Pathway in Gastric Epithelial Cells IL-8 Secretion Via the TLR4

Home , TLR5

The C-Terminal Disulfide Bonds of Helicobacter pylori GroES Are Critical for IL-8 Secretion via the TLR4-Dependent Pathway in Gastric Epithelial Cells This information is current as of September 26, 2021. Yu-Lin Su, Jyh-Chin Yang, Haur Lee, Fuu Sheu, Chun-Hua Hsu, Shuei-Liong Lin and Lu-Ping Chow J Immunol 2015; 194:3997-4007; Prepublished online 13 March 2015; doi: 10.4049/jimmunol.1401852 Downloaded from http://www.jimmunol.org/content/194/8/3997

References This article cites 66 articles, 25 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/194/8/3997.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 26, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

The C-Terminal Disulﬁde Bonds of Helicobacter pylori GroES Are Critical for IL-8 Secretion via the TLR4-Dependent Pathway in Gastric Epithelial Cells

Yu-Lin Su,* Jyh-Chin Yang,† Haur Lee,* Fuu Sheu,‡ Chun-Hua Hsu,x Shuei-Liong Lin,{ and Lu-Ping Chow*

Helicobacter pylori GroES (HpGroES), a potent immunogen, is a secreted virulence factor that stimulates production of proin- ﬂammatory cytokines and may contribute to gastric carcinogenesis. HpGroES is larger than other bacterial orthologs because of an additional C-terminal region, known as domain B. We found that the HpGroES-induced IL-8 release by human gastric epithelial cells was dependent on activation of the MAPK and NF-kB pathways. HpGroES lacking domain B was unable to

induce IL-8 release. Additionally, a TLR4 inhibitor significantly inhibited IL-8 secretion and reduced HpGroES-induced activa- Downloaded from tion of MAPKs. Furthermore, HpGroES-induced IL-8 release by primary gastric epithelial cells from TLR42/2 mice was significantly lower than from wild-type mice. We also found that HpGroES bound to TLR4 in cell lysates and colocalized with TLR4 on the cell membrane only when domain B was present. We then constructed two deletion mutants lacking C-terminal regions and mutants with point mutations of two of the four cysteine residues, C111 and C112, in domain B and found that the deletion mutants and a double mutant lacking the C94–C111 and C95–C112 disulfide bonds were unable to interact with TLR4 or induce IL-8 release. We conclude that HpGroES, in which a unique conformational structure, domain B, is generated by these two http://www.jimmunol.org/ disulfide bonds, induces IL-8 secretion via a TLR4-dependent mechanism. The Journal of Immunology, 2015, 194: 3997–4007.

elicobacter pylori is a Gram-negative, microaerophilic epithelial and immune cells (9–11). It is well recognized that re- bacterium that selectively colonizes the human stomach lease of these proinflammatory cytokines is closely linked to the H (1). H. pylori infection is a major cause of chronic pathogenesis of H. pylori–associated gastric cancer (12, 13). In- gastritis and peptic ulcer disease and is highly related to gastric duction of cytokine secretion by H. pylori depends on both the adenocarcinoma and MALT lymphoma (2). Despite a recent de- host genetic background and microbial virulence (9). Among cline in infection rates, the prevalence rate of H. pylori remains at these cytokines, IL-8 is the most highly expressed gene in nearly 50% of the world’s population, and the associated gastric H. pylori–infected gastric epithelial cells and appears to play a major by guest on September 26, 2021 cancer is the second leading cause of cancer-related death role in the epithelial cell response to H. pylori infection and in the worldwide. The most remarkable feature of persistent H. pylori pathological processes leading to gastric disease (14, 15). Several infection is that it causes inflammatory responses, which is an studies have demonstrated increased IL-8 production and secre- important risk factor for malignancy (3). H. pylori infection tion in response to H. pylori both in vivo and in vitro (16, 17). induces gastritis with infiltration of neutrophils, macrophages, Gastric mucosal IL-8 levels show a positive correlation with the dendritic cells, as well as T and B lymphocytes into the gastric degree of stomach corpus inflammation, and IL-8 expression is mucosa (4–8) in addition to the accumulation of various cyto- also highly increased in gastric cancer (18, 19). IL-8 production kines, including TNF-a, IL-1b, IL-6, and IL-8 secreted by gastric by gastric epithelial cells is enhanced by the virulence factor cytotoxin-associated gene (cag) pathogenicity island and other factors from H. pylori (20–23). We have previously reported that *Graduate Institute of Biochemistry and Molecular Biology, College of Medicine, † H. pylori Hp National Taiwan University, Taipei 100, Taiwan; Department of Internal Medicine, GroES ( GroES) can induce the production and se- Hospital and College of Medicine, National Taiwan University, Taipei 100, Taiwan; cretion of proinflammatory cytokines, including IL-8, IL-6, IL-1b, ‡ x Department of Horticulture, National Taiwan University, Taipei 106, Taiwan; De- and TNF-a, by human PBMCs and trigger IL-8 production and partment of Agricultural Chemistry, National Taiwan University, Taipei 106, Taiwan; and {Graduate Institute of Physiology, College of Medicine, National Taiwan Uni- secretion by gastric epithelial cells (24). versity, Taipei 100, Taiwan HpGroES, also referred to as heat shock protein (HSP) A, is an Received for publication July 21, 2014. Accepted for publication February 9, 2015. unusual homolog of the essential bacterial GroES chaperonin This work was supported by National Science Council Grant NSC 101-2325-B-002- family (25), the members of which serve as cochaperonins of 065, the Liver Disease Prevention and Treatment Research Foundation, and the the heptameric GroEL/GroES barrel complex, which mediates the Ministry of Education, Taiwan. refolding of a variety of nonnative proteins (26). In addition to Address correspondence and reprint requests to Prof. Lu-Ping Chow, Graduate Insti- tute of Biochemistry and Molecular Biology, College of Medicine, National Taiwan the highly conserved GroES chaperonin domain, named domain University, No. 1, Jen-Ai Road, Taipei 100, Taiwan. E-mail address: chowip@ntu. A, HpGroES contains a C-terminal extension domain named do- edu.tw main B (residues 91–118), which is absent in the other GroES Abbreviations used in this article: Cag, cytotoxin-associated gene; CD, circular di- members (25). Domain B consists of 28 aa, including 8 histi- chroism; 4CR, four-cysteine ring; EcGroES, Escherichia coli GroES; HA, hemagglutinin; HpGroES, Helicobacter pylori GroES; HSP, heat shock protein; KC, dine and 4 cysteine residues, with the latter forming two disul- keratinocyte-derived chemokine; OxPAPC, oxidized 1-palmitoyl-2-arachidonoyl-sn- fide bonds, C94–C111 and C95–C112, generating a closed-loop glycero-3-phosphocholine; PGN, peptidoglycan; WT, wild-type. structure (27). Apart from its cochaperone activity, HpGroES Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 plays important roles in nickel homeostasis and urease activation www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401852 3998 C TERMINI CYSTEINES NEEDED FOR HpGroES-INDUCED IL-8 VIA TLR4

(25, 28, 29). Besides these roles in intracellular locations (30), DMEM. All cells were grown at 37˚C with 5% CO2 in medium supple- HpGroES is able to gain access to the extracellular compartment mented with 10% FBS and 100 U/ml penicillin-streptomycin, whereas and is highly antigenic (29, 31) and is specifically recognized by HEK293/TLR4-HA cultures were also supplemented with 100 mg/ml blasticidin. serum Abs in H. pylori–infected patients and is considered a potential candidate for vaccine development and diagnosis of Immunofluorescence staining H. pylori infection (32, 33). KATO-III and HEK293/TLR4-HA cells were seeded on coverslips over- The increased production and secretion of proinflammatory night, then treated for 1 h with 0.05 mM HpGroES or domain A at 4˚C, then cytokines induced by HpGroES has been shown to contribute to fixed with 4% paraformaldehyde in PBS for 15 min at room temperature, gastric inflammation (7, 16, 34). However, the molecular mecha- blocked by incubation for 1 h at room temperature with 5% BSA, and immunostained with rabbit Abs against HpGroES (1:200) or mouse mAb nism of proinflammatory cytokine induction has not been eluci- against TLR4 (1:200) followed by FITC- or tetramethylrhodamine iso- dated. In this study, we found that HpGroES induced IL-8 thiocyanate–labeled secondary Abs (1:500) and staining of nuclei with secretion by gastric carcinoma cells by activating the MAPK and DAPI (1:1000), and then images were captured using a TCS SP5 confocal NF-kB pathways via a TLR4-dependent mechanism. In TLR4- microscope (Leica, Wetzlar, Germany). deficient mice, HpGroES failed to induce production of the Expression and purification of HpGroES, domain A, and mouse IL-8 homolog keratinocyte-derived chemokine (KC). We Escherichia coli GroES also demonstrated direct binding of HpGroES, but not domain A Hp E. coli Ec alone, to TLR4 on gastric epithelial cells and TLR4-overexpressing The genes encoding GroES, domain A, and GroES ( GroES) were cloned using the expression vector pQE30 as described previously cells. When the two disulfide bonds in domain B were disrupted (24); the primer sequences are listed in Table I. Expression of each protein, by site-directed mutagenesis, the mutated HpGroES failed to bind bearing a 6-His tag, was induced in E. coli strain M15 by incubation for 4 h Downloaded from to TLR4, leading to diminished IL-8 production. These data at 37˚C with 1 mM isopropyl b-D-1-thiogalactopyranoside. The soluble 2+ show that the unique domain B of HpGroES may contribute to recombinant proteins were purified on a Ni -chelating Sepharose column, with endotoxin being removed using 1% Triton X-114 before elution of the HpGroES-induced IL-8 secretion through the TLR4-dependent recombinant proteins. The purified recombinant proteins were dialyzed signaling pathway. against PBS and tested for endotoxin contamination by determining the effect of proteinase K digestion on the IL-8 releasing activity. All proteins were .95% pure as shown by Coomassie blue staining on SDS-PAGE. http://www.jimmunol.org/ Materials and Methods Protein concentrations were measured using the Bio-Rad protein assay. Inhibitors, Abs, and reagents Cell fractionation All media were purchased from HyClone (Cambridge, MA), and FBS, penicillin, and streptomycin were from Life Technologies (Gaithersburg, Cytosolic and nuclear fractions were prepared from KATO-III cells using MD). Blasticidin, Zeocin, LPS, peptidoglycan (PGN), oxidized 1-palmi- a Dounce homogenization protocol. All steps were at 4˚C. Briefly, the cells toyl-2-arachidonoyl-sn-glycero-3- phosphocholine (OxPAPC), and CLI- were scraped off the culture plates and resuspended and lysed in buffer A 095 were obtained from InvivoGen (Cayla, France). Triton X-114, (10 mM HEPES [pH 7.9], 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT) thioglycollate, and protease and phosphatase inhibitor cocktails were using a Dounce homogenizer. The lysates were then centrifuged at 1,000 3 g purchased from Sigma-Aldrich (St. Louis, MO). SuperScript II reverse for 10 min and the supernatant was cleared by further centrifugation at transcriptase, collagenase I, and dispase were from Life Technologies 20,000 3 g for 15 min (cytoplasmic fraction), whereas the pellet (nuclei by guest on September 26, 2021 (Gaithersburg, MD), the pQE30 plasmid was from Qiagen (Chatsworth, fraction) was washed and resuspended in buffer A and centrifuged at 2,500 3 g CA), and the Ni2+-chelating Sepharose column was from GE Healthcare for 10 min, then the pellet was resuspended in 3 ml S1 buffer (0.25 M (Kowloon, Hong Kong). SB203580, U0126, and SP600125 were from sucrose and 10 mM MgCl2) and the suspension was layered on top of a 3-ml Calbiochem Biosciences (Darmstadt, Germany). Quantikine ELISA assay cushion of S3 buffer (0.88 M sucrose and 0.5 mM MgCl2) and centrifuged at kits for human IL-8 and mouse KC were from R&D Systems (Minneap- 250 3 g for 10 min. olis, MN). The rabbit polyclonal Abs to HpGroES were produced in our laboratory (24) and bind to intact HpGroES and domain A, but not domain Cytokine ELISA B (data not shown). The mouse anti-human phospho-p38 (1:500), anti-p38 (1:500), anti-phospho-ERK1/2 (1:500), anti-ERK1/2 (1:500), anti- KATO-III cells were incubated for 16 h with various concentrations of hemagglutinin (HA) (1:2000), and anti-actin (1:5000) mAbs were pur- HpGroES, domain A, mutants, EcGroES, LPS, or PGN; in some ex- chased from Santa Cruz Biotechnology (Santa Cruz, CA). The mouse periments, KATO-III cells were preincubated for 1 h with inhibitors mAbs against human phospho-JNK (1:500), JNK (1:500), phospho-IkB (SB203580, U0126, SP600125, OxPAPC, CLI-095, or anti-TLR2 or anti- (1:500), IkB (1:500), NF-kB (1:1000), lamin B (1:1000), and tubulin TLR4 Ab) before incubation for 16 h with HpGroES. Cell culture super- (1:1000) were from Cell Signaling Technology (Cambridge, MA). The natants were then collected and IL-8 levels measured using Quantikine mouse mAbs against human TLR4 (HTA125) and TLR2 (TL2.1) were, ELISA assay kits. Primary mouse gastric epithelial cells were also stim- respectively, from BioLegend (San Diego, CA) and Hycult Biotech (Uden, ulated as described above, and the supernatants were collected and KC The Netherlands). The rabbit polyclonal Ab against human TLR4 (S441) levels in the supernatants measured using Quantikine ELISA assay kits. was from Bioworld Technology (St. Louis Park, MN). HRP-conjugated anti-mouse IgG (1:2000) and anti-rabbit IgG (1:2000) Abs were from Immunoblotting analysis Jackson ImmunoResearch Laboratories (West Grove, PA). The poly- KATO-III cells were incubated for various times with HpGroES or were vinylidene difluoride membrane, ECL reaction solution, tetramethylrhod- preincubated for 1 h with inhibitors (SB203580, U0126, SP600125, and amine isothiocyanate–conjugated anti-rabbit or anti-mouse IgG Abs, and CLI-095) followed by incubation with HpGroES in the presence of the FITC-conjugated anti-rabbit or anti-mouse IgG Abs were from Millipore inhibitors and then were washed with PBS and lysed at 4˚C in RIPA buffer (Bedford, MA). (25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS) containing protease and phosphatase inhibitors. Bacterial strain and growth conditions Protein extracts (20 mg/lane) were separated by SDS-PAGE and transferred The H. pylori strain was isolated from endoscopic biopsy samples from the to a polyvinylidene difluoride membrane, which was then blocked by in- stomach of a gastric cancer patient at National Taiwan University Hospital. cubation for 1 h with blocking buffer (5% nonfat milk in 50 mM Tris-HCl, The bacteria were inoculated onto Columbia agar containing 5% sheep blood 150 mM NaCl [pH 7.5]). The membrane was then incubated overnight at (Invitrogen, Grand Island, NY) and grown at 37˚C in a microaerophilic 4˚C with various primary Abs diluted in blocking buffer according to the manufacturer’s instructions, with anti-actin Abs being used as the loading chamber (West Yorkshire, U.K.) in 10% CO2,5%O2,and85%N2. control, and then, after six washing steps, were incubated for 1 h at 25˚C Cell culture with the appropriate secondary Abs diluted in blocking buffer. After an- other six washing steps, bound Abs were detected using ECL reaction The human gastric cancer cell line KATO-III (Japan Cancer Research Bank, solution and visualized on an LAS4000 luminescent image analyzer Tokyo, Japan) was cultured in RPMI 1640 medium, whereas HEK293/ (Fujifilm, Tokyo, Japan) and quantified using MultiGauge software (Fuji- TLR4-HA cells (InvivoGen) overexpressing TLR4-HA were grown in film). The intensity of the protein band of interest was normalized to that of The Journal of Immunology 3999 the actin band and then this value was normalized to the corresponding Results value for control cells. HpGroES, but not HpGroES lacking domain B, induces IL-8 Isolation of primary gastric epithelial cells from mouse production and binds to KATO-III cell membranes stomach In our previous study, we found that HpGroES induces proin- Five-week-old male specific pathogen-free C57BL/6J (wild-type [WT]) and flammatory cytokine release by KATO-III cells and PBMCs (24), 2 2 C57BL/10ScNJ (TLR4 / ) mice were used and housed at the Experi- whereas EcGroES is reported not to induce proinflammatory cy- mental Animal Center, National Taiwan University at a constant temper- tokine production (39). The lack of cytokine-inducing ability of ature (35). The experimental protocol was approved by the Ethics of Animal Experiments Committee of the National Taiwan University Col- EcGroES might be due to differences in the structure of the two lege of Medicine (permit no. 20120031). GroESs. As shown in Fig. 1A, comparison of the amino acid WT and TLR42/2 mouse primary gastric epithelial cells were obtained sequences of HpGroES from H. pylori strains HC1 (clinical iso- 2/2 as described previously (36, 37). In brief, WT and TLR4 mice were lated strain), J99, and 26695 with those of seven other bacteria sacrificed and stomachs were removed and immediately cleaned in colonizing the human gastrointestinal tract showed that all the DMEM/F-12 (1:1) medium. The stomachs were cut into small pieces (1–2 mm2) and incubated in 25 ml collagenase/dispase solution (3000 U col- HpGroES proteins have an additional 28 aa at their C termini, lagenase I, 31.25 U diapase, 31.25 mg BSA in DMEM/F-12) by stirring referred to as domain B (25), which is well conserved among constantly for 2 h at 37˚C. Twelve-well plates were precoated with 0.5 ml H. pylori strains. To determine whether domain B has functional Matrigel (5 ml/ml; Corning, Corning, NY) and incubated for 2 h at 37˚C. significance, recombinant HpGroES or HpGroES lacking domain The dispersed cells were centrifuged and resuspended in DMEM/F-12 with 20% FBS and 100 U/ml penicillin-streptomycin and seeded into Matrigel- B (domain A), and EcGroES were generated and purified (Fig. 1B, Table I). coated 12-well plates and incubated for 24 h at 37˚C in 5% CO2. Non- Downloaded from adherent cells were removed by PBS washes and then the adherent primary To determine whether the ability of HpGroES to stimulate IL- gastric epithelial cells were incubated for 24 h with HpGroES, domain A, 8 secretion by KATO-III cells was related to its C-terminal EcGroES, or LPS, the supernatants were collected, and KC was measured sequence, IL-8 levels in supernatants of cells incubated with by ELISA as described above. 0.05 mM HpGroES or HpGroES domain A or EcGroES were Immunoprecipitation assays measured, using LPS as a positive control. As shown in Fig. 1C, KATO-III cells were washed with PBS and lysed at 4˚C with RIPA buffer HpGroES and LPS significantly increasedIL-8productionin http://www.jimmunol.org/ containing protease inhibitors. After centrifugation, the supernatants were KATO-III cells, whereas domain A or EcGroES did not. Similar collected and incubated at 4˚C overnight with HpGroES, domain A, or results were also observed in MKN 45 and NCI-N87 cells (data mutants (all 6-His–tagged) immobilized on Ni-NTA beads or beads alone, not shown). To further examine whether the binding of and then the beads with the bound protein complexes were washed and resuspended in Laemmli buffer (4% SDS, 20% glycerol, 10% 2-ME, HpGroES to the surface of gastric epithelial cells is due to the 0.004% bromophenol blue, and 0.125 M Tris-HCl [pH 6.8]) and subjected presence of the C-terminal extension, KATO-III cells were in- to electrophoresis and immunoblotted with anti-HpGroES (1:5000) or anti- cubatedwith0.05mM HpGroES or domain A, fixed, and in- TLR4 (1:1000) Abs. cubated with anti-HpGroES Ab and then the presence of bound HEK293/TLR4-HA cell lysates were collected as described above and HpGroES was assessed by confocal microscopy. As shown in incubated at 4˚C overnight with HpGroES or domain A, and then protein bound to TLR4-HA was immunoprecipitated by incubation with anti-HA Fig. 1D, plasma membrane staining with rabbit anti-HpGroES by guest on September 26, 2021 Ab–coated beads for 1 h at 4˚C and the precipitates were immunoblotted Abs (24) that recognize HpGroES and domain A, but not do- using anti-HpGroES (1:5000) or anti-HA (1:2000) Abs. main B (data not shown), was detected when cells were incu- The H. pylori lysates were washed and sonicated at 4˚C in PBS con- bated with HpGroES, but not with domain A. These data show taining protease inhibitors. After centrifugation, the supernatants were collected and precleaned with protein G–Sepharose beads (GE Healthcare) that HpGroES, but not domain A, can bind to cells and stimulate at 4˚C for 1 h and then incubated with TLR4 (R&D Systems)/TLR4 IL-8 secretion. polyclonal Ab complex conjugated protein G beads, or beads alone at 4˚C for 16 h. The beads with bound proteins were washed and resuspended in HpGroES-induced IL-8 expression depends on MAPK Laemmli buffer and subjected to electrophoresis and immunoblotted with phosphorylation anti-HpGroES (1:5000) or anti-TLR4 (1:1000) Abs. MAPKs are crucial for cellular responses to many external stimuli Preparation of HpGroES mutants and are important signaling pathways for the production of The 6-His–tagged truncated mutants ΔH100 and ΔK106 of HpGroES were proinflammatory cytokines (40, 41). To investigate the role of prepared by PCR using pQE30-HpGroES as a template. The 6-His–tagged MAPK cascades in HpGroES-dependent IL-8 production, phos- point mutants were generated using the Phusion site-directed mutagenesis phorylation of p38, ERK, and JNK was investigated. As shown in system (Thermo Scientific, Waltham, MA). Primer sequences are listed in Fig. 2A, p38, ERK, and JNK were phosphorylated in KATO-III Table I. The PCR-amplified truncated and point mutant–containing vectors were then transformed into E. coli M15. To confirm the presence of the cells within 30 min of stimulation with 0.05 mM HpGroES and desired mutations, the DNA sequences of the mutants were determined. activation was maintained at this level for 60 min and then de- The methods for the expression and purification of the 6-His–tagged creased, whereas in cells stimulated with domain A, no activation HpGroES mutants were similar to those described above for HpGroES. was seen (data not shown). To determine the role of MAPKs in Circular dichroism measurements HpGroES-induced IL-8 production, MAPK inhibitors were used. When the cells were pretreated for 1 h with 5 mM SB203580 (p38 Circular dichroism (CD) spectra of 5 mM HpGroES, C111A, C112A, and inhibitor), 2.5 mM U0126 (ERK inhibitor), or 1 mM SP600125 C111A/C112A between 200 and 250 nm were measured at 4˚C in a 1-mm path length quartz cell using a JASCO J-810 spectropolarimeter (Japan (JNK inhibitor) before incubation for 1 h with 0.05 mM HpGroES Spectroscopic, Tokyo, Japan). The CD spectra were analyzed using CDPro in the presence of the inhibitor, phosphorylation of p38, ERK, software to determine the secondary structure (38). All spectra were av- and JNK was blocked, respectively, and an antagonistic effect eraged from three accumulations and were buffer-corrected. occurred slightly to ERK phosphorylation under SB203580 or Statistical analysis SP600125 treatment (Fig. 2B). The IL-8 levels produced in response to 16 h with 0.05 mM HpGroES stimulation after 1 h All experiments were performed three times and the data are expressed as the mean 6 SD. A Student t test was used to determine the significance of preincubation with MAPK inhibitors were significantly lower differences between treated and control samples. For all tests, a p value , (Fig. 2C). These data demonstrated that these three MAPK 0.05 was considered statistically significant. pathways are involved in HpGroES-induced IL-8 expression. 4000 C TERMINI CYSTEINES NEEDED FOR HpGroES-INDUCED IL-8 VIA TLR4 Downloaded from http://www.jimmunol.org/

FIGURE 1. HpGroES, but not domain A, induces IL-8 secretion and binds to KATO-III cells. (A) Sequence alignment of GroESs from various H. pylori strains (HC1; clinical isolated strain, J99, and 26695) with GroESs from other bacterial species. Conserved residues are shown in boxes. (B) SDS-PAGE of HpGroES, domain A, and EcGroES stained with Coomassie blue. (C) KATO-III cells were incubated for 16 h with 0.05 mM HpGroES, domain A,

EcGroES, or 1 mg/ml LPS and then IL-8 levels in the culture supernatant were measured by ELISA. Data are presented as mean 6 SD for ﬁve experiments. by guest on September 26, 2021 **p , 0.01. (D) KATO-III cells were incubated for 1 h with 0.05 mM HpGroES or domain A and then were ﬁxed and immunostained with anti-HpGroES Abs (red) and DNA was counterstained with DAPI (blue); the merged images are shown in the right. Scale bar, 10 mm. The data shown are representative of the results obtained in three independent experiments.

HpGroES promotes NF-kB translocation to the nucleus and kB activation. As shown in Fig. 3A, incubation of KATO-III cells activates the NF-kB signaling pathway with 0.05 mM HpGroES led to IkB phosphorylation, which in- Previous studies have shown that activation of the transcription creased from 30 to 120 min, and to IkB degradation at 30 and factor NF-kB plays a major role in H. pylori–induced IL-8 pro- 60 min with less at 120 min. The nuclear localization of NF-kB duction by gastric epithelial cells (18, 42). Activated NF-kB was also investigated. As shown in Fig. 3B, NF-kB was primarily translocates to the nucleus, where it regulates IL-8 gene tran- restricted to the cytoplasm in nonstimulated cells, whereas after scription. To examine the effect of HpGroES stimulation on NF- incubation of the cells for 1 h with 0.05 mM HpGroES, levels of kB activation, IkB degradation was assessed as a measure of NF- nuclear NF-kB were increased. Lamin B and tubulin were used as

Table I. Primer sequences used for GroES construct

Construct Primer Sequence F: 59-GGA TCC ATG AAG TTT CAG CCA TTA GGA GA-39 HpGroES R: 59-GGT ACC TTA GTG TTT TTT GTG ATC ATG ACA-39 HpGroES domain A R: 59-GTC GAC TTA GCC CAC AAT ACC TAG AAT-39 HpGroES ΔH100 R: 59-CTC GAG TTA ATG ATT ACC TGT ATG ACA-39 HpGroES ΔK106 R: 59-CTC GAG TTA TTT AGC ATG TTT ATG GTC-39 HpGroES C111A F: 59-GCT TGT CAT GAT CAC AAA AAA CAC-39 R: 59-AGC TTC ATG CTC TTT AGC ATG TTT-39 HpGroES C112A F: 59-GCT CAT GAT CAC AAA AAA CAC TAA-39 R: 59-GCA AGC TTC ATG CTC TTT AGC ATG-39 HpGroES C111A/C112A F: 59-GCT GCT CAT GAT CAC AAA AAA CAC TAA-39 R: 59-AGC TTC ATG CTC TTT AGC ATG TTT-39 EcGroES F: 59-GGA TCC ATG AAT ATT CGT CCA TTG CAT-39 R: 59-GGT ACC TTA CGC TTC AAC AAT TGC CAG-39 The mutation sites are underlined. F, forward; R, reverse. The Journal of Immunology 4001

by ∼50% at the higher concentration. Likewise, the IL-8 release induced by incubation for 16 h with 1 mg/ml of LPS (a TLR4 ligand) was signiﬁcantly reduced by pretreatment for 1 h with 5 mM CLI-095 or 30 mg/ml OxPAPC. To further clarify whether HpGroES induced IL-8 production through TLR2 or TLR4, control KATO-III cells and cells pretreated for 1 h with 10 mg/ml anti-TLR2 or anti-TLR4 Ab were incubated for 16 h with 0.05 mM HpGroES. As shown in Fig. 4B, HpGroES-induced IL-8 release was markedly reduced by treatment with anti-TLR4 Ab, but not anti-TLR2 Ab. The IL-8 release caused by incubation for 16 h with 1 mg/ml LPS or with 10 mg/ml TLR2 agonist PGN was markedly decreased by pretreatment with, respectively, anti-TLR4 Ab or anti-TLR2 Ab (Fig. 4B). These ﬁndings suggest that, in KATO-III cells, HpGroES induces IL-8 production via TLR4, but not TLR2. We also examined whether HpGroES activated MAPKs through TLR4 signaling. As shown in Fig. 4C, the HpGroES- induced phosphorylation of p38, ERK, and JNK in KATO-III cells was completely inhibited by preincubation for 1 h with CLI-095, showing that MAPK phosphorylation induced by Downloaded from HpGroES was dependent on TLR4. To further examine the role of TLR4 in HpGroES-induced IL-8 production, primary gastric epithelial cells obtained from WT and TLR4 2/2 mice were used. Cells from WT and TLR42/2 mice were incubated for 24 h with 0.05 mM HpGroES, domain A, or

EcGroES, or 1 mg/ml LPS, and then levels of KC in the culture http://www.jimmunol.org/ supernatants were measured by ELISA. As shown in Fig. 4D, HpGroES and LPS induced KC production by primary gastric epithelial cells from WT mice, whereas domain A or EcGroES did not, and no HpGroES- and LPS-induced KC production by primary gastric epithelial cells was seen in TLR42/2 mice. Similarly, FIGURE 2. Identification of the HpGroES-activated signaling path- HpGroES induced KC production by peritoneal macrophages ways required for IL-8 production. (A) KATO-III cells were incubated from WT mice, but not in TLR42/2 mice (data not shown). These with 0.05 mM HpGroES for the indicated time and then levels of total results show that HpGroES-induced IL-8 production is completely by guest on September 26, 2021 and phosphorylated p38, ERK, and JNK in the cell lysates were detected dependent on TLR4-mediated signaling. using the indicated Abs. (B) KATO-III cells were left untreated or were pretreated with SB203580 (5 mM), U0126 (2.5 mM), or SP600125 (1 HpGroES, but not domain A alone, binds to TLR4 mM) for 1 h prior to incubation for 1 h with PBS or 0.05 mM HpGroES in To ascertain whether HpGroES physically interacts with TLR4, an the presence of the inhibitors, and then levels of p38, ERK, and JNK in the cell lysates were measured as described above. In (A)and(B), the immunoprecipitation study was performed. KATO-III cell lysates values are expressed as the fold increase relative to the level in untreated were incubated for 16 h with 0.05 mM HpGroES or domain A cells and are indicated below the blot. The blots shown are representative immobilized on Ni-NTA beads. As shown in Fig. 5A, TLR4 was of those obtained in three separate experiments. (C)KATO-IIIcellswere immunoprecipitated by immobilized HpGroES, but not domain A. incubated for 1 h with or without inhibitors (5 mM SB203580, 2.5 mM We also used HEK293 cells stably overexpressing HA-tagged U0126, 1 mM SP600125) and then 0.05 mM HpGroES was added for TLR4 to examine the HpGroES–TLR4 interaction by testing a further 16 h and IL-8 levels in the culture supernatants were measured whether the TLR4-HA in cell lysates captured on anti-HA beads 6 by ELISA. The data are mean SD for the results of five experiments. was able to immunoprecipitate HpGroES or domain A and found **p , 0.01. that TLR4 was able to immunoprecipitate HpGroES, but not domain A (Fig. 5B). Similar results were obtained when the nuclear and cytosol loading controls, respectively. These findings experiments were performed in a reverse way that TLR4 could suggest that HpGroES markedly activates NF-kB signaling by immunoprecipitate HpGroES from H. pylori lysates (Fig. 5C). inducing IkB degradation and NF-kB nuclear translocation, The interaction between TLR4 and HpGroES was also exam- which, in turn, induces target gene transcription. ined by incubating HpGroES or domain A with KATO-III cells and determining the localization of HpGroES or domain A and HpGroES induces IL-8 secretion through a TLR4-mediated TLR4 by confocal microscopy. As shown in Fig. 5D, using mechanism HpGroES Abs that recognize both the HpGroES and domain A, Various bacterial proteins induce proinflammatory cytokine pro- HpGroES was shown to colocalize with TLR4 on the cell surface, duction through TLR2- or TLR4-dependent signals (43, 44). To whereas no binding of domain A was seen. Similarly, in HEK293/ determine whether HpGroES-induced IL-8 release was TLR- TLR4-HA cells, HpGroES, but not domain A, bound to, and colo- dependent, KATO-III cells were left untreated or were pre- calized with, TLR4 on the cell surface, confirming the physical treated for 1 h with 2.5 or 5 mM CLI-095 (a TLR4 inhibitor) or 15 interaction of HpGroES with TLR4 (Fig. 5E). or 30 mg/ml OxPAPC (TLR2 and TLR4 inhibitor) and then were incubated for 16 h with 0.05 mM HpGroES in the presence of the The two disulfide bonds in domain B are required for inhibitor. As shown in Fig. 4A, HpGroES-induced IL-8 secretion HpGroES-induced IL-8 release and binding to TLR4 was significantly inhibited by CLI-095 (∼60% at 2.5 mM and 90% As shown above, HpGroES induced IL-8 secretion through at 5 mM), but it was not inhibited by 15 mg/ml OxPAPC and only MAPKs and NF-kB activation by binding to TLR4, whereas 4002 C TERMINI CYSTEINES NEEDED FOR HpGroES-INDUCED IL-8 VIA TLR4

To further address the importance of disulfide bonds, the single mutants C111A and C112A and the double mutant C111A/C112A were constructed by site-directed mutagenesis and different mutant proteins were purified to homogeneity (Fig. 6A, Table I). As shown in Fig. 6C, mutants C111A and C112A, each retaining a different single disulfide bond, were both as effective as HpGroES in inducing IL-8 production, whereas mutant C111A/C112A, containing no disulfide bonds, had a much weaker effect. When KATO-III cell lysates were incubated with HpGroES or mutant C111A, C112A, or C111A/C112A immobilized on Ni-NTA beads FIGURE 3. HpGroES promotes NF-kB activation. (A) KATO-III cells were treated with 0.05 mM HpGroES for the indicated time and or beads alone, HpGroES and the mutants C111A and C112A, but then lysates were prepared and phosphorylation/degradation of IkB not mutant C111A/C112A, were able to immunoprecipitate TLR4 were measured by immunoblotting. The values are expressed as the (Fig. 6D). Similarly, Fig. 6E shows that HpGroES, C111A, and fold increase relative to the level in untreated cells and are indicated C112A were colocalized with TLR4 on the cell surface, whereas below the blot. (B) KATO-III cells were incubated for 1 h with 0.05 no binding of C111A/C112A was seen. The inability of the mM HpGroES and then cytoplasmic and nuclear fractions were ana- C111A/C112A mutant, lacking both disulfide bridges, to bind lyzed by immunoblotting using anti–NF-kB Ab. Lamin B and tubulin TLR4, as well as the ability of mutants C111A or C112A, lacking were used as nucleus and cytosol controls, respectively. The blots only one bridge, to bind TLR4, suggested that the HpGroES– shown in (A)and(B) are representative of those obtained in three TLR4 interaction was structure-dependent. CD spectroscopy was Downloaded from separate experiments. then used to further investigate structural differences between HpGroES and mutants C111A, C112A, and C111A/C112A. As domain A did not bind to TLR4 and was thus unable to activate shown in Fig. 6F, the far-UV CD spectra of mutants C111A and MAPKs and NF-kB and IL-8 production. This was not due to C112A, with two minima peaks at 208 and 223 nm, were similar a difference in structure of domain A and HpGroES, as gel fil- to those of HpGroES, whereas a large decrease in both was seen tration chromatography showed both eluted at a volume corre- with mutant C111A/C112A, indicating that the loss of the two http://www.jimmunol.org/ sponding to a molecular mass of ∼90 kDa, indicating that the disulfide bonds led to the disruption of the conformation of do- oligomeric states of domain A and WT were heptamers (data not main B. Analysis of these far-UV CD spectra using CDPro soft- shown). ware showed that the random coil content of mutant C111A/ To examine the structural requirements for domain B to induce C112A was .10% higher than that in HpGroES (data not IL-8 release, we first constructed two truncated HpGroES mutants shown). These findings indicate that the maintenance of confor- containing domain A, but each lacking a part of the C-terminal mation by the two disulfide bonds in domain B is crucial for region (Table I); their sequences and the purified proteins are HpGroES to induce IL-8 production via TLR4. shown in Fig. 6A. As shown in Fig. 6B, similar to domain A, the truncated mutants ΔH100 and ΔK106 were unable to induce IL-8 Discussion by guest on September 26, 2021 secretion. Because both of these mutants lacked the two disulfide H. pylori infection causes chronic inflammation and the produc- bonds C94–C111 and C95–C112, this suggested that a disulfide- tion of proinflammatory cytokines. Numerous studies have shown constrained conformation of domain B might be crucial for that H. pylori infection increases IL-8 production by gastric epi- HpGroES-mediated IL-8 induction. thelial cells (17, 18, 45). In this study, we investigated the sig-

FIGURE 4. HpGroES-induced IL-8 production is TLR4-dependent. (A) KATO-III cells were left untreated or were pretreated for 1 h with CLI-095 (TLR4 inhibitor) or OxPAPC (TLR2 and TLR4 inhibitor) at the indicated concentration before treatment for 16 h with 0.05 mM HpGroES, and then IL-8 levels in the culture supernatant were measured. Cells incubated for 16 h with LPS alone (1 mg/ml) or pretreated for 1 h with CLI-095 (5 mM) or OxPAPC (30 mg/ml) and then for 16 h with LPS were used as controls. (B) KATO-III cells were incubated for 1 h with or without 10 mg/ml anti-TLR2 or anti-TLR4 Ab before treatment for 16 h with 0.05 mM HpGroES, and then IL-8 levels in the culture supernatant were measured. Cells treated for 16 h with PGN (10 mg/ml) alone or after 1 h pretreatment with anti-TLR2 Ab (10 mg/ml) or for 16 h with LPS (1 mg/ml) alone or after 1 h pretreatment with anti-TLR4 Ab (10 mg/ml) were used as controls. The results in (A)and(B) are mean 6 SD. n =5.(C) KATO-III cells were incubated for 1 h with or without 5 mMCLI-095 and then for 1 h after addition of PBS or 0.05 mM HpGroES, and then cell lysates were prepared and levels of total and phosphorylated p38, ERK, and JNK were measured using the indicated Abs. The data shown are representative of those obtained in three independent experiments and the quantiﬁed results (fold increase compared with untreated controls) are indicated below the blots. (D) Primary gastric epithelial cells were isolated from WT (n = 3) and TLR42/2 mice (n = 3), after 24 h, isolated gastric epithelial cells were incubated for 24 h with 0.05 mM HpGroES, domain A, EcGroES, or 1 mg/ml LPS, and then KC in the culture supernatants was measured by ELISA. The data are mean 6 SD for three experiments. **p , 0.01. The Journal of Immunology 4003 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 5. HpGroES, but not domain A, binds to TLR4. (A) KATO-III cell lysates were incubated for 16 h with HpGroES or domain A immobilized on beads or beads alone, and then bound proteins were analyzed using anti-TLR4 Ab (top panel) or anti-HpGroES Ab (middle panel). (B) HEK293/TLR4-HA cell lysates were incubated with HpGroES or domain A and then TLR4-HA was immunoprecipitated with anti-HA Ab and bound proteins were analyzed using anti-HA Ab (top panel) or anti-HpGroES Ab (middle panel). (C) H. pylori lysates were incubated for 16 h with TLR4–TLR4 Ab immobilized on beads or beads alone and then bound proteins were analyzed using anti-TLR4 Ab (top panel) or anti-HpGroES Ab (middle panel). In (A)–(C), 10% of the input sample was subjected to immunoblotting (bottom panels). The quantiﬁed results are shown below the blot as the fold increase compared with beads alone. (D and E) KATO-III cells (D) or HEK293/TLR4-HA cells (E) were incubated for 1 h with PBS or 0.05 mM HpGroES or domain A and then ﬁxed and double-stained with anti-HpGroES Ab (green) and anti-TLR4 Ab (red), DNA was counterstained with DAPI (blue), and the merged images are shown on the right. The data shown are representative of those obtained in three independent experiments. Scale bars, 10 mm. 4004 C TERMINI CYSTEINES NEEDED FOR HpGroES-INDUCED IL-8 VIA TLR4 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 6. The disulfide bonds of domain B are required for HpGroES-induced IL-8 release through TLR4. (A) Top, Domain structure of HpGroES and model of the structure of domain B; disulfide bridges are displayed as sticks. Bottom, Sequences of domain B of HpGroES and the two truncated and three point mutants. The cysteines involved in disulfide bond formation are shown as linked, and mutated residues are underlined. Right, SDS-PAGE of the purified mutant proteins stained with Coomassie blue. (B) KATO-III cells were incubated for 16 h with 0.05 mM HpGroES, domain A, ΔH100, or ΔK106 and then released IL-8 was measured by ELISA (n = 5). **p , 0.01. (C) KATO-III cells were incubated for 16 h with 0.05 mM HpGroES, C111A, C112A, or C111A/C112A and then released IL-8 was measured by ELISA (n = 5). **p , 0.01. (D) KATO-III cell lysates were incubated for 16 h with HpGroES, C111A, C112A, or C111A/C112A immobilized on beads or beads alone. Bound proteins were analyzed using anti-TLR4 Ab (top panel) or anti-HpGroES Ab (middle panel). Ten percent of the input material was also subjected to immunoblotting (bottom panel). The quantified results for immunoprecipitated TLR4 compared with those with beads alone (fold value) are indicated below the blot. (E) KATO-III cells were incubated for 1 h with 0.05 mM HpGroES, C111A, C112A, or C111A/C112A and then fixed and stained with anti-HpGroES Ab (green) and anti-TLR4 Ab (red), DNA was counterstained with DAPI (blue), and the merged images are shown on the right. The data shown are representative of those obtained in three independent experiments. Scale bar, 10 mm. (F) Far-UV CD spectra of HpGroES and the cysteine mutants. The data shown are representative of those obtained in three independent experiments. naling pathway of HpGroES-induced IL-8 secretion in human that domain A was unable to bind to TLR4 and that domain B was gastric epithelial cells and showed that the MAPK and NF-kB required for TLR4 binding and induction of IL-8 production. signaling pathways were involved. Our data indicated that TLR4 Furthermore, we found that the structure of domain B generated was responsible for the HpGroES-induced signaling. We found by two disulfide bonds was crucial for TLR4 binding and IL-8 The Journal of Immunology 4005 secretion. To our knowledge, this is the first report that the and induces translocation of NF-kB to the nucleus, which may disulfide bond structure of HpGroES domain B is important for then lead to increased IL-8 expression. induction of inflammatory cytokine IL-8 production, and that IL-8 production in the gastric mucosa has been shown to be these inflammatory effects are mediated through the TLR4- induced by H. pylori infection and to correlate with the develop- dependent signaling pathway. ment of chronic gastritis (16). IL-8 production is also recognized The innate immune system is the first line of the host’s defense as an early response to H. pylori by the gastric epithelium (14). A against invading microbial pathogens. TLRs, pattern recognition number of H. pylori factors are known to cause gastric inflam- receptors that recognize pathogenic microbe-associated molecular mation and epithelial damage. CagA induces IL-8 production patterns, have been identified as an important component of the by CagA-positive H. pylori. The Cag type IV secretion system innate immune response to microbial pathogens in humans (46). translocates CagA into gastric epithelial cells and then tyrosine Activation of TLRs results in stimulation of intracellular signaling phosphorylation of CagA occurs, with subsequent induction of pathways leading to induction of inflammatory cytokine produc- IL-8 production (3). Vacuolating cytotoxin A increases IL-8 tion (47). Several reports have suggested that TLRs (TLR2, TLR4, production by U937 human macrophage cells by activation of and TLR5) are involved in H. pylori–induced signal transduction the p38-MAPK pathway (21). Urease induces IL-8 production in (46, 48, 49). Because chronic infection with H. pylori increases gastric epithelial cells (57). Outer inflammatory protein OipA is TLR expression in gastric epithelial cells (50), it was possible that known to enhance IL-8 production through activation of NF-kB, recognition of HpGroES by TLRs was involved in the initiation of AP-1, and CREB (3, 20). However, several studies reported that an innate immune response against this bacterium. It has been isogenic mutant strains lacking CagA, VacA, urease B, or OipA demonstrated that TLR4, TLR5, and TLR9 expressed on human still stimulate significant amounts of IL-8 production, indicating Downloaded from gastric epithelium are involved in the innate immune response to that other virulence factors are also essential for H. pylori–asso- H. pylori (50). Furthermore, several studies indicated that TLR2 ciated IL-8 production (17, 21, 57–60). Other secreted factor, such (51–53) and TLR4 (44, 49, 54) play a role in H. pylori infection. as peptidyl prolyl cis-trans isomerase, induces IL-6 release by However, a study showed that both TLR2 and TLR5, but not a TLR4-dependent pathway in human macrophages (44). Other TLR4, are involved in H. pylori–induced cytokine production investigators have reported similar findings while working on

(48). This indicates that the role of TLRs in H. pylori infection has strain-specific proteins that are found outside the cagPAI, such as http://www.jimmunol.org/ not been completely resolved. Future studies are needed to eval- dupA and JHP0940, which were shown to be able to induce IL-8 uate this aspect. secretion (61, 62). In our study, we found that HpGroES, a se- In this study, we found that HpGroES-induced IL-8 production creted protein, bound to TLR4 on the surface of gastric epithelial was markedly inhibited by a TLR4 inhibitor (CLI-095) or anti- cells and triggered IL-8 production and activation of the MAPK TLR4 Ab, showing that the TLR4 pathway was responsible for at and NF-kB pathways. least a large part of the HpGroES signaling. IL-8 secretion was HSPs are a family of proteins induced in prokaryotic and only slightly inhibited by anti-TLR2 Ab, and no synergistic effects eukaryotic cells by environmental stress. These proteins function of anti-TLR2 and anti-TLR4 Abs were seen. Additionally, our as chaperones to facilitate folding, unfolding, and translocation results indicated that HpGroES-induced IL-8 production through of intracellular polypeptides (26). Additionally, several bacterial by guest on September 26, 2021 TLR4 in primary gastric epithelial cells and macrophages. In the HSPs have been reported to activate proinflammatory cytokine human stomach, the apical and basolateral pole of gastric epi- production. GroEL and DnaK (HSP70) of E. coli induce IL-6 and thelium expressed TLR4, which gives it the possibility to interact TNF-a production by monocytes and endothelial cells (39), with HpGroES (50). In this study, we found that as a novel TLR4 and GroEL of Chlamydia pneumoniae activates inflammation and ligand, the HpGroES-induced IL-8 production provided a new contributes to coronary disease (63). H. pylori HSP60, also re- insight into H. pylori pathogenicity. ferred to as GroEL, plays roles in attachment to the gastric epi- Activation of TLR4 triggers both the NF-kB and MAPK sig- thelium and in inducing IL-8 secretion through the TLR2 pathway naling pathways (55). Our results demonstrated that p38, ERK, in human gastric epithelial cells and monocytes (43). These and JNK were rapidly phosphorylated in response to HpGroES studies indicate that bacterial HSPs are closely associated with stimulation. Phosphorylated p38, ERK, and JNK activate AP-1, inflammatory responses to bacterial infection. However, in the which then binds to the AP-1 binding site on the IL-8 promoter GroES family, HpGroES, but no other bacterial GroES (39), has and induces IL-8 expression. Our present results showed that the been reported to modulate inflammatory responses. The inability HpGroES-induced production of IL-8 was significantly reduced of EcGroES to induce cytokine production might be due to dif- by a p38 inhibitor (SB203580), an ERK inhibitor (U0126), or ferences in the structure of GroESs. HpGroES possesses a unique a JNK inhibitor (SP600125), indicating that the p38, ERK, and domain B at its C terminus that is absent from other known GroES JNK MAPK pathways are all involved in the HpGroES-induced homologs, including those of eukaryotic organisms (25). Its production of IL-8. It has been reported that p38, ERK, and JNK N-terminal domain (domain A) corresponds to the conventional regulate H. pylori–mediated IL-8 secretion from macrophages GroES chaperonin, whereas its C-terminal domain (domain B) is (42). The mechanisms of H. pylori–induced three MAPKs acti- a highly charged extension consisting of 28 aa and conferring vation have also been investigated in several studies, which a unique conformational structure generated by two disulfide suggested that they can be activated independently and simul- bonds, C94–C111 and C95–C112, which form a loop structure taneously (40). Our results suggest that HpGroES can activate between residues 96 and 110, also referred to as the four-cysteine three MAPKs subfamily members coordinately and can lead to IL-8 ring (4CR) motif. The two consecutive disulfide bridges form induction from gastric epithelial cells. Additional studies exam- a rigid scaffold that dramatically affects the topology of the loop ining this regulatory links between MAPKs are needed. Addi- structure of domain B (27). tionally, NF-kB is an important transcription factor that is The 4CR motif consists of disulfide bridges formed between two activated in TLR signaling (56). Previous studies have shown that pairs of two consecutive cysteine residues located in parallel or NF-kB activation controls the expression of genes encoding in- antiparallel peptide chains (64). HpGroES belongs to the parallel flammatory cytokines, such as TNF-a, IL-6, and IL-8 (56). Our 4CR motif family, with only two members in the Protein Data study showed that HpGroES reduces IkB levels in the cytoplasm Bank, the other being conotoxin ε-TxIX, in which the 4CR motif 4006 C TERMINI CYSTEINES NEEDED FOR HpGroES-INDUCED IL-8 VIA TLR4 forms a short loop generating a conformation optimal for receptor 13. Yamada, S., S. Kato, T. Matsuhisa, L. Makonkawkeyoon, M. Yoshida, recognition (65). Additionally, a loop formed by disulfide bonds in T. Chakrabandhu, N. Lertprasertsuk, P. Suttharat, B. Chakrabandhu, S. Nishiumi, et al. 2013. Predominant mucosal IL-8 mRNA expression in non-cagA Thais is chemokines, such as CXCL8, a member of the CXC chemokine risk for gastric cancer. World J. Gastroenterol. 19: 2941–2949. family, also affects their ability to bind to their receptors (66). 14. Eftang, L. L., Y. Esbensen, T. M. Tannæs, I. R. Bukholm, and G. Bukholm. 2012. Interleukin-8 is the single most up-regulated gene in whole genome profiling of Because HpGroES domain B contains four cysteine residues H. pylori exposed gastric epithelial cells. BMC Microbiol. 12: 9. forming two disulfide bonds, the role of the disulfide bonds in 15. Lee, K. E., P. N. Khoi, Y. Xia, J. S. Park, Y. E. Joo, K. K. Kim, S. Y. Choi, and TLR4 binding was investigated. To examine the relationship be- Y. D. Jung. 2013. Helicobacter pylori and interleukin-8 in gastric cancer. World J. Gastroenterol. 19: 8192–8202. tween the loop structure and receptor binding, we disrupted the 16. Naito, Y., M. Ito, T. Watanabe, and H. Suzuki. 2005. Biomarkers in patients with structure of domain B by generating cysteine single- and double- gastric inflammation: a systematic review. Digestion 72: 164–180. point mutants. We found that loss of a single S–S bond was tol- 17. Sharma, S. A., M. K. Tummuru, G. G. Miller, and M. J. Blaser. 1995. Interleukin-8 response of gastric epithelial cell lines to Helicobacter pylori erated, whereas when both disulfide bonds were lost, the structural stimulation in vitro. Infect. Immun. 63: 1681–1687. conformation of HpGroES domain B was disrupted and the pro- 18. Sharma, S. A., M. K. Tummuru, M. J. Blaser, and L. D. Kerr. 1998. Activation of tein could no longer bind to TLR4 or induce IL-8 production. The IL-8 gene expression by Helicobacter pylori is regulated by transcription factor nuclear factor-kB in gastric epithelial cells. J. Immunol. 160: 2401–2407. results indicate that the disulfide bonds in domain B help form 19. Ritter, B., P. Kilian, M. R. Reboll, K. Resch, J. K. DiStefano, R. Frank, W. Beil, a C-terminal loop structure that is important for HpGroES to bind and M. Nourbakhsh. 2011. Differential effects of multiplicity of infection on to TLR4 on gastric epithelial cells. Helicobacter pylori-induced signaling pathways and interleukin-8 gene transcription. J. Clin. Immunol. 31: 60–68. In conclusion, our results show that HpGroES is a potent TLR4 20. Yamaoka, Y., S. Kikuchi, H. M. el-Zimaity, O. Gutierrez, M. S. Osato, and activator and induces an innate immune response, namely TLR4- D. Y. Graham. 2002. Importance of Helicobacter pylori oipA in clinical pre- induced secretion of proinflammatory cytokine IL-8. HpGroES- sentation, gastric inflammation, and mucosal interleukin 8 production. Gastro- Downloaded from enterology 123: 414–424. induced IL-8 release is mediated by the MAPK and NF-kB sig- 21. Hisatsune, J., M. Nakayama, H. Isomoto, H. Kurazono, N. Mukaida, naling pathways. The disulfide bond structure of domain B is A. K. Mukhopadhyay, T. Azuma, Y. Yamaoka, J. Sap, E. Yamasaki, et al. 2008. crucial for HpGroES to interact with TLR4 and induce IL-8 re- Molecular characterization of Helicobacter pylori VacA induction of IL-8 in U937 cells reveals a prominent role for p38MAPK in activating transcription lease by gastric epithelial cells. Given the relationship between, factor-2, cAMP response element binding protein, and NF-kB activation. J. on the one hand, H. pylori infection and TLR signaling and, Immunol. 180: 5017–5027. 22. Harris, P. R., H. L. Mobley, G. I. Perez-Perez, M. J. Blaser, and P. D. Smith. on the other hand, inflammation and gastrointestinal diseases, this http://www.jimmunol.org/ 1996. Helicobacter pylori urease is a potent stimulus of mononuclear phagocyte study provides additional pathophysiologic insights into H. pylori activation and inflammatory cytokine production. Gastroenterology 111: 419– pathogenicity and reveals, to our knowledge for the first time, 425. a novel characteristic unique to HpGroES. 23. Brandt, S., T. Kwok, R. Hartig, W. Ko¨nig, and S. Backert. 2005. NF-kB activation and potentiation of proinflammatory responses by the Helicobacter pylori CagA protein. Proc. Natl. Acad. Sci. USA 102: 9300–9305. Acknowledgments 24. Lin, Y. F., M. S. Wu, C. C. Chang, S. W. Lin, J. T. Lin, Y. J. Sun, D. S. Chen, and L. P. Chow. 2006. Comparative immunoproteomics of identification and char- We thank Dr. Konan Peck (Institute of Biomedical Sciences, Academia acterization of virulence factors from Helicobacter pylori related to gastric Sinica, Taipei, Taiwan) for providing HEK293/TLR4-HA cells. cancer. Mol. Cell. Proteomics 5: 1484–1496. 25. Cun, S., H. Li, R. Ge, M. C. Lin, and H. Sun. 2008. A histidine-rich and cysteine-

rich metal-binding domain at the C terminus of heat shock protein A from by guest on September 26, 2021 Disclosures Helicobacter pylori: implication for nickel homeostasis and bismuth suscepti- The authors have no financial conflicts of interest. bility. J. Biol. Chem. 283: 15142–15151. 26. Tang, Y. C., H. C. Chang, A. Roeben, D. Wischnewski, N. Wischnewski, M. J. Kerner, F. U. Hartl, and M. Hayer-Hartl. 2006. Structural features of the GroEL-GroES nano-cage required for rapid folding of encapsulated protein. Cell References 125: 903–914. 1. Covacci, A., J. L. Telford, G. Del Giudice, J. Parsonnet, and R. Rappuoli. 1999. 27. Loguercio, S., C. Dian, A. Flagiello, A. Scannella, P. Pucci, L. Terradot, and Helicobacter pylori virulence and genetic geography. Science 284: 1328–1333. A. Zagari. 2008. In HspA from Helicobacter pylori vicinal disulfide bridges are 2. Polk, D. B., and R. M. Peek, Jr. 2010. Helicobacter pylori: gastric cancer and a key determinant of domain B structure. FEBS Lett. 582: 3537–3541. beyond. Nat. Rev. Cancer 10: 403–414. 28. Rowinska-Zyrek, M., D. Witkowska, D. Valensin, W. Kamysz, and H. Kozlowski. 3. Yamaoka, Y. 2010. Mechanisms of disease: Helicobacter pylori virulence fac- 2010. The C terminus of HspA—a potential target for native Ni(II) and Bi(III) tors. Nat. Rev. Gastroenterol. Hepatol. 7: 629–641. anti-ulcer drugs. Dalton Trans. 39: 5814–5826. 4. Peek, R. M., Jr., C. Fiske, and K. T. Wilson. 2010. Role of innate immunity in 29. Kansau, I., F. Guillain, J. M. Thiberge, and A. Labigne. 1996. Nickel binding and Helicobacter pylori-induced gastric malignancy. Physiol. Rev. 90: 831–858. immunological properties of the C-terminal domain of the Helicobacter pylori 5. Tsai, H. F., and P. N. Hsu. 2010. Interplay between Helicobacter pylori and GroES homologue (HspA). Mol. Microbiol. 22: 1013–1023. immune cells in immune pathogenesis of gastric inflammation and mucosal 30. Vanet, A., and A. Labigne. 1998. Evidence for specific secretion rather than pathology. Cell. Mol. Immunol. 7: 255–259. autolysis in the release of some Helicobacter pylori proteins. Infect. Immun. 66: 6. Andres, S., H. M. Schmidt, H. Mitchell, M. Rhen, M. Maeurer, and L. Engstrand. 1023–1027. 2011. Helicobacter pylori defines local immune response through interaction 31. Pe´rez-Pe´rez, G. I., J. M. Thiberge, A. Labigne, and M. J. Blaser. 1996. Rela- with dendritic cells. FEMS Immunol. Med. Microbiol. 61: 168–178. tionship of immune response to heat-shock protein A and characteristics of 7. Ibraghimov, A., and J. Pappo. 2000. The immune response against Helicobacter Helicobacter pylori-infected patients. J. Infect. Dis. 174: 1046–1050. pylori—a direct linkage to the development of gastroduodenal disease. Microbes 32. Todoroki, I., T. Joh, K. Watanabe, M. Miyashita, K. Seno, T. Nomura, H. Ohara, Infect. 2: 1073–1077. Y. Yokoyama, K. Tochikubo, and M. Itoh. 2000. Suppressive effects of DNA 8. Lee, J. Y., N. Kim, Y. J. Choi, R. H. Nam, Y. J. Choi, Y. H. Kwon, K. Yoon, vaccines encoding heat shock protein on Helicobacter pylori-induced gastritis in J. H. Suh, S. M. Lee, H. S. Lee, and D. H. Lee. 2014. Histologic findings and mice. Biochem. Biophys. Res. Commun. 277: 159–163. inflammatory reactions after long-term colonization of Helicobacter felis in 33. Wang, L., X. F. Liu, S. Yun, X. P. Yuan, X. H. Mao, C. Wu, W. J. Zhang, C57BL/6 mice. J Cancer Prev 19: 224–230. K. Y. Liu, G. Guo, D. S. Lu, et al. 2010. Protection against Helicobacter pylori 9. Russo, F., E. Jirillo, C. Clemente, C. Messa, M. Chiloiro, G. Riezzo, L. Amati, infection by a trivalent fusion vaccine based on a fragment of urease B-UreB414. L. Caradonna, and A. Di Leo. 2001. Circulating cytokines and gastrin levels in J. Microbiol. 48: 223–228. asymptomatic subjects infected by Helicobacter pylori (H. pylori). Immuno- 34. Moss, S. F., S. Legon, J. Davies, and J. Calam. 1994. Cytokine gene expression pharmacol. Immunotoxicol. 23: 13–24. in Helicobacter pylori associated antral gastritis. Gut 35: 1567–1570. 10. Lindholm, C., M. Quiding-Ja¨rbrink, H. Lo¨nroth, A. Hamlet, and A. M. Svennerholm. 35. Kuan, Y. C., W. T. Lee, C. L. Hung, C. Yang, and F. Sheu. 2012. Investigating the 1998. Local cytokine response in Helicobacter pylori-infected subjects. Infect. function of a novel protein from Anoectochilus formosanus which induced Immun. 66: 5964–5971. macrophage differentiation through TLR4-mediated NF-kB activation. Int. 11. Yamaoka, Y., M. Kita, T. Kodama, N. Sawai, K. Kashima, and J. Imanishi. 1995. Immunopharmacol. 14: 114–120. Expression of cytokine mRNA in gastric mucosa with Helicobacter pylori in- 36. Wang, Y. C., C. L. Chen, B. S. Sheu, Y. J. Yang, P. C. Tseng, C. Y. Hsieh, and fection. Scand. J. Gastroenterol. 30: 1153–1159. C. F. Lin. 2014. Helicobacter pylori infection activates Src homology-2 domain- 12. Shigematsu, Y., T. Niwa, E. Rehnberg, T. Toyoda, S. Yoshida, A. Mori, containing phosphatase 2 to suppress IFN-g signaling. J. Immunol. 193: 4149– M. Wakabayashi, Y. Iwakura, M. Ichinose, Y. J. Kim, and T. Ushijima. 2013. 4158. Interleukin-1b induced by Helicobacter pylori infection enhances mouse gastric 37. Bernhardt, A., D. Kuester, A. Roessner, T. Reinheckel, and S. Krueger. 2010. carcinogenesis. Cancer Lett. 340: 141–147. Cathepsin X-deficient gastric epithelial cells in co-culture with macrophages: The Journal of Immunology 4007

characterization of cytokine response and migration capability after Heli- pylori, Helicobacter felis, and Helicobacter hepaticus bacteria activate innate cobacter pylori infection. J. Biol. Chem. 285: 33691–33700. immunity via Toll-like receptor 2 but not Toll-like receptor 4. Infect. Immun. 72: 38. Sreerama, N., and R. W. Woody. 2000. Estimation of protein secondary structure 6446–6454. from circular dichroism spectra: comparison of CONTIN, SELCON, and 53. Obonyo, M., M. Sabet, S. P. Cole, J. Ebmeyer, S. Uematsu, S. Akira, and CDSSTR methods with an expanded reference set. Anal. Biochem. 287: 252– D. G. Guiney. 2007. Deﬁciencies of myeloid differentiation factor 88, Toll-like 260. receptor 2 (TLR2), or TLR4 produce speciﬁc defects in macrophage cytokine 39. Galdiero, M., G. C. de l’Ero, and A. Marcatili. 1997. Cytokine and adhesion secretion induced by Helicobacter pylori. Infect. Immun. 75: 2408–2414. molecule expression in human monocytes and endothelial cells stimulated with 54. Chochi, K., T. Ichikura, M. Kinoshita, T. Majima, N. Shinomiya, H. Tsujimoto, bacterial heat shock proteins. Infect. Immun. 65: 699–707. T. Kawabata, H. Sugasawa, S. Ono, S. Seki, and H. Mochizuki. 2008. Heli- 40. Yamada, H., T. Aihara, and S. Okabe. 2001. Mechanism for Helicobacter pylori cobacter pylori augments growth of gastric cancers via the lipopolysaccharide- stimulation of interleukin-8 production in a gastric epithelial cell line (MKN 28): Toll-like receptor 4 pathway whereas its lipopolysaccharide attenuates antitumor roles of mitogen-activated protein kinase and interleukin-1b. Biochem. Phar- activities of human mononuclear cells. Clin. Cancer Res. 14: 2909–2917. macol. 61: 1595–1604. 55. Brown, J., H. Wang, G. N. Hajishengallis, and M. Martin. 2011. TLR-signaling 41. Allison, C. C., T. A. Kufer, E. Kremmer, M. Kaparakis, and R. L. Ferrero. 2009. networks: an integration of adaptor molecules, kinases, and cross-talk. J. Dent. Helicobacter pylori induces MAPK phosphorylation and AP-1 activation via Res. 90: 417–427. a NOD1-dependent mechanism. J. Immunol. 183: 8099–8109. 56. Kawai, T., and S. Akira. 2007. Signaling to NF-kB by Toll-like receptors. Trends 42. Bhattacharyya, A., S. Pathak, S. Datta, S. Chattopadhyay, J. Basu, and Mol. Med. 13: 460–469. M. Kundu. 2002. Mitogen-activated protein kinases and nuclear factor-kB reg- 57. Beswick, E. J., I. V. Pinchuk, K. Minch, G. Suarez, J. C. Sierra, Y. Yamaoka, and ulate Helicobacter pylori-mediated interleukin-8 release from macrophages. V. E. Reyes. 2006. The Helicobacter pylori urease B subunit binds to CD74 on Biochem. J. 368: 121–129. gastric epithelial cells and induces NF-kB activation and interleukin-8 produc- 43. Zhao, Y., K. Yokota, K. Ayada, Y. Yamamoto, T. Okada, L. Shen, and K. Oguma. tion. Infect. Immun. 74: 1148–1155. 2007. Helicobacter pylori heat-shock protein 60 induces interleukin-8 via a Toll- 58. Yamaoka, Y., T. Kudo, H. Lu, A. Casola, A. R. Brasier, and D. Y. Graham. 2004. like receptor (TLR)2 and mitogen-activated protein (MAP) kinase pathway in Role of interferon-stimulated responsive element-like element in interleukin-8 human monocytes. J. Med. Microbiol. 56: 154–164. promoter in Helicobacter pylori infection. Gastroenterology 126: 1030–1043. 44. Pathak, S. K., S. Basu, A. Bhattacharyya, S. Pathak, A. Banerjee, J. Basu, and

59. Nozawa, Y., K. Nishihara, R. M. Peek, M. Nakano, T. Uji, H. Ajioka, Downloaded from M. Kundu. 2006. TLR4-dependent NF-kB activation and mitogen- and stress- N. Matsuura, and H. Miyake. 2002. Identiﬁcation of a signaling cascade for activated protein kinase 1-triggered phosphorylation events are central to Heli- interleukin-8 production by Helicobacter pylori in human gastric epithelial cells. cobacter pylori peptidyl prolyl cis-, trans-isomerase (HP0175)-mediated in- Biochem. Pharmacol. 64: 21–30. duction of IL-6 release from macrophages. J. Immunol. 177: 7950–7958. 60. Tabassam, F. H., D. Y. Graham, and Y. Yamaoka. 2011. Paxillin is a novel 45. Crabtree, J. E., and I. J. Lindley. 1994. Mucosal interleukin-8 and Helicobacter cellular target for converging Helicobacter pylori-induced cellular signaling. pylori-associated gastroduodenal disease. Eur. J. Gastroenterol. Hepatol. 6 (Suppl. 1): S33–S38. Am. J. Physiol. Gastrointest. Liver Physiol. 301: G601–G611. 46. Lu, Q., H. Ding, and W. Li. 2013. Role of Toll-like receptors in microbiota- 61. Rizwan, M., A. Alvi, and N. Ahmed. 2008. Novel protein antigen (JHP940) from the genomic plasticity region of Helicobacter pylori induces tumor necrosis

associated gastrointestinal cancer metastasis. J. Cancer Res. Ther. 9(Suppl.): http://www.jimmunol.org/ S142–S149. factor a and interleukin-8 secretion by human macrophages. J. Bacteriol. 190: 47. Kawai, T., and S. Akira. 2010. The role of pattern-recognition receptors in innate 1146–1151. immunity: update on Toll-like receptors. Nat. Immunol. 11: 373–384. 62. Hussein, N. R., R. H. Argent, C. K. Marx, S. R. Patel, K. Robinson, and 48. Smith, M. F., Jr., A. Mitchell, G. Li, S. Ding, A. M. Fitzmaurice, K. Ryan, J. C. Atherton. 2010. Helicobacter pylori dupA is polymorphic, and its active S. Crowe, and J. B. Goldberg. 2003. Toll-like receptor (TLR) 2 and TLR5, but form induces proinﬂammatory cytokine secretion by mononuclear cells. J. In- not TLR4, are required for Helicobacter pylori-induced NF-kB activation and fect. Dis. 202: 261–269. chemokine expression by epithelial cells. J. Biol. Chem. 278: 32552–32560. 63. Ausiello, C. M., G. Fedele, R. Palazzo, F. Spensieri, A. Ciervo, and A. Cassone. 49. Su, B., P. J. Ceponis, S. Lebel, H. Huynh, and P. M. Sherman. 2003. Heli- 2006. 60-kDa heat shock protein of Chlamydia pneumoniae promotes a T helper cobacter pylori activates Toll-like receptor 4 expression in gastrointestinal epi- type 1 immune response through IL-12/IL-23 production in monocyte-derived thelial cells. Infect. Immun. 71: 3496–3502. dendritic cells. Microbes Infect. 8: 714–720. 50. Schmausser, B., M. Andrulis, S. Endrich, S. K. Lee, C. Josenhans, H. K. Muller-€ 64. Zagari, A. 2009. The four cysteines ring motif in proteins. Biopolymers 91:

Hermelink, and M. Eck. 2004. Expression and subcellular distribution of Toll- 1048–1055. by guest on September 26, 2021 like receptors TLR4, TLR5 and TLR9 on the gastric epithelium in Helicobacter 65. Rigby, A. C., E. Lucas-Meunier, D. E. Kalume, E. Czerwiec, B. Hambe, pylori infection. Clin. Exp. Immunol. 136: 521–526. I. Dahlqvist, P. Fossier, G. Baux, P. Roepstorff, J. D. Baleja, et al. 1999. A 51. Chang, Y. J., M. S. Wu, J. T. Lin, B. S. Sheu, T. Muta, H. Inoue, and C. C. Chen. conotoxin from Conus textile with unusual posttranslational modifications 2004. Induction of cyclooxygenase-2 overexpression in human gastric epithelial reduces presynaptic Ca2+ influx. Proc. Natl. Acad. Sci. USA 96: 5758–5763. cells by Helicobacter pylori involves TLR2/TLR9 and c-Src-dependent nuclear 66. Joseph, P. R., J. M. Sarmiento, A. K. Mishra, S. T. Das, R. P. Garofalo, factor-kB activation. Mol. Pharmacol. 66: 1465–1477. J. Navarro, and K. Rajarathnam. 2010. Probing the role of CXC motif in che- 52. Mandell, L., A. P. Moran, A. Cocchiarella, J. Houghton, N. Taylor, J. G. Fox, mokine CXCL8 for high affinity binding and activation of CXCR1 and CXCR2 T. C. Wang, and E. A. Kurt-Jones. 2004. Intact Gram-negative Helicobacter receptors. J. Biol. Chem. 285: 29262–29269.