| Drug discovery is a time-consuming and high-cost process,which costs $1,778 million and takes over 13 years for per NME launch.Novel compounds are mainly achieved by three methods in the past: 1.Screening of natural products by experience;2.Screening large commercial or in-house databases of chemicals against the target;3.Chemical synthesis.These conventional methods showed low efficiency in the discovery of new drugs,so new strategies are required to accelerate the drug discovery process.Computer aided drug design(CADD)aims to improve the efficiency of drug discovery and reduce the cost,which has been widely used in academia and industry.Virtual screening accelerates the discovery of lead compounds.Quantum mechanics(QM)plays an import role in the calculation of non-covalent interactions,configuration identification of chiral compounds,and explanation of chemical reaction mechanism.Molecular dynamic can be applied to simulate the dynamics of proteins,which can explain biological mechanism and guide biological experiment.The main work of this thesis includes three parts: 1.QM calculations for the investigation of halogen bonding and the explanation of chemical reaction mechanism;2.Development of an efficient tool for detecting protein cavities;3.Identification of active compounds by virtual screening.The first section consists of chapter 2 and chapter 3.In chapter 2,QM calculations were carried out to study the halogen bonding,which exists in drug-protein complexes.Halogen bonding is a noncovalent interaction,the strength of which is comparable with hydrogen bonding.The electrostatic potential surfaces of halogen atoms contain a positive area along the extension of the covalent bond,which is called σ-hole.We demonstrated that halogen bonding with negatively charged donors is unstable or metastable in vacuum,but the interaction becomes attractive in protein environment.This kind of halogen bonding was called as negative halogen bonding.In the thesis,we studied the effect of the distance between the negative charge center of the organhalogen and the halogen atom on the strength of halogen bonding in different solvents.A series of model systems,composed of 4-halophenyl-conjugated polyene acids and ammonia,were designed as the negative halogen bonding complexes.QM calculations demonstrated that the longer the distance,the stronger the bonding.The solvents can enhance the weak halogen bonding and weaken the strong halogen bonding.Hence,the strength of halogen bonding involving negatively charged donors could be adjusted by changing the distance and the environment in which the bonding exists,which can be applied in material and drug design for tuning their function and activity.Transition metal-catalyzed C-H activation to construct C-C and C-X(X=N,O,S,et al)is one of the most powerful tools,which can be applied to synthesize natural products,drug molecules and materials.In chapter 3,QM calculations were carried out to study the the coupling mechanism of C(sp~2)-H/ C(sp~3)-H activation.Theoretical studies on the mechanisms of the reactions can help us gain insight into catalytic reaction and guide us design the reactions.Collaborator Zhou et al reported an unprecedented rhodium(III)-catalyzed regioselective redox-neutral annulations of 1-naphthylamine N-oxides with diazo compounds to synthesize various ~1H-benzo[g]indulines,which exhibit extra-ordinary biological and pharmaceutical properties.Density functional theory(DFT)calculations were performed with Gaussian09 to gain further insight into this reaction.We confirmed the possible reaction pathway from DFT calculations.Firstly,C(sp~2)-H activation most likely occurs through a carboxylate-mediated concerted metalation-deprotonation(CMD)pathway.Secondly,carbon inserts into metal carbon bond via metal-carbene migratory insertion.Thirdly,C(sp~3)-H activation most likely occurs through a similar carboxylate-mediated CMD pathway.Lastly,intermediate iminium is formed by an O-atom-transfer strategy.The second section is focused on the development of new method for detecting the protein cavities.Identification of protein-ligand binding sites is an essential step in the structure-based drug design(SBDD).In chapter 4,we developed some tools for ligand-binding site prediction.Various methods have been put forward for detecting the protein cavities,such as grid algorithm,sphere set algorithm,Voronoi Diagram algorithm.We developed a new algorithm called “sphere vector” on the basis of the above algorithms.Firstly,the two probe sphere with different radius will be used to roll over the protein surface to generate the cavities.Then,surface grid will be checked according to the vector method.We developed a new PyMOL plugin called “PyPocket” and the command tools.PyPocket can recognize the pocket boundaries more accurately than the current software Fpocket and PyRoll.We constructed a target-drug dataset including 75 PDB codes.We carried out the performance test for the PyPocket on the dataset.PyPocket achieves a high success rate of 98.6% while the success criterion is to detect an atom of the ligands.The binding strength of the ligand and the receptor is determined by its shape and physicochemical properties,so the characteristics of the pocket can help determine whether the pocket can be selected to design drug molecules.Then,we developed a plugin called “PyPocket-Descriptors”,which can calculate all kinds of descriptors on the pocket.We selected 56 protein pockets to perform calculation and analysis,which demonstrated that the volume of the pocket is mainly in the 500-1000 ?3 and the most popular residue is leucine while the least is cystine.In the last section,virtual screening technology was applied to discovery new active compounds.Virtual screening can improve the efficiency of lead compounds identification,which has been widely used in academia and industry.In chapter 5,virtual screening was carried out to find new active compounds for the treatment of S.japonicum disease.We screened the compounds in SPECS and in-house database against the aldose reductase.50 compounds were selected for the following bioassay.Two compounds,ALD3 and ALD5,could inhibit the aldose reductase with low activity.The docking studies showed that Trp111,Leu295 and Leu296 can form hydrogen bonds with ALD3.We also performed virtual screening against another target 3-oxoacyl-ACP reductase(sj-OAR).3D structure of sj-OAR was established by homology modeling.The docking studies showed that Ile18 can form hydrogen bonds with many compounds,which have high docking scores.In chapter 6,we screened the compounds in SPECS and in-house database against the transgelin-2 protein for the treatment of asthma disease.On the basis of binding mode,docking score and careful visual inspection,22 compounds were selected for the following bioassay.4 out of 22 compounds showed activity under 1 μM.Among them,TSG12 show the highest activity with EC50 values of 10.2 μM.Surface plasmon resonance analysis demonstrated that the compound TSG12 has physical interaction with transgelin-2.The docking study showed that residues Lys57,Gln89,Val42 of transgelin-2 were found to form hydrogen bonds with TSG12. |
PDF Full Text Request