A New Shiga Toxin 2 Variant (Stx2f) from Escherichia coli Isolated from Pigeons

Abstract

We have isolated Shiga toxin (Stx)-producing Escherichia coli (STEC) strains from the feces of feral pigeons which contained a new Stx2 variant gene designated stx_2f. This gene is most similar to sltIIva of patient E. coli O128:B12 isolate H.I.8. Stx2f reacted only weakly with commercial immunoassays. The prevalence of STEC organisms carrying the stx_2f gene in pigeon droppings was 12.5%. The occurrence of a new Stx2 variant in STEC from pigeons enlarges the pool of Stx2 variants and raises the question whether horizontal gene transfer to E. coli pathogenic to humans may occur.

Infections with Shiga toxin (Stx)-producing Escherichia coli (STEC) can cause a broad spectrum of disease ranging from mild diarrhea to hemorrhagic colitis, the latter of which can progress to the hemolytic uremic syndrome (6). Stx are thought to be the major pathogenicity factor of these organisms and comprise a family composed of Stx1, Stx2, and variants of Stx2 (10) which can be found in STEC strains isolated from either humans or animals (5, 7, 12, 13, 16). Members of the Stx family are differentiated by their biological activity or presence or lack of cross-reactivity with Stx1 or Stx2 antisera, and some of the variants are characterized by a lower binding affinity for the glycolipid receptor Gb3 (compared with that for Stx2) caused by amino acid exchanges in the B-subunit of the protein (10). The most prominent Stx2 variants are Stx2c (16), Stx2d (13), and Stx2e (18), the last of which is mainly found in isolates causing edema disease in pigs.

The intestinal tract of ruminants, in particular cattle, have long been regarded as the principal reservoir of STEC, particularly those belonging to serogroup O157, the main cause of illness in humans (6). STEC have also been isolated from other domestic animals, such as pigs, dogs, and cats (1), and recently from the feces of wild birds. Wallace et al. (17) reported the isolation of STEC O157 from gull droppings collected at an urban landfill site and at an intertidal area. Moreover, STEC strains possessing the intimin-coding eae gene associated with the ability to cause attaching-and-effacing lesion in the intestine of the host have been isolated from a relevant proportion of feral pigeons examined in the city of Rome, Italy (3). In this paper, we analyze the toxins produced by eae-positive STEC strains isolated from feral pigeons in both Germany and Italy and describe the characterization of a new Stx2 variant produced by such isolates.

E. coli control strains used in this study are described in Tables 1 and 2. E. coli O128:B12 strain H.I.8 was kindly provided by V. P. J. Gannon, Food Directorate Health Canada, Animal Diseases Research Institute, Lethbridge, Alberta, Canada. Plasmid pACYC184 was used as a cloning vector.

TABLE 1.

Results of PCR amplification of stx_2f from five E. coli strains from pigeons and different stx alleles from STEC reference strains with a panel of stx-specific primers^a

Primer pair	Result of PCR with primer pair for stx allele (reference strain)							Reference
	stx1 (EDL933)	stx2 (EDL933)	stx2c (E32511)	stx2d (EH250)	stx2e (ED-53)	stx2f
						(T4/97*)	(H.I.8)
LP43, LP44	−	+	+	+	+	−	−	2
GK3, GK4	−	+	+	−	−	−	−	15
LP30, LP31	+	−	−	−	−	−	−	2
KS7, KS8	+	−	−	−	−	−	−	14
MK1, MK2	+	+	+	+	+	−	−	8
FK1, FK2	−	−	−	−	+	−	−	4
VT2cm, VT2-f	−	−	−	+	−	−	−	13
128-1, 128-2	−	−	−	−	−	+	+	This study
up, down	+	+	+	+	+	+	+	9

^aThe PCR results for E. coli O128:H2 strain T4/97 are identical with those for strains E-D365 (O18a and O18b), E-D371 (O45), E-D373 (O45), and 5598/97 (O128:H2). 

TABLE 2.

Detection of Stx2f by a commercial EIA

Bacterial strain (Stx type)	Reference	Optical densitya		Detection under routine conditionsb
		NI	I	NI	I
C600 (none)	Lab strain	0.045	0.041	−	−
C600 (933W) (Stx2)	Lab strain	26.6	814	+	+
EDL933 (Stx1+Stx2)	11	29.2	254	+	+
E32511 (Stx2c)	16	12.0	105	+	+
E-D53 (Stx2e)	18	0.506	3.81	+	+
E57 (Stx2e)	4	0.073	0.083	−	−
T4/97 (Stx2f)	TSc	0.244	0.870	+	+
5998 (Stx2f)	TS	0.074	0.611	−	+
H.I.8 (SLTIIva, Stx2f)d	5	0.075	0.909	−	+
E-D365 (Stx2f)	TS	0.164	1.913	−	+
E-D371 (Stx2f)	TS	0.164	0.616	−	+
E-D373 (Stx2f)	TS	0.194	0.609	+	+

^aNI, not induced; I, induced with mitomycin C. 

^bNI, not induced; I, induced with mitomycin C; +, detection under routine conditions; −, no detection under routine conditions. Conditions were as described in the Meridian manual for performing and analyzing enzyme-linked immunosorbent assay experiments; breakpoint for + and − results is an optical density of 1.8. 

^cTS, this study. 

^dSince the nucleotide sequence of sltIIva is nearly identical with that of stx_2f and the gene could be amplified with the stx_2f-specific primers, we would propose to rename the sltIIva gene also as stx_2f. 

Specimens of feral pigeon (Columba livia) droppings were collected in the ring park and around the main station in Würzburg, Germany. Enrichment cultures in GN broth Hajna were examined by PCR for the presence of stx and eae genes with primers MK1 and MK2 (8) and primers SK1 and SK2 (14), respectively. PCR-positive cultures were streaked onto MacConkey agar, and lactose-fermenting isolated colonies resembling E. coli were tested for Stx production on Vero cells (15). STEC strains E-D65, E-D371, and E-D373 and 13 further STEC isolates were isolated from feral pigeon droppings in Rome, Italy (3). Serotyping of the STEC isolates was kindly performed at the Nationales Referenzzentrum für Enteritiserreger, Hygiene Institut Hamburg, for the German strains and at the Laboratory for Enteric Pathogens of the Public Health Laboratory Service in London, England, for the Italian strains.

PCR for amplification of stx-specific sequences was performed as previously described (15). Primers used for the detection of different stx genes are shown in Table 1. Primers used for the detection of the stx_2f gene described in this study were 128-1 (5′-AGA TTG GGC GTC ATT CAC TGG TTG-3′) and 128-2 (5′-TAC TTT AAT GGC CGC CCT GTC TCC-3′) and were constructed from the sequence determined in this study. The DNA was denatured at 94°C for 30 s, annealed at 57°C for 60 s, and then extended for 60 s at 72°C. After 30 cycles were completed, a final extension step of 5 min at 72°C was conducted. A PCR product of 428 bp covering a portion of the A-subunit gene was expected. Restriction of the PCR product with EcoRV was expected to yield 145- and 283-bp fragments.

Southern blot hybridization was performed as described earlier (15). Probes used included fragments of stx₁ (amplified from EDL933 by PCR with primers KS7 and KS8) and stx₂ (amplified from EDL933 by PCR with primers GK3 and GK4). The probe concentration used was 50 ng/ml.

Restriction, ligation, and transformation experiments were performed according to standard methods. For the selection of transformants, 30 μg of chloramphenicol per ml was added to the agar plates. Preparation of plasmids was performed with the plasmid Midi (Qiagen GmbH, Hilden, Germany). The nucleotide sequence of the stx_2f gene of STEC strain T4/97 was determined by Taq cycle sequencing. Initial sequences were determined with universal and reverse primers for pUC18 vectors. From these, internal primers were designed to create overlapping DNA fragments to sequence both strands of the whole gene. Each base was sequenced at least three times on each strand. As a control, the sltIIva gene of E. coli O128:B12 strain H.I.8 was sequenced in parallel.

An enzyme immunoassay (EIA) (Premier EHEC EIA; Meridian Diagnostics Inc., Cincinnati, Ohio) was used to detect the expression of Stx from isolated organisms according to the manufacturer's instructions with a minor modification: we inoculated 4 ml of tryptic soy broth without supplementation and, in parallel, 4 ml of tryptic soy broth supplemented with 0.05 μg of mitomycin C per ml. stx-positive E. coli O157:H7 strain EDL933 and stx-negative E. coli O157:H19 strain 693/91 were used as controls.

The reverse passive latex agglutination test for the detection of verocytotoxins 1 and 2 (VTEC-RPLA test; Unipath Limited, Basingstoke, United Kingdom) was performed according to the manufacturer's instructions.

Enrichment cultures of 19 specimens of feral pigeon droppings were examined by PCR for the presence of stx and eae genes. None of the samples reacted with primers MK1 and MK2, but seven of them yielded positive reactions with primers SK1 and SK2, indicating the presence of eae.

E. coli isolates from these eae-positive specimens were further examined for both the presence of eae and cytotoxicity to Vero cells. Two eae-positive strains inducing a cytopathic effect that was neutralized by a rabbit antiserum raised against Stx2 (3) were obtained. Both isolates belonged to serotype O128:H2 and were further subjected to PCR analysis using primers specific for different stx genes together with three other STEC strains isolated from pigeons in Italy (3) which belonged to serogroups O18ab and O45 (Table 1). In agreement with the results obtained with the enrichment cultures, negative results were obtained with primers MK1 and MK2. Amplification with primers LP30-LP31 and LP43-LP44, complementary to the A-subunit genes of stx₁ and stx₂, and with primers KS7-KS8 and GK3-GK4, complementary to the B-subunit genes of stx₁ and stx₂, failed as well to produce PCR products with any of the five strains (Table 1). Negative results were also obtained with primer pairs specific for the stx_2e and stx_2d genes (Table 2). However, it was possible to target the putative stx genes with primers described by Lin et al. (9) (Table 2). A PCR product of approximately 880 bp was obtained with all strains tested.

In order to further characterize the cytotoxin, chromosomal DNA of STEC isolate T4/97 (Table 1) was prepared, digested with restriction endonucleases SphI and BamHI, and separated on a 0.7% agarose gel. Hybridization with a stx₁ probe under conditions which allowed the annealing of sequences with 80% identity and higher was negative. By using a stx₂ probe, a signal could be detected under the same low-stringency conditions at a molecular weight of approximately 6 kb. This fragment was excised from the gel and cloned into pACYC184. Transformants were selected on chloramphenicol agar plates, and stx-positive clones were detected by colony blot hybridization. The 6-kb insert was then sequenced, and the sequence was analyzed. Two open reading frames of 960 and 264 bp were identified which shared extensive sequence similarity to A- and B-subunit genes of stx₂ and various stx₂ variants. Highest nucleotide sequence identity was found to be 99.8 and 100% to the A- and B-subunit genes of the stx2 variant sltIIva of E. coli O128:B12 strain H.I.8 described by Gannon et al. (5), whereas the sequence identity to stx₂ of E. coli O157:H7 strain EDL933 was only 63.4 and 57.4% for the A- and B-subunit genes, respectively. We proposed the name stx_2f for this stx₂ variant gene. The nucleotide sequences of stx from E. coli strains H.I.8 and T4/97 were determined in parallel. Whereas the B-subunit gene is identical with sltIIva, the A-subunit gene differed in two nucleotide residues, resulting in an exchange of amino acid residues from P⁶⁷ to L⁶⁷ and T¹¹⁰ to I¹¹⁰ in the predicted amino acid sequences of Stx2f and SltIIva, respectively. To prove the two mutations, the respective sites were sequenced three times on both strands by using different primers. In all cases, the same nucleotide exchanges could be observed. It is noteworthy to mention that the stx B-subunit genes of strains H.I.8 and T4/97 are six bases shorter than the B-subunit genes of other stx genes. Since both toxin genes are closely related, we propose to also rename sltIIva as stx_2f.

Two PCR primers were designed to amplify an internal part of the stx₂ variant gene. The sequence amplified covered positions 452 to 879 in the A-subunit gene. A 428-bp PCR product was produced when primers 128-1 and 128-2 were used. Regardless of the serotype, all five pigeon STEC isolates from Italy and Germany and the O128 strain H.I.8 yielded a PCR product with these primers (see Table 1) while negative results were obtained with E. coli O157:H7 strain EDL933 (stx₁ and stx₂), E. coli O157:H⁻ strain E32511 (stx_2c), E. coli Ont:H12 strain EH250 (stx_2d), and E. coli O101:H⁻ strain E-D53 (stx_2e) (Table 1). To further ensure that this PCR is specific for this stx₂ variant, restriction with EcoRV was employed. The PCR products from all isolates were cleaved in two fragments of 145 and 283 bp, thus demonstrating the presence of stx_2f.

In order to compare the newly discovered Stx with other Stx, we chose a commercial EIA and used bacterial cells which were either not induced or induced with mitomycin C. The results of this study are shown in Table 2. Whereas the control strains expressing stx₁, stx₂, stx_2c, and stx_2e could be clearly identified by the Stx EIA, most of the Stx produced by the pigeon STEC isolates were negative when the results were analyzed according to the kit instructions. That means that an optical density of <1.8 counted as a negative result. In addition, basic toxin production was low, and although the toxin production could be induced by mitomycin C, it was below the detection limit in most cases (Table 2), in accordance with the results obtained by Gannon et al. with strain H.I.8 (5). When the VTEC-RPLA was used, Stx2f yielded a response in dilutions of up to 1:8 that was weaker than those obtained with Stx2 (>1:128) and Stx2c (1:16).

To assess the prevalence of stx_2f, another 13 E. coli isolates from pigeons (3) which were shown to be toxic by the Vero cell test were subjected to the stx_2f PCR. All STEC isolates were positive in this PCR approach and did not carry the stx₁, stx₂, stx_2d, or stx_2e genes (Table 1). The prevalence of STEC carrying the stx_2f gene in pigeon droppings is therefore 12.5%, as compiled from Dell'Omo et al. (3).

In this report, we show that a variant of Stx2 almost identical to the toxin described by Gannon et al. (5) in the STEC O128:B12 strain H.I.8 isolated from a patient with diarrhea appears to be common among STEC isolates from feral pigeons. Following the alphabetical order designation used for Stx2c, Stx2d, and Stx2e, we propose to designate the toxin produced by the pigeon STEC isolates Stx2f. On the basis of sequence similarity, we also propose to include SltIIva in this group and rename it Stx2f.

It is interesting that stx_2f was present in E. coli strains from pigeons belonging to different serotypes and was isolated in different countries. This situation is analogous to that observed with pigs, in which STEC strains belonging to different serotypes all produced the same toxin variant, Stx2e.

The results of our PCR studies have implications for the DNA-based detection of STEC strains for either diagnostic or research purposes. A number of stx-specific PCR primers have been published, each of which has its special advantage or disadvantage. However, the emergence of new stx variants, in particular stx₂ variants, makes it difficult to decide which primer pair is the most appropriate for diagnostic purposes. In this study, we could show that only the primer pair described by Lin et al. (9) was able to target all the stx₂ variants used in this study. However, these primers have not been tested thoroughly in screening mixed cultures, such as enrichment broth or colony sweeps. The primer pair MK1 and MK2, which is able to target stx₁, stx₂, and the clinical relevant stx₂ variants stx_2c, stx_2d, and stx_2e, was evaluated but did not react with stx_2f sequences.

The usefulness of these primers for the detection of STEC in stool specimens should be better evaluated to ensure an optimal diagnostic spectrum. Future work should help to clarify whether the STEC isolates present in pigeons represent clonal lineages and should bring into light if this toxin variant is really specific for STEC strains adapted to birds.

Nucleotide sequence accession number.

The nucleotide sequence of the stx₂ variant gene stx_2f of STEC O128:H2 strain T4/97 will appear in the EMBL and Genbank database libraries under the accession no. AJ010730.