Abstract
The goal of this project is to study the protein interfaces characteristics, the area responsible for macromolecular communication and function. We propose to use an approach of studing the phyical-chemical and structural properties of the interface forming residues of known complexes with experimental resolved quartenary structure, extracting the parameters that describe in an objective way the differences between which complex classes. Protein-protein interactions are responsible for the control of, virtually, all the processes within cells, for example, signal transduction, metabolic and gene regulation, immunologic response, etc., which are very specific reactions. Even in intracellular crowded environment, the protein-protein interaction between two proteins are mediated through a specif area on the proteins surfaces. Many experimental methods are able to detect protein interaction from the cellular to the atomic resolution range. Nevertheless the combination of such experiments that give us a more complete understanding of the protein interaction network, are time costing, especially during the sample preparation. Such a bottleneck of the techniques makes the in silico approach very usefull, and it can guide the in vivo and in vitro experiments, reducing the time and money costs. In addition to difficulties while obtaining 3D structures (basically through the x-ray diffraction and nuclear magnetic resonance techniques), docking algorithms also are only able to predict atomic details of protein complexes if the sequence identity is higher than » 60%, bellow this mark (between 30 and 60%), only the structural similarity are conserved, and details such as residues pairs are not predicted correctly, and at last, when the similarity is under 30% similarity, no reliable model is obtained, and only the molecules relative orientation is predicted. To improve the quality and generation of more reliable in silico models, a better and accurate understanding of the protein-protein interfaces' principal characteristics is needed. We propose a study to determine the essential physical-chemical characteristics of the Interface Forming Residues (IFR) that distinguish them from the free surface residues, which are not involved in any inter-protein interactions. Using the entries in the PDB (Protein Data Bank) that are known to react with other proteins, we are going to use the BlueStar Sting software, as well as its data base (STING_DB) (developed by the Embrapa's Structural Bioinformatic's Group, and available at http://www.cbi.cnptia.embrapa.br/SMS) to determine the basic characteristics of the IFR. STING_DB has more than 700 descriptors for each amino acid of all the protein chains in the PDB, and due to facilities of its search method it becomes a powerful tool to select and identify the important residues, as we propose. For instance, if we consider a mean value of 250 amino acids per resolved protein in the PDB (with roughly 55,000 entries), with 2 chains per entry, there is a total of 2x1010 data in STING_DB! This show us the importance of using a computational biology/bioinformatic approach for analyzing successfully the macromolecules structure and interactions studies.In the first step of this project, there is a need to establish a non redundant dataset with oligomeric proteins. This is mainly due to existence of the high number of ambiguous entries in the PDB, where a specific protein may have other entries with single point mutations in its sequence (Alanine Scanning Studies, for example), or it may have some highly homologous protein (sequence similarity above 90%). After that, it is extremely important to differentiate between hetero- and homo-oligomers, because of the difference among these two interaction classes. Each class can also be separated into other subclasses, according to some characteristics (interface symmetry, for example. (AU)
|