Depression of T-cell Epitope Generation by Stabilizing Hen Lysozyme*

Conformational stability of proteins is an important factor that determines their resistance/susceptibility to proteolytic digestion. Intracellular proteolysis is the key step in antigen presentation events for protein antigens; hence, it is likely that increasing protein stability reduces the antigenicity of proteins. We prepared three hen egg white lysozyme derivatives possessing different stabilities by chemical modification to clarify the relationship between conformational stability and the antigenicity of the protein. One of the derivatives was conformationally unstabilized by removing one intramolecular disulfide bond, whereas the two others were stabilized by the addition of an intramolecular cross-link. The antigenicity of these derivatives was evaluated using hen egg white lysozyme-specific T-cell hybridoma cells and a B-lymphoma cell line, A20, as antigen-presenting cells. With an increase in conformational stability, the T-cell response decreased. However, the reduction was not derived from the inefficiency of internalization to A20 cells nor the alteration of antigenicity by chemical modifications. Moreover, from analyses of their susceptibility to proteolysis and the kinetics of presentation of the T-cell epitope, it was confirmed that increasing protein stability led to the depression of T-cell epitope generation by increasing resistance to proteolysis. These results have an important implication in devising a new strategy for manipulating T-cell response by the stability of protein antigen.

Conformational stability of proteins is an important factor that determines their resistance/susceptibility to proteolytic digestion. Intracellular proteolysis is the key step in antigen presentation events for protein antigens; hence, it is likely that increasing protein stability reduces the antigenicity of proteins. We prepared three hen egg white lysozyme derivatives possessing different stabilities by chemical modification to clarify the relationship between conformational stability and the antigenicity of the protein. One of the derivatives was conformationally unstabilized by removing one intramolecular disulfide bond, whereas the two others were stabilized by the addition of an intramolecular crosslink. The antigenicity of these derivatives was evaluated using hen egg white lysozyme-specific T-cell hybridoma cells and a B-lymphoma cell line, A20, as antigen-presenting cells. With an increase in conformational stability, the T-cell response decreased. However, the reduction was not derived from the inefficiency of internalization to A20 cells nor the alteration of antigenicity by chemical modifications. Moreover, from analyses of their susceptibility to proteolysis and the kinetics of presentation of the T-cell epitope, it was confirmed that increasing protein stability led to the depression of T-cell epitope generation by increasing resistance to proteolysis. These results have an important implication in devising a new strategy for manipulating T-cell response by the stability of protein antigen.
An antigen-specific CD4 ϩ T-cell recognizes an antigen-derived peptide that is mounted on a major histocompatibility complex class II molecule on a cell surface of antigen-presenting cells, via its antigen receptor (1,2). Therefore, the conformation of protein antigens is unlikely to have any role in the step of T-cell recognition. Prior to T-cell activation, however, antigen processing is necessary for a protein antigen to stimulate the specific T-cells; this processing consists of multiple steps of cellular events, i.e. internalization of proteins by antigen-presenting cells, reduction of the disulfide bond and unfolding of proteins, enzymatic digestion, and assembly of the generated peptides with major histocompatibility complex class II molecules (3,4). Proteases preferentially digest proteins in an unfolded state rather than those in a folded state (5-7); thus, the unfolding may be a crucial step for intracellular antigen processing. In this context, we can expect that depres-sion of protein unfolding by increasing protein stability would reduce the antigenicity of proteins for T-cells.
Several reports have demonstrated a relationship between increased antigenicity and the decreased stability of proteins (8 -10). However, the influence of protein stability on the antigenicity remains unknown. To address this issue, we prepared three derivatives of hen egg white lysozyme (HEL) 1 possessing different conformational stabilities (see Table I). A three-disulfide derivative of HEL, S-carboxymethylated HEL at Cys 6 and Cys 127 (6,127CM-HEL), was produced by selective reduction of a disulfide bond Cys 6 -Cys 127 in four original disulfide bridges (11). The loss of a single disulfide bond was reported not to give major three-dimensional structural change (11,12) but to dramatically decrease the conformational stability (13). On the other hand, HEL was stabilized by cross-linking between Lys 1 and His 15 with alkyl chain, 1-15CL-HEL (14), and by crosslinking between Glu 35 and Trp 108 through an ester bond, 35-108CL-HEL (15,16). These intramolecular bridges were shown not to cause structural constraint; therefore, they keep a similar native structure with enhancing protein stability (17,18). Using these derivatives, we evaluated the influence of the conformational stability of HEL on the steps of the antigen presentation pathway. We report that the increasing conformational stability of HEL resulted in reducing the antigenicity for HEL-specific T-cells. The results indicate that protein stability is an important factor in determining the dose of T-cell epitopes and consequently may determine the magnitude of T-cell response.

EXPERIMENTAL PROCEDURES
Antigen Preparations-HEL, purified by repeated recrystallization five times, was kindly donated by QP Co. (Tokyo, Japan). Three derivatives of HEL used in this study, 6,127CM-HEL, 1-15CL-HEL, and 35-108CL-HEL, were prepared by the methods of Radford et al. (11), Ueda et al. (14), and Imoto et al. (15), respectively. Unfolded-HEL and unfolded-1-15CL-HEL were prepared with an S-alkylating reagent of (3-bromopropyl)trimethylammonium bromide (TAP-Br) following reduction of HEL and 1-15CL-HEL with 2-mercaptoethanol as described previously (19). Because 35-108CL-HEL was highly stable, unfolding was carried out as follows: to hydrolyze the ester bond, 10 mg of 35-108CL-HEL was dissolved in 50 ml of 0.05 M sodium borate buffer (pH 10.0) and lyophilized. After dialyzing against distilled water, the soluble fraction was subjected to cation exchange chromatography using CM-Toyoperarl 650 M (Tosoh, Tokyo, Japan) and eluted with a gradient of 0.05 M sodium borate buffer (pH 10.0) and the same buffer containing 1 M NaCl to separate non-ester molecules from native 35-108CL-HEL. The protein fraction eluted earlier than native 35-108CL-HEL was collected, dialyzed against distilled water, and reduced with 2-mercaptoethanol, followed by S-alkylation with TAP-Br.
Conditions for Cell Culture-Cells were cultured in Dulbecco's modified Eagle's medium (Sigma) containing 4.5 g/liter glucose supple-* 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.
T-cell Hybridoma-T-cell hybridoma was established according to Adorini et al. (20). Briefly, lymph node cells were obtained from BALB/c mice (Seac Yoshitomi, Ltd., Fukuoka, Japan) 9 days after immunization with HEL (50 g/mouse) emulsified in complete Freund's adjuvant (Difco Laboratories, Detroit, MI). The cells were cultured in the presence of 150 g/ml HEL for 3 days, fused with a thymoma cell line, BW5147, using PEG-4000 (Merck, Darmstadt, Germany), and cultured in a medium containing hypoxanthine, aminopterin, and thymidine. Growing hybrids were screened for interleukin-2 (IL-2) producing capacity in response to HEL in a major histocompatibility complexrestricted manner using a syngenic mouse B-lymphoma cell line, A20 (21), as antigen-presenting cells. One representative clone established by repeated limiting dilution was used in this study.
T-cell Stimulation Assay-The T-cell hybridoma (2 ϫ 10 5 cells) were cultured with 5 ϫ 10 4 A20 cells in a well of microtiter plates (Falcon number 3072), in the presence of various concentration of HEL derivatives in a total volume of 240 l. After a 24-h incubation, IL-2 produced by the T-cell hybridoma was measured using a proliferative response of an IL-2-dependent cell line, CTLL-2, as already described (22). Cell-free supernatants of the T-cell hybridoma cultures (100 l) were transferred to other plates containing CTLL-2 (10 4 cells/well) in 100 l of culture medium. The cells were cultured for 24 h, followed by the colorimetric 3-[4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (Sigma) assay (23,24). CTLL-2 cultured with known concentrations of recombinant IL-2 (Genzyme, Cambridge, MA) are included in each experiment as an internal standard.
Quantification of Endocytosed HEL-Ten million A20 cells in 1 ml of complete culture medium were incubated with 25 M of various HEL derivatives at 37°C in triplicate dishes. After incubation for 30 min, the cells were washed three times with cold Dulbecco's modified Eagle's medium containing 1% fetal calf serum to remove free antigens. Cells were then treated once with 0.5 M NaCl in 0.2 M acetic acid for 10 min at 4°C to remove any proteins attached to the cell surface (25). After the acid treatment, cell pellets were stored at Ϫ80°C.
Amounts of HEL and the derivatives incorporated into A20 cells were determined using competitive enzyme-linked immunosorbent assay. Microtiter enzyme-linked immunosorbent assay plates (number 442404; Nunk, Roskilide, Denmark) were coated with 50 l of goat anti-mouse IgM ϩ IgG (Tago Inc., Burlingame, CA) in 0.1 M sodium carbonate buffer (pH 9.6) at 2 g/ml overnight at 4°C. After washing with phosphate buffered saline (PBS) containing 0.05% Tween 20 (PBST), a mouse anti-HEL polyclonal antibody diluted 1:1000 in PBST was incubated overnight at 4°C. Blocking was done with 2% nonfat dry milk in PBST. Cells stored at Ϫ80°C were thawed and lysed in the presence of 1% Triton X-100 in PBS (100 l) and diluted in the blocking buffer, and 50 l of the mixture was added to the precoated enzymelinked immunosorbent assay plates. The HEL derivatives were allowed to react with coated antibody overnight at 4°C. 50 l of biotinylated HEL (20 nM) in the blocking buffer was then added to each well and incubated for an additional 1 h at room temperature. The binding of biotinylated HEL was subsequently probed with an alkalinephosphatase-biotin/avidin complex (Vector Laboratories, Burlingame, CA). Coloring was obtained using a 1 mg/ml p-nitrophenyl phosphate (Wako Pure Chemical, Osaka, Japan) substrate solution with 1 mM MgCl 2 in 0.1 M sodium carbonate buffer (pH 9.6), and absorbance at 405 nm of each well was recorded at 24 h. Known concentrations of HEL derivatives are included in each plate as an internal standard. In preliminary experiments, no difference was observed between each of the HEL derivatives and native HEL in their reactivity to the polyclonal anti-HEL antibody.
Proteolytic Degradation of HEL Derivatives-Digestion of HEL derivatives was carried out following the method of Collins et al. (26) with some modifications. Briefly, HEL derivatives at 200 g/ml were digested either with 20 g of cathepsin B or cathepsin D (Sigma) in 1 ml of 0.1 M MES/acetate (pH 5.0) and 0.1 M Tris/citrate (pH 3.5), respectively, containing 1 mM EDTA and 3 mM 2-mercaptoethanol at 37°C. These digests were analyzed by SDS-polyacrylamide gel electrophoresis followed by a densitometric analysis as described previously (23).

RESULTS
T-cell-stimulating Capacity of HEL Derivatives-To examine the influence of protein stabilities on the antigenicity for an antigen-specific T-cell, a cloned T-cell hybridoma was established from BALB/c mice (H-2 d ). The sole T-cell epitope noted for H-2 d haplotype is located in the HEL sequence 107-116 and is presented in association with I-E d class II major histocompatibility complex molecules (20,27,28). In accordance, this hybridoma responded to HEL peptides 106 -129 and 98 -116 (data not shown).
The antigenicity of various HEL derivatives was evaluated by measuring the activities to induce IL-2 production in the T-cell hybridoma. The T-cell-stimulating capacity of the derivatives inversely correlated with their stabilities (Fig. 1). The order of antigenicity was 1) unfolded HEL, 2) 6,127CM-HEL, 3) native HEL, 4) 1-15CL-HEL, and 5) 35-108CL-HEL. Unfolded HEL was found to be approximately 100 times more potent than native HEL, whereas under the same conditions, 35-108CL-HEL did not activate the T-cells at all. None of the HEL derivatives were cytotoxic by themselves (data not shown); thus, conformational stability may be the factor that determines their differential antigenicity for the T-cell.
Internalization of HEL Derivatives to A20 Cells-There was a possibility that the differing T-cell-stimulating capacities of HEL derivatives might be due to their differential properties incorporated into antigen-presenting cells. To address this question, A20 cells were cultured with each derivative for 30 min, as the intracellular accumulation of antigen by fluidphase endocytosis in A20 cells was saturated by 20 min (29) and as the intracellular level of native HEL peaked at 30 min (data not shown). After the culture, antigens attached to the cells but not internalized were removed from the cell surface by treatment with an acidic high salt solution, and the amounts of endocytosed derivatives were assessed by competitive enzymelinked immunosorbent assay (Fig. 2). The intracellular level of stabilized derivatives, 1-15CL-HEL and 35-108CL-HEL, which possessed lowered antigenicities, was higher than that of their native counterpart. This indicates that the reduced antigenicity of stabilized HELs for the T-cell was not due to their inefficiency in being internalized into antigen-presenting cells. On the other hand, endocytosis of the unstabilized derivative, 6,127CM-HEL, was also increased compared with the native one. Its augmented antigenicity may be, in part, due to increased endocytosis by the antigen-presenting cells.
It is also suggested that the protein stabilization may lead to the enhancement of intracellular accumulation of protein antigen. During the time course of accumulation, the amount of 35-108CL-HEL was further increased by 60 min, whereas the maximum accumulation of 6,127CM-HEL was observed at 15-30 min, and then the level was gradually decreased (data not shown).
Recovery of T-cell-stimulating Activity of Stabilized HELs by Denaturation-There is another possibility that the reduced T-cell-stimulating capacities of 1-15CL-HEL and 35-108CL-HEL may be due to their antigenic alteration of the T-cell epitope region by chemical modifications. To evaluate this possibility, we measured T-cell responses against their unfolded forms (Fig. 3). In the native state, the T-cell-stimulating capacities of these HEL derivatives at 3 M were not so potent. When the stabilized derivatives were unfolded, they showed strong antigenicities comparable with the unfolded HEL prepared from the unmodified one. This means that the chemical modifications for cross-linking did not alter the antigenicity of the epitope region in HEL molecules for H-2 d T-cells. Thus, the importance of conformational stability in the antigenicity of HEL was again suggested.
Susceptibility of HEL Derivatives to Proteolytic Degradation-Restriction of the denaturation by increasing protein stability probably leads to blocking the generation of antigenic peptides, resulting in suppression of T-cell activation. To test this possibility, HEL derivatives were digested in vitro with two putative processing enzymes for HEL, cathepsin B and D. Both cathepsins completely degraded unfolded HEL within 1 h. On the other hand, stabilized HELs, 1-15CL-HEL, and 35-108CL-HEL were almost intact throughout the incubation period (Fig. 4). Although degradation profiles of native HEL and 6,127CM-HEL for each cathepsin were different, the result clearly indicated that the more stable one was more resistant to protease degradations. Thus, reduced antigenicity of stabilized HEL against T-cell may be derived from acquiring the resistance to intracellular proteases by protein stabilization.

Time Course of Antigen Processing for HEL Derivatives-The
kinetics of presentation of a T-cell determinant of HEL was examined by the pulsing of A20 cells with native HEL and various derivatives for an appropriate time period (Fig. 5). The epitope of HEL was more rapidly generated from unstabilized 6,127CM-HEL and was efficiently presented on the cell surface, compared with native HEL. On the contrary, the epitope presentation from more stable 1-15CL-HEL required longer incubation time and never reached to the same extent as the native HEL. The most stable 35-108CL-HEL was never processed during the incubation periods of 24 h. The efficiency of presentation showed a positive correlation with the susceptibilities to proteolysis (Fig. 4). Therefore, it is strongly suggested that stabilized HEL resists proteolysis by intracellular processing enzymes in the acidic compartments of antigen-presenting cells, which depresses the generation of the T-cell epitope. DISCUSSION The theory that the digestion of a small globular protein with a protease proceeds via the unfolded state rather than the folded state of the protein has been proposed by several investigators (5)(6)(7)30). The theory can be summarized by the following equation, where N and D are the folded and the unfolded proteins, respectively, and P is the protease (7,30). Prior to antigenspecific T-cell activation, protein antigens are taken up by antigen-presenting cells and then degraded within intracellular acidic organelle by proteases, such as cathepsin B and D (31)(32)(33). The process of proteolysis is a limiting step in transmitting information of antigen to T-cells. Thus, the theory suggests that the stability of protein antigen may be a critical parameter in determining the rate of processing and also that by controlling antigen stability, the dose of antigenic peptides and the resulting T-cell response could be artificially manipulated. For this reason, the conformational stability of a protein antigen may be a critical factor in determining the immune response of lymphocytes against the foreign molecule. However, the precise role of the stability of antigens in T-cell acti-vation has not been well defined, although its importance has been implied in several reports (8 -10).
To examine the relationship between the conformational stability of antigens and the susceptibility to proteolytic degradation, three derivatives of HEL with different conformational stabilities (Table I) but with retained similar native conformation and unfolded HELs were prepared. As was indicated by the previous theory, the resistance of each HEL derivative to proteolysis was correlated with the stability. In accordance with the results, T-cell epitope generation of HEL was suppressed by increasing HEL stability, which resulted in reduced T-cell activation. We found a similar result when HEL-specific polyclonal T-cells were stimulated with these HEL derivatives, 2 indicating that this result is not specific to the T-cell hybridoma used in this study. Rouas et al. reported that T-cell response for a subunit of human chorionic gonadotropin was dependent on the quaternary structure and that dissociation of the ␣/␤ subunit was required for the recognition (9). Also, Janssen et al. showed that a T-cell epitope of myelin basic protein inserted into the outer membrane protein PhoE of Escherichia coli was barely stimulative for the specific T-cell in the native trimeric form, whereas the denatured monomeric form was immunogenic and induced autoimmune encephalomyelitis (10). These reports suggest the critical role of the quaternary structure of protein antigen in depressing the T-cell triggering response. The conformation of an oligomeric protein is generally stabilized by intermolecular interactions of the subunits (34,35), and the previous two experiments suggest the role of intermolecular stabilization of a protein in depressing T-cell response. A favorable binding of a ligand to the folded state of the protein usually enhances the stability of the protein (36 -38). The stabilization by a ligand is expressed in the following equation, where X is a ligand to bind to the folded state of the protein. Therefore, the theory regarding stabilizing a protein by intermolecular interaction may be explained to be different from the present theory where an equilibrium shifts from the unfolded state to the folded one by intramolecular interaction. Stabilization of proteins is generally achieved by introducing an additional cross-link either by chemical modification (14,15,39) or by genetically introducing a disulfide bond (40,41). It is widely accepted that the stability conferred by the cross-link can be largely related to the entropic destabilization of the   (14). e Johnson et al. (16). f This value was calculated using the difference in free energy change between native HEL and 6,127CM-HEL at pH 3.8, which were reported in the literature (13). unfolded state (42,43). The strengthened conformational stability of 1-15CL-HEL and 35-108CL-HEL by the cross-link originates from this type of stabilization force. On the other hand, irreversible processes coupled with the unfolded state, such as protease digestions, can be effectively resisted by decreasing the unfolding rate as shown in the following equation, where k u and k f are the unfolding and the folding rate constants, respectively (44,45). In the case of 35-108CL-HEL, this kind of stabilization force also has an effect on enhancing the stability (46), which may be involved in the higher resistance for proteolytic digestion and the most depressed T-cell-stimulating capacity. Alternatively, protein stabilization can be accomplished by strengthening favorable interactions, such as hydrophobic interactions, electrostatic interactions, and hydrogen bonds (47)(48)(49). Also, conjugating the protein with a nonimmunogenic polymer, polyethylene glycol, is effective in stabilizing proteins against proteolytic degradation, which resulted in the abortion of the T-cell-stimulating capacity of HEL (23). These genetic or protein engineering strategies are promising for regulating the generation of stimulatory peptides from the proteins. Some forms of chemical reactions are coupled with the unfolded state of proteins. The reactions causing irreversible denaturation are suggested to be the deamidation of asparagine residues, the hydrolysis of peptide bonds at aspartic acid residues, and the destruction of disulfide bonds (50,51). Accumulation of these chemical reactions in the unfolded state leads to the formation of irreversibly unfolded molecular species, which are highly susceptible to proteolytic degradation. Thus, the T-cell-triggering capacity of these irreversible forms may be greater than the native form, and these processes may increase the antigenicity of proteins. Indeed, we observed that native HEL stored in PBS solution at 4°C for 1 month was more antigenic for T-cell hybridoma than the freshly purified one. 2 Therefore, HEL derivatives used in this study were subjected to assays immediately after purification using cation exchange chromatography. Moreover, it has been demonstrated by Weigle and his colleagues that an aggregated form of human gamma globulin (HGG) is highly antigenic for T-cells (52,53), whereas an aqueous preparation of HGG (deaggregated HGG) that is ultracentrifuged to remove small amounts of aggregates is highly tolerogenic for T-cells, i.e. injection of adult mice with deaggregated HGG renders them immunologically unresponsive to a subsequent injection of the immunogenic HGG (54,55). According to the present theory, the difference in their immunological properties between the deaggregated HGG and the untreated one may be attributed to the amount of irreversibly unfolded HGG (aggregated HGG) in the solutions, which may be extremely sensitive to protease degradation in vivo, resulting in the generation of greater quantities of antigenic peptides for the T-cells.
In conclusion, this article reports the first evidence to show the effect of the conformational stability of a protein antigen on T-cell-triggering response and the related mechanism. The finding that the production of the T-cell epitope can be controlled by the conformational stability of a protein antigen may have many ramifications for manipulating T-cell immunity.