Immunoinformatics Prediction and Structure-Based Modeling of HLA-II Binding Epitopes of Iron Surface Determinant B (IsdB) Protein of Staphylococcus aureus

Abstract

Most of the attempts to develop a vaccine to prevent Staphylococcus aureus infections to date have not translated into clinical success. To date, all vaccine attempts have been based upon the development of opsonic antibodies. Research information suggests that cell-mediated immunity (CMI) may be critical for protection against Staphylococcus aureus. Iron surface determinant B (IsdB) protein is identified as a potent immunogen against Staphylococcus aureus. This protein, which scavenges iron from hemoglobin and myoglobin, is highly conserved among Staphylococcus aureus isolates from diverse clinical and taxonomical backgrounds including MRSA and VRSA and is expressed on the surface of all isolates tested so far. Adoptive transfer of CD4+ cells protected SCID (severe combined immunodeficiency) mice challenged with a lethal dose of S. aureus via tail vein while neither CD8+ T cells nor plasmacytes were protective. Immunization with IsdB was shown to be critical for activating the transferred cells as T cells from nonimmunized mice were not protective. To dissect the immune response further, it was shown that Th17 cells were necessary for IsdB-generated protective immunity. Th17 cells are a class of T helper cells that are activated upon encounter of an antigenic peptide coupled to a HLA (human leukocyte antigen) class II molecule. The HLA class II molecules are highly polymorphic. Immunoinformatics tools enable the prediction of T cell epitopes which can bind to the HLA-DR, HLA-DQ, and HLA-DP alleles using artificial neural networks. Here we describe a combined immunoinformatics and structure-based modeling approach for the prediction of T cell epitopes in the IsdB protein that can bind to 26 different alleles. The IsdB protein from S. aureus was analyzed computationally for the presence of HLA-II binding peptides. All possible overlapping 15-mer peptide sequences were generated in silico and analyzed for their ability to bind to 14 HLA-DR alleles covering 9 HLA super types, 6 HLA-DP alleles, and 6 HLA-DQ alleles. Approximately Twenty four, eight and six percents of the generated peptides binds to HLA-DR, HLA-DQ and HLA-DP alleles respectively with an IC-50 value less then 50 nM. The structural basis for recognition of these high-affinity peptides was studied using structural modeling of HLA class II peptide complexes, and there exists a good correlation between structural analysis and binding prediction.