【作者】 曾娟；

【导师】 张增辉； 梅晔；

【作者基本信息】 华东师范大学 ， 原子与分子物理， 2014， 博士

【摘要】 分子动力学模拟可以用来研究生物大分子体系的各种性质,如：预测受体和配体之间的亲和力,模拟蛋白质的折叠过程,研究蛋白质与细胞膜之间的相互作用、离子通道以及酶的催化反应等等,因此逐渐受到科学家的重视。分子动力学模拟理论将每个原子看成一个球,原子与原子之间通过弹簧连接,模拟体系在高维的势能函数面上运动。目前,存在多种不同的势能函数,如：AMBER, CHARMM,OPLS等,所有的势能函数都可划分成两大类：通过共价键相互作用的势能和非共价键作用的势能,共价键作用能包含键的伸缩能,角的弯曲能,二面角的扭转势能,非共价键作用能包含静电作用能和范德华作用能,势能函数中不同的参数形成不同的分子力场。静电作用是一种长程作用力,对动力学模拟的结果具有重大的影响,静电作用力的正确与否直接关系到氢键、盐桥等作用力的合理性,而大量研究证明氢键和盐桥对蛋白质折叠、配体和受体相互作用等都具有非常重要的意义。目前所有的力场都是采用库伦定律来计算静电作用能,分子理论假设每个原子上的电荷都集中在原子的中心,通过拟合分子周围的电势得到每个原子上的电荷。传统分子力场中,每个氨基酸上的原子电荷为一个定值,不受周围环境的影响。而实际上,生物大分子的一切运动都离不开水溶液,水分子为极性分子与生物大分子之间存在明显的极化效应,因此我们在分子动力学模拟的时候必须考虑生物大分子与周围环境之间的极化作用。经过长期的努力,目前已发展了多种极化电荷模型,主要有：Drude振子模型,涨落电荷模型和诱导多极模型。但是这些方法仍然存在一些无法避免的缺陷,如：Drude振子模型引入了人为的参数,诱导多极模型的计算量非常大,目前根本无法与分子动力学模拟结合,而且力场的可迁移性也很难实现。我们课题组完全放弃了力场的可迁移性,运用分子碎片共轭帽法,将量化计算的方法引入到生物大分子的电势计算过程中,在计算电势的同时考虑了周围环境与生物大分子之间的极化效应,最后对电势进行拟合得到生物大分子的极化电荷。借助RESP方法拟合的蛋白质内部原子的极化电荷随构象的变化波动幅度较大,非常不利于分子动力学模拟,特别是在QM/MM模拟等过程中,可能存在大构象的变化,因此大幅度的电荷涨落,对模拟非常的不利。基于2011年张等人选择参考电荷将偶极矩进行分割的方法,本文在RESP方法的基础上发展了一套在数值上更稳定的拟合极化电荷的方法：dRESP方法。dRESP方法中,每个原子上的电荷被划分成两部分：参考电荷和需要拟合的电荷。此方法的前提是参考电荷能够近似描述原子的电荷分布,所以参考电荷一般为MBER、CHARM、OPLS等力场中常用的电荷。量化计算的电势与参考电荷在相同格点上产生的电势之间的差量就是我们需要拟合的电荷在这个格点上产生的电势,拟合这个电势就能得到一个差值电荷。在拟合的过程中,我们在每个原子上所加的限制都是不同,与其参考电荷平方的倒数成正比,最后参考电荷与拟合得到的差值电荷之和就是整个原子的电荷。通过这种方式拟合电荷的优势是：拟合的电荷比原子电荷小了几个数量级,因此造成的数值误差比直接拟合原子的电荷要小很多。计算结果证明参考电荷的选取不影响拟合电荷的波动性,这种分步拟合电荷的方法的确很大程度上避免了拟合过程中出现的数值问题。Streptavidin与biotin之间的结合自由能高达-18.1kcal/mol,是目前发现的自然界中亲和力最强的化合物,对合理药性物设计具有很高的研究价值。大量研究证明氢键网络对biotin与streptavidin之间的结合非常重要,然后Baugh等人发现streptavidin末端的130号氨基酸由PHE突变成LEU之后,结合自由能减少了4.2kcal/mol,但是结构却没有明显的变化,突变前后A链上Ca原子的RMSD值仅为0.377A。我们同时运用AMBER电荷和dRESP方法拟合的极化电荷对streptavidin突变前后与biotin的复合物进行分子动力学模拟,并借助MM/PBSA方法预测原生态和130氨基酸突变体系的结合自由能。AMBER电荷的计算结果发现biotin与streptavidin之间的亲和力为正值,两者无法自由结合,同时计算的突变引起的能量差与实验值相比也过低。dRESP极化电荷加强了biotin与streptavidin之间的结合能,计算的绝对结合能和原生态与突变态之间的相对能量都与实验值一致。我们的计算证明了Baugh等人的猜测：静电极化效应是引起130号氨基酸突变能量变化的主要原因。CYP是一大类单氧酶,涉及到70-80%药物的前期代谢反应,因此对药物-药物相互作用具有重大的影响。CYP2J2是CYP酶家族的一类子族,主要表达在小肠,心血管等肝外器官中,大量研究证明CYP2J2能够催化大部分的药物,是药物前期代谢必不可少的一种酶,也可能会涉及到药物-药物相互作用,但是CYP2J2药物代谢的机理还不明了。分子对接是计算机辅助药物设计的主要手段,结合实验、分子对接、分子动力学模拟等技术,我们从69种已知的药物中筛选到两种CYP2J2有效的抑制剂：telmisartan、flunarizine、telmisartan和flunarizine对CYP2J2的半抑制剂浓度分别为0.42声μM和0.94μM,而对CYP家族其它成员的活性几乎没有影响,同时我们还证明这两种抑制剂不是CYP2J2的底物,是研究CYP2J2药物代谢机理最理想的工具。分子对接和动力学模拟进一步证实了实验的结论,并在原子结构的级别上展示telmisartan是非竞争性抑制剂,flunarizine是竞争性的抑制剂。更多还原

【Abstract】 Molecular dynamics (MD) simulation is one of the popular tools to study all kinds of properties of biological macromolecules, such as:predicting the binding affinity between proteins and ligands, studying the mechanism of protein folding、the interactions between proteins and membrane、the mechanism of ion channels and the catalytic reactions of enzyme, etc. Then the technology of MD simulations gradually grasp the attention of scientists. The main idea of MD simulation is based on a mathematical model of a molecule as a collection of balls corresponding to the atoms, the springs are used to connect atoms. The movement of molecules is based on the high-dimensional potential energy surface. At present, there are various potential energy functions, such as:AMBER, Charmm, opls and so on. All these potential energy functions are composed of two parts:bonded interactional energy and non-bonded interactional energy. Bonded interactional energy includes:bond stretch, angle bending, dihedral terms, non-bonded interactional energy includes:Electrostatic and Van der waals (VDW) interactions. The different parameters of potential energy contribute to different force fields. Electrostatic interactions is long-range interactional force and play an important role during MD simulations. The fallacious description of hydrogen bond and salt bridge interactions are main due to the incorrect of electrostatic interactions. While there are much results can prove that the hydrogen bond and salt bridge interactions are very important for studying the mechanism of protein folding and interactions between protein and ligand, etc. The molecular mechanics (MM) calculated the electrostatic interactions based on the theory of coulomb and every atom carries point charge centered on the atom. The point charge is fitted the electric potential around molecules. Classic force fields view point charge as an unchanged value which will not affected by surrounding environment. In fact, the biological macromolecules are surrounded by waters. Water molecules are polar molecules that can polarize the macromolecules, the polarized macromolecules also can polarize the water molecules, then cause the redistribution of charge. So we must consider the effect of polarization between macromolecules and solvent.Now there are various models can include polarization, such as:Drude oscillator, fluctuating charge model and induced dipoles model. But there are still some inevitable defects:Drude oscillator introduce a fictitious particle connecting with every atom, induced dipoles model costs too much computation times, so the combination between induced dipoles model and MD simulation is not reality, the transformation of force field also is very hard to be reached. Our group completely discard the transformation of force field and develop a method to consider the polarization of environment. With the help of molecular fragmentation with conjugate caps method, we can divide the protein into based-residue fragment, then the electric potential of each fragment is calculated with quantum mechanics and the effect of environment is include as the background charge, at last the protein-specific polarized charge (PPC) is fitted.The PPC of buried atoms fitted with RESP method have large fluctuation along with the change of conformations. The large fluctuation will do harm to MD simulations. Inspired by the idea of charge decomposition in calculation of the dipole preserving and polarization consistent charges, we have proposed a numerically stable restrained electrostatic potential (ESP)-based charge fitting method for protein:dRESP. The atomic charge is composed of two parts. The dominant part is fixed to a predefined value (e.g., AMBER, Charmm, OPLS charge), and the residual part is to be determined by restrained fitting to residual ESP on grid points around the molecule. Nonuniform weighting factors as a function of the dominant charge are assigned to the atoms. Because the residual part is several folds to several orders smaller than the dominant part, the impact of ill-conditioning is alleviated. The results testify that the fluctuation can be decreased regardless of the predefined charge, so the method avoid the numerical problems.The binding affinity between streptavidin and biotin is-18.1kcal/mol which is the largest binding affinity found in nature. This complex is invaluable for rational drug design. A large number of research results testify that the network of hydrogen bond between streptavidin and biotin (BTN) is main force driving BTN binding to streptavidin. Recently, Baugh et al. discovered that a distal point mutation (F130L) in streptavidin causes no distinct variation to the structure of the binding pocket but a4.2kcal/mol reduction in biotin binding affinity, the RMSD value calculating through superimposing the A subunits of the wild type (WT) and mutant (F1301) structure using Ca atoms of the subunit cores is0.377A. We carry out molecular dynamics simulations with AMBER and polarized dRESP charge, then apply MM/PBSA method to calculate the binding free energies of biotin to WT and F130L. The absolute binding affinities based on AMBER charge are repulsive, and the mutation induced binding loss is underestimated. When using the polarized protein-specific charge, the absolute binding affinities are significantly enhanced. In particular both the absolute and relative binding affinities are in line with the experimental measurements. This work verifies Baugh’s conjecture that electrostatic polarization effect plays an essential role in modulating the binding affinity of biotin to the streptavidin through F130L mutation。Cytochrome P450(CYP) is a large group of enzymes that catalyze the oxidation of bound substances. It will involve the phase I metabolism of70-80%drug, so may play a key role between drug and drug interaction. CYP2J2mainly expressed in extrahepatic tissues, including intestine and cardiovascular systems is a subfamily of CYP enzyme. There are a lot of researches can proved that CYPJ22can metabolize a wide range of structurally diverse drugs and play an indispensable role in first-pass metabolism and drug-drug interactions, but the general role of CYP2J2in drug metabolism is not yet fully understood. Molecular docking is main tools for computer-aided drug design, we apply the technology of experiment、molecular docking and MD simulations to screened69know marketed drugs for inhibition of CYP2J2. We discovered two drugs as potent and selective CYP2J2inhibitors:telmisartan and flunarizine. They have CYP2J2inhibition IC50values of0.42mM and0.94mM, respectively, which are at least10-fold more selective against all other major metabolizing CYPs; moreover, they are not substrates of CYP2J2. So telmisartan and flunarizine are ideal tools for study the metabolic mechanism of CYP2J2. Molecular docking and MD simulation consolidate the results of experiments. The detailed structures information of conformations unveil that telmisartan is non-competitive inhibitor, while flunarizine compete with substrate for same binding site.更多还原