University of Nebraska - LincolnDigitalCommons@University of Nebraska - Lincoln

Papers in Veterinary and Biomedical Science Veterinary and Biomedical Sciences, Department of

3-1-2004

Escherichia coli 0 1 57:H7: an update on intestinalcolonization and virulence mechanismsRodney A. MoxleyUniversity of Nebraska - Lincoln, rmoxley1@unl.edu

This Article is brought to you for free and open access by the Veterinary and Biomedical Sciences, Department of at DigitalCommons@University ofNebraska - Lincoln. It has been accepted for inclusion in Papers in Veterinary and Biomedical Science by an authorized administrator ofDigitalCommons@University of Nebraska - Lincoln. For more information, please contact proyster@unl.edu.

Moxley, Rodney A., "Escherichia coli 0 1 57:H7: an update on intestinal colonization and virulence mechanisms" (2004). Papers inVeterinary and Biomedical Science. Paper 21.http://digitalcommons.unl.edu/vetscipapers/21

0 CAB International 2004 ISSN 1466-2523

Animal Health Research Reviews 511 ); 15-33

DOI: 10.1 079lAHRR200463

Escherichia coli 0 1 57:H7: an update on intestinal colonization and virulence mechanisms

Rodney A. Moxley

Department of Veterinary and Biomedical Sciences, University of Nebraska-Lincoln, Lincoln, NE, USA

Received 4 March 2004; Accepted 11 March 2004

Abstract

Cattle are a rnajor reservoir of Escherichia coli 0157:H7, an irnportant zoonotic pathogen that causes hernorrhagic colitis and hernolytic urernic syndrorne (HUS). Colonization of cattle occurs predorninantly in the large intestine, and rnay especially target follicle-associated epithelium (FAE) in the terminal rectum. Bacterial colonization involves induction of attach- ing-effacing (A/E) lesions, rnediated by type I11 secreted proteins and an outer rnernbrane protein called intirnin. ToxB, encoded on plasrnid p0157, contributes to adherence of E. coli 0157:H7 through prornotion of the production and/or secretion of type I11 secreted proteins. I'roduction of type 111 secreted proteins and intestinal colonization appear to involve quo- rum-sensing mechanisms. In the human host, E. coli 0157:H7 rnay have a preference for FAE in the distal srnall intestine. The H7 flagellum induces production of chernokines such as interleukin 8, and neutrophilic infiltration of the intestinal mucosa, which in turn rnay enhance Sl~iga toxin (Stx) uptake across the intestinal epithelium. Both Stx and cytokine responses play critical roles in the induction of the vascular lesions that underlie hernor- rl~agic colitis and HUS. In cattle, Stx binds to intestinal crypt cells and submucosal lymphocytes but not vascular endothelium. The role played by Stx in cattle rnay be to sup- press mucosal irnmunity, yet enhance other effects that promote intestinal colonization.

Keywords: Escherlchia coli; enterohemorrhagic strains; 0157:H7; colonization; virulence

Introduction

Escherichia coli 0157:H7 is an irnportant cause of hern- orrl~agic colitis and the potentially fatal post-diarrheal sequela known as hernolytic urernic syndrorne (HUS) (Levine, 1987; Nataro and Kaper, 1998). E. coli 0157:H7 is the prototype of E. coli strains belonging to the diar- rheagenic class known as enterohernorrhagic (EHEC) (Nataro and Kaper, 1998). E. coli 0157:H7 and other E. coli strains classified as EHEC or Shiga toxin-producing E. coli (STEC) express variants of Shiga toxin (Stx), also

Correspondence: Rodney A. Moxley, Department of Veterinary and Biomedical Sciences, University of Nebraska-Lincoln, Lincoln, NE 68583-0905, USA. E-mail: rmoxleyl @unl.edu

known as verotoxin or verocytotoxin and formerly known as Shiga-like toxin (Levine, 1987; O'Brien and Holmes, 1987; O'Brien and Kaper, 1998; I'aton and I'aton, 1998). Stx is produced by Shigella dysenteriae type 1; variants produced by E. coli 0157:H7 and other STEC rnay include Stxl, Stx2, Stx2c, Stx2d and Stx2e (Cantey, 1985; O'Brien and Holmes, 1987; Melton-Celsa and O'Brien, 1998; O'Brien and Kaper, 1998; I'aton and I'aton, 1998). The terrn EHEC was originally coined to denote strains that cause hernorrhagic colitis and HUS, express high levels of Stx, cause lesions in intestinal epithelial cells known as attachingyeffacing (A/E), and possess a specific 60-MDa EHEC plasrnid (Levine, 1987; Nataro and Kaper, 1998). Stx is a potent cytotoxin that plays a key role in the induction of vascular lesions and other pathogenetic events that characterize hernorrl~agic

Rodney A. Moxley

colitis and HUS. The term STEC refers to any E. coli strain that produces Stx, but not all STEC strains are believed to be pathogens (Nataro and Kaper, 1998; I'aton and I'aton, 1998). Enteropathogenic E. coli (EI'EC) strains constitute a separate class of diarrheagenic E. coli related to EHEC in that they cause A/E lesions, but do not express Stx (Nataro and Kaper, 1998). E. coli 0157:H7 is thought to have evolved from an EI'EC 055:H7 ancestor, first by acquiring Stx genes, and then by diverging into two branches. One branch became a P-glucuronidase- and sorbitol-negative 0157:H7 clone that spread worldwide, and the other lost motility, lead- ing to the 0157:H- clone that has become an increasing public health problem in Europe (Feng et al., 1998; Donnenberg and Whittarn, 2001).

In 1999, investigators at the Centers for Disease Control and I'revention (CDC) estimated that 73 480 people per year in the USA were infected with E. coli 0157:H7 and 61 of these cases were fatal (Mead et al., 1999). Most cases of E. coli 0157:H7 illness are attribut- able to food-borne infection (Armstrong et al., 1996); however, acquisition of disease via direct contact with anirnals and manure at petting zoos and dairy farms is of increasing concern (Crump et al., 2002; Heuvelink et al., 2002; Duffy, 2003). Arnong E. coli 0157:H7 food- borne outbreaks in 1999, one-third were due to beef; historically, undercooked ground beef is the most corn- rnon vehicle (Griffin et al., 2000). More recently, ground beef remains the most frequently identified vehicle, although produce is an increasingly recognized source of outbreaks (Breuer et al., 2001; Griffin, 2003). A case-control study of sporadic E. coli 0157:H7 infec- tions in five FoodNet sites (California, Connecticut, Georgia, Minnesota and Oregon) found that consump- tion of pink l~arnburgers at 11orne or in restaurants is a risk factor for E. coli 0157:H7 infection (Kassenborg et al., 2004). Microbiological testing of ground beef patties from a large outbreak that occurred in the I'acific Northwest between Novernber 1992 and February 1993 suggested that the infectious dose for E. coli 0157:H7 is fewer than 700 organisms (Tuttle et al., 1999). These findings presented a strong argument for enforcing zero tolerance for this organisrn in processed food and for rnarkedly decreasing contarnination of raw ground beef (Tuttle et al., 1999). A major source of the bacteria in ground beef is thought to be bovine feces, which con- taminates carcasses before evisceration; the organisrn is thought to be spread from contaminated hides to the surfaces of carcasses at slaughter (Elder et al., 2000; Barkocy-Gallagher et al., 2001, 2002; Keen and Elder, 2002; McEvoy et al., 2003; Nou et al., 2003). In addition to feces and hides, STEC 0157 has also been isolated from the oral cavities of cattle (Keen and Elder, 2002). A novel pen-sampling strategy for E. coli 0157:H7 was developed on the basis of cattle rubbing, chewing and licking devices place over feed bunks and water tanks (Smith et al., 2004a).

E. coli 0157:H7 is shed in bovine feces and, based on detection of the organisrn in fecal samples, the preva- lence in cattle is relatively high in the USA (Elder et al., 2000; Gansheroff and O'Brien, 2000; Srnit11 et al., 2001; Sargeant et al., 2003; LeJeune et al., 2004). Whereas epi- derniological studies conducted in the early 1990s had suggested that the prevalence of E. coli 0157:H7 in beef and dairy cattle was relatively low (0.3-2.2%; Griffin and Tauxe, 1991; Blanco et al., 19931, more recent studies using more sensitive culture rnethods have shown that the prevalence, at least in USA beef cattle, is much higher (Elder et al., 2000; Gansheroff and O'Brien, 2000; Srnit11 et al., 2001). In one study conducted during October and Novernber 1997, the prevalence in wean- ling range calves based on fecal culture was 87% of herds, and 7% in calves in positive herds (Laegreid et al., 1999). In the same study, serologic investigation showed that 83% of calves and 100% of herds tested positive for antibodies to E. coli 0157 (Laegreid et al., 1999). In Midwestern US beef feedlots, the prevalence during June to September 1999 based on fecal culture (rectal contents samples) was 23% (719 of 3163 sarnples positive; Srnit11 et al., 2001). In a study conducted in Midwestern US abattoirs during July and August 1999, 72% of lots (21 of 29) and 28% (91 of 327) of individual cattle (rectal contents samples) were positive (Elder et al., 2000). In a recent study conducted in Midwestern US beef feedlots involving 10 662 fecal samples, 10.2% were positive for E. coli 0157 (Sargeant et al., 2003). In another study conducted between April and September 2000, 13% (636 of 4790) of fecal pat sarnples from the pen floors were positive (LeJeune et al., 2004).

E. coli 0 1 57:H7 colonization of the intestinal tracts of cattle and other ruminants

E. coli 0157:H7 naturally colonizes the gastrointestinal tracts of cattle and is found in the highest numbers in the large intestine, and especially the rectum (Cray and Moon, 1995; Grauke et al., 2002; Naylor et al., 2003). The ability of the organisrn to colonize the intestine appears to require attachment to intestinal epithelial cells with induction of A/E lesions, a process dependent upon the expression of intirnin by the organisrn (Donnenberg et al., 1993; McKee et al., 1995; Dean- Nystrorn et al., 1998). E. coli 0157:H7 causes A/E lesions in a variety of anirnal models, e.g. infant rabbits, post- weaned rabbits, neonatal piglets (gnotobiotic and colostrum-deprived) and neonatal calves (gnotobiotic and colostrum-fed) (Dean-Nystrorn et al., 1997, 2002; Moxley and Francis, 1998; Woods et al., 2002). Deer, goats and sheep are also naturally infected with E. coli 0157:H7, and the last have been used as a model of ruminant infection (Kudva et al., 1995; Bielaszewska et al., 1997; Cl~aprnan and Ackroyd, 1997; Cornick et al., 2000, 2002; Fischer et al., 2001).

E. coli 0157:H7: intestinal colonization and virulence

The duration of tirne that naturally infected cattle shed detectable levels of E. coli 0157:H7 in the feces is highly variable. The length of detectable fecal shedding is influ- enced by the detection method, with sensitive rnetllods rnore likely to detect longer periods of shedding (Cray and Moon, 1995; Brown et al., 1997; Moxley, 2003). Besser et al. (1997) reported that fecal shedding of E. coli 0157:H7 by cattle, at least at detectable levels, is a relatively transient event. These authors reported that a typical duration was approximately 1 month, and 63% of cattle tested in their study had detectable excretion of the organisrn for less than 1 rnonth (Besser et al., 1997). Khaitsa et al. (20031, in a longitudinal study of naturally infected cattle in a feedlot, described the occurrence of pre-epidemic, epidernic and postepidernic periods. They found that the rnean duration of shedding was longest (4.5 weeks) during the epidernic period.

Naylor et al. (2003) found that the lyrnphoid follicle- dense mucosa in the terrninal rectum is a principle site of colonization in cattle by E. coli 0157:H7. They noted that the organisrn attaches to epithelial cells in this region and induces typical A/E lesions. Hence, the abil- ity of the organisrn to attach to rectal mucosal epithelial cells through A/E lesion developrnent is thought to enable colonization and a prolonged duration of shed- ding in the feces (Naylor et al., 2003). Inoculum dose, age, diet, irnmunity and other factors could also poten- tially affect colonization (Cray and Moon, 1995; Brown et al., 1997; Khaitsa et al., 2003). Rice et al. (2003) noted that anirnals are fecal culture-positive for longer (230 days) or shorter durations (<I0 days), and interpreted that longer durations represent colonized anirnals whereas shorter durations represent transient carriers. These authors developed a recto-anal mucosal swal-, cul- ture rnetllod to detect colonization of the terrninal rectum (Rice et al., 2003).

Cray and Moon (1995) reported that a relatively high inoculum dose [>lo' and probably 210' colony-forming units (c.f.u.11 of in uitmgrown E. coli 0157:H7 was required to infect 1- and 3-year-old adult cattle. I-'reweaned, 3- to 14-week-old calves inoculated with lo1' c.f.u. shed greater numbers of detectable organisms (c.f.u./g), and shed for a longer duration than adult cattle given the sarne level of inoculum (Cray and Moon, 1995). Calves were fecal culture-positive for 2-27 weeks after inoculation, whereas adults were positive for 2-14 weeks (Cray and Moon, 1995). The feces of naturally infected, culture-positive weanling range calves were found to con- tain relatively high numbers of E. coli 0157:H7, on the average, 4 x 10; to 1.3 x 10' c.f.u./g (Laegreid et al., 1999). Using a rnost probable number (MI'N) culture rnethod and automated irnmunornagnetic separation, Fegan et al. (2003) detected 3 MI-'N/g to 2.4 X 10' MI-'N/g in the feces of naturally infected cattle (<3 MI-'N/g were undetectable). The geometric rnean MI'N count of E. coli 0157 in grain- fed cattle was 130 MI-'N/g, whereas that for pasture-fed cattle was 13 MI-)N/g ( P = 0.06; Fegan et al., 2003).

Locus of enterocyte effacement

A/E lesion developrnent involves a cornplex series of signal transduction events, host cell cytoskeletal rearrangements, and forrnation of pedestals at the bacte- rial attachrnent sites (Valiance and Finlay, 2000). The effector rnechanisrn by which EHEC and EI'EC induce forrnation of A/E lesions involves proteins encoded by genes on a large pathogenicity island (I'AI) terrned the locus of enterocyte effacement (LEE) (Nataro and Kaper, 1998). The LEE I'AI and its protein products have been studied extensively. The LEE comprises 41 open reading frarnes organized in five rnajor operons, LEE1, LEE2, LEES, tir (LEES) and LEE4. Included in the LEE are sep and a c genes, encoding a type 111 secretion system; the eae gene, encoding intirnin; tir, encoding the translo- cated intirnin receptor; the apABD genes, which encode proteins that are secreted by the type 111 secretion sys- tern; and ler (LEE-encoded regulator), which encodes an H-NS-like protein that activates the expression of the LEE genes (Nataro and Kaper, 1998).

Genes on the LEE are subject to cornplex regulatory mechanisms. In EI'EC, the plasmid-encoded regulator I'er activates the LEE1, LEE2, LEE3 operons and mod- estly increases the expression of the LEE4 operon (Mellies et al., 1999). In EI'EC and EHEC, the first gene of the LEE1 operon encodes Ler, which is essential for the expression of the LEE genes (Mellies et al., 1999; Elliott et al., 2000; Sperandio et al., 2000). Ler directly binds to the upstream regulatory region of LEE2, which is downstream of LEES, to activate transcription of both the LEE2 and LEE3 operons (Sperandio et al., 2000). Ler up-regulates both intirnin and eae-independent adher- ence of STEC 0157 to epithelial cells (Ogierman et al., 2000). The transcription of ler required for LEE-encoded gene expression in E. coli 0157:H7 is promoted in response to NaHCO; in Lurid-Bertani broth (Abe et al., 2002). Since the concentration of NaHCO; in the lower intestinal tract is relatively high cornpared with that in the upper srnall intestine, the NaHCO; rnay be an irnportant signaling factor for promoting colonization of EHEC in the lower intestinal tract in humans (Abe et al., 2002). Quorum-sensing rnechanisrns also appear to play an irnportant role in the regulation of expression of fac- tors involved in colonization and virulence (discussed below).

One of the first LEE products recognized was the outer rnernbrane protein known as Eae (E. coli attaching and effacing) or intirnin (Jerse et al., 1990). A high per- centage of STEC strains in cattle harbor eae sequences (Wieler et al., 1996). Intirnin rnediates intimate bacterial attachrnent to intestinal epithelial cells and is produced by all bacteria that induce A/E lesions, e.g. E. coli 0157:H7, other EHEC, EI'EC, Hafnia aluei and Citrohacter rodentiurn (Donnenberg et al., 1993; McKee et al., 1995; Dean-Nystrorn et al., 1998; Nataro and Kaper, 1998). A eukaryotic cell-binding dornain is

18

located within a 280-arnino acid carboxy terminus of intirnin polypeptides (Int2801, and arnino acid sequence and antigenic variation exists within this cell-binding dornain expressed by different EI'EC and EHEC isolates (Adu-Bobie et al., 1998). Numerous intirnin subtypes have been described, e.g. al,a,, P, y, 6, K, E, 0, q, h, F, v, and 5; these subtypes are produced by specific EHEC and EI'EC serotypes (Adu-Bobie et al., 1998; Oswald et al., 2000; Zl~ang et al., 2002; Rarnachandran et al., 2003). E. coli 0157:H7 specifically produces intirnin- y (Adu-Bobie et al., 1998).

The particular intirnin subtype produced by a given A/E E. coli serotype influences tissue tropism in the intestine. In gnotobiotic piglets, EHEC strains expressing intirnin-a produced A/E lesions in both the srnall and large intestines, whereas strains expressing intirnin-y produced only large intestinal lesions (Tzipori et al., 1995). E. coli 0157:H7 causes A/E lesions in mucosal tis- sue explants frorn the ileum, colon and rectum of cattle, and the srnall intestines of humans (Baehler and Moxley, 2000, 2002; I-'hillips et al., 2000; Fitzhenry et al., 2002). E. coli 0157:H7 adherence to human srnall intestinal explants was restricted to follicle-associated epithelium (FAE) of ileal I'eyer's patches (I-'hillips et al., 2000). In human srnall intestinal explants, A/E E. coli strains expressing intirnin-a (e.g. EI'EC 0127:H6) adhere to and colonize srnall intestinal epithelium and I'eyer's patch FAE, whereas adherence and colonization by strains expressing intirnin-y (e.g. EHEC 0157:H7 and EI'EC 055:H7) is restricted to I'eyer's patch FAE (I-'hillips et al., 2000; Fitzhenry et al., 2002). A/E E. coli strains geneti- cally engineered to produce intirnin-y target I'eyer's patch FAE, whereas those engineered to produce intirnin-a adhere to and colonize srnall intestinal epithe- lium in human intestinal explants (I-'hillips and Frankel, 2000; I-'hillips et al., 2002).

EI'EC and EHEC bacteria secrete a protein called Tir (translocated intirnin receptor, also called EspE) which is injected into the host cell and becomes the receptor for intirnin (Kenny and Finlay, 1995; Kenny et al., 1997; Deibel et al., 1998; DeVinney et al., 1999). The critical role of the intirnin-Tir interaction in binding of the EHEC and EI'EC bacteria to the host cell during infection has been firrnly substantiated through studies involving inac- tivation of the tir gene (Kenny et al., 1997; DeVinney et al., 1999). However, the pattern of phosphorylation of EHEC 0157:H7 Tir after injection into the host cell is dif- ferent frorn that of EI'EC Tir, and this suggests differences in the function of these proteins within the eukaryotic cell (DeVinney et al., 1999).

Other studies have suggested that intirnin rnay bind to eukaryotic host cell surface proteins, and this interaction rnay possibly play a role in infection. Intirnin expressed by EI'EC binds to P1 integrins, and intirnin-y expressed by E. coli 0157:H7 binds to nucleolin (Frankel et al., 1996; Sinclair and O'Brien, 2002). Antibodies raised against nucleolin significantly reduced binding of EHEC

Rodney A. Moxley

0157:H7 to HEp-2 cells, with the reduction in adherence rnost evident when antibodies were added before or at the tirne of bacterial infection (Sinclair and O'Brien, 2002). These results were interpreted to rnean that the proposed interaction between intirnin-y and nucleolin could occur early in the infectious process. However, the role of these proposed interactions between intirnin and host cell surface proteins in intestinal colonization is currently unknown.

I'roteins called Esp (E. coli-secreted protein) are also expressed by genes on the LEE I'AI and play a rnajor role in the pathogenesis of A/E lesion development. Several proteins called Esc (E. coli secretion apparatus) that are encoded by genes on the LEE I'AI are compo- nents of a type I11 secretion systern through which EspA, EspB, EspD, EspF and Tir are secreted, and are also expressed by genes on the LEE I'AI (Nataro and Kaper, 1998; Roe et al., 2003a). Another protein, called I'as (protein associated with secretion), is required for secretion of the Esp proteins, although its exact role has yet to be elucidated (Kresse et al., 1998). EspA is a rnajor cornponent of a large filamentous organelle that is transiently present on the bacterial surface and inter- acts with the host cell during the early stage of A/E lesion forrnation (Frankel et al., 1998). EspA filarnents rnay contribute to bacterial adhesion, but their role in the forrnation of the translocation apparatus and the translocation of EspB, EspD and Tir to host cells appears to be rnore irnportant (Ebel et al., 1998; Frankel et al., 1998; Kenny et al., 1997; Knutton et al., 1998). EspA filarnents are produced transiently and their loss frorn the surface of the bacterial cell poten- tially allows the subsequent interaction of intirnin and Tir (Knutton et al., 1998). The highest proportions of EspA-filarnented bacteria are detected in late exponen- tial phase, after which they are lost rapidly frorn the bacterial cell surface (Roe et al., 2003b). The propor- tion of the bacterial population expressing EspA filarnents is associated with the level of EspD secretion, and heterogeneous surface expression of EspA fila- rnents is controlled at the post-transcriptional level (Roe et al., 2003b). While specific environmental condi- tions are required to induce expression of espADB, a post-transcriptional control rnechanisrn also exists before EspA filarnents are produced on the bacterial surface and effector proteins [e.g. Tir and mitochon- drion-associated protein (MAl')] are secreted (Roe et al., 2003b). On average, isolates of E. coli 0157:H7 frorn humans had 90-fold greater EspD secretion levels than those frorn cattle, suggesting that strains causing disease in humans represented a rnore virulent subset of the bovine population (McNally et al., 2001). This observation was taken as evidence in support of another study that had proposed different E. coli 0157 lineages in cattle and humans (Kirn et al., 1999).

Contact of the bacterium with the host cell stimulates expression of the LEE-encoded proteins and assernbly of

Rodney A. Moxley

epithelia of adult cattle are also still susceptible to A/E lesions induced i n uiuo by E. coli 0157:H7. To address this question, Baehler and Moxley (2000) prepared mucosal epithelial explants frorn the descending colon and rectum of 18-month-old adult steers and inoculated them with E. coli 0157:H7. A/E lesions were seen at the bacterial attachment sites, confirming the susceptibility of adult tissue to these lesions and providing evidence that the large intestinal mucosal epithelium may be a site of infection that contributes to carriage of E. coli 0157:H7 in adult cattle. I'hillips et al. (2000) found sirni- Par results in ileal mucosal organ cultures prepared frorn an adult cow, and also demonstrated that E. coli 0157:H7 induces A/E lesions in follicle-associated epithelium overlying the I'eyer's patches in human ileal tissue. Although these studies collectively suggested that adult cattle were susceptible to E. coli 0157:H7-induced A/E lesions, these lesions had not been detected i n uiuo in adult cattle infected with E. coli 0157:H7 until recently (Naylor et al., 20031.

Non-LEE-encoded colonization factors

Other factors besides LEE-encoded proteins play a role in intestinal colonization by EHEC and EI'EC, hut they are much less well understood (Torres and Kaper, 2003). The fact that intirnin-negative strains can still cause bloody diarrl~ea or HUS in human patients stitnulated the search for additional adherence factors expressed by E. coli 0157:H7 and other STEC strains. Several proteins have been proposed as novel adhesion factors, includ- ing 11122 (IrgA hornologue adhesin), implicated in adherence to HeLa cells when exposed to E. coli K-12 (Tarr et al., 2000); Saa (STEC autoagglutinating adhesin), produced by LEE-negative non-0157 STEC strains (I'aton et al., 2001); and Efa-1 (EHEC factor for adher- ence), an adhesin produced by non-0157 EHEC strains which is necessary for adhesion to Chinese hamster ovary (CHO) cells i n uitro, hernagglutination, autoaggre- gation, and colonization of the bovine intestine (Nicholls et al., 2000; Stevens et al., 2002~).

The efal gene is present in almost all A/E pathogens, but is not required for A/E lesion formation (Badea et al., 2003). The efal gene was originally detected in a clinical EHEC 0 1 1 l : H isolate during transposon muta- genesis studies and screening for reduced adherence to cultured CHO cells (Nicholls et al., 2000). In this study, 207 E. coli isolates representing a range of pathotypes and serotypes were investigated for the presence of efal gene sequences. None of the 91 eae-negative, but all 116 eae-positive E. coli isolates were positive for efal; however, the efal gene was not physically linked to the LEE I'AI in any of the latter isolates. E. coli 0157:H7 strains contain a truncated version of the efal gene encoding the N-terminal 433 amino acids of the protein on the chrornosorne, whereas sorbitol-positive

E. coli 0 1 5 7 : H strains contain the intact gene (I'erna et al., 2001; Hayashi et al., 2001; Janka et al., 2002; Stevens et al., 2002~). In addition, E. coli 0157:H7 strains encode a hornologue of Efal called ToxB [L7095 open reading frame (ORF)] on the pol57 virulence plasrnid (Burland et al., 1998; Tatsuno et al., 2001). The L7095 ORF displays 28% amino acid sequence identity to the l $ 4 gene in EI'EC, which encodes a lymphocyte inhibitory factor (LIF) called lyrnphostatin (Klapproth et al., 2000).

Efal, lyrnphostatin and ToxB are l~ornologues in EI'EC and EHEC that appear to mediate multiple functions, but most notably host cell adherence and irnmunosuppres- sive effects. Efal is required for efficient colonization of the bovine intestinal tract by STEC 0 5 and 0111, since @a1 deletion and insertion mutants were shed in the feces in significantly lower numbers in experimentally infected 4- and 11-day-old conventional calves (Stevens et al., 2002~). Lysates obtained frorn a prototypic B. coli 0157:H7 strain inhibited interleukin (1L)-2 and IL-4 pro- duction by mitogen-stimulated murine lymphoid cells (Malstrorn and James, 1998). EPEC lyrnphostatin inhib- ited human peripheral blood lymphocyte proliferation; in addition, it inhibited production of IL-2, -4 and -5 and interferon-y by human peripheral and lamina pro- pria mononuclear cells, and proliferation of these cells (Klapproth et al., 2000). An E. coli 0157:H7 strain cured of pol57 lacked the ability to inhibit IL-2 and IL-4 syn- thesis in mitogen-activated human peripheral blood mononuclear cells (Klapproth et al., 2000). Stevens et al. (2002b) noted that Efal, at least in EHEC 0 5 and 0111, rnay potentially influence bovine intestinal colonization in several ways; e.g. by potentially act as an adhesin, promoting EHEC survival in the bovine gut by modulat- ing mucosal immunity, or acting indirectly by influencing the expression and secretion of LEE-encoded proteins or other membrane-associated proteins that influence colonization.

Whether the Efal hornologue ToxB and the truncated @a1 gene on the chromosome of E. coli 0157:H7 play roles in intestinal colonization by this organism is uncer- tain; however, there is evidence to suggest an effect.

Transposon insertion upstream of the truncated @a1 gene in E. coli 0157:H7 reduced bacterial adherence to Caco-2 (human colonic carcinoma) cells (Tatsuno et al., 2000). A pOl57-cured strain of E. coli 0157:H7 had a reduction in the levels of production and secretion of EspA, EspB and Tir, concurrent with a reduction in the number of microcolonies adherent to Caco-2 cells (Tatsuno et al., 2001). Introduction of a mini-pol57 plasrnid composed of the toxB and ori regions restored the full production and secretion of the LEE-encoded proteins, and full adherence capacity to the p0157- cured strain. ToxB was interpreted to contribute to adherence of EHEC to epithelial cells thorough prorno- tion of the production and/or secretion of type I11 secreted proteins (Tatsuno et al., 2001).

E. coli 0157:H7: intestinal colonization and virulence

Effects of Shiga toxins and flagellin on colonization and virulence

The importance of Stxs in the causation of hernorrhagic colitis and HUS in humans is without question, but their contribution to EHEC-induced diarrhea and intestinal inflarnrnation is less well understood. Early studies by Keusch et al. (1972) showed that Stx (produced by Shigella dysenteriae type 1) has enterotoxigenic effects in rabbit ileal loops. Keenan et al. (1986) dernonstrated that purified Stx or Stxl (the latter produced by E. coli 0157:H7) injected directly into ligated ileal loops in adult rabbits caused apoptosis of mature absorptive epithelium with resultant villous atrophy. More recent studies have shown that Stxl causes apoptosis of HEp-2 (human laryngeal carcinorna) and HCT-8 (intestinal epithelial) cells (Ching et al., 2002; Srnit11 et al., 2003). However, parenterdl adrninistration of Stxs causes vdscu- lar lesions in different anirnal rnodels that resernble hernorrhagic colitis and HUS (Moxley and Francis, 1998; Siegler et al., 2001). Stxs given intravenously or by other parenteral routes of adrninistration in anirnal rnodels cause hernorrhagic infarction of the intestine, and this is another explanation for how they can cause intestinal cornprornise and diarrhea (Richardson et al., 1992; Moxley and Francis, 1998).

In susceptible cells, Stxs bind to globotriaosylce- rarnide (Gb3) receptors, are endocytosed, and cause inhibition of protein synthesis and cell death (O'Brien and Holmes, 1987; Lingwood et al., 1998). In humans, Gb3 is rnainly found on the surfaces of endothelial cells following stirnulation with inflarnrnatory rnediators and Stx has been shown to induce apoptosis of these stirnu- lated cells (van de Kar et al., 1992; van Setten et al., 1997; I'ijpers et al., 2001). In an adult rabbit model, radiolabeled Stxl injected intravenously localized in the vascular endothelium in the sarne organs in which Gb3 was expressed (Richardson et al., 1992). In a baboon model, intravenous adrninistration of Stxl induced the clinical features of HUS (Siegler et al., 2001). Greyhounds injected with purified Stxl or Stx2 devel- oped hernorrhagic necrosis in the srnall and large intestinal mucosa, renal lesions, and thrornbocytopenia and anernia without evidence of disseminated intra- vascular coagulation (Fenwick and Cowan, 1998).

The fact that E. coli 0157:H7 and other STEC infect the intestine has raised the question of how Stxs travel frorn the intestinal lumen to vascular targets in the brain, kidney and colon in humans (Acheson et al., 1996). In humans, Stxs rnay be translocated across the intestinal epithelium without apparent cellular disruption, and absorbed into the systernic circulation (Acheson et al., 1996, 1998). In one study, Stx2 was detected in nine of ten patients with epidemic HUS, and in these cases the toxin was bound exclusively to polymorphonuclear leukocytes (I'MN) (Maroeska et al., 2001). The detection of Stx2 bound to I'MN was associated with the presence

of diarrhea at the tirne the blood sarnples were obtained. Experirnentally, Stxl, when incubated in whole blood, binds rapidly and cornpletely to human I'MN and not to other components of blood (I\/laroeska et al., 2000). These studies support the hypothesis that, in humans, I'MN serve as a carrier for Stx in the circulation frorn the intestine to target organs.

Stxs induce CXC cl~ernokine production in human endothelial and intestinal mucosal epitl~elial cells, and these effects, through induction of I'MN infiltration into the intestinal epithelium, are thought to promote intes- tinal epitl~elial darnage and enhanced Stx uptake into the systernic circulation (Thorpe et al., 2001). Intestinal epitl~elial cells frorn mice, humans and certain irnmortal- ized cell lines (e.g. human colonic T84 cells) do not express Gb3 and are resistant to the cytotoxic effects of Stxs (Tesh et al., 1993). However, proinflarnrnatory cytokines, e.g. IL-1 and tumor necrosis factor a (TNF-a) can induce the expression of Gb3 on the surfaces of human glornerular epithelial cells. Hence, inflarnrnatory cytokines released during infection i n uiuo could result in epitl~elial Gb3 expression. Stxs induce and superin- duce IL-8 rnRNA and protein in intestinal epitl~elial cells i n uitro via a ribotoxic stress response (Thorpe et al., 1999). Multiple rnernbers of the CXC cl~ernokine farnily are induced in response to Stxl, including GRO-a, GRO-P, GRO-y and ENA-78 (Thorpe et al., 2001). In addition, exposure of intestinal epitl~elial cells i n uitro to Stxl induces the prirnary response genes c-jun and c-fos and activates the stress-activated protein kinases JNK/SAI'K and p38 (Smith et al., 2003). These data linked the Stxl-induced ribotoxic stress response with both cl~ernokine expression and apoptosis in the intes- tinal epitl~elial cell line HCT-8 and suggested that blocking host cell mitogen-activated protein kinases rnay prevent these Stx-associated events (Smith et al., 2003). Srnit11 et al. (2003) suggested that the blocking of specific host signal transduction pathways in the gut rnay be a rneans by which both Stx-induced host irnmune activation and apoptosis can be diminished.

Although Stxs are capable of inducing CXC cl~ernokine responses, the elevated responses seen in cells infected with certain STEC strains are largely attrib- utable to the production of flagellin (Rogers et al., 2003). Berin et al. (2002) dernonstrated that much of the IL-8 response elicited by an EHEC 0157:H7 strain they stud- ied was attributable to the presence of H7 flagellin. Rogers et al. (2003) found that H21 in LEE-negative strains (e.g. 0113:H21) elicited CXC cl~ernokine responses cornparable to those achieved by infection with the STEC strains themselves. Hence, through induc- tion of CXC cl~ernokine responses and I'MN infiltration, flagellin rnay indirectly enhance the uptake of Stxs across the intestinal epithelial layer.

The role of Stxs, if any, in intestinal colonization and disease in cattle is unknown. Unlike humans, cattle lack vascular receptors for E. coli 0157:H7 Stxs, and this is

Rodney A. Moxley

thought to explain why cattle are tolerant to infection (I-'ruirnboorn-Brees et al., 2000). However, bovine srnall and large intestinal crypt epithelia do contain receptors for E. coli 0157 Stxl (Hoey et al., 2002). Despite recep- tor expression, Stxs exhibit no cytotoxic activity against bovine epithelial cells, apparently due to exclusion of the toxin frorn the endoplasrnic reticulum and localiza- tion within lysosornes (Hoey et al., 2003). Bovine lyrnphocytes have receptors for Stx, and Stxl causes apoptosis of bovine B lyrnphocytes and inhibition of responses of CD8+ bovine T cells to rnitogens (Menge et al., 1999). Stxl rnay possibly contribute to the pathogen- esis of STEC infections in cattle by suppressing the mucosa-associated irnmune response (Menge et al., 1999). However, it does not appear to be involved in the induction of intestinal secretory and inflarnrnatory responses to EHEC 0103:H2 in bovine ligated ileal loops (Stevens et al., 20022). By contrast, stx2 alters how the infant rabbit rnodel responds to EHEC 0157:H7 infection by promoting heterophilic infiltration of the colonic epithelium and larnina propria (Ritchie et al., 2003). Furthermore, purified Stx2 induced enteric inflarn- rnation and diarrhea in the infant rabbit rnodel (Ritchie et al., 2003).

Quorum sensing mechanisms involved in intestinal colonization and virulence

Several virulence-associated genes in EHEC such as the LEE genes, Stx genes and the flagella regulon are acti- vated through the bacterial cell-to-cell signaling rnechanisrn known as quorum sensing (QS) (Sperandio et al., 1999, 2001). The QS rnechanisrn uses hormone- like compounds referred to as autoinducers to allow the organism to sense cellular density in the environment (Surette et al., 1999; Sperandio et al., 2003). In the case of the intestine, autoinducers are involved in activation of genes essential for colonization (Sperandio et al., 2003). The QS systern used by EHEC was initially reported to be the LuxS/autoinducer 2 (AI-2) systern, which is extensively involved in interspecies cornmuni- cation; AI-2 is a furanosyl borate diester, the synthesis of which requires LuxS (Surette et al., 1999; Schauder et al., 2001; Chen et al., 2002; Sperandio et al., 2003). LuxS is an enzyrne involved in the detoxification of S-adenosyl- methionine; in addition, it converts ribose-hornocysteine into hornocysteine and 4,5-dihydroxy-2,3-pentanedione, the precursor of AIL2 (Schauder et al., 2001; Sperandio et al., 2003). EHEC uses two novel regulatory systems in its QS cascade, narned QS E. coli regulator A (QseA) and QseB and C (Sperandio et al., 2002a, b, 2003). In their rnost recent report, Sperandio et al. (2003) found that an EHEC 1z~xS mutant, unable to produce AI-2, still responded to a eukaryotic cell signal to activate expres- sion of its virulence genes. They identified this signal as the horrnone epinephrine and dernonstrated that P- and

a-adrenergic antagonists can block the bacterial response to this hormone. Using purified and i n uitrcr synthesized AI-2, they showed that AI-2 is not the autoinducer involved in the bacterial signaling. They concluded that another, previously undescribed autoin- ducer (AI-31, the synthesis of which depends on the presence of LuxS, was involved. The results irnplied that potential cross-communication occurs between the luxS/AI-3 bacterial QS systern and the epinephrine host signaling systern (Sperandio et al., 2003).

Immune responses to E. coli 0 1 57:H7

The Esp and Tir proteins secreted by E. coli 0157:H7 play critical roles in the developrnent of the A/E lesion and are recognized serologically in human HUS patients. Antibodies to the intirnin, Esp and Tir proteins have been detected in HUS patients following infections with E. coli 0157:H7 and other EHEC (Brunder et al., 1997; Chart et al., 1998; DeVinney et al., 1999; Jarvis and Kaper, 1996; Li et al., 2000; I-'aton et al., 1998a). Sixty- two percent of human HUS patients seropositive for antibodies to E. coli 0157 lipopolysaccharide (LI-'S) also had antibodies to E. coli 0157:H7 secreted proteins of 80, 40, 39, 25 and 15 kDa (Chart et al., 1998). E. coli 0157:H7-secreted proteins of 104 (Espl-'), 37 (EspB) and 24 kDa (EspA) are recognized by rabbit antiserum raised against the Esp proteins (DeVinney et al., 1999; Jarvis and Kaper, 1996) and by sera of HUS patients convales- cent frorn E. coli 0157:H7 and/or other STEC infections, but not by sera frorn norrnal healthy individuals Qarvis and Kaper, 1996). Convalescent HUS patients have also been shown to contain serum antibodies to Tir (I-'aton et al., 1998a) and Espl-' (Brunder et al., 1997) proteins. Because the rnajority of human HUS patients are seropositive for antibodies to Esp and Tir proteins, this is also taken as evidence that these proteins play an irnportant role i n uiuo. A recent study by Li et al. (2000) found that HUS patients develop the highest antibody titers to the Tir protein. There was little reaction to EspA, EspB and Eae in the acute-phase sera, although there was sorne reactivity to Tir. By day 8, antibody titers to all four proteins were detected in all patients, with a very strong response to Tir (a titer of up to 1:256 000). The antibody titers to all four proteins at 60 days were decreased, but were still highest to Tir. Based on these results, Li et al. (2000) suggested that Tir is a promising vaccine candidate and that it rnay also be a useful vaccine candidate for cattle.

Immune responses to E. coli 0157:H7 have the poten- tial to reduce colonization of the intestine by blocking bacterial adherence and A/E lesion formation. Antibodies to LI-'S were shown in one study to block adherence of STEC to human intestinal (Henle 407) epithelial cells (I-'aton et al., 197812). Hyperirnmune serum raised against Tir blocked EHEC attachrnent to

E. coli 0157:H7: intestinal colonization and virulence

HeLa cells (B. B. Finlay, personal communication), a cell line cornrnonly used to study A/E lesions induced by EI'EC and EHEC (Finlay et al., 1992; Rosenshine et al., 1996). Gl~aern-Wagl~arni et al. (2001) found that EspA, EspB, Tir and intirnin-p were the targets of long-lived humoral irnmune responses in Citrohacter rodentiurnn- infected mice. In vaccination studies involving subcutaneous or intranasal exposure to different intirnin subtypes, rnice developed type-specific irnmunity that modulated the outcome of C. rodentiurn infection (Ghaern-Magl~arni et al., 2001). These results supported the idea that intirnin could be effectively used as a corn- ponent of an EI'EC or EHEC vaccine.

Few studies have addressed the irnmune responses of cattle to E. coli 0157:H7. Johnson et al. (1996) reported on the anti-0157 serological responses of cattle that had been experirnentally infected in another study (Cray and Moon, 1995). Cray and Moon (1995) had inoculated four 8- to 10-month-old calves with 1 x 10lo c.f.u. of the hornologous E. coli 0157:H7 strain 13-14 weeks after the last positive fecal sarnple had been collected. All shed the challenge organisrn for at least 2 weeks after inoculation, but the calves inoculated only once shed significantly rnore bacteria than did those after a second inoculation. Cray and Moon (1995) concluded that previ- ous infection does not prevent reinfection by the sarne strain, although they suggested that an irnmune response to the organisrn rnay have resulted in lower levels of shedding on the second challenge with the hornologous strain. Johnson et al. (1996) reported that the anirnals in the Cray and Moon (1995) study had precl~allenge antibodies to E. coli 0157 LI'S, and a chal- lenge dose of 1 x 0'' c.f.u. induced 'prompt and sustained increases in serum antibodies to 0157 LI'S' The authors, reportedly, did not find any correlation between the titers to 0157 LI'S and protection against rechallenge. These data confirm that the cattle had been previously exposed to E. coli 0157:H7 with the develop- rnent of a detectable irnmune response to 0157 LI'S; however, the challenge dose exceeded the level of pro- tective irnmunity that had been acquired, and an arnnestic irnmune response to 0157 LI'S was elicited. The anirnals were not tested for antibodies to intirnin, Esp proteins or Tir.

Laegreid et al. (1998) studied the prevalence of both anti-0157 antibodies in the serum and E. coli 0157:H7 in the feces of naturally infected range beef cattle. In all herds studied, the prevalence of these antibodies was high, 63-100% of individuals within herds being seropositive. The prevalence of E. coli 0157:H7 in the feces was also high; the authors did not address whether anti-0157 antibodies provided any degree of protection.

Wray et al. (2000) found that three of six 13- to 30- day-old calves developed an increased anti-0157 IgG response following experirnental oral inoculation with 2 X 10'"~f.u. E. coli 0157:H7. By contrast, none of six adult Friesian cows developed a significant increase in

their serum anti-0157 IgG levels following oral inocula- tion with 1 C lo9 c.f.u. of the sarne strain, although one cow did develop an increased serum anti-0157 IgM level. The authors noted that two of the six calves were found to be naturally infected with a wild-type E. coli 0157:H7 strain prior to challenge, and that the carriage of this strain did not appear to reduce shedding of the challenge inoculum. These authors concluded that 'local irnmunity to E. coli 0157 rnay not develop to any degree in the intestine and that immunization to reduce fecal shedding of this organisrn rnay not be effective'. However, these authors did not report on the antibody titers to Tir or any of the other secreted proteins known to rnediate A/E lesions.

Sanderson et al. (1999) suggested that cattle rnay develop 'sorne level of irnmunity that lirnits shedding' (following infection) as an alternative hypothesis to one stating that the 'bacterial flora of older calves rnay make thern rnore resistant to colonization'. These authors stud- ied the length of fecal shedding in four 1-week-old calves following three separate oral inoculations of E. coli 0157:H7. The calves progressively shed fewer organisms following each challenge. Although it was suggested that sorne level of irnmunity develops upon infection that lirnits shedding in subsequent exposures, irnmune responses to E. coli 0157:H7 were not reported in this study.

Although relatively little is currently known about the irnmune responses of cattle to E. coli 0157:H7 infection, there is scientific evidence that vaccination of cattle with antigenic bacterial proteins involved in colonization can significantly reduce fecal shedding and prevalence of E. coli 0157:H7 in cattle. Several bacterial proteins could potentially serve as vaccine targets; type 111 secreted pro- teins and intirnin are currently being investigated. Vaccination with E. coli 0157:H7 type I11 secreted pro- teins were recently shown to significantly reduce the numbers of E. coli 0157:H7 shed in the feces, the num- bers of anirnals that shed the organism, and the duration of shedding in experirnentally challenged cattle (I'otter et al., 2004). In the sarne study, vaccination of cattle with E. coli 0157:H7 type I11 secreted proteins signifi- cantly reduced the prevalence of E. coli 0157:H7 in a clinical trial conducted in a typical feedlot setting (I'otter et al., 2004).

Vaccination of pregnant sows with intirnin frorn E. coli 0157:H7 induced high intirnin-specific irnmune responses in the serum and colostrum, and suckling neonatal piglets had reduced bacterial colonization and intestinal lesions following experirnental challenge (Dean-Nystrorn et al., 2002). Recently, a proof-of-con- cept study for a plant-based intirnin vaccine intended for cattle was conducted in a mouse rnodel (Judge et al., 2004). Transgenic tobacco plant cells that express the carboxy-terminal host cell-binding dornain of E. coli 0157:H7 intirnin were developed. Mice were either irnmunized intraperitoneally with intirnin expressed frorn

Rodney A. Moxley

the plant cells, fed transgenic plant cells, or both. These mice developed an intirnin-specific mucosal immune response when primed parenterally and then boosted orally, and also exhibited a reduced duration of E. coli 0157:H7 fecal shedding after challenge.

Collectively, these studies suggest that vaccination may be a useful preharvest intervention strategy for reduction of the prevalence of E. coli 0157:H7 in cattle. Future research to elucidate factors involved in E. coli 0157:H7 colonization of the intestinal tracts of these ani- rnals should also help in the development of other intervention strategies.

Effects of diet and gastrointestinal cell renewal on intestinal colonization in ruminants

The effects of diet on E. coli 0157:H7 intestinal colo- nization in ruminants have been both an avenue for the development of intervention strategies and a subject of controversy. An issue related to diet in ruminants is its effects on acid resistance of E. coli 0157:H7 and other STEC (Russell et al., 2000). The acid-resistance proper- ties of this organisrn are thought to contribute to the low infectious dose for humans, which has been estimated to be fewer than 700 cells, and possibly as low as 10 cells (Armstrong et al., 1996; Lin et al., 1996; Shallow et al., 1997; Tuttle et al., 1999). The mean pH of the human stornach is typically 2.0, and the mean ingesta residence time in the stornach is about 1.5 hours (Russell et al., 2000). The mean pH of the abornasum of rumi- nants is typically 2.5-3.0 (Russell et al., 2000). Most isolates of E. coli, including non-pathogenic and labora- tory strains, are resistant to pH 2.5 for at least 2 hours, and several different rnechanisrns of acid resistance are found in these organisms (Gordon and Small, 1993; Lin et al., 1997). These rnechanisrns include an acid-induced oxidative systern, an acid-induced arginine-dependent systern and a glutarnate-dependent systern, and are also operative in E. coli 0157:H7 (Lin et al., 1997).

Although acid-inducible systerns are present, sus- tained acid tolerance in E. coli 0157:H7 is not dependent upon adaptation, but is induced by stationary phase or starvation, and this process requires RpoS (Cheville et al., 1996). RpoS is required for the E. coli 0157:H7 oxidative acid-resistance systern to function, and for survival of this organisrn in dry, fermented sausage (Cheville et al., 1996; Lin et al., 1997). I-'rice et al. (2000) inoculated calves with both a wild-type E. coli 0157:H7 and an rpoS mutant derivative of this strain; the fecal shedding levels of the rpoS mutant were 10- to 100-fold lower than that of the wild-type strain (I-' S 0.05). These authors concluded that RpoS plays a role in E. coli 0157:H7 shedding in calves, possibly by inducing expression of the glucose-repressed RpoS- dependent acid-resistance determinant, and thus increasing resistance to gastrointestinal stress. Lin et al.

(1997) hypothesized that the gastrointestinal environ- ment of cattle will induce one or more of these acid-resistance systerns in E. coli 0157:H7 and, hence, affect their acid resistance prior to introduction into food. They also found that once activated, the acid- resistance systerns will rernain active over prolonged periods of cold storage (Lin et al., 1997).

The diets cattle are fed can affect the populations of acid-resistant E. coli found in the gastrointestinal tract and feces of these anirnals (Russell et al., 2000). Cereal grains are fermented in the rumen at a faster rate than fiber, and significant arnounts of starch pass undigested to the colon and are fermented in this location (McAllister et al., 1990; Russell et al., 2000). Increased arnounts of cereal grains in the diet results in an increase in the arnount of starch in the colon, an increase in the production and accumulation of colonic volatile fatty acids (VFA), reduced colonic and fecal pH, and an induction of acid resistance in resident E. coli populations (Diez-Gonzalez et al., 1998; Diez-Gonzalez and Russell, 1999; Scott et al., 2000; Fu et al., 2003). Diez-Gonzalez and Russell (1999) reported that the 'extreme acid resistance' of E. coli [i.e. resistance to acute mineral acid (HC1) shock] is highly correlated with the concentration of undissociated volatile fatty acids in the growth medium, and this result indicated that pH itself was not the inducer. These results were supported by Fu et al. (20031, who conducted an i n uitro study and found that the acid resistance (i.e. percent viability after acid shock) of a fecal E. coli and E. coli 0157:H7 in pure culture showed quadratic responses to increased acetate and butyrate concentrations at pH 7.2. Acid resistance of E. coli was induced by acetate and butyrate, even though the environrnental pH was near neutral. Similar results were measured in an i n uiuo study, where via- bility after acid shock was more dependent on VFA concentration than on pH. Results from these studies demonstrated that culture pH and VFA affect acid resistance of E.coli (Fu et al., 2003).

Diez-Gonzalez et al. (1998) reported that hay feeding, which reduces the arnount of undigested starch in the colon, decreased the numbers of acid-resistant E. coli in the rumen and colon. Cattle fed a 90% grain diet had l0"fold more acid-resistant E. coli than cattle fed hay, but a brief period (5 days) of hay feeding decreased the acid-resistant count substantially (by 6 logs per gram) (Diez-Gonzalez et al., 1998). Based on these results, these authors hypothesized that if cattle were given a brief period of hay feeding irnrnediately before slaugh- ter, it would significantly reduce the risk of food-borne E. coli infection. Other researchers have found similar results when cattle are switched from high-concentrate to hay rations. Scott et al. (2000) found that in steers switched from a high concentrate ration to 5 days of alfalfa hay feeding, the acid-resistant E. coli populations decreased from 10 000 to 20 viable cells per gram in the feces; E. coli 0157:H7 populations were not measured

E. coli 0157:H7: intestinal colonization and virulence

in this study (Scott et al., 2000). However, Keen et al. (1999) conducted a study in 200 head of beef cattle that were divided into two groups; one was fed grain and the other was switched to hay. Fifty-two percent of the cattle that were rnaintained on grain shed E. coli 0157:H7, but only 18% of the cattle that were switched to hay were E. coli 0157:H7-positive ( P < 0.05). These results supported the previous report by Diez-Gonzalez et al. (1998) suggesting that hay feeding could poten- tially reduce EHEC entry into the food chain (Callaway et al., 2003). Krause et al. (2003) also studied the effects of diet on the numbers of naturally occurring E. coli in the intestinal tracts of cattle; they found that E. coli was more numerous in anirnals receiving a high-grain diet in anirnals on a high-forage diet ( P < 0.051, and the highest numbers were in the feces.

With regard to the effects of hay and grain feeding on the length of shedding of E. coli 0157:H7 in rumi- nants, different conclusions were found when sheep or cattle were experimentally inoculated with E. coli 0157:H7 (Kudva et al., 1995, 1997; Hovde et al., 1999; Tkalcic et al., 2000). In a study in 1- to 2-year-old Holstein steers inoculated with lo1' c.f.u. via a gastric tube placed directly in the rumen, hay-fed anirnals shed E. coli 0157 longer than did grain-fed anirnals, and, irrespective of diet, the bacteria were equally acid- resistant (Hovde et al., 1999). Russell et al. (2000) and Jarvis and Russell (2001) explained that these results occurred because of the differences in the pH gradient from the culture broth to the colon in the two different treatment groups (hay- and grain-fed anirnals), and also because E. coli grown aerobically in Lurid broth is rnarkedly resistant to mineral acid shock (exhibiting 'extreme acid resistance'). These authors hypothesized that in the hay-fed anirnals the pH gradient was negligi- ble and this allowed the cattle to shed for as long as 76 days; by contrast, in the grain-fed cattle the pH gra- dient from the broth to the colon was >1.5 units, and the cattle shed E. coli 0157:H7 for only 2-4 days (Russell et al., 2000). In another study, calves fed a high-roughage diet did not have increased or longer fecal shedding of E. coli 0157:H7 after inoculation corn- pared with calves fed a high-concentrate diet; however, two calves shedding the highest mean concentrations of E. coli 0157:H7 were both fed the high-concentrate diet (Tkalcic et al., 2000).

The effect of different grain diets on fecal shedding of E. coli 0157:H7 in experirnentally inoculated steers has also been studied. The number of anirnals that were cul- ture-positive for E. coli 0157:H7 during a 10-week postinoculation period was significantly higher for the barley-fed group than for the corn-fed group and the cottonseed and barley-fed group (Buchko et al., 2000). The fecal pH of the corn-fed anirnals was significantly lower than that of the anirnals fed the cottonseed and barley and barley diets. The authors hypothesized that this probably resulted in a less suitable environment for

E. coli 0157:H7 in the hindgut of the corn-fed anirnals (Buchko et al., 2000).

Although changing the diets of finishing cattle from grain to hay rnay reduce the shedding of acid-resistant E. coli in the feces of cattle, this strategy has rnarkedly detrirnental effects on cattle performance, and for this reason alternative intervention strategies have been sought; however, other diet-change strategies airned at reducing starch intake have had detrirnental effects on perforrnance as well. Folrner et al. (2003) conducted two experiments to evaluate the effects of three intervention strategies on the prevalence of E. coli 0157:H7 in feed- lot steers. In both experiments, 432 steers were assigned to one of 54 pens. Intervention strategies were two Lactobacillus acidophilus competitive exclusion (direct- fed microbial) products, and rnonthly pen cleaning. In the second experiment, a diet-change treatment was imposed prior to slaughter. Diet changes consisted of changing from a 33% high-moisture corn, 15% dry rolled corn, 40% wet corn gluten feed diet to a 44% corn bran and 44% wet corn gluten feed diet in a 2-day change period (Folmer et al., 2003). Alfalfa hay and supplement were included in both diets at rates of 7 and 5% respec- tively. No differences in perforrnance or carcass yield were observed for the competitive exclusion products or the pen-cleaning treatments cornpared with the control. However, changing the finishing diet prior to slaughter had large negative effects on steer performance. The 14- day diet change at the end of the feeding period significantly ( P < 0.01) decreased average daily gain by 5.6%, decreased hot carcass weight by 2.3%, decreased feed conversion by 5.1% and decreased diet change period dry matter intake by 10%. In addition, the diet change tended to decrease the rnarbling score. A non- significant decrease in the prevalence of E. coli 0157:H7 was observed with the direct-fed microbials, which was interpreted to be due to lack of power, because with additional observations a statistically significant reduc- tion in prevalence was detected (Smith et al., 2004b).

Magnuson et al. (2000) conducted studies of the effects of diet and gastrointestinal cell epithelial prolifer- ation on clearance of E. coli 0157:H7 in cattle and sheep. The most significant finding was that the rates of cell proliferation in the gastrointestinal tracts of cattle, especially the lower gastrointestinal tract, were associ- ated with the duration of the period for which anirnals rernained E. coli 0157:H7 culture-positive. Cattle with slower rates of intestinal cell proliferation in the cecum and distal colon were culture-positive significantly longer than cohort cattle with faster cell proliferation rates. The authors suggested that the rates of intestinal epithelial cell proliferation could be affected by diet, but they were not able to show a difference in effect with grain- versus forage-based diets, nor did they find a dif- ference in the duration of E. coli 0157:H7 shedding between these diets. However, the authors did find that a fasting-refeeding regimen increased the intestinal

Rodney A. Moxley

epithelial cell proliferation rates in sheep, and that cattle subjected to this regimen were E. coli 0157:H7 fecal cul- ture-positive for a shorter time than continuously fed cattle. These findings suggested that feeding strategies can play a role in the length of E. coli 0157:H7 fecal shedding in cattle, although the immune status to E. coli 0157 LI-'S or the secreted proteins in this study were not reported.

Cray and Moon (1995) noted that 10%.f.u./g is the threshold for recognition of adherent layers of bacteria in histological sections, suggesting that A/E lesions will not be detected unless bacterial counts in the gut lumen are quite large. One could also envisage a situation in which the presence of natural mucosal irnmunity to A/E proteins rnight prevent bacterial adherence even when bacterial counts are transiently high in the gut lumen. Magnuson et al. (2000) suggested 'The fact that E. coli 0157:H7 can persist in the colonic digesta of some anirnals for many rnonths suggests that the bacteria associate with the mucosa in some way'. These authors suggested that the organisrn may perl~aps persistently colonize the crypts in some manner other than by A/E attachment.

Direct-fed microbials, bacteriophage and colicins as preharvest intervention strategies

Several research groups have found that direct-fed rnicrobials (DFM), lytic bacteriophages and colicins active against EHEC may reduce fecal shedding and prevalence of E. coli 0157:H7 in cattle and sheep; the most work to date has been done with DFM (Murinda et al., 1996; Kudva et al., 1999; Stevens et al., 2002b; Brashears et al., 2003a, b; Elarn et al., 2003; Srnith et al., 2004b). DFM have been defined as feed products that contain a source of live, naturally occurring microorgan- isms (Elarn et al., 2003). 'I-'robiotic' is a more generic term, and has been defined as a live rnicrobial feed sup- plement which beneficially affects the host anirnal by improving its intestinal rnicrobial balance (Elarn et al., 2003). DFMs, when used to cornpete with a particular organisrn are referred to as competitive exclusion prod- ucts (Brashears et al., 2003b). Competitive exclusion involves the use of live rnicrobial cultures that exhibit antagonistic effects against specific groups of organisms, resulting in a decrease in their numbers in the intestinal tract (Brashears et al., 2003b). The rnechanisrns by which DFMs, such as Lactobacillus spp., could poten- tially cornpete with E. coli 0157:H7 in the intestine are numerous. Ogawa et al. (2001) noted that lactobacilli produce VFA; can enhance specific and total IgA secre- tion when used as an oral adjuvant; enhance the secretion of specific antibodies against Stxl, Stx2 and STEC cells; and have other effects.

Brashears et al. (2003b) isolated different species of lactic acid bacteria from the feces of cattle and found several isolates that significantly inhibited E. coli

0157:H7 in agar spot tests, yet were acid and bile toler- ant. Brashears et al. (200322) conducted clinical trials in feedlot cattle and found that a Lactobacillus acidophilz~s strain (NI'C747) significantly decreased ( P < 0.01) the shedding of E. coli in the feces of individual cattle dur- ing the feeding period. E. coli 0157:H7 was approxirnately twice as likely to be detected in control anirnal sarnples as in sarnples from anirnals receiving strain NI-'C747. In addition, DFM supplernentation decreased ( P < 0.05) the number of E. coli 0157:H7-pos- itive hide sarnples at harvest and the number of pens testing positive for the pathogen. In a 2-year study involving 448 steers, Srnith et al. (2004b) found that cat- tle fed strain NIX747 were 35% less likely to shed E. coli 0157:H7 than cattle in untreated pens ( P = 0.002). Elarn et al. (2003) found that cattle fed L. acidophilz~s strain LA51 had reduced E. coli 0157 fecal shedding preva- lence ( P < 0.1) and positive hide sarnples ( P < 0.05) cornpared with controls.

Conclusion

Since its first recognition as a cause of outbreaks of food-borne illnesses 22 years ago, E. coli 0157:H7 has been investigated extensively. The irnportance of car- riage and shedding of the organisrn by cattle has been firmly established and much has been learned about the prevalence of this organisrn in cattle and the mecha- nisms by which it colonizes. Efforts are under way to develop interventions to reduce the extent of coloniza- tion of the intestine of cattle by this bacterium. These interventions include the use of probiotics, bacterio- phages, other feed additives and vaccination. The proteins secreted by the type 111 secretion system encoded by the locus for enterocyte effacement (LEE) are logical targets for vaccination of cattle and some suc- cess has been reported in vaccine trials involving experimentally challenged cattle and naturally exposed feedlot cattle. E. coli 0157:H7 possesses several hundred proteins that are not present in non-pathogenic E. coli and a large number of these are probably involved directly or indirectly in survival of the bacterium in its major reservoir. Understanding the roles of these pro- teins in colonization is a daunting task, but genornic approaches are likely to facilitate the process of discov- ery. As understanding improves, the chances of controlling shedding by cattle and subsequent infection of humans will increase.

Acknowledgments

This work is a contribution of the University of Nebraska Agricultural Research Division, Lincoln (journal series no. 14523). Support was provided by a grant from the USDA, NRICGI-' (2001-02966).

E. coli 0157:H7: intestinal colonization and virulence

References

Abc A, clc Grado M, Pfuctzncr RA, Sanchez-Sanmartin C, UcVinncy R, Pucntc JL, Strynaclka NC ancl Finlay UU (1999). Entcropathogcnic Escherichia coli translocated intinlin receptor, Tir, requires a specific chaperone for sta- ble secretion. :2clolecz~lar:Wic1~ohiolo~): 33: 1162-1175.

Abc H, Tatsuno I, Tolx T, Okutani A ancl Sasakawa C (2002). Uicarl,onatc ion stimulates the expression of locus of cntc- rocytc effacement-cncoclccl genes in cntcrohcnlorrhagic Escherichia coli Ol57:H7. Itzfectiotz atzd Inznzzltzity 70: 3500-3509.

Achcson UWK, Moore R, Uc Urcuckcr S, Linciconlc L, Jaccwicz M, Skutclsky E and Kcusch GT (1996). Translocation of Shiga toxin across polarized intcstinal cclls in tissue cul- ture. Itzfectiotz atzd Inznzzltzity 64: 3294-3300.

Achcson UWK, Linciconlc LL, Jaccwicz MS and Kcusch GT (1998). Shiga toxin interaction with intcstinal epithelial cclls. In: Kapcr JU and O'Uricn AD, editors. Escherichia coli O157:H7 atzd Other Shiga Toxitz-prodzlcitzg E. coli Straitzs. Washington, UC: ASM Press, pp. 140-147.

Adu-Uobic J, Frankcl G, Uain C, Goncalvcs AG, Trabulsi LR, Uoucc G, Knutton S and Uougan G (1998). Uctcction of intinlins a, p, y ancl 6, four intinlin clcrivativcs cxprcssccl by attaching ancl effacing microbial pathogcns. .[oz~nzal of Qitzical :Wicrohiolo~): 36: 662-668.

Arnlstrong GL, Hollingsworth J and Morris JG (1996). Enlcrging food-borne pathogcns: Escherichia coli Ol57:H7 as a n~oclcl of entry of a new pathogen into the food supply of the clcvclopcd \vorlcl. Epidenziologic Rez~ieu~.s 18: 29-51,

Uadca L, Doughty S, Nicholls L, Sloan J, Robins-Urownc RM and Hartland EL (2003). Contl.ibution of Efal/LifA to the adherence of cntcropathogcnic Escherichia coli to cpithc- lial cclls. Microbial Pathogetze.sis 34: 205-215.

Uachlcr AA ancl Moxlcy RA (2000). Escherichia coli Ol57:H7 induces attaching-effacing lesions in largc intcstinal mucosal cxplants from adult cattlc. FE1WLC :Wicrohiolo~): Lette?,.~ 185: 239-242.

Uachlcr AA ancl Moxlcy RA (2002). Effect of culture conditions on Escherichia coli Ol57:H7-mccliatccl attaching-effacing lesions in a bovine largc intcstinal mucosal cxplant model. FEMC :Wicrohiolo~): Lette?,.~ 212: 107-1 10.

Uarkocy-Gallagher GA, Arthur TM, Siragusa GR, Kccn JE, Elder RO, Lacgrcicl WW ancl Koohn~araic M (2001). Genotypic analyses of Escherichia coli Ol57:H7 ancl 0157 nonnlotilc isolates recovered from beef cattlc ancl carcasses at processing plants in the Midwestern states of the IJnitcd States. Applied atzd Etzz~irotznzetztal :Wicrohiolo~): 67: 3810-3818.

Uarkocy-Gallaghcr GA, Ucrry ED, Rivera-Uctancourt M, Arthur TM, Nou X and Koohn~araic M (2002). Ucvclopnlcnt of n~cthocls for the recovery of Escherichia coli Ol57:H7 ancl Salnzotzella from beef carcass sponge sanlplcs ancl bovine fccal ancl hide samples. .[oz~nzal of Food Protectiotz 65: 1527-1534.

Ucrin MC, Uarfcuillc-Michaucl A, Egan LJ, Miyanloto Y and Kagnoff MF (2002). Role of EHEC Ol57:H7 virulence fac- tors in the activation of intcstinal cpithclial cclls NF-KU ancl MAP kinasc pathways ancl the uprcgulatccl expression of interleukin-8. Cellz~lar:Wicrohiolo~): 4: 635-647.

Ucsscr TE, Hancock UU, Pritcharcl LC, McRac EM, Rice UH and Tarr PI (1997). Duration of detection of fccal excretion of Escherichia coli Ol57:H7 in cattlc. .[oz~nzal of Itz fectiozls Disease.s 175: 726-729.

Uiclaszcwska M, Janda J, Ulahova K, Minarikova H, Jikova E, Karn~ali MA, Laubova J, Sikulova J, Preston MA, Khakhria R, Karch H, Klazarova H and Nyc 0 (1997). Human Escherichia coli Ol57:H7 infection associated with the

consumption of unpastcurizcd goat's milk. EpidenzioloS): atzd Itzfectiotz 119: 299-305.

Ulanco M, Ulanco J, Ulanco JE and Ranlos J (1993). Entcrotoxigcnic, vcrotoxigcnic, and nccrotoxigcnic Escherichia coli isolated from cattlc in Spain. Anlcrican Journal of Veterinary Research 54: 1446-1451.

Urashcars MM, Galycan ML, Loncragan GH, Mann JE and Killingcr-Mann K (2003a). Prevalence of Escherichia coli Ol57:H7 and pcrforn~ancc by beef feedlot cattlc given Lactobacillus direct-fed micro1,ials. .[oz~nzal of Food Protectiotz 66: 748-754.

Urashcars MM, Jaroni U and Trimblc J (20031>). Isolation, sclcc- tion, ancl characterization of lactic acid bacteria for a conlpctitivc exclusion product to reduce shcclcling of Escherichia coli Ol57:H7 in cattlc. .[oz~nzal of Food Protectiotz 66: 355-363.

Urown CA, Harnlon UG, Zhao T ancl Uoylc MP (1997). Expcrinlcntal Escherichia coli Ol57:H7 carriage in calves. Applied aad Etzz~irotznzetztal :Wicrohiolo~): 63: 27-32.

Urcucr T, Ucnkcl UH, Shapiro RL, Hall WN, Winnctt MM, Linn MJ, Ncin~ann J, Uarrctt TJ, Uictrich S, Uowncs FP, Toncy DM, Pcarson JL, Rolka H, Slutskcr L ancl Griffin PM (2001). A multistatc outbreak of Escherichia coli Ol57:H7 infections linked to alfalfa sprouts grown from contami- nated seeds. Enze~itzg Itzfectiozls Di.sease.s 7: 977-982.

Ur~mclcr W, Schnliclt H ancl Karch H. (1997). EspP, a novel extracellular scrinc protcasc of cntcrohacnlorrhagic Escherichia coli Ol57:H7 cleaves human coagulation fac- tor V. :2clolecz~lar:Wic1~ohiolo~): 24: 767-778.

Uuchko SJ, Hollcy RA, Olson WO, Gannon VP and Vcira UM (2000). The effect of different grain diets on fccal shccl- cling of Escherichia coli Ol57:H7 by steers. .[oz~nzal of Food Protectiotz 63: 1467-1474.

Uurlancl V, Shao Y, Pcrna NT, Plunkctt G, Sofia HJ ancl Ulattncr FR (1998). The conlplctc DNA sequence and analysis of the largc virulence plasnlicl of Escherichia coli Ol57:H7. Nzlcleic Acids Research 26: 4196-4204.

Callaway TR, Elder RO, Kccn JE, Anderson RC and Nisbct UJ (2003). Forage feeding to reduce prcharvcst Escherichia coli populations in cattlc, a review. .[oz~nzal of Dai?y Scietzce 86: 852-860.

Cantcy JR (1985). Shiga toxin-an expanding role in the patho- genesis of infectious discascs. .[oz~nzal of Itzfectiozls Disease.s 151: 766-771.

Chapn~an PA ancl Ackroycl HJ (1997). Farnlcd deer as a potcn- rial source of vcrocytotoxin-producing Escherichia coli Ol57:H7. Vete?-ltza?y Record 141: 314-315.

Chart H, Jenkins C, Snlith RR ancl Rowc U (1998). Serum anti- bodies to secreted proteins in patients infected with Escherichia coli 0157 ancl othcr VTEC. EpidenzioloS): atzd Itz fectiotz 120: 239-243.

Chcn X, Schauclcr S, Poticr N, Van Uorsscalacr A, Pclczcr I, Uasslcr UL and Hughson FM (2002). Structural idcntifica- tion of a bacterial quorum-sensing signal containing boron. Natzlm 415: 545-549.

Chcvillc AM, Arnold KW, Uuchricscr C, Chcng CM ancl Kaspar CW (1996). 1poS regulation of acid, hcat, ancl salt tolcr- ancc in Escherichia coli Ol57:H7. Applied aad Etzz~irotznzetztal :Wicrohiolo~): 62: 1822-1824.

Ching JCY, Jones NL, Ccponis PJM, Karn~ali MA and Shcrn~an PM (2002). Escherichia coli Shiga-like toxins induce apop- tosis and cleavage of poly(AUP-ribose) polynlcrasc itz zitro activation of caspascs. Itzfectiotz atzd Inznzzltzity 70: 4669-4677.

Cornick NA, Uoohcr SL ancl Moon HW (2000). Persistent colo- nization of sheep by E. coli Ol57:H7 ancl othcr pathotypcs of E. coli Applied aad Etzz ~irotznzetztal :Wicrohiolo~): 66: 4926-4934.

2 8

Cornick NA, Uoohcr SL ancl Moon HW (2002). Intinlin facili- tates colonization by Escherichia coli Ol57:H7 in adult ruminants. Infection atzd Inznzzltzity 70: 2704-2707.

Cray WC and Moon HW (1995). Expcrinlcntal infcction of calvcs ancl adult cattlc with Escherichia coli Ol57:H7. Applied aad Etzz~irotznzetztal :Wicrohioloqy 61: 1586-1590.

Crump JA, Sulka AC, Langcr AJ, Schabcn C, Criclly AS, Gage R, Uaysingcr M, Moll M, Withers G, Toncy DM, Hunter SU, Hockstra RM, Wong SK, Griffin PM and Van Gilder TJ (2002). An outbreak of Escherichia coli Ol57:H7 infcc- tions anlong visitors to a dairy farn~. New Etzglatzd.[oz~nzal of :Wedicitze 347: 6011-609.

Ucan-Nystrom EA, Uosworth UT, Cray Jr WC and Moon HW (1997). Pathogenicity of Escherichia coli Ol57:H7 in the intestines of neonatal calvcs. Infection atzd Inznzzltzity 65: 1842-1848.

Ucan-Nystrom EA, Uosworth UT, Moon HW, and O'Uricn AU (1998). Escherichia coli Ol57:H7 requires intinlin for cntcropathogcnicity in calvcs. Infection atzd Inznzzltzity 66: 4560-4563.

Ucan-Nystrom EA, Uosworth UT, O'Uricn AU and Moon HW (1999). Uovinc infcction with Escherichia coli Ol57:H7, In: Stewart CS and Flint HJ, editors. E. coli 0157 in Farm Atzinzals. CAUI, Wallingford, IJK, pp. 51-58,

Ucan-Nystmm EA, Ganshcroff LJ, Mills M, Moon HW and O'Uricn AU (2002). Vaccination of pregnant clanls with intiminoli7 protects suckling piglets from Escherichia coli Ol57:H7 infcction. Itzfectiotz atzd Inznzzltzity 70: 2414-2418.

Ucibcl C, Krimcr S, Chakralmrty T and Ebcl F (1998). EspE, a novel secreted protein of attaching ancl cffacing bacteria, is directly translocatccl into infected host cclls, where it appears as a tyrosinc-phosphorylatcd 90-kUa protein. :2clolecz~lar:Wic1~ohioloqy 28: 463-474.

UcVinncy R, Stein M, Rcischcicl U, Al,c A, Ruschkowski S ancl Finlay UU (1999). Entcrohcnlorrhagic Escherichia coli Ol57:H7 produces Tir, which is translocatccl to the host cell membrane but is not tyrosinc phosphorylatccl. Infection atzd Imnzzltzity 67: 2389-2398.

Uicz-Gonzalcz F and Russell JU (1999). Factors affecting the cxtrcnlc acid resistance of Escherichia coli Ol57:H7. Food :Wicrohioloqy 16: 367-374.

Uicz-Gonzalcz F, Callaway TR, Kizoulis MG and Russell JU (1998). Grain feeding and disscnlination of acid-resistant Escherichia coli from cattlc. Scietzce 281: 1666-1668.

Ujafari S, Ebcl F, Ucibcl C, Krimcr S, Hudcl M ancl Chakraborty T (1997). Characterization of an exported protcasc from Shiga toxin-producing Escherichia coli. :2(lolecz~lar :Wicrohioloqy 25: 771-784.

Uonncnbcrg MS ancl Whittan1 TS (2001). Pathogenesis ancl cvo- lution of virulence in cntcropathogcnic ancl cntcrohcnlorrhagic Escherichia coli .[oz~nzal of Qitzical Itzz~e.stigation 107: 539-548.

Uonncnbcrg MS, Gir6n JA, Nataro JP and Kapcr JU (1992). A plasmicl-cncoclccl typc IV fimbrial gene of cntcropatho- genic Escherichia coli associated with localized adherence. :2clolecz~lar:Wic1~ohioloqy 6: 342773437,

Uonncnbcrg MS, Tzipori S, McKcc M, O'Uricn AU Alroy J and Kapcr J (1993). The role of the eae gene of cntcrohcmor- rhagic E. coli in intin~atc attachnlcnt in zitro ancl in a porcine model. .[of1 ma1 of Qitzical Itzz ~e.$tigatiotz 92: 1418-1424.

Uuffy G (2003). Vcrocytotoxigcnic Escherichia coli in aninla1 facccs, manures ancl slurries. .[oz~nzal of Applied :Wicrohioloqy 94 (Supplement): 94s-103S.

Ebcl F, Ucibcl C, Krcssc AIJ, GuzmAn CA and Chakraborty T (1996). Temperature- ancl medium-clcpcnclcnt secretion of proteins by Shiga toxin-producing Escherichia coli. Infection atzd Imnzzltzity 64: 4472-4479.

Rodney A. Moxley

Ebcl F, Podzadcl T, Rohdc M, Krcssc AIJ, Kranlcr S, Ucibcl C, GuzmAn CA ancl Chakraborty T (1998). Initial binding of Shiga toxin-producing Escherichia coli to host cclls ancl subscqucnt induction of actin rcarrangcnlcnts clcpcncl on filamentous EspA-containing surface appendages. :2clolecz~lar:Wic1~ohioloqy 30: 147-161.

Elan1 NA, Glcghorn JF, Rivera JU, Galycan ML, Ucfoor PJ, Urashcars MM ancl Yo~mts-Uahl SM (2003). Effects of live cultures of Lactohacillzls acidophihls (strains NP44 ancl NP5 1) ancl Propiotzihacterizlnz fiezldetzreichii on pcrforn~- ancc, carcass, ancl intcstinal characteristics, ancl Escherichia coli 0157 shcclcling of finishing beef stccrs. .[oz~nzal of Atzinzal Scietzce 81: 2686-2698.

Elder RO, Kccn, JE, Siragusa GR, Uarkocy-Gallagher AG, Koohn~araic M ancl Lacgrcicl WW (2000). Correlation of cntcrohcnlorrhagic Escherichia coli Ol57:H7 prevalence in feces, hides, ancl carcasses of beef cattlc during pro- cessing. Proceeditzgs of the Natiotzal Academy of Scietzce.~ of the 1Jtzited State.$ ofAmerica 97: 2999-3003.

Elliott SJ, Spcrandio V, Girhn, Shin S, Mcllics JL, Wainwright LA, Shutchcson SW, McUanicl TK ancl Kapcr JU (2000). The locus of cntcrocytc cffaccnlcnt (LEE)-cncodcd regulator (Lcr) controls expression of both LEE- and non-LEE- cncoclccl virulence factors in cntcropathogcnic ancl cntcrohcnlorrhagic Escherichia coli. Infection atzd Imnzzltzity 68: 61 15-6126.

Fcgan N, Higgs G, Vandcrlindc P and Ucsn~archclicr P (2003). Enumeration of Escherichia coli 0157 in cattlc facccs ~rsing nlost probable number technique ancl automated immunomagnctic separation. Lette?,.~ in Applied :Wicrohioloqy 38: 56-59.

Fcng P, Lanlpcl KA, Karch H and Whittan1 T (1998). Genotypic ancl phenotypic changes in the emergence of E.schm-ichia coli 0 1 57:H7. .[oz~nzal of In fectiozls Di.$ea.$e.s 177: 1750-1753.

Fcnwick UW and Cowan LA (1998). Canine n~oclcl of hemolytic-uremic syndrome. In: Kapcr JU and O'Uricn AD, editors. Escherichia coli 0 1 57:H7 atzd Other Shiga Toxitz- Prodzlcitzg E. coli Straitzx Washington, UC: ASM Press, pp. 268-277.

Finlay UU, Roscnshinc I, Uonncnbcrg MS and Kapcr JU (1992). Cytoskclctal conlposition of attaching ancl cffacing lesions associated with cntcropathogcnic Escherichia coli aclhcr- cncc to HcLa cclls. Infection atzd Inznzzltzity 60: 2541-2543.

Fischcr J, Zhao T, Uoylc MP, Goldbcrg MR, Urown CA, Scbvcll CT, Kavanaugh UM and Uauman CU (2001). Expcrinlcntal ancl field studies of Escherichia coli Ol57:H7 in white- tailed deer. Applied aad Etzzirotznzetztal :Wicrohioloqy 67: 1218-1224.

Fitzhcnry RJ, Pickarcl UJ, Hartland EL, Rcccc S, Uougan G, Phillips AU ancl Frankcl G (2002). Intinlin typc influences the site of human intcstinal mucosal colonization by cntcrohacnlorrhagic Escherichia coli Ol57:H7. Gzlt 50: 180-185.

Folnlcr J, Mackcn C, Moxlcy R, Snlith U, Brashcars M, Hinklcy S, Erickson G and Klopfcnstcin T (2003). Intervention strategies for reduction of E. coli Ol57:H7 in feedlot stccrs. In: IJnivcrsity of Nebraska Cooperative Extension, Institute of Agriculture and Natural Resources, IJnivcrsity of Nebraska-Lincoln, Lincoln (NE): 2003 Nebraska Uccf Report, pp. 22-23.

Frankcl G, Liclcr 0, Hcrshkoviz R, Mould AP, Kachalsky SG, Candy UCA, Cahalon L, Humphrics MJ ancl Uougan G (1996). The cell-binding clon~ain of intinlin from cntcropathogcnic Escherichia coli binds to beta1 intcgrins. .[oz~nzal of Biological Chenzisty 271: 20359-20364.

Frankcl G, Phillips AD, Roscnshinc I, Uougan G, Kapcr JU ancl Knutton S (1998). Entcropathogcnic ancl cntcrohacmor-

E. coli 0157:H7: intestinal colonization and virulence

rhagic Escherichia coli: nlorc subvcrsivc elements. :2clolecz~lar:Wic1~ohiolo~): 30: 91 1-921.

FLI CJ, Porter JH, Fclton EEU, Lchmkuhlcr JW and Kcrlcy MS (2003). Prc-harvest factors influencing the acid-resistance of Escherichia coli and E. coli Ol57:H7. .[oz~nzal ofAtzinzal Scietzce 81: 1080-1087.

Ganshcroff LJ and O'Uricn AU (2000). Escherichia coli Ol57:H7 in beef cattlc presented for slaughter in the 1J.S.: higher prcvalcncc rates than previously cstimatccl. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ of Anzerica 97: 2959-2961.

Ghacm-Maghami M, Sinlnlons CP, Uanicll S, Pizza M, Lewis U, Frankcl G ancl Uougan G (2001). Intimin-specific immune responses prevent bacterial colonization by the attaching- cffacing pathogen Citrohacter rodetztizlnz. Itzfectiotz atzd Inznzz~tzity 69: 5597-5605.

Gordon J and Sn~all PLC (1993). Acid resistance in cntcric bac- teria. Itzfectiotz atzd Inznzzltzity 61: 364-367.

Graukc LJ, Kudva IT, Yoon JW, H~mt CW, Willianls CJ and Hovclc CJ. (2002) Gastrointestinal tract location of Escherichia coli Ol57:H7 in ruminants. Applied aad Etzz~irotznzetztal :Wicrohiolo~): 68: 2269-2277.

Griffin PM (2003). Epidcnliology of 283 outbreaks of Escherichia coli Ol57:H7 infections in the 20th century, IJnitcd States. In: VIEC 2003: Proceeditzgs of the 5th Itztenzatiotzal Synzposizlnz atzd Workhop otz Shiga Toxitz (Verocytotoxitz)-P~,odz~citzg E~cherichia coli Itzfectiotzs, Editzhzqh, Scotlatzd,.[zltze &ll. Abstract no. 0-2, p. 17.

Griffin PM and Tauxc RV (1991). The cpiclcnliology of infcc- tions caused by Escherichia coli Ol57:H7, other cntcrohcnlorrhagic E. coli ancl the associated hcnlolytic ~rrcmic syndrome. Epidemiologic Rez~ieu,:s 13: 60-98.

Griffin PM, Mead PS, Van Gilder T, Hunter SU, Strockbinc NA ancl Tauxc RV (2000). Shiga toxin-producing E. coli infcc- tions in the IJnitcd States: current status ancl challenge. Itz: VIEC 2000: Proceeditzgs of the 4th Itztertzatiotzal Synzposizlnz atzd Workhop otz Shiga Toxitz (Verocytotoxitz)-P~,odz~citzg E~cherichia coli Itzfectiotzs, Kyoto, .[apatz, October 29-Mz~enzher 2. Alxtract no. PI-1, p. 31.

Hartlund EL, Uanicll SJ, Uclahay RM, Ncvcs UC, Wallis T, Shaw RK, Hale C, Knutton S and Frankcl G (2000). The typc I11 protein translocation systcnl of cntcropathogcnic Escherichia coli involves EspA-EspU protein interactions. :2clolecz~lar:Wic1~ohiolo~): 35: 1483-1492.

Hayashi T, Makino K, Ohnishi M, Kurokawa K, Ishii K, Yokoyan~a K, Han CG, Ohtsubo E, Nakayan~a K, Murata T, Tanaka M, Tobc T, Iida T, Takami H, Honda T, Sasakawa C, Ogasawara N, Yasunaga T, Kuhara S, Shiba T, Hattori M ancl Shinagawa H (2001). Conlplctc gcnonlc sequence of cntcrohcnlorrhagic Escherichia coli Ol57:H7 ancl gcnonlic conlparison with a laboratory strain K-12. DNA Research 8: 11-22.

Hcuvclink AE, van Hccr\vaardcn C, Zwartkruis-Nahuis JT, van Oostcronl R, Eclink K, van Uuynhovcn YT ancl clc Uocr E (2002). Escherichia coli 0157 infcction associated with a petting zoo. EpidenzioloS): atzd Itzfectiotz 129: 295-302.

Hocy UE, Curric C, Else RW, Nutikka A, Lingwoocl CA, Gally UL ancl Snlith UG (2002). Expression of receptors for vcrotoxin 1 from Escherichia coli 0157 on bovine intcs- tinal epithelium. .[oz~nzal of Medical :Wicrohiolo~): 51: 143-149.

Hocy UE, Sharp L, Curric C, Lingwoocl CA, Gally UL ancl Snlith UG (2003). Vcrotoxin 1 binding to intestinal crypt cpithc- lial cclls results in localization to lysoson~cs ancl abrogation of toxicity. Cell :Wicrohiolo~): 5: 85-97.

Hovclc CJ, Austin PR, Cloud KA, Willianls CJ and Hunt CW (1999). Effect of cattle diet on E~cherichia coli Ol57:H7

acid rcsistancc. Applied aad Etzz~irotznzetztal :Wicrohiolo~): 65: 3233-3235.

Janka A, Uiclaszcwska M, Uobrindt IJ ancl Karch H (2002). Identification ancl distribution of the cntcrohcnlorrhagic Escherichia coli factor for aclhcrcncc (e fal) gene in sor- bitol-fermenting Escherichia coli Ol57:H-. Itztenzatiotzal .[oz~nzal of Medical :Wicrohiolo~): 292: 207-214.

Jarvis KG and Kapcr JU (1996). Secretion of extracellular pro- teins by cntcrohcnlorrhagic Escherichia coli via a putative typc I11 secretion system. Itzfectiotz atzd Inznzzltzity 64: 4826-4829.

Jarvis GN and Russell JU (2001). Uiffcrcnccs in Escherichia coli culture conditions can have a large inlpact on the induc- tion of cxtrcnlc acid rcsistancc. Cz~n,etzt :Wicl,ohiolo~): 43: 215-219.

Jcrsc AE, YLI J, Tall UU and Kapcr JU (1990). A genetic locus of cntcropathogcnic Escherichia coli necessary for the pro- duction of attaching ancl cffacing lesions on tissue culture cclls. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ oJ the 1Jtzited State.$ ofAmerica 87: 7839-7843.

Johnson RP, Cray Jr WC ancl Johnson ST (1996). Serum anti- body responses of cattlc following cxpcrinlcntal infcction with Escherichia coli Ol57:H7. Itzfectiotz atzd Inznzzltzity 64: 1879-1883.

Judge NA, Mason HS and O'Uricn AU (2004). Plant cell-based intinlin vaccine given orally to nlicc prinlccl with intinlin reduces tinlc of Escherichia coli Ol57:H7 shcclcling in feces. Itz fectiotz atzd Inznzzltzity 72: 16%175.

Kasscnborg HU, Hcdbcrg CW, Hockstra, Evans MC, Chin AE, Marcus R, V~rgia UJ, Snlith K, Ucsai Ahuja S, Slutskcr L ancl Griffin PM (2004) Farm visits ancl unclcrcookccl ham- burgers as n~ajor risk factors for sporadic Escherichia coli Ol57:H7 infcction: data from a cascxontrol study in 5 FoodNct sites. Qitzical Itzfectiozls Diseases 38 (Supplement 3): 5271-5278.

Kccn JE and Elder RO (2002). Isolation of Shiga-toxigcnic Escherichia coli 0157 from hide surfaces ancl the oral cav- ity of finished beef fcccllot cattlc. .[oz~nzal of the Anzel-lcatz Vetel-ltzaly Medical Associatiotz 220: 756-763.

Kccn JE, IJhlich GA and Elder RO (1999). Effects of hay- and grain-based diets on fccal shcclcling of naturally-acquired cntcrohcnlorrhagic E. coli (EHEC) Ol57:H7 in beef fcccllot cattlc. 80th Conference of Research Workers in Aninla1 Uiscascs, Abstract s'86, Nov. 7-9, 1999, Chicago, IL.

Kccnan KP, Sharpnack UU, Collins H, Forn~al SU ancl O'Uricn AU (1986). Morphologic evaluation of the effects of Shiga toxin ancl E. coli Shiga-like toxin on the rabbit intestine. Anlcrican Journal of Pathology 125: 69-36.

Kcnny U and Finlay UU (1995). Protein secretion by cntcropathogcnic Escherichia coli is essential for transcluc- ing signals to epithelial cclls. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ of Anzerica 92: 7991-7995.

Kcnny U, UcVinncy R, Stein M, Rcinschcicl UJ, Frcy EA and Finlay UU (1997). Entcropathogcnic Escherichia coli (EPEC) transfers its receptor for intin~atc aclhcrcncc into n~anln~alian cclls. Cell 91: 51 1-520.

Kcusch GT, Grady GF, Mata LJ and McIvcr J (1972). The patho- genesis of Shigclla diarrhea. I. Entcrotoxin production by Shigella dy.setzteriae 1. .[oz~nzal of Qitzical Itzz~e.stigatiotz 51: 2-3.

Khaitsa ML, Snlith UR, Stoner JA, Parkhurst AM, Hinklcy S, Klopfcnstcin TJ and Moxlcy RA (2003). Incidence, dura- tion, ancl prcvalcncc of Escherichia coli Ol57:H7 fccal shcclcling by fcccllot cattlc during the finishing period. .[oz~nzal of Food Protectiotz 66: 1972-1977.

Kin1 J, Nictfcldt J and Ucnson AK (1999). Octamcr-based gcnonlc scanning distinguishcs a ~miquc subpopulation of

Rodney A. Moxley

Escherichia coli Ol57:H7 in cattlc. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ of Anzel-lca 96: 13288-13293.

Klappmth JM, Scalctsky ICA, McNan~ara UP, Lai LC, Malstronl C, Janlcs SP and Uonncnlxrg MS (2000). A large toxin from pathogenic Escherichia coli strains that inhibits lynlphocytc activation. Itzjectiotz atzd Inznzzltzity 68: 214%2155.

Knutton, S, Roscnshinc I, Pallcn MJ, Nisan I, Ncvcs UC, Uain C, Wolff C, Uougan G and Frankcl G (1998). A novcl EspA- associated surface organelle of cntcropathogcnic E.sche?-ichia coli involved in protcin translocation into epithelial cclls. Ezlropeatz :2(lolecz~lar Biology Oqatzizatiotz .[oz~nzal 17: 2166-2176.

Krausc DO, Snlith WJM, Conlan LL, Gough JM, Willianlson MA and McSwccncy CS (2003). Uict influences the ecology of lactic acid bacteria ancl Escherichia coli along the digestive tract of cattlc: neural networks ancl 16s rUNA. :Wicrohioloqy 149: 57-65.

Krcssc AIJ, Schulzc K, Ucibcl C, Ebcl F, Rohdc M, Chakraborty T and Guzman CA (1998). Pas, a novcl protcin required for protcin secretion ancl attaching ancl effacing activities of cntcrohcnlorrhagic Escherichia coli. .[oz~nzal of Bacterioloqy 180: 4370-4379.

Kudva IT, Hatficld PG and Hovclc CJ (1995). Effect of diet on the shcclcling of Escherichia coli Ol57:H7 in a shccp model. Applied aad Etzz~irotznzetztal :Wicrohioloqy 61: 1363-1370.

Kudva IT, Hunt CW, Willianls CJ, Nancc IJM and Hovclc CJ (1997). Evaluation of dietary influences on Escherichia coli Ol57:H7 shcclcling by shccp. Applied aad Etzz~irotznzetztal :Wicrohioloqy 63: 3878-3886.

Kudva IT, Jclacic S, Tarr PI, Youndcrian P and Hovclc CJ (1999). Uiocontrol of E~cherichia coli 0157 with 0157- specific bacteriophages. Applied aad Etzz~irotznzetztal :Wicrohioloqy 65: 3767-3773.

Lacgrcid WW, Hoffn~an M, Kccn J, Elder R and Kwang J. (1998). Ucvclopnlcnt of a blocking enzyme-linked immunosorl~cnt assay for detection of serum antibodies to 0157 antigen of Escherichia coli. Qitzical atzd Diagtzostic Inznzzltzoloqy 5: 242-246.

Lacgrcid Elder RO and Kccn JE (1999). Prevalence of Escherichia coli Ol57:H7 in range beef calves at weaning. Epidenzioloqy atzd Itz fectiotz 123: 291-298.

LcJc~mc JT, Ucsscr TE, Rice UH, Ucrg JL, Stilborn RP and Hancock UU (2004). Longitudinal study of Escherichia coli Ol57:H7 in feedlot cattlc: prcdonlinancc and persistence of specific clonal types despite n~assivc cattlc population turnover. Applied aad Etzz~irotznzetztal :Wicrohiology 70: 377-384.

Lcvinc MM (1987). Escherichia coli that cause diarrhea: cntcro- toxigcnic, cntcropathogcnic, cntcroinvasivc, cntcrohcnlorrhagic ancl cntcroaclhcrcnt. .[oz~nzal of Itz fectiozls Disease.$ 155: 377-389.

Li Y, Frcy E, Mackcnzic AMR and Finlay UU (2000). Human response to Escherichia coli Ol57:H7 infection: antibodies to sccrctccl virulence factors. Itzfectiotz atzd Inznzzltzity 68: 5090-5095.

Lin J, Snlith MP, Chapin KC, Uaik HS, Ucnnctt GN ancl Foster JW (1996). Mcchanisnls of acid resistance in cntcrohcmor- rhagic Escherichia coli. Applied aad Etzz~irotznzetztal :Wicrohioloqy 62: 3094-3100.

Ling\vood CA, Mylvagananl M, Arab S, Khinc AA, Magnusson G, Grinstcin S and Nyholnl PG (1998). In: Kapcr JU and O'Uricn AD, editors. Escherichia coli 0 1 5ZH7 atzd Other Shiga Toxitz-Prodzlcitzg E. coli Straitzs Washington, UC: ASM Prcss, pp. 129-139.

Magnuson UA, Davis M, Hubclc S, Austin PR, Kudva IT, Willianls CJ, Hunt CW and Hovclc CJ (2000). Ruminant

gastrointestinal cell proliferation ancl clearance of Escherichia coli 0 1 57:H7. Itz fectiotz atzd Inznzzltzity 68: 3808-3814.

Malstrom C ancl Janlcs S (1998). Inhibition of murinc splenic ancl mucosal lynlphocytc f~rnction by cntcric bacterial products. Itzfectiotz atzd Inznzzltzity 66: 3120-3127.

Marocska U, tc Loo WM, Monncns LAH, van dcr Vcldcn TJAM, Vcrnlccr MA, Prcycrs F, Ucn~ackcr PNM, van den Hcuvcl LPWJ ancl van Hinsbcrgh VWM (2000). Binding ancl trans- fer of vcrocytotoxin by polymorphonuclcar leukocytes in hcnlolytic uremic syndrome. Blood 95: 3396-3402.

Marocska U, tc Loo WM, van Hinsbcrgh VWM, van den Hcuvcl LPWJ and Monncns LAH (2001). Uctcction of vcrocyto- toxin bound to circulating polymorphonuclcar leukocytes of patients with hcnlolytic uremic syndrome. .[oz~nzal of the Anzel-lcatz Society of Nephroloqy 12: 800-806.

McAllistcr TA, Chcn KJ, Rohdc LM, Forsbcrg CW and Costcrton JW (1990). Digestion of barley, n~aizc ancl wheat by sclcctccl species of ruminal bacteria. Applied aad Etzz~irotznzetztal :Wicrohiology 56: 31 46-31 53.

McEvoy JM, Uohcrty AM, Shcriclan JJ, Thomson-Carter FM, Garvcy P, McGuirc L, Ulair IS and McUowcll UA (2003). The prevalence ancl sprcacl of Escherichia coli Ol57:H7 at a conlnlcrcial beef abattoir. .[oz~nzal of Applied :Wicrohioloqy 95: 256-266.

McKcc ML, Mclton-Cclsa AR, Moxlcy RA, Francis UH and O'Uricn AU (1995). E.sche?-Ichia coli 0 1 57:H7 requires intinlin to col- onize the gnoto1,iotic pig intestine ancl to adhere to HEp-2 cclls. Itzjectiotz atzd Inznzzltzity 63: 3739-3744.

McNally A, Roc AJ, Sinlpson S, Thomson-Carter FM, Hocy DEE, Curric C, Chakraborty T, Snlith UGE ancl Gally UL (2001). Uiffcrcnccs in levels of sccrctccl locus of cntcrocytc cffaccnlcnt proteins between human disease-associated ancl bovine Escherichia coli Ol57:H7. Itzfectiotz atzd Inznzz~tzity69: 5107-5114.

Mead PS, Slutskcr L, Uictz V, McCaig LF, Urcscc JS, Shapiro C, Griffin PM and Tauxc RV (1999). Food-related illness and clcath in the IJnitcd States. Enze~itzg Itz fectiozls Di.$ease.s 5: 607-625.

Mcllics JL, Elliott SJ, Spcrandio V, Uonncbcrg MS and Kapcr JU (1999). The Per rcgulon of cntcropathogcnic Escherichia coli: identification of a regulatory cascade ancl a novcl transcriptional activator, the locus of cntcrocytc cffacc- nlcnt (LEE)-cncodcd regulator (Lcr). :2(lolecz~lar :Wicrohioloqy 33: 296-306.

Mclton-Cclsa AR and O'Uricn AU (1998). Structure, biology and relative toxicity of Shiga toxin fanlily members for cclls and animals. In: Kapcr JU and O'Uricn AD, editors. Escherichia coli 015ZH7 atzd Other Shiga Toxitz- Prodzlcitzg E. coli Straitzx Washington, UC: ASM Prcss, pp. 121-128.

Mcngc C, Wiclcr LH, Schlapp T ancl Ualjcr G (1999). Shiga toxin 1 from Escherichia coli blocks activation ancl prolif- eration of bovine lynlphocytc subpopulations itz z~itro. Itzfectiotz atzd Inznzzltzity 67: 2209-2217.

Moxlcy RA (2003). Uctcction ancl diagnosis of Escherichia coli Ol57:H7 in food-producing animals. In: Torrcncc ME and Isaacson RE, editors. Microbial Food Safety itz Atzinzal Ag?,iczlltzlre: Cz~rretzt Topics. AIIICS, Iowa: Iowa State IJnivcrsity Prcss, pp. 143-154.

Moxlcy RA and Francis UH (1998). Overview of aninla1 mod- els. In: Kapcr JU and O'Uricn AD, editors. Escherichia coli 0 1 5 7:H7 atzd Other Shiga Toxitz-Prodzlcitzg E. coli Straitzs Washington, UC: ASM Prcss, pp. 249-260.

Murinda SE, Roberts RF and Wilson RA (1996). Evaluation of colicins for inhibitory activity against cliarrhcagcnic Escherichia coli strains, including scrotypc Ol57:H7. Applied aad Etzz~irotznzetztal :Wicrohioloqy 62: 3196-3202.

Rodney A. Moxley

Roscnshinc I, Ruschkowski S ancl Finlay UU (1996). Expression of attaching/cffacing activity by cntcropathogcnic Escherichia coli clcpcncls on growth phase, temperature, ancl protein synthesis upon contact with cpithclial cclls. Infection atzd Imnzzltzity 64: 966-973.

Russell JU, Uicz-Gonzalcz F and Jarvis GN (2000). Effects of diet shifts on Escherichia coli in cattlc. .[oz~nzal of Daily Scietzce 83: 863-873.

Sandcrson, MW, Gay JM, Hancock UU, Gay CC, Fox LK and Ucsscr TE (1995). Sensitivity of bacteriologic culture for detection of Escherichia coli Ol57:H7 in bovine fcccs. .[oz~nzal of Qitzical :Wicrohiology 33: 2616-2619.

Sandcrson, MW, Ucsscr TE, Gay JM, Gay, CC and Hancock UU (1999). Fecal Escherichia coli Ol57:H7 shcclcling patterns of orally inoculated calves. Vetel-ltzaly :Wicrohiology 69: 199-205.

Sargcant JM, Sandcrson MW, Snlith RA and Griffin UU (2003). Escherichia coli 0157 in fcccllot cattlc fcccs ancl water in four n~ajor feeder-cattle states in the IJSA. Prez~etztiz~e Vetel-ltzaly 1Wedicilze 61: 127-135.

Schauclcr S, Shokat K, Surcttc MG ancl Uasslcr UL (2001). The LuxS fanlily of bacterial autoinduccrs: biosynthcsis of a novcl quorum-sensing signal molecule. :2clolecz~lar :Wicrohiology 41: 463-476.

Scott T, Wilson C, Uailcy U, Klopfcnstcin T, Milton T, Moxlcy R, Snlith U, Gray J ancl Hungcrford L (2000). Influence of diet on total ancl acid-resistant E. coli ancl colonic pH. In: IJnivcrsity of Nebraska Cooperative Extcnion, Institute of Agriculture and Natural Resources, IJnivcrsity of Ncbraska- Lincoln, Lincoln (NE): 2000 Nebraska Uccf Report, pp. 39-41.

Shallow S, Daily P ancl Rothrock G (1997). Foodborne cliscascs active surveillance network 1996. :2(lorhidity 12clortality Weekly Report 46: 258-261.

Sicglcr RL, Pyshcr TJ, Tcsh VL and Taylor Jr FU (2001). Response to single ancl cliviclccl closes of Shiga toxin-1 in a prin~atc nloclcl of hcnlolytic uremic syndrome. .[oz~nzal of the Anzel-lcatz Society of Nephrology 12: 1458-1467.

Sinclair JF and O'Uricn AU (2002). Cell surface-localized nuclc- olin is a cukaryotic receptor for the aclhcsin intimin-y of cntcrohcnlorrhagic Escherichia coli Ol57:H7. .[oz~nzal of Biological Chenzistly 277: 2876-2885.

Snlith UR, Ulackforcl M, Younts S, Moxlcy R, Gray J, Hungcrford L, Milton T and Klopfcnstcin T (2001). Ecological relationships between the prcvalcncc of cattlc shcclcling Escherichia coli Ol57:H7 ancl characteristic of the cattlc or conditions of the fcccllot pen. .[oz~nzal ofFood Protection 64: 1899-1903.

Snlith UR, Gray JT, Moxlcy RA, Younts-Uahl SM, Ulackforcl MP, Hinklcy S, H~mgcrford LL, Milton CT and Klopfcnstcin TJ (2004a). A diagnostic strategy to clctcrnlinc the Shiga toxin-producing Escherichia coli 0157 status of pens of fcccllot cattlc. Epidemiology atzd Infection 132: 297-302.

Snlith UR, Peterson RE, Klopfcnstcin TJ, Erickson GE, Moxlcy RA ancl Hinklcy S (20041,). Feeding direct-fed micro1,ials to reduce the prcvalcncc of Escherichia coli Ol57:H7 in feedlot cattlc. In: Proceeditzgs of the 23th Atztzz~al World Bzliatrics fitzgress, Qzm5ec City, Qzlehec, .[z~ly 11-1 6, 2004.

Snlith W E , Kanc AV, Campbell ST, Achcson UWK, Cochran UH ancl Thorpc CM (2003). Shiga toxin 1 triggers a ribotoxic stress rcsponsc leading to p38 and JNK activation and induction of apoptosis in intcstinal cpithclial cclls. Infection atzd Imnzzltzity 71: 1497-1504.

Spcrandio V, Mcllics JL, Nguycn W, Shin S and Kapcr JU (1999). QLIO~LIIII sensing controls expression of the typc I11 secretion gene transcription and protein secretion in cntcrohcnlorrhagic ancl cntcropathogcnic Escherichia coli.

Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ of America 96: 15196-1 5201.

Spcrandio V, Mcllics JL, Uclahay RM, Frankcl G, Crawford JA, Nguycn W and Kapcr JU (2000). Activation of cntcropath- ogcnic Escherichia coli (EPEC) LEE2 and LEE3 opcrons by Lcr. :2(lolecz~lar:Wic1~ohiology 38: 781-793.

Spcrandio V, Torrcs AG, Gir6n and Kapcr JU (2001). Quorum sensing is a global regulatory nlcchanisnl in cntcrohcmor- rhagic Escherichia coli Ol57:H7. .[oz~nzal of Bacteriology 183: 5187-5197.

Spcrandio V, Li CC and Kapcr JU (2002a). Quorum-sensing Escherichia coli regulator A: a regulator of the LysR fanlily involved in the regulation of the locus of cntcrocytc cffaccnlcnt pathogenicity island in cntcrohcnlorrhagic E. coli. Infection atzd Imnzzltzity 70: 3085-3093.

Spcrandio V, Torrcs AG and Kapcr JU (20021,). Quorum scns- ing Escherichia coli regulators U ancl C (QscUC): a novcl two-component regulatory systcnl involved in the regula- tion of flagella ancl nlotility by quorum sensing in E. coli. :2clolecz~lar:Wic1~ohiology 43: 809-821.

Spcrandio V, Torrcs AG, Jarvis U , Nataro JP and Kapcr JU (2003). Uactcria-host communication: the language of hor- mones. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ ofAmerica 100: 8951-8956.

Stcvcns MP, Marches 0, Campbell J, Hutcr V, Frankcl G, Phillips AD, Oswalcl E ancl Wallis TS (2002a). Intimin, Tir ancl Shiga toxin 1 do not influence cntcropathogcnic rcsponsc to Shiga toxin-producing Escherichia coli in bovine ligated intcstinal loops. Infection atzd Inznzzltzity 70: 945-952.

Stcvcns MP, vanUicnlcn PM, Uziva F, Jones PW and Wallis TS (2002b). Options for the control of cntcrohacnlorrhagic Escherichia coli in ruminants. :Wicrohiology 148: 3767-3778.

Stcvcns MP, vanUicnlcn PM, Frankcl G, Phillips AU ancl Wallis TS (2002~). Efal influences colonization of the bovine intcs- tine by Shiga toxin-pmducing Escherichia coli scrotypcs 0 5 ancl 0111. Itzfectiotz atzd Inznzzltzity70: 5158-5166.

Surcttc MG, Miller MU and Uasslcr UL (1999). QLIO~LIIII sensing in Escherichia coli, Salnzotzella t@5inzzlrizl nz, ancl Vihrio han~eyz a new fanlily of genes rcsponsiblc for autoin- cluccr production. Proceeditzgs of the Natiotzal Acadenzy of Scietzce.~ of the 1Jtzited State.$ ofAmerica 96: 1639-1644.

Tarr PI, Uilgc SS, Vary Jr JC, Jclacic S, Ha1,ccb RL, Ward TR, Uaylor MR and Ucsscr TE (2000). Iha: a novcl Escherichia coli Ol57:H7 adherence-conferring molecule cncoclccl on a recently acquired chronloson~al island of conserved structure. Infection atzd Inznzzltzity68: 1400-1407.

Tats~mo I, Kimura H, Okutani A, Kanamaru K, Abc H, Nagai S, Makino K, Shinagawa H, Yoshicla M, Sato K, Nakanloto J, Tolx T ancl Sasakawa C (2000). Isolation ancl charactcriza- tion of mini-TnKm2 insertion mutants of cntcrohcnlorhagic Escherichia coli Ol57:H7 deficient in aclhcrcncc to Caco-2 cclls. Itzjiectiotz atzd Imnzzltzity 68: 5943-5952.

Tatsuno I, Horic M, Abc H, Miki T, Makino K, Shinagawa H, Taguchi H, Kanliya S, Hayashi T ancl Sasakawa C (2001). toxB gene on pol57 of cntcrohcnlorrhagic Escherichia coli Ol57:H7 is required for full cpithclial cell aclhcrcncc phenotype. Infection atzd Imnzzltzity 69: 6660-6669.

Tcsh VL, Uurris JA, Owens JW, Gordon VM, Waclolkowski EA, O'Uricn AU and Samuel JE (1993). Conlparison of the rcla- tivc toxicities of Shiga-like toxins typc I ancl typc 11 for mice. Infection atzd Inznzzltzity 61: 3392-3402.

Tkalcic S, Urown CA, Harnlon UG, Jain AV, M~lcllcr EP, Parks A, Jacobscn KL, Martin SA, Zhao T and Uoylc MP (2000). Effects of diet on rumcn proliferation ancl fecal shcclcling of Escherichia coli Ol57:H7 in calves. .[oz~nzal of Food Protection 63: 1630-1636.

E. coli 0157:H7: intestinal colonization and virulence

Thorpc CM, Hurlcy UP, Linciconlc LL, Jaccwicz MS, Kcusch GT ancl Achcson UWK (1999). Shiga toxins stimulate secretion of interleukin-8 from intcstinal cpithclial cclls. Itzfectiotz atzd Inznzzltzity 67: 5985-5993.

Thorpc CM, Snlith WE, Hurlcy UP ancl Achcson UWK (2001). Shiga toxins induce, supcrinclucc ancl sta1,ilizc a variety of C-X-C chcnlokinc nlRNAs in intcstinal cpithclial cclls, resulting in increased chcnlokinc expression. Itzfectiotz atzd Inznzzltzity 69: 6140-6147.

Torres AG and Kapcr JU (2003). Multiple clcnlcnts controlling adherence of cntcrohcnlorrhagic Escherichia coli Ol57:H7 to HcLa cclls. Itzfectiotz atzd Inznzzltzity 71: 4985-4995.

Tuttlc J, Gonlcz T, Uoylc MP, Wells JG, Zhao T, Tauxc RV ancl Griffin PM (1999). Lessons fronl a large outbreak of Escherichia coli Ol57:H7 infections: insights into the infectious close ancl nlcthocl of wiclcsprcacl contanlination of hamburger patties. EpidenzioloS): atzd Itzfectiotz 122: 185-192.

Tzipori S, Gunzcr F, Uonncn1,crg MS, dc Montigny L, Kapcr JU and Uonohuc-Rolfc A (1995). The role of the eaeA gene in diarrhea ancl neurological conlplications in a gnotobi- otic piglet nloclcl of cntcrohcnlorrhagic Escherichia coli infcction. Itzfectiotz atzd Inznzzltzity 63: 3621-3627.

Vallancc UA and Finlay UU (2000). Exploitation of host cclls by cntcropathogcnic Escherichia coli Proceeditzgs of the Natiotzal Academy of Scietzce.~ of the 1Jtzited State.$ of Anzel-lca 97: 8799-8806.

van dc Kar NC, Monncns LAH, Karn~ali MA and van Hinsbcrgh VWM (1992). Tumor necrosis factor ancl interleukin-1

induce expression of the vcrocytotoxin receptor globo- triaosylccranliclc on human cnclothclial cclls: inlplications for the pathogenesis of the hcnlolytic uremic syndrome. Blood 80: 2755-2764.

van Scttcn PA, van Hinsbcrgh VWM, van dcr Vclclcn TJ, van clc Kar NC, Vcrnlccr M, Mahan JU, Assn~ann KJ, van den Hcuvcl LPWJ and Monncns LAH (1997). Effects of TNF alpha on vcrocytotoxin cytotoxicity in purified human glomcrular microvascular cnclothclial cclls. Kidtzey Itztenzatiotzal51: 1245-1256.

Wcilcr LH, Viclcr E, Erpcnstcin C, Schlapp T, Stcinriick, Uaucrfcind R, Uyonli A ancl Ualjcr G (1996). Shiga toxin- producing Escherichia coli strains from bovines: association of adhesion with carriage of eae ancl othcr gcncs. .[oz~nzal of Qitzical :Wicrohiology 34: 2980-2984.

Woods JU, Schnlitt CK, Uarncll SC, Mcysick KC ancl O'Uricn AU (2002). Ferrets as a n~oclcl systcnl for renal disease secondary to intcstinal infcction with Escherichia coli Ol57:H7 ancl othcr Shiga toxin-producing E. coli .[oz~nzal of Itz fectiozls Di.$ease.$ 185: 550-554.

Wray C, McLarcn IM, Randall LP and Pcarson GR (2000). Natural ancl cxpcrinlcntal infcction of norn~al cattle with Escherichia coli 0157. Vetel-ltzaly Record 147: 65-68.

Zhang WL, Klihlcr, Oswalcl E, Ucutin L, Karch H, Morabito S, Caprioli A, Sucrl~aum S ancl Schnliclt H (2002). Genetic diversity of intinlin gcncs of attaching ancl effacing Escherichia coli strains. .[oz~nzal of Qitzical :Wicrohiolo~): 40: 4486-4492.