pre)(2) = 0.864; M2: r(2) = 0.738, r((CV))(2) = 0.812, r(pre)(2) = 0.8605). The important descriptors were also explored to elucidate the main factors influencing the inhibition activities. The models suggested that the scoring functions (G_Score, -ECD, Dock_Score, and PMF) and docking-generated energy-based descriptors (S(hb_ext), DE(int), and Emodel) were significant. Some new compounds that are potential B-Raf inhibitors were obtained through virtual screening and theoretical predictions using the established models. Such information is useful in guiding the design of novel and robust B-Raf Type II inhibitors.