INFECTION AND IMMUNITY, Oct. 2008, p. 4669–4676 Vol. 76, No. 100019-9567/08/$08.00�0 doi:10.1128/IAI.00140-08Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Cortactin Recruitment by Enterohemorrhagic Escherichia coli O157:H7during Infection In Vitro and Ex Vivo�

Aurelie Mousnier,1† Andrew D. Whale,1† Stephanie Schuller,2 John M. Leong,3Alan D. Phillips,2 and Gad Frankel1*

Division of Cell and Molecular Biology, Imperial College London,1 and Centre for Paediatric Gastroenterology, Royal Free andUniversity College Medical School,2 London, United Kingdom, and Department of Molecular Genetics and Microbiology,

University of Massachusetts Medical School, Worcester, Massachusetts3

Received 31 January 2008/Returned for modification 7 March 2008/Accepted 21 July 2008

Enterohemorrhagic Escherichia coli (EHEC) O157:H7 is an important human pathogen that colonizes the gutmucosa via attaching and effacing (A/E) lesions; A/E lesion formation in vivo and ex vivo is dependent on the typeIII secretion system (T3SS) effector Tir. Infection of cultured cells by EHEC leads to induction of localized actinpolymerization, which is dependent on Tir and a second T3SS effector protein, TccP, also known as EspFU. Recently,cortactin was shown to bind both the N terminus of Tir and TccP via its SH3 domain and to play a role inEHEC-triggered actin polymerization in vitro. In this study, we investigated the recruitment of cortactin to the siteof EHEC adhesion during infection of in vitro-cultured cells and mucosal surfaces ex vivo (using human terminalileal in vitro organ cultures [IVOC]). We have shown that cortactin is recruited to the site of EHEC adhesion in vitrodownstream of TccP and N-WASP. Deletion of the entire N terminus of Tir or replacing the N-terminal polyprolineregion with alanines did not abrogate actin polymerization or cortactin recruitment. In contrast, recruitment ofcortactin to the site of EHEC adhesion in IVOC is TccP independent. These results imply that cortactin is recruitedto the site of EHEC adhesion in vitro and ex vivo by different mechanisms and suggest that cortactin might havea role during EHEC infection of mucosal surfaces.

Enterohemorrhagic Escherichia coli (EHEC), particularlyserotype O157:H7, is an important human pathogen that ad-heres closely to the gut epithelium and triggers localized ef-facement of the brush border microvilli, a phenomenon knownas attaching and effacing (A/E) (19). A/E lesion formationfollows translocation of effector proteins, notably Tir/EspE (9,18), through a type III secretion system (T3SS) (reviewed inreference 12). Following translocation, Tir is integrated intothe enterocyte plasma membrane in a hairpin topology (15);the surface-exposed loop functions as the receptor for theouter membrane bacterial adhesin intimin (reviewed in refer-ence 11).

During infection of epithelial cells in culture, EHEC triggerslocalized actin polymerization at the site of bacterial attach-ment (reviewed in references 5, 10, and 16), an activity com-pletely dependent on Tir. EHEC also encodes a second T3SSbacterial effector protein, TccP, also known as EspFU, whichbinds and activates the neuronal Wiskott-Aldrich syndromeprotein (N-WASP), leading to actin polymerization via theactin-related protein 2/3 (Arp2/3) complex (6, 13). Tir andTccP are necessary and sufficient for robust actin polymeriza-tion in vitro because transfection of Tir and TccP and artificialclustering of Tir by nonpathogenic E. coli K-12 expressingintimin or a T3SS EHEC mutant triggers actin polymerization(K. Campellone, unpublished data). To date, no direct inter-action between Tir and TccP has been reported. It is therefore

plausible that an adaptor protein encoded by the mammaliancell links the two bacterial proteins. Recruitment of TccP toTirEHEC is mediated by a C-terminal tripeptide NPY458 motif(3), suggesting that this sequence could be a target of a puta-tive host adaptor.

Although robust pedestal formation in vitro requires TccP,an EHEC O157:H7 tccP mutant can trigger inefficient actinpolymerization on cultured monolayers (4), suggesting thatEHEC triggers multiple pathways of actin assembly in hostcells. Consistent with this, EHEC O157:H7 tccP mutants caninduce A/E lesions on ex vivo human intestinal organ cultures(13) and in animal models (22, 26). These lesions do notappear dramatically different from A/E lesions formed by wild-type EHEC, and it is possible that alternative pathways leadingto A/E lesion formation function at a higher level during in-fection of primary intestinal epithelium than during infectionof immortalized cell lines.

The actin pedestals of EHEC O157:H7 are complex struc-tures that are enriched with focal adhesion (e.g., �-actinin andtalin), cytoskeletal (e.g., actin and cytokeratins 8 and 18), en-docytic (e.g., dynamin 2), and actin assembly-regulating (e.g.,N-WASP, Arp2/3, and cortactin) proteins (10, 14, 16). Cortac-tin is an F-actin binding protein and the substrate of numerouscellular kinases (reviewed in reference 25). Cortactin canweakly stimulate Arp2/3 directly or more potently via interac-tion with N-WASP. The activation of N-WASP is dependenton the SH3 domain of cortactin and is regulated by phosphor-ylation of serine and tyrosine residues in a proline-rich domainlocated upstream of the SH3 domain itself (25).

Given the propensity of cortactin to be present at sites ofdynamic actin remodeling, it is not surprising that cortactin isdetected in the actin-rich pedestals triggered by EHEC (7).

* Corresponding author. Mailing address: Division of Molecular andCellular Biology, Flowers Building, Imperial College London, LondonSW7 2AZ, United Kingdom. Phone: 44 020 2594 5253. Fax: 44 0205794 3069. E-mail: g.frankel@imperial.ac.uk.

† A.M. and A.D.W. contributed equally to this work.� Published ahead of print on 4 August 2008.

4669

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Furthermore, its importance in these structures is underscoredby the fact that truncated forms of cortactin have been re-ported to exert a dominant-negative effect on pedestal forma-tion, and depletion of cortactin from HeLa cells by small in-terfering RNA treatment abolished pedestal formation (7).Recently, Cantarelli et al. have shown that cortactin can bindthe N terminus of Tir and the proline-rich repeats of TccP (8).The fact that cortactin can activate N-WASP by itself, is able tobind Tir, and can interact with TccP led to the hypothesis thatit might be the adaptor protein connecting Tir and TccP duringinfection (8). Alternatively, cortactin may contribute to analternative pathway of actin assembly, such as the inefficientTccP-independent actin polymerization triggered by EHEC.The aim of this study was to investigate cortactin recruitmentin the context of EHEC-triggered actin polymerization in vitroand A/E lesion formation ex vivo.

MATERIALS AND METHODS

Bacterial strains. The bacterial strains used in this study are listed in Table 1.The bacteria were grown in Luria-Bertani (LB) medium or in Dulbecco’s mod-ified Eagle’s medium supplemented with kanamycin (50 �g � ml�1) and/or am-picillin (100 �g � ml�1) when necessary.

Cell culture and transfection. HeLa, Swiss 3T3 (ATCC), and N-WASP�/�

mouse embryo fibroblast cell lines (kindly supplied by S. Snapper) were grownunder conventional conditions. Cells were seeded on glass coverslips (13-mmdiameter) in 24-well plates at a density of 5 � 104 cells per well 48 h beforeinfection. Where appropriate, the cells were transfected 24 h after being seededusing Jet-PEI-RGD (Polyplus) in accordance with the manufacturer’s protocol.

Plasmids. The plasmids used in this study are listed in Table 1. For bacterialexpression of tir, primers F1 (5�-ATGCCTATTGGTAATCTTGGTC-3�) and R1(5�-AGGCTGCAGTTAGACGAAACGATGGGATC-3�) were used to amplify tirfrom wild-type EHEC genomic DNA strain TUV 93-0. PCR products were clonedbetween the SmaI and PstI restriction sites of pSA10 (23), a vector containingmultiple cloning sites downstream of the tac promoter, generating plasmid pICC421.Amino acid substitutions were introduced into TirEHEC using the QuikChangeII Site-Directed Mutagenesis kit (Stratagene), pICC421 as a template, and primersF2 (5�-AGGGACCGTGCAGAATCCGGCTGCTGATGTTAAAACATCG-3�),R2 (5�-CGATGTTTTAACATCAGCAGCCGGATTCTGCACGGTCCCT-3�),F3 (5�-TCCCAATGTGAATAATTCAATTGCTGCTGCAGCTGCATTAGCTTCACAAACCGACGG-3�), and R3 (5�-CCGTCGGTTTGTGAAGCTAATGCAGCTGCAGCAGCAATTGAATTATTCACATTGGGA-3�) to generate TirY458A

and TirP17-23A derivatives, plasmids pICC422 and pICC423, respectively.

Infection and immunofluorescent staining. EHEC cultures, grown in LB me-dium for 8 h, were diluted 1:500 in Dulbecco’s modified Eagle’s medium andgrown statically overnight at 37°C in a 5% CO2 atmosphere prior to infection.The cells were infected for 5 h, washed three times in phosphate-buffered saline,and fixed for 20 min in 4% paraformaldehyde. The infected monolayers werepermeabilized with 0.1% Triton for 4 min and labeled by indirect immunofluo-rescent staining. Goat polyclonal E. coli O157:H7 (Fitzgerald Industries Inter-national) and rabbit polyclonal TirEHEC (2) antisera were diluted 1:500. Mousemonoclonal anti-cortactin antibody and anti-hemagglutinin (HA) monoclonalantibody HA.11 (Convance) were diluted 1:200. Rhodamine- and Alexa 633-conjugated phalloidin (Invitrogen) were used at a dilution of 1:500; Oregongreen-conjugated phalloidin (Invitrogen) was used at a dilution of 1:100. Donkeyaminomethylcoumarin acetate-, Cy5-, rhodamine-, and Cy2-conjugated species-specific secondary antibodies (Jackson Immunoresearch Laboratories) were di-luted 1:200. Coverslips were mounted with Pro-Long Gold antifade reagent(Invitrogen) and analyzed using a Zeiss Axioimager fluorescence microscope.Representative images of multiple experiments (typically three) were processedusing axiovision software. Recruitment of cortactin and TccP-HA was quantifiedby counting a total of 4,400 adherent bacteria from two independent experi-ments.

Preparation of lysates for detection of exogenously expressed proteins byWestern blotting. Whole-cell extracts were prepared by scraping transfected cellmonolayers into protein-denaturing buffer and boiling them for 5 min prior topolyacrylamide gel electrophoresis and Western blotting. Primary antibodieswere diluted 1:1,000 and detected using porcine anti-rabbit or anti-mouse im-munoglobulin G (IgG) horseradish peroxidase-conjugated (Dako) secondaryantibodies and ECLplus detection reagent (GE Healthcare).

Human in vitro organ cultures (IVOC) and immunofluorescence staining ofcryosections. Pediatric tissue was obtained with fully informed parental consentand local ethical committee approval using grasp forceps during routine endo-scopic investigation of intestinal disorders. Mucosal biopsy specimens from theterminal ileum that appeared macroscopically normal were taken for organculture experiments as described previously (17). The biopsy specimens wereinfected with wild-type EHEC O157:H7, an isogenic tccP deletion mutant, orenteraggregative E. coli (EAEC) O42 for 8 h. An uninfected biopsy specimenwas included in each experiment to exclude endogenous bacterial infection.Adherence was examined using tissue from four patients (between 74 and 198months old) by cryosectioning and immunostaining as described previously (24).In each case, the comparison between different bacterial strains was made usingsamples from the same patient. For immunofluorescence, samples were embed-ded in OCT compound (Sakura), snap frozen in liquid nitrogen, and stored at�70°C until they were used. Serial sections 8 �m thick were cut with an MTEcryostat (SLEE Technik), picked up on poly-L-lysine-coated slides, and air dried.The tissue sections were fixed in formalin for 10 min and blocked with 0.5%bovine serum albumin, 2% normal goat serum in phosphate-buffered saline for20 min at room temperature. The slides were incubated with rabbit polyclonal

TABLE 1. Bacterial strains and plasmids used in this study

Name Description Source orreference

StrainsTUV 93-0 EHEC EDL933 O157:H7 �stx ATCCKC5 TUV 93-0 �tir 20ICC185 TUV 93-0 �tccP 13O42 Wild-type EAEC O44:H18 21

PlasmidspSA10 pKK177-3 derivative containing lacq 23pICC369 pSA10 derivative encoding TccP/EspFU-HA fusion protein 1pICC421 pSA10 derivative encoding EHEC Tir This studypICC422 pICC421 derivative encoding EHEC TirY458A This studypICC423 pICC421 derivative encoding EHEC TirP17-23A This studypBS1 pHM6 derivative encoding HN-TirEHEC-HA fusion protein 4pKC4�5-221 pHM6 derivative encoding HA-TirEHEC fusion protein with deletion of

amino acids 5 to 2214

pEGFP-N1 Vector encoding a GFPmut1 variant for expression in mammalian cells InvitrogenpGFP-N-WASP pEGFP-C1 derivative encoding murine N-WASP fused to GFPmut1 S. LommelpGFP-Cortactin pEGFP-C1 derivative encoding human cortactin fused to GFPmut1 L. Machesky

4670 MOUSNIER ET AL. INFECT. IMMUN.

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

anti-TirEHEC or mouse monoclonal anti-cortactin for 60 min at room tempera-ture, washed, and incubated in Alexa Fluor 647-conjugated goat anti-rabbit IgGor Alexa Fluor 488-conjugated goat anti-mouse IgG (Molecular Probes) for 30min. Counterstaining of bacteria and cell nuclei was performed using propidiumiodide (Sigma). The sections were analyzed with a Zeiss LSM 510 confocal laserscanning microscope.

RESULTS

Recruitment of cortactin to the site of EHEC adhesion invitro is dependent on TirY458. TirY458 has been implicated inactin polymerization and the recruitment of TccP (3, 4). Inorder to determine if TirY458 plays a role in the recruitment ofcortactin to sites of bacterial attachment, we infected Swiss 3T3

fibroblasts with wild-type EHEC, EHEC�tir, and EHEC�tircomplemented with plasmids encoding either wild-type Tir(pTirWT/pICC421) or TirY458A (pTirY458A/pICC422). Immu-nofluorescent staining confirmed that Tir was clustered be-neath adherent wild-type and EHEC�tir expressing eitherwild-type Tir or TirY458A (Fig. 1). Actin polymerization andcortactin were detected only beneath adherent bacteria in cellsinfected with wild-type EHEC and EHEC�tir expressing wild-type Tir, but not with EHEC�tir expressing TirY458A (Fig. 1).These results show that Y458 is essential for both actin poly-merization and the recruitment of cortactin.

Cortactin is recruited to the site of EHEC adhesion in vitrodownstream of TccP. In order to determine if overexpression of

FIG. 1. Recruitment of cortactin to the site of EHEC adhesion in vitro is dependent on TirY458 and TccP. Swiss 3T3 fibroblasts were infected withwild-type (WT) EHEC, EHEC�tir, EHEC�tccP, or EHEC�tir complemented with a plasmid encoding wild-type Tir or TirY458A. The bacteria weredetected in UV light using a polyclonal goat anti-E. coli O157 EHEC antibody, and Tir was detected in far-red light using a rabbit polyclonal anti-TirEHECantibody. Cortactin and actin were labeled green and red, respectively, using a mouse monoclonal anti-cortactin antibody and Rhodamine Red-X-conjugated phalloidin. Immunofluorescence showed Tir under all adherent EHEC strains except EHEC�tir. Actin polymerization and recruitment ofcortactin was seen under attached wild-type EHEC and EHEC�tir bacteria complemented with a plasmid encoding wild-type Tir. Neither actin norcortactin was detected beneath adherent EHEC�tir, EHEC�tir complemented with a plasmid encoding TirY458A,or EHEC�tccP bacteria. Shown aremonochrome images of the UV, green, red, and far-red fluorescent channels and a merged color image.

VOL. 76, 2008 RECRUITMENT OF CORTACTIN TO SITE OF EHEC CELL ADHESION 4671

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

cortactin can enhance the inefficient TirEHEC-dependent andTccP-independent actin polymerization, HeLa cells were tran-siently transfected with expression vectors encoding green fluo-rescent protein (GFP)-cortactin or GFP as a control. Expressionof both constructs was confirmed by Western blotting (Fig. 2A).GFP- and GFP-cortactin-expressing cells were infected for 5 hwith EHEC�tccP (strain ICC185), EHEC�tccP complemented

with an expression vector encoding HA-tagged TccP (pICC369),and wild-type EHEC as a control. Fixed monolayers were pro-cessed for immunofluorescent staining of bacteria and Tir. Mi-croscopic analysis revealed that Tir was concentrated beneathwild-type EHEC, EHEC�tccP, and complemented EHEC�tccPexpressing HA-tagged TccP (Fig. 2B), indicating that overexpres-sion of GFP-cortactin did not affect Tir recruitment underneath

FIG. 2. Cortactin recruitment in vitro is dependent on TccP. (A) Lysates of transfected HeLa cells were immunoblotted for expression of GFPor GFP-cortactin. The migratory position of GFP is indicated with an asterisk, and GFP-cortactin is indicated with an arrowhead. Tubulinimmunoblotting showed that equivalent protein levels were present within HeLa lysates. The positions of molecular mass standards are shown onthe left. (B) Expression of GFP-cortactin in HeLa cells does not impair Tir recruitment under adherent EHEC. GFP-cortactin-transfected HeLacells were infected with wild-type EHEC, EHEC�tccP (ICC185), and EHEC�tccP expressing HA-tagged TccP from a plasmid (pICC369). Theinfected cells were stained with goat anti-E. coli O157 EHEC to label bacteria (blue) and rabbit polyclonal anti-TirEHEC antibody (red). (C) GFP-and GFP-cortactin-transfected HeLa cells were infected with wild-type (WT) EHEC, EHEC�tccP, and EHEC�tccP expressing TccP-HA. Thebacteria were detected in UV light using a polyclonal goat anti-E. coli O157 EHEC antibody, and F-actin was stained using RhodamineRed-X-conjugated phalloidin. GFP-cortactin, colocalizing with actin, was detected beneath adherent wild-type EHEC and EHEC�tccPbacteria expressing TccP-HA, but not under EHEC�tccP bacteria. Shown are monochrome images of the UV, GFP, and red fluorescentchannels. Scale bar, 2 �m.

4672 MOUSNIER ET AL. INFECT. IMMUN.

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

adherent bacteria. Furthermore, staining of bacteria and F-actinrevealed that only wild-type EHEC and EHEC�tccP comple-mented with HA-tagged TccP recruited GFP-cortactin and trig-gered actin pedestal formation. No significant GFP signal wasseen in the pedestals triggered by these bacteria on cells express-ing GFP (Fig. 2C). Wild-type EHEC and EHEC�tccP expressingHA-tagged TccP triggered actin polymerization with similar effi-ciencies on cells expressing GFP and GFP-cortactin. As expected,EHEC�tccP was unable to efficiently trigger cortactin recruit-ment or actin pedestal formation on untransfected cells (Fig. 1) orcells transfected with GFP (Fig. 2C). Moreover, EHEC�tccP wasnot able to trigger formation of actin pedestals or recruit cortactinin GFP-cortactin-overexpressing cells (Fig. 2C). Together withthe data shown in Fig. 1, these results show that overexpression ofcortactin does not overcome the deficiency of EHEC�tccP intriggering efficient actin polymerization and that recruitment ofcortactin to the site of EHEC attachment occurs downstream ofTccP recruitment.

Cortactin is recruited to the site of EHEC adhesion in vitrodownstream of N-WASP. TccP is responsible for the recruit-ment of N-WASP during in vitro EHEC infection (6, 13). Inorder to determine if N-WASP is implicated in the recruitmentof cortactin, we infected an N-WASP-deficient fibroblast cellline (N-WASP�/� mouse embryo fibroblasts), transfected witheither GFP-tagged N-WASP or GFP alone as a control, withEHEC�tccP (ICC185) expressing HA-tagged TccP from aplasmid (pICC369). Western blotting of lysates of transfectedcells confirmed the expression of the two constructs and thatN-WASP expression did not alter cortactin levels in the cell(Fig. 3A).

N-WASP�/� cells were transfected with GFP, infected withEHEC�tccP expressing HA-tagged TccP, and immunostainedwith anti-HA and anti-cortactin antibodies. TccP-HA wasfound under 90% � 6% of adherent bacteria, and weak cor-tactin staining was seen under 9.5% � 2% of attached EHECbacteria (Fig. 3B, left and center, respectively). In contrast,

FIG. 3. Recruitment of TccP and cortactin during EHEC infection of an N-WASP-deficient cell line. (A) Lysates of transfected N-WASP�/�

mouse embryo fibroblasts were immunoblotted for GFP or GFP-tagged N-WASP using anti-GFP (left) and anti-N-WASP (right) antibodies. Themigratory position of GFP is indicated with an asterisk, and GFP–N-WASP is indicated with an arrowhead. Overexpression of N-WASP did notaffect cortactin levels. The positions of molecular mass standards are shown. (B) TccP, but not cortactin or F-actin, is efficiently recruited to sitesof EHEC adherence in the absence of N-WASP. GFP- or GFP–N-WASP-transfected N-WASP�/� cells were infected with EHEC�tccP (ICC185)expressing TccP-HA (pICC369). The infected cells were treated with anti-HA antibody to visualize, in red, epitope-tagged TccP (left); with mousemonoclonal anti-cortactin antibody (middle); and with phalloidin (right) to detect F-actin. The bacteria were labeled in UV light (pseudocolor,blue) using a goat polyclonal anti-E. coli O157 antibody.

VOL. 76, 2008 RECRUITMENT OF CORTACTIN TO SITE OF EHEC CELL ADHESION 4673

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

infection of N-WASP�/� cells transfected with GFP–N-WASPrestored intense cortactin recruitment beneath 95% � 0% ofadherent EHEC�tccP bacteria expressing HA-tagged TccP.Anti-HA staining confirmed that, similarly to cells transfectedwith GFP, TccP-HA was recruited beneath 92% � 7% ofadherent bacteria in cells transfected with GFP–N-WASP. Asexpected, no HA staining was detected beneath EHEC�tccPor wild-type EHEC (not shown) bacteria, nor were any of thestrains tested able to trigger the formation of distinct actinpedestals on N-WASP�/� cells transfected with GFP (Fig. 3B,right). Taken together, these data show that while TccP isrecruited to Tir in the absence of N-WASP, recruitment ofcortactin is N-WASP dependent. These results suggest thatcortactin is not involved in linking Tir and TccP during infec-tion of cultured cells in vitro.

The N terminus of Tir is dispensable for cortactin recruit-ment. In vitro, cortactin, via its SH3 domain, reportedly bindsdirectly to the N terminus of Tir (8), which contains a polypro-line region (amino acids 17 to 23) that is a putative SH3domain-binding site. We therefore investigated if the N termi-nus of Tir and the polyproline region contained within it havea role in cortactin recruitment under adherent bacteria. To thisend, we infected Swiss 3T3 fibroblasts with EHEC�tir comple-mented with a plasmid encoding Tir in which all five N-termi-nal proline residues were replaced with alanine, TirP17-23A

(pTirP17-23A/pICC423). In parallel, full-length HA-Tir and atruncated HA-Tir�5-221 derivative (4) lacking the entire N

terminus were expressed in HeLa cells via transfection prior toinfection with EHEC�tir. Phalloidin staining revealed thattranslocated TirP17-23A (Fig. 4A) and ectopically expressedHA-Tir and HA-Tir�5-221 (Fig. 4B) all complemented theability of EHEC�tir to induce actin-rich pedestals; in all cases,Tir was located at the tip of the pedestal (Fig. 4). Staining withan anti-cortactin antibody revealed, similarly, that cortactinwas recruited beneath adherent EHEC�tir bacteria expressingTirP17-23A on fibroblasts (Fig. 4A) and EHEC�tir bacteria fol-lowing infection of HA-Tir�5-221-transfected cells, but not onmock-transfected cells (Fig. 4B). Taken together, these dataindicate that the Tir polyproline region and, more generally,the whole N terminus of Tir are not necessary for efficient actinpedestal formation and cortactin recruitment.

Cortactin is recruited to the ex vivo site of EHEC adhesionindependently of TccP. We have recently shown, using ex vivoand in vivo infection models, that unlike the induction of actinpolymerization in vitro, TccP is dispensable for A/E lesionformation on mucosal surfaces (13, 22, 26). We therefore stud-ied cortactin recruitment to the site of EHEC adhesion duringinfection of human IVOC. To this end, intestinal biopsy spec-imens from the terminal ileum were infected with wild-typeEHEC and EHEC�tccP (ICC185). An uninfected biopsy spec-imen was used as a negative control. Immunofluorescent stain-ing of cryosections for cortactin showed general cytoplasmicand membrane staining of enterocytes in uninfected controlsamples (Fig. 5). In contrast, intense cortactin staining was de-

FIG. 4. The N terminus of Tir is dispensable for recruitment of cortactin. (A) Swiss 3T3 fibroblasts were infected with EHEC�tir complementedwith a plasmid encoding TirP17-23A. Staining as for Fig. 1 revealed actin polymerization and recruitment of cortactin under attached bacteria.(B) Mock-, HA-Tir-, or HA-Tir�5-221-transfected HeLa cells were infected with EHEC�tir. The infected cells were treated with polyclonal goatanti-E. coli O157 antibody to label bacteria (blue), with Oregon green phalloidin to detect F-actin (green), and with either a mouse monoclonalanti-HA antibody (left) to visualize epitope-tagged Tir (red) or a mouse monoclonal anti-cortactin antibody (right) (red). Shown are separatemonochrome images of the UV, red, and green fluorescence channels and a merged color image.

4674 MOUSNIER ET AL. INFECT. IMMUN.

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

tected under every adherent wild-type EHEC and EHEC�tccPbacterium (30/30) (Fig. 5), indicating that cortactin recruit-ment on mucosal surfaces is independent of TccP. AsEHEC�tir does not bind IVOC, we could not use this strain asa control. Instead, we employed EAEC O42, which binds toterminal ileal mucosa (21) by a Tir-independent mechanism.No cortactin was detected at the site of any attached EAECbacterium (10/10) (Fig. 5), suggesting that recruitment of cor-tactin to EHEC O157:H7 is specific.

DISCUSSION

A number of pathogenic bacteria target cortactin duringinfection. For example, Listeria monocytogenes and Shigellaflexneri exploit cortactin for invasion of mammalian host cells(reviewed in reference 25). In addition, cortactin is reported tobind the N terminus of Tir (8) and to play a role in triggeringactin polymerization at the site of EHEC adhesion to eukary-otic cells (7). Here, we have studied the recruitment of cortac-tin during EHEC infection of cultured cells and human IVOC.

We have shown that overexpression of cortactin does notenhance actin polymerization after infection of cultured cellswith EHEC�tccP, suggesting that cortactin is unlikely to con-tribute to the inefficient TccP-independent actin polymeriza-tion pathway triggered by TirEHEC in vitro. We have shownthat although wild-type Tir is focused under attachedEHEC�tccP bacteria and TirY458A is focused underEHEC�tir(pTirY458A) bacteria (which express TccP but areunable to recruit it under attached bacteria), this does not leadto recruitment of cortactin. These results suggest Tir does not

directly recruit cortactin, but rather, cortactin recruitment tothe site of bacterial attachment is TccP dependent. Moreover,cortactin was recruited to EHEC�tir expressing TirP17-23A andto EHEC�tir after infection of cells transfected to expressTir�5-221. These results show that despite the reported abilityof cortactin to bind the N terminus of Tir directly (8), thisbinding activity is not sufficient for recruitment of cortactin tothe site of intimate bacterial attachment or for actin polymer-ization.

Cortactin is able to bind to both Tir and TccP in vitro (8),thus implicating cortactin as a putative host adaptor linking thetwo bacterial effectors required for actin pedestal formation invitro. To test this hypothesis experimentally, we infected N-WASP-deficient cells complemented with either GFP–N-WASP or GFP alone. We established that TccP is recruited toTir in the absence of N-WASP and then reasoned that ifcortactin was the adaptor, it would also be recruited to Tir inthe absence of N-WASP. However, cortactin was recruitedonly in cells transfected with GFP–N-WASP. Together, theseresults indicate that in vitro, cortactin is recruited to the site ofEHEC adhesion, not directly by Tir but downstream of TccPand N-WASP. Although not involved in linking Tir and TccP,cortactin might have a role in stabilizing the actin polymeriza-tion complex, as inferred by Cantarelli et al. (8).

It has recently been shown that TccP is not needed forcolonization of mucosal surfaces ex vivo (1) and in vivo (22,26). In order to demonstrate if the data we gathered for cor-tactin in vitro could be extended to the infection of mucosalsurfaces, we infected human terminal ileal biopsy specimens

FIG. 5. Immunofluorescence staining of cryosections of human terminal ileum infected with wild-type EHEC, EHEC�tccP, or EAEC. Whereasgeneral cytoplasmic and membrane cortactin staining could be observed in enterocytes of noninfected and EAEC-infected samples, intensecortactin staining beneath Tir-expressing EHEC bacteria and an isogenic tccP deletion mutant was evident (upper row; green in merged images).Sections were also stained for Tir (blue in merged images) and counterstained with propidium iodide (red) to visualize bacteria and cell nuclei.Cortactin (green) and Tir (blue) staining overlap in the merged image.

VOL. 76, 2008 RECRUITMENT OF CORTACTIN TO SITE OF EHEC CELL ADHESION 4675

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

with wild-type EHEC O157:H7, EHEC�tccP, and EAEC O42as a control. In contrast to the recruitment of cortactin in vitro,which is TccP dependent, we found that cortactin recruitmentto the site of EHEC attachment in human IVOC was TccPindependent.

Previous data obtained from analysis of EHEC-infected in-testinal tissue using transmission electron microscopy revealedthat regions of electron-dense staining characteristic of local-ized actin assembly were often not associated with intimatelyattached EHEC�tccP bacteria (22) or an analogous tccP-mi-nus EPEC strain expressing EHEC O157-like Tir (1). Theseobservations suggest that EHEC�tccP has a diminished capac-ity to generate actin pedestals (but not A/E lesions) in vivo,despite retaining the ability to recruit cortactin. Importantly,the concentration of cortactin at the sites of EHEC A/E lesionsappears to be pathogen specific, as the protein was not re-cruited to sites of EAEC adhesion. One explanation for thesedata is that A/E lesions induced in ex vivo intestinal epithelialcells are not strictly equivalent to actin polymerization in cul-tured cells in vitro, both in terms of the signaling pathways (10)and in the host cell proteins that accumulate therein. Cortac-tin, as it is recruited to sites of bacterial adhesion ex vivo via aTccP- and N-WASP-independent pathway, might have a roleduring EHEC infection of mucosal surfaces that is more influ-ential than that of cultured cells in vitro, and its role in medi-ating A/E lesion formation is currently under investigation.

ACKNOWLEDGMENTS

We thank V. Koronakis (Cambridge University, United Kingdom) forkindly supplying anti-N-WASP antiserum, Scott Snapper of HarvardMedical School for the N-WASP knockout and control mouse embryofibroblast cell lines, S. Lommel (Institute for Cell Biology, University ofBonn, Bonn, Germany) for the GFP–N-WASP, and L. Machesky (Beat-son Institute, Glasgow, United Kingdom) for the cortactin expressionvectors. Monoclonal anti-tubulin antibody developed by M. Klymkowskywas obtained from the Developmental Studies Hybridoma Bank devel-oped under the auspices of the NICHD and maintained by the Depart-ment of Biological Sciences, University of Iowa, Iowa City, IA 52242.

J.M.L. is supported by NIH R01-AI46454. A.D.W. is funded by aBBSRC studentship. Work in the laboratory of A.D.P. was supportedby the NIH (grant R37AI21657 to J. B. Kaper). This project wassupported by the Wellcome Trust.

REFERENCES

1. Bai, L., S. Schuller, A. Whale, A. Mousnier, O. Marches, L. Wang, T. Ooka,R. Heuschkel, F. Torrente, J. B. Kaper, T. A. Gomes, J. Xu, A. D. Phillips,and G. Frankel. 2008. Enteropathogenic Escherichia coli (EPEC) O125:H6triggers attaching and effacing lesions on human intestinal biopsies indepen-dently of Nck and TccP/TccP2. Infect. Immun. 76:361–368.

2. Batchelor, M., J. Guignot, A. Patel, N. Cummings, J. Cleary, S. Knutton,D. W. Holden, I. Connerton, and G. Frankel. 2004. Involvement of theintermediate filament protein cytokeratin-18 in actin pedestal formationduring EPEC infection. EMBO Rep. 5:104–110.

3. Brady, M. J., K. G. Campellone, M. Ghildiyal, and J. M. Leong. 2007.Enterohaemorrhagic and enteropathogenic Escherichia coli Tir proteins trig-ger a common Nck-independent actin assembly pathway. Cell. Microbiol.9:2241–2253.

4. Campellone, K. G., M. J. Brady, J. G. Alamares, D. C. Rowe, B. M. Skehan,D. J. Tipper, and J. M. Leong. 2006. Enterohaemorrhagic Escherichia coli Tirrequires a C-terminal 12-residue peptide to initiate EspFU-mediated actinassembly and harbours N-terminal sequences that influence pedestal length.Cell. Microbiol. 8:1488–1503.

5. Campellone, K. G., and J. M. Leong. 2003. Tails of two Tirs: actin pedestal

formation by enteropathogenic E. coli and enterohemorrhagic E. coli O157:H7. Curr. Opin. Microbiol. 6:82–90.

6. Campellone, K. G., D. Robbins, and J. M. Leong. 2004. EspFU is a translo-cated EHEC effector that interacts with Tir and N-WASP and promotesNck-independent actin assembly. Dev. Cell 7:217–228.

7. Cantarelli, V. V., T. Kodama, N. Nijstad, S. K. Abolghait, T. Iida, and T.Honda. 2006. Cortactin is essential for F-actin assembly in enteropathogenicEscherichia coli (EPEC)- and enterohaemorrhagic E. coli (EHEC)-inducedpedestals and the alpha-helical region is involved in the localization ofcortactin to bacterial attachment sites. Cell. Microbiol. 8:769–780.

8. Cantarelli, V. V., T. Kodama, N. Nijstad, S. K. Abolghait, S. Nada, M.Okada, T. Iida, and T. Honda. 2007. Tyrosine phosphorylation controlscortactin binding to two enterohaemorrhagic Escherichia coli effectors: Tirand EspFu/TccP. Cell. Microbiol. 9:1782–1795.

9. Deibel, C., S. Kramer, T. Chakraborty, and F. Ebel. 1998. EspE, a novelsecreted protein of attaching and effacing bacteria, is directly translocatedinto infected host cells, where it appears as a tyrosine-phosphorylated 90 kDaprotein. Mol. Microbiol. 28:463–474.

10. Frankel, G., and A. D. Phillips. 2008. Attaching effacing E. coli and para-digms of Tir-triggered actin polymerization: getting off the pedestal. Cell.Microbiol. 10:549–556.

11. Frankel, G., A. D. Phillips, L. R. Trabulsi, S. Knutton, G. Dougan, and S. J.Matthews. 2001. Intimin and the host cell—is it bound to end in Tir(s)?Trends Microbiol. 9:214–218.

12. Garmendia, J., G. Frankel, and V. F. Crepin. 2005. Enteropathogenic andenterohaemorrhagic E. coli infections: translocation, translocation, translo-cation. Infect. Immun. 73:2586–2594.

13. Garmendia, J., A. Phillips, Y. Chong, S. Schuller, O. Marches, S. Dahan, E.Oswald, R. K. Shaw, S. Knutton, and G. Frankel. 2004. TccP is an entero-haemorrhagic E. coli O157:H7 type III effector protein that couples Tir tothe actin-cytoskeleton. Cell. Microbiol. 6:1167–1183.

14. Goosney, D. L., R. DeVinney, and B. B. Finlay. 2001. Recruitment of cy-toskeletal and signaling proteins to enteropathogenic and enterohemor-rhagic Escherichia coli pedestals. Infect. Immun. 69:3315–3322.

15. Hartland, E. L., M. Batchelor, R. M. Delahay, C. Hale, S. Matthews, G.Dougan, S. Knutton, I. Connerton, and G. Frankel. 1999. Binding of intiminfrom enteropathogenic Escherichia coli to Tir and to host cells. Mol. Micro-biol. 32:151–158.

16. Hayward, R. D., J. M. Leong, V. Koronakis, and K. G. Campellone. 2006.Exploiting pathogenic Escherichia coli to model transmembrane receptorsignalling. Nat. Rev. Microbiol. 4:358–370.

17. Hicks, S., G. Frankel, J. B. Kaper, G. Dougan, and A. D. Phillips. 1998. Roleof intimin and bundle forming pili in enteropathgenic Escherichia coli adhe-sion to pediatric intestine in vitro. Infect. Immun. 66:1570–1578.

18. Kenny, B., R. DeVinney, M. Stein, D. J. Reinscheid, E. A. Frey, and B. B.Finlay. 1997. Enteropathogenic E. coli (EPEC) transfers its receptor forintimate adherence into mammalian cells. Cell 91:511–520.

19. Moon, H. W., S. C. Whipp, R. A. Argenzio, M. M. Levine, and R. A. Gian-nella. 1983. Attaching and effacing activities of rabbit and human entero-pathogenic Escherichia coli in pig and rabbit intestines. Infect. Immun. 41:1340–1351.

20. Murphy, K. C., and K. G. Campellone. 2003. Lambda Red-mediated recom-binogenic engineering of enterohemorrhagic and enteropathogenic E. coli.BMC Mol. Biol. 4:11.

21. Nataro, J. P., Y. Deng, S. Cookson, A. Cravioto, S. J. Savarino, L. D. Guers,M. M. Levine, and C. O. Tacket. 1995. Heterogeneity of enteroaggregativeEscherichia coli virulence demonstrated in volunteers. J. Infect. Dis. 171:465–468.

22. Ritchie, J. M., M. J. Brady, K. N. Riley, T. D. Ho, K. G. Campellone, I. M.Herman, A. Donohue-Rolfe, S. Tzipori, M. K. Waldor, and J. M. Leong.2008. EspF(U), a type III-translocated effector of actin assembly, fostersepithelial association and late-stage intestinal colonization by E. coli O157:H7. Cell. Microbiol. doi: 10.1111/j.1462–5822.2007.01087.x.

23. Schlosser-Silverman, E., M. Elgrably-Weiss, I. Rosenshine, R. Kohen, and S.Altuvia. 2000. Characterization of Escherichia coli DNA lesions generatedwithin J774 macrophages. J. Bacteriol. 182:5225–5230.

24. Schuller, S., Y. Chong, J. Lewin, B. Kenny, G. Frankel, and A. D. Phillips. 2007.Tir phosphorylation and Nck/N-WASP recruitment by enteropathogenic andenterohaemorrhagic Escherichia coli during ex vivo colonization of human in-testinal mucosa is different to cell culture models. Cell. Microbiol. 9:1352–1364.

25. Selbach, M., and S. Backert. 2005. Cortactin: an Achilles’ heel of the actincytoskeleton targeted by pathogens. Trends Microbiol. 13:181–189.

26. Vlisidou, I., F. Dziva, R. M. La Ragione, A. Best, J. Garmendia, P. Hawes, P.Monaghan, S. A. Cawthraw, G. Frankel, M. J. Woodward, and M. P. Stevens.2006. Role of intimin-Tir interactions and the Tir-cytoskeleton couplingprotein in the colonization of calves and lambs by Escherichia coli O157:H7.Infect. Immun. 74:758–764.

Editor: J. B. Bliska

4676 MOUSNIER ET AL. INFECT. IMMUN.

on February 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from