Quantitative Structure-activity Relationship And Moiecular Docking Study Of Several Kinases Inhibitors As Antitumor Agents

The molecular design of kinase inhibitors is a hot spot in the research and development of antitumor drugs.In this paper,QSAR and molecular docking studies on several important antitumor kinase inhibitors were performed,aiming to provide theoretical guidance for the design of new kinase inhibitors.The paper is divided into four parts.In the first chapter,kinases and inhibitors,QSAR,Molecular Docking were summarized.In the QSAR part,we mainly introduced the molecular surface electrostatic potential and its derived descriptors,as well as various modeling methods developed in recent years,including support vector machine and Gaussian process.The second chapter focuses on the QSAR of indirubin analogues as kinases inhibitors.First,the 2D-QSAR models based on the fragment surface electrostatic potential of indirubin CDK1,CDK5,GSK-3β and DRAK2 inhibitors were built by MLR.The results are as followings:The descriptors of surface electrostatic potential derived from different substituents(or fragments)of indirubin analogues can be well used to construct their quantitative structure-activity relationship.For the inhibitory activity of CDK1,the electrostatic interaction tendency σ 2 from X fragment,R1 and R2 substituents have major impacts on the activity.For the inhibitory activity of CDK5,the average value of independent positive electrostatic potentials of X fragment,the separation degree of independent electrostatic potentials of R1 substituent and the average value of independent negative electrostatic potentials of R2 substituent are the most significant factors affecting the activity.For the inhibitoy activity of GSK-3β,the separation degree of independent electrostatic potentials from R1 and R2 substituents,and the separation degree of independent positive electrostatic potentials from R3 and R4 substituents are important for the activity.Relatively,the electrostatic potentials from X fragment are of little influence on the activity.For the inhibitory activity of DRAK2,the separation degree of independent negative electrostatic potentials from R4 substituent is the most important factor.Then,several linear and nonlinear models were constructed by modeling methods such as PLS,SVM,LSSVM and GP.The advantages and disadvantages of the models were analyzed by Monte Carlo cross validation.The results showed that the linear model re:flected well the relationship between the structure and activity of the compounds.In the third chapter,QSAR models based on the surface electrostatic potential of the molecular fragment were established for the maleimides analogues as GSK-3βinhibitors.The results showed the most positive electrostatic potentials from fragment X,the number of independent positive electrostatic potentials and the average value of electrostatic potential from fragment R’3,and the charge separation degree from fragment R3 had significant effects on the activity.Subsequently,we used the Glide module to conduct molecular docking.It was shown from the binding diagram that the hydrogen bond donor group such as the hydroxyl or carboxyl group should be introduced in the fragment R’3 region,which could form many hydrogen bonds with nearby residues.The interaction between the R3 and the nearby amino acids is mainly electrostatic,so polar groups should be introduced here,which is easy to have electrostatic interaction with the residues of the active sites.In the fourth chapter,Molecular Docking and 3D-QSAR research were carried out for carbazole amides as BTK inhibitors.The active conformations of molecules were obtained by Molecular Docking,and the 3D-QSAR models were built on the basis of superimposition.The model with 3 principal components showed best performance.By the analysis of the potential map and docking results,the structural modification and optimization of carbazole amides as BTK inhibitors should be mainly based on the ortho-and meta-substituents of benzene ring linked to carbazole.