中文摘要：

【目的】构建I型人类免疫缺陷病毒（human immunodeficiency virus type 1,HIV-1）非核苷类逆转录酶抑制剂（nonnucleoside reverse transcriptase inhibitors,NNRTIs）的药效团模型,通过对中药化学数据库（traditional Chinese medicine database,TCMD）的搜索,在中药中寻找新型抗耐药的NNRTIs。【方法】从已知的NNRTIs与逆转录酶复合物的晶体结构出发,通过构象分析和药效团识别等方法,构建NNRTIs的药效团模型并检验其可靠性;基于药效团模型对TCMD进行数据库搜索,发现新型潜在的NNRTIs。【结果】从PDB中检索出2010年至2014年RT与NNRTIs复合物的晶体结构30个,从中抽提出其活性配体,建立了一个包含30个活性配体的小分子数据库;通过对上市药物TMC278和TMC125及高活性抑制剂DJZ的构象叠合和药效特征基团分析构建了新的NNRTIs药效团模型,该模型包含了5个药效特征基团;以符合这5个药效特征基团中任意4个为条件,在活性配体小分子数据库中验证性搜索出18个化合物,检出率为60.0%;基于五点药效团模型对TCMD进行数据库搜索得到272个化合物。【结论】以TMC278、TMC125和DJZ构建的五点药效团模型,以符合其中任意4个为条件,在活性配体小分子数据库中的检出率达到60.0%,表明所建药效团模型是可靠的。

英文摘要：

【Objective】To construct a pharmacophore model of HIV-1 non-nucleoside reverse transcriptase inhibitors（NNRTIs） and search new anti-resistance NNRTIs among traditional Chinese medicine from Traditional Chinese Medicine Database（TCMD）.【Methods】Based on the known co-crystallized structures of HIV-1 reverse transcriptase（RT） and NNRTIs, the pharmacophore model of NNRTIs was constructed via conformation analysis and pharmacophore identification, and its reliability was tested. Then new types of potential NNRTIs were found via searching TCMD.【Results】A total of 30 co-crystallized structures of HIV-1 RT and NNRTIs from 2010 to 2014 were found via searching PDB. A small-molecule database containing 30 active extracted ligands was built. A new pharmacophore model of NNRTIs was constructed via analyzing conformation congruence and characteristic pharmacophore on marketed drugs TMC278 and TMC125 as well as high activity inhibitor DJZ. This model contained 5 characteristic pharmacophores. A total of 18 compounds were found from active ligand database for validation which matched 4 of the 5 features and the detection rate was 60.0%. There were 272 compounds found from TCMD based on 5 characteristic pharmacophores.【Conclusions】The pharmacophore model based on TMC278, TMC125 and DJZ was constructed.When matched 4 of the 5 features, the detection rate of searching active ligand database was 60.0%. It indicates that the constructed pharmacophore model is reliable.