摘要
针对56个环氧酮肽衍生物,分别采用比较分子场分析(comparative molecular field analysis,CoMFA)、比较分子相似性形状指数分析(comparative molecular similarity indices analysis,CoMSIA)、Topomer CoMFA、Holo-gram QSAR(HQSAR)以及基于一维和二维描述符的支持向量机(support vector machine,SVM)方法进行了细致的构效关系研究。研究显示:通过引入一维和二维描述符的SVM建模方法,避免了柔性分子在三维构效关系研究中的构象选择和叠合难题,亦可有效避免过拟合现象的发生。所建最优SVM模型的决定系数R2、均方根误差(RMS)、交互验证系数Q2和外部预测R2pred分别为0.681,0.436,0.572和0.641。分析结果显示:电性、拓扑特征、疏水性和分子体积是影响环氧酮肽蛋白酶体抑制活性的主要因素。在此基础上,以活性最高样本分子(CID:42638286的)为模板,基于相似性评价方法对其侧链进行设计,结合Lipinski"5规则"类药性筛选,共得到12个新颖目标分子,且预测活性均达到纳摩尔水平。
As oral proteasome inhibitors,epoxyketone peptide inhibitors showed strong inhibitory activities and specificities against proteasome, and had become one of the hotspots in anti-cancer drug research recently. In this paper, the comparative molecular field analysis （ CoMFA ）, comparative molecular similarity indices analysis （ CoMSIA ）, topomer CoMFA, the hologram QSAR （ HQSAR ） and SVM-based 2-D QSAR method were used to explore the relationships between structures and activities of 56 epoxyketone pep- tide inhibitors. The results showed that the best model was derived from SVM-based 2-D QSAR method, which avoided conformer a- lignments and overfitting problems. The coefficients of determination（ R2） , root mean squares（ RMS ）, cross-validated determination coefficients（Q2） and determination coefficients of external prediction （R2pred ）of the best SVM model are 0. 681,0. 436,0. 572, and 0. 641, respectively. The results showed that the electrical properties, topological properties, molecular volumes and hydrophobicities were the main factors affecting activities of epoxyketone peptide inhibitors. Furthermore ,the side chains of epoxyketone peptide were optimized and evaluated by molecular similarities to a template（ CID：42638286）with the highest activity. Screened by the Lipinski ＂s rule of five, 12 novel molecules were obtained, and the activities of which reached nanomolar levels.
出处
《化学研究与应用》
CAS
CSCD
北大核心
2012年第9期1376-1388,共13页
Chemical Research and Application
基金
重庆市自然科学基金重点项目(CSTC,2009BA5068)资助
中央高校基本科研业务费科研专项项目(CDJXS,11231177)资助
