ИСТИНА

Modulation of enzyme functional properties can be performed by applying two principally different approaches – by implementing changes in protein structure or due to binding of modulating molecules (ligands, effectors, stabilizing agents). Due to selective mutations we can extend substrate specificity, increase stereoselectivity or improve enzyme stability. Choice of hotspots for mutation can be rationalized nowadays by exploiting methods of systems biology, bioinformatics, molecular modeling and computational screening. The same set of methods can be applied to search for modulating molecules and their binding sites. Until recently, the major interest in drug design was focused on the active sites of enzymes and search of competitive inhibitors that prevent interaction with substrates and cofactors. However, computational structure analysis has revealed that enzymes along with the active sites have numerous unexplored binding pockets with unknown functional role. This well corresponds to the lately discussed suggestion that allostery – regulation of protein function at binding of low-molecular weight compounds in topologically independent regulatory sites – may be an inherent property of virtually all proteins. It becomes important therefore to identify and characterize new binding sites in protein structures and understand their role in modulation of protein function. In order to solve these problems we suggest to use methodology that combines a power of systems biology, bioinformatics, molecular modeling, theoretical chemistry and high-performance computing and allows to develop different approaches to modulate enzyme activity, selectivity and stability. This methodology was applied to understand structurefunction relationship in several enzyme families: Ntn-hydrolases, penicillin-binding proteins, α/β- hydrolases. Function-related positions in corresponding enzyme families were identified and used as hotspots for mutations to increase stability and synthetic activity of penicillin acylase from Escherichia coli, expand substrate specificity of D-aminopeptidase from Ochrobactrum anthropic and introduce amidase activity into Candida antarctica lipase B. Molecular modeling of in silico constructed mutants was used to evaluate effect of substitutions at function-related positions on stability as well as catalytic properties and to select the most promising variants for experimental evaluation. Produced mutants of penicillin acylase, D-aminopeptidase and lipase B demonstrated significantly improved functional properties. The methodology was also applied to search for previously unknown binding sites in the structure of the glyceraldehyde-3-phosphate dehydrogenase and the transketolase superfamilies and design of specific inhibitors of these enzymes with a novel mechanism of action. The experimental study has shown that compounds selected by the computer screening were selective inhibitors of glyceraldehyde-3-phosphate dehydrogenase and transketolase from Mycobacterium tuberculosis and did not suppress activity of their human homologs. The methodology can be used as a systematic tool to study catalytic mechanisms of enzymes, characterize and rank enzyme binding sites, search for selective inhibitors, select function-related positions and use them as hotspots for mutation to rationalize different protein engineering approaches and design enzymes with requested functional properties.