Intrinsic sequence features are important for T-cell recognition. T lymphocytes (T cells) recognize only a small number of segments (antigenic sites) of a foreign protein, and only when the segment is presented on the surface of an antigen presenting cell in association with a major histocompatibility (MHC) molecule. Previously, by statistical methods we found properties of the known antigenic sites that distinguish them from the rest of the protein. The property that was strongly correlated with T-cell antigenicity was alpha-helical amphipaticity, and we developed a predicitive algorithm, based mainly on this property. We are currently interested in studying the sequence features that determine the MHC restriction. We hope to incorporate those properties into the predictive algorithm, so that in addition to prediction of immunodominance it will identify the MHC restriction of a peptide. The ability to better predict T-cell sites from the primary sequence will be important for the design of synthetic and recombinant vaccines. [See also report Z01 CB 04020-09 MET of Jay A. Berzofsky, with whom this work is closely coordinated.] Our approach has been to simulate the complex of antigen-MHC class II molecule, based on the known structure and published coordinates of the class I molecule. By repeating the analysis for all antigens that are restricted to a specific MHC, we can look for consistent rules, and then test our hypothesis by simulating the same antigens with other MHC types for which they are known not to be restricted. Recently, we have also applied a new method, developed for protein folding problems, to generate all possible conformations of an amino acid sequence in the limited space of the binding region.
