Toll-Like Receptor 2 Enhances ZO-1â€“Associated Intestinal Epithelial

TE

E*G

Biic(sTtMPrpatCiuliTwoIstZldsasvpciit

Timp

GASTROENTEROLOGY 2004;127:224–238

oll-Like Receptor 2 Enhances ZO-1–Associated Intestinalpithelial Barrier Integrity Via Protein Kinase C

LKE CARIO,* GUIDO GERKEN,* and DANIEL K. PODOLSKY‡

Division of Gastroenterology and Hepatology, University Hospital of Essen, Essen, Germany; and ‡Gastrointestinal Unit, Massachusettseneral Hospital, Center for the Study of Inflammatory Bowel Disease, Harvard Medical School, Boston, Massachusetts

i(boP

eflthhettHhl

rtnpabhsatclT

flsP1Tr4

ackground & Aims: Protein kinase C (PKC) has beenmplicated in regulation of intestinal epithelial integrityn response to lumenal bacteria. Intestinal epithelialells (IECs) constitutively express Toll-like receptorTLR)2, which contains multiple potential PKC bindingites. The aim of this study was to determine whetherLR2 may activate PKC in response to specific ligands,hus potentially modulating barrier function in IECs.ethods: TLR2 agonist (synthetic bacterial lipopeptideam3CysSK4, peptidoglycan)–induced activation of PKC-elated signaling cascades were assessed by immuno-recipitation, Western blotting, immunofluorescence,nd kinase assays—combined with functional transfec-ion studies in the human model IEC lines HT-29 andaco-2. Transepithelial electrical resistance character-zed intestinal epithelial barrier function. Results: Stim-lation with TLR2 ligands led to activation (phosphory-ation, enzymatic activity, translocation) of specific PKCsoforms (PKC� and PKC�). Phosphorylation of PKC byLR2 ligands was blocked specifically by transfectionith a TLR2 deletion mutant. Ligand-induced activationf TLR2 greatly enhanced transepithelial resistance inECs, which was prevented by pretreatment with PKC-elective antagonists. This effect correlated with apicalightening and sealing of tight junction (TJ)-associatedO-1, which was mediated via PKC in response to TLR2igands, whereas morphologic changes of occludin, clau-in-1, or actin cytoskeleton were not evident. Down-tream the endogenous PKC substrate myristoylatedlanine-rich C kinase substrate (MARCKS), but not tran-criptional factor activator protein-1 (AP-1), was acti-ated significantly on stimulation. Conclusions: Theresent study provides evidence that PKC is an essentialomponent of the TLR2 signaling pathway with the phys-ologic consequence of directly enhancing intestinal ep-thelial integrity through translocation of ZO-1 on activa-ion.

he intestinal epithelium constitutes an anatomic aswell as immunologic barrier that forms a bipolar

nterface between the diverse populations of lumenalicrobes and immune cells of the underlying lamina

ropria. Barrier function is maintained by a complex

nterplay of numerous proteins within the tight-junctionTJ) complex.1 Biogenesis of TJ is regulated selectivelyy protein kinase C (PKC), and the TJ protein zonulaccludens-1 (ZO-1) appears to be a specific target ofKC.2

Disruption of the sensitive equilibrium of TJs may belicited by a number of pathogenic bacteria and proin-ammatory cytokines, exacerbating and perpetuating in-estinal inflammation in disease.3 To counteract potentialarmful effects of lumenal toxins and to protect barrieromeostasis, intestinal epithelial cells (IECs) exhibit sev-ral defensive features, including production of intestinalrefoil peptides and mucins.4 Commensals also may assisthe host to confer intestinal epithelial barrier integrity.5,6

owever, the underlying mechanisms of these beneficialost-microbe interactions that may tighten the paracel-ular seal are largely unclear.

Toll-like receptors (TLRs) comprise a class of pattern-ecognition receptors that specifically discriminate be-ween self- and microbial non–self-based on the recog-ition of broadly conserved molecular patterns.7 TLRslay a key role in microbial recognition, control ofdaptive immune responses, and induction of antimicro-ial effector pathways, leading to efficient elimination ofost-threatening pathogens. We and others recently havehown that IECs express several TLRs, including TLR2nd TLR4, in vitro and in vivo.8–14 As the frontline ofhe mucosal immune system, the intestinal epitheliumonstantly is exposed to large amounts of various TLRigands that appear to coexist in the intestinal mucosa.15

o maintain mucosal homeostasis, inflammatory re-

Abbreviations used in this paper: AP-1, activator protein-1; FITC,uorescein isothiocyanate; IEC, intestinal epithelial cell; LPS, lipopoly-accharide; MARCKS, myristoylated alanine-rich C kinase substrate;GN, peptidoglycan; PKC, protein kinase C; PMA, phorbol 12-myristate3-acetate; TER, transepithelial resistance; TJ, tight junction; TLR,oll-like receptor; TLR2DN, dominant-negative construct of Toll-likeeceptor 2; TLR4DN, dominant-negative construct of Toll-like receptor; ZO-1, zonula occludens-1.

© 2004 by the American Gastroenterological Association0016-5085/04/$30.00

doi:10.1053/j.gastro.2004.04.015

stcarlfldipt

1lslsbnmpaPphtoospPntdTm

bfBsbtTel

(

omttvafFpLupcBwmf

nfp(pa1MaandcpcmfafdpmafgIgNEItu

(lt

July 2004 TLR2 ENHANCES TER VIA PKC 225

ponses are suppressed toward commensals, leading tohe phenomenon of tolerance in the healthy gut. De-reased expression of TLRs and high levels of Tollipppear to correlate with inhibition of proinflammatoryesponses in the face of commensals.11,16,17 However,ipopolysaccharide (LPS) may acutely elicit several proin-ammatory responses in some IECs.8,18 Cytokine-in-uced imbalance in inflammation also might lead tontestinal intolerance toward LPS by altering TLR4 ex-ression and triggering exaggerated downstream eventshat promote disease.18,19

The PKC superfamily is known to consist of at least1 isozymes with selective tissue distribution and cellocalization as well as numerous activators and sub-trates, thus modulating a complex array of diverse cel-ular functions and immune responses. Based on theirtructure and cofactor regulation, PKC isoforms haveeen classified into 3 groups: conventional (�, �I, �II, �),ovel (�, �, �, �, �), and atypical (, /�). Differentialechanisms of PKC-isozyme regulation include phos-

horylation, which modulates the active enzymatic sitend subcellular trafficking.20,21 In Drosophila, atypicalKC is required for stimulation of the Toll-signalingathway, leading to activation of the nuclear factor � Bomologues Dif and Dorsal, controlling the transcrip-ional activity of the Drosomycin promoter.22 Inhibitionf various PKC isoforms using pseudosubstrate peptidesr pharmacologic inhibitors are thought to impair LPSignaling in human dendritic cells23 and murine macro-hages.24 Moreover, LPS-stimulated macrophages fromKC� / appear to be deficient in the induction ofitric oxide synthase, suggesting that PKC� is an essen-ial downstream target of LPS signaling.25 However, airect causal connection of PKC activation with theLR-signaling pathway has not been established in theammalian system so far.It recently has been suggested that IECs remain

roadly hyporesponsive to TLR2 ligands16 and, so far, nounctional role has been ascribed to TLR2 in IECs.ecause TLR2 contains multiple potential PKC-binding

ites, we speculated that specific isoforms of PKC mighte activated in response to TLR2 ligands. To understandhe specific biologic consequence of PKC activation viaLR2 in IECs, we analyzed PKC-induced signalingvents and morphologic changes in response to bacterialigands in this study.

Materials and MethodsReagents and Antibodies

The synthetic lipopeptide Pam3Cys-SKKKKx3HClPam CysSK4 [PCSK]; lots #D06, G05, K05, F06, G06) was
3
btained from EMC Microcollections GmbH (Tubingen, Ger-any) and were prepared as recommended by the manufac-

urer. In brief, Pam3CysSK4 was dissolved in sterile (endo-oxin � 0.005 EU/mL) H20 (ICN, Aurora, OH), thoroughlyortexed, stored in small aliquots for up to 8 weeks at 20°C,nd treated with ultrasonics before use. Peptidoglycan (PGN)rom Staphylococcus aureus (lot #14427/1) was purchased fromluka (Buchs, Switzerland) and dissolved in sterile Dulbecco’shosphate-buffered saline (PBS) without Ca2�/Mg2� (PAAaboratories, Linz, Austria), thoroughly sonicated for 5 min-tes on ice, and stored in small aliquots at 20°C. Highlyurified (99.9% free of DNA and protein) LPS from Escherichiaoli, serotype R515 (lot #L11290) was obtained from Alexisiochemicals (Grunberg, Germany). Rottlerin and Go6976ere purchased from Merck Biosciences (Schwalbach, Ger-any). Phorbol 12-myristate 13-acetate (PMA) was obtained

rom Sigma-Aldrich (Taufkirchen, Germany).Rabbit polyclonal antibodies to conventional PKC� (C-20),

ovel PKC� (C-17),26 and panPKC (H-300) were obtainedrom Santa Cruz Biotechnology (Santa Cruz, CA). Rabbitolyclonal antibodies to phospho-panPKC, phospho-PKC�/�II

threonine 638/641), phospho-PKC�/� (serine 643/676), andhospho–myristoyl-ted alanine-rich C kinase substrate (MARCKS) (serine 152/56) were purchased from Cell Signaling Technology (Beverly,A). As indicated by the manufacturer, phospho-panPKC

ntibody detects endogenous levels of PKC�, �I, �II, �, �, �,nd � isoforms only when phosphorylated at a carboxy-termi-al residue homologous to serine 660 of PKC�II and does notetect PKC phosphorylated at other sites. Extracts of U937ells treated with 0.2 �mol/L PMA for 30 minutes were kindlyrovided by Cell Signaling Technology as proven positiveontrol for phosphorylation of pan-PKC and PKC�.27 Mouseonoclonal antibodies to �-actin and flag-M2 were obtained

rom Sigma-Aldrich. Rabbit polyclonal antibodies to occludinnd claudin-1 and mouse monoclonal to ZO-1 were purchasedrom Zymed (South San Francisco, CA). Horseradish-peroxi-ase– conjugated anti-rabbit and anti-mouse antibodies wereurchased from Amersham Pharmacia Biotech (Freiburg, Ger-any). Fluorescein isothiocyanate (FITC)-conjugated goat

nti-rabbit immunoglobulin (Ig)G antibody was purchasedrom Vector Laboratories (Burlingame, CA). FITC-conjugatedoat anti-mouse IgG antibody was obtained from JacksonmmunoResearch Laboratories (West Grove, PA). CY5-conju-ated goat anti-rabbit IgG antibody was a gift from Dr. Jensurnberger (Division of Nephrology, University Hospital ofssen, Germany). Normal rabbit IgG (Santa Cruz) and mousegG (Ebioscience, San Diego, CA) were used as negative con-rols. All other reagents were obtained from Sigma-Aldrichnless otherwise specified.

Cell Culture

Two different IEC lines, Caco-2 (#5-23) and HT-29#3-20), were obtained from the American Type Culture Col-ection (Manassas, VA) through LGC Promochem, Tedding-on, Middlesex, United Kingdom, (lot #1537739 [HTB-37]

ahmEvratorcigh1ctetbaw

br�NTcmtmfiptaap(it(m(pmbdtm(si

nrtwse(cwwwdfai1aIt

1t�bmyaN2tpb(oalusmStfgotsu(Ar

c

226 CARIO ET AL. GASTROENTEROLOGY Vol. 127, No. 1

nd lot #1467609 [HTB-38], respectively) and kept in aumidified incubator at 37°C with 5% CO2. Caco-2 wereaintained in high-glucose (4.5 g/L) Dulbecco’s modifiedagle medium (PAA Laboratories), supplemented with 20%ol/vol non–heat-inactivated fetal calf serum (PAA Laborato-ies, endotoxin-free, lot #A01121-245), 100 U/mL penicillin,nd 100 �g/mL streptomycin (PAA Laboratories). When cul-ured on semipermeable membranes, Caco-2 cells spontane-usly differentiate into a highly functionalized epithelial bar-ier exhibiting similar structural and biochemicalharacteristics as mature enterocytes. HT-29 were maintainedn glucose (3.0 g/L)-containing McCoy’s 5A medium (Invitro-en, San Diego, CA), supplemented with 10% vol/vol non–eat-inactivated fetal calf serum, 100 U/mL penicillin, and00 �g/mL streptomycin. In contrast to Caco-2 cells, HT-29ells thus were kept undifferentiated with very low transepi-helial resistance after reaching confluence.28 When subconflu-nt, Caco-2 cells were passaged by gentle mechanical disrup-ion and HT-29 cells by treatment with enzyme-free, Hanks’-ased, cell dissociation buffer (Invitrogen). Media of Caco-2nd HT-29 cells was changed 2 and 3 times, respectively, pereek.

Western Blotting and Immunoprecipitation

Cells were subjected to last media change 24–36 hoursefore stimulation. After incubation with stimuli, cells wereinsed twice in cold PBS (without Ca2�/Mg2�) with 100mol/L Na3VO4, and then lysed in ice-cold lysis buffer (1%P-40 [Pierce, Rockford, IL], 50 mmol/L NaCl, 20 mmol/Lris-HCl, pH 7.4, 2 mmol/L ethylenediaminetetraacetic acid,ontaining 10 mmol/L NaF, 10 mmol/L dithiothreitol, 10mol/L Na3VO4, complete miniprotease inhibitor cocktail

ablet [Roche, Mannheim, Germany], and 2 mmol/L phenyl-ethyl sulfonyl fluoride plus [Roche]). Lysates were centri-

uged (12,000 � g, 15 min, at 4°C), and protein concentrationn each supernatant was determined by colorimetric Bradfordrotein assay (Bio-Rad, Hercules, CA). For immunoprecipita-ion, the supernatants were precleared at 4°C overnight bydding 2.5 �g of anti-rabbit IgG and 100 �L protein Agarose (3% vol/vol; Amersham) and then incubated withrimary antibody (1:200) and 100 �L of protein A agarose3%) overnight at 4°C. Beads were washed 4 times withce-cold lysis buffer and processed for further Western blot-ing. Proteins were heated in NuPAGE LDS sample bufferInvitrogen) after addition of 1 mmol/L dithiothreitol (85°C, 2in), subjected to sodium dodecyl sulfate–polyacrylamide gel

10-, 12-, or 15-well, 4%–12% Bis-Tris; Invitrogen) electro-horesis at 130 V, transferred onto a polyvinylidene difluorideembrane (Millipore, Eschborn, Germany) at 30 V, followed

y blocking (Tris-buffered saline tween-20 with 5% nonfatry milk or 1%–5% bovine serum albumin) for 1 hour at roomemperature, washing with Tris-buffered saline tween-20 for 5inutes 3 times, immunoblotting with primary antibody

1:500–1:1000 in 5% nonfat dry milk or 0.1%–5% bovineerum albumin) for overnight at 4°C, and then with horserad-sh-peroxidase– conjugated secondary antibody (1:8000 in 4%

onfat dry milk in Tris-buffered saline tween-20) for 1 hour atoom temperature. After washing with Tris-buffered salineween-20 for 5 minutes 3 times, the membrane was developedith the enhanced chemiluminescence detection kit Renais-

ance (NEN Life Science, Boston, MA) and exposed for differ-nt time periods (10 s–20 min) to Kodak BioMax Light filmKodak GmbH, Stuttgart, Germany) followed by manual pro-essing (Adefo Chemie, Nurnberg, Germany) in a standardizeday (developing, 10 s–2 min; rinsing, 30 s; fixation, 5 min;ashing, 5–10 min). To confirm equal loading, immunoblotsere stripped with 62.5 mmol/L Tris-HCl, pH 6.8, 2% so-ium dodecyl sulfate, containing 100 mmol/L 2-ME at 50°Cor 30 minutes and reprobed with anti–�-actin (1:10,000) ornti-panPKC (1:500). Images of Western blots were acquiredn a standardized way (800 dpi) using an Epson Perfection640SU-Photo scanner (Seiko Epson Corp., Nagano, Japan)nd digitized with Adobe Photoshop 5.0LE (Adobe Systems,nc., Palo Alto, CA). All experiments were repeated at least 3imes; representative results are shown for each experiment.

In Vitro Kinase Assay

HT-29 monolayers, grown at 70% confluence on00-mm tissue culture dishes, were stimulated and washedwice with cold PBS (without Ca2�/Mg2�) containing 100mol/L Na3VO4. Proteins were extracted by 10-minute incu-ation on ice with 750 �L of lysis buffer (1% NP-40, 50mol/L HEPES, pH 7.4, 100 mmol/L NaCl, 2 mmol/L eth-

lenediaminetetraacetic acid, 2 mmol/L ethylene glycol-bis(�-minoethyl ether)-N,N,N�,N�-tetraacetic acid, 50 mmol/LaF, 1 mmol/L Na3VO4, 5 mmol/L �-glycerophosphate, andmmol/L phenylmethyl sulfonyl fluoride). After centrifuga-

ion for 10 minutes at 12,000 � g at 4°C, and adjustment ofrotein concentrations, lysates (600 �L/sample) were incu-ated immediately with polyclonal antibodies against cPKC�3 �g) or nPKC� (3 �g) and protein-A Sepharose beads forvernight rotation at 4°C. After centrifugation for 10 minutest 12,000 � g at 4°C, beads were washed twice with ice-coldysis buffer, resuspended in 20 �L of prewarmed kinase buffersing a brightly colored, fluorescent PKC-specific peptideubstrate (PepTag nonradioactive protein kinase C assay; Pro-ega, Madison, WI), and incubated for 30 minutes at 30°C.

amples then were boiled for 10 minutes, vortexed, and cen-rifuged at 12,000 � g for 5 minutes at room temperature,ollowed by electrophoresis (100 V, 15 min) on a 0.8% agaroseel (50 mmol/L Tris-HCl, pH 8.0). Phosphorylation by PKCf the substrate alters the peptide’s net charge from �1 to 1,hereby allowing the active and inactive versions of the sub-trate to be separated rapidly by electrophoresis and visualizednder ultraviolet light. The PKC positive (active) and negativeinactive) controls were used as supplied by the manufacturer.ll experiments were repeated at least 3 times; representative

esults are shown for each experiment.

Confocal Immunofluorescence

Caco-2 cells were cultured on 4-well permanox slide-hambers (NalgeNunc, Naperville, IL) until differentiated

(a

ccbg6a(b4

c wtoIi

corMfiamscoOctbetbctcioemcmAli

oC

t0CcTbvfsiiwtvp

awCMsbGsemtasTitote(tfwEH(tm(dfs

lTw1


unless otherwise indicated in the Results section), stimulated,nd washed twice with PBS (without Ca2�/Mg2�) at 37°C.

For assessment of protein kinase C isoform translocation,ells were fixed with fresh, ice-cold 4% paraformaldehydeontaining 0.1% Triton X-100 for 60 minutes at 4°C followedy washing with ice-cold PBS. Cells were blocked with 1%oat serum (Vector) and 1% bovine serum albumin in PBS for0 minutes at room temperature and incubated with primaryntibody (cPKC� or nPKC�; 1:100) or normal rabbit IgGequivalent dilution) as negative control (data not shown) inlocking buffer containing 0.1% Triton X-100 overnight at°C.

For assessment of TJ assembly and transfection efficiency,ells were fixed with methanol/acetone (50:50) for 5 minutes at20°C, air-dried, and washed with TBST. Cells were blockedith normal goat serum (1:100 in PBS) for 60 minutes at room

emperature and incubated with primary antibody (claudin-1,ccludin, ZO-1, or flag-M2; 1:100) or normal rabbit/mousegG (equivalent dilution) as negative controls (data not shown)n PBS overnight at 4°C.

FITC-conjugated goat anti-rabbit, anti-mouse, and CY5-onjugated goat anti-rabbit IgG antibodies were used as sec-ndary antibody (1:250 or 1:200, 60 min, room temperature).hodamine- or Alexa Fluor647– conjugated phalloidin (1:40,

olecular Probes, Eugene, OR) were added to counterstainlamentous actin to mark cell boundaries (data not shown forll results). Samples were mounted (Vectashield mountingedia with or without DAPI; Vector Laboratories) and as-

essed within the next 24 hours by using a laser-scanningonfocal microscope (Plan-Apochromat 63�/1.40 (oil) DICbjective, Zeiss Axiovert 100M-LSM 510; Carl Zeiss,berkochen, Germany). At least 5 individual sites of image

apture were chosen randomly in areas of uniform monolayerhickness for each sample. To establish comparable conditionsetween individual cell monolayers, equivalent images ofqual number of horizontal slices (1024 � 1024 pixels) withhe same vertical depth from apical tip to basal membraneetween nonstimulated and stimulated monolayers (as indi-ated in the Results section) were acquired. Height of intes-inal epithelial monolayers was used as an indirect marker ofell differentiation.14,29,30 All images were captured underdentical laser settings. Results were considered significantnly if more than 70% of the scanned sections per fieldxhibited the observed effect. Single XY planes (parallel to cellonolayer) and reconstructed XZ/YZ planes (orthogonal to

ell monolayer) were processed using standardized 2-color oronochannel settings (software LSM510 v3.2) and exported todobe Photoshop 5.0LE. All experiments were repeated at

east 3 times; representative results are shown for each exper-ment.

Assessment of Intestinal Epithelial BarrierFunction

Transepithelial resistance (TER) was used as a measuref paracellular permeability and barrier function in confluentaco-2 monolayers that were maintained on 6-well cellulose

ranswell permeable inserts (BD Falcon, Franklin Lakes, NJ;.4-�mol/L filter pore size, high density). Confluency ofaco-2 monolayers usually was reached within 5–10 days andells were used for experiments 21–30 days after seeding.wenty-four–36 hours before stimulation, cells were equili-rated with fresh medium with an apical and basolateralolume of each 2 mL. Each individual stimulation was per-ormed in triplicate with matched negative controls on theame 6-well plate. For TER measurements, Millicell-ERS ep-thelial volt-ohm meter (Millipore) was used under standard-zed conditions with electrodes equivalently placed and washedith warm medium between each measurement. After sub-

raction of the media and filter resistance (�145 �), TERalues were adjusted for the filter surface (4.2 cm2) and ex-ressed as � � cm2.

Plasmids, Transfection, and LuciferaseAssay

The flag-tagged dominant-negative TLR2 (TLR2DN)nd dominant-negative TLR4 (TLR4DN) expression plasmidsere kindly provided by Tularik Inc. (South San Francisco,A) through Dr. Carsten J. Kirschning (Institute of Medicalicrobiology, Immunology, and Hygiene, Technical Univer-

ity of Munich, Munich, Germany) and both were confirmedy sequencing (Sequencing Core Facility, Institute of Humanenetics, University Hospital of Essen) and restriction analy-

is. Mammalian TLRs contain a very divergent ligand-bindingctodomain and a highly conserved cytoplasmic signaling do-ain (Toll/interleukin-1R domain). Mutagenesis and func-

ional studies previously have shown that human TLRs inter-ct via Toll/interleukin-1R domain with downstreamignaling partners to transmit immune responses.31 Therefore,LR2DN and TLR4DN were generated by deletion of this

ntracellular signaling domain with truncation of the carboxyl-erminal portion of the wild-type molecule, implying thatverexpression of these nonsignaling proteins impairs activa-ion of downstream signaling pathways in response to specificctodomain-binding ligands, as described earlier.32–34 pCMV-cytomegalovirus)-Flag1 (Sigma) was used as appropriate con-rol vector. pAP1(PMA)-TA-Luciferase Vector35 was obtainedrom BD Clontech and pSV-�-Galactosidase Control Vectoras obtained from Promega. Plasmids were prepared using thendoFree Maxi plasmid kit (Qiagen, Hilden, Germany).T-29 or Caco-2 cells were transfected transiently overnight

Lipofectamine 2000; Invitrogen), according to the manufac-urer’s instructions, with optimal concentrations of the plas-ids (as specified later) in serum-free Optimum I media

Invitrogen), which was changed to full media the followingay after transfection. Cells were stimulated 2 days after trans-ection and subsequently assayed as indicated in the Resultsection.

For assessment of PKC phosphorylation (as described ear-ier), HT-29 cells (1–5 � 105/mL) were transfected withLR2DN or control vector at a final concentration of 1 �g/ell the following day after seeding. Caco-2 cells grown to00% confluency on inserts of 6-well plates were transfected

wtscfp

mwscdcancdadt8

gwpfiSswccv

paEo

TotcRfebe

sc

Cptsiaawustcm

Ftw�mncidlr(


ith TLR2DN, TLR4DN, or control vector at final concen-rations of 0.1 ng–0.4 �g/chamber (as indicated in the Resultsection). To normalize for individual transfections, proteinoncentrations were equalized36 and confirmed by reprobingor constitutively expressed �-actin (data not shown) andan-PKC.For measurements of TER (as described earlier), Caco-2onolayers fully differentiated onto inserts were transfectedith TLR2DN, TLR4DN, or control vector 14–21 days after

eeding at a final concentration of 0.4 �g/chamber in tripli-ate. Directly after TER measurements, cells were fixed (asescribed earlier) and filters were cut out. Transfection effi-iency subsequently was visualized by immunostaining withnti–Flag M2 monoclonal, as described earlier. The number ofuclei (DAPI) of at least 5 identical field sizes per sample wereounted and the percentage of green fluorescing cells wasetermined.37 Transiently transfected cells presented optimalnd consistent expression of flag-tagged deletion mutants un-er pCMV-promoter at day 2 posttransfection. The transfec-ion efficiency was estimated at an approximate range from0%–90% (data not shown).For activator protein-1 (AP-1) luciferase assays, Caco-2 cells

rown onto 6-well plates at 90% confluency were cotransfectedith 0.5 �g/well of pAP1(PMA)-TA-Luc and 0.1 �g/well ofSV �-galactosidase in triplicate. Cell lysates were assayed forrefly luciferase activity using the Luciferase Reporter Assayystem (Promega) in a luminometer (TD20/20; Turner Bio-ystems, Sunnyvale, CA) under standardized conditions. Dataere normalized to �-galactosidase activity (as an internal

ontrol for transfection efficiency), measured by the lumines-ent �-galactosidase detection kit II (BD Clontech). All indi-idual experiments were performed in triplicate.

Statistical Analysis

Data are expressed as mean � SD of 3 or more inde-endent experiments. Differences between means were evalu-ted by using the heteroscedastic, 2-sided t test (Microsoftxcel; Microsoft, Redmond, WA) where appropriate. P valuesf �0.05 were considered significant.

ResultsTLR2 Ligands Rapidly Phosphorylate PKCComplex

As previously shown, IECs constitutively expressLR2.8 A synthetic, protein-free, and sterile preparationf a TLR2-specific lipopeptide,38 Pam3CysSK4, was usedo study TLR2 stimulation to exclude effects of potentialontaminants39 that might contribute to PKC activation.esults were confirmed using phenol-extracted PGN

rom S. aureus, as a second TLR2-binding ligand.40 Toxclude that PGN-responsive NOD2/CARD1541,42 maye involved in PKC signaling, Caco-2 cells, which do notxpress NOD2,43 were used in all experiments and re-

ults were compared with those in the NOD2-expressingell line HT-29.

Stimulation of the 2 different IEC lines (HT-29,aco-2) with the 2 TLR2 ligands resulted in significanthosphorylation of PKC complex in a time- and concen-ration-dependent manner (representative blots arehown in Figure 1A, B). Both Pam3CysSK4 and PGNnduced significant phosphorylation of PKC complexfter 15 minutes of stimulation. Pam3CysSK4-inducedctivation returned to resting levels after 60 minutes,hereas PGN-induced activation lasted up to 120 min-tes (Figure 1A). It recently has been shown that thepecific activity of lipopeptides varies with environmen-al conditions (e.g., oxidation, de-esterfication,44 and mi-elle formation [according to the manufacturer]), whichay have caused the short and variable duration of

igure 1. TLR2 ligands induce phosphorylation of PKC complex in aime- and concentration-dependent manner. (A) Caco-2 monolayersere treated with Pam3CysSK4 (20 �g/mL), PGN (20 �g/mL), LPS (1g/mL), or PMA (1 �g/mL) for the indicated time periods (15–120in). (B) HT-29 monolayers were treated with different doses (10g–20 �g/mL) of TLR2 ligands, Pam3CysSK4 or PGN, for the indi-ated time periods (30 or 120 min, respectively). Cell lysates weremmunoblotted and probed with anti–phospho-panPKC antibody, asescribed in the Materials and Methods section. To confirm equaloading and to exclude effects on total PKC expression, blots wereeprobed with either anti-panPKC (nonphosphorylated) or anti–�-actindata not shown). *PMA-treated U937 cells.

bomarPp1uspcp(gpd

pc

i2ntrcaPwtdtTsaddb(P

nasimiiPpPszs(iio

FPlmtppttptrc

Fotmanapa


iologic activity of Pam3CysSK4 compared with PGN inur experiments. In the absence of TLR2 stimulation,inimal baseline PKC activity was detectable, presum-

bly reflecting primarily physical stress. Concentrationesponse experiments indicated that 10 ng/mL of eitheram3CysSK4 or PGN was sufficient to cause significanthosphorylation of PKC complex after 30 minutes or20 minutes, respectively (Figure 1B). In contrast, stim-lation with the TLR4 ligand LPS did not lead toufficient phosphorylation of PKC complex at any timeoint tested (Figure 1A). PMA broadly activates PKComplex in U937 cells,27 detecting several isoforms asart of the PKC complex between 78 and 85 kilodaltonsFigure 1A). However, in comparison with TLR2 li-ands, our results show that PMA strongly stimulatedhosphorylation of other isoforms of PKC complex withifferent kinetics in IEC.

TLR2 Ligands Specifically Activate 2Isoforms: PKC� and PKC�

To identify the involved isoforms of PKC com-lex activated in response to Pam3CysSK4, immunopre-ipitation studies were performed with specific antibod-

igure 2. TLR2 ligands specifically phosphorylate conventionalKC�/�II and novel PKC�/�. (A) Caco-2 monolayers were treated with

ow-dose Pam3CysSK4 (100 ng/mL) for various time periods (5–120in) and subjected to immunoprecipitation followed by immunoblot-

ing with isoform-specific antibodies to phosphorylated PKC�/�II orhosphorylated PKC�/�. PMA (500 ng/mL) was used as presumedositive control (60-min stimulation). (B) Caco-2 monolayers werereated with Pam3CysSK4 (20 �g/mL) or PMA (1 �g/mL) for indicatedime periods (15–120 min). Cell lysates were immunoblotted androbed with anti–phospho-PKC�/� antibody, as described in the Ma-erials and Methods section. To confirm equal loading, blots wereeprobed with anti-PKC� (nonphosphorylated). *PMA-treated U937ells.

es to individual phosphorylated PKC isoforms (FigureA). Addition of small amounts of Pam3CysSK4 (100g/mL) led to significant time-dependent phosphoryla-ion of conventional PKC�/�II and novel PKC�/�, botheaching maximum peaks after 30 minutes and thusonfirming initial kinetic results with the antibodygainst the broad range of phosphorylated PKC isoforms.hosphorylation of both PKC isozyme groups alreadyas evident after 15 minutes of stimulation. By using

his sensitive method, phosphorylation of PKC�/�II wasetectable for up to 120 minutes, whereas phosphoryla-ion of PKC�/� no longer was evident after 45 minutes.he positive control (PMA-treated U937 cell extracts)

upplied by the manufacturer confirmed specificity of thentibody detecting phosphorylated PKC�/� at 78 kilo-altons (Figure 2B). In contrast to Pam3CysSK4, PMAid not lead to significant phosphorylation of PKC�/�etween 15 to 120 minutes of stimulation in IECsFigure 2A, B), implying selective activation of specificKC isoforms by different ligands in different cell lines.Specific ligand-induced activation of PKC isoforms is

ot only reflected by the status of phosphorylation butlso by the level of enzymatic activity and evidence ofubcellular translocation.45 To confirm that Pam3CysSK4ndeed activates specific PKC isoforms, 2 additionalethodologic approaches were included in our studies:

n vitro kinase assay for enzymatic activity and confocalmmunohistochemistry for subcellular localization ofKC isoforms. In vitro kinase assays confirmed thathosphorylation initiated enzymatic activation of the 2KC isoforms � and � in response to Pam3CysSK4. Ashown in Figure 3, Pam3CysSK4 led to significant en-ymatic activation of PKC� and � after 60 minutes oftimulation, which already was evident after 30 minutesdata not shown). Consistent with the preceding findingn Figure 2, stimulation with PMA (1 �g/mL, 60 min)nduced significant enzymatic activation of PKC�, butnly minimally of PKC�.

igure 3. The TLR2 ligand Pam3CysSK4 induces enzymatic activationf conventional PKC� and novel PKC�. HT-29 monolayers werereated with Pam3CysSK4 (20 �g/mL) or PMA (1 �g/mL) for 60inutes. Cell lysates were immunoprecipitated with isoform-specificntibodies and subjected to in vitro kinase reaction with the PepTagonradioactive PKC assay for assessment of PKC� and PKC� activity,s described in the Materials and Methods section. Negative andositive controls represent inactive and active PKC included in thessay.

umdaPwfaclt

iN

is

tHmaoPwmt

Fcvmt(iP


In parallel, translocation of both isoforms, PKC� (Fig-re 4A) and PKC� (Figure 4B), from cytosolic compart-ents to apical membrane and subapical cytoplasmic

omains was apparent at 30 minutes (data not shown)nd maximal at 60 minutes after addition ofam3CysSK4 in Caco-2 monolayers—regardless ofhether they were differentiated (Figure 4A) or nondif-

erentiated (Figure 4B). In nontreated monolayers (neg-tive control), both PKC isoforms were dispersed in theytosolic compartments without distinct apical or baso-ateral preference. In contrast, PMA induced transloca-ion of PKC�, but not of PKC�, in Caco-2 monolayers.

As shown in Figure 1A, the TLR4 ligand LPS did notnduce significant phosphorylation of PKC complex.either enzymatic activation nor translocation of PKC

igure 4. The TLR2 ligand Pam3CysSK4 induced apical translocatioonfluent and treated with Pam3CysSK4 (20 �g/mL) or PMA (1 �g/misualized by immunostaining with isoform-specific nonphosphorylatedicroscopy, as described in the Materials and Methods section. Nega

o outline cell boundaries (data not shown). Vertical height of cell moA) differentiated, (B) nondifferentiated.14,29,30 Representative imagendicate location of XZ/YZ-stacks in XY-stack, as well as location of XYKC-isoform. (63�/1.4 oil, scan zoom 1.0, total scanning depth of 1

soforms in response to LPS was detected (data nothown).

Activation of PKC Is Mediated Directly ViaTLR2To confirm that TLR2 ligand–induced PKC ac-

ivation was mediated directly via TLR2, Caco-2 orT-29 cells were transfected transiently with deletionutants of TLR2, TLR4, or appropriate control vector

nd incubated in the presence or absence of Pam3CysSK4r PGN (Figure 5). Although it has been shown thatMA-induced myeloid cell differentiation correlatesith up-regulation of TLR2 messenger RNA and TLR4essenger RNA expression after long-term stimula-

ion,46 PMA is not known to bind directly or activate

PKC� and PKC�. Caco-2 cells were grown on chamber slides until60 minutes and translocation of (A) PKC� or (B) PKC� isoforms wasbodies followed by FITC-conjugated secondary antibody and confocalontrols were left untreated. Cells were counterstained with phalloidiner was used as an indirect marker of the state of cell differentiation:XZ, YZ, and XY stacks are shown (monochannel: FITC). Grey linesk in XZ/YZ-stacks, respectively. White arrow indicates localization ofZ-stacks per monolayer is indicated per individual image.)

n ofL) for

antitive cnolays of-stac5–30

TPTvttpc�cspcct(igt

nfieTaTn

pMutgiPda

Fcnmat


LR2 or TLR4. Accordingly, PMA-induced activation ofKC complex neither was prevented nor altered byLR2DN or TLR4DN when compared with controlector (Figure 5A) or nontransfected cells (Figure 1A),hus excluding nonspecific inhibitory effects caused byransfection. However, Pam3CysSK4-induced rapidhosphorylation of PKC complex was blocked signifi-antly in TLR2DN-transfected cells (Figure 5B, 0.4g/chamber; Figure 5C, 0.01 �g/chamber). A lesser

oncentration of only 0.1 ng/chamber of TLR2DN wasuboptimal because Pam3CysSK4-induced PKC phos-horylation was not inhibited (Figure 5B), suggesting aoncentration-dependent mutant-specific effect. In theontrol experiment, Pam3CysSK4-induced phosphoryla-ion of PKC complex was not blocked by TLR4DNFigure 5C), implying that TLR2 is specifically mediat-ng PKC phosphorylation in response to the TLR2 li-and Pam3CysSK4. In addition, PGN-induced activa-ion of PKC was blocked significantly by TLR2DN, but

igure 5. TLR2 ligand–induced phosphorylation of PKC complex is bloell monolayers were transfected transiently with dominant-negativeg–1.0 �g/well), as described in the Materials and Methods section.in), (B, C) Pam3CysSK4 (20 �g/mL, 30 min), or (D) PGN (20 �g/mLnti–phospho-panPKC antibody, as described in the Materials andransfection and to confirm equal loading, blots were reprobed with a

ot by control vector in HT-29 cells (Figure 5D), con-rming that PKC activation is a TLR2-specific signalingffect. Kinetics of PKC phosphorylation induced byLR2 ligands were comparable between nontransfectednd control (TLR4DN or pCMV) transfected cells.ransfection did not alter total protein expression ofonphosphorylated PKC (Figure 5A–D).

Divergent Downstream Effects Via PKC

Downstream, Pam3CysSK4 led to rapid phos-horylation of the PKC-specific endogenous substrateARCKS after 15 and 30 minutes of stimulation (Fig-

re 6A). PKC activation is known to lead to activation ofhe transcriptional factor AP-1.47 However, TLR2 li-ands did not induce AP-1–dependent luciferase activityn Caco-2 cells (Figure 6B). In contrast, the phorbol esterMA (1 �g/mL) induced a 3-fold increase in AP-1–ependent luciferase activity (326% � 44% control)fter 4 hours of stimulation (Figure 6B), which was

specifically by dominant-negative TLR2. (A–C) Caco-2 and (D) HT-29ns of TLR2 (TLR2DN), TLR4 (TLR4DN), or pCMV vector control (0.1ays after transfection, cells were treated with (A) PMA (1 �g/mL, 30min), respectively. Cell lysates were immunoblotted and probed with

thods section. To exclude alteration of PKC expression owing toanPKC and anti–�-actin (data not shown), respectively.

ckedversioTwo d, 120

Menti-p

bcGp�ciPpfedd

l

FadPi�deDp#0§PGrmnAwDpa*

Fs(iM(wcaa(llnPt#P


locked almost completely by pretreating (30 min) theells with the conventional PKC isoform inhibitoro6976 (5 �mol/L48; 112% � 33% control), but onlyartially with the specific PKC� inhibitor rottlerin (10mol/L49; 169% � 30% control) (data not shown). In

ontrast to Pam3CysSK4, which may be inactivated eas-ly by, for example, oxidation in cell culture media,44

MA appears to exhibit a longer half-life, resulting inrolonged or delayed activation of certain PKC iso-orms.50 Thus, these results suggest that biologic differ-nces in kinetics of stimulus-dependent activation ofistinct PKC-isoform patterns may lead to contrastingownstream signaling effects in IEC.

igure 6. TLR2 ligands induce selective downstream effects via PKCubstrates. (A) Caco-2 monolayers were treated with Pam3CysSK420 �g/mL) for varying time periods (15–120 min) and subjected tommunoblotting with isoform-specific antibody to phosphorylatedARCKS. PMA (1 �g/mL) was used as presumed positive control

30-min stimulation). To confirm equal loading, blots were reprobedith anti–�-actin. (B) Caco-2 monolayers grown on 6-well plates wereotransfected with pAP1-specific luciferase construct (0.5 �g/well)nd pSV–�-galactosidase expression vector (0.1 �g/well). Two daysfter transfection, confluent cells were stimulated with Pam3CysSK410 �g/mL), PGN (10 �g/mL), or PMA (1 �g/mL) for 4 hours. Modu-ation of AP-1 activity was determined by luciferase assay in cellysates, normalized to �-galactosidase activity, and compared withegative controls, as described in the Materials and Methods section.resented data reflect 3 independent experiments, each performed inriplicate per condition. Data are expressed as means � SD% control.Pam3CysSK4 or PGN vs. negative control, P � 0.05; �PMA vs.am3CysSK4 or PGN, P � 0.001.

TLR2-Induced PKC Activation Leads toIncrease of TER in IECs

PKC activation has been implicated in the regu-ation of epithelial barrier integrity.2,51,52 To assess

igure 7. (A) The TLR2 ligand Pam3CysSK4 increases intestinal TER inPKC-dependent pathway. Caco-2 grown on inserts was used 21–30

ays after seeding. After apical and basolateral stimulation witham3CysSK4 (20 �g/mL), with or without pretreatment of specific PKC

soform antagonists rottlerin (10 �mol/L) or Go6976 (5 �mol/L), LPS (1g/mL) or PMA (1 �g/mL) TER was measured under standardized con-itions over indicated time periods with matched negative controls onach 6-well plate, as described in the Materials and Methods section.ata reflect at least 3 independent experiments per condition; each wereerformed in triplicate. Data are expressed as means � SD% control.Pam3CysSK4, 0 vs. 30, 120 min, P � 0.05; 0 vs. 60 minutes, P �.01. °Pam3CysSK4/rottlerin, 0 vs. 30, 60, 120 min, P � 0.17.Pam3CysSK4/Go6976, 0 vs. 30, 60, 120 min, P � 0.16. ■ ,AM3CysSK4; , PAM3CysSK4 � rottlerin; , PAM3CysSK4 �o6976; �, LPS; , PMA. (B) TLR2 ligand–increased transepithelial

esistance is blocked specifically by dominant-negative TLR2. Caco-2onolayers grown on permeable inserts were transfected with dominant-egative versions of TLR2 (TLR2DN) or TLR4 (TLR4DN) (0.4 �g/chamber).fter apical and basolateral stimulation with Pam3CysSK4 (20 �g/mL), TERas measured under standardized conditions over indicated time periods.ata comprise at least 3 independent experiments per condition, eacherformed in triplicate with matched negative controls. Data are expresseds means � SD% control. (TLR4DN- vs. TLR2DN-transfected monolayers:P � 0.01; #P � 0.001; °P � 0.05.) ■ , TLR2DN; X, TLR4DN.

wfwpwnPitc2mawnawc3nmpacs0swi1tit

cfalwthtrnuc(nHag

TidesflnsaePmfdsTta(t�w

fnlvoida1fcHaistcttd

n


hether Pam3CysSK4-induced PKC activation may af-ect intestinal epithelial barrier function, Caco-2 cellsere grown on inserts and TER was measured in theresence or absence of Pam3CysSK4 (20 �g/mL) with orithout pretreatment of PKC isoform–selective antago-ists. As shown in Figure 7A, stimulation witham3CysSK4 (20 �g/mL) rapidly led to an almost 2-fold

ncrease of TER after 30 minutes (165% � 36% con-rol), which peaked after 60 minutes (180% � 27%ontrol) in parallel with PKC activation (Figures 1 and). TER decreased back toward baseline levels after 120inutes of stimulation (133% � 14% control) in par-

llel with PKC inactivation. No further TER changesere observed when followed-up for over 24 hours (dataot shown). The Pam3CysSK4-induced TER increase waslmost completely abolished when cells were pretreatedith the PKC inhibitor rottlerin (60 min, 115% � 6%

ontrol) or Go6976 (60 min, 124% � 11% control) for0 minutes, suggesting a role for both conventional andovel PKC isoforms in Pam3CysSK4-induced enhance-ent of intestinal epithelial barrier integrity. In com-

arison, the PKC activator PMA (1 �g/mL) induced only1.4-fold increase of TER (60 min, 135% � 24%

ontrol; 120 min, 128% � 18% control), which was notignificant (0 vs. 60 min, P � 0.27; 0 vs. 120 min, P �.32) and was comparable with recently published re-ults by others.52 In contrast, the TLR4 ligand LPS,hich did not activate PKC complex, also did not affect

ntestinal epithelial barrier function (60 min, 105% �0% control; 0 vs. 60 min, P � 0.81) at any time pointested for up to 24 hours (data only shown up to 2 h),mplying selective specificity of TER increase in responseo TLR2 ligands via PKC.

To confirm that the Pam3CysSK4-induced TER in-rease was mediated via TLR2, Caco-2 cells were trans-ected with TLR2DN or TLR4DN and TER was assessedfter stimulation with Pam3CysSK4 (Figure 7B). Mono-ayers of the transfected cells showed steady-state TER,ith comparable baseline readings to age-matched non-

ransfected epithelia when followed-up for up to 120ours after transfection (data not shown). In TLR4DN-ransfected cells, the Pam3CysSK4-induced TER increaseeached a maximum of 148% � 17% compared withonstimulated TLR4DN-transfected cells after 60 min-tes of stimulation, which was not altered significantlyompared with that observed in nontransfected epitheliaPam3CysSK4-stimulation, TLR4DN-transfected vs.ontransfected, 0, 30, 60, and 120 min, P � 0.08).owever, the Pam3CysSK4-induced TER increase was

bolished completely in TLR2DN-transfected cells, sug-esting that this effect specifically requires TLR2.

TLR2-Induced TER Increase Correlates WithZO-1 Redistribution Via PKC

To assess changes of subcellular distribution ofJ-associated proteins that may occur during TLR2-

nduced increase of TER, confocal microscopy of highlyifferentiated Caco-2 monolayers was performed afterxposure to Pam3CysSK4 (20 �g/mL, 60 min). Aftertimulation, membrane-associated ZO-1 redistributed tourther apical TJ areas forming distinct and consolidatedateral cell-cell contacts in comparison with same-depthegative controls (85–90 �m) analyzed by (X/Y-)Z-ection confocal immunofluorescence (Figure 8A). Alter-tion of ZO-1 distribution was confirmed in all imagesxamined and correlated in a timely manner witham3CysSK4-induced TER increase. When pretreatingonolayers with PKC inhibitors (Go6976 and rottlerin)

or 30 minutes, Pam3CysSK4-induced alteration of ZO-1istribution was blocked significantly after 60 minutes oftimulation (Figure 8B), correlating with inhibition ofER increase via PKC (Figure 6). In contrast, stimula-

ion with Pam3CysSK4 did not induce any evident TJ-ssociated morphologic changes of occludin or claudin-1Figure 8A). Pam3CysSK4 also did not affect organiza-ion of the actin cytoskeleton. In contrast, PMA (1g/mL, 60 min) induced disruption of ZO-1 (Figure 9),hich correlated with a lack of significant TER increase.

DiscussionThe present study provides evidence that PKC

unctionally participates in the mammalian TLR2 sig-aling pathway in IECs. Our results show that TLR2igands lead to concentration- and time-dependent acti-ation (i.e., phosphorylation, translocation, and increasef enzymatic activity) of at least 2 PKC-specific isoformsn IECs: conventional PKC� and novel PKC�. In 2ifferent human IEC lines, activation of these kinasesfter stimulation with TLR2 ligands was maximal at5–30 minutes for phosphorylation and at 60 minutesor translocation, consistent with observations in otherell lines in response to the TLR4 ligand LPS.23,53,54

owever, in contrast, our results also suggest that PKCctivation is not involved in TLR4 signaling in IECs,mplying ligand-dependent TLRx-specific features ofignaling via distinct PKC isoforms in different cellypes. Phosphorylation of PKC complex was blockedompletely by transfection with a TLR2 deletion mu-ant, but not with a TLR4 deletion mutant, confirminghat TLR2-ligand–stimulated activation of PKC is me-iated specifically through TLR2 in IECs.

Based on the observation that TLR2 ligands induceeither nuclear factor � B activation nor interleukin 8

sglfrtAvu

noeitdPaPf

([tTbmpt

tdscdPwibls

Fd(acaXo s pe


ecretion in some IEC lines, it recently has been sug-ested that IECs are broadly hyporesponsive to TLR2igands.16 AP-1 is an abundantly expressed transcriptionactor that is central to several immune and inflammatoryesponses mediated via PKC.47 In this study, we showhat TLR2-induced PKC activation did not stimulateP-1 transcriptional activity in IECs. Instead, we pro-ide evidence that TLR2 may stimulate other, previouslynappreciated, functional responses via PKC.PKC has been implicated in the regulation of intesti-

al epithelial integrity.2,51,52 Ligand-specific activationf distinct patterns of PKC isoforms can exert paradoxicffects in intestinal epithelial monolayers: either increas-ng or decreasing intestinal permeability.2,48,52,55–57 Inhis study, Pam3CysSK4 treatment led to a time-depen-ent increase in TER via TLR2, which correlated withKC�/� activation. The TER increase was significantlybolished when pretreating monolayers with the specificKC isoform inhibitors: rottlerin (a specific antagonist

or PKC� in the concentrations used here58) and Go6976

igure 8. TLR2 ligand–induced increase of transepithelial resistancependent manner. (A, B) Caco-2 cells grown on chamber slides untilB) with or (A) without pretreatment of specific PKC-isoform antagonisssembly was assessed by specific antibodies to (A, B) ZO-1, (A) claonjugated secondary antibody and confocal microscopy, as describednd XY-stacks are shown. Red and green lines indicate location oZ/YZ-stacks, respectively. (B) Cells were counterstained with phalloidf ZO-1. (63�/1.4, oil, zoom 1.0, total scanning depth of 30 Z-stack

a selective inhibitor of the conventional PKC subfamilyPKC�, �, �]59), suggesting that TLR2-induced activa-ion of conventional and novel PKC isoforms increasesER. Loss of TLR2 function did not induce impairedarrier function in nonstimulated intestinal epithelialonolayers, as shown by steady-state TER readings com-

arable with those in nontransfected monolayers consti-utively expressing TLR2.

Little is known about the processes modulating dis-inct TJ assembly when TER increases in IECs. Stan-ardized confocal analysis by uniform Z sections to de-cribe morphologic changes of intestinal epithelial TJontacts mostly is used in studies when evaluating TERecreases rather than increases. Activation of certainKC isoforms may involve the release of calcium stores,hich is critical for TJ formation and tightness.51,60 PKC

soforms specifically may regulate the sorting and assem-ly of ZO-1, preserving proper development of TJs andeading to an increase of TER.2,61 This study showed thattimulation with the TLR2 ligand Pam CysSK4 results

rrelates with apical redistribution and tightening of ZO-1 in a PKC-rentiated were treated with Pam3CysSK4 (20 �g/mL) for 60 minutesttlerin (10 �mol/L) and Go6976 (5 �mol/L) for 30 minutes, and TJ

1, or (A) occludin followed by FITC- (ZO-1; claudin-1)/CY5- (occludin)e Materials and Methods section. Representative images of XZ-, YZ-,

/YZ-stacks in XY-stack, blue lines indicate location of XY-stack inodamine to outline cell boundaries. White arrow indicates localizationr monolayer is indicated per individual image.)

e codiffets, roudin-in th

f XZin-rh

3

iaveiPvtdilitttcfpHmma

wwp

tMptcTrim

deptgfluafttodlsiwrfwemtatstw

iieTnpiedcp

FiPsaaoXi1i


n apical tightening of ZO-1, which correlated withctivation of PKC�/� and PKC-mediated TER increaseia TLR2 in Caco-2 cells. Of note, a similar morphologicffect on ZO-1 distribution so far has not been describedn response to other PKC activators in IECs.am3CysSK4-induced translocation of ZO-1 was pre-ented by pretreatment with selective PKC inhibitorshat correlated with the inhibition of Pam3CysSK4-in-uced TER increase. This finding suggests that TLR2-nduced activation of distinct PKC-specific isoformseads to a sealing pattern of apical ZO-1, resulting in anncrease in TJ-associated barrier integrity. It is possiblehat TLR2 ligand-induced ZO-1 translocation couldrigger the tightening of other interconnecting TJ con-acts between adjacent epithelial cells resulting in in-reased TER. It recently has been suggested that barrierunction may be augmented by enhancing the TJ com-lex via recruitment49 and up-regulation62 of claudin-1.owever, any other Pam3CysSK4-induced TJ alterationsust be subtle, because no changes in distribution of 2ajor TJ proteins, claudin-1 and occludin, were detect-

ble.Moreover, reorganization of the actin cytoskeleton,

hich also may be regulated by specific PKC isoforms,51

as absent. However, we found that Pam3CysSK4 led tohosphorylation of MARCKS, one of the major cellular

igure 9. PMA-induced disruption of ZO-1 correlates with lack of TERncrease. Caco-2 cells grown on chamber slides were treated withMA (1 �g/mL, 60 min) or left untreated and ZO-1 was assessed bytaining with a specific antibody followed by FITC-conjugated second-ry antibody and confocal microscopy, as described in the Materialsnd Methods section. Cells were counterstained with phalloidin toutline cell boundaries (data not shown). Representative images ofZ-, YZ-, and XY-stacks are shown (monochannel: FITC). White arrow

ndicates disruption of ZO-1 in response to PMA stimulation. (63�/.4, oil, zoom 1.0, total scanning depth of 30 Z-stacks per monolayer

s indicated per individual image.)

argets of PKC isoforms. It recently has been shown thatARCKS also serves as an actin cross-linking protein,63

otentially participating in modulation of epithelial in-egrity. ZO-1 also may bind components of the TJomplex to the actin cytoskeleton.64 It is possible thatLR2-induced intestinal epithelial barrier enhancement

esults from potential convergence of various signalingnteractions and pathways downstream of PKC, whichay interact by bridging to the actin cytoskeleton.Because the most prominent intracellular targets of

iacylglycerol and of the functionally analogous phorbolsters belong to the PKC family,65 PMA has been used asresumed positive control of broad PKC isoform activa-ion in this study, although the current knowledge re-arding patterns and effects of PMA-induced PKC iso-orm activation is limited in IECs. In contrast to TLR2igands, PMA-induced phosphorylation of PKC complexsed other PKC isoforms with different kinetics. It islready well known that differences between PKC iso-orms for their ligand and substrate specificity may leado different physiologic responses in individual cellypes.66 Consequently, PMA-induced selective activationf PKC complex led to downstream effects that wereistinct from those seen after treatment with TLR2igands (e.g., AP-1 activation). It recently has beenhown that intestinal epithelial monolayer disassemblyn response to prolonged exposure to PMA correlatesith PKC-associated alterations in the perijunctional

ing of actin and myosin,55 which may have accountedor our observation of disruption of ZO-1 correlatingith a lack of TER increase. There also is emerging

vidence that several cellular processes induced by PMAay depend on non–PKC–triggered targets, disproving

he notion that all phorbol-ester effects must be medi-ted solely by PKC isozymes.65,67 PMA-specific induc-ion of different PKC- as well as non–PKC-associatedignaling events may have contributed significantly tohese contrasting downstream effects when comparedith TLR2 ligands in IECs.Taken together, our data suggest that a specific phys-

ologic consequence of the increase in certain PKCsozyme activation by TLR2 is enhancement of intestinalpithelial barrier function, which correlates with distinctJ-associated morphologic changes. Further studies areeeded to clarify in detail the TLR2-triggered Ca2�

athways that may regulate the complex TJ-associatednteractions of membrane signaling partners leading tonhancement of barrier integrity. It will be important toetermine whether PKC isoforms are central downstreamomponents of both the TIRAP/Mal- and MyD88-de-endent signaling pathways via TLR2.68,69 Because PKC

itpocctPoionmT

ipttcsbttitsriiet

tPrtmstid

1

1

1

1

1

1

1

1

1

1

2


s known to comprise a large family of enzymes, addi-ional isoforms may be involved in TLR2-induced phos-horylation of PKC complex. More importantly, parallelr subsequent activation of opposing PKC isoforms mayause differential downstream effects in epithelialells.48,51,70 In this context, further studies are requiredo identify whether other TLRx ligands activate differentKC isoforms, cooperatively or competitively increasingr decreasing intestinal epithelial barrier permeability. Its possible that bacterial ligands may counterbalance eachther via TLRs, maintaining barrier homeostasis of theormal intestinal epithelium. It remains to be deter-ined whether pathogenic bacteria-induced decreases inER also may be mediated via TLRx.Our results suggest that impaired function of TLR2

tself does not lead to increased intestinal epithelialermeability in vitro. In this context, it will be essentialo show whether lack of host-protective TLR2 ligands inhe lumen that may be present in the resident microfloraould lead to loss of barrier protection, facilitating inva-ion of pathogenic bacteria in disease. It recently haseen shown that probiotics, which may help some pa-ients with inflammatory bowel disease,71 enhance intes-inal epithelial barrier function.6 Although the underly-ng mechanisms remain to be determined, it is possiblehat certain probiotic compounds may contain TLR2-pecific immunostimulatory features, leading to amelio-ation of colitis by restoring intestinal epithelial barrierntegrity. Specific targeting of TLR2 possibly could helpn the design of novel adjuvant therapeutic means tonhance intestinal epithelial barrier function to protecthe underlying host.

In conclusion, the present study describes the induc-ion of a novel innate immune pathway through certainKC isoforms by distinct bacterial ligands, criticallyegulating intestinal epithelial barrier function by selec-ive rearrangement of TJ-associated ZO-1 via a specificechanism dependent from TLR2 signaling. Further

tudies are needed to investigate whether imbalance ofhe resident microflora may lead to barrier dysfunctionnduced by TJ disassembly through TLRx-PKC isozymeysregulation.

References1. Mitic LL, Van Itallie CM, Anderson JM. Molecular physiology and

pathophysiology of tight junctions I. Tight junction structure andfunction: lessons from mutant animals and proteins. Am J Physiol2000;279:G250–G254.

2. Stuart RO, Nigam SK. Regulated assembly of tight junctions byprotein kinase C. Proc Natl Acad Sci U S A 1995;92:6072–6076.

3. Berkes J, Viswanathan VK, Savkovic SD, Hecht G. Intestinalepithelial responses to enteric pathogens: effects on the tightjunction barrier, ion transport, and inflammation. Gut 2003;52:439–451.

4. Kindon H, Pothoulakis C, Thim L, Lynch-Devaney K, Podolsky DK.Trefoil peptide protection of intestinal epithelial barrier function:cooperative interaction with mucin glycoprotein. Gastroenterology1995;109:516–523.

5. Czerucka D, Dahan S, Mograbi B, Rossi B, Rampal P. Saccharomy-ces boulardii preserves the barrier function and modulates thesignal transduction pathway induced in enteropathogenic Esche-richia coli-infected T84 cells. Infect Immun 2000;68:5998–6004.

6. Madsen K, Cornish A, Soper P, McKaigney C, Jijon H, Yachimec C,Doyle J, Jewell L, De Simone C. Probiotic bacteria enhance mu-rine and human intestinal epithelial barrier function. Gastroenter-ology 2001;121:580–591.

7. Janeway CA Jr, Medzhitov R. Innate immune recognition. AnnuRev Immunol 2002;20:197–216.

8. Cario E, Rosenberg IM, Brandwein SL, Beck PL, Reinecker HC,Podolsky DK. Lipopolysaccharide activates distinct signalingpathways in intestinal epithelial cell lines expressing Toll-likereceptors. J Immunol 2000;164:966–972.

9. Cario E, Podolsky DK. Differential alteration in intestinal epithelialcell expression of toll-like receptor 3 (TLR3) and TLR4 in inflam-matory bowel disease. Infect Immun 2000;68:7010–7017.

0. Gewirtz AT, Navas TA, Lyons S, Godowski PJ, Madara JL. Cuttingedge: bacterial flagellin activates basolaterally expressed TLR5to induce epithelial proinflammatory gene expression. J Immunol2001;167:1882–1885.

1. Abreu MT, Vora P, Faure E, Thomas LS, Arnold ET, Arditi M.Decreased expression of Toll-like receptor-4 and MD-2 correlateswith intestinal epithelial cell protection against dysregulatedproinflammatory gene expression in response to bacterial lipo-polysaccharide. J Immunol 2001;167:1609–1616.

2. Fusunyan RD, Nanthakumar NN, Baldeon ME, Walker WA. Evi-dence for an innate immune response in the immature humanintestine: toll-like receptors on fetal enterocytes. Pediatr Res2001;49:589–593.

3. Haller D, Russo MP, Sartor RB, Jobin C. IKK beta and phospha-tidylinositol 3-kinase/Akt participate in non-pathogenic gram-neg-ative enteric bacteria-induced RelA phosphorylation and NF-kappa B activation in both primary and intestinal epithelial celllines. J Biol Chem 2002;277:38168–38178.

4. Cario E, Brown D, McKee M, Lynch-Devaney K, Gerken G, Podol-sky DK. Commensal-associated molecular patterns induce selec-tive toll-like receptor-trafficking from apical membrane to cyto-plasmic compartments in polarized intestinal epithelium. Am JPathol 2002;160:165–173.

5. Cario E. Toll-like receptors and gastrointestinal disease: frombench to bedside? Curr Opin Gastroenterol 2002;18:696–704.

6. Melmed G, Thomas LS, Lee N, Tesfay SY, Lukasek K, MichelsenKS, Zhou Y, Hu B, Arditi M, Abreu MT. Human intestinal epithelialcells are broadly unresponsive to Toll-like receptor 2-dependentbacterial ligands: implications for host-microbial interactions inthe gut. J Immunol 2003;170:1406–1415.

7. Otte JM, Cario E, Podolsky DK. Mechanisms of cross-hyporespon-siveness to Toll-like receptor ligands in intestinal epithelial cells.Gastroenterology 2004;126:1054–1070.

8. Suzuki M, Hisamatsu T, Podolsky DK. Gamma interferon aug-ments the intracellular pathway for lipopolysaccharide (LPS) rec-ognition in human intestinal epithelial cells through coordinatedup-regulation of LPS uptake and expression of the intracellularToll-like receptor 4-MD-2 complex. Infect Immun 2003;71:3503–3511.

9. Abreu MT, Arnold ET, Thomas LS, Gonsky R, Zhou Y, Hu B, ArditiM. TLR4 and MD-2 expression is regulated by immune-mediatedsignals in human intestinal epithelial cells. J Biol Chem 2002;277:20431–20437.

0. Newton AC. Protein kinase C: structure, function, and regulation.J Biol Chem 1995;270:28495–28498.

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

5

5

5

5


1. Dekker LV, Parker PJ. Protein kinase C—a question of specificity.Trends Biochem Sci 1994;19:73–77.

2. Avila A, Silverman N, Diaz-Meco MT, Moscat J. The Drosophilaatypical protein kinase C-ref(2)p complex constitutes a conservedmodule for signaling in the toll pathway. Mol Cell Biol 2002;22:8787–8795.

3. Aksoy E, Amraoui Z, Goriely S, Goldman M, Willems F. Critical roleof protein kinase C epsilon for lipopolysaccharide-induced IL-12synthesis in monocyte-derived dendritic cells. Eur J Immunol2002;32:3040–3049.

4. Rhee SH, Jones BW, Toshchakov V, Vogel SN, Fenton MJ. Toll-likereceptors 2 and 4 activate STAT1 serine phosphorylation bydistinct mechanisms in macrophages. J Biol Chem 2003;278:22506–22512.

5. Castrillo A, Pennington DJ, Otto F, Parker PJ, Owen MJ, Bosca L.Protein kinase Cepsilon is required for macrophage activationand defense against bacterial infection. J Exp Med 2001;194:1231–1242.

6. Franchi-Gazzola R, Visigalli R, Bussolati O, Gazzola GC. Involve-ment of protein kinase Cepsilon in the stimulation of anionicamino acid transport in cultured human fibroblasts. J Biol Chem1996;271:26124–26130.

7. Deszo EL, Brake DK, Cengel KA, Kelley KW, Freund GG. CD45negatively regulates monocytic cell differentiation by inhibitingphorbol 12-myristate 13-acetate-dependent activation and ty-rosine phosphorylation of protein kinase Cdelta. J Biol Chem2001;276:10212–10217.

8. Fantini J, Verrier B, Marvaldi J, Mauchamp J. In vitro differentiatedHT 29-D4 clonal cell line generates leakproof and electricallyactive monolayers when cultured in porous-bottom culturedishes. Biol Cell 1989;65:163–169.

9. Peterson MD, Mooseker MS. An in vitro model for the analysis ofintestinal brush border assembly. I. Ultrastructural analysis ofcell contact-induced brush border assembly in Caco-2BBe cells.J Cell Sci 1993;105:445–460.

0. Peterson MD, Bement WM, Mooseker MS. An in vitro model forthe analysis of intestinal brush border assembly. II. Changes inexpression and localization of brush border proteins during cellcontact-induced brush border assembly in Caco-2BBe cells. J CellSci 1993;105:461–472.

1. Xu Y, Tao X, Shen B, Horng T, Medzhitov R, Manley JL, Tong L.Structural basis for signal transduction by the Toll/interleukin-1receptor domains. Nature 2000;408:111–115.

2. Smith MF Jr, Mitchell A, Li G, Ding S, Fitzmaurice AM, Ryan K,Crowe S, Goldberg JB. Toll-like receptor (TLR) 2 and TLR5, but notTLR4, are required for Helicobacter pylori-induced NF-kappa Bactivation and chemokine expression by epithelial cells. J BiolChem 2003;278:32552–32560.

3. Faure E, Equils O, Sieling PA, Thomas L, Zhang FX, Kirschning CJ,Polentarutti N, Muzio M, Arditi M. Bacterial lipopolysaccharideactivates NF-kappaB through toll-like receptor 4 (TLR-4) in cul-tured human dermal endothelial cells. Differential expression ofTLR-4 and TLR-2 in endothelial cells. J Biol Chem 2000;275:11058–11063.

4. Bulut Y, Faure E, Thomas L, Karahashi H, Michelsen KS, EquilsO, Morrison SG, Morrison RP, Arditi M. Chlamydial heat shockprotein 60 activates macrophages and endothelial cells throughToll-like receptor 4 and MD2 in a MyD88-dependent pathway.J Immunol 2002;168:1435–1440.

5. Ahn JD, Morishita R, Kaneda Y, Lee SJ, Kwon KY, Choi SY, LeeKU, Park JY, Moon IJ, Park JG, Yoshizumi M, Ouchi Y, Lee IK.Inhibitory effects of novel AP-1 decoy oligodeoxynucleotides onvascular smooth muscle cell proliferation in vitro and neointimalformation in vivo. Circ Res 2002;90:1325–1332.

6. Heinz S, Haehnel V, Karaghiosoff M, Schwarzfischer L, Muller M,Krause SW, Rehli M. Species-specific regulation of Toll-like re-

ceptor 3 genes in men and mice. J Biol Chem 2003;278:21502–21509.

7. Knight A, Carvajal J, Schneider H, Coutelle C, Chamberlain S,Fairweather N. Non-viral neuronal gene delivery mediated by theHC fragment of tetanus toxin. Eur J Biochem 1999;259:762–769.

8. Sabroe I, Prince LR, Jones EC, Horsburgh MJ, Foster SJ, VogelSN, Dower SK, Whyte MK. Selective roles for Toll-like receptor(TLR)2 and TLR4 in the regulation of neutrophil activation and lifespan. J Immunol 2003;170:5268–5275.

9. Hirschfeld M, Ma Y, Weis JH, Vogel SN, Weis JJ. Cutting edge:repurification of lipopolysaccharide eliminates signaling throughboth human and murine toll-like receptor 2. J Immunol 2000;165:618–622.

0. Iwaki D, Mitsuzawa H, Murakami S, Sano H, Konishi M, Akino T,Kuroki Y. The extracellular toll-like receptor 2 domain directlybinds peptidoglycan derived from Staphylococcus aureus. J BiolChem 2002;277:24315–24320.

1. Inohara N, Ogura Y, Fontalba A, Gutierrez O, Pons F, Crespo J,Fukase K, Inamura S, Kusumoto S, Hashimoto M, Foster SJ,Moran AP, Fernandez-Luna JL, Nunez G. Host recognition ofbacterial muramyl dipeptide mediated through NOD2. Implica-tions for Crohn’s disease. J Biol Chem 2003;278:5509–5512.

2. Girardin SE, Boneca IG, Viala J, Chamaillard M, Labigne A,Thomas G, Philpott DJ, Sansonetti PJ. Nod2 is a general sensorof peptidoglycan through muramyl dipeptide (MDP) detection.J Biol Chem 2003;278:8869–8872.

3. Hisamatsu T, Suzuki M, Reinecker HC, Nadeau WJ, McCormickBA, Podolsky DK. CARD15/NOD2 functions as an antibacterialfactor in human intestinal epithelial cells. Gastroenterology2003;124:993–1000.

4. Morr M, Takeuchi O, Akira S, Simon MM, Muhlradt PF. Differentialrecognition of structural details of bacterial lipopeptides by toll-like receptors. Eur J Immunol 2002;32:3337–3347.

5. Shirai Y, Saito N. Activation mechanisms of protein kinase C:maturation, catalytic activation, and targeting. J Biochem (Tokyo)2002;132:663–668.

6. Li C, Wang Y, Gao L, Zhang J, Shao J, Wang S, Feng W, Wang X,Li M, Chang Z. Expression of toll-like receptors 2 and 4 and CD14during differentiation of HL-60 cells induced by phorbol 12-myris-tate 13-acetate and 1 alpha, 25-dihydroxy-vitamin D(3). CellGrowth Differ 2002;13:27–38.

7. Ways DK, Qin W, Garris TO, Chen J, Hao E, Cooper DR, Usala SJ,Parker PJ, Cook PP. Effects of chronic phorbol ester treatment onprotein kinase C activity, content, and gene expression in thehuman monoblastoid U937 cell. Cell Growth Differ 1994;5:161–169.

8. Song JC, Rangachari PK, Matthews JB. Opposing effects of PK-Calpha and PKCepsilon on basolateral membrane dynamics inintestinal epithelia. Am J Physiol 2002;283:C1548–C1556.

9. Yoo J, Nichols A, Song JC, Mammen J, Calvo I, Worrell RT,Cuppoletti J, Matlin K, Matthews JB. Bryostatin-1 attenuatesTNF-induced epithelial barrier dysfunction: role of novel PKCisozymes. Am J Physiol 2003;284:G703–G712.

0. Ahmed S, Kozma R, Lee J, Monfries C, Harden N, Lim L. Thecysteine-rich domain of human proteins, neuronal chimaerin,protein kinase C and diacylglycerol kinase binds zinc. Evidencefor the involvement of a zinc-dependent structure in phorbol esterbinding. Biochem J 1991;280:233–241.

1. Balda MS, Gonzalez-Mariscal L, Matter K, Cereijido M, AndersonJM. Assembly of the tight junction: the role of diacylglycerol. J CellBiol 1993;123:293–302.

2. Turner JR, Angle JM, Black ED, Joyal JL, Sacks DB, Madara JL.PKC-dependent regulation of transepithelial resistance: roles ofMLC and MLC kinase. Am J Physiol 1999;277:C554–C562.

3. Hsu YW, Chi KH, Huang WC, Lin WW. Ceramide inhibits lipo-polysaccharide-mediated nitric oxide synthase and cyclooxygen-

5

5

5

5

5

5

6

6

6

6

6

6

6

6

6

6

7

7

tge

(MgD

pDoTU


ase-2 induction in macrophages: effects on protein kinases andtranscription factors. J Immunol 2001;166:5388–5397.

4. Kontny E, Kurowska M, Szczepanska K, Maslinski W. Rottlerin, aPKC isozyme-selective inhibitor, affects signaling events and cy-tokine production in human monocytes. J Leukoc Biol 2000;67:249–258.

5. Hecht G, Robinson B, Koutsouris A. Reversible disassembly of anintestinal epithelial monolayer by prolonged exposure to phorbolester. Am J Physiol 1994;266:G214–G221.

6. Mullin JM, Laughlin KV, Ginanni N, Marano CW, Clarke HM,Peralta Soler A. Increased tight junction permeability can resultfrom protein kinase C activation/translocation and act as a tumorpromotional event in epithelial cancers. Ann N Y Acad Sci 2000;915:231–236.

7. Song JC, Hanson CM, Tsai V, Farokhzad OC, Lotz M, MatthewsJB. Regulation of epithelial transport and barrier function bydistinct protein kinase C isoforms. Am J Physiol 2001;281:C649–C661.

8. Gschwendt M, Muller HJ, Kielbassa K, Zang R, Kittstein W,Rincke G, Marks F. Rottlerin, a novel protein kinase inhibitor.Biochem Biophys Res Commun 1994;199:93–98.

9. Mukherjee JJ, Chung T, Ways DK, Kiss Z. Protein kinase Calphais a major mediator of the stimulatory effect of phorbol ester onphospholipase D-mediated hydrolysis of phosphatidylethanol-amine. J Biol Chem 1996;271:28912–28917.

0. Tai YH, Flick J, Levine SA, Madara JL, Sharp GW, Donowitz M.Regulation of tight junction resistance in T84 monolayers byelevation in intracellular Ca2�: a protein kinase C effect. JMembr Biol 1996;149:71–79.

1. Balda MS, Gonzalez-Mariscal L, Contreras RG, Macias-Silva M,Torres-Marquez ME, Garcia-Sainz JA, Cereijido M. Assembly andsealing of tight junctions: possible participation of G-proteins,phospholipase C, protein kinase C and calmodulin. J Membr Biol1991;122:193–202.

2. Kinugasa T, Sakaguchi T, Gu X, Reinecker HC. Claudins regulatethe intestinal barrier in response to immune mediators. Gastro-enterology 2000;118:1001–1011.

3. Hartwig JH, Thelen M, Rosen A, Janmey PA, Nairn AC, Aderem A.MARCKS is an actin filament crosslinking protein regulated byprotein kinase C and calcium-calmodulin. Nature 1992;356:618–622.

4. Fanning AS, Jameson BJ, Jesaitis LA, Anderson JM. The tight C

junction protein ZO-1 establishes a link between the transmem-brane protein occludin and the actin cytoskeleton. J Biol Chem1998;273:29745–29753.

5. Brose N, Rosenmund C. Move over protein kinase C, you’ve gotcompany: alternative cellular effectors of diacylglycerol and phor-bol esters. J Cell Sci 2002;115:4399–4411.

6. Nakashima S. Protein kinase C alpha (PKC alpha): regulation andbiological function. J Biochem (Tokyo) 2002;132:669–675.

7. Kazanietz MG. Novel “nonkinase” phorbol ester receptors: theC1 domain connection. Mol Pharmacol 2002;61:759–767.

8. Fitzgerald KA, Palsson-McDermott EM, Bowie AG, Jefferies CA,Mansell AS, Brady G, Brint E, Dunne A, Gray P, Harte MT, Mc-Murray D, Smith DE, Sims JE, Bird TA, O’Neill LA. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduc-tion. Nature 2001;413:78–83.

9. Horng T, Barton GM, Medzhitov R. TIRAP: an adapter molecule inthe Toll signaling pathway. Nat Immunol 2001;2:835–841.

0. Chen ML, Pothoulakis C, LaMont JT. Protein kinase C signalingregulates ZO-1 translocation and increased paracellular flux ofT84 colonocytes exposed to Clostridium difficile toxin A. J BiolChem 2002;277:4247–4254.

1. Podolsky DK. Inflammatory bowel disease. N Engl J Med 2002;347:417–429.

Received August 21, 2003. Accepted April 8, 2004.Address requests for reprints to: Elke Cario, M.D., Division of Gas-

roenterology and Hepatology, University Hospital of Essen, Instituts-ruppe I, Virchowstrasse 171, D-45147 Essen, Germany. e-mail:[email protected]; fax: (49) 211-495-7035.Supported by grants from the Deutsche Forschungsgemeinschaft

Ca226/4-1, Ca226/5-1, Priority Program Innate Immunity), and theedical Faculty (IFORES) at the University of Essen (to E.C.), and byrants from the National Institutes of Health (DK60049; DK43351) (to.K.P.).The authors thank Tularik Inc. (South San Francisco, CA) for kindlyroviding the dominant-negative constructs of TLR2 and TLR4 throughr. Carsten J. Kirschning (Institute of Medical Microbiology, Immunol-gy, and Hygiene, Technical University of Munich, Munich, Germany).he authors also thank Dr. Jens Nurnberger (Division of Nephrology,niversity Hospital of Essen, Essen, Germany) for the gift of the
Y5-conjugated goat anti-rabbit antibody.

Documents

Toll-Like Receptor 2 Enhances ZO-1â€“Associated Intestinal Epithelial