Preferred size of peptides that bind to H-2 Kb is sequence dependent

The identification of naturally processed viral peptides reveals that major histocompatibility complex (MHC) class I epitopes are composed of nine or eight amino acid residues. Peptides eluted from H-2 Kb MHC class I molecules have been suggested, as a class, to be eight amino acid residues long. To assay for peptide-class I interactions, a stabilization assay described for surface labeled "empty" class I molecules was employed, but on biosynthetically labeled class I molecules. The Sendai virus nucleoprotein-derived octapeptide APGNYPAL does not bind and stabilize Kb molecules, whereas other octameric Kb-restricted peptides and the nonameric peptide FAPGNYPAL interact stably. We attribute the failure of Sendai octamer binding to the presence of proline in position two: replacement of proline renders the resulting octamers as efficient as FAPGNYPAL for binding and stabilization of H-2 Kb. Substitution of glycine in position three of APGNYPAL slightly improves its Kb stabilizing capacity. Iodination of the tyrosine residue significantly alters the binding properties of the nonamer peptide. We conclude that the length of epitopes as selected by the class I Kb molecule is influenced by their sequence and suggest that proper positioning of the NH2 terminus of peptides is essential for class I stabilizing properties. The ability to stabilize newly synthesized "empty" class I molecules with peptide argues against an involvement of beta 2 microglobulin exchange in the experiments described here.