【作者】 秦晋；

【导师】 姚小军；

【作者基本信息】 兰州大学 ， 化学信息学， 2010， 博士

【摘要】 蛋白激酶(protein kinase,PK)是细胞内最大的蛋白家族之一,在真核细胞信号转导中扮演重要的角色。蛋白激酶是一类磷酸转移酶,其作用是将ATP分子的γ磷酸基转移到底物蛋白特定的氨基酸残基上,使底物蛋白磷酸化。人类基因组内共含有518个蛋白激酶基因,约占真核生物基因的1.7%。蛋白激酶活性异常通常会引发包括癌症、糖尿病、炎症在内的许多重大疾病。超过400种人类疾病与蛋白激酶有直接或间接的关系。因此蛋白激酶已成为继G蛋白偶联受体之后的第二大药物治疗靶标。据统计,目前全世界药物在研或开发项目中约三分之一均与蛋白激酶相关。蛋白激酶抑制剂(protein kinase inhibitor,PKI)可用于多种疾病的治疗,其中ATP-竞争性蛋白激酶抑制剂研究得最多,这类抑制剂已被研究和开发成为治疗多种复杂性疾病的新药。目前ATP-竞争性蛋白激酶抑制的研究主要集中在新骨架类型先导化合物的设计和发现,以及选择性抑制剂和多靶点抑制剂开发等方面。目前,计算机辅助药物设计(computer-aided drug design,CADD)作为药物研发中的重要技术和工具,被应用于蛋白激酶抑制剂的研究中,推动了这一研究领域的长足发展。定量构效关系、分子对接、药效团、同源模建和分子动力学模拟等多种分子模拟方法,以及量子化学等多种CADD方法均被应用于蛋白激酶抑制剂的设计和发现。本论文采用定量构效关系、分子对接和药效团建模等多种分子模拟技术和手段,研究多种蛋白激酶抑制剂分子结构与其活性之间的关系,抑制剂与蛋白之间的相互作用以及影响化合物活性的重要药效特征。旨在获得影响抑制剂活性的重要结构和药效特征、抑制剂与蛋白激酶间相互作用的机理,从而指导抑制剂的设计、结构优化以及活性预测,辅助设计和合成更高效的靶向激酶的治疗药物。在论文第一章中,我们对蛋白激酶及其抑制剂研究进展、计算机辅助药物设计中的分子模拟方法,如定量构效关系、分子对接及药效团模型方法做了介绍,并描述和总结了计算机辅助药物设计在蛋白激酶抑制剂研究中的应用。第二章中,我们对一系列噻唑氨类Aurora-A的抑制剂进行QSAR研究。2D-QSAR模型用遗传算法—多元线性回归方法(GA-MLR)建立。结果表明除了信息指数,GETAWAY和WHIM描述符对这类化合物抑制活性均有重要贡献。3D-QSAR研究中的CoMFA和CoMSIA模型给出的三维等势面图与这类化合物的结构特征以及X—射线晶体衍射结构信息相一致,并且指出对22号化合物苯胺部分的进一步修饰应综合考虑立体、疏水和氢键性质。第三章中,我们对一系列Rho激酶抑制剂进行分子对接和3D-QSAR研究。对接研究获得了整个数据集化合物的活性构象,并研究了化合物在质子化与非质子化两种状态下的结合模式。CoMFA和CoMSIA分析研究了影响化合物生物活性的关键结构因素。根据分子对接以及模型的三维等势面图结果发现1-H吲唑化合物优于异喹啉化合物,并提出了对这类抑制剂的修饰方案：(1)P区苯环4位用给电子基取代；(2)P区用体积较大疏水基团取代。根据分子模拟研究得到的信息,我们设计了一系列预测活性较高1-H吲唑化合物,说明通过适当的取代可以提高这类化合物的活性。第四章中,应用分子对接、药效团建模和3D-QSAR方法进行一系列EphB4激酶抑制剂的分子模拟研究。分子对接结果表明这类化合物与受体ATP结合口袋形成多个氢键相互作用：A环吡啶-N原子与铰链区Met696残基形成氢键；脲基的羰基与NH分别与Lys647和Asp758残基形成氢键。药效团模型给出了影响化合物活性的重要药效特征。CoMFA和CoMSIA分析研究了影响化合物生物活性的关键结构因素,模型的三维等势面图能够为这类抑制剂的修饰提供指导。第五章中,我们对一系列B-RAF激酶V600E突变体的抑制剂进行分子模拟研究。对接研究表明活性最高的代表化合物40与受体ATP结合口袋形成四个氢键相互作用：A环吡啶-N原子与铰链区Cys531残基形成氢键；脲基的羰基与NH分别与Asp593和Glu500残基形成氢键。药效团模型给出了影响化合物活性的5个重要药效特征：两个氢键受体、一个氢键给体、一个疏水特征及一个芳环特征。基于公共骨架、分子对接和药效团叠合的CoMFA和CoMSIA分析表明,利用分子对接叠合的构象得到的模型具有最佳的预测能力,模型的三维等势面图能够为这类抑制剂的修饰提供指导。第六章中,我们将分子对接和2D-QSAR方法应用于MK-2激酶抑制剂的研究中。分子对接结果表明这类化合物与受体ATP结合口袋形成四个氢键相互作用。2D-QSAR分析研究了影响化合物生物活性的关键结构因素,实验结果表明化合物包含的芳香环、氢键性质、拓扑信息以及NH基团对这类抑制剂的活性影响很大,这为今后这类抑制剂的设计和结构修饰提供了一定指导。第七章中,通过对一系列AP-1和NF-κB介导的转录活化作用的抑制剂进行CoMFA和CoMSIA研究,来考察这些抑制剂分子周围立体场、静电场、氢键给体场和氢键受体场对化合物活性的影响。所得模型可以成功预测这类化合物的活性。根据模型的三维等势面图给出的影响抑制活性的结构特征,提出应综合考虑立体、静电和氢键给体场性质来对喹唑环上苯环取代基进一步修饰以提高化合物活性。更多还原

【Abstract】 Protein kinases (PK) constitute one of the largest protein families in humans. Their function is to catalyze phosphorylation of serine, threonine, or tyrosine residues, and to regulate the majority of signal transduction pathways in cells. Thus they play important roles in cell growth, metabolism, differentiation, and apoptosis. There are 518 PKs are predicted in the human kinome based on the information from the human genome sequence, approximately 1.7% of all human genes. Deregulation of protein kinases is implicated in a number of diseases including cancer, diabetes, and inflammation. Thus, protein kinases make up the second largest group of pharmaceutically relevant protein targets. It is estimated that approximately one-third of drug discovery programs target protein kinases.Targeted inhibition of protein kinases has thereby become an attractive therapeutic strategy in the treatment of relevant diseases. Current drug discovery efforts typically focus on developing ATP-competitive small molecule protein kinases inhibitors (PKI), mainly selective and multi-targeted inhibitors.As an important technology and tool for drug design, computer-aided drug design (CADD) has been applied to PKI discovery. Molecular modeling approaches, such as quantitative structure-activity relationship (QSAR), molecular docking, pharmacophore, homology modeling, molecular dynamic simulation, and quantum chemistry methods have been applied to PKI design and discovery.This dissertation applied several techniques (QSAR, molecular docking and pharmacophore) to build the correlation between molecular structure features and their bioactivity, and to study the interaction between the targeted protein and their inhibitors. We aim at gaining insights into the key structural and phamacophore features affecting activity, and the interaction mechanism for inhibitor-protein binding, guiding the design, structural modification and activity prediction of PKI to aid the design and synthesize of highly active drugs targeted PK. In Chapter 1, we gave a general introduction of protein kinase, corresponding inhibitors, and molecular modeling methods used in this thesis, such as QSAR, molecular modeling and pharmacophore. The application of computer-aided drug design methods in PKI discovery is also described.In Chapter 2, QSAR study on a series of aminothiazole derivatives as Aurora-A kinase inhibitors was performed. The 2D-QSAR model was built using the genetic algorithm-multiple linear regression (GA-MLR) method. The obtained results indicated that 3D-GETAWAY and WHIM descriptors exhibited significant contributions to the inhibitory activity. The 3D-QSAR models were established by using CoMFA and CoMSIA methods. The obtained 3D contour maps were in accordance with the structural features of these inhibitors and the X-ray structure, which suggested that further modification of compound 22 on the aniline substructure should consider steric, hydrophobic and hydrogen bond properties.In Chapter 3, molecular docking and 3D-QSAR approches were applied to molecular modeling of a series of Rho kinase inhibitors. Docking studies were performed to obtain the active conformations for the whole dataset and normal bingding mode. The CoMFA and CoMSIA analyses gave some insights into the key structural factors affecting the bioactivity of these inhibitors. The obtained 3D contour maps along with the docking results suggested that the 1H-indazole derivatives may be superior to isoquinoline derivatives and highlight that 1) electron-donating substituents on 4-position of phenyl and 2) bulky and hydrophobic groups in the P region of the binding pocket increase potency. According to the results of our molecular modeling study, we designed a series of 1H-indazole derivatives that are possible to have high Rho kinase inhibitions.In Chapter 4, molecular modeling studies on a series of EphB4 kinase inhibitors were performed. Molecular docking results show that these inhibitors form four hydrogen bonds with binding pocket:pyridyl-N formed a hydrogen bond with Met696 residue of the hinge region, urea C=O and NH formed three hydrogen bonds with Lys647 and Asp758 residues. Pharmacophore model presented the most important pharmacophore features and their distributions. CoMFA and CoMSIA analyses gave some insights into the key structural factors affecting bioactivity. The obtained 3D contour maps can help further design and structural modification of these inhibitors.In Chapter 5, molecular docking, pharmacophore and 3D-QSAR studies are performed on a series of V600EB-RAF kinase inhibitors. Molecular docking explored the binding mode between ligands and receptor. Pharmacophore model presented five most important pharmacophore features:two hydrogen bond receptor, one hydrogen bond donor, one hydrophobic feature and one aromatic ring. CoMFA and CoMSIA analyses based on docked conformers obtained optimal predictivity and disclosed the key structural factors affecting bioactivity.In Chapter 6, a series of MK-2 inhibitors were analyzed using molecular modeling methods, including molecular docking and 2D-QSAR study. The docking results showed that these compounds can bind in ATP binding pocket by forming four hygrogen bonds.2D-QSAR study was used to analyze the critical factors influencing the inhibitory activity. The obtained results showed that the number of aromatic ratio, hydrogen bond properties, topological information and NH groups could greatly affect the activity. Our study can provide guidance for inhibitors design and modification.In Chapter 7, we analyzed a series of inhibitors of AP-1 and NF-κB mediated transcriptional activation using CoMFA and CoMSIA methods. The influence of steric, electrostatic, hydrogen bond donor and hydrogen bond acceptor field around molecules were investigated. The obtained models can be used for activity prediction of newly designed inhibitors and suggested that further structural modification should consider steric, electrostatic and hydrogen bond donor properties.更多还原