Bovine Non-O157 Shiga Toxin 2-Containing Escherichia coli Isolates Commonly Possess stx_2-EDL933 and/or stx_2vhb Subtypes

Abstract

stx₂ genes from 138 Shiga toxin-producing Escherichia coli (STEC) isolates, of which 127 were of bovine origin (58 serotypes) and 11 of human origin (one serotype; O113:H21), were subtyped. The bovine STEC isolates from Australian cattle carried ehxA and/or eaeA and predominantly possessed stx_2-EDL933 (103 of 127; 81.1%) either in combination with stx_2vhb (32 of 127; 25.2%) or on its own (52 of 127; 40.4%). Of 22 (90.9%) bovine isolates of serotype O113:H21, a serotype increasingly recovered from patients with hemolytic uremic syndrome (HUS) or hemorrhagic colitis, 20 contained both stx_2-EDL933 and stx_2vhb; 2 isolates contained stx_2vhb only. Although 7 of 11 (63.6%) human O113:H21 isolates associated with diarrhea possessed stx_2-EDL933, the remaining 4 isolates possessed a combination of stx_2-EDL933 and stx_2vhb. Three of the four were from separate sporadic cases of HUS, and one was from an unknown source.

Shiga toxin-producing Escherichia coli (STEC) is commonly carried by ruminants (especially cattle) and is an important enteropathogen, causing human diseases ranging from mild diarrhea to more severe conditions such as hemorrhagic colitis (HC) and hemolytic uremic syndrome (HUS) (7, 15, 22a, 25, 28, 47). Although over 150 serotypes of STEC have been associated with human disease (12, 20), serotype O157:H7 is believed to be responsible for the majority of outbreaks and sporadic cases of HUS in the United States, Japan, and Europe. However, non-O157 STEC is of greater significance in Australia (16) and Argentina (35), and its role in HUS and HC in the United States and Europe is gaining recognition (24).

The most critical factors of STEC associated with severe human disease are the Shiga toxins. STEC can produce one or both of these groups of toxins which comprise two immunologically non-cross-reactive groups (designated Stx1 and Stx2), although other virulence factors such as intimin and enterohemolysin either directly contribute to (or are implicated by association with) the pathogenicity of STEC (11, 21).

stx₂ comprises 11 distinct variants (5, 19, 23, 31, 38, 39, 42, 45, 46, 50) and is considered the most important STEC virulence factor associated with human disease (11, 21, 34, 41). More importantly, differences in the degree of pathogenicity of STEC serotypes have been associated with variations in the stx₂ subtype (27, 29, 30, 42). The most frequently reported stx₂ subtypes are represented by stx_2c, stx_2d, and stx_2e (42, 46, 50).

Although more than 200 different O:H serotypes of STEC have been isolated from cattle (1, 7, 12, 20, 24, 49, 51), there is a paucity of information regarding associations between serotype, stx₂ subtype, and virulence factor profiles among STEC isolates from meat-producing animals (9, 10, 22a, 26, 33, 43). stx₂ subtypes stx_2-EDL933 and stx_2c (stx_2vha and stx_2vhb) have been described in European studies of bovine STEC; however, their association with serotype has not been reported previously (8, 42, 46). Ramachandran et al. (44) demonstrated that stx₂-containing STEC from ovine feces usually belonged to the stx_2d subtypes (stx_{2d-Ount/O111/OX3a}).

The purpose of this study was to subtype stx₂ genes among a serologically diverse collection of 138 stx₂-containing STEC isolates primarily derived from bovine feces collected from the eastern states of Australia and to compare stx₂ subtypes of STEC isolates of serotype O113:H21 (an important serotype increasingly associated with HUS and HC in humans) of bovine and human origin.

A total of 138 STEC isolates, of which 134 were non-O157, were used in this study. These consisted of 127 isolates from cattle and 11 human isolates from individual cases of clinical infections (Table 1). All isolates were prepared and subjected to multiplex PCR for the detection of STEC virulence factors stx₁, stx₂, ehxA, and eaeA as described previously (14). Amplified PCR products were then resolved by gel electrophoresis through agarose (2% wt/vol) and stained with ethidium bromide (5 μl/ml). Visualization was undertaken using UV illumination, and images were captured using a GelDoc 1000 image analysis station (Bio-Rad).

TABLE 1.

Virulence factor profiles and stx₂ subtypes of STEC of bovine and human origin

Source	Serotype	No. of isolates of animal or human origin	Virulence factor				No. of isolates containing stx2 variant(s)
			stx1	stx2	ehxA	eaeA	stx2-EDL.933	stx2vhb	stx2d-Ount	stx2-EDL933/stx2vhb	stx2-EDL933/stx2vhb/stx1	stx2vhb/stx1	stx2-EDL933/stx1	stx2d-OX3a
Ba	O2:H8	1	−	+	+	−		1
B	O3:H7	2	+	+	+	−					1	1
B	O5:H−	2	+	+	+	−								2
B	O5:H7	2	+	+	+	−						1	1
B	O6:H8	1	−	+	+	−				1
B	O6:H28	1	−	+	+	−				1
B	O6:H34	3	−	+	+	−	3
B	O8(KA):H51	1	−	+	+	−	1
B	O8:H16	1	+	+	+	−					1
B	O8:H19	6	−	+	+	−	6
BDb	O21:H21	1	−	+	+	−	1
B	O28:H8	3	−	+	+	−	3
B, BD	O28:H40	2	−	+	+	−	2
B	O49:H−	1	−	+	+	+				1
B	O53:H2	1	−	+	+	−	1
B	O76:H7	1	−	+	+	+	1
B	O81:H21	1	−	+	+	−	1
B	O82:H8	5	+	+	+	−					1		4
B	O82:H40	2	+	+	+	−							2
B	O91:H−	1	+	+	+	−			1
B	O91:H21	1	−	+	+	+								1
B	O93:H19	1	−	+	+	−	1
B	O104:H7	1	−	+	+	−	1
B	O108:H7	2	+	+	+	−						2
B	O110:H40	1	+	+	+	−							1
B	O111:H−	1	−	+	+	+							1
B	O111:H−	1	+	+	+	+							1
BD	O111:H8	1	+	+	+	+							1
BD	O113:H−	1	−	+	+	−				1
B BD	O113:H21	20	−	+	+	−				20
B	O113:H21	2	+	+	−	−		2
Hc (HUS, HAd)	O113:H21	3	−	+	+	−				3
H (Diarrhea, HUS)	O113:H21	6	−	+	+	−	6
H (?)	O113:H21	1	−	+	−	−	1
H (?)	O113:H21	1	−	+	−	−				1
B	O116:H21	1	−	+	+	−	1
B	O130:H11	1	−	+	+	−							1
B	O130:H11	6	−	+	+	−	6
B	O130:H38	2	+	+	+	−						2
B	O141:H49	1	−	+	+	−	1
B	O153:H8	1	+	+	+	−			1
B	O153:HR	2	+	+	+	−							2
B, BD	O157:H8	2	−	+	+	−	2
B	O157:H−	1	+	+	+	+						1
B	O157:H7	1	−	+	+	+		1
B	O163:H−	1	+	+	+	−					1
B	O163:H19	1	−	+	+	−	1
B	O163:H19	1	+	+	+	−						1
B	Ont:H2	2	−	+	+	−	2
B	Ont:H5	1	−	+	+	−	1
B	Ont:H7	2	−	+	+	−	1			1
B	Ont:H8	1	+	+	+	−						1
B	Ont:H8	3	−	+	+	−	3
B	Ont:H11	2	+	+	+	−						1	1
B	Ont:H11	2	−	+	+	−		2
BD	Ont:H11	1	−	+	+	−		1
B	Ont:H16	1	+	+	+	−					1
B	Ont:H21	1	+	+	+	−							1
B	Ont:H28	1	−	+	+	−	1
B	Ont:H28	1	+	+	+	−							1
B	Ont:H30	1	+	+	+	−							1
B	Ont:H41	1	+	+	+	−							1
B	Ont:H49	3	−	+	+	−	2			1
B	Ont:H49	1	+	+	+	−	1
B	Ont:HR	1	−	+	+	−	1							/PICK>
B	OR:H−	1	−	+	+	−		1
B	OR:H3	1	−	+	+	−	1
B	OR:H8	1	−	+	+	−				1
B	OX3:H8	5	−	+	+	−	5
B	OX3:H8	1	+	+	+	−						1
B	OX3:H21	1	−	+	−	+	1
B	OX3:H40	1	−	+	+	−	1
Total		138					59	8	2	31	5	11	19	3

^aB, feces from healthy cattle unless otherwise specified.

^bBD, diagnostic fecal samples from cattle.

^cH, human feces.

^dHA, hemolytic anemia.

Bovine STEC isolates with stx₂ were subjected to subtyping using typing schemes previously reported (4, 42, 48). For this study, stx_2d subtypes are defined as nucleotide sequence variants of stx₂ (stx_2d-Ount, stx_2d-O111, and stx_2d-OX3a), as described previously (42), and are not the potentially mucin-activatable stx_2d variants described by Melton-Celsa et al. (30) (defined as stx_2vha and stx_2vhb in this study). stx₂ genes amplified with VT2-e and VT2-f primers and Lin F and Lin R primers were subjected to restriction fragment length polymorphism (RFLP) analysis with the enzymes HaeIII and PvuII (42) and the enzymes HincII and AccI (28), respectively. In circumstances in which a stx₂ gene could not be reliably subtyped using the above-described approach, the gene was amplified with Tyler F and Tyler R primers and subjected to RFLP analysis with the enzymes MspI, NciI, and RsaI (Table 2) (48). PCR products were digested for a minimum of 4 h at 37°C. Fragments were separated by electrophoresis through agarose gels (2% wt/vol), and subtypes were determined according to comparison with profiles reported previously (4, 42, 48).

TABLE 2.

Restriction fragment sizes used for analysis of stx₂

Primers used to amplify fragment	Restriction enzyme	Expected fragment size(s) for:				Reference
		stx2-EDL933	stx2vhb	stx2d-Ount	stx2d-OX3
VT2-e, VT2-f	HaeIII	348	216, 132	216, 132	167, 132, 49a	42
	PvuII	323, 25a	250, 73, 25a	200, 120, 28a	200, 120, 28a
Lin F, Lin R	HincII	555, 262, 62	555, 340	880, 15a	880, 15	28
	AccI	544, 351	544, 351	906	544, 351
Tyler F, Tyler R	MspI	232, 48a	108, 73, 51a		155, 125	48
	NciI	385	159, 126
	RsaI	216, 69	216, 69

^aFragment too small to visualize under the electrophoresis conditions used.

DNA sequence analysis of both the A and B subunits of two stx₂-containing STEC isolates with serotype O110:H40 and Ont:H11 were undertaken, since RFLP analyses confirmed these isolates possessed stx₂ subtypes representative of the majority of the collection. stx₂ was amplified using 10 pmol of each of the oligonucleotide primers Stx2F and Stx2R (Table 3). PCRs (50 μl) were carried out containing 2 μl of nucleic acid of a crude whole-cell DNA template prepared using Instagene matrix as described previously (17), 10 mM Tris-HCl (pH 8.3), 10 mM KCl, 2 mM MgCl₂, 200 μM each deoxynucleoside triphosphate, and 1 U of Taq DNA polymerase.

TABLE 3.

Primer pairs used to amplify stx₂

Primer	Sequence (5′ to 3′)	Size of fragment (bp)	Reference
stx2 typing
VT2-e	AATACATTATGGGAAAGTAATA	348	42
VT2-f	TAAACTGCACTTCAGCAAAT
Lin F	GAACGAAATAATTTATATGT	900a	28
Lin R	TTTGATTGTTACAGTCAT
Tyler F	AAGAAGATGTTTATGGCGGT	285	48
Tyler R	CACGAATCAGGTTATGCCTC
stx2 sequencing
StxF	TATCTGCGCCGGGTCT	1,280	44
StxR	CAAAKCCKGARCCTGA
Gannon F	CCATGACAACGGACAGCAGTT	779	18
Gannon R	CCTGTCAACTGAGCAGCACTTTG
Paton F	GGCACTGTCTGAAACTGCTCC	255	37
Paton R	TCGCCAGTTATCTGACATTCTG
ST F	AATGCAATGGCGG	200	This study
Stx2R	CAAATCCGGAGCCTGC
Stx2F2	AATCCAGTACAACGCGCCA	600	This study
ST R	AACGCAGAACTGCTCT
Tyler F	AAGAAGATGTTTATGGCGGT	285	48
Tyler R	CACGAATCAGGTTATGCCTC
KBStx2F	AATCCAGTACAACGCGCC	395	This study
KBStx2R	TTGCTGAATAATCAGACG

^aAmplicons differ by a few nucleotides depending on the variant.

Thermocycler steps involved an initial denaturation step (95°C for 5 min) followed by 35 cycles of denaturation (95°C for 30 s), annealing (60°C for 45 s), and extension (72°C for 90 s). A final extension step of 72°C for 5 min completed the PCR. The amplified PCR product was separated by 2% agarose gel electrophoresis. For DNA sequencing, PCR amplification products were purified using a QIAquick DNA purification kit (Qiagen, Hilden, Germany). Sequencing reactions were performed using the Big Dye terminator cycle sequencing ready reaction DNA sequencing kit and electrophoresed on an ABI prism 377 DNA sequencer (Perkin Elmer, Foster City, Calif.) as described previously (44). Auto Assembler software (Perkin Elmer) was used to compile and analyze the DNA sequences. Nucleotide and amino acid analysis was performed using programs accessed via the Australian National Genomic Information Service (www.angis.org.au). Sequences were compared with those deposited in public databases using the BlastN and BlastP algorithms (2).

All 138 isolates possessed stx₂ and comprised 58 serotypes. STEC serotypes and their virulence factor profiles are listed in Table 1. A significant proportion of bovine STEC isolates (32 of 127; 25.2%) possessed more than one stx₂ variant or a combination of an stx₂ variant(s) together with stx₁ (35 of 127; 27.6%). Isolates with two copies of stx₂ were shown to possess stx_2-EDL933 and stx_2vhb (31 of 127; 24.4%) (Table 1). STEC isolates possessing stx_2-EDL933 alone were also prevalent (59 of 138; 42.8%). These 59 isolates belonged to 28 serotypes (Table 4). Isolates with stx_2-EDL933 and stx₁ comprised the third most common group (19 of 138; 13.7%), representing 13 serotypes, followed by stx_2vhb and stx₁ (11 of 138; 7.9%), comprising 9 serotypes. STEC possessing the three Shiga toxin factors stx_2-EDL933, stx_2vhb, and stx₁ was identified with a frequency of 5 of 138 (3.6%) and was represented by 5 serotypes (Table 4). STEC containing stx_2vhb alone (8 of 138; 5.8%), stx_2d-Ount (2 of 138; 1.4%), or stx_2d-OX3a (3 of 138; 2.2%) was isolated infrequently (Table 4). Two bovine isolates of the serotype O5:H⁻, a serotype commonly recovered from ovine feces (14), possessed an stx_2d-OX3a subtype; this subtype is commonly associated with ovine STEC (44). Representative RFLP profiles produced by the digestion of PCR amplification products spanning different regions of stx₂ are depicted in Fig. 1 and 2. DNA sequence analysis of stx₂ derived from STEC serotype O110:H40 showed 99.8% sequence identity with sequences of stx_2-EDL933 (accession number Z37725.1) (36) and the p gene derived from phage 933W (13). The two nucleotide polymorphisms that differentiated the stx₂ sequence of O110:H40 with stx_2-EDL933 did not alter the predicted amino acid sequence. The stx₂ gene from an O110:H40 isolate was selected for DNA sequence analysis, because PCR-RFLP analysis indicated that it was indicative of stx₂ genes representative of the majority of bovine isolates used in this study (Fig. 1, lane 6). DNA sequence analysis of the stx₂ gene amplified from the Ont:H11 isolate showed 99.5% (5 nucleotide polymorphisms in the A subunit) sequence identity with stx_2vhb, a sequence variant of stx_2c (accession number X61283.1) (31). These data concur with the results of PCR-RFLP analyses used to type stx₂ variants in this study. Of the five nucleotide polymorphisms identified in the A-subunit sequence, only two polymorphisms resulted in a change in the amino acid sequence. The B subunit of stx_2vhb in Ont:H11 showed nucleotide and predicted amino acid sequences identical to those of stx_2vhb (accession number X61283.1) (31).

TABLE 4.

stx₂ variant(s) and association with serotype

stx2 variant(s)	No. (%) of isolatesa	Serotypes
stx2-EDL933	59 (42.8)	O6:H34, O8:(KA):H51, O8:H19, O21:H21, O28:H8, O28:H40, O53:H2, O76:H7, O81:H21, O93:H19, O104:H7, O113:H21, O116:H21, O130:H11, O141:H49, O157:H8, O163:H19, Ont:H2, Ont:H5, Ont:H7, Ont:H8, Ont:H28, Ont:H49, Ont:HR, OR:H3, OX3:H8, OX3:H21, OX3:H40
stx2-EDL933/stx2vhb	31 (32.6)	O6:H8, O6:H28, O49:H−, O113:H−, O113:H21, Ont:H7, Ont:H49, OR:H8
stx2-EDL933/stx1	19 (10.4)	O5:H7, O82:H8, O82:H40, O110:H40, O111:H−, O111:H8, O103:H11, O153:HR, Ont:H11, ONT:H21, Ont:H28, Ont:H30, Ont:H41
stx2vhb/stx1	11 (8.0)	O3:H7, O5:H7, O108:H7, O130:H38, O157:H−, O163:H19, Ont:H8, Ont:H11, OX3:8
stx2vhb	8 (5.8)	O2:H8, O113:H21, O157:H7, Ont:H11, OR:H−
stx2-EDL933/stx2vhb/stx1	5 (3.6)	O3:H7, O8:H16, O82:H8, O163:H−, Ont:H16
stx2d-OX3a	3 (2.2)	O5:H−, O91:H21
stx2d-Ount	2 (1.4)	O91:H−, O153:H8

^aValues are numbers and percentages compared to a total of 138 isolates.

FIG. 1.

HaeIII (A) and PvuII (B) digests of PCR products obtained with VT2-e and VT2-f primers and HincII (C) digests of PCR products obtained with Lin F and Lin R primers. Lanes: 1, 100-bp Plus marker; 2, stx_2d-Ount variant of serotype O91:H⁻; 3, stx_2-EDL933/stx_2vhb variant of serotype O113:H21; 4, stx_2-EDL933/stx_2vhb variant of serotype O113:H21; 5, stx_2-EDL933/stx_2vhb variant of serotype O113:H21; 6, stx_2-EDL933/stx_2vhb variant of serotype O110:H40; 7, stx_2-EDL933 variant of serotype Ont:H5; 8, stx_2vhb variant of serotype Ont:HR; 9, stx_2-OX3a variant of serotype O5:H⁻; 10, stx_2vhb variant of serotype Ont:H11.

FIG. 2.

MspI (A), NciI (B), and RsaI (C) digests of PCR products obtained with Tyler F and Tyler R primers. Lanes: 1, 100-bp Plus marker; 2, stx_2-EDL933/stx_2vhb variant of serotype O113:H21; 3, stx_2vhb variant of serotype Ont:HR; 4, stx_2-EDL933 variant of serotype Ont:H5; 5, stx_2-EDL933 variant of serotype Ont:H5.

In this study, we determined the stx₂ subtypes of 127 bovine STEC isolates belonging to 58 serotypes and showed that Australian bovine STEC which also possesses ehxA and/or eaeA predominantly possesses either stx_2-EDL933 or stx_2vhb subtypes (alone or in combination with one another). We also demonstrated that several serotypes, particularly O113:H21, simultaneously possess both these subtypes.

In a recent study of 167 stx₂-containing STEC isolates from healthy cattle in France, stx_2vhb, stx_2-EDL933, and stx_2vha subtypes were commonly identified (5). Our study identified stx_2-EDL933 and stx_2vhb as the predominant subtypes (122 of 127; 96.1%) among Australian bovine STEC isolates; stx_2vha-positive STEC isolates were not observed among any of STEC isolates in our study. Although the serotypes belonging to the stx₂-containing STEC isolates from France were not reported (5), a recent study of 186 STEC isolates from cattle during a 1-year study in France (43) showed that few isolates possessed serotypes in common with those identified in this study. Our data suggest that stx_2vha-positive STEC belongs to serotypes not commonly found in Australian cattle. Alternatively, stx_2vha may only be found in STEC isolates that lack the accessory virulence factors ehxA and/or eaeA (termed complex STEC [22]), since all but 4 of the 138 STEC isolates in this study possessed one or both of these factors. In support of the latter hypothesis, the study by Bertin et al. (5) identified enterohemolysin among STEC isolates harboring stx_2-EDL933 (78%) or a combination of stx_2-EDL933 and stx_2vhb (85%); only 6 of 43 (14%) stx_2vha-containing isolates also possessed ehxA, and the presence of eaeA was not reported. Furthermore, recent studies in our laboratories have identified stx_2vha-containing STEC in isolates that do not possess ehxA and/or eaeA (our unpublished data). STEC isolates simultaneously containing three stx₂ variants were not observed among the 138 Australian STEC isolates, unlike the results reported by Bertin et al. (5). Furthermore, the untypeable stx₂ variant (identified as stx_2-NV206) which comprised 24 of 167 (14.4%) of isolates in the study by Bertin et al. (5) was not observed in our Australian collection; 25% of STEC isolates possessing the stx_2-NV206 possessed ehxA (5). Collectively, stx₂-subtyping data suggest that different non-O157 STEC populations predominate in Australia and France. However, the fact that 134 of 138 (97.1%) of the Australian stx₂-containing STEC isolates and 54 of 130 (41.5%) of the stx₂-containing French isolates possessed ehxA and/or eaeA may bias comparisons between these two STEC populations. Further studies are required to validate these hypotheses.

Serotype O113:H21 is increasingly being isolated from patients with HUS (6, 17, 40) and was the most predominant serotype in stx₂-containing STEC isolates from healthy cattle in Australia (22a). Studies conducted in France (43), Japan (26), and Spain (9, 10, 33) all indicated that O113:H21 is among the most prevalent STEC serotypes recovered from cattle feces. Furthermore, O113:H21 was identified in 10 of 41 (24.4%) STEC isolates recovered from a Canadian study of naturally contaminated beef (3). All of the O113:H21 isolates recovered in our study (22 bovine and 11 human) were eaeA negative; 29 of 33 (87.9%) possessed the ehxA gene, and only 2 of 33 (6.1%) possessed stx₁. Of 22 (90.9%) bovine isolates of serotype O113:H21, 20 concomitantly possessed both stx_2-EDL933 and stx_2vhb while 2 (9.1%) only possessed stx_2vhb. Two O113:H21 isolates recovered from HUS patients and a third isolate from a patient with anemia also simultaneously possessed stx_2-EDL933 and stx_2vhb. However, four isolates of O113:H21 recovered from human cases of diarrhea, two from patients with HUS, and one isolate from an asymptomatic human individual possessed stx_2-EDL933 only. Of the 58 different serotypes identified in our study, 7 serotypes have been reported to cause HUS (O5:H⁻, O91:H⁻, O91:H21, O111:H⁻, O111:H8, O157:H7, and O163:H19).

Ramachandran et al. (44) reported that ovine STEC isolates that also contain ehxA and/or eaeA predominantly possess stx_2d (stx_2d-Ount, stx_2d-OX3a, and stx_2dO111) subtypes. In contrast, the cattle STEC isolates described in this study rarely possessed stx_2d subtypes (5 of 138; 3.6%) and three of these isolates possessed STEC serotypes (O5:H⁻ and O91:H⁻) typically found in sheep. Furthermore, none of the ovine isolates examined in the study of Ramachandran et al. (44) contained stx_2vhb and only a single isolate possessed stx_2EDL933, highlighting a dramatic contrast with results reported in the present study of bovine STEC. Bertin et al. (5) showed that 167 bovine stx₂-positive STEC isolates from healthy cattle rarely possessed an stx_2d subtype (8.5% of 167), but their serotype(s) was not reported. Beutin et al. (8) also found that with the exception of a single serotype (O90:H24), cattle and sheep differed with respect to the O:H types of their STEC floras. Together, these and other studies (14, 22, 22a, 32) suggest that genetically and serologically different populations of STEC inhabit the gastrointestinal tracts of cattle and sheep.