【作者】 黄小强； 薛靖； 林敏； 朱玉山；

【Author】 Xiaoqiang Huang;Jing Xue;Min Lin;Yushan Zhu;Department of Chemical Engineering,Tsinghua University;

【机构】 清华大学化学工程系；

【摘要】 计算酶设计可用于针对没有天然酶催化的化学反应来快速开发新酶。它的一个步骤是将一定的催化活性位点模型匹配到合适的蛋白质骨架中。为提高残基和过渡态安装的准确度,我们提出复杂活性位点模型(图1)。其他匹配算法无法克服多残基匹配的组合爆炸问题,我们基于ProdaMatch来研究此问题。表1为ProdaMatch算例测试结果,天然匹配都排名第一,计算的和晶体结构的过渡态构象吻合(图2),且运行时间短。此外,我们从1491个蛋白质骨架中搜索适合安装p-硝基苯乙酸酯水解反应(图3)的催化活性位点模型的骨架,活性位点模型如图4。每个骨架平均匹配时间不超过3小时,共有80个骨架可以匹配,其中有四个天然酶可以水解p-硝基苯乙酸酯,分别是1f6w、1jkm、1qe3和3m83。在这四个骨架上,天然匹配都被搜索到,且计算的残基构象与晶体结构的构象高度一致(图5)。综上,ProdaMatch算法和复杂活性位点模型的配合使用非常适合于从大规模的骨架库中搜索目的骨架用于活性位点模型的安装。更多还原

【Abstract】 Enzymes are efficient green catalysts.Computational enzyme design is a method that could be used for rapidly developing new enzymes,meaningful both in theoretical research and in chemical industry.In this work,a matching algorithm named ProdaMatch combined with a comlex active site model was developed for catalytic sites selection.The algorithm is suitable for catalytic sites selection problem with a large number of catalytic residues.The algorithm was tested using ten benchmark test cases and it performed with high accuracy and speed.Besides,ProdaMatch algorithm was also used to search scaffolds that can accommodate the complex active site model proposed for hydrolyzing p-nitrophenyl acetate from a scaffold library containing 1491 protein structures.Eighty scaffolds were identified and four of the native scaffolds were reported to hydrolyzing p-nitrophenyl acetate.The efficient algorithm can be of great significance in catalytic sites selection in computational enzyme design.更多还原