A common network of residue-residue contacts underlies interactions between peptides and HLA class II complex

Abstract

The formation of a stable antigenic peptide-HLA II (p-HLA class II) complex is a critical early step in the adaptive immune response. In this work, we identify the residue-residue contacts that ‘anchor’ the peptide between the alpha and beta chains of HLA II and examine whether the anchoring residue-residue contacts are shared among different p-HLA II complexes. We hypothesize that there are similarities between the contact map of the alpha and beta chains of HLA II with CLIP (the fragment of the invariant gamma chain that binds to newly synthesized HLA II molecules) and the contact maps between the different alpha and beta chains of HLA II molecules with various antigenic peptides. To test the hypothesis, 81 diverse peptide-HLA II DR and DQ complexes, including CLIP-HLA II complex, were selected from the PDB database, and ‘Unified Residue Numbering’ was introduced for all complexes. The Unified Residue Numbering enables us to compare residue contacts across complexes for each position, e.g., to identify the position in peptides occupied by residues with the highest number of contacts with HLA II similar to CLIP position with the highest number of contacts in all structures, and to define characteristics of residues for each position. We also identified all ‘similar contacts’ in the analyzed structures. ‘Similar contacts’ are defined as contacts between same-numbered residues in the peptide and HLA II and CLIP with HLA II structure independent of physicochemical properties of residues involved in the contact. We found that in the 81 analyzed structures, 90% of contacts between peptide and the alpha chains were ‘similar contacts’, as were 80% of contacts between peptides and the beta chains. Thus, our approach to sequence alignment, which is based on alignment of similar contacts rather than similar residues, allows one to define the common network of residues that underlies the interactions between peptide and HLA II. We also consider several criteria for the specificity of antigen peptide loading into HLA II based on the structural and physicochemical characteristics of the residues involved in the ‘anchoring’ contacts. These data may be useful for refining existing computational algorithms that predict peptide interactions with HLA II complexes.