TAP-Independent MHC Class I Peptide Antigen Presentation to ...

TAP-Independent MHC Class I Peptide Antigen Presentation to Alloreactive CTL Is Enhanced by Target Cell Incubation at Subphysiologic Temperatures1 Yan Shi,2*‡ Kelly D. Smith,3*† and Charles T. Lutz4*‡§ We investigated the peptide dependency of a group of CD81 anti-HLA-B7 alloreactive CTL. The CTL killed target cells after acid denaturation of more than 98% of target cell surface peptide/MHC class I complexes. The CTL also killed TAP2 HLA-B7transfected T2 (T2B7) cells. The killing was enhanced by target cell incubation at 26°C. Despite these findings, which suggested peptide-independent allorecognition, CTL-mediated cytolysis was reduced or abolished by several point mutations affecting the HLA-B7 peptide-binding groove. Acid denaturation of HLA complexes on T2B7 cells prohibited CTL recognition. CTL recognition was restored by T2B7 cell incubation with b2-microglobulin and a single HPLC fraction containing peptides extracted from TAP1HLA-B71 cells, but not by any of a panel of 17 synthetic HLA-B7-binding peptides. These findings indicated that CTL allorecognition was peptide specific. Sensitizing peptide was extracted from T2B7 cells only after incubation at 26°C. The amount of peptide detected in TAP1 cells was at least 10-fold and 100-fold greater than that detected in TAP2 cells incubated at 26°C and at 37°C, respectively. TAP-independent peptide epitope presentation was sensitive to treatment with brefeldin A, but not sensitive to treatment with chloroquine, consistent with an endogenous peptide source. We propose that subphysiologic temperature incubation can enhance peptide/MHC class I presentation in the total absence of TAP function. The Journal of Immunology, 1998, 160: 4305– 4312.

M

HC class I molecules are composed of heavy chain, b2-microglobulin (b2m),5 and peptide, and are assembled in the endoplasmic reticulum (ER) and transported to the cell surface for CD81 T cell recognition. TAP-1/ TAP-2 heterodimers translocate MHC class I-binding peptides from the cytosol into the ER (1). TAP-2-deficient mouse RMA-S cells and TAP2 human 721.174 and T2 cells express little cell surface MHC class I and fail to present most viral and other MHC class I-associated peptides (1). Furthermore, viral inhibitors of TAP function cause low MHC class I cell surface expression and poor peptide Ag presentation (2, 3). Mouse MHC class I molecules on TAP-deficient cells can be efficiently loaded with exogenous synthetic peptides (4, 5). For these reasons, MHC class I molecules on TAP-deficient cells appear to be “empty” or devoid of tightly bound peptides. Subphysiologic temperature incubation significantly enhances MHC class I expression on TAP-deficient cells,

Departments of *Pathology and †Microbiology and ‡Graduate Program in Immunology and §Molecular Biology, University of Iowa, Iowa City, IA 52242 Received for publication October 31, 1997. Accepted for publication January 6, 1998. 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 with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by Diabetes and Endocrinology Research Center Grant DK25295, National Institutes of Health Grant DE11139, and National Institutes of Health Medical Scientist Training Program Grant GM0737 to K.D.S. 2 Current address: Department of Pathology, University of Massachusetts Medical Center, 155 Lake Avenue N., Worchester, MA 01655. 3 Current address: Department of Pathology, University of Washington Medical Center, Seattle, WA 98195. 4 Address correspondence and reprint requests to Dr. Charles T. Lutz, University of Iowa, 153B MRC, Department of Pathology, Iowa City, IA 52242-1182. E-mail: [email protected]

Abbreviations used in this paper: b2m, b2-microglobulin; B7.221, 721.221 cells transfected with the unmutated HLA-B7 gene; BFA, brefeldin A; ER, endoplasmic reticulum; LCL, lymphoblastoid cell line; T2B7, T2 cells transfected with the unmutated HLA-B7 gene.

5

Copyright © 1998 by The American Association of Immunologists

especially for mouse MHC class I molecules (4). This is widely believed to be due to the low temperature stabilization of empty MHC class I molecules. Indeed, incubating TAP-deficient cells with anti-MHC mAbs or exogenous synthetic peptides in the cold stabilizes MHC class I molecules after the incubation temperature is raised to 37°C (4 – 8). Without MHC-binding mAbs or exogenous peptides, surface MHC class I molecules on TAP-deficient cells are thermolabile. Despite their obvious defects in MHC class I Ag processing, TAP-deficient cells do express cell surface class I-bound peptides. In certain instances, CD81 T cells recognize TAP-deficient cells that express viral proteins (9 –14) or MHC class I epitopes expressed by transfected minigene constructs (15, 16). Sequence analysis has revealed that HLA-A2-bound peptides in TAP2 cells are derived from protein signal sequences (17, 18). Peptides that are eluted from HLA-B7 molecules on transfected TAP2 T2 cells have a typical HLA-B7-binding sequence motif; the ratio of peptide to HLA-B7 heavy chain, measured spectrophotometrically, is the same in TAP2 and TAP1 cells (19). However, little or no radioactively labeled peptide has been eluted from most MHC class I molecules on metabolically labeled TAP2 cells (18, 19). This indicates that TAP2 cells process MHC-bound peptides by a slow and inefficient route (19). TAP-independent MHC class I Ag processing and presentation, and the induction of empty MHC class I molecules by subphysiologic temperature incubation are relevant to CD81 T cell recognition of alloantigen. TAP-deficient target cells have been used to assess whether T cells recognize allogeneic MHC class I molecules in a peptide-independent or peptide-nonspecific fashion. Many alloreactive CTL do not kill TAP-deficient cells unless the target cells have been incubated with specific cell extracts or HPLC fractions from TAP1 APC lysates, providing convincing evidence that many alloreactive CTL are peptide specific (20). In a few cases, the allopeptides have been identified (21–24). In contrast, some alloreactive CTL kill TAP-deficient target cells in the 0022-1767/98/$02.00

4306

LOW TEMPERATURE TAP-INDEPENDENT PEPTIDE PRESENTATION

absence of exogenous peptides (25, 26). CTL recognition of TAPdeficient cells has been cited to support peptide-independent MHC allorecognition. In this report, we present an analysis of TAP-independent allorecognition of HLA-B7. Previously, we characterized D7 and a group of related cloned CD81 T cells (D7-like CTL) that are specific for EBV EBNA-3c presented by HLA-B37. Here we show that D7-like CTL cross-react with allogeneic HLA-B7 molecules that are expressed by both TAP1 and TAP2 cells. We demonstrate that D7-like CTL are highly specific for a cellular peptide that is expressed on TAP1 cells. Using these cells, we investigate the hypothesis that TAP-independent MHC class I peptide Ag presentation is enhanced at subphysiologic temperatures.

Materials and Methods Cell lines D7-like CTL clones and the control peptide-specific, HLA-B7-alloreactive KOR-18 CTL clone have been described (27–30). TAP2 T2 cells; HLAA-, -B-, and -C-negative 721.221 cells; and lymphoblastoid cell lines (LCL) have been described (27–30). JB LCL was gift of Dr. A. McMichael, Oxford University, U.K. Mouse EL4 cells transfected with the HLA-A2 gene (EL4A2) or the HLA-B7 gene (EL4B7) were gifts of Dr. V. Engelhard, University of Virginia. The cell lines were maintained in culture medium (RPMI 1640, Life Technologies, Gaithersburg, MD; with 10% supplemented calf serum, HyClone, Logan, UT). For the serum-free cell culture, JB LCL and EL4B7 cells were grown in AIM-V medium (Life Technologies) for 2 wk. T2 and 721.221 cells were transfected with various HLA genes in the pHeBo vector and maintained in culture medium supplemented with 300 mg/ml of hygromycin B (Calbiochem, San Diego, CA) as described (19, 29 –31). The HLA-B7/Aw68 chimera was constructed by subcloning an a3 exon-containing BglII-SalI fragment from an HLA-Aw68 genomic clone (a gift of Dr. P. Parham, Stanford University) into a similar site in an HLA-B7 genomic clone in the pHeBo vector. Expression of transfected gene products in target cells lines was confirmed by flow cytometry using W6/32 mAb, as described (19, 29 –31).

Ag presentation inhibitors Brefeldin A (BFA; Sigma, St. Louis, MO) was dissolved in methanol at 20 mg/ml and diluted in culture medium to a final concentration of 1 mg/ml. Chloroquine (Sigma) was dissolved in culture medium immediately before assays and used at 20 mM. EDTA (Sigma) was used at a final concentration of 1 mM. For addition 1 h after initiation of the 51Cr release assay, 50 ml of 5 mM EDTA (in an aqueous solution of 140 mM NaCl and 15 mM HEPES) was added to 200 ml of assay mixture.

Cell extraction and HPLC separation HLA-B71 JY and T2B7 cells (109) were grown to a density of 106 cells/ml with less than 5% cell death. Some batches had been incubated at 26°C for 18 to 20 h before extraction. Cells were washed once with PBS and lysed in 0.1% trifluoroacetic acid (Fisher, Itasca, IL) in HPLC-grade water (Fisher) by vigorous vortexing for 1 min. Another 15 ml of 1% trifluoroacetic acid in water was added to the lysates. The lysates were kept on ice for 30 min with brief vortexing every 5 min. After centrifuging the lysates at 28,000 rpm at 4°C on a Beckman L8-80 M centrifuge (Beckman, Fullerton, CA) with a type 28 rotor, the supernatants were lyophilized. The lyophilates were dissolved in 0.15% trifluoroacetic acid in HPLC-grade water and centrifuged through Centricon 10 filters (Amicon, Beverly, MA) at 5000 3 g for at least 4 h. The filtrates were again lyophilized and dissolved in 1 ml of 0.15% trifluoroacetic acid. The suspensions were subjected to HPLC separation as previously described (19).

Cell surface MHC class I denaturation, reconstitution, and cytotoxicity assay 51 Cr-labeled HLA-B71 cells were washed with HBSS (Cancer Center, University of Iowa), and incubated for 90 s with acid-stripping medium (0.3 M glycine-HCl and 1% BSA in water, pH 2.4) at cell densities no higher than 2 3 107 cells/ml. Culture medium (100 3 vol) was added to neutralize pH. Cells were washed three times and suspended in assay medium (RPMI 1640 with 5% supplemented calf serum) at 106/ml. Completeness of acid stripping was assessed by flow cytometry, as described above. In some experiments, cell suspensions (50 ml) were mixed with equal volumes of each HPLC peptide fraction in the presence of 10 mg/ml

Table I. D7-like CTL recognize allogeneic HLA-B7 molecules a Expt.

E:T

CTL

1 2

10:1 10:1

3 4

10:1 1:1

5

3:1

D7 D5 D7 D7 D5 D7 D5 D7

Target Cells (% Specific Lysis) b

9013 (75) JY (77) JY (70) B7.221 (81) B7.221 (71) B7.221 (68) JY (63) JY (57)

9063 (0) 9034 (69) 9034 (70) 721.221 (1) A2.221 (2) A2.221 (0) EL4B7 (78) EL4B7 (79)

9103 (21) 9004 (21) 9004 (2)

EL4A2 (26) EL4A2 (21)

a D5 and D7 CTL were used in five separate experiments at the E:T ratios indicated. b Target cells are listed and HLA-B71 target cells are underlined. The numbers in parentheses are the percentage specific lysis of the adjacent target cells. B7.221 and A2.221 are human HLA-A, B, C-negative 721.221 cells transfected with HLA-B7 and HLA-A2, respectively. EL4B7 and EL4A2 are mouse EL-4 cells transfected with HLA-B7 and HLA-A2, respectively. The HLA types for the human LCL targets are: JY; HLA-A2, -B7, -Cw7; 9004: HLA-A2, -B27, -Cw1; 9013: HLA-A3, -B7, Cw7; 9034: HLA-A3, -B7, -Cw7; 9063: HLA-A32, -B44, -Cw5; 9103: HLA-A24, -A26, -B51, -B56, -Cw8.

of human b2m (Calbiochem). The mixtures were shaken at room temperature for 2 h and the volumes were expanded to 1 ml with the assay medium before the 51Cr release assay. HLA-B7-binding synthetic peptides were mixed at a final concentration of 1025 M with b2m (10 mg/ml) and 51 Cr-labeled acid-stripped T2B7 cells (5 3 104 cells in 1 ml) and were shaken at room temperature for 2 h. For all the cytotoxicity assays, peptide incubation mixtures were transferred to 96-well plates at 100 ml/well. An equal volume of effector cells was added to each well. 51Cr release assays were performed as previously described (31), at the E:T ratios indicated. Assays were 5 h long, except where otherwise indicated. Relative lysis used to estimate the effects of HLA-B7 mutations on CTL and was calculated as the percent specific lysis of 721.221 cells transfected with HLA-B7 variant genes divided by the percent specific lysis of 721.221 cells transfected with the parental HLA-B7 gene at the same E:T ratio.

Results EBV-specific CTL recognize allogeneic HLA-B7 molecules D7 and the other CD81 T cell clones used in this study (D7-like CTL) specifically recognize an EBV EBNA-3c peptide (amino acids 285-293) that is presented by autologous HLA-B37 class I molecules (28). D7-like CTL also killed HLA-B71, but not HLAB72 allogeneic LCL (Table I). D7-like CTL did not kill mouse EL4 target cells or human HLA-A-, -B-, or -C-negative 721.221 target cells. Recognition of both mouse and human target cells was achieved by transfection with the HLA-B7 gene, but not with the HLA-A2 gene, confirming cross-reactivity with allogeneic HLA-B7 molecules (Table I). As expected, the anti-HLA class I mAbs, MB40.5 and W6/32, blocked the alloreactive CTL killing, confirming MHC class I specificity (data not shown). To test the CD8 dependence of the allorecognition, we transfected 721.221 cells with a chimeric HLA gene encoding the a1 and a2 domains of HLA-B7, and the a3, transmembrane, and cytoplasmic domains of HLA-Aw68. The HLA-Aw68 molecule has an a3 mutation that disrupts CD8a subunit binding and efficient recognition by many CTL (32). D7-like CTL equivalently lysed 721.221 cells transfected with either unmutated HLA-B7 or the B7/Aw68 chimera (data not shown). CD8 dependency can be overcome by high ligand density and high TCR affinity (33, 34). This suggests that the specific antigenic determinant that was recognized by D7-like CTL on HLA-B71 cells was of high affinity or was highly expressed. D7 CTL kill target cells that express minimal HLA-B7-bound peptide To address whether HLA-B7 recognition by D7-like CTL is peptide-dependent, we acid-treated target cells to denature HLA class I complexes (35). Treatment of HLA-B7-transfected 721.221

The Journal of Immunology

4307 followed by the addition of EDTA after 1 h, acid stripping and continuous BFA treatment, or the combination of treatments severely diminished allorecognition by the peptide-specific KOR-18 CTL (Fig. 1B). This indicates that acid stripping of the target cells efficiently removed cell surface peptide/HLA-B7 complexes that are recognized by KOR-18 CTL. However, the same treatments did not significantly diminish killing by D7 CTL (Fig. 1B), indicating that D7 CTL recognized target cells after acid denaturation of 98.5% of surface HLA-B7 molecules. Thus D7 CTL may recognize extremely low levels of a specific peptide/HLA-B7 complex, a peptide/HLA-B7 complex that is resistant to acid denaturation, or “empty” HLA-B7 molecules that are stabilized by serum b2m after acid denaturation. HLA-B7 peptide-binding groove point mutations abolish recognition by D7 CTL

FIGURE 1. D7 CTL killing of TAP1HLA-B71 cells is resistant to acid stripping. A, B7.221 cells were either untreated (dotted line) or acid stripped (broken line) as described in Materials and Methods, and immediately stained on ice with W6/32 mAb followed by FITC-conjugated goat anti-mouse IgG. Untransfected parental 721.221 cells were also stained (continuous line). B, D7 CTL (filled bars) and peptide-specific alloreactive CTL clone KOR-18 (open bars) were used at an E:T ratio of 5:1 with B7.221 target cells in a 3-h 51Cr release assay. Cells were either untreated (NONE) or were treated with EDTA (1 mM), either throughout the 51Cr release assay (E) or beginning at 1 h after initiation of the assay (E1h). Other B7.221 target cells were acid stripped (AS) as described in Materials and Methods. Acid-stripped target cells received no further treatment, were incubated continuously in 1 mg/ml of BFA in assay medium both before and during the 51Cr release assay (B), were exposed to EDTA 1 h after initiation of the 51Cr release assay (E1h), or received a combination of treatments, as indicated.

(B7.221) cells with HCl/glycine (pH 2.4) medium removed 98.5% of cell surface trimeric HLA/b2m/peptide complexes, as assessed by the binding of conformation-sensitive W6/32 mAb (Fig. 1A). Because HLA-B7 is the only HLA-A, -B, or -C molecule expressed by B7.221 cells, this result indicates that acid denaturation removed the great majority of cell surface peptide/HLA-B7 complexes. Acid stripping minimally decreased the cytolytic activity by D7 CTL and control alloreactive KOR-18 CTL (Fig. 1B). KOR-18 CTL recognize allogeneic HLA-B7 molecules in a peptide-dependent manner (27, 29). We propose that KOR-18 recognition of acid-stripped B7.221 cells was due to target cell reexpression of peptide/HLA-B7 complexes during the 3-h 51Cr release assay. To test this hypothesis, we used two treatments to minimize new HLA expression following acid stripping. EDTA chelates Ca21 and Mg21, prevents cell-cell contact, and prohibits both perforin-mediated and Fas-mediated killing (36). EDTA treatment throughout the assay completely blocked CTL-mediated killing (Fig. 1B). EDTA added 1 h after the assay began did not reduce the cytolysis by either D7 or KOR-18 CTL (Fig. 1B). This indicates that both CTL caused considerable target cell damage in the first hour of the assay. BFA disrupts the Golgi complex and halts expression of HLA complexes (37). BFA treatment alone did not inhibit CTL-mediated killing (data not shown). Acid stripping

To further examine whether D7 CTL recognized peptide/HLA-B7 complexes, we used an indirect approach. HLA-B7 variants with point mutations affecting the peptide-binding groove were transfected into 721.221 cells. The expression of the HLA-B7 variants was comparable by mAb staining and each transfectant had been shown to be killed well by at least one HLA-B7-specific alloreactive CTL (29, 30, 35). Therefore, all variants had relatively unaltered tertiary structure. As predicted, D7 CTL recognition was abrogated by several point mutations in solvent-accessible residues (Fig. 2). Importantly, D7 CTL recognition also was abolished by multiple point mutations in the HLA-B7 peptide-binding groove (Fig. 2) that affected pockets A, B, C, D, and E (see Fig. 2 legend). The HLA-B7 A pocket preferentially binds peptide P1 arginine and alanine, the B pocket binds P2 proline, and the D pocket binds P3 arginine (38). Four point mutations in the peptide-binding groove pockets B (residue 67), F (residue 116), and A (residues 163 and 171), did not significantly affect D7 CTL recognition (Fig. 2). Lack of peptide-binding groove mutation effects has been interpreted to support peptide-independent allorecognition (26). However, a panel of peptide-specific allogeneic CTL clones (29, 30) that were studied in our laboratory, like D7 CTL, were variably affected by peptide-binding groove mutations. Expression of the D7 CTL epitope on TAP2 T2 cells is dependent on temperature and on HLA-B7-bound peptide MHC class I/peptides complexes are poorly expressed by TAPdeficient cells. Therefore, we investigated the ability of D7-like CTL to kill TAP2 T2 cells transfected with the HLA-B7 gene (T2B7). We also tested two HLA-B7 variants (E45A and S97R) that were not recognized by D7 CTL when expressed on TAP1 721.221 cells. Unlike many mutants, these two HLA-B7 variants are expressed at levels equivalent to or higher than unmutated HLA-B7 on transfected T2 cells (19). D7 and D5 CTL killed T2B7 cells (Fig. 3). Overnight T2B7 target cell culture at 26°C increased recognition by D7 and D5 CTL equivalent to that of TAP1 B7.221 target cells (Fig. 3). However, D7 and D5 CTL did not kill T2 cells that were transfected with either the E45A or the S97R HLA-B7 variant genes, even after target cell incubation at 26°C (Fig. 3 and data not shown). These data indicate that D7 CTL allorecognition is sensitive to changes in HLA-B7 peptide-binding groove residues, even on TAP2 target cells. To remove peptide/HLA-B7 complexes from the TAP2 target cells, we acid stripped T2B7 cells. As above, D7 and D5 CTL recognition was enhanced by a 26°C overnight incubation of T2B7 cells (Fig. 4). D7 and D5 CTL did not kill acid-stripped T2B7 cells (Fig. 4). Peptide processing by TAP2 T2B7 cells is slow (19), obviating the need for BFA to prevent peptide/HLA-B7 reexpression in the 5-h 51Cr release assay (Fig. 4).

4308


FIGURE 2. HLA-B7 peptide-binding groove variants affect recognition by D7 CTL. Shown is a top view of the HLA class I molecule. Filled residues denote relative cytolysis (see Materials and Methods) of the variant transfectant at less than 20% of the unmutated HLA-B7 transfectant, stippled residues denote 20 to 70% relative cytolysis, and open residues denote no significant decrease in cytolysis. The E45A mutation and the amino acids substituted at other residues have been described (29, 35). Cell surface expression of each variant transfectant was found to be within twofold of the unmutated HLA-B7 transfectant. D7 CTL recognition was reduced by mutations in the following peptide-binding pockets: residue 9 (pockets B, C); 45 (B); 63 (A, B); 66 (A, B); 70 (B, C); 73 (C); 74 (C); 97 (C, E); 99 (A, B, D); 152 (E); 155 (D); 156 (D, E); and 167 (A).

D7 CTL recognize a specific HLA-B7-bound peptide presented by TAP1 and TAP2 cells It has been proposed that some alloreactive T cells are peptide dependent but not peptide specific (39, 40). Therefore, we tested acid-stripped T2B7 cells that had been pulsed with human b2m and a panel of 17 synthetic HLA-B7-binding peptides. D7 (Table II) and D5 (data not shown) CTL did not kill the target cells pulsed with a high concentration (1025 M) of these synthetic peptides.

FIGURE 3. D7-like CTL do not recognize T2 cells transfected with a peptide-binding groove HLA-B7 mutant. T2B7 cells and T2 cells that were transfected with the E45A HLA-B7 variant were incubated at 37°C or 26°C overnight before the 51Cr release assay, as indicated (Temp). B7.221 cells were cultured continuously at 37°C. D5 and D7 CTL clones were used at an E:T ratio of 10:1.

Therefore, D7-like CTL do not promiscuously recognize peptide/ HLA-B7 complexes or an HLA-B7 structure that is stabilized by peptide binding. We next investigated the cellular HLA-B7-bound peptides that are recognized by D7-like CTL. To avoid potential loss of low affinity peptides during HLA-B7 purification, we extracted TAP1 HLA-B71 JY cells and HPLC-separated peptides as outlined in

FIGURE 4. D7-like CTL recognition of T2B7 cells is sensitive to acid stripping. B7.221 target cells were used without treatment. T2B7 target cells were incubated as indicated at 26°C (Low Temp) or at 37°C overnight, acid stripping was done immediately before the 51Cr release assay, and BFA was used at 1 mg/ml. CTL clones D5 and D7 were used at an E:T ratio of 10:1.

The Journal of Immunology

4309

Table II. D7-like CTL do not promiscuously recognize peptide-stabilized HLA-B7 molecules a Peptideb

838 858 899 909 937 971 979 995 999 1012 1026 1039 1080 1104 1127 EBNA-3A EBNA-3C

Sequence

Specific Lysis (%) c

APAPTVAVX APYGGPXAX APRTVALTA PRSGPGPEX APRXPXTGX MPRGVVVTX APRAXXXXX APRTLVLLL LVMAPRTVL APRTVALTAL RVMAPRAXX RPKSNIVLL APRAFXPXPV AASKERSGVSL SPRVIFTML RPPIFIRRL QPRAPIRPI

1 5 10 2 6 7 24 13 12 7 4 7 5 5 10 7 6

a T2B7 cells were incubated overnight at 26°C, acid stripped, and mixed with individual peptides and b2m as described in Materials and Methods. D7 CTL were tested at an E:T ratio of 10:1. D7 CTL killing of positive control JY cells was 79%, and 26°C-incubated T2B7 cells (without acid stripping) was 51%. b HLA-B7-binding peptides have been described previously (35, 38). c X indicates an equimolar mixture of L and I. Ambiguity in peptide sequences is due to the inability of mass spectroscopy to distinguish L and I.

Materials and Methods. Acid-stripped T2B7 target cells were incubated with human b2m and individual HPLC fractions. A single HPLC fraction (fraction 27) sensitized target cells for lysis by D7 CTL (Fig. 5A), consistent with the specific recognition of a single HLA-B7-bound peptide. Sensitizing peptide also was eluted from B7.221 cells (data not shown). Little or no D7 CTL lysis was induced by a similarly prepared extract from T2B7 cells that had

FIGURE 5. T2B7 cell peptide presentation is significantly enhanced by subphysiologic temperature incubation. Low m.w. molecules were acid extracted from TAP2 T2B7 cells that had been incubated at 37°C (T2B7) or 26°C (T2B7/26*C), or from TAP1 HLA-B71 JY cells (JY), as described in Materials and Methods. Extracts were HPLC fractionated and mixed with b2m and 51Cr-labeled acid-stripped T2B7 target cells as described in Materials and Methods. D7 CTL were tested at an E:T ratio of 10:1. A, Test of HPLC fractions 19 to 30. B, HPLC fraction 27 was diluted in PBS before mixing with b2m and acid-stripped T2B7 target cells.

been cultured at 37°C (Fig. 5A). However, considerable targeting activity, also present in HPLC fraction 27, was extracted from T2B7 cells after they had been cultured overnight at 26°C (Fig. 5A). This indicates that expression of the HLA-B7-bound peptide epitope was enhanced by subphysiologic temperature incubation. We quantitated the abundance of sensitizing peptide in each extract by dilution of HPLC fraction 27 (Fig. 5B). Even without dilution, sensitizing activity in the T2B7 cell (37°C) extract was not detectable. Fraction 27 of the T2B7 cell (26°C) extract sensitized target cells for lysis without dilution, but failed to do so after a 10-fold dilution (Fig. 5B). In contrast, fraction 27 of the JY cell extract sensitized target cells for lysis after a 100-fold dilution. Activity was above background after a 1000-fold dilution (Fig. 5B). These data indicate that TAP1 JY cells contained at least 100-fold more sensitizing peptide than did TAP2 T2B7 cells incubated at 37°C and at least 10-fold more sensitizing peptide than did T2B7 cells incubated at 26°C. D7 CTL recognize an HLA-B7-bound peptide that is derived from an endogenous cellular protein in TAP1 and TAP2 cells The observations that were presented above indicated that D7-like CTL recognition of T2 cells was peptide specific. We thus wanted to analyze the TAP-independent MHC class I peptide presentation pathway. We considered the possibility that the peptide was derived from calf serum present in cell culture media. HLAB71TAP1JB LCL and EL4B7 cells were cultured in serum-free AIM-V medium for 2 wk. T2B7 target cells died when cultured in AIM-V medium and were not tested. Traces of putative retained calf serum peptides were removed from JB LCL and EL4B7 cells by acid stripping. The acid-treated cells were further incubated in serum-free medium for 1 day and tested with D7 CTL in serumfree medium. D7 CTL did not show any decrease in killing of these target cells (data not shown), suggesting that the peptide recognized by D7 CTL is not derived from a calf serum protein. To test whether the HLA-B7-bound peptide on T2B7 cells was from an exogenous source, we incubated acid-stripped B7.221 cells and T2B7 cells overnight in chloroquine and tested the killing of these target cells by D7 CTL. Chloroquine inhibits endosomal compartment acidification and prevents exogenous Ag presentation. For comparison, we used BFA, which disrupts the Golgi complex. As expected, acid-stripped B7.221 cells were sensitive to D7 CTL killing in the presence of either inhibitor during the overnight incubation (Fig. 6). The killing of acid-stripped B7.221 cells in the presence of BFA was somewhat lower. This might be due to gradual loss of residual peptide/HLA-B7 complexes from acid-stripped B7.221 cells or to increased B7.221 target cell death observed after overnight culture in BFA at 37°C (data not shown). B7.221 target cells incubated for 6 h after acid stripping in the presence of BFA were killed well (data not shown). Overnight incubation of acidstripped T2B7 cells with BFA at 26°C was not toxic and prevented the recognition by D7 CTL (Fig. 6). Overnight treatment with chloroquine did not prohibit D7 CTL recognition. These data are consistent with the hypothesis that D7 and similar CTL recognized a peptide epitope that was derived from an endogenous cellular protein in TAP2 cells.

Discussion TAP molecules are required for efficient MHC class I peptide Ag presentation. In the absence of functional TAP heterodimers, cell surface MHC class I expression is low and presentation of endogenous cellular or viral Ags is poor. Nonetheless, TAP-deficient cell presentation of MHC class I molecule-bound peptides to CTL has

4310


FIGURE 6. Presentation of the specific peptide on TAP2 cells is inhibited by BFA, but not by chloroquine. Target cells were B7.221 (A) and T2B7 (B). Some cells were acid stripped either 24 h (Prior AS) or immediately (Imm AS) before the 51Cr release assay. Twenty-four hours before the assay, B7.221 and T2B7 target cells were incubated at 37°C and 26°C, respectively. BFA (1 mg/ml) and chloroquine (20 mM) inhibitors were added as indicated during the incubation and throughout the cytotoxicity assay. D7 CTL were used at an E:T ratio of 5:1. EDTA (1 mM) was added to each cytotoxicity assay mixture 1 h after initiation of the 51Cr release assay to prevent CTL recognition of newly regenerated cell surface peptide/MHC class I complexes. In this assay, BFA blocked the control KOR-18 CTL peptide epitope regeneration after acid stripping (data not shown).

been documented. We investigated EBV-specific CTL that crossreact with allogeneic HLA-B7 on TAP1 and TAP2 cells in a peptide-specific fashion. Targeting activity that was extracted from TAP2 T2B7 cells and from TAP1 B7.221 cells migrated in the same HPLC fraction, suggesting a single HLA-B7-bound peptide. At least 100-fold more targeting activity was extracted from TAP1 B7.221 cells grown at 37°C than from T2B7 cells grown at 37°C. This is consistent with the greater efficiency of TAP-dependent than TAP-independent peptide Ag presentation. Despite differences in presentation efficiency, it is clear that D7-like CTL recognize a peptide epitope in a TAP-independent fashion. TAP-independent presentation of alloreactive CTL has been observed in other systems (29, 41, 42). At least four potential mechanisms can explain TAP-independent MHC class I Ag presentation. First, Ags may be derived from extracellular sources. Extracellular peptides or proteins may be digested by serum proteases and then bind to cell surface MHC class I molecules (43). Ags from phagocytosed bacteria may be “regurgitated” and gain access to cell surface MHC class I molecules (44). Particulate Ags may be phagocytosed, processed in endosomal compartments, and presented to MHC class I molecules in a BFA-independent fashion (45). Second, the TAP-1 protein may form homodimers in cells that lack functional TAP-2 proteins. In cells with a defect in TAP gene expression, introducing a rat TAP-1 cDNA restored presentation of a viral epitope (46). A vesicular stomatitis virus peptide epitope is presented well by infected TAP-deficient RMA-S cells, but not by infected, H2Kbtransfected TAP2 T2 cells (10). Third, some peptides that are generated in the ER can directly bind nascent MHC class I molecules, without the need for TAP transport. Peptides from protein signal sequences have been isolated from HLA-A2 class I molecules in TAP2 T2 cells (17, 18). In some cases, signal sequences

can direct the cell surface presentation of MHC class I epitopes in TAP2 cells (15). Fourth, cytosolic peptides may be transported into the ER by a TAP-independent mechanism. The proteins that mediate this transport are not known. Some viral epitopes (9 –14) and peptides encoded by plasmid minigenes (15, 16) are presented in TAP-deficient cells, including TAP2 T2 cells. Common features of TAP-independent MHC class I peptide Ag transport are slow and inefficient peptide processing. In addition, the peptides that are presented have been generated in high quantity in the cytoplasm, often from viral or plasmid genes. It is unlikely that the HLA-B7-bound epitope that is recognized by D7-like CTL is derived from an exogenous protein or processed by serum proteases. T2 cells are not known to be phagocytic, and acid stripping and overnight chloroquine incubation did not prevent recognition of T2B7 cells. This is inconsistent with uptake of exogenous Ag via an endosomal route. Long-term target cell growth and assay in serum-free media combined with acid stripping to remove residual peptide/MHC complexes did not diminish recognition of B7.221 cells by D7 CTL. In regard to the second proposed TAP-independent MHC class I Ag presentation mechanism, T2B7 cells completely lack both TAP-1 and TAP-2 proteins, ruling out potential TAP-1 homodimer function, as suggested for mouse RMA-S cells. Our data do not exclude the possibility that the HLA-B7-bound peptide epitope is derived from a protein signal sequence, even though the peptide epitope was presented more efficiently in TAP1 cells than in TAP2 cells. Some signal sequences are not processed for MHC class I binding in the ER and must be exported to the cytoplasm for further processing and subsequent TAP-mediated importation into the ER (22, 47). Signal sequence peptide processing may be inefficient in the ER. Our peptide is presented at a low level in the absence of TAP and at higher levels in the presence of TAP. However, peptide synthesis is slower at subphysiologic temperatures, and it is expected that generation of HLA-binding signal sequence peptides also would be reduced. We speculate that the peptide epitope is transported from the cytoplasm into the ER efficiently by TAP and inefficiently by a TAP-independent mechanism. The relatively inefficient putative TAP-independent transport is consistent with the delayed and relatively feeble presentation of viral (9 –14) and plasmid minigene (15, 16) epitopes by TAP-deficient cells. Slow and relatively inefficient transport of peptides into the ER by a TAP-independent mechanism also is consistent with the observation that HLA-B7bound peptides are poorly labeled by 3H-labeled amino acids, but are detectable spectrophotometrically in T2B7 cells (19). Mouse MHC class I molecules on the surface of TAP-deficient cells may be largely empty or devoid of tightly bound peptides. The low level of surface MHC class I molecules is greatly enhanced by overnight incubation of TAP-deficient cells at subphysiologic temperatures. These molecules are unstable and gradually disappear upon raising the temperature to 37°C, unless the TAPdeficient cells are incubated with MHC class I-binding mAb or peptides (4, 5, 6 – 8). Subphysiologic temperatures stabilize empty MHC class I molecules in vitro and in vivo. However, incubation at subphysiologic temperatures induces relatively little expression of most human MHC class I molecules on either human TAP2 T2 cells or mouse TAP-deficient RMA-S cells (5, and data not shown). D7-like CTL killed TAP2 T2B7 cells and recognition was greatly enhanced by overnight target cell incubation at 26°C. Enhanced CTL recognition correlated with enhanced expression of CTL targeting activity at 26°C. To our knowledge, the current study is the first to quantitate TAP-independent MHC class

The Journal of Immunology I-bound peptide expression at different temperatures. Subphysiologic temperature-enhanced peptide presentation has been reported for mouse TAP-2-deficient RMA-S cells (41). Our results show the same phenomenon may be observed in the total absence of TAP proteins. Because very little is known about the mechanism of putative TAP-independent peptide transport, it is unclear why expression of the peptide epitope is enhanced at subphysiologic temperatures. However, several mechanisms can be envisioned. Decreased temperature may increase formation of trimeric peptide/b2m/HLA-B7 complexes in the ER. Subphysiologic temperatures may increase TAP-independent peptide transport, stabilize b2m/HLA-B7 heterodimers, or slow HLA-B7 heavy chain, b2m, or peptide degradation. It is known that unbound peptides are rapidly exported from the ER into the cytoplasm (48). The ER peptide export mechanism may be temperature sensitive. Alternatively, it is possible that slow turnover of cell surface peptide/MHC complexes at subphysiologic temperatures may account for enhanced expression of the peptide epitope. Our findings are relevant to whether alloreactive T cells can recognize MHC class I molecules in a peptide-independent or peptide-nonspecific fashion. It is clear that many alloreactive CTL recognize complexes of specific peptides bound to allogeneic MHC class I molecules (20). However, it has been proposed that many other CTL recognize MHC epitopes that are independent of specific bound peptides (25, 26, 39, 40, 49). Many of these studies have utilized TAP-deficient cells. Enhanced alloantigen presentation at subphysiologic temperatures has been interpreted as evidence of peptide-independent allorecognition (27). Some of our results had initially suggested that D7-like CTL might recognize HLA-B7 in a peptide-independent manner. D7-like CTL recognized and killed TAP2 T2B7 cells, and killing was enhanced by target cell incubation at subphysiologic temperatures. D7 and similar CTL killed TAP1HLA-B71 target cells, even after $98% of the peptide/HLA-B7 complexes had been removed from the cell surface and reexpression of new complexes was prohibited by BFA treatment. However, the suggestion that D7-like CTL allorecognition is independent of HLA-B7-bound peptide is not correct. In retrospect, it is clear that the ability of D7-like CTL to recognize acid-stripped target cells was due to the very high level of peptide epitope expression by TAP1 cells. CTL recognition exhibits a threshold pattern. Even when most peptide/MHC complexes are removed from target cells, CTL killing is efficient as long as the number of remaining peptide/MHC complexes are above threshold levels. For some T cells, this number can be very low, perhaps as few as one complex per cell (50). T2B7 cells grown at 26°C and 37°C contained at least 10-fold less and 100fold less sensitizing peptide, respectively, than did B7.221 cells grown at 37°C, yet the T2B7 target cells were recognized and killed by D7-like CTL. The high abundance of the peptide epitope on TAP1 cells suggests that the peptide is produced in large quantities in the cell. In parallel with viral (9 –14) and plasmid minigene (15, 16) epitopes, this presents a plausible explanation of why the peptide is presented, albeit inefficiently, by TAP2 T2B7 cells. As noted above, the reason for enhanced recognition of T2B7 cells after culture at subphysiologic temperatures was due to the unexpected enhanced expression of specific cell surface peptide/ HLA-B7 complexes. Our results show that T cell recognition of allogeneic MHC class I molecules on living target cells may depend upon a tiny amount of specific MHC-bound peptide. Previous studies (26, 51) have demonstrated allorecognition of purified, largely empty MHC class I molecules attached to plastic. These studies concluded that allorecognition can be peptide-independent and that some CTL may specifically recognize empty

4311 MHC class I molecules. In one of these studies (26), the alloreactive CTL also shared many properties of D7 CTL: recognition of acid-stripped target cells, lack of effect by some peptide-binding groove mutations, and recognition of TAP2 cells. We were able to generate suitable targets for peptide targeting studies by acid-stripping TAP2 T2B7 cells; we demonstrated that D7 and similar alloreactive CTL clones recognize specific peptides. This finding calls into question the physiologic significance of allorecognition of purified empty MHC class I molecules on plastic. Stimulation of alloreactive T cells by high density MHC on plastic surfaces may represent the inherent affinity of TCR for allogeneic MHC. CTL allorecognition of living cells may nonetheless involve specific peptides. Given that only tiny amounts of specific peptide are sufficient to stimulate some T cells, this possibility is difficult to exclude. Our results emphasize caution when interpreting apparently peptide-independent allorecognition.

Acknowledgments We thank Brian Mace for excellent technical support and Dr. John Harty for helpful discussions. HLA-B7-binding peptides were a kind gift of Dr. Victor H. Engelhard at the University of Virginia.

References 1. York, I. A., and K. L. Rock. 1996. Antigen processing and presentation by the class I major histocompatibility complex. Annu. Rev. Immunol. 14:369. 2. Ahn, K., A. Gruhler, B. Galocha, T. R. Jones, E. J. H. J. Wiertz, H. L. Ploegh, P. A. Peterson, Y. Yang, and K. Fru¨h. 1997. The ER-luminal domain of the HCMV glycoprotein US6 inhibits peptide translocation by TAP. Immunity 6:613. 3. Fru¨h, K., K. Ahn, and P. A. Peterson. 1997. Inhibition of MHC class I antigen presentation by viral proteins. J. Mol. Med. 75:18. ¨ hle´n, J. J. Neefjes, P. Ho¨glund, M.-T. 4. Ljunggren, H.-G., N. J. Stam, C. O Heemels, J. Bastin, T. N. M. Schumacher, A. Townsend, K. Ka¨rre, and H. L. Ploegh. 1990. Empty MHC class I molecules come out in the cold. Nature 346:476. 5. Anderson, K. S., J. Alexander, M. Wei, and P. Cresswell. 1993. Intracellular transport of class I MHC molecules in antigen processing mutant cell lines. J. Immunol. 151:3407. 6. Schumacher, T. N. M., M.-T. Heemels, J. J. Neefjes, W. M. Kast, C. J. M. Melief, and H. L. Ploegh. 1990. Direct binding of peptide to empty MHC class I molecules on intact cells and in vitro. Cell 62:563. 7. Cerundolo, V., T. Elliott, J. Elvin, J. Bastin, H.-G. Rammensee, and A. Townsend. 1991. The binding affinity and dissociation rates of peptides for class I major histocompatibility complex molecules. Eur. J. Immunol. 21:2069. 8. Ortiz-Navarrete, V., and G. J. Ha¨mmerling. 1991. Surface appearance and instability of empty H-2 class I molecules under physiological conditions. Proc. Natl. Acad. Sci. USA 88:3594. 9. Esquivel, F., J. Yewdell, and J. Bennink. 1992. RMA/S cells present endogenously synthesized cytosolic proteins to class I-restricted cytotoxic T lymphocytes. J. Exp. Med. 175:163. 10. Hosken, N. A., and M. J. Bevan. 1992. An endogenous antigenic peptide bypasses the class I antigen presentation defect in RMA-S. J. Exp. Med. 175:719. 11. Zhou, X., R. Glass, T. Liu, H.-G. Ljunggren, and M. Jondal. 1993. Antigen processing mutant T2 cells present viral antigen restricted through H-2Kd. Eur. J. Immunol. 23:1802. 12. Zhou, X., R. Glass, F. Momburg, G. J. Ha¨mmerling, M. Jondal, and H.-G. Ljunggren. 1993. TAP2-defective RMA-S cells present Sendai virus antigen to cytotoxic T lymphocytes. Eur. J. Immunol. 23:1796. 13. Khanna, R., S. R. Burrows, D. J. Moss, and S. L. Silins. 1996. Peptide transporter (TAP-1 and TAP-2)-independent endogenous processing of Epstein-Barr virus (EBV) latent membrane protein 2A: implications for cytotoxic T-lymphocyte control of EBV-associated malignancies. J. Virol. 70:5357. 14. De la Salle, H., E. Houssaint, M. A. Peyrat, D. Arnold, J. Salamero, D. Pinczon, S. Stevanovic, H. Bausinger, D. Fricker, E. Gomard, W. Biddison, P. Lehner, F. UytdeHaag, M. Sasportes, L. Donato, H. G. Rammensee, J. P. Cazenave, D. Hanau, M. M. Tongio, and M. Bonneville. 1997. Human peptide transporter deficiency: importance of HLA-B in the presentation of TAP-independent EBV antigens. J. Immunol. 158:4555. 15. Zweerink, H. J., M. C. Gammon, U. Utz, S. Y. Sauma, T. Harrer, J. C. Hawkins, R. P. Johnson, A. Sirotina, J. D. Hermes, B. D. Walker, and W. E. Biddison. 1993. Presentation of endogenous peptides to MHC class I-restricted cytotoxic T lymphocytes in transport deletion mutant T2 cells. J. Immunol. 150:1763. 16. Bacik, I., J. H. Cox, R. Anderson, J. W. Yewdell, and J. R. Bennink. 1994. TAP (transporter associated with antigen processing)-independent presentation of endogenously synthesized peptides is enhanced by endoplasmic reticulum insertion sequences located at the amino- but not carboxyl-terminus of the peptide. J. Immunol. 152:381. 17. Henderson, R. A., H. Michel, K. Sakaguchi, J. Shabanowitz, E. Appella, D. F. Hunt, and V. H. Engelhard. 1992. HLA-A2.1-associated peptides from a mutant cell line: a second pathway of antigen presentation. Science 255:1264.

4312


18. Wei, M. L., and P. Cresswell. 1992. HLA-A2 molecules in an antigen-processing mutant cell contain signal sequence-derived peptides. Nature 356:443. 19. Smith, K. D., and C. T. Lutz. 1996. Peptide-dependent expression of HLA-B7 on antigen processing deficient T2 cells. J. Immunol. 156:3755. 20. Sherman, L. A., and S. Chattopadhyay. 1993. The molecular basis of allorecognition. Annu. Rev. Immunol. 11:385. 21. Udaka, K., T. J. Tsomides, and H. N. Eisen. 1992. A naturally occurring peptide recognized by alloreactive CD81 cytotoxic T lymphocytes in association with class I MHC protein. Cell 69:989. 22. Aldrich, C. J., A. DeCloux, A. S. Woods, R. J. Cotter, M. J. Soloski, and J. Forman. 1994. Identification of a Tap-dependent leader peptide recognized by alloreactive T cells specific for a class Ib antigen. Cell 79:649. 23. Malarkannan, S., F. Gonzalez, V. Nguyen, G. Adair, and N. Shastri. 1996. Alloreactive CD81 T cells can recognize unusual, rare, and unique processed peptide/MHC complexes. J. Immunol. 157:4464. 24. Wang, W., P. H. Gulden, R. A. Pierce, J. A. Shabanowitz, S. T. Man, D. F. Hunt, and V. H. Engelhard. 1997. A naturally processed peptide presented by HLAA*0201 is expressed at low abundance and recognized by an alloreactive CD81 cytotoxic T cell with apparent high affinity. J. Immunol. 158:5797. ¨ hle´n, C., J. Bastin, H.-G. Ljunggren, L. Foster, E. Wolpert, G. Klein, 25. O A. R. M. Townsend, and K. M. Ka¨rre. 1990. Resistance to H-2-restricted but not to allo-H2-specific graft and cytotoxic T lymphocyte responses in lymphoma mutant. J. Immunol. 145:52. 26. Smith, P. A., A. Brunmark, M. R. Jackson, and T. A. Potter. 1997. Peptideindependent recognition by alloreactive cytotoxic T lymphocytes (CTL). J. Exp. Med. 185:1023. 27. Smith, K. D., D. F. Epperson, and C. T. Lutz. 1995. Alloreactive cytotoxic Tlymphocyte-defined HLA-B7 subtypes differ in peptide antigen presentation. Immunogenetics 43:27. 28. Shi, Y., K. D. Smith, M. G. Kurilla, and C. T. Lutz. 1997. Cytotoxic CD81 T cells recognize EBV antigen but poorly kill autologous EBV-infected B lymphoblasts: immunodominance is elicited by a peptide epitope that is presented at low levels in vitro. J. Immunol. 159:1844. 29. Smith, K. D., E. Huczko, V. H. Engelhard, Y.-Y. Li, and C. T. Lutz. 1997. Alloreactive cytotoxic T lymphocytes focus on specific major histocompatibility complex-bound peptides. Transplantation 64:351. 30. Smith, K. D., and C. T. Lutz. 1997. Alloreactive T cell recognition of MHC class I molecules: the T cell receptor interacts with limited regions of the MHC class I long a-helices. J. Immunol. 158:2805. 31. Kurago, Z. B., K. D. Smith, and C. T. Lutz. 1995. NK cell recognition of MHC class I: NK cells are sensitive to peptide-binding groove and surface a-helical mutations that affect T cells. J. Immunol. 154:2631. 32. Salter, R. D., A. M. Norment, B. P. Chen, C. Clayberger, A. M. Krensky, D. R. Littman, and P. Parham. 1989. Polymorphism in the a3 domain of HLA-A molecules affects binding to CD8. Nature 338:345. 33. Parnes, J. R. 1989. Molecular biology and function of CD4 and CD8. Adv. Immunol. 44:265. 34. Al-Ramadi, B. K., M. T. Jelonek, L. F. Boyd, D. H. Margulies, and A. L. M. Bothwell. 1995. Lack of strict correlation of functional sensitization with the apparent affinity of MHC/peptide complexes for the TCR. J. Immunol. 155:662. 35. Smith, K. D., B. E. Mace, A. Valenzuela, J. L. Vigna, J. A. McCutcheon, J. A. Barbosa, E. Huczko, V. H. Engelhard, and C. T. Lutz. 1996. Probing

36.

37. 38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

HLA-B7 conformational shifts induced by peptide-binding groove mutations and bound peptide with anti-HLA monoclonal antibodies. J. Immunol. 157:2470. Berke, G. 1993. The functions and mechanisms of action of cytolytic lymphocytes. In Fundamental Immunology, W. E. Paul, ed. Raven Press, Ltd., New York, p. 629. Yewdell, J. W., and J. R. Bennink. 1989. Brefeldin A specifically inhibits presentation of protein antigens to cytotoxic T lymphocytes. Science 244:1072. Huczko, E. L., W. M. Bodnar, D. Benjamin, K. Sakaguchi, N. Z. Zhu, J. Shabanowitz, R. A. Henderson, E. Appella, D. F. Hunt, and V. H. Engelhard. 1993. Characteristics of endogenous peptides eluted from the class I MHC molecule HLA-B7 determined by mass spectrometry and computer modeling. J. Immunol. 151:2572. Mullbacher, A., A. B. Hill, R. V. Blanden, W. B. Cowden, N. J. C. King, and R. T. Hla. 1991. Alloreactive cytotoxic T cells recognize MHC class I antigen without peptide specificity. J. Immunol. 147:1765. Bluestone, J. A., A. Kaliyaperumal, S. Jameson, S. Miller, and R. Dick II. 1993. Peptide-induced conformational changes in class I heavy chains alters allorecognition. J. Immunol. 151:3943. Ro¨tzschke, O., K. Falk, S. Faath, and H.-G. Rammensee. 1991. On the nature of peptides involved in T cell alloreactivity. J. Exp. Med. 174:1059. Crumpacker, D. B., J. Alexander, P. Cresswell, and V. H. M. Engelhard. 1992. Role of endogenous peptides in murine allogeneic cytotoxic T cell responses assessed using transfectants of the antigen-processing mutant 1743CEM. T2. J. Immunol. 148:3004. Sherman, L. A., T. A. Burke, and J. A. Biggs. 1992. Extracellular processing of peptide antigens that bind class I major histocompatibility molecules. J. Exp. Med. 175:1221. Pfeifer, J. D., M. J. Wick, R. L. Roberts, K. Findlay, S. J. Normark, and C. V. Harding. 1993. Phagocytic processing of bacterial antigens for class I MHC presentation to T cells. Nature 361:359. Liu, T., B. Chambers, A. D. Diehl, L. Van Kaer, M. Jondal, and H.-G. Ljunggren. 1997. TAP peptide transporter-independent presentation of heat-killed Sendai virus antigen on MHC class I molecules by splenic antigen-presenting cells. J. Immunol. 159:5364. Gabathuler, R., G. Reid, G. Kolaitis, J. Driscoll, and W. A. Jefferies. 1994. Comparison of cell lines deficient in antigen presentation reveals a functional role for TAP-1 alone in antigen processing. J. Exp. Med. 180:1415. Hombach, J., H. Pircher, S. Tonegawa, and R. M. Zinkernagel. 1995. Strictly transporter of antigen presentation (TAP)-dependent presentation of an immunodominant cytotoxic T lymphocyte epitope in the signal sequence of a virus protein. J. Exp. Med. 182:1615. Momburg, F., J. Roelse, G. J. Ha¨mmerling, and J. J. Neefjes. 1994. Peptide size selection by the major histocompatibility complex-encoded peptide transporter. J. Exp. Med. 179:1613. Killion, C. C., P.-J. Chen, J. M. Dadgari, and M. McMillan. 1995. Dissection of cross-reactivities using a panel of H-2Ld alloreactive T cell hybridomas. Cell. Immunol. 164:81. Sykulev, Y., M. Joo, I. Vturina, T. J. Tsomides, and H. N. Eisen. 1996. Evidence that a single peptide-MHC complex on a target cell can elicit a cytolytic T cell response. Immunity 4:565. Elliot, T. J., and H. N. Eisen. 1990. Cytotoxic T lymphocytes recognize a reconstituted class I histocompatibility antigen (HLA-A2) as an allogeneic target molecule. Proc. Natl. Acad. Sci. USA 87:5213.