INVESTIGATING AND PREDICTING THE DETERMINANTS OF PROTEIN-PROTEIN INTERACTIONS THROUGH COMPUTATIONAL-STRUCTURAL BIOLOGY APPROACHES: IMPLICATIONS FOR STRUCTURAL VACCINOLOGY

Clarifying the physico-chemical principles of protein-protein interactions is critically important to understand the relationships between biological structures and functions in all biochemical mechanisms. In this project we aim to develop, validate and apply new computational-theoretical methods to study and predict the binding regions of proteins starting from 3D structural information and from the analysis of the conformational and physico-chemical properties of the constituting amino acids. In particular, this project entails the integrated analysis of the energetic properties of different datasets of proteins solved at high resolution. In this context, we have focused on four main subjects with different, yet highly intertwined, objectives. The first subject will address the application of an energy-based computational predictor for the identification of possible antibody-binding surfaces (epitopes) of protein antigens from the pathogen Burkholderia pseudomallei, responsible for human melioidosis. The second will focus on the expansion of the same rationale, adapting the method towards different applications, and including as a novel functionality the prediction of MHC-II coupled epitopes to elicit the intervention of T helper cells. The third objective concerns the design and characterization of peptides and peptidomimetics to optimize the properties of the identified epitopes as better vaccine candidates. The fourth one will pursue the investigation of the energetic determinants of interacting proteins in a more general context (not limited to immunogenic epitopes), aiming at the identification of an energy-based property describing the interaction event at the atomistic level of resolution. This part of the project is aimed at the development of a computational tool based on such property to help improve the understanding of the determinants of protein interactions and help predict their binding interfaces and orientation. All four subjects have been investigated in the broad spectrum of activities of an academic consortium, devoted to the identification of antigens from B. pseudomallei showing sufficient immunogenic potential to be considered as components for a vaccine against the pathogen. The computational methods developed and tested within this framework have theoretical as well as practical implications, from the physico-chemical study and characterization of protein-protein interactions, to the design of biologically active molecules.