Monoclonal antibodies specific for Shigella flexneri

Vol. 53, No. 1

INFECTION AND IMMUNITY, JUlY 1986, P. 103-109 0019-9567/86/070103-07$02.00/0 Copyright ©D 1986, American Society for Microbiology

Monoclonal Antibodies Specific for Shigella flexneri Lipopolysaccharides: Clones Binding to Type I and Type 111:6,7,8 Antigens, Group 6 Antigen, and a Core Epitope NILS

I.

A. CARLIN' AND ALF A. LINDBERG2*

Department of Bacteriology, National Bacteriological Laboratory, S-105 21 Stockholm,1 and Karolinska Institute, Department of Clinical Bacteriology, Huddinge University Hospital, S-141 86 Huddinge,2 Sweden Received 25 November 1985/Accepted 2 April 1986

Monoclonal antibodies against the Shigellaflexneri lipopolysaccharide (LPS) were generated in two fusions by using the myeloma cell line Sp2/0 as a fusion partner with spleen cells from BALB/c mice immunized with S. flexneri serotypes lb and 3a bacteria. The antibodies were characterized by immunoblotting, enzyme-linked immunosorbent assay (ELISA), hemagglutination, and coagglutination. Four different types of monoclonal antibodies were isolated: (i) antibodies specific for the core antigen of the LPS, (ii) antibodies specific for the type I 0 antigen, (iii) antibodies specific for the group 6 0 antigen, and (iv) antibodies specific for the type 111:6,7,8 0 antigen. The core-specific antibodies were shown to be specific for the Escherichia coli R3 core, which all S. flexneri LPSs tested, except for S. flexneri serotype 6 LPS, have. The type I 0 antigen-specific antibodies were shown to bind exclusively to S. flexneri serotypes la and lb in ELISA. The type 111:6,7,8 0-antigen-specific antibodies were specific for S. flexneri serotype 3a in ELISA and hemagglutination. Two different group 6 0-antigen-specific antibodies were bound. One was bound in both ELISA and hemagglutination to LPSs of S. flexneri serotypes lb, 3a, 3b, and 4b, whereas the second was bound only to LPSs of serotypes 3a, 3b, and 4b in ELISA but to LPSs of all four serotypes in hemagglutination. The specificity of the isolated I, III:6,7,8, and group 6 monoclonal antibodies was verified by coagglutination of 363 S. flexneri clinical isolates.

and for the saccharide present in the complete core (MASF R3 core-1).

Shigella flexneri serotyping is based on the specificity which resides in the cell envelope 0 antigen. This is a part of the lipopolysaccharide (LPS) molecule, a major structural element in the outer membrane of gram-negative bacteria. The LPS consists of three parts, lipid A, the core, and the 0-antigen chain. Lipid A, anchoring the LPS molecule in the membrane, is linked to the core by an acid-labile 2-keto-3deoxyoctulosonic acid. The core is a single sequence of heptoses and hexoses, whereas the 0-antigen chain in S. flexneri is a polysaccharide built by a repetitive sequence of hexoses. Much interest has been focused on the structure of the S. flexneri 0 antigens (1, 37). Kenne et al. concluded that the basic repeating unit for all S. flexneri serotypes (except for serotype 6 [12]) is a tetrasaccharide: -2)-ct-L-Rhap-(1-2)-a-LRhap-(1-3)-a-L-Rhap-(1-3)-p-D-GlcpNAc-(1-. This structure is identical to the subserotype Y. The other serotypes are the result of substitutions of the repeating unit with a-D-glucosyl residues or O-acetyl groups or both, which add new or mask existing antigenic determinants (25-27) (Fig. 1). The close structural resemblance among different serotypes of S. flexneri readily explains the problems of raising specific antisera against a single antigenic determinant. Antisera are produced routinely by immunizing rabbits with whole heat-killed S. flexneri (14). The resulting immune sera are then absorbed with S. flexneri bacteria of cross-reacting serotypes. However, the resulting absorbed antisera still often show cross-reactivity or are of a low titer. The present paper reports on the production of characterization of monoclonal antibodies specific for S. flexneri type I 0 antigen (MASF I), type 111:6,7,8 0 antigen (MASF 111:6,7,8), and group 6 0 antigen (MASF 6-1 and MASF 6-2) *

MATERIALS AND METHODS Bacterial strains. Plesiomonas shigelloides (NBL 650), S. flexneri serotype la, lb, 2a, 3a, 4a, 4b, Sa, 5b, X, Y, 6, and 4bR (22), and Staphylococcus aureus (Cowan 1) were from the collection at the Department of Bacteriology, National Bacteriological Laboratory, Stockholm, Sweden. S. flexneri serotype 3b (a bacteriophage Sf6-lysogenized subserotype Y strain) was available from previous investigations (19, 31). For coagglutination, S. flexneri clinical isolates as well as strains from the collection at the Department of Bacteriology, National Bacteriological Laboratory, were used. Preparation of LPS. Bacteria were grown in submerged cultures, and LPSs were extracted with hot phenol-water (42) for smooth bacteria and with phenol-chloroform-light petroleum (17) for rough bacteria. The chemical characterization of the rough LPS used has been described elsewhere (21). Smooth S. flexneri LPS was subjected to sugar analysis as described by Sawardeker et al. (36). For proton nuclear magnetic resonance spectroscopy, a JEOL GX 270 instrument operated in the pulse Fourier transformation (PFT) mode was used. The spectra were recorded for solutions in D20 at 70 or 85°C with either tetramethylsilane as the external standard or H20 as the internal standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was run essentially as described previously (20, 30, 34). The separation gel (142 by 120 mm; 1.5-mm thick) contained 15% acrylamide and 0.45% bisacrylamide. The stacking gel (142 by 40 mm) contained 4.5% acrylamide and 1.2% bisacrylamide. Samples (5 ,ug per well) were heated for 5 min at 100°C in a buffer consisting of 0.0625 M Tris hydrochloride buffer (pH 6.8) containing 2%

Corresponding author. 103

104

INFECT. IMMUN.

CARLIN AND LINDBERG

Basic structure

1R

-(

3GIcNAc-

alh

p

I Substitutents

0

11

III

*acetyl group Glucosyl residue

immunoglobulin conjugate (diluted 1/400 in TBS-T) for 90 min at 20°C on a rocker. After being washed as described above, the NC paper was developed as described by Blake et al. (2) by a final wash in 0.15 M Veronal-acetate buffer (pH 9.6) for 5 min at 37°C and then incubated in a substrate buffer consisting of 0.15 M Veronal-acetate buffer (pH 9.6) containing 0.01% (wt/vol) Nitro Blue Tetrazolium (Sigma Chemical Co., St. Louis, Mo.), 0.005% 5-bromo-4-chloro-3-indolyl phosphage (Sigma), and 0.004 M MgCI2 at 37°C for 3 to 5 min.

la

2a

03a

lb

002b0

3

5a 00 05b

OX

Y

FIG. 1. S. flexneri 0-antigen structures. Abbreviations: Rha, rhamnose; GIcNAc, N-acetylglucosamine.

SDS, 10% glycerol, 2% mercaptoethanol, 0.1% disodium EDTA, and 0.001% phenol red as a tracking dye. Electrophoresis was run at a constant current of 40 mA until the tracking dye was approximately 2 mm from the end of the gel. Gels were stained by the silver method (39). Immunoblotting. Electrophoretic transfer from SDSPAGE gels was performed immediately after the electrophoresis was finished. The transfer was done by the method of Towbin et al. (38) in a buffer (pH 8.4) consisting of 20 mM Tris, 152 mM glycerin, and 20% (vol/vol) methanol for 3.5 h at 220 mA with 0.45-gm-pore-size nitrocellulose (NC) paper (BA 85; Schleicher & Schuell, Inc., Keene, N.H.) in a Trans-Blot system (Bio-Rad Laboratories, Richmond, Calif.). After transfer, the NC paper was incubated in 0.01 M Tris hydrochloride-0.15 M NaCl (pH 7.4)-0.05% Tween 20 (TBS-T) (2, 10) at 20°C for 1 h to coat residual binding sites. Next, the NC paper was incubated with the first antibody (ascitic fluid diluted 1/200 in TBS-T) at 4°C overnight on a rocker. After being washed in TBS-T, the NC paper was incubated with an alkaline phosphatase-rabbit anti-mouse

Cell culture medium. The cell culture medium for standard growth for hypoxanthine-aminopterin-thymidine selection and growth in serum-free supplemented medium was as described previously (5, 9). Immunization of mice. Female BALB/c mice were immunized on days 0 (intraperitoneally) and 21 (intravenously) with 108 heat-killed S. flexneri (suspended in an equal volume of Freund complete adjuvant for day-0 immunization and in an equal volume of phosphate-buffered saline [PBS] for day-21 immunization). Fusions were performed 4 days after the last immunization. Production of hybridomas. The procedure was essentially that of Kohler and Milstein (28) with modifications from De St. Groth and Scheidegger (11). Briefly, approximately 108 spleen cells from three pooled spleens of S. flexneriimmunized mice were fused with 5 x 107 Sp2/0-Ag 14 cells by using polyethylene glycol 4000 (E. Merck AG, Darmstadt, Federal Republic of Germany). After fusion, cells were grown in hypoxanthine-aminopterin-thymidine (Sigma) selection medium (32) in 96-well microtiter plates. Screening and selection of hybridomas. Putative hybrids were tested in an enzyme-linked immunosorbent assay (ELISA) with the homologous LPS as antigen (coating dose, 10 ,ug/ml). Positive clones (A405 at 100 min, >1.0) were further tested against a panel of S. flexneri LPSs of different serotypes (S. flexneri la, lb, 3a, 3b, 4b, X, and 4a) and against wells with only coating buffer and bovine serum albumin. Hybridomas secreting antibody of the desired specificity were recloned twice or until 100% cloning efficiency was accomplished by limiting dilution. Cell lines selected in this way were grown as ascites tumors in pristane-treated (33) BALB/c mice. Ascites fluids were tested in ELISA endpoint titration against a set of 15 different antigens, including LPSs from all S. flexneri serotypes, rough S. flexneri LPS, and LPS representing S. sonnei phase 1 antigen (P. shigelloides LPS) (15, 35). ELISA. The ELISA has been described in detail previously (5, 16, 40, 41). Briefly, serum, culture fluid, or ascites fluid diluted in PBS-0.05% Tween 20 (PBS-T) was added to washed (0.15 M NaCl-0.05% Tween 20) 96-well flatbottomed microtiter plates (Nunc, Roskilde, Denmark) that had previously been coated overnight at 20°C with 0.1 ml of LPS (10 ,ug/ml) in 0.05 M carbonate buffer (pH 9.6) followed by 1% bovine serum albumin in the same buffer. The first antibody incubation was for 4 h at 20°C, the plates were washed as described above, and an alkaline phosphataserabbit anti-mouse immunoglobulin conjugate diluted in PBST was added. Alternatively, an alkaline phosphatase-goat anti-rabbit immunoglobulin conjugate was used. Both conjugates detect immunoglobulin G (IgG), IgM, and IgA (data not shown). Trays were incubated overnight at 20°C. To develop trays, we washed them as described above and added 0.1 ml of 1 M diethanolamine-0.5 mM MgCl2 buffer (pH 9.8). Plates were incubated at 37°C for 100 min, and the A405 was read in a Titertek Multiscan photometer (Flow Laboratories Ltd., Irvine, Scotland).

VOL. 53, 1986

For endpoint titrations, 10-fold dilution steps from 10-2 to 10-6 were tested. Endpoint titers (8), defined as the reciprocal serum dilutions giving an A405 at 100 min and 37°C of 0.1, were calculated by linear regression on a PET Commodore CBM 8032 business computer interfaced with the Titertek Multiscan photometer by using an ELISA program developed by Meddata Digital AB, Solna, Sweden. Antibody classes and subclasses were determined in ELISA with serum-free culture supernatants of the antibody-producing cell lines as coating antigens (diluted 1/10 in PBS). The coated wells were then washed as described above and incubated with rabbit anti-mouse IgGl, IgG2a, IgG2b, IgG3, IgM, and IgA heavy-chain antisera and rabbit anti-mouse K and A light-chain antisera (Litton Bionetics, Kensington, Md.) diluted in PBS-T (1/8,000). Plates were incubated for 4 h at 20°C, washed as described above, and incubated overnight with an alkaline phosphatase-goat antirabbit IgG conjugate (Sigma). Plates were washed and developed as described above. Preparation of sensitized staphylococci and coagglutination. The coagglutination procedure has been described in detail elsewhere (6, 29). Briefly, 0.1 ml of ascites fluid was added to 1.0 ml of 10% formaldehyde and heat-treated staphylococci in 0.1 M NaPO4 buffer (pH 8.0). After incubation and washing, the reagent was brought to 2% (vol/vol) in the same buffer. Coagglutination was done with boiled S. flexneri cultures on a glass slide. Agglutination was recorded as + + when apparent to the naked eye within 30 s and as + if a magnifying glass was needed for observation. Passive hemagglutination. LPS (2.0 mg) was heated in 8.0 ml of PBS at 100°C for 30 min. After cooling, 0.2 ml of washed and packed sheep erythrocytes (SRBC) was added, and the mixture was gently blended and incubated for 20 min at 37°C. The reagent was washed in PBS and diluted to 0.25% SRBC per ml in PBS. For hemagglutination, 96-well round-bottomed mictotiter plates were used. Twofold dilutions of ascites fluid in 50 ,ul were made in PBS, and 50 ,ul of 0.25% LPS-coated SRBC was added. Uncoated SRBC were used as a negative control. Plates were incubated at 4°C and read after 18 h. The titer was defined as the reciprocal of the highest serum dilution producing agglutination visible to the naked eye. RESULTS Production and selection of monoclonal antibodies specific for S. flexneri 0 antigens. Fusion of pooled spleen cells from three BALB/c mice immunized with heat-killed S. flexneri serotype lb with Sp2/0 cells yielded 19 clones after initial screening with homologous LPS followed by testing with the panel; 2 hybridomas bound only to S. flexneri la and lb LPSs, 15 hybridomas produced antibodies which bound to S. flexneri lb, 3a, 3b, and 4b LPSs, and 2 hybridomas bound only to S. flexneri lb and 3b LPSs. Fusion with spleen cells from S. flexneri serotype 3aimmunized BALB/c mice yielded four hybridomas after initial screening and panel testing; one clone produced antibodies reactive exclusively with S. flexneri 3a LPS, and three clones produced antibodies binding to S. flexneri 3a, 3b, and 4b LPSs. The hybridomas from both fusions selected as described above were expanded and recloned by limiting dilution. After hybridomas had been recloned twice or when 100% cloning efficiency had been accomplished, a portion of cells was injected intraperitoneally into pristane-treated mice (33) for ascites production, and a portion was grown in serumfree medium as "suicidal'" cultures.

S. FLEXNERI-SPECIFIC MONOCLONAL ANTIBODIES

105

Evaluation of the specificity of the monoclonal antibodies by ELISA endpoint titration. The results of the ELISA endpoint titrations of three clones representative of the three different specificities found in the fusion experiment with the S. flexneri lb-primed spleen cells are shown in Fig. 2a to c. MASF I (IgGl) antibodies (Fig. 2a) bound only to S. flexneri la and lb LPSs, suggesting that the a-D-Glcp-(l-4)-1-DGlcpNAc disaccharide in the repeating unit of the 0 antigen (Fig. 1) is the immunodominant region to which these antibodies bind. MASF 6-1 (IgGl) antibodies (Fig. 2b) bound to S. flexneri lb, 3a, 3b, and 4b LPSs. These LPSs have one antigenic determinant in common, i.e., group 6 antigen (Fig. 1). It is therefore likely that the structure responsible for binding is the O-acetyl group on C-2 of rhamnose III. The result of the ELISA endpoint titration of the third type of monoclonal antibody, MASF R3 core-1 (IgG2a), is shown in Fig. 2c. It is evident that no 0-antigen specificity could be attributed to this monoclonal antibody. There was a strong binding to the rough LPS (4bR), indicating a specificity for the core region of the LPS. The antibodies were then tested in ELISA endpoint titrations against rough LPSs from E. coli and Salmonella typhimurium as well as S. flexneri 4bR LPS (Table 1). Binding of MASF R3 core-1 antibodies occurred exclusively to S. flexneri 4bR and E. coli F653 LPSs, both of which have the E. coli R3 core (21). To confirm the core specificity of the antibodies and to investigate whether the other S. flexneri LPSs tested have the R3 core, we subjected LPSs used in ELISA to SDSPAGE (Fig. 3a) and immunoblotting (Fig. 3b). As is evident from Fig. 3b, the MASF-R3 core-1 clone is specific for the core antigen of S. flexneri, since only the core band of smooth S. flexneri strains and the 4bR band were stained. Furthermore, all S. flexneri LPSs tested of serotypes 1 to 5 and variants X and Y were stained, indicating that they all

share the R3 core. S. flexneri serotype 6 and S. sonnei phase 1 LPSs were not stained, in complete agreement with published data, which showed that S. flexneri serotype 6 and S. sonnei LPSs have an E. coli Rl core (18, 21, 24). The results of the ELISA endpoint titrations of the ascites fluids from two hybridomas originating from the fusion with S. flexneri 3a-primed spleen cells are shown in Fig. 4a and b. MASF 111:6,7,8 antibodies showed exclusive specificity for S. flexneri 3a LPS, the only LPS carrying both the type III antigen and the group 6 and 7,8 antigens (Fig. 4a). These antibodies seemed to require the simultaneous presence of all three antigens for binding, since there was no binding to S. flexneri 3b LPS or to any LPS carrying the antigen 6 epitope (serotypes lb, 3b, and 4b) or the antigen 7,8 epitope (serotypes 2b, 5b, and X). The binding profile of MASF 6-2

antibodies resembled that of MASF 6-1 antibodies in that they bound to S. flexneri 3a, 3b, and 4b LPSs but differed in that they did not bind to S. flexneri lb LPS. Passive hemagglutination of LPS-coated SRBC with S. flexneri monoclonal antibodies. To further investigate the binding properties of the S. flexneri group 6 antigen- and type III:6,7,8 antigen-specific monoclonal antibodies and to see if MASF 6-1 and MASF 6-2 antibodies recognize two different antigenic epitopes or the same antigenic epitope, we tested serial twofold dilutions of ascites fluids by hemagglutination with LPS-coated SRBC (Table 2). MASF 111:6,7,8 antibodies bound to S. flexneri 3a LPS, in accordance with ELISA results (Fig. 4a). MASF 6-1 antibodies also caused hemagglutination, analogous to ELISA results (Fig. 2b). The hemagglutination pattern of the MASF 6-2 antibodies did, however, differ from the ELISA results, since they also

INFECT. IMMUN.

CARLIN AND LINDBERG

106

1000-

C', Co.

a

x

C~

E

KZK"11"7

100 -

0 0

I

x

0 0

10-

t)

-C

._

~0 1-

CO

Co w

la

lb 2a2b 3a3b4a4b5a5b X Y 644bR S.sor

LPS tested (10

Q

cn

E

pg/ml

BSA 17/)

1000-

0

b X

E x 0 0 x

100-

10'

a) c

.0 CL

a. a)

la lb

-

ornei

2a 2b 3a 3b 4a 4b 5a 5b X Y 6 4bR S.sc

LPS tested (10 yg/mI BSA 1i X)

11J

1000-

c

x

E

0

100

0

Coagglutination of S. flexneri type I, type 111:6,7,8, and group 6 0-antigen-specific monoclonal antibodies by using sensitized S. aureus Cowan 1. To determine whether the specificity seen in ELISA and hemagglutination with S. flexneri LPSs as antigens was of general validity, we tested the monoclonal antibodies by coagglutination. A total of 363 clinical isolates and reference typing strains of S. flexneri were tested (Table 3). The MASF I antibodies had excellent specificity for S. flexneri la and lb strains. None of the 304 S. flexneri strains with heterologous 0 antigens were agglutinated. MASF 111:6,7,8-specific antibodies correctly identified all 29 S.flexneri 3a strains. This result was in accordance with the ELISA and hemagglutination results (Fig. 4a and Table 2). Of the two group 6 0-antigen-binding clones, only MASF 6-2 antibodies were tested. They are of the IgG3 subclass and are therefore more suitable for coagglutination than are MASF 6-1 antibodies, which are of the IgGl subclass (6) (Table 3). MASF 6-2 antibodies agglutinated all S.flexneri lb strains as well as all S. flexneri 3a, 3b, and 4b strains. None of the 286 S. flexneri strains with heterologous 0 antigens were agglutinated. Thus, the coagglutination results for MASF 6-2 antibodies supported the hemagglutination results (Table 2) but were in contrast to the ELISA results (Fig. 4b).

DISCUSSION Monoclonal antibodies with specificities for four different epitopes of the polysaccharide of S. flexneri LPS were produced and characterized in this study. Two of these epitopes, the type I (MASF I antibodies) and group 6 (MASF 6-1 and MASF 6-2 antibodies) epitopes, were established earlier with absorbed rabbit antisera (14). The third type of epitope was recognized by antibodies specific for both type and group epitopes within S. flexneri serotype 3a (MASF III:6,7,8 antibodies) but nonreactive with all other S. flexneri LPSs studied. In this respect, the specificity of these antibodies is new and unique. The fourth epitope described belongs to the core saccharide of S. flexneri LPS and is recognized as the E. coli R3 epitope (MASF R3 core-1

antibodies), i.e., the rough phenotype (21).

MASF I antibodies were specific for epitopes in the 0 polysaccharide of S. flexneri la and lb, as shown in an ELISA with phenol-water-extracted LPS as antigen (Fig. 2a) and in coagglutination testing of 363 S. flexneri clinical isolates (Table 3). These data indicate that MASF I antibod-

LO

0

TABLE 1. ELISA endpoint titrations of MASF R3 core-1 antibodies obtained from ascites fluids against rough LPSs from S. flexneri, S. typhimurium, and E. coli

10L

Chemotype

LPS

ei

n

(IG2) BSA,

seru

three

ntibody albumin.17 tested MonoG1). (c) LP-cated MonoCloa attr core-1)

difeeatedwtS. flexneriseii hybrdms

256). Although the titer was almost 20-fold lower than that aigainst s. flexneri 3a, 3b, and 4b LPSs, it was still significanit ,

.

HF4704 EH100 4bR F653 2513 W3110 TV119 SL733 TV161 TV148 SL805 SL1032

SL1102 Lipid A'

E. E. E. E. E. E. S. S. S. S. S. S. S.

coli Ri coli R2 coli R3 coli R3 coli R4 coli K-12 typhimurium typhimurium typhimurium typhimurium typhimurium typhimurium typhimurium

Titer (103)

Ra Rbl Rb2 Rb3 Rc Rd Re