| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
FOOD SAFETY SYMPOSIUM |
Department of Pathobiology, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada, N1G 2W1
 |
Abstract |
|---|
Key Words: cattle • disease • Escherichia coli • serotype • Shiga toxin • subtype
|
INTRODUCTION |
|---|
). The E. coli that cause enteric disease have been divided into pathotypes, based on their virulence factors and mechanisms by which they cause disease (Nataro and Kaper, 1998; Kaper et al., 2004). One of these pathotypes, called Shiga toxin-producing E. coli (STEC), refers to those strains of E. coli that produce at least 1 member of a class of potent cytotoxins called Shiga toxin. The STEC are also called verotoxin-producing E. coli. The names Shiga toxin (Stx), derived from similarity to a cytotoxin produced by Shigella dysenteriae serotype 1 (O’Brien et al., 1982), and verotoxin (VT), based on cytotoxicity for Vero cells (Konowalchuk et al., 1977), are used interchangeably. Those STEC that cause hemorrhagic colitis and hemolytic uremic syndrome are called enterohemorrhagic E. coli (EHEC; Levine et al., 1987; Nataro and Kaper, 1998). Ruminants, especially cattle, constitute a vast reservoir of STEC, and it is not surprising that human infection can frequently be traced to contamination of food or water with cattle manure. Cattle production and processing of beef are, therefore, targeted as areas in which interventions may reduce contamination of food and the environment with pathogenic STEC shed by healthy cattle. Several methods are used to characterize STEC, investigate factors that influence shedding of these organisms by cattle, and understand the mechanisms by which STEC cause disease in humans. Data from these studies are likely to lead to improved methods of prevention and treatment of disease in humans.
|
CLASSIFICATION OF STEC BY SEROTYPING |
|---|
,b; Prager et al., 2005). The serotype of an E. coli isolate is based on the O (Ohne) antigen determined by the polysaccharide portion of cell wall lipopolysaccharide (LPS) and the H (Hauch) antigen due to flagella protein. There are 174 O antigens (numbered 1 to 181, with numbers 31, 47, 67, 72, 93, 94, and 122 deleted) and 53 H antigens in the international serotyping scheme, with E. coli isolates having various combinations of O and H antigens (Scheutz et al., 2004). A high percentage of STEC serotypes are nonmotile (NM) mutants of strains with an H antigen. Additional investigation can sometimes determine the serotype to which these NM strains belong or identify differences between what appear to be identical NM serotypes. For example, O157:NM STEC strains may be mutants of the commonly occurring O157:H7 STEC or they may be a distinctly different O157:NM STEC (Karch et al., 1993; Karch and Bielaszewska, 2001). The O157:H7 STEC are sorbitol-negative and are distributed worldwide, whereas the O157:NM STEC are sorbitol-positive and are usually found only in certain parts of Europe. Many E. coli isolates, including some STEC, have O or H antigens, or both, that are not in the international scheme and, therefore, cannot be serotyped. Furthermore, there is limited access to a small number of laboratories certified for E. coli serotyping. Nonetheless, serotyping is an important basis for differentiating STEC and is often the starting point in characterization of STEC. Because of the importance of serotype O157:H7 in human disease, it is common to consider STEC serotypes in 2 major categories, O157 and nonO157.
Association of serotypes with disease of varying severity in humans and with sporadic disease or outbreaks has led to the proposal that STEC be classified into 5 seropathotypes, A to E (Karmali et al., 2003; Table 1). Seropathotype A consists of O157:H7 and O157:NM, the serotypes considered to be most virulent. Seropathotype B consists of serotypes O26:H11, O103:H2, O111:NM, O121:H19, and O145:NM, that are similar to the O157 STEC in causing severe disease and outbreaks but occur at lower frequency. Seropathotype C is composed of serotypes that are infrequently implicated in sporadic HUS but are not associated with outbreaks and include O91:H21 and O113:H21. Seropathotype D is composed of numerous serotypes that have been implicated in sporadic cases of diarrhea, and seropathotype E is composed of the many STEC serotypes that have not been implicated in disease in humans. When certain putative virulence genes were examined as markers for the seropathotypes, the markers were shown to be effective, but the pattern was closer to one of a continuum rather than a clear delineation into discrete groups. The concept of seropathotypes is useful for investigation of bacterial factors that contribute to disease and transmissibility and may be refined to be of value in a diagnostic setting.
|
found that 312 STEC isolates obtained from Danish patients over a 2.5-yr period belonged to 50 O groups and 75 O:H serotypes. However, several other systems have been developed to subtype STEC. |
ALTERNATIVE METHODS OF SUBTYPING STEC |
|---|
 TopAbstractINTRODUCTIONCLASSIFICATION OF STEC BY... ALTERNATIVE METHODS OF SUBTYPING... RUMINANTS AS RESERVOIRS OF...HUMAN DISEASE DUE TO...PUTATIVE VIRULENCE FACTORSCONCLUSIONSLITERATURE CITED |
|---|
). In O157:H7, the diversity is due largely to insertions and deletions in O islands, the regions of the genome that are absent from E. coli K–12, and to bacteriophages (Kudva et al., 2002; Shaikh and Tarr, 2003). Epidemiological studies, outbreak investigations, and early detection of geographically dispersed foodborne disease outbreaks are dependent on subtyping methods that discriminate beyond the level of serotype. These methods are most extensively developed for O157:H7 EHEC.
One of the most widely applied methods of subtyping is pulsed-field gel electrophoresis (PFGE), a technique in which fragments of the bacterial chromosome generated by digestion with a restriction enzyme selected to cut the DNA into 20 to 25 pieces are separated by electrophoresis (Gerner-Smidt et al., 2006; Swaminathan et al., 2006; Terajima et al., 2006). Comparison of the patterns that are obtained allows investigators to determine the relationships of isolates to each other. In a comparison of PFGE, multilocus sequence typing, and repetitive-element PCR, Foley et al. (2004) reported that PFGE was the most discriminatory, identifying 72 distinct profiles among 92 O157:H7 isolates, whereas repetitive-element PCR identified 14 groups and multilocus sequence typing distinguished 5 groups. Rigorous standardization of a protocol involving digestion of genomic DNA with the enzyme XbaI has resulted in patterns that may be compared across the PulseNet system in the United States (Gerner-Smidt et al., 2006) and in other countries (Nielsen et al., 2006; Swaminathan et al., 2006; Terajima et al., 2006). The discriminatory power of PFGE can be increased by use of a second enzyme such as BlnI (Rivas et al., 2006). Use of PFGE has played a major role in identifying sources of infection by demonstrating the identity of isolates from foods, animals, or both with isolates from patients. This technique has also resulted in the detection of outbreaks that would otherwise have been missed. Tauxe (2006) noted that use of this method in 1 state resulted in a 67% increase in sensitivity of detection of outbreaks. Continuous PFGE analysis has been used to show that most STEC infections in Denmark are sporadic (Nielsen et al., 2006). Furthermore, O-rough or H- isolates could sometimes be associated with an O group, based on having PFGE profiles that were similar to those of STEC of a known O group.
Another method of subtyping O157:H7 STEC is called multiple locus variable-number tandem repeat analysis (Hyytia-Trees et al., 2006). This method is based on the occurrence of variable numbers of tandem duplications of short stretches of DNA at specific loci in the chromosome. Polymerase chain reactions are used to target these loci and have been very effective in determining relationships among O157:H7 STEC. Multiple locus variable-number tandem repeat analysis is a relatively simple, robust, and rapid technique that appears to be similar in sensitivity and more specific than PFGE (Noller et al., 2003; Lindstedt et al., 2003, 2004).
Multilocus sequence typing, a third molecular method of subtyping based on variations in nucleotide sequences of internal fragments of selected housekeeping genes, has not been effective. Noller et al. (2003) applied this procedure to 77 isolates of O157:H7 STEC that were diverse by PFGE. Their analysis involved 7 housekeeping genes and genes for the membrane proteins ompA and espA. They found that there was no diversity in the sequences of the housekeeping genes and espA and little diversity in ompA.
A phage typing scheme for O157:H7 STEC was developed by Ahmed et al. (1987) and later expanded from 62 to 88 phage types (Ahmed et al., 2000). This method has been a useful adjunct to serotyping and PFGE (Ahmed et al., 2000; Nielsen et al., 2006; Rivas et al., 2006). Phage typing has also been reported for subtyping of certain nonO157 STEC (Prager et al., 2002).
Analysis of virulence genes is a very useful method for subtyping STEC (Beutin et al., 1994). The Shiga toxin genes stx), the E. coli attaching and effacing gene (eae), and the gene for EHEC hemolysin (ehly or ehxA) are frequently investigated (Paton et al., 1993; Roldgaard et al., 2004; Nielsen et al., 2006; Rivas et al., 2006). In an interesting study, Roldgaard et al. (2004) used stx subtyping and phage typing to compare the subtypes of O157:H7 STEC from Danish cattle with those recovered from humans with severe disease. They examined 63 bovine and 86 human isolates and found that there was a marked overlap between the isolates from the 2 sources, but stx2-positive isolates that caused severe human disease were only a minor fraction of the isolates from the bovine sources. These findings are consistent with the notion that only a subset of bovine O157:H7 is implicated in human disease.
|
RUMINANTS AS RESERVOIRS OF STEC |
|---|
; Bettelheim, 2003; Blanco et al., 2004b). More than 470 serotypes have been reported from humans (Bettelheim, 2003; Beutin et al., 2004; Blanco et al., 2004a), most of these being serotypes that have been identified in cattle, beef, or both. In North America, cattle are of most significance as a reservoir of STEC, but in countries such as Australia, sheep are of greater significance.
Numerous investigations have determined prevalence rates of STEC in cattle, but it is difficult to make comparisons among the various studies because of differences in sampling and methodologies. Most studies involved examination of single samples of feces taken from cattle at slaughter or on the farm. Others were longitudinal studies. Some studies involved pooled samples, whereas others used individual animal samples. Reported prevalence rates range from 0% (Wilson et al., 1992) to 71% of animals (Cerqueira et al., 1999) and from 0% (Wilson et al., 1992) to 100% (Cobbold and Desmarchelier, 2000; Thran et al., 2001; Jenkins et al., 2002) of herds. Lowest prevalence rates were found in white veal calves (0 to 9.9% of calves; Wilson et al., 1992; McDonough et al., 1994). In beef cattle, the rates of excretion have ranged from 5.8% (Thran et al., 2001) to 70% (Pradel et al., 2000) of individual animals. There are recent reviews of the rates of excretion of STEC by dairy (Hussein and Sakuma, 2005) and beef cattle (Hussein and Bollinger, 2005).
Many studies of STEC in cattle have focused on serotype O157:H7 because of its importance in human disease. Highly sensitive techniques involving enrichment in broth culture followed by immunomagnetic separation (IMS) and plating (Chapman et al., 1994) permit fecal excretion to be detected in animals that are shedding relatively low numbers of this STEC. Use of IMS is now being employed in detection of STEC of other O serogroups such as O26, O103, O111, and O145 that are important in human disease (Jenkins et al., 2003a; Pearce et al., 2004). Inclusion of this step in isolation protocols has been reported to enhance the level of detection at least 100-fold (Chapman et al., 1994). However, LeJeune et al. (2006) showed that, even with IMS, the sensitivity of detection of O157:H7 from cattle feces was low when the concentration was less than 100 cfu per gram. They suggested that many samples with low concentrations go undetected.
The frequency with which O157:H7 STEC has been reported from cattle ranged from 0% of the animals tested (Conedera et al., 1997; Bonardi et al., 1999) to 41.5% (Mechie et al., 1997). In unreported studies, we recovered this serotype from as high as 85% of a group of feedlot cattle (C. Gyles and R. Johnson, Public Health Agency of Canada, Guelph, Ontario, unpublished data). Again, the lowest prevalences for O157:H7 (ranging from 0 to 0.5% of the calves) have been reported in white veal calves (Conedera et al., 1997; Heuvelink et al., 1998, Bonardi et al., 1999). The low prevalence in white veal calves may be due to their essentially nonruminant gastrointestinal tract, their diet, and the management system.
Shedding of O157:H7 STEC and STEC of other serotypes appears to be related to weaning and age. Lowest rates occur in calves before weaning, with highest rates in calves in the postweaning period and intermediate rates in adult cattle (Mechie et al., 1997; Shinagawa et al., 2000; Nielsen et al., 2002). Several studies have shown that STEC are excreted at higher frequency in the warmer months and at lower frequency in the cold months (Chapman et al., 1997; Jenkins et al., 2002; Dunn et al., 2004).
Serotype O157:H7 STEC have been detected in cattle feces at concentrations of 4 to 107 cfu/g, but in most cases the concentrations are less than 10 to 100 cfu/g (Fegan et al., 2004; Widiasih et al., 2004). There appears to be no difference in the prevalence rates or fecal concentrations of O157:H7 STEC that are shed by cattle on pasture and cattle in feedlots (Fegan et al., 2004). Some cattle shed O157:H7 STEC at high fecal concentration for several weeks and are called persistent shedders (Naylor et al., 2003). These animals are colonized by O157:H7 STEC at the terminal rectum and contribute disproportionately to contamination of beef and the environment. Matthews et al. (2006) used mathematical modeling and statistical analysis to identify sources of variation in shedding of O157 EHEC in Scottish cattle. Their analyses of 2 data sets concluded that the data are best explained by the existence of a small proportion of supershedders. In direct examinations of cattle populations to identify supersheders, Omisakin et al. (2003) found that among 44 slaughtered cattle that had O157 EHEC in their rectal feces, 9% had the organism at a concentration of >104 cfu/g. Similarly, Low and coworkers (2005) reported that among 35 slaughtered feedlot cattle that had O157 EHEC in their rectal feces, 3.7% were high-level carriers, having >103 cfu/g of feces.
Only a few studies have measured duration of shedding of STEC, although several studies of repeated samplings have established that duration of shedding is generally short and variable. In 1 study, cattle were shown to shed O157:H7 STEC for <1 to 10 wk and O26 STEC from <1 to <3 wk (Widiasih et al., 2004). In that study, O157:H7 STEC were shed by cattle not only for longer times but also at much higher concentrations compared with O26 STEC.
In 1 recent study, geographic location and diet were found to be highly associated with differences in prevalence of O157 STEC in market-ready feedlot cattle (Dewell et al., 2005). The odds of E. coli O157-positive samples were 6 times greater in cattle that were fed brewers grains compared with cattle that were not fed brewers grains and 9 times greater for cattle from Nebraska compared with cattle from Eastern Colorado.
More than 100 serotypes of STEC have been isolated from sheep, which are an important source of STEC for humans in some countries (Bettelheim, 2003; Blanco et al., 2003; Cookson et al., 2006). In pigs, the STEC serotypes that are usually detected are associated with edema disease (Gannon et al., 1988) and are specific for pigs, but O157:H7 EHEC have been isolated from pigs (Fratamico et al., 2000; Johnsen et al., 2001). The O157 E. coli that possess H antigens other than H7 and are Stx-negative are relatively common in pig feces (Kaufmann et al., 2006; C. Gyles, unpublished data). Shiga toxin-producing E. coli have also been recovered from goats, deer, horses, dogs, and birds.
|
HUMAN DISEASE DUE TO STEC |
|---|
), less than 10 O groups are responsible for the majority of cases (Figure 1). In most countries, O157:H7 is the serotype associated with most cases of disease and with most disease outbreaks. The O157 STEC were responsible for 26% of 312 STEC infections in Denmark (Nielsen et al., 2006), and 68% of the STEC isolates belonged to 8 O groups (O157, O103, O146, O26, O117, O145, O128, O111, in descending order of occurrence). Eight percent of the isolates were rough. The serogroups and serotypes that are implicated vary from country to country, although certain serogroups such as O26, O103, O111, and O145 tend to be frequent in most countries. For example, in a study of 103 STEC isolates from 99 children in Argentina with bloody diarrhea, HUS, and nonbloody diarrhea, 59% of the isolates were O157:H7, and the next most common serotypes, in descending order, were O145:NM, O26:H11, O113:H21, O174:H2, O8:H19, and O145:H25 (Rivas et al., 2006). In a thorough prospective study in the United States, Jelacic et al. (2003) found that 32% of 82 STEC isolates were O157:H7/H–, and the next most common serotypes, in descending order, were O26:H11, O121:H19, and O103:H2. Brooks et al. (2005) examined 940 nonO157 STEC isolated from patients in the United States during 1983 to 2002 and found that the dominant serogroups, in descending order, were O26, O111, O103, O121, O45, and O145.
|
; Karmali, 2004). Sources of infection include meat (especially undercooked beef hamburgers), ready-to-eat sausages, raw milk, cheese, unpasteurized apple cider and juice, lettuce, cantaloupes, alfalfa sprouts, radish sprouts, drinking water, water for bathing, and contact with animals (Figure 2). Interestingly, whereas hamburger has been a major source of O157:H7 infections in North America, it has never been identified as a source in continental Europe (Karch et al., 1999). Transmission also occurs by person-to-person spread. For O157:H7, the infectious dose is low, estimated to be less than 50 (Tilden et al., 1996) to a few hundred organisms (Bell et al., 1994).
|
; Nataro and Kaper, 1998; Welinder-Olsson and Kaijser, 2005). Although the kidneys are frequent targets, a wide range of organs, including the central nervous system, the lungs, pancreas, and heart, may be affected. The HUS develops in 5 to 10% of individuals infected with O157:H7 STEC and may be accompanied by long-term sequelae (Griffin and Tauxe, 1991). Children less than 5 yr old and the elderly are most susceptible to severe complications.
Intestinal Colonization by STEC
As with other enteric E. coli infections, the disease process is considered to involve colonization of the intestine and damage due to toxins. An overview of the disease process for EHEC is given in Figure 3. Colonization is the process by which STEC overcome host defense mechanisms and establish themselves in the intestine. Gastric acidity is an important host defense mechanism in the gastrointestinal tract, but acid resistance is a general feature of E. coli (Fukushima et al., 2000; Large et al., 2005) and has been demonstrated for O157 (Murinda et al., 2004) and other STEC serotypes (Waterman and Small, 1996; Large et al., 2005). Serotype O157:H7 STEC has been shown to survive in acidic foods such as apple juice and salami and to be capable of causing disease in humans who have ingested low numbers of the bacteria (Bell et al. 1994; Tilden et al., 1996). However, there is considerable strain-to-strain variation within this serotype. Exposure to weakly acidic environments induces an acid tolerance response, which enhances resistance to more acidic pH (de Jonge et al., 2003).
|
; Kaper et al., 2004). Although the AE lesion is not essential for bloody diarrhea and HUS in humans, the vast majority of strains implicated in these syndromes are eae-positive. Thus, most EHEC are eae-positive, and eae has been identified as a risk factor for HUS (Ethelberg et al., 2004).
The eae-positive STEC possess a pathogenicity island called the locus of enterocyte effacement (LEE), which encodes the bacterial proteins necessary for formation of the AE lesion. Similarities between the LEE of enteropathogenic E. coli (EPEC) and development of AE lesions in response to infection with EPEC have been exploited in understanding similar events in STEC (Finlay et al., 1996; Barba et al., 2005; Spears et al., 2006). The LEE is organized into 5 major polycistronic operons called LEE1 to LEE5. The products of the LEE are a type III secretion apparatus (LEE1 to LEE3); a protein translocation system (LEE4); an adherence system consisting of an outer membrane protein called intimin or Eae (E. coli attaching and effacing protein) and its receptor, translocated intimin receptor (TIR), both encoded by LEE5; and effector proteins that are translocated by the secretion system. The secretion apparatus is a molecular syringe structure that begins inside the bacterial cytoplasm, extends through the inner and outer membranes and passes through the host cell membrane. Secreted proteins are transferred from the bacterial cytoplasm to the host cell through this structure. The secreted proteins encoded by the LEE include Tir, mitochondrion-associated protein, EspF (E. coli secreted protein F), EspG, EspH, and EspZ. A number of nonLEE-encoded proteins are also translocated by the LEE secretion apparatus (Barba et al., 2005). The TIR protein becomes inserted into the host cell membrane and acts as the receptor for intimin on the bacterial surface, but certain host cell compounds also bind intimin. The TIR and other secreted proteins activate a number of signaling cascades that result in rearrangement of the intestinal epithelial cell architecture and in changes in the cell physiology.
Interestingly, 1 nonLEE-encoded secreted protein, EspJ, has been identified as an antivirulence factor. Its deletion results in prolonged survival in mice and lambs, and it has been suggested to promote host survival and pathogen transmission (Dahan et al., 2005).
There are at least 17 intimin types (
1, 2, ß1, ß2, 
1, 2/
, 
/
, 
, 
, 
, 2, 
, 
, µ, 
, 
, o; Garrido et al., 2006) associated with heterogeneity in the C-terminal part of the molecule that is involved in binding to Tir (Blanco et al., 2004b; Krause et al., 2005). These intimin variants are distributed among STEC and EPEC of human and animal origins. The 1 intimin is associated with highly pathogenic STEC serotypes such as O157:H7 and O145:H– (Blanco et al., 2005; Rivas et al., 2006). Replacement of eaeO157 with eae from an EPEC strain resulted in colonization by the modified O157:H7 STEC strain in the small intestine of experimentally infected gnotobiotic pigs (Tzipori et al., 1995). By contrast, the wild type O157:H7 STEC colonized the colon. This and other studies suggested that Tir also binds to some host cell structures. Frankel et al. (1996) showed that intimin from EPEC binds to a subset of beta-1 integrins. Recently, Sinclair and O’Brien (2002, 2004) showed that Tir from EHEC O157:H7 binds to nucleolin on the surface of HEp-2 cells and that there is competition between nucleolin and Tir. More recently, Sinclair et al. (2006) reported that nucleolin and beta-1 integrin were located in close association with adherent EHEC O157:H7 in infected pigs and calves.
The AE lesion involves structural changes to the epithelial cell and intimate adherence of the bacteria to the host cell surface. The structural changes include loss of microvilli, pedestal formation, and accumulation of cytoskeletal proteins beneath the adherent bacteria.
Much less is known about adherence of eae-negative STEC. Dytoc et al. (1994) investigated attachment of an eae-negative STEC of serotype O113:H21 to cultured epithelial cells (HEp-2) and to rabbit intestine in vivo. They showed that there were areas of microvillus effacement beneath the organism but that the cytoskeletal rearrangement characteristic of the AE lesion did not develop.
Although STEC strains have some capacity for invasion of enterocytes, infection appears to be localized without any evidence of septicemia. Systemic effects are attributable primarily to the action of toxin that is absorbed from the intestine.
Paton et al. (1997) provided data that support the notion that degree of in vitro adherence of STEC may be related to ability to cause disease. They compared the extent of attachment to Henle 407 cells of STEC recovered from mettwurst sausage but not from patients with that of STEC recovered from patients with HUS whose source of STEC infection was mettwurst. They showed that mean adherence to Henle 407 cells was significantly greater for STEC from outbreak-associated HUS cases than for STEC isolated from mettwurst only. There was no difference in degree of adherence between STEC isolates from sporadic HUS cases compared with STEC isolates from outbreak cases. However, isolates from HUS cases adhered to a greater extent than did isolates from nonhuman sources.
Role of Shiga Toxin
Shiga toxin is the critical virulence factor in STEC diseases. Two types of Stx, called Stx1 and Stx2, were originally recognized (Strockbine et al., 1986). The Stx1 molecule is a highly conserved structure that is identical to that of Shiga toxin of S. dysenteriae type 1 (Takao et al., 1988). Recently, a variant of Stx1, called Stx1c, was reported (Zhang et al., 2002). This variant is most commonly found in strains of ovine origin and may be found as the only Stx subtype or in combination with other subtypes (Brett et al., 2003). This toxin type was not found in eae-positive STEC and has been associated with mild or no disease in humans (Friedrich et al., 2003). In contrast, there are several antigenic variants of Stx2, named stx2c, stx2d, stx2d-activatable, stx2e, and stx2f, that differ in their biological activity and association with disease (Table 2). Other variants of Stx have been reported, but there is no information on their clinical significance. Friedrich et al. (2002) examined the association of stx2 gene variants with disease in 626 STEC isolates from humans. They determined that stx2d and stx2e were associated with mild disease or asymptomatic carriage, were produced by eae-negative strains, and were never present in 268 STEC isolates from patients with HUS. The stx2c gene was found at similarly low frequencies (about 5%) in isolates from patients with HUS and from patients with diarrhea. The stx2f gene was not present in any strain.
|
) or it may develop rapidly due to signaling by Stx (Cherla et al., 2003). The development of programmed cell death in response to Stx may vary with cell type. The cluster of B subunits of the Stx bind to specific glycolipid receptors on the surface of cells, permitting internalization of the toxin molecule. The Stx toxins bind to globotriaosylceramide (Gb3). The Stx (Stx2e) implicated in edema disease of pigs, uses globotetraosylceramide as its preferred receptor, but it can also bind Gb3.
The Stx toxins are approximately 70 kDa molecules, whose A subunits consist of an A1 fragment (27.5 kDa) that contains the enzymatic site and a 7.5 kDa A2 fragment that are linked through a disulfide bond. Proteolytic cleavage and reduction are needed to separate the 2 components (O’Brien and Holmes, 1987).
Following binding of toxin to its receptor, Stx appears to induce its transport to clathrin-coated pits from which the toxin molecule is taken up into the cell by receptor-mediated endocytosis (Sandvig and van Deurs, 2002; Sandvig et al., 2004). In this process, a fragment of cell membrane pinches off to produce a coated vesicle with toxin molecules on the internal surface of the membrane. The vesicles may fuse with lysosomal vesicles, resulting in destruction of the protein toxin, leading to protection of the cell. In cells that are susceptible, however, Stx in the vesicle is transported retrogradely to the Golgi apparatus and the endoplasmic reticulum, after which the A fragment enters the cytosol. A proteolytic enzyme nicks the A subunit leading to a molecule in which A1 and A2 fragments are linked by a disulfide bond, which is subsequently reduced to release both fragments. The fatty acid content of the Gb3 receptor may influence the interaction of Stx with the cell (Kiarash et al., 1994). Recently, a tyrosine kinase was shown to be involved in uptake and intracellular transport of Stx in HeLa cells (Lauvrak et al., 2006). Binding of Stx-induced signaling that resulted in Syk activation and an increase in Stx entry into the cell. The toxin thus appears to regulate its entry into cells. Further evidence that binding of the B subunit to cells may result in signal transduction comes from the observation that binding of Stx1-B to Gb3 on renal carcinoma cells causes cytoskeletal reorganization and morphological changes in the cells (Takenouchi et al., 2004).
Expression of the Stx receptor is a primary determinant of susceptibility to tissue injury. Both quantity and type of Gb3 present on epithelial or endothelial cells may therefore influence susceptibility to Stx. Sodium butyrate can upregulate expression of Gb3, and it has been suggested that sodium butyrate in the colon and the peripheral circulation may increase the susceptibility of cells to Stx (Louise and Obrig, 1995; Louise et al., 1995). The proinflammatory cytokines IL-1ß and tumor necrosis factor- may also upregulate expression of Gb3 on endothelial cells (Taneike et al., 2002).
There is evidence of an association of Stx2 with a higher risk of developing HUS and the presence of both eae and stx2 in an STEC isolate is considered to be a predictor of HUS (Boerlin et al., 1999; Ethelberg et al., 2004). It is not known whether the association of Stx2 with HUS is the result of the action of Stx2 or whether Stx2 is simply a marker for increased disease severity. However, Stx2 is about 1,000 times more toxic for human renal microvascular endothelial cells than is Stx1 (Louise and Obrig, 1995). Also, experimental support for the association between Stx2 and severe disease was provided by Siegler et al. (2003), who compared the effects of Stx1 and Stx2 in a primate animal model of HUS. Intravenous administration of Stx2 resulted in clinical and pathological developments characteristic of HUS, whereas administration of Stx1 failed to induce similar developments (Siegler et al., 2003). Comparison of the crystal structures of Stx2 and Shiga toxin resulted in the identification of 4 major structural differences that might account for the greater association of Stx2 with HUS (Fraser et al., 2004). These differences were greater accessibility of the active site of Stx2, possibly enhanced cytotoxicity due to the Stx2 A-subunit forming a short 2-turn alpha-helix structure after passing through the pore of the B-pentamer, 1 of the 3 receptor binding sites in the Stx2 B-pentamer having a different conformation than that in the Stx B-pentamer, and the carboxyl terminus of the A1 fragment of Stx2 binding at a receptor-binding site in Stx2, which is left unoccupied in Stx.
Most aspects of bloody diarrhea and HUS appear to be attributable to the action of Stx on vascular endothelial cells and thrombotic microangiopathy is a central feature of the disease (Ray and Liu, 2001). This is evident in the lesions and symptomatology of the STEC-associated diseases in animals and humans. Bloody diarrhea is associated with lesions in small blood vessels in the colon. The HUS is associated with renal glomerular lesions that are due to damage to endothelial cells, which become swollen and detach from the basement membrane. Fibrin thrombi develop, and there is narrowing or occlusion of the capillary lumen. The compromise in the blood supply to the glomeruli is the major contributor to loss of kidney function, but damage to glomerular epithelial and proximal tubular epithelium may also contribute to kidney damage.
Inflammation
Inflammation is a prominent feature of the intestinal lesion in infection with STEC O157:H7, and H7 flagellin appears to be a major contributor to inflammation (Berin et al., 2002; Miyamoto et al., 2006). Berin et al. (2002) reported that infection of 2 human colonic epithelial cell lines with wild-type EHEC resulted in activation of p38 and ERK MAP kinases, the nuclear translocation of NF-kappa, and increased expression of mRNA and protein for interleukin 8 (IL-8). Similar results were obtained with isogenic eae- and stx- mutants and with isolated EHEC H7 flagellin, and the researchers proposed that H7 flagellin played a major part in the activation of proinflammatory signals in human colon epithelial cells. Miyamoto et al. (2006) reported that upregulation of proinflammatory cytokines and attraction of subepithelial neutrophils were observed following addition of O157:H7 flagellin to the lumen of human colon xenografts. Flagellin binds to toll-like receptor 5 leading to activation of NF-kappaB and secretion of IL-8. Shiga toxin has also been shown to induce production of IL-8 by human colonic epithelial cells (Thorpe et al., 1999; Yamasaki et al., 1999). The inflammation that develops likely impairs the intestinal epithelial barrier function, thereby facilitating passage of Stx from the lumen into the submucosal area (Hurley et al., 2001).
The inflammatory response extends beyond the intestine, and HUS patients have been shown to have elevated urinary levels of IL-8 and monocyte chemoattractant protein 1 (van Setten et al., 1998). Guessous et al. (2005) reported that exposure of human microvascular endothelial cells to Stx2 and LPS resulted in the release of chemokines including IL-8 and other factors. Their findings support the view that platelet activation by these host products leads to the renal thrombosis characteristic of HUS. More recently, Stahl et al. (2006) showed that LPS was detectable on the surface of platelets of preHUS patients before HUS developed, that O157 LPS appeared to bind human platelets to a greater extent than did LPS from other EHEC strains, and that LPS binding to platelets resulted in their activation. They suggested that absorption of LPS in the early stages of O157 EHEC infection may lead to direct activation of platelets or to binding to endothelium followed by binding to platelets, resulting in platelet consumption. They note that other factors, such as activation of platelets by Stx or other compounds (Karpman et al., 2001), or endothelial cell injury, may also contribute to the thrombocytopenia and the thrombotic state observed in HUS. However, the inflammatory response may be moderated by other EHEC products; Hauf and Chakraborty (2003) showed that DNA binding of NF-kappaB in Hela cells was actively suppressed by STEC that secreted EspB and that the suppression of NF-kappaB activation required EspB. The researchers concluded that suppression of basal and signal-induced NF-kappaB DNA binding was a feature of attaching and effacing-inducing bacteria that likely favored colonization by counteracting host defense responses.
Regulation of Virulence
Regulation of virulence genes is critical if the bacterium is to deploy the various virulence factors in the right location, at the right time, and under appropriate conditions. There has been considerable research on regulation of 2 key virulence attributes, the LEE and Shiga toxin.
Regulation of the genes of the LEE is complex, involving several nonLEE-encoded and LEE-encoded regulators. A more detailed discussion of this subject is presented in Spears et al. (2006). Global nonLEE-encoded regulators include H-NS, which acts as a repressor, and IHF, which is an activator (Barba et al., 2005). Quorum-sensing E. coli regulator A also activates the LEE genes through quorum sensing (Clarke and Sperandio, 2005). Regulators encoded by the LEE include the H-NS-like transcriptional regulator Ler (LEE-encoded regulator), and GrlA (global regulator of LEE activator) that positively regulate the LEE genes. Barba et al. (2005) recently determined that Ler is necessary for the expression of grlA and that Ler and GrlA induce each other’s expression partly through counteracting H-NS-mediated repression.
Regulation through quorum sensing has been the subject of much recent investigation (reviewed by Clarke and Sperandio, 2005). Enterohemorrhagic E. coli appear to use a quorum sensing regulatory system to recognize the intestinal environment and activate genes required for colonization (Sperandio et al., 2003; Walters and Sperandio, 2006). These researchers have shown that the autoinducer by which EHEC sense the large intestine is a newly recognized autoinducer, AI-3, an analog of epinephrine and norepinephrine, and that EHEC also respond to epinephrine or norepinephrine by turning on genes for colonization. Kaper and Sperandio (2005) have reviewed evidence in support of their suggestion that EHEC use AI-3 produced by large numbers of bacteria in the colon to recognize their entry into the colon. Genes encoding flagella and motility are also regulated by this AI-3/epinephrine system, allowing for movement of the bacteria to the epithelium prior to turning on the LEE genes.
The stx genes are located in the late gene region of diverse lysogenic lambdoid phages and are highly expressed when the lytic cascade of the phage is activated (Wagner et al., 1999, 2001b, 2002; Herold et al., 2004; Tyler et al., 2004). Phages regulate Stx production through amplification of gene copy number, activity of phage gene promoters, and through release of Stx. Little is known about factors in the intestine that may promote induction of Stx phages, but human neutrophils do induce Stx production by EHEC (Wagner et al., 2001a). The stx1 genes are also regulated by iron concentration, with toxin synthesis repressed by high iron concentrations. In vitro toxin production has been shown to correlate with severity of disease symptoms due to infection with O157:H7 STEC (Muniesa et al., 2003). There is likely a very intricate interplay between toxin production and adherence, and Robinson et al. (2006) have recently shown in tissue culture studies that Stx induces an increase in a eukaryotic receptor for intimin, resulting in increased adherence of O157:H7 EHEC.
|
PUTATIVE VIRULENCE FACTORS |
|---|
). These products include adhesins, toxins, proteases, iron acquisition systems, LPS, and flagellin. The bases for considering these STEC products as putative virulence factors include homology to known virulence factors, displaying a phenotype consistent with a role in pathogenesis, and regulation by known regulators of virulence genes. The absence of a good animal model that mimics the disease in humans and the likelihood that many of the putative virulence factors contribute only slightly to virulence make it difficult to definitively determine whether they are virulence factors.
|
), and there are also nonfimbrial adhesins in this and other serotypes of STEC. Interestingly, Low et al. (2006) found that only a few of the 16 fimbrial gene clusters in EHEC O157:H7 were expressed. A number of these adhesins have been shown to mediate adherence of the bacteria to cultured cells. The nonfimbrial adhesins include EfaI, chromosomal Iha (iron-regulated gene A homologue; Tarr et al., 2000) encoded on O-islands (OI) 43 and 48, outer membrane protein A (OmpA; Torres and Kaper, 2003), STEC autoagglutinating adhesin (Saa; Paton et al., 2001), and plasmid-encoded ToxB (Tatsuno et al., 2001; Table 3
).
The Efa1 (EHEC factor for adherence) is encoded on OI-122, has homology with ToxB, and is found in EPEC as well as EHEC. This adhesin was detected by observation of a decrease in adherence to Chinese hamster ovary cells of a transposon mutant of a clinical EHEC O111:H– isolate (Nicholls et al., 2000). The mutant also showed reduced human red blood cell agglutination and autoaggregation. The EfaI adhesin is known to contribute to colonization of the bovine intestine (Stevens et al., 2002). The Iha is an outer membrane protein with homology to iron-regulated gene A of Vibrio cholerae that is widely distributed among STEC serotypes (Toma et al., 2004; Tatarczak et al., 2005). Transfer of the iha gene confers adherence to epithelial cells on a nonadhering E. coli. The outer membrane protein A is an adhesin for O157:H7 cells. Torres and Kaper (2003), working with HeLa and Caco2 cells, showed that a mutant deficient in OmpA had reduced adherence to these cell lines. The Saa (Paton et al., 2001) is produced by strains of certain serotypes of LEE-negative STEC (e.g., O113:H21 and O91:H2), including some strains that have been isolated from patients with HUS. This adhesin may be important in colonization of cattle; Jenkins et al. (2003b) showed that the gene for Saa was significantly more associated with STEC from cattle than with those from humans.
The ToxB protein, which has homology with with Clostridium difficile toxins A and B, was shown to contribute to adherence of EHEC to epithelial cells by demonstration that curing O157 Sakai of its pO157 plasmid resulted in a reduction of the number of microcolonies formed on Caco-2 cells and in the secretion of EspA, EspB, and Tir, and that these deficits were restored by introducing a mini-plasmid consisting of the toxB and ori regions of pO157 (Tatsuno et al., 2001). The gene for ToxB is present in a wide range of STEC serotypes (Toma et al., 2004; Tatarczak et al., 2005).
The fimbrial adhesins include long polar fimbriae (Lpf; Doughty et al., 2002; Torres et al., 2002, 2004; Torres and Kaper, 2003), SfaA (Spears et al., 2006), StpA (Friedrich et al., 2004), and StcA (Spears et al., 2006). Long polar fimbriae are encoded by 2 regions of the O157:H7 chromosome (OI-141 and OI-154) and are highly related to the long polar fimbriae of Salmonella enterica serovar Typhimurium. Escherichia coli O113:H21 has only a single copy of the lpf operon. The operons in O157:H7 are called lpf1 and lpf2, respectively. Toma et al. (2004) reported that there are at least 4 Lpf variants in STEC strains. The Lpf1 variant appears to be associated with adherence and microcolony formation and Lpf2 has been suggested to be involved in the early stages of adhesion (Torres et al., 2002, 2004). The Lpf appear to be present in a wide range of E. coli (Toma et al., 2006). An EPEC Lpf had no effect on adherence to HeLa cells or to human intestinal biopsy cells in an in vitro organ culture system and failed to induce antibody to the major subunit protein, LpfA, in infected volunteers (Tatsuno et al., 2006). These researchers also showed that Lpf in Citrobacter rodentium strain appeared to play no role in colonization or pathology in 2 mouse models of disease. These findings cast doubt on the likelihood that Lpf are virulence factors for EHEC. The Sfp fimbriae are encoded by an operon on the megaplasmid of sorbitol-positive O157:NM strains (Brunder et al., 2001). The sfpA gene that encodes the pilin appears to be uniquely associated with these strains (Friedrich et al., 2004; Toma et al., 2004). Shen et al. (2005) recently determined that fimbrial operons encoded in O islands 1, 47, 141, and 154 were highly specific for O157:H7 and O157:NM, but their role in disease has not been determined.
Although the critical importance of Shiga toxin in the pathogenesis of STEC diseases is well established, there are other toxins produced by STEC that may play roles in disease. Cytolethal distending toxin (CDT) was detected in a small proportion of eae-negative STEC and was associated with disease (Bielaszewska et al., 2004). More recently the toxin was reported in O157:H–STEC (Bielaszewska et al., 2005). The genes for CDT had homology with the cdt-III or the cdt-V family of cdt, and CDT-V from an O157:H– strain of STEC was shown to kill 2 lines of human vascular endothelial cells, leading to the suggestion that it may be a contributor to STEC-mediated disease processes.
The astA gene for enteroaggregative E. coli heat-stable enterotoxin (EAST1) is found in O157:H7 EHEC (Savarino et al., 1996). This gene is present in a wide range of diarrheagenic E. coli of human and animal origins. The contribution of EAST1 to STEC disease is not known but it has the potential to contribute to watery diarrhea.
The plasmid-encoded hemolysin (Ehly) is produced by both eae-positive and eae-negative STEC. Among strains from humans in Finland, 92% of the eae-positive STEC were positive for enterohemolysin compared with 35% of the eae-negative strains (Eklund et al., 2001). Ehly is a member of the RTX family of toxins and could contribute to disease through lysis of erythrocytes and release of hemoglobin as a potential source of iron for the bacteria. It could also contribute by its membrane damaging effect on a wide variety of cell types, its ability to induce production of proinflammatory cytokines, or both (Taneike et al., 2002).
Subtilase cytotoxin is a recently discovered toxin that has homology with subtilase of Bacillus anthracis (Paton et al., 2004). It was produced by an eae-negative O113:H21 STEC strain, and the gene was detected in a high percentage of other STEC strains. The toxin was lethal for mice and induced microvascular thrombosis and necrosis in several organs including the kidneys in mice. The toxin has an A1:B5 structure, is a serine protease, and may contribute to pathogenesis of disease.
Several other proteases (Table 3) are produced by STEC, and their activities suggest that they could contribute to disease (Brunder et al., 1996, 1997; Grys et al., 2005). These activities include mucinolysis, cleavage of coagulation factor V, degradation of apolipoproteins, reducing activity, and adherence. The fact that StcE production is regulated by Ler connects this product with major aspects of pathogenesis. The plasmid-encoded type 2 secretion system (Schmidt et al., 1997) may contribute to virulence through its secretion of StcE.
The role of H7 flagellin in inducing an inflammatory response has already been noted. Interestingly, there is also evidence that H21 flagellin plays a role in virulence, albeit through different mechanisms. Rogers et al. (2006) showed that a fliC deletion mutant of an eae-negative O113:H21 STEC was less virulent in a mouse model of infection. The only difference noted between the response to the mutant and wild type strains was that the wild type was closely associated with the colonic epithelium and the mutant was observed in the lumen and at some distance from the epithelium. There was no difference in neutrophil-specific antigen in colonic tissue.
The urease operon is present as 1 or 2 copies in O157:H7 and certain other STEC. In most strains of O157:H7 STEC, the urease-positive phenotype is not observed in vitro, but some serotypes, such as O5:NM, have a urease-positive phenotype (Heimer et al., 2002; Friedrich et al., 2006). Urease could be involved in acid resistance or cytotoxicity in vivo.
Kim et al. (1999) identified 2 lineages of O157:H7 EHEC. Lineage I was associated with human disease and lineage II with carriage by cattle, and it has been proposed that lineage I strains have features that promote their transmission to humans or increase their virulence. Subsequently, a number of studies have shown differences in expression of virulence-related genes between strains of the 2 lineages. Ritchie et al. (2003) showed that STEC strains from HUS produced higher levels of Stx than did strains of bovine origin. Dowd and Ishizaki (2006) used microarray analysis to compare levels of expression of 600 primarily virulence-related O157:H7 genes in lineage I and lineage II isolates grown anaerobically in vitro. Genes that were upregulated in lineage I isolates relative to lineage II isolates included stx2 and genes that regulated stx2 expression, genes for urease, curli fibers, and rfaH. Interestingly, lineage II strains had increased expression of a number of genes suggested to be virulence-related. These included toxB, etp genes, LPS, capsule, and flagella-related genes. Whereas such studies under selected in vitro conditions can point to potential virulence factors, interpretation is difficult given that the same virulence genes are likely upregulated and down-regulated at different times and locations in vivo.
|
CONCLUSIONS |
|---|
|
TopAbstractINTRODUCTIONCLASSIFICATION OF STEC BY...ALTERNATIVE METHODS OF SUBTYPING... |
|---|
