Modern bioinformatics and cheminformatics approaches are key approaches to structure-function relationships in biomolecules, and are now widely used in pharmaceutical and biotechnological companies and research institute. Among the computational tools, the molecular dynamics (MD) occupies a central position as it allows to investigate the dynamic properties of a protein and to extract precise information on biologically relevant motions, such as the flexibility of specific regions, the adaptability of the active site or the local structural rearrangement upon ligand binding. However, MD simulations are among the most CPU intensive calculations used to study molecular systems. As a consequence, the use of high-performance computing is crucial to carry out this kind of studies.
As an example of high-performance computing applied to the MD investigation of proteins we present a study in which the molecular basis of cold adaptation inside the specific enzymatic class of pancreatic elastases have been explored by MD simulations.
A comparative MD investigation reveals that specific loop regions are characterized by enhanced flexibility in the cold-adapted enzymes, leading to the conclusions that these differences play a crucial role for catalysis at low temperature. This observation fully supports the hypothesis suggesting that flexibility is the main adaptive character of psychrophilic enzymes. Remarkably, the corresponding mesophilic enzymes are characterized by enhanced flexibility, when compared to the cold-adapted ones, in scattered regions distant from the functional sites. Therefore, our results are also in agreement with a scenario in which local rigidity in regions far from the functional sites can be a positive factor in the adaptation of psychrophilic enzymes.