计算机辅助的头孢菌素C酰化酶的从头设计

In this project, the computer-aided de novo design methodology for enzyme catalyst will be developed based on superstructure modeling and systems optimization, and this method will be validated using the creation of a novel cephalosporin C (CPC) acylase which catalyzes the hydrolytic reaction of CPC to produce the important pharmaceutical intermediate 7-aminocephalosporanic acid (7-ACA). The key research contents include: (i) Develop protein scaffold selection algorithm and build scaffold library for target reaction based on three-dimensional database of proteins, i.e., PDB, and the size, shape, and simplified catalytic constraints of reaction transition state; (ii) Develop improved combinatorial optimization matching algorithm to anchor catalytic residues on selected scaffolds, the novel algorithm will consider no only the pre-organized catalytic residues but also the re-organization of the catalytic residues during different steps of the reaction; (iii) Develop improved algorithm for amino acid sequence selection at binding sites which will identify the Pareto solution of the multi-objective optimization problem for sequence selection where the two minimization targets are the folding energy of protein and binding energy between active site and small molecule transition state. Finally the scaffold library for de novo design of CPC acylase will be built and the artificial enzymes endowed with activity of catalyzing the hydrolysis of CPC will be used to validate the proposed computational strategy, and this integrated way of combining experiments and modeling will help us to investigate the physio-chemical rules behind the enzyme catalysis and enhance our enzyme design capability towards arbitrary target reaction so as to create great opportunities to develop green processes to replace the chemical processes with high emission.

本课题将以超结构建模及大系统优化为基础建立计算机辅助的酶分子从头设计计算策略，并以催化CPC水解制备7－ACA的反应过程为例设计CPC酰化酶进行实验验证。本课题在具体的建模过程中将重点研究（i）基于蛋白质三维数据库及目的反应过渡态小分子的大小、形状及简化催化约束开发骨架筛选算法及建立针对目的反应的骨架库；（ii）改进蛋白骨架上的催化残基匹配算法，匹配过程中不仅考虑催化残基的预组织还要考虑不同反应步骤之间催化残基的再组织约束；（iii）改进结合位点处的氨基酸序列选择优化算法，建立求解考虑蛋白折叠自由能最小及酶活性位点与小分子过渡态之间结合能最小的多目标优化模型。本课题将针对CPC水解反应建立蛋白骨架库及从头设计能够催化CPC水解的人工酶并进行实验验证，通过实验结果与模型计算相结合的方法探索酶分子从头设计的物理化学规律，从而为开发能够催化任意目的反应的人工酶打下理论基础。

本课题的背景是基于计算酶设计的方法创造人工酶催化非天然底物的反应从而开发制造精细化学品及药物中间体等的环境友好的绿色工艺。本课题以头孢类抗生素药物的酶法制备为具体的研究案例开发了一系列针对非天然反应的头孢菌素酰化酶。具体的研究内容包括：（i）开发了能够扫描结构数据库的匹配自组织酶活性位点模型的优化算法，该算法可以在天然骨架上将反应过渡态安装到<2Å的精度，从而为在惰性骨架上设计针对非天然反应的活性位点打下基础；（ii）针对头孢拉定的合成反应在蛋白质结构数据库中找到了能够催化头孢拉定水解及合成的新骨架，并设计了催化效率提高10倍的突变体；（iii）针对头孢菌素C水解制备7-氨基头孢烷酸的反应，在头孢菌素酰化酶N176的基础上设计了催化活性提高的突变体，在头孢菌素酰化酶AcyII的基础上设计了热稳定性提高的突变体；（iv）针对头孢拉定的合成反应，在青霉素酰化酶的基础上设计了动力学选择性提高10倍的突变体，并以这个突变体为基础开发了头孢拉定合成的酶法工艺，有望在工业上第一次实现头孢拉定的酶法合成。本课题开发的计算酶设计的研究方法可以用来设计催化头孢拉定以外的其它头孢类抗生素药物及非天然底物的酶催化剂，以此开发制备此类高附加值化学品及药物中间体的绿色工艺来取代传统的化学工艺，从而大大减少化工过程不可生物降解废物的排放，为解决我国目前面临的严重的环境危机带来机遇。