Presentation by Class I MHC Molecules Requires Cytoplasmic

Home , E1 (HCV), E2 (HCV)

Hepatitis C Virus Envelope Glycoprotein E1 Originates in the Endoplasmic Reticulum and Requires Cytoplasmic Processing for Presentation by Class I MHC Molecules This information is current as of September 26, 2021. Mark Selby, Ann Erickson, Christine Dong, Stewart Cooper, Peter Parham, Michael Houghton and Christopher M. Walker J Immunol 1999; 162:669-676; ; http://www.jimmunol.org/content/162/2/669 Downloaded from

References This article cites 45 articles, 22 of which you can access for free at:

http://www.jimmunol.org/content/162/2/669.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 26, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Hepatitis C Virus Envelope Glycoprotein E1 Originates in the Endoplasmic Reticulum and Requires Cytoplasmic Processing for Presentation by Class I MHC Molecules1

Mark Selby,* Ann Erickson,* Christine Dong,* Stewart Cooper,† Peter Parham,† Michael Houghton,* and Christopher M. Walker2*

We investigated whether hepatitis C virus envelope glycoprotein E1 is transported from the endoplasmic reticulum (ER) to the cytoplasm of infected cells for class I MHC processing. Target cells expressing E1 were killed by CTL lines from a hepatitis C virus-infected chimpanzee, and synthetic peptides were used to deﬁne an epitope (amino acids 233-GNASRCWVA-241) presented by the Patr-B*1601 class I MHC molecule. An unusually high concentration (>100 nM) of this nonameric peptide was required 234 for target cell lysis, but this could be reduced at least 1000-fold by replacing the asparagine at amino acid position 234 (Asn ) Downloaded from with aspartic acid (Asp), the anticipated anchor residue for NH2-terminal peptide binding to Patr-B*1601. Conspicuously, position 234 is part of an N-glycosylation motif (Asn-Xaa-Ser/Thr), suggesting that the Asn234 to Asp substitution might occur naturally within the cell due to deglycosylation/deamidation of this amino acid by the cytosolic enzyme peptide N-glycanase. In support of this model, we demonstrate that presentation of the epitope depended on 1) cotranslational synthesis of E1 in the ER, 2) glycosylation of the E1 molecule, and 3) a functional TAP transporter to shuttle peptide from the cytosolic to ER compartment. These results indicate for the ﬁrst time that during infection of the host, viral envelope glycoproteins originating in the ER are processed http://www.jimmunol.org/ in the cytoplasm for class I MHC presentation. That a posttranslational change in amino acid sequence from Asn to Asp alters -the repertoire of peptides presented to CD8؉ CTL has implications for the design of antiviral vaccines. The Journal of Immu nology, 1999, 162: 669–676.

pitopes presented by class I MHC molecules are gener- Processing pathways for transmembrane glycoproteins that are ated in a multistep process that begins with attachment of not normally present in the cytoplasm remain undefined. Under E ubiquitin (Ub)3 to cytoplasmic or nuclear proteins that are some circumstances processing can occur in the ER rather than the then targeted to the 26S proteasome for degradation into peptides. cytoplasm. For instance, signal peptidase (5–8) or other unidenti- by guest on September 26, 2021 Translocation of peptides from the cytoplasm into the endoplasmic fied proteases (6, 9) can generate peptides that bind directly to reticulum (ER) is mediated by TAP, a heterodimer encoded by class I MHC molecules in the ER, thus bypassing the cytoplasmic TAP-1 and TAP-2 genes located in the MHC (1). Class I MHC Ub/proteasome/TAP pathway. Most epitopes from transmembrane molecules in the ER are closely associated with TAP and various glycoproteins are TAP dependent, however, suggesting that at chaperones that facilitate loading of antigenic peptides (2). A sta- least some processing events occur in the cytoplasm. ␤ ble complex consisting of a class I molecule, 2m, and a peptide At least two different mechanisms could account for cytoplas- of about 8–10 amino acids, is transported to the cell surface via the ϩ mic processing of proteins that normally localize to the ER. In the exocytic pathway for surveillance by CD8 CTL (3). This classi- first mechanism, transmembrane glycoproteins may occasionally cal processing pathway might not, however, generate all class I ϩ be synthesized on free cytoplasmic rather than membrane-bound MHC-associated peptides displayed by a cell. CD8 CTL also ribosomes, and thus targeted for ubiquitination and destruction by recognize transmembrane glycoproteins, including viral envelope the proteasome. Studies of an HLA-B35-restricted epitope from proteins, tumor Ags, and certain alloantigens, that are cotransla- the HIV-1 gp120 envelope indicate that this pathway is functional tionally synthesized in the ER on membrane-bound ribosomes (4). in virus-infected cells (9, 10). The gp120 envelope contains a num- ber of signal motifs for Asn-linked glycosylation (Asn-Xaa-Ser/ *Chiron Corp., Emeryville, CA 94608; and †Department of Structural Biology, Stan- Thr, where Xaa is any amino acid except Pro), and all are known ford University, Stanford, CA 94305 to be modified by glycans during synthesis of the polyprotein in Received for publication August 10, 1998. Accepted for publication September the ER (11). One of these Asn residues is contained in the epitope, 24, 1998. but surprisingly was not glycosylated before presentation by the The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance HLA-B35 molecules (10). This result suggests that gp120 polypro- with 18 U.S.C. Section 1734 solely to indicate this fact. tein produced in the cytoplasm because of an error in translation 1 This work was supported by Chiron Corp., National Institutes of Health Grant (12–14) or possibly signal peptide-mediated targeting to the ER (9, AI31168 (to P.P.), and a postdoctoral fellowship from the Arthritis Foundation (to S.C.). 10) was the substrate for class I MHC processing. A second mechanism appears to involve export of defective 2 Address correspondence and reprint requests to Dr. Christopher Walker, Department of Pediatrics, Children’s Hospital, Room W503, 700 Children’s Dr., Columbus, OH transmembrane glycoproteins from the ER to the cytoplasm for 43205. E-mail address: [email protected] proteasome-mediated destruction (15–20). This pathway is sup- 3 Abbreviations used in this paper: Ub, ubiquitin; ER, endoplasmic reticulum; ported by recent studies of class I MHC-restricted epitopes in the PNGase, peptide:N-glycanase; E1, envelope glycoprotein 1 of hepatitis C virus; HCV, hepatitis C virus; NS3, nonstructural 3; moi, multiplicity of infection; LCL, lympho- influenza virus nucleoprotein (21) and the cellular glycoprotein blastoid cell line. tyrosinase that is expressed in melanocytes (22, 23). In the case of

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 670 AN ER TO CYTOSOL Ag PROCESSING PATHWAY FOR VIRAL GLYCOPROTEINS tyrosinase, it was demonstrated that efficient presentation of one der of E1 to amino acid 383. The internal primers were designated E1-M5 TAP-dependent epitope by HLA-A2 required posttranslational and E1-M3 (CGCGAGGGCGACGCCTCTAGATGTTGG and CGCCAC- conversion of an encoded Asn residue to Asp (22). This required CCAACATCTAGAGGCGTCGCC). The identity of the resulting clones was confirmed by sequencing. Subsequently, the E1-coding sequences cotranslational glycosylation of tyrosinase in the ER and subse- were transferred into pSC11 for generation of recombinant vaccinia vi- quent translocation of the Ag to the cytoplasm where it was deg- ruses. An 80% confluent monolayer of African green monkey BSC-40 cells lycosylated (23), probably by peptide:N-glycanase (PNGase) (24, on a 60-mm dish was infected with 0.05 plaque-forming units of wild-type vaccinia virus (WR strain)/cell. After2hofinfection at 37°C, the cells 25). Removal of glycans by this cytoplasmic enzyme causes a ϩ ϩ were transfected with 12.5 mg of pSC11 E1sϩ, pSC11 E1sϪ, and coding change from Asn to Asp by a process of deamidation, and ϩ pSC11 E1sϪD234 DNA using lipofectin reagent (Life Technologies, it is this modified form of tyrosinase that is processed by the Ub/ Grand Island, NY). Virions were harvested 2 days later, and serial dilutions proteasome/Tap pathway. Whether this processing pathway ap- (from 10Ϫ1 to 10Ϫ4) of each were used to infect 80% confluent monolayers plies generally to other glycoproteins is not yet known. of thymidine kinase-deficient human osteosarcoma 143 B cells on 60-mm In this study we demonstrate for the first time that the ER to dishes. After2hat37°C the inoculum was aspirated off the cells. An overlay that included 1% sea plaque agarose and 50 mg/ml 5-bromode- cytosol processing pathway is not restricted to cellular glycopro- oxyuridine was added to infected monolayers, and 3 days later the dishes teins such as tyrosinase, but is also operative for viral glycopro- were overlaid with 1% sea plaque agarose containing 0.03% 5-bromo-4- teins produced in infected cells. Class I MHC-restricted CTL spe- chloro-3-indolyl-␤-D-galactoside. Each blue recombinant plaque was cific for envelope glycoproteins E1 and E2 of hepatitis C virus picked from a dish as an agarose plug and transferred into DMEM. The virus recovered from the plaque plug was used to infect monolayers of (HCV) are present in the liver and peripheral blood of infected BSC-40 cells to generate a virus stock. humans (26–31) and chimpanzees (32–34). For at least one TAP- Recombinant vaccinia virus VV-ICP47 expressing the ICP47 protein of dependent E1 epitope, cotranslational glycosylation was strictly herpes simplex virus type 1 (35) was kindly provided by Dr. Barry Rouse, Downloaded from required for Ag presentation, presumably because an associated University of Tennessee (Knoxville, TN). Asn to Asp substitution facilitated peptide binding to class I MHC Synthetic peptides molecules. This posttranslational modification also appeared to influence the repertoire of HCV-specific CD8ϩ CTL in an infected HCV-1 peptides were synthesized by Chiron Mimotopes (Clayton, Aus- tralia) or Research Genetics (Huntsville, AL) using F-moc solid phase chimpanzee, and therefore has implications for vaccine design. methods. All peptides had free amino and carboxyl termini. http://www.jimmunol.org/ Materials and Methods Cytotoxicity assays CTL lines A B lymphoblastoid target cell line was established from the peripheral Chimpanzee Ross (CH-503) developed a persistent infection after experi- blood of chimpanzee CH-503 by transformation with supernatant from the mental challenge with 100 chimpanzee infectious doses of the HCV-1/910 marmoset cell line B95-8 as previously described (32). Target cells were (genotype 1a) virus stock. Plasma viremia has been consistently detected infected for 1 h with recombinant vaccinia viruses expressing HCV-1 Ags through 5 yr of follow-up study (32). The class I MHC genotype of this at a multiplicity of infection (moi) of 10. In experiments involving VV- chimpanzee is Patr-A*0401, -A*1401, -B*1601, -B*1701, -C*0501, ICP47, targets were first infected for 1 h with this virus or Vwt at an moi -C*0601 (33). HCV-specific CTL lines were generated from the liver of of 15 and then superinfected with recombinant viruses expressing the HCV this animal as described previously (32). CTL line 503/11.3 is specific for E1 glycoprotein for an additional2hatamoiof5.Allvaccinia virus- by guest on September 26, 2021 a Patr-A*0401-restricted epitope spanning amino acids 588–596 (KHP- infected targets were washed and cultured overnight before labeling with 51Cr. In some experiments target cells were sensitized with synthetic DATYSR) of the HCV E2 glycoprotein (32, 33). CTL lines 503/13.4 and 51 ϫ 3 503/10D recognize Patr-B*1601-restricted epitopes in E1 and nonstruc- HCV-1 peptides during Cr labeling. After three washes, 5 10 target cells were added to duplicate wells of a 96-well plate with varying numbers tural 3 (NS3) proteins of HCV-1, respectively (32, 33). CTL lines specific ϩ for the Patr-B*1601-restricted E1 epitope were also generated by incuba- of CD8 effector cells. Target cells cultured alone in medium or detergent ϫ 6 (1% Nonidet P-40) provided the minimum and maximum release values, tion of 4 10 Ficoll-Hypaque-separated PBMC in RPMI 1640 culture ␮ medium containing rIL-2 (20 U/ml), 10% FCS, tetanus toxoid (1 ␮g/ml), respectively. Plates were incubated for4hat37°C, then 50 l of culture and synthetic HCV E1 peptides at 10 ␮g/ml as previously described (30). supernatant was harvested onto 96-well Lumaplates (Packard, Downers ϫ Grove, IL), dried overnight, and counted in a Wallac 1450 Microbeta liquid After 7 days, lymphocytes were washed once and restimulated with 1 51 106 autologous irradiated (3000 rad) PBMC in culture medium supple- scintillation counter (Wallac, Gaithersburg, MD). The percent specific Cr ϩ release was calculated using the formula: [(test release Ϫ minimum re- mented with IL-2 and synthetic HCV peptide. After a further 7 days, CD8 Ϫ ϫ T cells were enriched from the cultures using immunoaffinity beads (32) lease)/(maximum release minimum release)] 100. and were tested for HCV-specific lytic activity as described below. Tunicamycin treatment of target cells Recombinant vaccinia viruses Tunicamycin was added at a concentration of 10 ␮g/ml of culture medium Vaccinia viruses Vwt, VC/E1, and VE1/2 were described previously (32, 1 h before infection with recombinant vaccinia viruses, and targets were 33). Vwt is a control virus that does not contain HCV genes. VC/E1 ex- maintained in the presence of drug during assays for Ag expression or presses amino acids 1–384 of HCV-1 spanning the core and E1 proteins. susceptibility to CTL lysis. The m.w. of HCV E1 and E2 Ags in drug- treated and untreated cells were assessed by Western blot analysis. Briefly, VE1/2 encodes amino acids 134–966 encompassing the NH2-terminus of core, E1, E2, p7, and most of NS2. Recombinant vaccinia viruses express- cells were pelleted and resuspended in Laemmli sample buffer. The lysates were sheared with an insulin hypodermic syringe, and the debris was pel- ing full-length (VE1sϩ) or modified (VE1sϪ and VE1sϪD234) E1 Ags were constructed as follows. Briefly, E1 with and without its natural leader were leted. Supernatants were loaded onto 14% SDS gels for electrophoresis and generated by PCR amplification of a genomic HCV-1 template and cloned then transferred to nitrocellulose in 20% methanol/electrophoresis buffer. into the pTM1 vector at the NcoI/BamHI sites. Two different 5Ј primers, Blots were blocked and then incubated with the appropriate mAb diluted E15 and E1-192 (GCAGTCATGATTTGCTCTTTCTCTATCTTCC and 1/1000 (anti-E1:3D5/C3; anti-E2:3E5–1) for1hatroom temperature in AACTCATGATCTACCAAGTGCGCAACTCCACG), were used to spec- 0.2% reconstituted dried nonfat milk and 0.1% Tween-20 in PBS. Blots ify an initiator MET codon preceding an isoleucine followed by residue were washed, and a 1/20,000 dilution of goat anti-mouse peroxidase 172 or 192, respectively. Both primers encode BspHI sites, which, after (Boehringer Mannheim, Indianapolis, IN) was added for1hatroom tem- digestion of the PCR product, are compatible with the vector NcoI site. A perature. After washing, enhanced chemiluminescence reagents (Amer- single 3Ј primer, E1383 (TGGTAGATCTTACGCGTCGACGCCGG sham, Arlington Heights, IL) were added, and the blots were exposed to CAAA), was used to specify a TAA termination codon after amino acid film. 383, followed by a BglII site, which, after digestion of the PCR product, is compatible with the vector BamHI site. Two PCR products were generated Results to create the N to D amino acid substitution at residue 234. The first cor- ϩ responded to amino acids 192–236 in which amino acid 234 was altered to A CD8 CTL line designated 503/13.4 was established from liver encode D rather than N; a unique XbaI site was created distally to facilitate tissue biopsied 28 wk after infection of chimpanzee Ross with the ligation with the second PCR product, XbaItoBglII encoding the remain- genotype 1a HCV-1 isolate (32). Autologous target cells infected The Journal of Immunology 671 Downloaded from http://www.jimmunol.org/

FIGURE 2. A, Comparison of Patr-B*1601- restricted epitopes from the HCV NS3 and E1 proteins. B, Titration of 58D and 58N peptides on LCL. Target cells were incubated with 51Cr and varying concentrations of peptides 58N (F) and 58D (Œ) for 1 h, washed three times, and cocultured with CTL line 13.4 at an E:T cell ratio of 20:1 for 4 h.

FIGURE 1. A, Recognition of HCV E1 by a liver-derived CD8ϩ CTL line. CTL line 13.4 was tested for lytic activity against autologous LCL that difference at amino acid position 2 that is probably the dominant by guest on September 26, 2021 were uninfected (X) or infected with vaccinia viruses expressing no HCV NH -terminal anchor residue for peptide binding to Patr-B*1601 f Œ F 2 Ag (VVwt; ), E1 and E2 (VE1/2; ), and core and E1 (VC/E1; ). The (Fig. 2A). We hypothesized that the Asp at position 2 of the NS3 percent specific 51Cr release was assessed after coincubation of effector and epitope facilitated efficient presentation by Patr-B*1601, and that target cells for 4 h. B, Fine mapping of an HCV E1 epitope. Autologous 234 LCL were pulsed with 51Cr and a 200-nM concentration of the indicated the E1 epitope would be better presented if Asn was replaced by D peptide for 1 h, washed three times, and incubated with CTL line 13.4 at Asp. As shown in Fig. 2B, the substituted peptide (designated 58 ; an E:T cell ratio of 20:1 for 4 h. GDASRCWVA) sensitized target cells for lysis over a wide range of concentrations. Half-maximal lysis of targets occurred at a peptide concentration of about 0.1 nM, an amount comparable to that with recombinant vaccinia viruses expressing the HCV-1 E1 en- required for the NS3 epitope restricted by Patr-B*1601 (36). By velope glycoprotein were lysed by this CTL line (Fig. 1A). The contrast, peptide 58N failed to sensitize targets at any of the con- epitope recognized by CTL line 503/13.4 was identified by pulsing centrations tested in this experiment. target cells with a series of overlapping decameric peptides that To assess whether CTL populations against either 58N or 58D spanned the HCV-1 E1-coding sequence. Only one peptide (amino sequences dominated in vivo, PBMC obtained from CH-503 at 263 acids 232-EGNASRCWVA-241) sensitized the target cells (data wk postinfection were stimulated twice at weekly intervals with not shown), and further truncation of NH2- and COOH-terminal each individual peptide. CTL activity was assessed after an addi- amino acids revealed that a nine-amino acid peptide (amino acid tional week of culture. CD8ϩ T cells enriched from all six replicate 233-GNASRCWVA-241) is the minimum optimal epitope (Fig. PBMC cultures stimulated with peptide 58N had low or no detect- 1B). This peptide was designated 58N, where the superscript indi- able lytic activity against targets sensitized with the homologous cates the amino acid residue at position 234 in single letter code. peptide compared with that against control targets (Fig. 3A). In A striking feature to emerge from these studies was the high contrast, high levels of 58D-specific lytic activity were detected in concentration of peptide 58N required to sensitize target cells for all six cultures stimulated with this Asp-substituted peptide (Fig. lysis; a peptide concentration exceeding 200 nM was needed to 3B). One representative cell line (B58D.3) efficiently killed targets match the levels of killing detected against targets infected with infected with VE1/2 (Fig. 3C). Moreover, when target cells were VE1/2. Epitope 58N was previously shown to be presented by the treated with peptides 58N and 58D, only the latter was recognized class I MHC allotype Patr-B*1601 (33). Notably, a second Patr- over a concentration range from 10–0.001 nM (Fig. 3D). Thus, B*1601-restricted CTL line, designated 503/10D, was derived CTL against the E1 epitope were still present in chimpanzee Ross from the liver of the same chimpanzee. It recognized a nonameric after 5 yr of persistent infection and showed the same preference epitope in the NS3 protein of HCV-1, and efficient lysis of target for the Asp-substituted peptide as the intrahepatic CTL line 503/ cells was achieved with peptide concentrations as low as 0.1 nM 13.4 established during the acute phase of hepatitis C. This result (Fig. 2A) (36). Comparison of the E1 and NS3 epitopes revealed a strongly suggests that the dominant form of this epitope eliciting 672 AN ER TO CYTOSOL Ag PROCESSING PATHWAY FOR VIRAL GLYCOPROTEINS

FIGURE 3. Characterization of 58D- and 58N-speciﬁc CTL lines derived from peripheral blood. PBMC were stimulated twice with peptide 58N (A) or peptide 58D (B) and enriched CD8ϩ T cells were tested for lysis of autologous, 51Cr- labeled LCL that were untreated or sensitized with a 10-nM concentration of homologous (i.e., 58N or 58D) peptide or irrelevant control peptide 162A. C, Autologous target cells were uninfected (f) or were

infected with vaccinia viruses ex- Downloaded from pressing no HCV Ag (Vwt; ؋), E1 and E2 (VE1/2; Œ), and core and E1 (VC/E1; F). The percent speciﬁc lysis by CTL line B58D.3 was assessed in a 4-h assay. D, Target cells were sensitized with the indi- N cated concentration of peptide 58 http://www.jimmunol.org/ (F)or58D (Œ) and cocultured with CTL line B58D.3 at an E:T cell ratio of 20:1 for 4 h. by guest on September 26, 2021

CD8ϩ CTL in the infected host contains an Asp rather than an Asn the signal peptide sequence should prevent target cell recognition residue at amino acid position 234. by CD8ϩ CTL. Studies with a panel of recombinant vaccinia vi- Notably, Asn234 is part of an Asn-Xaa-Ser/Thr consensus gly- ruses expressing modified E1 Ags (Fig. 6A) support this conten- cosylation motif in the E1 protein (Fig. 2A). Modification of this tion. Virus VE1sϩ expressed full-length E1, which consists of an residue by glycosylation during cotranslational synthesis of E1 might therefore influence presentation of the epitope. To address this question, target cells infected with VE1/2 were treated with tunicamycin, an inhibitor of N-linked glycosylation, and then ex- amined for the ability to present the E1 epitope. E1 expressed in untreated cells had an apparent molecular mass of 33 kDa (Fig. 4). After treatment with tunicamycin the size of E1 was reduced to 18 kDa, consistent with the value predicted for the unglycosylated E1 polyprotein. Untreated target cells expressing fully glycosylated E1 were efficiently killed by the liver-derived CTL line 503/13.4, but recognition was abrogated when tunicamycin was used to block addition of Asn-linked glycans (Fig. 5A). To control for other possible effects of tunicamycin on the Ag-processing apparatus including class I MHC synthesis, treated target cells infected with VE1/2 were tested for lysis by CTL line 503/11.3 that recognizes a Patr-A*0401-restricted E2 epitope (amino acid 588-KHPDATYSR-596; Fig. 5B). Tunicamycin-treated targets were lysed at levels 50–75% of those of untreated targets, indi- cating that the drug did not substantially interfere with the ability of FIGURE 4. Western blot of tunicamycin-treated, VVE1/E2-infected the cell to process or present epitopes derived from HCV envelope cells. Lysates of LCL infected with a control VV (Vwt; lanes 1 and 2)or proteins. VE1/2 (lanes 3 and 4) at an moi of 10 were loaded onto a 14% gel for 234 If cotranslational glycosylation of the Asn residue is impor- electrophoresis. After transfer to nitrocellulose, blots were reacted with an tant for presentation of the E1 epitope by Patr-B*1601, then lim- HCV E1-specific mAb. Cells in lanes 2 and 4 were infected in the presence iting Ag expression to the cytoplasm of target cells by removal of of tunicamycin at a concentration of 10 ␮g/ml of culture medium. The Journal of Immunology 673

initiation codon followed by the natural signal peptide (amino acids 171–191) and mature envelope glycoprotein (amino acids 192– 383; Fig. 6A). Target cells infected with this virus or pulsed with peptide 58D were killed at equivalent levels, conﬁrming that the normal synthetic pathway for E1 in the infected cell, which involves cotranslational glycosylation in the ER, leads to efﬁcient presentation of the Patr-B*1601-restricted epitope (Fig. 6B). As predicted, presentation of this epitope was prevented if glycosylation of the Asn234 residue was blocked by restricting E1 expres-

sion to the cytoplasm. Target cells infected with VE1sϪ, which expresses E1 without its natural leader peptide (Fig. 6A), were not susceptible to CTL lysis (Fig. 6B). We reasoned that cytoplasmic expression of E1 would be sufﬁcient for class I MHC presentation of this epitope if an Asn to Asp mutation was introduced at position 234, bypassing any normal requirement for glycosylation/deglycosylation of this residue to create an optimal anchor for Patr-

B*1601 binding. A virus designated VE1sϪD234 that carries this

mutation in E1 (Fig. 6A) sensitized target cells (Fig. 6B). Indeed, Downloaded from

equivalent levels of killing against target cells infected with VE1sϩ

and VE1sϪD234 were observed at all three E:T cell ratios, indicat- ing that an N234 to D substitution completely replaced the requirement for a functional signal peptide in presentation of this epitope. FIGURE 5. Lysis of tunicamycin-treated target cells by HCV-envelope Processing of the E1 glycoprotein in the cytoplasm after retrograde specific CTL lines. Autologous LCL were untreated (f) or were infected transport from the ER should require a functioning TAP apparatus to http://www.jimmunol.org/ with control Vwt (}) or VE1/2 in the presence (Œ) or the absence ()of tunicamycin at a concentration of 10 ␮g/ml of culture medium. Targets reintroduce the deglycosylated epitope into the ER for class I MHC were then assessed for lysis by E1-specific CTL line 13.4 (A) or E2-specific presentation. To investigate this issue, target cells were infected with line 11.3 (B) at the indicated E:T cell ratios in a 4-h assay. a recombinant vaccinia virus expressing the HSV-1 ICP47 protein by guest on September 26, 2021

FIGURE 6. A, Recombinant vaccinia viruses expressing HCV E1 envelope proteins. The amino acid at position 234 is shown in brackets in single letter code. B, Cytolysis of target cells expressing HCV E1 Ags. Autologous LCL were untreated (Œ), sensitized with peptide 58D (F), or infected with vaccinia viruses expressing no HCV Ag (Vwt; ✖), E1 and E2 (VE1/2; f), full- length E1 with signal peptide sequence (VE1sϩ; ࡗ ✸ ), E1 lacking signal sequence (VE1sϪ; ), and E1 lacking signal sequence with an N2343D substitution (VE1sϪD234; ). 674 AN ER TO CYTOSOL Ag PROCESSING PATHWAY FOR VIRAL GLYCOPROTEINS

mune recognition by cotranslational synthesis and/or modification by Asn-linked glycans. Presentation of the HCV E1 epitope by Patr-B*1601 was at least 1000-fold more efficient when the Asn234 residue encoded by HCV-1 was replaced by Asp. Modeling studies indicate that this amino acid switch is required for peptide binding to the B pocket of Patr-B*1601, and the presence of an Asp residue at position 2 of the HCV NS3 epitope (Fig. 2A) is consistent with this view. Given the poor presentation of synthetic peptide 58N by Patr- B*1601, it seemed improbable that an epitope containing Asn234 could elicit CTL in the infected host. Indeed, expansion of CD8ϩ CTL from the PBMC of this infected chimpanzee with peptide 58D, but not 58N, supports the view that an Asn234 to Asp substitution occurs naturally in HCV-1-infected cells. Remarkably, an overlapping E1 epitope (233-GNASRCWVAM-242) is also recognized by CD8ϩ CTL from an HCV-infected human subject (26) even though peptide binding to the HLA-B35 restriction element was reportedly several orders of magnitude less efficient than that

of other synthetic peptides considered immunogenic (27). An Downloaded from Asn234 to Asp substitution could also explain presentation of this epitope in humans, because peptide binding to HLA-B*3501 is sometimes mediated by a position 2 Asp instead of the usually dominant Pro anchor residue (38). Posttranslational changes in pri- mary amino acid sequence from Asn to Asp within class I MHC-

restricted epitopes of HCV E1, tyrosinase (22, 23), and HIV-1 http://www.jimmunol.org/ gp120 (R. Ferris and R. Siliciano, unpublished observations) suggest that this processing pathway, involving deglycosylation/deamidation, is not uncommon and could have implications for vaccine design. Synthetic peptide or protein immunogens are usually based on the genetically encoded amino acid sequence of an Ag. FIGURE 7. HCV E1 epitope is TAP dependent. A, Target cells were Moreover, priming of CD8ϩ CTL with plasmid DNA or recom- Œ D ● untreated ( ), sensitized with peptide 58 ( ), or infected sequentially with binant viral vaccine vectors might be enhanced by removal of sig- Vwt/Vwt (f) or VV-ICP47/Vwt (}) as described in Materials and Meth- f f nal peptide from the encoded envelope Ag (39). Our results sug- ods. Targets were also infected with VE1sϪD234 (B, ) or VE1/2 (C, ) by guest on September 26, 2021 alone or after infection with Vwt (F) or VV-ICP-47 (Œ). The percent gest that the specificity of CTL primed by these vaccines would be specific lysis was assessed after a 4-h coculture with CTL line B58D.3 overlapping, but not necessarily identical, with those primed by viral infection. A difference in the repertoire of epitopes presented by vaccination vs infection is a potentially important issue for viruses such as HCV and HIV-1 that have multiple Asn residues in (35), which is known to inhibit TAP function (37), and then super- envelope proteins that are targeted for glycosylation. infected with VE1/2 or VE1 . As shown in Fig. 7A, CD8ϩ T sϪD234 Presentation of the Patr-B*1601 E1 epitope begins with trans- cells lysed targets pulsed with peptide 58D, but not those infected with lation of the positive strand HCV genome into a single polyprotein Vwt and/or VV-ICP47. Class I MHC presentation of epitopes from that is cleaved into structural and nonstructural subunits by host cytoplasmic E1 should be strictly dependent on TAP, and thus we cell and viral proteases (40). E1 contains an NH -terminal signal evaluated whether ICP47 would prevent recognition of targets in- 2 peptide for cotranslational synthesis and a COOH-terminal type I fected with VE1 . The data presented in Fig. 7B clearly dem- sϪD234 membrane anchor for insertion into the lumenal face of the ER onstrate that cytoplasmic synthesis of E1 encoding the optimal Asp234 membrane (41, 42). Glycosylation of E1 was a prerequisite for residue resulted in efficient lysis of target cells, and this was prevented presentation of the Patr-B*1601 epitope because target cell recog- by prior infection with VV-ICP47 but not Vwt. Significantly, infec- nition was prevented by tunicamycin treatment or removal of the tion of the targets with VV-ICP47 before superinfection with VE1/2, signal peptide that directs ER synthesis. Strict dependence on a which expresses full-length E1 with signal peptide, also prevented functional TAP apparatus nevertheless suggests that the E1 gly- presentation of the epitope (Fig. 7C). These results demonstrate that coprotein is not processed in the ER, but instead is translocated to TAP is needed for Patr-B*1601-restricted recognition of this epitope the cytoplasm, where it enters the Ub/proteasome/TAP pathway. regardless of whether E1 is synthesized in the cytoplasm (VE1 ; sϪD234 Retrograde transport of Ags in mammalian and yeast cells is prob- Fig. 7B) or the ER (VE1/2; Fig. 7C). ably mediated by Sec61 (16, 43), a heterotrimeric complex com- posed of ␣, ␤, and ␥ subunits that form a protein-conducting chan- Discussion nel in the ER membrane (44, 45). The pore is normally used for This study demonstrates for the first time that viral envelope gly- extrusion of glycoproteins into the ER during cotranslational syn- coproteins originating in the ER are translocated to the cytoplasm thesis on membrane-bound ribosomes, but may also provide a for class I MHC processing, and thus appear to follow the same route for feeding defective glycoproteins back to the cytoplasm pathway recently described for cellular glycoproteins such as ty- where they are deglycosylated by PNGase (24, 25) and degraded rosinase (22, 23). Our results indicate that the classical class I by the proteasome (15). Proteasome inhibition should therefore MHC pathway, which is known for efficient presentation of cyto- result in accumulation of deglycosylated protein in the cytoplasm, plasmic or nuclear Ags, also processes membrane glycoproteins of which has been demonstrated for tyrosinase (23) and class I MHC intracellular pathogens or tumors that might otherwise avoid im- heavy chains (15). Preliminary results indicate that proteasome The Journal of Immunology 675 inhibitors also prevent presentation of the E1 epitope (C. M. 3. York, I., and K. L. Rock. 1996. Antigen processing and presentation by the class Walker, unpublished observations). We postulate that deglycosy- I major histocompatibility complex. Annu. Rev. Immunol. 14:369. 4. Siliciano, R. F., and M. J. Soloski, 1995. MHC class I-restricted processing of lation of E1 is particularly important for immune recognition of transmembrane proteins. J. Immunol. 95:2. HCV-infected cells. The most plausible mechanism is enzymatic 5. Wei, M. L., and P. Cresswell. 1992. HLA-A2 molecules in an antigen-processing hydrolysis of the ␤-aspartylglucosaminyl bond by PNGase; deami- mutant cell contain signal sequence-derived peptides. Nature 356:443. 234 6. Hammond, S. A., R. C. Bollinger, T. W. Tobery, and R. F. Siliciano. 1993. dation of the Asn residue appears to be an absolute requirement Transporter-independent processing of HIV-1 envelope protein for recognition by for peptide presentation by Patr-B*1601. It is conceivable that for CD8ϩ T cells. Nature 364:158. other epitopes, Asn to Asp substitutions cause more subtle alter- 7. Lee, S. P., W. A. Thomas, N. W. Blake, and A. B. Rickinson. 1996. Transporter (TAP)-independent processing of a multiple membrane-spanning protein, the Ep- ations in immune recognition, particularly if these residues are stein-Barr virus latent membrane protein 2. Eur. J. Immunol. 26:1875. important for TCR recognition instead of anchoring to class I 8. Henderson, R. A., A. L. Coss, K. Sakaguchi, E. Appellia, J. Shabanowitz, D. H. Hunt, and V. H. Engelhard. 1993. Direct identification of an endogenous MHC molecules. peptide recognized by multiple HLA-A2.1 specific cytotoxic T cells. Proc. Natl. Morrison and colleagues (46) originally demonstrated that the Acad. Sci. USA 90:10275. hemagglutinin protein of influenza virus must be synthesized in an 9. Hammond, S. A., R. P. Johnson, S. A. Kalams, B. D. Walker, M. Takiguchi, ϩ J. T. Safrit, R. A. Koup, and R. F. Siliciano. 1995. An epitope-selective, trans- infected cell for effective recognition by CD8 CTL. Hemagglu- porter associated with antigen presentation (TAP)-1/2-independent pathway and tinin synthesized in cells without a leader peptide for ER localiza- a more general TAP-1/2-dependent antigen-processing pathway allow recogni- ϩ tion is efficiently presented for CD8ϩ T cell recognition (47). tion of the HIV-1 envelope glycoprotein by CD8 CTL. J. Immunol. 154:6140. 10. Ferris, R. L., C. Buck, S. A. Hammond, A. S. Woods, R. J. Cotter, M. Takiguchi, These studies raised questions about whether glycoproteins pro- Y. Igarashi, Y. Ichikawa, and R. F. Siliciano. 1995. Class I-restricted presentation cessed for class I MHC presentation originate in the cytoplasm or of an HIV-1 gp41 epitope containing an N-linked glycosylation site. J. Immunol. the ER. Although the present study indicates that this can involve 156:834. Downloaded from 11. Leonard, C. K., M. W. Spellman, L. Riddle, R. J. Harris, J. N. Thomas, and retrograde transport of membrane glycoproteins from the ER to the T. J. Gregory. 1990. Assignment of intrachain disulfide bonds and characteriza- cytosol, it is probably not the only pathway in infected cells lead- tion of potential glycosylation sites of the type 1 recombinant human immuno- ing to class I MHC presentation of TAP-dependent epitopes. For deficiency virus envelope glycoprotein (gp120) expression Chinese hamster. J. Biol. Chem. 265:10373. instance, some epitopes of gp120 contain Asn-Xaa-Ser/Thr motifs 12. Boon, T., and A. Van Pel. 1989. T cell-recognized antigenic peptides derived but are neither glycosylated nor deglycosylated before class I from the cellular genome are not protein degradation products but can be generated directly by transcription and translation of short subgenic regions: a hy- MHC presentation (10). This suggests that a small fraction of en- http://www.jimmunol.org/ pothesis. Immunogenetics 29:75. velope proteins are aberrantly synthesized on free cytoplasmic ri- 13. Yewdell, J. W., J. C. Antoń, and J. R. Bennink. 1996. Defective ribosomal prod- bosomes because of a transient shortage of signal recognition par- ucts (DRiPs): a major source of antigenic peptides for MHC class I molecules? ticles that chaperone nascent polypeptides and attached ribosomes J. Immunol. 157:1823. 14. Yewdell, J. W., and J. R. Bennink. 1992. Cell biology of antigen processing and to the ER membrane (10) or because of errors in translation (13). presentation to major histocompatibility complex class I molecule-restricted T Taken together, studies of tyrosinase and transmembrane enve- lymphocytes. Adv. Immunol. 52:1. lopes of HCV and HIV-1 suggest that glycoproteins synthesized in 15. Wiertz, E. J. H., T. R. Jones, L. Sun, M. Bogyo, H. J. Geuze, and H. L. Ploegh. 1996. The human cytomegalovirus US11 gene product dislocates MHC class I the cytoplasm and ER feed into a final common Ub/proteasome/ heavy chains from the endoplasmic reticulum of the cytosol. Cell 84:769. TAP processing pathway. Indeed, recent studies involving a panel 16. Wiertz, E. J. H., D. Tortorella, M. Bogyo, J. Yu, W. Mothes, T.R. Jones,

T. A. Rapoport, and H. L. Ploegh. 1996. Sec61-mediated transfer of a membrane by guest on September 26, 2021 of epitopes from gp120 indicate that both sources of Ag are used protein from the endoplasmic reticulum to the proteasome for destruction. Nature in the same infected cell (R. Siliciano, unpublished observations). 384:432. Finally, it is noteworthy that Patr-B*1601-restricted CTL with 17. Hiller, M. M., A. Finger, M. Schweiger, and D. H. Wolf. 1996. ER degradation D of a misfolded luminal protein by the cytosolic ubiquitin-proteasome pathway. the same fine specificity for peptide 58 were derived from the Science 273:1725. liver during the acute phase of infection and from peripheral blood 18. Werner, E. D., J. L. Brodsky, and A. A. McCracken. 1996. Proteasome-depen- drawn 5 yr later when plasma viremia was still detectable. These dent endoplasmic reticulum-associated protein degradation: an unconventional route to a familiar fate. Proc. Natl. Acad. Sci. USA 93:13797. data provide confirmation in the chimpanzee model that HCV- 19. Hughes, E. A., C. Hammond, and P. Cresswell. 1997. Misfolded major histo- specific CTL circulate in the peripheral blood as previously de- compatibility complex class I heavy chains are translocated into the cytoplasm scribed for chronically infected humans (28–31). Comparison of and degraded by the proteasome. Proc. Natl. Acad. Sci. USA 94:1896. ϩ 20. Yu, H., G. Kahung, S. Kobayashi, and R. R. Kopito. 1997. Cytosolic degradation TCR sequences expressed by CD8 T cell lines established from of T-cell receptor ␣ chains by the proteasome. J. Biol. Chem. 272:20800. liver at week 28 or peripheral blood at week 263 postinfection is 21. Bacik, I., H. L. Snyder, L. C. Anton, G. Russ, W. Chen, J. R. Bennink, L. Urge, required to determine whether they belong to the same clonotype. L. Otvos, B. Dudkowska, L. Eisenlohr, et al. 1997. Introduction of a glycosylation site into a secreted protein provides evidence for an alternative antigen pro- Although data for this single Patr-B*1601-restricted T cell popu- cessing pathway: transport of precursors of major histocompatibility complex lation must be considered preliminary, they might suggest that class I-restricted peptides from the endoplasmic reticulum to the cytosol. J. Exp. clonal deletion caused by Ag-driven exhaustion or interaction with Med. 186:479. 22. Skipper, J. C. A., R. C. Hendrickson, P. H. Gulden, V. Brichard, A. Van Pel, infected hepatocytes that do not express costimulatory molecules Y. Chen, J. Shabanowitz, T. Wolfel, C. L. Slingluff, T. Bonn, et al. 1996. An is not the exclusive mechanism of immune evasion by HCV. HLA-A2-restricted tyrosinase antigen on melanoma cells results from post-trans- lational modification and suggests a novel pathway for processing of membrane proteins. J. Exp. Med. 183:527. Acknowledgments 23. Mosse, C. A., L. Meadows, C. J. Luckey, D. Kittlesen, E. L. Huczko, C. L. Slingluff, Jr., J. Shabanowitz, D. F. Hunt, and V. H. Engelhard. 1998. The We thank Drs. Bob Siliciano and R. Ferris (Johns Hopkins University, class I antigen-processing pathway for the membrane protein tyrosinase involves Baltimore, MD) for helpful discussions, Dr. Barry Rouse (University of translation in the endoplasmic reticulum and processing in the cytosol. J. Exp. Med. 187:37. Tennessee, Knoxville, TN) for providing a recombinant vaccinia virus ex- 24. Suzuki, T., A. Seko, K. Kitajima, Y. Inoue, S. Inoue. 1993. Identification of pressing the herpes simplex virus type 1 ICP47 protein, and Dr. Xavier peptide:N-glycanase activity in mammalian-derived culture cells. Biochem. Bio- Paliard for critical review of the manuscript. Nelle Cronen and Peter Ander- phys. Res. Commun. 194:1124. son provided expert assistance with the preparation of this manuscript. 25. Suzuki, T., A. Seko, K. Kitajima, Y. Knoue, and S. Inoue. 1994. Purification and enzymatic properties of peptide:N-glycanase from C3H mouse-derived L-929 fibroblast cells. J. Biol. Chem. 269:17611. References 26. Koziel, M. J., D. Dudley, J. T. Wong, J. Dienstag, M. Houghton, R. Ralston, and B. D. Walker. 1992. Intrahepatic cytotoxic T lymphocytes specific for hepatitis C 1. Heemels, M.-T., and H. Ploegh. 1995. Generation, translocation, and presentation virus in persons with chronic hepatitis. J. Immunol. 149:3339. of MHC class-I restricted peptides. Annu. Rev. Biochem. 64:463. 27. Koziel, M. J., and B. D. Walker. 1997. Characteristics of the intrahepatic cyto- 2. Elliott, T. 1997. Transporter associated with antigen processing. Adv. Immunol. toxic T lymphocyte response in chronic hepatitis C virus infection. Springer 65:47. Semin. Immunopathol. 19:69. 676 AN ER TO CYTOSOL Ag PROCESSING PATHWAY FOR VIRAL GLYCOPROTEINS

28. Kita, H., T. Moriyama, T. Kaneko, I. Harase, M. Normura, H. Miura, 37. Hill, A., P. Jugovic, I. York, G. Russ, J. Bennink, J. Yewdell, H. Ploegh, and I. Nakamura, Y. Yazaki, and M. Imawari. 1993. HLA B44-restricted cytotoxic T D. Johnson. 1995. Herpes simplex virus turns off the TAP to evade host immu- lymphocytes recognizing an epitope on hepatitis C virus nucleocapsid protein. nity. Nature 375:411. Hepatology 18:1039. 38. Rammensee, H.-G., T. Friede, and S. Stevanovic. 1995. MHC ligands and peptide 29. Shirai, M., H. Okada, M. Nishioka, T. Akatsuka, C. Wychowski, R. Houghten, motifs: first listing. Immunogenetics 41:178. C. D. Pendleton, S. Feinstone, and J. A. Berzofsky. 1994. An epitope in hepatitis 39. Tobery, T. W., and R. F. Siliciano. 1997. Targeting of HIV-1 antigens for rapid C core region recognized by cytotoxic T cells in mice and humans. J. Virol. intracellular degradation enhances cytotoxic T lymphocyte (CTL) recognition 68:3334. and the induction of de novo CTL responses in vivo after immunization. J. Exp. 30. Cerny, A., J. G. McHutchinson, C. Pasquinelli, M. E. Brown, M. A. Brothers, Med. 185:909. B. Grabscheid, P. Fowler, M. Houghton, and F. V. Chisari. 1995. Cytotoxic T 40. Choo, Q.-L., K. H. Richman, J. H. Han, K. Berger, C. Lee, C. Dong, C. Gallegos, lymphocyte response to hepatitis C virus-derived peptides containing the HLA D. Coit, A. Medina-Selby, P. J. Barr, et al. 1991. Genetic organization and di- A2.1 binding motif. J. Clin. Invest. 95:521. versity of the hepatitis C virus. Proc. Natl. Acad. Sci. USA 88:2451. 31. Battegay, M., J. Fikes, A. M. Di Bisceglie, P. A. Wentworth, A. Sette, E. Celis, 41. Ralston, R., K. Thudium, K. Berger, C. Kuo, B. Gervase, J. Hall, M. Selby, W. Ching, A. Grakoui, C. M. Rice, and K. Kurokohchi. 1995. Patient with G. Kuo, M. Houghton, and Q.-L. Choo. 1993. Characterization of hepatitis C chronic hepatitis C virus-encoded peptides binding to HLA-A2.1 molecules. virus envelope glycoprotein complexes expressed by recombinant vaccinia vi- J. Virol. 69:2462. ruses. J. Virol. 67:6753. 32. Erickson, A. L., M. Houghton, Q.-L. Choo, A. J. Weiner, R. Ralston, 42. Spaete, R. R., D. Alexander, M. E. Rugroden, Q.-L. Choo, K. Berger, E. Muchmore, and C. M. Walker. 1993. Hepatitis C virus-specific CTL responses K. Crawford, C. Kuo, S. Leng, C. Lee, R. Ralston, et al. 1992. Characterization in the liver of chimpanzees with acute and chronic hepatitis C. J. Immunol. of the hepatitis C virus E2/NS1 gene product expressed in mammalian cells. 151:4189. Virology 188:819. 33. Kowalski, H., A. L. Erickson, S. Cooper, J. D. Domena, P. Parham, and 43. Matlack, K. E., W. Mothes, and T. A. Rapoport. 1998. Protein translocation: C. M. Walker. 1996. Patr-A and B, the orthologues of HLA-A and B, present tunnel vision. Cell 92:381. hepatitis C virus epitopes to CD8ϩ cytotoxic T cells from two chronically in- 44. Rapoport, T. A., B. Jungnickel, and U. Kutay. 1996. Protein transport across the fected chimpanzees. J. Exp. Med. 183:1761. eukaryotic endoplasmic reticulum and bacterial inner membranes. Annu. Rev. 34. Walker, C. M. 1996. Cytotoxic T-lymphocyte responses to the hepatitis C virus Biochem. 65:271. in humans and chimpanzees. Semin. Virol. 7:13. 45. Hanein, D., K. E. S. Matlack, B. Jungnickel, K. Plath, K.-U. Kalies, K. R. Miller, Downloaded from 35. Banks, T. A., F. J. Jenkins, S. Kanangat, S. Nair, S. Dasgupta, C. M. Foster, and T. A. Rapoport, and C. W. Akey. 1996. Oligomeric rings of the Sec61p complex B. T. Rouse. 1994. Vaccination with the immediate-early protein ICP47 of herpes induced by ligands. Cell 87:721. simplex virus-type 1 (HSV-1) induces virus-specific lymphoproliferation, but 46. Morrison, L. A., A. E. Lukacher, V. L. Braciale, D. P. Fan, and T. J. Braciale. fails to protect against lethal challenge. Virology 200:236. 1986. Differences in antigen presentation to MHC class I- and class II-restricted 36. Weiner, A., A. L. Erickson, J. Kansopon, K. Crawford, E. Muchmore, influenza virus-specific cytolytic T lymphocyte clones. J. Immunol. 163:903. A. L. Hughes, M. Houghton, and C. M. Walker. 1995. Persistent hepatitis C virus 47. Townsend, A. R., J. Bastin, K. Gould, and G. G. Brownlee. 1986. Cytotoxic T infection in a chimpanzee is associated with emergence of a cytotoxic T lym- lymphocytes recognize influenza hemagglutinin that lacks a signal sequence. Na-

phocyte escape variant. Proc. Natl. Acad. Sci. USA 92:2755. ture 324:575. http://www.jimmunol.org/ by guest on September 26, 2021