Recognition of peptides bound to human leukocyte antigen (HLA) molecules by patrolling T-cells plays an important role in the body’s ability to detect and respond to potential threats such as disease or cancers. The repertoire of peptides presented by these HLA molecules is known as the immunopeptidome and are derived from endogenous proteins in HLA class I, and exogenous proteins in HLA class II pathways, respectively. Post-translational modifications (PTMs) such as glycosylation may affect the ability of a peptide to be presented by certain HLA allotypes or to be recognised by patrolling T-cells. Despite the ubiquity of glycosylation as a PTM for cellular proteins and the documented presentation of glycopeptides by class II HLA molecules, the presence of glycopeptides in the class I immunopeptidome has not been well characterised. This lack of characterisation of the class I glycoimmunopeptidome is primarily due to the significant technical challenges posed by analysis of glycoimmunopeptides using traditional mass spectrometry (MS) and computational methods.
We have developed a workflow to preprocess MS data, filtering out putative glycopeptide spectra based on oxonium ion presence and identifying the glycan modifications present in the dataset through computational methods we term “pseudoglycomic” analysis. This method allows for increased speed and accuracy in subsequent downstream searching of filtered spectra using traditional peptide search engine software. Use of this method facilitated a survey of multiple class I immunopeptide datasets which we conducted with the aim of exploring the glycopeptides presented by class I HLA molecules. This analysis found that not only do glycopeptides represent a significant and diverse repertoire within the class I immunopeptidome, but that the glycosylation of these peptides occurred primarily on the central amino acids of the immunopeptide rather than the terminal ends, suggesting these glycans are accessible for direct T cell receptor interactions. HLA-allotype specific effects on the glycan modifications present in the datasets were also observed, with the binding preferences of the HLA I allele influencing the types of glycosylation detected.
This analysis further establishes glycoimmunopeptides as a significant and important subset of the immunopeptidome, and our findings regarding the positional preference of HLA-peptide glycosylation and the effect of HLA allotype on glycopeptide presentation could have important implications for glycopeptide vaccine design or immunotherapies.