Soluble Tapasin Restores MHC Class I Expression and Function in the Tapasin-Negative Cell Line .220

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Immunity, Vol. 8, 221–231, February, 1998, Copyright 1998 by Cell Press Soluble Tapasin Restores MHC Class I Expression and Function in the Tapasin-Negative Cell Line .220

Paul J. Lehner,† Michael J. Surman, associate with the ER-resident chaperone calnexin and Peter Cresswell* (Noessner and Parham, 1995). Calnexin is not obligatory Section of Immunobiology for class I assembly; however, as in a calnexin-negative Howard Hughes Medical Institute cell line, peptide loading, assembly, and function of Yale University School of Medicine class I molecules appear to be normal (Sadasivan et al., New Haven, Connecticut 06510 1995; Scott and Dawson, 1995). In human cells, class I

binding to ␤2-microglobulin (␤2m) is thought to cause dissociation of calnexin (Sadasivan et al., 1996). The

Summary class I–␤2m heterodimer then associates with TAP and forms part of a large ER complex, “the TAP complex,” Tapasin forms a bridge between TAP (transporters the minimal components of which include TAP1 and associated with antigen processing) and MHC class I TAP2, MHC class I heavy chains, ␤2m, the soluble ER- molecules and plays a critical role in class I assembly. resident chaperone calreticulin, and the newly identified In its absence, TAP and class I do not associate, and protein tapasin (Sadasivan et al., 1996; Solheim et al., class I cell surface expression is reduced. We now 1997). Although the precise order of events governing identify two independent functions for tapasin. Ta- the assembly of this complex is not yet understood, pasin increases TAP levels and allows more peptide studies on mutant cell lines have been extremely useful to be translocated to the endoplasmic reticulum. Fur- in determining both how the complex fits together and thermore, when expressed in the tapasin-negative which proteins are redundant. In previous studies we .220 cell line, recombinant soluble tapasin retains its showed that in TAP-negative cell lines, MHC class I and association with class I and restores class I cell sur- ␤2m molecules associate with calreticulin and tapasin, face expression and function, even though it no longer while in ␤2m-deficient cells, free class I heavy chains no binds TAP or increases TAP levels. This finding sug- longer associate with TAP or calreticulin, but tapasin gests that the association of tapasin with class I is remains associated with TAP (Sadasivan et al., 1996). sufficient to facilitate loading and assembly of class I These studies suggested that tapasin plays an important molecules. role in bridging class I molecules to TAP. Compelling evidence for the importance of the TAP–class I associa- Introduction tion, and the critical role played by tapasin, was provided by analysis of the mutant .220 cell line, which lacks Recognition of peptide-loaded major histocompatibility tapasin (Greenwood et al., 1994; Grandea et al., 1995; complex (MHC) class I molecules at the cell surface Sadasivan et al., 1996). In this cell line, MHC class I by cytotoxic T lymphocytes (CTL) is the final step in a molecules do not associate with TAP, resulting in ineffi- complex intracellular process that begins when pep- cient peptide loading and impaired surface expression tides generated in the cytosol are translocated across of MHC class I molecules. Transfection of cDNA encod- the endoplasmic reticulum (ER) membrane. The majority ing tapasin corrects the phenotype of .220, restoring of class I–binding peptides are generated by the protea- TAP–class I association, cell surface class I expression, some (Rock et al., 1994), and are transported into the ER and recognition of virus-infected .220 cells by class via the transporter associated with antigen processing I–restricted CTL (Ortmann et al., 1997). (TAP) heterodimer (Howard, 1995). TAP is a member Cloning of the tapasin cDNA revealed the mature pro- of the adenosine triphosphate (ATP)–binding cassette tein to be a type I transmembrane protein of 428 amino family of transporters. TAP1 and TAP2 together form a acids, with a single N-linked glycosylation site and a combinatorial peptide-binding site (Androlewicz et al., probable ER retention signal (Li et al., 1997; Ortmann et 1994; Nijenhuis and Ha¨ mmerling, 1997). The binding of al., 1997). It is a member of the immunoglobulin super- peptide is ATP independent, whereas translocation re- family, and its gene is encoded within the MHC region. quires ATP hydrolysis (Androlewicz et al., 1993; Neefjes The stochiometry of the TAP complex suggests that et al., 1993). Once peptides have gained access to the tapasin and class I are present in a 1:1 ratio, and up to ER, they are loaded onto class I molecules, which are four tapasin–MHC class I molecules are able to associ- then transported through the secretory pathway to the ate with a single TAP1/2 heterodimer (Ortmann et al., cell surface. While the translocation of a peptide and 1997). the binding to its cognate class I molecule allows release While the critical elements of the class I presentation of that class I molecule from the ER (Ortmann et al., pathway have been defined, it remains unclear how ta- 1994; Suh et al., 1994), little is known about how translo- pasin facilitates peptide loading and assembly of MHC cation is linked to peptide loading. class I molecules. Here we show that the association Assembly of MHC class I molecules is initiated in the of tapasin with TAP increases steady-state TAP levels, ER, where nascent, unfolded MHC class I molecules allowing more peptide to be translocated into the ER and to become available for binding to class I molecules. We also show that a mutant tapasin protein lacking the * To whom correspondence should be addressed (e-mail: peter. [email protected]). transmembrane region and cytoplasmic tail no longer † Present address: Department of Medicine, University of Cam- associates detectably with TAP, but is able to facilitate bridge, Addenbrooke’s Hospital, Cambridge CB2 2QQ, England. peptide loading of class I molecules and restore surface Immunity 222

Figure 1. Tapasin Expression Increases TAP Levels and TAP-Mediated Peptide Translo- cation (A) Tapasin expression increases steady- state TAP levels. Control (Ϫ) and tapasin- transfected (ϩ) .220.B8 cells were extracted in 1% Triton X-100, and serial titrations (2-fold) of cell lysates, ranging from 1 ϫ 105 to 6.25 ϫ 103 cell equivalents, were separated by SDS-PAGE (12.5%), transferred to Immobilon-P membranes, and probed with R.RING4C (anti-TAP.1) and AF8 (anti-calnexin) (as an internal loading control) antibodies. Reactive bands were detectedby fluorescence and quantitated on a Fluorimager. (B) TAP photoaffinity labeling with the 125I- KB11-HSAB peptide is increased in tapasin- transfected .220.B8 cells. Photolabeling of TAP from wild-type (wt) tapasin– and control- transfected .220.B8 cells with the 125I-KB11- HSAB peptide was performed as described in Experimental Procedures. After solubilization in 1% Triton X-100, labeled TAP molecules were immunoprecipitated with R.RING4C (anti- TAP.1) antiserum, followed by protein A-Seph- arose beads, and counted on a ␥-counter. (C–E) TAP-mediated peptide translocation is increased in tapasin-expressing .220.B8 cells. SLO-permeabilized cells (C) or microsomal membrane preparations (D and E) from tapasin and control .220.B8 transfectants were incubated with the iodinated reporter peptides QVPLRNMTYK (C), RRYQNSTEL (D), and LLDGNATLRY (E) at 37ЊC for the indicated time periods. Translocation into the ER was assessed by binding of the glycosylated reporter peptide to concanavalin A–Sepha- rose beads and counting on a ␥-counter. expression of MHC class I. These results suggest that with ultraviolet light, and following detergent extraction, in the .220 cell line transfected with human leukocyte immunoprecipitated with a TAP1-specific antiserum and antigen B8 (HLA-B8) (.220.B8), it is the association of counted on a ␥-counter. As shown in Figure 1B, peptide tapasin with class I molecules that is critical for efficient binding to TAP was increased in the presence of tapasin. MHC class I assembly. At saturating concentrations of peptide there was an approximately 3-fold increase in peptide binding in the Results tapasin-positive .220.B8 as compared with the control transfectant. This is consistent with the observed 3-fold Tapasin Increases TAP Expression and Peptide increase in TAP levels seen with tapasin expression. Translocation into the ER TAP-mediated peptide translocation can be assessed To determine how tapasin–class I association with TAP in streptolysin-O (SLO)–permeabilized cells by introduc- augments class I assembly, we initially compared steady- tion of an iodinated reporter peptide containing an state TAP1 and TAP2 levels in the absence (.220.B8) N-linked glycosylation motif into the cytosol (Neefjes and presence (.220.B8 ϩ wild-type tapasin) of tapasin. et al., 1993). In the presence of a functional TAP1/2 Tapasin expression results in a 2.5- to 3-fold increase heterodimer, the peptide is translocated into the ER, in steady-state TAP1 (Figure 1A) and TAP2 levels (not where it is glycosylated and can be recovered by binding shown), as determined by immunoblot analysis and flu- to concanavalin A–Sepharose beads. Peptide translo- orimager quantitation. The increased TAP levels were cation rates in control and tapasin-transfected .220.B8 reproducible and seen in all tapasin-expressing .220.B8 cell lines were compared using three different reporter clones. To examine whether the observed increase in peptides. Both the absolute level and the rate of TAP- TAP levels affects TAP function, we compared peptide mediated peptide translocation were higher in the ta- binding and TAP-mediated peptide translocation in con- pasin-positive transfectant than in .220.B8 cells (Figures trol and tapasin-transfected cell lines. Peptide binding 1C–1E). This increase in peptide translocation rate was was determined using an iodinated, photoactivatable evident in assays performed with both SLO-permeabilized peptide to photolabel TAP molecules on membranes cells (Figure 1C) and microsomal membrane prepara- prepared from tapasin-expressing and control .220.B8 tions (Figures 1D and 1E). Therefore the tapasin-induced cells (Androlewicz et al., 1994). After incubation with increase in TAP levels results in enhanced transport of peptide for 30 min at 4ЊC, membranes were illuminated peptides into the ER for binding to class I molecules. Soluble Tapasin Restores Class I Expression 223

We have previously shown that in the ␤2m-deficient cell line Daudi, tapasin associates with TAP in the absence of MHC class I molecules (Sadasivan et al., 1996). In

Daudi and ␤2m-transfected Daudi cells, TAP levels and translocation rates were identical (data not shown), suggesting that the increased TAP level iscaused by tapasin alone and does not require tapasin in association with MHC class I.

Soluble Tapasin Does Not Associate with TAP and Is Secreted To analyze tapasin function further, we used an approach based on polymerase chain reaction (PCR) tech- niques to generatea cDNA encoding a truncated tapasin protein from which the 35 C-terminal amino acids of the proposed transmembrane region and cytosolic tail were absent. Expression of this truncated tapasin in .220.B8 cells revealed a 46 kDa protein by Western blot analysis with an N-terminal–specific tapasin antiserum (Figure 2A). Furthermore, this truncated tapasin is soluble and secreted. Glycoproteins were isolated with Lens culinaris lectin beads from the tissue culture supernatant of soluble and wild-type tapasin .220.B8 transfectants. Elution and subsequent immunoblot analysis revealed a 46 kDa band from transfectants expressing soluble tapasin, while no band was seen in transfected cells expressing wild-type tapasin (Figure 2C). The secreted tapasin acquired resistance to endoglycosidase H (endo H) digestion compared with a control glycoprotein, IP30, which although secreted into the tissue culture medium Figure 2. Soluble Tapasin Acquires Resistance to Endo H Digestion remains endo H sensitive as a result of mannose-6- and Is Secreted phosphate modification of its N-linked glycans (Aruna- A truncated tapasin protein from which the 35 C-terminal amino chalam and Cresswell, unpublished data). acids were absent was transfected into .220.B8 cells. Control and We were unable to detect the association of soluble tapasin-transfected .220.B8 cells were extracted in 1% Triton X-100, tapasin with either MHC molecules or TAP under condi- and lysates were separated by 12.5% SDS-PAGE, transferred to tions in which their interaction with wild-type tapasin Immobilon-P membranes, and probed with R.gp48N (anti-tapasin) was readily detectable (Sadasivan et al., 1996). To as- (A) and AF8 (anti-calnexin) (as a loading control) (B) antibodies. Glycoproteins from the tissue culture media of soluble (s) and wild- certain whether the truncated tapasin associated with type (wt) tapasin–transfected .220.B8 cells were isolated on L. culi- MHC molecules, ␤2m-associated proteins were immu- naris lectin beads. Eluted proteins were subjected to mock (Ϫ)or noprecipitated from metabolically labeled control, solu- endo H (ϩ) digestion, separated by 12.5% SDS-PAGE, and after ble tapasin–, and wild-type tapasin–expressing .220.B8 transfer to Immobilon-P membranes probed with R.gp48N (anti- cells extracted in digitonin containing the reducible tapasin) and anti-IP30 antiserum (C). The 35 kDa IP30 glycoprotein cross-linker 3,3Ј-dithio-bis(propionic acid N-hydroxy- remains endo H sensitive and serves as a positive control for endo succinimide ester) (DSP). Mature, assembled class I H enzymatic activity. molecules were first removed by immunoprecipitation with the monoclonal antibody (mAb) w6/32, which fails TAP. TAP immunoprecipitations of metabolically labeled to react with TAP- or tapasin-associated MHC class I digitonin extracts from tapasin transfectants were ana- molecules (Carreno et al., 1995; Neisig et al., 1996). This lyzed by two-dimensional polyacrylamide gel electro- reduced the background upon subsequent reprecipita- phoresis (PAGE) (Figures 3B and 3C). Even in the pres- tion of ␤2m-associated proteins following their release ence of cross-linking agents (data not shown) and by sodium dodecyl sulfate (SDS) under reducing condi- prolonged film exposure, neither tapasin, MHC class I, tions. Reprecipitation with the N-terminal specific ta- nor ␤2m could be immunoprecipitated with TAP-specific pasin antiserum readily identified both the soluble and antisera from .220.B8 cells expressing soluble tapasin. wild-type tapasin in the anti-␤2m immunoprecipitates As an additional approach to this question, we examined (Figure 3A). Because these experiments used an anti- TAP expression and function in soluble tapasin .220.B8 body to ␤2m to isolate the cross-linked complexes it is transfectants. Levels of TAP1 (Figure 3D) and TAP2 (data formally possible that free ␤2m is also associated with not shown) in control and .220.B8 cells expressing solu- tapasin, as suggested by Solheim et al. (1997). To date, ble tapasin were identical. Thus, TAP expression is not we have been unable to find a class I–specific mAb increased as in the wild-type tapasin transfectants. Sim- that reacts sufficiently well with the class I–tapasin–TAP ilarly, the TAP-mediated peptide translocation rate was complex to allow us to address this question further. not increased in the soluble tapasin .220.B8 transfec- In contrast to the results with class I molecules, solu- tant, as it was in .220.B8 cells expressing wild-type ta- ble tapasin does not appear to remain associated with pasin (Figure 3E). Overall, these results suggest that Immunity 224

Figure 3. Soluble Tapasin Associates with Class I Molecules but Not with TAP (A) Soluble tapasin associates with MHC class I molecules. Control, soluble (s) tapasin–, and wild-type (wt) tapasin– transfected .220.B8 cells were labeled with [35S]methionine and [35S]cysteine for 2 hr, and after cross-linking with 200 ␮M DSP, 1% digitonin extracts were precleared three times with w6/32 mAb to remove mature class I molecules and

immunoprecipitated (1Њ) with anti-␤2m antiserum. ␤2m associated proteins were eluted in SDS and dithiothreitol and reprecipitated (2Њ) with R.gp48N (anti-tapasin) or HC10 (free class I heavy chain) mAb and separated by 10% SDS-PAGE. The autoradiograph exposure time showing tapasin immunoprecipitation was three times longer than the class I immunoprecipitation. (B and C) Soluble tapasin does not associate with TAP. Wild- type tapasin– (B) and soluble tapasin– (C) transfected .220- B8 cells were labeled with [35S]methionine and [35S]cysteine for 20 min and 1% digitonin extracts immunoprecipitated with R.RING.4C (anti-TAP.1) antiserum and separated by two-dimensional PAGE. Both gels were deliberately over- exposed for identical times to maximize the possibility of tapasin and class I detection. The arrowhead at left indicates class I heavy chain; the arrowhead at right indicates tapasin;

and the arrowhead at bottom represents ␤2m. (D) Soluble tapasin does not affect TAP expression levels. Control, soluble tapasin–, and wild-type tapasin–expressing .220.B8 cells were extracted in 1% Triton X-100, and serial dilutions (3-fold) of cell equivalents (5 ϫ 104–5.6 ϫ 103) were separated by 12.5% SDS-PAGE, transferred to Immobilon-P membranes, and probed with R.RING4C (anti-TAP.1) and AF8 (anti-calnexin) (as a loading control) antibodies. Reac- tive bands were detected by enhanced chemiluminescence. (E) Soluble tapasin does not alter TAP-mediated peptide translocation. SLO-permeabilized cells from tapasin and control .220.B8 transfectants were incubated with the iodinated reporter peptide (QVPLRNMTYK) at 37ЊC for the indicated time periods. Translocation into the ER was assessed by binding of the glycosylated reporter peptide to concanavalin A–Sepharose beads and counting on a ␥-counter.

soluble tapasin associates with MHC class I molecules and wild-type tapasin transfectants attain endo H resis- but no longer associates with TAP and therefore has no tance. However, in soluble tapasin transfectants, MHC effect on TAP expression or peptide translocation. class I molecules mature twice as fast, taking 20 min for half-maximal acquisition of endo H resistance, as opposed to 40 min in wild-type tapasin transfectants Soluble Tapasin Rescues MHC Class I Cell (Figure 4E). Therefore, in the absence of TAP associa- Surface Expression tion, class I molecules from cells transfected with solu- To examine how the lack of TAP association by soluble ble tapasin cDNA mature faster than in wild-type cells tapasin affected MHC class I expression, cell surface and show comparable levels of MHC class I cell surface MHC class I was analyzed by flow cytometry with the expression. conformationally sensitive w6/32 antibody. Surprisingly, soluble tapasin restored class I expression to wild-type levels (Figures 4A and 4B). We next performed a pulse Surface MHC Class I Molecules Are Peptide Loaded chase analysis and compared class I maturation in in Cells Expressing Soluble Tapasin .220.B8 cells expressing soluble and wild-type tapasin. Transfection of soluble tapasin into .220.B8 cells re- Previous studies have shown that in the absence of stored expression of MHC class I molecules at the cell tapasin the majority of class I molecules do not mature surface. It was critical to determine whether the surface through the secretory pathway (Grandea et al., 1995; class I molecules in .220.B8 cells expressing soluble Grandea et al., 1997). As shown (Figures 4C–4E), a simi- tapasin were empty or peptide loaded. Class I molecules lar proportion of class I molecules from both soluble were therefore affinity purified fromradiolabeled control, Soluble Tapasin Restores Class I Expression 225

Figure 4. Soluble Tapasin Restores Cell Surface MHC Class I Expression (A and B) MHC class I molecule expression is rescued in soluble tapasin transfectants. Wild- type tapasin– (A) and soluble tapasin– (B) transfected .220.B8 cells were analyzed for class I surface expression by flow cytometry, using the mAb w6/32 and fluorescein-conjugated rabbit anti-mouse IgG. The control was a nonspecific isotype-matched mAb. (C–E) Class I heavy chains mature faster in soluble than in wild-type tapasin–transfected .220.B8 cells. Soluble (C) and wild-type (D) tapasin– transfected .220.B8 cells were radiolabeled with [35S]methionine and [35S]cysteine for 5 min and chased for the indicated time periods. 1% Triton X-100 lysates were immunoprecipitated with 3B10.7 (total class I) and w6/32 (conformational class I) mAb, digested (ϩ) or mock-digested (Ϫ) with endo H, and subjected to 10% SDS-PAGE. The ratio of endo H–resistant versus endo H–sensitive class I heavy chains present at each time point was quantitated by image analysis (E).

soluble tapasin, and wild-type tapasin .220.B8 transfec- et al., 1997; Lehner et al., 1997). Expression of US6 in tants, and peptides were isolated and characterized by soluble tapasin .220.B8 cells (Figure 6A) resulted in a reverse-phase (RP) high-pressure liquid chromatography dramatic reduction in class I expression (Figure 6B), (HPLC) (Figure 5). As expected, relatively few peptides indicating an absolute requirement for peptide in the were isolated from control .220.B8 cells. In contrast, assembly and transport of MHC class I molecules. Solu- peptide profiles from soluble tapasin– and wild-type ta- ble tapasin secretion is unaffected by US6 expression pasin–expressing cell lines show multiple peaks. An (Figure 6C) and thus appears to be independent of MHC overlay of HPLC profiles from these two cell lines indi- class I transport. Furthermore, no increase in free class cates that the dominant peaks lie at similar positions. I heavy chain was seen at the cell surface of these cells, Although the peptide profiles are a relatively crude rep- as assessed by heavy chain-specific mAbs (data not resentation of the complexity of class I bound peptides, shown). they indicate that cell surface MHC class I molecules are peptide loaded in cells expressing both soluble as well as wild-type tapasin. Soluble Tapasin Restores MHC Class An alternative approach to determine whether peptide I–Restricted CTL Recognition is required for class I assembly in soluble tapasin To assess the function of cells expressing soluble ta- transfectants is to block TAP-mediated peptide translo- pasin, we infected .220.B8 transfectants with influenza cation. The human cytomegalovirus (HCMV) US6 gene A virus and examined their susceptibility to lysis by HLA- product is a 23 kDa luminal glycoprotein that binds the B8–restricted, influenza A–specific CTL. Virus-infected TAP complex and inhibits peptide translocation in a .220.B8 cells are poorly susceptible to CTL lysis (Ort- tapasin-independent manner (Ahn et al., 1997; Hengel mann et al., 1997). However, .220.B8 cells expressing Immunity 226

Figure 5. Soluble and Wild-Type Tapasin .220.B8 Transfectants Show Similar MHC Class I Peptide Profiles .220.B8 cells and transfectants were labeled with 1 mCi each of L-[3,4,5-3H]leucine and L-[4, 5-3H]lysine for 5 hr. MHC class I molecules were isolated using a w6/32 affinity column and the associated peptides were separated by RP-HPLC.

soluble tapasin were lysed as efficiently as cells ex- class I molecules at the cell surface of soluble tapasin pressing wild-type tapasin (Figure 7). Soluble tapasin transfectants were correctly assembled and loaded with transfectants were therefore able to present endoge- peptide. The class I molecules are ␤2m associated, since nous antigen as efficiently as their wild-type counter- they were detected with a conformationally sensitive parts. antibody, and we found no increase in free class I heavy chains at the cell surface using the heavy chain specific Discussion mAb HC10 (Stam et al., 1986). The HPLC profiles of the class I–associated peptides in cells expressing soluble In this study we have identified two independent func- and wild-type tapasin were remarkably similar (Figure tions for tapasin. First, association of tapasin with TAP 5). An absolute peptide requirement for class I transport increases TAP levels, allowing more peptides to gain was demonstrated by blocking TAP-mediated peptide access to the ER for class I binding. Second, expression translocation in soluble tapasin transfectants with the of soluble tapasin, which associates with class I but not HCMV-derived US6 protein. Expression of US6 caused with TAP, restores cell surface class I to wild-type levels. a decrease in surface MHC class I expression, as de- This implies that tapasin has a fundamental TAP-inde- tected by both conformationally sensitive and free class pendent effect for promoting the assembly of MHC class I heavy chain–specific antibodies (Figure 6B). This dem- I molecules. onstrates that it is not simply association with soluble Following expression of truncated soluble tapasin in tapasin that allows class I transport to the cell surface. .220.B8 cells, formation of the TAP complex could no Furthermore, despite the impaired assembly of class I longer be visualized, nor was there functional evidence in .220.B8 cells expressing soluble tapasin and US6, for the TAP–tapasin interaction (Figure 3). This suggests soluble tapasin was still secreted. Finally, soluble ta- that the interaction of tapasin with TAP occurs primarily pasin restored the ability of .220.B8 cells to present viral through residues in either the transmembrane region or antigen to influenza A–specific CTL, showing that these cytoplasmic tail. It was therefore unexpected to find that cells are functional (Figure 7). Taken together, these soluble tapasin could restore cell surface expression of results suggest that class I molecules in soluble tapasin MHC class I molecules to wild-type levels (Figure 4). transfectants assemble and are loaded normally with Several lines of evidence suggested that the majority of peptides in the ER. We do not know whether the set of Soluble Tapasin Restores Class I Expression 227

Figure 6. Expression of the US6 Viral Glyco- protein, an Inhibitor of TAP-Mediated Peptide Translocation, Decreases Surface MHC Class I Expression in Soluble Tapasin Transfectants US6 and control soluble tapasin–transfected .220.B8 cells were extracted in 1% Triton X-100 and lysates separated by 12.5% SDS- PAGE,transferred toImmobilon-Pmembranes, and probed with the anti-US6 antibody (A). US6 and control soluble tapasin–transfected .220.B8 cells wereanalyzed for class I surface expression by flow cytometry, using the mAb w6/32 and fluorescein-conjugated rabbit anti-mouse IgG. The control was a nonspecific isotype-matched mAb (B). Glycoproteins from the tissue culture medium of US6 and control soluble tapasin–transfected .220.B8 cells were isolated on L. culinaris lectin beads; eluted proteins were separated by 12.5% SDS-PAGE; and after transfer to Im- mobilon-P membranes probed with R.gp48N (anti-tapasin) antiserum (C).

peptides presented in cells expressing soluble and wild- versus wild-type tapasin (Figure 4) suggests that soluble type tapasin are absolutely identical. The faster rate of tapasin may allow assembled class I molecules to es- transport of class I molecules in cells expressing soluble cape rapidlyfrom the ER, perhaps limiting their exposure to a certain set of peptides or a postulated ER editing mechanism (see below). The association of TAP with class I may allow some peptides to be loaded whereas otherwise they would be degraded in the ER. However, there appears to be an absolute requirement for peptide loading for class I expression in these cells. The data argue that class I molecules do not require association with TAP to be loaded with peptides efficiently. The requirement for TAP–class I association in the efficient assembly of class I molecules remains con- troversial. Neisig et al. (1996) have identified class I alleles that, despite binding TAP-translocated peptides, associate poorly or not at all with TAP and are expressed normally at the cell surface. Furthermore, membrane- Figure 7. SolubleTapasin Expression Restores CTL Lysis of .220.B8 bound HLA-G associates with TAP, while soluble HLA-G Cells appears not to do so but binds a qualitatively similar Peripheral blood mononuclear cells from an HLA-B8–positive donor set of peptides (Lee et al., 1995). In contrast, two reports were incubated in vitro with influenza A–infected cells and pulsed have demonstrated that mutant HLA-A*0201 molecules on day 8 with N380-88 peptide and feeder cells. On day 12 of bulk that do not associate with TAP (T134K) are poorly ex- culture, CTL lysis was tested in a 51Cr release assay against control, soluble tapasin–, and wild-type tapasin–transfected .220-B8 cells pressed at the cell surface and are unable to present that were infected with influenza A virus. Uninfected .220.B8 endogenous antigen to CTL (Lewis et al., 1996; Peace- transfectants expressing both soluble and wild-type tapasin gave Brewer et al., 1996). This, together with the evidence 51Cr release identical to that of the uninfected .220.B8 cells. from the .220 cell line, suggests that TAP association is Immunity 228

critical for class I presentation. Our data indicate that et al., 1995; Hill et al., 1995; Ahn et al., 1997; Hengel et class I molecules can bind peptide in the ER in the al., 1997; Lehner et al., 1997) decreases class I surface absence of TAP association when they interact with expression, while increased TAP levels can up-regulate soluble tapasin. In normal circumstances the associa- it (Russ et al., 1995). In the .220 cell line, the folding tion of class I with TAP presumably reflects an interac- effect of tapasin on class I molecules is clearly more tion involving all three components. important than its effect on TAP. Does this imply that the There are a number of possible explanations for the tapasin–TAP interaction is redundant? Since the mutant differential association of class I alleles with TAP. First, .220 cell line is Epstein-Barr virus (EBV) transformed, class I alleles that appear to associate poorly with TAP we believe that this result may be misleading. EBV- may in fact be less dependent on tapasin for their fold- transformed cells express high levels of TAP1 and TAP2 ing. Such a differential requirement could also be re- compared to nontransformed cells (Rowe et al., 1995) sponsible for the allelic variations in class I expression and may translocate sufficient peptide into the ER for in .220 cells (Greenwood et al., 1994; Grandea et al., class I loading, irrespective of tapasin expression. In 1997). Second, differential association with TAP may this situation, even in the absence of tapasin, peptide reflect the kinetics of class I assembly. Our studies on may not be limiting and the role of the tapasin–class I mutant cell lines have shown that class I can interact interaction therefore may be dominant. In a resting cell, with tapasin in the absence of TAP (Sadasivan et al., lower TAP levels may cause peptide to be a limiting 1996), although we do not know whether this event oc- factor. Under these circumstances a tapasin-induced curs physiologically. If the association of class I with increase in TAP expression and TAP-mediated peptide tapasin can precede TAP binding, then the likelihood of translocation might be as important as the tapasin–MHC the preformed class I–tapasin complex binding TAP may class I interaction. depend on peptide availability. A class I allele that has How tapasin increases TAP levels is not clear. Little a high concentration of suitable peptides available in is known about the assembly or degradation of the TAP the ER may bind TAP transiently, or not at all, before heterodimer. It exhibits slow turnover and a long half- binding peptide that promotes dissociation from tapasin life (data not shown), making kinetic analysis difficult. and transport to the cell surface. Conversely, a class We have not seen an effect of tapasin on TAP synthesis I–tapasin complex that does not immediately bind pep- and favor the idea that tapasin either promotes assem- tide may readily be found associated with TAP. Finally, bly or prevents degradation of TAP. This could occur different class I alleles may affect the affinity of tapasin in a number of ways. Tapasin may be a chaperone, for TAP, perhaps by modifying its conformation. These facilitating assembly of TAP1 and TAP2 heterodimers, possibilities remain to be investigated. which are more stable than the individual subunits. Alter- How soluble tapasin promotes class I assembly inde- natively, tapasin may prevent degradation of either the pendently of its association with TAP remains specula- folded heterodimer or the single-chain subunits, in- tive. In the absence of tapasin, class I molecules are creasing the likelihood of their correct assembly. These unstable in the ER (Grandea et al., 1997). By prolonging possibilities are under investigation. the half-life of assembled MHC class I molecules and preventing class I degradation, tapasin may simply in- Experimental Procedures crease the probability of peptide binding. It may also have an additional class I–specific “foldase” activity. Cell Lines and Viruses The .220.B8 cell line (Greenwood et al., 1994), and its stable tapasin- Alternatively, tapasin may have a specific peptide-edit- transfected derivatives were maintained in RPMI-1640 (GIBCO-BRL, ing function for class I molecules, as was recently identi- Gaithersburg, MD) with 10% fetal calf serum (GIBCO-BRL). Influenza fied for HLA-DM in peptide loading of MHC class II A (PR8) virus was a kind gift from Jonathan Yewdell. molecules (Kropshofer et al., 1996; Weber et al., 1996). This would ensure that optimal peptides, with both fast Antibodies “on-rates” and slow dissociation rates, would ultimately The following antibodies were used: 148.3, an anti-TAP.1 mAb bind class I molecules. Certain class I–peptide com- (Meyer et al., 1994); R.RING4C, a rabbit anti-peptide antibody to the C-terminal region of TAP.1 (Ortmann et al., 1994); 435.3, an anti- plexes retained within the ER/cis-Golgi have a shorter TAP.2 mAb (van Endert et al., 1994); w6/32, a ␤2m-dependent mono- half-life than peptides at the cell surface, suggesting morphic mouse anti-human class I murine mAb (Parham et al., 1979); that an editing mechanism may exist within the ER to 3B10.7, a conformation-independent rat anti-human class I mAb optimize the spectrum of class I–bound peptides (Sijts (Lutz and Cresswell, 1987); HC10, a murine mAb recognizing free and Pamer, 1997). Other possibilities are that tapasin class I heavy chains (Stam et al., 1986); R.gp48N, a rabbit anti- may protect the class I peptide-binding groove prior to peptide antibody to the N-terminal region of tapasin (Sadasivan et al., 1996); AF8, an anti-calnexin mAb (Hochstenbach et al., 1992); peptide loading or that it can itself bind peptides (Li et R.IP35i, a rabbit antiserum against an internal peptide of IP30 (B. al., 1997). Arunachalam and P. C., submitted); and R.US6C, a rabbit anti-pep- The primary function of TAP is to translocate newly tide antibody to the C-terminal region of US6 (Lehner et al., 1997). formed peptides from the cytosol into the ER. Tapasin increases TAP levels and allows more peptides to reach Peptides the ER for binding to class I molecules. Could peptide The influenza A nucleoprotein–derived, HLA-B8–binding peptide supply to the ER be limiting, such that the tapasin- N380–88 was used for the CTL experiments (Sutton et al., 1993). The reporter peptides used for the TAP translocation data were induced increase in TAP, and therefore peptide translo- a variant of an HLA-B27–binding peptide (Jardetzky et al., 1991) cation, regulate cell surface class I expression? Peptide (RRYQNSTEL); a variant of an HLA-A3–binding peptide (QVPLRNM supply to the ER can regulate cell surface expression TYK); and a variant of an HLA-A2–binding peptide (LLDGNATLRY) of class I. Virally induced inhibition of TAP function (Fru¨ h (Chang et al., 1996). The photopeptide analog was synthesized by Soluble Tapasin Restores Class I Expression 229

conjugation of the KB11 peptide (AKVPLRPMTYKA) with the photo- Flow Cytometric Analysis activatable cross-linker HSAB (Pierce) as described (Androlewicz Surface expression of MHC class I complexes on .220.B8 and their and Cresswell, 1994) and purified by HPLC.The peptides were iodin- tapasin transfectants was detected by flow cytometry, as previously ated to a specific activity of 75–90 cpm/fmol, as described (Andro- described (Lehner et al., 1995). lewicz et al., 1993). All peptides, other than LLDGNATLRY, were synthesized by the Keck Foundation Biotechnology Resource Labo- Immunoprecipitation and Endo H Digestion ratory (Yale University). For pulse-chase analysis cells were extracted for 30 min on ice at 2 ϫ 106 cells/ml in 1% digitonin (WAKO) or 1% Triton X-100 TBS Membrane Preparation and TAP Photoaffinity Labeling containing 0.5 mM PMSF, 0.1 mM N-␣-tosyl-L-lysyl-chloromethyl Membranes were prepared by freeze–thaw lysis of 108 cells in 2 ml ketone (Sigma), and 5.0 mM IAA. The postnuclear supernatant was buffer containing 10 mM Tris, 150 mM NaCl (Tris-buffered saline cleared overnight at 4ЊC with 5 ␮l normal rabbit serum and 50 ␮l [TBS]) (pH 7.5); 0.5 mM phenylmethylsulfonylfluoride (PMSF) (Sigma, Zysorbin (formalin-fixed) Staphylococcus aureus (Zymed Labora- St. Louis, MO); and 5.0 mM iodoacetamide (IAA) (Sigma). The cells tories, San Francisco, CA) per milliliter of extract. Aliquots were were pelleted for 5 min at 2500 ϫ g, the supernatant stored on ice, then incubated for 1 hr at 4ЊC with a mixture of 3B10.7 and w6/32 and the pellet washed in 1.5 ml hypotonic buffer containing 10 mM antibodies, followed by 30 min at 4ЊC with protein A-Sepharose. Tris (pH 7.5). After pelleting the remaining celldebris, the postnuclear The immunoprecipitates were washed and treated with endo H supernatants were pooled and centrifuged at 100,000 ϫ g for 1 hr (Boehringer Mannheim, Indianapolis, IN) or under control conditions at 4ЊC. The resulting membrane pellet was resuspended at 5 ϫ 107 as previously described (Androlewicz et al., 1993) before analysis cell equivalents/ml in intracellular transport buffer (50 mM HEPES by SDS-PAGE and autoradiography. Bands were quantitated with a Bio-Rad GS-250 Molecular Imager using the program Molecular [pH 7.0], 78 mM KCl, 4 mM MgCl2, 8.37 mM CaCl2, 10 mM EGTA, 1 mM dithiothreitol), and 2.5 ϫ 106 cell equivalents (50 ␮l) were incu- Analyst, Version 2.0.1 (Bio-Rad Laboratories). For detection of solu- 6 bated with peptide (50 ␮l, peptide concentration 0.4–20 ␮M) for 30 ble tapasin associated with MHC class I molecules, 5 ϫ 10 cells min on ice in a flat-bottomed 96-well tissue culture plate. Mem- were labeled for 90 min and solubilized in extraction buffer con- branes were exposed for 2.5 min to a shortwave ultraviolet lamp taining 1% digitonin and 200 ␮M DSP (Sigma). Assembled MHC (UVP, 254 nm) and immediately solubilized in a final volume of 200 class I molecules were removed by three successive immunoprecip- ␮l TBS containing 1% Triton X-100 (Sigma), 0.5 mM PMSF, and itations with w6/32 and protein A-Sepharose. The remaining ␤2m- 5.0 mM IAA. TAP proteins were immunoprecipitated with affinity- associated proteins were immunoprecipitated with ␤2m-specific an- purified R.RING4C and protein A-Sepharose beads. After immuno- tiserum andprotein A-Sepharose beads, washed three times in 0.1% precipitation the beads were washed three times in TBS containing digitonin in TBS, and heated for 5 min at 95ЊCin50␮l TBS containing 0.4% Triton X-100 and counted on a ␥-counter (LKB Compu- 0.2% SDS and 2 mM dithothreitol. After elution from the beads, gamma CS). dissociated proteins were diluted 30-fold in 1% Triton X-100 in TBS with 10 mM IAA; reprecipitated with either tapasin antiserum (Rgp48N) or class I heavy chain–specific antibody (3B10.7) and pro- Peptide Translocation Assay tein A-Sepharose beads; washed; eluted in sample buffer; and ana- Peptide translocation was carried out as previously described on lyzed by SDS-PAGE and autoradiography. Two-dimensional PAGE SLO (Murex, Norcross, GA)–permeabilized cells (Androlewicz et al., was performed as described (Sadasivan et al., 1996). 1993) or on microsomal membranes. The microsomes were generated as described above and resuspended incold intracellular trans- Immunoblot Analysis and Quantitation port buffer at a concentration of 2.5 ϫ 107 cell equivalents/ml. For Blots were performed as described (Sadasivan et al., 1996), and each time point, the translocation reaction was started by the addi- reactive bands were detected by chemiluminescence (Pierce, Rock- tion of 1 ϫ 107 cell equivalents (400 ␮l) containing 2 mM ATP to ford, IL). For quantitation of TAP1 and TAP2 levels on immunoblots, microfuge tubes containing 6 ϫ 10Ϫ7 M radiolabeled peptide, in a gels wereloaded with 2-fold decreasing dilutions of total celllysates, 37ЊC waterbath. The translocation reaction was stopped after the titrating from 1 ϫ 105–6.25 ϫ 103 cell equivalents. To normalize indicated time interval by the addition of 200 ␮l of cold (4ЊC) TBS for cell equivalents and gel loading, calnexin and TAP levels were containing a final concentration of 1% Triton X-100, 5 mM IAA, and quantitated on the same gel. Following electrophoretic transfer to 0.5 mM PMSF. The glycosylated peptide was isolated from the membranes and incubation with primary antibodies, bands were postnuclear extract (12,000 ϫ g, 5 min), using concanavalin visualized with goat anti-rabbit or goat anti-mouse alkaline phos- A–Sepharose beads (30 ␮l, 50% suspension in TBS; Sigma). The phatase–conjugated secondary reagent (Jackson Immunoresearch pellets were washed three times in TBS with 0.5% Triton X-100 and Lab). Reactive bands were detected by fluorescence (Vistra, Amer- counted in a ␥-counter. sham, IL), and area quantitation was performed on the relevant bands with a Molecular Dynamics 595 Fluorimager using Im- Cloning and Expression of Soluble Tapasin cDNA ageQuant software version 1.11. After it had been established that The truncated tapasin cDNA was amplified by PCR from the tapasin calnexin bands were detectable in the linear range, TAP levels were cDNA clone with the Pfu DNA polymerase (Stratagene) under stan- quantitated and compared between the .220.B8 transfectants. dard PCR conditions. The forward primer was 5Ј-GGGGTACCGCCG CCATGAAGTCCCTGTCTCTGCTCCT-3Ј, and the reverseprimer was MHC Class I Purification, Peptide Isolation, and Analysis 5Ј-CCATCGATTCAGTCCTCAAGGGAGGGCCCTG-3Ј. The 1258 bp of Class I–Associated Peptides amplification product was cloned into the polylinker of the pCR2.1 MHC class I molecules were purified using a w6/32 affinity column cloning vector (Invitrogen, Carlsbad, CA); the clone was sequenced from .220.B8 control and transfected cells labeled with [3H]leucine in both directions and then cloned into the pMCFR-PAC vector (a and [3H]lysine and peptides isolated as previously described (Wei kind gift from Tom Novak) using the 5Ј KpnI and 3Ј ClaI sites. and Cresswell, 1992; Hughes et al., 1996). Peptides were resolved on a reverse-phase column (Delta-Pak C18 3.9 ϫ 150 mm) using a Hewlett-Packard 1100 HPLC system. Gradients were generated us- Generation of Stable Transfectant Cell Lines ing an increasing concentration of acetonitrile in 0.1% hydrochloric Stable .220.B8 soluble tapasin clones were generatedby electropor- acid. ation of pMCFR-puro.soluble tapasin as described (Denzin et al., 1994; Ortmann et al., 1997) and selection in medium containing 500 ng/ml puromycin (Sigma). Stable .220.B8 ϩ soluble tapasin ϩ US6 Influenza Virus (PR8) and Peptide-Specific Cytotoxicity clones were generated by cotransfection of stable .220.B8 ϩ soluble Virus- and peptide-specific CTL lines were generated as described tapasin clones with 1 ␮g pREP8⌬ (Arnold et al., 1992) and 10 ␮g (Lehner et al., 1995). Control, soluble tapasin–, and wild-type ta- pMCFR-NEO.US6 at 210 V, 960 ␮F, and selection in medium con- pasin–transfected .220.B8 target cells were labeled and used in taining 5 mM histidinol (Sigma). cytotoxicity assays as described (Lehner et al., 1995). Immunity 230

Acknowledgments Howard, J.C. (1995). Supply and transport of peptides presented by class I MHC molecules. Curr. Opin. Immunol. 7, 69–76. We thank Balasubramanian Arunachalam, Naveen Bangia, and Hughes, E.A., Ortmann, B., Surman, M., and Cresswell, P. (1996). Jaana Karttunen for helpful discussions; Thomas Spies for the The proteinase inhibitor, N-actyl-L-leucyl-leucyl-L-norleucinal, de- .220.B8 cell line and the pREP8⌬ vector; Matt Androlewicz for the creases the pool of major histocompatibility complexclass I-binding photolabeled peptide; Jonathan Yewdell for the influenza virus; Tom peptidesand inhibits peptide trimming in the endoplasmic reticulum. Novak for the pMCFR.neo vector, and Nancy Dometios for help in J. Exp. Med. 183, 1569–1578. the preparation of this manuscript. This work was supported by the Jardetzky, T.S., Lane, W.S., Robinson, R.A., Madden, D.R., and Wi- Howard Hughes Medical Institute (P. C.) and The Wellcome Trust ley, D.C. (1991). Identification of self peptides bound to purified (P. J. L.) HLA-B27. Nature 353, 326–329. Kropshofer, H., Vogt, A.B., Moldenhauer, G., Hammer, J., Blum, J.S., Received November 24, 1997; revised January 7, 1998. Ha¨ mmerling, G.J. (1996). Editing of the HLA-DR-peptide repertoire by HLA-DM. EMBO J. 15, 6144–6154. Lee, N., Malacko, A.R., Ishitani, A., Chen, M.-C., Bajorath, J., Mar- References quardt, H., and Geraghty, D.E. (1995). The membrane-bound and soluble forms of HLA-G bind identical sets of endogenous peptides Ahn, K., Gruhler, A., Galocha, B., Jones, T.R., Wiertz, E.J.H.J., but differ with respect to TAP association. Immunity 3, 591–600. Ploegh, H.L., Peterson, P.A., Yang, Y., and Fru¨ h, K. (1997). The Lehner, P.J., Wang, E.C.Y., Moss, P.A.H., Williams, S., Platt, K., ER-luminal domain of the HCMV glycoprotein US6 inhibits peptide Friedman, S.F., Bell, J.I., and Borysiewicz, L.K. (1995). Human HLA- translocation by TAP. Immunity 6, 613–621. A0201 restricted CTL recognition of influenza A is dominated by T Androlewicz, M.J., Anderson, K.S., and Cresswell, P. (1993). Evi- cells bearing the V␤17 gene segment. J. Exp. Med. 181, 79–91. dence that transporters associated with antigen processing translo- Lehner, P.J., Karttunen, J.T., Wilkinson, G.W.G., and Cresswell, P. cate a major histocompatibility complex class I-binding peptide into (1997). The human cytomegalovirus US6 glycoprotein inhibits trans- the endoplasmic reticulum in an ATP-dependent manner. Proc. Natl. porter associated with antigen processing-dependent peptide Acad. Sci. USA 90, 9130–9134. translocation. Proc. Natl. Acad. Sci. USA 94, 6904–6909. Androlewicz, M.J., Ortmann, B., van Endert, P.M., Spies, T., and Lewis, J.W., Neisig, A., Neefjes, J., and Elliott, T. (1996). Point muta- Cresswell, P. (1994). Characteristics of peptide and major histocom- tions in the ␣2 domain of HLA-A2.1 define a functionally relevant patibility complex class I/␤2-microglobulin binding to the transport- interaction with TAP. Curr. Biol. 6, 873–883. ers associated with antigen processing (TAP1 and TAP2). Proc. Natl. Li, S., Sjogren, H., Hellman, U., Pettersson, R.F., and Wang, P. (1997). Acad. Sci. USA 91, 12716–12720. Cloning and functional characterization of a subunit of the trans- Arnold, D., Driscoll, J., Androlewicz, M., Hughes, E., Cresswell, P., porter associated with antigen processing. Proc. Natl. Acad. Sci. and Spies, T. (1992). Proteasome subunits encoded in the MHC are 94, 8708–8713. not generally required for the processing of peptide bound by MHC Lutz, P.M., and Cresswell, P. (1987). An epitope common to HLA class I molecules. Nature 360, 171–174. class I and class II antigens, Ig light chains, and ␤2 microglobulin. Carreno, B.M., Solheim, J.C., Harris, M., Stroynowski, I., Connolly, Immunogenetics 25, 228–233. J.M., and Hansen, T.H. (1995). TAP associates with a unique class Meyer, T.H., van Endert, P.M., Uebel, S., Ehring, B., and Tampe, R. I conformation, whereas calnexin associates with multiple class I (1994). Functional expressionand purification of the ABC transporter forms in mouse and man. J. Immunol. 155, 4726–4733. complex associated with antigen processing (TAP) in insect cells. Chang, S.A., Lacaille, V.G., Guttoh, D.S., and Androlewicz, M.J. FEBS Lett. 351, 443–447. (1996). Binding and transport of melanoma-specific antigenic pep- Neefjes, J.J., Momburg, F., and Ha¨mmerling, G.J. (1993). Selective tides by the transporter associated with antigen processing. Mol. and ATP-dependent translocation of peptides by the MHC-encoded Immunol. 33, 1165–1169. transporter. Science 261, 769–771. Denzin, L.K., Robbins, N.F., Carboy-Newcomb, C., and Cresswell, Neisig, A., Wubbolts, R., Zang, X., Melief, C., and Neefjes, J. (1996). P. (1994). HLA-DM and correction of the class II antigen-processing Allele-specific differences in the interaction of MHC class I mole- defect in T2 cells. Immunity 1, 595–606. cules with transporters associated with antigen processing. J. Im- Fru¨ h, K., Ahn, K., Djaballah, H., Sempe´ , P., van Endert, P.M., Tampe´ , munol. 156, 3196–3206. R., Peterson, P.A., and Yang, Y. (1995). A viral inhibitor of peptide Nijenhuis, M., and Ha¨mmerling, G.J. (1997). Multiple regions of the transporters for antigen presentation. Nature 375, 415–418. transporter associated with antigen processing (TAP) contribute to Grandea, I.A.G., Androlewicz, M.J., Athwal, R.S., Geraghty, D.E., its binding site. J. Immunol. 157, 5467–5477. and Spies, T. (1995). Dependence of peptide binding by MHC class Noessner, E., and Parham, P. (1995). Species-specific differences I molecules on their interaction with TAP. Science 270, 105–108. in chaperone interaction of human and mouse histocompatibility Grandea, A.G., Lehner, P.J., Cresswell, P., and Spies, T. (1997). complex class I molecules. J. Exp. Med. 181, 327–337. Regulation of MHC class I heterodimer stability and interaction with Ortmann, B., Androlewicz, M., and Cresswell, P. (1994). MHC class TAP by tapasin. Immunogenetics 46, 477–483. I/␤2-microglobulin complexes associate with TAP transporters be- Greenwood, R., Shimizu, Y., Sekhon, G.S., and DeMars, R. (1994). fore peptide binding. Nature 368, 864–867. Novel allele-specific, post translational reduction in HLA class I sur- Ortmann, B., Copeman, J., Lehner, P.J., Sadasivan, B., Herbert, J.A., face expression in a mutant human B cell line. J. Immunol. 153, Grandea, A.G., Riddell, S.R., Tampe, R., Spies, T., Trowsdale, J., et 5525–5536. al. (1997). A critical role for tapasin in the assembly and function of Hengel, H., Koopmann, J.-O., Flohr, T., Muranyi, W., Goulmy, E., multimeric MHC class I-TAP complexes. Science 277, 1306–1309. Ha¨ mmerling, G.J., and Koszinowski, U.H. (1997). A viral ER-resident Parham, P., Barnstable, C.J., and Bodmer, W.F. (1979). Use of mono- glycoprotein inactivates the MHC-encoded peptide transporter. Im- clonal antibody (w6/32) in structural studies of HLA-A, B, C antigens. munity 6, 623–632. J. Immunol. 123, 342–349. Hill, A., Jugovic, P., York, I., Russ, G., Bennink, J., Yewdell, J., Peace-Brewer, A.L., Tussey, L.G., Matsui, M., Li, G., Quinn, D.G., Ploegh, H., and Johnson, D. (1995). Herpes simplex virus turns off and Frelinger, J.A. (1996). A point mutation in HLA-A0201 results in the TAP to evade host immunity. Nature 375, 411–415. failure to bind the TAP complex and to present virus-derived pep- Hochstenbach, F., David, V., Watkins, S., and Brenner, M.B. (1992). tides to CTL. Immunity 4, 505–514. Endoplasmic reticulum resident protein of 90 kilodaltons associates Rock, K.L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., with the T-cell and B-cell antigen receptors and major histocompati- Hwang, D., and Goldberg, A.L. (1994). Inhibitors of the proteasome bility complex antigens during their assembly. Proc. Natl. Acad. Sci. block the degradation of most cell proteins and the generation of USA 89, 4734–4738. peptides presented on MHC class I molecules. Cell 78, 761–771. Soluble Tapasin Restores Class I Expression 231

Rowe, M., Khanna, R., Jacob, C.A., Argaet, V., Kelly, A., Powis, S., Belich, M., Croom-Carter, D., Lee, S., Burrows, S.R., et al. (1995). Restoration of endogenous antigen processing in Burkitt’s lym- phoma cells by Epstein-Barr virus latent membrane protein-1: coor- dinate up-regulation of peptide transporters and HLA-class I antigen expression. Eur. J. Immunol. 25, 1374–1384. Russ, G., Esquivel, F., Yewdell, J.W., Cresswell, P., Spies, T., and Bennink, J.R. (1995). Assembly, intracellular localization, and nucle- otide binding properties of the human peptide transporters TAP1 and TAP2 expressed by recombinant vaccinia viruses. J. Biol. Chem. 270, 21312–21318. Sadasivan, B.K., Cariappa, A., Waneck, G.L., and Cresswell, P. (1995). Assembly, peptide loading, and transport of MHC class I molecules in a calnexin-negative cell line. Cold Spring Harbor Symp. LV, 267–275. Sadasivan, B., Lehner, P.J., Ortmann, B., Spies, T., and Cresswell, P. (1996). Roles for calreticulin and a novel glycoprotein, tapasin, in the interaction of MHC class I molecules with TAP. Immunity 5, 103–114. Scott, J.E., and Dawson, J.R. (1995). MHC class I expression and transport in a calnexin-deficient cell line. J. Immunol. 155, 143–148. Sijts, A.J., and Pamer, E.G. (1997). Enhanced intracellular dissociation of major histocompatibility complex class I-associated peptides: a mechanism for optimizing the spectrum of cell surface-presented cytotoxic T lymphocyte epitopes. J. Exp. Med. 185, 1403–1411. Solheim, J.C., Harris, M.R., Kindle, C.S., and Hansen, T.H. (1997).

Prominence of ␤2-microglobulin, class I heavy chain conformation, and tapasin in the interactions of class I heavy chain with calreticulin and the transporter associated with antigen processing. J. Immunol. 158, 2236–2241. Stam, N.J., Spits, H., and Ploegh, H.L. (1986). Monoclonal antibodies raised against denatured HLA-B locus H-chains permit biochemical characterization of certain HLA-C locus products. J. Immunol. 137, 2299–2306. Suh, W.-K., Cohen-Doyle, M.F., Fru¨ h, K., Wang, K., Peterson, P.A., and Williams, D.B. (1994). Interaction of MHC class I molecules with the transporter associated with antigen processing. Science 264, 1322–1326. Sutton, J., Rowland-Jones, S., Rosenberg, W., Nixon, D., Gotch, F., Gao, X.-M., Murray, N., Spoonas, A., Driscoll, P., Smith, M., et al. (1993). A sequence pattern for peptides presented to cytotoxic T lymphocytes by HLA-B8 revealed by the analysis of epitopes and eluted peptides. Eur. J. Immunol. 23, 447–453. van Endert, P.M., Tampe, R., Meyer, T.H., Tisch, R., Bach, J.-F., and McDevitt, H.O. (1994). A sequential model for peptide binding and transport by the transporters associated with antigen processing. Immunity 1, 491–500. Weber, D.A., Evavold, B.D., and Jensen, P.E. (1996). Enhanced dissociation of HLA-DR-bound peptides in the presence of HLA-DM. Science 274, 618–620. Wei, M.L., and Cresswell, P. (1992). HLA-A2 molecules in an antigen- processing mutant cell contain signal sequence-derived peptides. Nature 356, 443–446.