节点文献
磷酸酶抑制剂设计的相关分子模拟理论与方法研究
Theoretical and Methodology Study on Computer-aided Design of Phosphatase Inhibitors
【作者】 武传杰;
【作者基本信息】 天津大学 , 应用化学, 2005, 博士
【摘要】 磷酸酶是细胞信号传导中的关键物质,已经成为癌症、糖尿病等疾病治疗的靶标。抑制剂活性位点处的α氟代效应是抑制剂的开发过程发现的一个重要现象,可以引起抑制剂活性提高三个数量级,其原因有待研究。本文对磷酸酶抑制剂的α氟代效应及其药物设计中的一些相关基本理论与方法作了研究。研究内容包括对α氟代效应的量子计算研究、基于全原子力场TEAM的全新蒙特卡罗程序设计、亨利常数的分子模拟计算研究和Widom方法程序设计及自由能计算中的可能影响因素分析、及磷酸酶与抑制剂结合能与结合自由能的分子动力学计算。研究结果发现:1.氟原子的引入主要是使得磷酸酶抑制剂的活性位点——磷酸基团上的电荷向与磷酸基团相联的原子上进行了分配,使磷酸基团的质子解离更为容易,从而改变了溶液中的离子比例,并大幅提高了抑制剂的活性。氟原子不是一个好的氢键受体。在计算离子过程时,考虑溶剂效应很重要。DPCM和IEFPCM可以比较好的用来计算溶剂化效应。2.基于TEAM全原子力场的蒙特卡罗程序设计与开发提供了一种可选择的分子模拟方法。程序可以基于TEAM力场进行正则系综下的蒙特卡罗计算。在程序中设计了四种基本的蒙特卡罗运动,分别是分子平动、分子整体刚性转动、原子随机小幅振动、主要二面角的扭动。四种运动依次按照Metropolis原理沿马尔可夫链进行。程序验证结果显示运行正常。3.Widom程序设计可以处理16种情况,它们包括:溶剂用分子动力学还是蒙特卡罗方法取样,取样系综是NVT系综还是NPT系综,粒子插入方法是格子法还是随机法,粒子插入后是否随机旋转。计算结果显示,Widom方法可以比较准确的计算气体在溶剂中的亨利常数。但计算过程中,组态空间的有效均匀取样问题仍然有待解决。使用联合原子力场在一定程度上可较好地避免这一问题。4.磷酸酶与其抑制剂的复合物的分子动力学模拟计算结果显示静电相互作用在酶与抑制剂间的相互作用中扮演了非常重要的角色。相对结合自由的计算结果显示,二氟代化合物的活性与酶的多肽底物的活性相近,远高于非氟代化合物。但是将计算结果与实验值直接定量比较还有些差距。因此,在开发此类抑制剂时,要更加注意引入原子的电荷效应,而非氢键效应。
【Abstract】 Phosphatase is a critical enzyme in regulating the signal transduction in cells. It has been proved to be a useful target for the therapy of cancer, diabetes and many other diseases.α-Fluorination effect existing in the active sites of the phosphatase inhibitor, which enhances the activity the inhibitors by 3 magnitudes, is an important phenomena and the underlying reasons are still undetermined. In this dissertation, we give an investigation to the theory and methodologies related to theα-fluorination effect and the drug design work. The content mainly covers four aspects, which are the ab initio calculation study on theα-fluorination effect, the program design of Monte Carlo (MC) method based on the TEAM all-atom force field, factors analysis influcing the efficiency of free energy perturbation method based on the molecular simulation calculation of Henry’s Law Constant (HLC) with Widom insertion method, binding mode and binding affinity calculation with molecular dynamics simulation (MD). Investigation results show that:1. The introduction of fluorine atoms on theα-position of the inhibitor active site mainly affects the charge distribution of the local sites. Fluorine is not a good hydrogen bond acceptor. Solvation effect calculation with DPCM and IEFPCM model was found to be good and important to the study.2. Four basic MC moves were designed for the MC program based on TEAM all-atom force field. They proceed successively along the Markov chain. The validation results tell that it works.3. Widom program was designed for 16 cases, which can be generated by combine the solvent sampling methods (MC or MD), the solvent ensemble (NVT or NPT), solute particle insertion method (random or grid insertion) and the rotation strategy (YES or NO). Calculation results give a strong hint that the free energy perturbation method can give a good results and enough attention should be paid to the sampling efficiency. United-atom force field is preferred by the sampling in some cases.4. The binding mode and binding affinity calculation prove that the electrostatic part of the molecular interaction energy play the key role when the inhibitor binding phosphatases. The difluorinated inhibitor posess the similar activity to the enzyme substrate, and is much more active than the non-fluorinated compound. Therefore, when developing this kind of inhibitors , we should pay more attention to the charge effect brought by the introduced atoms, but no hydrogen bonding effects.
【Key words】 phosphatase inhibitors; fluorination effect; drug design; ab intio; force field; free energy;