Toward a Definition of Self: Proteomic Evaluation of the Class I Peptide Repertoire

MHC class I molecules present host- and pathogen-derived peptides for immune surveillance. Much attention is given to the search for viral and tumor nonself peptide epitopes, yet the question remains, “What is self?” Analyses of Edman motifs and of small sets of individual peptides suggest that the class I self repertoire consists of thousands of different peptides. However, there exists no systematic characterization of this self-peptide backdrop, causing the definition of class I-presented self to remain largely hypothetical. To better understand the breadth and nature of self proteins sampled by class I HLA, we sequenced >200 endogenously loaded HLA-B*1801 peptides from a human B cell line. Peptide-source proteins, ranging from actin-related protein 6 to zinc finger protein 147, possessed an assortment of biological and molecular functions. Major categories included binding proteins, catalytic proteins, and proteins involved in cell metabolism, growth, and maintenance. Genetically, peptides encoded by all chromosomes were presented. Statistical comparison of proteins presented by class I vs the human proteome provides empiric evidence that the range of proteins sampled by class I is relatively unbiased, with the exception of RNA-binding proteins that are over-represented in the class I peptide repertoire. These data show that, in this cell line, class I-presented self peptides represent a comprehensive and balanced summary of the proteomic content of the cell. Importantly, virus- and tumor-induced changes in virtually any cellular compartment or to any chromosome can be expected to be presented by class I molecules for immune recognition.