Protein structure and function relationships: application of computational approaches to biological and biomedical problems

In this work we have studied several cases by means of different computational approaches for the analysis of the structure and function relationships. In chapter 2 we describe a method, based on multiple neural networks, that we developed for evaluate the accuracy of predicted threedimensional protein structures. This tool has been used in different studies described in this work, in which the prediction of the 3D structure of the protein under study, has been necessary. In chapter 3, the interaction among a new class of natural sweeteners (steviol glycosides) and the human sweet taste receptor, has been analyzed by means of an insilico docking study, which allowed to identify the preferential binding site for the steviol glycosides. In chapter 4 the relationship between the dynamical properties and the function of some psychrophilic enzyme has been studied. A comparative study (psychrophile vs mesophile) of the thermodynamic properties of two different enzymes belonging to the elastases and the uracilDNAglycosylases families has been done. This study, carried out with molecular dynamic simulations, revealed that the low temperature adaptation is related to the different flexibility of the psychrophilic compared to the mesophilic enzyme. In chapter 5, we have studied the structural and functional impact of point mutations on three different proteins which are involved in serious rare diseases which cause grave metabolic disorders.