Molecular modeling and design of invariant chain peptides with altered dissociation kinetics from class II MHC.

We have used molecular modeling to design substitutions in an invariant chain-derived peptide (CLIP), so as to alter the stability of its complex with class II major histocompatibility complex (MHC) proteins. We sought first to test whether CLIP binds in the same way to different class II MHC proteins. We designed destabilizing substitutions of two residues (Met 91 and Met 99) previously predicted to act as the major anchor residues for binding to all class II MHC and measured their effect on CLIP's dissociation rate from a series of three murine I-A MHC proteins. Even a conservative substitution preserving size and hydrophobicity but reducing flexibility (leucine, a branched residue) caused large accelerations in dissociation rates (up to 25-fold) at either position in all three MHC alleles, supporting the consistent role of these positions as the major anchors for MHC binding. These data also support the view that the special flexibility of the methionine side chains at these positions is essential for binding to diverse MHC molecules. We also used molecular modeling to design allele-specific enhancements of peptide binding. Designed substitutions of CLIP Pro 96 by Ala (for Ad), Glu (Ak), and Tyr (Au) each yielded strong enhancement of binding (up to 128-fold) for their targeted allele and only moderate or destabilizing effects to the other alleles. These results demonstrate the accuracy of the molecular models and the predictive value of this modeling. Moreover, they provide strong evidence for the proposed general model of invariant chain association, indicating that it binds to all class II MHC in the same conformation.