Systematic characterisation of cellular localisation and expression profiles of proteins containing MHC ligands

Abstract

Background: Presentation of peptides on Major Histocompatibility Complex (MHC) molecules is the cornerstone in immune system activation and increased knowledge of the characteristics of MHC ligands and their source proteins is highly desirable.

Methodology/principal finding: In the present large-scale study, we used a large data set of proteins containing experimentally identified MHC class I or II ligands and examined the proteins according to their expression profiles at the mRNA level and their Gene Ontology (GO) classification within the cellular component ontology. Proteins encoded by highly abundant mRNA were found to be much more likely to be the source of MHC ligands. Of the 2.5% most abundant mRNAs as much as 41% of the proteins encoded by these mRNAs contained MHC class I ligands. For proteins containing MHC class II ligands, the corresponding percentage was 11%. Furthermore, we found that most proteins containing MHC class I ligands were localised to the intracellular parts of the cell including the cytoplasm and nucleus. MHC class II ligand donors were, on the other hand, mostly membrane proteins.

Conclusions/significance: The results contribute to the ongoing debate concerning the nature of MHC ligand-containing proteins and can be used to extend the existing methods for MHC ligand predictions by including the source protein's localisation and expression profile. Improving the current methods is important in the growing quest for epitopes that can be used for vaccine or diagnostic purposes, especially when it comes to large DNA viruses and cancer.

Figure 1. Overview of GO cellular component terms enriched among proteins containing MHC ligands.

The results from an enrichment analysis were superimposed onto the GO tree structure; A: MHC class I ligand-containing proteins (MHCI data set) B: MHC class II ligand-containing proteins (MHCII data set). Only nodes representing the most significant terms are included (p-value cut-offs used as inclusion criteria is 1.00*10⁻¹³ for the MHCI data set and 0.01 for the MHCII data set). The significance level is reflected by the node colour, where red corresponds to the most significant p-values while grey indicates no enrichment. The size of the nodes reflects the number of proteins assigned to this term.

Figure 2. Distribution of MHC ligand-containing proteins relative to their mRNA expression level.

The proteins were grouped into bins of equal size, such that each bin contains 2.5% of all proteins in the data set. As a result, each bin comprises an equal number of proteins, increasing in mRNA expression level from left to right. The height of each bar represents the fraction of proteins that contain MHC ligands. A: Fraction of proteins that contain MHC class I ligands versus the mRNA expression level of the proteins according to the GNF gene expression database B: Fraction of proteins that contain MHC class II ligands versus the mRNA expression level of the proteins according to the GNF gene expression database.