A Chemoinformatic Analysis On Metabolic Network And The Application For Screening Of Drug Targets On Erwinia Carotovora

Since most metabolites are small molecules and metabolic processes are basically chemical reactions, we speculate that the metabolic network organization may have a chemical basis, which stimulated our interest to address this issue by chemoinformatics. Metabolic networks exhibit the same scale-free organization with other biological networks and non-biological networks. The scale-free organization of metabolic networks has been explained in terms of evolution that the new-recruited metabolite members attach preferentially to those that are already well connected. However, the molecular basis of preferential attachment principle has not been elucidated, as it is inexplicable how the new metabolites "know" which metabolites are well connected. Also we do not know how to select metabolites as hubs in network design. Besides, the evolutionary explanation to the metabolic network organization has no direct implications for practice, so we try to apply the correlations between network (topological) and chemical properties for metabolites to the screening of drug targets on Erwinia carotovora. And then to provide some new information of drug design on Erwinia carotovora, homology modeling and molecular docking were performed after screening.First and foremost, metabolic network reconstruction, visualization, graph theory and chemoinformatic analysis on metabolic networks of Escherichia coli, Saccharomyces cerevisiae and KEGG were performed. It is found that there exist qualitative and quantitative correlations between network (topological) and chemical properties for metabolites, suggesting that the organization of metabolic networks has a chemical basis. Since life originated from water environments, the primordial metabolites must be highly hydrophilic. With the evolution of organisms, more and more complex membrane systems evolved, which required hydrophobic metabolites to perform intercellular and intracellular communications. As a result, the evolutionary direction of metabolites is from hydrophilic to hydrophobic, which is clearly shown in the chemical evolution of Saccharomyces cerevisiae metabolome.Secondly, this chemical basis is further elucidated in terms of high concentrations required by metabolic hubs to drive a variety of reactions. For metabolites that participate in a large number of reactions as reactants, they must reserve high concentrations to drive the reactions. According to the correlation between concentration and polarity of metabolites, it is reasonable to infer that the early-originated metabolites have relatively higher concentrations than the late-recruited counterparts in organisms. Taken together, the present analysis reveals that metabolite concentration is a key factor to govern the metabolic network expansion. Although the late metabolites can not "know" which counterpart is well connected, they can "sense" which member is abundant, which provides a self-consistent explanation to the preferential attachment principle at molecular level. This elucidation has direct implications for metabolic network design, that is, the polarity of working environments must be considered during the design of metabolic networks. If the environments are polar, one should use hydrophilic molecules as hubs, while if the environments are non-polar, hydrophobic molecules should be selected as hubs. This conclusion is preliminarily supported by the fact that the hubs of network derived from synthetic chemistry are really much less polar than the hubs of metabolic networks, well reflecting the fact that synthetic chemistry mainly occurs in organic solvents which are less polar than water.Finally, the observed correlations between degree and chemical properties could be basically applied to the screening of drug targets on Erwinia carotovora. We have screened 17 targets of the initial 44 targets from DIGAP database by two criterions of Degree and AlogP, and determined murA to be the most potential drug target of Erwinia carotovora by many other conditions, such as not available on hosts, successful targets by experiments, chokepoint metabolites, etc. And then, homology modeling and molecular docking were performed to provide some new information of drug design on Erwinia carotovora.In summary, it brings forth several new ideas as follows. First, the metabolites were just seen as network nodes but their own chemical properties were ignored in previous researches. For this, we investigate the effect of chemical properties of nodes that may have contributed to analyzing structures of metabolic networks. And the present chemoinformatic analysis clearly indicates that the organization of metabolic networks has a chemical basis. The application of the correlations between network (topological) and chemical properties for metabolites to the screening of drug targets on Erwinia carotovora is performed. Second, as metabolic reactions mainly occur in non-membrane systems which are hydrophilic environments, the metabolic hubs must be water-soluble to reach high concentrations tend to be small and high-polar. The present finding not only provides a molecular-level explanation to the preferential attachment principle for metabolic network expansion, but also has direct implications for metabolic network design, that the criterion of metabolic hub partially depends on the polarity or non-polarity of environment.