目的 构建磷酸二酯酶4(PDE4)抑制剂药效团模型,为中草药的相关虚拟筛选研究提供新的方法。 方法 以22个具有PDE4抑制活性的化合物作为训练集化合物,其半数抑制浓度(IC50)范围在0.042~23000nmol·L^-1。利用Catalyst/HypoGen系统,对训练集化合物进行构象分析,通过对训练集化合物多个构象进行叠合,提取药效团特征及三维空间限制构建药效团模型。结合数据库搜索命中率筛选最优药效团。结果 最优药效团模型包含2个氢键受体、1个脂性疏水基团和1个芳香环特征,排除体积数为6个。模型相关系数、Totalcost值、Fixedcost值、Nullcost值、△cost值和Configuration cost值分别为0.9047、117.7、90.84、180.4、62.7和15.71。利用所得较优模型对MDL药物数据报道数据库进行搜索,该模型的命中已知活性化合物数与命中已知活性化合物总数比值为58.95%。 结论 利用Catalyst系统构建的PDE4抑制剂药效团具有较好的预测能力,可作为提问结构用于数据库的搜索,有助于中草药中具有抑制PDE4活性的化合物的虚拟筛选研究。
Objective To generate the pharmacophore model of PDE4 inhibitors and thus to guide the investigation on virtual screening of traditional Chinese medicines. Methods Based on the training set composed of 22 PDE4 inhibitors, pharmacophore models were generated by HypoGen program of the Catalyst software. The IC50 of the inhibitors varied from 0.042 to 23000 nmol·L-1. The pharmacophore models were a set of 3-D pharmacophore features, which were constructed by the generation of conformational models and alignments of conformations based on the training set. By screening the highest hit% of the models, the best one was confirmed. Results The best pharmacophore model consisted of two Hydrogen-bond acceptors, one Hydrophobic aliphatic region, one Ring aromatic feature, and six excluded volumes. Its correlation coefficient, and total, fixed, null, A, and configuration costs were 0.9047, 117.7, 90.84, 180.4, 62.7, and 15.71, respectively. According to the database screening, the radio of hits to active compounds from the MDL Drug Data Report Database was 58.95%. Conclusions The pharmacophore model of PDE4 inhibitors has a high predictive ability, and thus can be used as a query for database screening. By using such a model, we may find new PDE4 inhibitors in traditional Chinese medicine for developing traditional Chinese medicines and material medica.