Epe e Coli527.Full

Enteropathogenic Escherichia coli (EPEC)adhesion to intestinal epithelial cells: roleof bundle-forming pili (BFP), EspA filamentsand intimin

Jennifer Cleary,1 Li-Ching Lai,2 Robert K. Shaw,1

Anna Straatman-Iwanowska,1 Michael S. Donnenberg,2

Gad Frankel3 and Stuart Knutton1

Correspondence

Stuart Knutton

[email protected]

1Institute of Child Health, University of Birmingham, Birmingham, UK

2Division of Infectious Diseases, University of Maryland, Baltimore, MD, USA

3Centre for Molecular Microbiology and Infection, Department of Biosciences, Imperial College,London, UK

Received 28 August 2003

Revised 18 November 2003

Accepted 25 November 2003

Enteropathogenic Escherichia coli (EPEC), an important paediatric diarrhoeal pathogen,

employs multiple adhesins to colonize the small bowel and produces characteristic attaching

and effacing (A/E) lesions on small intestinal enterocytes. EPEC adhesins that have been

associated with A/E adhesion and intestinal colonization include bundle-forming pili (BFP),

EspA filaments and intimin. BFP are involved in bacteriabacteria interaction and microcolony

formation but their role in cell adhesion remains unclear; EspA filaments are components of the

EPEC type III secretion system but since they interact directly with host cells they may also

function as adhesins; intimin is the well characterized intimate EPEC adhesin which binds the

translocated intimin receptor, Tir. However, other uncharacterized host cell receptors have been

implicated in intimin-mediated adhesion. In this study, the role of BFP, EspA filaments and

intimin in EPEC adhesion to intestinal brush border cells was assessed by observing adhesion

of wild-type EPEC strain E2348/69 and a set of isogenic single, double and triple mutants in

bfpA, espA and eae (intimin gene) to differentiated human intestinal Caco-2 cells. E2348/69

(bfpA+ espA+ eae+) adhered rapidly (1 h) and with much

lower efficiency; confocal microscopy indicated BFP and EspA-mediated adhesion, respectively.

Strain UMD883 (bfpA espA eae+), which is unable to translocate Tir, was non-adherent

although this strain was able to form intimate attachment and A/E lesions when co-cultured

with strain CVD206 (bfpA+ espA+ eae) which supplied Tir to the membrane. Single mutant

strains CVD206 (bfpA+ espA+ eae) and UMD872 (bfpA+ espA eae+) showed adherence

characteristics of strain UMD880 (bfpA+ espA eae), whilst triple mutant strain UMD888

(bfpA espA eae) was totally non-adherent. These results support an adhesive role for BFP

and EspA in initial brush border cell attachment, and in typical EPEC which express both BFP

and EspA filaments suggest a predominant role for BFP; EspA filaments, however, could serve as

initial attachment factors in atypical EPEC which lacks BFP. The study found no evidence for an

independent host cell intimin receptor or for other adhesive factors able to support bacterial

adherence.

Abbreviations: A/E, attaching and effacing; BFP, bundle-forming pilus/pili; EPEC, enteropathogenic Escherichia coli ; LEE, locus of enterocyteeffacement; RBC, red blood cell; SEM, scanning electron microscopy; TTSS, type III secretion system; WGA, wheat germ agglutinin.

0002-6740 G 2004 SGM Printed in Great Britain 527

Microbiology (2004), 150, 527538 DOI 10.1099/mic.0.26740-0

INTRODUCTION

Enteropathogenic Escherichia coli (EPEC), an importantcause of infantile diarrhoea in the developing world (Nataro& Kaper, 1998), colonizes the small bowel and producesattaching and effacing (A/E) lesions on the brush bordersurface of small intestinal enterocytes characterized byeffacement of brush border microvilli, intimate attachmentof bacteria and formation of an actin-rich pedestal beneathintimately attached bacteria (Knutton et al., 1989; Moonet al., 1983). Current evidence, derived mainly from in vitrostudies using cultured epithelial cell lines, supports amodel of A/E lesion formation which involves initial non-intimate bacterial adhesion, type III secretion and effectorprotein translocation to the host cell followed by brushborder effacement, intimate bacterial attachment andpedestal formation (Frankel et al., 1998). Several adhesinshave been implicated in A/E EPEC adhesion to epithelialcells including the type IV bundle-forming pilus (BFP)(Giron et al., 1991), type III secretion system (TTSS) EspAfilaments (Knutton et al., 1998) and the outer-membraneadhesin, intimin (Donnenberg & Kaper, 1991).

BFP, encoded on a large ~80 kb EPEC adherence factorplasmid (EAF) (Nataro et al., 1987b) present in typical EPECstrains (Kaper, 1996), have been shown to be important forEPEC pathogenicity (Bieber et al., 1998), to be responsiblefor bacteriabacteria interaction and microcolony forma-tion (Giron et al., 1991) and also for dispersal of bacteriafrom microcolonies (Bieber et al., 1998; Knutton et al.,1999). Several studies have also implicated BFP in initialbinding of EPEC to host epithelial cells (Donnenberg et al.,1992; Giron et al., 1991; Tobe & Sasakawa, 2001, 2002)although other studies which used human intestinal organculture suggested that BFP were not involved in initial EPECadherence but only at a later stage to promote microcolonyformation (Hicks et al., 1998).

Genes encoding A/E lesion formation are localized to thelocus of enterocyte effacement pathogenicity island (LEEPAI) (McDaniel et al., 1995). The LEE encodes a TTSS,several secreted translocator and effector proteins andproteins involved in intimate bacterial attachment (Frankelet al., 1998). One secreted translocator protein, EspA, formsa long filamentous extension to the TTSS needle complex(Daniell et al., 2001; Sekiya et al., 2001), whereas two othertranslocator proteins, EspB and EspD, are thought to pro-vide a translocation pore in the host cell membrane, therebyallowing translocation of LEE effector proteins into thehost cell cytosol (Frankel et al., 1998). Since EspA filamentsform a direct interaction with the host cell during earlystages of A/E lesion formation (Knutton et al., 1998), theycould function as an initial EPEC attachment factor,particularly in the case of atypical EPEC which lacks EAFplasmids (Kaper, 1996).

Intimin is the well characterized EPEC adhesin that pro-motes intimate attachment to host cells. However, to forman intimate attachment and A/E lesion formation, the

LEE-encoded effector protein Tir (translocated intiminreceptor) has to be translocated and inserted into the hostcell membrane by the TTSS (Kenny et al., 1997); theintiminTir interaction triggers the assembly of actin intopedestals beneath intimately attached bacteria (Campellone& Leong, 2003). Whilst Tir is a well characterized intiminreceptor, additional uncharacterized host cell receptorshave been implicated in intimin-mediated EPEC adhesion(Frankel et al., 2001) and recently a specific host cell protein,nucleolin, was shown to bind a specific intimin subtype,intimin c, expressed by some EPEC and enterohaemorrhagicE. coli strains (Sinclair & OBrien, 2002).

The difficulty in assessing an adhesive role of BFP, EspAfilaments and intimin when they are being expressedsimultaneously in wild-type EPEC prompted us to generatea set of single, double and triple mutants in bfpA, espA andeae and to examine the role of these putative adhesinsindividually and in pairs. Human intestinal brush bordercells provide a better model of EPEC infection than doundifferentiated epithelial cells of non-intestinal origin.Hence, in this study, we examined EPEC adhesion toCaco-2 brush border cells. The study supports a role forBFP and EspA filaments in initial brush border attach-ment of EPEC and intiminTir interaction in intimateEPEC adhesion and A/E lesion formation. The studyfound no evidence for other adhesins able to promoteEPEC adhesion.

METHODS

Bacterial strains. The strains and plasmids used in the study areshown in Table 1. All mutants are derivatives of EPEC serotypeO127 : H6 strain E2348/69 (Levine et al., 1978). The followingmutant strains have been described previously: eae deletion mutantstrain CVD206, which does not produce intimin (Donnenberg &Kaper, 1991); espA deletion mutant strain UMD872, which doesnot produce EspA (Kenny et al., 1996); and bfpA mutant strainUMD901, which has a substitution of serine for cysteine at aminoacid 129 that renders the bundlin pilin protein unstable and rapidlydegraded and therefore does not produce BFP (Zhang &Donnenberg, 1996). The suicide vectors originally used to generatethese mutants were used to make the following by allelic exchange:UMD880 is an espA eae double mutant; UMD883 is a bfpA espAdouble mutant; UMD886 is a bfpA eae double mutant; andUMD888 is a bfpA espA eae triple mutant. Plasmids pRPA100,pMSD2 and pCVD438, used to transform bfp, espA and eae, respec-tively, have been described previously (Anantha et al., 2000;Donnenberg & Kaper, 1991; Kenny et al., 1996). Bacteria were rou-tinely grown overnight in Luria broth (LB) and then subcultured(1 : 100) in tissue culture media for cell adhesion studies. Bacterialmotility assays were performed in plastic vials in LB and DulbeccosModified Eagles Medium (DMEM) supplemented with 0?3 % agarand read after 18 h incubation at 37 uC.

Infection of HEp-2 cells. Adhesion to HEp-2 cells was carried outaccording to the method of Cravioto et al. (1979). Subconfluent cellcultures on glass coverslips were washed and incubated with bacteria[10 ml overnight LB culture per millilitre of tissue culture medium(DMEM/5 % fetal calf serum)] for 3 h at 37 uC. After thoroughwashing to remove non-adhering bacteria, cells were either fixedin methanol and stained with Giemsa to assess the pattern of cell

528 Microbiology 150

J. Cleary and others

adhesion or fixed in 4 % formalin and A/E lesion formation assessedusing the fluorescent actin staining test (Knutton et al., 1989). Semi-quantitative assessment of adhesion was made by counting 100 cellsin representative microscope fields and determining the percentageof cells with adherent bacteria. Adhesion was scored from 2 (non-adherent) to +++++ (80100 % cells with adherent bacteria).

Infection of Caco-2 cells. Caco-2 intestinal cells were grown inDMEM/20 % fetal calf serum on 24 mm diameter Transwell permeableculture supports for 12 days to ensure differentiation of the brushborder. EPEC infections were performed in HEPES-buffered ModifiedEagles Medium (MEM)/2 % fetal calf serum. Cell monolayers wereinfected with an overnight LB culture of EPEC [10 ml broth culture(ml culture medium)21] for up to 6 h at 37 uC; for long infections,the medium was changed after 3 h. In co-culture experiments, 5 mlbroth culture (ml medium)21 of each strain was used. At the end ofthe infection period, cells were washed three times in PBS for 1 minon an orbital shaker and fixed in either 4 % formalin (for immuno-fluorescence studies) or 3 % buffered glutaraldehyde [for scanningelectron microscopy (SEM)]. In some experiments, a gentle washingprocedure was used. In this case cells were washed three times for20 s with gentle rocking by hand. Quantitative assessment of adhe-sion was made from two separate infections by counting numbers ofbacteria in three representative microscope fields taken with a 636objective lens. Adhesion (mean number of bacteria per fieldSD) isexpressed relative to that of E2348/69 (100 %).

Infection of human red blood cell (RBC) monolayers. RBCmonolayers were produced as described previously (Knutton et al.,2002). EPEC infections were performed in HEPES-buffered DMEMand cell monolayers infected with an overnight LB culture of EPEC[10 ml broth culture (ml culture medium)21] for 3 h at 37 uC. Afterthree washes in PBS, monolayers were fixed in 4 % formalin.

Infection of human intestinal mucosa. Normal paediatric duo-denal mucosal biopsies taken with informed consent were trans-ported to the laboratory in ice-cold transport medium as describedpreviously (Knutton et al., 1987) and used immediately. Bacteria

grown for 3 h in DMEM express LEE genes and are primed readyto produce A/E lesions (Collington et al., 1998). Primed bacteriawere used to infect mucosal biopsies. Biopsies were placed in Bijouxbottles with 1?5 ml primed bacterial culture and 1?5 ml fresh culturemedium (DMEM : NCTC135/10 % FCS) and the bottles placed on arotary mixer inside an incubator for 1?5 h at 37 uC. Biopsies werewashed three times in warm PBS on the rotary mixer and fixed in3 % glutaraldehyde.

Antibodies and fluorescent stains. Polyclonal rabbit or mouseantibodies to BFP (Knutton et al., 1999), EspA (Knutton et al.,1998), intimin (Batchelor et al., 1999), Tir (Hartland et al., 1999)and H6 flagellin (Yona-Nadler et al., 2003) were used for immuno-fluorescence staining in conjunction with Alexa488 (green) andAlexa594 (red) goat anti-rabbit IgG or goat anti-mouse IgGsecondary antibody conjugates (Molecular Probes). In addition,bacteria were stained with propidium iodide (red) or DAPI (blue)(Molecular Probes), cell actin with fluorescein (green) or rhodamine(red) conjugated phalloidin (Sigma) and RBCs and the brush bordermembrane with rhodamine-conjugated wheat germ agglutinin(WGA) (Sigma).

Immunofluorescence microscopy. To stain Tir and cytoskeletalactin, Caco-2 cells were permeabilized with PBS containing 0?1 %Triton X-100 for 5 min and washed three times in PBS prior toimmunostaining. All antibody dilutions and immune reactionswere carried out in PBS containing 0?2 % bovine serum albumin(PBS/BSA). Formalin-fixed, permeabilized and washed cells wereincubated with antiserum (1 : 100) in PBS/BSA for 45 min at roomtemperature. After three 5 min washes in PBS, samples were stainedwith either Alexa488- or Alexa594-conjugated goat anti-rabbit IgGor goat anti-mouse IgG (Molecular Probes) diluted 1 : 100 in PBS/BSA for 45 min. Where appropriate, samples were simultaneouslystained with propidium iodide to stain bacteria and/or phalloidin tostain cytoskeletal actin- or rhodamine-conjugated WGA to stain thebrush border membrane. RBC samples were stained with rhodamine-conjugated WGA to stain the RBC membrane and with DAPI tostain bacteria. Preparations were washed a further three times in

Table 1. Strains and plasmids used in the study

Illustrated are strains, their bfpA espA eae genotypes, their phenotypes in HEp-2 and Caco-2 cell assays and their motility. HEp-2 cell

adhesion, 2 to +++++; LA, localized adherence; FAS, fluorescence actin staining test (+ indicates A/E lesion formation). Caco-2 cell

adhesion, meanSD [relative to wild-type E2348/69 (100 %)].

Strains Genotype Phenotype Motility Reference or source

bfpA espA eae HEp-2 cells Caco-2 cells

Adhesion LA FAS Adhesion LA FAS

E2348/69 + + + ++++ + + 1004 + + + Levine et al. (1978)

UMD880 + 2 2 ++++ + 2 1104 + 2 + This study

UMD886 2 + 2 + 2 2 81 2 2 +/2 This study

UMD883 2 2 + 2 2 2 00 2 2 + This study

UMD888 2 2 2 2 2 2 00 2 2 + This study

UMD901 2 + + + 2 + 71?5 2 + + Zhang & Donnenberg (1996)

UMD872 + 2 + ++++ + 2 724 + 2 + Kenny et al. (1996)

CVD206 + + 2 ++++ + 2 593?5 + 2 + Donnenberg & Kaper (1991)

Plasmids

pRPA100 Cloned bfpA Anantha et al. (2000)

pMSD2 Cloned espA Kenny et al. (1996)

pCVD438 Cloned eae Donnenberg & Kaper (1991)

http://mic.sgmjournals.org 529

EPEC adhesins

PBS and mounted in glycerol/PBS. HEp-2 and Caco-2 cell prepara-tions were examined by confocal microscopy using a Leica TCS SP2Spectral Confocal Microscope, equipped with an Argon (488 nm)and two Helium/Neon (543 nm, 633 nm) lasers and 406 or 636PL APO objectives. Confocal illustrations show either image sectionsor image projections through the brush border region of the cell.In triple-stained specimens using Alexa488, Alexa594 and propidiumiodide, the propidium iodide emission was artificially coloured blueto distinguish it from Alexa594 (red). RBC samples were examinedby conventional epifluorescence microscopy using a Leica DMRmicroscope equipped with a 636 PL APO objective and a LeicaDC200 digital camera.

SEM. Glutaraldehyde-fixed Caco-2 cell monolayers and intestinalmucosal biopsy tissue were post-fixed in 1 % osmium tetroxide,dehydrated through graded alcohol solutions and critical-point-dried.Mounted specimens were sputter-coated with platinum (Polaron)and examined in a Philips XL30 scanning electron microscope(Knutton, 1995).

Fig. 1. Horizontal xy (a) and vertical xz (b) confocal images ofuninfected Caco-2 cell monolayers stained for cellular actin(FITC, green) and a scanning electron micrograph (c). Caco-2cells grown on permeable filters have a well developed apicalbrush border surface. Bars, 5 mm.

Fig. 2. Confocal (ag) and SEM (h) images of Caco-2 cell monolayers infected with E2348/69 (bfpA+ espA+ eae+). Within10 min bacteria are seen adhering to the brush border (phalloidinFITC, green) as single bacteria (propidium iodide, red) (a);adherent bacteria express BFP (Alexa594, red), EspA filaments (Alexa488, green) and intimin (Alexa488, green, inset) (b).BFP fibrils (Alexa488, green) appear to promote attachment of bacteria (propidium iodide, coloured blue) to the brush border(phalloidinTRITC, red) and also bacteriabacteria aggregation (c). After 3 h, horizontal (d) and vertical (e) confocal imagesfollowing BFP (Alexa488, green), bacteria (propidium iodide, blue) and actin (phalloidinTRITC, red) staining revealedadherent three-dimensional microcolonies of bacteria connected by BFP. After 6 h, horizontal (f) and vertical (g) confocalimages following staining for bacteria (propidium iodide, red) and actin (phalloidinFITC, green) showed A/E lesion formation(actin accretion beneath bacteria) (f, g, arrows); bacterial microcolonies were no longer present. A/E lesion formation (h,arrow) and an absence of microcolonies was also seen by SEM (h). Bars, (ag) 5 mm; (h) 1 mm.



RESULTS

Characterization of mutants in HEp-2 cellassays

Prior to examining EPEC interaction with Caco-2 brushborder cells, the collection of mutants in bfpA, espA and eaewas tested in HEp-2 cell assays for extent of adhesion,pattern of adhesion (Nataro et al., 1987a) and fluore-scent actin staining (Knutton et al., 1989). Expressionof BFP, EspA filaments and intimin were assessed byimmunofluorescence. In each case the observed phenotypewas consistent with the genotype (Table 1).

Wild-type EPEC adhere to Caco-2 cells andform A/E lesions

Caco-2 cells grown on permeable filters for 12 days producea polarized sheet of columnar epithelial cells with a welldeveloped apical brush border surface that can be visualizedby confocal microscopy after staining the microvillousmembrane with WGA or the microvillous actin cytoskeletonwith phalloidin (Fig. 1a, b) or visualized by SEM (Fig. 1c).To examine initial EPEC attachment to Caco-2 cells, weexamined cell monolayers at early time points (10 min)when single bacteria were first seen bound to the brushborder surface (Fig. 2a). At this stage, fluorescence stainingof E2348/69 revealed expression of BFP fibrils and surfaceexpression of intimin; EspA filaments were expressed bysome but not all bacteria at this time point and they werefrequently shorter than mature EspA filaments seen atlater time points (Knutton et al., 1998). The longer andmore numerous BFP fibrils appeared to mediate bacterialattachment to the brush border surface (Fig. 2b). Actinaccretion, indicative of A/E lesion formation (Knuttonet al., 1989), was not observed at this stage. Adhesion ofindividual bacteria to Caco-2 cells was followed by bacterialaggregation giving rise first to small aggregates (Fig. 2c)

and later after 3 h to larger microcolonies (Fig. 2d) typicalof localized adherence seen with HEp-2 cells (Table 1).All bacteria within microcolonies expressed BFP and BFPappeared to mediate bacteriabacteria aggregation (Fig. 2d, e);actin accretion beneath bacteria in contact with the apicalcell surface indicated that they had formed A/E lesions(Fig. 2f, g). By 6 h only bacteria in contact with the cellsurface that had produced A/E lesions were observed,indicating that the three-dimensional bacterial micro-colonies had dispersed (Fig. 2g, h), a phenomenon wedemonstrated previously in E2348/69-infected HEp-2 cells(Knutton et al., 1999). Since BFP, EspA filaments andintimin were expressed when EPEC first adhered to Caco-2cells, we examined double deletion mutant strains UMD880,UMD886 and UMD883 (Table 1) to assess the role inadhesion of BFP, EspA filaments and intimin independently.

BFP mediates rapid adherence of EPEC toCaco-2 cells

Strain UMD880 (bfpA+ espA2 eae2) expressed BFP bundlestypical of the wild-type strain but there were no EspAfilaments and no surface intimin (Table 1). Strain UMD880adhered rapidly (10 min) to Caco-2 cells, initially as singlebacteria (Fig. 3a) and BFP fibrils appeared to promotebrush border attachment of bacteria (Fig. 3b, c). BFP-likefibrils connecting bacteria to brush border microvilli werealso seen by SEM (Fig. 3d). Subsequent bacterial aggrega-tion resulted after 3 h in typical localized microcolonieswith bacteria interlinked by BFP (Fig. 3e, f ). Strain UMD880did not induce actin accretion or A/E lesion formation butthe level of adhesion was comparable to that of the wild-type (Table 1). Interestingly, after 6 h UMD880 micro-colonies were still present and had not dispersed (data notshown), which suggests that the kinetics of dispersal aredifferent from those of the wild-type or that signals whichlead to dispersal are not produced by strain UMD880.

Fig. 3. Confocal (ac, e, f) and SEM (d) images of UMD880 (bfpA+ espA eae) infected Caco-2 cell monolayers. Within10 min individual bacteria (propidium iodide, red) adhere to the brush border (stained for actin, phalloidinFITC, green) (a).BFP fibrils (Alexa488, green) seen in horizontal (b, arrow) and vertical (c, arrow) confocal images appear to promote adhesionof UMD880 to the brush border (phalloidinTRITC, red). BFP-like fibrils connecting bacteria to brush border microvilli werealso seen by SEM (d, arrows). After 3 h, horizontal (e) and vertical (f) confocal images following BFP (Alexa488, green),bacteria (propidium iodide, coloured blue) and actin (phalloidinTRITC, red) staining revealed adherent three-dimensionalmicrocolonies of bacteria interlinked by BFP. Bars, (ac, e, f ) 5 mm; (d) 1 mm.


EPEC adhesins

Transformation of EspA (plasmid pMSD2) or intimin(plasmid pCVD438) into strain UMD880 did not alter theadherence characteristics of this strain.

EspA filaments mediate EPEC adherence toCaco-2 cells

Strain UMD886 (bfpA2 espA+ eae2) produced EspA fila-ments but lacked both BFP and surface intimin (Table 1).This strain adhered to Caco-2 cells but inefficiently com-pared to wild-type E2348/69 and UMD880. No adhesionwas detected after 10 min and only a low level of adhesionwas detected after 3 h (Fig. 4a, Table 1); bacteria adheredindividually and there was no aggregation into micro-colonies. We noted previously greatly improved EspAfilament-mediated bacterial adhesion to HEp-2 cells usinga gentle washing procedure to remove non-adherentbacteria following infection (Knutton et al., 1998). In asimilar manner, using the gentle washing procedure,adhesion of UMD886 to Caco-2 cells was greatly improved(Fig. 4b) and was ~10-fold greater than when using themore vigorous washing procedure. In each case EspAfilaments appeared to mediate brush border adhesion ofstrain UMD886 (Fig. 4c, d). Transformation of UMD886with pRPA100 restored BFP expression, localized adhesion,microcolony formation and good adhesion; transformationwith pCVD438 did not affect the overall level of adherencebut conferred A/E lesion formation (data not shown).

Intimin promotes Caco-2 cell adherence ofEPEC when Tir is present in the cell membrane

Strain UMD883 (bfpA2 espA2 eae+) did not produce BFPor EspA filaments but did express surface intimin (Table 1).This strain did not adhere to Caco-2 cells even after 6 h,irrespective of the washing procedure used (Table 1).Transformation of UMD883 with pRPA100 restored BFPexpression and microcolony formation whilst transforma-tion with pMSD2 restored adherence comparable to strainUMD886 and A/E lesion formation (data not shown).

Intimin is known to bind Tir following its translocationand insertion into the host cell membrane; strain UMD883lacks EspA filaments and thus a functional TTSS and sois unable to translocate Tir. To assess the adhesive pro-perties of UMD883 to cells possessing translocated Tir, weperformed co-culture infections with strain CVD206 orUMD886, which lack the intimin gene but which possess afunctional TTSS and can translocate Tir. In co-culture withCVD206 (Fig. 5) or UMD886, strain UMD883, identifiedby surface intimin staining, was able to adhere intimatelyto Caco-2 cells and produced typical A/E lesions. However,the extent of UMD883 adhesion and A/E lesion forma-tion was much more pronounced when co-cultured withCVD206 compared to UMD886, presumably due to the

Fig. 5. Confocal (ac) and SEM (d) images of Caco-2 cellsco-infected with UMD883 (bfpA espA eae+) and CVD206(bfpA+ espA+ eae) for 3 h and stained only for UMD883bacteria (intimin staining). In horizontal (a) and vertical (b) con-focal images of cells stained for actin (phalloidinFITC, green),UMD883 bacteria (Alexa594, red) were now adherent and pro-duced A/E lesions (actin accretion beneath bacteria) (b, arrow).Focused Tir (Alexa488, green) was also seen beneath intimatelyattached UMD883 bacteria (Alexa594, red) (c). A/E lesionsproduced by UMD883 (d, arrow) were also seen by SEM.Bars, 1 mm.

Fig. 4. Confocal images of UMD886 (bfpA espA+ eae) infected Caco-2 cell monolayers. UMD886 (propidium iodide, red)adhered poorly to the Caco-2 cell brush border (WGAFITC, green) even after 3 h (a). However, an~10-fold increase in thelevel of UMD886 adhesion was seen following a gentle wash procedure (b). In each case EspA filaments (Alexa488, green)seen in horizontal (c, arrow) and vertical (d, arrow) confocal images appeared to promote adhesion of bacteria (propidiumiodide, coloured blue) to the cell brush border (phalloidinTRITC, red). Bars, 5 mm.



much more efficient adhesion of CVD206 (Table 1).Focused Tir translocated by CVD206 (Fig. 5c) or UMD886was also detected beneath intimately attached UMD883bacteria.

Adhesion of single mutants in bfpA, espAand eae

To assess which might be the predominant EPEC adhesin,we compared adhesion of the double mutants just des-cribed with adhesion of single mutants in bfpA, espA andeae. Strain CVD206 (bfpA+ espA+ eae2) produced BFP andEspA filaments but lacked surface intimin; strain UMD872(bfpA+ espA2 eae+) produced BFP and expressed surfaceintimin but lacked EspA filaments; strain UMD901 (bfpA2

espA+ eae+) produced EspA filaments and expressed sur-face intimin but lacked BFP (Table 1). Strains CVD206and UMD872 exhibited Caco-2 cell adherence characteristicof UMD880, i.e. rapid (10 min) bacterial adhesion(Fig. 6a, b) followed by microcolony formation but therewas no A/E lesion formation. Adherence of strain UMD901

was characteristic of strain UMD886, i.e. no adhesion after10 min and poor adhesion after 3 h. However, unlike strainUMD886, adherent UMD901 bacteria do possess a func-tional TTSS and were able to translocate Tir and produceA/E lesions on Caco-2 cells after 6 h (Fig. 6c). As was thecase for UMD886, a mild washing procedure to removenon-adherent bacteria significantly increased the adherenceof strain UMD901.

Adhesion of a triple mutant in bfpA, espAand eae

Strain UMD888 (bfpA2 espA2 eae2) lacked BFP, EspAfilaments and intimin and was non-adherent to Caco-2cells after 6 h irrespective of which washing procedurewas used (Fig. 7d, Table 1). Transformation with plasmidpRPA100 restored adherence and microcolony formationtypical of UMD880; transformation with pMSD2 conferreda weak, non-localized cell adherence to cells typical ofUMD886; UMD888 transformed with pCVD438 remainednon-adherent (data not shown).

Fig. 7. Confocal images showing Caco-2 cell monolayers infected with EPEC strain E2348/69 (bfpA+ espA+ eae+) for10 min (a), 3 h (b) and 6 h (c) and with strain UMD888 (bfpA espA eae) for 6 h (d). Bacteria were stained with propidiumiodide (red) and flagella with an anti-flagellin antibody (Alexa488, green). All bacteria attached after 10 min possessedflagella (a) but with increasing time fewer and fewer flagellate bacteria were present (b, c). After 6 h, strain UMD888 wasnon-adherent (d) but flagellate bacteria were present in culture supernatants (d, inset). Bars, 5 mm.

Fig. 6. Confocal images showing Caco-2 cell monolayers infected with single mutant strains CVD206 (bfpA+ espA+ eae)(a), UMD872 (bfpA+ espA eae+) (b) and UMD901 (bfpA espA+ eae+) (c). Caco-2 cells were stained for actin(phalloidinFITC, green) and bacteria stained with propidium iodide (red). CVD206 adhered rapidly (10 min) to Caco-2 cells(a) and adherent bacteria expressed BFP (Alexa594, red) and EspA filaments (Alexa488, green) (a, inset). UMD872 similarlyinduced rapid Caco-2 cell adhesion (b) and adherent UMD872 produced BFP fibrils (Alexa594, red) but no EspA filaments(b, inset). By 6 h, some UMD901 bacteria (propidium iodide, red) had formed A/E lesions (c, arrow). Other adherent bacteriahad not induced actin accretion (c, arrowhead) and in such bacteria EspA filaments (Alexa488, green) seen in verticalconfocal images appeared to mediate brush border (phalloidinTRITC, red) adhesion of bacteria (c, inset). Bar, 5 mm; inset 1 mm.


EPEC adhesins

Flagella and EPEC adhesion to Caco-2 cells

Flagella were recently implicated in EPEC adhesion toepithelial cells (Giron et al., 2002). Since strain E2348/69produced flagella which might be contributing to adhesion,we examined motility and flagella expression of all thestrains; all the strains were motile when grown in LB and inthe DMEM used for cell adhesion assays (Table 1) and ineach strain flagella expression was detected by immuno-fluorescence using an anti-flagellin antibody.

Flagella expression was also examined during cell adhesion.Virtually all E2348/69 bacteria attaching initially to Caco-2cells expressed flagella (Fig. 7a) but as bacteria formedmicrocolonies and produced A/E lesions (3 h) the percen-tage of flagellated bacteria was greatly reduced (Fig. 7b);after 6 h few flagella were observed (Fig. 7c). A similardecrease in the number of flagellate bacteria followinginitial attachment was seen with the other adherentstrains (UMD880, UMD872, UMD901, CVD206). StrainsUMD883 and UMD888 (Fig. 7d) were non-adherent toCaco-2 cells but the presence of flagellate bacteria wasconfirmed by immunofluorescence of bacteria present inthe culture medium (Fig. 7d, inset).

EPEC adhesion to RBC monolayers

In recent studies we demonstrated EPEC adhesion to RBCmonolayers, initially by EspA filaments and subsequentlywith strains able to translocate Tir, by intiminTir intimate

interaction (Shaw et al., 2001, 2002). In this study we alsoscreened the collection of mutants in bfpA, espA and eaefor their ability to adhere to RBC monolayers. As reportedpreviously, wild-type E2348/69 adhered to RBC monolayers(Fig. 8a) and induced a high level (~80 %) of haemolysis(Shaw et al., 2001). Similar levels of RBC adhesion (andhaemolysis) were produced by strains UMD886 (Fig. 8d),UMD901 (Fig. 8f ) and CVD206 (Fig. 8h), whereas strainsUMD880 (Fig. 8b), UMD883 (Fig. 8c), UMD888 (Fig. 8e)and UMD872 (Fig. 8g) did not adhere to RBC monolayersand did not induce haemolysis. Thus, adhesion to RBCswas independent of BFP expression but correlated withexpression of EspA filaments; under the assay conditionsused, adhesion of the four positive strains was shown to bemediated by EspA filaments (data not shown).

EPEC adhesion to human intestinal mucosa

Studies using human intestinal organ culture suggested thatBFP was not involved in initial EPEC adherence but onlyat a later stage to promote microcolony formation (Hickset al., 1998). We showed previously that wild-type EPECstrain E2348/69 will colonize the surface of cultured humanintestinal mucosa and produce typical A/E lesions after a68 h incubation (Knutton et al., 1987). However, in thesestudies we did not attempt to examine early stages ofadhesion. Here we examined early stages of E2348/69adhesion to human intestinal mucosa (Fig. 9b) in com-parison with CVD206 (bfpA+ espA+ eae2), which cannotproduce A/E lesions but which does produce BFP and

Fig. 8. Fluorescence images showing EPEC adhesion to RBC monolayers. Bacteria were stained with DAPI (blue) and theRBC membrane with WGATRITC (red). Strains UMD886 (bfpA espA+ eae) (d), UMD901 (bfpA espA+ eae+) (f) andCVD206 (bfpA+ espA+ eae) (h) adhered to RBC monolayers in similar numbers to wild-type E2348/69 (bfpA+ espA+

eae+) (a), whereas strains UMD880 (bfpA+ espA eae) (b), UMD883 (bfpA espA eae+) (c), UMD888 (bfpA espA

eae) (e) and UMD872 (bfpA+ espA eae+) (g) did not adhere. Bar, 5 mm.



EspA filaments and may therefore be able to undergoinitial mucosal attachment if BFP and/or EspA filamentsare adhesins (Fig. 9c). After a shorter 1?5 h incubationwith E2348/69, SEM revealed adhesion of bacterial micro-colonies to mucosal cells in which brush border microvillihad become perturbed and elongated. Some bacteriaappeared deeply embedded in the brush border (Fig. 9b),typical of bacteria that have formed intimate attachmentand A/E lesions. CVD206 adhered in similar small micro-colonies and induced elongation of brush border microvillibut did not appear to be intimately attached to the mucosalsurface (Fig. 9c). Uninfected control tissue had a uniformlysmooth appearance and lacked any adherent bacteria(Fig. 9a).

DISCUSSION

Intestinal colonization and A/E lesion formation by EPECis a complex multi-stage process and several different wellcharacterized and putative adhesive factors involved in thisprocess have been described including BFP, EspA filamentsand intimin. Observation of the interaction with Caco-2cells of wild-type EPEC strain E2348/69 and a set of definedbfpA, espA and eae mutants and transformants supportsa role for BFP and EspA filaments in initial brush bordercell attachment of EPEC whereas intimin was required forintimate EPEC adhesion and A/E lesion formation provid-ing that Tir had been translocated to the host cell.

Several studies have implicated BFP in initial binding ofEPEC to host cells. The initial studies of Giron et al. (1991)demonstrated a significant reduction in localized adher-ence of EPEC in the presence of BFP antibodies; residualadherence can now be explained by EspA filament andintiminTir adherence. Tobe & Sasakawa (2001) implicatedBFP in host cell attachment by demonstrating preferentialbinding of BFP, producing EPEC to the Caco-2 cell surfacerather than to existing EPEC microcolonies formed during

an earlier infection. These authors also concluded thatdifferential EPEC adherence to cells of different speciesorigin was due to BFP (Tobe & Sasakawa, 2002). Thispresent study, which examined EPEC strains at an earlystage when they first adhere to cells, also supports such acell adhesive role for BFP. Early rapid adhesion of indivi-dual bacteria correlated with the expression of BFP and,by immunofluorescence, BFP fibrils appeared to connectbacteria to the brush border surface. Subsequently, indivi-dual bacteria aggregated to produce small and then largermicrocolonies with BFP appearing to link bacteria in theaggregates. Thus, these morphological data also support adirect role for BFP in initial cell attachment in additionto their recognized role in microcolony formation anddispersion. An adhesive role for BFP implies a specific hostcell BFP receptor. Numerous candidate BFP receptors havebeen proposed, including a variety of oligosaccharides and,most recently, the phospholipid phosphatidylethanolamine(PE) (Nougayrede et al., 2003). This study also confirmedour previous conclusion that EPEC adhesion to RBCs isindependent of BFP (Shaw et al., 2001) and indicates a BFPreceptor that is lacking on the RBC surface. A PE receptorwould be consistent with these RBC data because in theerythrocyte membrane PE is predominantly localized on thecytosolic side of the membrane.

A study by Hicks et al. (1998) employing in vitro intestinalorgan culture to examine EPEC adhesion to human smallintestinal mucosa concluded that BFP was not involved ininitial EPEC adherence but only at a later stage to promotemicrocolony formation. This conclusion was based onthe fact that strain JPN15, which lacks the large plasmidencoding BFP but expresses EspA and intimin, adheredto cultured paediatric small intestinal mucosa in two-dimensional microcolonies and produced A/E lesionswhereas strain CVD206 (bfpA+ espA+ eae2) did notadhere. We also demonstrated previously A/E adhesion ofa plasmid-cured derivative of E2348/69 to human intestinal

Fig. 9. SEM images of uninfected small intestinal mucosa (a) and following a 1?5 h infection with primed E2348/69 (bfpA+

espA+ eae+) (b) and CVD206 (bfpA+ espA+ eae) (c) bacteria. Uninfected mucosa had a smooth mucosal surface and noadherent bacteria were present (a). E2348/69 adhered in microcolonies to mucosal cells in which the brush border had beenperturbed and microvilli elongated. Some bacteria appeared deeply embedded in the brush border, typical of those that formintimate attachment and A/E lesions (b). CVD206 similarly adhered in microcolonies and induced elongation of brush bordermicrovilli but did not appear intimately attached to the mucosal surface (c). Bars, (a) 10 mm; (b, c) 5 mm.


EPEC adhesins

mucosa and noted that the extent of adhesion was highlyattenuated compared to that of the wild-type (Knuttonet al., 1987). This could now be explained by a lack ofBFP-mediated initial attachment with initial attachmentof the plasmid-cured strain being promoted by the muchless efficient EspA filaments. In this study we examined earlystages of EPEC adhesion to human intestinal mucosa and,in contrast to Hicks et al. (1998), we did observe adhesionof strain CVD206. CVD206, unlike E2348/69, lacks intiminand so cannot form an intiminTir intimate attachmentand A/E lesions but it does express BFP and EspA fila-ments which could be promoting non-intimate adhesion.Although the CVD206 adhesins involved were not identi-fied, the observation that CVD206 did adhere does castdoubt on the conclusions of Hicks et al. that BFP does notpromote initial mucosal attachment of EPEC. Failure todetect mucosal adhesion of CVD206 by Hicks et al. mayhave been due to differences in the adhesion assay protocolsused and/or to poor BFP expression by the CVD206 strainused. It was particularly noticeable in the study by Hickset al. that in a 3 h HEp-2 cell adhesion assay they onlydemonstrated adhesion of individual CVD206 bacteria tothe HEp-2 cell surface, whereas others had observed adhe-sion of large bacterial microcolonies by 3 h (Donnenberg& Kaper, 1991; Knutton et al., 1999).

EspA filaments are a component of the EPEC TTSS whoseprimary function is the delivery of virulence proteins intohost cells (Daniell et al., 2001). However, since EspAfilaments interact with host cells during the early stages ofA/E lesion formation, it has been proposed that they mayalso function as adhesins. We showed previously that EspAfilaments promoted attachment of strains lacking BFP toepithelial cells and to RBCs (Knutton et al., 1998; Shawet al., 2001) and similar results supporting a role of EspAfilaments as initial attachment factors were demonstratedwith Shiga toxin-producing E. coli (Ebel et al., 1998) whichalso lacks BFP. This present study also supports an adhesiverole for EspA filaments in EPEC adhesion to brush bordercells since adhesion correlated with EspA filament expres-sion and, by immunofluorescence, EspA filaments appearedto connect bacteria to the brush border surface. However,adhesion was much less efficient than with strains express-ing BFP. The extent of EspA-filament-mediated adhesionwas dependent on the washing procedure used, suggestingthat EspA-filament-mediated adhesion is weak comparedto BFP and intiminTir-mediated adhesion where thewashing procedure had no effect on levels of adhesion.Compared to BFP, the less efficient EspA-filament-mediatedadhesion probably reflects the small number of EspAfilaments produced (~12 EspA filaments per bacterium)(Daniell et al., 2001) and the nature of their interactionwith host cells which is currently unknown. EPEC havebeen divided into typical EPEC which possess EAF plasmidsand BFP and atypical EPEC which lack EAF plasmids andBFP (Kaper, 1996). The use of EspA filaments as initialattachment factors could explain why atypical EPEC isstill able to colonize the gut and produce A/E lesions but,

due to a lack of BFP, is a much less efficient colonizer thantypical EPEC (Levine et al., 1985).

Intimin expressed on the bacterial surface binds Tirfollowing its translocation and insertion into the host cellmembrane; intiminTir interaction produces the charac-teristic intimate EPEC attachment and triggers subsequentA/E lesion formation (Campellone & Leong, 2003; Frankelet al., 1998). Strain UMD883 expressing only intimin didnot adhere to Caco-2 cells although it was able to bind andform an intimate attachment and A/E lesions followingtranslocation of Tir into the Caco-2 cell membrane. Theadhesive role of intimin binding to translocated Tir iswell established (Kenny et al., 1997). However, there is alsoa considerable body of evidence for a host cell intiminreceptor, not least of which is the binding of the cell-bindingdomain of intimin to epithelial cells in the absence of Tir(Frankel et al., 2001) and the recent demonstration, in thecase of E. coli O157 : H7, of a host cell surface protein,nucleolin, which specifically binds intimin c expressed bythis serotype (Sinclair & OBrien, 2002). Although avail-able evidence supports the presence of host cell intiminreceptors, the inability of strain UMD883 to adhere to Caco-2 and HEp-2 cells in the absence of Tir clearly indicates thelack of a receptor able to support bacterial adhesion.

The present study examined three particular EPEC adhesivefactors. However, other putative EPEC adhesive factorshave been proposed, including other fimbrial antigens(Giron et al., 1993), Efa1/LifA (Badea et al., 2003) andflagella (Giron et al., 2002). Other than type 1 fimbriae,which have been demonstrated not to be involved inepithelial cell adhesion, strain E2348/69 has not beenreported to produce other recognized fimbriae (Elliott& Kaper, 1997). Efa1/LifA, a protein unique to EPEC(Klapproth et al., 2000) and other A/E pathogens, appearsto have cell-binding activity and might contribute toadhesion in some manner. However, the lack of adhesionof UMD888 (bfpA2 espA2 eae2) implies the absence ofother adhesive factors sufficient to promote adhesion ofEPEC to intestinal brush border cells. A direct role forflagella in EPEC adhesion was suggested recently (Gironet al., 2002). We thought previously that adherent EPECdid not express flagella because we did not see flagella bySEM. However, it is now clear from the immunofluores-cence staining performed as part of this study that adherentEPEC do possess flagella which could therefore play arole in adhesion. Nevertheless, this study found no directevidence for an adhesive role for flagella since strainsUMD883 (bfpA2 espA2 eae+) and UMD888 (bfpA2 espA2

eae2) produced flagella and yet were totally non-adherent.However, it is clear that EPEC flagella can interact withhost cells and stimulate host immune responses (Zhouet al., 2003) and thus a contribution to adhesion cannot beruled out although this study indicates that flagella, in theabsence of other adhesins, cannot by themselves supportbacterial adhesion.

In summary, the results of this study, using defined EPEC



E2348/69 mutants and Caco-2 brush border cells, areconsistent with data obtained previously using undiffer-entiated epithelial cells of non-intestinal origin and indicatethat BFP, EspA filaments and, in a Tir-dependent manner,intimin can each support bacterial adhesion to intestinalepithelial cells; no evidence was found for other adhesinswhich could support EPEC adhesion. BFP appeared tobe the predominant initial attachment factor but, in theabsence of BFP, EspA filaments were able to promote a lessefficient initial bacterial attachment. Intimin functioned asan adhesin in the presence of translocated Tir but we foundno evidence for an independent host cell intimin receptorcapable of promoting EPEC adhesion.

ACKNOWLEDGEMENTS

We thank Ilan Rosenshine for the H6 flagella antiserum and theWellcome Trust for financial support. M. S. D. and L.-C. L. weresupported by Public Health Service Awards AI37606 and AI32074 fromthe National Institutes of Health.

REFERENCES

Anantha, R. P., Stone, K. D. & Donnenberg, M. S. (2000). Effectsof bfp mutations on biogenesis of functional enteropathogenicEscherichia coli type IV pili. J Bacteriol 182, 24982506.

Badea, L., Doughty, S., Nicholls, L., Sloan, J., Robins-Browne, R. M.& Hartland, E. L. (2003). Contribution of Efa1/LifA to the adherenceof enteropathogenic Escherichia coli to epithelial cells. Microb Pathog34, 205215.

Batchelor, M., Knutton, S., Caprioli, A., Huter, V., Zanial, M.,Dougan, G. & Frankel, G. (1999). Development of a universalintimin antiserum and PCR primers. J Clin Microbiol 37, 38223827.

Bieber, D., Ramer, S. W., Wu, C. Y., Murray, W. J., Tobe, T.,Fernandez, R. & Schoolnik, G. K. (1998). Type IV pili, transientbacterial aggregates, and virulence of enteropathogenic Escherichiacoli. Science 280, 21142118.

Campellone, K. G. & Leong, J. M. (2003). Tails of two Tirs: actinpedestal formation by enteropathogenic E. coli and enterohemor-rhagic E. coli O157 : H7. Curr Opin Microbiol 6, 8290.

Collington, G. K., Booth, I. W. & Knutton, S. (1998). Rapidmodulation of electrolyte transport in Caco-2 cell monolayers byenteropathogenic Escherichia coli (EPEC) infection. Gut 42, 200207.

Cravioto, A., Gross, R. J., Scotland, S. M. & Rowe, B. (1979).An adhesive factor found in strains of Escherichia belonging to thetraditional infantile enteropathogenic serotype. Curr Microbiol 3,9599.

Daniell, S. J., Takahashi, N., Wilson, R. & 7 other authors (2001).The filamentous type III secretion translocon of enteropathogenicEscherichia coli. Cell Microbiol 3, 865871.

Donnenberg, M. S. & Kaper, J. B. (1991). Construction of an eaedeletion mutant of enteropathogenic Escherichia coli by using apositive-selection suicide vector. Infect Immun 59, 43104317.

Donnenberg, M. S., Giron, J. A., Nataro, J. P. & Kaper, J. B. (1992).A plasmid-encoded type IV fimbrial gene of enteropathogenicEscherichia coli associated with localized adherence. Mol Microbiol 6,34273437.

Ebel, F., Podzadel, T., Rohde, M., Kresse, A. U., Kramer, S.,Deibel, C., Guzman, C. A. & Chakraborty, T. (1998). Initial bindingof Shiga toxin-producing Escherichia coli to host cells and subsequent

induction of actin rearrangements depend on filamentous EspA-

containing surface appendages. Mol Microbiol 30, 147161.

Elliott, S. J. & Kaper, J. B. (1997). Role of type 1 fimbriae in EPECinfections. Microb Pathog 23, 113118.

Frankel, G., Phillips, A. D., Rosenshine, I., Dougan, G., Kaper, J. B.& Knutton, S. (1998). Enteropathogenic and enterohaemorrhagic

Escherichia coli: more subversive elements. Mol Microbiol 30, 911921.

Frankel, G., Phillips, A. D., Trabulsi, L. R., Knutton, S., Dougan, G. &Matthews, S. (2001). Intimin and the host cell is it bound to end

in Tir(s)? Trends Microbiol 9, 214218.

Giron, J. A., Ho, A. S. & Schoolnik, G. K. (1991). An induciblebundle-forming pilus of enteropathogenic Escherichia coli. Science

254, 710713.

Giron, J. A., Ho, A. S. & Schoolnik, G. K. (1993). Characterization of

fimbriae produced by enteropathogenic Escherichia coli. J Bacteriol

175, 73917403.

Giron, J. A., Torres, A. G., Freer, E. & Kaper, J. B. (2002). The flagella

of enteropathogenic Escherichia coli mediate adherence to epithelial

cells. Mol Microbiol 44, 361379.

Hartland, E. L., Batchelor, M., Delahay, R. M., Hale, C., Matthews, S.,

Dougan, G., Knutton, S., Connerton, I. & Frankel, G. (1999). Bindingof intimin from enteropathogenic Escherichia coli to Tir and to host

cells. Mol Microbiol 32, 151158.

Hicks, S., Frankel, G., Kaper, J. B., Dougan, G. & Phillips, A. D.

(1998). Role of intimin and bundle-forming pili in enteropathogenicEscherichia coli adhesion to pediatric intestinal tissue in vitro. Infect

Immun 66, 15701578.

Kaper, J. B. (1996). Defining EPEC. Rev Microbiol Sao Paulo 27,130133.

Kenny, B., Lai, L. C., Finlay, B. B. & Donnenberg, M. S. (1996). EspA,

a protein secreted by enteropathogenic Escherichia coli, is required to

induce signals in epithelial cells. Mol Microbiol 20, 313323.

Kenny, B., DeVinney, R., Stein, M., Reinscheid, D. J., Frey, E. A. &Finlay, B. B. (1997). Enteropathogenic E. coli (EPEC) transfers its

receptor for intimate adherence into mammalian cells. Cell 91, 511520.

Klapproth, J. M., Scaletsky, I. C., McNamara, B. P., Lai, L. C.,Malstrom, C., James, S. P. & Donnenberg, M. S. (2000). A

large toxin from pathogenic Escherichia coli strains that inhibits

lymphocyte activation. Infect Immun 68, 21482155.

Knutton, S. (1995). Electron microscopical methods in adhesion.Methods Enzymol 253, 145158.

Knutton, S., Lloyd, D. R. & McNeish, A. S. (1987). Adhesion of

enteropathogenic Escherichia coli to human intestinal enterocytes and

cultured human intestinal mucosa. Infect Immun 55, 6977.

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

enterohemorrhagic Escherichia coli. Infect Immun 57, 12901298.

Knutton, S., Rosenshine, I., Pallen, M. J., Nisan, I., Neves, B. C.,

Bain, C., Wolff, C., Dougan, G. & Frankel, G. (1998). A novel EspA-associated surface organelle of enteropathogenic Escherichia coli

involved in protein translocation into epithelial cells. EMBO J 17,

21662176.

Knutton, S., Shaw, R. K., Anantha, R. P., Donnenberg, M. S. &

Zorgani, A. A. (1999). The type IV bundle-forming pilus ofenteropathogenic Escherichia coli undergoes dramatic alterations in

structure associated with bacterial adherence, aggregation and

dispersal. Mol Microbiol 33, 499509.

Knutton, S., Shaw, R. & Frankel, G. (2002). Interaction of

enteropathogenic Escherichia coli with red blood cell monolayers.

Methods Enzymol 358, 350355.


EPEC adhesins

Levine, M. M., Berquist, E. J., Nalin, D. R., Waterman, D. H., Hornick,R. B., Young, C. R., Stoman, S. & Rowe, B. (1978). Escherichia colithat cause diarrhoea but do not produce heat-labile or heat-stableenterotoxins and are non-invasive. Lancet 1 (8074), 119122.

Levine, M. M., Nataro, J. P., Karch, H., Baldini, M. M., Kaper, J. B.,Black, R. E., Clements, M. L. & OBrien, A. D. (1985). The diarrhealresponse of humans to some classic serotypes of enteropathogenicEscherichia coli is dependent on a plasmid encoding an entero-adhesiveness factor. J Infect Dis 152, 550559.

McDaniel, T. K., Jarvis, K. G., Donnenberg, M. S. & Kaper, J. B.(1995). A genetic locus of enterocyte effacement conserved amongdiverse enterobacterial pathogens. Proc Natl Acad Sci U S A 92,16641668.

Moon, H. W., Whipp, S. C., Argenzio, R. A., Levine, M. M. &Giannella, R. A. (1983). Attaching and effacing activities of rabbitand human enteropathogenic Escherichia coli in pig and rabbitintestines. Infect Immun 41, 13401351.

Nataro, J. P. & Kaper, J. B. (1998). Diarrheagenic Escherichia coli.Clin Microbiol Rev 11, 142201.

Nataro, J. P., Kaper, J. B., Robins-Browne, R., Prado, V., Vial, P. &Levine, M. M. (1987a). Patterns of adherence of diarrheagenicEscherichia coli to HEp-2 cells. Pediatr Infect Dis J 6, 829831.

Nataro, J. P., Maher, K. O., Mackie, P. & Kaper, J. B. (1987b).Characterization of plasmids encoding the adherence factor ofenteropathogenic Escherichia coli. Infect Immun 55, 23702377.

Nougayrede, J. P., Fernandes, P. J. & Donnenberg, M. S. (2003).Adhesion of enteropathogenic Escherichia coli to host cells. CellMicrobiol 5, 359372.

Sekiya, K., Ohishi, M., Ogino, T., Tamano, K., Sasakawa, C. & Abe, A.(2001). Supermolecular structure of the enteropathogenic Escherichia

coli type III secretion system and its direct interaction with the EspA-sheath-like structure. Proc Natl Acad Sci U S A 98, 1163811643.

Shaw, R. K., Daniell, S., Ebel, F., Frankel, G. & Knutton, S. (2001).EspA-filament-mediated protein translocation into red blood cells.Cell Microbiol 3, 213222.

Shaw, R. K., Daniell, S., Frankel, G. & Knutton, S. (2002). Entero-pathogenic Escherichia coli translocate functional Tir and form anintimin-Tir intimate attachment to red blood cell membranes.Microbiology 148, 13551365.

Sinclair, H. B. & OBrien, A. D. (2002). Cell-surface localizednucleolin is a eukaryotic receptor for the adhesin intimin-gamma ofenterohemorrhagic Escherichia coli O157 : H7. J Biol Chem 277,28762885.

Tobe, T. & Sasakawa, C. (2001). Role of bundle-forming pilus ofenteropathogenic Escherichia coli in host cell adherence and inmicrocolony development. Cell Microbiol 3, 579585.

Tobe, T. & Sasakawa, C. (2002). Species-specific cell adhesion ofenteropathogenic Escherichia coli is mediated by type IV bundle-forming pili. Cell Microbiol 4, 2942.

Yona-Nadler, C., Umanski, T., Aizawa, S.-I., Friedberg, D. &Rosenshine, I. (2003). Integration host factor (IHF) mediatesrepression of flagella in enteropathogenic and enterohaemorrhagicEscherichia coli. Microbiology 149, 877884.

Zhang, H. Z. & Donnenberg, M. S. (1996). DsbA is required forstability of the type IV pilin of enteropathogenic Escherichia coli.Mol Microbiol 21, 787797.

Zhou, X., Giron, J. A., Torres, A. G., Crawford, J. A., Negrete, E.,Vogel, S. N. & Kaper, J. B. (2003). Flagellin of enteropathogenicEscherichia coli stimulates interleukin-8 production in T84 cells.Infect Immun 71, 21202129.



Epe e Coli527.Full

Documents

Transcript of Epe e Coli527.Full