The Journal of Immunology
Distinct Pathways Generate Peptides from Defective Ribosomal Products for CD8+ T Cell Immunosurveillance Brian P. Dolan,* Lily Li,* Charles A. Veltri,† Chris M. Ireland,† Jack R. Bennink,* and Jonathan W. Yewdell* To understand better the endogenous sources of MHC class I peptide ligands, we generated an antigenic reporter protein whose degradation is rapidly and reversibly controlled with Shield-1, a cell-permeant drug. Using this system, we demonstrate that defective ribosomal products (DRiPs) represent a major and highly efficient source of peptides and are completely resistant to our attempts to stabilize the protein. Although peptides also derive from nascent Shield-1–sensitive proteins and “retirees” created by Shield-1 withdrawal, these are much less efficient sources on a molar basis. We use this system to identify two drugs—each known to inhibit polyubiquitin chain disassembly—that selectively inhibit presentation of Shield-1–resistant DRiPs. These findings provide the initial evidence for distinct biochemical pathways for presentation of DRiPs versus retirees and implicate polyubiquitin chain disassembly or the actions of deubiquitylating enzymes as playing an important role in DRiP presentation. The Journal of Immunology, 2011, 186: 2065–2072. D8+ T cells constitute an important arm of the vertebrate adaptive immune system. Their major defined function is to detect and destroy cells expressing aberrant or pathogen-derived gene products. Detection is based on the display of oligopeptides by cell surface MHC class I (MHC I) molecules (1–3). Peptides are derived by proteolysis of proteins synthesized by the cell principally through the action of proteasomes and aminopeptidases, which generate and trim peptides to the 8- to 10residue length conferring high-affinity interaction with the MHC I peptide binding groove (4). There is great interest in harnessing the power of CD8+ T cells to treat malignancies and reduce the severity of infections with viruses and other intracellular pathogens. Improving such therapies and designing direct-primingbased vaccines that optimally induce CD8+ T cell responses entails a thorough understanding of how peptides are generated from biosynthesized Ags. Peptide Ags can be generated from two general sources of biosynthesized proteins. By combination of design and wear and tear, each gene product demonstrates a characteristic life span, which frequently follows a linear exponential decay curve, whose slope, k, can be expressed as a half-life, t = 0.693/k. In addition, because of unavoidable errors in protein synthesis, folding, assembly, or trafficking, a fraction of nascent proteins never reach
C
*Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892; and †Department of Medicinal Chemistry, University of Utah, Salt Lake City, UT 84112 Received for publication September 16, 2010. Accepted for publication December 1, 2010. This work was supported by the National Institute of Allergy and Infectious Diseases Division of Intramural Research and by National Institutes of Health Grant CA 36622 (to C.M.I.). Address correspondence and reprint requests to Dr. Jonathan W. Yewdell, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Room 2E13C.1, Building 33, Bethesda, MD 20892. E-mail address:
[email protected] Abbreviations used in this article: BFA, brefeldin A; BMDC, bone marrow-derived dendritic cell; CHX, cycloheximide; Cl-PGA2, 10-chloro-15-acetyl-1-methylesterPGA2; DRiP, defective ribosomal product; DUB, deubiquitinating; Eer1, eeyarestatin I; eGFP, enhanced GFP; ER, endoplasmic reticulum; MHC I, MHC class I; NP-40, Nonidet P-40; RDP, rapidly degraded polypeptide; rVV, recombinant vaccinia virus; SCRAP, Shield-1–controlled recombinant antigenic protein; TBS-T, TBS with 0.1% Tween 20. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1003096
a stable conformation or achieve the function for which they were intended and are typically destroyed rapidly. Such polypeptides are termed defective ribosomal products (DRiPs) (5), whereas standard protein substrates exhibiting classical decay are termed “retirees.” It must be stressed that any protein failing to achieve a metabolically stable native state is considered “defective,” regardless of cause. Although it is clear that DRiPs represent a major source of peptides in a number of experimental systems (6–10), the relative contributions of DRiPs versus retirees in the overall scheme of endogenous Ag presentation is uncertain. Quantitative study of Ag processing is greatly facilitated by the use of TCR-like Ab reagents to precisely quantitate peptide–MHC I complexes at the single-cell level via flow cytometry. Kinetic comparison of source protein versus peptide–MHC I complex expression provides unique and nearly unequivocal functional evidence for the derivation of peptides from retirees versus DRiPs (6, 10, 11). Here we use the 25-D1.16 mAb, specific for the Kb-SIINFEKL complex (12), in conjunction with a novel system based on the ProteoTuner methodology developed by Wandless and colleagues (13) for controlling protein stability. This system consists of a mutant form of the FKBP12 that confers a rapidly degraded phenotype. However, the protein stability can be regained upon the addition of a cell-permeable drug (Shield-1) with minimal offtarget effects (Fig. 1A). By fusing the destabilization domain to the well-studied SIINFEKL peptide and enhanced GFP (eGFP) as a reporter, this system uniquely enables temporal examination of peptide generation from a single gene product where the sole variable is the stability of the folded protein. This eliminates the many variables associated with comparing the generation of peptides from multiples forms of an Ag with different stabilities expressed by separate genetic constructs. Using the unique features of the ProteoTuner system, we demonstrate that DRiPs are processed for Ag presentation in a manner different from retired copies of a given protein. Additionally, we identify two small, cell-permeable molecules that selectively inhibit peptide generation from DRiPs, providing the initial nonkinetic evidence for the compartmentalization of presentation from new versus old proteins.
2066
Materials and Methods Abs and reagents Alexa 647-labeled 25D1.16 mAb against Kb-SIINFEKL was generated as previously described (9). Anti-GFP monoclonal Abs were from Roche, and rabbit polyclonal Abs against b-actin and cyclophilin B were from Abcam. Mouse mAb FK2 against polyubiquitin was from Biomol. The purified monoclonal Abs B22.249 (H-2Db) and HB.176 (H-2Kb) were detected with Alexa 647-coupled anti-mouse Abs from Molecular Probes (Eugene, OR). Alexa 680-coupled goat anti-rabbit Ab was from Molecular Probes, and goat anti-mouse IR 800 Ab was from Rockland. MG132 was from Calbiochem. Brefeldin A, cycloheximide, and BSA were from Sigma. Shield-1 was obtained either as a gift from Thomas J. Wandless (Stanford University) or from Clontech. Eerystatin-1 was synthesized by the National Institute of Diabetes and Digestive and Kidney Diseases chemical biology core. Cl-PGA2 was synthesized as previously described (14).
Constructs To generate the Shield-1–controlled recombinant antigenic protein (SCRAP), we PCR amplified the destabilization cassette from pBMN FKBP12 L106P-YFPHA IRES HcRed-tandem vector (13) using the upstream primer 59-GGAATTCCCACCATGGGAGTGCAGGTGGAAACC-39 and downstream primer 59-CTTTTCGAAGTTGATGATCGATTCCGGTTTTA-39. The SIINFEKL-
DISTINCT MHC I Ag PROCESSING PATHWAYS GFP cassette was amplified from pSC11-NP-S-GFP with the upstream primer 59-CTAAAACCGGAATCGATCATCAACTTCGAAAAG-39 and the downstream primer 59-GTCGACCTACTTGTACAGCTCGTCCATG-39. The PCR products were mixed and amplified with upstream primer from the destabilization cassette and the downstream primer from the SIINFEKL-GFP cassette. The resulting PCR product was cloned using a TOPO TA cloning into pCR4 TOPO and sequenced. The SCRAP construct was cloned into pCDNA3.1 by digestion of the PCR cloning vector with EcoRI and into pSC11 using KpnI and NotI. When necessary, sequencing was used to confirm the proper orientation of the SCRAP construct in the vector.
Cells, transfections, virus production, and infections EL4 cells (American Type Culture Collection) were grown in RPMI 1640 containing 7.5% FCS at 37˚C in 6% CO2. Stable transfectants expressing SCRAP were generated by digesting 5 mg pCDNA3.1/SCRAP with PvuI followed by ethanol purification of the DNA. DNA was resuspended in 5 ml water and mixed with 2 3 106 EL4 cells in 100 ml Amaxa solution L. The cells were electroporated using the Amaxa nucleofector program C-009. Cells were cultured overnight and then exposed to G418 at 400 mg/ml. G418 resistant cells were cultured with Shield-1 overnight and GFP+ cells sorted from the overall population using a BD FACSAria. Sorted cells were subsequently cultured in the absence of G418. Bone marrow-derived dendritic cells (BMDCs) were derived as follows: bone marrow was harvested from the femurs of 6- to 24-wk-old female C57Bl6
FIGURE 1. Shield-1 stabilizes SCRAP expressed by EL4 cells. A, Schematic illustration of the SCRAP protein containing the mutant FKBP12 destabilization domain (DD), SIINFEKL peptide, and GFP reporter. Shield-1 allows SCRAP to fold properly. In the absence of Shield-1, SCRAP is rapidly degraded into peptide fragments. B, Left panel, EL4/SCRAP cells were treated overnight with EtOH (solid black line) or 2.5 mM Shield-1 (gray histogram) and analyzed the following day for eGFP expression by flow cytometry and compared with non-transfected EL4 cells (red line). Right panel, EL4/SCRAP cells treated overnight with Shield-1 (gray histogram) were washed in media and cultured without Shield-1 (dashed line) for 1.5 h prior to flow cytometry analysis. EL4/SCRAP cells treated with ethanol alone (black line) shows background staining. C, Dose-response curves of EL4/SCRAP cells treated with different concentrations of Shield-1. At the indicated times, cells were analyzed for eGFP expression by flow cytometry. D, EL4 or EL4/SCRAP cells were treated with 5 mM Shield-1 for the indicated times (left), and total cell lysates were analyzed for expression of SCRAP by SDS-PAGE and immunoblotting for eGFP. EL4/SCRAP cells were also cultured overnight with 2.5 mM Shield-1 (right), washed to remove Shield-1, and total cell lysates prepared at the indicated times. E, EL4/SCRAP cells were labeled with [35S]methionine for 10 min and then chased by the addition of cold medium in the presence EtOH or 2.5 mM Shield-1. At indicated times, total cell lysates were prepared, and SCRAP was immunoprecipitated with anti-eGFP Abs and subsequently analyzed by SDS-PAGE followed by autoradiography. F, Decay curves for SCRAP protein were determined after Shield-1 removal by flow cytometry (black line) or immunoblotting (blue line) by determining how much SCRAP remained as a percentage of the starting levels prior to Shield-1 removal. A decay curve for 35 S-labeled SCRAP (red line) in the absence of Shield-1 is also plotted. The relative calculated half-lives from three independent experiments are listed.
The Journal of Immunology mice, and 8 3 105 bone marrow cells were plated in 5 ml RPMI 1640 with 7.5% FCS and 20 ng/ml GM-CSF in each well of a 6-well tissue culture plate. Medium was exchanged every 2 d, and cells were used for infections on days 8 and 9. The cell line DC2.4 was cultured in DMEM with 7.5% FCS. Recombinant vaccinia virus (rVV) expressing SCRAP was generated following standard procedures. Cells were infected with rVV in 0.1% BSA in balanced salt solution at 107cells/ml with a multiplicity of infection of 10 for 30 min at 37˚C. Cells were then washed and plated in nonadherent plastic dishes and harvested at indicated times and kept at 4˚C until flow cytometry analysis. In some instances, rVV was irradiated by exposing virus to UV light for 10 s prior to infection.
Ag presentation assays To study the kinetics of presentation of cellular proteins, EL4/SCRAP cells were washed in ice-cold citric acid buffer (pH 3) at 1 3 107–2 3 107 cells/ ml for 90 s. Cells were washed and resuspended in culture media at 106 cells/ml. At indicated times, an aliquot of cells (generally 105) was removed and stained with Alexa 647-labeled 25D1.16 mAb and analyzed by flow cytometry analysis. In some instances, inhibitors of Ag presentation were added to cells after the acid wash. For Shield-1 treatment, cells were prepared as described and incubated with indicated amounts of Shield-1.
Western blot and pulse-chase experiments For Western blot analysis, cells were treated as described above, and, at indicated times, 106 cells were harvested, washed in excess PBS, and lysed by boiling in 100 ml SDS-PAGE buffer (Quality Biologicals) containing Roche Protease Inhibitor Cocktail for 20 min with occasional vortexing. After complete lysis of cells, 100 ml water containing DTT was added to each sample, and samples were run on a 4–12% NuPAGE gel (Invitrogen) and blotted onto nitrocellulose membranes. Membranes were blocked with a 4% milk solution in TBS with 0.1% Tween 20 (TBS-T). After a 1-h block, membranes were incubated with primary Abs in 0.5% milk in TBST for 1 h. Membranes were washed with TBS-T for 5 min and incubated with infrared detectable secondary Abs in 0.5% milk in TBS-T for 1 h. After two washes in TBS-T and water, membranes were analyzed using an Odyssey infrared imager (LI-COR). For pulse-chase experiments, 107 EL4/SCRAP cells were harvested, washed once in warm methioninedeficient (Met2) medium, and resuspended to 900 ml in warm Met2 medium. One hundred microliters of 35S-labeled methionine (Amersham) was then added to cells, and cells were incubated at 37˚C for 10 min. Cells
2067 were spun down and resuspended in RPMI 1640 with 7.5% serum and cultured for 0, 10, 30, 60, or 130 min. Cells (2 3 106) were harvested, washed once in cold PBS, and lysed in 200 ml of a PBS solution containing 1% SDS by boiling and vortexing for 20 min. The cell lysate was cooled to room temperature, and 600 ml 1.5% Nonidet P-40 (NP-40)/TBS buffer was added to the lysate. The cell lysate was then precleared overnight at 4˚C with an irrelevant Ab and g bind plus beads (Amersham). Precleared lysates were incubated with ∼2 mg anti-GFP Abs for 1 h at 4˚C followed by 50 ml of a 50% slurry of g bind plus Sepharose beads in 0.1% NP-40/ TBS for an additional hour at 4˚C. Beads were washed three times with 0.1% NP-40/TBS and resuspended in 40 ml SDS-PAGE sample buffer with DTT. Beads were boiled for 10 min, and samples were run on a 4–12% NuPAGE gel. The gel was fixed in 50% methanol, 7% acetic acid for 10 min and washed in 10% methanol, 7% acetic acid twice. The gel was then dried using a Bio-Rad gel dryer for 2 h and exposed to a phosphoimaging screen overnight. The screen was imaged using a Typhoon imager, and images were analyzed using Image Quant. In some experiments, Shield-1 was added to cells just prior to methionine pulse and kept in the culture media during the chase.
Statistics and protein half-life calculation Statistical and regression analyses and half-life determinations were performed with GraphPad Prism software.
Results Creation and expression of SCRAP in EL4 cells To study independently the generation of peptides from DRiPs versus retirees, we genetically fused the ProteoTuner destabilization domain to SIINFEKL followed by eGFP to create SCRAP (Fig. 1A). We selected stable EL4 cell transfectants expressing SCRAP under the control of the CMV immediate early promoter. EL4/SCRAP cells exhibit minimal eGFP fluorescence detected above control EL4 autofluorescence levels (Fig. 1B, 1C). Overnight treatment with Shield-1 greatly increases eGFP fluorescence in a dose-dependent manner (Fig. 1C), with saturation attained
FIGURE 2. MHC class I Ag presentation occurs in the presence or absence of stabilized substrates. A, EL4/SCRAP cells were washed briefly in cold citric acid buffer to remove existing Kb-SIINFEKL complexes from the cell surface. Cells were then cultured in the presence of increasing concentrations of Shield-1 and analyzed for Kb-SIINFEKL expression by flow cytometry. Regression analysis was performed, and the best fit line is shown. B, EL4/SCRAP cells were treated as in A and cultured with EtOH alone, BFA, CHX, or MG132. Cells were harvested at indicated times and analyzed for Kb-SIINFEKL expression by flow cytometry and plotted as in A. C, EL4/SCRAP cells were washed in citric acid buffer and resuspended in culture media containing EtOH, Shield-1, MG132, or BFA. Five hours after acid-stripping, cells were analyzed for total Kb and Db expression by flow cytometry. D, EL4/SCRAP cells were cultured in 5 mM Shield-1 or 10 mM MG132 and total cell lysates prepared for Western blot analysis at indicated times with anti-eGFP Abs. The GFP signal from the entire lane was determined by densitometry and compared with a cyclophilin B loading control.
2068
DISTINCT MHC I Ag PROCESSING PATHWAYS
FIGURE 3. Retired SCRAP protein minimally contributes to antigenic peptide presentation. EL4/SCRAP cells were washed in citric acid buffer and cultured with or without 5 mM Shield-1 for 3.5 h and then washed again in citric acid buffer followed by treatment with CHX and either DMSO or MG132. Cells were collected at indicated times and analyzed for both eGFP (A) and Kb-SIINFEKL (B) expression by flow cytometry. The background signal for KbSIINFEKL (determined immediately after each acid wash for each respective treatment) was subtracted from each time point. C, Kb-SIINFEKL generated from retired SCRAP was also measured in the absence of CHX to ensure that any complications from the addition of CHX did not impact retiree presentation. The experiments were repeated in triplicate and analyzed by Student t test for significance.
at 2.5 mM. (Note that at lower concentrations, the effect of Shield1 is transient; this is probably due to metabolism of the drug.) The stabilizing effect of Shield-1 is completely reversible; when EL4/ SCRAP cells with a large pool of Shield-1–stabilized SCRAP were washed and cultured in the absence of Shield-1, GFP fluorescence decreased to background SCRAP levels (Fig. 1B, right panel). Although eGFP is a convenient tag for flow cytometric monitoring of protein expression and catabolism, not all SCRAP-gene products will fluoresce due to misfolding and mistranslation. We therefore examined SCRAP expression in the presence and absence of Shield-1 by immunoblotting with a GFP-specific mAb. This revealed a species with the predicted mobility of SCRAP (∼42 kDa), whose levels increase with Shield-1 addition (Fig. 1D) and falls dramatically with Shield-1 removal. We next labeled EL4/ SCRAP cells with [35S]methionine for 10 min in the presence or absence of Shield-1 and chased for up to 130 min in the presence or absence of Shield-1. Total cell lysates were prepared using 95˚C SDS-containing extraction buffer to maximize Ag recovery, and after neutralizing SDS denaturing activity, we analyzed species that interact with a-GFP Abs by SDS-PAGE. SCRAP levels
FIGURE 4. SCRAP expression by rVV results in rapid Ag presentation regardless of protein stabilization. DC2.4 cells were infected with rVV/ SCRAP at a multiplicity of infection of 10 for 30 min and then cultured in 5 mM Shield-1 or equivalent amount of EtOH, or CHX, or MG132 and harvested at the indicated times for flow cytometry analysis of eGFP (top) or Kb-SIINFEKL (bottom) expression.
rapidly decline in the absence of Shield-1 but remain virtually unchanged for up to 2 h in the presence of Shield-1 (Fig. 1E). Another GFP-reactive species present in the immunoprecipitation, GFP 21, decayed with rapid half-life as previously reported (11). Notably, the initial amount of GFP 21 detected was increased by Shield-1, demonstrating that it is derived from the folded fusion protein and is not the result of downstream initiation or other
FIGURE 5. Eer1 inhibits Ag presentation of SCRAP DRiPs. A, EL4/ SCRAP cells were prepared as in Fig. 2 and cultured with ethanol or 5 mM Shield-1 combined with DMSO or 5 mM Eer1. At indicated times, cells were harvested and analyzed for Kb-SIINFEKL expression by flow cytometry. B, After acid wash, cells were cultured for 100 and 300 min in the presence of Eer1, and the relative recovery of both Kb (top) and Db (bottom) was determined by flow cytometry. Relative recovery was set to 1 for mock-treated cells and 0 for BFA-treated cells. CHX and MG132 are shown as a positive control. Staining was repeated in triplicate with SE reported. C, EL4/SCRAP cells were treated with Shield-1 for 3 h while being treated with either DMSO, MG132, or Eer1 and total cell lysates analyzed for SCRAP expression by immunoblotting. EL4/SCRAP cells were also treated with Shield-1 for 3 h then washed and cultured in DMSO, MG132, or Eer1 for 3 h. Again, total cell lysates were analyzed for SCRAP expression by immunoblotting.
The Journal of Immunology forms of DRiPs. This species likely results from autocatalytic cleavage of GFP during active site cyclization (15). Fig. 1F displays decay curves of full-length SCRAP as determined by flow cytometry, immunoblotting, and pulse-chase analyses. SCRAP unfolded due to Shield-1 removal is degraded with a half-life of 30 min (as determined by immunoblotting and flow cytometry analyses), whereas nascent SCRAP is degraded twice as rapidly (t = 16 min), demonstrating that nascent and retired SCRAP are handled in a distinct manner. Taken together, these data indicate that by using Shield-1, we can control the stability of both nascent and mature, native SCRAP. Presentation of SIINFEKL is only partially affected by protein stability We next used Alexa 647-conjugated 25-D1.16 to examine the effect of Shield-1 on generation of Kb-SIINFEKL complexes from SCRAP via flow cytometry. To increase the resolution of detection, we briefly exposed cells to pH 3 to destroy preexisting KbSIINFEKL complexes. We then incubated cells at 37˚C in growth medium and sampled every 60 min. Kb-SIINFEKL complexes appear on the cell surface within 60 min after acid stripping and increase linearly over the next 5 h. Importantly, in the presence of a saturating Shield-1 concentration (5 mM), the generation of Kb-SIINFEKL complexes is reduced by as little as 30% in some experiments (Fig. 2A), as deduced from rates determined by the slope of the linear regression curves. Over five experiments, the average reduction was 36% (3% SE) in five experiments. Complex generation requires both nascent protein synthesis and proteasome activity (Fig. 2B), as it is completely blocked by treating cells with either cycloheximide (CHX) or MG132, which inhibit, respectively, protein synthesis and proteasome activity within seconds of their addition. As expected, Kb-SIINFEKL complex expression was completely blocked by brefeldin A (BFA), which prevents egress of complexes from the
FIGURE 6. Eer1 partially inhibits DRiP Ag presentation but has a minimal effect on presentation of other forms of Ags during rVV infection. A, DC2.4 cells were infected with either NP-S-GFP, eGFP-UbS, or partially UV-inactivated eGFP-Ub-S expressing rVV for 30 min. Cells were then cultured with or without Eer1 and harvested at indicated times postinfection and Kb-SIINFEKL expression determined by flow cytometry analysis. B, DC2.4 cells were infected as in A, and total cell lysates were prepared 3.5 h postinfection. Immunoblotting with GFP-specific monoclonal Abs was used to determine the relative levels of antigenic protein expressed by infected cells. C, BMDCs were infected with rVV-SCRAP and cultured with EtOH or Shield-1 in the presence of either DMSO or Eer1 and harvested at indicated times for flow cytometry analysis. Background Kb-SIINFEKL (determined by staining uninfected cells) was subtracted from the reported values and the regression analysis plotted.
2069 endoplasmic reticulum (ER) (Fig. 2B). The specificity of Shield-1 treatment is clearly demonstrated by its lack of effect on overall cell surface Kb or Db molecules, whereas both MG132 and BFA, as expected, greatly impeded recovery of total MHC I after acidwashing (Fig. 2C). The generation of Kb-SIINFEKL complexes in the presence of 5 mM Shield-1 cannot be attributed to residual proteasome degradation of protein detected by standard immunoblotting methodology, as levels of SCRAP rescued by Shield-1 and MG132 are indistinguishable (Fig. 2D). These data indicate that the substrate responsible for DRiP presentation cannot be readily determined by standard biochemical means and likely represents a small fraction of the SCRAP synthesized. These data indicate that although decreasing nascent protein stability enhances peptide presentation, the effect is partial. This is due to the existence of a translation pool of DRiPs that are resistant to SCRAP stabilization and serve as the source of .50% of the antigenic peptides. As we failed to detect the DRiP pool by standard biochemical analysis, it suggests that this source of Ags constitutes a minor fraction of total SCRAP protein translated by the cell. In any event, the ability of Shield-1 to allow selectively peptide generation from DRiPs provides a tool for teasing out differences between the processing of DRiPs versus other sources of antigenic peptides. Retirement is a much less efficient source of antigenic peptides We next cultured acid-washed EL4/SCRAP cells with Shield-1 for 3.5 h to generate a native protein pool that we could “retire” by removing Shield-1. Cells were then rewashed in acid and resuspended in media lacking Shield-1 but containing CHX to prevent synthesis of new SCRAP protein. Degradation of SCRAP retirees resulted in a slight but statistically significant (p , 0.05) increase in Kb-SIINFEKL complexes compared with that of cells treated with ethanol vehicle alone or with MG132 to prevent SCRAP degradation (Fig. 3). Complexes presented in the presence of MG132
2070 represent complexes present in the secretory pathway during acid treatment. Thus, complexes generated from the SCRAP retirement pool represent the difference between MG132-treated and -untreated cells. Extended CHX treatment may impact protein degradation (16) and thus mask the true number of complexes generated by Shield-1 removal. We therefore also cultured cells in the absence of CHX after Shield-1 removal. An increase in the number of Kb-SIINFEKL complexes similar in magnitude to cells treated with CHX was observed after retirement of SCRAP proteins (Fig. 3C) indicating that CHX treatment did not impact SCRAP degradation or Ag presentation. The number of complexes generated from the SCRAP pool accumulated over 235 min is equivalent to the number of complexes generated during ∼36 min of SCRAP synthesis in the presence of Shield-1. By this simple time-of-synthesis analysis, generating peptides from DRiPs is 6.5 times as efficient as generating them from retirees, providing a clear example of the importance of DRiPs in generating peptides from cellular gene products. Presentation of SCRAP peptide expressed from rVV Little is known about the differential contribution of processing pathways of a given gene product in different translational contexts. To address this issue, we inserted SCRAP into vaccinia virus under the control of the early/late p7.5 promoter. We infected DC2.4 cells with rVVs in the presence or absence of Shield-1 and measured GFP and Kb-SIINFEKL complexes over time. As with transfected cells, the GFP signal remained low in the absence of Shield-1 in DC2.4 cells, increasing nearly linearly in the presence of Shield-1, consistent with a steady rate of translation and negligible degradation relative to the synthesis time window (Fig. 4). Complex formation required proper trafficking through the secretory pathway, de novo protein synthesis, and a functional proteasome, as the respective inhibitors prevented Ag presentation (Fig. 4). The rate of Kb-SIINFEKL complex generation was decreased 28% by Shield-1 (Fig. 4), remarkably similar to the findings with EL4 transfectants. These findings demonstrate that SCRAP Ag processing is highly similar when expressed by viral versus cellular translation machinery.
DISTINCT MHC I Ag PROCESSING PATHWAYS inhibit DRiP Ag presentation by blocking protein synthesis or proteasome-mediated degradation. To extend these results to other Ags, we infected DC2.4 cells with rVVs expressing a fusion protein consisting of influenza A virus nucleoprotein, SIINFEKL, and GFP (NP-S-GFP). Eer1 treatment of infected cells resulted in an ∼50% decrease in KbSIINFEKL levels (Fig. 6A). By contrast, Eer1 had little effect on the generation of Kb-SIINFEKL complexes from rVV-GFP-UbSIINFEKL, which expresses preprocessed cytosolic SIINFEKL liberated co-translationally by ubiquitin hydrolases from the fusion protein. (Note that we equalized the amount of Kb-SIINFEKL complexes generated by the two viruses by partially inactivating the latter virus by UV irradiation, as otherwise it generates saturating amounts of SIINFEKL.) The failure of Eer1 to interfere with SIINFEKL presentation from GFP-Ub-SIINFEKL demonstrates that Eer1 does not block presentation by interfering with Kb biogenesis per se, which is critical given the recent report that Eer1 can inhibit the import of some proteins into the ER (19). Eer1 treatment of vaccinia virus-infected cells did not significantly affect expression of NP-S-GFP or GFP-Ub-S determined by immunoblotting with anti-GFP Abs (Fig. 6B). Eer1 treatment of BMDCs infected with rVV-SCRAP in the presence or absence of Shield-1 resulted in a similar trend of inhibited DRiP Ag presentation (Fig. 6C). Together, these data demonstrate that Eer1 inhibits DRiP presentation in a manner independent of protein synthesis, proteasome activity, TAP-mediated peptide transport into the ER, or peptide
Eeyarestatin I selectively inhibits DRiP Ag presentation Uniquely, our SCRAP system allows us to dissect processing of DRiPs versus other rapidly degraded polypeptides (RDPs) (11) or retirees by the simple addition of Shield-1 to limit presentation to DRiPs. In screening drugs that selectivity modulate DRiP processing, we tested eeyarestatin I (Eer1), originally described as an inhibitor of ER-associated degradation (17, 18). We found that Eer1 reduces overall generation of Kb-SIINFEKL by ∼50%. In the presence of Shield-1, Eer1 reduced Kb-SIINFEKL generation to nearly background levels, demonstrating that it inhibits processing of SIINFEKL from SCRAP DRiPs while sparing processing from Shield-1–responsive RDPs (Fig. 5A). After acid stripping of class I molecules, Eer1 treatment reduced Kb or Db recovery over the next 100 or 300 min by ∼20% (Fig. 5B). This effect must be gauged by the partial effects of CHX and MG132 on recovery and is consistent with a significant effect of Eer1 on a sizeable pool of class I peptide ligands. How does Eer1 interfere with Ag processing? We cultured EL4/ SCRAP cells with Shield-1, removed Shield-1, and treated cells with either MG132 or Eer1. Eer1 had no significant effect on SCRAP accumulation or proteasome-mediated degradation (Fig. 5C), which is consistent with its lack of effect on Shield-1–dependent presentation. This finding demonstrates that Eer1 does not
FIGURE 7. Cl-PGA2 and Eer1 inhibit DRiP Ag presentation and increase levels of polyubiquitinated proteins in EL4/SCRAP cells. A, EL4/ SCRAP cells were treated as in Fig. 5 except Cl-PGA2 was used in place of Eer1. B, EL4/SCRAP cells were treated with either DMSO, MG132, Eer1, or Cl-PGA2 for 3 h and total cell lysates prepared for analysis by immunoblotting with Abs for cyB (top) or polyUb (bottom). C, The ratio of polyUb to cyB as determined by Western blot. D, A model for DRiP-specific Ag presentation where newly synthesized DRiPs containing a polyubiquitin chain are deubiquitinated by the action of DUBs and processed by the 20S proteasome for peptide generation and Ag presentation.
The Journal of Immunology loading onto class I, therefore defining a novel step of the DRiP Ag presentation pathway. Both Eer1 and Cl-PGA2 increase levels of polyubiquitylated proteins and inhibit DRiP presentation Eer1 is believed to inhibit ER-associated degradation by binding to the AAA ATPase p97, leading to inhibition of substrate deubiquitylation by one or more p97-associated deubiquitinating (DUB) enzymes. Accordingly, Eer1 treatment is known to increase the levels of polyubiquitylated proteins (17). We examined DUB participation in DRiP processing by treating EL4/SCRAP cells with another DUB inhibitor, 10-chloro-15-acetyl-1-methylesterPGA2 (hereafter termed Cl-PGA2) (20). Cl-PGA2 behaved nearly identically to Eer1 in selectively inhibiting Ag presentation from the Shield-1–resistant SCRAP pool (Fig. 7A). Treating cells with either Eer1 or Cl-PGA2 increased cellular levels of polyubiquitylated proteins without inhibiting proteasomemediated SCRAP degradation, consistent with the participation of DUBs in Ag processing (Fig. 7B, 7C). It is important to note that Eer1 and Cl-PGA2 have widely divergent chemical structures, greatly reducing the chance of their acting on a common non-DUB target. Rather, their highly similar effects on Ag processing strongly suggests that DUBs are necessary for the generation of peptides from SCRAP DRiPs whereas they are dispensable for generating peptides from Shield-1–responsive RDPs.
Discussion The DRiP hypothesis was formulated largely to explain how peptides derived from stable viral proteins are presented so rapidly after viral infection (5). The key concept is that gene products possess multiple half-lives (7, 21), with the rapidly degraded cohort providing an immediate source of peptide Ags. Mounting evidence from multiple systems supports the conclusion that DRiPs provide a major, even the principal source of peptides for viral and host Ags (22, 23). The physical nature of DRiPs remains elusive. Although upwards of 20% of nascent proteins are degraded with an average t1/2 of ∼7 min (11, 21), not all of these RDPs are necessarily relevant for Ag processing. Qian et al. (11) reported that antigenic peptides derive from a minor fraction of RDPs that are degraded in a Ub-independent manner by proteasomes. This finding along with current and past (22) findings that retirees are an inefficient source of peptides strongly imply that proteasome substrates are highly heterogeneous with regard to their access to the class I processing pathway. To dissect the complexities of the class I processing pathway, we applied the ProteoTuner system to generate a gene product that can be targeted for immediate proteasome destruction as either a nascent or folded protein. This system is ideal for identifying drugs (or small interfering RNAs) that selectively interfere with distinct pathways that generate class I peptide ligands. We used this system to implicate DUBs in the generation of peptides from DRiPs. How might DUBs contribute to Ag processing? We cannot rule out the possibility that DUB inhibitors inhibit DRiP processing by their downstream effects on other cellular pathways, which, given the involvement of DUBs in myriad cellular processes, are likely to occur. Their selective targeting of DRiPs versus other RDPs is far more consistent, however, with the direct participation of DUBs in DRiP processing, as is the insensitivity of SIINFEKL presentation to DUB inhibitors. In Qian et al. (11), using cells lacking functional E1 Ubactivating enzyme or 19S proteasome regulators, we reported that SIINFEKL is liberated by 20S proteasomes from NP-S-GFP
2071 DRiPs independently of ubiquitylation. Our current findings that presentation of DRiPs requires DUB activity is consistent with the idea that in cells with a functional E1, DRiP Ag processing entails proteasome degradation of substrates generated by DUB action (Fig. 7D). What of the inherently rapidly degraded SCRAP (i.e., Shieldsensitive) and the subset of NP-SIINFEKL-GFP based substrates that are apparently unaffected by the DUB inhibitors? It is possible that degradation of these pools is Ub dependent, and substrates are processed by the 26S proteasome. Our observation that Kb and Db recovery after acid stripping is partially reduced by DUB inhibitors (Fig. 5B) is consistent with the idea that many peptides are generated by a DUB-independent pathway. It is possible that DUB-dependent peptides predominate when DUBs are fully functional but can be supplemented by DUB-independent peptides when DUBs are inhibited. It is important to note that dissecting the contribution of individual DUBs to Ag processing, although of obvious importance in future studies, is greatly complicated by the existence of ∼100 genes with known or suspected DUB activity (24). The efficiency of peptide generation from the various SCRAP pools is a question of obvious interest. The pool of true SCRAP DRiPs (insensitive to Shield-1) appears to be a small fraction of nascent chains based on the equivalent recovery of SCRAP in total cell lysates in immunoblots when cells are treated with MG132 versus Shield-1. Despite their biochemical invisibility, the Shield1–resistant pool generates peptides at 2.5 times the rate of the Shield-1–sensitive pool of nascent proteins and at 6.5 times the rate of Shield-1–withdrawal retirees. If Shield-1–resistant DRiPs represent 10% of total translation products (arguably, the minimal level of detection by immunoblotting), then they are, respectively, a 25-fold and 65-fold more efficient source of peptides than Shield-1–responsive RDPs and retirees (with the true efficiency adjusted depending on the actual prevalence of Shield-1–resistant DRiPs). Our demonstration that RDPs are compartmentalized into efficient versus inefficient sources of peptides is consistent with the idea that co-translational peptide generation is channelized to maximize efficiency. Our recent findings that cytosolic peptides do not compete with DRiP-derived peptides lend further support to this concept (25). The nonstoichiometric representation of peptides versus translated mRNAs first noted by Boon and Van Pel (26) and observed in many other systems clearly indicates that many immunologically important peptides derive from a specialized translation compartment, as proposed by the immunoribosome hypothesis (27).
Acknowledgments We thank Thomas J. Wandless (Stanford University) for the gift of the pBMN FKBP12 L106P vector and initial quantities of Shield-1 and Yihong Ye and William Trenkle (National Institute of Diabetes and Digestive and Kidney Diseases) for Eer1.
Disclosures The authors have no financial conflicts of interest.
References 1. Yewdell, J. W., and S. M. Haeryfar. 2005. Understanding presentation of viral antigens to CD8+ T cells in vivo: the key to rational vaccine design. Annu. Rev. Immunol. 23: 651–682. 2. Vyas, J. M., A. G. Van der Veen, and H. L. Ploegh. 2008. The known unknowns of antigen processing and presentation. Nat. Rev. Immunol. 8: 607–618. 3. Hansen, T. H., and M. Bouvier. 2009. MHC class I antigen presentation: learning from viral evasion strategies. Nat. Rev. Immunol. 9: 503–513.
2072 4. Rock, K. L., D. J. Farfa´n-Arribas, and L. Shen. 2010. Proteases in MHC class I presentation and cross-presentation. J. Immunol. 184: 9–15. 5. Yewdell, J. W., L. C. Anto´n, and J. R. Bennink. 1996. Defective ribosomal products (DRiPs): a major source of antigenic peptides for MHC class I molecules? J. Immunol. 157: 1823–1826. 6. Princiotta, M. F., D. Finzi, S. B. Qian, J. Gibbs, S. Schuchmann, F. Buttgereit, J. R. Bennink, and J. W. Yewdell. 2003. Quantitating protein synthesis, degradation, and endogenous antigen processing. Immunity 18: 343–354. 7. Schubert, U., L. C. Anto´n, J. Gibbs, C. C. Norbury, J. W. Yewdell, and J. R. Bennink. 2000. Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 404: 770–774. 8. Khan, S., R. de Giuli, G. Schmidtke, M. Bruns, M. Buchmeier, M. van den Broek, and M. Groettrup. 2001. Cutting edge: neosynthesis is required for the presentation of a T cell epitope from a long-lived viral protein. J. Immunol. 167: 4801–4804. 9. Dolan, B. P., L. Li, K. Takeda, J. R. Bennink, and J. W. Yewdell. 2010. Defective ribosomal products are the major source of antigenic peptides endogenously generated from influenza A virus neuraminidase. J. Immunol. 184: 1419– 1424. 10. Qian, S. B., E. Reits, J. Neefjes, J. M. Deslich, J. R. Bennink, and J. W. Yewdell. 2006. Tight linkage between translation and MHC class I peptide ligand generation implies specialized antigen processing for defective ribosomal products. J. Immunol. 177: 227–233. 11. Qian, S. B., M. F. Princiotta, J. R. Bennink, and J. W. Yewdell. 2006. Characterization of rapidly degraded polypeptides in mammalian cells reveals a novel layer of nascent protein quality control. J. Biol. Chem. 281: 392–400. 12. Porgador, A., J. W. Yewdell, Y. Deng, J. R. Bennink, and R. N. Germain. 1997. Localization, quantitation, and in situ detection of specific peptide-MHC class I complexes using a monoclonal antibody. Immunity 6: 715–726. 13. Banaszynski, L. A., L. C. Chen, L. A. Maynard-Smith, A. G. Ooi, and T. J. Wandless. 2006. A rapid, reversible, and tunable method to regulate protein function in living cells using synthetic small molecules. Cell 126: 995–1004. 14. Ratnayake, A. S., T. S. Bugni, C. A. Veltri, J. J. Skalicky, and C. M. Ireland. 2006. Chemical transformation of prostaglandin-A2: a novel series of C-10 halogenated, C-12 hydroxylated prostaglandin-A2 analogues. Org. Lett. 8: 2171–2174. 15. Barondeau, D. P., C. J. Kassmann, J. A. Tainer, and E. D. Getzoff. 2006. Understanding GFP posttranslational chemistry: structures of designed variants that
DISTINCT MHC I Ag PROCESSING PATHWAYS
16.
17. 18.
19.
20.
21.
22. 23. 24.
25.
26.
27.
achieve backbone fragmentation, hydrolysis, and decarboxylation. J. Am. Chem. Soc. 128: 4685–4693. Borrelli, M. J., Y. J. Lee, H. E. Frey, J. P. Ofenstein, and J. R. Lepock. 1991. Cycloheximide increases the thermostability of proteins in Chinese hamster ovary cells. Biochem. Biophys. Res. Commun. 177: 575–581. Wang, Q., L. Li, and Y. Ye. 2008. Inhibition of p97-dependent protein degradation by eeyarestatin I. J. Biol. Chem. 283: 7445–7454. Fiebiger, E., C. Hirsch, J. M. Vyas, E. Gordon, H. L. Ploegh, and D. Tortorella. 2004. Dissection of the dislocation pathway for type I membrane proteins with a new small molecule inhibitor, eeyarestatin. Mol. Biol. Cell 15: 1635–1646. Cross, B. C. S., C. McKibbin, A. C. Callan, P. Roboti, M. Piacenti, C. Rabu, C. M. Wilson, R. Whitehead, S. L. Flitsch, M. R. Pool, et al. 2009. Eeyarestatin I inhibits Sec61-mediated protein translocation at the endoplasmic reticulum. J. Cell Sci. 122: 4393–4400. Verbitski, S. M., J. E. Mullally, F. A. Fitzpatrick, and C. M. Ireland. 2004. Punaglandins, chlorinated prostaglandins, function as potent Michael receptors to inhibit ubiquitin isopeptidase activity. J. Med. Chem. 47: 2062–2070. Wheatley, D. N., M. R. Giddings, and M. S. Inglis. 1980. Kinetics of degradation of “short-” and “long-lived” proteins in cultured mammalian cells. Cell Biol. Int. Rep. 4: 1081–1090. Yewdell, J. W., E. Reits, and J. Neefjes. 2003. Making sense of mass destruction: quantitating MHC class I antigen presentation. Nat. Rev. Immunol. 3: 952–961. Yewdell, J. W., and C. V. Nicchitta. 2006. The DRiP hypothesis decennial: support, controversy, refinement and extension. Trends Immunol. 27: 368–373. Reyes-Turcu, F. E., K. H. Ventii, and K. D. Wilkinson. 2009. Regulation and cellular roles of ubiquitin-specific deubiquitinating enzymes. Annu. Rev. Biochem. 78: 363–397. Lev, A., M. F. Princiotta, D. Zanker, K. Takeda, J. S. Gibbs, C. Kumagai, E. Waffarn, B. P. Dolan, A. Burgevin, P. Van Endert, et al. 2010. Compartmentalized MHC class I antigen processing enhances immunosurveillance by circumventing the law of mass action. Proc. Natl. Acad. Sci. USA 107: 6964– 6969. Boon, T., and A. Van Pel. 1989. T cell-recognized antigenic peptides derived from the cellular genome are not protein degradation products but can be generated directly by transcription and translation of short subgenic regions. A hypothesis. Immunogenetics 29: 75–79. Yewdell, J. W. 2005. The seven dirty little secrets of major histocompatibility complex class I antigen processing. Immunol. Rev. 207: 8–18.
