Functionally active recombinant alpha and beta chain-peptide complexes of human major histocompatibility class II molecules.

Major histocompatibility (MHC) class II molecules are cell surface heterodimeric (alphabeta) glycoproteins that display processed antigens to T cell receptors (TCRs) of CD4-positive T cells. The present study describes that individual recombinant alpha and beta chains of human MHC class II molecules lacking the transmembrane region (alpha-Tm and beta-Tm) are capable of binding antigenic peptide and that these complexes of chain-peptide are recognized by TCRs to induce antigen-specific apoptosis in restricted T cells. The alpha-Tm and the beta-Tm of human HLA-DR2 (DRB5*0101) were cloned, expressed in Escherichia coli, and purified in large scale by conventional chromatographic methods. The in vitro binding of an immunodominant epitope from the myelin basic protein (MBP-(83-102)Y83) to purified DR2 alpha-Tm and DR2 beta-Tm was demonstrated with biotinylated and fluoresceinated MBP-(83-102)Y83 peptide. The specificity of the MBP-(83-102)Y83 peptide binding to both DR2 alpha-Tm and DR2 beta-Tm was demonstrated in a competitive peptide binding assay. When exposed to a transformed T cell clone (SS8T) restricted to DR2(DRB5*0101) and MBP-(84-102) peptide, complexes of DR2 alpha-Tm and DR2 beta-Tm with MBP-(83-102)Y83 peptide were able to specifically recognize TCRs as measured by the increase in gamma-interferon (gamma-IFN) cytokine. Such recognition of TCRs by soluble alpha-MBP-(83-102)Y83 and beta-MBP-(83-102)Y83 complexes led to the induction of antigen-specific apoptosis in SS8T cells as measured by double fluorescence flow cytometry and electron microscopy. These results provide the first evidence that soluble complexes of antigenic peptide and individual chains of human MHC class II molecules lacking the transmembrane region can recognize TCRs and induce antigen-specific apoptosis in T cells. Since activated CD4-positive T cells are involved in pathogenesis of various autoimmune diseases, the apoptosis triggered by individual soluble chain-peptide complexes has significant potential for eliminating autoreactive T cells.

MHC 1 class II proteins are heterodimeric glycoproteins that bind peptides within the cell and present them at the cell surface for interaction with T cells (1,2). Several in vitro studies have demonstrated that peptides can bind to affinity purified MHC class II molecules (3)(4)(5)(6)(7) and that these complexes stimulate specific T cell responses (8 -10). In general, MHC class II molecules consist of a 34-kDa ␣ polypeptide and a 28 -30-kDa ␤ polypeptide chain noncovalently associated with each other. Furthermore, each polypeptide contains two distinct extracellular domains (␣1/␣2 in the ␣ chain and ␤1/␤2 in the ␤ chain), a transmembrane region and a small cytoplasmic C terminus region. The crystal structure of MHC class II molecules reveals that the extracellular ␣1 and ␤1 domains of both chains are involved in creating the peptide binding groove of MHC class II molecules (11)(12)(13).
The first observation for the binding of antigenic peptide to individual ␣ and ␤ polypeptide of MHC class II molecules appeared in a study where fluorescence peptide was found to be associated with both chains when murine class II-peptide complexes were subjected to SDS-gel electrophoresis under reduced conditions (14). Recent results from our laboratory have showed that electroeluted purified native individual ␣ and ␤ polypeptide chains isolated from affinity-purified murine MHC class II proteins are capable of binding antigenic peptide (15) and that purified chain-peptide complexes can stimulate T cells in vitro as measured by an increase in extracellular acidification rate in a sensor-based assay (16). In earlier studies, the possibility of T cell activation by MHC chain-peptide complexes was also suggested by the ability of alloreactive IA k -specific cytotoxic T lymphocytes to specifically lyse transfected L cells expressing either A k b1/D d c2 (17) or Ak a 1/Dd c 2 (18) MHC class II/class I hybrid molecules.
Although these studies show that native individual chains of murine MHC class II containing the transmembrane region are capable of binding peptide and can trigger T cells, no evidence of peptide binding to individual polypeptides of human MHC class II exists. In this report we describe that Escherichia coli expressed individual recombinant ␣ and ␤ polypeptides of human HLA-DR2 (DRB5*0101) lacking the transmembrane region are capable of binding immunodominant epitopes from MBP. In addition, monomeric complexes of DR2␣-Tm and DR2␤-Tm chains and MBP peptide can induce antigen-specific apoptosis in cloned T cells to a degree comparable with that of native ␣/␤ dimer-peptide complexes.

MATERIALS AND METHODS
Cell Lines, Antibodies, and Chemicals-The hybridoma cell line L243, producing monoclonal antibodies against monomorphic human HLA DR molecules, was obtained from American Type Culture Collection, Bethesda, MD. Homozygous lymphoblastoid cell lines, GMO 3107 expressing HLA DR2 and GMO 8067 expressing HLA DR3, were obtained from the National Institute of General Medical Sciences (NIGMS) human genetic mutant cell repository (Coriell Institute of Medical Research, NJ). Immunopure biotinylated bovine serum albumin containing known amount of biotin molecules was purchased from Pierce Chemicals. Anti-human ␥-IFN monoclonal antibody and rabbit anti-human ␥-IFN polyclonal antibody were obtained from Endogen, Inc. Peroxidase-conjugated rabbit IgG was purchased from Jackson Immunoresearch Laboratories. Human ␥-IFN was obtained from Boeh-* 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.
Cloning, Expression, and Purification of DR2␣-Tm and DR2␤-Tm-Cloning and expression of ␣-Tm and ␤-Tm of DR2 (DRB5 * 0101) in E. coli were described in our earlier report (19). Briefly, the plasmids p329129 and p33425 expressing DR2␣-Tm and DR2␤-Tm chains were transfected into the E. coli expression host W3310/DE3. Cultures were grown at 37°C in L-broth containing 0.4% glucose, 100 g/ml ampicillin, and 15 g/ml tetracycline. Cells were induced in mid-log growth by addition of isopropyl-␤-D-thiogalactopyranoside to a final concentration of 0.4 mM and were harvested for inclusion body preparations.
Purification of DR2 ␣ and ␤ Chains Lacking the Transmembrane Region-The detailed procedure for the purification of ␣ and ␤ chains lacking the transmembrane region from E. coli inclusion body preparations has been described recently (20). Briefly, the ␣ chain E. coli inclusion bodies were solubilized in 25 mM phosphate buffer, pH 7.4, containing 8 M urea and 20 mM dithiothreitol and purified by ionexchange chromatography using High Q-50 resin (Bio-Rad). The recombinant ␤ chain was purified by one-step gel filtration chromatography using Sephacryl S-100 resin packed in an Pharmacia XK50 (2.5-cm diameter ϫ 100-cm height) column. Both ␣ and ␤ chain fractions were collected and analyzed by SDS-gel electrophoresis using LabLogix silver staining kit (Belmont, CA). Individually pooled ␣ and ␤ chains showed purity levels greater than 95% with recovery of 52 and 86%, respectively.
Purification of Human HLA DR2 and DR3 from Lymphoblastoid Cells-Purification of HLA DR2 from Epstein-Barr virus transformed lymphoblastoid cells was carried out as described earlier (21) with some minor modifications. Triton X-100 cell lysate was applied onto L243 coupled Sepharose-4B column, and the bound DR2 was eluted in phosphate buffer containing 0.05% n-dodecyl ␤-D-maltoside detergent at pH 11.3. Fractions were immediately neutralized with 1 M acetic acid, and the DR2 pool was collected through a DEAE ion-exchange column in a phosphate buffer containing 0.5 M NaCl and 0.05% n-dodecyl ␤-D-maltoside, pH 6.0. Purified protein was then filtered through a 180-kDa membrane, dialyzed against PBS for 24 h at 4°C, and characterized by 13.5% SDS-polyacrylamide gel electrophoresis followed by silver staining. Affinity-purified HLA DR3 was obtained by similar method in 0.01% Tween 80 detergent.
Biotinylation of various peptides was carried out as described previously (22). For the synthesis of carboxyfluoresceinated MBP-(83-102)Y 83 (CF-MBP-(83-102)Y 83 ) peptide, 220 mg of 6-carboxyfluorescein (Molecular Probes) was dissolved in 10 ml of dimethylformamide, and 0.12 mmol of peptide resin was added to this solution. After mixing the suspension for 1 min, 75 l of diisopropylcarbodiimide was added, followed by gently mixing the slurry at room temperature for 2 h. The resin was then filtered and washed with dimethylformamide and methanol alternately, followed by washing with dichloromethane and vacuum drying for 2 h. The modified peptide was cleaved from the resin by suspending the resin in 10 ml of trifluoroacetic acid containing 0.7 g of 4-methylmercaptophenol and 1 ml of 4-methoxybenzenethiol. After 2 h, the resin was separated by filtration, and the filtrate was collected in 1 liter of pentane:acetone (8:1, v/v) mixture. The precipitated CF-peptide was separated by decantation and centrifugation and washed with pentane/acetone followed by pentane. Crude peptide was purified by reverse-phase HPLC using C18 column (Vydac, CA) with acetonitrile gradient (0.1% trifluoroacetic acid in water to 0.1% trifluoroacetic acid in 70% aqueous acetonitrile) and characterized by mass spectroscopy.
Complex Preparation and Peptide Binding Assay-For the quantitative detection of bound peptide, affinity-purified HLA-DR2 at a concentration of 2 g/ml was incubated with biotinylated MBP peptides at 37°C for 96 h at optimized pH. The resulting complex preparations were analyzed by antibody capture plate assay using an enzyme-conjugated avidin system as described recently (7) with minor modifications. Briefly, purified polyclonal antibodies against HLA-DR2, DR2␣-Tm chain, and DR2␤-Tm chain at a concentration of 20 g/ml were immobilized in a 96-well microtiter plate. Bovine serum albuminbiotin with 8 biotin molecules per bovine serum albumin was used as a standard with a concentration range of 0.014 -1.8 pmol (0.117-15 ng). The bound biotinylated peptide in complex preparations were detected colorimetrically using alkaline phosphatase-conjugated streptavidin and p-nitrophenyl phosphate in 0.1 M diethanolamine as a substrate.
Dissociation Kinetics of Chain-Peptide Complexes by SDS-Gel Electrophoresis-Stability of single chain-peptide complexes was measured by SDS-polyacrylamide gel analysis of various complexes prepared with radioiodinated MBP peptide. Complexes of HLA-DR2 and 125 I-MBP peptides were prepared under optimized binding conditions and purified from unbound peptide by G-75 size exclusion gel filtration chromatography. Resulting complexes were then incubated at 4 and 37°C, and at various time points complex samples were removed and frozen at Ϫ20°C. At the end of 72 h, the complexes were applied on 13.5% SDS gels under nonreduced conditions. Gels were stained, dried, and autoradiographed. Each lane containing the ␣␤ dimer, ␣-Tm, or ␤-Tm polypeptide chain was cut and counted in a ␥ counter.
Size-exclusion HPLC Analysis and Fluorescence Peptide Binding-Solutions of 5 M ␣-Tm or ␤-Tm and 50-fold molar excess of CF-MBP-(83-102)Y 83 peptide in PBS buffer, pH 9, were incubated at 37°C for 72 h. Complexes were purified from the excess of peptide by SE-HPLC using a Toso Haas TSK SW3000 column (0.75 ϫ 60 cm) at the flow rate of 0.5 ml/min in PBS buffer at pH 9. The protein fraction has been collected from 20 to 40 ml, concentrated using a Centricon-10 microconcentrator (10-kDa cut-off membrane), and then analyzed by size-exclusion HPLC using the same column connected to UV diode array and fluorescence detectors in series. Two signals have been collected for each SE-HPLC chromatogram: fluorescence with excitation wavelength of 448 nm and emission wavelength of 525 nm and a UV signal at 278 nm. The fluorescence signal was used to measure the bound peptide, and the UV detection was used to calculate the protein concentration.
T Cell Receptor Occupancy Assay-The herpes saimiri virus-transformed SS8T human T cell clone restricted to DR2 (DRB5*0101) and MBP-(84 -102) was cultured in RPMI 1640 medium supplemented with 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, 10% fetal bovine serum (Hyclone) and 50 units/ml human recombinant IL-2 (rIL-2) at 37°C. Every alternate day cells were transferred to fresh media. Various complex preparations were incubated in a microtiter tissue culture plate at a density of 20,000 cells/200 l/well in the absence of rIL-2. After 48 h of incubation at 37°C, the supernatants were collected from each well to test for the increase in ␥-IFN cytokine level. The detection of ␥-IFN levels was performed by antibody enzymelinked immunosorbent assay as described recently (19).
In Situ Terminal Deoxynucleotidyl Transferase Assay and Flow Cytometry-The quantitative detection of DNA strand break was performed by end labeling of 3Ј-OH end of fragmented DNA using biotinylated dUTP followed by a fluorescein isothiocyanate-conjugated avidin detection system in a flow cytometer as described earlier with some modifications (23). Briefly, the transformed SS8T cloned T cells at a density of 1 ϫ 10 6 cells/ml were incubated with equimolar amounts of DR2 dimer-peptide or chain-peptide complex preparations. Cells were fixed in 1% buffered formaldehyde and stored at Ϫ20°C in 70% ethanol. Cells were rehydrated in Hanks' balanced salt solution and resuspended in 50 l of terminal deoxynucleoside transferase reaction mixture containing 10 l of terminal deoxynucleoside transferase buffer (1 M potassium cacodylate, 125 mM Tris-HCl, pH 6.6, 1.25 mg/ml bovine serum albumin), 0.2 l terminal deoxynucleotide transferase, 5 l of CoCl 2 , and 0.5 nmol of biotin-16-dUTP. The reaction mixture was incubated at 37°C for 30 min, and cells were rinsed in Hanks' balanced salt solution and resuspended in 100 l of staining solution containing 2.5 g/ml fluorescein isothiocyanate-avidin in saline sodium citrate buffer. Cells were incubated for an additional 30 min at room temperature in the dark and resuspended in 1 ml of Hanks' balanced salt solution containing 5 g/ml propidium iodide in the presence of 0.1% RNase. Double fluorescence measurements were carried out in Becton-Dickinson flow cytometer using LYSYS II software.
Transmission Electron Microscopy-SS8T cells at a density of 2 ϫ 10 6 cells/ml were incubated with 50 g/ml of freshly prepared native DR2-MBP-(83-102)Y 83 complexes or 25 g/ml of DR2␣-MBP-(83-102)Y 83 or DR2␤-MBP-(83-102)Y 83 complexes at 37°C for 18 h. T cells were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, at 25°C for 2 h. Following a rinse in the buffer, cells were post fixed in 1% osmium tetraoxide at 25°C for 1 hour. Cells were then dehydrated in a graded series of ethanol, treated with propylene oxide, embedded in a quick-mix epoxy resin, and polymerized at 60°C for 16 h. The sample blocks were thin-sectioned with a diamond knife and sections were stained with 2% uranyl acetate in the presence of 0.4% lead citrate. Cells were characterized using JEOL-100CX TEM at an accelerating voltage of 80 kV.

RESULTS AND DISCUSSION
MHC class II molecules consist of two individual polypeptide chains (an ␣ and a ␤) of similar size that are noncovalently associated with each other. Recently reported crystal structure of MHC class II molecules shows that the extracellular ␣1 and ␤1 domains are involved in creating the peptide binding domain that can accommodate peptides of varied length (11)(12)(13). In the heterodimeric structure, it has been shown that the electrophoretically resolved ␣ and ␤ chains independently bind the same peptides, although the two chains may or may not bind identical amino acid residues in the peptides (14,24). In our previous reports, we demonstrated that electroeluted purified ␣ and ␤ chains of murine MHC class II polypeptides can bind antigenic peptides like the native heterodimer, and equimolar amounts of single chain-peptide complexes can trigger T cell response as measured by a sensor-based assay (15,16,25).
The limited availability of purified native chains by the tedious electroelution method led us to investigate recombinant MHC class II chains for further studies. In this report we describe our successful effort to demonstrate that (i) E. coli  (26). Similarly, the peptide of MBP (MBP-(84 -102)) used in this study is considered a major immunodominant epitope for human MS (27). The peptide analog of MBP used in our study contains a tyrosine residue at position 83 and was found to have increased binding affinity to HLA-DR2 without loss of TCR recognition. 2 The increased binding of N terminus tyrosine containing peptide was also observed in several other MHC class II-peptide complexes. 3 The cloning and expression of human HLA-DR2 ␣-Tm and ␤-Tm was carried out as described recently (19). The expression of recombinant individual ␣ and ␤ chain of HLA-DR2 represent 30% of the total cell protein. In contrast to individual ␣ and ␤ polypeptide chain, the recombinant expression of heterodimeric MHC class II molecules in E. coli was totally unsuccessful. The insoluble denatured inclusion body preparations were solubilized in 8 M urea, purified in a scale of 50 -100 mg by conventional chromatography methods as described earlier (20), and stored in PBS containing an 8 M urea solution. Prior to peptide loading, the ␣ and ␤ chains were dialyzed against PBS buffer. Binding of various biotinylated MBP peptides to purified ␣ and ␤ polypeptide chains were carried out by antibody captured plate assay using chain-specific rabbit polyclonal antibodies and enzyme-conjugated streptavidin as described recently (7).
The optimum pH for maximum peptide occupancy was measured by incubating a known amount of each chain with 50-fold molar excess of MBP peptides. Three MBP peptides were selected for the optimization of peptide binding to individual chains. Their affinities toward HLA-DR2 have been shown in the order of MBP-(84 -102) Ͼ MBP-(124 -143) Ͼ MBP-(1-14) (28). The epitope MBP- (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14) had no affinity toward HLA-DR2 and was used as a negative control. As shown in Fig. 1, the binding of MBP-(83-102)Y 83 peptide to both ␣ and ␤ chain was maximum at pH 9.0. In contrast, the binding of MBP-(83-102)Y 83 peptide to native DR2 heterodimer was found to be maximum at acidic pH (7). In case of the ␣ chain, the binding pH consistently appears to be critical below or above pH 9. The second high binding epitope of MBP, (MBP-(124 -143)) bound strongly to the ␤ chain at basic pH and weakly to the ␣ chain. Further increase in peptide concentration beyond 50-fold molar excess as well as increase in incubation period did not provide additional binding (data not shown). In negative controls, the MBP-(1-14) peptide did not bind to either the ␣ or ␤ chain like the native ␣␤ DR2 heterodimer. Although the maximum binding of the MBP peptide to each individual chain was at pH 9.0, a significant amount of peptide remained bound at around physiological pH between 7 and 8.
The  Further characterization of single chain-peptide complexes with respect to aggregation level was performed by size-exclusion HPLC analysis. Although the recombinant ␣ and ␤ polypeptide chains used in our study lack the hydrophobic transmembrane region, due to inclusion body preparation of these polypeptides in E. coli, purified proteins tend to aggregate in solution in the absence of denaturing agent. The HPLC result presented in Fig. 3D shows that, prior to the peptide loading, almost all of ␤ polypeptide preparation appeared in the aggregated state (600 kDa). In contrast, most of the ␣ polypeptide appears to be in a state of ␣-␣ homodimers (60 kDa) as shown in Fig. 3A. The existence of ␣-␣ and ␤-␤ homodimers in purified native polypeptide chains was also observed in our earlier studies (15,16). Upon peptide binding, however, a significant amount of purified complexes of both ␣ and ␤ chainpeptide shifted to the monomeric state with a molecular size of ϳ30 kDa (Fig. 3, B and E). In these experiments, carboxyfluorescein labeled MBP-(83-102)Y 83 peptide was used to monitor the bound peptide associated with various molecular size protein fractions. Results presented in Fig. 3, C and F, show that almost no fluorescence intensity was associated with highly aggregated proteins or with homodimers. In fact, the fluorescence intensity was only found to be associated with the monomeric form of both ␣and ␤-MBP-(83-102)Y 83 peptide complexes. These results clearly demonstrate that bound peptide prevents aggregation of purified chains. Such prevention of aggregation of MHC class II dimers on cell surface by peptide binding has been reported recently (29). Similarly, in a separate study we have observed that bound peptide significantly inhibits aggregation of purified MHC class II heterodimers in solution. 4 Calculated percent aggregation and associated bound peptide by size-exclusion HPLC of individual chains and their complexes are also presented in Fig. 3. The molar percent of total bound peptide data observed with CF-MBP-(83-102)Y 83 peptide in HPLC experiment correlates well with the biotinylated peptide binding results obtained with the antibody captured plate assay.
The recognition of ␣and ␤-MBP-(83-102)Y 83 peptide complexes by TCR was performed using herpes saimiri virus-transformed SS8T cloned T cells. The SS8T cell clone was generated from an MS patient and was fully characterized for its restriction to HLA-DR2 (DRB5*0101) and MBP-(84 -102) peptide (30). The TCR engagement by soluble ␣and ␤-MBP-(83-102)Y 83 peptide complexes was monitored by an increase in ␥-IFN cytokine in a dose-dependent manner. Such increase in ␥-IFN production by T cells was correlated with the occupancy of TCRs on the surface of T cells in an earlier study (30) and was adopted for SS8T cells in our laboratory (19). As shown in Fig. 4, specific increase in ␥-IFN was observed when SS8T cells were exposed to complexes of native DR2, ␣ or onstrate that the observed level of increased ␥-IFN is not due to the release of bound peptide in the culture medium, the MBP-(83-102)Y 83 peptide was complexed with irrelevant HLA-DR3 as a control and showed no increase in ␥-IFN level (Fig. 4A). Similarly, in a mock experiment, equivalent amount of MBP-(83-102)Y 83 peptide incubated and passed through Sephadex G-75 column under identical purification conditions in the absence of chains, did not show any increase in ␥-IFN level (data not shown).
Prolonged incubation of SS8T cells with relevant complexes of individual ␣ and ␤ chains led to the induction of apoptosis. Typically apoptosis is characterized by chromatin condensation and is associated with endonuclease activity. The endonuclease activity in apoptotic cells can be demonstrated by cleavage of cellular DNA. The quantitative detection of DNA strand breaks in this study was demonstrated by labeling the 3Ј-OH end of the fragmented DNA with biotinylated dUTP followed by fluorescein isothiocyanate-conjugated The calculated percent T cell apoptosis by relevant chain-peptide complexes with respect to various controls is presented in Fig. 6A. Apoptosis of T cells by chain-peptide complexes appeared to be time-dependent as shown in Fig. 6B. Finally, the chromatin condensation and cell shrinkage characteristics of apoptotic cells were demonstrated by transmission electron microscopy (Fig. 7). As compared with untreated T cells (Fig. 7A), SS8T cells treated with native DR2-MBP-(83-102)Y 83, ␣-MBP-(83-102)Y 83 , and ␤-MBP-(83-102)Y 83 complexes showed typical apoptotic cells (Fig. 7, B-D).
In summary, results presented in this report describe that individual recombinant polypeptide chains of human MHC class II molecules are capable of binding antigenic peptide and that complexes of single-chain peptide can recognize TCR to induce antigen-specific apoptosis in T cells. Furthermore, this study demonstrates that the transmembrane region of human MHC class II molecules are not involved in either peptide binding or TCR recognition. The physiological significance of single chain-peptide complexes is unknown at present and requires further investigation. In a parallel study, we observed that recombinant murine IA s ␣ chain complexed with rat MBP-(90 -101) peptide was highly effective in preventing experimental allergic encephalomyelitis in mice, an animal model for human MS. 5 Prevention and treatment of several autoimmune diseases in animal models by soluble native MHC class IIpeptide complexes were demonstrated in our laboratory (33,34). 6 The T cell apoptosis reported here by recombinant soluble chain-peptide complexes may have significant clinical relevance in developing therapeutics for the elimination of autoreactive T cells in various autoimmune diseases in an antigenspecific manner.