MHC class I (MHC I) molecules are constitutively expressed on the surface of nearly all nucleated cells in jawed vertebrates. MHC I molecules are noncovalently associated trimers consisting of a polymorphic heavy chain, β2m, and an oligopeptide. The MHC heavy chain folds into two alpha helices and a beta sheet, creating the walls and floor, respectively, of a peptide-binding cleft . Hydrogen bonding between conserved heavy-chain cleft residues and the N- and C-termini of associated peptides typically limit their length to 8–13 residues (though exceptions abound ), and strictly dictate peptide directionality. Additionally, six defined pockets (termed A–F) in the cleft confer specificity for peptide side chains oriented toward the groove . Extensive polymorphism between the binding grooves of different MHC allomorphs (>7000 known alleles (and climbing) in human populations with up to six allomorphs expressed per person from the HLA-A, -B, and -C loci (http://hla.alleles.org/nomenclature/stats.html)), ensures that a wide spectrum of peptides is presented to the immune system, essentially preventing pathogen escape at the population level.
Despite allelic preferences, common themes guide peptide/MHC (pMHC) binding. Pooled aa sequencing of peptides eluted from many different individual class I allomorphs revealed residues overrepresented at each position [4, 5], though notably, these allele-specific peptide-binding motifs are also influenced by peptide liberation, transport, and trimming (reviewed in ). Most peptide pools exhibit highly dominant specific aas (or chemically similar aas) at or near their N- and C-termini [4, 5]. These “anchor” residues greatly influence peptide-binding affinity. The more N-terminal anchor is typically located at peptide position 2 or 3 (denoted as p2 and p3) and is accommodated by the B-pocket of the peptide-binding cleft, though it can be located up to p5 (C-pocket, as is the case for the mouse H-2Db allomorph ). The deep F-pocket cradles the C-terminal anchor, typically an aliphatic or aromatic residue for mouse class I allomorphs (some human allomorphs also favor basic C-terminal anchors). Predictably, detailed peptide mapping  and high-throughput mass spectrometry  identify numerous high-affinity peptides that break these simple rules, increasing both the size of the immunopeptidome and the difficulty of in silico peptide-binding prediction. Class I molecules present tens of thousands of different self-peptides among approximately 105 pMHC complexes on the surface of each cell , consistent with their role in tumor immunosurveillance.
How does the cell supply this diverse array of pMHCs? Most MHC class I peptides arise from rapidly degraded polypeptides, ensuring representation among the translatome independently of protein stability and minimizing the time to detect viral translation . To enhance immunosurveillance of tumor-associated Ags (TAAs), ribosome subpopulation sampling [11, 12] likely enables surveillance of low abundance bona fide and defective mRNAs [13, 14]. TAA–peptide abundance is critical, since many TAAs derive from nonmutated genes and are thus recognized by low-affinity T cells that escape self-tolerance .
The affinity of peptides for MHC governs the stability of complexes and hence levels of cell surface expression . Consequently, the immunogenicity of pMHCs can potentially be improved by increasing peptide affinity . Although peptide-binding algorithms have greatly enhanced rational peptide design, they are far from perfect. Further, despite their orientation away from the T-cell receptor (TCR), anchor residue substitutions can change pMHC conformation to negatively impact TCR recognition. What is needed then is a bit of magic: a general method for increasing peptide affinity while minimizing changes in TCR specificity.
In this issue of the European Journal of Immunology, while seeking to improve the CD8+ T-cell response to the melanocyte differentiation Ag Gp100, Uchtenhagen et al.  appear to achieve the impossible, or at least the improbable. Gp100 expression is greatly enhanced in melanoma, making it an attractive therapeutic vaccine target. Human Gp10025–33 peptide (KVPRNQDWL (KVP)) presented by the mouse class I Db allomorph elicits self-reactive mouse CD8+ T cells, while the orthologous mouse peptide (EGSRNQDWL (EGS)) does not (Fig. 1) [19, 20]. Both peptides possess canonical p5N and p9L anchor residues for Db (which has a motif of XXXX NXXX[IML], where X represents any aa, N = asparagine, I = isoleucine, M = methionine, L = leucine) . Despite identical anchors, EGS binds Db with 100-fold lower affinity than KVP , evincing the contribution of nonanchor residues to Db binding [21, 22]. Systematic crosswise substitution of p1–3 between KVP and EVS revealed greatly enhanced peptide binding ) and pMHC stability when simply replacing p3 of EGS with proline (Pro; EGPRNQDWL (EGP)) . Immunization with EGP elicited higher numbers of EGS-specific CD8+ T cells than EGS itself, and critically, protected against tumor challenge while the homologous peptide did not .
Uchtenhagen et al.  scrutinized the structural basis for enhanced EGP peptide affinity with surprising and potentially generally applicable findings. X-ray crystallography of Db complexed with Gp100 peptides KVP, EGS, or EGP revealed a conserved peptide conformation and similar peptide- Db hydrogen bonding in each complex . Thus, the EGP's increased affinity was not due to large structural alterations in the complex. Notably, in EGP, the pyrrilodine ring of p3P and the hydroxyphenyl group of Db-Y159 formed CH-π interactions, which affords substantial intermolecular-binding energy [24, 25] (and see http://www.tim.hi-ho.ne.jp/dionisio/page/whatis.html).
To examine the contribution of CH-π interactions (which occur with aromatic residues) to EGP/Db stability, Uchtenhagen et al. substituted Y159 with either another aromatic (F) or aliphatic residues with a short (A) or long (L) side chain. Intriguingly, the enhanced pMHC stability of EGP versus EGS was abrogated with Db-Y159A or Db-Y159L. An intermediate effect was observed with Y159F, consistent with reduced energetic stabilization of Phe-Pro CH-π interactions compared with that of Tyr-Pro. To confirm CH-π interactions between p3P and Y159F, the authors crystallized EGP/Db–Y159F and found nearly identical peptide and 159 side-chain orientation as in WT Db.
The most important aspect of the study is the effect of the CH-π interaction on TCR recognition of the modified peptide. EGP/Db complexes bind better to the cognate TCRs than complexes with WT peptide, providing a double advantage of the Pro substitution. To gain insight into this effect, Uchtenhagen et al. used high-powered computers to simulate the simultaneous movements of individual atoms of the structure. Such “molecular dynamics” analysis suggests that increased TCR affinity results from increased rigidity of the peptide within the Db cleft.
As with all good science, discoveries beget questions. Most pragmatically, can the increased pMHC affinity, pMHC stabilization, and TCR recognition afforded by the p3P substitution be generally extended to other peptide/MHC combinations for enhanced vaccine efficacy? Previous work by Achour and colleagues suggests that p3P altered peptides bind to Db or Kb with increased affinity . Since Y159 is highly conserved among human HLA genes and alleles, this likely applies to human pMHC complexes, particularly for those unusual allomorphs that do not bind with strong p2 anchors (such as B*0801). Can other aromatic residues within the peptide-binding cleft be exploited for CH-π interactions, and if so, will tertiary structure be preserved to maintain TCR recognition? Is increased peptide rigidity generally positive for TCR recognition? Does increased binding uniformly extend to endogenous peptides when they are loaded on to class I in the ER by the peptide-loading complex? Although binding of exogenous peptides to class I is generally considered to precisely mimic the binding of endogenous peptides, peptides can bind to class II in multiple conformations, depending on how they are loaded, with major biological consequences .
The work of Uchtenhagen et al.  beautifully illustrates the importance of continued research on problems thought to be “solved”. It is essential for young scientists in particular to appreciate that nature's secrets are boundless, and that the critical information for practical applications often springs from surprising sources that are best accessed by curiosity-driven research.