Template-based peptide modeling for celiac risk assessment of newly expressed proteins in GM crops

Newly expressed proteins in genetically modified (GM) crops are subject to celiac disease risk assessment according to EFSA guidelines. Amino acid identity matches between short peptides (9aa) and known celiac restricted epitopes are required to be further evaluated through peptide modeling; however, validated methods and criteria are not yet available. In this investigation, several structures of HLA-DQ2.5/peptide/TCR (T-cell receptor) complexes were analyzed and two template-based peptide molding software packages were evaluated using various peptides including ones not associated with celiac disease. Structural characterization indicates that residues at P(position)1, P2, P5, P8, and P9 in the 9aa restricted epitopes also contribute to the binding of celiac peptides to the HLA-DQ2.5 antigen in addition to the presence of the motif Q/EX₁PX₂ starting at P4 or P6. The recognition of the HLA-DQ2.5/peptide complex by TCR is through specific interactions between the residues in the restricted epitopes and some loop structures in the TCR. The template-based software package GalaxyPepDock seems to be suitable for the application of peptide modeling when an estimated accuracy value of >0.95 combined with >160 interaction similarity score are used as a threshold for biologically meaningful in silico binding. Nevertheless, caution should be exercised when applying peptide modeling to celiac disease risk assessment until methods are rigorously validated and further evaluated to demonstrate its value in the risk assessment of newly expressed proteins in GM crops.