Study On Nucelic Acid-Ligand Molecular Docking Methods

Molecular docking method, as a reliable tool of drug design, has been widely applied in the field of drug discovery and design. To be a robust and reliable molecular docking method, an efficient molecular conformational searching algorithm and an accurate scoring function are two crutial issues of it. In particular, scoring function is more important because it not only determines the way of molecular conformational searching, but also plays a role of discriminating the active conformation from inactive conformations of a molecule. Therefore, developing accurate scoring functions and molecular docking methods is one important filed of drug discovery and design. Nucleic acids, including deoxyribonucleic acid (DNA) and ribonucleic acid (RNA), are the crucial carriers and mediators of genetic information in organism and control all biological processes of life. In recent years, mechanisms of nucleic acid-related diseases have been gradually understood by researchers and thus drug design targeting nucleic acids has aroused growing attention. Although significant progresses have been made over the last decades, current molecular docking methods are mostly designed for protein-ligand systems, and the nucleic acid-ligand molecular docking methods are still unavailable. In addition, scoring functions and docking methods designed for protein-ligand systems are not suitable for nucleic acid-ligand systems due to the difference of structures and physicochemical properties between protein and nucleic acids. As a consequence, there is still an urgent need to develop new scoring functions and docking methods specific to nucleic acid-ligand systems including DNA-ligand and RNA-ligand systems.In view of this, this thesis mainly focuses on studying and developing of scoring functions and docking methods specific to DNA-ligand and RNA-ligand respectively. In Chapter2, focusing on DNA-ligand systems, we performed a study on assessing computation accuracy of our in-house DNA-ligand docking program zDNASBinder and other popular protein-ligand docking programs including AutoDock and Glide. The results show that/DNASBinder performs much better than the other docking programs in both accuracy of predicting DNA-ligand native binding poses and sensibility of scoring function for ranking binding poses, which provide significant basis of chosing methods for drug discovery and design aginst DNA targets. In Chapter3, focusing on RNA-ligand systems, we developed a RNA-ligand molecular docking program named RNABinder. Based on the inverse Boltzmann law, we developed a knowledge-based scoring function for RNA-ligand through statistics of atom pairs informations from RNA-ligand complexed crystal structures. Meanwhile, a force field-based scoring function for RNA-ligand according to the potential energy functions of AMBER force field and the corresponding parameter definition for RNA was developed by this work. Then, a multi-objective optimization model was designed by integrating these two types of scoring functions, and a RNA-ligand docking tool named RNABinder was successfully developed. To validate RNABinder, an accuracy test was performed among RNABinder, AutoDock and Glide. As a result, RNABinder was found to be able to accurately predict40.00%of RNA-ligand native binding poses from the test set, which is superior to Glide SP(35.56%) and Glide XP(33.33%), but a little bit inferior to AutoDock(68.89%). It is confirmed that RNABinder has an advantage in predicting native binding modes of RNA-ligand and is a reliable method for drug discovery against RNA targets.In conclusion, focusing on nucleic acids systems, DNA-ligand docking programs have been further studied, and a RNA-ligand docking program called RNABinder has been developed, which provide powerful tools for drug discovery and design targeting nucleic acids.