Bordner, Andrew J.

Abstract Self-tolerant T cells can cross-react weakly on self, but are usually incapable of activation. However, when activated these T cells can cross-react effectively on cells expressing self antigens. We have observed that mutants of the MHC class I molecule H2-Kb have structural differences resulting in altered antigen presentation. We hypothesize that structural manipulation of MHC class I can increase stabile amino acid contacts with the TCR. Increasing stabile contacts may decrease the energy barrier required to overcome the threshold for activation, breaking tolerance for self antigens. Using computer modeling techniques, amino acid substitutions at specific sites of MHC class I were predicted to result in stabilization of the TCR:pMHC complex. These altered MHC molecules were made and expressed in cell lines. We have shown that altered MHC enhanced TCR binding. Altered MHC enhanced activation and proliferation of naïve T cells in response to a non-stimulatory, weak antigen. Altered MHC also induced CTL killing against self-antigens. In an in vivo transplant model, protective immunity was induced against a native tumor after an altered MHC expressing tumor challenge. By altering the structure of MHC class I we have activated normally tolerant T cells that are capable of cross-reacting on self.

Significance Provision of sufficient activating signals can drive self-tolerant T cells to cross-react on self-peptide-MHC in a breakdown of tolerance. This approach has been applied in the development of heteroclitic peptide vaccines, which retain sufficient similarity to self to allow reactivation of effector cells on unaltered self-epitopes but require detailed knowledge about the behavior of individual altered self-peptide antigens. We designed altered-self MHC mutant antigen-presenting molecules to drive CTL capable of cross-reacting on self-peptide-MHC and demonstrate the breaking of self-tolerance to defined and undefined self-antigens. These findings bear particular importance in the field of cancer immunotherapy in which breaking tolerance to weakly antigenic cancer-associated antigens is fundamental to vaccination strategies.

Abstract Motivation: Specific non-covalent binding of metal ions and ligands, such as nucleotides and cofactors, is essential for the function of many proteins. Computational methods are useful for predicting the location of such binding sites when experimental information is lacking. Methods that use structural information, when available, are particularly promising since they can potentially identify non-contiguous binding motifs that cannot be found using only the amino acid sequence. Furthermore, a prediction method that can utilize low-resolution models is advantageous because high-resolution structures are available for only a relatively small fraction of proteins. Results: SitePredict is a machine learning-based method for predicting binding sites in protein structures for specific metal ions or small molecules. The method uses Random Forest classifiers trained on diverse residue-based site properties including spatial clustering of residue types and evolutionary conservation. SitePredict was tested by cross-validation on a set of known binding sites for six different metal ions and five different small molecules in a non-redundant set of protein–ligand complex structures. The prediction performance was good for all ligands considered, as reflected by AUC values of at least 0.8. Furthermore, a more realistic test on unbound structures showed only a slight decrease in the accuracy. The properties that contribute the most to the prediction accuracy of each ligand were also examined. Finally, examples of predicted binding sites in homology models and uncharacterized proteins are discussed. Availability: Binding site prediction results for all PDB protein structures and human protein homology models are available at http://sitepredict.org/. Contact: bordner.andrew@mayo.edu Supplementary information: Supplementary data are available at Bioinformatics online.