Molecular descriptors for the structural analysis of enzyme active sites

Enzymes are increasingly used to perform a range of chemical reactions. These catalysts from nature are sustainable, selective, and efficient and offer a variety of benefits such as environmentally friendly manufacturing processes, reduced use of solvents, lower energy requirement, high atom efficiency, and reduced cost. However, natural biocatalysts are often not optimally suited for industrial applications. To boost the use of enzymes in industrial processes, it is important to expand the range of reactions catalyzed by enzymes and to improve their properties for industrial applications. Traditionally, in the past, new enzymes for desired reactions were obtained by tedious and time-consuming screening of microbial cultures, often following enrichment and isolation of new cultures. Due to the genomics revolution, massive sequencing combined with appropriate use of databases and efficient predictive bioinformatics tools have the potential to replace the current laborious screening approaches. The technological advances in the field offer an array of tools, which nowadays still have to express their full applicative potential. Actually, time-consuming, expensive, and investment-intensive screening in the laboratory is expected to be replaced by in silico screening using computer programs, ranking, design, and automated DNA synthesis, thus allowing a much shorter time from process idea to feasibility judgment with considerable savings on research costs. To fully exploit the enormous developments in life sciences, technologies and information must be used according to more effective and integrated strategies so that designing, developing, and applying new and better enzymes for industrial processes become a faster and more effective practice. The achievement of this goal is of crucial importance for the technological and economic competitiveness of industrial biotechnological processes. During the last 40 years, rigorous quantum mechanics (QM)-based computational methodologies have been developed and applied for the investigation of the physical-chemical features, thermodynamic parameters, and electrostatic contributions of enzyme active site in order to fully understand the source of the catalytic power of enzymes. QM simulations result to be very expensive in terms of computational power required because of the system definition with its high level of theory. Therefore, the enzyme system is usually QM defined just in its catalytic machinery or in a limited portion of the enzyme corresponding to the active site, while the remaining part of the system is defined with the molecular mechanics (MM) theory level [1]. While the oversimplification of the former methods makes quantitative predictions unfeasible, the latter are definitely too much time consuming to be attractive as predicting tools, and above all, they often still provide unsatisfactory quantitative accuracy . Recent advances in computer sciences have led to sophisticated and refined molecular descriptors able to describe quantitatively the features of target molecules and macromolecules.