Veterinary Microbiology, 31 ( 1992 ) 251-261 Elsevier Science Publishers B.V., Amsterdam
251
Detection of entero- and verocyto-toxin genes in Escherichia coli from diarrhoeal disease in animals using the polymerase chain reaction M.J. Woodward a, P.J. Carroll b and C. W r a y b aMolecular Genetics Unit and Department of Bacteriology, New Haw, Weybridge, UK bCentral Veterinary Laboratory, New Haw, Weybridge, UK (Accepted 16 October 1991 )
ABSTRACT Woodward, M.J. Carroll, P.J. and Wray, C., 1992. Detection of entero- and verocyto-toxin genes in Escherichia coli from diarrhoeal disease in animals using the polymerase chain reaction. Vet. Microbiol., 31: 251-261. Oligonucleotide primers were designed for the specific polymerase chain reaction (PCR) amplification of the enterotoxins STIa and LTI and of the verocytotoxins VT 1 and VT2. All of 184 E. coli isolates from cases of diarrhoea from pigs, cattle and sheep gave identical toxin profiles by PCR and gene probe. Differentiation between VT2 and VT2v was achieved using two oligonucleotide primers pairs in PCR and showed that all of 34 VT2 + porcine isolates, of which 23 were 0138: K1, harboured VT2v whereas 20 VT2 + bovine and ovine isolates harboured VT2. No isolate harboured both VT2 polymorphs. Simplified methods for sample preparation for PCR were examined and showed that PCR was not inhibited by direct addition of broth culture to the reaction mixture.
INTRODUCTION
Although Escherichia coli is recognised as a normal member of the lower gut bacterial flora, some strains are potentially pathogenic (Sussman, 1985 ). Enterotoxigenic E. coli (ETEC) is an important cause of diarrhoeal disease in man and animals (Sojka, 1971; Sack, 1975; Robins-Browne, 1985 ). The heat-labile toxins LTI and LTII, and the heat stable toxins STIa, STIb and STII, are well characterised functionally, antigenically and genetically (for review see Sussman, 1985 ). Verocytotoxin-producing E. coli (VTEC) are associated with sporadic diarrhoea, haemorrhagic colitis, haemolytic uraemic syndrome and thrombocytopaenic purpurea in man (Karmali, 1989 ). In animals, VTEC have been shown to be associated with diarrhoea and oedema disease (Linggood and Thompson, 1987 ). The verocytotoxins, of which three antigenically distinct types have been defined as VT 1, VT2 and VT2 variants, VT2v, vtx2ha and vtx2hb have been characterised functionally and geneti0378-1135/92/$05.00
© 1992 Elsevier Science Publishers B.V. All rights reserved.
252
M.J, WOODWARD ET AL.
cally (for reviews see Smith and Scotland, 1988; Karmali, 1989; Ito et al, 1990). The VT2v toxin has been shown to produce oedema disease in pigs (MacLeod et al., 1991 ). There is a need to develop rapid, specific and sensitive methods for the detection of toxin-producing isolates orE. coli because of their clinical significance. Conventional techniques for toxin detection require either bioassays for ST (Dean et al., 1972; Smith and Halls, 1967) or cell culture assays for LT and VT (Guerrant et al., 1974; Sack and Sack, 1975; Konowalchuk et al., 1977 ) and are time-consuming and unsuitable for routine diagnosis. The use of gene probes to detect toxins is a well established technology but not used routinely because considerable safety precautions are required when radioisotopic labels are used. Immunological methods for the detection of LT and ST toxins by latex agglutination and EIA tests are available commercially and in routine use (Carroll et al., 1990). However a rapid, single method which can be used for the detection of all E. coli toxins is desirable. The polymerase chain reaction (PCR) (Saiki et al., 1988) involves the enzymic amplification ofa DNA sequence initiated from a pair of short DNA fragments (primers) which bind either side of the chosen target sequence. The amplified DNA product of this reaction can be easily detected by gel electrophoresis followed by ethidium bromide staining and UV transillumination. The PCR is rapid, sensitive and highly specific and has been developed to detect enterotoxin genes (Olive, 1989; Furrer et al, 1990; Victor et al., 1991 ) and verocytotoxin genes (Pollard et al., 1990; Johnson et al., 1990; Kobayashi et al., 1990) and gives results within a working day. Sophisticated developments of the PCR have been used to simultaneously detect several virulence genes within a single reaction (Fraenkel et al., 1989). We report here our initial findings on the development of a simple protocol for PCR tests for LT, ST 1a, VT 1 VT2 and VT2v as a preliminary to its use for clinical samples. MATERIALS AND METHODS
Bacterial strains E. coli strains isolated from cattle, pigs and sheep suffering from enteric disease were stored on Dorset egg slopes (PM5, Oxoid) at room temperature. All isolates were sub-cultured every six months. Serotyping was by standard methods (Sojka, 1965 ) and detection of adhesins was by latex agglutination (Thorns et al., 1989) using FIMBREX (CVL reagents) following manufacturer's recommendations. Detection of toxin genes by D N A - D N A hybridisation followed the methods of Woodward et al., (1990). A panel of reference strains which produced toxins as assessed by standard methods have been described (Sack and Sack, 1975; Nagy et al, 1976; Moseley et al., 1980; Willshaw et al., 1987; Woodward et al, 1990) and were used as positive controls
DETECTION OF ENTERO- AND VEROCYTO-TOXIN GENES IN E. COLI
253
in PCR experiments; the control strains were B44 09:K30:K99 F41 STIa, G7 08:K87:K88 LT STII, E57 0138:K81 STIa VT2v, H19 026:H19 VT1 and E3511 0 1 5 7 : H - VT2. Cultures were grown in Luria-Bertani 'LB' broth (Maniatis et al., 1982). Polymerase chain reaction assay
Oligonucleotide primers design. Primers were designed following the principles of Innis and Gelfand (1990). Primers were synthesised by an ABI PCRmate D N A synthesiser (PCR-MATE EP 391, Applied Biosystems Inc) following manufacturer's recommendations and are listed in Table 1. (b) Polymerase chain reaction. The reaction mixtures (50/tl) comprised 50 mM KC1, 10 mM Tris-HC1 pH 8.4, 0.01% (w/v) gelatin, MgC12 titrated in the range 1.5 to 8.0 mM; nucleotides 200/~M for each of dATP, dCTP, dGTP and dTTP (Pharmacia); 0.2/tM for each primer; and 2.5 U ofthermostable DNA polymerase (amplitaq, Perkin Elmer Cetus ). The reaction mixtures were overlaid with an equal volume of mineral oil. Samples were pipetted through the mineral oil into the mixture. Parameters for amplification were denaturation at 94°C for 2 min for 1 cycle followed by 30 cycles of 94°C for 1.5 min, annealing at 59°C for 1.5 min, extension at 72°C for 2 min, followed by a final extension by incubation at 72°C for 5 min. A reagent blank (negative control) was as described above but without addition of sample. TABLE1
Nucleotide Sequences of Synthetic Oligonucleotide Primers Target gene
LTI
Primer code
LT. 1 LT.2 STIa Sta. 1 STa. 2 VT1 VT1.1 VT 1.2 VT2al 1 VT2.1 VT2.2 VT2v VT2.3 VT2.4 VT2 VT2-3 VT2-5
Primer sequence 5'-3'
Primer location ( b p ) in published sequence
Size of amplified product (bp)
Source of DNA sequence used for primer design
ATTTACGGCGTTACTATCCTC TTTTGGTCTCGGTCAGATATG TCTTTCCCCTCTTTTAGTCAG ACAGGCAGGATTACAACAAAG TTAGACTTCTCGACTGCAAAG TGTTGTACGAAATCCCCTCTG CTATATCTGCGCCGGGTCTG AGACGAAGATGGTCAAAACG CCGTCAGGACTGTCTGAAAC GGACGCGATAATTAAACCG CCGTCAGGACTGTCTGAAAC GAGTCTGACAGGCAACTGTC
27-48 286-307 39-60 184-205 239-260 748-769 183-203 490-510 838-858 1481-1501 838-858 1544-1564
280
Olive ( 1989 )
166
Moseley et al. ( 1983 ) Strockbine et al. ( 1988 ) Ito et al. (1990) Ito el al. ( 1990 ) Ito et al. ( 1990 )
530 327 663 726
254
M.J. WOODWARD ET AL.
Sample preparation. Cultures were grown overnight in LB broth at 37 °C with gentle shaking and 500/tl samples were centrifuged. The cell pellet was resuspended in 500/tl TE ( 10 m M Tris-HC1 pH 8.0, 1 m M EDTA) to which an equal volume of phenol-chloroform ( 1 : 1 ) was added and the mixture was mixed thoroughly for at least 30 s. The phases were separated by centrifugation and the upper aqueous layer was extracted twice more. The aqueous phase was adjusted to 300 m M with sodium acetate and 2.5 vols of ice-cold ethanol was added to precipitate the nucleic acids. The nucleic acids were centrifuged and the pellet resuspended in 100 #ml TE containing RNase A ( 1/zg m l - l ). Dilutions of this preparation were made in TE and samples ( 1 ¢tl containing 200 ng DNA) added to the PCR. In later experiments, 1 /zl samples of the aqueous phase from the first phenol-chloroform extraction and 1 ¢tl samples from broth cultures and individual colonies picked from agar plates followed by homogenisation in 100/tl TE were used as substrates for the PCR.
Genetic methods. Samples ( 10 #1 ) from PCR mixtures were analysed by submarine gel electrophoresis ( 1.5% agarose; 4 V cm-1 ) and the amplified products were visualised after staining with ethidium bromide (0.5/~g m l - 1) in running buffer (40 m M Tris-HC1 pH 8.0, 5 m M sodium acetate, 1 m M EDTA). To confirm amplified products were the desired sequences, DNA species were transferred from agarose gels to nylon filters (Hybond-N, Amersham) by capillary action using the method of Southern (1975) followed by hybridisation with the relevant gene probe using methods previously described (Woodward et al., 1990). Concentrations of DNA were determined using DNA dipsticks (National Diagnostics) following manufacturer's recommendations. RESULTS
Design and evaluation of primer pairs Synthetic oligonucleotide primers were made according to the schedule shown in Table 1. Each primer pair was designed homologous with published sequences. However, for STIa and LTI primers, it is possible that sufficient homology exists with STIb and LTII, respectively, to give successful amplification. In the specific case of STIa primers there is a 4/21 mismatch in each primer with respect to the STIb sequence. The unavailability of confirmed STIb and LTII reference strains precluded the relevant tests to evaluate specificity. For VT2 detection by PCR, primer design took into consideration the four known sequences of the VT2 A and B sub-unit gene sequences (Ito et al., 1990). Primers VT2.1 and VT2.2 were designed homologous with highly conserved regions of the A subunit gene sequence and recognise VT2 (SLTII), VT2v (SLT.IIv), vtx2ha and vtx2hb. However, for our epidemiological purposes a test to differentiate VT2 from VT2v was needed. Primer VT2.3 was
255
DETECTION OF ENTERO- AND VEROCYTO-TOXIN GENES IN E. COLI
TABLE 2 Results of PCR amplification using target DNA from reference strains Reference strain
Toxin profile
LTI
STIa
VT 1
VT2
VT2v
VT2
all B44 09:K30:K99
STIa
-
+
G 7 08 : K 8 7 : K 8 8
LTI STII
+
.
E57 0138:K81
STIa VT2v
-
+
-
+
+
-
H19026:H19
VT1
-
-
+
-
-
-
E32511 0157:H-
VT2
-
-
-
+
-
+
6.7
5.5
1.5
8.0
5.5
3.0
optimum
F41
M g 2+ c o n c e n t r a t i o n s ( m M )
were
. .
.
.
.
.
. .
designed homologous with a highly conserved region of the A sub-unit sequence. This primer was to be used in conjunction with primers VT2.4 and VT2.5 which were designed homologous with heterologous regions just beyond the 3' coding region of the B subunit gene and were specific for VT2v and VT2 (including vtx2ha and vtx2hb) respectively. In the preliminary experiments, each primer pair was used in PCR experiments using 200 ng target DNA isolated from each of the control strains listed in methods. The Mg 2+ concentration was titrated for each primer pair to give o p t i m u m amplification under the thermal cycling conditions described previously. The results of preliminary evaluation and Mg 2+ titration experiments are given in Table 2. For each primer pair, a single DNA species of the predicted size was obtained from known positive reference strains only. Southern hybridisation experiments confirmed that the amplified products were homologous with the relevant toxin specific gene probe (data not shown). To confirm specificity of each primer pair, 200 ng of DNA extracted from each of Salmonella typhimurium, S. enteritidis, S. dublin, Klebsiella pneumoniae, Citrobacter freundii, Proteus vulgaris, E. cloacae and E. coli K12 were used as targets in PCR experiments using optimised Mg 2+ concentrations. In no experiments, were any DNA species produced whereas the relevant positive control did produce the predicted product.
PCR ofE. coli isolates The next objective was to compare the PCR tests with alternative methods of toxin or toxin gene detection. A panel of E. coli isolates from diarrhoeal disease in animals had been collected by this laboratory during 1989 and toxins identified by gene probe and immunological tests (Carroll et al., 1990; Woodward et al., 1990; Woodward and Wray, 1990). A selection of these characterised strains (55 STIa positive, 39 LT positive, 2 VT1 positive, 48 VT2 positive and 41 lacking any known toxin ) were used for PCR. DNA was extracted as described previously and neat samples and 10- 2 dilutions ( 1 ,ul ) were used in PCR experiments. The results are shown in Table 3 and exam-
256
M.J. WOODWARD ET AL.
TABLE 3 Detection o f e n t e r o t o x i n s LTI a n d STIa a n d v e r o c y t o t o x i n s V T 1, V T 2 a n d V T 2 v by P C R O K antigen serogroup (if known)
F i m b r i a l A n i m a l a N u m b e r T o x i n profile a d h e s i o n source tested b y gene p r o b e
STIa LT VTI
08 08 09 09 020 074 064 064 073 0101 0101 0112 0138 0141 u u u u 08 045 0147 0149 0t49 0149
K99 K99 K99 K99 K99 K99 K99 K99 K88 K88 K88 K88 K88 K88 F41 K88 K88 -
2 4 1 1 4 3 1 I l 12 2 2 6 2 2 6 4 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 3 11 14 1 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
7 0 0 0 0 0 0 0 0 0 0
0 1 0 i 0 0 0 0 1 0 0
0 I 2 0 1 23 1 18 0 1 0
0 1 0 0 0 0 1 18 0 0 0
0 0 2 0 1 23 0 0 0 1 0
0157 018 026 026 065 0138 0149 u u u Various b
K85ab
KI01 K'VI42' K'VI42' K'VI42' K28 K30 K81 K85
K87 K'E65' K89 K91 K91 K91 K'VI7'
K'M326/81' K81 K91
-
bov por por por por por por por por bov pot por por boy por por por por por por pot pot pot por
2 4 1 1 4 3 1 1 1 12 2 2 6 2 2 6 4 1 2 3 11 14 1 1
STIa STIa STII STla STII STIa STII VT2 STIa STIa STIa STIa STII STIa STIa STIa STIa STIa STII V T 2 STIa STIa STIa STIa STII STIa STII LT STII LT STII LT STII ITSTII LT STII LT
pot bov pot bov por pot ovi bov por por por/bov
7 1 2 1 1 23 1 18 1 1 41
LT STII VTIVT2 VT2 VTI VT2 VT2 VT2 VT2 VTI VT2 None
VT2 (all)
VT2 V T 2 v
"E. coliisolates tested were o f b o v i n e ( b o v ) p o r c i n e ( p o r ) a n d o v i n e ( o v i ) origin. bNegative control isolates lacking toxins o f porcine origin belonged to 0 serogroups 01 ( 1 ), 04 ( 3 ), 06 ( 2 ) , 035 ( 1 ), 071 ( 2 ) , 078 ( 1 ), 088 ( 2 ) , 0 1 3 9 ( 1 ) a n d u n t y p a b l e ( 12 ) a n d o f b o v i n e origin belonged to 0 s e r o g r o u p s 06 ( 1 ), 0 8 ( 2 ) , 0 2 5 ( 1 ) , 0 7 8 ( 1 ) , 0 1 0 1 ( 3 ) , 0 1 1 5 ( 1 ) , 0 1 5 7 ( 1 ) a n d u n t y p a b l e ( 6 ) ; where the n u m b e r s in brackets are the n u m b e r s o f isolates tested.
ples of visualisation o f amplified products are shown in Fig. 1 a and b. The correlation between toxin profile as determined by gene probe and by PCR was 100%. Furthermore, differentiation between VT2 and VT2v was achieved. All of 34 porcine derived E. coli isolates determined to be VT2 + by gene probe encoded VT2v (SLTIIv). No isolate encoded both polymorphs o f VT2.
DETECTIONOF ENTERO-AND VEROCYTO-TOXINGENESIN E. w
I
2
3
w
4
5
6w
1
257
COLI
W
2
3 4 5
6
W
1636
251
Detection of entero- and verocyto-toxin genes in Escherichia coli from diarrhoeal disease in animals using the polymerase chain reaction M.J. Woodward a, P.J. Carroll b and C. W r a y b aMolecular Genetics Unit and Department of Bacteriology, New Haw, Weybridge, UK bCentral Veterinary Laboratory, New Haw, Weybridge, UK (Accepted 16 October 1991 )
ABSTRACT Woodward, M.J. Carroll, P.J. and Wray, C., 1992. Detection of entero- and verocyto-toxin genes in Escherichia coli from diarrhoeal disease in animals using the polymerase chain reaction. Vet. Microbiol., 31: 251-261. Oligonucleotide primers were designed for the specific polymerase chain reaction (PCR) amplification of the enterotoxins STIa and LTI and of the verocytotoxins VT 1 and VT2. All of 184 E. coli isolates from cases of diarrhoea from pigs, cattle and sheep gave identical toxin profiles by PCR and gene probe. Differentiation between VT2 and VT2v was achieved using two oligonucleotide primers pairs in PCR and showed that all of 34 VT2 + porcine isolates, of which 23 were 0138: K1, harboured VT2v whereas 20 VT2 + bovine and ovine isolates harboured VT2. No isolate harboured both VT2 polymorphs. Simplified methods for sample preparation for PCR were examined and showed that PCR was not inhibited by direct addition of broth culture to the reaction mixture.
INTRODUCTION
Although Escherichia coli is recognised as a normal member of the lower gut bacterial flora, some strains are potentially pathogenic (Sussman, 1985 ). Enterotoxigenic E. coli (ETEC) is an important cause of diarrhoeal disease in man and animals (Sojka, 1971; Sack, 1975; Robins-Browne, 1985 ). The heat-labile toxins LTI and LTII, and the heat stable toxins STIa, STIb and STII, are well characterised functionally, antigenically and genetically (for review see Sussman, 1985 ). Verocytotoxin-producing E. coli (VTEC) are associated with sporadic diarrhoea, haemorrhagic colitis, haemolytic uraemic syndrome and thrombocytopaenic purpurea in man (Karmali, 1989 ). In animals, VTEC have been shown to be associated with diarrhoea and oedema disease (Linggood and Thompson, 1987 ). The verocytotoxins, of which three antigenically distinct types have been defined as VT 1, VT2 and VT2 variants, VT2v, vtx2ha and vtx2hb have been characterised functionally and geneti0378-1135/92/$05.00
© 1992 Elsevier Science Publishers B.V. All rights reserved.
252
M.J, WOODWARD ET AL.
cally (for reviews see Smith and Scotland, 1988; Karmali, 1989; Ito et al, 1990). The VT2v toxin has been shown to produce oedema disease in pigs (MacLeod et al., 1991 ). There is a need to develop rapid, specific and sensitive methods for the detection of toxin-producing isolates orE. coli because of their clinical significance. Conventional techniques for toxin detection require either bioassays for ST (Dean et al., 1972; Smith and Halls, 1967) or cell culture assays for LT and VT (Guerrant et al., 1974; Sack and Sack, 1975; Konowalchuk et al., 1977 ) and are time-consuming and unsuitable for routine diagnosis. The use of gene probes to detect toxins is a well established technology but not used routinely because considerable safety precautions are required when radioisotopic labels are used. Immunological methods for the detection of LT and ST toxins by latex agglutination and EIA tests are available commercially and in routine use (Carroll et al., 1990). However a rapid, single method which can be used for the detection of all E. coli toxins is desirable. The polymerase chain reaction (PCR) (Saiki et al., 1988) involves the enzymic amplification ofa DNA sequence initiated from a pair of short DNA fragments (primers) which bind either side of the chosen target sequence. The amplified DNA product of this reaction can be easily detected by gel electrophoresis followed by ethidium bromide staining and UV transillumination. The PCR is rapid, sensitive and highly specific and has been developed to detect enterotoxin genes (Olive, 1989; Furrer et al, 1990; Victor et al., 1991 ) and verocytotoxin genes (Pollard et al., 1990; Johnson et al., 1990; Kobayashi et al., 1990) and gives results within a working day. Sophisticated developments of the PCR have been used to simultaneously detect several virulence genes within a single reaction (Fraenkel et al., 1989). We report here our initial findings on the development of a simple protocol for PCR tests for LT, ST 1a, VT 1 VT2 and VT2v as a preliminary to its use for clinical samples. MATERIALS AND METHODS
Bacterial strains E. coli strains isolated from cattle, pigs and sheep suffering from enteric disease were stored on Dorset egg slopes (PM5, Oxoid) at room temperature. All isolates were sub-cultured every six months. Serotyping was by standard methods (Sojka, 1965 ) and detection of adhesins was by latex agglutination (Thorns et al., 1989) using FIMBREX (CVL reagents) following manufacturer's recommendations. Detection of toxin genes by D N A - D N A hybridisation followed the methods of Woodward et al., (1990). A panel of reference strains which produced toxins as assessed by standard methods have been described (Sack and Sack, 1975; Nagy et al, 1976; Moseley et al., 1980; Willshaw et al., 1987; Woodward et al, 1990) and were used as positive controls
DETECTION OF ENTERO- AND VEROCYTO-TOXIN GENES IN E. COLI
253
in PCR experiments; the control strains were B44 09:K30:K99 F41 STIa, G7 08:K87:K88 LT STII, E57 0138:K81 STIa VT2v, H19 026:H19 VT1 and E3511 0 1 5 7 : H - VT2. Cultures were grown in Luria-Bertani 'LB' broth (Maniatis et al., 1982). Polymerase chain reaction assay
Oligonucleotide primers design. Primers were designed following the principles of Innis and Gelfand (1990). Primers were synthesised by an ABI PCRmate D N A synthesiser (PCR-MATE EP 391, Applied Biosystems Inc) following manufacturer's recommendations and are listed in Table 1. (b) Polymerase chain reaction. The reaction mixtures (50/tl) comprised 50 mM KC1, 10 mM Tris-HC1 pH 8.4, 0.01% (w/v) gelatin, MgC12 titrated in the range 1.5 to 8.0 mM; nucleotides 200/~M for each of dATP, dCTP, dGTP and dTTP (Pharmacia); 0.2/tM for each primer; and 2.5 U ofthermostable DNA polymerase (amplitaq, Perkin Elmer Cetus ). The reaction mixtures were overlaid with an equal volume of mineral oil. Samples were pipetted through the mineral oil into the mixture. Parameters for amplification were denaturation at 94°C for 2 min for 1 cycle followed by 30 cycles of 94°C for 1.5 min, annealing at 59°C for 1.5 min, extension at 72°C for 2 min, followed by a final extension by incubation at 72°C for 5 min. A reagent blank (negative control) was as described above but without addition of sample. TABLE1
Nucleotide Sequences of Synthetic Oligonucleotide Primers Target gene
LTI
Primer code
LT. 1 LT.2 STIa Sta. 1 STa. 2 VT1 VT1.1 VT 1.2 VT2al 1 VT2.1 VT2.2 VT2v VT2.3 VT2.4 VT2 VT2-3 VT2-5
Primer sequence 5'-3'
Primer location ( b p ) in published sequence
Size of amplified product (bp)
Source of DNA sequence used for primer design
ATTTACGGCGTTACTATCCTC TTTTGGTCTCGGTCAGATATG TCTTTCCCCTCTTTTAGTCAG ACAGGCAGGATTACAACAAAG TTAGACTTCTCGACTGCAAAG TGTTGTACGAAATCCCCTCTG CTATATCTGCGCCGGGTCTG AGACGAAGATGGTCAAAACG CCGTCAGGACTGTCTGAAAC GGACGCGATAATTAAACCG CCGTCAGGACTGTCTGAAAC GAGTCTGACAGGCAACTGTC
27-48 286-307 39-60 184-205 239-260 748-769 183-203 490-510 838-858 1481-1501 838-858 1544-1564
280
Olive ( 1989 )
166
Moseley et al. ( 1983 ) Strockbine et al. ( 1988 ) Ito et al. (1990) Ito el al. ( 1990 ) Ito et al. ( 1990 )
530 327 663 726
254
M.J. WOODWARD ET AL.
Sample preparation. Cultures were grown overnight in LB broth at 37 °C with gentle shaking and 500/tl samples were centrifuged. The cell pellet was resuspended in 500/tl TE ( 10 m M Tris-HC1 pH 8.0, 1 m M EDTA) to which an equal volume of phenol-chloroform ( 1 : 1 ) was added and the mixture was mixed thoroughly for at least 30 s. The phases were separated by centrifugation and the upper aqueous layer was extracted twice more. The aqueous phase was adjusted to 300 m M with sodium acetate and 2.5 vols of ice-cold ethanol was added to precipitate the nucleic acids. The nucleic acids were centrifuged and the pellet resuspended in 100 #ml TE containing RNase A ( 1/zg m l - l ). Dilutions of this preparation were made in TE and samples ( 1 ¢tl containing 200 ng DNA) added to the PCR. In later experiments, 1 /zl samples of the aqueous phase from the first phenol-chloroform extraction and 1 ¢tl samples from broth cultures and individual colonies picked from agar plates followed by homogenisation in 100/tl TE were used as substrates for the PCR.
Genetic methods. Samples ( 10 #1 ) from PCR mixtures were analysed by submarine gel electrophoresis ( 1.5% agarose; 4 V cm-1 ) and the amplified products were visualised after staining with ethidium bromide (0.5/~g m l - 1) in running buffer (40 m M Tris-HC1 pH 8.0, 5 m M sodium acetate, 1 m M EDTA). To confirm amplified products were the desired sequences, DNA species were transferred from agarose gels to nylon filters (Hybond-N, Amersham) by capillary action using the method of Southern (1975) followed by hybridisation with the relevant gene probe using methods previously described (Woodward et al., 1990). Concentrations of DNA were determined using DNA dipsticks (National Diagnostics) following manufacturer's recommendations. RESULTS
Design and evaluation of primer pairs Synthetic oligonucleotide primers were made according to the schedule shown in Table 1. Each primer pair was designed homologous with published sequences. However, for STIa and LTI primers, it is possible that sufficient homology exists with STIb and LTII, respectively, to give successful amplification. In the specific case of STIa primers there is a 4/21 mismatch in each primer with respect to the STIb sequence. The unavailability of confirmed STIb and LTII reference strains precluded the relevant tests to evaluate specificity. For VT2 detection by PCR, primer design took into consideration the four known sequences of the VT2 A and B sub-unit gene sequences (Ito et al., 1990). Primers VT2.1 and VT2.2 were designed homologous with highly conserved regions of the A subunit gene sequence and recognise VT2 (SLTII), VT2v (SLT.IIv), vtx2ha and vtx2hb. However, for our epidemiological purposes a test to differentiate VT2 from VT2v was needed. Primer VT2.3 was
255
DETECTION OF ENTERO- AND VEROCYTO-TOXIN GENES IN E. COLI
TABLE 2 Results of PCR amplification using target DNA from reference strains Reference strain
Toxin profile
LTI
STIa
VT 1
VT2
VT2v
VT2
all B44 09:K30:K99
STIa
-
+
G 7 08 : K 8 7 : K 8 8
LTI STII
+
.
E57 0138:K81
STIa VT2v
-
+
-
+
+
-
H19026:H19
VT1
-
-
+
-
-
-
E32511 0157:H-
VT2
-
-
-
+
-
+
6.7
5.5
1.5
8.0
5.5
3.0
optimum
F41
M g 2+ c o n c e n t r a t i o n s ( m M )
were
. .
.
.
.
.
. .
designed homologous with a highly conserved region of the A sub-unit sequence. This primer was to be used in conjunction with primers VT2.4 and VT2.5 which were designed homologous with heterologous regions just beyond the 3' coding region of the B subunit gene and were specific for VT2v and VT2 (including vtx2ha and vtx2hb) respectively. In the preliminary experiments, each primer pair was used in PCR experiments using 200 ng target DNA isolated from each of the control strains listed in methods. The Mg 2+ concentration was titrated for each primer pair to give o p t i m u m amplification under the thermal cycling conditions described previously. The results of preliminary evaluation and Mg 2+ titration experiments are given in Table 2. For each primer pair, a single DNA species of the predicted size was obtained from known positive reference strains only. Southern hybridisation experiments confirmed that the amplified products were homologous with the relevant toxin specific gene probe (data not shown). To confirm specificity of each primer pair, 200 ng of DNA extracted from each of Salmonella typhimurium, S. enteritidis, S. dublin, Klebsiella pneumoniae, Citrobacter freundii, Proteus vulgaris, E. cloacae and E. coli K12 were used as targets in PCR experiments using optimised Mg 2+ concentrations. In no experiments, were any DNA species produced whereas the relevant positive control did produce the predicted product.
PCR ofE. coli isolates The next objective was to compare the PCR tests with alternative methods of toxin or toxin gene detection. A panel of E. coli isolates from diarrhoeal disease in animals had been collected by this laboratory during 1989 and toxins identified by gene probe and immunological tests (Carroll et al., 1990; Woodward et al., 1990; Woodward and Wray, 1990). A selection of these characterised strains (55 STIa positive, 39 LT positive, 2 VT1 positive, 48 VT2 positive and 41 lacking any known toxin ) were used for PCR. DNA was extracted as described previously and neat samples and 10- 2 dilutions ( 1 ,ul ) were used in PCR experiments. The results are shown in Table 3 and exam-
256
M.J. WOODWARD ET AL.
TABLE 3 Detection o f e n t e r o t o x i n s LTI a n d STIa a n d v e r o c y t o t o x i n s V T 1, V T 2 a n d V T 2 v by P C R O K antigen serogroup (if known)
F i m b r i a l A n i m a l a N u m b e r T o x i n profile a d h e s i o n source tested b y gene p r o b e
STIa LT VTI
08 08 09 09 020 074 064 064 073 0101 0101 0112 0138 0141 u u u u 08 045 0147 0149 0t49 0149
K99 K99 K99 K99 K99 K99 K99 K99 K88 K88 K88 K88 K88 K88 F41 K88 K88 -
2 4 1 1 4 3 1 I l 12 2 2 6 2 2 6 4 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 3 11 14 1 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
7 0 0 0 0 0 0 0 0 0 0
0 1 0 i 0 0 0 0 1 0 0
0 I 2 0 1 23 1 18 0 1 0
0 1 0 0 0 0 1 18 0 0 0
0 0 2 0 1 23 0 0 0 1 0
0157 018 026 026 065 0138 0149 u u u Various b
K85ab
KI01 K'VI42' K'VI42' K'VI42' K28 K30 K81 K85
K87 K'E65' K89 K91 K91 K91 K'VI7'
K'M326/81' K81 K91
-
bov por por por por por por por por bov pot por por boy por por por por por por pot pot pot por
2 4 1 1 4 3 1 1 1 12 2 2 6 2 2 6 4 1 2 3 11 14 1 1
STIa STIa STII STla STII STIa STII VT2 STIa STIa STIa STIa STII STIa STIa STIa STIa STIa STII V T 2 STIa STIa STIa STIa STII STIa STII LT STII LT STII LT STII ITSTII LT STII LT
pot bov pot bov por pot ovi bov por por por/bov
7 1 2 1 1 23 1 18 1 1 41
LT STII VTIVT2 VT2 VTI VT2 VT2 VT2 VT2 VTI VT2 None
VT2 (all)
VT2 V T 2 v
"E. coliisolates tested were o f b o v i n e ( b o v ) p o r c i n e ( p o r ) a n d o v i n e ( o v i ) origin. bNegative control isolates lacking toxins o f porcine origin belonged to 0 serogroups 01 ( 1 ), 04 ( 3 ), 06 ( 2 ) , 035 ( 1 ), 071 ( 2 ) , 078 ( 1 ), 088 ( 2 ) , 0 1 3 9 ( 1 ) a n d u n t y p a b l e ( 12 ) a n d o f b o v i n e origin belonged to 0 s e r o g r o u p s 06 ( 1 ), 0 8 ( 2 ) , 0 2 5 ( 1 ) , 0 7 8 ( 1 ) , 0 1 0 1 ( 3 ) , 0 1 1 5 ( 1 ) , 0 1 5 7 ( 1 ) a n d u n t y p a b l e ( 6 ) ; where the n u m b e r s in brackets are the n u m b e r s o f isolates tested.
ples of visualisation o f amplified products are shown in Fig. 1 a and b. The correlation between toxin profile as determined by gene probe and by PCR was 100%. Furthermore, differentiation between VT2 and VT2v was achieved. All of 34 porcine derived E. coli isolates determined to be VT2 + by gene probe encoded VT2v (SLTIIv). No isolate encoded both polymorphs o f VT2.
DETECTIONOF ENTERO-AND VEROCYTO-TOXINGENESIN E. w
I
2
3
w
4
5
6w
1
257
COLI
W
2
3 4 5
6
W
1636
