By tying peptide fragments originally distant in parental proteins, the proteasome can generate spliced peptides that are recognized by CTL. This occurs by transpeptidation involving a peptide-acyl-enzyme intermediate and another peptide fragment present in the catalytic chamber. Four main subtypes of proteasomes exist: the standard proteasome (SP), the immunoproteasome, and intermediate proteasomes β1-β2-β5i (single intermediate proteasome) and β1i-β2-β5i (double intermediate proteasome). In this study, we use a tandem mass tag–quantification approach to study the production of six spliced human antigenic peptides by the four proteasome subtypes. Peptides fibroblast growth factor-5172-176/217-220, tyrosinase368-373/336-340, and gp10040-42/47-52 are better produced by the SP than the other proteasome subtypes. The peptides SP110296-301/286-289, gp100195-202/191or192, and gp10047-52/40-42 are better produced by the immunoproteasome and double intermediate proteasome. The current model of proteasome-catalyzed peptide splicing suggests that the production of a spliced peptide depends on the abundance of the peptide splicing partners. Surprisingly, we found that despite the fact that reciprocal peptides RTK_QLYPEW (gp10040-42/47-52) and QLYPEW_RTK (gp10047-52/40-42) are composed of identical splicing partners, their production varies differently according to the proteasome subtype. These differences were maintained after in vitro digestions involving identical amounts of the splicing fragments. Our results indicate that the amount of splicing partner is not the only factor driving peptide splicing and suggest that peptide splicing efficiency also relies on other factors, such as the affinity of the C-terminal splice reactant for the primed binding site of the catalytic subunit.

Visual Abstract

Cytotoxic T lymphocytes are essential players of the adaptive immune response because they can kill virally infected cells and tumor cells. On the cell surface, CTL recognize peptides of 8–11 aa loaded onto MHC class I molecules (1). These peptides are mostly derived from the degradation of cellular proteins by the proteasome, a large multiprotease complex located in the cytosol. The 20S core of the proteasome is composed by four stacked heptameric rings (2, 3). The two outer rings are made of seven structural α subunits (α1–7), whereas the two inner rings, which delimit the catalytic chamber, contain seven β subunits (β1–7), three of which (β1, β2, and β5) are catalytically active.

Originally, antigenic peptides produced by the proteasome and recognized by CTL were believed to solely derive from linear fragments of proteins. However, some years ago, this idea was challenged by the identification of several antigenic peptides that were produced after splicing of two noncontiguous fragments of the parental protein (49). Peptide splicing was shown to occur in the catalytic chamber of the proteasome through a transpeptidation reaction (4). This reaction requires the production of an acyl-enzyme intermediate, composed of a peptide fragment bound to the proteasome through the N-terminal threonine of the catalytic subunit. This acyl-enzyme intermediate is subjected to a nucleophilic attack by the N-terminal amino group of another peptide fragment present in the catalytic chamber of the enzyme, leading to the production of a spliced peptide (4). This process involves the excision of the intervening sequence and the creation of a new peptide bond between the spliced fragments. So far, six human tumor antigenic peptides, recognized by CTL, were found to be produced by peptide splicing (49). Four of these peptides contain peptide fragments that are spliced in the reverse order to that in which they occur in the parental protein (69). In addition, one of these peptides was shown to contain two aspartate residues resulting from posttranslational deamidation of asparagine residues (7). Although peptide splicing is a low-efficiency process (4, 10, 11), the extreme sensitivity of CTL, which can recognize target cells presenting <10 MHC/peptide complexes (12), likely explains how such a low-efficiency process is immunologically relevant and can elicit clinically meaningful CD8 T cell responses. It is, for example, the case for CD8 T cells isolated against a spliced tyrosinase peptide, which were used for adoptive cell transfer in a melanoma patient and were shown to induce dramatic and lasting tumor regressions (7, 13). The recent developments in peptidomic approaches have led to the detailed characterization of peptides presented by the MHC class I molecules at the cell surface, based on the mass spectrometry (MS) analysis of peptides eluted from MHC molecules (14). Using this approach, several groups have tried to estimate the proportion of spliced peptides effectively expressed at the surface of cells (1517). Because identification of spliced peptides using this approach is based on the development of very complex algorithms, estimation of the proportion of spliced peptides varies from 0.1% to up to 30% between the investigators and therefore remains an intense matter of debate (1521). Overall, this confirms that isolation of CTL remains a crucial step for the validation of spliced peptides.

Based on their ability to cleave fluorogenic peptides and to degrade long precursor peptides, the proteasome catalytic subunits were shown to display cleavage specificities: β1 cleaves mostly after acidic residues (caspase-like activity), β2 after basic residues (trypsin-like activity), and β5 after hydrophobic residues (chymotrypsin-like activity) (22). In cells exposed to IFN-γ and in some lymphoid cells, these three constitutive catalytic subunits can be replaced by their inducible counterparts β1i (LMP2), β2i (MECL1), and β5i (LMP7) to form the immunoproteasome (IP) (1). Because they contain alternative catalytic subunits, IP display lower caspase-like activity and higher trypsin-like and chymotrypsin-like activities than standard proteasomes (SP) (23, 24). Our group identified two intermediate proteasome subtypes that have incorporated only one or two of the three catalytic subunits of the IP: the single intermediate proteasome (SIP), which contains β1, β2, and β5i, and the double intermediate proteasome (DIP), composed of β1i, β2, and β5i. These two intermediate proteasomes are present in normal tissues (human liver, kidney, and gut), where they represent ∼20–30% of the proteasome content. Intermediate proteasomes also represent half of the proteasome content of immature and mature dendritic cells (25). Although the role of intermediate proteasomes is not yet clear, their different catalytic activity leads to the production of a unique peptide repertoire, with some peptides being solely produced by one or the other of these proteasome subtypes. Identifying peptides that are efficiently processed by intermediate proteasomes SIP and DIP is of great interest for the development of immunotherapeutic cancer vaccines, as these two proteasome subtypes are shared between tumor cells and dendritic cells, which are instrumental to vaccination approaches.

Analyzing proteasome digests performed with long peptide precursors, we previously observed that both the SP and IP are able to splice peptides and that some spliced peptides are better produced by the SP, whereas others are better produced by the IP (7, 26). Our results suggested that this was related to differences in cleavage specificity resulting in differential production of the fragments to splice. However, thus far, the ability of intermediate proteasomes to produce spliced peptides has not been investigated. In this study, we compared the four different proteasome subtypes for their ability to produce the five different spliced peptides: fibroblast growth factor (FGF)-5172-176/217-220, sp110296-301/286-289, tyrosinase368-373/336-340, gp100195-202/191or192, and gp10040-42/47-52, against which we have previously isolated CTL clones. We used a tandem mass tag (TMT) labeling approach to quantitatively compare, by MS, the fragments produced during in vitro digestion. By studying the splicing of peptide RTK_QLYPEW and its reverse spliced reciprocal counterpart QLYPEW_RTK by the four proteasome subtypes, we show that the efficiency of peptide splicing does not only depend on the abundance of the peptide splicing partners.

The HEK-293–EBNA clonal cell lines expressing different proteasomes subtypes were previously described (25). These cells were cultured in IMDM medium (Thermo Fisher Scientific) supplemented with 10% FBS (Sigma Life Science), GlutaMAX (2 mM l-alanyl l-glutamine dipeptide; Thermo Fisher Scientific), and 100 U/ml penicillin and 100 µg/ml streptomycin (Thermo Fisher Scientific) (IMDM complete medium). 293-SIP was supplemented with hygromycin (600 µg/ml) (Roche), and 293-DIP and 293-IP were supplemented with both hygromycin (600 µg/ml) and puromycin (5 µg/ml; Invivogen). WEHI 164 clone 13 cells were cultured in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 5% FBS HyClone (GE Healthcare), GlutaMAX (2 mM l-alanyl l-glutamine dipeptide), 100 U/ml penicillin, and 100 µg/ml streptomycin (Thermo Fisher Scientific). EBV-transformed B cells LG2-EBV and CAN-EBV were grown in IMDM medium supplemented with 10% FBS, GlutaMAX (2 mM l-alanyl l-glutamine dipeptide), 100 U/ml penicillin, and 100 μg/ml streptomycin. The CTL clone 14 recognizes the HLA-A32–restricted spliced peptide gp10040-42/47-52 RTK_QLYPEW (4). CTL clone C2 recognizes the spliced peptide FGF-5172-176/217-220 NTYAS_PRFK associated with the HLA-A3 molecule (5). CTL clone DRN-7 recognizes the HLA-A3–restricted spliced peptides SP110 SLPRGT_STPK, SLPRGT_STPKR, and SLPRGT_STPKRR (6). CTL 888 clone 62 recognizes the HLA-A24–restricted spliced peptide tyrosinase368-373/336-340 IYMDGT_ADFSF (7). These four CTL were expanded by stimulation every 14 d with anti-CD3 OKT3 mAb (10 ng/ml; BioXCell) and irradiated allogeneic PBLs and LG2-EBV cells (100 Gy) (27). CTL clone M45-3B recognizes the HLA-A3–restricted spliced peptide gp100195-202/191or192 RSYVPLAH_R (8). This CTL clone was propagated by stimulation every 14 d with 1 µg/ml PHA (Sigma-Aldrich) using irradiated allogeneic PBMCs (100 Gy) and CAN-EBV cells. All CTL were cultured in IMDM (Thermo Fisher Scientific) containing 10% human serum, rIL-2 (150 U/ml; Proleukin; Sanofi), and IDO inhibitor 1-methyl-l-tryptophan (200 µM; Sigma-Aldrich) (28).

Twelve hours before transfection, 293-SP, 293-SIP, 293-DIP, and 293-IP were plated in 96-well plates (30,000 cells/well; Corning) in 100 µl of IMDM complete medium. Cells were transfected using LT transit reagent (MIR2360; Sopachem) with 50 ng of pcDNA3 plasmid encoding the appropriate HLA molecule and the indicated amount of pcDNA3 plasmid encoding the parental antigenic protein. Control cells were transfected with an empty pcDNA3 and loaded or not with the appropriate antigenic peptide (1 µM). Twelve hours after transfection, the relevant CTL was added (20,000 cells/well) in 200 µl of CTL medium. After an overnight coculture, the supernatants were collected to measure their TNF-β content using WEHI cells.

The capture proteasome assay (CAPA) was performed as previously described (29, 30). Briefly, 96-well plates (Greiner) were coated for 2 h at room temperature with the mAb MCP21 (5 μg/ml) recognizing the proteasome subunit α2 (31) and then blocked for 1 h with PBS containing 2% BSA (Roth). Cell pellets were lysed on ice at a cell density of 107 cells/ml of 50 mM Tris-HCl and 0.1% Nonidet P-40 (pH 7.5; lysis buffer). The amount of proteins in the lysate was measured using a BCA protein assay kit (Thermo Fisher Scientific), and protein concentration was adjusted to 4 mg/ml using lysis buffer. A total of 50 μl of cell lysate was then added in each well, and the plate was incubated overnight at 4°C. The plate was washed first in 10 mM Tris-HCl and 0.1% Nonidet P-40 (pH 7.5) and then in 10 mM Tris-HCl (pH 7.5). After proteasome capture, 50 μl of the indicated precursor peptide was added at a concentration of 20 μg/ml in 10 mM Tris-HCl. Plates were then sealed with a plastic cover and incubated at 37°C. For each time point, supernatants were collected (45 μl) in Protein LoBind tubes (Eppendorf), and digestions were stopped by the addition of 2 μl 10% trifluoroacetic acid. All samples were lyophilized and resuspended in 20 μl of water. A total of 5 μl of each digest was pulsed in 50 μl X-VIVO 10 serum-free medium (Lonza) onto the indicated EBV-B cells (40,000 cells/well). After 1-h incubation at 37°C, CTL were added (20,000 cells/well) to CTL medium containing IL-2 (25 U/ml). IFN-γ produced by the CTL was then measured by ELISA (4).

Purification of 20S proteasome was performed as previously described (32, 33) using frozen pellets of 1 × 109 293 EBNA cells. Digestions of precursor peptides (>95% pure) were performed by incubating at 37°C in 20 μl of phosphate buffer (KH2PO4 10 mM [pH 7.5])/time point, 1 μg of SP, SIP, DIP, or IP/time point with 5 μg of RTKAWNRQLYPEWTEAQR (gp100), 7.5 μg of NTYASAIHRTSTHFLPRFKQSEQP (FGF-5), 4 μg of STPKRRHKKKSLPRGTASSR (sp110), 5 μg of ADFSFRNTLEHNALHIYMDGTMSQV (tyrosinase), or 5 μg of RRGSRSYVPLAHSSSAFT (gp100)/time point. Reactions were stopped at the indicated time point using 2 μl 10% trifloroacetic acid. All samples were lyophilized and resuspended in 20 μl of water. A total of 5 μl of each digest was pulsed in 50 μl X-VIVO 10 serum-free medium (Lonza) onto the indicated EBV-B cells (20,000 cells/well). After 1-h incubation at 37°C, CTL were added (15,000 cells/well) to CTL medium containing IL-2 (25 U/ml). After an overnight coculture, IFN-γ produced by the CTL was measured by ELISA (4).

In Fig. 7, 1 μg of SP, SIP, DIP, or IP/time point with 1 μg of RTKQLYPEWTEA, RTKQLYPEWTEAQR, QLYPEWRTKA, and QLYPEWRTKAW was used. In Fig. 8, an equimolar concentration of each of the combined peptides was used with 1 μg of SP, SIP, DIP, or IP/time point. A no-digestion control (no proteasome) was performed for each reaction.

All precursor peptides were synthesized on a solid phase using fluorenylmethoxycarbonyl chemistry. Synthesized peptides were purified by reverse-phase HPLC to >95% purity and finally characterized by MS. The lyophilized peptides were diluted to a final concentration of 20 mg/ml in DMSO and stored at −80°C.

A total of 10 μl of each purified proteasome digestion (one half) was lyophilized and solubilized in 10 μl of water. Five microliters of 200 mM HEPES buffer (pH 8.5) and 2 μl of 6-plex TMT (20 μg/μl in acetonitrile [ACN]) (Thermo Fisher Scientific) were added to 2 μl aliquots of digestions. The solution was gently mixed, centrifuged, and incubated for 90 min without agitation at room temperature. The labeled reaction was quenched by adding 1 μl of 50% hydroxylamine, incubated for 30 min, and stored at −80°C. Digestions with SP, SIP, DIP, IP, and the no-digestion control (no proteasome) were labeled with TMT-126, TMT-127, TMT-128, TMT-129, and TMT-131 (Figs. 25) or TMT-130 (Fig. 1), respectively.

For TMT analysis and epitope detection (Orbitrap)

The digestions were analyzed by liquid chromatography-tandem MS (MS/MS) with an UltiMate 3000 RSLCnano system (Thermo Fisher Scientific) connected in-line to an Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo Fisher Scientific). Peptides were directly loaded onto a reversed-phase trap column (Acclaim PepMap 100; 0.3 × 5 mm, C18, 5 μm, 100 A; Thermo Fisher Scientific) at 10 μl/min with aqueous solution containing 0.1% (v/v) trifluoroacetic acid and 2% ACN. After 3 min, the trap column was set on-line in backflush mode with a reversed-phase analytical column (Acclaim PepMap RSLC; 0.075 × 250 mm, C18, 2 μm, 100 A; Thermo Fisher Scientific). Solvent A was HPLC grade water with 0.1% (v/v) formic acid (FA), and solvent B was HPLC grade 80% ACN with 0.1% (v/v) FA. Separations were performed by applying a linear gradient of 4–45% solvent B over 40 min, followed by a linear gradient of 45–60% solvent B over 5 min, followed by a washing step (10 min at 95% solvent B) and an equilibration step (15 min at 4% solvent B) at a constant flow rate of 300 nl/min.

For TMT analysis, intact peptides were detected in the Orbitrap at a resolution of 120,000 over a scan range of 280–1500 mass-to-charge ratio (m/z) with an automatic gain control target of 4 × 105 ions and a maximum injection time of 50 ms. A data-dependent procedure of MS/MS scans was applied for the top precursor ions with 15-s dynamic exclusion. The total cycle time was set to 3 s. Precursor ions were accumulated with a 0.7-m/z isolation window, and their HCD fragmentation was performed with normalized collision energy of 38. Fragment ions were transferred to the Orbitrap at a resolution of 50,000 with an automatic gain control target of 5 × 104 ions and a maximum injection time of 50 ms.

For the targeted quantification of epitopes, the mass acquisition method combined a full scan method (same parameters as above) with a parallel reaction monitoring method. This parallel reaction monitoring method employed an isolation of the target precursor ions on the basis of the theoretical masses of the epitope sequences (considering the charge states +2 or +3) with a 0.7-m/z window. Collision-induced dissociation fragmentation was performed with a normalized collision energy of 35, and fragment ions were detected in the Orbitrap at a resolution of 50,000 with an automatic gain control target of 5 × 104 ions and a maximum injection time of 50 ms.

Extended spliced peptides digestion (LTQ)

The digestions were analyzed by liquid chromatography-MS/MS using an UltiMate 3000 RSLCnano HPLC system (Thermo Fisher Scientific) coupled to an LTQ XL ion-trap mass spectrometer (Thermo Finnigan). Peptides were separated on a reversed-phase analytical column (50-cm µPAC C18; Pharmafluidics) with a 34-min linear gradient of 3–52% ACN (0.1% FA), followed by 6 min at 70% ACN (0.1% FA), and 20-min reequilibration at 3% ACN (0.1% FA), at a constant flow rate of 300 nL/min.

MS analysis was performed on-line with the mass spectrometer equipped with a nanoelectrospray source and operated in positive mode with default parameters and active automatic gain control. Mass spectra were acquired in a mode that alternated single MS scans (300–2000 m/z) with MS/MS scans using low-energy collision-induced dissociation with a relative collision energy of 35%.

TMT-labeled peptide identification and quantification

Peptide identifications were made using SearchGUI 3.3.17 (34), which was set to use the X! Tandem, MS-GF+, Open Mass Spectrometry Search Algorithm, and Comet search engines. Precursor tolerance was set at 7.0 ppm and fragment tolerance at 20.0 ppm. Resulting search engine files were then harmonized and preprocessed using PeptideShaker 2.0.31 (35). Project files were then exported to Reporter 0.7.20 for relative TMT quantifications (https://compomics.github.io/projects/reporter.html). Peptides displayed in the figures were identified with 100% (Figs. 1, 2, 5), 90% (Fig. 3), and 80% confidence (Fig. 4) and originate, in most cases, from a single cleavage occurring either at one of the borders or inside one of the peptide splicing partners. Peptide identifications and quantifications were manually confirmed using the Thermo Scientific Xcalibur 4.1.50.

Several years ago, Hanada et al. (5) showed that a human anti-tumor CD8+ T lymphocyte clone (CTL C2) isolated from a patient with renal cell carcinoma recognized a peptide presented by HLA-A3 and composed of two noncontiguous fragments of the FGF-5 protein. Production of this peptide required the removal of a 40-aa sequence located between fragments NTYAS and PRFK, which compose the spliced peptide NTYAS_PRFK recognized by the CTL (5). We subsequently showed that production of this peptide was mediated by the proteasome in a transpeptidation reaction (11). To analyze the processing of this spliced peptide by the four proteasome subtypes, we used a set of engineered HEK-293–EBNA cells, which exclusively express either SP (293-SP), SIP β5i (293-SIP), DIP β1iβ5i (293-DIP), or IP (293-IP). These cell lines were previously obtained by successive transfections of parental HEK-293–EBNA cells, which exclusively express the SP (hereafter called 293-SP), with strong expression vectors encoding the different inducible subunits. The overexpressed subunits completely replaced the endogenous subunits in the 20S proteasome particles, and LC–MS/MS analysis of proteasomes purified from the four cell lines 293-SP, 293-SIP, 293-DIP, and 293-IP confirmed that they contained exclusively SP, SIP, DIP, and IP, respectively (25, 33). This occurs despite continued transcription of the genes encoding the constitutive subunits, because of the preferential and cooperative incorporation of inducible subunits into nascent proteasome particles, as determined from the study of IP assembly (36). Moreover, the lack of incorporation of constitutive subunits in the proteasomes from lines 293-SIP, 293-DIP, and 293-IP appears to trigger their degradation, as a Western blot of the total lysates did not show any trace of unincorporated subunits (33). These HEK-293–EBNA cells were transiently transfected with plasmid constructs encoding the entire FGF-5 protein and HLA-A3. Transfected cells were then tested for their ability to activate CTL C2. 293-SP stimulated CTL C2 much more efficiently than 293-SIP, 293-DIP, or 293-IP, suggesting that the SP produced the spliced peptide better than the three other proteasome subtypes (Fig. 1A). To study in detail the processing of the FGF-5 peptide, we performed in vitro proteasome digestions using two different methods. The CAPA is a technique that involves the capture of proteasomes on a microplate using the α2-specific Ab MCP21 (29, 30). Lysates originating from the four different HEK-293–EBNA cell lines were added to wells coated with MCP21. Plates were then extensively washed to remove unbound proteins, and the synthetic precursor peptide NTYASAIHRTSTHFLPRFKQSEQP was added to the plate and digested by the captured proteasomes. Digestions were stopped at different time points and pulsed on HLA-A3+ cells. In agreement with the results of the transfection assay, digests performed with SP were much better recognized by CTL C2 than those obtained from SIP, DIP, and IP (Fig. 1B). Similar results were obtained when the precursor peptide was digested with purified 20S proteasomes (Fig. 1C). These results were in line with our previous results showing that SP better produced this FGF-5 spliced peptide than IP (26). Degradation of the precursor peptide by purified proteasomes was estimated by MS, and digests displaying similar rate of precursor degradation were analyzed by HPLC-MS. Peptide fragments were quantified using a 6-plex TMT mass tag labeling approach to obtain quantitative information about the relative abundance of each peptide in the digests. Using this approach, peptides contained in each of the four digests were labeled with a given TMT mass tag. TMT tags are isobaric mass tags that contain defined isotopic substitutions in their mass reporter region, which can therefore be distinguished and quantified following MS/MS fragmentation. Similar amounts of the four labeled digests, displaying similar percentages of precursor degradation (Fig. 1D, top), were pooled together and loaded on the mass spectrometer in the same run. Upon MS/MS fragmentation, the mass reporter tags are cleaved and their amount precisely quantified, enabling the precise comparison of the abundance of given peptides in each digest. Unfortunately, using this approach, we were not able to detect the NTYAS fragment, which composes the N-terminal part of the spliced peptide NTYAS_PRFK. However, the complementary fragment AIHRTSTHFLPRFKQSEQP could be found in digests made with SP, DIP, and IP and to a much lesser extent in those made with SIP. Although we could not directly estimate the efficiency at which the acyl-enzyme intermediate involving NTYAS was produced, these results show that SP, DIP, and IP (and to a lesser extent SIP) are able to produce the cleavage between S176 and A177, which is necessary for the production of the corresponding acyl-enzyme intermediate. Interestingly, the other splice reactant PRFK, which composes the C terminus of the antigenic peptide, was much more abundant in the SP digest. This is likely due to the fact that in DIP and IP digests (and to a lesser extent in the SIP digest), the PRFK fragment is destroyed by an internal destructive cleavage, which occurs between F219 and K220 (fragments PRF, NTYASAIHRTSTHFLPRF, and KQSEQP) (Fig. 1D). Furthermore, two other destructive cleavages between A175 and S176 (fragments NTYA and SAIHRTSTHFLPRFKQSEQP) and between Y174 and A175 (fragments NTY and ASAIHRTSTHFLPRFKQSEQP) were more intense in DIP/IP digests and in SIP/DIP/IP digests, respectively, and therefore impacted the production of the spliced peptide partner NTYAS (Fig. 1D). Interestingly, the destructive cleavages Y174/A175 and F219/K220 both occurred after large hydrophobic amino acids (Y and F). The improved ability of SIP, DIP, and IP to produce these cleavages could therefore result from the presence of the β5i subunit, for which the S1 pocket accommodates larger amino acids than that of the β5 subunit found in SP. In conclusion, our results suggest that the increased production of spliced peptide NTYAS_PRFK by the SP is related to the increased ability of the SP to produce the nucleophile peptide PRFK, combined with destructive cleavages inside the NTYAS splicing partner by the other proteasome subtypes.

FIGURE 1.

Impact of the four proteasome subtypes on the processing of spliced peptide FGF-5172-176/217-220 NTYAS_PRFK. (A) Presentation of spliced peptide NTYAS_PRFK by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding the FGF-5. As control, cells were pulsed with 1 μM NTYAS_PRFK peptide. FGF-5–specific CTL clone C2 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide NTYAS_PRFK in digests obtained by incubating precursor peptide NTYASAIHRTSTHFLPRFKQSEQP with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). In the precursor peptide, the size of the intervening sequence was shortened from 40 to 10 aa to facilitate the mass spectrometry analysis of the produced fragments. Digests were loaded on HLA-A3+ CAN-EBV cells, and the FGF-5–specific CTL C2 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 210-min SP digest, 90-min SIP and DIP digests, and 120-min IP digest, representing ∼50% of precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

FIGURE 1.

Impact of the four proteasome subtypes on the processing of spliced peptide FGF-5172-176/217-220 NTYAS_PRFK. (A) Presentation of spliced peptide NTYAS_PRFK by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding the FGF-5. As control, cells were pulsed with 1 μM NTYAS_PRFK peptide. FGF-5–specific CTL clone C2 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide NTYAS_PRFK in digests obtained by incubating precursor peptide NTYASAIHRTSTHFLPRFKQSEQP with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). In the precursor peptide, the size of the intervening sequence was shortened from 40 to 10 aa to facilitate the mass spectrometry analysis of the produced fragments. Digests were loaded on HLA-A3+ CAN-EBV cells, and the FGF-5–specific CTL C2 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 210-min SP digest, 90-min SIP and DIP digests, and 120-min IP digest, representing ∼50% of precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

Close modal

The spliced peptide SLPRGT_STPK is a minor histocompatibility Ag created by a polymorphism in the SP110 gene (6). This peptide is presented by HLA-A3 and recognized by CTL clone DRN-7 isolated from a myeloma patient after MHC-matched hematopoietic cell transplantation. This peptide is produced after excision of a 6-aa sequence and splicing of the fragments SLPRGT and STPK, which are assembled together in the reverse order to that of the parental protein (6). To study the processing of this peptide, we transfected HEK-293–EBNA cells with plasmids encoding SP110 and HLA-A3. Transfected cells expressing IP were much better recognized by CTL DRN-7 than those expressing DIP or SIP. Cells expressing SP did not stimulate CTL DRN-7 at all (Fig. 2A). In vitro digestions were then performed by incubating synthetic precursor peptide STPKRRHKKKSLPRGTASSR with immunocaptured or purified proteasomes obtained from the four HEK-293–EBNA cell lines. Digests were pulsed on HLA-A3+ EBV-B cells and tested for their ability to activate CTL DRN-7. As expected, IP digests stimulated CTL DRN-7 better than the DIP, SIP, and SP digests (Fig. 2B, 2C). These results were in line with the results obtained in the cellular assay and with previous analyses showing that the IP better produced this peptide than the SP (26).

FIGURE 2.

Impact of the four proteasome subtypes on the processing of spliced peptide SP110296-301/286-289 SLPRGT_STPK. (A) Presentation of spliced peptide SLPRGT_STPK by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding SP110. As control, cells were pulsed with 1 μM SLPRGT_STPK peptide. SP110-specific CTL DRN-7 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide SLPRGT_STPK in digests obtained by incubating precursor peptide STPKRRHKKKSLPRGTASSR with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A3+ CAN-EBV cells, and the SP110-specific CTL DRN-7 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 480-min SP, SIP, and IP digests and the 300-min DIP digest, corresponding to ∼50% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

FIGURE 2.

Impact of the four proteasome subtypes on the processing of spliced peptide SP110296-301/286-289 SLPRGT_STPK. (A) Presentation of spliced peptide SLPRGT_STPK by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding SP110. As control, cells were pulsed with 1 μM SLPRGT_STPK peptide. SP110-specific CTL DRN-7 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide SLPRGT_STPK in digests obtained by incubating precursor peptide STPKRRHKKKSLPRGTASSR with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A3+ CAN-EBV cells, and the SP110-specific CTL DRN-7 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 480-min SP, SIP, and IP digests and the 300-min DIP digest, corresponding to ∼50% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

Close modal

Four digests with approximately similar percentages of precursor degradation were labeled with TMT, pooled, and analyzed by HPLC-MS/MS. Fragment SLPRGT, which composes the N-terminal part of the antigenic peptide, was more abundant in the DIP and IP digests (Fig. 2D). This fragment is produced by two different cleavages: an N-terminal cleavage between K295 and S296, which was more intense in the DIP and IP digests (as demonstrated by the quantification of peptides STPKRRHKKK and SLPRGTASSR), and a C-terminal cleavage occurring between T301 and A302, which was also more intense in the DIP and IP digests as reported by fragment STPKRRHKKKSLPRGT. Two destructive cleavages were observed at S296/L297 and L297/P298. These destructive cleavages were stronger in the DIP digest, but not sufficient to decrease production of the N-terminal splice reactant SLPRGT by DIP when compared with SP and SIP. We detected STPK, STPKR, and STPKRR, three fragments that could be used as splicing partners for the production of spliced peptide SLPRGTSTPK, SLPRGTSTPKR, and SLPRGTSTPKRR, respectively. Although the natural peptide eluted from HLA-A3 molecules is SLPRGT_STPK, we previously observed that both peptides SLPRGT_STPKR and SLPRGT_STPKRR were also recognized by CTL DRN-7 (6). Moreover, these extended spliced peptides might be further trimmed by the proteasome to produce the final antigenic peptide (8). Unfortunately, we could not detect, after TMT labeling, any of the spliced peptides SLPRGT_STPK, SLPRGT_STPKR, or SLPRGT_STPKRR. Although STPK is better produced by the SIP, DIP, and IP, STPKR and STPKRR were more abundant in the IP digest, suggesting that the higher propensity of IP to produce the spliced peptide recognized by CTL DRN-7 might also be at least partly linked to its increased ability to produce the corresponding extended spliced peptides (Fig. 2D). Overall, the increased ability of IP to produce the SP110-spliced peptide appears related to its increased propensity to produce both splicing partners when compared with the other proteasome subtypes, particularly the SP.

The HLA-A24–restricted peptide IYMDGT_ADFSF is derived from tyrosinase and recognized by melanoma-infiltrating CTL that induced dramatic and lasting tumor regressions after adoptive cell therapy (13). In addition to the reverse splicing reaction in the proteasome, processing of this peptide involves the deamidation of two asparagine residues (7). After transfection of the four HEK-293–EBNA cells with plasmids containing the cDNA encoding tyrosinase and HLA-A24, only 293-SP cells could stimulate the CTL (Fig. 3A). In agreement with this, only digests obtained with SP and the long peptide precursor ADFSFRNTLEHNALHIYMDGTMSQV could activate the CTL, when pulsed on HLA-A24+ cells (Fig. 3B, 3C). Digests corresponding to ∼40% of precursor degradation were labeled with TMT and analyzed by HPLC-MS/MS (Fig. 3D). Peptide IYMDGT, which constitutes the N-terminal splicing partner, could not be detected, even in digests obtained with SP. However, fragment MSQV, which is released following cleavage of the peptide bond between T373 and M374, was more abundant in SP digests, suggesting that the SP better produced the acyl-enzyme intermediate needed for the peptide splicing reaction. Production of the other splicing partner ADFSF was more effective in digests from SP and IP, as compared with SIP or DIP. Several internal cleavages were observed in both splicing partners. In particular, cleavages between I368 and Y369 and between M370 and D371, which likely decreased the production of the acyl enzyme-intermediate required for the production of the antigenic peptide, were more intense in SIP, DIP, and IP than in SP digests. However, a cleavage between D371 and G372 was more intensively produced by SP and SIP, likely because these proteasomes share the β1 subunit, which is responsible for the caspase-like activity of the proteasome.

FIGURE 3.

Impact of the four proteasome subtypes on the processing of spliced peptide tyrosinase368-373/336-340 IYMDGT_ADFSF. (A) Presentation of spliced peptide IYMDGT_ADFSF by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A24 and the indicated amount of plasmid encoding tyrosinase. As control, cells were pulsed with 1 μM IYMDGT_ADFSF peptide. Tyrosinase-specific CTL 888 clone 62 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide IYMDGT_ADFSF in digests obtained by incubating precursor peptide ADFSFRNTLEHNALHIYMDGTMSQV with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). In the precursor peptide, the size of the intervening fragment was shortened from 27 to 10 aa to facilitate the mass spectrometry analysis of the produced peptide fragments. Digests were loaded on HLA-A24+ LG2-EBV cells, and tyrosinase-specific CTL 888 clone 62 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼40% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 60-min SP and IP digests and 90-min SIP and DIP digests, corresponding to ∼40% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

FIGURE 3.

Impact of the four proteasome subtypes on the processing of spliced peptide tyrosinase368-373/336-340 IYMDGT_ADFSF. (A) Presentation of spliced peptide IYMDGT_ADFSF by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A24 and the indicated amount of plasmid encoding tyrosinase. As control, cells were pulsed with 1 μM IYMDGT_ADFSF peptide. Tyrosinase-specific CTL 888 clone 62 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide IYMDGT_ADFSF in digests obtained by incubating precursor peptide ADFSFRNTLEHNALHIYMDGTMSQV with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). In the precursor peptide, the size of the intervening fragment was shortened from 27 to 10 aa to facilitate the mass spectrometry analysis of the produced peptide fragments. Digests were loaded on HLA-A24+ LG2-EBV cells, and tyrosinase-specific CTL 888 clone 62 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼40% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 60-min SP and IP digests and 90-min SIP and DIP digests, corresponding to ∼40% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

Close modal

Overall, these observations suggest that spliced peptide tyrosinase368-373/336-340 is better produced by the SP in part because the SP better produces the cleavage between T373 and M374, which is required for the production of the acyl-enzyme intermediate involving IYMDGT. Again, destructive cleavages within peptide splicing partners likely also influence the peptide splicing reaction. However, our inability to detect IYMDGT prevents us from ruling out additional factors potentially involved.

Peptide RSYVPLAH_R is derived from the melanosomal protein gp100 and presented by HLA-A3 (8). In contrast to the previously described spliced peptides, which are produced by the association of fragments of 3 to 6 aa, peptide RSYVPLAH_R results from the association of an 8-aa fragment with a single arginine residue. The production of this spliced peptide is a two-step process involving the reverse splicing of a peptide fragment containing at least 3 aa followed by the trimming of the C-terminally extended spliced peptide to produce the final antigenic peptide (8). To compare the processing of this peptide, we cotransfected 293-SP, 293-SIP, 293-DIP, and 293-IP cells with plasmids encoding gp100 and HLA-A3. Transfected 293-IP and 293-DIP cells were recognized by CTL M45-3B, whereas transfected 293-SIP and 293-SP cells were not or very poorly (Fig. 4A). We then digested synthetic precursor peptide RRGSRSYVPLAHSSSAFT with plate-captured or purified proteasomes. When loaded onto HLA-A3+ target cells, the digests performed with IP and DIP were recognized by CTL M45-3B, whereas digests produced by SIP or SP were not or only slightly recognized (Fig. 4B, 4C), confirming the results of the transfection assay. Four digests presenting a similar level of precursor degradation were labeled with TMT as above, pooled together, and analyzed by HPLC-MS. In DIP and IP digests, we detected higher amounts of peptide RSYVPLAH, which composes the N-terminal part of the antigenic peptide recognized by the CTL. Production of RSYVPLAH is dependent on two different cleavages. A first cleavage is between H202 and S203, which can be produced by all four proteasome subtypes as demonstrated by the quantification of the peptide RRGSRSYVPLAH and its complementary fragment SSSAFT (Fig. 4D). The N-terminal cleavage between S194 and R195, which is required for the release of RSYVPLAH, can be detected in both the SP and IP digests (fragment RSYVPLAHSSSAFT). The increased production of the RSYVPLAH peptide by DIP and IP is likely due to the occurrence of a destructive cleavage between A201 and H202 in the SP and SIP digests (as demonstrated by the increased production of peptides RRGSRSYVPLA and HSSSAFT). This destructive cleavage likely relies on the activity of subunit β1, which is found in both SP and SIP but not in DIP and IP. Indeed, although subunit β1 is mainly linked to the caspase-like activity, it was suggested to display additional activities targeting small neutral amino acids or branched amino acids (37), activities that are enhanced by the presence of a proline residue in P3 (38). Note that the IP also makes another destructive cleavage, between Y197 and V198, which, however, does not limit its efficiency to produce the RSYVPLAH splicing partner.

FIGURE 4.

Impact of the four proteasome subtypes on the processing of spliced peptide gp100195-202/191or192 RSYVPLAH_R. (A) Presentation of spliced peptide RSYVPLAH_R by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding gp100. As control, cells were pulsed with 1 μM RSYVPLAH_R peptide. Gp100-specific CTL clone M45-3B was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide RSYVPLAH_R in digests obtained by incubating precursor peptide RRGSRSYVPLAHSSSAFT with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A3+ CAN-EBV cells and the gp100-specific CTL M45-3B was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 240-min SP and SIP digests and 30-min DIP and IP digests, corresponding to ∼50% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

FIGURE 4.

Impact of the four proteasome subtypes on the processing of spliced peptide gp100195-202/191or192 RSYVPLAH_R. (A) Presentation of spliced peptide RSYVPLAH_R by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A3 and the indicated amount of plasmid encoding gp100. As control, cells were pulsed with 1 μM RSYVPLAH_R peptide. Gp100-specific CTL clone M45-3B was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide RSYVPLAH_R in digests obtained by incubating precursor peptide RRGSRSYVPLAHSSSAFT with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A3+ CAN-EBV cells and the gp100-specific CTL M45-3B was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼50% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 240-min SP and SIP digests and 30-min DIP and IP digests, corresponding to ∼50% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide was evaluated by HPLC-MS/MS.

Close modal

We previously showed that production of the spliced peptide RSYVPLAH_R does not occur through the splicing of a single arginine by transpeptidation (8). Rather, splicing of a longer peptide fragment leads to the production of a C-terminally extended spliced peptide that will then be further trimmed by the proteasome to produce the final spliced peptide. The arginine involved in the splicing reaction was previously shown to correspond to R191 or R192. Unfortunately, our experimental conditions did not allow us to detect any fragment containing R191 or R192.

Altogether, our results suggest that the differential processing of spliced peptide RSYVPLAH_R can be explained by a greater capacity of the DIP and IP to produce RSYVPLAH, which composes the acyl enzyme intermediate. Although, we were unable to evaluate the role of the C-terminal spliced reactant, efficiency of the splicing reaction appeared to depend on the production of the RSYVPLAH-enzyme intermediate.

The HLA-A32–restricted spliced peptide derived from the melanoma differentiation Ag gp100 is recognized by an anti-tumor CTL clone isolated from a melanoma patient and was the first spliced peptide shown to be produced by the proteasome in a transpeptidation reaction (4). Production of this spliced peptide RTK_QLYPEW involves the excision of 4 aa and the splicing of the fragments RTK and QLYPEW by the proteasome (4). We assessed the processing of this spliced peptide by transfecting the four 293 cell lines 293-SP, 293-SIP, 293-DIP, and 293-IP with plasmids encoding gp100 and HLA-A32. Transfected cells were then tested for CTL recognition. After transfection, 293-SP were the most efficient at producing the peptide (Fig. 5A), whereas 293-SIP, DIP, and IP were barely recognized. We then digested peptide RTKAWNRQLYPEWTEAQR using immunocaptured proteasomes or purified proteasomes. Digests were then pulsed onto HLA-A32+ cells and tested for their ability to sensitize the CTL. In line with the cell transfection assay, the SP digest efficiently stimulated CTL 14, whereas the IP digest did not and the SIP and DIP were in between (Fig. 5B, 5C).

FIGURE 5.

Impact of the four proteasome subtypes on the processing of spliced peptide gp10040-42/47-52 RTK_QLYPEW. (A) Presentation of spliced peptide RTK_QLYPEW by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A32 and the indicated amount of plasmid encoding gp100. As control, cells were pulsed with 1 μM RTK_QLYPEW peptide. Gp100-specific CTL clone 14 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide RTK_QLYPEW in digests obtained by incubating precursor peptide RTKAWNRQLYPEWTEAQR with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A32+ LG2-EBV cells, and gp100-specific CTL 14 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼40% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 120-min SP, 90-min SIP and DIP digests, and 190-min IP digest, corresponding to ∼40% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide were evaluated by HPLC-MS/MS.

FIGURE 5.

Impact of the four proteasome subtypes on the processing of spliced peptide gp10040-42/47-52 RTK_QLYPEW. (A) Presentation of spliced peptide RTK_QLYPEW by HEK-293–EBNA cells expressing different proteasome subtypes. Cells were transfected with HLA-A32 and the indicated amount of plasmid encoding gp100. As control, cells were pulsed with 1 μM RTK_QLYPEW peptide. Gp100-specific CTL clone 14 was added, and TNF production was measured after 16-h coculture. Error bars show SDs of duplicate samples. One representative out of three experiments is shown. Detection of peptide RTK_QLYPEW in digests obtained by incubating precursor peptide RTKAWNRQLYPEWTEAQR with immunocaptured proteasomes (CAPA) (B) or purified 20S proteasomes (C). Digests were loaded on HLA-A32+ LG2-EBV cells, and gp100-specific CTL 14 was added. IFN-γ was measured after an overnight coculture. Error bars show SDs of duplicates. One of three representative experiments is shown. In (C), recognition by the CTL of digests displaying ∼40% of precursor degradation (quantified in D) is shown. (D) Quantitative analysis of the fragments present in the 120-min SP, 90-min SIP and DIP digests, and 190-min IP digest, corresponding to ∼40% of the precursor degradation. Digests were differentially labeled with TMT, pooled together, and the relative abundance of the precursor peptide and the fragments that are relevant for the production of the antigenic peptide were evaluated by HPLC-MS/MS.

Close modal

Digests produced with purified proteasomes and corresponding to ∼40% of precursor degradation were labeled with TMT and analyzed by HPLC-MS/MS (Fig. 5D, top). Fragment RTK, which composes the N-terminal part of the antigenic peptide, was present in all four proteasome digests, although slightly less in the IP digest and slightly more in the SP and DIP digests (Fig. 5D). The C-terminal splicing partner QLYPEW was found in all proteasome digests but was slightly more abundant in the SP digest. Two internal destructive cleavages were more intense either in the IP digest (Y49/P50) or in the SIP, DIP, and IP digests (L48/Y49). The increased abundance of spliced peptide RTK_QLYPEW in the SP digest was confirmed by searching the digest for that peptide by MS/MS (Fig. 6A). Increased production of spliced peptide RTK_QLYPEW by the SP might be partially explained by an increased abundance of the peptide QLYPEW in the SP digest. However, the amounts of QLYPEW were only modestly increased, so that this factor was unlikely to be the only explanation. Moreover, when we searched the digest for the presence of the reversed spliced peptide QLYPEW_RTK, which was previously described by Ebstein et al. (9), we surprisingly found that this peptide, which involves the same splicing partners as RTK_QLYPEW, was produced in an opposite manner by the four proteasome subtypes, being preferentially processed by the DIP and IP (Fig. 6B). This indicates that the splicing efficiency does not solely depend on the abundance of the splicing partners but also relies on additional factors.

FIGURE 6.

MS/MS detection of the reciprocal spliced peptides RTK_QLYPEW and QLYPEW_RTK in proteasome digests. MS/MS detection of RTK_QLYPEW (A) or QLYPEW_RTK (B) in the digests of (Fig. 5D.

FIGURE 6.

MS/MS detection of the reciprocal spliced peptides RTK_QLYPEW and QLYPEW_RTK in proteasome digests. MS/MS detection of RTK_QLYPEW (A) or QLYPEW_RTK (B) in the digests of (Fig. 5D.

Close modal

We then explored whether differential splicing of these two peptides could potentially be explained by the fact that the splicing reaction might use C-terminal splice reactants for which the size is longer than the minimal C-terminal splicing partner. C-terminally extended spliced peptides would be preferentially produced by one proteasome subtype and would then be further cleaved by that proteasome to produce the final antigenic peptide.

Three peptides starting with QLYPEW and extended at their C terminus were detected in the digests: QLYPEWTE, QLYPEWTEA, and QLYPEWTEAQR (Fig. 5D). Peptide QLYPEWTEAQR was found in all digests, whereas slightly more was found in SIP. Fragments QLYPEWTEA and QLYPEWTE were highly released by the SP, but not by the other proteasome subtypes. In line with the presence of QLYPEWTEA only in the SP digest (Fig. 5D), the corresponding extended spliced peptide RTK_QLYPEWTEA was solely produced by the SP (Fig. 7A). In addition, just as SIP was the most efficient at producing the QLYPEWTEAQR peptide (Fig. 5D), SIP was also the most efficient at producing the extended spliced peptide RTK_QLYPEWTEAQR (Fig. 7A). So, the increased production of the final antigenic peptide RTK_QLYPEW by SP and SIP (Fig. 6A) might result from a better production of the spliced extended peptide fragments RTK_QLYPEWTEAQR or RTK_QLYPEWTEA, which would be further trimmed by these proteasomes to release the final antigenic peptide. Such a phenomenon was previously shown to be involved in the production of the peptide RSYVPLAH_R (8). To verify whether SP and SIP are able to produce the final trimming step, we incubated the extended peptides RTK_QLYPEWTEA and RTK_QLYPEWTEAQR with SP and SIP, respectively, and we could detect the final spliced peptide RTK_QLYPEW in both digests (Fig. 7B). Likewise, when analyzing the spliced reactants potentially involved in the production of the reverse spliced peptide QLYPEW_RTK from the long precursor peptide, we observed that increased production of the peptide by DIP and IP might be explained by an increased production of two extended spliced reactants, RTKA and RTKAW (Fig. 5D). Again, the corresponding extended spliced peptides QLYPEW_RTKA and QLYPEW_RTKAW were also more abundant in the DIP and IP digests (Fig. 7C), and the final trimming could be produced by both DIP and IP (Fig. 7D). However, we also observed that digestion of extended peptides RTK_QLYPEWTEA, RTK_QLYPEWTEAQR, QLYPEW_RTKA, and QLYPEW_RTKAW with their respective proteasome subtypes induced major destructive cleavages (Fig. 7B, 7D). These cleavages likely impair the production of the final spliced peptides from these extended precursors, suggesting that trimming of extended spliced precursors is not the principal mechanism involved and that other parameters likely regulate the splicing reaction itself.

FIGURE 7.

Production of RTK_QLYPEW and QLYPEW_RTK from extended spliced peptide precursors. (A) MS/MS detection of RTK_QLYPEWTEA and RTK_QLYPEWTEAQR in the digests of (Fig. 5D. (B) MS detection of the indicated peptides in digests of precursor peptides RTKQLYPEWTEA and RTKQLYPEWTEAQR by SP and SIP respectively (60-min digests). (C) MS/MS detection of QLYPEWRTKA and QLYPEWRTKAW in the digests of (Fig. 5D. (D) MS detection of the peptides QLYPEW_RTK and QLYPEW in digests of precursor peptides QLYPEWRTKA and QLYPEWRTKAW by DIP and IP (60-min digests).

FIGURE 7.

Production of RTK_QLYPEW and QLYPEW_RTK from extended spliced peptide precursors. (A) MS/MS detection of RTK_QLYPEWTEA and RTK_QLYPEWTEAQR in the digests of (Fig. 5D. (B) MS detection of the indicated peptides in digests of precursor peptides RTKQLYPEWTEA and RTKQLYPEWTEAQR by SP and SIP respectively (60-min digests). (C) MS/MS detection of QLYPEWRTKA and QLYPEWRTKAW in the digests of (Fig. 5D. (D) MS detection of the peptides QLYPEW_RTK and QLYPEW in digests of precursor peptides QLYPEWRTKA and QLYPEWRTKAW by DIP and IP (60-min digests).

Close modal

Another explanation for the differential production of these two spliced peptides could be that, in a given proteasome subtype, the C-terminal splicing partner interacts more readily with the primed binding site (i.e., the site accommodating the amino acids located downstream of the cleavage site) of the catalytic subunit carrying the acyl-enzyme intermediate, thereby facilitating the splicing reaction. To test this hypothesis, we performed in vitro digestions with pairs of peptides incubated with proteasomes. Because the first step of the splicing reaction is based on the production of the acyl-enzyme intermediate, we used the extended peptide fragment RTKAWNR to produce the acyl-enzyme intermediate required for the splicing of the RTK_QLYPEW peptide. Because most of the acyl-enzyme intermediates end up being hydrolyzed, whereas only a minority undergo peptide splicing, we postulated that the amount of acyl-enzyme intermediate RTK bound to the proteasome catalytic subunit was directly proportional to the amount of fragment RTK released by hydrolysis. We therefore measured the amount of RTK released in the course of digestion (Fig. 8A). Digestion times were chosen so as to obtain similar releases of fragment RTK and consequently similar amounts of acyl-enzyme intermediate RTK in the digests. Furthermore, at the start of the digestion, we added identical amounts of the C-terminal spliced reactant QLYPEW, which remained stable in the course of the digestion (Fig. 8A). Therefore, in the conditions of this assay, equal amounts of both peptide splicing partners were present in the digests, so as to rule out potential influence of the amounts of splicing partners available in the reaction. Production of the spliced peptide was then measured by MS and in a T cell assay, after pulsing the digests on HLA-A32+ cells (Fig. 8A). Interestingly, we observed that in conditions in which the amounts of acyl-enzyme intermediate RTK and QLYPEW peptide were roughly the same, SP and SIP were much more efficient at producing the spliced peptide when compared with DIP and IP. We then performed a similar experiment to evaluate the production of the reciprocal spliced peptide QLYPEW_RTK (Fig. 8B). We digested peptide QLYPEWTEAQR, adjusted the digestion times to obtain similar amounts of QLYPEW enzyme intermediates, and added identical amounts of C-terminal splice reactant RTK. We observed that, despite the presence of similar amounts of splice reactants in the four digests (Fig. 8B), the IP was much more efficient at producing spliced peptide QLYPEW_RTK when compared with the other proteasome subtypes (Fig. 8B). These results are in line with the differential processing observed for the two spliced peptides (RTK_QLYPEW being better produced by SP and SIP and QLYPEW_RTK being better processed by the IP) (Fig. 6). Therefore, in conditions in which the amounts of splicing partners are identical, splicing efficiency varies according to the proteasome subtype. This suggests that the splicing reaction might be influenced by the ability of the nucleophile peptide to interact with the primed binding site of the catalytic subunit, where the C-terminal splice reactant likely binds during the peptide splicing reaction.

FIGURE 8.

Splicing efficiency of RTK_QLYPEW and QLYPEW_RTK. (A) Quantification of RTK and QLYPEW in SP, SIP, DIP, and IP digests (90-min SP and DIP, 30-min SIP, and 210-min IP) performed with peptide pair RTKAWNR + QLYPEW. Digestion times were chosen so that release of fragment RTK was similar between the four digests. Digests were differentially labeled with TMT, pooled together, and the relative abundance of RTK (top panel), QLYPEW (second panel from top), and RTK_QLYPEW (second panel from bottom) was evaluated by HPLC‐MS/MS. The digests were also loaded on HLA-A32+ LG2-EBV cells, and CTL 14 was added. IFN-γ was measured by ELISA after overnight coculture (bottom panel). (B) Quantification of QLYPEW and RTK in SP, SIP, DIP, and IP digests (240-min SP, 60-min SIP, 360-min DIP, and 210-min IP) performed with peptide pair QLYPEWTEAQR + RTK. Digestion times were chosen so that release of fragment QLYPEW was similar among the four digests. Digests were differentially labeled with TMT, pooled together, and the relative abundance of QLYPEW (top panel), RTK (middle panel), and QLYPEW_RTK (bottom panel) was evaluated by HPLC-MS/MS.

FIGURE 8.

Splicing efficiency of RTK_QLYPEW and QLYPEW_RTK. (A) Quantification of RTK and QLYPEW in SP, SIP, DIP, and IP digests (90-min SP and DIP, 30-min SIP, and 210-min IP) performed with peptide pair RTKAWNR + QLYPEW. Digestion times were chosen so that release of fragment RTK was similar between the four digests. Digests were differentially labeled with TMT, pooled together, and the relative abundance of RTK (top panel), QLYPEW (second panel from top), and RTK_QLYPEW (second panel from bottom) was evaluated by HPLC‐MS/MS. The digests were also loaded on HLA-A32+ LG2-EBV cells, and CTL 14 was added. IFN-γ was measured by ELISA after overnight coculture (bottom panel). (B) Quantification of QLYPEW and RTK in SP, SIP, DIP, and IP digests (240-min SP, 60-min SIP, 360-min DIP, and 210-min IP) performed with peptide pair QLYPEWTEAQR + RTK. Digestion times were chosen so that release of fragment QLYPEW was similar among the four digests. Digests were differentially labeled with TMT, pooled together, and the relative abundance of QLYPEW (top panel), RTK (middle panel), and QLYPEW_RTK (bottom panel) was evaluated by HPLC-MS/MS.

Close modal

Understanding the rules dictating the establishment of the MHC class I peptide repertoire is key to the development of cancer immunotherapy based on vaccination. This can help in defining the most efficient vaccination modality to use for a given target Ag. In that sense, the ability of intermediate proteasomes to process tumor Ags is valuable, as these proteasomes are present in significant amounts in both tumors and dendritic cells, which are instrumental to the vaccination approach. Therefore, such intermediate proteasome–dependent Ags can be efficiently targeted by vaccines encoding full-length proteins, which might be more efficient than single epitope vaccines as they can trigger responses to multiple tumor Ags.

Some years ago, proteasomes were shown to produce antigenic peptides by splicing of peptide fragments originally distant in the parental protein (49). In this study, we studied the production of six spliced peptides (49) by the four proteasome subtypes. We confirmed previous results obtained by our laboratory (26) showing that the SP produces better peptides FGF-5172-176/217-220 (Fig. 1), tyrosinase368-373/336-340 (Fig. 3), and gp10040-42/47-52 (Figs. 5, 6A), whereas the IP processes the SP110296-301/286-289 better than the SP (Fig. 2). The poor processing of the peptides FGF-5172-176/217-220, gp10040-42/47-52, and tyrosinase368-373/336-340 by IP and intermediate proteasomes suggests that vaccine strategies using these Ags should bypass processing by dendritic cells, which mostly contain these proteasomes. In contrast, the IP and DIP process peptides SP110296-301/286-289 (Fig. 2), gp100195-202/191or192 (Fig. 4), and gp10047-52/40-42 (Fig. 6B). This suggests that, in the frame of cancer immunotherapy, both peptides gp100195-202/191or192 and gp10047-52/40-42 could be targeted by vaccines based on full-length constructs, which require processing by dendritic cells. In particular, this might be important for the design of full-length vaccines based on viral vectors or mRNA, which are known to produce CD8+ T cell responses and might therefore be key to the design of efficient immunotherapeutic vaccines (39, 40).

The most unexpected observation of our study is the fact that two reciprocal spliced peptides, RTK_QLYPEW and QLYPEW_RTK, which result from the splicing of the very same fragments in the reciprocal orientations, are produced by the four proteasome subtypes in a totally opposite manner: peptide RTK_QLYPEW is better produced by the SP, whereas peptide QLYPEW_RTK is more efficiently produced by the IP and DIP. This suggests that splicing efficiency does not solely depend on the amount of the peptide splicing partners found in the digest, but might also be ruled by other constraints. In line with this, other groups previously observed that, in vitro, splicing could also take place between splicing partners released from minor cleavages (10, 41). To investigate this further, two hypotheses were foreseen. First, we envisioned the possibility that the C-terminal splice reactant might be extended by a few amino acids at its C terminus. This would lead to a C-terminally extended spliced peptide, which would then need to be further cleaved by the proteasome to produce the final antigenic peptide, as previously observed for peptide RSYPLAH_R (8). We did detect extended versions of the spliced peptides in the relevant digests and confirmed the ability of the relevant proteasome to trim those extended peptides. However, we also observed major destructive cleavages inside the extended spliced precursor peptides, suggesting that this mechanism does not entirely account for the production of the spliced peptides.

Another mechanism to explain the differential processing of peptides RTK_QLYPEW and QLYPEW_RTK by the various proteasome subtypes could be that the efficiency of the splicing reaction varies according to the splicing partners and the proteasome subtype involved. Whether peptide splicing follows specific rules is a matter of intense debate (10, 41, 42). Based on the fact that peptide splicing occurs by transpeptidation and involves the nucleophilic attack of an acyl-enzyme intermediate by a peptide present in the proteasome chamber, it is expected that splicing efficiency depends on the speed of the reaction, which is linked to the concentration of both splicing partners. Surprisingly, Mishto et al. and others (10, 41) previously observed that in in vitro digests, peptide splicing often relies on minor proteasome cleavages, suggesting that proteasome-catalyzed peptide splicing might obey specific rules that depend on the nature of the peptide involved in the peptide splicing reaction. Only a few studies investigated the biochemical constraints of proteasome-catalyzed peptide splicing. Berkers et al. (42) digested pairs of peptides in which they sequentially modified 1 aa either in the acyl-enzyme intermediate or in the sequence of the C-terminal splice reactant. Splicing of the fragment was quantified by measuring the ability of the spliced peptide to bind to HLA-A2 in a fluorescence polarization assay. Although the results were limited to the HLA-A2 allele, their work suggested that splicing mostly depends on the sequence of the N-terminal reactant and on the concentration of the C-terminal reactant (42). Mishto et al. (10) also suggested that a longer retention time of the N-terminal reactant at the nonprimed site in the catalytic subunit might be compulsory for the peptide splicing reaction. More recently, another group studied the 20S SP splicing efficiency after digestion of 25 polypeptides of different lengths (41). Using an MS-based de novo sequencing approach, they identified and characterized forward and reverse spliced peptides obtained after digestion with SP and determined specific splicing signatures based on the length and sequence of splice reactants.

To date, no specific rules have been defined regarding the ability of the different proteasome subtypes to splice peptides. Surprisingly, our experiments showed that when incubating peptide RTKAWNR and QLYPEW with SP, SIP, DIP, or IP, in conditions in which a similar level of RTK hydrolysis occurred (and therefore similar amounts of RTK acyl-enzyme intermediate were produced), spliced peptide RTK_QLYPEW was produced by SP and SIP but not by DIP or IP, even though similar amounts of the C-terminal splice reactant were available for the reaction (Fig. 8). In a remarkable mirror image, when we similarly incubated peptides QLYPEWTEAQR and RTK in conditions in which the amount of QLYPEW and RTK remained identical among the four proteasome subtypes, the spliced peptide QLYPEW_RTK was mostly produced by the IP (Fig. 8). These results suggest that the ability of a C-terminal reactant to splice is influenced by the type of proteasome involved, likely through differences in the amino acid composition and/or structure of the catalytic pockets found in those proteasomes. By essence, transpeptidation involves the nucleophilic attack of an acyl-enzyme intermediate by a C-terminal peptide reactant found in the proteasome chamber. To perform the nucleophilic attack, the C-terminal peptide reactant has to localize in close vicinity to the acyl-enzyme intermediate. Previously, it was suggested that peptide splicing might be facilitated by the existence of a nucleophile binding site (site δ), distinct from the primed binding site, that would be located in close vicinity to the catalytic N-terminal threonine and would be occupied by the C-terminal splice reactant. However, no clear evidence for the existence of such an alternative binding site was put forward thus far, and it is likely that the C-terminal splice reactant rather binds to the primed binding site of the catalytic subunit harboring the acyl-enzyme and is therefore adequately positioned for the nucleophilic attack. Binding of the C-terminal splice reactant to the primed binding site is likely influenced by the nature of the amino acids lining that site. In agreement with this, we previously showed that the peptide causing the nucleophilic attack on the acyl-enzyme intermediate must comprise at least 3 aa to give rise to a spliced peptide, suggesting that the C-terminal splice reactant must indeed be sufficiently long to properly bind to the primed binding site (8). The increased ability of the SP and SIP to produce the spliced peptide RTK_QLYPEW is likely linked to the presence of the β1 subunit found in the SP and SIP but not in the DIP or IP. In contrast, increased production of peptide QLYPEW_RTK by IP likely depends on the presence of the β2i subunit found in IP but not in SP, SIP, or DIP. If one admits that the β1 (for RTK_QLYPEW) and β2i (for QLYPEW_RTK) subunits are the catalytic subunits where the splicing reaction takes place, it follows that, in conditions in which the amounts of acyl-enzyme intermediates are identical, increased splicing is likely related to an increased availability of the C-terminal splice reactant at the primed binding site of that catalytic subunit as a result of an increased affinity of the nucleophile for that site.

It is important to note, however, that splicing of peptide RTK_QLYPEW might rather require the trypsin-like activity of the β2 or β2i subunit (and not the β1 subunit) for the production of the acyl-enzyme intermediate involving RTK. Likewise, production of the spliced peptide QLYPEW_RTK likely relies on the chymotrypsin-like activity of the β5 or β5i subunit. Thus, the catalytic subunits that favor peptide splicing do not appear to correspond to those expected to produce the acyl-enzyme intermediate. One explanation might be that incorporation of the β1i or β2 subunit induces a modification of the structure of the primed binding site of the β2 or β5i subunit, respectively, thereby impairing the binding of the QLYPEW and RTK peptides. Alternatively, incorporation of the β1i or β2 subunit could partly prevent peptides QLYPEW and RTK from reaching the primed binding site of the catalytic subunit harboring the acyl-enzyme intermediate, thereby decreasing the local concentration of these peptides at the splicing site. This could happen if the C-terminal spliced reactants preferentially interacted with β1i or β2, rather than with the primed binding site of the catalytic subunit where peptide splicing takes place. Both models suggest that affinity of the C-terminal splicing partner for the primed site of the catalytic subunit harboring the acyl-enzyme might be another factor determining the efficiency of peptide splicing.

Crystallographic analysis of yeast proteasome bound to the inhibitor homobelactosin C helped defining the primed binding sites of subunits β5 and β5i and showed striking differences in their sequence and structure. Similar differences were later observed by comparing the crystal structure of murine constitutive and IP (43, 44). The putative primed binding sites of β1 and β2 were modeled based on that of subunit β5. Analysis of their crystal structure in murine proteasomes revealed that the overall structure of the β2 and β2i primed sites appeared identical, whereas the primed site of β1i was shortened when compared with that of β1 (44). In both cases, however, the primed sites exhibited significant sequence differences that might affect their potency to interact with peptides. A better definition of the primed binding site of the catalytic subunits of the four proteasome subtypes might, in the future, help to better understand the origin of these differences in splicing efficiency.

Finally, although the role of the N-terminal splicing partner was not directly investigated in this study, it is another crucial parameter to take into account, because it will define which of the three catalytic subunits will be involved in the formation of the acyl-enzyme intermediate and consequently the nature of the primed site to which the C-terminal splice reactant will bind. The affinity of the N-terminal splicing partner for the nonprimed site of the catalytic pocket could also play a role in the splicing reaction, but in the experiments reported in this article, we suppressed this effect by controlling the amount of N-terminal splicing partner produced.

Altogether, our results suggest that the efficiency of a splicing reaction depends on several parameters, among which are: 1) the amount of peptide fragments available for the splicing reaction including extended fragments that may be subsequently trimmed; 2) internal destructive cleavages that could destroy these fragments or the spliced peptide itself; and 3) the affinity of the C-terminal splice reactant for the primed binding site of the catalytic subunit. It is likely the conjunction of these parameters that modulates the production of spliced peptides.

We thank Rui Cheng, Thérèse Aerts, and Giuseppe Attardo for excellent technical assistance and Isabelle Grisse for editorial assistance.

V.F. is supported by a fellowship from the Fonds National de la Recherche Scientifique (FNRS)–Formation à la Recherche dans l’Industrie et dans l’Agriculture (Grant 1.E091.14). J.A.H. is supported by a fellowship from the FNRS (TELEVIE Grant 7455115F). S.N. is supported by an Excellence of Science (EOS) grant (FNRS/Research Foundation – Flanders) (EOS O000518F). This work received support from the FNRS Belgium (Grants EOS O000518F and PDR T.0091.18), Walloon Excellence in Lifesciences and Biotechnology (WELBIO-CR-2010-11R), and the Fonds Joseph Maisin, Belgium.

Abbreviations used in this article:

ACN

acetonitrile

CAPA

capture proteasome assay

DIP

double intermediate proteasome

FA

formic acid

FGF

fibroblast growth factor

IP

immunoproteasome

MS

mass spectrometry

MS/MS

tandem mass spectrometry

m/z

mass-to-charge ratio

SIP

single intermediate proteasome

SP

standard proteasome

TMT

tandem mass tag

1.
Rock
K. L.
,
A. L.
Goldberg
.
1999
.
Degradation of cell proteins and the generation of MHC class I-presented peptides.
Annu. Rev. Immunol.
17
:
739
779
.
2.
Löwe
J.
,
D.
Stock
,
B.
Jap
,
P.
Zwickl
,
W.
Baumeister
,
R.
Huber
.
1995
.
Crystal structure of the 20S proteasome from the archaeon T. acidophilum at 3.4 A resolution.
Science
268
:
533
539
.
3.
Unno
M.
,
T.
Mizushima
,
Y.
Morimoto
,
Y.
Tomisugi
,
K.
Tanaka
,
N.
Yasuoka
,
T.
Tsukihara
.
2002
.
The structure of the mammalian 20S proteasome at 2.75 A resolution.
Structure
10
:
609
618
.
4.
Vigneron
N.
,
V.
Stroobant
,
J.
Chapiro
,
A.
Ooms
,
G.
Degiovanni
,
S.
Morel
,
P.
van der Bruggen
,
T.
Boon
,
B. J.
Van den Eynde
.
2004
.
An antigenic peptide produced by peptide splicing in the proteasome.
Science
304
:
587
590
.
5.
Hanada
K.
,
J. W.
Yewdell
,
J. C.
Yang
.
2004
.
Immune recognition of a human renal cancer antigen through post-translational protein splicing.
Nature
427
:
252
256
.
6.
Warren
E. H.
,
N. J.
Vigneron
,
M. A.
Gavin
,
P. G.
Coulie
,
V.
Stroobant
,
A.
Dalet
,
S. S.
Tykodi
,
S. M.
Xuereb
,
J. K.
Mito
,
S. R.
Riddell
,
B. J.
Van den Eynde
.
2006
.
An antigen produced by splicing of noncontiguous peptides in the reverse order.
Science
313
:
1444
1447
.
7.
Dalet
A.
,
P. F.
Robbins
,
V.
Stroobant
,
N.
Vigneron
,
Y. F.
Li
,
M.
El-Gamil
,
K.
Hanada
,
J. C.
Yang
,
S. A.
Rosenberg
,
B. J.
Van den Eynde
.
2011
.
An antigenic peptide produced by reverse splicing and double asparagine deamidation.
Proc. Natl. Acad. Sci. USA
108
:
E323
E331
.
8.
Michaux
A.
,
P.
Larrieu
,
V.
Stroobant
,
J. F.
Fonteneau
,
F.
Jotereau
,
B. J.
Van den Eynde
,
A.
Moreau-Aubry
,
N.
Vigneron
.
2014
.
A spliced antigenic peptide comprising a single spliced amino acid is produced in the proteasome by reverse splicing of a longer peptide fragment followed by trimming.
J. Immunol.
192
:
1962
1971
.
9.
Ebstein
F.
,
K.
Textoris-Taube
,
C.
Keller
,
R.
Golnik
,
N.
Vigneron
,
B. J.
Van den Eynde
,
B.
Schuler-Thurner
,
D.
Schadendorf
,
F. K.
Lorenz
,
W.
Uckert
, et al
2016
.
Proteasomes generate spliced epitopes by two different mechanisms and as efficiently as non-spliced epitopes.
Sci. Rep.
6
:
24032
.
10.
Mishto
M.
,
A.
Goede
,
K. T.
Taube
,
C.
Keller
,
K.
Janek
,
P.
Henklein
,
A.
Niewienda
,
A.
Kloss
,
S.
Gohlke
,
B.
Dahlmann
, et al
2012
.
Driving forces of proteasome-catalyzed peptide splicing in yeast and humans.
Mol. Cell. Proteomics
11
:
1008
1023
.
11.
Dalet
A.
,
N.
Vigneron
,
V.
Stroobant
,
K.
Hanada
,
B. J.
Van den Eynde
.
2010
.
Splicing of distant peptide fragments occurs in the proteasome by transpeptidation and produces the spliced antigenic peptide derived from fibroblast growth factor-5.
J. Immunol.
184
:
3016
3024
.
12.
Sykulev
Y.
,
M.
Joo
,
I.
Vturina
,
T. J.
Tsomides
,
H. N.
Eisen
.
1996
.
Evidence that a single peptide-MHC complex on a target cell can elicit a cytolytic T cell response.
Immunity
4
:
565
571
.
13.
Robbins
P. F.
,
M.
el-Gamil
,
Y.
Kawakami
,
E.
Stevens
,
J. R.
Yannelli
,
S. A.
Rosenberg
.
1994
.
Recognition of tyrosinase by tumor-infiltrating lymphocytes from a patient responding to immunotherapy.
Cancer Res.
54
:
3124
3126
.
14.
Chong
C.
,
F.
Marino
,
H.
Pak
,
J.
Racle
,
R. T.
Daniel
,
M.
Müller
,
D.
Gfeller
,
G.
Coukos
,
M.
Bassani-Sternberg
.
2018
.
High-throughput and sensitive immunopeptidomics platform reveals profound interferon γ-mediated remodeling of the human leukocyte antigen (HLA) ligandome.
Mol. Cell. Proteomics
17
:
533
548
.
15.
Liepe
J.
,
F.
Marino
,
J.
Sidney
,
A.
Jeko
,
D. E.
Bunting
,
A.
Sette
,
P. M.
Kloetzel
,
M. P.
Stumpf
,
A. J.
Heck
,
M.
Mishto
.
2016
.
A large fraction of HLA class I ligands are proteasome-generated spliced peptides.
Science
354
:
354
358
.
16.
Faridi
P.
,
C.
Li
,
S. H.
Ramarathinam
,
J. P.
Vivian
,
P. T.
Illing
,
N. A.
Mifsud
,
R.
Ayala
,
J.
Song
,
L. J.
Gearing
,
P. J.
Hertzog
, et al
2018
.
A subset of HLA-I peptides are not genomically templated: evidence for cis- and trans-spliced peptide ligands.
Sci. Immunol.
3
:
eaar3947
.
17.
Mylonas
R.
,
I.
Beer
,
C.
Iseli
,
C.
Chong
,
H. S.
Pak
,
D.
Gfeller
,
G.
Coukos
,
I.
Xenarios
,
M.
Müller
,
M.
Bassani-Sternberg
.
2018
.
Estimating the contribution of proteasomal spliced peptides to the HLA-I ligandome.
Mol. Cell. Proteomics
17
:
2347
2357
.
18.
Liepe
J.
,
J.
Sidney
,
F. K. M.
Lorenz
,
A.
Sette
,
M.
Mishto
.
2019
.
Mapping the MHC class I-spliced immunopeptidome of cancer cells.
Cancer Immunol. Res.
7
:
62
76
.
19.
Paes
W.
,
G.
Leonov
,
T.
Partridge
,
T.
Chikata
,
H.
Murakoshi
,
A.
Frangou
,
S.
Brackenridge
,
A.
Nicastri
,
A. G.
Smith
,
G. H.
Learn
, et al
2019
.
Contribution of proteasome-catalyzed peptide cis-splicing to viral targeting by CD8+ T cells in HIV-1 infection.
Proc. Natl. Acad. Sci. USA
116
:
24748
24759
.
20.
Lichti
C. F.
,
N.
Vigneron
,
K. R.
Clauser
,
B. J.
Van den Eynde
,
M.
Bassani-Sternberg
.
2022
.
Navigating critical challenges associated with immunopeptidomics-based detection of proteasomal spliced peptide candidates.
Cancer Immunol. Res.
10
:
275
284
.
21.
Erhard
F.
,
L.
Dölken
,
B.
Schilling
,
A.
Schlosser
.
2020
.
Identification of the cryptic HLA-I immunopeptidome.
Cancer Immunol. Res.
8
:
1018
1026
.
22.
Dick
T. P.
,
A. K.
Nussbaum
,
M.
Deeg
,
W.
Heinemeyer
,
M.
Groll
,
M.
Schirle
,
W.
Keilholz
,
S.
Stevanović
,
D. H.
Wolf
,
R.
Huber
, et al
1998
.
Contribution of proteasomal beta-subunits to the cleavage of peptide substrates analyzed with yeast mutants.
J. Biol. Chem.
273
:
25637
25646
.
23.
Gaczynska
M.
,
K. L.
Rock
,
A. L.
Goldberg
.
1993
.
Gamma-interferon and expression of MHC genes regulate peptide hydrolysis by proteasomes. [Published erratum appears in 1995 Nature 374: 290.]
Nature
365
:
264
267
.
24.
Gaczynska
M.
,
K. L.
Rock
,
T.
Spies
,
A. L.
Goldberg
.
1994
.
Peptidase activities of proteasomes are differentially regulated by the major histocompatibility complex-encoded genes for LMP2 and LMP7.
Proc. Natl. Acad. Sci. USA
91
:
9213
9217
.
25.
Guillaume
B.
,
J.
Chapiro
,
V.
Stroobant
,
D.
Colau
,
B.
Van Holle
,
G.
Parvizi
,
M. P.
Bousquet-Dubouch
,
I.
Théate
,
N.
Parmentier
,
B. J.
Van den Eynde
.
2010
.
Two abundant proteasome subtypes that uniquely process some antigens presented by HLA class I molecules.
Proc. Natl. Acad. Sci. USA
107
:
18599
18604
.
26.
Dalet
A.
,
V.
Stroobant
,
N.
Vigneron
,
B. J.
Van den Eynde
.
2011
.
Differences in the production of spliced antigenic peptides by the standard proteasome and the immunoproteasome.
Eur. J. Immunol.
41
:
39
46
.
27.
Murata
M.
,
E. H.
Warren
,
S. R.
Riddell
.
2003
.
A human minor histocompatibility antigen resulting from differential expression due to a gene deletion.
J. Exp. Med.
197
:
1279
1289
.
28.
Uyttenhove
C.
,
L.
Pilotte
,
I.
Théate
,
V.
Stroobant
,
D.
Colau
,
N.
Parmentier
,
T.
Boon
,
B. J.
Van den Eynde
.
2003
.
Evidence for a tumoral immune resistance mechanism based on tryptophan degradation by indoleamine 2,3-dioxygenase.
Nat. Med.
9
:
1269
1274
.
29.
Vigneron
N.
,
J.
Abi Habib
,
B. J.
Van den Eynde
.
2015
.
The capture proteasome assay: a method to measure proteasome activity in vitro.
Anal. Biochem.
482
:
7
15
.
30.
Vigneron
N.
,
J.
Abi Habib
,
B. J.
Van den Eynde
.
2015
.
The capture proteasome assay (CAPA) to evaluate subtype-specific proteasome inhibitors.
Data Brief
4
:
146
151
.
31.
Hendil
K. B.
,
W.
Uerkvitz
.
1991
.
The human multicatalytic proteinase: affinity purification using a monoclonal antibody.
J. Biochem. Biophys. Methods
22
:
159
165
.
32.
Schmidtke
G.
,
S.
Emch
,
M.
Groettrup
,
H. G.
Holzhutter
.
2000
.
Evidence for the existence of a non-catalytic modifier site of peptide hydrolysis by the 20 S proteasome.
J. Biol. Chem.
275
:
22056
22063
.
33.
Abi Habib
J.
,
E.
De Plaen
,
V.
Stroobant
,
D.
Zivkovic
,
M. P.
Bousquet
,
B.
Guillaume
,
K.
Wahni
,
J.
Messens
,
A.
Busse
,
N.
Vigneron
,
B. J.
Van den Eynde
.
2020
.
Efficiency of the four proteasome subtypes to degrade ubiquitinated or oxidized proteins.
Sci. Rep.
10
:
15765
.
34.
Barsnes
H.
,
M.
Vaudel
.
2018
.
SearchGUI: a highly adaptable common interface for proteomics search and de novo engines.
J. Proteome Res.
17
:
2552
2555
.
35.
Vaudel
M.
,
J. M.
Burkhart
,
R. P.
Zahedi
,
E.
Oveland
,
F. S.
Berven
,
A.
Sickmann
,
L.
Martens
,
H.
Barsnes
.
2015
.
PeptideShaker enables reanalysis of MS-derived proteomics data sets.
Nat. Biotechnol.
33
:
22
24
.
36.
Griffin
T. A.
,
D.
Nandi
,
M.
Cruz
,
H. J.
Fehling
,
L. V.
Kaer
,
J. J.
Monaco
,
R. A.
Colbert
.
1998
.
Immunoproteasome assembly: cooperative incorporation of interferon gamma (IFN-gamma)-inducible subunits.
J. Exp. Med.
187
:
97
104
.
37.
Valmori
D.
,
U.
Gileadi
,
C.
Servis
,
P. R.
Dunbar
,
J. C.
Cerottini
,
P.
Romero
,
V.
Cerundolo
,
F.
Lévy
.
1999
.
Modulation of proteasomal activity required for the generation of a cytotoxic T lymphocyte-defined peptide derived from the tumor antigen MAGE-3.
J. Exp. Med.
189
:
895
906
.
38.
Cardozo
C.
,
A.
Vinitsky
,
C.
Michaud
,
M.
Orlowski
.
1994
.
Evidence that the nature of amino acid residues in the P3 position directs substrates to distinct catalytic sites of the pituitary multicatalytic proteinase complex (proteasome).
Biochemistry
33
:
6483
6489
.
39.
McAuliffe
J.
,
H. F.
Chan
,
L.
Noblecourt
,
R. A.
Ramirez-Valdez
,
V.
Pereira-Almeida
,
Y.
Zhou
,
E.
Pollock
,
F.
Cappuccini
,
I.
Redchenko
,
A. V.
Hill
, et al
2021
.
Heterologous prime-boost vaccination targeting MAGE-type antigens promotes tumor T-cell infiltration and improves checkpoint blockade therapy.
J. Immunother. Cancer
9
:
e003218
.
40.
Sahin
U.
,
P.
Oehm
,
E.
Derhovanessian
,
R. A.
Jabulowsky
,
M.
Vormehr
,
M.
Gold
,
D.
Maurus
,
D.
Schwarck-Kokarakis
,
A. N.
Kuhn
,
T.
Omokoko
, et al
2020
.
An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma.
Nature
585
:
107
112
.
41.
Paes
W.
,
G.
Leonov
,
T.
Partridge
,
A.
Nicastri
,
N.
Ternette
,
P.
Borrow
.
2020
.
Elucidation of the signatures of proteasome-catalyzed peptide splicing. [Published erratum appears in 2021 Front. Immunol. 12: 755022.]
Front. Immunol.
11
:
563800
.
42.
Berkers
C. R.
,
A.
de Jong
,
K. G.
Schuurman
,
C.
Linnemann
,
H. D.
Meiring
,
L.
Janssen
,
J. J.
Neefjes
,
T. N.
Schumacher
,
B.
Rodenko
,
H.
Ovaa
.
2015
.
Definition of proteasomal peptide splicing rules for high-efficiency spliced peptide presentation by MHC class I molecules.
J. Immunol.
195
:
4085
4095
.
43.
Groll
M.
,
O. V.
Larionov
,
R.
Huber
,
A.
de Meijere
.
2006
.
Inhibitor-binding mode of homobelactosin C to proteasomes: new insights into class I MHC ligand generation.
Proc. Natl. Acad. Sci. USA
103
:
4576
4579
.
44.
Huber
E. M.
,
M.
Basler
,
R.
Schwab
,
W.
Heinemeyer
,
C. J.
Kirk
,
M.
Groettrup
,
M.
Groll
.
2012
.
Immuno- and constitutive proteasome crystal structures reveal differences in substrate and inhibitor specificity.
Cell
148
:
727
738
.

The authors have no financial conflicts of interest.