Genome-wide Characterization of a Viral Cytotoxic T Lymphocyte Epitope Repertoire

Weimin Zhong,Pedro A Reche,Char-Chang Lai,Bruce Reinhold,Ellis L Reinherz

doi:10.1074/jbc.m307417200

Abstract

A genome-wide search using major histocompatibility complex (MHC) class I binding and proteosome cleavage site algorithms identified 101 influenza A PR8 virus-derived peptides as potential epitopes for CD8+ T cell recognition in the H-2b mouse. Cytokine-based flow cytometry, ELISPOT, and cytotoxic T lymphocyte assays reveal that 16 are recognized by CD8+ T cells recovered directly ex vivo from infected animals, accounting for greater than 70% of CD8+ T cells recruited to lung after primary infection. Only six of the 22 highest affinity MHC class I binding peptides comprise cytotoxic T lymphocyte epitopes. The remaining non-immunogenic peptides have equivalent MHC affinity and MHC-peptide complex half-lives, eliciting T cell responses when given in adjuvant and with T cell receptor-ligand avidity comparable with their immunogenic counterparts. As revealed by a novel high sensitivity nanospray tandem mass spectrometry methodology, failure to process those predicted epitopes may contribute significantly to the absent response. These results have important implications for rationale design of CD8+ T cell vaccines.

Highlights

Antigen-specific recognition of peptides bound to major histocompatibility complex class I (MHCI)1 molecules is mediated by CD8ϩ T lymphocytes
Computational Prediction of All Potential H-2 Db- and Kbrestricted cytotoxic T lymphocyte (CTL) Epitopes Encoded by the Influenza PR8 Virus Genome—Two important biological processes determine whether a peptide segment within a native protein can be naturally presented by a given MHCI molecule: 1) proteolytic processing within the cell required to liberate the peptide [22]; and 2) binding of the excised peptide to the MHCI molecule with sufficient affinity to create a stable epitope as a cell surface pMHC for CTL recognition [23]
A Peptides in bold indicate the known immunogenic CD8ϩ T cell epitopes. b Proteasomal cleavage possibility predicted by Probabilistic Model of Proteosomal Cleavage (PMPC) algorithm. ϩ, cleavage predicted; Ϫ, cleavage not predicted; Not considered, not among the top 1.6% of the potential MHCI binders when examined by either SYFPEITHI or RANKPEP algorithm, because no Kb 9-mers were assessed. c Peptide concentration that stabilize a half-maximal number of MHCI molecules, as determined by RMA-S assay

Summary

EXPERIMENTAL PROCEDURES

CTL Epitope Prediction—All potential T cell epitopes encoded within the influenza A virus (strain A/Puerto Rico/8/34; PR8) proteome were predicted using the motif matrix-based algorithm SYFPEITHI [16] (syfpeithi.bmi-heidelberg.com/scripts/MHCServer.dll/home.htm). BAL cells recovered from B6 mice infected with the virus for 10 days were stained with phycoerythrin-conjugated NP366–374/Db and PA224–233/Db tetrameric reagents (The Trudeau Institute, Saranac Lake, NY) and Cy-Chrome-conjugated anti-mouse CD8 ␣ mAb (BD Biosciences). To generate peptide-pulsed target cells, 1–2 ϫ 106 EL-4 cells were incubated with 20 ␮g/ml of individual influenza peptides in 500 ␮l of complete RPMI medium for 1 h at 37 °C. Immunoprecipitation—RMA-S cells and EL-4 cells were either pulsed with PR8 virus-derived synthetic peptides or infected with a high dose of live PR8 viral particles as described above. Nanospray Tandem Mass Spectrometry on a Quadrapole Time of Flight Spectrometer—The nanospray MS/MS used for identification of PR8 virus-derived peptides from the virally infected EL-4 cells in the present study is described in detail in the Supplemental Material. The MS spectrum of a typical peptide extract is characterized by a set of TABLE I Db- and Kb-restricted CTL epitopes predicted from influenza PR8 virus

No of potentially cleaved bindersb

RESULTS

Not considered

Proteasomal cleavagec

DISCUSSION