Endoplasmic reticulum aminopeptidase associated with antigen processing defines the composition and structure of MHC class I peptide repertoire in normal and virus-infected cells

Abstract

The MHC class I (MHC-I) molecules ferry a cargo of peptides to the cell surface as potential ligands for CD8(+) cytotoxic T cells. For nearly 20 years, the cargo has been described as a collection of short 8-9 mer peptides, whose length and sequences were believed to be primarily determined by the peptide-binding groove of MHC-I molecules. Yet the mechanisms for producing peptides of such optimal length and composition have remained unclear. In this study, using mass spectrometry, we determined the amino acid sequences of a large number of naturally processed peptides in mice lacking the endoplasmic reticulum aminopeptidase associated with Ag processing (ERAAP). We find that ERAAP-deficiency changed the oeuvre and caused a marked increase in the length of peptides normally presented by MHC-I. Furthermore, we observed similar changes in the length of viral peptides recognized by CD8(+) T cells in mouse CMV-infected ERAAP-deficient mice. In these mice, a distinct CD8(+) T cell population was elicited with specificity for an N-terminally extended epitope. Thus, the characteristic length, as well as the composition of MHC-I peptide cargo, is determined not only by the MHC-I peptide-binding groove but also by ERAAP proteolysis in the endoplasmic reticulum.

FIGURE 1

Unusually long peptides are presented by MHC-I in the absence of ERAAP. A, Nonoverlapping B6 or ERAAP^−/− specific peptides were represented as single vertical bars and plotted according to their amino acid length. Indicated p values between two groups were calculated by unpaired two-tailed t test based on the amino acid length. B, A different representation of the length distribution from all samples shown in A.

FIGURE 2

ERAAP deficiency leads to presentation of longer bulging peptides. A and B, Logo representation of unique H-2D^b–bound (A) B6 or (B) ERAAP^−/− peptide sequences analyzed independently according to their lengths shown on the x-axis. The height of each column is proportional to the degree of amino acid conservation and the height of each letter composing the column is proportional to its frequency at the given position. Numbers between parentheses indicate the number of peptide sequences analyzed. Groups containing less than 8 peptides (arbitrary choice) were not analyzed. Amino acids are colored as follows: acidic (red), basic (blue), hydrophobic (black), neutral (purple), and the others (green). C, Unusually long and unique peptides were tentatively sorted into two categories according to the location of the bulge: between the N terminus and the internal asparagine residue (N-bulge) or between the internal asparagine residue and the C terminus (C-bulge). See Supplemental Table 1 for peptide annotation. Bars show the fraction of peptides in each category. As some peptides could be classified in two categories, the sum of percentages may be higher than 100. Numbers between parentheses represent the number of peptides showing an apparent consensus motif (therefore included in the analysis) over the total number of longer and unique H-2D^b–bound peptides.

FIGURE 3

Peptide specificity of CD8⁺ T cell response to mCMV is altered in ERAAP-deficient mice. A, Seven days after i.p. infection, splenocytes from H-2^d ERAAP^+/− and ERAAP^−/− mice were costained with CD8 and the YL9-H-2L^d tetramer (upper panels) or the AI9-H-2D^d tetramer (lower panels). The percentage of tetramer⁺ among all CD8⁺ cells is indicated on the dot plots. B, Frequencies of CD8⁺ cells stained with NP118-H-2L^d (a control tetramer), YL9-H-2L^d or AI9-H-2D^d tetramers. Each point represents an individual mouse. Data are pooled from two separate experiments. C, Seven days postinfection, IFN-γ production by splenocytes from H-2^d ERAAP^+/− and ERAAP^−/− mice was measured after restimulation with YL9 (upper panels) or AI9 (lower panels). Numbers on the dot plots represent the percentage of IFN-γ⁺ among all CD8⁺ cells. D, Frequencies of CD8⁺ cells producing IFN-γ in response to TL9 (an irrelevant H-2L^d binding peptide), YL9 or AI9. Each point represents an individual mouse. Data are pooled from two separate experiments. *p < 0.05; **p < 0.005.

FIGURE 4

CD8⁺ T cell responses toward N-terminally extended peptides differ between ERAAP^−/− and ERAAP^+/+ mice The amino acid sequences of peptide analogs of the (A) YL9 and (B) AI9 series, with the natural N-terminal extensions encoded by viral proteins IE1 and m164 are shown in bold and underlined. C and D, Synthetic peptides with these sequences were tested for their ability to stimulate IFN-γ production by splenocytes from WT or ERAAP-deficient mice infected 7 d (top) or 50 d (bottom) earlier with mCMV. Bars show the mean ± SEM of three spleens per genotype at day 50 and the value from five pooled spleens per genotype at day 7. Results represent three independent experiments. E, Seven days postinfection, IFN-γ production by splenocytes was measured in response to endogenously synthesized AI9 or DI10 peptides. APCs were ERAAP^−/− fibroblasts cotransfected with H-2D^d and a construct encoding ubiquitin fusion proteins for releasing AI9 or DI10 peptides in the cytoplasm (see Materials and Methods). Data represent two separate experiments with two to three mice per genotype. *p < 0.05.

FIGURE 5

Generation of T cell hybridomas specific for AI9-H-2D^d and DI10-H-2D^d. A and B, LacZ responses of CD8⁺ T cell hybridomas derived from mCMV-infected ERAAP^+/− (A) and ERAAP^−/− (B) mice. The hybridoma cells (AI9Z.~ or DI10Z.~) were cultured overnight with H-2^d J774 macrophages in absence or presence of synthetic AI9 or DI10 peptides. The lacZ activity was measured as the absorbance at 595 nm (A595) of a cleavage product of the CPRG substrate (29). A, A representative panel of 11 AI9-responding hybridomas (of 32 total) is shown. B, Only hybridomas with a higher reactivity to DI10 over AI9 are shown. C and D, The lacZ responses of (C) AI9Z.3 and (D) DI10Z.1 hybridomas to indicated concentrations of synthetic AI9 and DI10 peptides presented exogenously by J774 macrophages used as APCs. E and F, The lacZ responses of (E) AI9Z.3 and (F) DI10Z.1 hybridomas to ERAAP^−/− fibroblasts expressing endogenously synthesized AI9 and DI10 precursors. Data (mean ± SEM) are from three different experiments and representative of the hybridomas shown in A and B.

FIGURE 6

The DI10-H-2D^d—specific CD8⁺ T cells are uniquely elicited in ERAAP^−/− mice. A, Indirect quantification of DI10-H-2D^d–specific CD8⁺ T cells at day 7 postinfection. IFN-γ response by spleen cells from infected mice assessed after stimulation with exogenous AI9 or a mix of AI9 and DI10. The difference between both measurements corresponds to the frequency of genuine DI10-H-2D^d–specific CD8⁺ T cells. Data representative of three experiments. B–E, Direct visualization of DI10-H-2D^d–specific CD8⁺ T cells ex vivo. Seven days postinfection, splenocytes from (B) ERAAP^−/− and (C) ERAAP^+/− mice were costained with CD8 and the AI9-H-2D^d tetramer. CD8⁺ AI9-H-2D^d tetramer⁻ cells were FACS-sorted and further stained with DI10-H-2D^d tetramer (right panels). Numbers on dot plots show the percentage of tetramer⁺ of CD8⁺ cells. Results representative of two separate experiments with three spleens pooled per genotype. Splenocytes from mCMV-infected (D) ERAAP^−/− and (E) ERAAP^+/− mice were stimulated with synthetic AI9 and IFN-γ–producing cells were labeled, without fixation, using the Miltenyi IFN-γ secretion assay (left panels) (see Materials and Methods). IFN-γ–negative cells were sorted using magnetic beads. Following further stimulation with synthetic AI9 or DI10, IFN-γ production was monitored using a regular intracellular staining. Numbers show the percentage of IFN-γ⁺ of CD8⁺ cells. Data represent two experiments with three spleens pooled per genotype.

FIGURE 7

Similar numbers of DI10-specific CD8⁺ T cells induced by immunization but larger amounts of DI10 peptide produced in infected ERAAP^−/− macrophages. A, Absolute numbers of DI10-H-2D^d–specific CD8⁺ cells in lymph nodes 7 d after immunization with DI10-pulsed BMDCs. Numbers of IFN-γ–producing CD8⁺ cells in response to AI9 alone or to a mix of AI9 and DI10 were measured. Bars show the difference between both responses. Mean ± SEM (n = 3 mice). Data representative of two independent experiments with 3–5 mice per group. B, Induction of β-galactosidase in DI10Z.1 hybridoma in response to mCMV-infected BMMs peptide extracts fractionated by reverse phase-HPLC. BMMs were infected for 24 h at multiplicity of infection 0.5. Each HPLC fraction was pulsed onto J774 macrophages that served as APCs. Buffer B (right axis) is acetonitrile containing 0.1% trifluoroacetic acid. C, LacZ response of DI10Z.1 hybridoma after fractionation of 1.5 pmoles of synthetic AI9 or DI10 using the same HPLC conditions as in B.