In silico method development for epitope localization

Allergic diseases are a major worldwide health threat. A key step in the allergic type-I reaction cascade is the cross-linking of granulocyte-bound IgE antibodies by allergen molecules, therefore the characterization of the antibody-binding epitope regions on the surfaces of allergenic proteins is a major research aim, providing strategies both for diagnosis and therapy of type-I hypersensitivity. X-ray crystallography is a powerful tool for the structural elucidation of epitopes despite the lack of crystal structures for allergens in complex with IgE. For instance, structural models of allergens can be used for homology modeling of related allergens and to map linear (T-cell) epitopes to the allergen surfaces. Computational approaches include the development of epitope databases and the prediction of epitope features based on sequence motifs. The proposed project will be dedicated to the development and scientific application of a software tool for structure-based quantitative protein surface comparison. Thereby it will add a novel approach to the palette of prediction tools for conformational epitopes, which we call in-silico epitope localization. In a short-term project the program core was written and a feasibility study was performed. Currently the software is able to compare two protein surfaces on the basis of geometric parameters (triangular face distances and directional vectors). Preliminary results for the comparison of several PR-5 (TLP) family allergens confirm the principal functionality of the method. In the next stage electrostatic and hydrophobicity properties will be added to the surface analysis and the underlying global structure alignment will be complemented by a more sophisticated algorithm of multiple local superpositions. Furthermore, three-dimensional visualization of the comparison results will be provided, including the mapping of similarity scores to the surface models. In the future research phase of the project, the PR-5 allergen studies will be intensified and extended to other allergen families with several crystal structures available (Bet v 1 family and EF-hand proteins). Our investigations aim to find a correlation between cross-reactivity and local surface similarities of allergens within a family. The resulting similarity filter would enable us to define epitopes as very similar surface regions exclusive to cross- reactive allergens. Collaborating groups shall assess our predictions with experimental methods such as NMR, mass spectrometry, mutation studies and ultimately by crystallographic antibody complex characterization. To our knowledge, the approach of this project will be the first one to fully acknowledge the importance of structural surface analysis for the in-silico prediction of epitopes. Furthermore, the evaluation of the resulting epitope characteristics may reveal general rules of epitope composition and features, which would finally facilitate a method for direct and family-independent epitope prediction. The applicant is very experienced in crystallographic software development and well suited to add innovation in terms of in-silico methodology to the group of Prof. Keller, who has otherwise a great expertise in allergen crystallography.