Discriminated selection among viral peptides with the appropriate anchor residues: implications for the size of the cytotoxic T-lymphocyte repertoire and control of viral infection.

Structural characterization of peptides restricted by major histocompatibility complex (MHC) class I molecules has identified residues critical for MHC class I binding and for T-cell receptor recognition. For example, optimal peptides fitting into the murine MHC class I Db groove are 9 to 11 amino acids long and require as MHC anchor residues an Asn (N) at position 5 and also either a hydrophobic residue, a Met (M) or a Cys (C), at the carboxy terminus. The three known Db-restricted peptides of lymphocytic choriomeningitis virus (LCMV) are glycoproteins GP1 (amino acids [aa] 33 KAVYNFATC), GP2 (aa 276 SGVENPGGYCL), and nucleoprotein NP (aa 396 FQPQNGQFI). In addition to these two GP and one NP peptides, computer search revealed 11 other GP peptide sequences and 20 additional NP sequences that contained the Db binding motif. By Db competitive binding analysis, only two of these 11 GP peptides and 1 of these 20 NP peptides bound to the MHC Db molecule with an affinity equivalent to the measured affinities for the three known GP1, GP2, and NP cytotoxic T-lymphocyte (CTL) epitopes. No CTL specific for these three peptides were generated when H-2b mice were inoculated with viral variants in which either the two known GP epitopes (GP1 and GP2; termed GPV) or the GPV and NP epitopes (termed GPV + NPV) were mutated. However, a novel CD8+ anti-LCMV CTL response ordinarily not seen in H-2b mice inoculated with wild-type virus was noted when such mice were inoculated with the GPV + NPV-mutated variant. This result indicates that (i) despite large numbers of peptides containing the appropriate anchor residues within a viral protein, only a restricted number induce CTL, thereby maintaining a limited CTL repertoire, (ii) despite the limited repertoire, the immune system retains the flexibility to generate an immune response(s) to a previously silent protein(s), suggesting a hierarchial control mechanism, and (iii) identification of a primary amino acid sequence is not sufficient, per se, to predict CTL epitopes, and peptide conformations are likely more complex than indicated by simple linear sequence comparisons.