B-cell repertoire specific for an unfolded self-determinant of mouse lysozyme escape tolerance and dominantly participate in the autoantibody response

Abstract

We previously found that autoantibodies against mouse lysozyme (ML) were strongly induced in normal BALB/c mice when immunized with mutant ML that has triple mutations rendering the dominant T-cell epitope of hen egg lysozyme (HEL), HEL 107–116. As T cells specific for HEL 107–116 were primed in these mice, the anti-ML immunoglobulin G (IgG) responses would be the result of collaborations between autoreactive B cells specific for ML and T cells specific for HEL 107–116. Serum IgG responses against ML were dominantly focused on the ML 14–69 region, indicating that B cells responding to the epitope escape tolerance. In the present study, we prepared several monoclonal antibodies (mAbs) specific for ML 14–69 and examined their antigen specificities in detail, to characterize the nature of the remaining B-cell repertoire specific for ML. mAbs specific for ML 14–69 interacted weakly with soluble, native ML, but the interactions were strengthened by denaturation of ML. The apparent affinity constants between these mAbs and ML showed an increase, ranging from six- to 80-fold, by denaturation of ML. Therefore, these mAbs were more specific for the denatured determinant than for the determinant in the native structure. These results indicate that a substantial number of autoreactive B cells, specific for the unfolded conformation of ML, escape tolerance and are dominantly involved in the autoantibody response to ML. Our finding provides important information to understand the naturally occurring autoreactive B-cell repertoire in normal mice.

Introduction

B-cell tolerance to soluble self-antigens in vivo has been extensively studied using double-transgenic mice expressing hen-egg lysozyme (HEL) as a neo-self protein along with surface immunoglobulin transgenic receptors specific for HEL.¹ In a transgenic model, tolerance induction in self-reactive B cells is not absolute and is limited by thresholds determined by factors such as antigen (Ag) concentration and form of Ag, which affect the affinity and avidity between antigens (Ags) and their specific immunoglobulin receptors. B cells with a high avidity to soluble self-protein are subjected to functional silencing (anergy), while B cells with a low avidity to soluble self-Ag escape tolerance and are present in the periphery.

Self-proteins carrying a foreign T-cell epitope often induce autoantibodies in normal mice. For example, self-proteins coupled to foreign antigenic peptides or proteins,^2,3 self-proteins with foreign T-cell epitopes inserted into its sequence^4,5 and foreign proteins which resemble self-proteins,^6,7 induced autoantibodies specific for each self-Ag. The phenomena can be explained by the collaboration of autoreactive B cells and T cells specific for the foreign T-cell epitope. These experiments clearly show that B-cell tolerance is incomplete in normal mice. However, the detailed nature of the remaining autoreactive B-cell repertoire is not well understood. It is important to examine the naturally occurring autoreactive B-cell repertoire in normal mice for further understanding of B-cell tolerance.

Mouse lysozyme (ML) is a self-protein present in the blood, and the protein is widely distributed in body fluids and tissues.^8,9 In a foregoing study, we selected ML as a model self-protein, and prepared mutant mouse lysozyme (mutML) that has triple mutations (Ala114 to Asn, His115 to Arg, Gln117 to Lys), on the most potent T-cell epitope of HEL, HEL 107–116. mutML has both self-B-cell epitopes and a minimal foreign T-cell epitope (HEL 107–116) in the molecule. We found that mutML strongly induced a self-reactive antibody [anti-ML immunoglobulin G (IgG)] response in BALB/c mice.¹⁰ As T cells specific for HEL 107–116 were primed in these mice, the anti-ML IgG responses would be the result of collaborations between autoreactive B cells specific for ML and T cells specific for HEL 107–116. Surprisingly, mutML generated a 10⁴ times higher autoantibody titre than did the heterodimer of ML and HEL. Thus, we concluded that mutation in self-proteins may cause a high-titre autoantibody response if the mutated region can bind major histocompatibility complex (MHC) molecules and is recognized by T cells. The major epitope of anti-ML IgG was the ML 14–69 region – the original sequence of ML – and is separate from the mutated region. These results indicate that B cells responding to the epitope were not tolerant and were mainly involved in anti-ML IgG responses.

To characterize the nature of the remaining B-cell repertoire specific for ML, we prepared several monoclonal antibodies (mAbs) specific for the major epitope, ML 14–69, and examined their antigen specificities in detail. The mAbs were more specific for the denatured ML than for the native ML. The apparent affinity constants (K_A) of mAbs to native ML were determined to range from K_A = 3·0 × 10⁵ to 3·5 × 10⁶, while their constants to peptide ML 14–69 ranged from K_A = 5·1 × 10⁶ to 8·4 × 10⁷. The critical residue for mAb binding was glutamine at position 50 in the ML sequence. This residue is located on the surface loop region in native ML, but is surrounded by polar and bulky side-chains of neighbouring residues. The location of Gln50 in the native structure would explain why denaturation of ML strengthens the binding between anti-ML autoantibodies and ML. These results indicate that a substantial number of autoreactive B cells specific for the unfolded conformation of ML escape tolerance and produce autoantibodies to ML, when T-cell help is provided to them.

Materials and methods

Expression of ML and mutML in yeast

Construction of ML cDNA was carried out as described previously.¹⁰ Site-directed mutagenesis of ML cDNA was performed by polymerase chain reaction (PCR) (Expand High Fidelity PCR system; Boehringer Mannheim GmbH, Germany), using the megaprimer method.¹¹ Codon changes in positions 114 (GCA to AAC), 115 (CAC to AGA) and 117 (CAA to AAG) were introduced into ML cDNA using one oligonucleotide. Introduction of the desired mutations was confirmed by DNA sequencing. ML and mutML were expressed in yeast Pichia pastoris GS115 (Invitrogen, San Diego, CA), as described previously.¹⁰ Human lysozyme (HL) cDNA was constructed, as described previously,¹² and expressed in P. pastoris GS115 (Invitrogen). ML, mutML and HL were purified by cation-exchange chromatography (4·0 × 15 cm column) (CM-Toyopearl 650M; TOSOH, Tokyo, Japan), and the protein fraction was dialysed exhaustively against distilled water, then lyophilized. The purities of ML, mutML and HL were verified by sodium dodecyl sulphate–polyacryamine gel electrophoresis (SDS–PAGE), by mass spectrometry (MALDI-TOF/MS Voyager; PerSeptive Biosystems Inc., Framingham, MA) and by amino acid composition analysis (Hitachi 835 amino acid analyzer; Hitachi, Tokyo, Japan). Mass spectrometry was used to confirm that no glycosylation had occurred by the mutations. To prepare denatured ML, disulphide bonds were reduced with 2-mercaptoethanol (2-ME) and S-alkylated by using N-(3-bromopropyl)-N,N,N′,N′,N′-pentamethyl-1,3-propanedi(ammonium bromide) (TAP2-Br), as described previously.¹³

Expression of ML, R47A ML and Q50A ML in Escherichia coli

The cDNA fragment encoding ML was amplified by PCR using oligonucleotides, 5′-CGCAAGCTTCATATGTCTAAGGTCTATGAACGTTGTGAGT-3′ and 5′-GCCGAATTCAGACTCCGCAGTTCCGAA-3′. A codon change in position 47 (CGT to GCG) or in position 50 (CAA to GCG) was introduced into ML cDNA.¹¹ The constructed cDNAs were introduced into vector pET-22b(+). ML, R47A ML and Q50A ML were expressed in E. coli BL21. Isopropyl-β-d-thiogalactopyranoside was added to the culture at a final concentration of 1 mm. The culture was then further incubated at 37° for 4 hr. The cells were harvested by centrifugation and broken by a freeze–thaw process followed by sonication. The inclusion bodies containing the recombinant lysozyme were collected by centrifugation (2,000 g, 20 min, 4°). The wild-type ML and mutant ML thus obtained were dissolved in 0·58 m Tris–HCl (pH 8·6) containing 8 m urea and 5·25 mm ethylenediaminetetraacetic acid, reduced with 2-ME and subjected to gel filtration (Sephadex G-75; Amersham Pharmacia UK Ltd, Bucks, UK) with 10% aqueous acetic acid. The purified lysozyme was lyophilized, and the disulphide bonds reduced and S-alkylated using TAP2-Br.¹³

Preparation of peptides

Lysyl endopeptidase fragments of ML and HL were prepared, as described previously.¹⁰ Briefly, disulphide bonds were reduced with 2-ME and S-alkylated using TAP2-Br.¹³ Five milligrams of TAP2-ML was dissolved in 25 mm Tris–HCl (pH 9·0) containing 4 m urea, and 50 µg of lysyl endopeptidase (Wako, Osaka, Japan) was added to the solution. After 18 hr of incubation at 30°, the resultant peptides were separated by reverse-phase high-performance liquid chromatography (HPLC) on a column of Mightysil RP-18 GP (250 × 4·6 mm; Kanto Chemical Co., Tokyo, Japan) at a flow rate of 0·8 ml/min with a gradient of acetonitrile (containing 0·1% HCl) ranging from 1 to 40% over 100 min. To prepare tryptic peptides of ML, 5 mg of TAP2-ML was dissolved in 0·1 m phosphate buffer (pH 8·0), and 50 µg of l-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Worthington Biochemical Co., Lakewood, NJ) was added to the solution. After 2 hr of incubation at 40°, the resultant peptides were separated by reverse-phase HPLC on a column of Mightysil RP-18 GP (250 × 4·6 mm) at a flow rate of 0·8 ml/min with a gradient of acetonitrile (containing 0·1% HCl) ranging from 1 to 40% over 100 min. Purity and concentration of peptides were analyzed by amino acid analysis (Hitachi 835 amino acid analyzer; Hitachi) after hydrolysis in 6 m HCl, under vacuum at 110° for 20 hr.

CD measurements

Far-UV circular dichroic (CD) spectra were recorded on a Jasco J-720 spectropolarimeter at 25° and pH 7·2 in a 0·1-cm path-length quartz cell (Hellma, Forest Hills, NY), at a protein concentration of 10 µm. The temperature of the cell was controlled by circulating water in a jacket around the cell to within ± 0·1°. Results were expressed as the mean residue ellipticity (molecular ellipticity), which is defined as:

where θ_obs is the observed ellipticity (in degrees), c is the concentration (moles/ l) in residue, and l is the length (cm) of the light-pass. Spectra were recorded with 0·1-nm resolution and data represent the average of 10 scans. The final spectra were corrected by subtracting the corresponding baseline spectrum obtained under identical conditions.

Immunization

BALB/c mice of either sex, 8–12 weeks of age, were obtained from Japan SLC Inc. (Shizuoka, Japan). Group of five mice were immunized by giving subcutaneous (s.c.) injections of 3·5 nmol of mutML or HEL, with compete Freund's adjuvant (CFA) emulsion (Sigma, St. Louis, MO). Each mouse was bled on days 9 and 28 from the orbital sinus into capillary tubes. In another experiment, five BALB/c mice were injected, three times weekly for 2 weeks, intraperitoneally (i.p.) with 3·5 nmol of mutML dissolved in phosphate-buffered saline (PBS), and the mouse sera were obtained on day 21. Sera were stored at −20° until used. All experiments involving the use of mice were carried out in accordance with protocols approved by the Animal Care and Use Committee of Graduate School of Pharmaceutical Sciences, Kyushu University.

ML-specific mAb production and characterization

Five BALB/c mice were immunized s.c. into both hind footpads with 3·5 nmol of mutML emulsified 1 : 1 in CFA. Eight days later, these mice were killed and their popliteal lymph nodes removed. Pooled lymph node cells were immortalized by fusion with P3X63Ag8U.1 myeloma cells (American Type Culture Collection, Manassas, VA) using polyethylene glycol.¹⁴ The fused cells were cultured in Dulbecco's modified Eagle's minimal essential medium (DMEM; Sigma) supplemented with l-glutamine (216 µg/ml), l-asparagine (36 µg/ml), l-arginine–HCl (116 µg/ml), folic acid (6 µg/ml), HEPES (10 mm), 2-mercaptoethanol (5 × 10⁻⁵ m), penicillin (100 U/ml), streptomycin (100 µg/ml), 10% heat-inactivated fetal calf serum (FCS) (Bio Whittaker, Walkersville, MD) and 10% hybridoma cloning factor (IGEN Inc., Gaithersburg, MD), in a humidified atmosphere at 37° and 5% CO₂. Anti-ML Abs producing hybridomas were screened by solid-phase enzyme-linked immunosorbent assay (ELISA) (see below), subcloned in limiting dilution, expanded and placed in liquid nitrogen for later use. Antibodies present in the culture supernatant of cloned hybridomas were examined for antigen specificity.

ELISA

To characterize anti-ML mAbs, ELISA plates (Nunc, Roskilde, Denmark) were coated overnight at 4° with 50 µl of native ML and peptides of ML dissolved in 0·1 m carbonate buffer (pH 9·6) at 2 µg/ml and 10 µg/ml, respectively. Plates were washed three times with PBS containing 0·05% Tween-20 (PBST), and residual binding sites were blocked overnight at 4° with 100 µl of 2% non-fat dry milk in PBST (blocking buffer). After washing with PBST, the culture fluids of cloned hybridoma diluted 1 : 100 in blocking buffer were added to each well, followed by incubation for 1 hr at room temperature. After washing with PBST, 50 µl of alkaline phosphatase-conjugated goat anti-mouse IgG, IgM or IgG1 (Zymed, San Francisco, CA), diluted 1 : 1000 in blocking buffer, was added to each well, followed by incubation for 1 hr. After washing with PBST, the final reaction was visualized by incubation with p-nitrophenyl phosphate (Wako) in 0·1 m carbonate buffer (pH 9·6) containing 1 mm MgCl₂. The absorbance was measured at 405 nm (A₄₀₅).

In inhibition experiments, ELISA plates were coated overnight at 4° with 50 µl of native ML dissolved in carbonate buffer to a final concentration of 2 µg/ml. After washing and blocking, 1 : 10 000-diluted culture supernatant of cloned hybridomas, preincubated with serial dilutions of native ML, denatured ML, peptide ML 14–69 or native HEL, were added to each well, followed by incubation for 1 hr at room temperature. After washing with PBST, 50 µl of alkaline phosphatase-conjugated goat anti-mouse IgG, diluted 1 : 1000 in blocking buffer, was added. After washing with PBST, the final reaction was visualized by incubation with p-nitrophenyl phosphate. The absorbance was measured at 405 nm.

Affinity measurements

The mAb concentrations in culture supernatant were determined by capture ELISA. Briefly, microtitre plates were coated with goat anti-mouse µ, γ, and L chain (Tago, Burlingame, CA), followed by addition of serial dilutions of the mAb culture fluids. The captured mAbs were detected by alkaline phosphatase-conjugated goat anti-mouse IgG. Concentrations were standardized to the concentration of affinity-purified anti-HEL IgG, which was determined spectrophotometrically by absorbance at 280 nm, using an extinction coefficient (1 mg/ml) of 1·4. The association rate constants (k_ass.), dissociation rate constants (k_diss.), and K_A of the mAbs were determined by surface plasmon resonance, using IAsis (Affinity Sensors, Cambridge, UK). ML or peptide ML 14–69 was coupled to the carboxymethyl dextran on the cuvette surface at a concentration of 3·5 µm in 10 mm sodium acetate, pH 6·5, using coupling agents N-hydroxysuccinimide and N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride followed by blocking with 1 m ethanolamine, pH 8·5. Binding measurements were obtained using concentrations of mAbs from 80 µg/ml to 0·31 µg/ml diluted with 10 mm sodium phosphate buffer (pH 7·4) containing 2·7 mm KCl, 138 mm NaCl and 0·05% Tween-20. HCl 100 mm was used as a regeneration buffer. Data were analyzed using iasis fast-fit software (Affinity sensors).

Results

Characterization of mutML

The amino acid sequence of ML, HEL and HL, and an illustration of mutML, are shown in Fig. 1. We mutated amino acids of ML at Ala114, His115 and Gln117 into Asn, Arg and Lys, respectively, and the core immunodominant T-cell epitope of HEL 107–116 restricted to I-E^d MHC class II molecule was introduced into ML. The conformation of the B-cell epitope in mutML was almost similar to that of native ML, as determined using affinity-purified rabbit anti-ML polyclonal IgG (Fig. 2a). mutML retained similar enzymatic activity to that seen in the native ML (data not shown). Far-UV CD spectra showed that mutML took on the native-like secondary structure (Fig. 2b). Therefore, a drastic change in global conformation does not occur in mutML when triple amino acid substitution occurs at positions 114, 115 and 117 in ML.

Figure 1.

(a)Comparison of the amino acid sequence of mouse lysozyme (ML), hen egg lysozyme (HEL) and human lysozyme (HL). The identical amino acid residue is indicated by (−) and deletion by (*). Only the amino acids of HEL and HL that differ from those of ML in sequence are shown. Enzymatically relevant residues are surrounded by boxes. (b) Illustrations of ML and mutML. mutML has the minimal immunogenic T-cell epitope HEL 107–116 and the immunogenic B-cell epitope ML 14–69.

Figure 2.

Conformation of B-cell epitope and secondary structure of mutant mouse lysozyme (mutML). (a) The antigenicity of mutML was evaluated by a competitive enzyme-linked immunosorbent assay (ELISA). Serially diluted native ML, mutML and denatured ML were added to ELISA plates (precoated with native ML), in the presence of affinity-purified rabbit anti-ML immunoglobulin G (IgG). The absorbance was measured at 405 nm (A₄₀₅). (b) Far-ultraviolet circular dichroic (CD) spectra of native ML, mutML and denatured ML at 25°. Spectra were measured at 10 µm in 50 mm phosphate buffer, pH 7·2.

Preparation of mAbs specific for major epitope on mouse lysozyme

We previously reported that mutML emulsified in CFA evoked an HEL 107–116-specific T-cell response and a high-titre anti-ML IgG autoantibody response in BALB/c mice (I-A^d, I-E^d), while native ML plus CFA did not induce such autoantibody responses.¹⁰ Thus, the collaboration between T cells specific for HEL and autoreactive B cells specific for ML occurred in these mice. Serum IgG responses against ML were dominantly focused on ML 14–69, when mice were immunized with mutML emulsified in CFA (Table 1). When mice were immunized with HEL plus CFA, no anti-ML Ab and no anti-ML peptide Ab responses were induced (Table 1). A dominant anti-ML 14–69 response was not derived from the use of CFA plus antigen emulsion, because mutML dissolved in PBS could evoke a similar autoantibody response against ML 14–69 (Table 1). These results indicate that an anti-ML B-cell repertoire specific for the ML 14–69 region dominantly escapes tolerance.

Table 1. Reactivity of mutant mouse lysozyme (mutML) and hen-egg lysozyme (HEL) polyclonal immunoglobulin G (IgG) antibodies against native ML and ML peptides.

	Antigen
Polyclonal immune sera	ML	ML 1–13	ML 14–69	ML 70–97	ML 98–130
Day 9 (HEL + CFA)	−	−	−	−	−
Day 28 (HEL + CFA)	−	−	−	−	−
Day 9 (mutML + CFA)	++	−	++	−	−
Day 28 (mutML + CFA)	+++	−	+++	+	+
Day 21 (mutML in PBS)	++	−	++	−	−

Sera obtained from five BALB/c mice immunized with HEL emulsified in complete Freund's adjuvant (CFA), with mutML emulsified in CFA (on days 9 and 28 after immunization) and with mutML dissolved in phosphate-buffered saline (PBS) (on day 21) were tested for reactivity against native ML and lysyl endopeptidase-digested ML peptides. Sera diluted 1 : 5 (CFA on day 9 and PBS on day 21) and 1 : 100 (CFA on day 28) were used. The mean values of duplicate wells were taken.

The reactivity of polyclonal immunoglobulin G (IgG) to each antigen was graded as follows: A₄₀₅ of < 0·1, −; 0·1−0·5, +; 0·5−1·0, ++; and >1·0, +++.

To understand the nature of naturally occurring ML-specific autoreactive B cells, we prepared mAbs specific for ML. Nine B-cell hybridoma cells specific for ML could be cloned and their antigen specificities are summarized in Table 2. All mAbs were IgG1 and showed a strong reactivity against native ML adherent to ELISA plates. Seven out of the nine mAbs were specific for peptide ML 14–69. The majority of anti-ML mAbs specific for ML 14–69 in established anti-ML mAbs suggest that ML 14–69 is the major epitope on ML, even at a clonal level. Similar results were obtained when the same protocol was performed at a different time (data not shown). Anti-ML mAbs specific for ML 14–69 did not react with HEL, denatured HEL or HEL 98–116 (data not shown). The remaining two anti-ML mAbs, 4E11D and 4H8E, hardly bound to peptides of ML, but strongly to native ML, indicating that these mAbs recognized conformational epitopes.

Table 2. Reactivity of anti-mouse lysozyme (ML) monoclonal antibodies (mAbs) against native ML, denatured ML and ML peptides.

	Antigen
mAb	ML	Denat. ML	ML 1–13	ML 14–69	ML 70–97	ML 98–130
8F7A	+++	+++	−	+++	−	−
8F10C	+++	+++	−	+++	−	−
3G3A	+++	+++	−	+++	−	−
3G1G	+++	+++	−	+++	−	−
7C5C	+++	+++	−	+++	−	−
7C5A	+++	+++	−	+++	−	−
9G9C	+++	+++	−	+++	−	−
4E11D	+++	−	−	−	−	−
4H8E	+++	+	−	+	−	+

Nine anti-ML immunoglobulin G1 (IgG1) mAbs were established and their reactivity to ML, denatured ML (denat. ML) and lysyl endopeptidase-digested ML peptides were measured using solid-phase enzyme-linked immunosorbent assay (ELISA). The mean values of duplicate wells were used.

The reactivity of Ab to each antigen was graded as follows: A₄₀₅ of < 0·1, −; 0·1−0·5, +; 0·5−1·0, ++; and >1·0, +++.

The results are representative data of two independent experiments.

Effect of denaturation of ML on antibody binding

We next examined the precise interaction between mAbs specific for the major epitope (ML 14–69) and ML by solution-phase ELISA. In a competitive binding assay using soluble native ML, native mutML, denatured ML and peptide ML 14–69, mAb 8F7A showed weak interactions with native ML and native mutML, whereas the mAb showed strong interactions with denatured ML and peptide ML 14–69 (Fig. 3). The binding between mAb and denatured ML was specific, as control HEL and denatured HEL had no inhibitory effect on the binding (Fig. 3). Similar results were obtained from the same experiments using six other mAbs specific for ML 14–69 (data not shown). These results indicate that anti-ML 14–69 mAbs are more specific for the unfolded structure of ML than for the native structure.

Figure 3.

Evaluation of the antigen specificity of anti-mouse lysozyme (anti-ML) monoclonal antibody (mAb) 8F7A by competitive enzyme-linked immunosorbent assay (ELISA). Serially diluted native ML, mutant mouse lysozyme (mutML), denatured ML, peptide ML 14–69, native hen-egg lysozyme (HEL) and denatured HEL were added to ELISA plates precoated with native ML in the presence of anti-ML mAb 8F7A, diluted 1 : 10 000. The absorbance was measured at 405 nm (A₄₀₅).

Measurement of affinity of anti-ML IgG by surface plasmon resonance

To analyze the interaction between ML and anti-ML mAb in greater detail, k_ass. and k_diss. rates of mAbs to ML were measured by surface plasmon resonance, and K_A values were calculated directly from these rates. Native ML or peptide ML 14–69 was covalently attached to the cuvette surface and each intact monoclonal IgG was added to the cuvette. As reported by Mueller & Jemmerson,¹⁵k_diss. values were lower for intact monoclonal IgG than for the Fab fragment, but k_ass. values were similar. As intact monoclonal IgG was used instead of Fab fragment in this experiment, the k_diss. values obtained in our study may be underestimated. The sensogram of anti-ML mAb (8F7A) association with and dissociation from native ML and peptide ML 14–69 is shown in Fig. 4. The surface plasmon resonance of 8F7A association with peptide ML 14–69 was significantly different from that found with native ML. The lower response between 8F7A and native ML was specific, because anti-HEL mAb showed no interaction with either native ML or ML 14–69. The k_ass., k_diss. and K_A values of mAbs against native ML and peptide ML 14–69 are summarized in Table 3. The apparent affinity constants of mAbs against native ML ranged from 3·0 × 10⁵ to 3·5 × 10⁶, whereas those against peptides ML 14–69 ranged from 5·1 × 10⁶ to 8·4 × 10⁷. The apparent affinity constants between mAbs and peptide ML 14–69 were six- to 80-fold higher than those between mAbs and native ML.

Figure 4.

Sensogram of anti-mouse lysozyme (anti-ML) monoclonal antibody (mAb) 8F7A association with and dissociation from native ML and peptide ML 14–69, as measured by surface plasmon resonance. ML or peptide ML 14–69 was immobilized to the cuvette and incubated with 133 nm anti-ML mAb 8F7A or control anti-hen-egg lysozyme (HEL) mAb diluted with 10 mm sodium phosphate buffer (pH 7·4) containing 2·7 mm KCl, 138 mm NaCl and 0·05% Tween-20 (PBST). The association kinetics was followed for 20 min prior to washing with the PBST to follow the time-course of dissociation. Sharp swings in signals during the change of buffers reflect the refractive index difference of buffers and are not caused by binding/unbinding of the mAb. An enlarged figure, which represents low but certain interactions between 8F7A and native ML, is shown on the right.

Table 3. Kinetic parameters of interactions between anti-mouse lysozyme (ML) 14–69 monoclonal antibodies (mAbs) and native ML and between anti-ML 14–69 mAbs and peptide ML 14–69.

	Native ML			Peptide ML 14–69
mAb	kass. (104/m/second)	kdiss. (10−3/second)	KA (10−7/m)	kass. (104/m/second)	kdiss. (10−3/second)	KA (107/m)
8F7A	1·0	10	0·10	5·0	0·59	8·4
8F10C	2·1	6·0	0·35	2·9	1·3	2·2
3G3A	0·28	9·8	0·030	1·7	1·4	1·3
3G1G	0·55	9·6	0·060	1·1	0·56	2·0
7C5C	1·3	10	0·13	1·2	1·4	0·84
7C5A	0·39	6·1	0·060	2·5	0·67	3·7
9G9C	ND	ND	ND	0·87	1·7	0·51

The association rate constant kinetic parameter, k_ass., was the slope of the association plot using the k_on value at different mAb concentrations. The disassociation rate constant kinetic parameter, k_diss., was the intercept of the association plot. The apparent affinity constant, K_A, was the value k_ass. ÷ k_diss.. The values obtained for the same mAb from experiments performed on a different day varied with an average of less than 15%.

ND, not determined.

Determination of the precise epitope recognized by mAbs specific for ML 14–69

The dominant B-cell epitope was located in ML 14–69 (Table 1) and denaturation of ML may facilitate exposure of some critical residues from the folded conformation (Fig. 3). To determine more precise epitope regions, peptide ML 14–69 was further digested with trypsin and the binding capacity of mAbs to the tryptic peptides was evaluated. All seven mAbs specific for peptide ML 14–69 reacted strongly to peptide ML 48–62 (data not shown). To determine the critical residues for mAb binding within ML 48–62, we used HL, which has a 79% amino acid sequence identity with ML (Fig. 1a). The amino acid sequence of ML 48–62 differs from HL 48–62 only at position 50 (Fig. 1a). When reactivities of the anti-ML mAbs to denatured HL and peptide HL 35–69 were examined, no interactions were observed (data not shown). The results indicated the significant contribution of Gln50 to autoantibody recognition. To test this possibility, we prepared mutant ML, substituting Ala for Gln50 (Q50A ML), and evaluated the importance of Gln50 for mAb binding. In addition, from a comparison between ML and HL sequences (Fig. 1a), another mutant ML (R47A ML) was also prepared to assess the role of different positions on mAb binding. As shown in Fig. 5, the inhibitory activity of denatured Q50A ML for binding between immobilized native ML and mAb 8F7A was about 200-fold less than that of denatured ML and R47A ML. Similar experiments using other mAbs and antisera gave similar results (data not shown). The results clearly show that Gln50 in ML is one of the critical residues for binding of autoantibodies induced by mutML.

Figure 5.

Determination of the precise epitope on mouse lysozyme (ML). Serially diluted denatured ML, denatured R47A ML and denatured Q50A ML were added to enzyme-linked immunosorbent assay (ELISA) plates, precoated with native ML, in the presence of anti-ML monoclonal antibody (mAb) (8F7A) diluted 1 : 10 000. The absorbance was measured at 405 nm (A₄₀₅).

Discussion

When BALB/c mice were immunized with mutML, high-titre autoantibody responses against ML 14–69 were detected in the sera (Table 1). We previously showed that autoantibody responses against ML 14–69 and ML 98–130 were also induced when B10.A and C3H/HeN mice were immunized with another mutant ML carrying the dominant T-cell epitope, HEL 50–61.¹⁷ Thus, incomplete B-cell tolerance against ML is not limited to BALB/c mice; rather it is seen in other strains of mice, and autoreactive B cells specific for ML can be activated when T-cell help is provided.

Most antibodies elicited by a globular native protein were thought to recognize antigenic determinants that reflect the native structure of the protein, because they reacted poorly with denatured forms of the Ag^18,19 and peptide fragments were poor substitutes for the native antigen.^20,21 Using polyclonal sera from BALB/c mice immunized with foreign Ag (native HEL plus CFA), we tested the reactivity of polyclonal anti-HEL sera to lysylendopeptidase fragment of HEL, i.e. HEL 1–13, HEL 14–33, HEL 34–97, HEL 98–116 and HEL 117–129. However, we could not detect any binding of antibodies against HEL peptides. These results indicate that strong anti-peptide Ab responses rarely occurred when mice were immunized with conformational lysozyme. In contrast, polyclonal autoantibodies induced by native mutML reacted strongly with peptide ML 14–69 (Table 1). In addition, among the nine B-cell hybridoma clones that responded strongly to native ML attached to ELISA plates, seven showed strong reactivity to denatured ML and peptide ML 14–69, while only two recognized the conformational epitope of ML (Table 2). mutML retained enzymatic activity similar to ML, indicating that denaturation of mutML did not occur by the inserted mutation. Denaturation of immunogen (mutML) as a result of emulsification in CFA could not explain the phenomenon, because aqueous native mutML solution evoked similar responses (Table 1). Why were dominant anti-denatured or anti-peptide autoantibody responses induced by immunization with native mutML? We do not have any data to explain the phenomenon, but we consider that different tolerance status among B cells specific for native ML and B cells specific for denatured ML may explain such an event. As the native form of ML should be more abundant than the denatured form in body fluid, B cells specific for the native form may be more strongly anergized than those specific for the denatured form, this thinking being based on the results of double transgenic mice.¹ Therefore, most B cells specific for a conformational epitope of ML are hardly involved in Ab responses to mutML, whereas B cells specific for the denatured ML may escape tolerance and can dominantly participate in the Ab production to mutML. mAbs specific for denatured forms of ML showed low, but certain, interaction with native mutML (Fig. 3). This observation indicates that autoreactive B cells specific for the denatured form of ML can recognize native mutant ML with low affinity, and they can present the T-helper epitope and receive help from HEL-specific T cells. More importantly, in vivo denaturation of mutant ML could allow the strong interaction between mutant ML and such B cells.

Established anti-ML mAbs specific for ML 14–69 were directed against a single epitope, ML 48–62, and are sensitive to a single amino acid substitution from glutamine to alanine at position 50 (Fig. 5). These results indicate that Gln50 in ML is a critical residue for binding with anti-ML mAbs. A computer model of ML indicates that Gln50 is located on the surface loop between the β-sheet and is exposed to the solvent at 48·7% in the folded conformation of ML (Fig. 6b). In this regard, it seems that anti-ML mAbs can interact with both native and denatured ML. However, as shown in Fig. 3, anti-ML mAbs interacted much more strongly with denatured ML than with native ML. It is not clear why the binding affinity is increased by the denaturation of ML in this state. One of the possible explanations for this is steric hindrance by other residues around Gln50 in folded ML, because Gln50 is suggested to be surrounded by polar and bulky side-chains of Tyr45, Arg47, Arg62, Lys69 and Pro71 in the native structure (Fig. 6a). Denaturation may result in conformational change around Gln50, allowing exposure of Gln50 to solvent and a strong interaction between anti-ML autoantibodies and ML.

Figure 6.

Computer-generated space-filling model of mouse lysozyme (ML). The ML molecule was derived from the X-ray structure of human lysozyme (HL)¹⁶ using the program Insight-Discover. Sequence identity between HL and ML is 79%. The residues of HL that differ from ML were replaced with that of ML. Energy minimization was performed at pH 4·5 after addition of H₂O within 5Å layers around the molecule until the maximum derivative became less than 0.1 kcal/mol.Ų. (a) Top view; (b)side view.

The autoreactive B-cell repertoire to mouse cytochrome c (MCyt c) in normal mice has been studied in detail by Jemmerson's group.^22,23 In their experiments, using the conjugate of MCyt c and ovalbumin as an immunogen, major autoantibodies against MCyt c were specific for the native structure and interacted strongly with soluble, native MCyt c.²² However, the antigen specificities of major autoantibodies to ML differed. Major autoantibodies against ML were specific for the denatured structure and interacted weakly with soluble, native ML (Fig. 3). The different specificities of major autoantibodies between MCyt c and ML can be explained by different Ag concentrations in blood. As MCyt c is an intracellular protein, while ML mainly exists extracellularly, the serum concentration of ML should be higher than that of MCyt c, and B-cell tolerance to ML may be more profound than that to MCyt c. We examined the concentration of ML in serum by sandwich ELISA using rabbit anti-ML polyclonal Abs and mouse anti-ML mAb 4E11D; the results are shown in Table 2. The serum concentration of ML was ≈460 ± 11 ng/ml, a concentration which is sufficient to induce B-cell tolerance.²⁴ In our study, a minor population of mAbs, which bound strongly to native ML, were also obtained (Table 2). The apparent K_A values of 4E11D and 4H8E against native ML were 2·0 × 10⁸ and 8·9 × 10⁷, respectively. We were unable to obtain more detailed information on the nature of 4E11D and 4H8E in this study. The epitope recognized by the mAbs is under investigation.

In double-transgenic mice expressing soluble HEL as a neo-self protein, low-affinity B cells specific for HEL escape tolerance.¹ As anti-ML 14–69 mAbs interacted weakly with soluble, native ML (Fig. 3), and the average K_A value of established anti-ML mAb to native ML was low (1·2 × 10⁶) (Table 3), B cells specific for ML 14–69 would escape tolerance as a result of low interactions with native ML. Low-affinity autoantibodies produced by such B cells seem harmless to individuals. However, some recent reports indicate that B cells specific for denatured self-proteins may play a key role in autoimmunity. Autoantibodies to complement components were detected in serum from patients with systemic lupus erythematosus (SLE) and other hypocomplementaemic states.²⁵ In SLE, the occurrence of anti-C1q antibodies is strongly associated with a decline in renal function, and rises in the titre of anti-C1q autoantibodies can predict renal relapses in SLE. Various studies have shown that anti-C1q antibodies in diseases such as SLE react primarily with solid-phase C1q and do not react with solution-phase C1q. The results indicate that B cells specific for denatured C1q would be involved in the anti-C1q autoantibody response, and the autoantibodies produced by these B cells would be unfavourable factors. In SLE patients, anti-peptide autoantibody responses were also found against nuclear ribonucleoprotein Ro60²⁶ and apolipoprotein A-I.²⁷ These results also indicate the implication of B cells specific for denatured self-protein in autoimmune diseases.

In conclusion, we showed that the majority of anti-ML autoantibodies were more specific for denatured ML than for native ML. These results indicate that B cells specific for the unfolded conformation of ML escape tolerance and participate mainly in the autoantibody response to ML. Our findings contribute to a better understanding of B-cell tolerance and to the type of B-cell repertoire involved in autoantibody production.

Abbreviations

Ag

antigen

CFA

complete Freund's adjuvant

ELISA

enzyme-linked immunosorbent assay

HEL

hen-egg lysozyme

HL

human lysozyme

K_A

affinity constant (k_ass/k_diss)

k_ass.

association rate constant

k_diss.

dissociation rate constant

mAbs

monoclonal antibodies

ML

mouse lysozyme

mutML

mutant ML with triple mutations rendering the most potent T-cell epitope of HEL (sequence 107–116)

PBS

phosphate-buffered saline

PBST

PBS containing 0·05% Tween-20

Q50A ML

mutant ML with a single mutation at position 50 in ML (Gln to Ala)

R47A ML

mutant ML with a single mutation at position 47 in ML (Arg to Ala)