|
|
||||||||
1Laboratory of Enteric Pathogens, Central Public Health Laboratory, Specialist and Reference Microbiology Division, Health Protection Agency, 61 Colindale Avenue, London, NW9 5HT, UK 2Scottish Agricultural College, Veterinary Science Division, Drummondhill, Stratherrick Rd, Inverness, IV2 4JZ, UK 3Centre for Tropical and Veterinary Medicine, University of Edinburgh, Roslin, Midlothian, EH25 9RG, UK 4Centre for Molecular Microbiology and Immunology, Flowers Building, Imperial College, London, SW7 2AZ, UK
Correspondence C. Jenkins claire.jenkins{at}hpa.org.uk
Received July 29, 2003
Accepted July 29, 2003
Verocytotoxin-producing Escherichia coli (VTEC) causes a wide spectrum of disease in humans, from mild diarrhoea to haemolytic uraemic syndrome (HUS). The verocytotoxin (vtx) and intimin (eae) genes of VTEC strains, other than those of serogroup O157, were subtyped to identify common properties that may be associated with increased pathogenicity. Strains were isolated from patients with HUS, those with diarrhoea or from asymptomatic individuals. Strains of VTEC that carried vtx2 gene subtypes vtx2 and vtx2c were most commonly associated with HUS, whereas strains from patients with less severe disease and from the healthy control group were more likely to have vtx1c or vtx2d genes. The eae gene was detected more frequently in strains isolated from HUS patients than in those associated with cases of diarrhoea; ß-intimin was the most common intimin subtype in strains isolated from both groups of patients. None of the strains from the healthy control group carried the eae gene.
| INTRODUCTION |
|---|
|
|
|---|
The capacity of VTEC to cause disease in humans is associated with its ability to express verocytotoxin (VT) (Melton-Celsa & O'Brien, 1998) and to form attaching and effacing lesions in the intestine (Frankel et al., 1998). Certain strains of VTEC produce an enterohaemolysin that may also contribute to VTEC pathogenicity (Schmidt et al., 1995). There are two major types of VT (VT1 and VT2) and VT2 can be divided into at least five subtypes [VT2, VT2c, VT2d, VT2e and VT2f (Scheutz et al., 2001)]. In this paper, the VT2d subtype we refer to is that described by Piérard et al. (1998), not the activatable VT2d toxin described by Melton-Celsa & O'Brien (1998). There are a number of different methods for differentiation of these subtypes (Bastian et al., 1998) and the nomenclature used to describe them is currently under review (Scheutz et al., 2001). VT1 genes (vtx1) appear to be more homogeneous than vtx2, although variants have been described (Paton et al., 1995; Koch et al., 2001; Zhang et al., 2002a). The variant vtx1 gene, vtx1OX3 or vtx1c, was initially described in VTEC strains that were isolated from sheep (Paton et al., 1995) and was termed vtx1OX3, but has also been detected in human strains (Koch et al., 2001; Zhang et al., 2002a). Zhang et al. (2002a) have designated this variant VT1c, in keeping with the nomenclature applied to the VT2 variants.
The locus of enterocyte effacement (LEE) is located on the chromosome and contains genes that encode proteins required for the attaching/effacing phenotype. These include the eae gene, which encodes intimin, an outer-membrane protein that mediates close contact between bacteria and the target cell (Frankel et al., 1998). There are five well- established intimin types:
, ß,
,
and
(Adu-Bobie et al., 1998; Oswald et al., 2000); other types have also been described, including
(Tarr & Whittam, 2002; Zhang et al., 2002b),
(Tarr & Whittam, 2002),
,
and
(Zhang et al., 2002b).
The aims of this study were to subtype the vtx1, vtx2 and eae genes carried by strains of non-O157 VTEC isolated from patients with diarrhoeal disease and from individuals in a healthy control group and to try to identify common properties in strains associated with HUS patients, compared to those isolated from patients with diarrhoea who did not develop HUS and from asymptomatic individuals.
| METHODS |
|---|
|
|
|---|
Control strains used in PCR and PCR-RFLP tests for subtyping of the vtx1 and vtx2 genes were: VT1, E3787 (O26 : H11) (Smith & Linggood, 1971); VT1c, E105991 (O26 : H11) (this study); VT2, EDL933 (O157 : H7) (O'Brien et al., 1984); VT2c, E32511 (O157 : H-) (Schmitt et al., 1991); VT2d, E25702 (O118 : H12) (Willshaw et al., 1992); VT2e, E88949 (O9ab : H-) (Thomas et al., 1994); and VT2f, E75191 (O128 : H-) (Gannon et al., 1990). Control strains used in the tests for subtyping of intimin genes included:
-intimin, E2347 (O127 : H6) (Cravioto et al., 1979); ß-intimin, E3787 (O26 : H11) (Smith & Linggood, 1971);
-intimin, E32511 (O157 : H-) (O'Brien et al., 1984);
-intimin, E52849 (O111 : H-) (Willshaw et al., 1992);
-intimin, ICC95 (O86 : H34) (Adu-Bobie et al., 1998);
-intimin, E5276 (O103 : H2) (Dorn et al., 1989); and
-intimin, E147779 (E7477 : H25, a provisional new serotype) (Jenkins et al., 2002).
Serotyping.
Each isolate was serotyped by using the LEP serotyping scheme, which depends on identification of the heat-stable LPS somatic or O antigens and the flagella H antigen (Gross & Rowe, 1985). Strains that could not be serogrouped by using the serotyping scheme described above, i.e. those designated O? and O rough', were typed by using PCR-RFLP as described by Coimbra et al. (2000). Rough strains do not express the O antigen and therefore cannot be typed by using a phenotypic serotyping scheme.
Typing of vtx and intimin genes and detection of VTEC enterohaemolysin (ehxA).
VT typing was carried out by using digoxigenin-labelled DNA probes for vtx1 and vtx2 (Thomas et al., 1991). The vtx1 and vtx2 genes were subtyped by PCR with the following primer pairs: VT1c, Stx1c-1/Stx1c-2 (Zhang et al., 2002); VT2 (subtype), P2/PVT2 (Thomas et al., 1994); VT2c, P2/PVT2vh (Thomas et al., 1994); VT2d, VT2-cm/VT2-f (Piérard et al., 1998); VT2e, P2/PVT2e (Thomas et al., 1994); and VT2f, 128-1/128-2 (Schmidt et al., 2000). vtx genes that could not be typed by using these primer pairs were analysed by using the PCR-RFLP method described by Lin et al. (1993).
The eae gene and intimin subtypes
, ß,
and
were detected by using PCR (Oswald et al., 2000). Intimin genes that could not be subtyped by PCR were identified by using the PCR-RFLP method described by Oswald et al. (2000). VTEC enterohaemolysin (ehxA) genes were detected by using PCR (Schmidt et al., 1995).
Statistical analysis.
The distributions of each of the vtx variant genes, the eae gene and the ehxA genes between patients with HUS and individuals who did not develop HUS, i.e. patients with diarrhoea and asymptomatic subjects, were determined in order to identify genes associated with HUS. Fisher's exact tests were performed to test the significance of the association between detection of particular genes and disease symptoms. A S2 test of association was performed to assess the relationship between age-group and disease symptoms. In all cases, a P value of < 0.05 was taken to indicate significance.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
Serotyping
In all three groups of strains, 33 different recognized serotypes were identified and 14 strains were designated O? or O rough (Tables
13). VTEC of serotypes O145 and O26 were isolated most frequently from groups 1 and 2, respectively. Outside the UK, USA and Canada, the most frequently reported non-O157 VTEC serotypes have been O26 : H11/H-, O103 : H2, O111 : H-and O145 : H- (Paton et al., 1996; Caprioli et al., 1997; Schmidt et al., 1999). VTEC of serogroup O26 has emerged as a particularly significant cause of human disease (Caprioli et al., 1997; Schmidt et al., 1999) and has been associated with outbreaks in the Republic of Ireland (McMaster et al., 2001). VTEC O26 has been isolated frequently from cattle and has the same characteristics as human strains of the same serotype (Wieler et al., 1996; Zhang et al., 2000) although, to date, there has been no direct evidence that bovine VTEC O26 causes human infection. VTEC of serogroup O128ab was detected in all three groups, although O128ab : H2 was the only serotype common to all three (Tables
13). VTEC O128ab : H2 that harbours the vtx1c and vtx2d genes has been isolated from sheep (Koch et al., 2001; Ramachandran et al., 2001).
|
|
|
Fourteen strains in this study could not be serogrouped by using the current serotyping scheme. Further analysis of the rfb gene showed that each strain had a different PCR-RFLP pattern, designated P1P14 (Tables
13), indicating that they probably belonged to different serogroups and that the O? strains do not represent a single newly emerging serogroup. PCR-RFLP rfb patterns have not been determined for all 173 established serogroups in the LEP serotyping scheme, so the O rough strains could be variants of established O-types.
Typing of vtx genes
vtx subtyping results are shown in Tables
13. Of strains isolated from patients with HUS, 82 % carried vtx2, of which 46 % had subtype vtx2 and 32 % had vtx2c (three strains had both vtx2 and vtx2c subtypes) (Table 1). vtx1c genes were present in 32 % of strains isolated from patients with less severe diarrhoeal disease and 44 % of strains had vtx2d genes (Table 2). Of the 15 strains of VTEC isolated from healthy individuals in the control group, 67 % carried the gene that encodes the VT1c variant and the vtx2d gene was detected in 80 % of strains (Table 3). Statistical analysis of the results in our study also showed that vtx2 and vtx2c were associated with HUS (P < 0.0001 and P < 0.003, respectively) and that the vtx1c and vtx2d genes were associated significantly with individuals who did not develop HUS (P < 0.001 for both vtx1c and vtx2d).
Other studies have shown that VT2 subtypes VT2 and VT2c were associated with HUS, but that strains that harbour the vtx2d gene were rarely associated with this disease (Bonnet et al., 1998; Piérard et al., 1998; Boerlin et al., 1999; Friedrich et al., 2002). In this study, two strains with vtx2d genes and two that carried vtx2e genes were isolated from patients with HUS. However, it is not clear whether these isolates were the cause of the patients HUS disease, or whether they represent side-findings without any aetiological association with the underlying HUS and the true cause of the patients disease was a strain of VTEC more commonly associated with HUS, such as VTEC O157 : H7 or O26 : H11. Detection of antibodies to the LPS of VTEC O157 or O26 in the patients sera may help to clarify the cause of disease but, unfortunately, sera from these patients were not available.
vtx2 genes from the O rough : H45 strain were untypable by using the primer pairs in this study. A fragment (868 bp) of the vtx2 gene in this strain was amplified by using the primers described by Lin et al. (1993); further analysis showed that this vtx2 gene had a novel PCR-RFLP pattern (data not shown).
One strain of serotype O26 : H11, isolated from a patient with diarrhoea, had the vtx1c and ß-intimin genes (Table 2). Human VTEC that contains vtx1c is usually eae-negative and does not belong to the major non-O157 serotypes (Koch et al., 2001; Zhang et al., 2002). In order to clarify this unusual result, the serotype of this strain was retested by the E. coli reference laboratory at the LEP by using the serotyping scheme described by Gross & Rowe (1985) and was confirmed as E. coli O26 : H11. In addition, the vtx gene was analysed by using PCR-RFLP as described by Lin et al. (1993), using the restriction enzymes HindII and HhaI. The patterns were consistent with those of the vtx1c control strain.
Intimin typing and VTEC enterohaemolysin
The intimin gene, important in the pathogenesis of human VTEC infection (Frankel et al., 1998), was detected more frequently in strains isolated from patients with HUS than from those with less severe disease (50 and 32 %, respectively) and was not detected in any of the strains from the healthy control group (Tables
13). Statistical analysis showed that the intimin gene was significantly associated with cases of HUS (P = 0.04). Other studies have also shown an association between HUS and the presence of the eae gene in strains of VTEC (Boerlin et al., 1999; Eklund et al., 2002). In this study, ß-intimin was the intimin type detected most frequently in both groups 1 and 2. No single intimin subtype has been associated with HUS (Schmidt et al., 1999; Oswald et al., 2000), although there is evidence that intimin subtype mediates both tissue tropism and host specificity (Phillips et al., 2000; Reece et al., 2001). Other pathogenicity factors associated with attachment are being investigated, as 59 % of strains in this study did not carry the intimin gene. Paton et al. (2001) suggested that an outer-membrane protein, designated STEC autoagglutinating adhesin (Saa), may have a role in attachment in strains of VTEC that do not have the intimin gene.
Genes that encode VTEC enterohaemolysin were detected in 64 % of strains from HUS patients, in 68 % of those from patients with less severe disease and in 66 % of strains from the healthy control group (Tables
13). There was no association between the presence of ehxA and VTEC isolated from patients with HUS (P = 0.87).
Subtyping of genes associated with the pathogenicity of VTEC infection, such as the vtx and eae genes, can provide information on the association of each strain with severe disease and on the bacteriumhost relationship, e.g. certain vtx genes and intimin subtypes may be host-specific. In this study, strains of VTEC that carried the eae gene and the VT2 gene subtypes vtx2 and vtx2c were most likely to be associated with HUS. There was no association between the presence of ehxA and symptoms of HUS and the role of VTEC enterohaemolysin in the pathogenesis of disease remains unclear. None of the intimin types was found to be significantly associated with HUS. However, data on intimin subtyping may be important to clarify whether intimin type can influence tissue tropism and determine host specificity.
| ACKNOWLEDGEMENTS |
|---|
|
|
|---|
| REFERENCES |
|---|
|
|
|---|
, ß,
, and
, four intimin derivatives expressed by attaching and effacing microbial pathogens. J Clin Microbiol 36, 662668.
modulates intimin-mediated tissue tropism and host specificity. Mol Microbiol 40, 8698.[CrossRef][Medline]This article has been cited by other articles:
![]() |
M. G. Pedersen, C. Hansen, E. Riise, S. Persson, and K. E. P. Olsen Subtype-Specific Suppression of Shiga Toxin 2 Released from Escherichia coli upon Exposure to Protein Synthesis Inhibitors J. Clin. Microbiol., September 1, 2008; 46(9): 2987 - 2991. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Persson, K. E. P. Olsen, S. Ethelberg, and F. Scheutz Subtyping Method for Escherichia coli Shiga Toxin (Verocytotoxin) 2 Variants and Correlations to Clinical Manifestations J. Clin. Microbiol., June 1, 2007; 45(6): 2020 - 2024. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. S. Milnes, G. C. Pritchard, K. Sprigings, R. Futter, and G. A. Willshaw Verocytotoxin-producing E coli O26 infection in calves Vet Rec., July 29, 2006; 159(5): 159 - 159. [Full Text] [PDF] |
||||
![]() |
M. Noris and G. Remuzzi Hemolytic Uremic Syndrome J. Am. Soc. Nephrol., April 1, 2005; 16(4): 1035 - 1050. [Full Text] [PDF] |
||||
![]() |
C. Toma, E. Martinez Espinosa, T. Song, E. Miliwebsky, I. Chinen, S. Iyoda, M. Iwanaga, and M. Rivas Distribution of Putative Adhesins in Different Seropathotypes of Shiga Toxin-Producing Escherichia coli J. Clin. Microbiol., November 1, 2004; 42(11): 4937 - 4946. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Mundy, C. Jenkins, J. Yu, H. Smith, and G. Frankel Distribution of espI among clinical enterohaemorrhagic and enteropathogenic Escherichia coli isolates J. Med. Microbiol., November 1, 2004; 53(11): 1145 - 1149. [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 |