Production and characterization of rabbit polyclonal sera against Shiga toxins Stx1 and Stx2 for detection of Shiga toxin-producing Escherichia coli
Article first published online: 24 SEP 2008
© 2008 The Societies and Blackwell Publishing Asia Pty Ltd
Microbiology and Immunology
Volume 52, Issue 10, pages 484–491, October 2008
How to Cite
Mendes-Ledesma, M. R. B., Rocha, L. B., Bueris, V., Krause, G., Beutin, L., Franzolin, M. R., Trabulsi, L. R., Elias, W. P. and Piazza, R. M. F. (2008), Production and characterization of rabbit polyclonal sera against Shiga toxins Stx1 and Stx2 for detection of Shiga toxin-producing Escherichia coli. Microbiology and Immunology, 52: 484–491. doi: 10.1111/j.1348-0421.2008.00068.x
- Issue published online: 24 SEP 2008
- Article first published online: 24 SEP 2008
- Received 28 August 2007; accepted 18 June 2008
- polyclonal rabbit antisera;
- Shiga toxin-producing E. coli
STEC has emerged as an important group of enteric pathogens worldwide. In this study, rabbit polyclonal Stx1 and Stx2 antisera were raised and employed in the standardization of immunoassays for STEC detection. Using their respective antisera, the limit of detection of the toxin was 35.0 pg for Stx1 and 5.4 pg for Stx2. By immunoblotting, these antisera recognized both toxin subunits. Cross-reactivity was observed in the A subunit, but only Stx2 antiserum was able to neutralize the cytotoxicity of both toxins in the Vero cell assay. Six stx-harboring E. coli isolates were analyzed for their virulence traits. They belonged to different serotypes, including the O48:H7, described for the first time in Brazil. Only three strains harbored eae, and the e-hly gene and hemolytic activity was detected in five strains. Three isolates showed new stx2 variants (stx2v-ha and stx2vb-hb). The ELISA assay detected all six isolates, including one VCA-negative isolate, while the immunodot assay failed to detect one isolate, which was VCA-positive. In contrast, the colony-immunoblot assay detected only one VCA-positive isolate. Our results demonstrate that among the immunoassays developed in this study, the immunodot, and particularly the ELISA, appear as perspective for STEC detection in developing countries.
bovine serum albumin
diarrheagenic E. coli
- E. coli
enterohemorrhagic E. coli
enzyme-linked immunosorbent assay
nitro blue tetrazolium chloride/ 5-bromo-4-chloro-3-indolyl phosphate toluidine salt
Shiga toxin-producing Escherichia coli
Shiga toxin 1
Shiga toxin 2
tryptic soy agar
tryptic soy broth
Vero cell assay
STEC constitutes an important group of emerging enteric pathogens. STEC was first discovered in 1977 (1) and first associated with HUS in 1983 (2). Since then, STEC has been identified as a major food-borne pathogen in different countries all around the world. Epidemiological studies have shown that STEC are present in numerous serotypes in cattle and other domestic animals worldwide, independent of geographic region, species, husbandry methods and climate (3–5). Prevalence rates and types of STEC in humans also vary among different regions, even in the same country (4).
Among the DEC pathotypes, STEC has been considered emergent in Brazil. In different regions of our country, the isolation rate of STEC strains ranges from 12 to 71% in animals from dairy farms, beef farms and slaughterhouses for cattle (6–11) and also for sheep (12). In spite of the low prevalence (13–16), the occurrence of clinically important STEC strains associated with disease in humans, including bloody diarrhea, hemolytic anemia and HUS, has been demonstrated in the last few years in Brazil (17, 18).
The expression of Stx1 and/or Stx2 is characteristic of STEC strains, and they function as major virulence factors. Both toxins are similar in biological activity but immunologically distinct (19). The only way to identify all types of STEC in any kind of test sample is by the detection of these toxins produced by the bacteria, or of the genes associated with Stx production. Additional factors produced mainly by strains that are directly associated with human disease, and collectively referred to as EHEC, include intimin and enterohemolysin (20, 21).
Numerous assays for the diagnosis of STEC have been developed (reviewed in Bettelheim and Beutin ). The Vero cell toxicity test detects functionally active toxin and is often used as the gold standard for evaluation of immunoserological tests. Stx-specific PCR detects gene sequences whether or not they are expressed (22). Commercially available immunological test kits are offered by different companies (23), and in different reference laboratories some of the evaluated kits have shown variability in sensitivity and specificity. One of these, the Premier-EHEC kit, showed a high specificity but lower sensitivity compared to the Vero cell test and PCR in two investigations (24, 25). Beutin et al. (4) have evaluated the Ridascreen-EIA, which is based on ELISA detection. The authors found this kit to be suitable for detecting all known types of Stx, but less applicable in testing samples where low amounts of Stx are expected, such as mixed cultures and certain Stx2 variants. These commercial kits are economically unaffordable in developing countries.
The emergence of these pathogens worldwide, including in our country, has made it necessary to be able to identify them in routine diagnosis. For this reason, we raised polyclonal anti-Stx1 and anti-Stx2 antisera in rabbits. The performance of these antisera was evaluated by ELISA, immunoblotting and neutralization assays against purified toxins, and by ELISA, immunodot and colony immunoblot assays with Shiga toxin producing or non-producing E. coli. In addition, we further characterized the virulence factors of the STEC strains recently isolated in Brazil (14).
MATERIAL AND METHODS
Purified Stx1 and Stx2
Anti-Stx1 and Stx2 polyclonal antibodies
New Zealand rabbits were immunized intradermically with 1.5 ml of 20 μg/ml of Stx1 or Stx2 toxoids in Montanide ISA 50V (SEPPIC, Paris, France). After 14 days, the same dose of antigens was injected intramuscularly, followed by an intravenous injection 40 days after the first injection with 0.5 ml of 15 μg/ml of the antigens in Montanide ISA 50V (SEPPIC). Serum samples were obtained seven days after the last antigen injection (28). Serum samples, to be used as negative controls in specific antibody evaluation, were obtained just before immunization by auricular venipuncture.
Sera reactivity and limit of detection
The reactivity and limit of detection of the immune sera were tested by ELISA (29) with some modifications. For immune serum reactivity, microplates (Nunc-Immuno MaxiSorp, Rochester, NY, USA) were coated overnight at 4°C with a solution of 0.01 M, pH 7.4, PBS containing 1 μg/ml of Stx1 or Stx2. After blocking (1% BSA in PBS) for 30 min at 37°C, the microplates were incubated for 1 hr at 37°C with anti-Stx1 or anti-Stx2 polyclonal antiserum diluted from 1:50 to 1:6400. For limit of detection assays, the microplates were coated with the toxins at different concentrations (from 5.0 μg to 5.4 pg). After blocking (1% BSA in PBS) for 30 min at 37°C, the microplates were incubated with anti-Stx1 or anti-Stx2 polyclonal antiserum diluted 1:50. Both reactions were followed by incubation with goat anti-rabbit peroxidase-conjugated antibodies (Zymed, San Francisco, CA, USA), diluted 1:5000, for 1 hr at room temperature. ELISA was developed with 0.5 mg/ml of OPD (Sigma Aldrich, St Louis, MO, USA) plus H2O2, and the reaction was stopped by the addition of 1N HCl. The absorbance was measured at 492 nm in a Multiskan EX ELISA reader (Labsystems, Milford, MA, USA).
Stx1 and Stx2 characterization by immunoblotting
The reactivity of antibodies to purified Stx1 and Stx2 toxins and toxoids was tested by immunoblotting using polyclonal anti-Stx. Briefly, 5 μg per slot of toxins or toxoids were applied to a 15% SDS-polyacrylamide gel (30, 31). After electrophoresis, the separated proteins were transferred to a nitrocellulose membrane (Hybond C-Extra, Amersham Biosciences, Little Chalfont, UK) at 100V for 18 hr at 4°C. The membrane was blocked with 3% BSA for 2 hr and reacted with 1:50 dilution of Stx1 or Stx2. The membrane was then washed and incubated for 2 hr with alkaline phosphatase goat anti-rabbit IgG (1:10 000). After washing, the substrate NBT/BCIP was added and the reaction stopped by adding distilled water.
Vero cell cytotoxicity
The VCA was performed as described previously (32). Neutralization of Shiga toxins was done by incubating culture supernatant of STEC producing Stx1 and/or Stx2 with 1:500 dilutions of Stx1- and Stx2-specific rabbit antisera for 2 hr at 37°C or 1 × 10−6μg of purified toxins with different serum dilutions (from 1:50 to 1:6400) for 2 hr at 37°C. After incubation, 500 μl of serum-treated samples were tested as described by Beutin et al. (32). Non-specific rabbit antiserum was used as negative control. Cytotoxicity of VCA positive isolates and neutralization of this activity were determined as described by Gentry and Dalrymple (33), according to which the percentage of cytotoxicity was measured after staining the cells with crystal violet and calculated using the formula control A620 nm minus sample A620 nm divided by control A620 nm.
We employed 12 E. coli isolates from our bacterial collection; six of them harbored the stx gene, and the other six isolates were E. coli from microbiota with no virulence markers for diarrheagenic E. coli pathotypes (14). In addition, C7–88 (20) and EDL933 (34) strains were included as positive controls for Stx1 and Stx2 respectively, and E. coli DH5α (35) was used as a negative control.
For the Vero cell assay, immunodot assay and ELISA, the strains were cultivated in TSB at 37°C and 180 rpm for 18 hr. The cultures were centrifuged at 13 000 g for 15 min, and the supernatant kept at −20°C until use. For the colony immunoblot assay, bacterial isolates were cultivated in TSA at 37°C for 18 hr.
Development and standardization of Stx-specific immunoassays
Stx-specific antisera were adsorbed against E. coli DH5α to remove possible cross-reactivity with other E. coli proteins.
Immunodot assay was performed as described before (36) according to which 100 μl of bacterial supernatant was applied onto a nitrocellulose membrane by Dot Blot Filtration Manifolds (Pharmacia, San Francisco, CA, USA). After blocking with 3% BSA in PBS containing 0.05% Tween 20 for 1 hr, the membranes were incubated for 1 hr with Stx1 or Stx2 antiserum individually (diluted 1:50) or in a mixture of anti-Stx1 and anti-Stx2 antisera diluted 1:50, followed by incubation for 1 hr with goat anti-rabbit peroxidase-conjugated antibodies, diluted 1:15 000. The reactions were developed with 0.16 mg/ml of DAB, (Sigma Aldrich) plus H2O2, and stopped by the addition of distilled water.
For the colony immunoblot assay, bacterial growths were transferred to a nitrocellulose membrane after incubation with 0.01 M PBS, pH 7.2, containing 1 mg/ml of polymyxin B at 37°C for 30 min. After blocking with 3% BSA in PBS containing 0.05% Tween 20 for 1 hr, the reaction was performed under the same conditions as described for the immunodot assay.
For the ELISA, flat-bottom wells of polystyrene plates (Nunc-Immuno PolySorp) were coated for 16–18 hr at 4°C with 100 μl of bacterial supernatants. After blocking with 3% BSA in PBS containing 0.05% Tween 20 for 30 min, the wells were incubated for 30 min with Stx1 or Stx2 antiserum individually (diluted 1:50) or in a mixture of anti-Stx1 and anti-Stx2 antisera diluted 1:50, followed by incubation for 30 min with goat anti-rabbit peroxidase-conjugated antibodies diluted 1:15 000. The reaction was developed with 0.05 M phosphate–citrate buffer (pH 5.0) containing OPD solution at 0.5 mg/ml plus H2O2 as substrate. The absorbance was measured at 492 nm on a Multiskan EX ELISA reader (LabsystemsMilford, MA, USA). Wells incubated with pre-immune sera were used as negative control. ELISA absorbance (A492 nm) data were analyzed by mean and standard error using GraphPad Prism 3.00 (GraphPad Software, La Jolla, Ca, USA) using data from quadruplicates of five independent experiments.
Virulence characteristics of Shiga toxin-producing E. coli strains
The strains harboring stx were further analyzed for their serotype, hemolytic activity, biochemical characteristics (sorbitol- and lactose-fermenting ability and β-glucuronidase production) and the presence of eae and e-hly encoding genes. Serotyping of O:H antigens was performed as described earlier (37). The hemolytic phenotype was investigated on washed sheep blood agar (enterohemolysin agar; Oxoid, Wesel, Germany). Production of Shiga toxins was also confirmed by the Ridascreen-EIA Verotoxin enzyme immunoassay (R-Biopharm AG, Darmstadt, Germany, Art. N° C2201), following the manufacturer's instructions. The virulence genes eae, E-hlyA, and the stx genetic variants were investigated by PCR. DNA of STEC reference strains, Stx-negative controls and test strains were prepared with the Dynabeads DNA Direct universal kit (Invitrogen, Karlsruhe, Germany) according to the protocol supplied by the manufacturer. Aliquots of 1–2 μl of purified DNA were used as template in the PCR assays. Primers, size of amplicons and amplification conditions for eae and E-hlyA were as described by Leomil et al. (38). Detection of stx genetic variants were performed employing the primers, amplification conditions and size of amplicons as described by Beutin et al. (4), and the nomenclature for describing stx genes and their products is according to Scheutz et al. (23).
Antibody titers against the toxins showed that the mean absorbances at 492 nm reached 1 at serum dilutions of 1:800 and 1:6400 for Stx1 and Stx2 respectively. The limit of detection for the antisera was defined as an absorbance of 0.1, where pre-immune rabbit serum served as the negative control. The lowest amount of toxin detected by ELISA was 35.0 pg for Stx1, using anti-Stx1 antiserum (Fig. 1a), and 5.4 pg for Stx2, using anti-Stx2 antiserum (Fig. 1b). By immunoblotting analysis, both antisera recognized their respective A (35 kDa) and B (7.5 kDa) toxin subunits (Fig. 2, lanes 1 and 7). Also, both antisera recognized their own toxoids as a smear with apparent molecular weight of 80 kDa (Fig. 2, lanes 2 and 8). Cross-reactivity between antisera occurred only with the A subunit of the toxins (Fig. 2, lines 3 and 5).
Stx1 antiserum was able to inhibit 1 × 10−6 μg of purified Stx1 toxin in a 1:6400 dilution while, in contrast, it did not inhibit Stx2 toxin at any dilution (Fig. 3, rows A and C). On the other hand, Stx2 specific antiserum was found to neutralize Stx1 toxin and Stx2 toxin at dilutions of 1:800 and 1:6400, respectively (Fig 3, rows B and D). On testing the culture supernatant from six stx-harboring E. coli isolates in the Vero cell assay, five of them showed a cytotoxic effect, and regardless of the stx subtype, both antisera inhibited up to 85% of the cytotoxicity of the isolates (Table 1).
|Strain||Serotype||Stx subtype||Gene detection||E-hly expression||VCA||Neutralization assay (%)||Colony immunoblot||Immunodot||ELISA|
The polyclonal antisera obtained in this study were employed in the development and standardization of immunoserological assays. Both antisera individually, or in a mixture of equal volumes of Stx1 and Stx2, did not recognize any of the E. coli from microbiota. By colony immunoblot the mixture of antisera detected only one isolate (Table 1). In contrast, by immunodot assay, Stx1 plus Stx2 antisera failed to detect one isolate, and by ELISA this mixture was able to detect all isolates (Table 1). Figure 4 shows the mean values of all Stx-producing, non-producing strains and controls. Statistical analysis showed that the difference between Stx producing- and non-producing strains by mean and standard error were considered significant (P = 0.04, R2 = 0.92), also variances between groups were significant (P = 0.0096). On the other hand, the reactivity of each antiserum individually was variable (data not shown).
The six isolates which showed cytotoxic effects on Vero cells and/or stx were subject to further characterizations. First of all, the production of Shiga toxins was confirmed by Ridascreen immunoassay (R-Biopharm AG, Darmstadt, Germany), in which only isolate 53 showed a weak reaction. Serotyping analyses showed that they belonged to different serotypes; but only two of them were characterized as the most common EHEC serotypes O157:H7 and O26:H11 (Table 1). All of them were lactose- and sorbitol-fermenting strains, and with the exception of isolate 53, the others were β-glucuronidase-producing strains. Isolate 4123 being the exception, the e-hly gene and hemolytic activity on blood agar were detected in the other five isolates (Table 1). Data from stx subtyping showed that three of them demonstrated new variants in their subtypes of toxin (stx2v-ha and stx2vb-hb). Three isolates, two of them belonging to the serotypes O157:H7 and O26:H11, showed the intimin-encoding gene (Table 1).
In the past, human STEC infections were restricted to sporadic cases of non-bloody diarrhea in Brazil (18, 39). STEC has recently been implicated as the cause of HUS and bloody and severe diarrhea (17). A number of reports have demonstrated the prevalence of several STEC serotypes in cattle (6–12) and humans in Brazil (16). Reliable and low-cost STEC detection assays are not currently available in our country, which is similar to the situation in other developing countries. The above-mentioned information led to this study, in which we raised rabbit polyclonal Stx1 and Stx2 antisera. Concerning the reactivity of these antisera with the toxins, both of them showed a high detection limit and cross-reactivity occurred only with the A subunit of the toxins. In addition, Stx2 antiserum neutralized both toxins.
Despite having a high capacity to detect purified Stx, the two antisera, both individually and mixed, showed weak reactivity in the colony immunoblot assay. The possibility that this was due to the incubation period of 24 hr, which would have allowed secretion of the toxin into the medium was considered. In order to check this possibility, we also tested this assay after 3 hr of bacterial incubation, and obtained the same results (data not shown). The lower reactivity observed in the colony immunoblot assay confirms previous data from Strockbine et al. (40) and Perera et al. (41), which showed that only E. coli expressing moderate or high levels of toxins can be detected by this method. In contrast, when we used culture supernatant, either on nitrocellulose or on microtiter plates, reactivity was increased for both immunodot assay and ELISA, due to toxin exposure in the culture supernatant, which promotes antibody accessibility.
Variability in sensitivity among immunoassays has been previously described (24, 25, 42–45), but one should note that those immunoassays used both monoclonal and polyclonal antibodies for Stx detection. Despite the use of rabbit polyclonal antisera in this study, we observed that the Stx1 and Stx2 antisera recognized the STEC isolates by ELISA.
Six isolates which harbored the stx1 and/or stx2 genes, and were recently isolated in our country, were used in this study (14). They showed interesting phenotypic and genotypic features in regard to STEC characteristics. The O157:H7 serotype strain was the only one in which we did not observe a cytotoxic effect; in addition, this one was isolated from a child belonging to the control group (i.e., not displaying gastrointestinal symptoms). The other five strains were isolated from patients with diarrhea. In addition to the O157:H7 strain, the O26:H11 and the OR:NM are also intimin-expressing strains. Comparing the studies conducted by Vaz et al. (16, 46) and Cergole-Novella et al. (47), the finding of a STEC strain belonging to the serotype O48:H7 is here described for the first time in Brazil.
Concerning stx subtypes, the stx2vb-hb variant, which had previously been described among animal strains (48–50), was found in this study in human isolates. In contrast, Cergolle-Novella et al. (47) described finding this genotype in isolates from cattle in Brazil. The other genotypes found are frequently isolated from humans (51).
Taken together, our results demonstrate that the immunodot and ELISA methods employing a mixture of rabbit anti-Stx1 and Stx2 antisera, and standardized in this study, are appropriate techniques for the detection of STEC.
In spite of the better performance of the ELISA for STEC detection, this assay has not yet been evaluated in terms of industrial quality control and commercially availability. This is our future goal. However, one can conclude that this assay is reproducible, fast, easy to perform and reliable. After bacterial supernatant coating of the microplates the development of this assay takes 105 min, and the results show inter and intra-test reproducibility demonstrated by error bars. Considering only the cost of all necessary materials, the estimated cost of the assay would be U$70 per 96 detections, which is realistically affordable for developing countries.
We are indebted to Dr. Cheleste M. Thorpe and the staff of Tufts-New England Medical Center (Tufts University School of Medicine) for kindly providing purified Stx1 and Stx2. This work was supported by FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo, Grants 99/09485-0 to R.M.F.P. and 04/12136-5 to W.P.E.) and FINEP (Financiadora de Estudos e Projetos).
- 42006) Comparative evaluation of the Ridascreen® Verotoxin enzyme immunoassay for detection of Shiga-toxin producing strains of Escherichia coli (STEC) from food and other sources. J Appl Microbiol 102: 630–39., , , , , , (
- 232001) Nomenclature of verotoxins. In: DuffyG., GarveyP., McDowellD.A. eds. Verocytotoxigenic E. coli. Trumbull : Food and Nutrition Press, pp. 447–52., , , (
- 351989) Molecular cloning: a laboratory manual. New York : Cold Spring Harbor Laboratory Press., , (
- 382005) Characterization of two major groups of diarrheagenic Escherichia coli O26 strains which are globally spread in human patients and domestic animals of different species. FEMS Microbiol Lett 15: 335–42., , , , (