Detailed and atypical HLA-E peptide binding motifs revealed by a novel peptide exchange binding assay.

Lucy C Walters,Geraldine M Gillespie,Andrew J Mcmichael

doi:10.1002/eji.202048719

Abstract

Diverse SIV and HIV epitopes that bind the rhesus homolog of HLA-E, Mamu-E, have recently been identified in SIVvaccine studies using a recombinant Rhesus cytomegalovirus (RhCMV 68-1) vector, where unprecedented protection against SIV challenge was achieved. Additionally, several Mycobacterial peptides identified both algorithmically and following elution from infected cells, are presented to CD8+ T cells by HLA-E in humans. Yet, a comparative and comprehensive analysis of relative HLA-E peptide binding strength via a reliable, high throughput in vitro assay is currently lacking. To address this, we developed and optimized a novel, highly sensitive peptide exchange ELISA-based assay that relatively quantitates peptide binding to HLA-E. Using this approach, we screened multiple peptides, including peptide panels derived from HIV, SIV, and Mtb predicted to bind HLA-E. Our results indicate that although HLA-E preferentially accommodates canonical MHC class I leader peptides, many non-canonical, sequence diverse, pathogen-derived peptides also bind HLA-E, albeit generally with lower relative binding strength. Additionally, our screens demonstrate that the majority of peptides tested, including some key Mtb and SIV epitopes that have been shown to elicit strong Mamu-E-restricted T cell responses, either bind HLA-E extremely weakly or give signals that are indistinguishable from the negative, peptide-free controls.

Highlights

It is well established that HLA-E preferentially accommodates a signal peptide comprising residues 3–11 of MHC class Ia leader sequences, typically VMAPRT (L/V)(V/L/F)L (VL9), in its binding groove and that these peptides dominate the HLA-E-presented ligandome in the steady state [1,2]
In the original peptide binding studies, the HLA-E sequence motif was explored via Ala and Gly substitution experiments along the HLA-B*0801 leader peptide (VMAPRTVLL), for which strong sequence selectivity was uncovered at primary anchor positions 2 (Met) and 9 (Leu), with weaker preferences observed for secondary anchor positions 7 (Val) and 3 (Ala) [1]
Three distinct buffer systems were trialed in the photo-assisted peptide exchange reaction including a Lutrol-Tris-maleate-based buffer previously used in the micro-refolding peptide binding assay [22,7,20], a Tris-based buffer previously used for UV-induced peptide exchange with classical MHC class I molecules [23], www.eji-journal.eu

Summary

Introduction

It is well established that HLA-E preferentially accommodates a signal peptide comprising residues 3–11 of MHC class Ia leader sequences, typically VMAPRT (L/V)(V/L/F)L (VL9), in its binding groove and that these peptides dominate the HLA-E-presented ligandome in the steady state [1,2]. HLA-E- and Qa-1-restricted peptides from pathogens including Salmonella [10,11], Hepatitis C virus [12], EBV [13], and Influenza virus [13], have been reported. These findings extend to self-derived peptides from proteins such as Fam49B [14], Hsp60 [15], and Prdx5 [16] in the context of defective antigen processing, cellular stress, or autoimmune disease

Methods

Results

Conclusion