International Journal of Food Microbiology. 13 (1991) 31-40 ~ 1991 Elsevier Science Publishers B.V. 0168-1605/91/$03.50
31
FOOD 00395
D N A hybridization and latex agglutination for detection of heat-labile- and shiga-like toxin-producing Escherichia coli in meat S. Notermans i K. Wernars 1, P.S. Soentoro 1, j. Dufrenne and W. Jansen 2 t Laboratory for Water and Food Microbiology and : Laborator)."for Bacteriology, National Institute of Public Health and Environmental Protection, Bihhooen. The Netherlands (Received 20 August 1990; accepted 31 December 1990)
DNA-hybridization and latex-agglutination tests were used for screening of a group of Escherichia coli isolates for heat-labile enterotoxin (LT)- and ship-like toxin (SLTI or VT1) -producing strains, respectively. Strains tested originated from 162 meat samples (poultry, pigs and beef) chosen at random. Additionally LT- and SLTl-producing reference strains were tested. The DNA-hybridization technique allowed screening of large numbers of strains, whereas large scale testing of strains by latex agglutination was laborious. Of 800 E. coil strains tested by DNA hybridization none contained the gene encoding LT. Production of LT as tested by latex agglutination was not found. The gene encoding SLTI was detected m 10 of the 800 isolates tested. None of these swains, however, showed cytotoxicity on Veto cells. Serotyping was done with sorbitol-negative E. coli strains, first by using the latex-agglutination test for 0157 followed by complete serotyping. No E. coli of serogroup 0157 were found. Therefore the results obtained also indicate that routine screening of £. coli isolated randomly from food for toxin production is not useful and should be limited to food-borne disease outbreaks with an etiology resembling an E. coli infection. Key words: DNA hybridization: Latex agglutination: Escherichia coli; Heat-labile enterotoxin. Ship-like toxin
Introduction Escherichia coli is a Gram-negative, facultative anaerobic microorganism and is part of the normal flora of the human and animal intestine. The presence of E. coil in food is usually caused by faecal contamination and their numbers in a food product are a measure for the success of 'good manufacturing practices' during processing. Although the majority of strains within this species is non-pathogenic,
Corregpomdence address: S. Notermans, Laboratory. for Water and Food Microbiology, National Institute of Public Health and Environmental Protection. P.O. Box 1, 3720 BA Bilthoven, The Netherlands.
32 some E. coil strains are able to cause infections in humans and the majority of these strains cause diarrhoea upon ingestion of contaminated food and water. These diarrhoeagenic E. coil have been divided into different classes: enterotoxinogenic (ETEC), enterohemorrhagic (EHEC), enteroinvasive (EIEC) and enteropathogenic (EPEC). They display distinct clinical symptoms upon infection and differ in their pathogenesis and sero groups. The epidemiology and pathogenesis of these diarrhoeagenic E. coil categories has recently been reviewed by Levine (1987). ETEC strains are an important cause of infant diarrhoea in developing countries, particularly in the tropics, and in persons travelling to these countries (Gross and Rowe, 1985: Levine, 1987). The strains produce either a heat-stable enterotoxin (ST), a heat-labile enterotoxin (LT), or both. Until recently, the most commonly used test for detecting LT-producing E. coil was testing of culture filtrate for the presence of the toxin by tissue culture tests (Linggood, 1982). Now immunassays such as the latex-agglutination assay are commercially available for detection of LT-producing strains (Scotland et al., 1988; Bettelheim et al., 1989). As well as using the latex-agglutination assay, LT-positive E. coil can be detected by D N A hybridi32 zation. Almost a decade ago Moseley et al. (1980) demonstrated that a P-labelled fragment from the E. coli LT gene could be used as a probe to identify ETEC colonies. Since then many studies have been published describing the application of DNA hybridization for the detection of ETEC strains (Seriwatana et al., 1988: Hill, 1981; Lanata et al., 1985; Medon et al., 1988). EHEC strains provide a significant risk of life-threathening haemorrhagic colitis (Riley et al.. 1983) and are a public health problem of great concern in developed countries. E. coli strains implicated in cases of haemorrhagic colitis have been shown to produce a cytotoxin (verotoxin) with biological properties similar to shiga toxin from Shigella dysenteriae (Konowalchuk et al., 1977: O'Brien and La Veck, 1983). As with shiga toxin, there is no direct proof that E. coil verotoxins play a role in the disease. Their putative clinical relevance is suggested by epidemiologicai studies, indicating a close association between infection by verotoxin-producing E. coli and diarrhoeal disease (Karmali, 1989). The E. coli serotype most frequently isolated from haemorrhagic colitis is 0157:H7. Isolation of this serotype from diarrhoeal stool especially with blood is indicative of a shiga-like toxin (SLT1)-producing strain (Karmali. 1989), although no absolute link between the serotype and SLTI production has been found. For identification of putative EHEC strains serotyping can be carried out. Since E. coh 0 1 5 7 : H 7 strains do not ferment sorbitol, only non-sorbitol-fermenting colonies need to be tested. For detection of E. coil 0157 serotypes a commercial latex agglutination test is available. It should also be possible to detect putative EHEC strains by DNA hybridization using a DNA probe encoding SLT1. Since simple latex immunoassays for detection of ETEC and putative E H E C strains are n o b commercially available these assays can be used for routine purposes. In this study the value of such routine tests is evaluated. For this, E. coli, isolated from randomly selected raw meat products, were tested for the presence of genes encoding LT and SLT1. In addition some of the strains were tested with the commercial available latex agglutination assays.
33 Materials and Methods !
.
Meat samples E. coil was isolated from raw meat products. In total 64 samples of poultry meat. 47 samples of ground pork and 51 samples of ground beef were tested. The meat samples were purchased from local butchers. Detection of E. coil The number of colony-forming units per gram (CFU/g) was determined according to ISO/DIS 6391. However, instead of the mineral modified glutamate agar, tryptose soya agar (oxoid CM 131) was used as repair medium. Samples of 25 g were added to 225 ml buffered peptone water, pH 7.0 (BPW). The mixture was blended in a stomacher for 1 rain. Portions of 0.1 ml undiluted fluid and of decimal dilutions were plated on tryptose soya agar overlaid by a membrane (millipore 30, pore size 0.45 pro) and incubated at 37°C. After 3 h the membranes were transferred onto tryptose bile agar (Oxoid CM595) and incubated at 44°C for 18 h. All colonies present on the membranes were tested for indole formation, lndole-producing strains were regarded as E. coll. For sorbitol fermentation colonies were transferred onto sorbitol MacConkey agar (oxoid CM 813) and incubated at 37 °C for 18 h. Non-sorbitol-fermenting strains are colourless while sorbitol-fermenting strains have a yellow appearance. Enrichment of samples was carried out if less than 100 CFU of E. coil were present per gram of product. For this, 1 g of product was transferred into tubes containing 9 ml BPW and incubated at 37°C for 18 h. The enrichment fluid was tested for the presence of E. coil by streaking dilutions of culture fluid onto tryptose soya agar overlaid by a membrane. Incubation and other test procedures were carried out as described above. Latex immunoassay for LT-producing E. coil For detection of LT-producing E. coil the commercially available latex agglutination kit from Oxoid, Diagnostic Reagents (TD920) was applied. Following the instructions of the manufacturer, E. coil strains were incubated in 10 ml Mundelrs medium (Mundell et al., 1976). The cultures were incubated at 37°C with shaking (110 rpm). After 20 h polymixin B was added to a concentration of 10000 units/ml, followed by an incubation at 34°C. After 4 h the culture was centrifuged at 900 g for 20 rain at 4°C. The supernatant was used as test sample. The latex test procedure was used following the manufacturer's instructions. Latex immunoassay for identification of E. coli sero group 0157 For detection of putative SLTl-producing E. coil strains non-sorbitol-fermenting strains were tested by the E. coli 0157 latex test (Oxoid, Diagnostic Reagents, DR 620). Non-sorbitol-fermenting colomes were taken from the sorbitol MacConkey agar plates and tested following the recommendations of the manufacturer.
34 Screening L T-posltwe and SLTl-positive E. coil by DNA hybridization E. coh strains were tested for the presence of genes encoding LT and SET1 b~ colony hybridization. The LT probe consisted of a 0.85 kb HindIII insert from recombinant plasmid pWD299 encoding the B and part of the A LT-subunit (Dallas et al., 1979). The probe for SET1 was an 1.35 kb H i n d I I I - E c o R V insert from recombinant plasmid pHS7040 encoding the SLT1 A subunit (Bohnert et al., 1988). Probes were labelled with 32p using a random primed DNA labelling kit (Boehringer Mannheim) according to the manufacturer's instructions. Preparation of the membranes for colony hybridization and the hybridization procedure have been described previously (Notermans et al., 1988). After hybridization the membranes were washed in a buffer containing 0.2 x saline sodium citrate buffer (SSC: 1 x SSC contains 0.15 M NaC1 and 0.015 M sodium citrate), 0.1% sodium pyrophosphate and 1% sodium dodecylsulphate for 2 x 15 min at 56°C. Positive colonies were detected by autoradiography. Total DNA from E. coil was isolated according to the method of Boom et al. (1990). The D N A was digested to completion with restriction endonucleases and fractionated on a 0.8% agarose gel. The gel was then blotted onto Gene-Screen Plus hybridization membrane (Dupont, NEN Research Products) and used for hybridization. Serotyping of E. coli All non-sorbitol-fermenting strains as well as all strains with a positive DNA-hybridization signal for the SLT1 gene were serotyped as described by Guin~e et al. (1979). Vero cell assay Vero cell assays were carried out as described by Guin~e et al. (1981) with one minor modification: strains were grown in trypton soya broth (Oxoid CM 129) instead of Evan's synthetic medium. Reference strains The following reference strains were used: E. coli (SLTl-positive) strain H868/88, H199/88, H120/88, H122/88 and H 2 4 3 / 8 8 all of serotype 0157, k-negative and strain H19WS of serotype 026, K60. E. coli (LT-positive) strain H 6 7 6 / 8 7 serotype O159, H662/87 serotype 06, H 9 0 / 8 8 and H 9 2 / 8 8 both of serotype O6, K15, H104/87 serotype 078, K80 and strain Bile, serotype 06, K15.
Results
Frequency of distribution of E. coli present on meat samples The frequency distribution of the presence of E. coli on different meat samples is presented in Fig. 1. E. coli was present in most samples tested; however, the numbers of colonies were small. Only in 9% of the samples tested were more than 10 colonies found on the plates inoculated with 0.1 ml of the blended samples. Non-sorbitol-fermenting E. coli were found to be present in 13% of the samples
35 % 6050-
I I poultry n=6t, k-x\\\-~ pork n=t.7
/.0"
t:.:;:;:::;:::::!
13~@f
/3 = '~ 1
30' 20' 10" : !
31
FOOD 00395
D N A hybridization and latex agglutination for detection of heat-labile- and shiga-like toxin-producing Escherichia coli in meat S. Notermans i K. Wernars 1, P.S. Soentoro 1, j. Dufrenne and W. Jansen 2 t Laboratory for Water and Food Microbiology and : Laborator)."for Bacteriology, National Institute of Public Health and Environmental Protection, Bihhooen. The Netherlands (Received 20 August 1990; accepted 31 December 1990)
DNA-hybridization and latex-agglutination tests were used for screening of a group of Escherichia coli isolates for heat-labile enterotoxin (LT)- and ship-like toxin (SLTI or VT1) -producing strains, respectively. Strains tested originated from 162 meat samples (poultry, pigs and beef) chosen at random. Additionally LT- and SLTl-producing reference strains were tested. The DNA-hybridization technique allowed screening of large numbers of strains, whereas large scale testing of strains by latex agglutination was laborious. Of 800 E. coil strains tested by DNA hybridization none contained the gene encoding LT. Production of LT as tested by latex agglutination was not found. The gene encoding SLTI was detected m 10 of the 800 isolates tested. None of these swains, however, showed cytotoxicity on Veto cells. Serotyping was done with sorbitol-negative E. coli strains, first by using the latex-agglutination test for 0157 followed by complete serotyping. No E. coli of serogroup 0157 were found. Therefore the results obtained also indicate that routine screening of £. coli isolated randomly from food for toxin production is not useful and should be limited to food-borne disease outbreaks with an etiology resembling an E. coli infection. Key words: DNA hybridization: Latex agglutination: Escherichia coli; Heat-labile enterotoxin. Ship-like toxin
Introduction Escherichia coli is a Gram-negative, facultative anaerobic microorganism and is part of the normal flora of the human and animal intestine. The presence of E. coil in food is usually caused by faecal contamination and their numbers in a food product are a measure for the success of 'good manufacturing practices' during processing. Although the majority of strains within this species is non-pathogenic,
Corregpomdence address: S. Notermans, Laboratory. for Water and Food Microbiology, National Institute of Public Health and Environmental Protection. P.O. Box 1, 3720 BA Bilthoven, The Netherlands.
32 some E. coil strains are able to cause infections in humans and the majority of these strains cause diarrhoea upon ingestion of contaminated food and water. These diarrhoeagenic E. coil have been divided into different classes: enterotoxinogenic (ETEC), enterohemorrhagic (EHEC), enteroinvasive (EIEC) and enteropathogenic (EPEC). They display distinct clinical symptoms upon infection and differ in their pathogenesis and sero groups. The epidemiology and pathogenesis of these diarrhoeagenic E. coil categories has recently been reviewed by Levine (1987). ETEC strains are an important cause of infant diarrhoea in developing countries, particularly in the tropics, and in persons travelling to these countries (Gross and Rowe, 1985: Levine, 1987). The strains produce either a heat-stable enterotoxin (ST), a heat-labile enterotoxin (LT), or both. Until recently, the most commonly used test for detecting LT-producing E. coil was testing of culture filtrate for the presence of the toxin by tissue culture tests (Linggood, 1982). Now immunassays such as the latex-agglutination assay are commercially available for detection of LT-producing strains (Scotland et al., 1988; Bettelheim et al., 1989). As well as using the latex-agglutination assay, LT-positive E. coil can be detected by D N A hybridi32 zation. Almost a decade ago Moseley et al. (1980) demonstrated that a P-labelled fragment from the E. coli LT gene could be used as a probe to identify ETEC colonies. Since then many studies have been published describing the application of DNA hybridization for the detection of ETEC strains (Seriwatana et al., 1988: Hill, 1981; Lanata et al., 1985; Medon et al., 1988). EHEC strains provide a significant risk of life-threathening haemorrhagic colitis (Riley et al.. 1983) and are a public health problem of great concern in developed countries. E. coli strains implicated in cases of haemorrhagic colitis have been shown to produce a cytotoxin (verotoxin) with biological properties similar to shiga toxin from Shigella dysenteriae (Konowalchuk et al., 1977: O'Brien and La Veck, 1983). As with shiga toxin, there is no direct proof that E. coil verotoxins play a role in the disease. Their putative clinical relevance is suggested by epidemiologicai studies, indicating a close association between infection by verotoxin-producing E. coli and diarrhoeal disease (Karmali, 1989). The E. coli serotype most frequently isolated from haemorrhagic colitis is 0157:H7. Isolation of this serotype from diarrhoeal stool especially with blood is indicative of a shiga-like toxin (SLT1)-producing strain (Karmali. 1989), although no absolute link between the serotype and SLTI production has been found. For identification of putative EHEC strains serotyping can be carried out. Since E. coh 0 1 5 7 : H 7 strains do not ferment sorbitol, only non-sorbitol-fermenting colonies need to be tested. For detection of E. coil 0157 serotypes a commercial latex agglutination test is available. It should also be possible to detect putative EHEC strains by DNA hybridization using a DNA probe encoding SLT1. Since simple latex immunoassays for detection of ETEC and putative E H E C strains are n o b commercially available these assays can be used for routine purposes. In this study the value of such routine tests is evaluated. For this, E. coli, isolated from randomly selected raw meat products, were tested for the presence of genes encoding LT and SLT1. In addition some of the strains were tested with the commercial available latex agglutination assays.
33 Materials and Methods !
.
Meat samples E. coil was isolated from raw meat products. In total 64 samples of poultry meat. 47 samples of ground pork and 51 samples of ground beef were tested. The meat samples were purchased from local butchers. Detection of E. coil The number of colony-forming units per gram (CFU/g) was determined according to ISO/DIS 6391. However, instead of the mineral modified glutamate agar, tryptose soya agar (oxoid CM 131) was used as repair medium. Samples of 25 g were added to 225 ml buffered peptone water, pH 7.0 (BPW). The mixture was blended in a stomacher for 1 rain. Portions of 0.1 ml undiluted fluid and of decimal dilutions were plated on tryptose soya agar overlaid by a membrane (millipore 30, pore size 0.45 pro) and incubated at 37°C. After 3 h the membranes were transferred onto tryptose bile agar (Oxoid CM595) and incubated at 44°C for 18 h. All colonies present on the membranes were tested for indole formation, lndole-producing strains were regarded as E. coll. For sorbitol fermentation colonies were transferred onto sorbitol MacConkey agar (oxoid CM 813) and incubated at 37 °C for 18 h. Non-sorbitol-fermenting strains are colourless while sorbitol-fermenting strains have a yellow appearance. Enrichment of samples was carried out if less than 100 CFU of E. coil were present per gram of product. For this, 1 g of product was transferred into tubes containing 9 ml BPW and incubated at 37°C for 18 h. The enrichment fluid was tested for the presence of E. coil by streaking dilutions of culture fluid onto tryptose soya agar overlaid by a membrane. Incubation and other test procedures were carried out as described above. Latex immunoassay for LT-producing E. coil For detection of LT-producing E. coil the commercially available latex agglutination kit from Oxoid, Diagnostic Reagents (TD920) was applied. Following the instructions of the manufacturer, E. coil strains were incubated in 10 ml Mundelrs medium (Mundell et al., 1976). The cultures were incubated at 37°C with shaking (110 rpm). After 20 h polymixin B was added to a concentration of 10000 units/ml, followed by an incubation at 34°C. After 4 h the culture was centrifuged at 900 g for 20 rain at 4°C. The supernatant was used as test sample. The latex test procedure was used following the manufacturer's instructions. Latex immunoassay for identification of E. coli sero group 0157 For detection of putative SLTl-producing E. coil strains non-sorbitol-fermenting strains were tested by the E. coli 0157 latex test (Oxoid, Diagnostic Reagents, DR 620). Non-sorbitol-fermenting colomes were taken from the sorbitol MacConkey agar plates and tested following the recommendations of the manufacturer.
34 Screening L T-posltwe and SLTl-positive E. coil by DNA hybridization E. coh strains were tested for the presence of genes encoding LT and SET1 b~ colony hybridization. The LT probe consisted of a 0.85 kb HindIII insert from recombinant plasmid pWD299 encoding the B and part of the A LT-subunit (Dallas et al., 1979). The probe for SET1 was an 1.35 kb H i n d I I I - E c o R V insert from recombinant plasmid pHS7040 encoding the SLT1 A subunit (Bohnert et al., 1988). Probes were labelled with 32p using a random primed DNA labelling kit (Boehringer Mannheim) according to the manufacturer's instructions. Preparation of the membranes for colony hybridization and the hybridization procedure have been described previously (Notermans et al., 1988). After hybridization the membranes were washed in a buffer containing 0.2 x saline sodium citrate buffer (SSC: 1 x SSC contains 0.15 M NaC1 and 0.015 M sodium citrate), 0.1% sodium pyrophosphate and 1% sodium dodecylsulphate for 2 x 15 min at 56°C. Positive colonies were detected by autoradiography. Total DNA from E. coil was isolated according to the method of Boom et al. (1990). The D N A was digested to completion with restriction endonucleases and fractionated on a 0.8% agarose gel. The gel was then blotted onto Gene-Screen Plus hybridization membrane (Dupont, NEN Research Products) and used for hybridization. Serotyping of E. coli All non-sorbitol-fermenting strains as well as all strains with a positive DNA-hybridization signal for the SLT1 gene were serotyped as described by Guin~e et al. (1979). Vero cell assay Vero cell assays were carried out as described by Guin~e et al. (1981) with one minor modification: strains were grown in trypton soya broth (Oxoid CM 129) instead of Evan's synthetic medium. Reference strains The following reference strains were used: E. coli (SLTl-positive) strain H868/88, H199/88, H120/88, H122/88 and H 2 4 3 / 8 8 all of serotype 0157, k-negative and strain H19WS of serotype 026, K60. E. coli (LT-positive) strain H 6 7 6 / 8 7 serotype O159, H662/87 serotype 06, H 9 0 / 8 8 and H 9 2 / 8 8 both of serotype O6, K15, H104/87 serotype 078, K80 and strain Bile, serotype 06, K15.
Results
Frequency of distribution of E. coli present on meat samples The frequency distribution of the presence of E. coli on different meat samples is presented in Fig. 1. E. coli was present in most samples tested; however, the numbers of colonies were small. Only in 9% of the samples tested were more than 10 colonies found on the plates inoculated with 0.1 ml of the blended samples. Non-sorbitol-fermenting E. coli were found to be present in 13% of the samples
35 % 6050-
I I poultry n=6t, k-x\\\-~ pork n=t.7
/.0"
t:.:;:;:::;:::::!
13~@f
/3 = '~ 1
30' 20' 10" : !
