| Proteins are the important substance in the life activities. The interactions of proteins with all kinds of biomolecules form the molecular foundation to guarantee the normal operation of the self-organizing complicating system of lives, which draw the constant attention of the biological studies and become the focal and hot themes of the life science field in the 21st century. Following Human Genome Plan, the life science has entered the post-genomic era. Proteomics, centering at the subject of proteins interaction and recognition, has burgeoned soon, been extensively applied in many fields of the life science, and hold the critical strategic position. As the growing progress of computers'processing ability, theoretical simulation methods have been rapidly developed and extensively applied. Utilizing molecular docking method to predict the structures of protein complexes has been a powerful means to study interactions and recognition of proteins, and could provide instructive hints for experimental studies. How to obtain more right docked decoys and then how to efficiently yet accurately screen out the near native structures remain to be the key question and challenge facing to the protein-protein docking algorithms. Therefore, the first part of this dissertation focused on these two major difficult points and made further exploration and discussion. In the study of protein structure and function, a myriad of proteins were found to be connected with the occurrence and protection of diseases, which can be ideal targets for the novel drug design. Three enzymes play the key roles in the life cycle of human immunodeficiency virus (HIV), the pathogen of AIDS. Integrase catalyzes the integration of DNA replication of the viral genome into the host genome and there is no known analogue in the human cells. Therefore, integrase is a potential target for the anti-HIV drug design. However, its poor solubility, lack of the complete enzyme's structure and the limitation of its screening method for inhibitors in vitro, retarded the development of integrase inhibitors. Up to now, none of integrase inhibitors has been commercially available. Consequently, in the second part of this dissertation integrase was selected as the studying target, employed the computer-aided drug design method to virtually screen the HIV-1 integrase inhibitors to make rational design and modification. On this basis, in order to obtain the bioactive HIV-1 integrase inhibitors, the enzymological activity assay platform is being set up to carry out a series of biochemical experiments such as HIV-1 integrase expression and purification (see the 3rd part of the dissertation). This dissertation mainly includes three aspects as follows. 1. Study on Protein-Protein Docking Approaches The goal of this work is to present an efficient biological-enhanced sampling docking method. The binding site information of complexes was incorporated in the rotation-translation scanning process of FTDock algorithm. The docking predictions of 10 targets by this method showed that a larger number of near-native structures (hits) were obtained, and the ligand root mean square deviations (RMSD) of the best docking decoys were obviously reduced. The multi-conformational superposition scheme was also adopted to represent the backbone flexibility of protein, which upgraded the performance of the docking procedure. The protein-complex-type-dependent scoring functions were proposed here. In accord with the variation of interface features, 64 complexes were categorized into four subsets, i.e. protease/inhibitor, antibody/antigen, other enzyme/inhibitor and others complexes. Four combinatorial complex-type-dependent scoring functions were derived and optimized by the multiple regression method for every group. The scoring functions incorporate energies terms upon binding, embodied by the interface residue pairing preferences (RP), the atomic contact energy (ACE), electrostatic and van der Waals'interactions. As showed in the docking results, all the four scores demonstrate stronger discriminative ability than the individual energy term. Moreoverfor the protease/inhibitor, antigen/antibody and enzyme/inhibitor classes, the performance was elevated to a higher level in comparison with some general scoring functions. 2. HIV-1 Integrase Inhibitors Design As a potent and selective HIV-1 integrase inhibitors, three-dimensional quantitative structure-activity relationship (3D-QSAR) of styrylquinoline derivatives (SQLs) and their binding modes with integrase were investigated in the present work, and rational structural modifications for them were also completed. Comparative molecular field analysis (CoMFA) was performed to analyze 3D-QSAR of SQLs. The resulting CoMFA model can reasonably predict the biological activity of SQLs and visualize structural requirements for the biological activity of these inhibitors. Docking results show that a carboxyl group at C-7 and a hydroxyl group at C-8 in the quinoline subunit bound closely to the magnesium ion in the integrase catalytic site. And there is a linear correlation between the binding energy of the inhibitors and their inhibitory effects. According to the results above, a de novo drug design method was employed to generate new compounds. Then, the compounds were evaluated by docking simulation and clustering. Four different series of compounds with the structural features not found in known SQLs were obtained in this work. Simultaneously, the interfacial peptides were designed by using computational methods. Molecular dynamics simulation was performed to investigate the intramolecular interactions within the integrase dimer and the interactions between interfacial peptides and the monomer. Amino acids virtual mutations were performed on interfacial peptides ( α1 and α5 helix). The helix-forming tendencies and affinities of peptides with integrase were investigated subsequently to test our design using the methods of secondary structure prediction, conformational simulation and molecular docking. The secondary structure prediction and conformational simulation both showed favorable results for the designed peptides. Some designed peptides exhibited obvious improvement on both the helix properties and binding free energies compared...
|
PDF Full Text Request