Central role of PAFR signalling in ExoU-induced NF-κB activation

Central role of PAFR signalling in ExoU-inducedNF-κB activation

Carolina Diettrich Mallet de Lima,1

Jessica da Conceição Costa,1

Sabrina Alves de Oliveira Lima Santos,1

Simone Carvalho,2 Laís de Carvalho,2

Rodolpho Mattos Albano,3 Mauro Martins Teixeira,4

Maria Cristina Maciel Plotkowski1 andAlessandra Mattos Saliba1*1Departamento de Microbiologia, Imunologia eParasitologia, Universidade do Estado do Rio deJaneiro, Rio de Janeiro, RJ, Brazil.2Departamento de Histologia e Embriologia,Universidade do Estado do Rio de Janeiro, Rio deJaneiro, RJ, Brazil.3Departamento de Bioquímica, Universidade do Estadodo Rio de Janeiro, Rio de Janeiro, RJ, Brazil.4Departamento de Bioquímica e Imunologia,Universidade Federal de Minas Gerais, Minas Gerais,MG, Brazil.

Summary

ExoU is an important virulence factor in acutePseudomonas aeruginosa infections. Here, weunveiled the mechanisms of ExoU-driven NF-κBactivation by using human airway cells andmice infected with P. aeruginosa strains. Severalapproaches showed that PAFR was crucially impli-cated in the activation of the canonical NF-κBpathway. Confocal microscopy of lungs frominfected mice revealed that PAFR-dependent NF-κBactivation occurred mainly in respiratory epithelialcells, and reduced p65 nuclear translocation wasdetected in mice PAFR−/− or treated with the PAFRantagonist WEB 2086. Several evidences showedthat ExoU-induced NF-κB activation regulatedPAFR expression. First, ExoU increased p65 occu-pation of PAFR promoter, as assessed by ChIP.Second, luciferase assays in cultures transfectedwith different plasmid constructs revealed thatExoU promoted p65 binding to the three κB sites in

PAFR promoter. Third, treatment of cell cultureswith the NF-κB inhibitor Bay 11–7082, or trans-fection with IκBα negative-dominant, significantlydecreased PAFR mRNA. Finally, reduction in PAFRexpression was observed in mice treated with Bay11–7082 or WEB 2086 prior to infection. Together,our data demonstrate that ExoU activates NF-κB byPAFR signalling, which in turns enhances PAFRexpression, highlighting an important mechanismof amplification of response to this P. aeruginosatoxin.

Introduction

Pseudomonas aeruginosa is a major agent of ventilator-associated pneumonia and sepsis in critically ill patients(Fujitani et al., 2011). P. aeruginosa pathogenicity is mul-tifactorial and is supported by virulence factors encodedby genes within the core genome. In addition, productsencoded by genes within the accessory genome canincrease bacterial virulence (Kung et al., 2010). ExoU is apotent toxin with phospholipase A2 (PLA2) activity (Satoet al., 2003), encoded by a pathogenicity island (He et al.,2004), that is associated with bacterial persistence in thelung of infected animals (Shaver and Hauser, 2004;Howell et al., 2013), blood invasiveness (Wareham andCurtis, 2007) and poor clinical outcomes of patientswith acute pneumonia (Roy-Burman et al., 2001) andbacteremia (El-Solh et al., 2012). Once injected into hostcell cytosol by the type III secretion system, ExoU istargeted to plasma membranes (Rabin et al., 2006) whereit cleaves membrane phospholipids, resulting in celllysis. Hydrolysis of the sn-2 ester bond of membraneglycerophospholipid can also result in the release ofarachidonic acid and lysophospholipids that can be usedfor the synthesis of pro-inflammatory mediators, suchas eicosanoids and platelet-activating factor (PAF)(Sitkiewicz et al., 2007).

PAF is a potent bioactive lipid and can exert severalbiological effects by binding to the PAF receptor (PAFR),a specific G-protein coupled receptor (GPCR). Signallingdownstream PAFR participates of diverse physiologicalphenomena and is also involved in a variety of patho-logical disorders (Stafforini et al., 2003), including infec-tious process such as malaria (Lacerda-Queiroz et al.,2013), Influenza A pneumonia (Garcia et al., 2010) and

Received 19 November, 2013; revised 30 January, 2014; accepted 14February, 2014. *For correspondence. E-mail [email protected]; Tel. (+55) 21 2868 8280; Fax (+55) 21 2868 8376.

Cellular Microbiology (2014) 16(8), 1244–1254 doi:10.1111/cmi.12280First published online 21 March 2014

© 2014 John Wiley & Sons Ltd

cellular microbiology

mailto:[email protected]

mailto:[email protected]

polymicrobial sepsis (Moreno et al., 2006). In addition,PAF-PAFR interaction contributes to inflammatoryresponses as it stimulates the leucocyte recruitment bythe activation of the transcription factors NF-κB, AP1,NFAT and NF-IL6, and the induction of chemokine secre-tion (Kravchenko et al., 1995; Venkatesha et al., 2004;Borthakur et al., 2010; Park et al., 2013).

Studies using murine models of acute pneumonia haveshown that ExoU induces an intense inflammatory infil-trate rich in neutrophils in lungs of infected mice (Salibaet al., 2005; Diaz et al., 2008) and we have shown thatmice treatment with a NF-κB inhibitor prior to infection byExoU-producing P. aeruginosa reduced the secretion ofthe murine IL-8 homologue KC and the neutrophil recruit-ment into the lungs of infected animals, pointing to a majorrole of this transcription factor in ExoU-mediated inflam-matory response (de Lima et al., 2012). Since activatedNF-κB can regulate a number of genes, and represent animportant step in the pathogenesis of infectious disease,the aim of this study was to gain insight into the mecha-nisms of NF-κB activation triggered by ExoU. Our findingsindicate that NF-κB is activated downstream of PAFR thatbinds to PAF produced by the ExoU-injected cells. Moreo-ver, NF-κB enhances PAFR expression, amplifying theeffects of ExoU.

Results and discussion

ExoU-driven NF-κB activation occurs in infected and innon-infected cells of the neighbourhood as well

To evaluate the molecular mechanisms underlying NF-κBactivation by ExoU, we first investigated whether NF-κBactivity depended on the direct effect of ExoU or resultedfrom the action of soluble factors released by the infectedcells. To address this question, human epithelial respira-tory cells from the A549 line, cultured in microplate wellsand transwell inserts, were transfected with the p6κB-LUCreporter plasmid. Cells cultured in microplate wells wereinfected for 1 h with the ExoU-producing PA103 strain,with the exoU-deficient PA103ΔexoU mutant or withPA103ΔUT/S142A, an exoU-deficient bacteria comple-mented with site-directed mutated exoU serine catalyticmotif, which secretes ExoU without PLA2 activity. Aftertreatment of infected cultures with gentamicin, to kill bac-teria, cultures in inserts were positioned above the wellsto be exposed to soluble factors eventually released bythe infected cells. All cultures were incubated for addi-tional 24 h to investigate NF-κB-dependent transcriptionalactivity by luciferase reporter assay. As expected,luciferase activity was significantly higher in culturesinfected with the ExoU-producing PA103 strain than innon-infected cultures or in cultures infected with themutants. More importantly, no significant differences in

luciferase activity were observed when cultures inmicroplate wells were compared with cultures in inserts,indicating that the ExoU-induced NF-κB activation wasmediated by soluble factors released by the ExoU-injected cells (Fig. 1).

PAF is the soluble factor released by the ExoU-infectedcells that activates NF-κB

The effects of ExoU include a potent cytotoxicity against arange of eukaryotic cells (Finck-Barbançon et al., 1997;Saliba et al., 2006; Diaz and Hauser, 2010), activationof the coagulation cascade (Plotkowski et al., 2008;Machado et al., 2010) and stimulus of a robust pro-inflammatory response (Saliba et al., 2005; Diaz et al.,2008). Whereas ExoU PLA2 activity has a direct effect thatleads to disruption of membrane integrity and death ofthe infected cell, the release of arachidonic acid andlysophospholipids from host membranes supports thesynthesis of eicosanoids and PAF, which may produce anumber of effects in the neighbour cells. In fact, previousworks carried out in our laboratory showed that PAF con-tributes to host response to ExoU by activating the coagu-lation cascade and inhibiting the fibrinolytic pathway(Machado et al., 2011).

Since PAF-PAFR interaction can activate NF-κB, wenext investigated the role of this signalling in ExoU-mediated NF-κB activation. First, we evaluated theNF-κB-dependent transcriptional activity in A549 cells

Fig. 1. ExoU activates NF-κB by stimulating the release ofeukaryotic soluble factors. A549 cells cultured in microplate wellsand transwell inserts were transfected with p6kB-LUC andpRL-CMV plasmids. Cells in microplate wells were infected withP. aeruginosa strains while cells in inserts were exposed toproducts released by the infected cells. Control cells were treatedwith culture medium or exposed to products from uninfected cells.After 24 h, all cultures were evaluated by luciferase reporter assay.Graph represents the means of luciferase activity (in arbitraryvalues) ± SEM obtained in three different assays. ***P < 0.001when the values obtained from cultures infected with PA103 orexposed to ExoU indirect effects were compared with the othercultures.

Central role of PAFR in ExoU-induced NF-κB activation 1245

© 2014 John Wiley & Sons Ltd, Cellular Microbiology, 16, 1244–1254

treated with WEB 2086, a PAF antagonist, or anti-PAFRantibody prior to infection. Luciferase reporter assaysshowed that both treatments caused a steep decrease inthe luciferase activity detected in cultures infected with theExoU-producing PA103 strain (Fig. 2A and B). Second,we investigated by Western blot whether anti-PAFR treat-ment could interfere in IκBα degradation and p65 nucleartranslocation induced by ExoU. As observed in Fig. 2Cand D, infection with ExoU-producing strains reducedthe levels of IκBα and increased p65 accumulation innucleus, and both phenomena were inhibited when cellswere infected with the ExoU mutant strains or treated withanti-PAFR before PA103 infection. Third, by confocalmicroscopy we observed increased p65 nuclear translo-cation in PA103-infected cells than in cells infected withthe mutant strain or treated with anti-PAFR or WEB 2086before PA103 infection (Fig. 2E).

Some studies have shown that PAF is a proximalinducer of the transcription factor NF-κB in responseto inflammatory stimuli (Kravchenko et al., 1995;Venkatesha et al., 2004; Borthakur et al., 2010; Parket al., 2013), but the signalling events that initiate theseresponses are not fully understood. Since NF-κB hasthe ability to cross-talk with other pathways includingPI3-kinase/AKT (Hussain et al., 2012; Dan et al., 2014),we investigated the effect of AKT inhibition in IκBαphosphorylation. The results obtained by ELISA showed asignificant increase of phospho-IκBα in PA103-infectedcultures that was completely inhibited by treatment withanti-PAFR, WEB 2086 and Bay 11–7082, and partiallyinhibited by an AKT inhibitor (Fig. 2F). The treatment withMAPK inhibitors did not affect this pathway (data notshown). These results suggest that Akt contributes toNF-κB activation by ExoU, although other factors seemto be involved. A candidate is Bcl10, which contributes toNF-κB activation by PAF in intestinal epithelial cells(Borthakur et al., 2010).

Since we had previously reported that NF-κB activationby ExoU induces IL-8 expression, we next verified theeffect of PAFR blocking on IL-8 mRNA levels. Asassessed by real-time RT-PCR, anti-PAFR treatmentcompletely inhibited the ExoU-induced expression of IL-8gene (Fig. 2G). Together, these results indicate that ExoUactivates the canonical NF-κB pathway by inducing PAFRsignalling.

NF-κB also regulates PAFR expressionin response to ExoU

The promoter region of one of the two functional PAFRmRNA species has three consensus κB sites (Mutohet al., 1993) and once PAFR binds to PAF, it can activateNF-κB and regulate its own expression (Mutoh et al.,1994).

To evaluate the role of the ExoU-mediated NF-κB acti-vation in PAFR expression, we first performed chromatinimmunoprecipitation (ChIP) assays. As shown in Fig. 3A,infection with the ExoU-producing PA103 strain led to asignificant increase of p65 binding to PAFR promoter. Incontrast, infection with the PA103ΔexoU mutant led to alow NF-κB binding that did not differ from the observed incontrol cells. Then, we aimed to identify the relative role ofeach of the three κB sites that exist in the PAFR promoter,in the ExoU-dependent PAFR expression. Comparison ofluciferase activity in cultures transfected with a panel ofplasmids expressing luciferase gene under control of wildtype or mutated PAFR promoter revealed that ExoUstimulates NF-κB binding to the three κB sites in thePAFR promoter. Moreover, all sites are required to fullPAFR gene expression (Fig. 3B). Finally, treatment ofcultures with the NF-κB inhibitor Bay 11–7082 (Fig. 3C)or transfection with IκBα negative-dominant plasmid(Fig. 3D) led to a significant reduction in PAFR mRNAlevels, as assessed by real time qRT-PCR. Together,these results demonstrate that NF-κB also regulatesPAFR expression during infection by ExoU-producingP. aeruginosa strain.

Exogenous PAF at 1 μM activates NF-κB and increasesPAFR and IL-8 expression at the same levels inducedby ExoU

To determine the concentration of PAF needed to inducethe same levels of NF-κB transcriptional activity observedin PA103-infected cultures, we evaluated the effect ofdifferent concentrations of PAF by luciferase reporterassays. These assays showed that exogenous PAF at1 μM led to the same levels of NF-κB transcriptional activ-ity induced by ExoU (Fig. 4A), and this effect was inhibitedby anti-PAFR antibody or WEB 2086 treatment (Fig. 4B).Furthermore, qRT-PCR assays showed that ExoU-producing P. aeruginosa strain and exogenous PAF at1 μM induced the same significant increase of IL-8(Fig. 4C) and PAFR (Fig. 4D) expression, and both weredecreased with prior treatment with WEB 2086.

Central role of PAFR in NF-κB activation and increasedPAFR expression were also detected in the airways ofmice infected with the ExoU-producing strain

The importance of PAFR signalling in ExoU-mediatedNF-κB activation was confirmed in a mouse model ofacute pneumonia. Both PAFR(−/−) mice and mice treatedwith the PAF antagonist WEB 2086 prior to PA103 infec-tion exhibited a drastic reduction in p65 nuclear translo-cation in lung cells, as assessed by ELISA carried out withlung extracts (Fig. 5A and B). Additionally, analysis byconfocal microscopy showed that p65 accumulation in

1246 C. D. Mallet de Lima et al.


Fig. 2. PAF has a central role in NF-κB activation by ExoU. A549 cells were treated with 10 μg ml−1 WEB 2086 or 2 μg ml−1 anti-PAFRantibody prior to infection with P. aeruginosa strains or treatment with culture medium (control).A and B. Cells had already been transfected with p6kB-LUC and pRL-CMV plasmids to assess luciferase activity 24 h post infection.C and D. IκB-α (C) and p65 (D) expression in cells infected for 3 h was assessed by Western blotting.E. Immunofluorescence localization of p65 in cultures infected for 3 h were analysed using confocal microscopy. The photographs arerepresentative of 3 assays.F. A549 cells were pre-treated for 1 h with different inhibitors, infected for 3 h, lysed and evaluated by ELISA to detect IκBα phosphorylation.G. IL-8 mRNA levels in cultures infected for 18 h were quantified using real time RT-PCR. Graphs represent the means ± SEM of valuesobtained in three different assays. ***P < 0.001 when the values obtained from PA103-infected cultures were compared with those obtainedfrom the other cultures.



nucleus was mainly observed in airway epithelial cells ofmice infected with PA103 strain (Fig. 5C). Accordingly,qRT-PCR assays showed that NF-κB inhibition by treat-ment with Bay 11–7082 (Fig. 6A), as well as inhibition ofPAF effect by treatment with WEB 2086 (Fig. 6B), abol-ished PAFR expression induced by ExoU in mice lungs,confirming that NF-κB and PAFR are key factors in theinduction of PAFR expression during the response toExoU.

During P. aeruginosa infection, membrane hydrolysisby ExoU PLA2 activity not only accounts to cell death,which can impair the immune system and destroy tissuebarriers, but also allows the production of a number ofmediators that can act in all cells of the neighbourhoodand play an important role in pathogenesis. Here weshowed that PAF produced from lysophospholipidsreleased from cell membranes hydrolysed by ExoU binds

to PAFR, activating NF-κB in non-infected adjacent cells.In turn, PAF-PAFR interaction stimulates the expressionof many genes such as IL-8 and PAFR itself, maintaininga regulatory loop of gene activation (Fig. 7).

Since NF-κB is a transcriptional factor that regulates anumber of genes involved in inflammatory and immuneresponses, and ExoU plays an important role inP. aeruginosa pathogenesis, this pathway may representan important mechanism triggered during P. aeruginosaacute infections and a potential target to be explored.

Experimental procedures

Ethics statement

All animal experiments were conducted under prior approval bythe Comissão de Ética no Uso de Animais (CEUA) of theUniversidade do Estado do Rio de Janeiro (protocol # CEUA/

Fig. 3. NF-κB activates PAFR expressionduring response to ExoU.A. A549 cells were infected or treated withculture medium for 14 h to evaluate p65occupation in PAFR promoter by ChIP assay.B. Cells were transfected with plasmidscontaining luciferase gene under control ofwild type (WT PAFR) or mutated PAFRpromoter (in κB site 1, κB site 2, κB site 3 orall 3 κB sites), and then infected for 24 h toassess luciferase activity.C and D. Determination, by real timeRT-PCR, of PAFR expression by A549 cellsinfected for 18 h, previously treated or notwith 10 mM Bay 11–7082 for 1 h (C) ortransfected with a dominant-negative pIκB-αS32/36A or empty vector (D). Graph showsthe means ± SEM of values obtained in threeindependent assays. ***P < 0.001 when thevalues obtained from cultures infected withthe ExoU-producing PA103 strain werecompared with those obtained from the othercultures.



022/2011), according to Brazilian guidelines on animal work ofthe Conselho Nacional de Controle de Experimentação Animal(CONCEA) (Law No. 11.794). Surgeries were performed underanesthesia (ketamine and xylazine), and all efforts were made tominimize suffering.

Bacterial strains

All experiments were performed with the ExoU-producingP. aeruginosa PA103 strain and the exoU deficient mutantPA103ΔexoU (Saliba et al., 2005). In some assays, exoU defi-cient mutants complemented with the wild type gene (PA103ΔUT/exoU) or with gene coding for site-mutated ExoU serine catalyticmotif (PA103ΔUT/S142A) were used to confirm ExoU effect andto evaluate ExoU PLA2 activity respectively (Rabin and Hauser,2005). PA103ΔUT/exoU and PA103ΔUT/S142A were a kinddonation of Dr. Alan Hauser (Northwestern University, Chicago).

Cell infection

Human airway epithelial cells from A549 line (purchased from theRio de Janeiro Cell Bank, BCRJ 0033) were cultured in F12nutrient culture medium (Life Technologies) containing 10% fetalcalf serum (FCS) and antibiotics. Cells were infected withP. aeruginosa strains at moi of 100, centrifuged (1000 g for10 min) to favour the close contact between bacteria and hostcells and incubated for 1 h at 37°C. Control cultures wereexposed to culture medium only. To reduce bacterial challengeand assess host response in later periods, cell cultures were

treated with gentamicin at 300 μg ml−1, to kill extracellular bacte-ria, and incubated for additional time. To evaluate the activation ofNF-κB by PAFR, cells were treated for 1 h prior to infection witha polyclonal anti-PAFR antibody (Cayman Chemical) at 2 μg ml−1

or WEB 2086 (Biomol) at 10 μg ml−1. To assess the role of NF-κBpathway in PAFR induction, cells were treated with the inhibitorof Ikk complex Bay 11–7082 (Enzo Life Sciences) at 10 μM(Park et al., 2009). To investigate other pathways that couldcontribute to NF-κB activation by ExoU, cells were treatedwith an Akt inhibitor (Calbiochem) at 10 μM, the MEK inhibitorUO126 (Calbiochem) at 20 μM, the p38 inhibitor SB203580(Calbiochem) at 10 μM or the JNK inhibitor SP600125(Calbiochem) at 10 μM, 1 h before infection. The effect ofexogenous PAF was assessed by treating cells with PAF(Calbiochem) at different concentrations.

Luciferase reporter assay

To evaluate NF-κB-dependent transcriptional activity, A549 cellson microplate or 0.4 μM pore polycarbonate membrane transwellinserts (Nalge Nunc International) were co-transfected with770 ng p6κB-LUC (kindly provided by Dr. Patrick Baeuerle, Uni-versity of Munich, Germany) and 30 ng pRL-CMV (Promega)plasmids using Lipofectamine 2000 reagent (Invitrogen), for 24 h.Cells in microplates were directly infected while cells in insertswere exposed to supernatants from infected cells. After 24 h,cultures were washed with PBS and lysed to measure firefly andRenilla luciferases using the Dual Luciferase Reporter System(Promega) and a Glomax® multi-detection system: Luminometer(Promega).

Fig. 4. Exogenous PAF at 1 μM activatesNF-κB and increases PAFR and IL-8expression at the same levels induced byExoU.A. A549 cells transfected with p6kB-LUC andpRL-CMV plasmids were infected withP. aeruginosa strains or exposed toexogenous PAF at 10 μM, 1 μM, 100 nM,1 nM to determine luciferase activity (A).B. Luciferase was measured in cells treatedwith 10 μg ml−1 WEB 2086 or 2 μg ml−1

anti-PAFR antibody prior to infection oraddition of exogenous PAF 1 μM.C and D. IL-8 (C) and PAFR (D) expressionwas assessed by real time RT-PCR of A549cells treated or not with 10 μg ml−1 WEB 2086and then infected with P. aeruginosa strains orexposed to PAF 1 μM for 18 h. Graphrepresents the means ± SEM obtained inthree different assays. ***P < 0.001 when thevalues obtained from cultures infected withPA103 or exposed to PAF were comparedwith the other cultures.



Western blotting analysis

Western blotting assays were performed to evaluate IκB-α deg-radation and p65 translocation to nucleus. For total proteinextraction, A549 cells were washed twice with ice-cold PBS andthen lysed in 60 μl of lysis buffer (50 mM Tris-HCl, pH 7.5, 5 mMEDTA, 10 mM EGTA, 50 mM NaF, 20 mM 3-glycerophosphate,250 mM NaCl, 0.1% Triton X-100, 1 μg ml−1 BSA and proteaseinhibitor cocktail II [Calbiochem] at 1:1000). For nuclear protein

extraction, A549 cells were washed and scraped in PBS andcentrifuged at 14 000 g, 1 min, at 4°C. Cell pellets wereresuspended in 900 μl of lysis buffer A (20 mM Hepes pH 7.9,1 mM EDTA, 10 mM NaCl, 0.25% Nonidet P-40 [NP-40], 1 mMDTT and protease inhibitor cocktail II at 1:1000) and incubated onice for 10 min. Cell membranes were then pelleted by centrifu-gation 14 000 g, 1 min, at 4°C, and the cytoplasmic extracts wereremoved. The pelleted nuclei were immediately resuspended in50 μl of buffer C (20 mM Hepes pH 7.9, 420 mM NaCl, 1 mM

Fig. 5. ExoU-induced NF-κB activation inmice airway epithelial cells also depends onPAF-PAFR interaction. The role of PAF andPAFR in ExoU-induced NF-κB in vivo wasassessed in C57BL/6 mice pretreated or notwith 7.5 mg kg−1 WEB 2086 (A, B and C) andin C57BL/6 PAFR(−/−) mice (B). Mice wereintratracheally inoculated with P. aeruginosastrains or saline and lungs were recovered forquantification of nuclear p65 by ELISA (A andB) or evaluation of p65 expression byconfocal microscopy (C).A and B. Graph show the means ± SEM ofvalues obtained by ELISA assays of nuclearextracts of lungs from mice instilled withPA103, PA103ΔexoU or saline. **P < 0.01 or***P < 0.001 when the values obtained fromPA103-infected mice were compared withthose obtained from the other groups.C. All micrographs show p65 expression (red)in bronchiolar epithelial cells and nucleistained with DAPI. Note that untreated miceinfected with PA103 showed higher p65expression in the cytoplasm, as well asnumerous points of p65 staining ineuchromatin regions throughout the nuclei(inset). Bar = 20 μm.



EDTA, 25% glycerol, 2 mM DTT, protease inhibitor cocktail II1:1000) and the nuclei were extracted on ice for 20 min withoccasional vortexing. The nuclei were then pelleted at 14 000 g,5 min, 4°C, to collect the supernatants as nuclear extracts.Protein concentration was quantified using BCATM Protein AssayReagent (Thermo Scientific), 35 μg of proteins were separatedby 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE)and transferred to nitrocellulose membranes. The membraneswere blocked in TBST containing 5% powdered milk for 1 h andincubated overnight with IκBα or p65 antibody (Santa Cruz Bio-technology) at 4°C. Blots were washed, probed with secondaryantibody conjugated to HRP, at 25°C for 1 h, developed with ECLsubstrate (Thermo Scientific) and exposed to X-ray. After strip-ping, membranes were re-probed with antibody against αTubulin(Santa Cruz Biotechnology) or lamin A/C (Cell Signaling), for totaland nuclear proteins respectively.

In vitro immunofluorescence localization of p65

Pre-treated and infected cells grown on glass coverslips werewashed twice with PBS buffer and fixed for 10 min at 37°C inPBS-4% (w/v) paraformaldehyde. Monolayers were washedthree times with PBS-1% BSA, blocked with PBS-5% BSA for30 min and then incubated with the primary p65 antibody (SantaCruz Biotechnology) diluted 1:100 in PBS-0.1% saponin at 4°C,overnight. The cells were washed three times with PBS-1% BSAand incubated for 1 h at room temperature with secondary anti-body Alexa Fluor® 594-conjugated anti-rabbit IgG (Invitrogen)diluted 1:500 in PBS. After exhaustive washing, the cells were

then stained with DAPI nuclear dye (Sigma-Aldrich) and thecoverslips were mounted in Prolong Gold antifade media (LifeTechnologies) for immunofluorescence examination. Confocalanalysis was performed using the LSM META 510 (Carl Zeiss,Germany) laser scanning confocal microscope. Images obtainedby LSM META 510 were analysed using ZEN software (version2012 lite) (Carl Zeiss, Germany).

IκBα phosphorylation ELISA assay

A549 whole proteins were extracted using the FunctionELISAIκBα Kit (Active Motif), 100 μg were used to quantify IκBαphosphorylation and the chemiluminescence was measuredwith a luminometer (Promega), according to manufacturer’sinstruction.

Real-time qRT-PCR

A549 cells infected for 18h were trypsinized, washed with PBSand total RNA was isolated using the Qiagen Rneasy kit. cDNAwas synthesized from 3 μg of total RNA by reverse transcriptionwith the SuperScript™ III First-Strand synthesis system forRT-PCR (Invitrogen) and diluted (1:200). Real-time qRT-PCRwas performed in the Applied Biosystems StepOne detectionsystem (Applied Biosystems) using GoTaq qPCR Master Mix(Promega). Experimental qRT-PCR data were normalized byGAPDH as an endogenous control. Primers used in humanPAFR reactions were 5′-GTG GGC TGG ACT TGG CTG AT-3′(sense) and 5′-CCT GGT CCC TCA GCA GGA AA-3′ (antisense)and in IL-8 reactions were 5′-AAG AAA CCA CCG GAA GGAAC-3′ (sense) and 5′-AGC ACT CCT TGG CAA AAC TG-3′ (anti-sense), whereas primers used in GAPDH reactions were 5′-TGCACC ACC AAC TGC TTA GC-3′ (sense) and 5′-GGC ATG GACTGT GGT CAT GAG-3′ (antisense). All expression ratios weredetermined by the relative gene expression ΔΔCt method viaStepOne 2.0 software 2.0.6 (Applied Biosystems).

In addition to the Bay 11–7082 treatment, the role of NF-κBon PAFR expression was also evaluated in A549 cells co-transfected with 3.85 μg pCDNA3 (Invitrogen) or pIκB-α S32/36A(kindly provided by Dr Ulisses Gazos Lopes, Federal Universityof Rio de Janeiro, Brazil) using Lipofectamine 2000 reagent(Invitrogen). The pIκB-α S32/36A encodes IκB-α with a substitu-tion of serine 32 and 36 alanine residues that cannot bephosphorylated by Iκκ complex and overexpression of this domi-nant negative (ND IκB-α) prevents NF-κB translocation to thenucleus. Transfected cells were infected for 24 h and real-timeqRT-PCR was performed, as described above.

Chromatin immunoprecipitation (ChIP)

ChIP assays were performed in A549 cells infected for 14 h usingthe SimpleChIP Enzimatic Chromatin IP kit (Cell Signaling),according to manufacturer’s instructions. Briefly, protein–DNAcomplexes were cross-linked with formaldehyde, cells were lysedand chromatins were treated with Micrococcal Nucleaseand sonicated. An aliquot of the total input was stored and20 μg DNA of each cross-linked chromatin preparation wereimmunoprecipitated with 10 μg of anti-p65 antibody (Santa CruzBiotechnology). A pool of all cross-linked chromatin preparationswas immunoprecipitated with rabbit IgG isotype control to ensure

Fig. 6. NF-κB and PAF also regulate the ExoU-induced PAFRexpression in mice lungs. C57BL/6 mice were treated or not with20 mg kg−1 Bay 11–7082 (A) or and 7.5 mg kg−1 WEB 2086 (B) for1 h, and intratracheally inoculated with PA103, PA103ΔexoU orsaline for quantification of PAFR expression in lungs by real timeRT-PCR. Graphs show the means ± SEM of values obtained byqRT-PCR assays carried out in triplicate. ***P < 0.001 when thevalues obtained from untreated mice infected with theExoU-producing PA103 strain were compared with those obtainedfrom the other groups.



Fig. 7. PAF-PAFR signalling in ExoU-induced NF-κB activation. ExoU is injected by the type III secretion system into the cytosol of airwayepithelial cells infected by P. aeruginosa. ExoU localizes to host membranes and exerts its phospholipase A2 activity, which leads to theincreased production of PAF. PAF binds to PAFR in non-infected cells of the neighbourhood and activates NF-κB by the canonical pathwaywith participation of Akt. Once in nucleus, p65/p50 binds to the three κB sites in PAFR promoter 1 and induces PAFR expression, which canbind to more molecules of PAF, amplifying the effects of ExoU in non-infected cells.



specificity. After washing, chromatin-DNA cross-linking ofimmunoprecipitated and total input samples was reversed, DNAwas purified and then amplified by Real Time qPCR using thefollowing primers against the NF-κB site in PAFR promoter5′-CAG GGA GAC CCT ATC TCT ACA A-3′ (sense) and 5′-TGGTCA GGG GCA TGA CAG-3′ (antisense).

Plasmid constructs

To construct wild-type and mutated PAFR promoters, DNA frag-ments of A549 cells were amplified by PCR using Pfu DNApolymerase (Promega). Primers used to construct PAFR wild-type promoter were 5′-GTG CTA GCC ACA AGG AAG CAA TCAACC AAG TCC AGA A-3′ (sense) and 5′-GTA AGC TTT TCCCCC TTA CCT GTA GCA G-3′ (antisense). Primers used tointroduce site-directed mutagenesis in the 3 κB sites were MUTκB1: 5′-AAA CAC TTA GAC CCT ATC TCT ACA AAA CAC GG-3′(sense) and 5′-AGA TAG GGT CTA AGT GTT TTG CCT AGGTTG GTC TC-3′ (antisense), MUT κB2: 5′-TCT CAC TTA GCTTCC ACC TCC ACC CCC ACA GA-3′ (sense) and 5′-AGG TGGAAG CTA AGT GAG ACC GTG TTT TGT AGA GA-3′ (antisense),and MUT κB3: 5′-CTC CAC TTA GGC CCT GCC CTG TCATGC-3′ (sense) and 5′-GGG CAG GGC CTA AGT GGA GACCAA GAA A-3′ (antisense). PCR products were purified usingWizard, A-tailed and subcloned into pGEM®-T (Promega).Recombinant plasmids were digested with NheI and HindIII andinserts were cloned into pGL2-basic (Promega). All constructswere validated by complete DNA sequencing on a MegaBACE1000 automated DNA sequencer (GE Healthcare) using theprimers pGL2.1 5′-TGT ATC TTA TGG TAC TGT AAC TG-3′ andpGL2. 2 5′-CCT TAT GTT TTT GGC GTC TTC CA-3′.

Mice infection model

Female isogenic C57BL/6 wild-type and PAFR (−/−) mice (8–12weeks old) anesthetized with a mixture of ketamine (65 mg kg−1)and xylazine (13 mg kg−1) were infected intratracheally with 104

colony-forming units of PA103 or PA103ΔexoU in 50 μl oflipopolysaccharide-free saline. As control, mice were instilled withsaline only. Some mice were intraperitoneally inoculated with theNF-κB inhibitor Bay 11–7082 at 20 mg kg−1, PAF antagonist WEB2086 (Biomol) at 7.5 mg kg−1 or DMSO (vehicle), 1 h before bac-terial infection. At 24 h post infection, animals were euthanized byintraperitoneal injection of sodium pentobarbital and their lungswere recovered.

Murine real-time qRT-PCR

RNA in lungs was immediately stabilized with RNAlater (Qiagen).Tissues were disrupted, total RNA was isolated, cDNA was syn-thesized and PAFR qRT-PCR was performed with the primers5′-ACG GCT GAG CTC CTC CTA C-3′ (sense) and 5′-CGA AACTCA GAA TCC ACA CG-3′ (antisense). GAPDH was used asendogenous control.

p65 nuclear extract and ELISA assay

Lung proteins were extracted using the Nuclear Extract kit (ActiveMotif) and 20 μg were used to evaluate p65 concentration in

nucleus by TransAM® NFκB p65 Transcription Factor ELISA,according to manufacturer’s instruction (Active Motif).

Murine immunofluorescence localization of p65

Lung samples from experimental groups were immersed inTissue-Tek® O.C.T. Compound (Sakura® Finetek) and frozen in−70°C. Sections 5 μm thick were fixed in −20°C acetone andprocessed for immunofluorescence as follows: after blocking in5% BSA (bovine serum albumin, Sigma-Aldrich) PBS solution,sections were incubated with anti-mouse p65 primary monoclo-nal antibody (Santa Cruz Biotechnology) overnight, followed byincubation with Alexa 555-conjugated anti-rabbit secondary anti-body (Life Technologies). Slides were then stained with DAPInuclear dye (Sigma-Aldrich) and mounted in ProLong Goldanti-fade media (Life Technologies) for further analyses in theLSM META 510 (Carl Zeiss, Germany) laser scanning confocalmicroscope.

Statistical analysis

Statistical significance was accepted at the P < 0.05 level byone-way ANOVA for multiple group analysis with a Bonferroniadjustment.

Acknowledgements

We thank Maria Angelica Pereira da Silva for her technical assis-tance and the web designer Nathalia Lepsch for the illustration inFigure 7. This work was supported by grants from FAPERJ,CNPq and CAPES.

References

Borthakur, A., Bhattacharyya, S., Alrefai, W.A., Tobacman,J.K., Ramaswamy, K., and Dudeja, P.K. (2010) Platelet-activating factor-induced NF-κB activation and IL-8 produc-tion in intestinal epithelial cells are Bcl10-dependent.Inflamm Bowel Dis 16: 593–603.

Dan, H., Cooper, M.J., Cogswell, P.C., Duncan, J.A., Ting,J.P., and Baldwin, A.S. (2014) Akt-dependent regulation ofNF-kB is controlled by mTOR and Raptor in associationwith IKK. Genes Dev 22: 1490–1500.

Diaz, M.H., and Hauser, A.R. (2010) Pseudomonasaeruginosa cytotoxin ExoU is injected into phagocytic cellsduring acute pneumonia. Infect Immun 78: 1447–1456.

Diaz, M.H., Shaver, C.M., King, J.D., Musunuri, S., Kazzaz,J.A., and Hauser, A.R. (2008) Pseudomonas aeruginosainduces localized immunosuppression during pneumonia.Infect Immun 76: 4414–4421.

El-Solh, A.A., Hattemer, A., Hauser, A.R., Alhajhusain, A.,and Vora, H. (2012) Clinical outcomes of type IIIPseudomonas aeruginosa bacteremia. Crit Care Med 40:1157–1163.

Finck-Barbançon, V., Goranson, J., Zhu, L., Sawa, T.,Wiener-Kronish, J.P., Fleiszig, S.M., et al. (1997) ExoUexpression by Pseudomonas aeruginosa correlates withacute cytotoxicity and epithelial injury. Mol Microbiol 25:547–557.



Fujitani, S., Sun, H.Y., Yu, V.L., and Weingarten, J.A. (2011)Pneumonia due to Pseudomonas aeruginosa: part I: epi-demiology, clinical diagnosis, and source. Chest 139: 909–919.

Garcia, C.C., Russo, R.C., Guabiraba, R., Fagundes, C.T.,Polidoro, R.B., Tavares, L.P., et al. (2010) Platelet-activating factor receptor plays a role in lung injury anddeath caused by Influenza A in mice. PLoS Pathog 6:e1001171.

He, J., Baldini, R.L., Déziel, E., Saucier, M., Zhang, Q.,Liberati, N.T., et al. (2004) The broad host range pathogenPseudomonas aeruginosa strain PA14 carries two patho-genicity islands harboring plant and animal virulencegenes. Proc Natl Acad Sci USA 101: 2530–2535.

Howell, H.A., Logan, L.K., and Hauser, A.R. (2013) Type IIIsecretion of ExoU is critical during early Pseudomonasaeruginosa pneumonia. mBio 4: e00032–13.

Hussain, A.R., Ahmed, S.O., Ahmed, M., Khan, O.S.,AbdulMohsen, S.A., Platanias, L.C., et al. (2012) Cross-Talk between NFkB and the PI3-Kinase/AKT Pathway CanBe Targeted in Primary Effusion Lymphoma (PEL) CellLines for Efficient Apoptosis. PLoS ONE 7: e39945.

Kravchenko, V.V., Pan, Z., Han, J., Herbert, J.M., Ulevitch,R.J., and Ye, R.D. (1995) Platelet-activating factor inducesNF-kappa B activation through a G protein-coupledpathway. J Biol Chem 270: 14928–14934.

Kung, V.L., Ozer, E.A., and Hauser, A.R. (2010) The acces-sory genome of Pseudomonas aeruginosa. Microbiol MolBiol Rev 74: 621–641.

Lacerda-Queiroz, N., Rachid, M.A., Teixeira, M.M., andTeixeira, A.L. (2013) The role of platelet-activating factorreceptor (PAFR) in lung pathology during experimentalmalaria. Int J Parasitol 43: 11–15.

de Lima, C.D.M., Calegari-Silva, T.C., Pereira, R.M., Santos,S.A., Lopes, U.G., Plotkowski, M.C., and Saliba, A.M.(2012) ExoU activates NF-κB and increases IL-8/KCsecretion during Pseudomonas aeruginosa infection. PLoSONE 7: e41772.

Machado, G.B., de Assis, M.C., Leão, R., Saliba, A.M., Silva,M.C., Suassuna, J.H., et al. (2010) ExoU-induced vascularhyperpermeability and platelet activation in the course ofexperimental Pseudomonas aeruginosa pneumosepsis.Shock 33: 315–321.

Machado, G.B., Oliveira, A.V., Saliba, A.M., Lima,C.D.M., Suassuna, J.H.R., and Plotkowski, M.C. (2011)Pseudomonas aeruginosa toxin ExoU induces aPAF-dependent impairment of alveolar fibrin turnoversecondary to enhanced activation of coagulation andincreased expression of plasminogen activator inhibitor-1in the course of mice pneumosepsis. Respir Res 12: 104.

Moreno, S.E., Alves-Filho, J.C., Rios-Santos, F., Silva, J.S.,Ferreira, S.H., Cunha, F.Q., and Teixeira, M.M. (2006)Signaling via platelet-activating factor receptors accountsfor the impairment of neutrophil migration in polymicrobialsepsis. J Immunol 177: 1264–1271.

Mutoh, H., Bito, H., Minami, M., Nakamura, M., Honda, Z.,Izumi, T., et al. (1993) Two different promoters directexpression of two distinct forms of mRNAs of humanplatelet-activating factor receptor. FEBS Lett 22: 129–134.

Mutoh, H., Ishii, S., Izumi, T., Kato, S., and Shimizu, T. (1994)Platelet-activating factor (PAF) positively auto-regulates

the expression of human PAF receptor transcript 1(leukocyte-type) through NF-kappa B. Biochem BiophysRes Commun 205: 1137–1142.

Park, A.-S., Na, H.-K., Kim, E.-H., Cha, Y.N., and Surh,Y.J. (2009) 4-Hydroxyestradiol Induces Anchorage-Independent Growth of Human Mammary Epithelial Cellsvia Activation of I kB Kinase: potential Role of ReactiveOxygen Species. Cancer Res 69: 2416–2424.

Park, O.J., Han, J.Y., Baik, J.E., Jeon, J.H., Kang, S.S., Yun,C.H., et al. (2013) Lipoteichoic acid of Enterococcusfaecalis induces the expression of chemokines via TLR2and PAFR signaling pathways. J Leukoc Biol 94: 1275–1284.

Plotkowski, M.C., Plotkowski, M.C., Feliciano, L.F., Machado,G.B., Cunha, L.G., Jr, Freitas, C., et al. (2008) ExoU-induced procoagulant activity in Pseudomonas aeruginosa-infected airway cells. Eur Respir J 32: 1591–1598.

Rabin, S.D., Veesenmeyer, J.L., Bieging, K.T., and Hauser,A.R. (2006) A C-terminal domain targets the Pseudomonasaeruginosa cytotoxin ExoU to the plasma membrane ofhost cells. Infect Immun 74: 2552–2561.

Rabin, S.D.P., and Hauser, A.R. (2005) Functional regions ofthe Pseudomonas aeruginosa cytotoxin ExoU. InfectImmun 73: 573–582.

Roy-Burman, A., Savel, R.H., Racine, S., Swanson, B.L.,Revadigar, N.S., Fujimoto, J., et al. (2001) Type III proteinsecretion is associated with death in lower respiratory andsystemic Pseudomonas aeruginosa infections. J Infect Dis183: 1767–1774.

Saliba, A.M., Nascimento, D.O., Silva, M.C.A., Assis, M.C.,Gayer, C.R.M., Raymond, B., et al. (2005) Eicosanoid-mediated proinflammatory activity of Pseudomonasaeruginosa ExoU. Cell Microbiol 7: 1811–1822.

Saliba, A.M., Assis, M.C., Nishi, R., Raymond, B.,Marques Ede, A., Lopes, U.G., et al. (2006) Implications ofoxidative stress in the cytotoxicity of Pseudomonasaeruginosa ExoU. Microbes Infect 8: 450–459.

Sato, H., Frank, D.W., Hillard, C.J., Feix, J.B., Pankhaniya,R.R., Pankhaniya, R.R., et al. (2003) The mechanism ofaction of the Pseudomonas aeruginosa encoded type IIIcytotoxin, ExoU. EMBO J 22: 2959–2969.

Shaver, C.M., and Hauser, A.R. (2004) Relative contributionsof Pseudomonas aeruginosa ExoU, ExoS, and ExoT tovirulence in the lung. Infect Immun 72: 6969–6977.

Sitkiewicz, I., Stockbauer, K.E., and Musser, J.M. (2007)Secreted bacterial phospholipase A2 enzymes: better livingthrough phospholipolysis. Trends Microbiol 15: 63–69.

Stafforini, D.M., McIntyre, T.M., Zimmerman, G.A., andPrescott, S.M. (2003) Platelet-activating factor, apleiotrophic mediator of physiological and pathological pro-cesses. Crit Rev Clin Lab Sci 40: 643–672.

Venkatesha, R.T., Ahamed, J., Nuesch, C., Zaidi, A.K., andAli, H. (2004) Platelet-activating factor-induced chemokinegene expression requires NF-kappaB activation and Ca2+/.calcineurin signaling pathways. Inhibition by receptorphosphorylation and beta-arrestin recruitment. J Biol Chem279: 44606–44612.

Wareham, D.W., and Curtis, M.A. (2007) A genotypic andphenotypic comparison of type III secretion profiles ofPseudomonas aeruginosa cystic fibrosis and bacteremiaisolates. Int J Med Microbiol 297: 227–234.



Central role of PAFR signalling in ExoU-induced NF-κB activation

Documents

Transcript of Central role of PAFR signalling in ExoU-induced NF-κB activation