General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (Rev. 3/06) Page 1 of 47

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code OZO713

2. Project title

The role of the native host gut flora and innate immune status upon the colonisation of E. coli O157 in ruminants

3. Contractororganisation(s)

Veterinary Laboratories Agency,Woodham Lane,New Haw,Weybridge,Surrey,KT15 3NB.

54. Total Defra project costs £ 272,864(agreed fixed price)

5. Project: start date................ 01 October 2004

end date................. 30 September 2007

SID 5 (Rev. 3/06) Page 2 of 47

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Project OZ0713 has run over the past three years with VLA as the lead organisation of this highly collaborative set of studies that involved numerous partners, notably Bristol University, Imperial, Texas University, Institute of Animal Health, the Defra Fellowship at Edinburgh University and the Moredun Institute, the later being funded directly through this project. The overall goal was to understand the mechanisms of colonisation and persistence of Escherichia coli O157:H7 and Non-O157 attaching and effacing E. coli (AEEC) in sheep and to further elucidate how non-bacterial factors may effect colonisation and persistence of AEEC. The specific aims were:-

To determine by directed and signature tagged mutagenesis the contribution of factors other than LEE encoded of EHEC O157:H7 in colonisation and persistence in the sheep To define the role of parasites in enhancing EHEC O157:H7 colonisation and persistence in the sheep models.

To evaluate the host responses in the various models used above in collaboration with Prof. David Smith at the Moredun Research Institute.

Objective 1Much of the basic pathogenesis of E. coli O157:H7 has been elucidated recently, especially with the publication of several entire genome sequences that have given insights into new targets for study. Potential virulence determinants were studied in a previous Defra project OZO706 and their role elucidated. Briefly, intimin (eae) and the translocated intimin receptor (tir) were demonstrated to be key factors in the colonisation of sheep with O157, whereas the role of the O157 H7 flagellum (fliC) remained equivocal. To determine what other O157 factors contributed to colonisation in the sheep defined targeted mutants based on newly available sequence data were constructed by various collaborators and were tested in the sheep model at the VLA and signature tagged mutagenesis studies were conducted at IAH Compton. A signature tagged mutant library of O157 was made at IAH (Mark Stevens). The principle of this approach is that rather than testing one mutant at a time, a randomly constructed library of mutants is made and pools of mutants tested. In this study, sheep were dosed orally with pools of mutants and the tagged O157 mutants were recovered from faeces. The concept being that

SID 5 (Rev. 3/06) Page 3 of 47

‘fit’ organisms colonise and are shed normally whereas ‘unfit’ mutants would be lost. By searching the recovered by tag (presence absent) it should be possibly to identify ‘unfit’ mutants from the starting pool and then analyse the mutation specifically. In our studies the input dose yielded unexpected low faecal outputs of tagged O157 giving an unrealistically high ratio of presumed ‘unfit’ mutants. Being an expensive procedure, this approach was not taken further especially as sequence data had yielded new targets for analysis. Thus, to extend our understanding of those O157 factors that make it such a successful coloniser of ruminants we performed competitive index challenge studies by co-infection of lambs with wild-type and a specific mutant of O157. In this series of studies, O157 mutants defective for TccP, long polar fimbriae (lpf), non-LEE encoded effector genes (nleD) and ‘O’ island 50 (OI50) were used. Each was shown to contribute to colonisation with the most profound impact mediated by OI50. The role of this island is now under detailed investigation by David Gally (Edinburgh). The other genes could be considered to have a minor but positive contribution toward O157 colonisation of ruminants.

Objective 2The second objective was to study non-bacterial factors influencing E. coli O157 colonisation. This objective arose from the serendipitous observation that in an O157 infection lamb model, unexpectedly high shedding of O157 was noted with concurrent infection with Cryptosporidia. Thus, we proposed that prior infections may be a risk factor in colonisation and shedding of O157. A number of studies were performed that addressed this issue.Numerous experimental infections have been undertaken by many researchers but all have used healthy, well managed animals which have little or no evidence of stress of or other health status issues. Thus, a further aim was to investigate the role of parasites in the colonisation and persistence of O157 especially as the prevalence of parasitic infections of sheep are considered as high. In these studies, young lambs infected with Cryptosporidium parvum were subsequently infected with E. coli O157. It was observed that lambs pre-disposed by Cryptosporidium infection shed higher numbers of O157 in their faeces and at necropsy tissues were found to be colonised to a higher level than lambs solely inoculated with O157. Furthermore O157 AE lesions were observed in association with Cryptosporidium. We concluded that gut perturbation by coincidental infection may be another significant risk factor.Colostrum has been shown to be protective against many entero-pathogens and its benefits to gut maturation and overall health are well documented. Studies conducted in this project aimed to elucidate the role of colostrum in O157 infection in lambs. We showed that young lambs deprived of colostrum and then challenged with O157 at 6 weeks of age were highly susceptible to colonisation with O157 compared to conventionally reared lambs. Additionally, in this study AE lesions induced by O157 were considerably more readily detected than in conventionally reared lambs (ie/ given colostrums) in which AE lesions were small and very sparse. We conclude that deprivation of colostrums which often happens in petting farms is a serious risk factor that leads to ready colonisation of lambs by O157Continuing this theme of predisposing factors, we investigated the role of E. coli O26, a very commonly found attaching effacing E. coli (AEEC) strain in sheep, upon O157 colonisation. In these studies, we showed O26 reduced adherence of O157 in tissue culture studies. However, when co-infection studies were performed in lambs we observed that O157 suppressed colonisation and shedding of O26, a counter intuitive finding from tissue culture results. We concluded that O157 is particularly well adapted to the colonisation of sheep (ruminants) even to the exclusion of other prevalent serotypes. Whether there are other members of the naïve flora capable of excluding O157 needs to be assessed.

Objective 3The primary aim of this objective was to ascertain the immune responses to O157. Previous studies performed at Moredun showed that immune responses to O157 LPS, H7 Flagella and ST toxin were weak and variable. However, infections of 14 day old calves suggested that O157 preferentially colonised the terminal rectal, specifically the recto-anal junction that is a site with profuse lymphoid (immune) cells. The detailed work of Arvind Mahajan at Moredun suggested the H7 flagella of O157 specifically adhered the lymphoid tissue of calves. Thus, for this objective we studied the interaction of O157 with the distal gut of lambs to establish whether the findings for calves were also true for sheep.From all the oral ovine O157 challenge studies we have performed prior to and including early in

SID 5 (Rev. 3/06) Page 4 of 47

this project, it was clear that O157 has a tropism for the large intestine, but there was no concrete evidence from these studies in sheep to support the hypothesis that the terminal rectum and specifically recto-anal junction was the preferred site of colonisation, as suggested for bovines. However, rectal inoculation of young lambs undertaken in this project showed that the terminal rectum was readily colonised by O157 but, not unsurprisingly, unlike orally challenge animals O157 was not found distributed throughout the intestinal tract at necropsy. In the control oral inoculation experiments O157 was found throughout the gut. The question remains how valid is rectal inoculation as this is a very biased route of inoculation and delivers a high dose to a specific site in the gut? Whilst this data supports the concept of the terminal rectum can be colonised, we have prior evidence that O157 can associate with the small intestine too especially in those animals that show long term and persistent shedding. Indeed in such animals the colonisation at the terminal rectum is modest.Finally, it has been a long term goal to use surrogate in vivo models to comply with 3Rs HO requirements. We have developed successfully a mouse model in which O157 behaves similarly to the behaviour found in ruminants in terms of colonisation and shedding. Whilst we recognise the biological ‘distance’ between cattle and sheep on the one hand and the mouse on the other, we consider this a valuable tool for screening purposes that could be used by the wider scientific community. Immune responses in the mouse where not observed.Thus we conclude that O157 is well adapted to the gut environment and does not induce consistent high level immune responses. The specific tropism for lymphoid tissue in ruminants needs further analysis in terms of immune development and the differential behaviour on bovines and ovines.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

Scientific objectives

The scientific objectives as set out in the Project proposal were:-

Obj. No. Completed Description

SID 5 (Rev. 3/06) Page 5 of 47

by date

01 31:03:2007 To determine by directed and signature tagged mutagenesis the contribution of factors other than LEE encoded of EHEC O157:H7 in colonisation and persistence in the sheep models. [This work was undertaken in collaboration with Tim Wallis and Mark Stevens at the Institute for Animal Health, Compton; Prof. Gad Frankel at Imperial College London, Prof. David Smith at the Moredun Research Institute and Dr David Gally at Edinburgh University].

02 31:06:2006 To define the role of parasites in enhancing EHEC O157:H7 colonisation and persistence in the sheep models

03 31.03.07 To evaluate the host responses in the various models used above in collaboration with Prof. David Smith at the Moredun Research Institute

04 31:04:2007 To complete all reports and submit papers relating to these studies

OBJECTIVE 01

In this section of the report we describe a series of experiments in which mutants constructed in putative virulence determinants, Eae, Tir, TccP, NleC, NleD, NleH, Lpf and OI-50, were tested to assess their contribution toward the biology of E. coli O157:H7. Here we give reasons for the targeting of these putative determinants and then describe the behaviour of the mutant in the relevant animal model.

Background to the selection of target genes for studyAttaching and effacing strains of Escherichia coli (AEEC) comprise a group of gastrointestinal pathogens

of humans and animals that induce distinctive attaching and effacing (A/E) lesions within the host intestinal mucosa. A/E lesions are characterized by intimate attachment of the bacteria to the host cell surface, the localized effacement of intestinal microvilli and the rearrangement of host cytoskeletal proteins beneath the adherent bacteria. A/E lesion formation is mediated by the products of the locus for enterocyte effacement (LEE) pathogenicity island (PAI) and the LEE of enteropathogenic E. coli (EPEC) contains 41 open reading frames (ORFs) of which approximately half encode a type III secretion system (TTSS), which directs the secretion of several LEE-encoded translocator (EspA, EspB and EspD) and effector (EspG, EspH, Map, Tir, EspF) proteins. However only Tir plays a direct role in A/E lesion formation. The LEE PAI is well conserved among AEEC including the closely related human pathogen enterohemorrhagic E. coli (EHEC) and the animal pathogens, rabbit-specific enteropathogenic E. coli (REPEC) and Citrobacter rodentium. For all AEEC the primary effector molecule for intimate attachment is the eae gene with the translocated intimin receptor, tir. Whilst there roles are established individually, the role of both together has yet to be fully interrogated.

Recently, a number of novel effector proteins of AEEC were described that are not encoded within the LEE but are secreted and translocated into host cells by the LEE-encoded TTSS. The cycle inhibiting factor, Cif, was the first non-LEE encoded effector to be identified and is found in a subset of human and animal isolates of EPEC. Although it is not required for A/E lesion formation, Cif induces host cell cycle arrest and reorganization of the actin cytoskeleton. The second non-LEE encoded effector to be described was EspI/NleA (Z6024). EspI/NleA is also not required for A/E lesion formation but is essential for full virulence in the C. rodentium mouse model of infection. Screening of a large collection of Shiga toxin producing E. coli strains isolated from clinical cases and from controls revealed that nleA/espI is highly associated with the presence of the LEE (p < 0.0001). Moreover, nleA/espI was more frequently associated with strains isolated from symptomatic than from asymptomatic patients; both from those with severe disease (HUS) (p < 0.006) and from those with diarrhoea (p < 0.002). An effector protein carried on prophage CP933U termed EspJ (Z3071) has homology to the type III effector HopF from Pseudomonas syringae. In the murine model of infection, an espJ deletion mutant of C. rodentium showed altered colonization and clearance dynamics, although mice suffered similar disease symptoms to animals infected with wild type C. rodentium, suggesting that EspJ does not play a major role in AEEC pathogenesis. EspG2, which is encoded on the EspC PAI, shares similarity with EspG and was recently shown to interact with tubulin and trigger the dissociation of microtubules beneath adherent bacteria in a manner similar to EspG. TccP (Tir-cytoskeleton coupling protein)/EspFU described in EHEC O157:H7 is located downstream to EspJ on prophage CP933U and shows a low homology to EspF; TccP/ EspFU is essential for A/E lesion formation. In EHEC infected host cells, TccP/ EspFU functions as a linker between Tir and N-WASP and compensates functionally for the absence of the host adapter protein, Nck, which in EPEC infected cells is recruited to the pedestal upon tyrosine phosphorylation of Tir. Therefore unlike EPEC, strains of EHEC induce A/E lesion

SID 5 (Rev. 3/06) Page 6 of 47

formation and actin polymerization in an Nck-independent manner and without tyrosine phosphorylation of Tir using the additional prophage carried effector protein, TccP/ EspFU. Therefore, the role of TccP of EHEC in the ruminant model requires examination. Several other secreted proteins, dotted along the chromosome, have been identified through proteomic analysis of the secretome of C. rodentium and shown to be secreted by the LEE-encoded TTSS, although it is still unclear if these proteins constitute novel translocated effectors. These include NleB, -C, -D, -E, –F and -H). Recently, the gene encoding NleD was identified during a signature tagged mutagenesis screen as essential for full colonization of the bovine gut by EHEC O157:H7 strain EDL933 (see below). NleC/D/H were selected for study as their contribution is unknown and with respect to NleH, the genome sequences of EHEC O157:H7 strain EDL933 and Sakai, EPEC strain 2348/69 and C. rodentium (strain ICC168) revealed that EHEC and EPEC contain two nleH alleles, while C. rodentium harbors only one nleH gene. NleH shares approximately 49% sequence identity with the Shigella flexneri T3SS serine/threonine kinase effector protein OspG (outer Shigella protein G). OspG prevents ubiquitination and subsequent degradation of phospho-IκBα and downstream activation of the transcriptional factor NF-κB possibly via the phosphorylation of the E3 ubiquitin ligase SCFβ-TrCP (Skp-Cullin-F box protein) complex (14). Nf-κB proteins are transcription factors that when activated control the transcription of a large number of genes, many of which are involved in the immune (inflammatory) response. The similarities between NleH and OspG prompted us to investigate its role in colonization and activation of the Nf-κB pathway in vivo.

The first step of E. coli O157:H7 pathogenesis involves the initial adherence of the bacterium to the intestinal epithelium and the subsequent formation of intimin-mediated attaching and effacing (A/E) lesions on intestinal epithelial cells. As described above much of the regulatory mechanisms and the function of the E. coli O157:H7 virulence determinants associated with A/E lesion formation have been studied and documented, however, the identity and role of the determinants associated with initial bacterial interaction with enterocytes or persistence within the intestine remain poorly understood. The Long Polar Fimbriae (LPF) represents one adherence determinant in E. coli O157:H7 that was originally found important for the virulence and pathogenesis of Salmonella enterica serovar Typhimurium. E. coli O157:H7 contains two non-identical lpf loci homologous to lpf of serovar Typhimurium. While expression of the E. coli O157:H7 lpf operon 1 (lpf1) in E. coli K-12 increase adherence to tissue cultured cells (by c. 60%) and has been associated with the appearance of peritrichous long fimbriae, a clear role of the cloned lpf operon 2 (lpf2) in adherence has not been fully elucidated. However, it has been shown that E. coli O157:H7 mutant strains harbouring insertions in the lpfA1 or lpfA2 genes, predicted to encode the major fimbrial subunits encoded in loci 1 and 2, respectively, exhibit a reduction in adherence to culture cells. An area for investigation is the contribution of LPF to persistence in the intestine of infected animals and what the in vivo function of these fimbrial clusters are. Therefore, we generated a collection of lpf mutants and performed an exploratory study intended to provide the initial cues of how LPF is influencing persistence in the intestine during infection of a relevant 6-week old lamb model.

Another important finding from genome sequence comparisons between E. coli O157:H7 and laboratory strain E. coli K12 was the identification of regions of the O157:H7 genome that seemed to be inserted into what might be described as an E. coli backbone common to both K12 and O157. Given that K12 is a commensal and O157:H7 is a pathogen, it stands to reason that these insertions, or ‘O’ islands contribute to the biology of O157 and probably contain pathogenic determinants. Islands O-7, O-36 and O-50 contain regions of homology with other known pathogenic determinants. The aim of this study was to determine by rectal faecal sampling the diversity in persistence of a wild-type and isogenic ‘O’ island deletion mutant of an Stx-negative attaching and effacing E. coli O157:H7 isolate using the 6 week lamb model.

Approaches used for VLA components of these studies.

Bacterial Strains and growth conditions. All mutants were constructed by our collaborators using standard techniques (nleC and nleD, Frankel, Imperial College; lpf, Torres, Texas A&M; O-50 deletion, Gally, Edinburgh). Bacterial strains were grown in Luria-Bertani (LB) broth and incubated aerobically with shaking at 37ºC overnight or in Dulbecco’s modified Eagle medium (DMEM) buffered with 25 mM HEPES and supplemented with 10% fetal calf serum (FCS) and 2 mM glutamine at 37ºC 5% CO2

when cultured for IVOC and tissue culture studies. HeLa and HEp-2 cells were cultivated in DMEM supplemented with 10% fetal calf serum (FCS), 2 mM glutamine at 37ºC 5% CO2.

Fluorescent actin staining (FAS) test. FAS test was performed on infected HEp-2 cell as described by Knutton et al. (1989) which is an established technique and for brevity not repeated here.

Lamb in vitro organ cultures. For IVOC, proximal small Light microscopy subsequently showed no histological abnormality. The ability of the E. coli O157:H7 wild-type and mutant strains to colonize was assessed by IVOC. Intestinal mucosal biopsies from selected parts of the gastro-intestinal tract (duodenum, jejunum, ileum, caecum, ascending colon, spiral colon, rectum and rectal anal junction (RAJ)) of six week old conventional lambs which appeared macroscopically normal were taken for organ culture experiments. Briefly, approx. 2cm2 tissue samples were obtained and placed into Duran bottles containing tissue culture medium (see above). The tissues were infected with bacteria to a final concentration of approximately 1 x 108 CFU/ml. After 3 h of incubation at 37ºC (5% CO2), the tissue samples were disrupted and serial dilutions were plated to obtain CFU/mL. In each experiment an un-inoculated sample (to exclude endogenous bacterial adhesion) and a positive control (IVOC with parent strain to exclude host factors) were included. Samples were fixed with 2.5% glutaraldehyde.

SID 5 (Rev. 3/06) Page 7 of 47

Oral inoculation of lambs. Standard methods have been developed for the use of the six week old lamb model in previous studies (e.g. OZ0703 and OZ0706). An appropriate number of 6-week-old cross bred lambs were randomly divided into 3 equal groups, supplied with food and water ad libitum and confirmed to be free of EHEC O157 by enrichment and O157 IMS (immuno magnetic separation). All lambs were housed in biosecure CL2 accommodation. Each group was housed in a separate room with its own air handling. Animals were visited only by experienced staff who changed clothing between each group. Lambs were each dosed orally with either 1 x 109 CFU of the test strain resuspended in 10 ml of PBS (pH 7.4). Approximately 24 hours after dosing, and as required thereafter for up to 27 days, rectal faecal samples from each lamb were collected for direct plating onto Sorbitol MacConkey (SMAC, Oxoid, Basingstoke, UK) plates supplemented with either 15 g/ml nalidixic acid or 25 g/ml kanamycin (Sigma). Samples which were negative on direct plating were enriched in BPW for 6 hrs at 37C and then plated onto SMAC supplemented with the appropriate antibiotic. Representative colonies were confirmed to be E. coli O157 by latex agglutination (Oxoid, Basingstoke, UK). For co-infection studies, the competitive index (CI) was calculated and is defined as the ratio between the mutant and wild-type strains within the output (bacteria recovered from the host after infection) divided by their ratio within the input (initial inoculum). For this experiment the input ratio was 1:1. The null hypothesis that CI = 1 was tested by a 2-sided t-test.Histopathology. Segments of the gut tissues covering all parts of the gut were collected at post mortem at times during studies as indicated for each study. Each tissue was rinsed of their content and fixed in 10% buffered formalin for microscopic examination. Formalin-fixed tissues were then processed, paraffin-embedded, sectioned at 5 m, and stained with hematoxylin and eosin (H&E) according to standard techniques. Sections were examined by light microscopy for the presence of intimately adhering bacteria on intestinal cells, as previously described (10). Crypts length was also evaluated and the length of at least 4 well-oriented crypts has been measured on each section. A non-parametric ANOVA with a posteriori comparisons was performed using commercially available GraphPad InStat v3.06 software (GraphPad Software Inc., San Diego California USA). P values ≤0.05 were considered significant.

Immunohistochemistry. Snap-frozen colonic tissues, embedded in OCT mounting medium (VWR BDH, Lutterworth, UK) were sectioned using a cryostat to a thickness of 5µm. Sections were mounted on polysine slides (VWR BDH) and air dried overnight before fixing in acetone at room temperature for 20 min. After air drying for 1 h, sections were rehydrated in Tris-buffered saline (TBS) for 5 min then incubated with antibodies against CD3, CD4 and CD8 (Serotec, Oxford, UK) at a 1:50-1:100 dilution for 1 h. Sections were gently washed with TBS 3 times before addition of biotinylated anti-rat IgG (Serotec) at a dilution of 1:200 with 4% (v/v) normal murine serum for blocking (Sera Laboratories International, Horsted Keynes, UK) for 30 min. After washing, a 1:200 dilution of 0.1% avidin-peroxidase (Sigma-Aldrich, Dorset, UK) was added for 30 min before further washing and the addition of diaminobenzadine substrate (Sigma-Aldrich) for 5 min. The reaction was stopped with excess TBS and sections were counterstained with Mayers hematoxylin (Sigma-Aldrich) for 30 seconds, dipped in acid alcohol and washed in tap water for 5 min. Sections were dehydrated through an ethanol gradient of 70%, 90% and 100% solutions (2 minutes each) followed by clearing in histoclear (VWR BDH) and mounting in DPX (VWR BDH). A control slide using no primary antibody was also performed to show endogenous peroxidase-containing cells. Stained cell populations were counted in 5 randomly selected fields per section and data was expressed as the number of T cells per 250μm2 of lamina propria.

Scanning electron microscopy. Intestinal segments were fixed in 2.5% glutaraldehyde and processed for scanning electron microscopy (SEM) as previously described (9). SEM samples were examined blindly, at 25 kV using a JEOL JSM-5300 scanning electron microscope (JEOL (UK) Ltd., Herts, United Kingdom).

Indirect immunofluorescence. An indirect immunofluorescence assay (IFA) was used for the detection of E. coli O157 in formalin-fixed, paraffin-embedded (FFPE) sections as previously described. Tetramethyl rhodamine iso-thiocyanate (TRITC)-conjugated donkey anti-rabbit (Jackson ImmunoResearch Europe Ltd., Soham, Cambridgeshire, UK) secondary antibody was used to visualise O152-positive bacteria, while DNA of both bacteria and epithelial cells was counterstained with Hoechst 33342 (Sigma-Aldrich Co, United Kingdom). Sections were examined with an Axio Imager M1 microscope (Carl Zeiss MicroImaging GmbH, Germany). Images were acquired using an AxioCam MRm monochrome camera, and computer-processed using AxioVision (Carl Zeiss MicroImaging GmbH, Germany), Adobe Photoshop 5.0 and Adobe Illustrator 8.0 software (Adobe Systems Incorporated, California, USA).

Statistical analysis. The non-parametric Kruskall-Wallis test was used to compare the mean counts at each time-point in the animal studies and the comparisons with the parent strain NCTC12900 were done by Dunn’s test.

RESULTS OBJECTIVE 01

Tir Data

SID 5 (Rev. 3/06) Page 8 of 47

E. coli O157:H7 strain NCTC 12900 nalR eae::CamR and tir::StrR mutants exhibited marked defects in their ability to colonise the ovine gastrointestinal tract. Significant reductions in shedding of both mutants relative to the wild-type were detected from day 2 post-inoculation {p values 0.0006 and < 0.0001 respectively}. Consistent with the results obtained using the calf model (performed simultaneously at IAH but reported here), the NCTC 12900 nalR tir::StrR mutant was shed in lower numbers and for a shorter duration than the isogenic intimin mutant. Indeed, the NCTC 12900 nalR tir::StrR could not be detected even after enrichment from day 3 post-inoculation, whereas the NCTC 12900 nalR eae::CamR mutant was eliminated on day 16 post-inoculation (Fig. 1a and 1b).

Fig 1a and 1b - Course of fecal excretion of E. coli O157:H7 eae and tir mutants in 14-day-old calves (A) and six-week-old lambs (B). The level of colonization is indicated by the mean log-transformed viable count (CFU/g) of E. coli O157:H7 in fecal samples taken twice daily for 12 days (A) or at intervals for 27 days (B).

TccP data

As described in the introduction to this section, mutation of tccP uncoupled actin assembly from the intimin-Tir-mediated adherence of EPEC and EHEC in vitro but did not impair intestinal colonization in calves of the EHEC strain, implying that pedestal formation may not be necessary for persistence at least in that system. In work carried out at the VLA, the tccP mutant and its parent E. coli O157:H7 strain were assessed in tissue culture, ligated gut loop and the in vivo lamb model. We demonstrated that an E. coli O157:H7 TccP mutant induced typical attaching and effacing lesions in tissue culture and an ovine ligated ileal loop model of infection, suggesting that TccP-independent mechanisms of actin assembly may operate in vivo. To test this hypothesis a competitive index experiment in the six week old model was performed and the results are shown below (Fig 2a and 2 b). There were no statistically significant differences between the persistence and shedding pattern of either the parent or mutant tccP strain. These data suggest the role for TccP in EPEC is different from that in EHEC and that in EHEC TccP plays at most a minor role in gastro0intestinal colonization of ruminants.

SID 5 (Rev. 3/06) Page 9 of 47

Fig 2a Faecal shedding of wild-type E. coli O157 and Tccp KO mutant from orally infected six week old conventionally reared lambs. Solid symbols represent the wild-type (85-170) and the open symbols represent the TCCP mutant.

Fig 2b - Faecal shedding of wild-type E. coli O157 and Tccp KO mutant in a oral competitive index infection study undertaken in six week old conventionally reared lambs. Solid symbols represent the wild-type (85-170) and the open symbols represent the TCCP mutant.

NleC and NleD dataIn vitro analysis of nleC and nleD mutants of EPEC and EHEC. Although the LEE of EPEC encodes all the proteins required for A/E lesion formation, the prophage-carried effector TccP is additionally required for A/E lesion formation in EHEC. Therefore to examine the contribution of NleC and NleD to A/E lesion formation and adherence in vitro, nleC and nleD deletion mutants of EPEC E2348/69 (ICC193 and ICC194 respectively) and EHEC 85-170 (ICC195 and ICC196 respectively) were used to infected HEp-2 cell monolayers with wild type and mutant strains. In a fluorescent actin staining (FAS) test, there was no difference in the ability of wild type and nleC and nleD mutant strains of either EPEC or EHEC to induce actin polymerization (data not shown). In addition, there was no difference in the pattern of adherence of wild type and mutant EPEC strains to HEp-2 cells which all exhibited localized adherence, or in the pattern of adherence of wild type EHEC and mutant strains which all exhibited diffuse adherence. NleC and NleD are therefore not required for adherence to tissue culture cells or actin accumulation in EPEC and EHEC.

Colonization of lambs by nleC and nleD mutants of EHEC O157:H7. Although not required for colonization of mice by C. rodentium, we determined the contribution of NleC and NleD to colonization in other models of AEEC infection. Here we used wild-type and nleC and nleD mutant strains of EHEC 85-170 (wild-type), ICC195 (nleC,

SID 5 (Rev. 3/06) Page 10 of 47

KO) and ICC196 (nleD KO), to infect conventional 6 week old lambs. The ability of the test strains to establish and persist in lambs was investigated by monitoring viable bacteria recovered in stools collected per rectum. In this model wild-type EHEC 85-170 produced a typical shedding pattern, persisting in high numbers in the early stages of infection, then declining until undetectable by day 10 post-infection. Wild-type EHEC 85-170 then remained undetectable by direct plating or with enrichment for the remainder of the study (27 days in total, data not shown). In contrast, while ICC195 and ICC196 showed similar levels of colonization to wild-type EHEC 85-170 in the early stages of infection, one animal continued to shed ICC196 (as detected by enrichment) until 14 days post-infection. These data indicate that nleC and nleD do not play a role in colonization.

NleH data Contribution of NleH1 and NleH2 to colonization of conventional 6-week-old-lambs. A 6-week-old lamb model was next used to compare the persistence of an E. coli O157:H7 nleH1/nleH2 double mutant and the isogenic wild-type strain. The ability of the mutant to establish itself and persist in lambs was investigated by monitoring the viable counts recovered in faecal pellets collected per animal. When given as a single inoculum, the wild-type E. coli O157:H7 isolate produced the classical shedding pattern in lambs as described previously (Woodward et al., 2003: La Ragione et al., 2005), persisting in relatively high numbers during the early stages of infection and then declining by day 11 post-inoculation (Fig. 3a). The ΔnleH1/nleH2 mutant demonstrated a similar shedding pattern, except that positive faecal samples were only noted until day 12 post-inoculation (Fig. 3a). When both isolates were administered in the same inoculum, the wild-type persisted for 4 days longer than the ΔnleH1/nleH2 mutant (data not shown). The mean competitive index was significantly less than 1 for all time points where it could be calculated except day 1 (Fig. 3a) demonstrating that the wild-type strain outcompeted the ΔnleH1nleH2 mutant in the ovine model, indicating that both NleH1 and NleH2 play a role in colonization.

Fig 3a. Faecal shedding of wild-type E. coli O157 and NleH1/NleH2 mutant in a competitive index oral inoculation study in conventionally reared six week old lambs. Solid symbols represent the wild-type (85-170) and the open symbols represent the NKeH1/2 mutant.

Long Polar Fimbriae data

Environmental regulation of lpf1 is affected by the osmolarity and pH of the growth medium. Regulation of fimbriae gene expression by environmental signals in enteric pathogens is well documented. For example, expression of the bundle-forming pili genes in enteropathogenic E. coli, another A/E lesion-forming pathogen, is modulated by the ammonium concentration and by temperature. In the case of Salmonella lpf genes, the expression is induced when bacteria are grown in either static LB (pH 5.1) broth or static CFA broth. Because temperature was the only environmental condition that we previously established to stimulate E. coli O157:H7 lpf1 gene expression, the role of multiple environmental factors as stimulus of transcription for lpf genes was examined using the -galactosidase assay. At the conditions tested (37°C and late logarithmic phase of growth), induction of the expression in D-MEM medium was not affected by ammonium sulfate or ferric chloride concentration or by iron limitation (data not shown). In contrast, the pH of the medium did influence the expression of the lpfA1p::lacZ fusion, with the greatest induction levels observed at pH 6.5 (P < 0.05) and low levels occurring at pH 5.5. Interestingly, induction of the lpf1 loci was modulated when the media was supplemented with NaCl. Our results showed that lpf1 gene expression was reduced to almost half as compared with the Beta-galactosidase activity obtained when the bacteria was grown in unadjusted D-MEM medium. Increasing concentrations of NaCl did not further increase or decrease the lpf1 expression but did negatively affect the growth rate of the culture. When NaCl was replaced with KCl, similar expression patterns were observed,

SID 5 (Rev. 3/06) Page 11 of 47

indicating that the regulatory effect is not specific to NaCl (data not shown). Although our experiments were performed using a plasmid-based reporter system which produce higher Beta-galactosidase enzymatic activity than single copy fusions, our differences displayed similar patterns and were statistically significant. Therefore, we considered that our results contribute to reveal important aspects of the environmental regulation, in particular the observation suggesting that expression of the lpf1 loci responds to temperature, NaCl, and pH. Due to the difficulty of obtaining sufficient purified LPF from the wild type strain to perform protein experiments, we were unable to confirm our transcriptional data by Western blot analysis. However, we were able to establish additional conditions modulating the expression of the E. coli O157:H7 lpf1 loci and showed that these environmental conditions are quite different to those modulating the expression of the Salmonella lpf operon. These findings argue in favor of the presence of different environmental signals controlling LPF expression during colonization by these two enteric pathogens.

Persistence of E. coli O157:H7 and lpf mutants in six week old cross bred lambs. An interesting feature of the two LP fimbrial clusters of E. coli O157:H7 is that they contained genes with premature stop codons. These will imply that the two LP fimbrial clusters are unable to produce fully functional fimbriae although expression may be possible by using complementary fimbrial proteins from other clusters. This could be a likely possibility because our previous studies demonstrated that these two fimbrial clusters have functional roles in vivo using animal models of infection, or ex vivo utilizing the IVOC (in vitro organ culture) system. However, intact LP fimbrial operons do not guarantee functional fimbriae expression, as we demonstrated for enteropathogenic E. coli (EPEC) and Citrobacter rodentium, where the presence of intact LPF clusters neither play an apparent role in EPEC pathogenesis nor they are required for C. rodentium virulence in either the C3H/HeJ or C57BL/6 mouse models (Tatsuno et al. 2006). Despite the problems cited above and the fact that nobody has been able to consistently visualize and isolate fimbrial structures from the surface of E. coli O157:H7 strains grown in vitro, it does not mean that LP fimbriae do not play an important role in vivo; therefore, we investigated whether LPF contribute to the colonization of experimentally inoculated lambs by E. coli O157:H7. Our approach compared the ability to colonize of the lpf mutant strains JGG01, JGG03 and JGG05 to that of the parent strain (Fig 4), E. coli O157:H7 NCTC12900 (Torres et al ., 2007). To determine that the genetic manipulation did not have an effect in the growth of the isogenic mutants as compared with the wild-type strain, the bacteria were diluted in either LB or D-MEM and their growth rate monitored by OD600 measurements taken every hour for a period of 8 h. Furthermore, aliquots of 100 µl of the bacterial cultures were taken, diluted and plated on LB agar plates to obtain colony forming units. Overall, the wild type and lpf mutant strains bacterial strains grew at the same rate regardless of the media tested.

Fig 4. Faecal shedding of Wild-type E. coli O157 NCTC12900 (black) and LPF KO mutants (red, yellow and green) in orally inoculated six week old conventionally reared lambs.

Eighteen 6-week old conventionally reared cross breed lambs were dosed orally with 109 CFU of either the parent strain (5 animals), JGG01 (4 animals), JGG03 (4 animals), or JGG05 (5 animals). Rectal fecal samples from each group were taken for direct plating onto selective media plates. All inoculated strains were recovered from the feces of all 18 lambs during the initial period of the study. At days 2 and 3, the parent, the JGG03 ( lpfA2-) and the JGG05 (lpfA1- lpfA2-) strains were recovered from the faeces at significantly higher levels than JGG01 ( lpfA1-) (P < 0.05 when comparing the wild-type to the lpfA1 mutant). However, at day 6 and then day 8, the mutant JGG01 strain was recovered at comparable levels as the parent or the other two mutant strains. The proportion of lambs shedding each strain declined starting at day 11; however, at least three of the strains (parent, JGG01 and

SID 5 (Rev. 3/06) Page 12 of 47

JGG03) were recovered from the lambs throughout the study. Nevertheless and starting at day 11, the JGG03 was consistently recovered at lower levels than the parent or the JGG01 strain (however, no statistical significance was observed). Interestingly, the double lpf mutant strain JGG05 was recovered at lower levels than the other three strains and at days 15 and 21, this mutant strain was not recovered from the feces of four of the five infected animals.

This study confirmed the role in vivo of LP fimbriae during colonization and persistence experimentally infected lambs by E. coli O157:H7 strain NCTC12900 (Woodward et al. 2003) and suggest that long term colonization might be facilitated by LPF. Although the reduction in the bacteria recovered from the feces of the single lpf mutants was significant, particularly lpfA1 in the first days post infection, the effect of the double mutation of the LPF genes was more pronounced, e.g., JGG05 was recovered at lower levels than the parent strain; particularly at the late time points. Previously, it has been demonstrated in rabbits, sheep and pigs that AEEC organisms carrying a mutation in the lpf genes are recovered at lower levels than their parent strains and this was demonstrated best at early time points during colonization (Jordan et al. 2004; Newton et al. 2004). In our current study, we observed that recovery of JGG05 from the lambs at later time points was intermittent and reduced compared with the recovery of the parent strain.

Association of E. coli O157:H7 strains NCTC12900 and JGG05 to lamb in vitro organ cultures. Because the strain JGG05 was cleared more rapidly from the intestinal lumen than the parent or either lpf single mutant, we hypothesize that the lpf double mutant could display a reduced adherence to most sites in the lamb gastrointestinal tract. Therefore and to determine whether the reduction in faecal shedding of the JGG05 double mutant strain was due to a defect in their ability to colonize different tissues, we performed a preliminary comparative analysis to the parent strain for their ability to colonize the lamb duodenum, jejunum, ileum, cecum, ascending colon, spiral colon, rectum and rectal anal junction (RAJ) using the in vitro organ culture (IVOC) model (Torres et al., 2007). The results with the parent strain confirmed our previous published observations (Woodward et al. 2003), indicating that E. coli O157:H7 strain NCTC12900 does not display tissue tropism for the cecum or very terminal rectum and it appears that the mid-rectum is the place of preferred colonization. However, the parent strain was also recovered from duodenum, jejunum, ileum and colon. The double mutant strain JGG05 adheres as well as the wild type and in some cases appears to adhere more effectively to the mid rectum and RAJ (association of JGG05 to these tissues is 0.5-1.0 log higher than parent strain) and display similar levels of recovery than parent strain to the small intestine and colon. Even though the IVOC experiment was only performed once, the trends observed suggested that the reduction in the persistence observed with the lpf double mutant was not due to its inability to colonize the lamb gastrointestinal tract during early stages of infection if we assume the IVOC system mimics actual colonization in vivo. Previous studies have demonstrated that strain NCTC12900 causes A/E lesions and intimin-dependent persistence in the 6-week old sheep model (Woodward et al. 2003) and this is mainly due to the binding of intimin to its bacterial receptor Tir (Vlisidou et al. 2006). Based on our preliminary results, we can speculate that the presence of LPF may hinder interactions with the mucosa as the primary role is probably not adherence to the lamb intestinal tract, but instead stabilizing a formed/forming A/E lesion. Thus, inactivation of LP fimbriae may remove a potentially interfering factor for host pathogen interactions that cause the slight increase in adherence to the preferred colonization sites. Upon time, the presence of LPF becomes evident and the bacteria that lack these fimbriae may not persist in the intestine. However, additional experiments are needed to clearly define the role of LPF in persistence in the intestine and this is one of our goals for future investigation.

OI-50 data

The aim of this study was to determine by rectal faecal sampling the diversity in persistence of a wild-type and isogenic ‘O’ island deletion mutants (O-7, O-36 and O-50) of an Stx-negative attaching and effacing E. coli O157:H7 isolate using the 6 week lamb model. Groups of 6-week-old lambs were inoculated with a 1 x 10 10

CFU/ml dual inoculum (Competitive Index study) consisting of 5 x 109 CFU/ml E. coli O157:H7 TUV (wild-type) and either 5 x 109 CFU/ml O-7, O-36, or O-50 mutants. The results demonstrated that all isolates persisted in lambs for 28 days post inoculation, with the exception of the O-50 mutant, which only persisted for 9 days post inoculation. Furthermore, statistical analysis demonstrated that significantly less O50 (p<0.0001) and O36 (p=0.0026), but not O7 (p=0.0944) were recovered from faecal samples as compared to the TUV wild-type. Further analysis of the genetic make up of O50 O-island is in hand at Edinburgh University. An OI-50 knockout mutant did not persist for as long or at as high numbers as the wild-type progenitor strain (Fig. 5).

SID 5 (Rev. 3/06) Page 13 of 47

Fig 5. Faecal shedding of Wild-type O157 (black) and OI-50 mutant (red) in orally inoculated six week old conventionally reared lambs. Black symbols represent the wild-type (TUV) and the red symbols represent the OI-50 mutant.

Signature Tagged Mutagenesis approach

A signature tagged mutant library of O157 was made at IAH by Mark Stevens and his team as described in Dziva, F., P. M. van Diemen, M. P. Stevens, A. J. Smith, and Wallis. TS. 2004. Identification of Escherichia coli O157 : H7 genes influencing colonization of the bovine gastrointestinal tract using signature-tagged mutagenesis. Microbiology 150:3631-3645. For brevity this report will not report on the work at IAH but introduce the concepts of the study. The principle of this approach is that rather than testing one mutant at a time, a randomly constructed library of mutants is made and pools of mutants tested. The unique feature of the mutation system is that each mutant is individually tagged so it can be readily and uniquely identified. Thus, a pool of mutants, say 100 at a time, can be put through a test, such as passage through an animal, and the surviving mutants recovered. Because of the unique tagging in each mutant, it is possible probe recovered mutants against the original mutant library to see which have not been recovered. Those not recovered were clearly unfit to survive the test and, therefore, the mutation in that mutant may contribute to characteristics of survival in the test system. This approach was used in cattle and identified seventy-nine E. coli O157:H7 mutants impaired in their ability to colonize calves and 59 different genes required for intestinal colonization were identified by cloning and sequencing of the transposon insertion sites. Thirteen transposon insertions were clustered in the locus of enterocyte effacement (LEE), which encodes a type III protein secretion system required for the formation of attaching and effacing lesions on intestinal epithelia. A putative structural component of the apparatus (EscN) is essential for intestinal colonization; however, the type III secreted effector protein Map plays only a minor role. Other Type III secretion-associated genes were implicated in colonization of calves by E. coli O157 : H7, including z0990 (ecs0850), which encodes the non-LEE-encoded type III secreted effector NleD and the closely related z3023 (ecs2672) and z3026 (ecs2674) genes which encode homologues of Shigella IpaH proteins. We also identified a novel fimbrial locus required for intestinal colonization in calves by E. coli O157 : H7 (z2199-z2206; ecs2114-ecs2107/locus 8) and demonstrated that a mutant harbouring a deletion of the putative major fimbrial subunit gene is rapidly out-competed by the parent strain in co-infection studies.

Thus, the objective in this study was to repeat the in vivo phase of the study but in sheep to examine whether new genes may be identified that contribute to colonisation in sheep as opposed to calves. The six week old lamb model which is now well established was used. Individual lambs were dosed with different pools of mutants, each pool containing 100 tagged mutants. The total oral dose per animal was 109 cfu comprising 107 cfu of each mutant. Recovery of O157:H7 organisms was made on day 7 after oral dosing because this was considered the minimum time for the effect of mutations impairing gut colonisation in ‘unfit’ mutants being lost. This time point was based on prior studies by this author (see Woodward et al 2003). The recovery of O157:H7 organisms from all animals was sub-optimal for reasons that we could not account for. Usually by this time wild type parental strains are recovered at between 104 to 107 cfu per gram of faeces. In these experiments the highest yield was 5x105 cfu. However, whole cell lysates from the selection plates were made and the DNA used to probe the original libraries. From each experiment between 48 and 26 mutants per 100 in each pool did not hybridise with the recovered bacteria. This was an unexpectedly high rate of loss and not comparable with previous studies performed by IAH colleagues in cattle. Upon interrogation, even mutants with mutations in house

SID 5 (Rev. 3/06) Page 14 of 47

TUV versus O-50

1.00E+00

1.00E+01

1.00E+02

1.00E+03

1.00E+04

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1 2 3 4 5 6 7 9 14 16 21 23 28

Day Post Inoculation

CFU

/g/F

aece

s

TUV-B39

TUV-B41

TUV-B44

TUV-B48

TUV-B50

TUV-B61

O-50-B39

O-50-B41

O-50-B44

O-50-B48

O-50-B50

O-50-B61

hold genes known to have no effect on colonisation were lost. This finding implied that the recovery was sub-optimal as both fit and unfit mutants were lost. This approach was not taken further especially as sequence data had yielded new targets for analysis.

General discussion for objective 01

The aim of this objective was to determine those bacterial factors that contributed to colonisation and persistence in the host. The VLA has built expertise with sheep as a model system whilst IAH and SAC/Moredun have expertise in cattle. Thus, one of our aims has been to work closely with our colleagues at these other institutes in order to maximise the efforts and to ensure the best possible comparable analysis between these two ruminant species. There are differences in the behaviour of O157 in cattle and sheep and these have been discussed more fully in a review article that we have just had accepted for publication (La Ragione, R. M. Best, A., Woodward, M. J. and Wales, A. 2009. E. coli O157 in small ruminants. Here we focus on the major findings arising from the work performed at the VLA within his project.

It is well established that intimin (Eae) is key to AE lesion formation and that the translocated intimin receptor (Tir) is known to act by binding bacterial surface exposed intimin. What has not been interrogated is the significance of Tir and its regulator in vivo. Intimin facilitates intestinal colonization by enterohemorrhagic Escherichia coli O157:H7, however the importance of intimin binding to its translocated receptor (Tir) as opposed to cellular co-receptors is unknown. Intimin-Tir interaction is needed for optimal actin assembly under adherent bacteria in vitro, a process which requires the Tir cytoskeleton coupling protein (TccP/EspFU) in E. coli O157:H7. Here we report that E. coli O157:H7 tir mutants are at least as attenuated as isogenic eae mutants in calves (IAH work) and lambs (VLA work), implying that the role of intimin in colonization of reservoir hosts can largely be explained by it binding to Tir.

Mutation of tccP uncoupled actin assembly from intimin-Tir-mediated adherence of E. coli O157:H7 in vitro but did not impair intestinal colonization in calves and lambs, implying that pedestal formation may not be necessary for persistence. This observation may be in accordance with the extreme difficulties that we and many other workers have had in detecting ‘classic’ AE lesions in ruminants colonised by E. coli O157. Certainly, we have suggested previously that AE lesion formation is easier to detect in neonates and six week old lambs as opposed to three month old or older sheep. It is possible that AE lesion formation is an age related phenomenon in ruminants. However, an E. coli O157:H7 tccP mutant induced typical attaching and effacing lesions in a bovine ligated ileal loop model of infection, implying that TccP-independent mechanisms of actin assembly may operate in vivo also.

The LEE-encoded TTSS of AEEC is essential for virulence and translocates at least 10 proteins from the bacterial cytoplasm into the host cell cytosol. Some of the translocated effectors have been shown to be essential for colonization and disease in various animal models of infection, including EspB and EspD, which together with EspA form the EPEC/EHEC translocon required for the delivery of other effector proteins into host cells. Tir, which is essential for intimin binding and intimate attachment of the bacteria to the host cell surface, is also essential for virulence in several models of infection.

Until recently, the only effector proteins known to be translocated by the LEE-encoded TTSS were also encoded within the LEE. However, now several novel effector proteins have been identified that are not encoded by LEE yet they are secreted and translocated into cells by the LEE-encoded TTSS. Two of these putative non-LEE encoded (Nle) effectors, NleC and NleD, were identified in C. rodentium by proteomic analysis of secreted proteins by Gadi Frankel and co-workers. The genes encoding NleC and NleD are present in other AEEC, including EPEC and EHEC, and are located in O-island 36 of the EHEC O157:H7 EDL933 genome. While NleC has no similarity with proteins from non-AEEC, NleD shares some similarity with the TTSS effector, HopPtoH from the plant pathogen, P. syringae. Although the contribution of HopPtoH to the virulence of P. syringae is unknown, both proteins share a common zinc binding motif found in the neurotoxin of C. botulinum which may be important for their function. Frankel demonstrated that NleC and NleD are true translocated effectors of the LEE encoded TTSS. An intact LEE was required for the secretion of NleC-TEM-1 and NleD-TEM-1 fusion proteins and for the translocation of the fusion proteins into host cells. The observations that NleC and NleD could enter cells suggested that these proteins may play a direct role in host-pathogen interactions and to determine the contribution of the proteins to virulence, nleC and nleD deletion mutants were tested in several in vitro and animal models of infection. The results showed that neither NleC nor NleD played a role in AE lesion formation, adherence to human intestinal tissue or adherence to tissue culture cells, or carriage and virulence in lambs (VLA work), mice (IC work) or calves (IAH work). At this stage the precise role of NleC and NleD in the pathogenesis of infections with AEEC is unclear. Although in this study we were unable to find a phenotype for NleC or NleD, the fact that they are translocated into host cells suggests the proteins have the potential to influence host-pathogen interactions.

In this study we investigated the role of NleH, another translocator effector molecule that is non-LEE encoded, in colonization, competitiveness and activation of the NF-κB pathway in vivo. Genomic analysis of EHEC O157:H7 EDL933 shows that it both contains two, almost identical, copies of nleH. The functional consequences of this gene duplication are not known. Interestingly, a recent shotgun sequencing of a number of EHEC O157:H7 isolates has shown that they contain, in addition to nleH, a gene that shares 90.8% sequence identity with OspG (NCBI BLASTp) which is yet to be investigated fully. Investigating the contribution of NleH towards colonization and competitiveness of EHEC O157:H7 in the ovine (VLA) and bovine (IAH) hosts has

SID 5 (Rev. 3/06) Page 15 of 47

shown that the EHEC O157 ΔnleH/1nleH2 double mutant was shed in greater numbers than the parental strain from orally challenged calves, significantly from day 10 post-inoculation. The precise effect of deleting nleH on factors or regulatory mechanisms involved in E. coli O157:H7 colonization in bovine intestines remains to be investigated. In single infection studies in lambs there was no statistical difference in shedding after oral inoculation with the same strains. However, competitive indices measured following oral inoculation of lambs with a mixture of EHEC O157:H7 wild-type and ΔnleH1/nleH2 double mutant revealed that the mutant was significantly out-competed. The reasons for the different phenotypes observed in the bovine and ovine models are currently not known, although animal age might be an important factor. Here is evidence that the ovine and ovine host are significantly different with respect to colonisation by E. coli O157:H7

Fimbriae play an important role in colonisation as many have specific binding motifs that recognise host cell surface moieties. E. coli O157:H7 is unusal in that it appears to possess a number of major fimbrial types that possess gene defects and are not operable, the classic example being a 12bp deletion in the type 1 fimbrial operon. The discovery of genes sharing homology with the long polar fimbriae in S. Typhimurium that do contribute to adherence to the gut at least in the mouse model is intriguing. In summary, we report that expression of the lpf1 and lpf2 loci occurs in response to growth phase and different environmental conditions. In the case of lpf1, expression seems to be affected by the osmolarity and pH of the growth medium and lpf2 by iron depletion conditions. Furthermore, we provide data indicating that LPF is contributing to the colonization of lambs by E. coli O157:H7. Mutations in both lpf fimbrial operons diminished the ability of E. coli O157:H7 to persist in the intestine for two weeks but did not eliminate the capacity of the strain to colonize the GI tract. Our results strongly support the fact that redundant mechanisms (e.g. presence of two lpf loci) enable bacteria to thrive and survive within many different environmental conditions and therefore, mutations in both systems may eliminate the maintenance capacity of the bacterial population. Furthermore, our study underlines the importance of studying putative bacterial virulence factors in target animal host wherever possible.

Sequence analysis of E. coli O157:H7 compared with E. coli K12 demonstrates that O157 carries significantly more DNA on what can be described as a common core backbone genome shared by both strain types. It stands to reason that many of the virulence traits of O157 are encoded within these regions, the classic examples being the LEE island of some 44kb in size and the shiga-like toxin genes harboured within bacteriophages. Three O island knock-out mutants were create in the Gally (Edinburgh) laboratory and these were assessed in the six week old lamb model. The O-50 mutant was shed significantly less that the parental strain and it can be deduced that O-50 influences the ability of E. coli O157:H7 to colonise and persistence in the lamb model. Further work is now required to tease out what genes are contributing to which activity of colonisation and persistence.

The signature tagged mutagenesis approach is an elegant technique that can make a random mutant library from which large pools of mutants can be tested simultaneously. The method is very complex and time consuming. Furthermore, now that genome sequences are available, it is possible to interrogate specific genes based on pragmatic selection (ie/ sequence homology with known virulence determinants in other species or genes present within an O island) rather than random chance. The failure of the method was disappointing but the decision was taken not to repeat expensive animal model studies but rather focus on targeted analysis of the range of mutants that have been identified by ourselves and collaborators. This approach yielded analysis of 10 discrete gene (gene clusters) and delivered definitive data that showed the contribution to gastro-intestinal colonisation of five of them. This is an astonishingly effective yield of research activity from such a small programme of work!

OBJECTIVE 02

The primary objective of this section of the study was to define the role of parasites in enhancing EHEC O157:H7 colonisation and persistence in the sheep models. This arose from observations of high O157 shedding from an animal co-infected with Cryptosporidium. The question arose as to whether this was a reproducible phenomenon. If so, it was clearly a potentially significant risk factor given the high prevalence of Cryptosporidia in farmed animal ruminant species.

Following on from this objective, and not part of the original project, we investigated the role of colostrums and the native bacterial flora in influencing the shedding patterns of O157.

Introduction to the impact of Colostrum and Cryptosporidia on O157 colonisationTwo pathogens whose incidence in human disease has increased significantly over the last decade are

E. coli O157 and C. parvum. Both pathogens are of zoonotic importance and have potential economic implications to farmers.

Transmission of E. coli O157:H7 is faecal-oral with cattle considered to be the primary reservoir, although sheep and goats also are also recognised as significant reservoirs. The transfer of E. coli O157:H7 within adult animal species has been the subject of much research with few firm conclusions, however for young animals, colostrum deprivation is a risk factor with all gastro-intestinal infective agents (Kelleher & Lonnerdahl, 2001) and specifically O157:H7 in cattle (Dean-Nystrom et al., 1997; Rugsberg et al., 2003), pigs (Dean-Nystrom et al., 2002) and goats (La Ragione et al., 2005). Indeed, orphan lambs which are deprived of colostrum and are bottle

SID 5 (Rev. 3/06) Page 16 of 47

fed with milk replacer, on petting farms are a known source of human infection (Chapman et al., 2000; Pritchard et al., 2000; Payne et al., 2003). Hence the pertinence of this study.

Intestinal cryptosporidiosis is caused by C. parvum, a widely distributed protozoan parasite, which infects both wild and many domesticated animals, as well as humans, chiefly immunocompromised individuals (O’Dnoghue, 1995 and Griffiths, 1998). Since first its first description in cattle in Australia, (Barker & Carbonell, 1974), Cryptosporidium has been attributed an increasingly important role in neonatal diarrhoea syndrome in domestic livestock. A study by Munoz-Fernandez et al. (1996) showed C. parvum to be the most frequent aetiological agent involved in outbreaks of diarrhoea in lambs in Europe. Cryptosporidiosis is most common in young lambs (Foreyt, 1990) and may result in poor feed to weight conversion and occasionally death. During the incubation period and clinical course of the disease, the parasite proliferates mainly in the jejunum and the ileum. However, after the start of oocyst shedding the lesions can spread to other parts of the small and the large intestine (de Graaf et al., 1999).

Cryptosporidium has been shown to be a common agent associated with diarrhoeal disease in young calves in Europe and in a study by De la Fuente (1999) mixed infections with other entero-pathogens were reported in approximately 50% of cases of diarrhoea. Although, mixed infections are more commonly detected in clinically affected animals (Morin et al., 1980; Reynolds et al., 1986), the pathogenic significance of concurrent infections is unclear, particularly with regard to colonisation and shedding of E. coli O157:H7. Recently a goat experimentally infected with E. coli O157H7, shed high levels of the organism (107 g-1 faeces) and was found at post mortem examination to have a concurrent Cryptosporidium infection. The E. coli O157:H7 AE lesions and attached cryptosporidia were co-located on the mucosa of the large intestine (La Ragione et al., 2005). Also, the incidence of Cryptosporidium in orphan lambs in the UK has been reported as high (Pritchard, personal communication) the risk to human health is increased especially on petting farms.

Given that neonates and young animals are protected from most gastro-intestinal infections when sucking (Altmann & Mukkur, 1983), we wished to assess the longer term impact of deprivation of colostrum and ewes milk upon colonisation of lambs beyond weaning age with E. coli O157:H7. Additionally, we wished to assess the potential contribution of pre-infection with C. parvum in the lamb model used in these studies. Here we report our findings.

Materials and MethodsBacteria, parasite and inocula. A derivative of E. coli O157:H7 strain NCTC12900 that does not possess either stx1 or stx2 verocytotoxin genes (NCTC, Health Protection Agency, Colindale, London) was made resistant to nalidixic acid at 15 g ml-1 (Sigma) by passage on complex medium supplemented with the antibiotic and designated E. coli O157:H7 strain NCTC12900 Nalr (Best et al., 2005; La Ragione et al., 2005).

Strain NCTC12900 Nalr was stored in Heart Infusion Broth (HIB), (Oxoid) medium supplemented with 30% (w/v) glycerol on beads at –80oC and working stocks were stored at room temperature on Dorset’s egg medium.

NCTC12900 Nalr was streaked from Dorset’s egg medium onto SMAC plates (Oxoid) containing nalidixic acid (15 g ml-1) (Sigma) and well-isolated colonies were inoculated separately into 100 ml aliquots of LB broth (Oxoid) in 250 ml conical flasks. After incubation for 16 hours at 37˚C with gentle agitation the bacterial cells were harvested by centrifugation (3000 g for 10 min) and resuspended in PBS (0.1M, pH 7.4). The bacterial suspensions contained approximately 1 x 109 cfu. ml-1 as determined by serial dilution and plating on SMAC plates.

The C. parvum isolate used in these studies was from a field case of diarrhoea in a calf confirmed by standard VLA testing protocols to have been associated with C. parvum. The inocula for dosing lambs were made directly from the faeces from this calf by resuspension in PBS (0.1M, pH7.4) and centrifugation to remove faecal debris. Oocysts were suspended in PBS at 1 x 106 oocysts ml-1 and stored at +4oC until required.

Animals: Post partum, nine cross bred lambs from multiple births (leaving a least one lamb with its mother) were separated from their mothers immediately after birth and placed into two groups of three and six, respectively (A & B). These animals were bottle fed with milk replacer. A further twelve neonatal lambs were allowed to suckle and after weaning at approximately 4 weeks of age were separated from their mothers and placed in two groups (C & D) of five and seven animals, respectively. Each animal was identified with duplicate ear tags encoding a unique four digit identification number. The animals were housed indoors and provided with standard rations and water ad libitum. At 4 weeks of age and prior to inoculation with E. coli O157:H7 or Cryptosporidium, faeces were taken per rectum from each animal and cultured for E. coli O157, as described below. The same samples were tested for cryptosporidia by a modified Ziehl Neelsen and Indirect Fluorescent Antibody Tests (see below). At five weeks of age the lambs in groups B and D were dosed orally with 5 x 105 and 1 x 106 cryptosporidia oocysts, respectively. At six weeks of age all groups (A-D) were dosed orally with 1 x 10 10 cfu. of E. coli O157:H7. All inocula were delivered in a 10ml volume using a worming gun (Novartis Animal Health) ensuring that the whole inoculum was delivered directly to the pharynx. All procedures were conducted under the jurisdiction of Home Office licence 70/5441 granted under the Animals (Scientific Procedures) Act (1986).

Preparation and necropsy procedures for animals studied post mortem. Two colostrum deprived lambs (tagged 1458 & 1470, both group B) and two conventional lambs (tagged 1486 & 1489, both group D) that had been

SID 5 (Rev. 3/06) Page 17 of 47

challenged with both C. parvum and E. coli O157:H7 were euthanased and necropsied as described previously (Wales et al., 2001a, b). Briefly, one lambs from each group was examined at 24 (1458 & 1486) and 96 (1470 & 1489) hours after challenge with E. coli O157:H7. These animals were euthanased with a barbiturate overdose and tissue samples were collected immediately thereafter. Tissues samples (1g) were collected from the rumen, duodenum, jejunum, ileum, caecum, ascending colon, spiral colon, recto-anal junction (RAJ) and from six sites excised at approximately 2cm apart measured from the RAJ, along the rectum toward the distal colon. The samples were submitted for bacteriological, parasitological, histological examination and for confocal and transmission electron microscopy as described below. Sterile plastic clips and bags were used to enclose the cut ends of the intestinal tract as dissection proceeded, as described previously (Wales et al., 2001a), to prevent cross-contamination of samples.

Bacteriological examination. Faecal samples taken prior to oral inoculation were examined for the presence of E. coli O157 following previously described methods (Wales et al., 2001a, b; 2005; Woodward et al., 2003). Briefly, faeces (1g) were re-suspended in 9ml Buffered Peptone Water (BPW), incubated at 37oC for six hours, statically and E. coli O157:H7 organisms recovered by O157 specific IMS (Dynal) and plated onto CT-SMAC plates (Oxoid, Basingstoke, UK).

To detect the inoculated strains, previously described methods were followed (Wales et al., 2001a, b; 2005; Woodward et al., 2003; La Ragione et al., 2005). Faeces (1g) were re-suspended in 9 ml BPW (Oxoid, Basingstoke, UK) by vortexing and serial dilutions were plated directly onto SMAC plates containing nalidixic acid (15 g ml-1). Additionally, dilutions were retained overnight at 4C and, if there were no direct counts of confirmed E. coli O157 organisms, the dilutions were incubated at 37C for six hours and samples were then plated on SMAC plates containing nalidixic acid (15g ml-1). The serogroup of bacteria recovered by these processes was verified by E. coli O157-specific latex agglutination (Oxoid, Basingstoke, UK). Tissue samples collected from lambs for bacteriological examination were homogenised in BPW (Oxoid, Basingstoke, UK) (1 g in 9 ml) and processed as described above. Parasitological studies. Tissues and faeces were examined for the presence of cryptosporidia by both the modified Ziehl Nielson (mZN) and Indirect-fluorescent antibody test (IFAT). The methods used were briefly as follows : Modified Ziehl Neelsen - A smear of fresh faeces or a deep tissue scrape was prepared on a clean grease-free microscope slide, air dried and then passed through a flame to fix. The fixed slide, was then immersed in cold Carbol-Fuchsin for 25 minutes, rinsed thoroughly in H2O, decolourised in 5% sulphuric acid for 90 seconds, rinsed in H2O. The preparation was then counter stained in 5% malachite green for 3 minutes, rinsed in tap water, air dried and examined at x400. Indirect Fluorescent Antibody Test – The fresh faecal sample or deep tissue scrape was smeared within the well of a 4-well microscope slide and fixed by carefully applying 50μl of methanol to each well and then allowed to air dry. Once dry 30μl anti-Cryptosporidium Immuno-Fluorescent Antibody (IFA) was applied and the slide placed in a humidified staining chamber, and incubated for 45 minutes at 37oC. Excess IFA stain was carefully aspirated from each well and replaced with distilled H2O for approximately 1 minute. Excess H2O was aspirated and the smear air dried. Once dry 20µl of UV-compatible mounting fluid was added and a cover slip placed over the preparation. The cover slip was then sealed with DPX. Preparations were examined using a fluorescence microscope fitted with a 450-490 nm excitation filter and x20 or x40 objectives. Cryptosporidium were identified as apple-green epi-fluorescing, slightly ovoid to spherical oocysts, 4 to 6µm in diameter. Light microscopy At necropsy, tissues were placed immediately into 10% neutral buffered formalin at room temperature and left to fix for at least 24 hours. Trimmed tissues were processed routinely to paraffin wax. Sections were cut at 4μm and stained with Haematoxylin and Eosin (H & E). Selected tissues were also stained using the Giemsa method. All tissues were observed using an Olympus CX41 microscope. Immunohistochemistry Tissue blocks were fixed in 10% neutral buffered formalin, processed to wax and sectioned at 4μm on a rotary microtome in preparation for immunohistochemistry. Tissue sections were de-waxed in xylene, and dehydrated in absolute alcohol before immersion in freshly prepared 3% hydrogen peroxide-methanol for 10 minutes to inhibit endogenous peroxidase activity.

Sections were rehydrated in running tap water prior to assembly into the Shandon Sequenza ® staining system. The slides were washed with 2 x sodium chloride tris buffered Saline (0.005M TBS, pH 7.6 1.7% NaCl) for 5 minutes before incubation at room temperature with normal goat serum (Vector Labs, UK) for 20 minutes. E. coli O157 specific polyclonal antibody, raised in rabbits (VLA, Weybridge), was then applied (1/1000 and 1/5000 diluted in 2 x NaCl TBS supplemented with 5% normal rabbit serum) to the sections for 1 hour at room temperature (Wales et al., 2001b).

SID 5 (Rev. 3/06) Page 18 of 47

E. coli antigens were visualised following incubation with biotinylated goat anti-rabbit IgG (1/200 dilution, with normal goat serum at a 1/66 dilution in 2 x NaCl TBS supplemented with 3% normal sheep serum) for 30 minutes at room temperature, an avidin–biotin-peroxidase conjugate (Vector Labs, UK) for 30 minutes at room temperature and citrate-buffered diaminobenzidine (DAB) for 10 minutes at room temperature. The sections were then counter-stained in Meyer’s Haematoxylin, before being dehydrated in absolute alcohol, cleared in xylene and cover-slipped using DPX mountant. The stained tissue sections were then examined by light microscopy, using an Olympus CX41 microscope. Transmission electron microscopy. At necropsy, tissues were immediately fixed in 3% gluteraldehyde prepared in a 0.1M phosphate buffer. Following microscopical examination of the H and E sections, at sites where lesions were identified, corresponding gluteraldehyde fixed tissues were cut to 1-2mm in thickness for transmission electron microscopical (TEM) examination. Briefly the tissues were then washed in 0.1M phosphate buffer, post fixed in 1% osmium tetroxide, dehydrated by immersion in a series of alcohol solutions increasing to 100% alcohol and placed in propylene oxide prior to embedding in araldite resin. The resin was polymerised at 60C for 48 hours. One-micron sections were stained with toluidine blue for light microscopical examination. Ultra-thin sections at 70-90nm thickness were then prepared onto copper grids using a diamond knife and stained with uranyl acetate and lead citrate prior to examination using a Phillips CM10 TEM.

Confocal microscopy. After de-waxing and rehydration, tissues were placed in PBS prior to preparation for examination by confocal microscopy. Briefly, sections were permeabilised in PBS containing 0.1% Triton X-100 followed by detection of E. coli using a fluorescein isothiocyanate (FITC)-labelled affinity purified antibody to E. coli O157:H7 produced in goats. The sections were then washed thoroughly using PBS and mounted in Vectashield containing DAPI (Vector Laboratories). Images of the FITC-labelled E. coli were obtained by confocal laser scanning microscopy using a Leica TCS SP2 AOBS confocal system attached to a Leica DM IRE2 microscope equipped with ArKr laser excitation (488nm) and a diode laser (405nm). An oil-immersion objective lens (63x, N.A. 1.32) was used, and imaging parameters were selected to optimise resolution.Statistical analyses. The sensitivity of detection by direct plating was approximately 500 organisms per gram of faeces. Samples positive by enrichment were considered to have up to 500 organisms per gram, and those samples in which no organisms were detected were given an arbitrary value of 1 to avoid the issue of a zero value giving results to infinity. T-tests were used to compare the mean counts transformed to log10 (count + 1) at each time point.

RESULTSClinical findings

All animals in all study groups remained clinically normal throughout the experiment, with no evidence of pyrexia or diarrhoea. However, a change in faecal consistency was observed for the animals in groups B and D challenged with Cryptosporidium between 5-8 days after oral dosing, whereby the faeces became smooth and sticky in consistency. Faecal shedding of Cryptosporidium

Prior to experimental inoculation with cryptosporidia, all animals were confirmed to be free from cryptosporidia by analysis of faeces by mZN and IFAT. The Cryptospridium dose was approximately 5 x 105

oocysts for the colostrum-deprived lambs (Group B) and 1 x 106 oocysts for the conventional lambs (Group D). In both the conventionally reared and colostrum-deprived lambs Cryptosporidium was detected in the faeces 24 hours after inoculation in all animals except 1493 (group D) which remained negative throughout the study. In addition, cryptosporidia were only detected on day one post inoculation in lamb 1487. In all other lambs the organism persisted for between 8 and 12 days with the notable exceptions of lambs 1492 and 1496 (both group D) where cryptosporidia were detected in the faeces 39 days post inoculation.

Faecal shedding of E. coli O157:H7 All faeces samples collected from the animals prior to the experimental procedure were negative for

E. coli O157 as assessed by immuno-magnetic separation. The oral inoculation doses of E. coli O157 were determined to be 1 x 1010 cfu. for each lamb in each of the four groups. At 24 hours after oral inoculation only two animals (1487 and 1492, both Group D) were not shedding E. coli O157:H7 organisms although by 72 hours after oral inoculation all animals in all groups were shedding in the range 103 to 108 cfu. g -1 faeces. For the colostrum-deprived lambs on day 3 the results indicated a significant difference (p=0.038) with the counts of E. coli O157:H7 organisms for lambs dosed with E. coli alone being higher than those pre-dosed with C. parvum. Thereafter the O157:H7 counts were consistently lower than those from lambs pre dosed with C. parvum. For the conventionally reared lambs on day 3 the results indicated a significant difference (p=0.019) with the counts of E. coli O157:H7 organisms for lambs dosed with E. coli alone being higher than those pre-dosed with C. parvum. Thereafter the E. coli O157:H7 counts were consistently lower than those from lambs pre dosed with C. parvum.

We have shown previously (Wales et al., 2001b; Woodward et al., 2003) that in this age of conventional lamb, faecal shedding of E. coli O157:H7 after deliberate oral dosing declines gradually from 107 cfu. g-1 faeces to 103 cfu. g-1 faeces over the first 10-14 days and thereafter faecal shedding is intermittent. Similar findings were shown here also for early sampling (days 1, 3, 8 and 11 or 12). From day 15 onward, as anticipated the numbers

SID 5 (Rev. 3/06) Page 19 of 47

of animals shedding in each group was low and the statistical analysis indicated that differences between the groups were not significant. However, the shedding frequency including and beyond day 15 (expressed as the fraction of the number of positive faecal samples over the number of faecal samples examined per test group) was 0.16 for group C [conventional], 0.33 for group D [conventional plus Cryptosporidium], 0.46 for group A [colostrum-deprived] and 0.625 for group B [colostrum-deprived plus Cryptosporidium].

Bacteriological and parasitological findings at necropsyNo overt gross pathological changes were observed at necropsy in any of the four animals examined at

24 or 96 hours after oral inoculation, although moderate congestion of the intestines was noted in all animals examined.

There were differences in the numbers of E. coli O157:H7 strain NCTC12900 Nalr recovered from tissues in individual animals (Fig. 2). E. coli O157:H7 was not detected in the duodenum of any animals examined or the jejunum of animals 1458, 1486 and 1489. E. coli O157:H7 was only recovered from the ileum and rumen of animals 1470 and 1486. By contrast in the large intestine, E. coli O157:H7 was detected in all sites examined in all four animals and in high numbers (5 x 106 cfu. g-1 tissue) in tissues from the rectum and spiral colon of animal 1470 the E. coli O157:H7 counts for the caecum of animal 1458 were particularly high (1 x 10 7 cfu. g-1 tissue) and abundant AE lesions were observed in these tissues (see below). The overall numbers of E. coli O157:H7 recovered from the gastro-intestinal tract was notably higher (approx. 2 logs) in the colostrum-deprived lambs compared to the conventionally reared lambs. I

In lamb 1458 (colostrum-deprived plus Cryptosporidium), necropsied 24h after E. coli O157:H7 infection, cryptosporidia were detected in the duodenum, jejunum, ileum, rumen, ascending colon, spiral colon, caecum and proximal rectum. Cryptosporidia were found in the duodenum, rumen and spiral colon of lamb 1486 (conventional plus Cryptosporidium) necropsied 24h after E. coli infection. No Cryptosporidia were detected in tissues collected at necropsy of animals 1470 (colostrum deprived plus Cryptosporidium) and 1489 (conventional plus Cryptosporidium) necropsied 96h after E. coli infection.

Histopathological findingsOn histopathological examination significant changes were confined to the caecum, rectum and RAJ of

animals 1458 (colostrum-deprived plus Cryptosporidium, 24h), 1470 (colostrum-deprived plus Cryptosporidium, 96h p.i.) and 1489 (conventional plus Cryptosporidium, 96h p.i.). The mucosa of the large intestine was moderately well preserved, with only small areas having some loss of epithelium in some individual sections. The majority of the mucosa at all sites appeared normal. However, some degenerative cellular changes and visible numbers of scattered foci of closely adherent bacteria were seen in the rectum, RAJ and caecum. AE lesions were not identified in animal 1486 (conventional plus Cryptosporidium) examined 24h p.i.), however, E. coli O157:H7 positive bacteria were identified by IHC in the lumen of the caecum and rectum of this animal. Cryptosporidia were not observed co-located with bacteria on the mucosal surface. However, where Cryptosporidium lesions occurred these were associated with inflammatory infiltrates in the lamina propria with moderate numbers of mixed lymphoid cells, eosinophils and polymorphonuclear leukocytes. The remaining large intestine (ascending colon and spiral colon) were negative for cryptosporidia and attached bacteria in all animals.

Sections of small intestine (duodenum, jejunum and ileum) examined were less well preserved than the large intestine and some artefactual separation of the underlying lamina propria had occurred.. The epithelium in the majority of samples was separating from the villi, although representative epithelium remained. Cryptosporidium was only identified in the ileum of animal 1458 (colostrum-deprived, 24h). Small numbers of coccidia were identified in the ascending colon of animal 1486 (conventional plus Cryptosporidium, 24h). Abnormalities were not detected in the rumen.

Immunohistochemistry & confocal microscopyMultifocal, variable sized colonies of E. coli O157:H7 organisms were identified in three of the four

animals examined (1458, 1470 & 1489). E. coli O157:H7 colonies were identified most frequently in the rectum, although E. coli O157:H7 colonies were also identified in the caecum and RAJ. Lesions tended to involve 1 or 2, to several adjacent epithelial cells. Lesions identified in the mucosa at the terminal rectum were found on both lymphoid-associated epithelium and non-lymphoid-associated epithelium. Transmission electron microscopy

Attaching-effacing lesions associated with bacteria were confirmed on the intestinal mucosa. Cryptosporidia were not observed co-located with intimately attached bacteria, however they were observed in the ileum of one animal (1458, colostrum-deprived plus Cryptosporidium). confirming the light microscopic findings.

SID 5 (Rev. 3/06) Page 20 of 47

a) b) c)

Fig 6. Confocal (a) illustrating O157 attached to the gut mucosa (green = O157, blue nuclei of gut cells) and (b) transmission electron microscopy illustrating E. coli O157 specific attaching and effacing lesions and (c)Cryptosporidium in the distal intestine of experimentally orally infected (Cryptosporidium and E. coli O157) six week old conventionally reared lambs.

DISCUSSIONIt is established that E. coli O157:H7 colonises the intestinal tract of ruminants often leading to persistent

infection. In a conventional six-week-old lamb model it has been previously reported that the intimin of E. coli O157:H7 is a significant factor for colonisation and persistence, but that AE lesions for which intimin is essential, were small and extremely sparse in infected animals. The most striking feature of the present study was the frequency, density and distribution of E. coli O157:H7 induced AE lesions observed in the animals. The experimental differences between studies that may have contributed to the observed outcomes were the deprivation of colostrum and ewes milk and the pre-infection with C. parvum.

It has been demonstrated in neonatal calves that colostrum is highly protective against O157:H7 challenge and that AE lesions are readily induced in the distal gastro-intestinal tract if colostrum is with-held (Dean-Nystrom et al., 1997). In conventional, six-day-old lambs it has been shown that small and sparse AE lesions were induced by E. coli O157:H7 (Wales et al., 2001b). Given this data, it seems reasonable to assume that neonatal lambs deprived of colostrum would be more susceptible to colonisation by E. coli O157:H7 with extensive AE lesion formation. Of concern, and relevant to this study was the issue of longer term sequelae for lambs deprived of colostrum and ewes milk, such as orphan lambs that are hand reared and found frequently on petting farms. The lamb model used in this study therefore attempted to mimic this scenario. The experimental evidence indicated that E. coli O157:H7 induced AE lesions were detected readily in the distal gastro-intestinal tract of lambs deprived of colostrum and ewes milk and that faecal shedding of E. coli O157:H7 organisms was greater than for conventional animals of the same age. By comparison with a previous study, in which six-week-old conventional lambs were challenged with the same E. coli O157:H7 strain, and in which a similar number of tissues were examined (Woodward et al., 2003), only two lesions were identified, in one animal, one in the caecum and one in the rectum. The present data also support the concept that colostrum is essential for the longer term health of the gastro-intestinal tract, even beyond weaning (Butler, 1979).

We questioned the impact of C. parvum on E. coli O157:H7 infection following the incidental observation of high faecal shedding of E. coli O157 and readily identified multifocal AE lesions associated with C. parvum in the large intestine of a deliberately orally infected animal (La Ragione et al., 2005). This raised the question whether C. parvum may have an effect on O157:H7 colonisation and shedding in the lamb models used in this study. Concurrent infections with two or more enteropathogens, including cryptosporidia in association with E. coli, in naturally infected diarrhoeic and asymptomatic calves and goats have been described previously (Moon et al., 1978; Janke et al., 1990; de la Fuente et al., 1999; Gunning et al., 2001). In natural and experimental infections in goats and lambs, C. parvum causes severe clinical disease often with high morbidity and mortality (Tzipori et al., 1981) and severe lesions are induced often in the posterior jejunum and ileum (Koudela & Jiri, 1997). However, asymptomatic carriage of cryptosporidia has been described in adult goats (Noordeen et al., 2002; La Ragione et al., 2005). In the study reported here, lambs challenged with C. parvum remained asymptomatic, but a few animals shed cryptosporidia for up to 39 days post inoculation. On microscopical examination of the four lambs examined post-mortem, cryptosporidia were only found associated with the mucosa of the ileum in one lamb and not co-located with O157:H7 induced AE lesions in the large intestine. It is well documented that the formation of O157:H7 AE lesions is the result of actin polymerisation (Shaner et al., 2005) and similarly Cryptosporidium utilises actin polymerisation to initiate infection. This may explain the opportunistic co-localisation that has been described (Elliott and Clark, 2000; Elliott et al., 2001) previously between E. coli O157:H7 and Cryptosporidium in a goat (La Ragione et al., 2005). The absence of cryptosporidia in the large intestine may indicate the model used in this study was not appropriate, especially with regard to the timing of

SID 5 (Rev. 3/06) Page 21 of 47

super-infection with E. coli O157:H7, to reproduce the incidental finding in the goat study (La Ragione et al., 2005). However, the relative frequency with which O57:H7 induced AE lesions were observed in one conventional animal inoculated with C. parvum prior to E. coli O157:H7 may indicate predisposition to colonisation.

The shedding profiles of O157:H7 NCTC12900 irrespective of the study group were broadly similar to each other, and consistent with shedding profiles shown in previous studies (Wales et al., 2001b). The numbers of O157:H7 organisms that were shed declined over the first 10-14 days after inoculation followed by intermittent shedding thereafter for twenty days or so. With the exception of two time points early after oral inoculation, there were no statistically significant differences between groups. However, a trend was shown that suggested colostrum deprivation and prior C. parvum challenge increased the number of positive faecal samples compared to the conventional animals. The duration of shedding however, was not increased.

Collectively, the data produced in this study suggest that lambs may be predisposed to O157:H7 colonisation by colostrum deprivation and concurrent infection with C. parvum. This raises the question as to whether other pathogenic agents may predispose to O157:H7 colonisation. The number of animals in the study was a limiting factor and trends, rather than statistically confirmed differences between treatment groups, were observed. These trends need to be examined further before firm conclusions can be drawn. However, the potential public health implications of the findings of this study are of particular importance. Lambs on petting farms are frequently orphans which may not have received adequate colostrum (La Ragione et al., 2005a). A consequence, may be more susceptible to infection with Cryptosporidium, E. coli O157:H7 or other gastro-intestinal pathogens. It is widely accepted that ruminants are the primary source of E. coli O157:H7 and Cryptosporidium infections for people and so a greater understanding of the relationship between these two enteropathogens may lead to effective intervention strategies.

Interaction between Attaching-Effacing Escherichia coli O26:K60 and O157:H7 in young lambs: O157:H7 suppresses faecal shedding of O26:K60.

The administration of feed supplements such as Lactobacillus spp. and multiple probiotic bacteria have been shown to reduce the carriage and faecal shedding of various pathogenic E. coli in cattle (Ogawa et al., 2001; Tkalcic et al., 2003) and lambs (Lema et al., 2001). However, Zhao et al. (2003) found no significant difference in faecal shedding of E. coli O157:H7 in calves between probiotic-treated and untreated groups. Competition and dominance of certain E. coli serotypes is suggested to be one explanation for the shedding of one predominant E. coli strain over time (Midgley et al., 1999). Interestingly, it has been shown that E. coli serogroup O26 was prevalent in ruminants at slaughter, whereas the prevalence of O157 in the same studies were very low (Aktan et al., 2004; Fukushima & Seki, 2004). More recently, it has been shown that E. coli O26:K60 colonised 6-week-old lambs after oral dosing, with persistent shedding for well over a calendar month and with the induction of AE lesions that were small and sparse in the distal GIT (Aktan et al., 2007). These findings were not dissimilar to the behaviour of EHEC O157:H7 in the same models (Wales et al., 2002a, b; Woodward et al., 2003) even though the O26:K60 strain harboured the -intimin and the O157:H7 harboured the -intimin, which is considered to effect the tropism to different locations in the GIT (Phillips et al., 1998. Also, it was reported on the interactions between O26:K60 and O157:H7 in tissue culture adherence assays that showed pre-incubation of tissue culture cells with either one strain reduced significantly the extent of adherence of the strain that was applied second (La Ragione et al., 2004). Given these data, we wished to test whether the colonisation and shedding of lambs by E. coli O157:H7 may be altered, possibly reduced, if the lambs are previously colonised by E. coli O26:K60. Here competitive in vitro tests were described prior to performing a dual oral inoculation study with 6-week-old conventional lambs.

Materials and methods

Bacterial strains, growth and culture conditions. E. coli O26:K60 strain EC335/98, obtained from the Veterinary Laboratories Agency (VLA; Weybridge, United Kingdom) culture collection, was of sheep origin, lacked stx1 or stx2

toxin genes, possessed the β1-intimin gene, was naturally resistant to streptomycin (>250ug/ml) and has been described previously. E. coli O157:H7 NCTC12900nalr strain lacks stx1 or stx2 toxin genes, possesses the γ-intimin gene and has been fully characterised and described previously. Bacterial inocula for adhesion and invasion assays and for in vivo studies were prepared as described previously. Confirmation of bacterial numbers was done by plating ten-fold serial dilutions onto Sorbitol MacConkey (SMAC) agar plates supplemented with antibiotics as appropriate.

To test for inhibitions between the O26:K60 and O157:H7, the strains were cross-streaked onto separate 5% blood and LB agar plates. Also, overnight broth cultures of each strain were diluted in fresh pre-warmed medium to a concentration of circa 1 x 105 CFU ml-1 each and equal volumes were mixed and grown at 37oC and CFU ml-1 estimated by plating serial dilutions onto LB agar plates at 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 hours. Each strain was inoculated into fresh broth, filter sterilised (0.2μm) spent broth from overnight growth of the homologous strain and filter sterilised (0.2μm) spent broth from overnight growth of the heterologous strain.

Recovery and enumeration of bacteria from in vivo studies.

SID 5 (Rev. 3/06) Page 22 of 47

(a) direct counting. Faeces, intestinal contents and tissue samples (1g) were collected as described (see below) and immediately placed into 9 ml volumes of buffered peptone water (BPW). The samples were homogenised and plated directly onto SMAC agar supplemented with the appropriate antibiotic. (b) immunomagnetic bead separation (IMS). The BPW homogenates collected as described in (a) above were incubated at 37°C for 6 h and tested by IMS using anti-E. coli O26 or O157 Dynabeads® (Dynal, Oslo) using the automated BeadRetriever (Dynal, Oslo), as directed by the user manual. One hundred microlitres of the final washed bead suspensions were streaked onto SMAC agar. Selected single colonies were tested by O157-specific latex agglutination (LA; E. coli O157:H7 Test, Oxoid) or by O26:K60-specific antisera (VLA).

Bacterial adherence and invasion assays and confocal microscopy. Bacterial adherence and invasion methods and confocal microscopy were performed as described previously.

Animals. All animal studies were performed in accordance with the Animals (Scientific Procedures) Act (1986) and were approved by the local Ethical Review Committee. All animals used in these studies were screened for excretion of E. coli O157 and O26 serogroups by IMS on two separate occasions prior to experimentation.

In vivo lamb experiment. An oral inoculation ovine model was used as described previously (31) with modifications. A total of 16 six-week-old conventional cross bred lambs were housed in three groups. On day 1, lambs in group A (n =8) and control group B (n = 5) were dosed orally with 1010 CFU of E. coli O26 strain EC335/98 strain just before morning feeding. On day 4, lambs in group B and control group C (n = 3) were dosed orally with 1010 CFU of E. coli O157:H7 NCTC12900nalr strain. On day 6, two lambs selected at random from group A were euthanased by intravenous injection of barbiturate (Somulose®). The tissue samples collected for bacteriological analysis (see above) and histology analysis (see below). Tissue samples included rumen, duodenum, jejunum, ileum (six sections of ileum; were sampled starting approximately 5cm from ileo-caecal junction), ascending colon, spiral colon, caecum, rectum (six parts of rectum; sections were taken starting approximately 2 cm from RAJ) and RAJ.

Pathological studies. Light microscopy and Immunohistochemistry were performed essentially as described previously (14, 31). Briefly, tissues were fixed in 10% neutral buffered Formalin, processed to wax and sectioned at 4μm. Sections were rehydrated prior to assembly into the Shandon Sequenza® staining system. The slides were washed with 2 x sodium chloride Tris Buffered Saline (0.005M TBS, pH 7.6 1.7% NaCl) for 5 minutes before incubation at room temperature with a normal goat serum (Vector Labs, UK). E. coli O26 or O157- polyclonal antibodies, raised in rabbits (VLA; Weybridge, United Kingdom), were then applied (1/1000 and 1/5000 diluted in 2 x NaCl TBS supplemented with 5% normal rabbit serum). E. coli antigens were visualised through incubation with biotinylated goat anti-rabbit Immunoglobulin G. The sections were then counter-stained in Meyer’s Haematoxylin.

Statistical analysis. For statistical analyses of assays counts were transformed to log 10, one being added if there were zero counts, and analyses of variance (ANOVA) performed. Counts for O26:K60 were compared in Group A (6 lambs) and Group B (5 lambs) and counts for O157:H7 were compared in Group A and Group C (3 lambs).

RESULTS

Interaction between O157:H7 and O26:K60 when cultured in vitro E. coli strains O26:K60 and O157:H7 were cross-streaked on 5% Sheep’s blood and LB agar plates and no inhibition of growth of either strain was observed. Spent medium growth inhibition assays were performed and the E. coli O26:K60 spent medium culture reduced both the growth rate and final density of E. coli O157:H7 by approximately one log10 at the termination of the experiment. By contrast, the E. coli O157:H7 spent medium culture did not cause growth inhibition of E. coli O26. Both strains were susceptible to homologous growth suppression; the O26:K60 strain more so than the O157:H7 strain.

SID 5 (Rev. 3/06) Page 23 of 47

Fig 6. Confocal microscopy illustrting in vitro E. coli O157 (Red) and O26 (green) interaction studies, illustrating that E. coli O26 reduces O157 colonisation of cultured epithelial cells. (C = O26 alone and D, E and F represent competitive exclusion assays with O26 and O157.

O157:H7 and O26:K60 mutually suppress adherence to HEp-2 cells When HEp-2 cells were inoculated with E. coli O26:K60 and incubated for 30 min prior to the addition of E. coli O157:H7, the numbers of E. coli O157:H7 that adhered and invaded were lower than when O157:H7 was assayed alone (P = 0.002, P = 0.024, respectively). In the converse experiment, when HEp-2 cells were inoculated with E. coli O157:H7 and incubated for 30 min prior to the addition of E. coli O26:K60, the numbers of E. coli O26:K60 that adhered was lower than when the organisms were assayed alone (P = 0.039). There was no significant difference for invasion (P > 0.05).

Oral inoculation of lambs with O157:H7 was associated with reduced shedding of O26:K60 All lambs were clinically normal, healthy and free from disease at the onset of the study. None were shedding E. coli O26 or O157 as tested by IMS. No clinical signs were observed in any lamb in any group after oral inoculation with the challenge strains and the shedding results are summarised in Figures 7 and 8.

At 24h after oral inoculation with E. coli O26:K60, the numbers of O26:K60 organisms in the faeces in all animals were in the region of 109 CFU/g faeces. For the animals that were dosed with E. coli O26:K60 alone, the numbers of O26:K60 organisms recovered from faeces stabilised at 105 to 106 CFU/g faeces until the termination of the experiment at day 35. However, for the animals that were dosed with E. coli O26:K60 and then with E. coli O157 four days later, the numbers of O26:K60 organisms recovered from faeces declined steadily such that from day 12 the shedding was sporadic and the mean count was approximately 103 CFU/g faeces. From day 27 onwards, O26:K60 organisms were not detected from any animal on the three occasions tested. The differences between recoveries between these two study groups were statistically significant. Specifically, the linear component (days 12-35) differed significantly between the two groups (P = 0.010) with a lower rate of decline for animals dosed with E. coli O26:K60 alone.

At 24h after oral inoculation with E. coli O157:H7, the numbers of O157:H7 organisms in the faeces in all animals were in the region of 105 to 106 CFU/g faeces. For the animals that were dosed with E. coli O157:H7 alone or animals pre-dosed with E. coli O26:K60, the numbers of O157:H7 organisms recovered from faeces stabilised at about a mean value of 104 CFU/g faeces until the termination of the experiment at day 35 although from day 14 onwards shedding became sporadic. There was no evidence of any consistent difference between these two groups (P = 0.850). However, the shedding of O157:H7 organisms in animals pre-dosed with E. coli O26:K60 declined rapidly between days 5 and 7 compared with the animals dosed with E. coli O157:H7 alone but this trend was not significantly different between the groups (P = 0.190), (Figs 7 and 8.).

SID 5 (Rev. 3/06) Page 24 of 47

-1

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 17 18 19 20 21 24 25 27 28 31 32 34 35

Days post inoculation

CFU

log1

0/gr

am fe

ces

5141651417514185148051488514045140651408514225142151486Mean GroupBMean GroupA

Fig 7. Shedding pattern of E. coli O26:K60 from lambs orally dosed with O26:K60 alone at day 1 (individual lamb scores given as blue, mean shown as full line) and lambs orally dosed with both O26 at day 1 and O157:H7 at day 4 (individual lamb scores given as red, mean shown as dotted line).

-1

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 17 18 19 20 21 24 25 27 28 31 32 34 35

Days post inoculation

CFU

log1

0/gr

am fe

ces

514045140651408514225142151486514055141551484Mean GroupCMean GroupA

Fig 8. Shedding pattern of E. coli O157:H7 from lambs orally dosed with O157:H7 alone at day 4 (individual lamb scores given as blue, mean shown as full line) and lambs orally dosed with both O26 at day 1 and O157:H7 at day 4 (individual lamb scores given as red, mean shown as dotted line). Arrow points to day 4 where selected animals were dosed with E. coli O157:H7.

In conclusion, the statistical examination of shedding data indicated that the counts for O157:H7 were unaffected by the competition from O26:K60, but the O26:K60 counts were lower when competing with O157:H7.

Two lambs (51410 and 51413) selected at random from the group of animals that were dosed with E. coli O26:K60 and 4 days later with E. coli O157:H7 were examined at post-mortem on day 6. E. coli O26:K60 bacteria were recovered by direct counts from all sites tested except for rumen, duodenum, jejunum and caecum of one lamb and rumen and jejunum of the second lamb. E. coli O157:H7 bacteria were recovered by direct counts from caecum, descending colon, rectum 4, rectum 3, rectum 1 and RAJ of one lamb only. IHC analysis showed O26 stained bacteria intimately associated with the mucosa of the descending colon of lamb 51410, producing small sparse focal AE lesions. No evidence for intimate association of O157 stained bacteria was gained by IHC in the same animals.

DISCUSSION

The data generated in this study indicated that challenging 6-week-old lambs previously colonised with E. coli O26:K60 with an oral dose of E. coli O157:H7 was associated with a statistically significant reduction of shedding of the O26:K60 strain. This observation is, to the best of our knowledge, unique and points to complex interactions between these two strains and the host.

Growth of organisms in spent culture media assays has been undertaken by numerous groups in order to test inhibitory activities (Helling et al., 1987; Mack et al., 2003; Pérez et al., 2001). For example, Lactobacillus spp. spent culture supernatant has been reported to exert an antimicrobial effect against Salmonella enterica

SID 5 (Rev. 3/06) Page 25 of 47

serovar Typhimurium and other intestinal pathogens (De Keersmaecker et al., 2006). The studies presented here demonstrated that the spent medium from E. coli O26:K60 suppressed the rate of growth of E. coli O157:H7. This effect may also be due to quorum sensing (Falcao et al., 2004; Reading & Sperandio, 2006; Sperandio et al., 2003) or another more overt inhibitor such as a colicin, although cross streaking gave no evidence to support this. Also, it was shown that pre-incubation of HEp-2 cells with either strain of E. coli reduced significantly the extent of adherence of the strain that was applied second confirming previous findings of La Ragione et al. (2004a; b). It is known that the presence of one microorganism in a specific niche in a host suppresses the colonisation of that niche by another (Brogden et al., 2005) for which several mechanisms have been suggested including the production of metabolic inhibitors and the competition for attachment sites (Hakkinen & Schneitz, 1996; Stavric et al., 1993). Etcheverria et al. (2006) suggested that the presence of other E. coli strains influence the localization of E. coli O157:H7 in the GIT. Collectively, the in vitro findings reported in this study and previously (La Ragione et al., 2004), supports the hypothesis that primary colonisers displace secondary challenge strains in vivo. Whilst this may seem obvious, this assumes similar modes of adherence and tropism to the same target on the host cells between test strains and absence of other competitive mechanisms such as bacteriocins, lytic bacteriophages and so forth. One explanation for these data is that there may be a limited number of sites on the host cell surface to which AEEC, irrespective of intimin type, may associate with and therefore there may be competition for these sites in a dual inoculum study. However in a previous study, confocal microscopy demonstrated that the extent of microcolony formation by O157 when applied to cultured epithelial cells after O26 seemed to be impaired. The post-mortem data at day 6 suggests that E. coli O157:H7 organisms were not readily recovered from sites along the GIT previously colonised by E. coli O26:K60 which were recovered at levels of 105

to 107 CFU g-1 tissue at the majority of sites examined. The post-mortem data are limited to one time point soon after challenge with E. coli O157:H7, and it is possible that this density of O26:K60 organisms were sufficient to induce quorum signalling or block binding sites of O157:H7 organisms in the proximal GIT, if these exist. However, over the duration of the study and notably from day 12 onwards, O157:H7 organisms were associated with reduction of O26:K60 shedding. Thus, it must be concluded that the suppressive effects seen in vitro were not repeated in vivo.

In the present study, the pattern of shedding of E. coli O26:K60 from animals dosed with this organism alone was similar to that shown in an earlier study (Aktan et al., 2007) in which the lambs were well colonised with high level shedding persisting for beyond a calendar month. Small sparse AE lesions were identified in the distal GIT early after inoculation and, by day 35, E. coli O26:K60 organisms were located primarily in the ileum and not the distal GIT (Aktan et al., 2007). Given this knowledge, the reduced shedding of E. coli O26:K60 upon challenge with E. coli O157:H7 may be by displacement from the lower regions of the GIT. This assumes greater avidity of E. coli O157:H7 for the distal GIT compared with the resident E. coli O26:K60. It is understood that and intimins are associated with different tropisms in the GIT, proximal and distal respectively (Phillips et al., 1998). Furthermore, it has been suggested that transient high numbers of AEEC, E. coli O26:K60 in this case, may be luminal and not intimately adhered (Moxley, 2004) and possibly more susceptible to displacement.

The long term shedding of E. coli O157:H7 was similar for animals dosed with E. coli O157:H7 alone or from animals pre-dosed with E. coli O26:K60. Therefore, it must be concluded E. coli O157:H7 colonised sufficiently well to establish and persist and that there were more than enough sites for both organisms to colonise irrespective of the presence of E. coli O26:K60. Also, it is possible that the presence of E. coli O157:H7 induced host innate and/or immune responses that contributed to the elimination of E. coli O26:K60.

Data from various surveillance studies (Aktan et al., 2004; Hussein & Sakuma, 2005; Jenkins et al., 2002) indicated that O26 organisms are significantly more prevalent in ruminants than O157 and the objective of this study was to test whether colonisation of lambs with O26:K60 bacteria reduced colonisation by E. coli O157:H7 when challenged. This study shows that pre-inoculation with O26:K60 had no effect on O157:H7 shedding but conversely the shedding of O26:K60 bacteria were significantly reduced in O157:H7 challenged animals. This finding may indicate that O157:H7 organisms may be more fit than other AEEC with regard to colonisation of ruminants.

OBJECTIVE O3

The primary aim of this section of the project was to ascertain the immune responses to O157 and this was undertaken by Prof David Smith and his team at the Moredun Institute. Early studies quickly indicated that circulating antibody responses to LPS, Flagellin (H7) and shiga-like toxns were at best very weak and highly variable. It was decided not undertake further study on this aspect of the immune responses as it seemed highly likely that no useful diagnostic would be forthcoming. However, it was considered pertinent to understand how O157 may be subverting immune responses and to do this, it was important to establish the nature of the interaction between O157 and the recto-anal junction which is a site of preferred colonisation in cattle and is densely populated with lymphoid (immune response) cells. Thus the focus of the research at Moredun was upon this interaction.

As a follow on to the apparent O157-recto-anal junction tropism in cattle, it was important to establish whether a similar tropism was observed in sheep. This study was undertaken at VLA (Weybridge)

SID 5 (Rev. 3/06) Page 26 of 47

Interaction of E. coli O157:H7 with bovine terminal rectum: role of the flagellum.

Cattle are the principal reservoir of E. coli O157:H7 and most of the disease outbreaks have been associated with the consumption of meat or dairy products as well as water, fruit and vegetables contaminated with cattle faeces or manure (Borczyk et al., 1987; Hussein, 2007; Locking et al., 2001; O'Brien et al., 2001; Orskov et al., 1987). Recent surveys of cattle presented for slaughter in the United States and UK have revealed faecal prevalence rates of E. coli O157:H7 ranging from approx. 5% to approx. 30% (Elder et al., 2000; Paiba et al., 2002) indicating that there is high risk for contamination of the food chain and environment. Recent studies have shown that E. coli O157:H7 colonisation of the bovine gastro-intestinal tract is primarily associated with the last few centimetres of the terminal rectal mucosa (Cobbold et al., 2007; Low et al., 2005; Naylor et al., 2003; Rice et al., 2003). Expression of specific bacterial adhesin(s) or cell receptor(s) may account for this tropism. A knowledge of such factors could help in designing effective control strategies.

E. coli O157:H7 possesses an array of adhesins that could play a role in intestinal colonization. Although a substantial amount of data has been generated in recent years regarding the interaction of E. coli O157:H7 with host cells so far type III secretory proteins are the primary O157:H7 virulence determinants demonstrated to play a role in intestinal colonization of cattle and sheep in vivo (see objective 01). The role of many T3SS-dependent factors is likely to be limited to intimate adherence and effacement but those factors conferring initial interaction of EHEC with intestinal epithelium remain to be clearly defined.

A role in adherence to epithelial cells has also been described in certain bacteria for flagella, primarily a motility organelle. For example, Clostridium difficile (Tasteyre et al., 2001), Burkholderia pseudomallei (Inglis et al., 2003), Aeromonas spp. (Kirov et al., 2004), and Listeria monocytogenes (Dons et al., 2004). Recently, evidence has been presented to support a role for E. coli flagella in epithelial colonisation not merely via motility/chemotaxis but directly as an adhesin. Specifically, flagella of Enteropathogenic E. coli (EPEC) of several serotypes (including O127:H6 (E2348/69), O119:H6, O128:H2, and O127:H40), were directly implicated in promotion of adherence and formation of microcolonies on HeLa or HEp-2 cells (Giron et al., 2002). Of note, these H-types are present on various E. coli pathotypes including EHEC (http://www.lugo.usc.es/ecoli/). Together, these findings indicate that the contribution of flagella as adherence organelles cannot be overlooked.

In this investigation, we focused specifically on E. coli O157:H7, the major EHEC serotype in many countries, to determine the extent to which H7 flagella play a role in adherence to terminal rectum epithelium of cattle, the site for which this bacterium shows preferential tropism. We show that H7 flagella directly confer adherence to terminal rectum epithelial cells not merely through motility/chemotaxis but as a primary adherence organelle. Deletion of fliCH7 from E. coli O157 significantly reduced bacterial adherence to rectal epithelium as did incubation of the parental strain with H7-specific antibodies. Since H7 flagella were down-regulated following contact with epithelium as microcolonies formed and conversely T3SS up-regulated, this work shows that these flagella contribute to the initial colonisation process of the bovine host.

Materials and Methods

Bacterial Strains, Flagella preparation and Antibodies. EHEC strains: MCI24 (Stx-negative E. coli O157:H7 strain NCTC 12900), MCI25 (fliC- isogenic mutant derived from strain NCTC 12900; (Best et al., 2005); MCI10 (Stx-negative E. coli O157:H7) and MCI66 (Stx2-positive E. coli O157:H7) (Ostroff et al., 1990), ZAP 116 (O26:H11) (Whittam et al., 1993) and ZAP 244 (O113:H21) (Luck et al., 2006). Flagella from MCI24 (O157:H7), ZAP116 (O26:H11) and ZAP244 (O113:H21) were isolated as described (DePamphilis and Adler, 1971). Antibodies used in this study were Rabbit anti-H7, -H11, -H21, -O157 polyclonal antibody (Mast Diagnostics), Anti-rabbit IgG FITC/ TRITC-conjugated antibodies (Sigma), 10-nm Gold-labelled anti-rabbit IgG (British Biocell International), Horseradish peroxidase-conjugated goat anti-rabbit IgG (Dako A/S, Denmark).

Interaction with primary bovine rectal epithelial cells and In Vitro Organ Culture (IVOC). Bovine rectal primary epithelial cells cultured from terminal rectum (details in supporting Text) were used for the adherence, immunofluorescent and ultrastructural studies. The bacterial adherence assays were done as described. (Dziva et al., 2007) The adherence inhibition experiments were done as described previously (Giron et al., 2002). For IVOC, tissue specimens obtained from adult cattle at a local abattoir were processed and prepared for ex vivo binding studies as described previously (Baehler and Moxley, 2000). Full details are provided in the Supplementary Materials and Methods.

SID 5 (Rev. 3/06) Page 27 of 47

Fluorescence microscopy, In-Cell Western Assays and Electron Microscopy. To quantify adherent bacteria and in situ expression of flagella during course of E. coli O157:H7 (MCI10, MCI24) interaction to bovine rectal primary epithelial cells ‘In-Cell Western Assays’ were carried out modified from the method of (Arredondo et al., 2005). Immunofluoresence microscopy was used to detect adherent E. coli, to monitor expression of flagella during adherence of E. coli O157 strains and to assess binding of isolated flagella to bovine rectal primary epithelial cells and was carried out essentially as described by (Giron et al., 2002). The specimens were visualised with a Leica DMLB epifluoresence microscope and images were analysed using different microscopy software. Scanning electron microscopy using the Hitachi 4700 Field Emission Scanning Electron Microscope was as described previously (Mahajan et al., 2005). Immuno-gold staining was performed using anti-rabbit IgG conjugated with 10-nm gold particles (British Biocell International) as described (Giron et al., 2002). Details of these Methods are provided in the Supporting text.

Statistical analysis. Count data were analysed using generalised linear mixed models, (Brown et al., 1999), modelling the response as a Poisson variable, using a logarithmic link function and fitting the log of assay dilution as an offset. Experimental replicate and other experimental blocking factors were fitted as random effects where appropriate, and other terms as fixed effects. Where possible, residual variability was modelled as a random effect, otherwise a dispersion parameter was estimated. The statistical significance of fixed effects were evaluated using approximate F-tests (Kenward and Roger, 1997). The in-cell Western Assays were analysed as continuous variables using linear mixed models (Brown et al., 1999), fitting adherence as an explanatory variable, expression as a normally distributed response variable, and time and strain as factors. Where multiple comparisons were made between mean values from a single analysis, the false discovery rate was controlled using the method described by (Benjamini and Hotchberg, 1995).

RESULTS

E. coli O157 lacking flagella exhibit diminished adherence to bovine rectal epithelium. Following demonstration of terminal rectal mucosa as a principal site of E. coli O157:H7 colonisation in

cattle by us and others (Cobbold et al., 2007; Low et al., 2005; Naylor et al., 2003; Rice et al., 2003) we developed a primary epithelial cell culture in vitro system to investigate factors that contribute to the colonisation of terminal rectal epithelium of cattle. To examine the possible role of H7 as an adhesin, the adherence of an E. coli O157:H7 flagellate strain (NCTC 12900, termed MCI24), and its fliC mutant (MCI25) (Best et al., 2005) to bovine rectal primary epithelial cells was compared. The average growth rate of wild type and mutant strains were indistinguishable, phenotypic characterisation of these strains for the expression of flagella was confirmed via motility assay, microscopy and immunoblotting (Best et al., 2005).

At 3h the parental strain MCI24 showed localized adherence, with abundant expression of flagella and substantial microcolony formation whilst the aflagellate MCI25 strain adhered sparsely and expressed no flagella. Quantification of colony forming units (cfu) revealed that wild type MCI24 adhered at statistically significant higher mean levels than the fliC mutant MCI25 at both1h and 3h post infection (p<0.001).

To determine whether this lower initial adherence of the fliC-mutant was due to loss of motility, binding assays were carried out in which bacterial cells were centrifuged onto bovine rectal primary epithelial cells. Following slow speed centrifugation coupled with a short incubation of 15 minutes, adherent bacteria were enumerated. Compared to adherence without centrifugation, the mild centrifugation significantly enhanced mean binding of both the flagellate parent strain MCI24 and the fliC mutant MCI25 strain (p<0.001) but the mean adherence for MCI24 remained significantly higher than that for the fliC mutant (MCI25; p<0.001) indicative that H7 flagellin actively contributes to adherence.

To further confirm the role of flagella in adherence of E. coli O157:H7 to bovine gut, ex-vivo binding studies were carried out using tissue explants from the terminal rectum mucosa. On explants the parent flagellate strain MCI24 formed large and compact microcolonies in comparison to the fliC mutant strain (MCI25) which exhibited sparse adherence and only occasional microcolonies. For strain MCI24 (flagellate), flagella expression was apparent for some of the bacteria present as single cells but not within microcolonies.

Flagella antiserum inhibits E. coli O157:H7 binding to Bovine rectal primary epithelial cell.

To confirm the role of H7 in adherence, binding inhibition assays were carried out in the presence of anti-H7 antibody. MCI24 and its fliC mutant MCI25 were treated with anti-H7 polyclonal antibody (1:10 dilution; Mast Diagnostics) for 30 min prior to infection of cells, and after 1h incubation with cells the adherent bacteria were enumerated. Rabbit H7-antiserum statistically significantly inhibited the mean binding of flagellate parental strain MCI24 (p<0.001) but had no effect on the fliC mutant MCI25 (p=0.29). In the presence of anti-H7 bovine antisera the mean binding of both strains was statistically significantly reduced, but the reduction in mean binding for the flagellate strain (MCI24) was statistically significantly greater than that for the aflagellate mutant (MCI25) (p=0.02).

Isolated flagella possess adhesive properties and inhibit bacterial binding. On the basis of the above data, it was postulated that the flagella of E. coli O157 may possess adhesive

properties per se. To examine this, purified flagella from EHEC serotypes O157:H7, O26:H11 and O113:H21

SID 5 (Rev. 3/06) Page 28 of 47

were incubated with bovine rectal primary epithelial cells for 3h and bound flagella were detected by immunofluorescence (IF) assay using type specific anti-H antibodies after thorough washing to remove loosely associated flagella. Of the three flagella types examined, only the H7 flagella bound to the bovine rectal epithelial cells at detectable levels. Neither H11 nor H21 flagella could be observed in association with rectal epithelium demonstrating that, of these 3 flagella types, only H7 flagella directly adheres to rectal epithelium.

To further confirm that H7 acts as an adhesin, and that the inhibition exhibited by anti-H7 reagents is not due to occlusion of other adhesins, the capacity of purified H7 flagella to inhibit E. coli O157:H7 binding to rectal primary epithelial cells was determined. After pre-treatment of cells for 30 min with purified flagella, adherence of E. coli O157:H7 was decreased. Although there was much variability associated with the different flagella doses in different experimental replicates, and no apparent dose-dependent pattern, there was a statistically significant decrease in mean adherence observed at a concentration of 0.25µg/ml when compared with mock flagella (prepared from MCI25) -treated control cells (p=0.007).

E. coli O157:H7 express flagella during attachment to rectal epithelium. Upon adherence to bovine rectal primary epithelial cells E. coli O157:H7 strain (MCI66) abundantly

expressed flagella. In these immunofluorescent micrographs flagella can be seen as filamentous structures on the cell surface extending from bacterial poles and forming an extensive network between adjacent bacteria. Flagella appeared to mediate direct interaction with the epithelial cells, forming contact points directly at the apical surface of epithelial cells as observed in co-localization (dual labelling) studies. Some of the bacterial cells were observed apparently using these flagellar appendages to anchor to epithelial cells therefore presumably playing a role during early bacterial-epithelial interaction(s).

Interaction of flagella with infected bovine rectal epithelial cells was further examined by scanning electron microscopy (SEM). The long filament resembling flagella can be seen intercalated within the micro-villi. The identity of these structures was confirmed to be H7 flagella by immunogold labelling and high resolution field Scanning Electron Microscopy (SEM) using anti-H7 antibodies and anti-rabbit IgG conjugated 10nm gold particles. At the contact points with the epithelial cell the flagellar surface was not apparently accessible for immuno-gold labelling, and were evident only as unlabelled intercepts on the otherwise uniformly labelled flagella. Expression of flagella in situ was then investigated to determine whether flagella were expressed during interaction of E. coli O157:H7 with the bovine gut. Terminal rectal tissue from a calf colonized with E. coli O157:H7 following experimental infection was processed for electron microscopy. Individual bacteria expressing flagella-like appendages at the poles were visible adhering to the rectal epithelium. Together these fluorescent and electron microscopic observations provide support to the postulate that H7 flagella act as an adherence organelle for bovine rectal epithelial cells (Fig 9.).

A B

C D

SID 5 (Rev. 3/06) Page 29 of 47

Fig 9. Confocal micrographs of E. coli O157:H7, adhering to bovine rectal primary epithelial cells (A and D), with H7 flagella specific staining illustrating H7 flagella as an adhesion (B and C, blue and red staining, respectively).

Temporal expression of flagella during E. coli O157:H7 adherenceInitial observations at 3 h post-infection showed flagella expression by individual or small clusters of E. coli

O157:H7 but not larger compact groups or microcolonies on adherence to bovine rectal primary epithelial cells. Therefore, we reasoned that expression of flagella might vary during stages of infection. Expression of flagella by E. coli O157:H7 strains MCI10, MCI24 and MCI66 was examined during the course of adherence to bovine rectal primary epithelial cells qualitatively and quantitatively using IF staining and In-Cell Western Assays, respectively. After 1h of infection the majority of individual adherent bacteria expressed flagella; by 3h of infection bacteria formed visible micro-colonies and at this stage only few of the adherent bacteria in microcolonies expressed flagella. By 8h post-infection adherent bacteria formed typical A/E lesions characterised by subjacent actin-pedestals and absence of flagella, whereas attached bacteria not associated with actin-pedestals, remained flagellated.

In-Cell Western Assays were used to quantify the total bacterial adherence vis a vis flagellar expression during the course of E. coli O157:H7 interaction with bovine rectal epithelial cells. The relationship between adherence and expression varied at different times. For both strains (MCI10 and MCI24) the mean level of adherence significantly increased (p<0.001) by 5 h post infection whilst mean H7-expression significantly decreased (p<0.001), Analysis of the relationships between adherence and H7 expression showed that at time 1h there was a statistically significant relationship between mean levels of bacterial adherence and expression of flagella (p<0.001). At times 3 h and 5h, the relationship between H7-expression and adherence was statistically significantly weaker (p<0.001), and there was no evidence of any difference in the relationship at the two times (p=0.54). At 5 h post-infection, the overall mean level of H7 expression was depressed. There was no statistically significant evidence of strain specific effects on the relationship between H7-expression and adherence.

DISCUSSIONWe and others (Cobbold et al., 2007; Low et al., 2005; Naylor et al., 2003; Rice et al., 2003) have previously

reported the terminal rectum as a principal site of E. coli O157:H7 colonisation in cattle. To characterise factors contributing to such a tropism we developed a primary cell culture system for terminal rectum epithelium cells of cattle and have used it as an in vitro model to assess adherence (Dziva et al., 2007; Low et al., 2006). Initial immunofluorescent microscopic investigations of E. coli O157:H7 adherence to rectal primary epithelium cells in culture, revealed abundant expression of H7 flagella on contact with epithelium at 1 hour of E. coli O157:H7 infection. Notably, there appeared to be patent contact points between flagella and epithelial cells, an observation supported by SEM which revealed the presence of flagella-like structures intercalating with the microvilli (data not shown). Although flagella are primarily motility organelles, a role in adherence to epithelial cells has also been described for certain bacterial flagella and indeed for H6 flagella of EPEC (Giron et al., 2002). Although the latter publication ruled out a role for H7 flagella in attachment of E. coli O157:H7, the model systems employed in that work were human epithelial HeLa and Hep-2 cells (neither of which are intestinal in origin), and hence may not apply to intestinal cells of bovine origin. Therefore, as a consequence of the observed intercalation of H7 flagella with microvilli of bovine enterocytes, we set out to test the postulate that H7 flagella may act as an adherence factor in attachment to bovine intestinal epithelial cells.

Firstly, adherence of parental and flagellin (fliC) mutant E. coli O157:H7 was compared using bovine rectum epithelial cell cultures. The fliC mutant strain showed significantly reduced adherence (p<0.001) indicating that flagella do play a role in the binding of E. coli O157:H7 to bovine rectal epithelial cells. The involvement of H7 in epithelial interactions was also indicated through pre-treatment of bacteria with rabbit anti-H7 polyclonal antiserum which significantly reduced the mean binding of E. coli O157:H7 but not of its fliC mutant strain. Notably, bovine sera containing antibodies to H7 flagella also significantly inhibited average adherence of E. coli O157:H7. Antibodies purified from mucosal extracts from cattle immunized with purified H7 (McNeilly et al., 2007) also reduced mean

SID 5 (Rev. 3/06) Page 30 of 47

bacterial attachment to primary bovine rectal epithelial cells thus further corroborating these findings. Flagella are potent immunogens and antibodies to H7 have also been detected in mucosal secretions from animals colonised with E. coli O157:H7. To determine whether flagella filaments per se were required for adherence of E. coli O157:H7 to epithelial cells, the motility defect of the fliC mutant was overcome through gentle centrifugation of bacteria onto the epithelial monolayer. As expected this procedure increased average adherence of both parent and fliC mutant strains although average adherence of the latter remained significantly lower than that of the parent strain. Thus, presence of intact flagella on the bacterial surface promotes adherence and is a further indication that flagella directly contribute to host cell attachment.

In assessing the capacity of purified flagella of different EHEC O157:H7, O26:H11 and O103:H21 serotypes to adhere, only isolated H7 flagella bound to bovine rectal primary epithelial cells at detectable levels confirming that these flagella per se can confer attachment presumably via a specific ligand-receptor interaction. One of the hallmark features of ligand-receptor interaction is that purified ligand can competitively inhibit binding to its cognate receptor in dose-dependent manner. Indeed, purified H7 flagella tended to inhibit E. coli O157:H7 binding to bovine rectal primary epithelial cells, showing a statistically significant decrease in average adherence at a flagella concentration of 0.25µg/ml. These data support our contention that H7 flagella perform an initiating role in attachment to rectal epithelium.

The molecular basis of H7 flagella binding remains elusive however it is conceivable that H7 flagella possess domains that confer attachment. We propose that TLR5 is unlikely to be the initial receptor for flagella since the flagellin domain that binds TLR5 forms part of the core of the flagellar filament (Eaves-Pyles et al., 2001) and thus is not exposed in the native structure and TLR5 is usually located on the basolateral surface of enterocytes (Gewirtz et al., 2001). Other epithelial receptors or co-receptors for flagella have been proposed, including TLR2 (Adamo et al., 2004), gangliosides (Feldman et al., 1998) and mucin (Lillehoj et al., 2002). Recently it has been shown that H7 flagella bind mucin (Erdem et al., 2007) confirming the key role that these flagella play in adherence.

In assessing E. coli O157:H7 attachment to intestinal epithelial cells, it was evident that flagella were expressed in a time-dependent manner: at early time points post-infection flagella were expressed extensively by attaching bacterial cells whereas by later time points flagella expression was apparent only by isolated bacterial cells. Notably, flagella expression diminished as bacterial microcolonies were formed and as aggregation of actin became apparent beneath attached bacteria. This observation suggests that H7 flagella are important in early stages of interaction of E. coli O157:H7 with rectal epithelium whilst LEE-dependent factors assume greater significance following initial attachment, taking on the role as major adhesive factors and modifying host cellular function. An inverse relationship between expression of flagella and LEE-endoded determinants is consistent with in vitro investigations of expression of these loci (Iyoda et al., 2006). Another potentially significant consequence of flagella down-regulation following attachment is evasion of host inflammatory responses. Flagella, including H7 are recognised as potent inflammatory ligands, and indeed they can evoke inflammatory responses through engagement of each of the receptors mentioned above (Berin et al., 2002; Zhou et al., 2003). Thus, down-regulation of this potent pro-inflammatory signal after initiating attachment, and while other contact-dependent factors assume their roles in adherence, represents a further stratagem to evade activation of immune responses.

Specifically, we have demonstrated that H7 flagella facilitate attachment of E. coli O157:H7 to bovine intestinal (rectal) epithelial cells. The role of H7 flagella is two-fold, firstly through its role as the organelle of motility and secondly by conferring adherence to bovine rectal epithelial cells in culture. Thus, H7 flagella may be employed by E. coli O157:H7 to “browse” the epithelial surface, thence tethering the organism to accessible receptor(s). Having performed this role in initiating contact and establishing a nidus of infection, other adhesins can engage with their receptors and the contribution of flagella to this process is superseded. Colonisation of epithelium then progresses into a secondary phase, corresponding to decreased expression of flagella and increased expression of LEE and associated factors. This regulatory and functional switch promotes persistence resulting from intimate attachment and modulation of inflammatory responses. Although H7 flagella are not the sole factors responsible for conferring epithlelial tropism, their involvement offers potential as a further candidate to target in intervention approaches and experiments are under way to assess H7 in immunological strategies to reduce E. coli O157 in cattle. Overall, this study re-emphasises the multifactorial nature of infection caused by E. coli O157:H7 in the reservoir host and further demonstrates the complex, multiphasic basis of bacterium-epithelium interaction during colonisation.

The Recto-Anal Junction is not the primary site of colonisation in the six week old lamb model

Various studies in which cattle have been inoculated orally with E. coli O157:H7 have shown that as a general principal, pre-weaned calves shed greater numbers of E. coli O157:H7 for a longer time period (up to 42 days) than adult cattle, whereas, 4-month-old calves and year-old steers can stop shedding E. coli O157:H7 as early as day 7 and 9 P.I., respectively (Cray & Moon, 1995; Buchko et al., 2000; Ohya et al., 2000). In similar studies performed in sheep, E. coli O157:H7 was shed in the faeces for up to 48 days for young lambs and 60 days in young adult sheep (Cornick et al., 2000; Woodward et al., 2003). For both bovine and ovine models, recovery of E. coli O157:H7 by faecal isolation is variable in terms of numbers of organisms shed, duration of shedding and numbers of animals positive in any test cohort.

SID 5 (Rev. 3/06) Page 31 of 47

The early experimental findings arising from oral inoculation studies in cattle appear to be at odds with the rapid spread and rise of E. coli O157:H7 as a major food-borne zoonosis. Recent data has demonstrated that the combined effects of higher numbers of E. coli O157:H7 in the faeces and more persistent excretion in particular infected cattle, “super-shedders”, results in elevated transmission of E. coli O157:H7 (Chase-Topping et al., 2007; Matthews et al., 2006a, 2006b).

Grauke et al., (2002) suggested that the lower GIT of orally inoculated cattle, including the colon, was one location for E. coli O157:H7 colonisation associated with persistence. More recently, Naylor et al., (2003) showed the lymphoid follicle-dense recto-anal mucosa at the terminal rectum, as the primary site for E. coli O157:H7 localisation in young calves following oral inoculation. This hypothesis was further assessed by Sheng et al., (2004) who demonstrated that rectal swab administration of E. coli O157:H7 resulted in persistent long-term colonisation, thus strongly supporting the terminal rectum as the preferred site for E. coli O157:H7 colonisation in cattle. A recent feedlot E. coli O157:H7 infection study generated supportive data suggesting that “super-shedder” status is linked to greater E. coli O157:H7 colonisation of the terminal rectum (Cobbold et al., 2007). Furthermore, necropsy of cattle that were defined as long-term colonised by E. coli O157:H7, revealed that of all gastrointestinal (GIT) sites examined, only the terminal rectum was culture positive for E. coli O157:H7 (Lim et al., 2007). Collectively these data are compelling that in cattle, E. coli O157:H7 colonises the lower GIT effectively, but that long term and high level shedding is dependent upon colonisation at the terminal rectum.

Significantly less is known about the primary site of colonisation for E. coli O157:H7 in sheep, should a single site exist. The interpretation of data from the various oral inoculation studies that have been performed (Grauke et al., 2002; La Ragione et al., 2006; Wales et al., 2001a: Woodward et al., 2003) have not identified the terminal rectum as the primary site of colonisation. Indeed, small sparse attaching and effacing (A/E) lesions induced by E. coli O157:H7 have been identified in the colon more readily than at other sites (Woodward et al., 2003). Furthermore, in addition to the generally sporadic nature of shedding and persistence, some animals were faecal culture negative for E. coli O157:H7 from as early as day 3 after oral inoculation. It is possible that E. coli O157:H7 is less well adapted to the colonisation of sheep than cattle, because the ovine terminal rectum is less susceptible to E. coli O157:H7 colonisation. To test this, in vitro and in vivo studies were performed to assess whether the terminal rectum of sheep, as in cattle, may be the primary localisation site associated with persistence.

Materials and Methods

Bacterial strain, inocula and reagents.

E. coli O157:H7 strain NCTC12900nalr has been described previously (La Ragione et al., 2005) It does not possess stx1 and stx2, but is does induce attaching and effacing (A/E) rearrangements on HEp-2, and porcine cell lines and in vivo ovine models (La Ragione et al. 2006; Wales et al. 2001a; Best et al. 2006). NCTC12900nalr was stored in heart infusion broth (HIB) supplemented with 30% glycerol (v/v) at –80oC and was streaked to single colonies on 5% sheep blood agar when required. For bovine and ovine IVOC assays and ovine in vivo studies, bacterial inocula was prepared by growth in 20ml or 100ml LB broths for 16 h, at 37 oC with shaking (225rpm). For IVOC assays, 25l of overnight culture, corresponding to approximately 2 x 107 CFU, was added to each tissue explant. For in vivo studies, overnight cultures, corresponding to approximately 1 x 109 CFU/ml were administered as described below.

Bovine and ovine in vitro organ culture (IVOC). The bovine and ovine intestinal IVOC system was adapted from the bovine and porcine intestinal IVOC system described previously by Girard et al. (2007). Briefly, the terminal rectum and the last 2cm of the rectum immediately before the terminal rectum, termed rectum-1, were excised from two conventionally reared 6-week-old lambs and 12-week-old calves and placed in cold complete RPMI medium without D-mannose. Approximately 12mm2 explants were cut and placed mucosal side up onto biopsy foam pads, in sterile 24-well Nunclon Delta surface tissue culture plates, with 4 explants on each biopsy foam pad. Complete RPMI 1640 medium (10% Foetal Bovine serum, 0.25% lactalbumin hydrolsylate, 0.2g/ml hydrocortisone, 0.1g/ml insulin, 75mM -mercaptoethanol, 2mM L-glutamine, 2mM L-aspartate) was added to each well (1ml). Explants were then placed on a see-saw rocker and maintained at 37oC (5% CO2) for 8 hours. Complete RPMI medium was replaced after the first 2 hours post incubation and hourly thereafter.

Animals for in vivo studies, challenge methods and isolation of E. coli O157:H7. Two ovine in vivo studies were performed. For experiment 1, twelve, 5-week-old cross-bred conventionally reared lambs were randomly separated into two equal groups of six animals each (A and B) and housed indoors with food and water ad libitum. The animals were allowed to acclimatise to their environment for one week and prior to challenge, faecal samples were taken from each lamb and confirmed negative for E. coli O157:H7 by immunomagnetic separation (IMS) and culture. At 6-weeks of age, all lambs in group A were challenged with 1 x 10 10 CFU. of E. coli O157:H7 isolate NCTC12900nalr. Inocula were delivered in a 10ml volume using a worming gun (Novartis Animal Health) ensuring that the whole inoculum was delivered directly to the pharynx. Group B lambs acted as negative control for analysis of immune cell populations in the epithelial cells of the terminal rectum. Rectal faecal samples were taken by digital insertion from lambs in group A only on days 3, 5 and 7 after inoculation.

SID 5 (Rev. 3/06) Page 32 of 47

For experiment 2, eighteen cross-bred conventionally reared 5-week-old lambs were randomly divided into two groups, group C comprising of twelve animals and group D comprising of six animals, respectively. Lambs were housed indoors and were provided with food and water ad libitum. Prior to challenge, faecal samples were taken from each lamb and confirmed negative for E. coli O157:H7 by immunomagnetic separation (IMS). At 6-weeks of age all lambs in group C were dosed per rectum following manual faecal evacuation with 1 x 1010 CFU. of E. coli O157:H7. Inocula were delivered in a 10ml volume using a syringe and flexible catheter tube (Agrihealth, lamb reviver). Lambs were restrained for 2 minutes to ensure the inocula remained in the rectum. All group D lambs were dosed orally with 1 x 1010 CFU of E. coli O157:H7. Inocula were delivered in a 10ml volume using a worming gun (Novartis Animal Health) ensuring that the whole inoculum was delivered directly to the pharynx. Faecal samples were taken per rectum by digital insertion. All procedures were conducted under the jurisdiction of Home Office licence 70/6250 granted under the Animals (Scientific Procedures) Act (1986).

Necropsy of E. coli O157:H7 infected lambs. All lambs from experiment 1 and from group C of experiment 2 check group labelling were euthanased and necropsied as described previously (Wales et al., 2001a, ). Briefly, for experiment 1, two lambs were examined at days 3, 5 and 7 after inoculation, respectively. Tissue samples, either for bacterial enumeration or IHC, were collected from the duodenum, jejunum, ileum, caecum, ascending colon, spiral colon, the last 2 cm of the distal rectum and the terminal rectum. For experiment 2, three lambs from group A were examined at day 1 and day 3 after inoculation and the remaining six animals were examined on day 14 after inoculation. Tissue samples for bacterial enumeration, IHC or confocal analysis, were collected from the duodenum, jejunum, ileum, caecum, ascending colon, spiral colon, terminal rectum and from six sites excised at approximately 2cm apart measured from the terminal rectum, along the rectum toward the distal colon, (La Ragione et al., 2006).

Bacterial enumeration. For experiments 1 and 2, all E. coli O157 isolates were detected using methods described previously (La Ragione et al. 2006). Briefly, lamb faeces (1g) and necropsy tissue samples (1g) were separately re-suspended in 9ml of BPW, mashed with sterile forceps and vortexed. Ten-fold-serial dilutions were plated directly onto sorbitol MacConkey agar (SMAC) plates supplemented with nalidixic acid (15g/ml). If no direct counts were observed after overnight incubation of the plates at 37oC, the BPW homogenates were enriched by incubation at 37oC for 6h, and then plated onto SMAC supplemented with nalidixic acid (15g/ml). The serogroup of bacteria recovered was verified by E. coli O157-specific latex agglutination (Oxoid, Basingstoke, UK).

Light microscopy. At necropsy, tissues were placed immediately into 10% neutral buffered formalin at room temperature and left to fix for at least 24 hours. Trimmed tissues were processed routinely to paraffin wax. Sections were cut at 4μm and stained with haematoxylin and eosin (HE). Immunohistochemistry (IHC). Methods were as described previously (La Ragione et al., 2005). Briefly, tissue blocks were fixed in 10% neutral buffered formalin, processed to wax and sectioned at 4m. Sections were rehydrated prior to assembly into the Shandon Sequenza staining system. Slides were washed with 2x sodium chloride Tris-buffered saline (0.005M Tris-buffered saline (TBS), pH 7.6, 1.7% NaCl) before incubation at room temperature with normal goat serum (Vector Laboratories). E. coli O157:H7 specific polyclonal antibody (VLA, Weybridge), was then applied (1:1000 and 1:5000 diluted in 2x NaCl TBS supplemented with 5% normal rabbit serum). E. coli antigens were visualised following incubation with biotinylated goat anti-rabbit IgG. Sections were counter stained in Meyer’s haematoxylin. Confocal microscopy. After de-waxing and rehydration, tissues were placed in PBS prior to preparation for examination by confocal microscopy. Briefly, sections were permeabilised in PBS containing 0.1% Triton X-100 followed by detection of E. coli using a fluorescein isothiocyanate (FITC)-labelled affinity purified antibody to E. coli O157:H7 produced in goats. The sections were then washed thoroughly using PBS and mounted in Vectashield containing DAPI (Vector Laboratories). Images of the FITC-labelled E. coli were obtained by confocal laser scanning microscopy using a Leica TCS SP2 AOBS confocal system attached to a Leica DM IRE2 microscope equipped with ArKr laser excitation (488nm) and a diode laser (405nm). An oil-immersion objective lens (63x, N.A. 1.32) was used, and imaging parameters were selected to optimise resolution.

Statistical analysis. Analysis of variance (Anova) was calculated on the log2(counts) for oral and rectal shedding data from in vivo experiment 2. Bacterial enumeration shedding data was analysed using 95% confidence intervals to calculate the difference between counts for each day and a p-value was generated. 

RESULTS

E. coli O157:H7 NCTC12900nalr adherence patterns on ovine rectum-1 and terminal rectum explants.The ability of E. coli O157:H7 NCTC12900nalr to associate intimately with ovine and bovine rectum-1 and

terminal rectum explants was assessed using the IVOC method of Girard et al. (2007). Bovine rectum-1 and terminal rectum explants were used as a control as previous reports have demonstrated that E. coli O157:H7 intimately associates with these tissue sites (Naylor et al., 2003; Girard et al., 2007).

SID 5 (Rev. 3/06) Page 33 of 47

Fig 10, Enumeration of E. coli O157 present in specific areas of the ovine gut and or associated with lymphoid tissue following experimental inoculation.

After staining both ovine explants with anti-O157-FITC and analysis by confocal microscopy, diffuse attachment was observed on all explants as were small clusters (<10 bacteria) of E. coli O157. Larger, more densely packed micro-colonies of NCTC12900nalr were associated with the terminal rectum explants, but not rectum-1 explants and similar findings were made for the bovine control tissues.

Localistion of E. coli O157:H7 NCTC12900nalr in the ovine gastrointestinal tract after oral inoculation of young lambs.

Lymphoid and non-lymphoid GIT tissues from two, randomly selected orally inoculated 6-week-old lambs were taken at post-mortem on days 3, 5 and 7 after oral inoculation. Samples were analysed by bacterial enumeration and tissues were also stained for CD3 (T-cell), CD79a (B-cell) and CD68 (monocyte macrophage) cells. In terms of association with the ovine GIT, all lambs were positive for E. coli O157:H7 NCTC12900nalr, but not all lambs were positive for every tissue cultured. Furthermore, there was variation in the numbers of E. coli O157:H7 recovered from each lamb at each time point. However, on days 5 and 9 after inoculation, approximately a 100-fold more E. coli O157:H7 NCTC12900nalr were recovered from the rectum-1 and terminal rectum sites compared with all other GIT sites cultured. Comparison of lymphoid and non-lymphoid tissue was inconclusive with higher numbers recovered from lymphoid tissues in lambs-C and –F and from non-lymphoid tissues in lambs-D and -E. The highest numbers of E. coli O157:H7 NCTC12900nalr were recovered consistently from the faeces of lamb-D.

Analysis of immune-cell infiltration in all lambs from days 5 and 7 after inoculation, revealed similar levels of T-cell, B-cell and monocyte macrophage specific staining in terminal rectum associated lymphoid tissue of lambs infected with E. coli O157:H7 as in unaffected controls (data not shown).

Duration and extent of E. coli O157:H7 NCTC12900nal r faecal shedding from young lambs inoculated orally and rectally. Faecal samples were taken daily over a 2-week period from six orally and six rectally inoculated 6-week-old lambs, and plated onto the appropriate media. Twenty-four hours after rectal and oral inoculation, six and five lambs were positive, respectively, with recoveries at 104-105 and 102-105 CFU/g faeces-1, respectively. Throughout the 2-week period of the study, variable shedding of E. coli O157:H7 NCTC12900nalr was noted with only one significant difference between groups observed at day 1 after inoculation in which recovery of E. coli O157:H7 NCTC12900nalr from animals inoculated rectally was higher (p=0.044). Over the 2-week period of the study, 66/84 and 56/84 faecal samples were culture-positive for E. coli O157:H7 NCTC12900nalr from rectally and orally inoculated lambs, respectively. At the end point of the study (day 14) four rectally inoculated lambs were shedding NCTC12900nalr, whereas only one lamb from the orally inoculated group was shedding NCTC12900nalr.

Recovery of E. coli O157:H7 NCTC12900nalr from tissues of the gastrointestinal tract of rectally inoculated 6-week-old lambs.

Necropsy examinations were performed after rectal inoculation on three lambs on day 1, three on day 3 and six on day 14, respectively. The numbers of E. coli O157:H7 NCTC12900nalr recovered from different sites along the distal GIT are shown below. Recovery of E. coli O157:H7 NCTC12900nalr from the faecal samples taken immediately before necropsy is also shown. Only one lamb was culture-positive for E. coli O157:H7 NCTC12900nalr for all rectum, spiral and ascending colon, and caecum samples, although numbers were low, and that was on day 1. Otherwise, E. coli O157:H7 NCTC12900nalr was not recovered from the colon, caecum

SID 5 (Rev. 3/06) Page 34 of 47

and proximal rectum samples from any of the remaining 11 animals. No lambs were culture positive for E. coli O157:H7 recovery from the duodenum, jejunum and ileum. However, the terminal rectum was positive in 9 of 12 animals examined whereas rectum 1 and rectum 2 samples were positive in 5 of 12 and 4 of 12 animals, respectively.

(A)

(B)

Fig 11. Shedding of E. coli O157 from lambs either orally (A) or rectally (B) inoculated with E. coli O157:H7 at six weeks of age.

Fig 12. Bacterial enumeration of E. coli NCTC12900 recovered from the gastronintestinal tissues of conventionally reared lambs, rectally inoculated with E. coli O157 NCTC12900 at six weeks of age.

Three lambs cultured after necropsy that had more than 1x104 CFU/g tissue-1 of E. coli O157:H7 NCTC12900nalr associated with the terminal rectum, shed NCTC12900nal r at 105 to 106 CFU/g faeces-1. By

SID 5 (Rev. 3/06) Page 35 of 47

contrast, six lambs with less than 1x104 CFU/g tissue-1 of NCTC12900nalr associated with the terminal rectum only shed lower numbers.

Visualisation of E. coli O157:H7 NCTC12900nalr association to the terminal rectum. Tissue sections taken from the GIT were initially fixed in formalin. Taking into consideration the

bacteriological data at necrospy, only tissues from rectum-1 and terminal rectum were selected for staining with anti-O157-FITC for examination by Confocal microscopy. Seven of the 12 animals examined gave tissues that showed diffuse adherence and small clusters (<10 bacteria) of E. coli O157:H7 NCTC12900nalr to rectum-1 and terminal rectum. Of these, four exhibited large, densely packed micro-colonies at the terminal rectum. By day 14, only one of the six rectally inoculated animals was well colonised at the rectum-1 and terminal rectum sites by E. coli O157:H7 NCTC12900nalr.

A B

Fig 13. E. coli O157 specific IHC (brown staining) demonstrating E. coli O157 AE lesions at the rectal anal junction mucosa (a) and association with lymphoid tissue (b) (lymphoid aggregate directly below IHC brown staining).

DISCUSSION

The overall aim of this study was to assess whether the terminal rectum of sheep is a preferred site of colonisation for E. coli O157:H7 as indicated for cattle (Naylor et al., 2003, Sheng et al., 2004; Cobbold et al., 2007; Lim et al., 2007). Our experimental data which used both in vitro and in vivo studies indicates that the terminal rectum may be colonised by E. coli O157:H7.

Using previously described IVOC techniques (Girard et al., 2007) we have demonstrated for the first time that E. coli O157:H7 can adhere and associate intimately, albeit sparsely as small micro-colonies, on the mucosal surface of explants taken from the terminal rectum of 6-week-old lambs. Association with the distal biopsy portion of the rectum (rectum-1) was observed, but the evidence for micro-colonies was equivocal. A similar pattern of E. coli O157:H7 association was observed with the respective bovine control tissues. These data suggest that E. coli O157:H7 has the ability to attach and aggregate as micro-colonies at the ovine terminal rectum. That micro-colonies were only observed at the ovine terminal rectum, but not the rectum-1 mucosa, may suggest a tissue tropism, although the numbers of samples evaluated were limited and care should be taken in drawing firm conclusions especially as the adherence patterns in these assays may be influenced by many factors including the 8 hour incubation period, the fact that the biopsies are not under the same physiological conditions as in vivo, and the epithelial cell cycle and the metabolism of the biopsies may have be disrupted. Indeed, epithelial cell rate

SID 5 (Rev. 3/06) Page 36 of 47

proliferation, which will cease in biopsy, may be linked to the number of E. coli O157:H7 micro-colonies formed as indicated by the cell-proliferation studies of Magnuson et al., (2000) and this may impact negatively on the significance of the IVOC data.

Examination at necropsy of ovines orally inoculated with E. coli O157:H7 (Wales et al., 2001a; La Ragione et al., 2006; Woodward et al., 2003) showed that neonates and six-week-old lambs were readily and consistently colonised but that very small and very sparse AE lesions induced by E. coli O157:H7 were identified at various sites in the distal GIT. The presence of E. coli O157:H7 at the rectum was considered to be as a result of the passage of inoculum rather than colonisation at this site as no lesions were observed in those early studies, possibly because too few tissues were examined. Here, we compared the outcome of oral and rectal inoculation of 6-week-old lambs and there were no differences in faecal shedding of E. coli O157:H7 between the six rectally and six orally inoculated lambs over a fourteen day period except for day one after inoculation, where greater numbers were recovered from the faeces of rectally inoculated lambs. An unsurprising finding given the oral inoculum has to travel through the GIT before being shed. That the rectally inoculated lambs shed E. coli O157:H7 for as long as the orally inoculated lambs indicates that sufficient colonisation must have occurred at the very distal GIT to sustain shedding. Thus the identification of AE lesions at the caecum and colon in previous studies (Wales et al., 2001a; La Ragione et al., 2006; Woodward et al., 2003) may indicate more promiscuous colonisation in sheep and that multiple sites can be colonised to sustain shedding.

More faecal samples were culture positive for E. coli O157:H7 from rectally inoculated lambs than orally inoculated lambs. At the end point of this study (day 14) the faeces of only one orally inoculated lamb were culture positive for E. coli O157:H7 as compared to four rectally inoculated lambs. A lack of statistically significant differences for E. coli O157:H7 shedding between rectally and orally inoculated lambs, may reflect the large variability in numbers of E. coli O157:H7 from faecal samples which has been reported previously for orally inoculated 6-week-old lambs (Cornick et al. 2007; Torres et al., 2006; La Ragione et al., 2006). If the terminal rectum was the site of preferred colonisation, the expectation would be to see significant differences between orally and rectally inoculated groups. After inoculation, only one lamb inoculated rectally established a higher level of shedding of E. coli O157:H7 from day 7 compared to other lambs from both inoculation groups. Given the apparent consistency of high level shedding from rectally inoculated calves (Sheng et al., 2004; Cobbold et al., 2007; Lim et al., 2007), our data may indicate the ovine terminal rectum may not be the preferred site of colonisation or at least not associated with consistent high shedding.

Another factor that may influence intermittent colonisation by E. coli O157:H7 at the ovine terminal rectum is the formation of sheep faecal pellets. The spiral colon, despite its similarity in ovine and bovine species is the site of pellet formation in sheep. This peculiarity is associated with a permanent and localized contractile activity of the ovine spiral colon instead of five minute periods of continuous activity migrating slowly, aborally in cattle (Ruckebusch & Fioramonti, 1980). These mechanical differences in the formation of faeces and the physiological differences between the faeces of ovine and bovine (hard versus soft) may in part explain why colonisation by E. coli O157:H7 at the terminal rectum of infected sheep does not occur as consistently as at the same site in cattle.

In conclusion, the studies conducted here are the first to report that E. coli O157:H7 can form micro-colonies on ovine terminal rectum explants and colonise the terminal rectum of rectally inoculated young lambs. Although, the in vivo data does suggest that the ovine terminal rectum may be colonised by E. coli O157:H7, colonisation is less efficient than in cattle. Furthermore, and in contrast to rectally inoculated cattle (Sheng et al., 2004; Cobbold et al., 2007; Lim et al., 2007) the numbers of E. coli O157:H7 shed in ovine faeces over the time-course of the study were not significantly different to those of orally inoculated lambs. We suggest that a more comprehensive understanding of the factors that promote effective colonisation at the ovine terminal rectum requires further analysis in order to assess the appropriateness of intervention strategies (Mathews et al., 2006a) to prevent transmission from an ovine herd need to be considered.

Development of a surrogate O157 mouse model for immune and other studies

Oral ModelC. rodentium is a natural mouse pathogen that, while causing colonic hyperplasia, shares many virulence

factors with EPEC and EHEC. Following inoculation via the oral route bacteria colonise the colon typically peaking at day 9 before clearance at around day 17. It seemed reasonable to attempt to use the mouse as a surrogate for EHEC infections. The benefits are clearly tangible in terms of cost and numbers of animals that can be used for statistical robustness.

SID 5 (Rev. 3/06) Page 37 of 47

Fig 14. Faecal shedding of two wildtye E. coli O157 isolates (NCTC12900 and 85-170) from ICR mice orally inoculated with either E. coli O157:H7 (NCTC12900) or E. coli O157:H7 (85-170).

ICR:SLC mice, which are susceptible to EHEC infection, were maintained at the VLA. All animal studies were performed in accordance with the Animals (Scientific Procedures) Act 1986 (HO Licence 70/6103) and approved by the local Ethical Review Committee. All animals were housed in individually HEPA filtered cages with sterile bedding and free access to sterilized food and water. Bacterial inocula were prepared by culturing bacteria overnight at 37oC in LB-broth (containing the appropriate antibiotics if required). Cultures were harvested by centrifugation and re-suspended to the appropriate concentration in phosphate buffered saline (PBS). For challenge studies 5 week old mice were dosed with cimetidine prior to challenge with approx by oral gavage with approximately 1 x 107cfu. E. coli O157 in 200 l. The viable count of the inoculum was determined by retrospective serial dilution and plating on LB-agar containing the appropriate antibiotic. Stool samples were recovered aseptically at various time points after inoculation and the number of viable bacteria per gram of stool determined by plating onto LB agar containing the appropriate antibiotics. At selected time points post-infection, mice were killed by cervical dislocation. The colon and caecum were aseptically removed and weighed after the removal of faecal pellets and caecal contents. The organs were then homogenized mechanically in 5 ml of sterile PBS using a Seward 80 stomacher (Seward, London, United Kingdom) and the number of viable bacteria per gram of organ homogenate determined by plating onto LB agar containing the appropriate antibiotics. Independent experiments were performed at least twice using groups of at least 4 mice per strain. In the first experiment, the persistence of O157:H7 was determined by faecal recovery of the organisms and serial dilution plating. An example of the shedding profile is given in the figure above.

Murine ligated ileal gut loop to study intimate host microbe interactions

In order to study the intimate relationship between E. coli O157 and the gut Mucosa a ligated murine gut loop model was developed. This model provided far superior tissue quality than post-mortem tissues and so allowed for detailed ultrastructural analysis of tissues and for specific cell markers to be used on cryo-preserved tissues. The model was developed as follows : Eight week old female SLC:ICR mice were used for these studies.   Feed was withdrawn for eight hours prior to studies.  Mice were placed in an induction chamber (small anaesthetic induction chamber; Vet TECH Solutions Ltd., Cheshire, UK) for anaesthetic induction with isoflurane in oxygen. Anaesthesia was then maintained with isoflurane in oxygen delivered via coaxial face mask (rodent mask delivery system, size 1; Vet TECH solutions Ltd., Cheshire, UK). Analgesia was provided with 5mg/kg butorphanol (Torbugesic; Fort Dodge Animal Health, Southampton, UK) administered subcutaneously. Intra-operative fluid therapy consisted of 1 millilitre per mouse of warmed Hartmann’s solution (Aqupharm number 11; Animal Care Ltd., York, UK) injected subcutaneously. Rectal temperature was monitored continually throughout the procedure (TES 1319 K-Type thermometer with thermocouple probe TP-K02; TES, Ontario, Canada), and warming with (using a hot water bottle), or cooling (with a small hand-held rotary fan) instigated to maintain as near normothermia (37.4 degrees celcius) as possible. Eyes were protected with a bland ophthalmic ointment. A mid-line laparotomy was performed to exteriorise the distal ileum.  Once the ileum was isolated four loops were made (approx. 2cm and incorporating a Peyers patch), each with spacer loops to avoid any cross contamination from one loop to another.  The loops were inoculated with 100 microlites (1x109 cfu) of test bacteria and inoculated loops were reinstated into the abdominal cavity.  The vital signs of the mouse were continuously monitored (temperature, colour and respiration).  Following a 3hr incubation period the loops were harvested under terminal anaesthesia and the mouse euthanased.  The tissues were fixed immediately for EM and histopathology or frozen in OCT using ice-cold isopentane. This model is now being used to study a number of intervention agents for use against a range of zoonotic pathogens.

SID 5 (Rev. 3/06) Page 38 of 47

In preliminary studies, we demonstrated intimate attachment of E. coli O157:H7 to the gut mucosa as shown in the figure below. Thus we have a model that may be suitable for the study of interventions aimed at reducing AE mediated carriage and persistence.

A) B)

Fig 15. Transmission electron microscopy of murine ileal loops illustrating PBS a) and O157 b) inoculated loops. In the O157 infected loop (b) the brush border is effaced and bacteria are intimately adhered to the mucosa, whereas in the PBS control the mucosa remains intact (a).

SID 5 (Rev. 3/06) Page 39 of 47

REFERENCES

Adamo R, Sokol S, Soong G, Gomez MI, Prince A. (2004). Pseudomonas aeruginosa flagella activate airway epithelial cells through asialoGM1 and toll-like receptor 2 as well as toll-like receptor 5.Am J Respir Cell Mol Biol. 30(5):627-34.

Aktan, I., K. A. Sprigings, R. M. La Ragione, L. M. Faulkner, G. A. Paiba, and M. J. Woodward. 2004. Characterisation of attaching-effacing Escherichia coli isolates from animals at slaughter in England and Wales. Vet. Microbiol. 102:43–53.

Aktan, I. La Ragione, R. M. & Woodward, M. J. 2007. Colonisation, persistence and tissue tropism of Escherichia coli O26 in conventionally reared weaned lambs. Appl. & Enviro. Micro. 73, 691-698.

Altmann, K. & Mukkur, T. K. (1983). Passive immunisation of neonatal lambs against infection with enteropathogenic Escherichia coli via colostrum of ewes immunised with crude and purified K99 pili. Res Vet Sci 35, 234–239.

Arredondo J, Chernyavsky AI, Marubio LM, Beaudet AL, Jolkovsky DL, Pinkerton KE, Grando SA. (2005).Receptor-mediated tobacco toxicity: regulation of gene expression through alpha3beta2 nicotinic receptor in oral epithelial cells. Am J Pathol. 166(2):597-613.

Baehler AA, Moxley RA. (2000). Escherichia coli O157:H7 induces attaching-effacing lesions in large intestinal mucosal explants from adult cattle. FEMS Microbiol Lett. 15;185(2):239-42.

Barker, I. K. & Carbonell, P. L. (1974). Cryptosporidium agni spp. N. from lambs and Cryptosporidium bovis sp. N. from a calf with observations on the oocyst. Z Parasitenkd 44, 289–298.

Benjamini, Y. and Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. Roy. Statist. Soc. Ser. B 57 289-300.

Berin MC, Darfeuille-Michaud A, Egan LJ, Miyamoto Y, Kagnoff MF. (2002). Role of EHEC O157:H7 virulence factors in the activation of intestinal epithelial cell NF-kappaB and MAP kinase pathways and the upregulated expression of interleukin 8. Cell Microbiol. 2002 Oct;4(10):635-48.

Best, A., La Ragione, R. M., Sayers, A. R. & Woodward, M. J. (2005). Role for flagella but not intimin in the persistent infection of the gastro-intestinal tissues of specific pathogen free chicks by Stx negative Escherichia coli O157 : H7. Infect Immun 73, 1836–1846.

Borczyk A.A., Karmali M.A., Lior H., Duncan L.M. (1987). Bovine reservoir for verotoxin-producing Escherichia coli O157:H7, Lancet 1:98.

Brogden KA. (2005). Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria?Nat Rev Microbiol. 3(3):238-50. Review.

Brown, BW., Lovato, J. and Russell, K. (1999). Asymptotic power calculations: description, examples, computer code. Statistics in Medicine, 18, 22,  3137 – 3151.Buchko SJ, Holley RA, Olson WO, Gannon VP, Veira DM. (2000). The effect of different grain diets on fecal shedding of Escherichia coli O157:H7 by steers. J Food Prot. 63(11):1467-74.

Butler, J. E. (1979). Immunologic aspects of breast feeding, antiinfectious activity of breast milk. Semin Perinatol 3, 255–270.

Chapman, P. A., Cornell, J. & Green, C. (2000). Infection with verocytotoxin-producing Escherichia coli O157 :H7 during a visit to an inner city open farm. Epidemiol Infect 125, 553–536.

Chase-Topping ME, McKendrick IJ, Pearce MC, MacDonald P, Matthews L, Halliday J, Allison L, Fenlon D, Low JC, Gunn G, Woolhouse ME. (2007). Risk factors for the presence of high-level shedders of Escherichia coli O157 on Scottish farms. J Clin Microbiol. 45(5):1594-603.

Cobbold R.N., Hancock D.D., Rice D.H., Berg J., Stilborn R., Hovde C.J., Besser T.E. (2007). Rectoanal junction colonisation of feedlot cattle by Escherichia coli O157:H7 and its association with supershedders and excretion dynamics, Appl. Environ. Microbiol. 73:1563–1568.

SID 5 (Rev. 3/06) Page 40 of 47

Cornick NA, Booher SL, Casey TA, Moon HW. (2000). Persistent colonization of sheep by Escherichia coli O157:H7 and other E. coli pathotypes. Appl Environ Microbiol. 66(11):4926-34.

Cornick NA, Helgerson AF, Sharma V. (2007). Shiga toxin and Shiga toxin-encoding phage do not facilitate

Escherichia coli O157:H7 colonization in sheep. Appl Environ Microbiol. 73(1):344-6.

Cray WC Jr, Moon HW. (1995). Experimental infection of calves and adult cattle with Escherichia coli O157:H7. Appl Environ Microbiol. 61(4):1586-90.

De Keersmaecker SC, Verhoeven TL, Desair J, Marchal K, Vanderleyden J, Nagy I. (2006).Strong antimicrobial activity of Lactobacillus rhamnosus GG against Salmonella typhimurium is due to accumulation of lactic acid. FEMS Microbiol Lett. 59(1):89-96.

Dean-Nystrom, E. A., Bosworth, B. T. & Moon, H. W. (1997). Pathogenesis of O157 :H7 Escherichia coli infection in neonatal calves. Adv Exp Med Biol 412, 47–51.

Dean-Nystrom, E. A., Bosworth, B. T., O’Brien, A. D. & Moon, H. W. (1999). Bovine infection with Escherichia coli O157 : H7. In E. Coli O157 in Farm Animals. Edited by C. S. Stewart & H. J. Flint. Wallingford, UK: CAB International.

Dean-Nystrom, E. A., Gansheroff, L. J., Mills, M., Moon, H. W. & O’Brien, A. D. (2002). Vaccination of pregnant dams with intimin (O157) protects suckling piglets from Escherichia coli O157 :H7 infection. Infect Immun 70, 2414–2418.

De Graff, D. C., Vanopdenbosch, E., Ortega-Mora, L. M., Abbassi, H. & Peeters, J. E. (1999). A review of the importance of cryptosporidiosis in farm animals. Int J Parasitol 29, 1269–1287.

De la Fuente, R., Luzon, M., Ruiz-Santa-Quiteria, J. A., Garcia, A., Cid, D., Orden, J. A., Garcia, S., Sanz, R. & Gomez-Baustista, M. (1999). Cryptosporidium and concurrent infections with other major enteropathogens in 1 to 30 day old diarrheic dairy calves in central Spain. Vet Parasitol 80, 179–185.

DePamphilis ML, Adler J. (1971). Fine structure and isolation of the hook-basal body complex of flagella from Escherichia coli and Bacillus subtilis. J Bacteriol. 1971 Jan;105(1):384-95.

Dons L, Eriksson E, Jin Y, Rottenberg ME, Kristensson K, Larsen CN, Bresciani J, Olsen JE. (2004). Role of flagellin and the two-component CheA/CheY system of Listeria monocytogenes in host cell invasion and virulence. Infect Immun. 72(6):3237-44.

Dziva F, van Diemen PM, Stevens MP, Smith AJ, Wallis TS. 2004. Identification of Escherichia coli O157 : H7 genes influencing colonization of the bovine gastrointestinal tract using signature-tagged mutagenesis. Microbiology. 2004 Nov;150(Pt 11):3631-45.

Dziva F, Mahajan A, Cameron P, Currie C, McKendrick IJ, Wallis TS, Smith DG, Stevens MP. (2007). EspP, a Type V-secreted serine protease of enterohaemorrhagic Escherichia coli O157:H7, influences intestinal colonization of calves and adherence to bovine primary intestinal epithelial cells. FEMS Microbiol Lett. 271(2):258-64.

Eaves-Pyles T, Murthy K, Liaudet L, Virág L, Ross G, Soriano FG, Szabó C, Salzman AL. (2001). Flagellin, a novel mediator of Salmonella-induced epithelial activation and systemic inflammation: I kappa B alpha degradation, induction of nitric oxide synthase, induction of proinflammatory mediators, and cardiovascular dysfunction. J Immunol. 15;166(2):1248-60.

Elder RO, Keen JE, Siragusa GR, Barkocy-Gallagher GA, Koohmaraie M, Laegreid WW. (2002).Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing. Proc Natl Acad Sci U S A. 28;97(7):2999-3003.

Elliott, D. A. & Clark, D. P. (2000). Cryptosporidium parvum induces host cell actin accumulation at the host–parasite interface. Infect Immun 68, 2315–2322.

Elliott, D. A., Coleman, D. J., Lane, M. A., May, R. C., Machesky, L. M. & Clark, D. P. (2001). Cryptosporidium parvum infection requires host cell actin polymerization. Infect Immun 69, 5940–5942.

SID 5 (Rev. 3/06) Page 41 of 47

Erdem AL, Avelino F, Xicohtencatl-Cortes J, Girón JA. (2007). Host protein binding and adhesive properties of H6 and H7 flagella of attaching and effacing Escherichia coli. J Bacteriol. 189(20):7426-35.

Etcheverría AI, Arroyo GH, Perdigón G, Parma AE. (2006). Escherichia coli with anti-O157:H7 activity isolated from bovine colon. J Appl Microbiol. 100(2):384-9.

Falcão JP, Sharp F, Sperandio V. (2004). Cell-to-cell signaling in intestinal pathogens. Curr Issues Intest Microbiol. 2004 Mar;5(1):9-17.

Feldman M, Bryan R, Rajan S, Scheffler L, Brunnert S, Tang H, Prince A. (1998). Role of flagella in pathogenesis of Pseudomonas aeruginosa pulmonary infection. Infect Immun. 66(1):43-51.

Foreyt, W. J. (1990). Coccidiosis and cryptosporidiosis in sheep and goats. Vet Clin North Am Food Animal Pract 6, 655–670.

Fukushima H, Seki R. (2004). High numbers of Shiga toxin-producing Escherichia coli found in bovine faeces collected at slaughter in Japan. FEMS Microbiol Lett. 238(1):189-97.

Gewirtz AT, Simon PO Jr, Schmitt CK, Taylor LJ, Hagedorn CH, O'Brien AD, Neish AS, Madara JL. (2001). Salmonella typhimurium translocates flagellin across intestinal epithelia, inducing a proinflammatory response. J Clin Invest. 107(1):99-109.

Girard F, Dziva F, van Diemen P, Phillips AD, Stevens MP, Frankel G. (2007). Adherence of enterohemorrhagic Escherichia coli O157, O26, and O111 strains to bovine intestinal explants ex vivo. Appl Environ Microbiol. 2007 73(9):3084-90.

Girón JA, Torres AG, Freer E, Kaper JB. (2002). The flagella of enteropathogenic Escherichia coli mediate adherence to epithelial cells. Mol Microbiol. 44(2):361-79.

Grauke LJ, Kudva IT, Yoon JW, Hunt CW, Williams CJ, Hovde CJ. (2002). Gastrointestinal tract location of Escherichia coli O157:H7 in ruminants. Appl Environ Microbiol. ;68(5):2269-77.

Griffiths, J. K. (1998). Human cryptosporidiosis: epidemiology, transmission, clinical disease, treatment, and diagnosis. Adv Parasitol 40, 37–85.

Gunning, R. F., Wales, A. D., Pearson, G. R., Done, E., Cookson, A. L. & Woodward, M. J. (2001). Attaching and effacing lesions in the intestines of two calves associated with natural infection with Escherichia coli O26 : H11. Vet Rec 148, 780–782.

Hakkinen M, Schneitz C. (1996). Efficacy of a commercial competitive exclusion product against a chicken pathogenic Escherichia coli and E coli O157:H7. Vet Rec. 10;139(6):139-41.

Helling RB, Vargas CN, Adams J. (1987). Evolution of Escherichia coli during growth in a constant environment. Genetics. 116(3):349-58.

Hussein HS, Sakuma T. (2005). Shiga toxin-producing Escherichia coli: pre- and postharvest control measures to ensure safety of dairy cattle products.J Food Prot. 68(1):199-207. Review.

Hussein HS. (2007). Prevalence and pathogenicity of Shiga toxin-producing Escherichia coli in beef cattle and their products. J Anim Sci. 2007 Mar;85(13 Suppl):E63-72. Epub 2006 Oct 23. Review.

Inglis TJ, Robertson T, Woods DE, Dutton N, Chang BJ. (2003). Flagellum-mediated adhesion by Burkholderia pseudomallei precedes invasion of Acanthamoeba astronyxis. Infect Immun. 71(4):2280-2.

Janke, B. H., Francis, D. H., Collins, J. E., Libal, M. C., Zeman, D. H., Johnson, D. D. & Neiger, R. D. (1990). Attaching and effacing Escherichia coli infection as a cause of diarrhea in young calves. J Am Vet Med Assoc 196, 897–901.

Jenkins C, Pearce MC, Chart H, Cheasty T, Willshaw GA, Gunn GJ, Dougan G, Smith HR, Synge BA, Frankel G. (2002). An eight-month study of a population of verocytotoxigenic Escherichia coli (VTEC) in a Scottish cattle herd. J Appl Microbiol. 93(6):944-53.

Kelleher, S. L. & Lonnerdal, B. (2001). Immunological activities associated with milk. Adv Nutr Res 10, 39–65.

SID 5 (Rev. 3/06) Page 42 of 47

Kenward MG, Roger JH. (1997). Small sample inference for fixed effects from restricted maximum likelihood.Biometrics. 53(3):983-97.

Kirov SM, Castrisios M, Shaw JG. (2004). Aeromonas flagella (polar and lateral) are enterocyte adhesins that contribute to biofilm formation on surfaces. Infect Immun. 72(4):1939-45.

Knutton, S., Baldwin, T., Williams, P. H. & McNeish, A. S. (1989). Actin accumulation at sites of bacterial adhesion to tissue culture cells: basis of a new diagnostic test for enteropathogenic and enterohaemorrhagic E. coli. Infect Immun 57, 1290–1298.

Koudela, B. & Jiri, V. (1997). Experimental cryptosporidiosis in kids. Vet Parasitol 71, 273–281.

La Ragione, R. M., A. Best, K. A. Sprigings, W. A. Cooley, M. A. Jepson, and M. J. Woodward. 2004. Interaction between attaching and effacing Escherichia coli serotypes O157:H7 and O26:K60 in cell culture. Vet. Microbiol. 104:119–124.

La Ragione, R. M., Narbad, A., Gasson, M. J. & Woodward, M. J. 2004a. In vivo characterisation of Lactobacilli johnsonii F19785 for use as a defined competitive exclusion agent against bacterial pathogens in poultry. Letts Appl. Micro. 38, 197-205.

La Ragione, R. M., Ahmed, N. M. Y., Best, A., Clifford, D., Weyer, U., Cooley, W. A., Johnson, L., Pearson, G. R. & Woodward, M. J. (2005). Colonisation of eight-week-old conventionally-reared goats by Escherichia coli O157 :H7 after oral inoculation. J Med Microbiol 54, 485–492.

La Ragione, R. M., Best, A., Clifford, D., Weyer, U. & Woodward, M. J. (2005a). Failure to detect transmission of E. coli O157 :H7 from orally dosed nannies to suckling kids at foot. Vet Rec 157, 659–661.

La Ragione, R. M., Best, A., Clifford, D., Weyer, U., Johnson, L., Marshall, R. M., Marshall, J., Cooley, W. A., Pearson, G. R. & Woodward, M. J. 2006. Influence of colostrum deprivation and concurrent Cryptosporidium parvum infection on the colonisation and persistence of Escherichia coli O157:H7 in young lambs. J. Med. Micro. 55, 819-828.

La Ragione RM, Best A, Woodward MJ, Wales AD. 2009. Escherichia coli O157:H7 colonization in small domestic ruminants. FEMS Microbiol Rev. 33(2):394-410. Review.

Lema M, Williams L, Rao DR. (2001). Reduction of faecal shedding of enterohemorrhagic Escherichia coli O157:H7 in lambs by feeding microbial feed supplement. Small Rumin Res. 39(1):31-39.

Lillehoj EP, Kim BT, Kim KC. (2002). Identification of Pseudomonas aeruginosa flagellin as an adhesin for Muc1 mucin. Am J Physiol Lung Cell Mol Physiol. 282(4):L751-6.

Lim JY, Li J, Sheng H, Besser TE, Potter K, Hovde CJ. (2006). Escherichia coli O157:H7 colonization at the rectoanal junction of long-duration culture-positive cattle. Appl Environ Microbiol. ;73(4):1380-2.

Lim JY, Li J, Sheng H, Besser TE, Potter K, Hovde CJ. (2006). Escherichia coli O157:H7 colonization at the rectoanal junction of long-duration culture-positive cattle. Appl Environ Microbiol. 3(4):1380-2.

Locking, M. E., O’Brien, S. J., Reilly, W. J., Wright, E. M., Campbell, D. M., Coia, J. E., Browning, L. M. & Ramsay, C. N. (2001). Risk factors for sporadic cases of Escherichia coli O157 infection: the importance of contact with animal excreta. Epidemiol Infect 127, 215–220.

Low JC, McKendrick IJ, McKechnie C, Fenlon D, Naylor SW, Currie C, Smith DG, Allison L, Gally DL. 2005. Rectal carriage of enterohemorrhagic Escherichia coli O157 in slaughtered cattle. Appl Environ Microbiol. 71(1):93-7.

Low AS, Dziva F, Torres AG, Martinez JL, Rosser T, Naylor S, Spears K, Holden N, Mahajan A, Findlay J, Sales J, Smith DG, Low JC, Stevens MP, Gally DL. (2006). Cloning, expression, and characterization of fimbrial operon F9 from enterohemorrhagic Escherichia coli O157:H7. Infect Immun. 2006 Apr;74(4):2233-44.

Luck SN, Badea L, Bennett-Wood V, Robins-Browne R, Hartland EL. (2006). Contribution of FliC to epithelial cell invasion by enterohemorrhagic Escherichia coli O113:H21.Infect Immun. 2006 Dec;74(12):6999-7004. Epub 2006 Sep 18.

SID 5 (Rev. 3/06) Page 43 of 47

Iyoda S, Koizumi N, Satou H, Lu Y, Saitoh T, Ohnishi M, Watanabe H. (2006). The GrlR-GrlA regulatory system coordinately controls the expression of flagellar and LEE-encoded type III protein secretion systems in enterohemorrhagic Escherichia coli. J Bacteriol. 188(16):5682-92.

Mack DR, Ahrne S, Hyde L, Wei S, Hollingsworth MA. 2003. Extracellular MUC3 mucin secretion follows adherence of Lactobacillus strains to intestinal epithelial cells in vitro.Gut. 52(6):827-33.

Mahajan A, Naylor S, Mills AD, Low JC, Mackellar A, Hoey DE, Currie CG, Gally DL, Huntley J, Smith DG. (2005). Phenotypic and functional characterisation of follicle-associated epithelium of rectal lymphoid tissue. Cell Tissue Res. 321(3):365-74. Matthews L, McKendrick IJ, Ternent H, Gunn GJ, Synge B, Woolhouse ME. (2006a). Super-shedding cattle and the transmission dynamics of Escherichia coli O157. Epidemiol Infect. 2006 Feb;134(1):131-42.

Matthews L, Low JC, Gally DL, Pearce MC, Mellor DJ, Heesterbeek JA, Chase-Topping M, Naylor SW, Shaw DJ, Reid SW, Gunn GJ, Woolhouse ME. (2006b). Heterogeneous shedding of Escherichia coli O157 in cattle and its implications for control.Proc Natl Acad Sci U S A. 17;103(3):547-52. Epub 2006 Jan 9.

McNeilly TN, Naylor SW, Mitchell MC, McAteer S, Mahajan A, Smith DG, Gally DL, Low JC, Huntley JF. (2007). Simple methods for measurement of bovine mucosal antibody responses in vivo. Vet Immunol Immunopathol. 15;118(1-2):160-7.

Midgley J, Fegan N, Desmarchelier P. (1999). Dynamics of Shiga toxin-producing Escherichia coli (STEC) in feedlot cattle. Lett Appl Microbiol. 29(2):85-9.

Moon, H. W., McClurkin, A. W., Isaacson, R. E., Pohlenz, J., Skartvedt, S. M., Gillette, K. G. & Baetz, A. L. (1978). Pathogenic relationships of rotavirus, Escherichia coli, and other agents in mixed infections in calves. J Am Vet Med Assoc 173, 577–583.

Morin, M., Lariviee, S., Lallier, R., Begin, M. E., Ethier, R., Roy, R. S. & Tremblay, A. (1980). Diarrhoea of newborn calves. II. Agents responsible for the disease on Quebec dairy farm. Med Vet Québec 10, 60–65.

Moxley RA. (2004). Escherichia coli 0157:H7: an update on intestinal colonization and virulence mechanisms. Anim Health Res Rev. 5(1):15-33. Review.

Munoz-Fernandez, M., Alvarez, M., Lanza, I. & Carmenes, P. (1996). Role of enteric pathogens in the aetiology of neonatal diarrhoea in lambs and goat kids in Spain. Epidemiol Infect 117, 203–211.

Naylor, S. W., Low, J. C., Besser, T. E., Mahajan, A., Gunn, G. J., Pearce, M. C., McKendrick, I. J., Smith, D. G. & Gally, D. L. (2003). Lymphoid follicle-dense mucosa at the terminal rectum is the principal site of colonization of enterohemorrhagic Escherichia coli O157 :H7 in the bovine host. Infect Immun 71, 1505–1512.

Noordeen, F., Horaldagoda, N. U., Faizal, A. C. M., Rajapakse, R. P. V. J., Razak, M. A. A. & Arulkanthan, A. (2002). Infectivity of Cryptosporidium parvum isolated from asymptomatic adult goats to mice and goat kids. Vet Parasitol 103, 217–225.

O'Brien SJ, Adak GK, Gilham C. (2001). Contact with farming environment as a major risk factor for Shiga toxin (Vero cytotoxin)-producing Escherichia coli O157 infection in humans.Emerg Infect Dis. 7(6):1049-51.

O’Donoghue, P. J. (1995). Cryptosporidium and cryptosporidiosis in man and animals. Int J Parasitol 25, 139–195.

Ogawa M, Shimizu K, Nomoto K, Takahashi M, Watanuki M, Tanaka R, Tanaka T, Hamabata T, Yamasaki S, Takeda Y. (2001). Protective effect of Lactobacillus casei strain Shirota on Shiga toxin-producing Escherichia coli O157:H7 infection in infant rabbits. Infect Immun. 69(2):1101-8.

Ohya T, Marubashi T, Ito H. (2000). Significance of fecal volatile fatty acids in shedding of Escherichia coli O157 from calves: experimental infection and preliminary use of a probiotic product. J Vet Med Sci. 62(11):1151-5.

Orskov F, Orskov I, Villar JA. (1987). Cattle as reservoir of verotoxin-producing Escherichia coli O157:H7. Lancet. 1987 Aug 1;2(8553):276.

SID 5 (Rev. 3/06) Page 44 of 47

Ostroff SM, Tarr PI, Neill MA, Lewis JH, Hargrett-Bean N, Kobayashi JM. (1989). Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coli O157:H7 infections. J Infect Dis. 1989 Dec;160(6):994-8.

Paiba GA, Gibbens JC, Pascoe SJ, Wilesmith JW, Kidd SA, Byrne C, Ryan JB, Smith RP, McLaren M, Futter RJ, Kay AC, Jones YE, Chappell SA, Willshaw GA, Cheasty T. 2002. Faecal carriage of verocytotoxin-producing Escherichia coli O157 in cattle and sheep at slaughter in Great Britain. Vet Rec. 11;150(19):593-8.

Payne, C. J., Petrovic, M., Roberts, R. J. & 9 other authors (2003). Vero cytotoxin-producing Escherichia coli O157 gastroenteritis in farm visitors, North Wales. Emerg Infect Dis 9, 526–530. Pepin, M., Russo, P. & Pardon, P. (1997). Public health hazards from small ruminant meat products in Europe. Rev Sci Technol 16, 415–425.

Pérez PF, Minnaard J, Rouvet M, Knabenhans C, Brassart D, De Antoni GL, Schiffrin EJ. (2001).Inhibition of Giardia intestinalis by extracellular factors from Lactobacilli: an in vitro study. Appl Environ Microbiol. 67(11):5037-42.

Frankel G, Philips AD, Novakova M, Batchelor M, Hicks S, Dougan G. (1998). Generation of Escherichia coli intimin derivatives with differing biological activities using site-directed mutagenesis of the intimin C-terminus domain. Mol Microbiol. 1998 Jul;29(2):559-70.

Pritchard, G. C., Willshaw, G. A., Bailey, J. R., Carson, T. & Cheasty, T. (2000). Verocytotoxin-producing Escherichia coli O157 on a farm open to the public: outbreak investigation and longitudinal bacteriological study. Vet Rec 147, 259–264.

Reading NC, Sperandio V. (2006). Quorum sensing: the many languages of bacteria. FEMS Microbiol Lett. 54(1):1-11.

Reynolds, D. J., Morgan, J. H., Chanter, N., Jones, P. W., Bridger, J. C., Debney, T. G. & Bunch, K. L. (1986). Microbiology of calf diarrhoea in southern Britain. Vet Rec 119, 34–39.

Rice DH, Sheng HQ, Wynia SA, Hovde CJ. 2003. Rectoanal mucosal swab culture is more sensitive than fecal culture and distinguishes Escherichia coli O157:H7-colonized cattle and those transiently shedding the same organism. J Clin Microbiol. 41(11):4924-9.

Ruckebusch Y, Fioramonti J. (1980). Colonic myoelectrical spiking activity: major patterns and significance in six different species. Zentralbl Veterinarmed A. 1980 27(1):1-8.

Rugsbjerg, H., Nielsen, E. M. & Andersen, J. S. (2003). Risk factors associated with faecal shedding of verocytotoxin-producing Escherichia coli O157 in eight known-infected Danish dairy herds. Prev Vet Med 58, 101–113.

Shaner, N. C., Sanger, J. W. & Sanger, J. M. (2005). Actin and alphaactinin dynamics in the adhesion and motility of EPEC and EHEC on host cells. Cell Motil Cytoskeleton 60, 104–120.

Sheng H, Davis MA, Knecht HJ, Hovde CJ. (2004). Rectal administration of Escherichia coli O157:H7: novel model for colonization of ruminants. Appl Environ Microbiol. 2004 Aug;70(8):4588-95.

Sperandio V, Torres AG, Jarvis B, Nataro JP, Kaper JB. (2003). Bacteria-host communication: the language of hormones. Proc Natl Acad Sci U S A. 100(15):8951-6. Epub 2003 Jul 7.

Stavric S, Buchanan B, Gleeson TM. (1993). Intestinal colonization of young chicks with Escherichia coli O157:H7 and other verotoxin-producing serotypes. J Appl Bacteriol. 74(5):557-63.

Tasteyre A, Barc MC, Collignon A, Boureau H, Karjalainen T. (2001). Role of FliC and FliD flagellar proteins of Clostridium difficile in adherence and gut colonization. Infect Immun. 69(12):7937-40.

Tkalcic S, Zhao T, Harmon BG, Doyle MP, Brown CA, Zhao P. (2003). Fecal shedding of enterohemorrhagic Escherichia coli in weaned calves following treatment with probiotic Escherichia coli. J Food Prot. 66(7):1184-9.

Torres, A. G., Milflores-Flores, L., Garcia-Gallegos, J. G., Best, A., La Ragione, R. M., Martinez-Laguna, Y. & Woodward, M. J. 2007. Environmental regulation and colonisation attributes of the Long Polar fimbriae (LPF) of Escherichia coli O157:H7. Int. J. Med. Micro. 297, 177-185.

Tzipori, S., Angus, K. W., Gray, E. W., Campbell, I. & Allan, F. (1981). Diarrhoea in lambs experimentally infected with Cryptosporidium isolated from calves. Am J Vet Res 42, 1400–1404.

SID 5 (Rev. 3/06) Page 45 of 47

Tatsuno, I., Mundy, R., Frankel, G., Chong, Y., Phillips, A.D., Torres, A.G., Kaper, J.B., 2006. The lpf gene cluster for long polar fimbriae is not involved in adherence of enteropathogenic Escherichia coli or virulence of Citrobacter rodentium. Infect. Immun. 74, 265–272.

Vlisidou I, Dziva F, La Ragione RM, Best A, Garmendia J, Hawes P, Monaghan P, Cawthraw SA, Frankel G, Woodward MJ, Stevens MP. 2006. Role of intimin-tir interactions and the tir-cytoskeleton coupling protein in the colonization of calves and lambs by Escherichia coli O157:H7. Infect Immun. 74(1):758-64.

Wales, A. D., Clifton-Hadley, F. A., Cookson, A. L., Dibb-Fuller, M. P., La Ragione, R. M., Sprigings, K. A., Pearson, G. R. & Woodward, M. J. (2001a). Experimental infection of six-month old lambs with Escherichia coli O157 : H7. Vet Rec 148, 630–631.

Wales, A. D., Pearson, G. R., Skuse, A. M., Roe, J. M., Hayes, C. M., Cookson, A. L. & Woodward, M. J. (2001b). Attaching and effacing lesions caused by Escherichia coli O157 :H7 in experimentally inoculated neonatal lambs. J Med Microbiol 50, 752–758.

Wales, A. D., Pearson, G. R., Roe, J. M., Hayes, C. M., La Ragione, R. M. & Woodward, M. J. (2005). Attaching-effacing lesions associated with Escherichia coli O157 :H7 and other bacteria in experimentally inoculated conventional neonatal goat kids. J CompPathol 132, 185–194.

Whittam TS, Wolfe ML, Wachsmuth IK, Orskov F, Orskov I, Wilson RA. (1993). Clonal relationships among Escherichia coli strains that cause hemorrhagic colitis and infantile diarrhea. Infect Immun. 61(5):1619-29.

Woodward, M. J., Best, A., Sprigings, K. A. & 7 other authors. (2003). Non-toxigenic Escherichia coli O157 :H7 strain NCTC12900 causes attaching-effacing lesions and eae dependent persistence in weaned sheep. Int J Med Microbiol 293, 299–308.

Zhao T, Tkalcic S, Doyle MP, Harmon BG, Brown CA, Zhao P. (2003). Pathogenicity of enterohemorrhagic Escherichia coli in neonatal calves and evaluation of fecal shedding by treatment with probiotic Escherichia coli. J Food Prot. 66(6):924-30.

The project leader wishes to acknowledge the input from:-

Prof Mark Stevens and his team at the Institute for Animal HealthProf Gadi Frankel and his team at Imperial CollegeProf David Smith and his team at the Moredun InstituteProf David Gally and his team at Edinburgh UniversityDr Alfredo Torres and his team at Texas UniversityDr Geoff Pearson and his team at Bristol University

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

SID 5 (Rev. 3/06) Page 46 of 47

Papers from Objective 01Olivier Marches, Siouxsie Wiles, Francis Dziva, Roberto, M. La Ragione, Yuwen Chong, Angus Best , Eric Oswald, Alan, D. Phillips, Elizabeth L. Hartland, Martin J. Woodward, Mark P. Stevens

and Gad Frankel. 2005. Characterization of two non-LEE encoded type III effectors, NleC and NleD, in attaching and effacing pathogens: role in virulence. I & I. 73, 8411-8417.Torres, A. G. Lorena Milflores-Flores, J. Gerardo Garcia-Gallegos, Angus Best, Roberto, M. La Ragione, Ygnacio Martinez-Laguna and Martin J. Woodward 2007. Environmental regulation and virulence attributes of the Long Polar fimbriae of Escherichia coli O157:H7. International journal of Med. Micro. 297, 177-185.Tadasuke Ooka, Mõnica A. M. Vieira, Yoshitoshi Ogura, Lothar Beutin, Roberto La Ragione, Pauline M. van Diemen, Mark P. Stevens, Ilknur Aktan, Shaun Cawthraw, Angus Best, Rodrigo T. Hernandes, Gladys Krause, Tania Gomes, Tetsuya Hayashi, Gad Frankel. 2007. Characterisation of tccP2 of atypical enteropathogenic Escherichia coli of human and animal origin. FEMS Micro. Letts. 271, 126-35.Vlisidou, I., Dziva, F., La Ragione, R. M., Best, A., Garmendia, J., Hawes, P., Cawthraw, S. A., Monaghan, P., Frankel, G., Woodward, M.J. & Stevens, M.P. 2006. Role of intimin-Tir interactions and Tir cytoskeleton coupling protein in the colonization of calves and lambs by Escherichia coli O157:H7. I & I. 74, 758-764.Hemrajani, C., Marches, O., Wiles, S., Girard, F., Dennis, A., Dviza, F.m Best, A., Phillips, AD., Mousnier, A., Crepin, VF., Krudenier, L., Woodward. MJ., Stevens, MP., La Ragione, RM., MacDonald, TT. and Frankel, G. 2008. Role of NleH, a type III secreted effector from attaching and effacing pathogens, in colonization of the bovine, ovine and murine gut. Infect Immun. 2008 Aug 25. [Epub ahead of print].

Papers from Objective 02La Ragione, R. M., Best, A., Clifford, D., Weyer, U., Johnson, L., Marshall, R. M., Marshall, J., Cooley, W. A., Pearson, G. R. and Woodward, M. J. 2006. Influence of Cryptosporidium on Escherichia coli O157:H7 colonisation and persistence in young lambs. Journal of Medical Microbiology. 55, 819-828.

Papers from Objective 03Best, A., Woodward, M. J. La Ragione, R. M. 2007. Rectal colonisation of young sheep. Infect and Immun. Submitted.Girard, F., Frankel, G., Phillips, AD., Weyer, U., Dugdale, AH., Woodward, MJ. and La Ragione, RM. 2008. Interaction of enterohemorrhagic Escherichia coli O157:H7 with mouse intestinal mucosa. FEMS Microbiol Lett. 283(2):196-202. Mahajan, A., Currie, C., Mackie, S., Wilson, L., McKendrick, I., McNeilly, T., Roe, A., La Ragione, R., Best, A., Woodward, M.J., Gally, D., Smith, D. 2008. Role of flagella in adherence of E. coli O157:H7 to intestinal epithelium. Cellular Microbiology. (e-pub ahead of print)

General papers with contributions from this project (added value)La Ragione, R. M. Best, A., Woodward, M. J. and Wales, A. 2008 . E. coli O157 in small ruminants. Review. FEMS microbiology Letters. accepted.Cooley WA; La Ragione RM; Jepson MA; Woodward MJ 2005. "A confocal and scanning electron microscopy study on the interaction of attaching and effacing E. coli." Scanning. 27, 72-73.Alfredo Caprioli, Stefano Morabito, Flemming Scheutz, Henrik Chart, Eric Oswald, Maurizio Brigotti, Leo Monnens, Anna Aspan, Roberto La Ragione, Chris Low, and Diane Newell. Pathogenesis of Verocytotoxin/Shiga toxin-producing E. coli infection. 12, 8 Emerging Infectious Diseases.Aktan, I. La Ragione, R. M. & Woodward, M. J. 2007. Colonisation, persistence and tissue tropism of Escherichia coli O26 in conventionally reared weaned lambs. AEM. 73, 691-698.Aktan İ., La Ragione R. M., Pritchard, G. C. and Woodward, M. J. 2007. Prevalence of attaching and effacing E. coli serogroup O103 in orphan lambs. Vet Record.161, 386-388.Aktan, I., La Ragione, R. M., Henddrik Wilking, Lothar Wieler, Woodward, M. J. & Anjum, M. F. 2007. The influence of geography, host animal and stx gene on the virulence characteristics of Escherichia coli O26 strains. J. Med. Micro. 56:1431-9..

Posters – SGM, Zoonoses conference, Med-Vet-Net and VTEC 2006Oral presentations - AstraZeneca, Med-Vet-Net Royal Veterinary College, Surrey University, IAH, and The European coccidiosis discussion group,

SID 5 (Rev. 3/06) Page 47 of 47