Non-T cell activation linker regulates ERK activation in Helicobacter pylori-infected epithelial cells

Cellular Signalling 22 (2010) 395–403

Contents lists available at ScienceDirect

Cellular Signalling j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c e l l s i g

Non-T cell activation linker regulates ERK activation in Helicobacter pylori-infected epithelial cells Cornelia Rieke a, Thilo Kähne a, Katrin Schweitzer a, Burkhart Schraven b, Jürgen Wienands c, Michael Engelke c, Michael Naumann a,⁎ a b c

Institute of Experimental Internal Medicine, Otto von Guericke University, Magdeburg, Germany Institute of Molecular and Clinical Immunology, Otto von Guericke University, Magdeburg, Germany Institute of Cellular and Molecular Immunology, Georg August University, Göttingen, Germany

a r t i c l e

i n f o

Article history: Received 11 September 2009 Received in revised form 18 October 2009 Accepted 19 October 2009 Available online 28 October 2009 Keywords: Detergent-resistant membrane microdomain Transmembrane adapter protein Phosphorylation c-Met Growth factor receptor-bound protein 2 Cytosolic phospholipase A2

a b s t r a c t It is supposed that human pathogens, e.g. Helicobacter pylori abuse lipid raft domains on the host cell plasma membrane to infect the cell. Investigating DRM-associated molecules we identified the transmembrane adapter proteins (TRAPs), non-T cell activation linker (NTAL) and lymphocyte-specific protein tyrosine kinase (Lck)-interacting membrane protein (LIME) to be regulated by H. pylori in the human epithelial cell line HCA-7. Up to now, raft-associated TRAPs were exclusively described to mediate signal propagation downstream of antigen receptors. Our results posed the question whether these proteins adopt a role in H. pylori-infected epithelial cells too. Our studies revealed that H. pylori induces tyrosine phosphorylation of NTAL as well as LIME within 15 min of infection. We observed that activated NTAL and LIME bind to the Src homology 2 (SH2)-domain of growth factor receptor-bound protein 2 (Grb2) within 15 to 30 min of infection and associate with the c-Met receptor. Further, NTAL has a contributory role in regulating H. pyloriinduced extracellular signal-regulated kinase (ERK) activation. After suppression of NTAL protein levels by siRNA, ERK phosphorylation was reduced to approximately 50%. Additionally, the knockdown of NTAL suppressed the phosphorylation of cytosolic phospholipase A2 (cPLA2). Activated cPLA2 catalyzes the release of arachidonic acid (AA), whose metabolites are pivotal mediators in the H. pylori-induced inflammatory response. Thus, we propose that NTAL participates in the activation of the c-Met-Grb2-ERK-cPLA2 signalling cascade at early stages of H. pylori infection. © 2009 Elsevier Inc. All rights reserved.

1. Introduction The gram-negative, microaerophilic bacterium Helicobacter pylori is one of the most successful human microbial pathogens [1]. It persistently colonizes the stomach of half the world's population increasing the risk of peptic ulcer disease, gastric adenocarcinoma and mucosa-associated lymphoid tissue (MALT) lymphomas [1]. The cag Abbreviations: TRAP, transmembrane adapter protein; NTAL, non-T cell activation linker; LIME, lymphocyte-specific protein tyrosine kinase (Lck)-interacting membrane protein; LAT, linker for activation of T cells; PAG, protein associated with glycosphingolipid-enriched microdomains; SH2, Src homology 2; Grb2, growth factor receptor-bound protein 2; ERK, extracellular signal-regulated kinase; cPLA2, cytosolic phospholipase A2; AA, arachidonic acid; PAI, cag pathogenicity island; T4SS, type IV secretion system; Cag A, cytotoxin associated gene A protein; VacA, vacuolating toxin; SFK, Src-family kinase; Csk, carboxyl-terminal Src kinase; HGF, hepatocyte growth factor; EGFR, epidermal growth factor receptor; siRNA, small interfering RNA; DRM, detergent-resistant membrane microdomain; PBMC, peripheric blood mononuclear cell; COX-2, cyclooxygenase-2; PGE2, prostaglandin E2. ⁎ Corresponding author. Institute of Experimental Internal Medicine, Otto von Guericke University, Leipziger Str. 44, 39120 Magdeburg, Germany. Tel.: +49 391 6713227; fax: +49 391 6713312. E-mail address: [email protected] (M. Naumann). 0898-6568/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2009.10.010

pathogenicity island (PAI), a cluster of up to 28 genes, is one of the best characterized virulence factors. It encodes a type IV secretion system (T4SS), that translocates cytotoxin associated gene A protein (CagA), muropeptides and possibly other molecules into epithelial host cells [2]. H. pylori is attracted to lipid rafts by sensing a cholesterol gradient [3]. Contact with lipid rafts was shown to be required for the H. pylori virulence factor vacuolating toxin (VacA) [4] to enter the cell and for the T4SS-mediated translocation of CagA [5]. Numerous studies have demonstrated that pathogens abuse lipid raft domains in the host cell plasma membrane as concentration devices and signalling platforms [6]. Lipid rafts are liquid-ordered (lo) phase microdomains proposed to exist in biological membranes. They have been widely studied by isolating detergent-resistant membrane microdomains (DRMs). Although DRMs isolated from cells do not correspond precisely to pre-existing rafts in living cells, enrichment of a protein in DRM fractions indicates that it is raftophilic and tends to associate with rafts when they form [7]. In hematopoietic cells, lipid raft-associated transmembrane adapter proteins (TRAPs) are involved in modulating signal transduction by recruiting Src homology 2 (SH2)-domain containing cytosolic molecules to the cell membrane via multiple tyrosine-

396

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

based signalling motifs. TRAPs are substrates of Src-family kinases (SFKs). Four raft-associated TRAPs have been identified: linker for activation of T cells (LAT), protein associated with glycosphingolipidenriched microdomains (PAG), non-T cell activation linker (NTAL) and lymphocyte-specific protein tyrosine kinase (Lck)-interacting membrane protein (LIME). Activated PAG and LIME are capable of recruiting the catalytic SFK inhibitor carboxyl-terminal Src kinase (Csk) to the membrane [8]. Among the lipid raft-associated TRAPs LAT, NTAL and LIME are implicated in the assembly of growth factor receptor-bound protein 2 (Grb2)-containing complexes stimulating extracellular signal-regulated kinase (ERK) activation in different hematopoietic cell types [9–11]. Grb2 binds to the specific phosphopeptide motif pYXNX via its central SH2-domain. TRAPs may recruit Grb2 from the cytosol to the membrane and thereby deliver Grb2-bound Son of sevenless homolog 1 (Sos1) in proximity to the small GTPase Ras, which then activates ERK [8]. So far TRAPs have not been described to be involved in tyrosine kinase receptor signalling in epithelial cells. However, Grb2 is known to associate with tyrosine kinase receptors, including hepatocyte growth factor (HGF) receptor (c-Met). HGF treatment of epithelial cells induces ERK phosphorylation [12]. It was shown by Churin et al. [13] that H. pylori activates c-Met independent of the PAI. ERK is also activated by H. pylori in a PAI-independent manner, but PAI-positive strains induce moderately stronger ERK phosphorylation [14]. In addition to c-Met, other receptors, e.g. epidermal growth factor receptor (EGFR) [15] or human EGFR2 (Her2/Neu) [13], are capable to induce ERK activation. In the present study we analysed lipid raft-associated TRAPs in H. pylori-infected human epithelial cells. Herein, we provide evidence that NTAL and LIME are regulated by H. pylori. An inducible interaction of NTAL and LIME with the Grb2-SH2-domain as well as with the c-Met receptor is demonstrated. Furthermore, we show that NTAL exerts a functional role in H. pylori-induced phosphorylation of ERK and cPLA2.

sion plasmids using Effectene Transfection Reagent (Qiagen, Hilden, Germany) or with siRNA using siLentFect™ Lipid Reagent (Bio-Rad, Munich, Germany) according to the manufacturer's instructions. 2.3. Plasmid construction Expression plasmids coding for NTAL or LIME respectively were constructed by cloning the sequences into the eukaryotic expression vector pcDNA3 (invitrogen, Paisley, UK). In addition, the T7- and the His6-tag sequences were included in the pcDNA3 vector at the 5′ end of the MCS. 2.4. Small interfering RNA Knockdown of NTAL and LIME expression was achieved by transfection of cells with custom-synthesized small interfering RNA (siRNA) duplexes. The target sequences were as follows: 5′-GGUGCAAAGAGGUCAGAGA-3′ (NTAL) or 5′-GGGUCUGCAAGCCUAAAAG-3′ (LIME). Scrambled siRNA (siNeg) was used as a negative control. 2.5. RNA isolation, reverse transcription-PCR and quantitative real timePCR Total RNA was extracted with RNeasy Mini Kit (Qiagen). cDNA was synthesized from 2 µg of RNA using oligo(dT)18 primers, dNTPs, Ribonuclease inhibitor (Fermentas, St. Leon-Rot, Germany) and AMV Reverse Transcriptase (Promega, Mannheim, Germany). For quantitative real time-PCR primers (0.25 µM), fluorescein (1: 1 × 105, BioRad) and SensiMix (Quantace, Berlin, Germany) were added in a dilution of 1: 20 to the cDNA mixture. Reactions were performed in the iCycler (Bio-Rad) using the following protocol: denaturation (95 °C for 10 min), amplification (40 cycles: 95 °C for 15 s, 60 °C for 30 s, 72 °C for 30 s), melting curve program (72–95 °C; 0.5 °C), and cooling to 22 °C. 2.6. H. pylori cultivation and infection

2. Materials and methods 2.1. Materials The following antibodies were used to detect the respective proteins: Caveolin (Abcam, Cambridge, UK), Clathrin Heavy Chain, PTP1D/SHP2 (BD Biosciences, San Jose, CA, USA), EGFR 1005, ERK, Flotillin1, Grb2 E-1, GST, LIME-06, Met C-28, phospho-Tyr99, phospho-Tyr20 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), phospho-cPLA2 (Ser505), cPLA2, phospho-ERK1/2, phosphoMet (Tyr1234/1235), NTAL/LAB, phospho-Tyr100 (Cell Signalling Technology Inc., Danvers, MA, USA), NAP-08 (V. Horejsi, Prague, Czech Republic), phospho-Tyr4G10 (Upstate/Millipore, Schwalbach, Germany), T7-tag (Novagen, Merck Chemicals Ltd., Beeston, UK). EGF was purchased from Sigma (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany), HGF from Calbiochem (Merck Chemicals Ltd., Beeston, UK) and tumour necrosis factor (TNF)α from R&D Systems (WiesbadenNordenstadt, Germany). Glutathione S-transferase (GST) fusion proteins of Grb2-SH2 were prepared and used as described previously by Grabbe and Wienands [16]. 2.2. Cell culture and transfection The epithelial human colon carcinoma cell line HCA-7 (European Collection of Cell Cultures, Salisbury, UK) was cultured in RPMI 1640 complete medium (PAA Laboratories, Cölbe, Germany) containing 10% fetal calf serum, penicillin (100 U/ml) and streptomycin (100 µg/ ml) in a humidified 5% CO2 atmosphere. For transfection cells were cultured free of serum and antibiotics in either 60 mm- or 100 mm-dishes and were transfected with expres-

The H. pylori strain P1 wild type (wt), its isogenic mutant strains cagA and virB7 lacking a functional T4SS [17] as well as H. pylori P14 possessing an inactivated vacA gene (kindly provided by R. Haas, Max von Pettenkofer-Institute for Hygiene and Medical Microbiology, Munich, Germany) were cultured on agar plates containing 10% horse serum for 48–72 h in a microaerophilic atmosphere (generated by Campy Gen, Oxoid, Basingstoke, UK) at 37 °C. HCA-7 cells were grown in complete medium to reach 60% of confluency. 16 h before infection, the medium was replaced by fresh RPMI 1640 medium. For infection, the cell monolayer (60–80% confluence) was incubated with the bacteria at a multiplicity of infection (MOI) of 100 for different periods of time. 2.7. Preparation of detergent-resistant membrane microdomain (DRM)fractions All steps were conducted at 4 °C. 0.8–1 × 108 cells washed in PBS were lysed in 0.5 ml TNE buffer (50 mM Tris, 100 mM NaCl, 5 mM EDTA) supplemented with 1 mM Na3VO4, 50 mM NaF, 10 mM Na4P2O7 and 1 mM AEBSF in the presence of 3% Brij-58 (Pierce/ Perbio Science, Bonn, Germany). Cell lysates were homogenized with 10 strokes of a Potter-Elvehjem homogenizer, mixed with an equal volume of 80% sucrose (w/v) in MNE (25 mM MES, 150 mM NaCl and 5 mM EDTA) and placed in Beckmann ultracentrifuge tubes. The samples were immediately overlaid with 1 ml 30% sucrose and 0.5 ml 5% sucrose (w/v) in MNE. The gradient was centrifuged for 20 h at 200.000 g at 4 °C. Membrane domains were harvested by collecting 5 fractions of 250 µl (R1–R5) and 4 fractions of 310 µl (F1–F4) from the top of the gradient.

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

397

2.8. Cell lysis, immunoprecipitation and GST-SH2 pull down Cells were washed once in PBS and then lysed in RIPA buffer [50 mM Tris (pH 7.5), 150 mM NaCl, 5 mM EDTA, 10 mM K2HPO4, 10% (v/v) glycerol, 1% (v/v) Triton X-100, 0.15% SDS, 1 mM Na3VO4, 1 mM Na2MoO4, 20 mM NaF, 10 mM Na4P2O7, 0.1 mM PMSF, 20 mM glycerol 2-phosphate and EDTA-free protease inhibitor cocktail (Roche, Mannheim, Germany)] on ice. Cells were disrupted by passing the cell suspensions through a 27-gauge needle. For immunoprecipitation lysates were incubated with appropriate antibodies for 2 h to overnight at 4 °C. The immune complexes were bound to protein G-Sepharose beads (GE Healthcare, Munich, Germany), washed, boiled in sample buffer, and subjected to immunoblot analysis. For GST-SH2 pull downs GST fusion proteins of Grb2-SH2 were bound to glutathione (GSH)-Sepharose beads (GE Healthcare). Beads were washed and incubated with cleared lysates for 4 h at 4 °C. Proteins were eluted with 50 mM Tris (pH 8.0) containing 40 mM glutathione and 20 mM Hepes. 2.9. Immunoblot analysis Proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electro-transferred to Immobilon-P transfer polyvinylidene fluoride membrane (Millipore, Schwalbach, Germany). Membranes were incubated with specific antibodies overnight at 4 °C. Immunoreactivity was detected using the chemiluminescence detection kits ECL™ Western Blotting Detection Reagents (GE Healthcare) or Super Signal® West Dura Extended Duration Substrate (Pierce/Perbio Science). Either the protein bands on x-ray films were scanned or the signal intensities were determined directly by using a Luminescent Imaging System (Versa Doc®, BioRad). The values were analysed with QuantityOne® software (BioRad) including local background subtraction. Statistics are presented as means ± S.E. of 3 independent experiments.

Fig. 1. H. pylori induces tyrosine phosphorylation of DRM-localized proteins. (A) Schematic illustration of the DRM preparation. DRMs were isolated by cell lysis in Brij-58 followed by flotation on a sucrose step gradient. Fractions from top to the bottom of the gradient were collected. For better resolution of the DRM fractions we distinguished between two different volumes (R1–R5 à 250 µl, F1–F4 à 310 µl). (B) In H. pylori-infected epithelial cells DRM fractions are enriched in tyrosine-phosphorylated proteins of low molecular weight. HCA-7 cells were left untreated or infected with H. pylori for 15, 30 or 60 min. Fractions from the top of the gradients (R1–R3) were loaded on a SDS-polyacrylamide gel and analysed by immunoblotting using antiphospho-tyrosine20, anti-Flotillin1 or anti-Caveolin1 antibodies. To compare the protein amounts in the DRM fractions the lipid raft marker proteins Flotillin and Caveolin were detected.

2.10. Statistical analysis Statistical analysis of the data in bar charts was done using Student's t test. p < 0.05 was considered significant. 3. Results 3.1. H. pylori induces tyrosine phosphorylation of proteins in DRMs H. pylori is attracted to lipid rafts by sensing a cholesterol gradient [3]. Thus, we studied lipid raft-associated proteins in H. pyloriinfected epithelial cells. In our studies we used HCA-7 cells, which retain some of the morphological and functional polarity exhibited by normal epithelium and are able to form a polarized epithelial sheet when grown on tissue culture plastic [18]. Raft-associated proteins were analysed by isolation of DRMs (Fig. 1A). Investigating DRMs of H. pylori-infected epithelial HCA-7 cells we observed enhanced tyrosine phosphorylation levels of low molecular weight proteins within the DRM fractions R1–R3. This was accompanied by a redistribution of the lipid raft marker proteins Caveolin1 and Flotillin1 to the uppermost fraction of the sucrose density gradient (Fig. 1B). Tyrosine phosphorylation reaches maximal levels already 15 min after co-culture with H. pylori (Fig. 1B). Thus, we suggest that H. pylori rapidly induces tyrosine phosphorylation of DRM-associated proteins. 3.2. Expression of TRAPs in epithelial cells By a screen with a number of different antibodies we interestingly recognized TRAPs within the DRM fractions R1–R3 of H. pyloriinfected HCA-7 cells (Fig. 2A). TRAPs represent a group of adapter molecules, which in hematopoietic cells transmit signals from

Fig. 2. Expression of lipid raft-associated TRAPs in epithelial cells. (A) After sucrose density gradient centrifugation PAG, LAT, NTAL and LIME were detected in DRM fractions (R1–R4) of HCA-7 cells infected with H. pylori for 60 min. (B) Expression of TRAPs was determined by quantitative real time-PCR. cDNAs obtained from human PBMCs or from HCA-7 cells were used to amplify the PAG, LAT, NTAL or LIME transcripts. The housekeeping gene Tubulin was used as an internal reference. The mRNA expression levels of HCA-7 cells were set in relation to those of PBMCs (= 1.0).

398

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

activated receptors to cytosolic proteins [8]. After lysis of the cells in Brij-58 and sucrose density step gradient ultracentrifugation we detected protein expression of PAG, LAT, NTAL and LIME in the DRM fractions by an immunoblot analysis (Fig. 2A). To show that the protocol applied here definitely separates the detergent-resistant raft fractions, we used Caveolin as a raft marker and Clathrin as a non-raft marker protein. Further, similar results were received when we analysed expression of TRAPs in the gastric adenocarcinoma cell line AGS, or in the gastric carcinoma cell line NCI-N87 (data not shown). Studying the expression of mRNAs of PAG, LAT, LIME and NTAL in the human epithelial cell line HCA-7 we observed that the total expression levels are low compared to peripheral blood mononuclear cells (PBMCs) ranging from 1.0% (LAT) to 5.4% (LIME) (Fig. 2B). 3.3. H. pylori enhances tyrosine phosphorylation of NTAL and LIME Based on our observations that TRAPs are expressed in epithelial cells and localized in DRMs we addressed the question, whether these proteins are regulated upon H. pylori infection. To investigate tyrosine phosphorylation in H. pylori-infected epithelial cells, we overexpressed T7-tagged NTAL as well as LIME to facilitate the performance of immunoprecipitations (IPs). Overexpression of NTAL did not interfere with the DRM localization. Like the endogenous protein, T7-tagged NTAL could be detected in the DRM fractions R2–R4 (Fig. 3A). Note that the amount of HCA-7 cells used in this experiment was too low to detect endogenous NTAL protein. Immunoprecipitations of T7-tagged NTAL or LIME revealed an inducible tyrosine phosphorylation of NTAL (Fig. 3B) and LIME (Fig. 3C) within 5–15 min of H. pylori infection. 3.4. NTAL and LIME interact with the SH2-domain of Grb2 Encouraged by our observation that NTAL and LIME are both inducibly tyrosine-phosphorylated by H. pylori, we investigated putative interacting molecules. Within the tyrosine-based signalling motifs YXNX-sequences represent potential Grb2 binding sites. In hematopoietic cells Grb2 is the major interaction partner of NTAL and LIME regulating ERK activation [9,11]. Analysis of Grb2 revealed that upon H. pylori infection NTAL and LIME interact with an expressed GST-tagged SH2-domain of Grb2 (Fig. 4A and B). Using isogenic

mutant strains (cagA and virB7) we could show that the interaction between TRAPs and Grb2 is neither CagA- nor T4SS-dependent (Fig. 4A and B). The LIME-specific antibody as well as the pTyr4G10 antibody recognized the GST-SH2 protein. To distinguish between GST-SH2 and LIME in the immunoblot non-tagged LIME was overexpressed (Fig. 4B and D). Additionally we show that endogenous Grb2 is localized in DRMs and that the association with DRM fractions R1–R4 is marginally enhanced in H. pylori-infected cells (Fig. 4G). To reveal putative upstream effectors driving NTAL and LIME regulation, we treated the cells with different stimuli activating receptor-stimulated pathways, which are also induced by H. pylori infection. Here, we show that HGF, the ligand for c-Met, similar to H. pylori infection, was able to stimulate association of NTAL and LIME with Grb2-SH2 (Fig. 4C, 5 min of stimulation, and D). In the immunoblot of the GST-SH2 pull down eluates the tyrosine-specific antibody recognized co-precipitated proteins corresponding with the molecular weight of NTAL and LIME, respectively (Fig. 4C and D, upper panel). As there are no antibodies commercially available that recognize phosphorylated NTAL or phosphorylated LIME, respectively, common phospho-tyrosine antibodies have been used. These data suggest that the tyrosine phosphorylation of NTAL and LIME enables interaction with the SH2-domain of Grb2. In contrast to H. pylori and HGF, TNFα as well as EGF treatment could only slightly induce NTALGrb2 interaction (Fig. 4C). 3.5. c-Met receptor associates with NTAL as well as LIME in H. pyloriinfected cells As previously reported by Churin et al. [13], c-Met activation and tyrosine auto-phosphorylation is induced by H. pylori wt as well as cagA and virB11 strains within 30 minutes of infection. Complementary to these data we show that also a vacA mutant strain induces phosphorylation of c-Met (Tyr1234/1235) in HCA-7 cells (Fig. 5A). To investigate the role of c-Met in the regulation of TRAPs we immunoprecipitated c-Met from H. pylori-infected cells and analysed co-precipitated proteins. Upon 15 min of H. pylori infection NTAL as well as LIME were co-immunoprecipitated in an inducible manner (Fig. 5B–D). The association of NTAL and LIME with c-Met is not dependent on the T4SS, CagA (Fig. 5B and C) or VacA (Fig. 5D). To provide evidence that c-Met is localized in DRMs we analysed the distribution of the receptor in DRM fractions. c-Met was detected in DRMs of unstimulated cells and further accumulated in the fractions R2–R4 of the sucrose gradient upon H. pylori infection (Fig. 5G). These data suggest that upon H. pylori infection the activated c-Met receptor, due to multifunctional docking sites, binds to several adapter proteins, among them Grb2 [12] and the DRMassociated TRAPs NTAL and LIME. By supporting the translocation of Grb2-bound Sos1 to DRMs in the plasma membrane NTAL and LIME could be involved in spatially regulated c-Met receptor-mediated ERK activation. 3.6. NTAL contributes to H. pylori-induced ERK activation

Fig. 3. H. pylori enhances tyrosine phosphorylation of NTAL and LIME. (A) HCA-7 cells were transfected with T7-tagged NTAL and the subcellular localization of NTAL was studied in 8 × 106 cells subjected to preparation of DRMs. (B) Cells were transiently transfected with T7-tagged NTAL. After 24 h cells were infected with H. pylori for the indicated periods of time, lysed and subjected to immunoprecipitations (IPs) with an anti-T7 antibody. Immunoblot analysis of immunoprecipitates was performed using anti-phospho-tyrosine100 antibody. (B) HCA-7 cells, which were transiently transfected with T7-tagged LIME for 24 h, were infected with H. pylori for 5 or 15 min. Lysates were subjected to IPs with an anti-T7 antibody. Immunoblot analysis of immunoprecipitates was performed using antibodies as indicated.

Studying the c-Met receptor-regulated kinase ERK we observed that ERK was recruited to DRM fractions R1–R3 with a peak after 30 min of H. pylori infection (Fig. 6A). The transient accumulation of ERK was accompanied by an increased phosphorylation of ERK (Fig. 6A, 30 min, fraction R3, ERK-P immunostaining). In order to investigate the impact of TRAPs on ERK activity we suppressed NTAL and LIME protein expression by small interfering RNA (siRNA) technology. To show efficient decrease of both proteins in RIPA lysates we additionally transfected plasmids encoding T7-tagged NTAL or LIME, respectively. After siRNA transfection very efficient knockdowns of both TRAPs compared to mock transfection were achieved (Fig. 6B). The absence of NTAL led to a significant reduction of ERK phosphorylation upon 30 min of H. pylori infection (to

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

399

Fig. 4. H. pylori stimulation as well as HGF treatment allow interaction of NTAL and LIME with the Grb2-SH2-domain. HCA-7 cells were transiently transfected (A, C) with T7-tagged NTAL or (B, D) with non-tagged LIME. After 24 h, cells were infected either with H. pylori wt, cagA or virB7 mutants or treated with HGF (20 ng/ml), EGF (20 ng/ml) or TNFα (20 ng/ ml). Total cell lysates were prepared and incubated with GST-Grb2-SH2 bound to GSH-Sepharose. Immunoblot analysis of NTAL and LIME affinity-purified (AP) by GST-Grb2-SH2 was performed using antibodies as indicated. (E, F) Quantitative analysis of affinity-purified NTAL and LIME. Graphs represent the mean-fold changes of band intensity from three separate experiments ± S.E. with the value for the control as 1 arbitrarily. *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus non-infected cells were calculated by Student's t test. (G) Endogenous Grb2 was studied in DRM fractions of unstimulated cells and of cells infected with H. pylori for 60 min, and analysed by immunoblotting.

approximately 50%). In contrast to NTAL, LIME had a minor influence on ERK phosphorylation (Fig. 6C). No further reduction of ERK activation could be achieved by knocking down the expression of both proteins. These data suggest that besides NTAL and LIME other regulatory proteins contribute to the control of ERK activity.

3.7. NTAL regulates the phosphorylation of the ERK target molecule cPLA2 In a previous study we demonstrated that the colonization of epithelial cells by H. pylori leads to an activation of the ERK target

400

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

Fig. 5. H. pylori-activated c-Met interacts with NTAL and LIME. (A) c-Met tyrosine phosphorylation at Y1234/1235 after HGF treatment as well as upon H. pylori infection (wt and vacAdeficient strain P14) was recognized after IP by immunoblotting. (B–D) c-Met was immunoprecipitated from lysates of HCA-7 cells overexpressing T7-tagged NTAL or LIME, respectively and prepared at the indicated times of H. pylori infection (wt, virB7, cagA, vacA) or HGF stimulation (20 ng/ml). Lysates were subjected to c-Met IPs. Immunoblot analysis of immunoprecipitates was performed using anti-NTAL antibody NAP-08 or anti-LIME antibody. (E, F) Relative protein levels of co-precipitated NTAL or LIME respectively were quantitatively analysed as described under Materials and methods section. *, p < 0.05; ***, p < 0.001 (G) DRM fractions of unstimulated cells and of cells infected with H. pylori for 30 min were prepared and analysed by immunoblotting using an anti-c-Met antibody.

molecule cPLA2 [19]. Dependent on the subcellular localization of Ras, ERK preferentially targets specific substrates. Phosphorylation of cPLA2 is most prominent, when ERK activation is associated with lipid

rafts [20]. In addition to our data, that upon H. pylori infection ERK is fast and transiently active in DRM fractions (Fig. 6A), we demonstrate that in the same period of time (30–60 min) cPLA2 becomes activated

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

401

Fig. 6. NTAL contributes to H. pylori-induced ERK phosphorylation. HCA-7 cells were left untreated or were infected with H. pylori for the indicated periods of time. (A) DRM fractions R1–R3 were subjected to immunoblot analysis and the abundance of ERK as well as ERK phosphorylation were studied. (B) NTAL and LIME protein levels were efficiently downregulated by specific siRNA sequences (each 20 nM) compared to negative control siRNA (siNeg) sequences. For immunoblot analysis the cells were transiently transfected with plasmids encoding T7-tagged NTAL or LIME. ERK was immunodetected to show equal protein amounts in the samples. (C) HCA-7 cells, which were transiently transfected with negative control siRNA (siNeg) or NTAL/LIME-specific siRNAs (siNTAL/siLIME), were infected with H. pylori for 30 min. Lysates were analysed by immunoblotting with antiphospho-ERK (ERK-P) or anti-ERK antibodies. Luminescent imager (Versa Doc®) signals of phospho-ERK and ERK were determined and quantified using QuantityOne® analysis software. Subsequently pERK/ERK-ratios were calculated, related to non-infected cells and additionally set in proportion to those of cells transfected with negative control siRNA. Quantitative analysis of NTAL/LIME siRNA effects on H. pylori-induced ERK phosphorylation was done as described under Materials and methods section. ***, p < 0.001 versus siNegtreated cells was calculated by Student's t test.

by phosphorylation of Ser505 (Fig. 7A, cPLA2-P immunostaining). Treatment of cells with NTAL-specific siRNA suppressed cPLA2 phosphorylation at early stages of H. pylori infection (Fig. 7A and B). However, at later time points (>90 min) the knockdown of NTAL is not suppressive for cPLA2 phosphorylation (Fig. 7A). These data suggest that apart from the early response regulated by NTAL other signalling processes (e.g. non-DRM originated) contribute to cPLA2 phosphorylation at later time points of infection.

Fig. 7. NTAL contributes to the activation of the ERK target molecule cPLA2. HCA-7 cells were transfected with negative control siRNA (siNeg) or NTAL-specific siRNA (siNTAL). After 48 h, cells were left untreated or were infected with H. pylori for the indicated periods of time. (A) Lysates were analysed by immunoblotting with anti-phosphocPLA2 or anti-cPLA2 antibodies. (B) Luminescent imager (Versa Doc®) signals of phospho-cPLA2 and cPLA2 were determined and quantified using QuantityOne® analysis software. Afterwards phospho-cPLA2/cPLA2-ratios were calculated and related to untreated cells. pcPLA2/cPLA2-ratios of NTAL knockdown cells were set in proportion to those of cells transfected with negative control siRNA. *, p < 0.05; ***, p < 0.001, relative to siNeg-treated cells.

4. Discussion Lipid rafts represent dynamic structures that might play a role in signal transduction. Upon stimulation of cells individual rafts may cluster together to connect lipid raft-associated proteins and interacting molecules into well defined protein complexes. Thereby, spatially regulated signalling processes can be induced [21]. Accumulating studies support the view that pathogens co-opt lipid rafts for initial contact and subsequent attacks [6]. Investigating DRM fractions of AGS cells by isobaric tag for relative and absolute protein quantification (iTRAQ)-based proteomic analysis Zeaiter et al. [22] could identify 21 host cell proteins to be increased or decreased during infection with H. pylori. Using CagA- and VacA-deficient isogenic mutant strains they could conclude that H. pylori might induce cellular responses via contact to lipid rafts mainly by CagA-independent mechanisms. Here, we show that H. pylori augments protein tyrosine phosphorylation in DRM fractions within 15 min of infection. Tyrosine phosphorylation is a hallmark of transmembrane signalling. Receptors with intrinsic tyrosine kinase activities, like c-Met, stimulate a plethora of signalling pathways through phospho-tyrosine-mediated recruitment of PTB- or SH2-domain containing adapter proteins to the cytoplasmic side of the plasma membrane [23]. This view was extended by Simons and Toomre [21] who introduced lipid rafts as platforms for the assembly of individual receptors. Immunoglobulin E signalling was the first process that has been shown to involve lipid rafts. The conversion of immune receptor-mediated signals into appropriate cellular responses is supported by a highly specialized group of integral membrane proteins known as TRAPs [8]. Upon tyrosine phosphorylation these molecules bind SH2-domain containing cytosolic signalling and effector molecules. For the first time, we have demonstrated that lipid raft-associated TRAPs are also expressed in human epithelial cells. We observed that NTAL and LIME are phosphorylated early after H. pylori infection. Our results suggest that NTAL and LIME could be involved in organizing signalling complexes downstream of the c-Met receptor, which is activated by H. pylori in a PAI-independent manner [13]. Performing immunoprecipitations of c-Met, we could demonstrate that association of the c-Met receptor with NTAL as well as with LIME can be

402

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

stimulated by H. pylori infection independently of the bacterial T4SS or the effector proteins CagA and VacA. In accordance with the data of Seveau et al. [24] we could detect cMet in DRM fractions of human epithelial cells. Activation of c-Met results in phosphorylation of two carboxyl-terminal tyrosine residues (Y1349VHV, Y1356VNV) providing multifunctional docking sites for effector proteins with PTB- or SH2-domains. Among those are SH2domain containing transforming protein (Shc) or c-Src, whose HGFinduced kinase activity correlates with its ability to associate with c-Met [12,25]. In hematopoietic cells TRAPs are phosphorylated by SFKs [8]. Thus, we suppose, that c-Src could be a candidate kinase catalyzing the tyrosine phosphorylation of NTAL and LIME upon H. pylori infection (Fig. 8). Phosphorylated tyrosine residues within the cytosolic parts of NTAL and LIME are likely to mediate their interaction with Grb2. Several studies indicate a central role for Ras in c-Met-mediated responses. Ras can be activated via Grb2/Sos1 association with phosphorylated c-Met. It is supposed that Grb2 binds exclusively to the phosphorylated Y1356VHV-motif of c-Met [12]. Interestingly, interactions of SH2-

domains with the activated c-Met receptor are characterized by fast association and dissociation rates allowing multiple effector proteins to bind to the same docking site [26]. Recently McDonald et al. [27] provided evidence that soluble Grb2 exists in a monomer-dimer equilibrium, which in quiescent cells is in favour of the monomer, but after stimulation may be shifted to the dimer resulting in enhanced stability and greater specificity. Moreover, Harmer et al. [28] demonstrated that in B-lymphocytes both Grb2 binding sites are required for the efficient formation of Shc/ Grb2/Sos1 complexes. They assumed that two Grb2 molecules are needed for stable binding of one molecule Sos1 to one molecule Shc. However, in HGF-treated HepG2 cells direct binding of Grb2 to c-Met is required for strong ERK activation [29]. Hence, NTAL may support the recruitment of Grb2-dimers to the activated c-Met receptor and stabilize the c-Met/Grb2/Sos1 complex within DRMs. Subsequently, Grb2-bound Sos1 could activate DRMlocalized Ras resulting in ERK activation. This hypothesis is substantiated by our finding that H. pylori-induced ERK phosphorylation is

Fig. 8. Model of H. pylori-activated, lipid raft-coupled c-Met receptor signalling. ① H. pylori rapidly induces activation of the c-Met receptor. Phosphorylated tyrosine residues of c-Met represent docking sites for SH2-domain containing signalling molecules, e.g. c-Src or Grb2. NTAL could be phosphorylated either by the kinase domain of the receptor or by c-Src. ② Grb2 can bind to phosphorylated YXNX-motifs of NTAL. Thereby NTAL could support the recruitment of Grb2 to lipid raft-domains of the plasma membrane, and may stabilize the c-Met–Grb2–Sos1-protein complex. ③ When guanine nucleotide-exchange factor Sos1 is delivered in close proximity to the small GTPase Ras, within 30 min of infection Ras-Raf-1-MEK1/2-ERK pathway can be activated by converting membrane-bound Ras-GDP into Ras-GTP. ④ cPLA2 is primarily regulated by lipid raft-associated Ras. For cPLA2 activation, ERK interacts with the scaffold protein Kinase Suppressor of Ras (KSR). ⑤ Further, the bacterial effector protein CagA is translocated into epithelial cells via the T4SS and phosphorylated by SFKs. CagA interacts with tyrosine-phosphorylated c-Met and could represent an adapter protein, which contributes to ERK activation at late stages of infection (> 60 min) via Ras-independent mechanisms.

C. Rieke et al. / Cellular Signalling 22 (2010) 395–403

affected by NTAL. We demonstrate that NTAL is required for full ERK activation, while LIME seems to have no impact. In addition to c-Met signalling several other ERK-activating pathways are induced by H. pylori infection, e.g. via EGFR [15] or Her2/Neu [13]. This could explain why diminishing NTAL protein levels only partially reduced phosphorylation of ERK. In contrast to NTAL, LIME is able to associate with Csk [8] suggesting that LIME rather is involved in negative feedback loop mechanisms by supporting Csk-mediated SFK inactivation. It has been proposed by Pillinger et al. [30] that H. pylori stimulates ERK phosphorylation via early (<30 min), CagA-independent and late (1–2 h), CagA-dependent pathways. From our results we draw the conclusion that NTAL modulates c-Met-mediated ERK activation independent of the T4SS or CagA at early stages (30 min) of H. pylori infection by supporting the activation of DRM-associated Ras isoforms (Fig. 8). At later time points of infection (>60 min) ERK may be also stimulated via CagA-dependent mechanisms. In H. pylori-infected epithelial cells CagA translocation as well as ERK signalling are critical for the induction of the motogenic response [31]. The nature of ERK signalling strongly depends on its subcellular localization. With the help of different scaffold proteins ERK can be distinctively coupled to various substrates. For instance, in response to growth factor treatment kinase suppressor of Ras (KSR) translocates from the cytoplasm to lipid raft domains enabling the co-localization of MEK, Raf and ERK [20,32]. Moreover, Casar et al. [20] showed that phosphorylation of cPLA2 is mainly stimulated by Ras signals emanating from lipid rafts. According to this we demonstrate that reduction of NTAL protein levels results in diminished cPLA2 activation. The ubiquitously distributed cPLA2 catalyzes the hydrolysis of sn-2 ester bonds of glycerolipids to release arachidonic acid (AA). Metabolites of AA as the cyclooxygenase-2 (COX-2)-produced prostaglandin E2 (PGE2) might play a crucial role in the initiation and propagation of H. pylori-induced gastric inflammation [19]. 5. Conclusion For the first time we investigated the function of NTAL and LIME in H. pylori-infected epithelial cells. Our data indicate that H. pylori rapidly induces tyrosine phosphorylation of NTAL and LIME and enables their association with the Grb2-SH2-domain. In addition, we show that after H. pylori infection as well as after HGF treatment NTAL and LIME co-immunoprecipitate with c-Met. We have also established that NTAL has a contributory role in regulating H. pylori-induced ERK activation. Further, we demonstrate that the knockdown of NTAL suppresses the phosphorylation of the ERK target molecule cPLA2. Activated cPLA2 generates AA and thereby is involved in the initiation and propagation of the H. pylori-induced inflammatory response. Overall, our results disclose a novel function of the adapter protein NTAL in human epithelial cells. Here, DRM-associated NTAL might be involved in the c-Met-Grb2-ERK-cPLA2 signalling cascade at early stages of H. pylori infection.

403

Acknowledgments The work was funded by the Deutsche Forschungsgemeinschaft (GRK 1167/1), and the “Centre of Behavioural Brain Sciences” (CBBS) within the program networks of scientific excellence in SaxonyAnhalt and the Förderprogramm “Biotechnologie - Chancen nutzen und gestalten” Fördermodul: Forschungseinheiten der Systembiologie-FORSYS Bundesministerium für Bildung und Forschung (BMBF) (0313922) by grants to M.N.

References [1] M. Naumann, J.E. Crabtree, Trends Microbiol. 12 (2004) 29. [2] M. Naumann, Int. J. Med. Microbiol. 295 (2005) 335. [3] C. Wunder, Y. Churin, F. Winau, D. Warnecke, M. Vieth, B. Lindner, U. Zahringer, H.J. Mollenkopf, E. Heinz, T.F. Meyer, Nat. Med. 12 (2006) 1030. [4] C.H. Kuo, W.C. Wang, Biochem. Biophys. Res. Commun. 303 (2003) 640. [5] C.H. Lai, Y.C. Chang, S.Y. Du, H.J. Wang, C.H. Kuo, S.H. Fang, H.W. Fu, H.H. Lin, A.S. Chiang, W.C. Wang, Infect. Immun. 76 (2008) 3293. [6] F.G. van der Goot, T. Harder, Semin. Immunol. 13 (2001) 89. [7] D.A. Brown, Physiology 21 (2006) 430. [8] V. Horejsi, W. Zhang, B. Schraven, Nat. Rev. Immunol. 4 (2004) 603. [9] E. Ahn, H. Lee, Y. Yun, Blood 107 (2006) 1521. [10] W. Zhang, B.J. Irvin, R.P. Trible, R.T. Abraham, L.E. Samelson, Int. Immunol. 11 (1999) 943. [11] E. Janssen, M. Zhu, W. Zhang, S. Koonpaew, W. Zhang, Nat. Immunol. 4 (2003) 117. [12] K.A. Furge, Y.W. Zhang, G.F. Vande Woude, Oncogene 19 (2000) 5582. [13] Y. Churin, L. Al-Ghoul, O. Kepp, T.F. Meyer, W. Birchmeier, M. Naumann, J. Cell Biol. 161 (2003) 249. [14] S. Wessler, M. Hocker, W. Fischer, T.C. Wang, S. Rosewicz, R. Haas, B. Wiedenmann, T.F. Meyer, M. Naumann, J. Biol. Chem. 275 (2000) 3629. [15] C. Wallasch, J.E. Crabtree, D. Bevec, P.A. Robinson, H. Wagner, A. Ullrich, Biochem. Biophys. Res. Commun. 295 (2002) 695. [16] A. Grabbe, J. Wienands, Blood 108 (2006) 3761. [17] M. Naumann, S. Wessler, C. Bartsch, B. Wieland, A. Covacci, R. Haas, T.F. Meyer, J. Biol. Chem. 274 (1999) 31655. [18] S.C. Kirkland, Cancer Res. 45 (1985) 3790. [19] T. Pomorski, T.F. Meyer, M. Naumann, J. Biol. Chem. 276 (2001) 804. [20] B. Casar, I. Arozarena, V. Sanz-Moreno, A. Pinto, L. Gudo-Ibanez, R. Marais, R.E. Lewis, M.T. Berciano, P. Crespo, Mol. Cell. Biol. 29 (2009) 1338. [21] K. Simons, D. Toomre, Nat. Rev., Mol. Cell Biol. 1 (2000) 31. [22] Z. Zeaiter, D. Cohen, A. Musch, F. Bagnoli, A. Covacci, M. Stein, Cell. Microbiol. 10 (2008) 781. [23] T. Hunter, Keio J. Med. 51 (2002) 61. [24] S. Seveau, H. Bierne, S. Giroux, M.C. Prevost, P. Cossart, J. Cell Biol. 166 (2004) 743. [25] N. Rahimi, W. Hung, E. Tremblay, R. Saulnier, B. Elliott, J. Biol. Chem. 273 (1998) 33714. [26] C. Ponzetto, A. Bardelli, Z. Zhen, F. Maina, Z.P. Dalla, S. Giordano, A. Graziani, G. Panayotou, P.M. Comoglio, Cell 77 (1994) 261. [27] C.B. McDonald, K.L. Seldeen, B.J. Deegan, M.S. Lewis, A. Farooq, Arch. Biochem. Biophys. 475 (2008) 25. [28] S.L. Harmer, A.L. DeFranco, Mol. Cell. Biol. 17 (1997) 4087. [29] A. Kondo, N. Hirayama, Y. Sugito, M. Shono, T. Tanaka, N. Kitamura, J. Biol. Chem. 283 (2008) 1428. [30] M.H. Pillinger, N. Marjanovic, S.Y. Kim, Y.C. Lee, J.U. Scher, J. Roper, A.M. Abeles, P.I. Izmirly, M. Axelrod, M.Y. Pillinger, S. Tolani, V. Dinsell, S.B. Abramson, M.J. Blaser, J. Biol. Chem. 282 (2007) 18722. [31] R. Franke, M. Müller, N. Wundrack, E.D. Gilles, S. Klamt, T. Kähne, M. Naumann, BMC Syst. Biol. 2 (4) (2008) 4. [32] J. Müller, S. Ory, T. Copeland, H. Piwnica-Worms, D.K. Morrison, Mol. Cell 8 (2001) 983.