Large-scale characterization of natural ligands explains the unique gluten-binding properties of HLA-DQ2.

Celiac disease is an enteropathy caused by intolerance to dietary gluten. The disorder is strongly associated with DQA1*0501/DQB1*0201 (HLA-DQ2) as approximately 95% of celiac patients express this molecule. HLA-DQ2 has unique Ag-binding properties that allow it to present a diverse set of gluten peptides to gluten-reactive CD4+ T cells so instigating an inflammatory reaction. Previous work has indicated that the presence of negatively charged amino acids within gluten peptides is required for specific binding. This, however, only partly explains the scale of the interaction. We have now characterized 432 natural ligands of HLA-DQ2 representing length variants of 155 distinct sequences. The sequences were aligned and the binding cores were inferred. Analysis of the amino acid distribution of these cores demonstrated that negatively charged residues in HLA-DQ2-bound peptides are favored at virtually all positions. This contrasts with a more restricted presence of such amino acids in T cell epitopes from gluten. Yet, HLA-DQ2 was also found to display a strong preference for proline at several anchor and nonanchor positions that largely match the position of proline in gluten T cell epitopes. Consequently, the bias for proline at p6 and p8 facilitates the enzymatic conversion of glutamine into glutamic acid in gluten peptides at p4 and p6, two important anchor sites. These observations provide new insights in the unique ability of HLA-DQ2 to bind a large repertoire of glutamine- and proline-rich gluten peptides. This knowledge may be an important asset in the development of future treatment strategies.