基于药效团的BRD4靶向小分子抑制剂虚拟筛选

Virtual screening of small molecule targeted inhibitors of BRD4 based on pharmacophore

BRD4靶点和多种肿瘤密切相关,是具有良好成药性的热门靶点。本文选取活性较好且结构差异较大的BRD4小分子抑制剂作为训练集分子,基于配体小分子共同特征(HipHop)方法使用Discovery Studio 3.0分子模拟软件构建了药效团。药效团通过测试集验证、ROC曲线验证(SE(sensitivity)=0.93765、SP(specificity)=0.89500、(AUC)=0.956),结果表明构建得到的药效团具有较强的可靠性和较高的可信度。药效团模型含有1个芳环中心、1个疏水基团、2个氢键受体四个药效特征元素。此药效团被用于ZINC数据库进行虚拟筛选,共筛选了861203个分子,命中率为0.782%。再对筛选得到的分子经过分子对接、ADMET成药性预测、构象分析并讨论分子-蛋白相互作用模式,最终得到了21个有潜力的BRD4小分子抑制剂。

The BRD4 target is closely related to a variety of tumors and is a popular target with good druggability.In this paper,BRD4 small molecule inhibitors with better activity and larger structural differences were selected as training set molecules,and the pharmacophore was constructed using Discovery Studio 3.0 molecular simulation software based on the common feature of ligand small molecules(HipHop)method.The pharmacophore passed the test set verification and ROC curve verification(SE(sensitivity)=0.93765,SP(specificity)=0.89500,(AUC)=0.956),the results showed that the constructed pharmacophore had strong reliability and high credibility.The pharmacophore model contained four pharmacodynamic characteristic elements:an aromatic ring center,a hydrophobic group and two hydrogen bond receptors.This pharmacophore was used for virtual screening of 861203 molecules in the ZINC database,with a hit rate of 0.782%.After molecular docking,prediction of ADMET druggability,conformation analysis and discussion of molecular-protein interaction mode on the selected molecules,21 potential BRD4 small molecule inhibitors were finally obtained.