- Email: [email protected]
PCR-based detection of verotoxin-producing Escherichia coli (VTEC) in ground beef
(O I NSTITUT PASH!UI¢/I~I.SI-!VIt-R Paris 1998
Res. Microbiol. 1998, 149, 145-154
PCR-based detection of verotoxin-producing Escherichia coli (VTEC) in ground beef M. Gilgen .,I E Htibner '*). C. H6felein, J. Li.ithy and U. Candrian ~'' Laboratory ~tl"[~od Clwmistry. Dep, rmu'nt of Chemistry am/Biochemisto', University o.['Berne. Freiestrasse 3, CH-3012 Berne
SUMMARY
Pathogenic strains of Escherichia coil producing verotoxins (VTs) have been recognized as a cause of human disease, and rapid and sensitive detection tests are urgently needed to ensure the safety of food, especially ground beef. We applied two nested polymerase chain reaction (PCR) assays to detect the genes encoding VT1 and VT2 irrespective of the bacterial serotype, in combination with a direct sample preparation protocol, we were able to uncover the presence of about 110 CFU of verotoxinogenic E. coil (V'rEC) in 10 g of ground beef. When a six-hour enrichment was included, we found the detection limit to be in the range of 1 to 10 bacterial cells per 10 g of ground beef. To evaluate our detection system, we tested 30 ground beef samples originating from butcher shops in Berne, Switzerland. One sample yielded positive PCR results for both the VT1 and VT2 genes, indicating the presence of verotoxinogenic E. coil Finally, 20 food homogenates, shown to contain E. coil strains by standard culture, were analysed with our method, and the gene encoding VT2 was detected in one cheese sample. The results suggest that the described PCR method can serve as a valuable tool for the surveillance of VTEC contamination of foods.
Key-words: Food, Escherichia coli; O157:H7, VTEC, PCR, Ground beef, Detection.
INTRODUCTION Four classes of enterovirulent Escherichia coil have been recognized to cause gastroenteritis in humans" enterotoxigenic (ETEC), enteroinvasive (EIEC), enteropathogenic (EPEC) and enterohaemorrhagic (EHEC) strains. EHECs have as a common feature the production of protein toxins, cytolethal to Vero cells in culture (Konwalchuk et aL, 1977) and therefore referred to as verotoxins
(VTs), and the VT-producing E. coil strains verocytotoxin-producing E. coli (VTEC). VTEC strains produce two irnmunologically distinct VTs (VTI and VT2) either exclusively or in combination, both of which are encoded by bacteriophages (Scotland et aL, 1983; Smith et al., 1983). Since VTs ale closely related to Shiga toxin, they are often referred to as Shiga-like toxins (SLT-I and SLT-II) (Strockbine et al., 1986). VTEC strains ate implicated in the pathogenesis of the
Submitted January 3, 1997, accepted October I0, 1997. (*) Corresponding attthor. " Present address: Central Laboratory BTS SRC. PCR Diagnostics. Wankdorl'stra.,,se 10, P.O. Box, CH-3000 Berne 22, Switzerland. Cb~Presenl address: Swiss Federal Office of Public Health, Division of Biologicals. Section for Vaccines, CH-3003 Berne. Switzerland.
M. G I L G E N E T A L .
146
diarrhoea-associated form of haemolytic-uraemic syndrome (HUS), which is the most common cause of acute renal failure in childhood (Kaplan et aL. 1990). The prototype of the VT-producing E. coli iS the enterohaemorrhagic strain O157:H7, associated with a spectrum of diseases including bloody diarrhoea, haemorrhagic colitis and thrombotic thrombocytopenic purpura (Karmali, 1989). The infective dose of this pathogen is probably very low (Willshaw et al., 1994). E. coil OI57:H7 and the diseases it causes have only recently become prominent, first being identified as a cause of haemorrhagic colitis in Canada in the late 1970's (Johnson et al., 1983; Riley et al., 1983). This organism is not only a prototype for the EHEC group, but also tbr new and emerging bacterial pathogens in general (Hughes and La Montagne, 1994). Outbreaks and isolated cases of disease with E. coil O157:H7 have been associated with the
consumption of raw or undercooked hamburger patties (Bell et al., 1994; Wells et al., 1983; Willshaw et al., 1994), milk (Borczyk et aL, 1987; Wright et ai., 1994) and apple cider (Besser et ai., 1993). Waterborne (Swerdlow et aL, 1992) and person-to-person (Reida et al., 1994) transmission are additional modes of transmission for this pathogen. In addition to O157:H7 VTEC, at least 50 serotypes have been recognized. Although no outbreaks of bloody stools caused by non-O157 strains have been reported in the United States, Canada and the Un;,ted Kingdom, several investigators have isolated such VT-producing strains from patients with diarrhoea, HUS, from cattle herds and from foods of animal origin (Blanco et aL, 1994; Griffin and Tauxe, 1991 ; L0scher et aL, 1992; Ricotti et aL, 1994; Zhao et aL, 1995). Mo~t detection met~-ods for verotoxinogenic E. coil ~re based on the cultivation of serotype O157. Because E. coil O157 does not ferment sorbitol rapidly, sorbitol MacConkey agar can be
BSA CFU EHEC HUS LB mEC+n
= = = = = =
bovine serum albumin. colony-forming unit. enterohaemorrhagic E. coll. haemolytic uraemic syndrome. Luria-Bertani (medium~. modified E. coli broth+novobiocine.
used to rapidly screen clinical or environmental isolates (Kleanthous et ai., 1988; March and Ratham, 1986). Improved media additionally contain cefixime (Chapman et al., 1991) and tellurite (Zadik et al., 1993). Identification can be done by direct agglutination of the colourless colonies with commercially av~i~able latex-conjugated antisera. This method and other diagnostic tests like ELISA, immunomagnetic separation and "Petrifilm" test kit HEC are limited to the isolation and detection of serotype O157. However, the detection of other serotypes of VT-producing E. coil (which do ferment sorbitol and do not agglutinate O157 antisera) is possible by use of DNA probes to the VT genes (Newland and Neill, 1988; Wiilshaw et al., 1987) or VT genespecific PCR (Begum and Jackson, 1995; Gannon et al., 1992; Johnson et al., 1990; Karch and Meyer, 1989; Olsvik et al., 1991 ; Paton et al., 1993; Pollard et aL, 1990; Witham et al., 1996). PCR not only allows detection of VTECs irrespective of their serotype, but also of different variants of VT genes. Since the virulence potential of non-O157 VTEC strains is not yet known, it is important to have detection methods recognizing all E. coil strains carrying VT genes. Based on the PCR assays described by Gannon et al. (1992), we developed nested PCR systems for the detection of the VTI and VT2 genes of VTEC. Additionally, we established an isolation method that enables the detection of low numbers of VTEC in ground beef.
MATERIALS
AND METHODS
Bacterial strains
Several enteric bacterial strains including VTproducing E. coil of various serotypes were used for testing the specificity of the developed PCR
PCR SDS SLT-! SLT-II VT VTEC
= = = = = =
p o l y m e r a s e chain reaction. stadium dodecyl sulphate. Shiga-like toxin I. Shiga-like toxin II. verotoxin. verotoxinogenic E. coli.
"vn ~t_ LPL T E C T i O N i N G R O U N D B E E k
systems and for the evaluation of isolation methods. All used VTEC strains had been involved in cases of severe diarrhoea. Strains containing the V T I gene w e r e : O 2 6 : H 1 1 ( E 3 7 8 7 , B. R o w e , PHLS, Colindale, London, England), O60:H(MM 730, M. Altwegg, Department o f Medical Microbiology, University of Ziirich, Switzerland) and O! 17:H- (MM 731, Altwegg). The follo',;ving V T E C s h a r b o u r e d the V T 2 g e n e : O I 5 7 : H (E32511, B. R o w e l O79:H not typable (MM732, Altwegg) and O30:H21 (MM734, Altwegg). Both genes were present in two strains: O157:H7 (CDC Atlanta 933. A l t w e g g ) and O 7 4 : H 3 4 ( M M 7 3 6 , Altwegg). Other bacterial strains used were E. coli L M C 30 ( h e a t - l a b i l e e n t e r o t o x i n [LTI], heatstable enterotoxin [STILL LMC 31 (LTI), L M C 41, I~MC 45 (LTI, STI), LMC 46. Listeria mottoc y t o g e n e s LMC 74, S a l m o n e l l a enterica (subsp. e n t e r i c a ) serotypc Laridi LMC 48, S. e n t e r i c a (subsp. e n t e r i c a ) serotype Enteritidis L M C 49. Caml~ylobacter j e j t m i LMC 6 and C. coli LMC 16 (LMC, Laboratory o f Food Chemistry, Department of Chemistry and Biochemistry, University of Berne, Switzerland).
147
60 rain at 60°C. One ml of the iysate was centrifuged in a 1.5-ml Eppendorf tube for 10 min at 14.500 g and 450 Jal of the supernatant were used for nucleic acid isolation.
Protocol with emqchment
In order to increase the sensitivity of detection, bacteria were cultivated for 6 h in modified E. coli broth with novobiocine (mEC+n) prior to the isolation o f nucleic acids. Ten g of ground b e e f and 100 ml of mEC+n (20 g tryptone, 1.12 g bile salts no. 3 IOxoid L561, 5.0 g lactose, 4.0 g K~HPO 4. 1.5 KH_,PO~ and 5.0 g NaCI per liter H~O,-pH 6.9 _0. ! : prior to use, novobiocine ISigma N-16281 was added at a final concentration of 20 nag/l) were put into a stomacher bag (Merck BA6041) and homogenized for 30 s at low blender speed. The homogenate was decanted into a 250-ml flask and incubated in a "Gallcnkamp Rotatory Shaker" at 120 rpm and 37°C for 6 h. Subsequently, the bacteria from 50 ml were isolated and lysed as described above. Again, 450 ~1 of the supernatant were used for bacterial DNA isolation.
Meat samples Initial experiments for the development of the isolation protocol werc done with lreshly minced ground beef purchased in a local grocery store. For further evaluation of this isolation protocol, ground b e e f (100-g samples) was bought from different stores in Berne.
Isolation of bacteria from ground beef Protocol withmtt emqchment
Ten g of ground beef or hamburger patties were m i x e d with 75 mi o f h o m o g e n i z a t i o n b u f f e r (10 mmol/l Tris-HCl pH 8.0, 150 mmol/I NaCI, 2 mmol/1 EDTA, 0 . 1 % SDS, 0.5 mg/ml pronase [Boehringer, Mannheim, G e r m a n y l ) in a stomacher bag with mesh insert (Merck BA6041ISTR). Homogenization was achieved by using a "Colworth stomacher blender 400" at low speed lbr 30 s followed by incubation of the bags in a water bath for 30 rain at 37°C to release adsorbed bacteria from the meat surface. Fifty ml of the buffer were transferred to a 50-ml tube and centrifuged at 30,000 g for 30 min. The supernatant was removed and the pellet was lysed in a total volume of 2 ml o f l y s i s b u f f e r ( 1 0 m m o l / l T r i s - H C l pH 8.0, 50 mmol/l KCI, 1.5 g mmol/I MgCIO, containing 4 mg/ml lysozyme (Sigma L6876, St~ Louis, MO) and 0.2 mg/ml proteinase K (Boehringer 24568) for 20 min at r o o m t e m p e r a t u r e f o l l o w e d by
Isolation of bacteria from E. coli-positive food samples Food samples (cream, butter, pastry, cheese) confirmed to be E. co/i-positive by standard bacterial culture were provided by the official state control laboratory of the Berne Canton. Five to 50 ml of 1:10 NaCI-peptone Ik)od homogenates prepared by the control laboratory were decanted into a 50-ml tube and centrifuged as described for ground beef. Lysis of the bacterial pellet was performed using a 2-ml lysis buffer solution, as specified above, in combination with 200 ~1 of 5 M GuHCI ( S i g m a G - 4 5 0 5 ) . F o l l o w i n g c e n t r i f u g a t i o n at 14,500 g, 450 [al of the supernatant were used for DNA isolation.
Isolation of bacterial nucleic acids B a c t e r i a l DNA was isolated from bacterial lysates with the "Wizard" DNA clean-up system ( P r o m e g a , M a d i s o n , W I ) , a c c o r d i n g to the m a n u f a c t u r e r ' s instructions, briefly, 450 gtl of lysed bacteria (supernatant) were added to 1 ml of Wizard DNA purification resin aqd mixed by gentle inversion. The mixture was given onto a Wizard column, the column was washed with 2 ml of isopropanol 80% and dried at room temperature for 15 min. DNA was eluted with 50 lal of sterile distilled wate,- preheated to 70°C: 10 ~.tl of this DNA solution were used for PCR.
148
M. GILGEN ET AL.
Spiking experiments
Analysis o f PCR proehwts
Ten tal of 10-fold serial dilutions of overnight cultures of strain O157:H7 (CDC Atlanta 933) were inoculated into 10 g of ground beef or 10 g of hamburger meat and mixed Vigorously. Spiked samples were kept at 4°C for one day a n d a t - 2 0 ° C for up to 30 days to simulate conditions as they would exist in stored beef or hamburgers. Colony-tbrming units (CFU) of the culture were determined by plating on LB agar.
Aliquots of reactions (10 or 20 lal) were mixed with 6xxylene sample buffer (0.25 % xylene cyanol, 15% Ficoil IPharmacial in H,O) and separated through 1.5 % agarose gels. The ]lmplification products were visualized b'y ethidium bromide staining and UV transillumination.
RESULTS
PCR Oligonttt4eotides
Oligonucleotides used for PCR were synthesized by -Microsynth" (Wangen, Switzerland). Primers of the first PCR systems were based on Gannon et al. (1992). Primers for additional nested PCRs were chosen based on the sequences of Calderwood et al. (1987), Jackson et al. (1987) and Meyer et aL (1992). The primer pairs I-1 ( A C A C T G G A T G A T C T CAGTGG)/I-2 (CTGAATCCCCCTCCATTATG) and I-3 ( T T G T C A T C A T C A T G C A T C G C ) / I-4 (AGTTACACAATCAGGCGTCG) were used to amplify a region of the VTI gene. Primer pair I-1/1-2 results in an amplicon of 614 bp, whereas the length of the fragment of the nested system I-3/1-4 is 247 bp. Primers for the amplification of VT2 sequences, including genetic variants (Meyer et aL, 1992), were II-I (CCATGACRACGGACAGCAGTT) and 11-2 ( C C T G T C A R C T G A G C A C T T T G ) resulting in a PCR product of 779 bp; the nested system !I-3 (GTTCTGCGTTTTGTC ACTGT)/ I I-4 ( A GCTGT A T r A C ~ C C C A T A A ) yielded a fragment of 372 bp; maIB PCR was performed with the primers and condilions de~ribed previously (Meyer et aL, ! 991). Reaction conditions
Reaction volumes were 100 lal and conditions were I0 mmol/l Tris-HCI pH 8.8, 50 mmol/i KCI, 0.1% Triton X-100, 2 lag/ml BSA, 200 lamol/I each dNTP (Promega U1240), 2 U of Taq polymerase (Promega M1865), 3 mmol/1 MgCI~ and 0.25 [.tmol/! primer tbr the first round of PCR frith the primer pairs 1-1/I-2 and I1-11II-2, and 1.5 mmol/l MgCI~ and 0.5 lamol/i primer for the second, nested PCR'with the primer pairs !-3/I-4 or II-3/II-4, respectively. The reaction mixes were overlaid with 80 gl of mineral oil (Sigma M 5 ~ ) , Cycling was done on a "Techne cycler PHC3" with ! min at 94°C, |bllowed by 25 cycles at 94°C for 35 s, 60°C for 90 s, 72°C tbr 100 s with the outer primer pairs and 40 cycles with the nested PCR systems at 94°C for 5 s, 55°C for 30 s and 72°C for 40 s. A final extension step of 3 min at 72°C was included.
Specificity and sensitivity of nested PCR systems Spec~ficiO'
The nested PCR systems for the detection of V T I and V T 2 genes were tested with V T E C strains o f different serotypes and with various other E. coli strains and enteric bacteria (fig. 1). The data indicate that the PCR systems are specific, with the length of the products being as expected, 247 bp for the V T I PCR system and 372 bp lbr the VT2 system. No amplicons were obtained from r,trains that did not contain either the V T ! or the VT2 gene. Since nested P C R systems were used, further identification of the a m p l i c o n s by Southern hybridization was not necessary.
Sensitivity
To determine the detection limits of the two used nested PCR systems, serial dilutions o f a microscopically counted, lysed culture of strain O157:H7 were subjected to PCR. The sensitivity was the s a m e for both s y s t e m s , e n a b l i n g the detection of DNA corresponding to about 10 bacterial cells (table I).
Sensitivity of VTEC detection from ground beef Ground beef samples (10 g) were spiked with 10-fold serial dilutions of an overnight culture of a VT-producing E. coii strain (O157:H7) and bacterial DNA was isolated as described in "Materi-
'v't'~C D E T E C T I O N IN G R O U N D B E E F
2 3
149
4:5 6 7 8 9 1011!213141516171819L
A 247
1 2 3 4:5:6:7:8
9 101112:13141516!71819L
B 372 -
Fig. 1. Specificity of tile.used nested PCR systems. Agarose gel with ethidium-bromide-stained amplicons. DNA equivalent to 10 3 CFU from different VTEC strains and other enteric bacteria was subjected to our nested PCR systems. A) Amplicons produced during PCR with the nested PCR system Ibr VTI: B) amplicons produced during PCR with the nested PCR system for VT2. Lanes: I =E. coli O26:H! ! (VTI); 2=E. coli OI57:H- (VT2); 3=E. coli O60:H- (VTI); 4=E. coil O117:H (VTI); 5=E. coli O79:H not typable (VT2): 6=E. coil O30:H21 (VT2); 7=E. coli O157:H7 (VTI/VT2); 8=O74:H34 (VTI/VT2); 9=E. coli LMC 30; 10=E. coil LMC 31: i I=E. coli LMC 41: 12=E. coli LMC 45; 13=E. coli LMC 46; 14=L. monotTtogenes: 15 =S. enterica (subsp. enterica) serotype Laridi; 16=S. enterica (subsp. enteric'a) serotype Enteritidis; 17=C. jejuni; 18=C. coli; 19=PCR-negative control; L= 100-bp DNA standard (Gibco BRL). Amplicons of low molecular weight (< 100 bp) are artefacts that may occur in the absence of target DNA. Additional bands in positive samples are due to high target and high cycle numbers (Bell and DeMarini, 1991 ).
als and Methods"~ As outlined in figure 2, detection was possible up to a dilution of I × 106, corresponding to 110 CFU per 10 g. The use of a 6-h enrichment step enhanced the detection limit by a factor of 100, enabling the detection of 1 to 10 CFU in 10 g of ground beef. The sensitivity of VT detection from spiked manufactured hamburger patties was found to be reproducibly in the same range (data not shown).
Detection of VT-producing E. coli in ground beef samples of various origin Thirty ground beef samples obtained fi'om different butcher shops located in Berne were analysed in December 1994 for the presence of VTECs of different serotypes. Both sample preparation protocols were applied with each sample. However, to increase efficacy and efficiency,
150
M. GiLGEN ET AL.
Table L Sensitivities of nested PCR systems alone and in combination with isolation procedures of E. coli O157:H7 from 10 g of ground beef (without and with a 6-h cultivation in mEC+n).
Detection procedure
Detection limit
~ R alone !*~ Isolation+ PCR Cultivation + isolation + PCR
10 cells/reaction 110 CFU/IO g 1.1 CFU/10 g
(*) PCR alone was done with suspensionsof microscopically countedbacterialcells.
A
Detection of VT-producing E. coli in various E. co//-positive food samples In c o l l a b o r a t i o n with the official control laboratory of the Berne Canton, we analysed food samples found to be E. coli-positive by culture for the presence of VTEC. All samples were first analysed by a malB-specific PCR in order to confirm the positive culture results for E. coli and to assess the presence of PCR inhibitors. However, only 13 of 20 samples were positive by malB PCR, indicating that not all culture-positive results could be confirmed by PCR. Some P C R - n e g a t i v e samples such as pastry were shown to contain PCR inhibitors, most likely originating from the food matrix. Subsequently, the samples were tested with the PCR assays specific for the VT1 and V T 2 s e q u e n c e s (table II). One cheese sample was found to be positive for VT2 sequences.
DISCUSSION
B
Ground beef samples (1:0 g)were inoculated with different amounts 0fEi Coil OI57!H71 Detection was done with the nested ~ R system for VTI, A) Amplicons produced by PCR tbllowing the direct isolation and detection procedure without cultivation; B) amplicons produced by PCR following the isolation and detection procedure using a 6-h enrichment step. Lanes: 1= 1.1 × 10 6 CFU ; 2=l.ix105 CFU; 3=1.1 x104 CFU; 4=l.lX 103 CFU; 5=110 CFU; 6=11 CFU; 7=1.1 CFU; 8=0 CFU: 9= PCR-negative control ; i 0= PCR-positive control ; L = 100-bp DNA standard (Gibco BRL).
DNA isolates were combined and subjected to PCR analysis with either the VT 1 or VT2 system. We found one sample to contain VT1 and VT2 sequences; all other samples yielded negative results (table II).
Since outbreaks with VTECs are most often associated with the consumption of ground beef (Bell et al., 1994; Wells et al., 1983; Willshaw et aL, 1994), the purpose of this study was to assess the capacity of the PCR technique for the detection of these human pathogens in ground beef samples and hamburgers. In this study, the amplification reactions were designed to detect genes coding for either VT1 or VT2. Experiments with heat-labile and heat-stable enterotoxins from E. coil indicate that a good correlation between the phenotype, i.e. the capacity to actually produce the toxin, and the genotype can be expected (Candrian et aL, 1991 ). The high sensitivity and specificity of the PCR tests were a c h i e v e d by c h o o s i n g n e s t e d format';. We selected rather high cycle numbers in our PCR assays, which resulted in additional fragments of a higher molecular weight (fig. I B). This well known phenomenon (Bell and DeMarini, 1991) can be eliminated by reducing the number of cycles performed in the second PCR from forty to thirty without loss of sensitivity (data not shown).
I~111l.~-.
L . J L . , I ~,q~, l i l , . J J ¥
~..JlILq.Jlll~Lj
D~I~I~
L.~ L ~ £ ~ I '
i ~1 ! .2. I
Table II. Analysis of lood samples for the presence of verocytotoxin-producing E. coli.
Food tested
No. tested
malB-positive ~*~
No. confirmed
No. positive for VTEC
~, positive for VTEC
Ground beef
30 20
ND 13
! (VTI +VT2) 1' (VT2)
3.3% 5%
E. coli-positive
food samples (*) m a l B PCR (Meyer et al., 1991) was done to determine the presence of E. coli. ND=not done.
For sample preparation, two approaches were investigated. First, direct isolation of bacterial cells without culturing steps enabled the detection of culturable and non-culturable but viable cells. At this time, it is not clear to what extent dead bacteria are also recognized. This d e p e n d s on how fast cells that c o m p l e t e l y stopped any metabolic activity are lysed in the food matrix or during sample preparation. Secondly, a sample preparation protocol that includes an enrichment step was worked out. The culture step was limited to 6 h, since longer incubation times did not increase sensitivity (data not shown). With this approach, a hundred-fold increase in the detection limit could be realized even when samples spiked with highly viable bacteria from an overnight culture were stored at - 2 0 ° C to mimic common storage conditions potentially reducing viability. However, it cannot be excluded that this improvement in detection limit may not be observed with naturally contaminated samples (Niederhauser et al., 1992). As shown in table I, the direct sample preparation protocol yields a detection limit that is dependent exclusively on the detection limit of the PCR alone (10 cells per reaction). This protocol enables analysis of a 3 % fraction of the original food sample in each PCR reaction. Assuming no loss during sample preparation, one would theretbre expect a detection limit of about 330 cells per analysed sample volume. The lower value of 110 CFU per analysed sample that we found in our experiments is due to non-culturable cells in the bacterial culture used
for spiking, but clearly demonstrates the effectiveness of our sample preparation protocol. According to our experience with the Wizard DNA clean-up resin, the very effective sample preparation was mainly due to the use of this resin. When m a l B PCR was used to detect E. coli in food samples previously found to be positive by culture, only 13 of 20 samples were recognized. This PCR assay is not in the nested format and exhibits a sensitivity of about 100 cells per amplification reaction without Southern hybridization (Candrian et al., 1991). We therefore conclude that either the PCR-negative samples were contaminated with only low nurnbers or that PCR inhibitors were present in some samples that were coisolated with the DNA. From 30 analysed ground beef samples in this study, 3.3 % were V T E C - p o s i t i v e (table II). Similar results were reported by other investigators. Analysing ground beef samples, Doyle and Schoeni (1987) found 6 of 164 ('~.7 %) and B a u m g a r t n e r and Grand (1995) 4 of 166 (2.4%) samples to be positive. The PCR-positive results of the latter could not be confirmed by culture indicating the presence of non-culturable bacterial cells. Additionally, we have shown that VTEC may be present at low percentages in dairy and bakery products (table II). hz conclusion, the results of this study prove that analytical tests based on in v i o v amplification of bacterial nucleic acids are valuable tools for monitoring VTEC contamination. However,
152
M. G ! L G E N E T AL.
several challenges remain for future research. For example, not all diarrhoeagenic VTEC strains may have the same potential to induce HUS, and further virulence genes may be important. In addition, the significance of non-culturable but viable bacterial cells in food and water for the occurrence of human infectious diseases needs to be further investigated.
Acknowledgements We thank M. Altwcgg for providing bacterial sirains. This work was in part supported by grant 316-91I)342 from the Swiss Federal Office of Public Health.
D~tection par PCR de Escherichia coli produisant des v~rotoxines dans la viande hach~e Les souches pathog~nes de Escherichia coli prod u i s a n t des v~rotoxines (VTs) sont r e c o n n u e s comme causes de maladies humaines. Des tests de d~tection rapides et sensibles sont alors demand6s d'urgence pour assurer la s6curit6 des denr6es alimentaires, surtout de la viande hach6e. Nous avons utilis6 deux essais de la ,~ nested PCR,, pour d6tecter les g~nes codants les VTI et VT2 ind6pendemment du s6rotype bact6rien. A l'aide d'un protocole de preparation directe de 1'6chantillon, nous avons d~cel6 1 i 0 C F U de E. c o l i v 6 r o t o x i n o g ~ n e s (ECVTs) dans 10g de viande hachfie. Apr~s un enri¢hissement de six heures, la limite de d6tection est de 1 "~ l0 cellules bact~riennes par 10 g de viande hach~e. Pour 6valuer notre syst~me de d6teetion, nous avons examin8 30 6chantillons de v i a , d e hach~e provenant de boucheries ~t Berne, Suisse. Un ~chantillon a rfv616 un r6sultat positif de PCR pour les deux g~nes, VTI et VT2, indiquant la presence de ECVTs, Finalement, 20 6chantillons homog6n6isSs de denr6es alimentaires dans lesquels la presence de souches de E. coli a ~t8 dSmontr~e par culture standard, ont 6t6 analys6s au moyen de notre m6thode, Le g~ne codant VT2 a 6t8 d6tect6 dans un ~chantillon de f r o m a g e . Les rfisultats sugg~rent que la mgthode de PCR d6crite peut servir d'outil valable pour la surveillance des contaminations de denr~es alimentaires par ECVTs. Mots-cl~s: Aliment, Escherichia coli; O157:H7, ECVT, PCR, Viande hachSe, D6tection.
References Baumgarmer. A. & G~and. M. (1995). Detection ot verotoxin-producing E~'cherichia coil in minced beef and raw hamburgers: comparison of polymerase chain reaction (PCR) and immunomagnetic beads. Arch. Lebensmittelhyg., 46, 127-130. Begum, D. & Jackson, M.P. (1995), Direct detection of Shiga-like toxin-producing Escherichia coli in ground beef using the polymerase chain reaction. MoL Cell, Probes. 9, 259-264. Bell, B.P., Goldofl, M,, Griffin, P.M., Davis, M,A., Gordon, D.C., Tarr, P.I,, Bartleson, C.A., Lewis° J.H., Barrett, 1".3., Wells, ].G., Baron, R. & Kobayashi, J. (1994L A multistate outbreak of Escherichia coli OI57:H7-associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. ,i. Am. Med. Assoc.. 212, 1349-1353. Bell, D.A. & DeMarini, D.M. (1991), Excessive cycling converts PCR products to random-length higher molecular weight fragments. Nucleic Acids Res.. 19, 5(179, Besser, R.E., Lett, S.M., Weber, J.T., Doyle, M.P., Barrett, T.J., Wells, J.G. & Griffin, P.M. (1993), An outbreak of diarrhea and hemolytic uremic syndrome fi'om Escherichia coil OI57:H7 in fresh pressed apple cider, J. Am. Med. Assoc., 269, 2217-2220. Blanco, M., Blanco, J., Blanco, J.E., Gonzalez, E.A., Gomes, T.A.T., Zerbini, L.F., Yano, T. & de Castro, A.F.P. (1994), Genes coding for Shiga-like toxins in bovine verotoxin-producing Escherichia coli (VTEC) strains belonging t.o different O:K:H serotypes. Vet. MicrobioL. 42, 105- I I0. Borczyk, A.A., Karmali, M,A., Lior, H. & Duncan. L.M.C. (1987), Bovine reservoir for verotoxia-producing Escherichkt coil O!57:H7. Lancet, I, 9g. Calderwood, S.B., Auclair, F., Donohue-Rolfe, A., Keusch, G.T. & Mekalanos, J.J. (1987), Nucleotide sequence of the Shiga-like toxin genes of EscheHchia coll. Proc. Nat, Acad. Sci. USA, 84, 4364-4368. Candrian, U., Furrer, B., H6felein, C., Meyer, R., Jermini, M. & Ltithy, J. (1991), Detection of Escherichia coli and identification of enterotoxigenic strains by primer-directed enzymatic amplification of specific DNA sequences, hit. J. Food Mictvbiol., 12, 339352. ,- Chapman, P.A., Siddons, C.A., Zadik, P.M. & Jewes, L. (1991), An improved selective medium for the isolation of Escherichia coli OI57. J. Med. Microbiol.. 35, 107-110. Doyle, M.P. & Schoeni, J.L. (1987), [solation of Escherichia coli O157:H7 from retail fresh meats and poultry. AppL Environ. MicrobioL, 53, 2394-2396. Gannon, V.P.J., King, R.K., Kim, J.Y. & Golsteyn Thomas, E.J. (1992), Rapid and sensitive method for detection of Shiga-like toxin-producing Escherichia coli in ground beef using the polymerase chain reaction. Appi. Environ. MicrobioL, 58, 3809-3815. Griffin, P.M. & Tauxe, R.V. (1991), The epidemiology of infections caused by Escherichia coil OI57:H7, other enterohemorrhagic E. coil. and the associated hemolytic uremic syndrome. EpMemioL Rev.. 13, 60-98. Hughes, J.M. & La Montagne, J.R. (1994), Emerging Infectious Diseases. J. b!fect. Dis., 170. 263-264. Jackson, M.P., Neill, R.J., O'Brien, A.D., Holmes, R.K. &
l l ' l ' ] ~ ' l "~ I ' ) I - . ~ T F I ' ~ T l l ' ) I k l
Newland. J.W. (1987), Nucleotide sequence analysis and comparison of the structural genes for Shiga-like toxin I and Shiga-like toxin II encoded by bacteriophages from Escherichia ,,'o!i 933. H','MS MicrohiM. Lett., 44, 109- I 14. Johnson, W.M., Lior, H. & Benzanson, G.S. 119831, Cytotoxic Escherichia colt O157:H7 associated with hemorrhagic colitis in Canada, Lancet. 1, 76. Johnson, W.M,, Pollard, D.R., Lior, H., Tyler, S.D. & Rozee, K.R. (1990), Differentiation of genes coding for Escherichia colt Verotoxin 2 and the Verotoxin associated with porcine edema disease (VTe) by the polymerase chain reaction. J. Clin. Microbiol.. 28, 2351-2353. Kaplan. B.S., Cleary, T.G. & Obrig, T.G. (1990), Recent advances in understanding the pathogenesis of the hemolytic-uremic syndrome. Pediatr. Nephrol.. 4, 276-283. Karch, H. & Meyer, T. (19891, Single primer pair for amplifying segments of distinct Shiga-like-toxin genes by the polymerase chain reaction. J. Clin. Mk'robiol.. 27, 2751-2757. Karmali, M.A. (1989), Infection by verocytot¢~xin-prt~ducing Escherichia coil Clio. Microhiol. Rev.. 2, 15-38. Kleanthous, H., Fry, N.K., Smith, H.R., Gross, R.J. & Rowe, B. (1988), The use of sorbitol-MacConkey agar in conjunction with a specific antiserum tot" the detection of Veto cytotoxin-producing strains of Escherichia coli O 157. Epidemiol. h{t'ec't.. I 0 I, 327335. Konwalchuk, J., Speirs, J.l. & Stavric, S. (1977). Vero response to a cytotoxin of E,scherichia coil h~t?ct. Immun.. 18. 775-779. Ltischer, D., Graf Settah, S. & Altwegg, M. (1992), Bakterielle Durchfallerreger: Nachweis wm Verotoxinproduzierenden Escherichia colt mittels eolymeraseKettenreaktion. Schweiz. Med. Wochenschr.. 122, 1911-1918. March, S.B. & Ratnam, S. (1986), Sorbitol-M:tcConkcy medium for detection of Escherichiu colt OI57 associated with hernorrhagic colitis. J. Clin. Microbiol., 23, 869-872, Meyer, R., Liithy, J. & Candrian, U. ( 1991 ), Direct detection by polymerase chain reaction (PCR} of Escherichia colt in water and soft cheese and identification of enterotoxigenic strains. Lett. AppL Microbiol.. 13, 268-2"71. Meyer, T., Karch, H., Hacker, J., Bocklage. H. & Heesemann, J. (1992), Cloning and sequencing of a Shigalike toxin-ll-related gene from Escherichia colt OI57:H7 strain 7279. Zentralbl. BakterioL Mikrohiol. Hyg., 276, 176-188. Newland, J.W. & Neill, R.J. 11988), DNA probes for Shiga-like toxins I and I1 and for toxin-converting bacteriophages. J. Clin. Microbiol., 26, 1292-1297. Niederhauser, C., Candrian, U.. H~felein, C.. Jermini, M., Btihler, H.-P. & Ltithy, J. (1992), Use of polymerase chain reaction tbr detection of Listeria monotTtogenes in food. AppL Environ. Microbiol., 58, 15641568. Olsvik, O., Rimstad, E., Homes, E., Strockbine, N., Wasteson, Y., Lund, A. & Wachsmuth, K. (1991), A nested PCR followed by magnetic separation of amplified fragments for detection of Eschcrwhia colt Shiga-like toxin genes. Mol. Cell. Probes. 5, 429-435.
IAI
l'7.1~l)llAll~
K*K't~r
1 ~'~
Paton, A.W., Paton, J.C., Goldwater, PN. & Manning, P.A. (1993), Direct detection of Escherichia co!i Shiga-like toxin genes in primary fecal cultures by polyme.'ase chain reaction. J. CIm. Microhiol.. 31. 3063-3(}67. Pollard, D.R., Johnson, W.M., l,ior, H., Tyler. S.D. & Rozee, K.R. (1990), Rapid and specific detection of Verotoxin genes in Escheriehia colt by the polymerase chain reaction. J. CliJt. Mic~vbiol., 28, 540545. Reida, P.. Wolff, M., P6hls, H.-W., Kuhlmann, W., L e h m a c h e r . A., A l e k s i c , S., Karch, H. & Bockemtihl, J. (1994), An outbreak due to enteroh a e m o r r h a g i c E s c h e r i c h i a colt O I 5 7 : H 7 in a children's day care center characterized by personto-person transmission and environmental contamination, Zentralhl. Bakteriol. Mikrobiol. Hyg.. 281, 534-543. Ricotti, G.C., Buonomini, M.I.. Merlitti, A., Karch, H., Luzzi, 1. & Caprioli, A. (1994), A fatal case of hemorrhagic colitis, thrombocytopenia, and renal failure associated wqh verotoxin-producing, nonO 157 E.wherichia coil Clin. In.feet. Dis.. 19, 815816. Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B.. Wells, J.G., Davis, B.R., Hebert, R.J., Olcott, E.S., Johnson, L.M., Hargrett, N.T., Blake, P.A. & Cohen, M.L. (1983), Hemorrhagic colitis associated with a rare Escherichia colt serotype. N. EngL J. Med.. 308, 681-685. Scotland, S.M., Smith, H.R., Willshaw, G.A. & Rowe. B. (19831, Vero cytotoxin production in strain of E.scherichia colt is determined by genes tattled on bacteriophage. L,:4~cet. 2, 216. Smitl;. W.H.. (a een, P. & Parsell. Z. ( 1983~, Veto cell toxins in ~M'herichkl colt and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chicken and pigs. J. Gen. Microhiol., 129, 3121-3137. Strockbine, N.A., Marques, L.R., Newland, J.W., Smith, H.W., Holmes, R.K. & O'Brien, A.D. (1986), Two t o x i n - c o n v e m n g phages from E s c h e r i c h i a colt OI57:H7 strain 933 encode antigenically distinct toxins with similar biologic activities, htfect, hnmlm.. 53, 135-14(}. Swerdlow, D.L., Woodruff. B.A., Brady, R.C., Griffin, P.M., Tippen, S., Donell, H.D., Geldrcich, E., Payne, B.J., Meyer, A., Wells, J.G., Greene, K.D., Bright, M., Bean, N. & Blake, P.A. (1992), A waterborne outbreak in Missouri of Escherichia colt OI57:H7 associated with bloody diarrhea and death. Am,. buern. Med.. 117, 812-819. Wells. I.G., Davis, B.R., Wachsmuth, [.K.. Riley, L.W., Remis, R.S.. Sokolow, R. & Morris, G.K. (1983), Laboratory investigation of hem{wrhagic colitis outbreaks associated with a rare lzM'herichia coli seretype. ,/. Clin. Microhiol.. 18, 5 i 2-520. Willshaw, G.A., Smith, H.R.. Scotland, S.M.. Field, A.M. & Rowe, B. (1987), Heterogeneity of Escherichia colt phages encoding Vero cytotoxins : comparison of cloned sequences determining VTI and VT2 and development of specific gene probes../. Getr. Microbiol.. 133, 1309-1317. Willshaw. G.A., Thirwell, J., Jones. A.P., Parry. S., Salmon, R.L. & Hickey. M. (1994), Verocylotoxin-
154
M. G I L G E N ET AL.
producing Escherichia coli O157 in beef burgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolitic uraemic syndrome in Britain. Lett. Appl. MicrobioL, 19, 304-307. Witham, P.K., Yamashiro, C.T., Livak, K.J. &Batt, C.A. (t 996), A PCR~based assay for the detection of Escherichia coli Shiga-like toxin genes in ground beef. AppL Environ, Microbiol., 62, 1347-1353. Wright, D.J., Chapman, P.A. & Siddons, C.A. (1994), Immunomagnetic separation as a sensitive method
tbr isolating Escherichht coli OI57 from food sampies. Epidemiol. hzfect., 113, 31-39. Zadik, P.M., Chapman, P.A. & Siddons, C.A. (1993), Use of tellurite for the selection of verocytotoxigenic Escherichia t'oli O157. J. Med. MicrobioL, 39, 155158. Zhao, T., Doyle, M.P., Shere, J. & Garber, L. (1995), Prevalence of enterohemorrhagic Escherichia coli OI57:H7 in a survey of dairy herds. Appl. Environ. MicrobioL, 61, 1290-1293.
Res. Microbiol. 1998, 149, 145-154
PCR-based detection of verotoxin-producing Escherichia coli (VTEC) in ground beef M. Gilgen .,I E Htibner '*). C. H6felein, J. Li.ithy and U. Candrian ~'' Laboratory ~tl"[~od Clwmistry. Dep, rmu'nt of Chemistry am/Biochemisto', University o.['Berne. Freiestrasse 3, CH-3012 Berne
SUMMARY
Pathogenic strains of Escherichia coil producing verotoxins (VTs) have been recognized as a cause of human disease, and rapid and sensitive detection tests are urgently needed to ensure the safety of food, especially ground beef. We applied two nested polymerase chain reaction (PCR) assays to detect the genes encoding VT1 and VT2 irrespective of the bacterial serotype, in combination with a direct sample preparation protocol, we were able to uncover the presence of about 110 CFU of verotoxinogenic E. coil (V'rEC) in 10 g of ground beef. When a six-hour enrichment was included, we found the detection limit to be in the range of 1 to 10 bacterial cells per 10 g of ground beef. To evaluate our detection system, we tested 30 ground beef samples originating from butcher shops in Berne, Switzerland. One sample yielded positive PCR results for both the VT1 and VT2 genes, indicating the presence of verotoxinogenic E. coil Finally, 20 food homogenates, shown to contain E. coil strains by standard culture, were analysed with our method, and the gene encoding VT2 was detected in one cheese sample. The results suggest that the described PCR method can serve as a valuable tool for the surveillance of VTEC contamination of foods.
Key-words: Food, Escherichia coli; O157:H7, VTEC, PCR, Ground beef, Detection.
INTRODUCTION Four classes of enterovirulent Escherichia coil have been recognized to cause gastroenteritis in humans" enterotoxigenic (ETEC), enteroinvasive (EIEC), enteropathogenic (EPEC) and enterohaemorrhagic (EHEC) strains. EHECs have as a common feature the production of protein toxins, cytolethal to Vero cells in culture (Konwalchuk et aL, 1977) and therefore referred to as verotoxins
(VTs), and the VT-producing E. coil strains verocytotoxin-producing E. coli (VTEC). VTEC strains produce two irnmunologically distinct VTs (VTI and VT2) either exclusively or in combination, both of which are encoded by bacteriophages (Scotland et aL, 1983; Smith et al., 1983). Since VTs ale closely related to Shiga toxin, they are often referred to as Shiga-like toxins (SLT-I and SLT-II) (Strockbine et al., 1986). VTEC strains ate implicated in the pathogenesis of the
Submitted January 3, 1997, accepted October I0, 1997. (*) Corresponding attthor. " Present address: Central Laboratory BTS SRC. PCR Diagnostics. Wankdorl'stra.,,se 10, P.O. Box, CH-3000 Berne 22, Switzerland. Cb~Presenl address: Swiss Federal Office of Public Health, Division of Biologicals. Section for Vaccines, CH-3003 Berne. Switzerland.
M. G I L G E N E T A L .
146
diarrhoea-associated form of haemolytic-uraemic syndrome (HUS), which is the most common cause of acute renal failure in childhood (Kaplan et aL. 1990). The prototype of the VT-producing E. coli iS the enterohaemorrhagic strain O157:H7, associated with a spectrum of diseases including bloody diarrhoea, haemorrhagic colitis and thrombotic thrombocytopenic purpura (Karmali, 1989). The infective dose of this pathogen is probably very low (Willshaw et al., 1994). E. coil OI57:H7 and the diseases it causes have only recently become prominent, first being identified as a cause of haemorrhagic colitis in Canada in the late 1970's (Johnson et al., 1983; Riley et al., 1983). This organism is not only a prototype for the EHEC group, but also tbr new and emerging bacterial pathogens in general (Hughes and La Montagne, 1994). Outbreaks and isolated cases of disease with E. coil O157:H7 have been associated with the
consumption of raw or undercooked hamburger patties (Bell et al., 1994; Wells et al., 1983; Willshaw et al., 1994), milk (Borczyk et aL, 1987; Wright et ai., 1994) and apple cider (Besser et ai., 1993). Waterborne (Swerdlow et aL, 1992) and person-to-person (Reida et al., 1994) transmission are additional modes of transmission for this pathogen. In addition to O157:H7 VTEC, at least 50 serotypes have been recognized. Although no outbreaks of bloody stools caused by non-O157 strains have been reported in the United States, Canada and the Un;,ted Kingdom, several investigators have isolated such VT-producing strains from patients with diarrhoea, HUS, from cattle herds and from foods of animal origin (Blanco et aL, 1994; Griffin and Tauxe, 1991 ; L0scher et aL, 1992; Ricotti et aL, 1994; Zhao et aL, 1995). Mo~t detection met~-ods for verotoxinogenic E. coil ~re based on the cultivation of serotype O157. Because E. coil O157 does not ferment sorbitol rapidly, sorbitol MacConkey agar can be
BSA CFU EHEC HUS LB mEC+n
= = = = = =
bovine serum albumin. colony-forming unit. enterohaemorrhagic E. coll. haemolytic uraemic syndrome. Luria-Bertani (medium~. modified E. coli broth+novobiocine.
used to rapidly screen clinical or environmental isolates (Kleanthous et ai., 1988; March and Ratham, 1986). Improved media additionally contain cefixime (Chapman et al., 1991) and tellurite (Zadik et al., 1993). Identification can be done by direct agglutination of the colourless colonies with commercially av~i~able latex-conjugated antisera. This method and other diagnostic tests like ELISA, immunomagnetic separation and "Petrifilm" test kit HEC are limited to the isolation and detection of serotype O157. However, the detection of other serotypes of VT-producing E. coil (which do ferment sorbitol and do not agglutinate O157 antisera) is possible by use of DNA probes to the VT genes (Newland and Neill, 1988; Wiilshaw et al., 1987) or VT genespecific PCR (Begum and Jackson, 1995; Gannon et al., 1992; Johnson et al., 1990; Karch and Meyer, 1989; Olsvik et al., 1991 ; Paton et al., 1993; Pollard et aL, 1990; Witham et al., 1996). PCR not only allows detection of VTECs irrespective of their serotype, but also of different variants of VT genes. Since the virulence potential of non-O157 VTEC strains is not yet known, it is important to have detection methods recognizing all E. coil strains carrying VT genes. Based on the PCR assays described by Gannon et al. (1992), we developed nested PCR systems for the detection of the VTI and VT2 genes of VTEC. Additionally, we established an isolation method that enables the detection of low numbers of VTEC in ground beef.
MATERIALS
AND METHODS
Bacterial strains
Several enteric bacterial strains including VTproducing E. coil of various serotypes were used for testing the specificity of the developed PCR
PCR SDS SLT-! SLT-II VT VTEC
= = = = = =
p o l y m e r a s e chain reaction. stadium dodecyl sulphate. Shiga-like toxin I. Shiga-like toxin II. verotoxin. verotoxinogenic E. coli.
"vn ~t_ LPL T E C T i O N i N G R O U N D B E E k
systems and for the evaluation of isolation methods. All used VTEC strains had been involved in cases of severe diarrhoea. Strains containing the V T I gene w e r e : O 2 6 : H 1 1 ( E 3 7 8 7 , B. R o w e , PHLS, Colindale, London, England), O60:H(MM 730, M. Altwegg, Department o f Medical Microbiology, University of Ziirich, Switzerland) and O! 17:H- (MM 731, Altwegg). The follo',;ving V T E C s h a r b o u r e d the V T 2 g e n e : O I 5 7 : H (E32511, B. R o w e l O79:H not typable (MM732, Altwegg) and O30:H21 (MM734, Altwegg). Both genes were present in two strains: O157:H7 (CDC Atlanta 933. A l t w e g g ) and O 7 4 : H 3 4 ( M M 7 3 6 , Altwegg). Other bacterial strains used were E. coli L M C 30 ( h e a t - l a b i l e e n t e r o t o x i n [LTI], heatstable enterotoxin [STILL LMC 31 (LTI), L M C 41, I~MC 45 (LTI, STI), LMC 46. Listeria mottoc y t o g e n e s LMC 74, S a l m o n e l l a enterica (subsp. e n t e r i c a ) serotypc Laridi LMC 48, S. e n t e r i c a (subsp. e n t e r i c a ) serotype Enteritidis L M C 49. Caml~ylobacter j e j t m i LMC 6 and C. coli LMC 16 (LMC, Laboratory o f Food Chemistry, Department of Chemistry and Biochemistry, University of Berne, Switzerland).
147
60 rain at 60°C. One ml of the iysate was centrifuged in a 1.5-ml Eppendorf tube for 10 min at 14.500 g and 450 Jal of the supernatant were used for nucleic acid isolation.
Protocol with emqchment
In order to increase the sensitivity of detection, bacteria were cultivated for 6 h in modified E. coli broth with novobiocine (mEC+n) prior to the isolation o f nucleic acids. Ten g of ground b e e f and 100 ml of mEC+n (20 g tryptone, 1.12 g bile salts no. 3 IOxoid L561, 5.0 g lactose, 4.0 g K~HPO 4. 1.5 KH_,PO~ and 5.0 g NaCI per liter H~O,-pH 6.9 _0. ! : prior to use, novobiocine ISigma N-16281 was added at a final concentration of 20 nag/l) were put into a stomacher bag (Merck BA6041) and homogenized for 30 s at low blender speed. The homogenate was decanted into a 250-ml flask and incubated in a "Gallcnkamp Rotatory Shaker" at 120 rpm and 37°C for 6 h. Subsequently, the bacteria from 50 ml were isolated and lysed as described above. Again, 450 ~1 of the supernatant were used for bacterial DNA isolation.
Meat samples Initial experiments for the development of the isolation protocol werc done with lreshly minced ground beef purchased in a local grocery store. For further evaluation of this isolation protocol, ground b e e f (100-g samples) was bought from different stores in Berne.
Isolation of bacteria from ground beef Protocol withmtt emqchment
Ten g of ground beef or hamburger patties were m i x e d with 75 mi o f h o m o g e n i z a t i o n b u f f e r (10 mmol/l Tris-HCl pH 8.0, 150 mmol/I NaCI, 2 mmol/1 EDTA, 0 . 1 % SDS, 0.5 mg/ml pronase [Boehringer, Mannheim, G e r m a n y l ) in a stomacher bag with mesh insert (Merck BA6041ISTR). Homogenization was achieved by using a "Colworth stomacher blender 400" at low speed lbr 30 s followed by incubation of the bags in a water bath for 30 rain at 37°C to release adsorbed bacteria from the meat surface. Fifty ml of the buffer were transferred to a 50-ml tube and centrifuged at 30,000 g for 30 min. The supernatant was removed and the pellet was lysed in a total volume of 2 ml o f l y s i s b u f f e r ( 1 0 m m o l / l T r i s - H C l pH 8.0, 50 mmol/l KCI, 1.5 g mmol/I MgCIO, containing 4 mg/ml lysozyme (Sigma L6876, St~ Louis, MO) and 0.2 mg/ml proteinase K (Boehringer 24568) for 20 min at r o o m t e m p e r a t u r e f o l l o w e d by
Isolation of bacteria from E. coli-positive food samples Food samples (cream, butter, pastry, cheese) confirmed to be E. co/i-positive by standard bacterial culture were provided by the official state control laboratory of the Berne Canton. Five to 50 ml of 1:10 NaCI-peptone Ik)od homogenates prepared by the control laboratory were decanted into a 50-ml tube and centrifuged as described for ground beef. Lysis of the bacterial pellet was performed using a 2-ml lysis buffer solution, as specified above, in combination with 200 ~1 of 5 M GuHCI ( S i g m a G - 4 5 0 5 ) . F o l l o w i n g c e n t r i f u g a t i o n at 14,500 g, 450 [al of the supernatant were used for DNA isolation.
Isolation of bacterial nucleic acids B a c t e r i a l DNA was isolated from bacterial lysates with the "Wizard" DNA clean-up system ( P r o m e g a , M a d i s o n , W I ) , a c c o r d i n g to the m a n u f a c t u r e r ' s instructions, briefly, 450 gtl of lysed bacteria (supernatant) were added to 1 ml of Wizard DNA purification resin aqd mixed by gentle inversion. The mixture was given onto a Wizard column, the column was washed with 2 ml of isopropanol 80% and dried at room temperature for 15 min. DNA was eluted with 50 lal of sterile distilled wate,- preheated to 70°C: 10 ~.tl of this DNA solution were used for PCR.
148
M. GILGEN ET AL.
Spiking experiments
Analysis o f PCR proehwts
Ten tal of 10-fold serial dilutions of overnight cultures of strain O157:H7 (CDC Atlanta 933) were inoculated into 10 g of ground beef or 10 g of hamburger meat and mixed Vigorously. Spiked samples were kept at 4°C for one day a n d a t - 2 0 ° C for up to 30 days to simulate conditions as they would exist in stored beef or hamburgers. Colony-tbrming units (CFU) of the culture were determined by plating on LB agar.
Aliquots of reactions (10 or 20 lal) were mixed with 6xxylene sample buffer (0.25 % xylene cyanol, 15% Ficoil IPharmacial in H,O) and separated through 1.5 % agarose gels. The ]lmplification products were visualized b'y ethidium bromide staining and UV transillumination.
RESULTS
PCR Oligonttt4eotides
Oligonucleotides used for PCR were synthesized by -Microsynth" (Wangen, Switzerland). Primers of the first PCR systems were based on Gannon et al. (1992). Primers for additional nested PCRs were chosen based on the sequences of Calderwood et al. (1987), Jackson et al. (1987) and Meyer et aL (1992). The primer pairs I-1 ( A C A C T G G A T G A T C T CAGTGG)/I-2 (CTGAATCCCCCTCCATTATG) and I-3 ( T T G T C A T C A T C A T G C A T C G C ) / I-4 (AGTTACACAATCAGGCGTCG) were used to amplify a region of the VTI gene. Primer pair I-1/1-2 results in an amplicon of 614 bp, whereas the length of the fragment of the nested system I-3/1-4 is 247 bp. Primers for the amplification of VT2 sequences, including genetic variants (Meyer et aL, 1992), were II-I (CCATGACRACGGACAGCAGTT) and 11-2 ( C C T G T C A R C T G A G C A C T T T G ) resulting in a PCR product of 779 bp; the nested system !I-3 (GTTCTGCGTTTTGTC ACTGT)/ I I-4 ( A GCTGT A T r A C ~ C C C A T A A ) yielded a fragment of 372 bp; maIB PCR was performed with the primers and condilions de~ribed previously (Meyer et aL, ! 991). Reaction conditions
Reaction volumes were 100 lal and conditions were I0 mmol/l Tris-HCI pH 8.8, 50 mmol/i KCI, 0.1% Triton X-100, 2 lag/ml BSA, 200 lamol/I each dNTP (Promega U1240), 2 U of Taq polymerase (Promega M1865), 3 mmol/1 MgCI~ and 0.25 [.tmol/! primer tbr the first round of PCR frith the primer pairs 1-1/I-2 and I1-11II-2, and 1.5 mmol/l MgCI~ and 0.5 lamol/i primer for the second, nested PCR'with the primer pairs !-3/I-4 or II-3/II-4, respectively. The reaction mixes were overlaid with 80 gl of mineral oil (Sigma M 5 ~ ) , Cycling was done on a "Techne cycler PHC3" with ! min at 94°C, |bllowed by 25 cycles at 94°C for 35 s, 60°C for 90 s, 72°C tbr 100 s with the outer primer pairs and 40 cycles with the nested PCR systems at 94°C for 5 s, 55°C for 30 s and 72°C for 40 s. A final extension step of 3 min at 72°C was included.
Specificity and sensitivity of nested PCR systems Spec~ficiO'
The nested PCR systems for the detection of V T I and V T 2 genes were tested with V T E C strains o f different serotypes and with various other E. coli strains and enteric bacteria (fig. 1). The data indicate that the PCR systems are specific, with the length of the products being as expected, 247 bp for the V T I PCR system and 372 bp lbr the VT2 system. No amplicons were obtained from r,trains that did not contain either the V T ! or the VT2 gene. Since nested P C R systems were used, further identification of the a m p l i c o n s by Southern hybridization was not necessary.
Sensitivity
To determine the detection limits of the two used nested PCR systems, serial dilutions o f a microscopically counted, lysed culture of strain O157:H7 were subjected to PCR. The sensitivity was the s a m e for both s y s t e m s , e n a b l i n g the detection of DNA corresponding to about 10 bacterial cells (table I).
Sensitivity of VTEC detection from ground beef Ground beef samples (10 g) were spiked with 10-fold serial dilutions of an overnight culture of a VT-producing E. coii strain (O157:H7) and bacterial DNA was isolated as described in "Materi-
'v't'~C D E T E C T I O N IN G R O U N D B E E F
2 3
149
4:5 6 7 8 9 1011!213141516171819L
A 247
1 2 3 4:5:6:7:8
9 101112:13141516!71819L
B 372 -
Fig. 1. Specificity of tile.used nested PCR systems. Agarose gel with ethidium-bromide-stained amplicons. DNA equivalent to 10 3 CFU from different VTEC strains and other enteric bacteria was subjected to our nested PCR systems. A) Amplicons produced during PCR with the nested PCR system Ibr VTI: B) amplicons produced during PCR with the nested PCR system for VT2. Lanes: I =E. coli O26:H! ! (VTI); 2=E. coli OI57:H- (VT2); 3=E. coli O60:H- (VTI); 4=E. coil O117:H (VTI); 5=E. coli O79:H not typable (VT2): 6=E. coil O30:H21 (VT2); 7=E. coli O157:H7 (VTI/VT2); 8=O74:H34 (VTI/VT2); 9=E. coli LMC 30; 10=E. coil LMC 31: i I=E. coli LMC 41: 12=E. coli LMC 45; 13=E. coli LMC 46; 14=L. monotTtogenes: 15 =S. enterica (subsp. enterica) serotype Laridi; 16=S. enterica (subsp. enteric'a) serotype Enteritidis; 17=C. jejuni; 18=C. coli; 19=PCR-negative control; L= 100-bp DNA standard (Gibco BRL). Amplicons of low molecular weight (< 100 bp) are artefacts that may occur in the absence of target DNA. Additional bands in positive samples are due to high target and high cycle numbers (Bell and DeMarini, 1991 ).
als and Methods"~ As outlined in figure 2, detection was possible up to a dilution of I × 106, corresponding to 110 CFU per 10 g. The use of a 6-h enrichment step enhanced the detection limit by a factor of 100, enabling the detection of 1 to 10 CFU in 10 g of ground beef. The sensitivity of VT detection from spiked manufactured hamburger patties was found to be reproducibly in the same range (data not shown).
Detection of VT-producing E. coli in ground beef samples of various origin Thirty ground beef samples obtained fi'om different butcher shops located in Berne were analysed in December 1994 for the presence of VTECs of different serotypes. Both sample preparation protocols were applied with each sample. However, to increase efficacy and efficiency,
150
M. GiLGEN ET AL.
Table L Sensitivities of nested PCR systems alone and in combination with isolation procedures of E. coli O157:H7 from 10 g of ground beef (without and with a 6-h cultivation in mEC+n).
Detection procedure
Detection limit
~ R alone !*~ Isolation+ PCR Cultivation + isolation + PCR
10 cells/reaction 110 CFU/IO g 1.1 CFU/10 g
(*) PCR alone was done with suspensionsof microscopically countedbacterialcells.
A
Detection of VT-producing E. coli in various E. co//-positive food samples In c o l l a b o r a t i o n with the official control laboratory of the Berne Canton, we analysed food samples found to be E. coli-positive by culture for the presence of VTEC. All samples were first analysed by a malB-specific PCR in order to confirm the positive culture results for E. coli and to assess the presence of PCR inhibitors. However, only 13 of 20 samples were positive by malB PCR, indicating that not all culture-positive results could be confirmed by PCR. Some P C R - n e g a t i v e samples such as pastry were shown to contain PCR inhibitors, most likely originating from the food matrix. Subsequently, the samples were tested with the PCR assays specific for the VT1 and V T 2 s e q u e n c e s (table II). One cheese sample was found to be positive for VT2 sequences.
DISCUSSION
B
Ground beef samples (1:0 g)were inoculated with different amounts 0fEi Coil OI57!H71 Detection was done with the nested ~ R system for VTI, A) Amplicons produced by PCR tbllowing the direct isolation and detection procedure without cultivation; B) amplicons produced by PCR following the isolation and detection procedure using a 6-h enrichment step. Lanes: 1= 1.1 × 10 6 CFU ; 2=l.ix105 CFU; 3=1.1 x104 CFU; 4=l.lX 103 CFU; 5=110 CFU; 6=11 CFU; 7=1.1 CFU; 8=0 CFU: 9= PCR-negative control ; i 0= PCR-positive control ; L = 100-bp DNA standard (Gibco BRL).
DNA isolates were combined and subjected to PCR analysis with either the VT 1 or VT2 system. We found one sample to contain VT1 and VT2 sequences; all other samples yielded negative results (table II).
Since outbreaks with VTECs are most often associated with the consumption of ground beef (Bell et al., 1994; Wells et al., 1983; Willshaw et aL, 1994), the purpose of this study was to assess the capacity of the PCR technique for the detection of these human pathogens in ground beef samples and hamburgers. In this study, the amplification reactions were designed to detect genes coding for either VT1 or VT2. Experiments with heat-labile and heat-stable enterotoxins from E. coil indicate that a good correlation between the phenotype, i.e. the capacity to actually produce the toxin, and the genotype can be expected (Candrian et aL, 1991 ). The high sensitivity and specificity of the PCR tests were a c h i e v e d by c h o o s i n g n e s t e d format';. We selected rather high cycle numbers in our PCR assays, which resulted in additional fragments of a higher molecular weight (fig. I B). This well known phenomenon (Bell and DeMarini, 1991) can be eliminated by reducing the number of cycles performed in the second PCR from forty to thirty without loss of sensitivity (data not shown).
I~111l.~-.
L . J L . , I ~,q~, l i l , . J J ¥
~..JlILq.Jlll~Lj
D~I~I~
L.~ L ~ £ ~ I '
i ~1 ! .2. I
Table II. Analysis of lood samples for the presence of verocytotoxin-producing E. coli.
Food tested
No. tested
malB-positive ~*~
No. confirmed
No. positive for VTEC
~, positive for VTEC
Ground beef
30 20
ND 13
! (VTI +VT2) 1' (VT2)
3.3% 5%
E. coli-positive
food samples (*) m a l B PCR (Meyer et al., 1991) was done to determine the presence of E. coli. ND=not done.
For sample preparation, two approaches were investigated. First, direct isolation of bacterial cells without culturing steps enabled the detection of culturable and non-culturable but viable cells. At this time, it is not clear to what extent dead bacteria are also recognized. This d e p e n d s on how fast cells that c o m p l e t e l y stopped any metabolic activity are lysed in the food matrix or during sample preparation. Secondly, a sample preparation protocol that includes an enrichment step was worked out. The culture step was limited to 6 h, since longer incubation times did not increase sensitivity (data not shown). With this approach, a hundred-fold increase in the detection limit could be realized even when samples spiked with highly viable bacteria from an overnight culture were stored at - 2 0 ° C to mimic common storage conditions potentially reducing viability. However, it cannot be excluded that this improvement in detection limit may not be observed with naturally contaminated samples (Niederhauser et al., 1992). As shown in table I, the direct sample preparation protocol yields a detection limit that is dependent exclusively on the detection limit of the PCR alone (10 cells per reaction). This protocol enables analysis of a 3 % fraction of the original food sample in each PCR reaction. Assuming no loss during sample preparation, one would theretbre expect a detection limit of about 330 cells per analysed sample volume. The lower value of 110 CFU per analysed sample that we found in our experiments is due to non-culturable cells in the bacterial culture used
for spiking, but clearly demonstrates the effectiveness of our sample preparation protocol. According to our experience with the Wizard DNA clean-up resin, the very effective sample preparation was mainly due to the use of this resin. When m a l B PCR was used to detect E. coli in food samples previously found to be positive by culture, only 13 of 20 samples were recognized. This PCR assay is not in the nested format and exhibits a sensitivity of about 100 cells per amplification reaction without Southern hybridization (Candrian et al., 1991). We therefore conclude that either the PCR-negative samples were contaminated with only low nurnbers or that PCR inhibitors were present in some samples that were coisolated with the DNA. From 30 analysed ground beef samples in this study, 3.3 % were V T E C - p o s i t i v e (table II). Similar results were reported by other investigators. Analysing ground beef samples, Doyle and Schoeni (1987) found 6 of 164 ('~.7 %) and B a u m g a r t n e r and Grand (1995) 4 of 166 (2.4%) samples to be positive. The PCR-positive results of the latter could not be confirmed by culture indicating the presence of non-culturable bacterial cells. Additionally, we have shown that VTEC may be present at low percentages in dairy and bakery products (table II). hz conclusion, the results of this study prove that analytical tests based on in v i o v amplification of bacterial nucleic acids are valuable tools for monitoring VTEC contamination. However,
152
M. G ! L G E N E T AL.
several challenges remain for future research. For example, not all diarrhoeagenic VTEC strains may have the same potential to induce HUS, and further virulence genes may be important. In addition, the significance of non-culturable but viable bacterial cells in food and water for the occurrence of human infectious diseases needs to be further investigated.
Acknowledgements We thank M. Altwcgg for providing bacterial sirains. This work was in part supported by grant 316-91I)342 from the Swiss Federal Office of Public Health.
D~tection par PCR de Escherichia coli produisant des v~rotoxines dans la viande hach~e Les souches pathog~nes de Escherichia coli prod u i s a n t des v~rotoxines (VTs) sont r e c o n n u e s comme causes de maladies humaines. Des tests de d~tection rapides et sensibles sont alors demand6s d'urgence pour assurer la s6curit6 des denr6es alimentaires, surtout de la viande hach6e. Nous avons utilis6 deux essais de la ,~ nested PCR,, pour d6tecter les g~nes codants les VTI et VT2 ind6pendemment du s6rotype bact6rien. A l'aide d'un protocole de preparation directe de 1'6chantillon, nous avons d~cel6 1 i 0 C F U de E. c o l i v 6 r o t o x i n o g ~ n e s (ECVTs) dans 10g de viande hachfie. Apr~s un enri¢hissement de six heures, la limite de d6tection est de 1 "~ l0 cellules bact~riennes par 10 g de viande hach~e. Pour 6valuer notre syst~me de d6teetion, nous avons examin8 30 6chantillons de v i a , d e hach~e provenant de boucheries ~t Berne, Suisse. Un ~chantillon a rfv616 un r6sultat positif de PCR pour les deux g~nes, VTI et VT2, indiquant la presence de ECVTs, Finalement, 20 6chantillons homog6n6isSs de denr6es alimentaires dans lesquels la presence de souches de E. coli a ~t8 dSmontr~e par culture standard, ont 6t6 analys6s au moyen de notre m6thode, Le g~ne codant VT2 a 6t8 d6tect6 dans un ~chantillon de f r o m a g e . Les rfisultats sugg~rent que la mgthode de PCR d6crite peut servir d'outil valable pour la surveillance des contaminations de denr~es alimentaires par ECVTs. Mots-cl~s: Aliment, Escherichia coli; O157:H7, ECVT, PCR, Viande hachSe, D6tection.
References Baumgarmer. A. & G~and. M. (1995). Detection ot verotoxin-producing E~'cherichia coil in minced beef and raw hamburgers: comparison of polymerase chain reaction (PCR) and immunomagnetic beads. Arch. Lebensmittelhyg., 46, 127-130. Begum, D. & Jackson, M.P. (1995), Direct detection of Shiga-like toxin-producing Escherichia coli in ground beef using the polymerase chain reaction. MoL Cell, Probes. 9, 259-264. Bell, B.P., Goldofl, M,, Griffin, P.M., Davis, M,A., Gordon, D.C., Tarr, P.I,, Bartleson, C.A., Lewis° J.H., Barrett, 1".3., Wells, ].G., Baron, R. & Kobayashi, J. (1994L A multistate outbreak of Escherichia coli OI57:H7-associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. ,i. Am. Med. Assoc.. 212, 1349-1353. Bell, D.A. & DeMarini, D.M. (1991), Excessive cycling converts PCR products to random-length higher molecular weight fragments. Nucleic Acids Res.. 19, 5(179, Besser, R.E., Lett, S.M., Weber, J.T., Doyle, M.P., Barrett, T.J., Wells, J.G. & Griffin, P.M. (1993), An outbreak of diarrhea and hemolytic uremic syndrome fi'om Escherichia coil OI57:H7 in fresh pressed apple cider, J. Am. Med. Assoc., 269, 2217-2220. Blanco, M., Blanco, J., Blanco, J.E., Gonzalez, E.A., Gomes, T.A.T., Zerbini, L.F., Yano, T. & de Castro, A.F.P. (1994), Genes coding for Shiga-like toxins in bovine verotoxin-producing Escherichia coli (VTEC) strains belonging t.o different O:K:H serotypes. Vet. MicrobioL. 42, 105- I I0. Borczyk, A.A., Karmali, M,A., Lior, H. & Duncan. L.M.C. (1987), Bovine reservoir for verotoxia-producing Escherichkt coil O!57:H7. Lancet, I, 9g. Calderwood, S.B., Auclair, F., Donohue-Rolfe, A., Keusch, G.T. & Mekalanos, J.J. (1987), Nucleotide sequence of the Shiga-like toxin genes of EscheHchia coll. Proc. Nat, Acad. Sci. USA, 84, 4364-4368. Candrian, U., Furrer, B., H6felein, C., Meyer, R., Jermini, M. & Ltithy, J. (1991), Detection of Escherichia coli and identification of enterotoxigenic strains by primer-directed enzymatic amplification of specific DNA sequences, hit. J. Food Mictvbiol., 12, 339352. ,- Chapman, P.A., Siddons, C.A., Zadik, P.M. & Jewes, L. (1991), An improved selective medium for the isolation of Escherichia coli OI57. J. Med. Microbiol.. 35, 107-110. Doyle, M.P. & Schoeni, J.L. (1987), [solation of Escherichia coli O157:H7 from retail fresh meats and poultry. AppL Environ. MicrobioL, 53, 2394-2396. Gannon, V.P.J., King, R.K., Kim, J.Y. & Golsteyn Thomas, E.J. (1992), Rapid and sensitive method for detection of Shiga-like toxin-producing Escherichia coli in ground beef using the polymerase chain reaction. Appi. Environ. MicrobioL, 58, 3809-3815. Griffin, P.M. & Tauxe, R.V. (1991), The epidemiology of infections caused by Escherichia coil OI57:H7, other enterohemorrhagic E. coil. and the associated hemolytic uremic syndrome. EpMemioL Rev.. 13, 60-98. Hughes, J.M. & La Montagne, J.R. (1994), Emerging Infectious Diseases. J. b!fect. Dis., 170. 263-264. Jackson, M.P., Neill, R.J., O'Brien, A.D., Holmes, R.K. &
l l ' l ' ] ~ ' l "~ I ' ) I - . ~ T F I ' ~ T l l ' ) I k l
Newland. J.W. (1987), Nucleotide sequence analysis and comparison of the structural genes for Shiga-like toxin I and Shiga-like toxin II encoded by bacteriophages from Escherichia ,,'o!i 933. H','MS MicrohiM. Lett., 44, 109- I 14. Johnson, W.M., Lior, H. & Benzanson, G.S. 119831, Cytotoxic Escherichia colt O157:H7 associated with hemorrhagic colitis in Canada, Lancet. 1, 76. Johnson, W.M,, Pollard, D.R., Lior, H., Tyler, S.D. & Rozee, K.R. (1990), Differentiation of genes coding for Escherichia colt Verotoxin 2 and the Verotoxin associated with porcine edema disease (VTe) by the polymerase chain reaction. J. Clin. Microbiol.. 28, 2351-2353. Kaplan. B.S., Cleary, T.G. & Obrig, T.G. (1990), Recent advances in understanding the pathogenesis of the hemolytic-uremic syndrome. Pediatr. Nephrol.. 4, 276-283. Karch, H. & Meyer, T. (19891, Single primer pair for amplifying segments of distinct Shiga-like-toxin genes by the polymerase chain reaction. J. Clin. Mk'robiol.. 27, 2751-2757. Karmali, M.A. (1989), Infection by verocytot¢~xin-prt~ducing Escherichia coil Clio. Microhiol. Rev.. 2, 15-38. Kleanthous, H., Fry, N.K., Smith, H.R., Gross, R.J. & Rowe, B. (1988), The use of sorbitol-MacConkey agar in conjunction with a specific antiserum tot" the detection of Veto cytotoxin-producing strains of Escherichia coli O 157. Epidemiol. h{t'ec't.. I 0 I, 327335. Konwalchuk, J., Speirs, J.l. & Stavric, S. (1977). Vero response to a cytotoxin of E,scherichia coil h~t?ct. Immun.. 18. 775-779. Ltischer, D., Graf Settah, S. & Altwegg, M. (1992), Bakterielle Durchfallerreger: Nachweis wm Verotoxinproduzierenden Escherichia colt mittels eolymeraseKettenreaktion. Schweiz. Med. Wochenschr.. 122, 1911-1918. March, S.B. & Ratnam, S. (1986), Sorbitol-M:tcConkcy medium for detection of Escherichiu colt OI57 associated with hernorrhagic colitis. J. Clin. Microbiol., 23, 869-872, Meyer, R., Liithy, J. & Candrian, U. ( 1991 ), Direct detection by polymerase chain reaction (PCR} of Escherichia colt in water and soft cheese and identification of enterotoxigenic strains. Lett. AppL Microbiol.. 13, 268-2"71. Meyer, T., Karch, H., Hacker, J., Bocklage. H. & Heesemann, J. (1992), Cloning and sequencing of a Shigalike toxin-ll-related gene from Escherichia colt OI57:H7 strain 7279. Zentralbl. BakterioL Mikrohiol. Hyg., 276, 176-188. Newland, J.W. & Neill, R.J. 11988), DNA probes for Shiga-like toxins I and I1 and for toxin-converting bacteriophages. J. Clin. Microbiol., 26, 1292-1297. Niederhauser, C., Candrian, U.. H~felein, C.. Jermini, M., Btihler, H.-P. & Ltithy, J. (1992), Use of polymerase chain reaction tbr detection of Listeria monotTtogenes in food. AppL Environ. Microbiol., 58, 15641568. Olsvik, O., Rimstad, E., Homes, E., Strockbine, N., Wasteson, Y., Lund, A. & Wachsmuth, K. (1991), A nested PCR followed by magnetic separation of amplified fragments for detection of Eschcrwhia colt Shiga-like toxin genes. Mol. Cell. Probes. 5, 429-435.
IAI
l'7.1~l)llAll~
K*K't~r
1 ~'~
Paton, A.W., Paton, J.C., Goldwater, PN. & Manning, P.A. (1993), Direct detection of Escherichia co!i Shiga-like toxin genes in primary fecal cultures by polyme.'ase chain reaction. J. CIm. Microhiol.. 31. 3063-3(}67. Pollard, D.R., Johnson, W.M., l,ior, H., Tyler. S.D. & Rozee, K.R. (1990), Rapid and specific detection of Verotoxin genes in Escheriehia colt by the polymerase chain reaction. J. CliJt. Mic~vbiol., 28, 540545. Reida, P.. Wolff, M., P6hls, H.-W., Kuhlmann, W., L e h m a c h e r . A., A l e k s i c , S., Karch, H. & Bockemtihl, J. (1994), An outbreak due to enteroh a e m o r r h a g i c E s c h e r i c h i a colt O I 5 7 : H 7 in a children's day care center characterized by personto-person transmission and environmental contamination, Zentralhl. Bakteriol. Mikrobiol. Hyg.. 281, 534-543. Ricotti, G.C., Buonomini, M.I.. Merlitti, A., Karch, H., Luzzi, 1. & Caprioli, A. (1994), A fatal case of hemorrhagic colitis, thrombocytopenia, and renal failure associated wqh verotoxin-producing, nonO 157 E.wherichia coil Clin. In.feet. Dis.. 19, 815816. Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B.. Wells, J.G., Davis, B.R., Hebert, R.J., Olcott, E.S., Johnson, L.M., Hargrett, N.T., Blake, P.A. & Cohen, M.L. (1983), Hemorrhagic colitis associated with a rare Escherichia colt serotype. N. EngL J. Med.. 308, 681-685. Scotland, S.M., Smith, H.R., Willshaw, G.A. & Rowe. B. (19831, Vero cytotoxin production in strain of E.scherichia colt is determined by genes tattled on bacteriophage. L,:4~cet. 2, 216. Smitl;. W.H.. (a een, P. & Parsell. Z. ( 1983~, Veto cell toxins in ~M'herichkl colt and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chicken and pigs. J. Gen. Microhiol., 129, 3121-3137. Strockbine, N.A., Marques, L.R., Newland, J.W., Smith, H.W., Holmes, R.K. & O'Brien, A.D. (1986), Two t o x i n - c o n v e m n g phages from E s c h e r i c h i a colt OI57:H7 strain 933 encode antigenically distinct toxins with similar biologic activities, htfect, hnmlm.. 53, 135-14(}. Swerdlow, D.L., Woodruff. B.A., Brady, R.C., Griffin, P.M., Tippen, S., Donell, H.D., Geldrcich, E., Payne, B.J., Meyer, A., Wells, J.G., Greene, K.D., Bright, M., Bean, N. & Blake, P.A. (1992), A waterborne outbreak in Missouri of Escherichia colt OI57:H7 associated with bloody diarrhea and death. Am,. buern. Med.. 117, 812-819. Wells. I.G., Davis, B.R., Wachsmuth, [.K.. Riley, L.W., Remis, R.S.. Sokolow, R. & Morris, G.K. (1983), Laboratory investigation of hem{wrhagic colitis outbreaks associated with a rare lzM'herichia coli seretype. ,/. Clin. Microhiol.. 18, 5 i 2-520. Willshaw, G.A., Smith, H.R.. Scotland, S.M.. Field, A.M. & Rowe, B. (1987), Heterogeneity of Escherichia colt phages encoding Vero cytotoxins : comparison of cloned sequences determining VTI and VT2 and development of specific gene probes../. Getr. Microbiol.. 133, 1309-1317. Willshaw. G.A., Thirwell, J., Jones. A.P., Parry. S., Salmon, R.L. & Hickey. M. (1994), Verocylotoxin-
154
M. G I L G E N ET AL.
producing Escherichia coli O157 in beef burgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolitic uraemic syndrome in Britain. Lett. Appl. MicrobioL, 19, 304-307. Witham, P.K., Yamashiro, C.T., Livak, K.J. &Batt, C.A. (t 996), A PCR~based assay for the detection of Escherichia coli Shiga-like toxin genes in ground beef. AppL Environ, Microbiol., 62, 1347-1353. Wright, D.J., Chapman, P.A. & Siddons, C.A. (1994), Immunomagnetic separation as a sensitive method
tbr isolating Escherichht coli OI57 from food sampies. Epidemiol. hzfect., 113, 31-39. Zadik, P.M., Chapman, P.A. & Siddons, C.A. (1993), Use of tellurite for the selection of verocytotoxigenic Escherichia t'oli O157. J. Med. MicrobioL, 39, 155158. Zhao, T., Doyle, M.P., Shere, J. & Garber, L. (1995), Prevalence of enterohemorrhagic Escherichia coli OI57:H7 in a survey of dairy herds. Appl. Environ. MicrobioL, 61, 1290-1293.