| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Veterinary Clinical Science and Animal Husbandry, University of Liverpool, Neston, South Wirral CH64 7TE, UK
2 Department of Food and Environmental Safety, Veterinary Laboratories Agency (VLA), Weybridge, Woodham Lane, New Haw, Addlestone, Surrey KT15 3NB, UK
3 Institute of Microbiology and Epizootics, Free University Berlin, D-10115 Berlin, Germany
Correspondence
Muna F. Anjum
m.anjum{at}vla.defra.gsi.gov.uk
Received 26 March 2007
Accepted 11 July 2007
Abbreviations: aEPEC, atypical enteropathogenic E. coli; EHEC, enterohaemorrhagic E. coli; HPI, high-pathogenicity island; HUS, haemolytic uraemic syndrome; STEC, Shiga toxin-producing E. coli; Stx, Shiga toxin.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Enterohaemorrhagic E. coli (EHEC) O26 is a subset of Shiga toxin (Stx)-producing E. coli (STEC), which can cause a broad spectrum of disease from diarrhoea to haemolytic uraemic syndrome (HUS; Nataro & Kaper, 1998). STEC O26 is important as a cause of human infection in central and southern Europe (Tozzi et al., 2003; Zhang et al., 2000a) and is the major cause of HUS and haemorrhagic colitis in Italy, surpassing STEC O157 (Tozzi et al., 2003). In contrast, the number of O26 infections reported in the UK, North America and Japan is much smaller than for E. coli O157 (Banatvala et al., 2001; Brooks et al., 2005; Hiramatsu et al., 2002; McMaster et al., 2001; Pearce et al., 2006; Smith et al., 1998). However, it is possible that the incidence of non-O157 : H7 STECs such as STEC O26 has been under-reported and thus overlooked as a human pathogen as there is no convenient medium, such as sorbitol MacConkey's agar used for O157 : H7, that will reliably screen for STEC O26 in a cost-effective manner.
E. coli O26 strains that lack the stx genes have been classified as atypical enteropathogenic E. coli (aEPEC), as they have the ability to form attaching and effacing lesions on the intestinal epithelium but lack the EPEC adherence factor plasmid (Nataro & Kaper, 1998). Investigation of aEPEC O26 strains from humans and animals has concluded that two major groups exist, one group comprising EHEC/STEC and aEPEC strains with very similar genotypes and virulence markers, differentiated by the stx gene, and the other comprising only aEPEC O26 strains with characteristics unlike those of the former group (Leomil et al., 2005).
These data, taken collectively, indicate that O26 strains isolated from different geographical regions and host animals may harbour diverse virulence characteristics. Therefore, the aim of this study was to compare the virulence profiles of STEC O26 strains from ruminants, the primary reservoir for these pathogens, from the UK with those of strains from other regions, mainly in central Europe, to determine the influence of geographical origin on virulence characteristics. In addition, the virulence profiles of aEPEC and STEC O26 ruminant strains mainly from the UK were compared to determine whether an association of risk factors such as host species and virulence could be linked to distinguish STEC-like O26 strains from other aEPEC O26 strains.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
and 3. The remaining O26 strains were collected between 1989 and 1995 at the Institute of Microbiology and Epizootics (Berlin, Germany). Biochemical characterization was established using the API 20E strip (bioMérieux) on all strains following the manufacturer's protocol.
|
|
, -
, -
and -
), enterohaemolysin (E-hlyA), serine protease (espP), type III secreted proteins (espB and espBß), periplasmic catalase–peroxidase (katP), bundle-forming pili (bfpA), high-pathogenicity island (HPI) genes (irp2 and fyuA) and flagella (fliC) genes were as described previously (Cookson et al., 2002; La Ragione et al., 2002; Sandhu et al., 1997; Schubert et al., 1998). The somatic O-type was also confirmed for all O26 strains using O26-specific PCR amplification of the wzx and wzy genes as described previously (DebRoy et al., 2004).
The remaining PCR primers used in this study (espA, fimB, fimE and yagP) were designed within the conserved region identified from multiple sequence alignments within each gene family, using sequences available in GenBank. For espA, which encompasses a large and diverse gene family, degenerate primers were designed to account for allelic variation. The GC content, presence of hairpin loops and the annealing temperature of primer pairs were checked using CLONE 3 software (Peterson & Ward, 1990). BLASTN searches were performed to ensure high specificity of the primers. The PCR amplification conditions and fragment sizes are given in Table 1. Control strains EDL933, MG1655, LT2 and PT4 were used to determine the reliability and specificity of the primers. For espA, the primer concentration in the PCR mix was varied between 0.25 and 2 µM and PCR was performed at three annealing temperatures for the control strains. For test strains, the espA PCR was performed at 55 °C, whilst the remaining PCR conditions were as outlined in Table 1.
|
Data analysis. A 2x2 contingency table was used and a two-sided Fisher's exact test was carried out to test the association between the risk factors (host, country and stx1 status) and virulence markers (Fleiss, 1981).
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
) and 55 °C (data not shown). However, at a lower annealing temperature (54 °C), a faint non-specific band was detected in the negative control samples (Fig. 1b). For amplification of the fimE and yagP genes, MG1655 was used as the positive control and the sequenced Salmonella Typhimurium LT2 strain as the negative control. In both instances, bands of the appropriate size were detected only in MG1655 (data not shown). For fimB amplification, MG1655 was used as the positive control and the sequenced Salmonella Enteritidis strain PT4 strain as the negative control; as expected, a band of approximately 600 bp was detected only in MG1655 (data not shown). These results were reproduced in two further experiments.
|
Characteristics of STEC O26 strains from different regions
Genotypic and phenotypic profiles were compared for 28 UK and non-UK STEC strains of bovine origin (Table 2
). All strains were negative for rhamnose fermentation, were type 1 fimbriated and harboured the stx1, espA and espBß genes (data not shown). With the exception of strain IMT3753, all of the STEC O26 strains also lacked the 5.2 kb chromosomal yagP–yagT region.
In addition, the STEC strains from the UK showed an identical profile for all genotypes and phenotypes tested except for three strains, EC1702/00, EC690/02 and EC537/01. These three strains were motile, whilst the remaining O26 UK strains were non-motile. Also, EC537/01 harboured the stx1 and stx2 genes. In contrast, only eight of the 16 stx1-positive non-UK strains showed identical profiles. This profile was identical to that of the UK STEC strains in all characteristics except motility. The greatest variability in non-UK STEC O26 strains was seen in the plasmid-encoded E-hlyA, espP and katP genes. The strains could be divided into three groups according to their plasmid gene-related profile. One group was positive for both E-hlyA and espP genes but negative for the katP gene (IMT4400, IMT3837 and IMT3887); strains in this group were from Belgium, Switzerland and Germany. The second group (IMT3853, IMT477, IMT3870 and IMT4040) showed absence of all three plasmid-encoded genes and comprised only strains from Germany. The third group, containing the only strain from the USA (IMT3753), was negative for espP and katP and positive for the E-hlyA gene. The majority of non-UK strains were motile.
When an association between the region of isolation (UK or non-UK) and the presence of virulence markers in STEC strains was tested using a two-sided Fisher's exact test, motility (P=0.0035) and the presence or absence of the katP gene (P=0.0081) presented strong evidence of an association with the region of isolation, with UK strains being less likely to be motile but more likely to possess the katP gene. The espP (P=0.1060) and E-hlyA (P=0.2308) genes presented weak evidence of an association with the region of isolation.
Characteristics of aEPEC O26 strains
The aEPEC strains in our study comprised strains of both ovine and bovine origin of which 21 were from the UK and one from Germany. All aEPEC O26 strains harboured the yagP–yagT region with the exception of the German strain, IMT3862, and the majority (18/22) fermented rhamnose (Table 3). The strains showed mixed profiles for the remaining genotypic and phenotypic characteristics, which were generally very different from the consensus profiles for STEC O26 strains (Table 2). Within the UK bovine aEPEC strains, the profile was varied, with three of the nine strains harbouring O26 plasmid-encoded genes, of which two were unable to ferment rhamnose. Four strains within this group (EC38/99, EC459/01, EC1001/01 and EC1002/01) showed profiles identical to each other apart from motility. Within the ovine strains, five (EC740/01, EC746/01, EC738/01, EC744/01 and EC747/01) showed identical profiles to each other. These strains were non-motile and fermented rhamnose but had the E-hlyA gene. Two other ovine aEPEC strains, EC460/02 and EC417/02, showed identical profiles; they were both motile but did not possess any O26 plasmid-associated genes. The only stx-negative non-UK strain present in our study, IMT3862, was identical in genotypic and phenotypic profile to the consensus profile for STEC O26 German strains (Table 2), except for the absence of the stx1 gene. Several aEPEC strains showed absence of the locus of enterocyte effacement (LEE)-encoded espA and espBß genes, possibly indicating the presence of a truncated LEE region within these strains that needs to be confirmed in future studies.
Using a two-sided Fisher's exact test to compare traits with the host species, the combined fimBE and haemagglutination marker exhibited an association (P=0.0498) with the host animal (ovine).
Comparison of UK O26 strains and predicted profiles for each group
Genotypic and phenotypic profiles from STEC and aEPEC O26 strains from the UK were compared. The results showed that, whilst the profiles for UK STEC strains were nearly identical, the profiles of the aEPEC strains were diverse (Tables 2 and 3). A correlation matrix was constructed to determine whether the presence or absence of the stx1 gene was associated with any other virulence determinant present in this study (Table 4). It was found that the stx1, espP and katP genes were positively associated, whereas yagP–yagT and rhamnose fermentation were both negatively associated with the presence of the stx1 gene. A weak association was also seen between the presence of the E-hlyA gene and the presence of the combined markers fimBE and haemagglutination (Table 4).
|
). All four groups harboured unique profiles, although features such as absence of the yagP–yagT region and inability to ferment rhamnose generally distinguished STEC O26 from aEPEC strains. Comparison of the predicted consensus profiles for the STEC non-UK and UK strains showed motility to be the only variable within the two predicted consensus profiles. Expression of type 1 fimbriae and possession of the E-hlyA gene were variable features between bovine and ovine aEPEC from the predicted consensus profiles. However, the reliability of the predicted consensus profile for the aEPEC group is questionable due to the variability shown by this group, especially within bovine strains.
|
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The major differences in characteristics between the UK and non-UK STEC O26 strains were motility status and the presence or absence of plasmid-encoded genes, with the remaining chromosomally encoded genes and phenotypes being indistinguishable between the two groups. For UK STEC O26 strains, the profile was essentially identical for all determinants except for motility, where the majority of strains were non-motile. However, the non-UK strains were highly likely to be motile and to show variation in plasmid-encoded genes (katP, espP and E-hlyA). Motility has previously been attributed to many strains of the O26 : [H11] complex, especially those forming part of the group 1 complex (i.e. O26 STEC and aEPEC strains; Durso et al., 2005; Leomil et al., 2005; Zhang et al., 2000a). It should be noted that UK strains were collected between 1999 and 2002, whilst non-UK strains were collected between 1989 and 1995. The difference in time of collection of the UK and non-UK strains may have influenced the genotypic and phenotypic differences seen between the two groups, and further studies of O26 isolates collected during the same time span from these regions are required to verify categorically the differences reported here.
Variations in the sizes and combinations of plasmid-encoded genes have been noted previously in STEC O26 human strains from central Europe (Zhang et al., 2000b) and complement our data from ruminant strains gathered from this region. These studies also noted differences in PFGE patterns in STEC O26 human strains, indicating the presence of diverse clonal groups. This is in contrast to studies of Australian and North American STEC O26 strains, which have shown limited clonal diversity (Cobbold & Desmarchelier, 2001; Murinda et al., 2004; Whittam et al., 1993). These inconsistencies are probably due to differences in the genomic composition of STEC O26 strains from different regions and may be key to differences seen in the incidence of O26-associated human disease between the UK/USA and central and southern Europe. However, the similarities seen in this study between UK O26 STEC isolates from ruminants need to be confirmed with similar studies using strains from human infections in the UK. Data gathered from such studies will also establish whether ruminants are the main source of O26 STEC infections in humans, or whether human-to-human contact plays a greater role than anticipated.
This and previous studies have shown that strains within the O26 serogroup share many common characteristics. These include a core genome (Anjum et al., 2003) and possession of common virulence determinants such as the LEE-encoded ß-intimin gene (eaeß; Kaper, 1998) and the HPI fyuA and irp2 genes (Bielaszewska et al., 2005; Karch et al., 1999). In contrast, characteristics that consistently distinguish STEC and aEPEC O26 strains are difficult to find. In our study, STEC O26 strains were easily differentiated from aEPEC O26 strains due to the absence of the chromosomally encoded yagP–yagT region in all but one STEC strain present in our panel.
Absence of the 5.2 kb yagP–yagT region, encoding open reading frames of unknown function in STEC O26 strains, was first identified by comparative genomic microarray studies that showed this region to be present in other pathogenic and non-pathogenic E. coli strains including STEC O157, EPEC O111, aEPEC O26 and MG1655 (K12), but absent in STEC O26 strains (Anjum et al., 2003). Although the significance of the absence of this region is not yet known, it may nevertheless be a useful marker for identifying STEC O26 strains. However, the only STEC O26 strain included in this study from North America, IMT3753, harboured this region. Further studies are required to determine whether strain IMT3753 is an exception or whether it is a variant group representing North American STEC O26 strains.
Similarly, although all STEC O26 strains included in our study were unable to ferment rhamnose, as expected (Hiramatsu et al., 2002; Wieler et al., 1995), approximately 18 % of the aEPEC O26 strains were also unable to ferment rhamnose. Previous studies have found such exceptions in 40–82 % of aEPEC O26 strains tested (Bielaszewska et al., 2005; Murinda et al., 2004). In fact, the inability to ferment rhamnose has previously been associated with the presence of the E-hlyA gene in both STEC and aEPEC O26 strains (group 1), whereas the ability to ferment rhamnose was associated with aEPEC O26 strains carrying the plasmid encoding the -haemolysin gene (Leomil et al., 2005; Wieler et al., 1995). However, no significant association was found between the rhamnose and E-hlyA markers in this study, as several aEPEC strains that fermented rhamnose also harboured the E-hlyA gene, but lacked the hlyD gene encoding -haemolysin (data not shown).
All STEC O26 strains in our panel were of bovine origin and harboured the stx1 gene; only one STEC strain also harboured the stx2 gene. Indeed, STEC strains of bovine origin, including STEC O26, are more likely to harbour stx1 than stx2 or stx1 and stx2 together (Pearce et al., 2004, 2006; Wieler et al., 1996a; Zschock et al., 2000). In contrast, STEC strains, especially O26, harbour either the stx1 or the stx2 gene, although a recent shift from stx1 to stx2 has been detected in human STEC O26 strains (Bielaszewska et al., 2005). It has been hypothesized that the emergence of stx2-harbouring strains may be due to better adaptation to the host or to changes in food consumption (Zhang et al., 2000b). Therefore, the question arises as to the origin of the stx2-harbouring O26 strains found in human infection and whether cattle products and manure are the source of infection.
Comparison of genotypic and phenotypic profiles of UK O26 stx+ and stx– strains showed that aEPEC strains harboured diverse profiles with limited consensus among strains, and that the host animal had no influence on any marker except expression of type 1 fimbriae. The latter was more likely to be present and expressed in strains of ovine origin. A weak association was seen between the presence of the E-hlyA gene and expression of type 1 fimbriae, although several strains expressed type 1 fimbriae but lacked the E-hlyA gene and vice versa. Curiously, in a previous study (Anjum et al., 2003), the fimB gene, which regulates expression of type 1 fimbriae (Sohanpal et al., 2004), was found to be absent in all O26 strains that lacked the E-hlyA gene.
In contrast to the aEPEC strains, the UK STEC O26 strains showed extremely similar profiles, with the presence of the stx1 gene being highly correlated with possession of the espP, katP and E-hlyA genes and expression of type 1 fimbriae. Only one aEPEC strain shared the profile of its STEC counterparts. This strain, IMT3862, was unusual in that it was also the only aEPEC O26 strain that lacked the yagP–yagT region. It is possible that this strain was a STEC that had lost the stx-encoding phage. Loss of the lambdoid stx-encoding prophage from STEC strains is a common phenomenon, especially upon subcultivation (Karch et al., 1992; Mellmann et al., 2005), and is the probable cause of aEPEC-associated HUS cases (Sayers et al., 2006). Overall, the results of our aEPEC study are in contrast to previous work, which has shown aEPEC strains of both human and animal origin (including ruminants) to be divided into two clear groups, with one aEPEC group showing identical characteristics to STEC O26 strains with exception of the stx gene (Leomil et al., 2005). As ruminants are a likely source of human infection, it is expected that the diversity seen within UK aEPEC strains in this study will be reflected in human O26 isolates from the UK, which remain to be studied for these determinants. Such studies will also provide information on the proportion of O26 isolates passing through the food chain and causing human infection that are indeed from ruminants.
Currently, phylogenetic studies are being performed on the panel of UK and non-UK O26 strains to determine whether aEPEC and STEC O26 strains have evolved through different clonal lineages, and to see how closely related different aEPEC O26 strains from this study are to each other. Such studies will also determine whether a recent clonal spread of O26 isolates has occurred in the UK, especially within the STEC pathotype.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. L. Cookson, J. Bennett, C. Nicol, F. Thomson-Carter, and G. T. Attwood Molecular Subtyping and Distribution of the Serine Protease from Shiga Toxin-Producing Escherichia coli among Atypical Enteropathogenic E. coli Strains Appl. Envir. Microbiol., April 1, 2009; 75(7): 2246 - 2249. [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 |
