Structural Mapping of MHCII-Eluted Peptides to their Source Proteins: A Preliminary Survey of the Effect of Structure on Immunodominance

Abstract

The prediction of which peptides would bind to major histocompatibility complex class II molecules (MHCII) has long been of interest as a method of predicting immunodominance. Most approaches for MHCII binding prediction have been sequence-based (i.e. predicting which amino acids are optimal binders to the MHCII binding groove). Alternatively, a structure-based view for MHCII-peptide binding may be taken. Recent studies suggest that the population of peptides available for binding MHCII may be dependent on exposed protease cleavage sites in the protein structure at different stages of antigen processing. Thus, we hypothesized that a survey of MHCII-binding peptides would reveal them to be clustered at a particular depth in their respective undigested source proteins as these areas would be exposed at appropriate times for loading into MHCII. A population of 127 documented MHC-II binding peptide sequences with available source protein structure files in the RCSB Protein Data Bank was used for this study. An algorithm was constructed to determine the relative position of these peptides in their source protein. The surface of each source protein was estimated by computing the convex hull of their alpha-carbon coordinates. The distance of each eluted peptide from its hull was then computed (d: distance of peptide to protein surface) and normalized against the distance from the center of the hull (S: percent submergence in source protein). In our preliminary survey, 53% of the samples had d-values of 1A or less, with 51% having S-values of less that 15%, indicating that peptides that are selected as MHCII binders are relatively near the surface of their undigested source proteins. Our preliminary results indicate that timing, as well as affinity, may be equally important in MHCII binding.