It has been established that CD1 molecules are capable of binding microbial lipopeptides, such as CD1a-presented dideoxymycobactin, to mediate lipopeptide-specific T-cell responses. Recent evidence, however, raised the possibility that a fraction of MHC class I molecules has evolved the ability to bind viral lipopeptides and activate lipopeptide-specific CTLs. Rhesus MHC class I molecules, Mamu-B*098 and Mamu-B*05104, bind N-terminal lipopeptide fragments derived from the N-myristoylated SIV Nef protein and elicit lipopeptide-specific CTL responses. A series of X-ray co-crystallographic analyses of these lipopeptide-presenting MHC class I molecules, collectively termed LP1, provided a clue to how lipopeptides are captured by LP1 and how TCRs recognize lipopeptides. Mamu-B*05104, for example, bind N-myristoylated 4-mer lipopeptides (C14-Gly1-Gly2-Ala3-Ile4). The overall 6-pocket structure constructed in the antigen-binding groove is comparable with that of conventional MHC class I molecules; however, unusually large, hydrophobic B pocket accommodates the acyl chain while the C-terminal Ile4 residue anchors at the F pocket. TCRs interact specifically with the amide bond of C14-Gly1, which is unique to lipopeptide ligands.
We now set out to address whether such LP1 molecules may exist in humans. We observed that ligand-induced stable complex formation occurred for HLA-A*24:02 in the presence of not only peptides but also lipopeptides. X-ray crystallographic analyses of HLA-A*24:02 in a form complexed with either peptides or lipopeptides indicated that its B pocket exhibited plasticity upon ligand binding, and hydrogen-bond network and van der Waals interactions were optimized for binding each of the anchoring elements (P2 of peptides and the myristic acid of lipopeptides). The capacity of HLA-A*24:02 to bind both peptides and lipopeptides appears advantageous for host defense, which may account for its marked prevalence worldwide.