Portal de Periódicos da CAPES

A Dominant Negative Mutant β2-Microglobulin Blocks the Extracellular Folding of a Major Histocompatibility Complex Class I Heavy Chain

2003; Elsevier BV; Volume: 278; Issue: 8 Linguagem: Inglês

10.1074/jbc.m208381200

1083-351X

Dawn M. Hill, Tina Kasliwal, E. J. Schwarz, Andrea Hebert, Trina Chen, Elena Gubina, Lei Zhang, Steven Kozlowski,

Immunotherapy and Immune Responses

The major histocompatibility complex class I (MHC1) molecule plays a crucial role in cytotoxic lymphocyte function. β2-Microglobulin (β2m) has been demonstrated to be both a structural component of the MHC1 complex and a chaperone-like molecule for MHC1 folding. β2m binding to an isolated α3 domain of MHC1 heavy chain at micromolar concentrations has been shown to accurately model the biochemistry and thermodynamics of β2m-driven MHC1 folding. These results suggested a model in which the chaperone-like role of β2m is dependent on initial binding to the α3 domain interface of MHC1 with β2m. Such a model predicts that a mutant β2m molecule with an intact MHC1 α3 domain interaction but a defective MHC1 α1α2 domain interaction would block β2m-driven folding of MHC1. In this study we generated such a β2m mutant and demonstrated that it blocks MHC1 folding by normal β2m at the expected micromolar concentrations. Our data support an initial interaction of β2m with the MHC1 α3 domain in MHC1 folding. In addition, the dominant negative mutant β2m can block T-cell functional responses to antigenic peptide and MHC1. The major histocompatibility complex class I (MHC1) molecule plays a crucial role in cytotoxic lymphocyte function. β2-Microglobulin (β2m) has been demonstrated to be both a structural component of the MHC1 complex and a chaperone-like molecule for MHC1 folding. β2m binding to an isolated α3 domain of MHC1 heavy chain at micromolar concentrations has been shown to accurately model the biochemistry and thermodynamics of β2m-driven MHC1 folding. These results suggested a model in which the chaperone-like role of β2m is dependent on initial binding to the α3 domain interface of MHC1 with β2m. Such a model predicts that a mutant β2m molecule with an intact MHC1 α3 domain interaction but a defective MHC1 α1α2 domain interaction would block β2m-driven folding of MHC1. In this study we generated such a β2m mutant and demonstrated that it blocks MHC1 folding by normal β2m at the expected micromolar concentrations. Our data support an initial interaction of β2m with the MHC1 α3 domain in MHC1 folding. In addition, the dominant negative mutant β2m can block T-cell functional responses to antigenic peptide and MHC1. The major histocompatibility complex class I (MHC1) 1The abbreviations used are: MHC1, major histocompatibility class I molecules; β2m, β2-microglobulin; CTL, cytotoxic T-lymphocyte; TCR, T-cell receptor; hβ2m, human β2-microglobulin; APC, antigen-presenting cells; SPR, surface plasmon resonance; RU, response units; ER, endoplasmic reticulum; PBS, phosphate-buffered saline; BSA, bovine serum albumin; IL, interleukin molecule and antigenic peptide are recognized by CD8+ cytotoxic T-lymphocytes (CTL) in CTL activation and lysis of targets (1Townsend A. Bodmer H. Annu. Rev. Immunol. 1989; 7: 601-624Google Scholar). The heavy chain of the MHC1 molecule can interact noncovalently with a number of other molecules in the formation of a CTL activating complex. These include the MHC1 light chain or β2m, the antigenic peptide fragment, the T-cell receptor (TCR), and the CD8 molecule (2Jones E.Y. Tormo J. Reid S.W. Stuart D.I. Immunol. Rev. 1998; 163: 121-128Google Scholar). The specificity of the CTL response resides in the selective MHC1 binding of specific antigenic peptide fragments and in the TCR recognition of these antigenic peptides and MHC1 (3Germain R.N. Margulies D.H. Annu. Rev. Immunol. 1993; 11: 403-450Google Scholar, 4Ljunggren H.G. Thorpe C.J. Histol. Histopathol. 1996; 11: 267-274Google Scholar). The MHC1 contact surface for TCR and peptide binding is formed by the α1 and α2 domains of the three-domain MHC1 heavy chain (2Jones E.Y. Tormo J. Reid S.W. Stuart D.I. Immunol. Rev. 1998; 163: 121-128Google Scholar, 5Ozato K. Evans G.A. Shykind B. Margulies D.H. Seidman J.G. Proc. Natl. Acad. Sci. U. S. A. 1983; 80: 2040-2043Google Scholar, 6Bjorkman P.J. Saper M.A. Samraoui B. Bennett W.S. Strominger J.L. Wiley D.C. Nature. 1987; 329: 512-518Google Scholar, 7Davis M.M. Boniface J.J. Reich Z. Lyons D. Hampl J. Arden B. Chien Y. Annu. Rev. Immunol. 1998; 16: 523-544Google Scholar). The MHC1 heavy chain α1 and α2 domains, as well as the immunoglobulin-like α3 domain, have been shown by x-ray crystallography (8Saper M.A. Bjorkman P.J. Wiley D.C. J. Mol. Biol. 1991; 219: 277-319Google Scholar) to interact with β2m, the nonpolymorphic component of the MHC1 complex. Mutations in the α1 (9Santos-Aguado J. Biro P.A. Fuhrmann U. Strominger J.L. Barbosa J.A. Mol. Cell. Biol. 1987; 7: 982-990Google Scholar, 10Ribaudo R.K. Margulies D.H. J. Immunol. 1995; 155: 3481-3493Google Scholar) or α3 (11Salter R.D. Hum. Immunol. 1992; 35: 40-49Google Scholar) domains of the MHC1 heavy chain lead to changes in β2m binding. These studies demonstrate that the functional interaction of the MHC1 heavy chain with β2m occurs at multiple surfaces on different domains. In the absence of β2m, most MHC1 molecules are not expressed efficiently on the surface of cells (12Seong R.H. Clayberger C.A. Krensky A.M. Parnes J.R. J. Exp. Med. 1988; 167: 288-299Google Scholar, 13Zijlstra M. Bix M. Simister N.E. Loring J.M. Raulet D.H. Jaenisch R. Nature. 1990; 344: 742-746Google Scholar). Although some MHC1 molecules, such as murine H-2Ld and H-2Db, are transported to the cell surface without β2m, they have diminished levels of expression (14Allen H. Fraser J. Flyer D. Calvin S. Flavell R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 7447-7451Google Scholar, 15Bix M. Raulet D. J. Exp. Med. 1992; 176: 829-834Google Scholar). This decreased MHC1 expression is not simply because of an export requirement for fully assembled MHC1 complexes. Transfection of β2m-negative cells with ER-retained β2m was able to salvage MHC1 cell surface expression (16Solheim J.C. Johnson N.A. Carreno B.M. Lie W.R. Hansen T.H. Eur. J. Immunol. 1995; 25: 3011-3016Google Scholar). MHC1 folded in the presence of this ER-retained β2m was exported to the cell surface without bound β2m. Thus β2m, which promotes protein folding through a transient interaction, fits the definition of a chaperone (17Fink A.L. Physiol. Rev. 1999; 79: 425-449Google Scholar). Therefore, β2m plays two roles in MHC1, first, as a structural subunit of the assembled complex and second, as a chaperone for the folding of the MHC1 heavy chain. A possible mechanism for β2m as a chaperone is facilitation of the interaction of MHC1 heavy chain with other chaperones, such as calreticulin, tapasin, transporter associated with antigen processing, and Erp57 (18Solheim J.C. Immunol. Rev. 1999; 172: 11-19Google Scholar, 19Cresswell P. Bangia N. Dick T. Diedrich G. Immunol. Rev. 1999; 172: 21-28Google Scholar). However, because β2m has been shown to promote stabilization or folding of MHC1 on the cell surface in the absence of ER chaperones (20Rock K.L. Gramm C. Benacerraf B. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 4200-4204Google Scholar, 21Otten G.R. Bikoff E. Ribaudo R.K. Kozlowski S. Margulies D.H. Germain R.N. J. Immunol. 1992; 148: 3723-3732Google Scholar), it is likely that β2m also has a direct effect on MHC1 folding. Although high concentrations of high affinity peptides can promote the folding of MHC1 in the absence of β2m (15Bix M. Raulet D. J. Exp. Med. 1992; 176: 829-834Google Scholar), these same peptides can stabilize MHC1 folded with β2m at significantly lower concentrations (22Elliott T. Cerundolo V. Elvin J. Townsend A. Nature. 1991; 351: 402-406Google Scholar, 23Elliott T. Elvin J. Cerundolo V. Allen H. Townsend A. Eur. J. Immunol. 1992; 22: 2085-2091Google Scholar). Therefore, with physiologic concentrations of high affinity peptides or any concentration of lower affinity peptides, β2m levels are limiting for the folding of MHC1 molecules. The two roles of β2m, as structural subunit and chaperone, do not depend equally on β2m concentration. β2m binds to MHC1 heavy chain with an equilibrium dissociation constant (K d) in the nanomolar range (24Hyafil F. Strominger J.L. Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 5834-5838Google Scholar, 25Parker K.C. Strominger J.L. Biochemistry. 1985; 24: 5543-5550Google Scholar, 26Hochman J.H. Shimizu Y. DeMars R. Edidin M. J. Immunol. 1988; 140: 2322-2329Google Scholar, 27Pedersen L.O. Hansen A.S. Olsen A.C. Gerwien J. Nissen M.H. Buus S. Scand. J. Immunol. 1994; 39: 64-72Google Scholar) while it folds or stabilizes cell surface MHC1 at micromolar concentrations (20Rock K.L. Gramm C. Benacerraf B. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 4200-4204Google Scholar, 21Otten G.R. Bikoff E. Ribaudo R.K. Kozlowski S. Margulies D.H. Germain R.N. J. Immunol. 1992; 148: 3723-3732Google Scholar, 28Shields M.J. Moffat L.E. Ribaudo R.K. Mol. Immunol. 1998; 35: 919-928Google Scholar). We have demonstrated previously (29Whitman M.C. Strohmaier J. O'Boyle K. Tingem J.M. Wilkinson Y. Goldstein J. Chen T. Brorson K. Brunswick M. Kozlowski S. Mol. Immunol. 2000; 37: 141-149Google Scholar) that human β2m (hβ2m) binds the isolated α3 domain of the MHC1 heavy chain with a K d in that same micromolar range and this binding has the same species dependence and thermodynamics as the β2m-driven refolding of the MHC1 heavy chain (30Hebert A.M. Strohmaier J. Whitman M.C. Chen T. Gubina E. Hill D.M. Lewis M.S. Kozlowski S. Biochemistry. 2001; 40: 5233-5242Google Scholar). This suggested that β2m folding of the complete MHC1 heavy chain may be nucleated by a β2m-α3 interaction. Although the biochemical characteristics of folding match those of the predicted limiting initial β2m-α3 interaction (30Hebert A.M. Strohmaier J. Whitman M.C. Chen T. Gubina E. Hill D.M. Lewis M.S. Kozlowski S. Biochemistry. 2001; 40: 5233-5242Google Scholar), it is formally possible that the similar binding characteristics of β2m-α3 binding and β2m folding of MHC1 are coincidental. To resolve this issue, we have generated β2m mutants with predicted defects in interactions with MHC1 α1/α2 domains and evaluated the mutant β2m effects on native β2m-driven folding of MHC1. Because the β2m-driven folding of the α1 and α2 domains is likely to be dependent on the β2m-α1/α2 interaction, many possible models would predict poor folding of MHC1 by these mutants. However, if the initial limiting interaction of β2m were with the α3 domain, these β2m mutants would still have an intact initial interaction and be able to compete with native β2m for this initial interaction. Thus β2m mutants with diminished α1/α2 interactions would be competitive inhibitors of native β2m-driven MHC1 folding. Although such a mutant β2m protein would be predicted to inhibit extracellular MHC1 folding, its design and mode of inhibition are similar to dominant negative mutations used in the study of intracellular signaling and cytoskeletal structures (31Herskowitz I. Nature. 1987; 329: 219-222Google Scholar, 32Perlmutter R.M. Alberola-Ila J. Curr. Opin. Immunol. 1996; 8: 285-290Google Scholar). It would be expected that this dominant negative effect on MHC1 folding would have a concentration dependence similar to that of the β2m-α3 interaction. The H-2Dd α3 domain sequence was generated by PCR amplification of an H-Dd cDNA as described previously (29Whitman M.C. Strohmaier J. O'Boyle K. Tingem J.M. Wilkinson Y. Goldstein J. Chen T. Brorson K. Brunswick M. Kozlowski S. Mol. Immunol. 2000; 37: 141-149Google Scholar) with the upper primer ACTCCATGGCAACAGATCCCCCAAAGGCCC and the lower primer GATGAATTCGACCCGGAAGGAGGAGGTTC. The α3 domain sequence was inserted between the NcoI andEcoR1 sites of a modified pET21d vector (Novagen, Madison, WI). The vector was modified by ligating a synthetic oligonucleotide (GAGGAATTCTGGAATTTCGCAAGCTGTACATGCTGCACACGCTGAAATTAACGAAGCAGGAAGAGCACTCGAGCAC) between the EcoR1 and XhoI sites of the pET21d bacterial expression vector. The completed construct had the correct sequence and when transfected and induced, generated a 15-kDa fusion protein consisting of the H-Dd α3 domain fused to vector expressed sequence, a 17-amino acid peptide sequence from ovalbumin, and a polyhistidine tag. The expression construct was transfected into BL21(DE3) bacteria (Novagen, Madison, WI). 400-ml cultures of transfected bacteria in LB broth with 200 μg/ml carbenicillin (Sigma) were grown to an absorbance of 0.6 at 600 nm. The cultures were then induced with 1 mm isopropyl-1-thio-β-d-galactopyranoside (Sigma), and the cells were harvested by centrifugation after overnight induction at 28 °C. The bacteria were washed with PBS, and the pellet was frozen at −70 °C. The frozen pellet was thawed in 0.5m NaCl, 10 mm Tris, pH 8.0, with 1 mg/ml lysozyme (Sigma). After addition of imidazole (Sigma) to a concentration of 5 mm and Triton X-100 (Roche Molecular Biochemicals) to a concentration of 1%, the bacteria were sonicated in a Brinkmann homogenizer (Brinkmann, Westbury, CT) for 3 × 30 s at a setting of 4. The homogenate was treated with ∼500 units of Benzonase (Sigma) in the presence of 5 mm MgCl. Inclusion bodies were pelleted by spinning at 15,000 × g, and the soluble fraction was loaded on to buffer-equilibrated nickel-nitrilotriacetic acid resin (Qiagen) for 1 h at 4 °C. The loaded nickel-nitrilotriacetic acid resin was washed three times with 0.5 m NaCl, 10 mm Tris, pH 8.0, 1% Triton X 100, 5 mm imidazole buffer and then three times with 0.5m NaCl, 10 mm Tris, pH 8.0, buffer. The fusion protein was eluted with high concentration imidazole (150 to 500 mm). The imidazole was removed by dialysis, and the protein was further purified by size exclusion chromatography. The protein concentration was measured by 280-nm absorbance. The extinction coefficient at 280 nm was calculated from the primary amino acid sequence (33Gill S.C. von Hippel P.H. Anal. Biochem. 1989; 182: 319-326Google Scholar). The recombinant native hβ2m was expressed using a construct in a PET21d vector generously supplied by Randall K. Ribaudo (34Shields M.J. Assefi N. Hodgson W. Kim E.J. Ribaudo R.K. J. Immunol. 1998; 160: 2297-2307Google Scholar, 35Shields M.J. Kubota R. Hodgson W. Jacobson S. Biddison W.E. Ribaudo R.K. J. Biol. Chem. 1998; 273: 28010-28018Google Scholar). Mutant hβ2m genes were generated by PCR with the Pfu-1 polymerase (Stratagene) using splice overlap extension (36Horton R.M. Cai Z.L. Ho S.N. Pease L.R. Biotechniques. 1990; 8: 528-535Google Scholar) with mutant oligonucleotides. The D53K mutation was generated with the upper GGAGCATTCAAAATTGTCTTTCA oligonucleotide primer (the base pairs in mutated codons are underlined) and a complementary lower primer. The D53R mutation was generated with the upper GGAGCATTCAAGATTGTCTTTCA oligonucleotide primer and a complementary lower primer. The W60A mutation was generated with the upper ATAGAAAGACGCGTCCTTGCT oligonucleotide primer and a complementary lower primer. The hβ2m upper GACGGAGCTCGAATTCGGATC primer and hβ2m lower AGGAGATATATCATGATCCAGCGT primer were used with the corresponding mutant oligos to amplify the mutated gene fragments in the first amplification step and for generating an intact gene the second amplification step. The intact gene fragments were digested withBspHI and BamHI and ligated into PET21d vector that was digested with NcoI and BamHI. Double mutants such as W60A/D53K were sequentially generated using the same process. The mutant genes were verified by sequencing. The constructs were transfected into BL21(DE3) bacteria (Novagen). 100-ml cultures of transfected bacteria in LB broth with 250 μg/ml carbenicillin with were grown at 37 °C to an absorbance of 0.8 at 600 nm and then induced with 1 mmisopropyl-1-thio-β-d-galactopyranoside for 2 h. The bacteria were harvested by centrifugation and washed and resuspended with 0.1 m Tris, pH 8.0, with 2 mmEDTA. Lysozyme (Sigma) was added at 0.5 mg/ml, and the bacteria were incubated overnight at 4 °C. Deoxycholate was added to a final concentration of 0.1%, and the mixture was sonicated on a Brinkmann homogenizer for 4 × 30 s. The inclusion bodies were pelleted by centrifugation at 15,000 × g, and the soluble fraction was discarded. Inclusion bodies were washed three times with 0.1 m Tris, pH 8.0, with 2 mm EDTA and 0.1% deoxycholate followed by a wash with 0.1 m Tris, pH 8.0, 2 mm EDTA. The pellet was resuspended with 2 ml of 6m guanidine HCl, 0.1 mm dithiothreitol, 0.1 m Tris, pH 8.0, 2 mm EDTA. The dissolved protein was added to 50 ml of pre-chilled 0.4 m arginine, 0.1 m Tris, 2 mm EDTA, 5 mmoxidized glutathione, 0.5 mm reduced glutathione and left to refold in the cold room with gentle agitation for a minimum of 3 days. The protein was then dialyzed against 1 liter of HEPES-buffered saline, 3 mm EDTA three times and then concentrated using an Amicon ultrafiltration cell and MicroSep 3K omega filters (Pall Corporation, Ann Arbor, MI). The concentrated β2m was then purified by size exclusion chromatography. Both native and mutant β2m were purified in the same manner. The protein concentrations were measured by 280-nm absorbance. The extinction coefficient at 280 nm was calculated from the primary amino acid sequence (33Gill S.C. von Hippel P.H. Anal. Biochem. 1989; 182: 319-326Google Scholar). For β2m and for H-2Dd α3 domain purification, 500–1000 μl of protein was loaded onto a Superdex 75 column (Pharmacia LKB, AmershamBiosciences, Uppsala, Sweden) using the P-500 pump (Pharmacia LKB) at 0.5 ml/min buffer. Half-ml fractions were collected, and the protein concentration was determined by 280-nm absorbance. Protein used for surface plasmon resonance experiments was purified in HBS-EP buffer (HEPES-buffered saline with 3 mm EDTA and 0.005% polysorbate 20) or PBS. Proteins used in flow cytometry and functional experiments were purified in PBS. Anti-tetra His antibody was purchased from Qiagen. An anti-H-2Dd antibody (34.5.8S), an anti-H-2Kd antibody (SF1–1.1), and an IgG2aκ control antibody were purchased from Pharmingen. The LKD8 cell line, a peptide transport-deficient EE2H3 embryonic cell line transfected with H-2Dd (37Bikoff E.K. Otten G.R. Robertson E.J. Eur. J. Immunol. 1991; 21: 1997-2004Google Scholar), was the generous gift of David H. Margulies (NIH, Bethesda, MD). RMAs-Kd, a peptide transport-deficient cell line transfected with H-2Kd (38Chen W. Yewdell J.W. Levine R.L. Bennink J.R. J. Exp. Med. 1999; 189: 1757-1764Google Scholar), was generously provided by Jonathan Yewdell (NIH, Bethesda, MD). The B4.2.3 T-cell hybridoma (reactive with gp160 p18-I10 in the context of H-2Dd) and H-2Dd L-cell transfectant cell lines were also used (39Kozlowski S. Takeshita T. Boehncke W.H. Takahashi H. Boyd L.F. Germain R.N. Berzofsky J.A. Margulies D.H. Nature. 1991; 349: 74-77Google Scholar). The H-2Dd-expressing gp160 transfectant 3T3 cell line (40Takahashi H. Cohen J. Hosmalin A. Cease K.B. Houghten R. Cornette J.L. DeLisi C. Moss B. Germain R.N. Berzofsky J.A. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 3105-3109Google Scholar), 15–12, and its control cell line, 18neo, were generously provided by Jay A. Berzofsky (NIH, Bethesda, MD). The p18-I10 peptide (RGPGRAFVTI) (41Kozlowski S. Corr M. Takeshita T. Boyd L.F. Pendleton C.D. Germain R.N. Berzofsky J.A. Margulies D.H. J. Exp. Med. 1992; 175: 1417-1422Google Scholar, 42Takeshita T. Takahashi H. Kozlowski S. Ahlers J.D. Pendleton C.D. Moore R.L. Nakagawa Y. Yokomuro K. Fox B.S. Margulies D.H. et al.J. Immunol. 1995; 154: 1973-1986Google Scholar) was obtained from the Center for Biologics Evaluation and Research Facility for Biotechnology Resources (Bethesda, MD). The peptide was synthesized on an ABI 433 peptide synthesizer (Applied Biosystems, Foster City, CA) and characterized by matrix-assisted laser desorption ionization time-of-flight mass spectrometry analysis (Voyager; Applied Biosystems). All SPR experiments were performed on the BIAcore 3000 biosensor (BIAcore AB, Uppsala, Sweden). Anti-His antibody, diluted in 10 mm acetate buffer, pH 4.5, was covalently coupled to the carboxymethylated dextran matrix on a CM5 sensor chip (Biacore AB) by using the amine coupling kit as described previously (43Johnsson B. Lofas S. Lindquist G. Anal. Biochem. 1991; 198: 268-277Google Scholar). Experiments were performed in HBS-EP buffer, and regeneration of the anti-His surface was achieved with 20 mm HCl. Equilibrium binding data for β2m were obtained by averaging a 5–10-s interval of normalized signal after reaching equilibrium. The normalized signal was obtained by subtracting the control surface signal from α3 surface signal. The equilibrium binding data was analyzed by nonlinear curve fitting of the Langmuir isotherm to the data. The curve fitting was performed using the BIAevaluation 3.0 software (BIAcore AB). β2m and β2m mutants were titrated at the indicated concentrations in 24- or 12-well tissue culture plates containing 1 or 2 ml of OPTI-MEM medium (Invitrogen) with 0.5% BSA (Sigma) and 0.5 × 106LKD8 cells (for H-2Dd folding) or 0.5 × 106 RMAs-Kd cells (for H-2Kdfolding) per well. The assay was carried out in a 7.5% CO2incubator at 26–28 °C for low temperature-induced folding or at 37 °C, with the indicated concentration of peptide, for peptide-induced folding. After an overnight incubation, the cells were spun down at 4 °C. One-half μg of the biotinylated anti-H-2Dd antibody, 34.5.8S (Pharmingen), or biotinylated anti-H-2Kd antibody, SF1–1.1 (Pharmingen), was added to the cells. Cells receiving no primary antibody or a biotinylated isotype-matched antibody (Pharmingen) were used as controls. After an hour of incubation on ice, the cells were washed with 0.5% BSA/OPTI-MEM medium and then 50 μl of streptavidin-fluorescein isothiocyanate (Pharmingen), diluted 1:50 in 0.5% BSA/OPTI-MEM, was added to the cells for an additional 30-min incubation on ice. The cells were then washed with 0.5% BSA/OPTI-MEM and analyzed on a FACScalibur (BD Biosciences) flow cytometer. The cells were size-gated, and generally 2,500 to 10,000 cells were counted for each data point. Data analysis was performed using Cell Quest Software (BD Biosciences). β2m and β2m mutants were added at the indicated concentrations to wells in 96-well round bottom tissue culture plates with 3.6 × 105 of the APC. The p18-I10 peptide was added as indicated to wells. The final volume was 200 μl/well of three parts 5% BSA/OPTI-MEM and one part PBS. After an overnight incubation at 37 °C in 7.5% CO2, the cells were washed two times with 5% BSA/OPTI-MEM and resuspended in 150 μl of Dulbecco's modified Eagle's medium with 10% fetal calf serum, 2 mml-glutamine, nonessential amino acids, penicillin/streptomycin (100 units/ml penicillin), and 5 × 10−5m β-mercaptoethanol (complete medium). Fifty μl of washed APC were titrated by threes into 100 μl of complete medium in 96-well flat bottom tissue culture plates as indicated. B4.2.3 T-cells were added at the indicated concentrations in 50 μl of complete medium, and the plates were incubated overnight at 37 °C in 7.5% CO2. For IL2 cytokine determinations, 25 μl of culture supernatants were assayed with commercially available kits (Endogen, Boston, MA) according to the manufacturer's specifications. Horseradish peroxidase-conjugated streptavidin (Zymed Laboratories Inc., San Francisco, CA) and tetramethylbenzidine (Dako, Carpinteria, CA) were used as developers. The absorbance was read on a Bio-Rad model 3550 microplate reader (Hercules, CA) at 655 nm with background subtraction. Cytokine standards were run with each experiment. A linear fit of these IL2 standard values was used to extrapolate the scales for IL2 levels. In some experiments the cells were pulsed with 1 μCi/well of [3H]thymidine (PerkinElmer Life Sciences) for 3–6 h and harvested and counted to assess growth inhibition (44Ashwell J.D. Cunningham R.E. Noguchi P.D. Hernandez D. J. Exp. Med. 1987; 165: 173-194Google Scholar). Human and murine β2m are ∼70% homologous (34Shields M.J. Assefi N. Hodgson W. Kim E.J. Ribaudo R.K. J. Immunol. 1998; 160: 2297-2307Google Scholar), and critical β2m contact residues with MHC1 heavy chain are conserved (8Saper M.A. Bjorkman P.J. Wiley D.C. J. Mol. Biol. 1991; 219: 277-319Google Scholar). Human β2m is more effective than murine β2m at folding murine MHC1 heavy chain and at binding an MHC1 α3 domain (30Hebert A.M. Strohmaier J. Whitman M.C. Chen T. Gubina E. Hill D.M. Lewis M.S. Kozlowski S. Biochemistry. 2001; 40: 5233-5242Google Scholar). Because hβ2m has a higher affinity for an MHC1 α3 domain than murine β2m, we chose to mutate hβ2m. This higher affinity for MHC1 α3 would predictably lower the concentration necessary to observe a mutant dominant negative effect. We limited our mutation of β2m residues to those that interact with the α1 and α2 interface of the MHC1 heavy chain in MHC1 crystal structures. HLA-A2 and H-2Dd MHC1 crystal structures (8Saper M.A. Bjorkman P.J. Wiley D.C. J. Mol. Biol. 1991; 219: 277-319Google Scholar, 45Li H. Natarajan K. Malchiodi E.L. Margulies D.H. Mariuzza R.A. J. Mol. Biol. 1998; 283: 179-191Google Scholar) implicate the tryptophan at position 60 of β2m as a critical residue for multiple contacts between β2m and the α2 domain of the MHC1 heavy chain. Mutations of β2m at position 60 have also been shown to interfere with β2m exchange onto MHC1 (46Fukazawa T. Hermann E. Edidin M. Wen J. Huang F. Kellner H. Floege J. Farahmandian D. Williams K.M. Yu D.T. J. Immunol. 1994; 153: 3543-3550Google Scholar). To create a defect in β2m folding of MHC1, we mutated the tryptophan at β2m position 60 to an alanine residue, a non-conservative change. The HLA-A2 MHC1 crystal structure also predicts a number of interactions between the aspartate at position 53 of β2m and residues of the α1 domain of the MHC1 heavy chain. However, crystal structures of H-2Dd (45Li H. Natarajan K. Malchiodi E.L. Margulies D.H. Mariuzza R.A. J. Mol. Biol. 1998; 283: 179-191Google Scholar, 47Achour A. Persson K. Harris R.A. Sundback J. Sentman C.L. Lindqvist Y. Schneider G. Karre K. Immunity. 1998; 9: 199-208Google Scholar, 48Tormo J. Natarajan K. Margulies D.H. Mariuzza R.A. Nature. 1999; 402: 623-631Google Scholar, 49Shields M.J. Hodgson W. Ribaudo R.K. Mol. Immunol. 1999; 36: 561-573Google Scholar) suggest weaker interactions between the aspartate at position 53 of β2m and the α1 domain. Despite this, a previous study (34Shields M.J. Assefi N. Hodgson W. Kim E.J. Ribaudo R.K. J. Immunol. 1998; 160: 2297-2307Google Scholar) has demonstrated that changing the negatively charged aspartate residue at position 53 to a neutral valine decreased the ability of hβ2m to fold murine MHC1 molecules, including H-2Dd. Based on this study, we generated charge reversal mutants, changing position 53 of β2m to positively charged lysine or arginine, predicting that these mutants would have greater defects in MHC1 folding than the neutral D53V mutation. In addition, we generated a β2m mutant with non-conservative changes at both positions, 60 and 53. Based on amino acid sequence data, these residues are conserved across murine and human β2m molecules (8Saper M.A. Bjorkman P.J. Wiley D.C. J. Mol. Biol. 1991; 219: 277-319Google Scholar). The locations of the two sites we mutated in β2m, as related to an example MHC1 crystal structure (50Garboczi D.N. Madden D.R. Wiley D.C. J. Mol. Biol. 1994; 239: 581-587Google Scholar), are illustrated in Fig.1. Based on structural data, our β2m mutants at positions 60 and 53 are likely to be defective in their interactions with MHC1 α1/α2 and not MHC1 α3. However, to rule out an unexpected global effect of the mutations, we verified that the mutant β2m-α3 interactions were not compromised. We expressed the β2m proteins using constructs in bacterial expression vectors (34Shields M.J. Assefi N. Hodgson W. Kim E.J. Ribaudo R.K. J. Immunol. 1998; 160: 2297-2307Google Scholar, 35Shields M.J. Kubota R. Hodgson W. Jacobson S. Biddison W.E. Ribaudo R.K. J. Biol. Chem. 1998; 273: 28010-28018Google Scholar). Purified mutant β2m molecules were evaluated for binding to isolated MHC1 H-2Dd α3 domain molecules using a biosensor for SPR (30Hebert A.M. Strohmaier J. Whitman M.C. Chen T. Gubina E. Hill D.M. Lewis M.S. Kozlowski S. Biochemistry. 2001; 40: 5233-5242Google Scholar). Antibody to the polyhistidine tail of the α3 protein was directly coupled to the carboxymethylated dextran surface of a biosensor chip, allowing for the capture of α3 protein on this surface. β2m molecules were injected across the surface of the immobilized α3 protein, and the binding was assessed by mass-related changes in the sensor chip matrix refractive index and quantified as response units (RU). β2m was also injected across a control surface consisting of the directly coupled capture antibody without α3 protein. We injected a control protein, ovalbumin, and there was no detectable α3 domain binding (data not shown). This experimental system was sensitive to a 2-fold difference in affinity between human and mouse β2m (30Hebert A.M. Strohmaier J. Whitman M.C. Chen T. Gubina E. Hill D.M. Lewis M.S. Kozlowski S. Biochemistry. 2001; 40: 5233-5242Google Scholar). This allowed protein concentration effects on the refractive index response to be subtracted out. An example of such an SPR experiment with the W60A/D53K double mutant β2m is shown in Fig.2 A. The responses of the α3 domain and control surfaces were normalized to each other immediately prior to the injection of the β2m to allow comparison of the active and control surfaces in this figure. Increasing concentrations of β2m were then evaluated for binding to purified monomeric α3 domains. The binding curves after subtraction of the control surface are shown in Fig. 2 B. Equilibrium values taken from the subtracted data were used to generate the plot in Fig. 2 C. This plot was fit with the nonlinear Langmuir isotherm for calculation of the equilibrium binding constant. The equilibrium constants for α3 binding were generated in this manner from three experiments with each of the β2m proteins and are shown in TableI.