微生物生物合成中常用跨物种酶来重构特定代谢途径并采用定向进化策略获得大量突变体，巨量的实验数据迫使蛋白质分子设计技术亟待突破：1)如何在分子水平上快速整合高通量实验平台所获得海量突变体的酶催化效率实验数据？特别是如何理解远离酶催化中心的氨基酸残基突变效应？2) 如何突破人工进化技术的实验瓶颈，开展全局突变甚至多点组合突变推进酶分子的高效优化？3) 能否在人工进化的基础上，从反应机制上拓新底物谱，从头设计全新的酶催化反应？本项目拟利用申请人课题组前期开发的PRS预反应态模型，运用量子力学、分子动力学和QM/MM等多层次计算化学方法系统深入地阐明聚酮合酶中硫酯酶的全局突变效应和底物多样性问题，特别是在电子结构、原子和分子的微观水平上重现和突破匹马霉素硫酯酶的底物识别机制，以期在此基础上实现羧基甲基化新代谢产物的高效生物合成。

In microbial biosynthesis, cross-species enzymes are commonly used to reconstruct specific metabolic pathways and then adopt directed evolution strategy to obtain a large number of mutants. The huge amount of experimental data pushes protein molecular design techniques to a new frontier requirement: 1) How to quickly integrate catalytic efficiency of massive mutants obtained from the high-throughput platform? In particular, how to understand mutation effects of amino acid residues far away from its catalytic site? 2) How to break through the experimental bottleneck of artificial evolution, and carry out a global mutation or even multi-points combinated mutation to further promote enzyme efficiency? 3) Can we develop a brand-new substrate based on catalytic mechanism with big data from artificial evolution and de novo design an enzyme for a unnatural chemical reaction? This project intends to employ the pre-reaction state (PRS) model that was developed by our group, using the multi-level computational chemistry method such as quantum mechanics, molecular dynamics and QM/MM to clarify the global mutation effect and substrate diversity of PKS thioesterase. In particular, we will illustrate molecular recognition mechanism of pimaricin thioesterase in polyketide synthase at the microscopic level of electronic, atomic and molecular levels. On this basis, more efficient biosynthesis will be realized for new drugable metabolites such as decarboxylated methyl compounds.