This issue’s Pillars of Immunology article (1) describes the characterization by mass spectrometry of endogenous peptides eluted from the class I MHC molecule HLA-A2.1 (A*0201). This was the first contribution from the collaboration between Don Hunt and Vic Engelhard on the identification of MHC-bound peptides, which soon would yield seminal work on tumor-specific CTL recognition (2, 3, 4), nonclassical MHC proteins (5), posttranslational modification of MHC-bound Ags (6, 7, 8), and minor histocompatibility Ags and graft-vs-host disease (9, 10).
The work was a tour de force in microchemical analysis and set the standard for ultrasensitive detection and characterization of biological peptides. The method used relied on a combination of several recent technical advances in mass spectrometry. First and most important was the development of tandem mass spectrometry (11). In this technique, two mass analyzers are used in series, the first to select particular molecules from a mixture by virtue of their exact masses, and the second to analyze the pattern of ions resulting from fragmentation of the selected molecule. Don Hunt had recently adapted this technique for peptide sequencing by building a triple quadrupole instrument that interposed a middle chamber filled with argon atoms to fragment peptide ions selected in the first chamber that could then be analyzed in the third chamber (12). Fragmentation occurred mostly at amide bonds, as in earlier chemical fragmentation approaches used for mass spectrometry sequencing, so that the sequence could be read from the pattern of fragment ions detected in the third chamber. The high sensitivity of mass spectrometric detection facilitated the analysis of minute quantities of molecules such as low abundance peptides eluted from MHC proteins. The second development was that of electrospray ionization, for which John B. Fenn would later be awarded the 2002 Nobel Prize in Chemistry (13). This technique allowed samples to be introduced continuously into the mass spectrometer. In earlier ionization methods, samples were individually mixed with volatilization-promoting compounds and introduced one at a time into the spectrometer. The volatile acidic solvents typically used for reverse phase HPLC proved to be a perfect match for electrospray ionization. Complex mixtures of peptides separated by HPLC could be fed directly into the tandem electrospray ionization mass spectrometer, allowing mixtures of unprecedented complexity to be analyzed. To fractionate the small quantities of peptide eluted from MHC molecules with minimal loss, Hunt et al. developed a microcapillary reverse phase HPLC column (75-μm diameter) for use in-line with the mass spectrometer. Immunologists who heard Don Hunt’s lectures from that time might recall marveling at his descriptions of these hair-thin columns and their resolving power. The same group would later develop a microcapillary effluent splitter in which a fraction of the HPLC eluent was diverted in-line to a microtiter plate for CTL assay (3). This approach allowed the same fraction to be analyzed for biological activity and molecular mass, and led to the identification of the peptides recognized by CTL specific to tumor Ags, alloantigens, and viral and bacterial epitopes (reviewed in Ref. 14). In this article, it was not T cell epitopes that were of interest but the set of naturally processed peptides presented by HLA-A2.1 on a normal, noninfected cell. Endogenous protein sources for these peptides were identified by matching the experimentally determined peptide sequences to protein and nucleic acid sequence databases, the third development on which this Pillar of Immunology rests. Only four of the 19 fully or partially sequenced peptides identified by Hunt et al. corresponded to peptides in the then-available database. At the time, GenBank had <0.1% of its current entries (15), basic local alignment search tool (BLAST) had just been developed (16), and the Human Genome Project was just getting off the ground (17). The use of database analysis in the identification of proteins from partial sequence data was to explode in a few years with increased sequence coverage in databases and with the development of two-dimensional gel tryptic peptide analysis (18).
This work helped clarify aspects of MHC-peptide interactions and Ag processing pathways, the understanding of which was then rapidly evolving. The year before (another Pillars of Immunology article, Ref. 19), Hans-Georg Rammensee’s group reported pooled sequencing of peptides eluted from several class I MHC proteins, including HLA-A2.1, with identification of allele-specific peptide binding motifs (20). Consistent with the Rammensee motif, Hunt et al. found Leu (or Ile, which is indistinguishable from Leu by mass) at the P2 position of each sequenced peptide, but at the C-terminal position they found Leu and Ala in addition to Val, which had been identified previously as the dominant anchor residue, and they observed no preferences at all in the other positions. Thus, even this small set of sequenced peptides considerably broadened the understanding of MHC peptide-binding specificity. Hunt et al. noted that their sequencing data were consistent with the pockets observed in the refined 2.6-Å crystallographic model of HLA-A2.1 bound to a mixture of peptides that had appeared the previous year (21). A high-resolution structure sufficient to characterize the molecular basis for tight peptide-MHC binding would not appear until later that year, aided by the strict peptide binding motif for HLA-B27, which restricted the complexity of the bound peptide mixture (22). In addition to helping characterize MHC sequence preferences, the identification of the protein sources for the endogenous peptides helped to illuminate class I MHC Ag processing pathways. Six months earlier, Don Wiley’s group had reported the identification of 11 self-peptides bound to HLA-B27 that had been sequenced using conventional Edman degradation (23). All were from abundant cytosolic or nuclear proteins. With one exception, the peptides identified by Hunt et al. also derived from cytosolic and nuclear proteins, consistent with the developing understanding of the importance of TAP (24, 25) and proteasomal processing in the generation of class I MHC ligands (26, 27, 28).
One of the identified naturally processed peptides that Hunt et al. found associated with HLA-A2.1 derived from the endoplasmic reticulum-targeting signal sequence of IP-30. IP-30 (a.k.a.GILT, a γ-IFN-induced lysosomal thiol reductase) was neither cytosolic nor nuclear and was thus at odds with contemporary understanding of the Ag processing pathways that supply peptides to nascent class I MHC proteins. A companion paper in the same issue of Science (29) reported that all HLA-A2.1-bound peptides identified in the Ag processing-deficient cell line T2 were derived from such signal sequences and suggested a TAP-independent alternate peptide-loading pathway for class I MHC proteins. This work presaged innate recognition by NK cells of signal sequence-derived peptides bound to the nonclassical class I MHC proteins HLA-E and Qa-1 (30, 31). Moreover, many of the peptides presented in the mutant cells were longer than nine residues, potentially broadening further the set of potential CD8+ T cell epitopes. Presentation of such noncanonical long peptides on class I MHC proteins continues to be a topic of much interest (32, 33).
The work of Hunt et al. also provided the first quantitative estimates of the complexity of the repertoire of peptides presented by class I MHC molecules to the immune system. Two hundred individual peptides were detected from ∼15 pmol (10–20 ng) of eluted material. Based on the prevalence of the identified peptides and the fraction of the total amount of material in the eluted peptide mixture, Hunt et al. were able to estimate that the total number of different peptides presented by HLA-A2 could easily exceed 1000, with most of the peptides present in 100 or fewer copies per cell. This suggested a large and complex “MHC-restricted immunome” that continues to this day to be identified, catalogued (34, 35), and mined in many aspects of basic and applied immunological research.
