Hybridomas secreting immunoglobulin A (IgA) monoclonal antibodies (MAbs) against Salmonella enteritidis lipopolysaccharide (LPS) were generated after mucosal immunization of BALB/c mice with heat killed bacteria. Antigen binding properties and specificity of the produced MAbs were studied in ELISA and immunoblotting with purified LPS. Two IgA MAbs agglutinated all Salmonella OD1 strains and all S. enteritidis clinical isolates. MAb 178H11 recognized O:9 antigen of subserogroup OD1 LPS. MAb 177E6/A9 reacted also with OD3 LPS antigen and agglutinated OD3 strains. These data suggest the existence of different O:9 antigen subspecificities, one presented in subgroup OD1 and the other common for OD1 and OD3. Thus the produced IgA MAbs prove to be useful reagents, which could differentiate OD1 and OD3 from OD2 strains.
Serological identification and classification of Salmonella is based on their O and H antigens according to the scheme of Kauffmann–White . Salmonella species are subdivided in different serogroups by their lipopolysaccharide (LPS) antigen specificities . A variety of O antigens had been determined by cross absorption of rabbit agglutinating sera. Salmonella isolates can be identified and grouped by slide agglutination test . Preparation of MAbs directed to O and H antigens has been reported [4–8]. MAbs with their monoepitopic specificity have many advantages over monospecific polyclonal sera. Salmonella strains which present LPS O:9 antigen are unified in group OD. MAbs strongly reacting with LPS of group OD1 Salmonella were established as useful reagents for identification of this serogroup and for differentiation of OD1 and OD2 strains, which have identical flagellar H antigens . Important pathogens Salmonella typhi (9,12[Vi]; d-) and Salmonella enteritidis (1,9,12; g,m-) belong to serogroup OD and different assays using MAbs are developed to detect their antigens [8–10]. The MAbs used for diagnostic purposes are mainly of IgG and IgM isotypes. There are several reports about production of hybridoma cell lines secreting IgA MAbs, which bound bacterial surface antigens [11,12]. IgA antibodies are the principal mucosal antibody class produced by local plasma cells, which enter as dimeric or polymeric forms into the secretions via specific polymeric immunoglobulin receptor-mediated transport. They form the first line of immune defense against infection, including Salmonella infection, by their ability to inhibit attachment of microorganisms and their toxins to the mucosal surfaces . Thus IgA MAbs are used mainly to clarify the mechanisms of local immune defense [14–18]. However, their possible application as serotyping reagents has not been assessed and their diagnostic potential has not been compared with that of the available IgG and IgM diagnostic MAbs. There are two main strategies to obtain IgA isotype MAbs: by fusion of myeloma line cells with Payer's patches lymphoblasts after mucosal immunization with live microorganisms and by isolation of IgA isotype switch variants of IgG and IgM hybridomas [15,19]. In this study we characterized two IgA MAbs generated after fusion between spleen lymphoblasts and myeloma cells after intragastral immunization with heat killed bacteria.
2Materials and methods
2.1Bacterial strains and antigen preparation
Salmonella laboratory strains used in this study were obtained from the Collection of Institute Pasteur, Paris, France and the National Microbiological Collection, Sofia, Bulgaria. Fifty-seven S. enteritidis clinical isolates were confirmed by routine biochemical and serological methods. The bacteria were grown overnight on trypticase soy agar. Highly virulent for mice S. enteritidis clinical isolate (LD50<10 bacteria after intraperitoneal (i.p.) application) was used for immunization.
S-form LPS were extracted by the hot phenol–water method of Westphal  from Salmonella strains of different serogroups. Purified LPS from Salmonella typhimurium (O:1,4,5,12), S. enteritidis (O:1,9,12), Salmonella minnesota Ra (strain R60) and lipid A were kindly provided by C. Galanos, Max-Planck-Institute of Immunobiology, Freiburg, Germany. All LPS preparations were dissolved at a concentration of 2 mg ml−1 in distilled water and neutralized with triethylamine if necessary.
2.2Production of MAbs
Eight weeks old BALB/c mice were immunized intragastrally (i.g.) daily for 4 weeks with 5×109 heat killed bacteria in 0.5 ml carbonate bicarbonate buffer pH 9.6. After the last application the animals were bled from the retroorbital venous plexus and sera were tested by ELISA. Two weeks later one mouse was boosted intravenously (i.v.) with 1×109 bacteria and another one was boosted i.g. with 10×109 bacteria. Three days afterwards spleen cells were harvested and fused with P3X63-Ag8.653 myeloma cells in two separated fusions according to the method of Köhler and Milstein [21,22]. Supernatants of produced hybridomas were tested by ELISA and isotyped by antigen-mediated ELISA using ISO-2 kit (Sigma). Two IgA producing hybridomas reacting in slide agglutination test were grown in serum-free medium (Sigma). Ascitic fluids were obtained after i.p. injection of 10×106 hybridoma cells in pristine primed BALB/c mice.
2.3Purification of the IgA MAbs
MAbs from ascites were partially purified by ammonium sulfate precipitation by 45% (vol/vol) saturation and by cold ethanol precipitation. Precipitated MAbs were dissolved in 20 mM Tris–HCl buffer and fractionated by anion-exchange chromatography on Mono Q HR 5/5 column (Pharmacia). Purified MAbs were equilibrated to 1.3 mg ml−1 concentration and stabilized by addition of 1% BSA. Serum-free supernatants were precipitated via dropwise addition of saturated ammonium sulfate to 45% (vol/vol) and dialyzed against phosphate-buffered saline (PBS). Production of monomeric, dimeric and polymeric IgA forms was confirmed by SDS–PAGE under reducing and non-reducing conditions and by immunoblotting.
2.4Slide agglutination test (SAT)
Slide agglutination was performed by mixing 20 μl of antibody solution (undiluted or serially diluted) with bacteria grown on solid media and the results were read in 1 min.
2.5Passive hemagglutination and passive hemolysis
Microtest version of the passive hemolysis (PH) described by Galanos  was used. Hemolysis titers (the last dilution at which approximately 50% hemolysis occurred) were determined after incubation for 1 h at 37°C and overnight at 4°C. Passive hemagglutination (PHA) was performed in the same way as PH but without adding of complement. The last dilution with hemagglutination was considered as a titer.
2.6Enzyme-linked immunosorbent assay (ELISA)
A sensitive antigen-mediated ELISA  with purified LPS was performed. Goat anti-mouse Ig (G, A, M) peroxidase conjugated antibody (Sigma) was used as a secondary antibody. MAbs were isotyped by antigen-mediated ELISA performed according to the ISO-2 kit producer's protocol (Sigma).
2.7Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotting
SDS–PAGE of LPS was carried out according to the method of Laemmli  and the separated LPS were stained by silver staining . Immunoblotting of fractionated LPS was performed by the method of Towbin . LPS were transferred from separating gels to nitrocellulose membranes (pores of 0.45 μm, Schleicher and Schuell) for 1 h at 0.8 mÅ cm−2 with Bio-Rad semi-dry blotting device. After blocking with 3% bovine serum albumin in PBS the membranes were incubated with supernatants overnight at 4°C. IgA MAbs bound to the LPS bands were visualized with goat anti-mouse (α-chain specific) peroxidase conjugated antibody and 4-chloro-1-naphtol as a substrate. To detect monomeric and polymeric IgA forms purified MAbs were resolved by SDS–PAGE under reducing and non-reducing conditions and the separated proteins were visualized by Coomassie staining. Immunoblotting of purified IgA MAbs was performed using the same protocol.
2.8Indirect immunofluorescent assay (IFA)
Microscope slides were coated with S. enteritidis, air-dried and fixed in cold (−20°C) acetone. The slides were incubated with purified MAbs for 1 h at 37°C in humid atmosphere. After three washes in PBS, FITC-labeled goat anti-mouse antibody (Sigma) was added for 30 min and the samples were examined with fluorescent microscope.
2.9Tests for detection of secretory IgA (sIgA)
Hybridoma 177E6 was grown as ‘backpack’ tumor by subcutaneous injection of 5×106 cells in PBS on the upper back of BALB/c mice as described [14,15]. Two weeks later serum, bile and gut samples from mice with visible tumors were collected and tested by SAT and by antigen-mediated ELISA using purified 177E6 MAb as a standard. In the other set of experiments purified MAbs 177E6 and 178H11 were injected i.v., the samples were collected 8 h later and tested in the same way.
2.10Tests for stability
Purified IgA MAbs and supernatants containing 0.1% NaN3 were tested by SAT, PHA and ELISA after exposure to different temperatures: 56°C for 1 h, 37°C for 2 weeks, 2–8°C or at room temperature (18°C) for 1 year. Repeated fast freezing and thawing was used to test purified MAbs stabilized with 1% BSA in PBS. Diluted in PBS ascites or undiluted supernatants were mixed 1:1 (vol/vol) with freshly prepared 10 mM dithiothreitol (DTT) solution for detection of antibody stability in the presence of reducing agents. The samples were incubated at 37°C for 2 h and the specific activity was measured by PHA and SAT. IgA stability was examined after exposure to different pH values. Purified MAbs were diluted in acetate buffer pH 5.2 and carbonate bicarbonate buffer pH 10.0, stored for several days at 2–8°C and after overnight dialyzation against PBS the antibody activity was measured.
3.1Production of MAbs
After the first period of oral immunization sera of the immunized mice demonstrated specific IgA reactivity only in ELISA with S. enteritidis LPS. Following reimmunization course, increasing titers of the specific serum IgA antibodies were observed in five animals. Two of them with agglutinating titer 1:8 were selected and boostered. Fusion with spleen cells of mouse with final i.g. boost gave no rise of antigen specific hybridomas. In contrast, high yield of S. enteritidis LPS-reactive hybridomas was obtained after fusion with spleen cells of the mouse with final i.v. boost. Culture supernatants were screened by ELISA and 12 hybridoma clones were isolated and isotyped. Eight clones secreted IgA, two clones IgM, one clone IgG2a and one clone IgG2b MAbs. Two of the IgA hybridomas (177E6 and 178H11) secreting agglutinating antibodies were selected for detailed characterization.
3.2Purification of IgA MAbs
The first precipitation step led to partial purification of IgA from ascites. Ammonium sulfate precipitation decreased agglutination activity more than two times. In contrast cold ethanol precipitation did not lead to decrease of activity. On the anion-exchange Mono Q column IgA was fractionated in one peak and the separation of the monomeric, dimeric and polymeric IgA forms was not possible. Single step ammonium sulfate purification of serum-free supernatants was efficient without significant waste of activity. Analysis by SDS–PAGE under reducing condition showed only two bands of heavy and light chains after Coomassie staining (Fig. 1). Under non-reducing condition the polymeric IgA formed two bands with high molecular mass (MM over 205 kDa marker band) and the monomeric form migrated as a single band (MM approximately 110 kDa). Subsequent immunoblotting with mouse α-chain specific peroxidase conjugated antibody confirmed that bands were monomeric and polymeric IgA forms (Fig. 2A and B). Under reducing condition only one band corresponding to heavy α-chain (MM approximately 55 kDa) was visualized (Fig. 2C).
3.3Characterization of IgA MAbs
IgA MAbs were found to react with LPS from Salmonella serogroup OD1 in PHA and ELISA and to agglutinate all OD1 strains, including 57 clinical isolates of S. enteritidis (Table 1). Antibody 177E6 reacted in addition with strains and LPS from subserogroup OD3. The MAbs did not recognize Salmonella OD2 antigen. Cross-reactivity with LPS of other serogroups (A to O:67) and LPS of complete core rough mutant S. minnesota R60 (Ra chemotype) and with purified lipid A was not detected. The capacity of IgA MAbs to agglutinate bacteria was demonstrated in SAT using purified MAbs in serial dilutions starting from 10 μg ml−1. Concentration of 1.25 μg ml−1 of IgA MAb 177E6 was sufficient to agglutinate bacteria. Both MAbs produced bright fluorescence with all OD1 strains used in IFA (Table 1).
Table 1. Specificities of IgA MAbs as determined by ELISA, slide agglutination and immunofluorescent assay
Slide agglutination testd
(F to O:67)
aMAbs from culture supernatants.
bSupernatants diluted 1:10; optical densities at 492 nm less than 0.100 are defined as negative (−).
cLPS from serogroup OD1, S. enteritidis (1,9,12); OD2, S. plymouth (O:9,46); and OD3, Salmonella 149/66II (O:1,9,12,46,27).
dNumber of agglutinated strains of different serovars/number of used strains.
epassive hemolysis and passive hemagglutination reciprocal endpoint titers (1/T) with S. enteritidis LPS coated SRBC.
fIFA with Salmonella miami and two S. enteritidis clinical isolates.
LPS from smooth strains analyzed by SDS–PAGE showed characteristic ladder-like pattern after silver staining (Fig. 3A). After transfer to nitrocellulose membrane and subsequent immunoblotting both agglutinating IgA MAbs reacted with high molecular mass bands corresponding to the O antigen of group OD1 (Fig. 3B, C). MAb 177E6 reacted also with OD3 LPS antigen (Fig. 3C). Both IgA MAbs were strongly specific for S. enteritidis in immunofluorescent assay. Although IgA MAbs showed considerable agglutinating capacity in PHA they did not react in PH (Table 1).
Specific sIgA MAb was found in serum, bile and gut samples of mice bearing hybridoma 177E6 as subcutaneous tumor as well as in the samples after i.v. injection of both IgA MAbs.
Exposure of the IgA MAbs to different temperatures did not decrease significantly their activity compared to the control IgG MAb as measured by PHA (Table 2). In contrast, the agglutinating activity of control IgM MAb was significantly decreased after exposure for 1 h at 56°C and at room temperature (18°C) for 6 months. Different pH values and repeated freezing and thawing procedure also did not affect IgA activity. The reducing agent DTT completely abolished the activity of IgA and IgM but not of IgG MAbs.
Table 2. Stability of IgA MAbs after exposure at different temperatures, pH values and reducing agent as measured by passive hemagglutinationa
10 mM DTTf
aSRBC coated with S. enteritidis LPS and LPS from S. typhimurium in PHA for MAb 13G9 (specific for Salmonella O:42 antigen).
bNon-treated supernatants stored at −80°C.
cReciprocal endpoint titer.
dAcetate buffer, pH 5.2.
eCarbonate bicarbonate buffer, pH 10.0.
f10 mM dithiothreitol in PBS for 2 h.
The presented data demonstrate the production and characterization of two anti-LPS MAbs of IgA isotype. Production of IgA hybridomas requires mucosal application of antigen and the classical immunization protocols need modification. To produce IgA secreting hybridomas live low virulent or avirulent for mice strains are used for i.g. immunization and Payer's patches were the source of cells for fusion [12,15,17]. In the present work modified immunization protocol was applied because highly virulent for mice S. enteritidis strain was used. The mice were immunized i.g. with heat killed bacteria for prolonged time. The induced immune response shows that this approach is reasonable when highly virulent strain is used for immunization. Preliminary experiments indicated that immunization with killed bacteria did not evoke significant stimulation and enlargement of the Payer's patches as in the case of immunization with live bacteria. Thus the Payer's patches were not an adequate cell source for fusion in this case. As a possible alternative, spleen cells were used for fusion after i.v. or i.g. final boost. Obtaining of hybridomas with prevalence of IgA isotype producing clones proved the adequacy of this approach. No specific hybridoma was generated after i.g. boost. The good yield of specific hybridomas and prevalence of IgA secreting clones after i.v. boost supports the concept that IgA switched B-cells settle in the spleen after mucosal contact with antigen. Thus spleen cells are reasonable fusion partners when IgA secreting hybridomas are the aim, but final i.v. immunization is necessary.
The produced IgA MAbs were specific for the OD group antigen as shown in all immunological reactions with LPS of S. enteritidis (Table 1). Two of the MAbs with strong agglutinating ability were interesting for detailed characterization of their specificity and biological properties, including protective activity. The absence of cross-reactivity with strains from serogroups A to O:67, Ra LPS and lipid A indicate that MAbs carry O:9 specificity. Antigen O:9 is defined as serogroup OD specific factor. The other presented O antigens subdivide the OD strains into three subserogroups: OD1 (1,9,12), OD2 (9,46) and OD3 (1,9,12,27,46). Production and possible diagnostic applications of anti-OD MAbs have been reported . However their specificities for subgroups OD1, OD2 and OD3 and the possible existence of O:9 subspecificities are not discussed. Subspecificities are described in some Salmonella serotypes such as factor O:3, O:4, O:6 and recently for O:7 [5,28]. Our findings suggest expression of at least three different subspecificities (epitopes) of 0:9 antigen. One subspecificity is presented only in OD1 O:9 antigen as shown by reaction of MAb 178H11 with strains and LPS of serogroup OD1 but not of OD2 and OD3. Reaction of MAb 177E6 with strains and LPS of subgroups OD1 and OD3 suggests the existence of common OD1 and OD3 subspecificity. The common Salmonella OD epitope was demonstrated by MAbs in another study by us (manuscript in preparation).
Evaluation of the IgA forms secreted from both hybridomas and preparation of IgA standard need precise purification. IgA purification is more difficult than IgG and some procedures are intended for that. In our hand cold ethanol precipitation was suitable as a first step for IgA purification from ascites and revealed minimal losses of MAb activity in comparison with ammonium sulfate precipitation. This method is simple to perform but needs a subsequent purification step by other techniques such as affinity-, gel- or ion-exchange chromatography . Anion-exchange columns Mono Q supply good yield of pure IgA MAbs, which is confirmed by electrophoresis under reducing condition (Fig. 1). However, separation of monomeric and polymeric IgA is not possible in this way. Single step purification of IgA MAbs from serum-free supernatants by ammonium sulfate precipitation is efficient. SDS–PAGE and subsequent immunoblotting of the purified IgA MAbs confirmed the production of dimeric and polymeric IgA forms. Molecular structure of IgA shows some differences with conventional immunoglobulin structure. In BALB/c mice, the light chains might be disulfide bonded to each other but not to the heavy chains  and this could explain the existence of bands corresponding to heavy-heavy chain dimers with MM 110 kDa (Fig. 2A and B). Under used experimental conditions both IgA MAbs showed high stability, comparable to IgG and greater than IgM Salmonella specific agglutinating MAbs (Table 2). These findings suggest that the IgA MAbs are suitable for preparation of reagents for serotyping of Salmonella isolates.
Preliminary experiments to define biological function were performed as a part of the complete characterization of the produced MAbs. The negative results of PH test confirmed the lack of complement activation properties of IgA antibodies. Specific secretory IgA detected in biological fluids after growing of hybridoma 177E6 as ‘backpack’ tumor and after i.v. injection of purified MAbs verified that both hybridomas secrete polymeric IgA. This makes them suitable for examination of biological properties of sIgA, including their protective potential in experimental Salmonella infection.
Thanks are due to V. Paskova and A. Cherneva for valuable technical assistance.