Regulation of MyD88-Dependent Signaling Events by S Nitrosylation Retards Toll-Like Receptor Signal Transduction and Initiation of Acute-Phase Immune Responses

Reagents and cell culture. N ^G-Monomethyl-l-arginine monoacetate (l-NMMA), S-nitrosoglutathione (GSNO), glutathione (GSH), N-ethylmaleimide, coumermycin A, N-acetyl l-cysteine (NAC), ascorbic acid, and diphenyleneiodonium (DPI) were obtained from Sigma-Aldrich. SNAP (5-nitroso-N-acetyl-d,l-penicillamine) was purchased from Cayman Chemical. ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one) and KT5823 were obtained from Calbiochem. Preparation of TLR ligands, including highly purified Escherichia coli LPS, Salmonella LPS, Pam₃CSK₄, macrophage-activating lipopeptide 2 (MALP-2), and Salmonella enterica serovar Typhimurium flagellin, was as described previously (18). Recombinant human IL-1β was from R&D Systems. Human aortic endothelial cells (HAECs) and human embryonic kidney 293 (HEK293) cells were maintained as described previously (18). HEK293 cells stably expressing human TLR4, MD2, and CD14 (293-TLR4 cells) and HEK293 cells stably expressing human TLR2 and CD14 (293-TLR2 cells) were obtained from InvivoGen.

Mice.iNOS-deficient (iNOS^−/−) mice and endothelial-NOS (eNOS)-deficient (eNOS^−/−) mice were from The Jackson Laboratories. C57BL/6J control (wild-type) mice were obtained from Japan SLC. All mice were kept under specific pathogen-free conditions. Male mice between 6 and 10 weeks of age were used for all of experiments. All animal protocols were approved by the National Institute for Longevity Sciences Animal Experimentation Committee at the National Center for Geriatrics and Gerontology (Aichi, Japan).

For LPS-induced acute lung injury, anesthetized mice received Escherichia coli LPS dissolved in pyrogen-free phosphate-buffered saline (PBS) containing 1 mg/ml Evans Blue intratracheally immediately after mechanical ventilation. After 30 min of administration, lung was excised and then lysed for immunoblot analysis. The febrile responses in mice treated with E. coli LPS were tested according to a protocol described previously (45, 49). Mice (n = 6) were maintained at a neutral ambient temperature of 31°C and challenged by intraperitoneal (i.p.) injection of 5 mg LPS/kg of body weight dissolved in pyrogen-free PBS. A high dose of LPS (more than 50 mg/kg) was fatal within 90 min in eNOS^−/− mice. A colonic thermocouple was inserted and fixed to the base of the tail with adhesive tape. The change in temperature was monitored at 5-min intervals during a period of 120 min after LPS administration. All of the tests were performed at the temperature of 31°C. After 2 h or 12 h of LPS administration, 2 ml of PBS was injected into the abdominal cavity of each mouse. Then, fluids were collected and centrifuged for assessment of cytokine production by an enzyme-linked immunosorbent assay (ELISA). Preparation of peritoneal macrophages was as described previously (28).

Plasmids.The DNA construct encoding 3× Flag-tagged MyD88 fused to the B subunit of the bacterial DNA gyrase (MyD88-GyrB) was as described previously (11). Plasmids encoding human MyD88 and TIRAP were kind gifts from Margaret K. Offermann (Emory University School of Medicine). The cDNAs of N-terminal Flag-tagged and Myc-tagged MyD88, Myc-tagged TIRAP, and IRAK-1 were amplified by PCR and cloned into the pcDNA3.1 vector (Invitrogen). The construct encoding human TLR2 was as described previously (17). Constructs encoding mutated Flag-MyD88 were obtained using a QuikChange II site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions.

Protein purification.Recombinant Flag-MyD88 proteins were prepared using a FLAG M purification kit (Sigma-Aldrich) from HEK293 cells stably expressing Flag-MyD88 constructs, according to the manufacturer's instructions. Purity of recombinant proteins was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), followed by silver staining and immunoblotting with anti-Flag antibody.

Detection of S-nitrosylated proteins.To detect S-nitrosylated MyD88 from lung lysates from wild-type mice and eNOS^−/− mice, we referred to the protocol described by Jaffrey et al. (21). Several experiments were performed using a NitroGlo nitrosylation detection kit (PerkinElmer) according to the manufacturer's instructions. Lung lysates from wild-type and eNOS^−/− mice were subjected to the biotin switching S-nitrosylation assay, and then biotinylated proteins were purified on streptavidin-agarose. Purified proteins eluted by 2-mercaptoethanol were detected by immunoblotting with anti-MyD88 antibody.

The quantitative measurement of S-nitrosylated recombinant MyD88 by ELISA was performed as follows. Briefly, recombinant Flag-MyD88 (150 μg) was treated with or without SNAP for 30 min at 37°C in the dark. Then, the free sulfides of Flag-MyD88 were blocked with 4 mM methylmethanethionsulfonate for 15 min. After purification by using Micro Bio-Spin chromatography columns (Bio-Rad Laboratories), Flag-MyD88 was reacted with 25 mM ascorbate to be completely denitrosylated. Free sulfides were then labeled with a biotin-conjugated maleimide, using a biotin labeling kit (SH; Dojindo Laboratories) according to the manufacturer's instructions. The diluents of biotinylated Flag-MyD88 proteins dissolved in Tris-buffered saline (pH 7.2) were stabilized in the wells of immobilizer streptavidin plates (Nunc). Flag-MyD88 proteins in the wells were detected by using anti-Flag antibody and a secondary antibody conjugated with horseradish peroxidase. Colorimetric reaction was detected by absorbance on a spectrophotometer at 450 nm. Results were expressed as means ± standard deviations (SD) of three determinations.

Photolysis of S-nitrosylated proteins.Mouse lung lysates were exposed for 3 min to a UV-visible light mercury vapor lamp according to a protocol recently described (8). The samples were then subjected to the biotin switch technique as described above.

Luciferase reporter assay.293-TLR2 cells were transiently transfected with wild-type MyD88-GyrB or MyD88-GyrB mutants, each with a cysteine residue replaced with a serine residue, together with 50 ng of an NF-κB (5×) luciferase reporter plasmid (pNF-κB-Luc; Stratagene) and 5 ng of an internal control luciferase reporter plasmid (pRL-TK; Promega) and incubated for 16 h. At 6 h before the end of incubation, cells were treated with or without 250 μM SNAP. Cells were then stimulated with 100 ng/ml Pam₃CSK₄ for 6 h. HEK293 cells stably expressing MyD88-GyrB were transfected with pNF-κB-Luc and pRL-TK. After 24 h of incubation, cells were stimulated with coumermycin A in the presence and absence of 250 μM SNAP. The dual luciferase activity was measured as described previously (19).

Immunoblot analysis of IRAK-1 and IκBα.HAECs were stimulated with 10 ng/ml of LPS for 0 to 90 min. HEK293 cells stably expressing MyD88-GyrB were stimulated with 1 μM coumermycin for 20 min. Cells were lysed in the presence of protease inhibitor and phosphatase inhibitor cocktails (Roche) at 4°C. Cell lysates or lysates from the mouse lungs were separated by SDS-PAGE, followed by immunoblot analyses using anti-IRAK-1, anti-IκBα, and phosphorylation-specific anti-IκBα (Ser32/Ser36) antibodies (Cell Signaling Technology).

RNA extraction and reverse transcription-PCR.Total RNA was isolated from mouse peritoneal macrophages stimulated with 100 ng/ml LPS and 10 ng/ml gamma interferon, and transcripts were quantified by real-time quantitative reverse transcription-PCR on a LightCycler ST300 system (Roche). All values were normalized to the level of β-actin mRNA. The primer sets used are as follows: for mouse macrophage inflammatory protein 2 (MIP-2), 5′-ATCCAGAGCTTGAGTGTGACGC-3′ (sense) and 5′-AAGGCAAACTTTTTGACCGAA-3′ (antisense); for mouse IL-6, 5′-CCACGGCCTTCCCTAC-3′ (sense) and 5′-AGTGCATCATCGTTGTTC-3′ (antisense); and for mouse β-actin, 5′-AAATCGTGCGTGACATCAAA-3′ (sense) and 5′-AAGGAAGGCTGGAAAAGAGC-3′ (antisense).

Cytokine ELISA.Concentrations of human IL-8, mouse MIP-2, and mouse IL-6 were determined using a Cytoset ELISA kit (Biosource) according to the manufacturer's instructions.

Subcellular fractionation.Subcellular fractionation of HEK293 cells stably expressing Flag-MyD88 and 293-TLR4 cells stably expressing Flag-MyD88-GyrB was performed using a ProteoExtract subcellular proteome extraction kit (Calbiochem) according to the manufacturer's instructions. This kit enables extraction of different subcellular fractions of the cytoplasm, plasma membrane, nuclei, and cytoskeleton from mammalian cells. 293-TLR4 cells stably expressing Flag-MyD88-GyrB were maintained in serum-free Dulbecco's modified Eagle's medium containing 5% PANEXIN H cell growth supplement (PAN Biotech GmbH) to avoid nonspecific cell activation by animal serum components. The whole-cell lysate was obtained using NuPAGE LDS sample buffer containing 2-mercaptoethanol. Each fraction was mixed with NuPAGE LDS sample buffer and boiled for 5 min, followed by SDS-PAGE and immunoblot analyses using anti-Flag (Sigma), anti-MyD88 (Santa Cruz Biotechnology), anti-IRAK-1, and anti-vimentin (BD Biosciences) antibodies.

Blue native PAGE and immunoprecipitation.HEK293 cells stably expressing Flag-MyD88 were treated with SNAP for 1 h, washed twice with PBS, and lysed with HEPES buffer (pH 7.2) containing 1% Triton X-100, 1% Nonidet-P40, and proteinase inhibitor cocktail (Roche). Cell lysates were separated by blue native PAGE according to the protocol provided by Invitrogen and then immunoblotted with anti-Flag antibody. HEK293 cells transiently transfected combinatorially with Flag-MyD88 and Myc-TIRAP or Flag-MyD88 and IRAK-1 were treated with GSNO or GSH for 1 h. HEK293 cells stably expressing 3× Flag-MyD88-GyrB were stimulated with 1 μM coumermycin A for 20 min. Cells were lysed with HEPES buffer (pH 7.2) containing 1% Triton X-100 and proteinase inhibitor cocktail (Roche). Clarified cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblot analysis using anti-Flag, anti-Myc, and anti-TRAF6 (StressGen) antibodies. All experiments were performed at least three times, and representative results are shown.

Immunofluorescent cell staining.HAECs were fixed at −20°C with methanol, and double immunostaining was then carried out with anti-β-actin monoclonal antibody (Santa Cruz Biotechnology) and Alexa 488-conjugated immunoglobulin G secondary antibody (Invitrogen) and then with anti-MyD88 rabbit polyclonal antibody (Santa Cruz Biotechnology) and Alexa 564-conjugated immunoglobulin G secondary antibody (Invitrogen). Cell nuclei were also stained with 2.5 μg/ml of Hoechst 33342 for 30 min.

Statistical analysis.Probability (P) values were calculated by Student's t test and analysis of variance and were considered significant at 0.05 or 0.01.

TLR signaling components are S nitrosylated in vivo.To examine whether TLR signal components are physiologically S nitrosylated in vivo, we detected S-nitrosylated proteins in the lung lysates from wild-type and eNOS^−/− mice by utilizing the biotin switch method (21). Interestingly, this method facilitated the detection of MyD88 as an S-nitrosylated protein (Fig. 1A). To exclude the possibility that the detection of S-nitrosylated MyD88 is an experimental artifact, we utilized the biotin switch method combined with photolysis of S nitrosylation, which has recently been reported as a useful method for confirming the specificity of S nitrosylation (8). The detectable S-nitrosylated MyD88 protein was reduced after exposure of the samples to a UV lamp (Fig. 1B), suggesting that the result is not false positive. Thus, our result at least suggests that MyD88 is potentially S nitrosylated in addition to other signaling molecules, including NF-κBs, ASK1, and caspase-1.

• Open in new tab
• Download powerpoint

FIG. 1.

S nitrosylaton of MyD88. (A) Lung lysates from wild-type (WT) and eNOS^−/− mice and SNAP-treated recombinant human MyD88 (rhMyD88) were subjected to the biotin switching S-nitrosylation assay, and then biotinylated proteins were purified on streptavidin-agarose. Purified proteins were detected by immunoblotting with anti-MyD88 antibody (upper). MyD88 proteins in lung lysates and rhMyD88 were also shown as loading controls (lower). (B) Mouse lung lysates were exposed for 3 min to a UV-visible light mercury vapor lamp. The samples were then subjected to the biotin switch method. (C) A recombinant Flag-MyD88 protein was treated with or without SNAP (100, 200, 500, and 1,000 μM) for 30 min. For the denitrosylation study, SNAP-treated proteins were incubated with 1 mM ascorbic acid, 1 mM GSH, or 1 mM HgCl₂ for 5 min before the blockade of free thiols by methylmethanethionsulfonate. Then, S-nitrosylated residues of MyD88 were switched into biotins and proteins were fixed on streptavidin-coated plates, followed by ELISA with anti-Flag antibody. Each value is the mean ± SD (n = 3). See text for details. (*, P < 0.01 for comparison with the group of 500 μM SNAP). (D) Schematic of human MyD88. (E) Recombinant Flag-MyD88 wild-type proteins and mutants with each cysteine residue replaced with a serine residue were treated with or without 500 μM SNAP for 30 min. Then, S-nitrosylated MyD88 fixed on streptavidin-coated plates was detected by ELISA using anti-Flag antibody. Each value is the mean ± SD (n = 3). See text for details. (*, P < 0.01 for comparison with the wild-type group). (F) Sequence alignment of the region around nine cysteine residues of MyD88. (G) Sequence alignment of human TIR domain-containing molecules. The regions corresponding to that around the residues of Pro200, a critical residue for TIR-TIR interaction, and Cys216 of MyD88 are shown. (H) HEK293 cells stably expressing TLR2 were transiently transfected with GyrB-fused wild-type MyD88 or mutant MyD88 with each cysteine residue replaced with a serine residue together with the NF-κB-driven luciferase gene and incubated for 16 h. At 6 h before the end of incubation, cells were treated with 200 μM SNAP. Cells were stimulated with 100 ng/ml Pam₃CSK₄ for 6 h, and then luciferase activity was measured. Each value is the mean ± SD (n = 3). (*, P < 0.05 for comparison with the wild-type group).

We further examined the details of S-nitrosylated MyD88 in vitro by utilizing a quantitative method for detecting S nitrosylation of recombinant proteins. We could detect S nitrosylation of recombinant MyD88, which increased, accompanied by an increase in the concentration of the NO donor SNAP (Fig. 1C). It has been known that S-nitrosylated proteins are reversibly denitrosylated by antioxidants or oxidoreductases, by which substantial protein functions are restored (14, 42). Indeed, the detectable S-nitrosylated MyD88 protein was reduced when the NO donor-treated protein was reacted with ascorbate, HgCl₂, or GSH (Fig. 1C). We further determined the site of S nitrosylation because the modification is effected toward particular cysteine residues (14). Mammalian MyD88 contains a total of nine cysteine residues: one in a short linker region and the other eight in the TIR domain (Fig. 1D). These residues are thought not to be involved in the formation of intramolecular disulfide bonds. Among vertebrates, all cysteine residues are highly conserved (data not shown). We prepared recombinant MyD88 proteins of nine individual mutants, each with one of the nine cysteine residues replaced with a serine residue. We found that the degrees of S nitrosylation of Cys113 and Cys216 were significantly reduced compared with those of wild-type MyD88 (Fig. 1E). These cysteine residues partially fulfill the predictive site of the S nitrosylation “acid-base motif” that comprises flanking acidic and basic residues (14) (Fig. 1F). Interestingly, the cysteine residue equivalent of Cys216 is conserved even in invertebrates, while others are not (data not shown). Among other TIR domain-containing adaptor molecules, only TIRAP and SIGIRR have a cysteine residue corresponding to the position of Cys216 (Fig. 1G).

To determine the requirement of cysteine residues for functioning of MyD88, we utilized MyD88 fused to the B subunit of the bacterial DNA gyrase (MyD88-GyrB). The Streptomyces-derived bivalent antibiotic coumermycin binds GyrB with a stoichiometry of 1:2, acting as a natural dimerizer of GyrB (7). Although overexpressed MyD88 is known to reveal TLR stimulation-independent nonspecific activation of downstream signaling through self-dimerization (11, 36), MyD88-GyrB does not reveal such nonspecific activation unless cells are exposed to TLR stimulation or coumermycin treatment (11). We prepared GyrB-fused wild-type MyD88 and MyD88 mutants, each with one of the nine cysteine residues replaced with a serine residue, and examined the NF-κB-activating properties in the TLR2 ligand Pam₃CSK₄-stimulated HEK293 cells stably expressing TLR2. None of the cysteine replacement mutants abrogated the NF-κB-activating property of MyD88 (Fig. 1H). However, the Cys216Ser mutant significantly increased the activity compared with that of wild-type MyD88 (Fig. 1H). Additionally, similar results were found in the cells treated with SNAP (Fig. 1H). Thus, it is possible that Cys216 of MyD88 mediates the suppressive effect of NO.

S nitrosylation alters MyD88-mediated signaling events.We next explored how S nitrosylation of signaling components alters TLR signaling events. To examine this in vivo, we utilized an animal model of acute lung injury induced by intratracheal administration of LPS. We investigated degradation of IRAK-1 and IκBα, hallmarks of MyD88-dependent and IKKβ-dependent signaling events, in the lungs 30 min after LPS administration. Interestingly, degradation of IRAK-1 and IκBα was apparently promoted in eNOS^−/− mice compared with that in wild-type mice (Fig. 2A). We also examined whether NO alters degradation of IRAK-1 and IκBα in cultivated vascular endothelial cells. In HAECs, LPS induced degradation of IRAK-1 and IκBα within 30 min after stimulation (Fig. 2B). Degradation of IRAK-1 and IκBα was promoted and occurred within 15 min after stimulation when endogenous NO was predepleted by the l-arginine analog l-NMMA (Fig. 2B). In addition, IRAK-1 degradation was also promoted when confluent HAECs were maintained in culture media without the eNOS activator vascular endothelial growth factor or the phosphatidylinositol 3-kinase inhibitor LY294002 (data not shown). In contrast to these results, degradation was delayed and residual proteins were observed even at 45 min after stimulation when cells were pretreated with SNAP (Fig. 2B). The effect of NO was not altered in the presence of the guanylate cyclase inhibitor ODQ or the cyclic-GMP-dependent protein kinase inhibitor KT5823 (data not shown). Notably, LPS-induced degradation of IRAK-1 and IκBα in HAECs was prevented by the irreversible thiol modification by N-ethylmaleimide (Fig. 2C), implying that the effect of NO on the signaling events depends on modification of cysteine residues.

• Open in new tab
• Download powerpoint

FIG. 2.

NO suppresses LPS-induced degradation of IRAK-1 and IκBα. (A) Anesthetized wild-type (WT) and eNOS^−/− mice intratracheally received E. coli LPS and mechanical ventilation. After 30 min of administration, lung was excised and then lysed for immunoblotting with anti-IRAK-1, anti-IκBα, and anti-GAPDH antibodies. (B) HAECs pretreated with 1 mM l-NMMA for 12 h or 0.25 mM SNAP for 1 h were stimulated with 10 ng/ml E. coli LPS for the indicated periods. The expression levels of IRAK-1 and IκBα were determined by immunoblot analysis. (C) HAECs pretreated with 0.1 mM N-ethylmaleimide for 10 min were stimulated with 10 ng/ml E. coli LPS for the indicated periods. The expression levels of IRAK-1 and IκBα were determined by immunoblot analysis.

To test whether NO alters the MyD88-dependent signal events, we utilized the MyD88-GyrB construct. Under coumermycin treatment of cells stably expressing MyD88-GyrB, MyD88-GyrB undergoes dimerization and mimics TLR-triggered typical MyD88-dependent functions, such as the activation of MAPKs and IKKs and secretion of proinflammatory cytokines (11). We found that coumermycin-dependent induction of NF-κB activation in HEK293 cells was suppressed by pretreatment with cells with NO donors (Fig. 3A). Coumermycin could induce phosphorylation of IκBα at Ser32 and Ser36, the target residues of IKKβ involved in degradation of IκBα (51), but SNAP pretreatment could suppress induction of the response (Fig. 3B). SNAP also suppressed coumermycin-induced phosphorylation of MAPKs (data not shown). The coumermycin-dependent dimerization of MyD88-GyrB induced interaction with TRAF6, consistent with a previous study (10), and we found that this interaction was reduced by SNAP treatment (Fig. 3C).

• Open in new tab
• Download powerpoint

FIG. 3.

Effects of NO on TLR-mediated signaling events. (A) HEK293 cells stably expressing MyD88-GyrB were transiently transfected with an NF-κB-driven luciferase gene and incubated for 24 h. Cells were pretreated with or without 0.25 mM SNAP for 1 h and then treated with 1 μM coumermycin for 3 h. Then, luciferase activity was measured. Each value is the mean ± SD (n = 3). (B) HEK293 cells stably expressing MyD88-GyrB were pretreated with or without 0.25 mM SNAP for 1 h and then treated with 1 μM coumermycin for the indicated periods. The phosphorylation of IκBα at Ser32/Ser36 was detected by immunoblot analysis. (C) HEK293 cells stably expressing MyD88-GyrB were pretreated with or without 0.25 mM SNAP for 1 h and then treated with 1 μM coumermycin for 20 min. Then, cell lysates were immunoprecipitated (IP) with anti-Flag antibody, followed by immunoblotting with anti-Flag and anti-TRAF6 antibodies. (D) HEK293 cells transiently expressing Flag-tagged wild-type (WT) or Cys residue (113 or 216) replacement MyD88 together with IRAK-1 were treated with 500 μM SNAP for 1 h. Then, cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with anti-Flag and anti-IRAK-1 antibodies. (E) 293-TLR4/MD2-CD14 cells stably expressing Flag-MyD88-GyrB were treated with or without 500 μM SNAP for 1 h and then stimulated with 100 ng/ml LPS for 20 min. Then, cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with anti-IRAK-1 and anti-Flag antibodies. (F) HEK293 cells transiently expressing Flag-tagged MyD88 together with Myc-tagged TIRAP were treated with GSH or GSNO (0, 100, and 500 μM) for 1 h. Then, cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblotting (IB) with anti-Flag and anti-Myc antibodies. (G) HEK293 cells transiently expressing Flag-tagged wild-type or Cys residue (113 or 216) replacement MyD88 together with Myc-tagged TIRAP were treated with 500 μM SNAP for 1 h. Then, cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with anti-Flag and anti-IRAK-1 antibodies.

We found that NO and alteration of Cys113 and Cys216 residues of MyD88 did not alter the interaction of overexpressed MyD88 with IRAK-1 in HEK293 cells (Fig. 3D). However, NO clearly attenuated TLR4 stimulus (LPS)-dependent induction of MyD88-IRAK-1 interaction (Fig. 3E), suggesting that NO targets upstream signaling events of IRAK-1. It has been known that the recruitment of MyD88 to TLR2 or TLR4 is mediated by binding of the sorting adaptor TIRAP to the membrane phosphatidylinositol 4,5-bisphosphate, followed by interaction of MyD88 with TIRAP through TIR-TIR interaction (23). We examined whether NO affects the interaction of MyD88 with TIRAP. Overexpressed MyD88 interacted with TIRAP in HEK 293 cells (Fig. 3F). We found that treatment of the cells with the NO donor GSNO attenuated the interaction (Fig. 3F). We further investigated whether S nitrosylation of residues 113 and 216 is involved in MyD88-TIRAP interaction. Alteration of Cys residues did not affect the interaction (Fig. 3G). However, SNAP-induced attenuation of the interaction was reduced in the Cys216 mutant (Fig. 3G).

In HAECs, MyD88 was enriched with filamentous cytoskeletal structures and partly colocalized with β-actin (Fig. 4A). Additionally, a large part of MyD88 stably expressed in HEK293 cells was found in the cytoskeletal fraction (Fig. 4B). These findings are consistent with results of a previous study showing that MyD88 associates with β-actin in HeLa cells (22). Cytoskeletal MyD88 was separated from IRAK-1, which was found only in the cytoplasm (Fig. 4B), suggesting that MyD88 is maintained as an inactive state in cytoskeleton. We found that SNAP treatment altered such cytoskeletal localization of MyD88 into the cytoplasm (Fig. 4C). We further investigated the subcellular localization of MyD88 after TLR4 stimulation in 293-TLR4/MD2-CD14 cells stably expressing Flag-MyD88-GyrB. After LPS treatment, a part of MyD88 was transported to the cytoplasmic membrane from the cytoskeleton (Fig. 4D). However, SNAP treatment retarded such LPS-induced transportation of MyD88 (Fig. 4D). Although native PAGE analysis revealed that MyD88 formed a protein complex (more than 480 kDa) in HEK293 cells, SNAP treatment resulted in reduction in the size of the complex to approximately 450 kDa or 250 kDa, accompanied by an increase in concentration (Fig. 4E). Moreover, higher concentrations of SNAP further altered the complex to render a monomer (approximately 35 kDa) (Fig. 4E), implying disruption of functional MyD88 protein complex by NO.

• Open in new tab
• Download powerpoint

FIG. 4.

Reverse of the suppressive effect of NO by GSH. (A) HAECs were fixed and stained immunofluorescently (IF) with anti-β-actin antibody (green, left), anti-MyD88 antibody (red, middle), and Höechst33342 (blue, right). (B) Parental HEK293 cells and HEK293 cells stably expressing Flag-MyD88 were fractionated into the cytoplasm (Cp), cytoplasmic membrane (M), nucleus (N), and cytoskeleton (Cs). Whole-cell lysates (W) were also obtained. The fractions were assessed by immunoblotting (IB) with anti-Flag, anti-MyD88, anti-IRAK-1, and anti-vimentin antibodies. (C) HEK293 cells stably expressing Flag-MyD88 were treated with the indicated concentration of SNAP for 1 h and fractionated into each fraction. The fractions were assessed by immunoblotting with anti-Flag antibody. (D) 293-TLR4/MD2-CD14 cells stably expressing Flag-MyD88-GyrB were treated with or without 500μM SNAP for 1 h and then stimulated with 100 ng/ml LPS for 20 min. Cells were then fractionated into each fraction. The fractions were assessed by immunoblotting with anti-Flag antibody. (E) HEK293 cells stably expressing Flag-MyD88 were treated with or without SNAP (10, 50, 125, 250, and 500 μM) for 1 h. Then, cell lysates were assessed by blue native PAGE and immunoblotting with anti-Flag antibody. (F) HEK293 cells transiently expressing Flag-MyD88 and Myc-tagged TIRAP were treated with 500 μM GSNO for 1 h and then with or without 500 μM GSH for 15 min. Then, cell lysates were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with anti-Flag and anti-Myc antibodies. (G) HEK293 cells stably expressing Flag-MyD88 were treated with the 500 μM GSNO for 1 h and then with or without 500 μM GSH for 15 min. The cells were fractionated into each fraction, followed by immunoblotting with anti-Flag antibody.

Thus, NO has a capability to obstruct the MyD88 signing pathway through disruption of the multiple steps of protein interactions.

NO reversibly suppresses the MyD88 signaling events.S-nitrosylated proteins are known to undergo denitrosylation, by which regulatory effects of NO are conferred to control protein functions. We therefore investigated how S nitrosylation and denitrosylation affect MyD88-mediated signaling events. For this purpose, we utilized GSH because GSH had a capability to denitrosylate MyD88 (Fig. 1C). We found that GSH restored NO-induced impaired interaction of MyD88 with TIRAP (Fig. 4F). Furthermore, the NO-induced cytoplasmic localization of cytoskeletal MyD88 was restored by treatment of the cells with GSH (Fig. 4G). Thus, these results suggest that S nitrosylation alters the MyD88 pathway, and antioxidants or oxidoreductases restore such NO-derived actions, probably through denitrosylation.

NO reversibly suppresses TLR-mediated cellular responses.HAECs responded to multiple bacterial TLR agonistic molecules, Pam₃CSK₄ (for TLR1/TLR2), MALP-2 (for TLR2/TLR6), LPS (for TLR4), flagellin (for TLR5), and IL-1β, all of which are known to activate MyD88-dependnet signaling to induce production of the NF-κB-driven chemokine IL-8 after stimulation for 3 h (Fig. 5A). Predepletion of endogenous NO by l-NMMA resulted in a significant increase in IL-8 production induced by each stimulator (Fig. 5A), indicating that endogenous NO has a suppressive effect on the TLR-mediated cellular response. Furthermore, IL-8 production by each stimulator was suppressed in the presence of SNAP (Fig. 5A). The effect of NO donors was not altered in the presence of ODQ or KT5823 (data not shown). We investigated whether such a suppressive effect of NO can be restored because S nitrosylation is a reversible protein modification. However, it is difficult to examine the effects of antioxidants or oxidoreductases because the TLR signaling pathway is greatly affected by reactive oxygen species generated from NADPH oxidases (27, 34, 48). Indeed, treatment of cells with ascorbic acid, GSH, NAC, or the NADPH oxidase inhibitor DPI greatly impaired LPS-induced IL-8 production in HAECs (Fig. 5B). We therefore attempted to address whether the effect of NO is transient or persistent. For this purpose, HAECs were pretreated with SNAP for 1 h and washed two times to remove the NO donor. Then, at various times afterwards, the cells were stimulated with LPS and IL-8 production was measured. NO suppression of IL-8 production was gradually neutralized or restored in a time-dependent manner, although the restrictive effect continues for several hours (Fig. 5C).

• Open in new tab
• Download powerpoint

FIG. 5.

NO reversibly suppresses TLR-mediated cellular responses. (A) HAECs pretreated with 1 mM l-NMMA for 12 h or 0.25 mM SNAP for 1 h were stimulated with 1 μg/ml Pam₃CSK₄, 100 nM MALP-2, 10 ng/ml E. coli LPS, 10 ng/ml Salmonella LPS, 5 μg/ml flagellin, and 10 ng/ml IL-1β for 3 h. Then, production of IL-8 was determined by ELISA. Each value is the mean ± SD (n = 3). *, P < 0.01 for comparison with the vehicle group. (B) HAECs were stimulated with 10 ng/ml E. coli LPS for 3 h in the presence or absence of 1 mM ascorbic acid (AA), 1 mM GSH, 1 mM NAC, or 20 μM DPI. Then production of IL-8 was determined by ELISA. Each value is the mean ± SD (n = 3). (*, P < 0.01 for comparison with the vehicle group). (C) HAECs were treated with 0.25 mM SNAP for 1 h and then at various times afterwards stimulated with 10 ng/ml E. coli LPS for 3 h. Then, production of IL-8 was determined by ELISA. Each value is the mean ± SD (n = 3). (*, P < 0.01 for comparison with the vehicle group).

NO suppresses acute-phase immune responses to LPS in vivo.To explore how NO regulation of MyD88-dependent signaling reflects innate immune or proinflammatory responses in vivo, we utilized a popular animal model of sepsis induced by i.p. administration of LPS. We first investigated the cytokine responses as the major hallmark of innate immune responses. MIP-2 is known as one of the early LPS-responsive genes, the mRNA expression of which indeed showed a rapid rise and reached a peak within 1 h after LPS stimulation in mouse peritoneal macrophages (Fig. 6A). In contrast, IL-6 is known as a late LPS-responsive gene, the expression of which showed a gradual rise and reached a peak more than 4 h after stimulation (Fig. 6A). We determined the amounts of MIP-2 and IL-6 produced in the abdominal cavity 2 h after LPS administration in wild-type, eNOS^−/−, and iNOS^−/− mice. Interestingly, eNOS^−/− mice exhibited the most intensive production of MIP-2 (Fig. 6B). On the other hand, the most prominent production of IL-6 was observed in iNOS^−/− mice (Fig. 6C). In contrast to these results, there was no significant difference in the amounts of MIP-2 and IL-6 production when the fluids were collected 12 h after LPS administration (data not shown). Thus, eNOS and iNOS at least exert a suppressive effect on early cytokine responses in vivo.

• Open in new tab
• Download powerpoint

FIG. 6.

Roles of eNOS and iNOS in early innate immune responses in vivo. (A) Peritoneal macrophages from wild-type mice were stimulated with 100 ng/ml LPS and 10 ng/ml gamma interferon for the indicated periods. Then, expression levels of mRNAs of Mip-2 and Il-6 were determined by quantitative PCR. Percent mRNA expression was calculated by taking the maximum values of mRNA levels of Mip-2 and Il-6 as 100%. (B, C) Wild-type (WT), eNOS^−/−, and iNOS^−/− mice were i.p. treated with LPS. After 2 h, PBS was injected into the abdominal cavity and fluids were collected for measurement of the amounts of MIP-2 (B) and IL-6 (C) by ELISA. Each value is the mean ± SD (n = 6). (*, P < 0.01 for comparison with wild-type mice). (D) Wild-type, eNOS^−/−, and iNOS^−/− mice were i.p. treated with LPS. The colonic temperature was monitored at 5-min intervals during a period of 10 min before and 120 min after LPS administration. Each value is the mean ± SD (n = 6). (*, P < 0.01 for comparison with wild-type mice).

We further examined LPS-induced febrile response as a hallmark of acute-phase responses of inflammation. LPS is known to act as a pyrogen to induce TLR4-dependnet polyphasic fever (44). The major initiator of LPS fever is generated prostaglandin E2, which stimulates thermoregulatory neurons and elevates body core temperature (3). LPS can directly induce prostaglandin E2 generation through the MyD88-dependent signaling pathway (47). The first-phase febrile response in eNOS^−/− mice occurred significantly earlier than that in wild-type or iNOS^−/− mice (Fig. 6D), indicating that NO from eNOS suppresses the initiation of the response. However, eNOS deficiency did not alter the magnitude of febrile response compared with that for wild-type mice, suggesting that the suppressive effect of NO is not persistent. On the other hand, a transient decrease in fever was found in wild-type and eNOS^−/− mice at about 70 min after LPS administration but not in iNOS^−/− mice (Fig. 6D), indicating that NO from iNOS suppresses promotion of the response. Thus, these results suggest that NO generated from eNOS and iNOS exerts a suppressive effect on acute-phase inflammatory responses to LPS in vivo, probably through S nitrosylation.