CD1c bypasses lysosomes to present a lipopeptide antigen with 12 amino acids

Abstract

The recent discovery of dideoxymycobactin (DDM) as a ligand for CD1a demonstrates how a nonribosomal lipopeptide antigen is presented to T cells. DDM contains an unusual acylation motif and a peptide sequence present only in mycobacteria, but its discovery raises the possibility that ribosomally produced viral or mammalian proteins that commonly undergo lipidation might also function as antigens. To test this, we measured T cell responses to synthetic acylpeptides that mimic lipoproteins produced by cells and viruses. CD1c presented an N-acyl glycine dodecamer peptide (lipo-12) to human T cells, and the response was specific for the acyl linkage as well as the peptide length and sequence. Thus, CD1c represents the second member of the CD1 family to present lipopeptides. lipo-12 was efficiently recognized when presented by intact cells, and unlike DDM, it was inactivated by proteases and augmented by protease inhibitors. Although lysosomes often promote antigen presentation by CD1, rerouting CD1c to lysosomes by mutating CD1 tail sequences caused reduction in lipo-12 presentation. Thus, although certain antigens require antigen processing in lysosomes, others are destroyed there, providing a hypothesis for the evolutionary conservation of large CD1 families containing isoforms that survey early endosomal pathways.

Figure 1.

CD1c presents lipopeptide antigens to T cells. (A) [³H]Thymidine incorporation by the IL-2–sensitive HT-2 cell line was used to measure the response of polyclonal 1A3 T cells stimulated with DC and antigen. mAbs against CD1a, CD1b, CD1c, CD1d, or isotype-matched control were added at 20 µg/ml before adding antigen mixture at 0.5 µg/ml. This experiment was performed twice. (B) CD1-deficient C1R B lymphoblastoid cells transfected with cDNAs encoding human CD1a, CD1b, CD1c, CD1d, or no insert (mock) were treated with the antigen mixture before adding polyclonal 1A3 T cells and measuring IL-2 release. This experiment was performed twice. (C) Three major compounds were purified from a mixture of synthetic lipopeptide antigens using an effluent flow splitter on an HPLC interfaced to an ESI-MS. Total ion current is the basis for the relative abundance that is plotted on the y axis, and positive mode spectra corresponding to the time interval of each peak are shown. (D) The amount of each of the three major compounds was normalized for absorbance at 280 nm and used to stimulate 5 × 10⁴ 1A3 T cells with 3 × 10⁴ DC, as measured by [³H]thymidine incorporation. Error bars represent SEM.

Figure 2.

Mass spectrometric analysis identifies antigens as acylated 12-mer peptides. (A) The three major components eluting in LC-MS experiments were named according to their HPLC elution times as lipopeptide 1, 2, or 3. The structures were deduced on the basis of low- and high-resolution MS and MSⁿ experiments. X corresponds to a mass shift consistent with O-linked kynurenine, likely resulting from oxidation of tryptophan during the synthesis process. (B) QIT MS² and MS³ data support the structure of lipopeptide 3. Assignments were also based on the accurate mass values measured by FTICR MS analyses that are presented in Table I. The b_9–11 ion series at m/z 1,316.8, 1,502.9, and 1,589.9 corresponds to N-terminal fragments that provide the sequence WSK. The y₉ ion at m/z 1,227.6 and its water loss peak at m/z 1,209.5 are consistent with the sequence C18:0-(GGK)[1,227.6] that is more fully defined by the MS³ spectrum shown in the inset. Internal ions that bracket the O-Kyn residue (a_m/z_n and b_my_n) are also present, as indicated by assignments on the MS² spectrum. The presence of an ester linkage in the peptide backbone is confirmed by the MS³ spectrum obtained after isolation and decomposition of the c₆ fragment ion at m/z 928.6. In the MS³ spectrum, the b-type fragments at m/z 509, 695, 782, and 910 indicate the sequence WSK and the presence of an ester-derived C-terminal -OH rather than the -NH₂ that would have resulted from N-C_α cleavage of an amide linkage. In addition, y-type fragments observed at m/z 420, 548, and 605 define the N terminus as C18:0-GGK. Spectra obtained during comparable experiments leading to the deduced structures of lipopeptides 1 and 2 are shown in Fig. S2, and the corresponding FTICR MS accurate mass assignments are listed in Table I. (C) Secretion of IL-2 by 1A3 T cells was measured in response to lipopeptide 3 (C18-GGKWSKXSKWSK) or synthetic analogues generated with C18 fatty acids carried on the N-terminal glycine or isoleucine residues presented by DCs. (D) ELISPOT detection of IFN-γ capture in response to DCs treated with lipopeptide 3 or an analogue containing tryptophan in place of kynurenine (C18-GGKWSKWSKWSK). This experiment was performed three times with essentially the same results. *, >1,500 spots per well. (E) Recombinant CD1c-Ig fusion proteins were bound to a protein G–coated plate and treated with the indicated lipopeptide in molar excess before MPM antigen was added. After washing, the MPM-specific, CD1c-restricted T cell line CD8-1 was added and IFN-γ release was measured by ELISA. One out of three independently performed experiments is shown. Error bars represent SEM.

Figure 3.

Proteases influence recognition of lipopeptides. (A and B) lipo-12 and DDM were treated with proteinase K or pronase, or were mock treated in the same buffer without the proteases, and were used to stimulate 1A3 T cells and CD1a-restricted J.RT3/CD8-2 cells, respectively. Error bars represent SEM. (C) To test whether protease-treated lipo-12 itself or the inactivated proteases present in the digestion mixture were toxic to 1A3 T cells, high concentrations (10 µg/ml) of these mixtures were tested in the presence of a suboptimal amount of untreated lipo-12. (D) lipo-12 was added to DCs in the continuous absence or presence of 10 ng/ml of the protease inhibitor LHVS, and proliferation of 1A3 T cells was measured. This experiment was performed twice with essentially the same result.

Figure 4.

Redirected trafficking of CD1c to lysosomes reduces lipopeptide antigen presentation. (A) To enable side-by-side comparison of expression patterns of wild-type and chimeric proteins, C1R lympoblastoid cells were doubly transfected with cDNAs as indicated, permeabilized, and stained with an antibody against the extracellular part of CD1b (BCD1b.3; blue) or CD1c (F10/21A3.1; red) Pseudocolored confocal laser scanning microscopy images are shown. Bar, 10 µm. (B) Colocalization of CD1c wild-type or CD1c_extra/CD1b_tail with LAMP1 was studied using directly labeled antibodies against the extracellular part of CD1c (F10/21A3.1; red) and LAMP1 (H4A3; transformed into green) antibodies. Colocalization of CD1c and LAMP1 is shown (yellow). Bar, 10 µm. Insets show a higher magnification of a single cell. (C) Stably transfected C1R cells were cloned, and individual clones were matched for equivalent surface expression of CD1c (white) as compared with an isotype control (gray). The expression of CD1c_extra/CD1b_tail is shown with a thin line (MFI 585), and wild-type CD1c is shown with a thick line (569 MFI). The C1R transfectants were used to present lipo-12 to 1A3 T cells or MPM to CD8-1 T cells. The experiment shown is representative of five independently performed experiments. (D) C1R cells transfected with CD1b or with mutant CD1b with the cytoplasmic tail of CD1c were used to present C₈₀ GMM, C₅₄ GMM, or C₃₂ GMM presented to LDN5 T cells. Supernatants were tested for the presence of IL-2 using the HT-2 bioassay. Error bars represent SEM.