The Prediction Research Of Metabolic Components Of Chinese Medicine Based Dealkylation And Hydroxylation Reaction

Drug metabolism is defined as the enzymatic change of chemical structure that is absorbed by the human body. Most drugs lost or reduced activity, a small number of drugs produces more reactive or toxic metabolites through the metabolism. In new drug development, poor metabolites reduce the bioavailability, which has become a major factor for the candidate compounds to be eliminated. A recent U.S. report shows that only 10% of the new candidate compounds can enter the market, and about 40% of them have been eliminated because of no activity or poor pharmacokinetic parameters in vivo. Therefore, at present we urgently need to study its metabolic characteristics, avoiding the generation of toxic products, reducing expenses and time in an early stage of drug development.With genomics and chemical synthesis development, more and more lead compounds and targets are found. Traditional experimental approach to drug metabolism properties time-consuming and labor-consuming, can not meet our needs. Increasing the cost of new drug, we urgently need a rapid and efficient way to solve this problem. Therefore, the computer virtual screening technology has aroused world wide attention from pharmaceutical companies. The technology QSAR, molecular docking, pharmacophore and other methods have been involved in drug metabolism research, and have achieved some success.Chinese medicine contains a variety of active ingredients which have complex metabolism mechanism in vivo, which has been the key influential factor for new drug development. Computer virtual screening technology can assist to have a preliminary understanding of the metabolites of the active ingredients, which can also simulate the structure of effective metabolites to access new drugs.In this paper, the classification models of ten common metabolic reactions were constructed, optimized and improved. Moreover, final models can be used to predict TCMD (Traditional Chinese Medicine Database), and explain the metabolism mechanism of Chinese medicine in vivo.The main content of this paper can be devided into the following two aspects:1. Constructing metabolic reaction related classification modelsChoosing six oxidation reactions and four deglycosidations which related metabolism from literatures as research objects, the classification models of these reactions were constructed, including O-dealkalytion reaction, S-dealkalytion reaction, N-dealkalytion reaction, Aliphatic-hydroxylation reaction, Aromatic-hydroxylation reaction, N-hydroxylation reaction, O-deglycosidation reaction, S-deglycosidation reaction, N-deglycosidation reaction, and C-deglycosidation reaction.2. In order to achieve the classification models more reasonable and reliable, the followingthree aspects were considered. (1) From the perspective of drug and enzyme's reactic mechanism, metabolic site and the matching of the drug and enzyme's pocket was considered. Molecular shape discriminative model is constructed with P450 enzyme catalytic substrate and non-substrate molecules to determine whether molecules react with P450 enzymes. Then the metabolic sites of the substrate in the molecule were also used to build the metabolic site discriminative model for the predicting of metabolic site.(2) In this experiment, the accuracy of different methods were compared. As an example,the results showed that SVM and KStar models of P450 enzyme-catalyzed o-dealkalytion reaction were superior to the others. The data sets of P450 enzyme-catalyzed o-dealkalytion reaction in the body,P450 enzyme-catalyzed o-dealkalytion reaction,and o-dealkalytion reaction in the body were also used to construct models by SVM and KStar methods, The result showed that the model of P450 enzyme-catalyzed o-dealkalytion reaction in the body is superior to other data set.(3) Traditional Chinese medicine molecules that have metabolites were used to validate the models.3. Predicting metabolite of Chinese medicine from TCMDFirstly, the TCMD database was searched to find out the moleculars which meet the reaction conditions. Secondly, molecular shape discrimination model and metabolic site discriminative model were used to predict the metabolic sites, that are metabolic atoms or chemical bonds. Finally, the rationality of metabolites based on the reported pharmacological experiments was analyzed.This paper choosed six oxidation reactions and four deglycosidations as research objects, constructed molecular shape discriminative model and metabolic site discriminative model. It was discussed how to improve models'reliability. After that, the models were used to predict TCMD, find metabolites, and explain the metabolic mechanism of Chinese medicine in the body.