|
|
||||||||
BACTERIAL PATHOGENICITY |





*Department of Bacterial Diseases, Veterinary Laboratories Agency (Weybridge), Woodham Lane, New Haw, Addlestone, Surrey KT15 3NB,
Department of Clinical Veterinary Science and
Department of Pathology and Microbiology, University of Bristol, Langford, Bristol BS40 3DU
Corresponding author: Professor M. J. Woodward (e-mail: m.j.woodward{at}vla.defra.gsi.gov.uk).
Present address: AgResearch Ltd, Grasslands Research Centre, Tennent Drive, Private Bag 11008, Palmerston North, New Zealand.
Received 6 Dec. 2001; accepted 16 May 2002.
| Abstract |
|---|
|
|
|---|
) subtype of intimin. Supernates of both O115:H- isolates induced cytopathic effects on Vero cell monolayers, and PCR analysis verified that both isolates encoded EAST1, CNF1 and CNF2 toxins but not Shiga-like toxins. Both isolates harboured similar sized plasmids but PCR analysis indicated that only one of the O115:H- isolates (CO) possessed the plasmid-associated virulence determinants ehxA and etpD. Neither strain possessed the espP, katP or bfpA plasmid-associated virulence determinants. These E. coli O115:H- strains exhibited a novel combination of virulence determinants and are the first isolates found to possess both CNF1 and CNF2. | Introduction |
|---|
|
|
|---|
This report describes the recovery of novel attaching and effacing E. coli (AEEC) O115 strains from a lamb that had been experimentally inoculated with E. coli O157:H7 18 days previously, at 6 weeks of age. This is the first AEEC to be described that causes natural lesions in a weaned sheep. As there is concern regarding the emergence and dissemination of new types of E. coli as potential food-borne pathogens, the genotype and phenotype of these ovine AEEC isolates are described in detail.
| Materials and methods |
|---|
|
|
|---|
|
Animal procedures
The procedures have been described in detail previously [1719]. Briefly, the affected animal was one of four conventionally reared 6-week-old weaned cross-bred lambs, inoculated orally with 109 cfu of a single human-derived strain of E. coli O157:H7 (140065 Nalr) [19]. All lambs were maintained in a closed pen and received water and pelleted complete feed ad libitum. Faecal samples were taken daily from each animal for bacteriological analysis. Tissues were taken under terminal anaesthesia for pathological examination and bacteriological analysis.
All procedures complied with the Animals (Scientific Procedures) Act 1986 and were performed under Home Office Licence 70/4987.
Light microscopy
Tissues were processed for staining as described in previous studies [1719]. Briefly, tissues were fixed in neutral buffered formalin 10% immediately after excision and 4-µm paraffin wax-embedded sections were prepared and stained with haematoxylin and eosin (H&E) for light microscopy. Selected sections were immunostained by an indirect peroxidase-antiperoxidase (PAP) stain, as described previously [18, 19], with polyclonal primary antisera (Veterinary Laboratories Agency, Weybridge) specific for E. coli O somatic antigens.
Transmission electron microscopy
From one AE lesion identified by light microscopy, adjacent tissue was excised from the wax embedded specimen. This was de-waxed, post-fixed with osmium tetroxide, embedded in epoxy resin, and thin sections were prepared and examined in a Philips 201 electron microscope.
Bacterial recovery and identification
Faecal and tissue samples were processed as described previously [1719] for bacteriological analysis by direct plating and immunomagnetic separation (IMS). Enriched cultures and washed beads (Dynal) were spread on to CHROMagar O157 plates (CHROMagar) supplemented with nalidixic acid 15 µg/ml. Lilac-coloured colonies were tested by O157-specific latex agglutination (Oxoid) and representative O157 and non-O157 (latex agglutination-negative) isolates were stored on Dorset's egg slopes for future analysis. Both serotyping and toxin typing of the isolates were performed at the Salmonella and E. coli Serotyping Section, Veterinary Laboratories Agency (Weybridge), with standard conditions and reagents.
PCR analysis
The PCR primers used in this study are defined in Table 2 [13, 2023]. Template DNA was purified from bacterial strains by phenol/chloroform extractions. Amplification was performed in a total volume of 50 µl containing template DNA (500 ng), 25 pmol of each primer, 100 µM of each dNTP, 5 µl of 10xPCR buffer, 20 mM MgCl2 and 2 units of Taq DNA polymerase (Promega). After an initial denaturation step of 5 min at 94°C, 30 cycles of denaturation at 94°C for 1 min, annealing at 56°C for 1 min and extension for 2 min at 72°C were performed. The reaction was completed with a final extension step of 5 min at 72°C. PCR products were analysed by agarose gel electrophoresis and ethidium bromide staining.
|
Plasmid profile and Southern blot analysis
The plasmid profiles of the O115:H- isolates were determined by standard methods as described previously [24]. DNA was transferred to nitrocellulose filters and probed with chemiluminescent-labelled probes (Amersham Pharmacia Biotech) derived from ehxA, CNF1 and CNF2 PCR amplicons as described in the manufacturer's instructions.
Bacterial adhesion to HEp-2 cells
Approximately 107 cfu of bacteria were added to HEp-2 cells and incubated for 6 h by the method of Donnenberg and Nataro [25]. Non-adherent bacteria were removed by washing with diluent. Bacterial microcolony formation was observed by scanning electron microscopy of HEp-2 cells that had been seeded on to 13-mm diameter glass coverslips and fixed in glutaraldehyde 3% v/v. The fluorescence actin staining (FAS) test was performed as described previously [26].
| Results |
|---|
|
|
|---|
Bacterial recovery and identification
E. coli O157:H7 140065 Nalr strain was recovered from the colon contents of the affected animal [19]. Two further isolates were recovered by direct plating from the colon and rectum, respectively, of the same lamb. These isolates had the lilac colony colour on CHROMagar O157 which is typical of E. coli O157:H7 but they failed to agglutinate in the O157 latex test. They were confirmed as serotype O115:H- and toxin typing by analysis of cytopathic effect on Vero cells indicated that they expressed cyto-necrotising factor (CNF) toxin but not Shiga-like toxin. These isolates were designated O115:H- (CO) and O115:H- (RC) for the colon and rectum isolates, respectively.
Light microscopy
Multiple foci of bacteria closely adherent to the mucosal surface, which appeared to be excavated in places (Fig. 1), were seen on the luminal epithelium in the caecum, the proximal loop of the ascending colon and the rectum. Bacteria associated with the lesions did not immunostain for the O157 antigen, but did stain positively with an O115 antiserum (Fig. 2), which was selected following recovery by culture of this serogroup.
|
|
Transmission electron microscopy
In the lesion identified in wax-embedded tissue, bacteria were observed to be closely adherent to the host cell membrane, and frequently on pedestals, in areas of microvillus effacement (Fig. 3).
|
PCR analysis
When PCR primers specific for stx1, stx2 and the conserved region of eaeA were used in a multiplex PCR, the O115:H- isolates gave positive results for the presence of the eaeA gene only. With additional oligonucleotide primers for the specific amplification of each of the five intimin (eaeA) subtypes thus far described, the O115:H- isolates were determined to be of the
subtype. The RC O115:H- isolate gave negative results in the PCR analyses for the EHEC plasmid markers ehxA (enterohaemolysin), katP (catalase/peroxidase), espP (serine protease) and the type II secretion system (etp). However, the CO isolate gave positive PCR amplification with the primer pairs that amplify ehxA and etpD. PCR analyses of both isolates indicated the possession of both CNF1 and CNF2. Both isolates were also positive for EAST1 (astA), but were negative for the amplification of bundle-forming pili (bfpA). Table 1 summarises the characteristics of the E. coli O115:H- isolates in comparison with other reference strains used as controls in the PCR analyses.
Plasmid profile and Southern blot analysis
Both the E. coli O115:H- isolates had identical plasmid profiles, each harbouring a high and low mol. wt plasmid (Fig. 4). The low mol. wt plasmid appeared to be a high copy number plasmid, as indicated by the intensity of ethidium bromide staining. The predominant DNA species of the low mol. wt plasmid was visible as covalently closed circular (ccc) DNA with a secondary species visible as a slower migrating DNA band, most probably the open circular (oc) form. Southern hybridisation analysis of the plasmid profiles from each isolate with labelled CNF1, CNF2 and ehxA probes in separate experiments indicated that all three probes hybridised with the residual chromosomal DNA of O115:H- (CO) but with neither of the two plasmids from this strain. Both the CNF1 and CNF2 probes hybridised with chromosomal DNA from strain RC, but the ehxA probe did not hybridise with any of the genetic material derived from the RC isolate (Fig. 5).
|
|
Bacterial adhesion to HEp-2 cells
The two O115:H- isolates showed localised adherence to HEp-2 cells, forming small, sparse microcolonies after incubation for 6 h (Fig. 6) and were positive in the FAS test for the detection of AE lesion formation.
|
| Discussion |
|---|
|
|
|---|
The recovery of the O115:H-organisms was serendipitous in that bacteriological analysis of samples from the animal was intended to select for the O157:H7 strain 140065 Nalr by plating on CHROMagar O157 supplemented with nalidixic acid. Whether the strains were resistant to nalidixic acid before plating is unknown but, if not, sufficiently high numbers must have been plated in order to select and recover spontaneous resistant mutants. The fact that the O115 strains had the typical lilac appearance of O157:H7 organisms on CHROMagar O157 and grew despite two selective methods to eliminate non-O157 organisms (CHROMagar O157 and nalidixic acid supplementation), emphasises the importance of serological confirmation of isolates believed to be O157. The occurrence of E. coli strains that mimic the growth and appearance of E. coli O157:H7 on CHROMagar O157 has been established previously [28].
The induction of AE lesions by E. coli O115:H- in sheep is a novel observation. The lamb was conventionally reared and therefore it suckled from its mother and was exposed to natural challenge by environmental bacteria before the O157:H7 experimental inoculation at 6 weeks of age. Therefore, it is most likely that the O115:H- isolates were of natural origin.
To the best of our knowledge the combinations of virulence factors harboured by the two O115:H- isolates have not been described previously. Several studies have demonstrated that astA may be present in human EPEC isolates in addition to eaeA. In one study, only 14 (22%) of 65 EPEC isolates examined were astA-positive [6], but in an outbreak of diarrhoeal disease, an O39:H- isolate that was astA- and eaeA-positive was isolated from >100 cases [29]. An O111 isolate that was also astA- and eaeA-positive was implicated in an outbreak of diarrhoea in Finnish adults and schoolchildren [30]. Interestingly, EPEC strain 2348/69, the prototype EPEC strain used for volunteer studies, contains two copies of the astA gene, one in the chromosome and one in the EPEC adherence factor (EAF) plasmid, which also encodes bundle-forming pili on the bfp operon [6].
The two O115:H- isolates described in this study are eaeA-positive, and bfpA- and stx-negative and as such these isolates may be described as atypical EPEC, as they lack the EAF plasmid [31]. Several case-control studies including eaeA-positive EPEC indicated that EAF plasmid-positive isolates were significantly associated with diarrhoea [32, 33]. However, atypical EPEC have also been reported as being the sole causative agent of acute diarrhoeal disease in several countries [32, 3436].
In addition to astA, CNF toxin activity was demonstrated by the Vero cell assay and both isolates were confirmed as being positive for both CNF1 and CNF2. The possession of both CNF1 and CNF2 has not previously been noted in E. coli. Hybridisation analysis indicated that both the CNF1 and CNF2 genes were chromosomally encoded by both O115:H- isolates. While CNF1 is normally chromosomally encoded [37], CNF2 has hitherto only been reported to be encoded by a transferable F-like plasmid, Vir [38]. CNF1 and CNF2 toxins are expressed by E. coli strains isolated from intestinal infections and induce multinucleation of several eukaryotic cell types in culture [39]. CNF1 production has been demonstrated in enterotoxigenic E. coli (ETEC) strains from cases of diarrhoea, and from E. coli strains associated with urinary tract infections and bacteraemia in man, extra-intestinal infections in cats and dogs, enteritis in piglets, and diarrhoea or bacteraemia, or both, in calves [40]. CNF2-producing strains have been isolated from calves and lambs with diarrhoea or bacteraemia, or both [40]. Experimental inoculation of an E. coli strain with proven CNF2 toxin expression into colostrum-restricted newborn calves resulted in intestinal colonisation, causing long-lasting diarrhoea and bacteraemia with localisation in various internal organs [41].
To date, at least five different intimin subtypes have been described from EPEC and EHEC isolates [20] and their distribution may account for the ability of the intimin-producing strains to colonise different tissues and hosts. The
intimin subtype is the most recent form of the eaeA gene to be described [20]. This intimin subtype has been identified among several STEC serotypes of humans and cattle, with the majority (19 of 25) being serogroup O103 [20]. Only 2 of the 25
intimin isolates (both O103:Hnd) were from non-human (cattle) sources [20]. In man, the EHEC O103:H2 serotype has been associated with HUS in Europe [4244], the USA [45] and Canada [46], although it is only rarely associated with diarrhoeagenic disease in animals [47]. Clonal analysis of the pathogenic STEC serogroup O103 determined that it has a unique profile of virulence traits, including a distinct eaeA sequence [47], that differ from those of other STEC serotypes including O157:H7 [48].
AE lesions were formed on HEp-2 cells by the O115 isolates in the present study, but the lesions were not as developed as those observed in vivo in the lamb. This may reflect differences in the time periods over which the lesions developed, in the specificity of the bacteriahost interaction or features of the expression and regulation of
intimin in vivo that may promote the formation of more extensive AE lesions. It is also possible that the isolates tested in vitro were attenuated by the selection on nalidixic acid.
O115:H- (CO), but not O115:H- (RC), was positive for the typically plasmid-borne elements ehxA and the etp locus by PCR, despite there being similar plasmid profiles for both strains. Southern blot analysis indicated that the ehxA gene was chromosomally located. The large plasmids of diarrhoeagenic E. coli are highly variable in their genetic composition [4]. EHEC O157:H7 possesses a large plasmid, pO157, that commonly encodes the ehxA-D genes necessary for the expression of enterohaemolysin [49], katP for expression of a catalase/peroxidase system [21], the production of a serine protease encoded by espP [50], and a type II secretion system encoded by the etp locus [51]. The role of pO157-associated virulence determinants has not been fully established. The carriage of these plasmid-borne determinants is less common in non-O157 EHEC isolates [4] and they are rarely associated with stx-negative E. coli [48, 5255]. Hybridisation experiments with a specific etp probe previously determined that the etp genes were found in 100% of the EHEC O157:H7 and 60% of the EHEC non-O157 isolates [51]. Of isolates from bovine faeces, only 10% of STEC carried the etp operon, and other stx-negative E. coli groups examined (EAggEC, EPEC, ETEC, EIEC) were all negative for the etp gene cluster [51]. Therefore, strain O115:H- (CO) is unusual in being an stx-negative isolate carrying ehxA and the etp gene cluster and, to our knowledge, this is the first report of the presence of the latter in an stx-negative E. coli of animal origin. Furthermore, the location of these genes is unusual, as Southern blot analysis showed that neither of the O115:H- plasmids hybridised with the ehxA probe, but only reacted with residual chromosomal material of the CO O115:H- isolate. The ehx and etp loci lie directly adjacent to each other on pO157, and may have been acquired together by the CO isolate or were lost from the RC isolate upon selection [56, 57].
The O115:H- isolates described in this report present a combination of virulence factors associated with EHEC, EPEC and ETEC isolates that may cause disease in man and animals. It is not known whether the mild clinical signs exhibited by the lamb infected with E. coli O115 were due to the O115:H- isolate(s) or the inoculated O157:H7, or some other factor. Similar clinical signs were also seen with the three other O157:H7-inoculated lambs in the group, in which O115 was not detected. However, EPEC-like isolates lacking the EAF locus, similar to these O115 strains, are associated with diarrhoea in calves and sheep. In one report a Shiga toxin-negative E. coli isolate was positively identified as being associated with AE lesion formation in the small and large intestines in a calf with severe enteritis [58]. This isolate was subsequently found to be eaeA-positive and stx- and bfpA-negative and of serotype O80:H- [59]. E. coli O115 has been associated with neonatal septicaemia in sheep [60], but the virulence factors associated with the septicaemic strains were not reported. The potential of the O115:H- isolates for causing human disease cannot be excluded, considering the combination of virulence factors they possess. The occurrence of novel, potentially pathogenic combinations of virulence factors in E. coli serovars that have hitherto not been associated with pathogenicity demonstrates the value of screening for specific virulence factors in routine surveillance.
| Acknowledgments |
|---|
| References |
|---|
|
|
|---|
, ß,
, and
, four intimin derivatives expressed by attaching and effacing microbial pathogens. J Clin Microbiol 1998; 36: 662668.This article has been cited by other articles:
![]() |
I. Aktan, B. Carter, H. Wilking, R. M. La Ragione, L. Wieler, M. J. Woodward, and M. F. Anjum Influence of geographical origin, host animal and stx gene on the virulence characteristics of Escherichia coli O26 strains J. Med. Microbiol., November 1, 2007; 56(11): 1431 - 1439. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. L. Cookson, J. Bennett, F. Thomson-Carter, and G. T. Attwood Molecular Subtyping and Genetic Analysis of the Enterohemolysin Gene (ehxA) from Shiga Toxin-Producing Escherichia coli and Atypical Enteropathogenic E. coli Appl. Envir. Microbiol., October 15, 2007; 73(20): 6360 - 6369. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. L. Cookson, D. Croucher, C. Pope, J. Bennett, F. Thomson-Carter, and G. T. Attwood Isolation, Characterization, and Epidemiological Assessment of Shiga Toxin-Producing Escherichia coli O84 Isolates from New Zealand. J. Clin. Microbiol., May 1, 2006; 44(5): 1863 - 1866. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. L. COOKSON, A. D. WALES, J. M. ROE, C. M. HAYES, G. R. PEARSON, and M. J. WOODWARD Variation in the persistence of Escherichia coli O157:H7 in experimentally inoculated 6-week-old conventional lambs J. Med. Microbiol., December 1, 2002; 51(12): 1032 - 1040. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |