Molecular Modeling And Docking Of HAb18G/CD147 With Its Monoclonal Antibody And Antibody Humanization Design

Along with vigorous development of computer assisted molecule simulation methodology, Computational biology is increasingly used in all fields of life sciences. Under such situation, many theoretical methods and application software have presented in protein structure analysis and prediction and molecule docking. The application of these methods and tools in the study about antibody structure and interaction of antigen-antibody will provide strong theoretical guidance and foundation data, which are required by experiments, for antibody engineering-related research.The first purpose of our study was to establish a method of 3-D structure prediction and simulation of mouse and human antibody variable regions based on public-accessible bioinformatics tools and databases. Furthermore, put methods into available software tools. Using this software, an anti-HAB18G/CD147 monoclonal antibody is modeled. The second purpose was to verify the epitope and the key residues by analyzed the complex model of HAb18G/CD147 extracellular portion and its MAb based on computer assisted molecule docking. The third purpose was to perform a design of humanizing murine antibody with the method of residue replacement based on analyzing of antibody spatial structure and interaction among residues by means of computer assisted molecule design.There were three parts in our study: 1) Antibody structure modeling. A local database of antibody 3-D structure was constructed by extracting antibody resource from PDB. Sequence alignment and homology template search against the local database for query antibody was performed by a Java program and modeling was realized by Modeller 9v1. 2) Epitope prediction. The model of extracellular portion of HAb18G/CD147 molecule was constructed by a serial of steps that included fold recognition, DOPE evaluation and model refinement using AMBER6. The docking of HAb18G/CD147 extracellular portion and its MAb was performed by InsightII. 3) Humanization design of MAb HAb18 based on variable domain resurfacing. Identification of the surface exposed non-human like residues of MAb variable region and calculation of inter-molecular and intra-molecular hydrogen bond interaction by sequence and structure analysis in terms of the homology modeled variable region of mAb determined the residues that need to be mutated to their human counterpart.In this study, we proposed a method of predicting antibody 3-D structure and developed a software tool named by AbModeller using to construct antibody model that is easy to be used and contain user interface and interactive function. In 2? of RMSD, AbModeller can predict accurately antibody spatial structure. Moreover, to simulate the complex model of HAb18G/CD147 extracellular portion and its MAb by molecule docking and identify the key residues of epitope that can be a theoretical evidence for experiment. Finally, the antibody variable domain resurfacing method was applied to the humanization of the monoclonal antibodies HAb18 against human hepatoma cell. The candidate mutation sites were put into three categories in experiment for maintaining an appropriate balance between the biological function and reduced immunogenicity.In conclusion, the methods and software of modeling and docking of antigen-antibody can support effectively the research of antibody biology and provide theoretical guidance for humanization and epitope identification and drug design based on the sites of antigen-antibody interaction.