INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacillus cereus is known to cause a variety of nongastrointestinal diseases (12) as well as two different types of food poisoning (for reviews, see references 17, 19, and 23), which are characterized by either emesis or diarrhea. The diarrheal type of intoxication has been related to single proteins (1, 28, 29) as well as protein complexes (10, 30) as causative agents. Currently two different enterotoxin complexes, each consisting of three exoproteins, are discussed extensively. One of these, a nonhemolytic enterotoxin (NHE) consisting of three components with molecular masses of 39, 45, and 105 kDa, was recently described by Lund and Granum (24). This complex, however, is not fully characterized, whereas the enterotoxic hemolysin BL (HBL) has been studied extensively (3-5, 7, 8). HBL contains the protein components B (37.5 kDa), L₁ (38.2 kDa), and L₂ (43.5 kDa), and all three components are required to produce maximum biological activity. It could be demonstrated that HBL is lethal to mice, cytotoxic to CHO cells, and positive in both the ileal loop test and the vascular permeability reaction (5, 7). The genes encoding for the components of HBL have been cloned and characterized, and it has been shown that they are transcribed from the same operon in one mRNA (22, 27).

At present, immunochemical characterization of the proteins constituting the HBL and NHE complexes is limited by the nonavailability of specific antibodies. Most research groups used in-house polyclonal antisera (7, 10, 30), which usually show reactivity with several proteins when used for immunoblotting. A reversed passive latex agglutination assay (Oxoid RPLA), which detects mainly the L₂ component of HBL, also uses polyclonal antisera (6, 20). The only commercial enzyme-linked immunosorbent assay (Tecra visual immunoassay), however, reacts with two nontoxic proteins (6), one of these probably representing a component of NHE (24). Also, the specificities of two monoclonal antibodies against HBL components, which were mentioned in an earlier study (3), have not been fully defined.

Due to these problems, the immunochemical detection of the B. cereus enterotoxins is still not satisfactory, and a range of in vivo and in vitro tests is required to estimate the toxicity of culture filtrates, e.g., the mouse lethality test, the rabbit ileal loop test, the vascular permeability reaction, and cell culture assays (7, 11, 28, 30). Since all these methods show limitations, particularly with regard to specificity and sensitivity, and are hardly applicable for the detection of B. cereus enterotoxins in food, we attempted to produce monoclonal antibodies to improve the specific detection of the components of HBL and facilitate the screening of B. cereus isolates for enterotoxic activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

B. cereus strains, culture medium, and culture conditions. Enterotoxic strains B-4ac (16) and F837-76 (31) were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSM), Braunschweig, Germany (DSM 4384 and DSM 4222); strain WSBC 10204 was obtained from S. Scherer, Freising, Germany (14), and strain 0075/95 was obtained from P. E. Granum, Oslo, Norway (24). All other B. cereus strains (prefix MHI for Milch-Hygiene-Institut) were isolated from infant food (2). Unless otherwise indicated, cells were grown in CGY medium (5) supplemented with 1% glucose for 6 h at 32°C with shaking. EDTA (1 mmol/liter) was added at the time of harvesting. Cell-free supernatants obtained by centrifugation (10,000 × g at 4°C for 20 min) and filtration through 0.2-µm-pore-size Millipore filters were used for purification of proteins and as coating antigens in the indirect enzyme immunoassays (EIA), respectively.

Protein preparations used for immunization of mice. Strain B-4ac was grown for 8 h at 32°C by the sac-culture method exactly as described by Parker and Goepfert (26). Proteins were precipitated by the addition of solid ammonium sulfate (518 g/1,000 ml), resuspended in Tris-HCl buffer (5 mmol/liter, pH 8.6), and further concentrated by Centriprep-30 concentrators (Millipore). The resulting dark brown-colored retentate (2 ml) was fractionated by gel filtration on Sephadex G-75 superfine (2.6 by 90 cm; flow rate, 12 ml/h) equilibrated with Tris-HCl buffer. Selected fractions were pooled, designated as Sephadex G-75 superfine (A, B1, B2, and C), and concentrated by ultrafiltration to approximately 0.5 ml.

Purification of HBL. Strain B-4ac grown in CGY medium was used for the production of HBL. The single components were purified according to the procedure described by Beecher and Wong (5), except the last step on the Resource Q column was omitted.

Production of monoclonal antibodies. The exoprotein preparations Sephadex G-75 superfine B1, B2, and C were used as immunogens. Three groups of 12-week-old female mice (three BALB/c strain and a hybrid strain of BALB/c × [NZW × NZB] per group) were immunized by intraperitoneal injection of 30 µg of the respective protein preparation, dissolved in Tris-HCl buffer, and emulsified in Freund's complete adjuvant (1:3). At day 93, the animals received a booster injection of the same amount of immunogen in incomplete Freund's adjuvant. Finally, at day 142, 3 days before fusion, the animals got a final booster injection of 45 µg of antigen dissolved in Tris-HCl. For fusion, a single-cell suspension from spleen and axillary lymph nodes of a hyperimmunized mouse was made and fused with myeloma cells (X63-Ag8.653) according to Fazekas de St. Groth and Scheidegger (13) at a ratio of 2:1. Hybridoma and myeloma cells were maintained in Dulbecco's modified Eagle medium supplemented as previously described (21). Culture supernatants were tested for specific antibodies 12 days after fusion in an indirect EIA, using the respective Sephadex G-75 exoprotein preparation as coating antigen. Positive hybridomas were cloned at least three times by limiting dilution technique. A mouse-hybridoma subtyping kit (Roche Diagnostics, Mannheim, Germany) was used according to the instructions of the manufacturer for the determination of class, subclass, and light chain type of the monoclonal antibodies. Mass production of the antibodies was performed in a Mini-Perm bioreactor (In Vitro Systems, Osterode, Germany). The antibody preparations were purified by affinity chromatography on protein A-agarose (Bio-Rad, Munich, Germany).

Indirect EIA. To determine the relative antibody titers of the mouse sera and to screen for antibody secreting hybridomas, an indirect EIA system was used. Microtiter plates were coated overnight at room temperature either with the Sephadex G-75 superfine (A to C) exoprotein preparations, the purified HBL components, or sphingomyelinase (Roche Diagnostics) at a concentration of 1 µg/ml (100 µl per well) in carbonate-bicarbonate buffer (0.05 mol/liter, pH 9.6). Free protein binding sites of the plates were blocked with phosphate-buffered saline containing sodium caseinate (30 g/liter) for 30 min. Then the plate was washed three times with Tween 20 solution (0.25 ml/liter of 0.15 mol/liter of sodium chloride solution) and made semidry. Serial dilutions of mouse sera or hybridoma culture supernatants (100 µl/well) were added and incubated for 1 h. After a washing step, rabbit anti-mouse immunoglobulin G (IgG) labeled with horseradish peroxidase (1:3,000 in phosphate-buffered saline containing sodium caseinate [10 g/liter]) was added and incubated for 1 h at room temperature. Then the plate was washed again, and 100 µl of substrate-chromogen solution (1 mmol of 3,3',5,5'-tetramethylbenzidine, 3 mmol of H₂O₂ per liter of potassium citrate buffer [pH 3.9]) per well was added (15). After 20 min, the color development was stopped with 1 mol of H₂SO₄ (100 µl/well) per liter, and the absorbance was measured at 450 nm.

Immunoblot. To further characterize the specificity of the monoclonal antibodies, exoprotein preparations from B. cereus were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretically transferred to nitrocellulose paper (pore size, 0.2 µm; Schleicher & Schuell). Immunochemical staining was performed by blocking the paper with Tris-buffered saline (0.05 mol/liter of Tris-HCl, 0.15 mol/liter of NaCl [pH 7.5]) containing sodium caseinate (10 g/liter) for 30 min and then incubating it with protein A-purified monoclonal antibodies (2 µg/ml). After a washing step, bound antibodies were detected with alkaline phosphatase-labeled anti-mouse IgG by using 4-nitroblue tetrazolium chloride-5-bromo-4-chloro-3-indolylphosphate (NBT-BCIP) as the chromogenic substrate according to the instructions of the manufacturer (Roche Diagnostics).

Cytotoxicity tests. The cytotoxicities of sample materials (cell-free supernatants of B. cereus, exoprotein preparations, and components of the HBL enterotoxin complex) were determined by measuring cell proliferation and cell viability by using Vero cells. The growth medium and diluent consisted of Eagle minimum essential medium (Biochrom KG, Berlin, Germany) with Earle salts supplemented with 1% fetal calf serum and 2 mmol of glutamine per liter. The cytotoxic activity was tested by placing serial dilutions of each sample (0.1 ml) into microtiter plates. Cell suspensions (0.1 ml; 10³ cells/well) were added immediately afterwards, and the plates were incubated for 24 h at 37°C in a 5% CO₂ atmosphere. After removing 0.1 ml of medium, 10 µl of cell proliferation reagent WST-1 (Roche Diagnostics) was added to each well. After incubation of the plates for 1 h under the same conditions, the absorbance of the soluble formazan dye formed after cleavage of the tetrazolium salt WST-1 by mitochondrial enzymes of viable cells was determined at 450 nm. The resulting dose-response curve was used to calculate the 50% inhibitory concentration (expressed as the reciprocal value of the dilution that resulted in 50% loss of mitochondrial activity) by linear interpolation.

SDS-PAGE. SDS-PAGE was carried out by using the PhastSystem (Amersham Pharmacia Biotech, Freiburg, Germany) and precast minigels (PhastGel gradient 10 to 15). Separated proteins were stained with Coomassie brilliant blue.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Exoprotein preparations. The B. cereus reference strain (B-4ac) was used throughout this study for the production of the different exoprotein preparations. Fractions B1, B2, and C obtained from centrifuged culture supernatants, after separation on Sephadex G-75 superfine, contained mainly proteins in the 25- to 50-kDa range according to SDS-PAGE (Fig. 1). Fraction B1 contained two major bands between 43 and 47 kDa, whereas fraction B2 showed a strong band at 39 kDa and a number of other proteins. Fraction C consisted mainly of one band of approximately 33 kDa, together with some impurities of low molecular weight. In addition, the protein components B, L₁, and L₂ of the HBL complex were purified exactly as described by Beecher and Wong (5), except for the last step of the procedure, which was omitted. SDS-PAGE (Fig. 2) revealed that each of the respective preparations contained a single major component, accompanied by some minor impurities. The cytotoxic activity of these preparations was tested by using the WST-1 assay. The results (Table 1) indicated that the single proteins were only slightly cytotoxic, while a respective mixture of L₁, L₂, and B resulted in a more than 30-fold increase in cytotoxicity. All combinations of two compounds showed no increase in toxicity, compared to the single components.

View larger version (106K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE of pooled fractions from Sephadex G-75 superfine chromatography. Lane M, Pharmacia low-molecular-weight markers (from the top: phosphorylase b, 94.0 kDa; bovine serum albumin, 67.0 kDa; ovalbumin, 43.0 kDa; bovine carbonic anhydrase, 30.0 kDa; trypsin inhibitor, 20.1 kDa; and alpha-lactalbumin, 14.4 kDa); lane 1, fraction C; lane 2, fraction B2; lane 3, fraction B1; lane 4, fraction A. Fractions B1, B2, and C were used for the immunization of mice.

View larger version (111K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE of HBL components isolated from reference strain B-4ac as described in the text. Molecular weight markers were run in lanes M; sizes are shown to the left. Lane 1, B component; lane 2, L2 component; lane 3, L1 component.

Antibody production. Groups of six mice were each immunized with Sephadex G-75 superfine fractions B1, B2, and C. After an immunization period of 5 months, including two booster injections, the specific antibody titers, checked by indirect EIA with the respective Sephadex G-75 superfine fractions as coating antigens, were determined in the sera of the mice. The animals showing the highest serum antibody titers against the B. cereus exoprotein preparations were used for the fusion experiments. A total of 40 hybridoma cell lines, secreting specific antibodies against B. cereus exoproteins, were obtained from four different fusions of mouse splenocytes with myeloma cells. For further characterization of the cell lines, antibody-containing supernatants were tested for reactivity with the HBL complex by using purified B, L₁, and L₂ preparations, as well as whole culture supernatants as test antigens in the indirect EIA. From these studies, five monoclonal antibodies were chosen for further characterization. The determination of antibody class revealed that all these antibodies were composed of the IgG heavy chain and the kappa light chain; the IgG subtypes are listed in Table 2.

Antibody characterization. Western blot analyses were done to characterize the specificity of the monoclonal antibodies. A distinct reactivity pattern of the five antibodies was obtained if a crude culture supernatant of strain B-4ac was used for the blotting experiments (Fig. 3). In detail, antibody 2A3 reacted only with the band at approximately 37 kDa, and both 8B12 and 1A12 reacted with the band at 43 kDa. Antibody 1C2, on the other hand, gave a strong reactivity with the band at 38 kDa but also showed a clear reaction with a top protein of approximately 93 kDa. Antibody 2A12 finally showed a distinct reactivity with a protein that had an M_r of 33,000, which was identical with a commercial B. cereus sphingomyelinase preparation (Roche Diagnostics). By using the purified proteins of the HBL complex for Western blot analyses, it could be proved that monoclonal antibody 2A3 reacts with the B component, that 1A12 and 8B12 react with the L₂ component, and that 1C2 reacts with the L₁ protein of this enterotoxin complex. If, however, the monoclonal antibodies were tested by using strain 0075/95, which has been found to be HBL negative by PCR (24), antibodies 2A3, 1A12, and 8B12 gave clearly negative results, whereas 1C2 showed the same pattern of bands in the immunoblot as the HBL-positive strain B-4ac. These findings underline the specificity of the antibodies against the B and L₂ components of HBL and indicate that antibody 1C2 cross-reacts with another protein, possibly the 39-kDa protein of the NHE complex. The specificities of all antibodies, as apparent from the immunoblot analyses, are summarized in Table 2.

View larger version (49K): [in this window] [in a new window]   FIG. 3.   Reactivity of the monoclonal antibodies with exoproteins of B. cereus B-4ac. Crude, cell-free culture supernatants were concentrated (100 times), separated by SDS-PAGE, and transferred to nitrocellulose. Proteins were reversibly stained with Ponceau S solution. The molecular weight standards were marked with a pencil (on each blot at the left or right side) and subsequently probed with the monoclonal antibodies (2 µg/ml) under study. Lane 1, antibody 2A3 against the B component of HBL; lane 2, antibody 1A12; lane 3, antibody 8B12 against the L2 component of HBL; lane 4, antibody 1C2 against the L1 component of HBL; lane 5, antibody 2A12 against sphingomyelinase.

Characteristics of B. cereus strains. A total of 26 different strains of B. cereus, including the reference strains, were screened for cytotoxic activity and reactivity in the indirect EIA (Table 3). According to the reactivity pattern obtained in the assays, the B. cereus strains may be divided into three groups. About 50% of the strains tested reacted in a manner similar to that of the enterotoxic reference strains B-4ac and F837-76, i.e., they proved to be cytotoxic and positive in the indirect EIA for HBL components. Furthermore, nearly 40% behaved like the HBL-negative, highly cytotoxic strain 0075/95, which was isolated during a food poisoning outbreak in Norway (24). Only about 10% of the strains under study proved to be HBL negative and exhibited low cytotoxic activity. With one exception (strain MHI 147b), all highly cytotoxic strains also produced a significant amount of sphingomyelinase.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

There is still some controversy about the components involved in the diarrheal type of B. cereus food poisoning. The work of Beecher and Wong (5) presented strong evidence that HBL is a major pathogenicity factor. The work described herein was therefore focused on this toxin complex, although the potential role of other protein complexes (10, 30) or even a single component toxin (28) had to be generally considered. Since purification on DEAE columns possibly leads to biological inactivation (9), we decided to use crude exoprotein preparations of B. cereus, obtained by gel filtration on Sephadex G-75 superfine, for the immunization procedure. The resulting chromatograms showed a peak pattern similar to that described by Thompson et al. (30), and the fractions were therefore accordingly designated A to C, the B fraction being further divided into two pools (B1 and B2). Considering the reactivity of the resulting monoclonal antibodies, it can be concluded that fraction B1 contained a mixture of the L₂ component of HBL and a nontoxic exoprotein with a molecular mass of approximately 47 kDa. The B and L₁ components of HBL were contained in fraction B2, together with other exoproteins. Finally, fraction C contained B. cereus sphingomyelinase, probably together with other phospholipases. Since sphingomyelinase copurifies with enterotoxin preparations (18), fraction C was also included in the immunization experiments, with the intention to obtain a monoclonal antibody against sphingomyelinase, such as 2A12, to verify the purity of the specific enterotoxin preparations.

All the exoprotein preparations from the Sephadex G-75 superfine column were highly immunogenic, and the antibody titers determined in the sera of the immunized mice were usually higher than 1:250,000. After fusion of mouse splenocytes with myeloma cells, a broad range of antibodies specific for various B. cereus exoproteins could be established. Since the immunogens were crude preparations, a hybridoma selection strategy had to be established by using a panel of test systems to identify and characterize clones with the desired properties. A preselection of suitable clones was done by screening for antibodies in the indirect EIA. Though this method represents a nonoptimized technique, it was interesting to note that cell-free culture supernatants of B. cereus could be diluted up to 1:2,560 and still give positive results (Table 3). This finding may be interpreted as a sign for the high affinity of the selected antibodies.

To verify the reactivity of the monoclonal antibodies, the preparation of highly purified HBL components was essential. The results obtained from SDS-PAGE (Fig. 2) and the cytotoxicity data presented in Table 1, which were in full agreement with the data published by Beecher and Wong (5), proved that the isolated proteins represent the components of the HBL enterotoxin complex. Since a variation in the N-terminal sequence has been described for the L₂ component (25), we additionally chose this compound for sequencing after purification by immunoaffinity chromatography (unpublished data). The N-terminal sequence found was E T Q Q E G M D I S, which differs at position 6 (glycine instead of asparagine) from the original sequence deduced for the L₂ component from B. cereus F837-76 (5, 27) but is in accordance with the slightly modified sequence found for L₂ produced by strain 1230-88 (25). In addition, none of the purified HBL preparations contained detectable amounts of sphingomyelinase when tested in the indirect EIA with antibody 2A12.

Regardless if crude culture supernatants or purified components were used for the immunoblots, it could be proved that the monoclonal antibodies against the B (2A3) and the L₂ (1A12 and 8B12) components were highly specific for the respective protein and showed no unexpected cross-reactivity. Although the two monoclonal antibodies described by Beecher and Macmillan (3) have not been characterized in detail, there was evidence that the antibody which was supposed to react with the L₂ component also bound to a 36-kDa protein and that the antibody against the B component cross-reacted with a protein having a molecular mass of approximately 45 kDa. In a recent publication (27), it was additionally stated that the latter antibody reacts with an extracellular protein of approximately 100 kDa. This protein was already detectable in early log-phase cultures of B. cereus, and it was suggested that it might be a precursor to the B component or even to all three components of HBL. This theory could also be an explanation for the unexpected cross-reactivity of the antibody against the L₁ component (1C2) with a not-yet-identified exoprotein showing an M_r of 93,000 (Fig. 3), which was observed in our experiments. It is interesting to note that this antibody also gave positive results in the indirect EIA if HBL-negative but highly cytotoxic strains were tested (Table 3). This is a further indication that this antibody reacts also with the 39-kDa protein of the enterotoxin complex described by Lund and Granum (24, 25). Although the same authors (25) reported some degree of amino acid homology between L₁ and the 39-kDa component, which might explain the cross-reactivity of antibody 1C2, further studies have to be performed to fully clarify the relevance of these findings.

To demonstrate the applicability of the monoclonal antibodies for the characterization of the enterotoxic activity of B. cereus strains, cell-free culture supernatants were tested for reactivity in the indirect EIA and the results were compared to the cytotoxic activities (Table 3). As expected, it could be demonstrated that all strains producing the HBL complex were also highly cytotoxic but also that cytotoxicity was not necessarily correlated with the production of the HBL enterotoxin complex. B. cereus strains, such as 0075/95, MHI Hi 56, and others, were highly cytotoxic but produced no detectable amounts of the B and L₂ components. Since it has been postulated that all three components of HBL are cotranscribed and constitute an operon (22, 27), it might be concluded that the positive results obtained for these strains in the assay for L₁ are due to the reactivity of antibody 1C2 with an unrelated substance, e.g., the 39-kDa component of NHE, rather than to the presence of L₁. Another interesting finding was the different sensitivities of the indirect EIA with antibodies 1A12 and 8B12, which both react with the L₂ component. A possible explanation could be that the affinity of antibody 1A12 for L₂ is much lower than the affinity of antibody 8B12, but steric hindrance in the indirect test system, where the antigen is coated to the plastic surface, could also be a reason for that difference.

In conclusion, this study describes the production and characterization of the first complete set of high-affinity monoclonal antibodies against all three components of the HBL enterotoxin produced by B. cereus. The antibodies enable the specific and sensitive detection of these compounds in culture supernatants and may be used for the quantitative evaluation of the toxin profile of B. cereus strains. In addition, these antibodies represent versatile tools for studies on the mode of action of B. cereus enterotoxins and on the kinetics of toxin production and facilitate the purification of the single toxin components by applying immunoaffinity chromatography.