【作者】 庄树林；

【导师】 俞庆森；

【作者基本信息】 浙江大学 ， 化学生物学， 2006， 博士

【摘要】 巨噬细胞移动抑制因子(macrophage migration inhibitory factor，MIF)，是一种重要的先天性免疫系统的调节因子，在宿主抗微生物防卫系统和应激反应中起重要作用，涉及败血症、肿瘤炎症和自身免疫性疾病等多种疾病的发病机制，并且具有酶的活性，能够催化苯丙酮酸、D-多巴色素的异构化反应以及硫醇蛋白的氧化还原反应。随着MIF的特殊生物学功能及其在多种疾病中的重要作用日益为人们所确认，MIF已逐渐成为新的研究热点。但目前关于MIF的研究主要采用实验手段在组织细胞水平进行，较少从原子水平上进一步研究MIF的结构、酶催化机理以及与抑制剂的相互作用。不同类型抑制剂与MIF具体的结合模式，同类抑制剂活性的差异，以及MIF突变体与这些分子作用模式的差异都没有从原子水平得到合适的解释。MIF与生物活性分子的结合过程中MIF构象的动态变化，MIF重要氨基酸在催化反应中的重要角色及其在MIF与不同底物、抑制剂结合过程中的作用需要进一步探讨。基于以上问题，在实验研究基础上从理论计算角度对一些实验无法解释的现象从原子水平上深入分析，进一步探讨MIF酶的催化机理，了解MIF的重要氨基酸在催化中的作用，研究抑制剂、底物等不同小分子与MIF及其突变体的特异性识别与结合亲和性，动态研究MIF在与生物活性分子的结合过程中的构象变化，比较MIF及其突变体与生物活性分子作用的差异。本论文主要对MIF与8个对羟基苯丙烯酸类似物、5个底物分子、2个ISO-1类似物、5个香豆素衍生物分子，以及MIF蛋白突变体P1G、K32A、N97A等与8个对羟基苯丙烯酸类似物的相互作用，利用分子对接和分子动力学模拟等计算工具进行了详细的分析。1．针对MIF对(E)-2-氟-对羟基苯丙烯酸类似物的立体化学选择性，对MIF与6个苯丙烯酸类似物的复合物结构分别进行了分子动力学模拟，并采用MM-PBSA分析了复合物的结合自由能。MIF蛋白构象非常稳定，与苯丙烯酸类似物的结合并没有对其构象产生较大影响。MIF主要通过Asn-97’、Pro-1、Lys-32与苯丙烯酸类似物形成氢键或盐桥。顺反类似物与MIF结合模式的最大区别在于，反式类似物与残基Pro-1形成一个N-H…O氢键，而顺式类似物无法与Pro-1形成任何氢键，这是造成顺反类似物亲和性差别的最主要原因，并且MM-PBSA分析也表明静电作用是造成这种结合亲和性差别的重要原因。2．为研究MIF、MIF突变体与丙烯酸类似物相互作用的差别，以及重要残基突变对蛋白构象的影响，利用分子对接方法构建了苯丙烯酸类似物与MIF、P1G、K32A、N97A的复合物结构，并对这些复合物进行了分子动力学模拟。氢键分析发现反式苯丙烯酸类似物与MIF、P1G的结合模式存在很大区别，反式类似物与MIF通过残基Pro-1形成一个N-H…O氢键，而与P1G则通过残基Ile-64一个N-H…O氢键。顺式类似物与MIF和P1G的结合模式没有明显区别，并没有与Pro-1或Ile-64形成氢键。N97A突变体都通过残基Ile-64与顺式、反式类似物形成N-H…0氢键，并与反式类似物通过残基Lys-32形成一个盐桥，而与顺式类似物不形成此盐桥。顺式类似物仅与K32A的Asn-97’残基形成一个O-H…O氢键，而反式类似物与K32A突变体的Asn-97’形成一个O-H…O氢键，与Ile-64形成一个N-H…O氢键。计算结合自由能的差别定性反映了顺反类似物对P1G、K32A、NP7A突变蛋白可能的抑制活性的差异，预示着P1G、K32A、NP7A可能对苯丙烯酸类似物仍然存在一定的选择性。对MIF蛋白三聚体、单体、N97A、K32A突变体结构也进行了动力学模拟，构象并没有太大变化，在纳秒时间尺度上呈现为刚性，显示重要残基的突变对MIF蛋白的构象影响较小。3．为分析对羟基苯丙酮酸(HPP)、R-多巴色素(RDP)、R-和S-多巴色素甲酯(RDPM，SDPM)等底物与MIF结合的差异性，通过分子对接方法构建了MIF与这些底物的复合物结构，并进行了多重动力学模拟和MM-PBSA结合自由能分析。HPP主要与MIF的残基Asn-97’和Ile-64分别形成O-H…O氢键，N-H…O氢键，并且与残基LVs-32形成盐桥。RDP与Pro-1和Ile-64分别形成一个N-H…O氢键。SDPM与Pro-1和Asn-97’分别形成一个N-H…N氢键，O-H…O氢键，而RDPM仅与Pro-1形成一个稳定性不强的N-H…O氢键。MM-PBSA分析表明底物与MIF的结合亲和力从强到弱依次为RDPM，HPP，SDPM，RDP，这一分析与实验结论一致说明了RDPM和SDPM是比RDP更好的MIF底物。同时MM-PBSA表明了范德华作用很大程度上决定了结合自由能大小，是MIF与这4个底物最重要的相互作用。4．对MIF与ISO一1等2个类似物(R1，S1)的复合物进行了分子动力学模拟，以研究它们具体的结合模式，并采用MM-PBSA计算方法衡量了R型、S型ISO-1与MIF的结合自由能。R1、S1都只与MIF的残基Asn-97’形成一个O-H…O氢键。MM-PBSA分析发现R型和S型的ISO-1的结合亲和性存在明显差别，静电作用是造成R1、S1结合亲和性差异的一个最重要因素。5．针对一系列高活性MIF抑制剂香豆素衍生物(A，B，C，D，E)，采用分子对接方法构建了MIF与香豆素衍生物的复合物结构。并对5个复合物进行了分子动力学模拟。分析发现5个香豆素衍生物分子均与MIF的残基Asn-97’形成一个O-H…O氢键，分子A和E分别与残基Ile-64形成一个N-H…O氢键，而分子B、C、D都与残基Pro-1形成一个N-H…O氢键。这5个分子处在疏水性氨基酸周围，相互之间的疏水力能够增强小分子与MIF之间的结合亲和力。以上研究将有利于进一步了解MIF的催化机理，促进深入了解MIF的催化活性，并且为基于结构的MIF相关疾病的药物设计提供有益的信息。更多还原

【Abstract】 Macrophage migration inhibitory factor (MIF) is a pivotal regulator of innate immunity and plays an important role in the host antimicrobial alarm system and stress response that promotes the pro-inflammatory functions of immune cells. It has been implicated in the pathogenesis of sepsis, tumors, inflammatory and autoimmune diseases. MIF can also act as a phenylpyruvate tautomerase, D-dopachrome tautomerase and thiol-protein oxidoreductase. Since the accumulated acknowledgement for the special biological function and its role in many diseases, MIF has now become a research hotspot.The current research related to MIF is mainly limited to experiments at the cell level, while little research is performed for further study of the structure of MIF, enzymatic mechanism and interactions with inhibitors at atomic level. The detailed binding mode of MIF with different types of inhibitors, the bioactivity difference within certain a series of inhibitor analogues and the effects on the binding modes brought about by the mutation of MIF residues remain to be unknown. The possible induced conformational change of MIF during its binding with bioactive molecules, the role of key residues in the catalytic reactions and the interaction of MIF with different substrates, inhibitors need to be further investigated.Theoretical study at the atomic level was performed to tackle questions that beyond the reach of experiments. The importance of several key residues for the catalytic reactions as well as the specific recognition and binding affinity of MIF, MIF mutants with different inhibitors and substrates were further investigated. The conformational change of MIF during the complex formation was dynamically studied and the comparisons for the interactions of bioactive molecules with MIF and MIF mutants were also made.The interaction of eight hydroxycinnamate analogues, five substrate molecules, two ISO-1 analogues, five coumarin derivative molecules molecules with MIF and the interaction between eight hydroxycinnamate analogues and three MIF mutants, P1G; K32A, N97A were fully studied using molecular docking and molecular dynamics simulation.1. To elucidate the stereochemistry preference of MIF for (E)-2-fluoro-p-hydroxycinnamate and its analogues, the molecular dynamics simulations were performed on these six cinnamate analogues and the MM-PBSA analysis was also made for the evaluation of the binding free energy. The conformation of MIF is very steady and it shows no large change during the binding of cinnamate analogues. MIF has formed hydrogen bonds or salt bridge with cinnamate analogues mainly through residue Asn-97’, Pro-1, Lys-32. The binding modes of E-ligands are much different from those of Z-ligands, i. e. E-ligands have hydrogen bonds with Pro-1 while no hydrogen bond was found between Z-ligands and Pro-1. This is the main cause of large difference in the binding affinities and the MM-PBSA analysis shows that the electrostatic interaction mostly contributes to this difference.2. To study the difference in the interaction of cinnamate analogues between MIF and its mutants and to investigate the mutational effects of MIF on its conformation, the complex structures of cinnamate analogues with P1G, K32A, NP7A mutants were constructed with molecular docking and these structures were then studied by molecular dynamics simulations. The hydrogen bond analysis revealed that the binding mode of E-ligands to MIF or P1G is different from that of Z-ligands. For E-ligand, it has a N-H…O hydrogen bond with Pro-1 of MIF or with Ile-64 of P1G; however, no obvious binding difference was found between Z-ligands and the proteins. There is no any hydrogen bond of Z-ligands formed with Pro-1 of MIF or with Ile-64 of P1G.. N97A mutant can form a hydrogen bond with both Eand Z-ligands via its residue Ile-64, while it have a salt bridge only with E-ligands through its residue Lys-32. Besides forming a O-H…O hydrogen bond with residue Asn-97’ of K32A mutant, the E-ligands have a N-H…O hydrogen bond with Ile-64 of K32A mutant, however, the Z-ligands only have a O-H…O hydrogen bond with Asn-97’ of K32A mutant. The calculated binding free energy difference shows that there still exist a difference in the inhibition between the E- and Z-ligands for P1G, K32A, N97A mutants. This lead us a suggestion that P1G, K32A, NP7A may still have certain stereochemistry preference for these analogues. Molecular dynamics simulations have been also performed on the trimer, monomer of MIF, N97A, K32A mutants and no large conformational change were found. The conformation shows some rigidity on the nanosecond time scale and it was not affected much by the residue mutations.3. To investigate the binding specificity of MIF with p-hydroxyphenylpyruvate(HPP), R-dopachrome(RDP), R- and S-dopachrome methyl ester(RDPM, SDPM), molecular docking was performed to construct their structures. The multiple molecular dynamics simulations were then carried out on these structures and MM-PBSA analysis was also performed to evaluate the binding free energy. HPP forms two hydrogen bonds with MIF, one O-H…O hydrogen bond with Asn-97’ and one N-H…O hydrogen bond with Ile-64. There also exists a salt bridge between HPP and Lys-32 of MIF. RDP has two N-H…O hydrogen bonds with Pro-1 and Ile-64 of MIF. The N-H…N and O-H…O hydrogen bond are formed between SDPM and Asn-97’, Pro-1 of MIF, However, only one N-H…O hydrogen bond with limited stability exists between RDPM and Pro-1 of MIF. MM-PBSA analysis gave a correct ranking binding affinity of these MIF substrates that RDPM＞HPP＞SDPM＞RDP. This order is in agreement with the experiment that, RDPM and SDPM are better substrates than RDP. The MM-PBSA analysis also revealed that the van der Waals interaction contributes the most part of the binding free energy and this interaction is of most significance to the whole interaction between MIF and these four substrates.4. Molecular dynamics simulations were performed on the complexes of MIF with two ISO-1 analogues(R1, S1) to investigate the detailed binding mode. MM-PBSA analysis was also carried out to evaluate the binding free energy of R-and S types of ISO-1. R1, S1 each has a O-H…O hydrogen bond with MIF. MM-PBSA analysis revealed that there is an obvious difference in the binding affinity between R and S ISO-1 and the electrostatic interactions are the predominant factor to this binding affinity difference.5. For a series of coumarin derivatives(A, B, C, D, E) as potent inhibitors of MIF, molecular docking was used to construct their complexes with MIF and molecular dynamics simulations were subsequently performed on these five complexes. Five coumarin derivative molecules all forms one O-H…O hydrogen bond with Asn-97’ of MIF. Molecule A and E each has another N-H…O hydrogen bond with Ile-64, however, molecule B, C, D each forms another N-H…O hydrogen bond with Pro-1. These coumarin derivative molecules are surrounded by the hydrophobic residues and the hydrophobic force can enforce their binding affinities.The information obtained from this study could help to clarify enzymatic mechanism for MIF and may shed light on its enzymatic activity. It could provide instructive information for the structure-based drug design in the treatment of a variety of inflammatory and immune-related diseases associated with MIF.更多还原