Jagiellonian University – Medical College, Krakow, Poland
Irena Roterman-Konieczna heads the bioinformatics group at the Jagiellonian University Medical College and the Faculty of Physics, Astronomy and Applied Computer Science, Poland. Her background is the theoretical chemistry. She worked for Harold Scheraga group as the postdoc (Cornell University – Ithaca NY in 1987-1989). Her research focuses on bioinformatics with a specific orientation on protein structure prediction. Particularly the application of fuzzy oil drop model (Irena Roterman and Leszek Konieczny are the Authors) appears to be successful for 3D protein structure analysis including the active site recognition, protein-protein complexation and recently for recognition of amyloidogenesis mechanism.
Model to represent the influence of water environment on the protein structure is expressed by fuzzy oil drop model. The main assumption is that the idealized distribution of hydrophobicity in protein molecule can be expressed by 3D Gauss function (Fig.1. – blue line). According to this function the protein molecule encapsulated in ellipsoid represents the highest hydrophobicity in the center of molecule with gradual decrease together with the increase of distance versus the center reaching the zero-level on the surface of molecule in the distance of 3sigma in any direction. This organization of hydrophobicity makes the protein molecule perfectly well soluble however deprived of any form of activity. Activity requires the local elasticity (fluctuations) and generation of local force field for ligand (substrate) to be bound to protein molecule. This is why the local discrepancies versus the idealized hydrophobicity distribution are observed mainly in the area of biological activity. The discrepancies - local hydrophobicity excess or local hydrophobicity deficiency is interpreted as coded form for protein-protein complexation (mediated by exposed hydrophobicity) in the first case and ligand binding area in the second.
The local discrepancy versus the idealized distribution can be measured quantitatively applying the Kullback-Leibler entropy formula expressing the distance between two distributions: target one (idealized 3D Gauss function distribution) and the observed one as it appears in certain protein molecule (inter-residual hydrophobic interaction). This measurement allows comparison of structures of high topological similarity with however high dissimilarity in a sense of hydrophobicity distribution. This was shown in the cease of immunoglobulin and flavodoxin families. The fuzzy oil drop model was successfully applied for the construction of the model of amyloidogenecity. The encapsulation of the protein molecule in the shell of 3D Gauss function form ensures the generation of individual soluble units. The specific disorder versus the ordered one generates hydrophobicity distribution supporting the propagation of omplexation in one-dimensional form of fibrils.