A PCR–ELISA for detecting Shiga toxin-producing Escherichia coli

Microbes and Infection 4 (2002) 285–290 www.elsevier.com/locate/micinf

Original article

A PCR–ELISA for detecting Shiga toxin-producing Escherichia coli Beilei Ge a, Shaohua Zhao a,1, Robert Hall b, Jianghong Meng a,* b

a Department of Nutrition and Food Science, University of Maryland, College Park, MD 20742, USA Center for Food Safety and Applied Nutrition, Food and Drug Administration, 200 C Street, SW Washington, DC 20204, USA

Received 25 June 2001; accepted 20 November 2001 First published online 14 February 2002

Abstract A sensitive and specific PCR–ELISA was developed to detect Escherichia coli O157:H7 and other Shiga toxin-producing E. coli (STEC) in food. The assay was based on the incorporation of digoxigenin-labeled dUTP and a biotin-labeled primer specific for Shiga toxin genes during PCR amplification. The labeled PCR products were bound to streptavidin-coated wells of a microtiter plate and detected by an ELISA. The specificity of the PCR was determined using 39 bacterial strains, including STEC, enteropathogenic E. coli, E. coli K12, and Salmonella. All of the STEC strains were positive, and non-STEC organisms were negative. The ELISA detecting system was able to increase the sensitivity of the PCR assay by up to 100-fold, compared with a conventional gel electrophoresis. The detection limit of the PCR–ELISA was 0.1–10 CFU dependent upon STEC serotypes, and genotypes of Shiga toxins. With the aid of a simple DNA extraction system, PrepMan, the PCR–ELISA was able to detect ca. 105 CFU of STEC per gram of ground beef without any culture enrichment. The entire procedure took about 6 h. Because of its microtiter plate format, PCR–ELISA is particularly suitable for large-scale screening and compatible with future automation. © 2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: PCR; ELISA; Shiga toxin-producing E. coli

1. Introduction Shiga toxin-producing Escherichia coli (STEC) were first recognized as human pathogens in 1982 when E. coli O157:H7 caused two outbreaks of hemorrhagic colitis associated with consumption of undercooked ground beef [1]. Since then, more than 200 serotypes of STEC have been isolated from animals, food and other sources [2,3]. Although not all STEC have been shown to cause human illness, O157:H7, O26:H11, O111:NM and several other serotypes have caused outbreaks of hemorrhagic colitis and hemolytic uremic syndrome (HUS) worldwide [4]. The Center for Disease Control and Prevention (CDC) estimates that in the United States, E. coli O157:H7 causes 73,480 illnesses and 61 deaths each year, and non-O157 STEC

* Corresponding author. Tel.: +1-301-405-1399; fax: +1-301-314-9327. E-mail address: [email protected] (J. Meng). 1 Current address: Division of Animal and Food Microbiology, Office of Research/Center for Veterinary Medicine, Food and Drug Administration, Laurel, MD 20708 © 2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. PII: S 1 2 8 6 - 4 5 7 9 ( 0 2 ) 0 1 5 4 0 - X

account for an additional 37,740 cases with 30 deaths [5]. Eighty-five percent of these cases are attributed to foodborne transmission. Among the 196 outbreaks reported in the United States for which a vehicle has been identified, 48 (33.1%) were associated with ground beef, 4 (2.8%) with raw milk, and 3 (2.1%) with roast beef [6]. Other foods and environmental vehicles have also been implicated in the transmission of STEC infections. With the increase in reports of infections with E. coli O157:H7 and other STEC, great attention has been given to the development of methods for detecting these pathogens, including culture isolation, serological tests, DNA probes, and PCR assays. In contrast to culture isolation and serology, PCR provides a rapid and sensitive alternative. DNA sequences used as targets in PCR assays include genes encoding Shiga toxin(s) (Stx) [7], intimin [8], enterohemorrhagic E. coli hemolysin [9] and β-glucuronidase [10], and a unique fragment upstream of the eae gene [11]. Conventional PCR procedures followed by gel electrophoresis or Southern hybridization for detecting pathogenic microorganisms limit the number of samples that can be

286

B. Ge et al. / Microbes and Infection 4 (2002) 285–290

analyzed during one electrophoresis run [12]. Simple and rapid methods for the analysis of PCR products have been developed to facilitate large-scale screening, including enzyme-linked immunosorbent assay (ELISA). A digoxigenin (DIG)-ELISA kit manufactured by Roche Diagnostics Corporation (Indianapolis, IN) provides a convenient, nonradioactive detection solution for PCR products in a microtiter plate format. A biotin-labeled primer is used together with DIG-11'-dUTP, and they are both incorporated into PCR products during amplification. The amplified products are immobilized onto the streptavidin-coated surface of a microtiter plate via the strong affinity of the avidin–biotin interaction, and then the amplicons are detected with an anti-DIG–peroxidase conjugate through the substrate 2,2'azino-di-[3-ethylbenzthiazoline sulfonate] (ABTSt). Shiga toxin-producing E. coli all produce one or both Shiga toxins (Stx1 and Stx2), which are key virulence factors causing human disease. Molecular studies on Stx1 from different E. coli strains revealed that Stx1 is either

completely identical to the Stx of Shigella dysenteriae type 1 or differs by only one amino acid. Unlike Stx1, toxins of the Stx2 group are more divergent, and several subgroups have been identified, including Stx2, Stx2c, Stx2d, Stx2e, and Stx2f [13]. The variability primarily reflects sequence diversity in the B subunit, which in turn may alter receptor binding preference or affinity. In this study, we used the sequences of A subunits of stx genes as the target and combined PCR and ELISA procedures to develop a rapid and sensitive assay that offers a potential for large-scale screening of STEC.

2. Materials and methods 2.1. Bacterial strains and preparation of DNA templates Table 1 identifies the study organisms selected from our culture collection at the Department of Nutrition and Food

Table 1 Virotypes/species, serotypes, sources and Shiga toxin (Stx) genotypes of E. coli and other bacteria used in this study. Strain ID

Virotype/Species

Serotype

Source

Stx genotype

PCR result

umd 11* umd 12 umd 13 umd 19* umd 21 umd 95 umd 200* umd 201 umd 263 umd 82 umd 84 umd 174 umd 173 umd 170 umd 167 umd 166 umd 165 umd 168* umd 164 umd 161 umd 162 umd 204 umd 160 umd 159 umd 145 umd 144 umd 141* umd 142 umd 140 umd 206 umd 135 umd 205 umd 131 umd 146 umd 147 umd 139 umd 130 umd 348 umd 196 .

STEC

O157:H7 O157:H7 O157:H7 O157:H7 O157:H7 O157:H7 O157:H7 O157:H7 O157:H7 O157:NM O157:NM O126:H8 O125:NM O113:K75:H21 O111:H11 O111:H11 O111:H8 O111:NM O111:NM O103:H2 O103:H2 O91:H21 O91:H21 O88:H49 O46:H38 O45:H2 O26:H11 O26:H11 O22:H8 O5:NM O5:NM U:NM OR:H9 O55:H7 O55:H7 O18:K1:H7 K-12 typhimurium typhimurium

ground beef ground beef ground beef calf call human sheep sheep human calf calf unknown human human calf calf calf calf human human steer dairy cow sheep ground beef beef human calf human ground beef human sheep sheep human unknown unknown unknown unknown unknown unknown

1, 2 1, 2 1, 2 2 2 1 1 1 1, 2 1, 2 1, 2 1 1 2 1 1 1 1, 2 1 1, 2 1 1 1, 2 2 1, 2 1 1 1 2 2 1 1, 2 2 none none none none none none

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + – – – – – –

EPEC

E. coli Salmonella

* Strains were also used for the sensitivity test.

B. Ge et al. / Microbes and Infection 4 (2002) 285–290

287

Fig. 1. Sequence alignment of Shiga toxin genes 1 (stx1) and 2 (stx2). Oligonucleotide sequences in bold were used as primers of a PCR assay for detecting STEC.

Science, University of Maryland. E. coli O157:H7 (9 strains), STEC of various serotype (24 strains), enteropathogenic E. coli (3 strains), E. coli K-12 (1 strain) and Salmonella (2 strains) represented the range of relevant genotypes and controls. The STEC strains included ten serotypes isolated from humans, food, cattle, and sheep. The Shiga toxin genotypes are listed in Table 1. The bacteria were routinely cultured using tryptic soy broth or agar (Difco, Detroit, MI). To prepare a DNA template for PCR, a single colony on a tryptic soy agar plate was selected, and suspended in 500 µl of Tris–EDTA (TE: 10 mM Tris, pH 8.0, 1 mM EDTA) buffer. The bacterial suspension was then heated for 10 min at 95 °C in a dry bath and stored at –30 °C until use.

dCTP, and dGTP, 190 µM dTTP, 10 µM DIG-11'-dUTP, 1 U of Taq DNA polymerase, 1× PCR buffer (consisting of 10 mM Tris–HCl, 50 mM KCl, pH 8.3), 1.5 mM MgCl2, and 10 µl of DNA template. The PCR assay consisting of 30 cycles of denaturation for 45 s at 94 °C, primer annealing for 45 s at 50 °C, and extension for 45 s at 72 °C was performed using a DNA thermal cycler (GeneAmp PCR System 9600; Perkin-Elmer Corp., Branchburg, NJ). A

2.2. Target, PCR primers and reaction conditions The genes encoding Shiga toxins 1 and 2 were selected as the target of the assay. The sequences used for designing PCR primers specific for Shiga toxin genes were determined based on the homologous sequences of the two genes (Fig. 1) obtained from GenBank (Accession numbers M19473 and X07865). Forward (F1) and reverse (R1) primers (bold sequences in Fig. 2) were synthesized by Life Technologies (Gaithersburg, MD). Biotin was incorporated onto the 5' end of primer R1. The PCR assay was performed using the reagents provided in the DIG-labeling kit (Roche, Indianapolis, IN) in a final reaction volume of 50 µl. Each reaction mixture consisted of 0.3 µM of each primer, 200 µM each of dATP,

Fig. 2. Gel picture (A) and ELISA reading (B) of PCR amplicons of 19 bacterial strains. Strain umd 130 is E. coli K-12 (negative control). Strains umd 139, 196, and 348 are E. coli O18:K1:H7, Salmonella typhimurium, and S. typhimurium, respectively; the remainder are STEC strains. The dash line in B indicates the cut-off value which is two times greater than that of the negative control strain.

288

B. Ge et al. / Microbes and Infection 4 (2002) 285–290

reagent control (sterile water) was included for surveillance of interior contamination. 2.3. Detection of PCR products 2.3.1. Agarose gel electrophoresis Ten microliters of each PCR product was loaded into a well of a 2% agarose gel containing 0.5 µg/ml ethidium bromide. A 1-kb ladder (Life Technologies) was used as the molecular weight marker. PCR products were electrophoresed and visualized under UV light, and gel images were stored using a gel documentation system (Gel Doc 1000; BioRad, Hercules, CA). 2.3.2. ELISA ELISA reagents were provided in the DIG-detection kit (Roche). The manufacturer’s protocol was modified by using a biotin-labeled primer during PCR instead of a biotin-labeled capture probe that was added during detection. Each amplicon (5 µl) was diluted in 445 µl of hybridization buffer, and 200 µl of the amplicon suspension were added in duplicate to streptavidin-coated wells of a microtiter plate. The plate was incubated at 42 °C with shaking for 1 h. After five washes with washing buffer, 200 µl of 10 µU/µl anti-DIG–peroxidase conjugate was added to each well and incubated at 42 °C with shaking for 30 min. The plate was washed five times with washing buffer, and 200 µl of substrate solution (1 mg/ml 2,2'-azino-di-[3ethylbenzthiazoline sulfonate] (ABTSt)) were added. Color was allowed to develop at 42 °C in the dark without shaking for 5–10 min. ELISA absorbency was measured at 405 nm using an ELISA reader (Bio-Tek Instruments Inc. ELX 800; Winooski, VT). 2.4. Specificity and sensitivity The 39 bacterial strains in Table 1 were used to determine the specificity of the assay. Three E. coli O157:H7 (umd11, umd19, and umd200), one E. coli O111:NM (umd168), and one E. coli O26:H11 strain (umd141) were selected to determine its sensitivity. Strain umd11 and umd168 produced both Shiga toxins 1 and 2, whereas umd141 and umd200 produced only Stx1, and umd19 produced only Stx2. Tenfold serial dilutions of overnight broth cultures prepared in sterile saline solution were quantified using standard plating methods. 2.5. Preparation of ground beef samples for PCR Strain umd263, a nalidixic acid-resistant E. coli O157:H7, was used to evaluate the assay for its capability to detect STEC in ground beef. Tenfold serial dilutions of an overnight culture of umd263 were made in 0.1% buffered peptone water, resulting in bacterial concentrations ranging from approximately 108 CFU/ml to 1 CFU/ml. Standard spread plates were used to determine the number of CFU. A

novel purification system, PrepMan™ Sample Preparation Reagent (Perkin-Elmer), was used to extract STEC DNA from meat samples according to the manufacturer’s instructions, with modification. Briefly, 25 g ground beef in a 50-ml centrifuge tube were inoculated with different dilutions of E. coli O157:H7 and suspended in 25 ml buffered peptone water. The sample was homogenized and centrifuged at 150 g for 5 min. One milliliter of middle layer liquid (aqueous non-fatty suspension) was transferred into a 1.5-ml microcentrifuge tube, and centrifuged at 13,800 g for 2 min to separate the bacterial cells. The supernatant was discarded, and 200 µl PrepMan solution added. The tube was vortexed vigorously, and heated for 10 min in a boiling water bath. After removal from the bath, the tube was cooled and centrifuged at 13,800 g for 2 min. To 50 µl of the supernatant were added 50 µl of distilled water, resulting in a template ready to be used for PCR reactions.

3. Results and discussion Conventional PCR has been used to detect stx genes for screening STEC in clinical specimens [14] as well as in food samples including ground beef and raw milk [15,16]. The technique has proven to be both sensitive and specific. Most PCR assays for detecting STEC employ two pairs of primers (multiplex) specific for stx1 and stx2, respectively. The advantage of such a design includes simultaneous determination of Shiga toxin genotypes of STEC. However, a multiplex PCR assay is often complex and more sensitive to reaction conditions including Mg2+ and Taq polymerase concentrations. In addition, a PCR–ELISA procedure is not suitable for multiplex PCR because the ELISA reporting system was not designed for more than one target. In the present study, two primers (F1 and R1) specific for both stx1 and stx2 were designed based on the homologous sequences of the two genes (Fig. 1). The primers generated an approximately 220-bp fragment during the PCR amplification. Although there were a few mismatches of nucleotide bases in the primer sequences between stx1 and stx2 (3 out of 22 in F1 and 2 out of 21 in R1), special attention was paid to ensure a complete match at the 3' ends of the primers. Therefore, the primers enabled the amplification of both genes although they were based on the stx2 sequence. An ELISA system was selected to detect PCR products in order to develop a method that was suitable for rapid large-scale screening for STEC and also more sensitive than conventional gel electrophoresis. DIG-labeling and DIGdetection kits were used for the PCR amplification and direct analysis of PCR products, which provided results within 6 h. Several studies have shown that PCR–ELISA is a useful technique for detecting microbial pathogens, particularly for large-scale screening [17–20]. Luk et al. [12] reported a PCR–ELISA assay for screening of Salmonella serogroup D in clinical specimens. With overnight enrichment, the assay was able to correctly identify all isolates of

B. Ge et al. / Microbes and Infection 4 (2002) 285–290

serogroup D Salmonella within 10–22 h. In contrast, the conventional culture method took several days and could not differentiate serogroup D strains from other Salmonella. A PCR–microplate capture hybridization method was developed for detecting STEC [21]. Biotin-labeled probes were used to capture PCR amplicons of stx genes. The assay had excellent specificity with a detection limit of 200 cells. A quantitative PCR–ELISA was able to detect one cell of E. coli per gram of oyster [22]. A DIG-labeled primer and biotinylated probe were used in this PCR–ELISA system. In the present study, unlike other PCR–ELISA assays, we modified the manufacturer’s protocol by using a biotinlabeled primer during PCR instead of a biotin-labeled capture probe that is added during detection. The omission of the probe hybridization step reduces the experimental time by at least 1 h in performing the assay, and in addition eliminates a manipulation step from the method. The specificity of the primers was determined using 39 bacterial strains consisting of STEC, enteropathogenic E. coli, E. coli K12, and Salmonella, which were previously examined for stx genes [23]. STEC strains with different combinations of stx genotypes were selected (Table 1). All of the STEC were amplified, and none of the non-STEC strains were positive according to standard gel electrophoresis. In the ELISA detection, a clear-cut positive colorimetric signal (optical density [OD] of ≥ 0.9; A405 = 1.138 ± 0.24; n = 33) was obtained with the PCR products from the STEC strains, whereas the OD of non-STEC was 0.188 ± 0.07. Fig. 2 shows a gel of PCR amplicons from 19 bacterial strains and their corresponding ELISA readings. To further evaluate the performance of the PCR–ELISA method, the sensitivity of this assay was compared with standard agarose gel electrophoresis for detecting PCR products amplified from both pure bacterial cultures, and ground beef samples artificially contaminated with STEC. Table 2 compares the respective detection limits of gel electrophoresis and ELISA for detecting PCR products of pure bacterial cultures. The sensitivity of ELISA was at least tenfold higher than the gel electrophoresis, with up to 100-fold increase observed. Less than 1 CFU of umd11 gave a positive result, which could be explained by the fact that the PCR was designed to detect both stx1 and stx2, and that theoretically it would be more sensitive in detecting STEC strains that contain both genes. Fig. 3 compares the sensitivity between gel electrophoresis and ELISA. Noticeably, there were several weak or absent bands in the gel, whose ELISA readings were considerably higher than that of the negative control. However, this PCR–ELISA was not a quantitative assay, since there were only 10 µM DIG-11'dUTP but 190 µM dTTP in each reaction. The incorporation of DIG-11'-dUTP into amplicons was random rather than proportional, explaining why the ELISA readings in the sensitivity test did not correlate well with the cell concentrations. Replacing all the dTTP with DIG-11'-dUTP would be expected to render this assay quantitative, but the cost of DIG-11'-dUTP is significant. Furthermore, the presence of

289

Table 2 Comparison of detection limits of gel electrophoresis and ELISA for detecting PCR products using STEC cultures Strain ID

Serotype

Stx genotype

Detection limit (CFU/reaction) Gel electrophoresis

umd141 umd200 umd19 umd11 umd168 .

O26:H11 O157:H7 O157:H7 O157:H7 O111:NM

1 1 2 1,2 1,2

2

10 102 102 10 10

ELISA 10 10 1.0 0.1 1.0

too many DIG-11'-dUTP on amplicons may interfere with the amplification process as a result of the change of configuration. In this study, we found that 5 µl of PCR product diluted in 445 µl buffer of ELISA were sufficient to obtain a sensitive result, and that increasing the volume of PCR products actually lowered the sensitivity of the assay (data not shown). Similar observations have also been reported in other studies [12]. One explanation is the competitive binding of the biotin-labeled primer to the streptavidincoated wells, which may decrease the signal considerably. Other possible background signals include the non-specific binding of biotin-labeled single-strand DNA or the enzyme conjugate antibody to the microtiter plate. With the aid of PrepMan, the PCR–ELISA assay was able to detect 105 cells of STEC per gram in ground beef samples without any enrichment procedure. Obviously, the assay alone is not useful to identify food samples that are potentially contaminated with STEC. Most studies that apply PCR assays for detecting pathogens in food have included an enrichment procedure. Some studies also applied immunomagnetic beads to selectively concentrate target organisms in the enrichment culture [24]. The present study indicates that even using recent separation technology

Fig. 3. Sensitivity of gel electrophoresis (A) and ELISA (B) for detecting the STEC PCR products. Strains umd11 and umd168 are STEC serotypes O157:H7 and O111:NM, respectively. The dash line in B indicates the cutoff value which is two times greater than that of the negative control (E. coli K-12).

290

B. Ge et al. / Microbes and Infection 4 (2002) 285–290

direct detection of STEC without enrichment does not offer a useful sensitivity, and that enrichment remains necessary. Although not all STEC are proven human pathogens, and the presence of STEC may not establish a threat to healthy populations, identification of STEC in the food supply should be considered hazardous to susceptible populations including the very young and the elderly. In summary, the PCR–ELISA assay was highly specific, and using the ELISA detection system instead of conventional gel electrophoresis significantly increased the sensitivity of the assay. Automation of the PCR and ELISA procedures with robotic equipment will enable use of the assay for rapid and large-scale screening of STEC.

Acknowledgements

[10] P. Feng, Identification of Escherichia coli serotype O157:H7 by DNA probe specific for an allele of uid A gene, Mol. Cell. Probes 7 (1993) 151–154. [11] J. Meng, S. Zhao, M. Doyle, S. Mitchell, S. Kresovich, Polymerase chain reaction for detecting Escherichia coli O157:H7, Int. J. Food Microbiol. 32 (1996) 103–114. [12] J.M. Luk, U. Kongmuang, R.S. Tsang, A.A. Lindberg, An enzymelinked immunosorbent assay to detect PCR products of the rfbS gene from serogroup D salmonellae: a rapid screening prototype, J. Clin. Microbiol. 35 (1997) 714–718. [13] A.R. Melton-Celsa, A.D. O’Brien, Structure, biology, and relative toxicity of Shiga toxin family members for cells and animals, in: J. Kaper, A. O’Brien (Eds.), Escherichia coli O157:H7 and other Shiga Toxin-producing E. coli Strains, ASM Press, Washington, DC, 1998, pp. 121–128. [14] K. Ramotar, B. Waldhart, D. Church, R. Szumski, T.J. Louie, Direct detection of verotoxin-producing Escherichia coli in stool samples by PCR, J. Clin. Microbiol. 33 (1995) 519–524.

This study was supported in part by USDA-NRI grant 9735201 and by Odwalla Food Safety Funds.

[15] M. Uyttendaele, S. van Boxstael, J. Debevere, PCR assay for detection of the E. coli O157:H7 eae-gene and effect of the sample preparation method on PCR detection of heat-killed E. coli O157:H7 in ground beef, Int. J. Food Microbiol. 52 (1999) 85–95.

References

[16] R. Lindqvist, Preparation of PCR samples from food by a rapid and simple centrifugation technique evaluated by detection of Escherichia coli O157:H7, Int. J. Food Microbiol. 37 (1997) 73–82.

[1] L.W. Riley, R.S. Remis, S.D. Helgerson, H.B. McGee, J.G. Wells, B.R. Davis, R.J. Hebert, E.S. Olcott, L.M. Johnson, N.T. Hargrett, P.A. Blake, M.L. Cohen, Hemorrhagic colitis associated with a rare Escherichia coli serotype, N. Engl. J. Med. 308 (1983) 681–685. [2] P.M. Griffin, Escherichia coli O157:H7 and other enterohemorrhagic Escherichia coli, in: M.J. Blaser, P.D. Smith, J.I. Ravdin, H.B. Greenberg, R.L. Guerrant (Eds.), Infections of Gastrointestinal Tract, Raven Press, New York, 1995, pp. 739–761. [3] World Health Organization, Zoonotic non-O157 Shiga toxinproducing Escherichia coli (STEC), 1998. [4] J. Meng, M.P. Doyle, Microbiology of Shiga toxin-producing Escherichia coli in food, in: J. Kaper, A. O’Brien (Eds.), Escherichia coli O157:H7 and other Shiga Toxin-producing E. coli Strains, ASM Press, Washington, DC, 1998, pp. 92–108. [5] P.S. Mead, L. Slutsker, V. Dietz, L.F. McCaig, J.S. Bresee, C. Shapiro, P.M. Griffin, R.V. Tauxe, Food-related illness and death in the United States, Emerg. Infect. Dis. 5 (1999) 607–625. [6] Centers for Disease Control and Prevention, Surveillance for outbreak of Escherichia coli O157:H7 infection, summary of 1982-1998 data, Personal communication, Atlanta, 1999. [7] M.J. Brian, M. Frosolono, B.E. Murray, A. Miranda, E.L. Lopez, H.F. Gomez, T.G. Cleary, Polymerase chain reaction for diagnosis of enterohemorrhagic Escherichia coli infection and hemolytic-uremic syndrome, J. Clin. Microbiol. 30 (1992) 1801–1806. [8] V.P. Gannon, M. Rashed, R.K. King, E.J. Thomas, Detection and characterization of the eae gene of Shiga-like toxin-producing Escherichia coli using polymerase chain reaction, J. Clin. Microbiol. 31 (1993) 1268–1274. [9] R.H. Hall, J.G. Xu, Rapid and sensitive detection of O157:H7 and other enterohemorrhagic E. coli, United States Patent, May 26, 1998, Patent number 5,756,293 (1998).

[17] M. Ritzler, M. Altwegg, Sensitivity and specificity of a commercially available enzyme-linked immunoassay for the detection of polymerase chain reaction amplified DNA, J. Microbiol. Methods 27 (1996) 233–238. [18] R. Gutierrez, T. Garcia, I. Gonzalez, B. Sanz, P.E. Hernandez, R. Martin, A quantitative PCR-ELISA for the rapid enumeration of bacteria in refrigerated raw milk, J. Appl. Microbiol. 83 (1997) 518–523. [19] J.C. Knight, W. McGuire, M.M. Kortok, D. Kwiatkowski, Accuracy of genotyping of single-nucleotide polymorphisms by PCR-ELISA allele-specific oligonucleotide hybridization typing and by amplification refractory mutation system, Clin. Chem. 45 (1999) 1860–1863. [20] P. Scheu, A. Gasch, K. Berghof, Rapid detection of Listeria monocytogenes by PCR-ELISA, Lett. Appl. Microbiol. 29 (1999) 416–420. [21] L. Cocolin, G. Astori, M. Manzano, C. Cantoni, G. Comi, Development and evaluation of a PCR-microplate capture hybridization method for direct detection of verotoxigenic Escherichia coli strains in artificially contaminated food samples, Int. J. Food Microbiol. 54 (2000) 1–8. [22] I. Gonzalez, T. Garcia, A. Fernandez, B. Sanz, P.E. Hernandez, R. Martin, Rapid enumeration of Escherichia coli in oysters by a quantitative PCR-ELISA, J. Appl. Microbiol. 86 (1999) 231–236. [23] J. Meng, S. Zhao, M.P. Doyle, Virulence genes of Shiga toxinproducing Escherichia coli isolated from food, animals and humans, Int. J. Food Microbiol. 45 (1998) 229–235. [24] P.M. Fratamico, L.K. Bagi, T. Pepe, A multiplex polymerase chain reaction assay for rapid detection and identification of Escherichia coli O157:H7 in foods and bovine feces, J. Food Prot. 63 (2000) 1032–1037.