Abstract
Protein kinase D (PKD), also called protein kinase C (PKC)μ, is a serine-threonine kinase that is involved in diverse areas of cellular function such as lymphocyte signaling, oxidative stress, and protein secretion. After identifying a putative PKD phosphorylation site in the Toll/IL-1R domain of TLR5, we explored the role of this kinase in the interaction between human TLR5 and enteroaggregative Escherichia coli flagellin in human epithelial cell lines. We report several lines of evidence that implicate PKD in TLR5 signaling. First, PKD phosphorylated the TLR5-derived target peptide in vitro, and phosphorylation of the putative target serine 805 in HEK 293T cell-derived TLR5 was identified by mass spectrometry. Furthermore, mutation of serine 805 to alanine abrogated responses of transfected HEK 293T cells to flagellin. Second, TLR5 interacted with PKD in coimmunoprecipitation experiments, and this association was rapidly enhanced by flagellin treatment. Third, pharmacologic inhibition of PKC or PKD with Gö6976 resulted in reduced expression and secretion of IL-8 and prevented the flagellin-induced activation of p38 MAPK, but treatment with the PKC inhibitor Gö6983 had no significant effects on these phenotypes. Finally, involvement of PKD in the p38-mediated IL-8 response to flagellin was confirmed by small hairpin RNA-mediated gene silencing. Together, these results suggest that phosphorylation of TLR5 by PKD may be one of the proximal elements in the cellular response to flagellin, and that this event contributes to p38 MAPK activation and production of inflammatory cytokines in epithelial cells.
Since the discovery that TLR4 initiates an inflammatory response after recognition of bacterial LPS (1), the interest in these pattern recognition receptors and their signaling pathways has been intense. In homodimeric or heterodimeric form, TLRs respond to peptidoglycans, lipoproteins, LPS, dsRNA, unmethylated CpG-DNA, and bacterial flagellin. Ligation of these innate immune receptors, which are poised on surface or organellar membranes of many cell types, including epithelial cells, macrophages and dendritic cells, initiates signaling cascades that trigger secretion of inflammatory cytokines and chemokines as well as expression of costimulatory signals for adaptive immunity.
TLRs are characterized by N-terminal leucine-rich repeats, a central transmembrane domain, and a cytosolic C-terminal Toll/IL-1R (TIR)3 domain, which links the receptor to proximal signaling adaptors such as MyD88 or TIR domain-containing adaptor-inducing IFN-β. In the MyD88-dependent pathway common to most TLRs, receptor ligation results in phosphorylation and activation of IL-1R-associated kinase family members, allowing the recruitment of the adaptor TNFR-associated factor 6, which activates the TGF-β-activated kinase 1. The latter then activates the transcription factors NF-κB (via I-κB degradation) and AP-1 (via p38 MAPK and JNK), leading to cytokine and chemokine expression (reviewed in Ref. 2).
TLR5, which recognizes bacterial flagellin, is thought to transduce signals via the MyD88-dependent pathway only (3, 4). Activation of IL-1R-associated kinase (5), nuclear translocation of NF-κB (6, 7), and activation of MEK (8), p38 MAPK, and ERK (6) have all been demonstrated after flagellin treatment in many cell types, including professional APCs and epithelial cells. These pathways lead to the secretion of β-defensin (9), cytokines IL-6 and TNF-α (10), and/or chemokines MIP3α and IL-8 (3, 4), which recruit and activate dendritic cells and neutrophils, respectively. However, the characteristic features of TLR5 signaling that confer specific responses to flagellin as opposed to other TLR agonists have yet to be identified.
Using an in silico analysis to identify potential protein interaction sites in the TIR domain of TLR5, we identified a putative protein kinase D (PKD) phosphorylation site (see Fig. 1 A). PKD is a serine-threonine kinase of the calmodulin-calcium-dependent kinase family, and contains diacylglycerol-binding and pleckstrin homology domains within its regulatory N-terminal region. The three known PKD subtypes PKD1 (also called protein kinase C (PKC)μ), PKD2, and PKD3 (PKCν) are involved in diverse aspects of cellular function including membrane trafficking, differentiation, proliferation, and apoptosis as well as the responses to BCR and TCR ligation, regulatory peptides, and oxidative stress (see reviews Refs. 11, 12). Using the interaction between the enteroaggregative Escherichia coli H18 flagellin (FliC) and human TLR5 in epithelial cell lines as a model, we report that PKD interacts with TLR5, and that this kinase is required for transducing flagellin-stimulated inflammatory responses.
Materials and Methods
Abs and vectors
The following Abs were used: anti-PK (Serotec), anti-PKD1, anti-phospho-PKD1 (S916), anti-phospho-p38 MAPK (T180/Y182), anti-total p38 (Cell Signaling Technology), and anti-GAPDH (Fitzgerald Industries International).
TLR5 cloned into pEF6-V5His-Topo (Invitrogen Life Technologies) was a gift from A. Aderem (Institute for Systems Biology, Seattle, WA). pEGFP was from BD Clontech. Hemagglutinin (HA)-PKD1 and pSPKD1-1 were provided by A. Toker (Harvard University, Boston, MA). The IL-8pLuc reporter vector (13) was provided by B. Salh (University of British Columbia, Vancouver, British Columbia, Canada).
The TLR5 S805A mutant was produced by the quick-change circular mutagenesis method (Stratagene), using the following primers: 5′-gtaccagttgatgaaacatcaagcccatcagaggctttgtacag and 5′-ctgtacaaagcctctgatgg cttgatgtttcatcaactggtac.
TLR2 in pcDNA3.1 was provided by M. Smith (University of Virginia, Charlottesville, VA); it was cloned into the KpnI and XbaI sites of pEF6/V5-HisTOPO (Invitrogen Life Technologies) using the following primers: TLR2V5F, 5′-aaaaggtaccatgccacatactttgtggatggtgtggg and 5′-aaaatctagactctttatcgcagctctcagatttacccaaaatcc to create TLR2-V5.
The pSuper-based small hairpin RNA (shRNA) vector for PKD1 knockdown pSPKD1-2 as well as the scrambled version of pSPKD1-1 (pSscr1-1) was created by annealing the following oligonucleotides and cloning them into the pSuper vector (OligoEngine): pPKD1–8 top 5′-gatccccgcagattcaactgccataaacttcaagagagtttatggcagttgaatctgcttttta, bottom 5′-agcttaaaaagcagattcaactgccataaactctcttgaagtttatggcagttgaatctgcggg; and pscrPKD1-1A top 5′-gatcccctgatcgggtgctgagctggttcaagagaccagctcagcacccgactatttta, bottom 5′-agcttaaaaatagtcgggtgctgagctggtctcttgaaccagctcagcacccgactaggg. All oligonucleotides were obtained from Operon Biotechnologies.
Cell culture and transfections
Caco-2 cells were obtained from the American Type Culture Collection (ATCC) and grown in HyQ DMEM-high glucose (HyClone) supplemented with 10% FBS (Invitrogen Life Technologies), nonessential amino acids (HyClone), penicillin (100 U/ml), and streptomycin (100 μg/ml) (Sigma-Aldrich). Cells were seeded at 1 × 106 cells/ml in polystyrene culture dishes and used for experiments 7–14 days after becoming confluent. Human embryonic kidney (HEK) 293T cells also obtained from ATCC were grown in HyQ DMEM-high glucose (HyClone Laboratories) supplemented with 10% heat-inactivated FBS (HyClone), 25 mM HEPES (StemCell Technologies), 2 mM glutamine (StemCell Technologies), and antibiotics as earlier described. Caco-2 cells were passaged every 1–2 wk, while the HEK cells were passaged two to three times per week.
Cell transfection
HEK cells were seeded at 2 × 105 cells/ml in growth medium without penicillin or streptomycin in plates coated with 0.01% poly-l-lysine (Sigma-Aldrich). The following day, cells were transfected with LipofectAMINE 2000 (Invitrogen Life Technologies) according to the manufacturer’s instructions. Medium was changed the following day, and cells were stimulated after an additional 18 h.
Caco cells were transfected by electroporation as follows: cells were trypsinized, washed in PBS, and suspended at 107/ml in Opti-Mem (Invitrogen Life Technologies). Plasmid DNA and sheared salmon sperm DNA (Sigma-Aldrich) totaling 25–35 μg was added to 400 μl of cell suspension. After brief incubation, cells were transferred to a 2-mm cuvette and pulsed in an ECM 630 electroporator (BTX) using capacitance of 1000 μF, resistance 50 Ohm, and potential 250 V. Cells were allowed to rest 5 min at room temperature, transferred into growth medium, and seeded in culture plates or dishes.
Stimulation and inhibition of cells
The H18 flagellin from enteroaggregative E. coli strain 042 was expressed in pCR-NT-T7-Topo (Invitrogen Life Technologies) in BL21 (DE3) pLysS cells and purified by metal affinity chromatography followed by polymyxin B chromatography as previously described (14).
Pharmacological inhibitors (Gö6976, Gö6983) were obtained from Calbiochem, and phorbol 12,13-dibutyrate (PDB) and IL-1β were from Sigma-Aldrich. Gö6976 was used at a final concentration of 3 μM, Gö6983 at 12 μM, PDB at 200 nM, and IL-1β at 10 ng/ml. FliC (flagellin) was used at a final concentration of 0.5–1 μg/ml, while Pam3CysSK4 (EMC Microcollections) was used at 10 μg/ml. Cells were pretreated with inhibitors or DMSO vehicle for 1 h, followed by stimulation for varied amounts of time, depending on cell type and readout.
For detection of PKD, cells were lysed in 50 mM Tris-HCl (pH 7.4), 1% Triton X-100, 150 mM NaCl, 5 mM EDTA (pH 7.4), and Protease Inhibitor Cocktail for mammalian cells (Sigma-Aldrich). For detection of phosphorylated proteins, cells were lysed in 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Nonidet P-40, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 2 mM Na3VO4, and Protease Inhibitor Cocktail. Equal amounts of proteins were separated by SDS-PAGE and blotted for Western blot analysis.
For analysis of secreted IL-8, in the case of Caco-2 cells, culture medium was removed after 3 h of stimulus. For HEK 293T cells, longer time points (6 or 24 h) were used due to the slower kinetics of IL-8 production. The IL-8 contents of supernatants were determined by ELISA (OptEIA; BD Biosciences).
In vitro kinase assays
The synthetic peptides purchased from GenScript and dissolved at 3 mM in DMSO were PKD (AALVRQMSVAFFFK), TLR5 (YQLMKHQSIRGFVQ), and S805A (YQLMKHQAIRGFVQ). Peptides were added to a final concentration of 150 μM in a reaction mixture of 0.03% Triton X-100, 20 mM HEPES (pH 7.4) with 25 ng of active recombinant PKD1 (Upstate Biotechnology). A kinase mixture containing [γ-32P]ATP (1 μCi/reaction) in 75 mM MgCl2, 500 μM ATP, 20 mM MOPS (pH 7.2), 25 mM β-glycerophosphate, 5 mM EGTA, 1 mM Na3VO4, and 1 mM DTT was added, and the reaction allowed to proceed for 10 min at 30°C. The reaction was stopped by the addition of 100 μl of 0.75% phosphoric acid, and the entire reaction spotted on squares of P81 phosphocellulose paper. Papers were washed three times in 0.75% phosphoric acid and once in acetone, air-dried, and counted for radioactivity. Samples containing no target peptide were used to obtain background counts.
Coimmunoprecipitation
HEK 293T cells plated at a density of 2 × 105 cells/ml in 6-well plates were transiently transfected with a total of 3 μg of DNA consisting of HA-PKD, TLR5-V5, and/or TLR2-V5. After stimulation for the indicated times with 1 μg/ml flagellin, cells were washed with cold PBS and lysed with 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Octyl β-d-glucopyranoside, 1 mM β-glycerophosphate, 2.5 mM sodium pyrophosphate, 2 mM Na3VO4, and protease inhibitor mixture (Sigma-Aldrich).
After preclearing 0.75–1.5 mg of total protein in 400 μl volume with protein A-agarose (Bio-Rad) for 10 min at 4°C, 7 μl of anti-V5 was added to the supernatant and the solution was nutated overnight at 4°C. Protein A-agarose beads were blocked overnight with 5% BSA in lysis buffer, washed, and added to the samples the next morning. After 1 h, the immune complexes were washed four times in modified radioimmunoprecipitation assay buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.25% sodium desoxycholate, 2 mM Na3VO4, and protease inhibitor mixture). Beads were eluted by boiling in 2× SDS-PAGE loading buffer. Proteins were then separated by SDS-PAGE, blotted onto polyvinylidene difluoride membranes, and probed with anti-PKD1. After stripping, blots were probed with anti-V5 to detect TLR2 and TLR5.
Quantitative RT-PCR
For mRNA quantitation, cells were lysed 1 h (for IL-8) or 3 h (for CCL20) after flagellin stimulation. Total RNA was isolated using RNeasy mini columns (Qiagen), and equal amounts of RNA from each sample were reverse transcribed using RevertAid H Minus First-Strand cDNA Synthesis kit (Fermentas). Real-time PCR was performed using SYBR Green master mix (Applied Biosystems) and run on an Opticon thermal cycler (Bio-Rad). Equal volumes from each reaction were run in triplicate for 40 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 45 s, with fluorescent detection after each cycle. Melting curves were performed after 40 cycles and correct product size was confirmed by agarose electrophoresis. Primers used were: GAPDH (forward) 5′-gaaggtgaaggtcggagtc, (reverse) 5′-gagggatctcgctcctggaaga; CCL20 (forward) 5′-ccaagagtttgctcctggct, (reverse) 5′-tgcttgctgcttctgattcg; and IL-8 (forward) 5′-atgacttccaagctggccgtggct, (reverse) 5′-tctc agccctcttcaaaaacttctc.
Relative mRNA amounts were calculated as follows: the cycle threshold (Ct) for fluorescent detection was measured for each sample, and the difference between the IL-8 and GAPDH cycle threshold values was calculated (dCt). The difference in cycle threshold for untreated Caco-2 cells was subtracted from the difference in cycle threshold for each experimental condition to yield ddCt. The fold increase in mRNA expression for each condition compared with the control was then calculated as 2−ddCt.
EMSA analysis
Caco-2 cells were lysed 30 min after stimulation as described above. Nuclear protein extracts were isolated, tested for protein concentration, and diluted in nuclear extract buffer to 1 mg/ml. A double-stranded NF-κB-binding probe of sequence 5′-agttgaggggactttcccaggc was end-labeled with [γ-32P]ATP using polynucleotide kinase (Fermentas) and purified by Sephadex G-25 chromatography and ethanol precipitation. To measure NF-κB activity, a reaction mixture containing 5 μg of nuclear extract, and 7.5 μl of binding buffer (20 mM HEPES, 50 mM KCl, 0.1 mM EDTA, 5% glycerol, 200 μg/ml BSA, and 1 mM DTT (pH 7.9), and 1.5 μg of poly(deoxyinosinic-deoxycytidylic acid) (Amersham Biosciences) in 20 μl total volume was incubated 20 min on ice, and 1 μl of labeled NF-κB probe was added. Following a 15-min incubation at room temperature, reactions were separated by 4% nondenaturing PAGE, and gels were dried onto filter paper and examined by autoradiography. NF-κB activity was measured by delayed mobility, and specificity was verified by including a 100-fold excess of unlabeled NF-κB-binding oligonucleotide in some samples.
IL-8 reporter assays
Caco cells were transfected as described with 1 μg of pEGFP, 7.5 μg of IL-8pLuc, and sheared salmon sperm DNA to a total of 30 μg, diluted in 1.2 ml of growth medium, and seeded at 100 μl/well in 96-well plates. Medium was changed the following day and fluorescence verified by microscopy. Transfection efficiencies of 20% were routinely achieved. After an additional 5 days, cells were pretreated with inhibitors for 1 h and then stimulated with flagellin or IL-1β for 6 h. Supernatants were removed and tested for IL-8 concentration, and cells were lysed with a 1:1 mixture of growth medium and Bright-Glo reagent (Promega). After 2 min, cells were read in a microplate luminometer (Berthold). The lysates were recovered and GFP fluorescence measured in a Fluoroskan Ascent FL (Thermo Electron). The ratio of luminescence to fluorescence in arbitrary units (to correct for cell number and transfection efficiency) was calculated for each sample, and the fold increase in this value compared with controls within the same experiment was defined as the fold increase in expression.
Structural modeling of TLR5
The TIR domain of TLR5 was mapped with Molecular Operating Environment version 2005.10 (Chemical Computing Group) software, using the crystal structure of the TLR2 TIR domain as a scaffold (15). To quantify the degree of surface exposure of individual residues in the structure of the TLR5 domain of interest, we calculated the steric hindrance parameters (Rs) for all alpha-carbon atoms of the TLR5 TIR domain, as previously described (16, 17).
Mass spectrometric identification of TLR5 phosphorylation
HEK cells transfected in 15-cm dishes were treated with flagellin (1 μg/ml) for 10 min, washed in PBS, and lysed by freeze-thaw in 10 mM HEPES (pH 7.4). Lysates were spun at 15,000 × g for 10 min to pellet insoluble material, and supernatants were then spun at 100,000 × g for 1 h to pellet membranes. Membrane proteins were immunoprecipitated with anti-V5 as described (18) and separated by SDS-PAGE. Gels were stained with SYPRO Ruby (Invitrogen Life Technologies) and the band corresponding to TLR5 (verified by its absence in identically treated cells not expressing TLR5) was excised, trypsinized in situ, and analyzed by liquid chromatography-single mass spectrometry at the Nucleic Acids and Protein Synthesis Core Mass Spectrometry facility at the University of British Columbia (Vancouver, British Columbia, Canada). Predicted tryptic peptides of TLR5 with a mass shift of 80 kDa (the size of a phosphate group) were specifically sought.
Statistical analysis
Statistics were performed on original data using the VassarStats statistical computation website (http://vassun.vassar.edu/∼lowry/VassarStats.htm). Except where noted, multiple groups were analyzed by ANOVA to verify the presence of significant differences. This analysis was followed by paired testing by Student’s t test or Tukey HSD post hoc test.
Results
The putative PKD target serine in TLR5 is predicted to be surface-exposed
Residues 800–806 in the TIR domain of TLR5 contain a putative PKD phosphorylation motif (Fig. 1 A). Interestingly, although there is a similar region in TLR4 that also fits the conserved PKD recognition/phosphorylation site, none are found in other TLRs. Phosphorylation of degenerate peptide library arrays have indicated that PKD strongly prefers a leucine at −5 and an arginine or lysine at −3 relative to the target serine or threonine (19, 20). Typically, the residue at +1 is hydrophobic. Surrounding residues are less important but do affect kinase assays to varying degrees as shown in detailed analyses by Doppler et al. (20).
Structural modeling of TLR5 suggests that the region containing this PKD recognition site lies at the junction of the DD loop and the αD helix, in a surface-accessible location (Fig. 1 B). An in silico steric hindrance analysis assigned steric hindrance values of 1.74 and 1.88 to K802 and S805, respectively, which fall within the lowest third of steric hindrance values, suggesting surface availability on TLR5.
S805 is phosphorylated in flagellin-treated cells
To measure phosphorylation of TLR5 in vitro, we immunoprecipitated TLR5-V5 from overexpressing HEK 293T cells that had been treated with flagellin for 10 min. The immunoprecipitates were separated by SDS-PAGE, and the band corresponding to TLR5 (absent in cells not expressing TLR5) was evident after SYPRO Ruby staining. This band was excised and analyzed by mass spectrometry. As shown in Fig. 1 C, the pentapeptide fragment containing S805 had a mass shift of 80 kDa, corresponding to the mass of a phosphate group.
PKD phosphorylates the predicted TLR5-derived consensus peptide in vitro
To gather evidence for the phosphorylation of S805 by PKD, we looked at the ability of recombinant PKD1 to phosphorylate TLR-derived peptides in direct kinase assays. A 14-mer peptide containing the PKD motif in TLR5 was synthesized, as was as the optimal PKD consensus peptide (21) as a positive control and TLR5-mutant S805A (target serine mutated to alanine) as a negative control. These peptide sequences are listed in Materials and Methods. In vitro kinase assays with active, recombinant PKD1 were used to measure the ability of this enzyme to incorporate 32P from ATP. As shown in Fig. 2, the PKD consensus peptide displayed the highest incorporation of 32P and the S805A-derived peptide displayed only background radioactivity, as would be expected. The TLR5 peptide showed 26 ± 19% of the 32P incorporation displayed with the consensus peptide, confirming the identity of this TLR5 sequence as a PKD substrate.
Treatment of epithelial cells with flagellin increases interaction between TLR5 and PKD1
Direct interaction between TLR5 and PKD1 was examined by coimmunoprecipitation using HEK 293T cells. Cells were transiently transfected with HA-PKD and TLR5 containing a V5 epitope tag (TLR5-V5), treated briefly (1–10 min) with 1 μg/ml flagellin, and lysed with a Tris-based buffer containing 1% octyl β-d-glucopyranoside. One milligram of total soluble cell lysate was immunoprecipitated with anti-PK (which binds the V5-tag) and immunoblotted with anti-PKD1. Although there was nonspecific pulldown of PKD as shown by the presence of PKD in cells not transfected with TLR5, interaction between TLR5 and PKD1 clearly increased with exposure to flagellin only in cells overexpressing both proteins (Fig. 3). To determine the specificity of this ligand-dependent interaction, we repeated experiments using TLR2 in place of TLR5. As seen in Fig. 3, there was background or nonspecific interaction resulting in PKD pulldown that did not change after cells were treated with the TLR2 agonist Pam3CysSK4.
S805 of TLR5 is required for IL-8 production
To investigate the importance of S805 in TLR5 signaling, we mutated this residue to an alanine to eliminate the putative phosphorylation motif. This construct was transiently expressed in HEK 293T cells and IL-8 release in response to flagellin was measured. As shown in Fig. 4, transfection of cells with wild-type TLR5 but not the S805A mutant caused a strong increase in IL-8 production after treatment with flagellin. Stimulation with IL-1β induced equal IL-8 production in both mutants, whereas Western blot analysis of V5-tagged proteins from total cell lysates confirmed comparable receptor expression in cells transfected with both constructs (data not shown).
Pharmacological inhibition of PKD reduces flagellin-stimulated inflammatory responses in Caco-2 cells
To determine the relevance of PKD in the inflammatory TLR5 signaling of intestinal epithelial cells, we looked at the effect of pharmacological inhibition of PKD on the flagellin-induced production of IL-8 in Caco-2 cells. Although there is no known substance that specifically inhibits PKD, a combination of staurosporine derivatives Gö6976 and Gö6983 can uncover PKD involvement. The former selectively inhibits classical PKC isoforms as well as PKD, whereas the latter suppresses the activity of all three PKC subgroups (classic, novel, and atypical) but not PKD (22). As shown in Fig. 5, A and B, pretreatment of Caco-2 cells with Gö6976 but not with Gö6983 reduced both IL-8 mRNA concentrations (p < 0.05 by Student’s t test) and IL-8 promoter activity (p < 0.005 by Student’s t test) after flagellin treatment compared with those measured after pretreatment with DMSO vehicle alone. A similar response pattern was seen when mRNA concentrations of the dendritic cell chemokine CCL20 (MIP3α), which is also produced by Caco-2 cells in response to TLR5 ligation (23), were examined (Fig. 5,C). Moreover, Gö6976 but not Gö6983 significantly inhibited flagellin-induced IL-8 secretion from Caco-2 cells (Fig. 5 D; mean reduction 75 ± 9%; p < 0.0001 by Student’s t test). In contrast, IL-8 release from cells stimulated with PDB esters, which activates PKD and various isoforms of PKC, was reduced by both inhibitors, demonstrating the efficacy of Gö6983.
Pharmacological PKD inhibition before flagellin treatment has no effect on NF-κB activation but prevents activation of p38 MAPK in Caco-2 cells
MyD88-dependent TLR signaling is thought to branch into at least two separate pathways, one involving the activation of NF-κB and one leading to activation of MAPK (2). To better understand the downstream effects of PKD inhibition, we looked at the effect of Gö6976 and Gö6983 on elements of both pathways in Caco-2 cells. Flagellin, like other TLR agonists, signals through I-κB kinase-mediated I-κB degradation and translocation of NF-κB to the nucleus where it activates proinflammatory cytokines and chemokines (7). EMSA with nuclear extracts from inhibitor-treated, flagellin-stimulated cells showed that neither PKD nor PKC inhibition had an effect on NF-κB activation after flagellin stimulation (Fig. 6 A), indicating that PKD does not modulate this branch of TLR5 signaling in Caco-2 cells.
Activation of TGFβ-activated kinase 1 through TNFR-associated factor 6 after TLR ligation leads not only to I-κB kinase activation but also to phosphorylation of MAPKs. Both ERK and p38 MAPKs have been implicated in TLR5 signaling (6, 7). Interestingly, pretreatment of Caco cells with Gö6976 but not Gö6983 abrogated the activation of p38 MAPK by flagellin as measured by phosphorylation of T180/Y182 (Fig. 6 B). To determine whether this effect is specific for TLR5-mediated inflammation, the experiments were repeated with cells stimulated with 10 ng/ml IL-1β. Although densitometric analysis of three pooled experiments showed a small reduction in phospho-p38 MAPK levels in Gö6976-treated cells stimulated with IL-1β (92 ± 14% decreased to 63 ± 11%; p = 0.026), the effect was far more pronounced in FliC-treated cells, in which the relative intensity of 96 ± 4% was reduced to 15 ± 3% (p < 0.001). At both 30 and 60 min of stimulation, Gö6976 inhibited p38 phosphorylation in response to FliC significantly more than it inhibited the IL-1β response (p < 0.001).
Targeted knockdown of PKD1 in TLR5-transfected HEK 293T cells reduces IL-8 production and p38 MAPK activation in response to flagellin
As the specificity of pharmacological inhibitors can be a contentious issue, we took steps to verify PKD involvement in TLR5 signaling by shRNA-mediated silencing of PKD1. Because of the poor transfection rate of Caco-2 cells (∼20%), these experiments were done exclusively in HEK 293T cells. Treatment of TLR5-transfected, flagellin-stimulated HEK 293T cells with Gö6976 caused a significant reduction in secreted IL-8 (51 ± 4%, p < 0.01 by ANOVA), whereas treatment with Gö6983 had no effect, indicating that a PKD-dependent pathway is involved downstream of TLR5 ligation in these cells (Fig. 7,A). Again, PDB was used as a positive control for the activity of Gö6983. For shRNA experiments, HEK 293T cells were transiently transfected with TLR5 and a pSuper-based shRNA vector containing PKD1-derived sequences. One sequence, pSPKD1-1, has been reported to specifically silence expression of PKD1 (24). Another shRNA construct, pSPKD1-2, was generated to ensure that effects were not due to nonspecific interactions of the small interfering RNA (siRNA). As an additional control, we generated an shRNA construct with a scrambled version of the PKD1-1 siRNA sequence, pSscr1-1. Western blot analysis shows that HEK 293T cells expressing the knockdown constructs had reduced PKD expression in comparison with cells expressing the empty pSuper vector (Fig. 7,B). FliC challenge of cells expressing pSPKD1-1 reduced IL-8 production in a consistent but modest fashion compared with cells expressing pSuper alone (61 ± 11%, p < 0.02 by Student’s t test), whereas cells expressing pSPKD1-2 were more substantially impaired in their IL-8 response (17 ± 6%, p < 0.001 by Student’s t test). The negative control pSscr1-1 failed to knock down PKD expression (data not shown) or flagellin-induced IL-8 release (Fig. 7 C). As PKD isoforms are reportedly involved in protein trafficking (25), the above experiments were repeated with IL-1β-induced IL-8 production to rule out the possibility of a nonspecific effect on IL-8 secretion. Although significant reduction of IL-8 in response to IL-1β stimulation was observed in the supernatants of cells expressing both pSPKD1-1 and pSPKD1-2 (p < 0.01), the effect was much less than that exhibited after FliC stimulation (p < 0.001, FliC vs IL-1β). This demonstrates that PKD inhibition has specific effects on TLR5 signaling that are not shared by signaling pathways downstream of IL-1. More importantly, the reduction in secreted IL-8 after exposure to flagellin cannot be the sole consequence of nonspecific effects on protein secretion.
As previous experiments in Caco cells indicated that at least one contribution of PKD to TLR5-linked inflammatory signaling may be through the activation of p38 MAPK, we sought to verify this finding using the PKD1 knockdown vectors. Cells were transfected with TLR5 and either pSuper or pSPKD1-2 and then stimulated with 0.5 μg/ml FliC. To control for specificity, 10 ng/ml IL-1β was added to cells transfected with knockdown vector or control (without exogenous TLR5). As shown in Fig. 7,D, PKD shRNA significantly reduced p38 phosphorylation in response to flagellin. This response was confirmed by densitometry (n = 4, p < 0.05 by Student’s t test). In contrast, IL-1β-induced p38 phosphorylation was minimally inhibited (Fig. 7 D).
Discussion
As TLRs are positioned at the apex of the immune response, a complete understanding of their signaling is required to manipulate downstream events that control both innate and acquired immunity. Consequently, much effort has gone into the elucidation of TLR signaling, especially in the case of TLR2 and TLR4. Much less is known about TLR5 signaling, although this receptor is highly expressed on epithelial, endothelial, and professional immune cells. In particular, TLR5 is expressed by several cell types in the intestinal mucosa, a major site of exchange with the external environment and home to an enormous population of commensal bacteria. Hence, TLR5 is thought to play an important role in normal immune homeostasis; moreover, flagellin recognition by TLR5 has recently been shown to be involved in disease states such as Crohn’s disease (26), systemic lupus erythematosus (27), Salmonella gastroenteritis (28) and Legionnaire’s disease (29).
Like all TLRs, TLR5 contains a TIR domain that is responsible for interaction with intracellular signaling elements. The structure/function relationships within the TLR5 TIR domain have yet to be elucidated, but differences in cellular responses to various TLR ligands suggest that the TIR domains of TLRs mediate distinct activities.
We examined the TIR domain of TLR5 to identify potential binding sites for signaling partners, and identified a putative PKD recognition/phosphorylation site targeting S805. We therefore hypothesized that the phosphorylation of TLR5 by PKD is required for cellular responses to flagellin. In support of this, mass spectroscopic analysis of flagellin-treated TLR5 showed that the putative target serine was phosphorylated. Moreover, recombinant PKD was able to phosphorylate a TLR5-derived peptide containing the putative PKD recognition sequence.
More direct evidence of PKD phosphorylation of TLR5 was provided by coimmunoprecipitation, which indicated flagellin-dependent interaction of PKD and TLR5. Although the results showed some nonspecific interaction between the two proteins, the strength of the interaction clearly increased within 5–10 min of flagellin treatment. Although we also found some nonspecific interaction between PKD and TLR2, this activity did not consistently increase after receptor ligation. It is likely that ligation of TLR5 induces conformational changes that enhance the direct or indirect interaction with PKD. Although these results provide evidence for phosphorylation of TLR5 by PKD, more conclusive studies are needed.
The physiological relevance of S805 phosphorylation was established by mutation of the target serine to alanine, which eliminated the IL-8 response to flagellin. Furthermore, reduction of PKD activity, either through pharmacologic inhibition or shRNA-mediated gene knockdown, also reduced IL-8 release. We also found a smaller but significant effect of PKD inhibition on IL-1β responses. This suggests that PKD is involved in more than one aspect of inflammatory signaling. However, flagellin responses were consistently more dependent on PKD activity, which suggests a specific role for PKD in TLR5 signaling, rather than a general role in protein secretion; the latter is of special importance as PKD isoforms are involved in protein secretion (25).
In the majority of reports concerning PKD, activation of PKD is PKC-dependent. In these cases, stimulation is followed by activation of phospholipase C, resulting in the production of diacylglycerol and activation of novel PKC isoforms that then phosphorylate activation loop serines (S738/S742 in human PKD) within the kinase domain of PKD (30). However PKC-independent activation, which correlates with a lack of phosphorylation of the activation loop serines, has been described (31, 32). We found that the addition of flagellin to Caco-2 cells does not result in phosphorylation of the activation loop serines, although addition of PDB ester does (data not shown). The inability of the PKC inhibitor Gö6983 to block flagellin-mediated IL-8 responses in Caco-2 and HEK 293T cells may indicate that PKD is constitutively active in these cells. In contrast, the PKC inhibitor Gö6983 was able to block flagellin-stimulated IL-8 release in HeLa cells (data not shown), indicating that the mechanism of PKD activation could depend on the cell type.
To ascertain which downstream signaling elements are involved in PKD modulation of TLR5 signaling, we examined the effect of inhibitors on two separate components of the pathway. The best characterized element of TLR signaling is NF-κB. PKD has been shown to contribute to NF-κB activation through interaction with the I-κB kinase complex, leading to degradation of I-κBα, in response to oxidative stress in HeLa cells (24). In our experiments with Caco-2 cells, PKD inhibition did not prevent activation of NF-κB (as assessed by EMSA) in response to flagellin. It still remains to be seen whether PKD plays a role in NF-κB signaling through alteration of either the phosphorylation status of the transcription factor itself or the conformation of dependent promoter sites as described in other systems (33).
Although NF-κB activation is generally held to be the central component of TLR signaling, we have previously reported that inhibition of NF-κB activation by Bay11–7082 only reduced flagellin-dependent IL-8 production by 50%, whereas inhibition of p38 MAPK with SB203580 reduced it by 90% (6). In this study, we show that pharmacological and shRNA-mediated PKD inhibition abrogated flagellin-mediated activation of p38. A similar mechanism has been described for the activation of p38 MAPK by the bone morphogenic protein, BMP-2, in mouse cell lines; the requirement of PKD was confirmed with siRNA-mediated PKD silencing (31), indicating that the phenotype seen after treatment with Gö6976 was not due to inhibition of alternative proteins upstream of p38 MAPK. It is notable that BMP-2 related PKD signaling seems to be independent of PKC, as may be the case in Caco-2 and HEK 293T cells. It is still not clear how p38 MAPK activation facilitates IL-8 production. In IL-1β signaling, p38 MAPK activation contributes to both AU-rich element-mediated IL-8 mRNA stabilization (34, 35) and to up-regulation of IL-8 promoter activity (13). Although loss of mRNA stabilization could be responsible for the reduction of IL-8 mRNA after Gö6976 treatment, the decrease in promoter activity suggests that p38 MAPK is involved in other aspects of flagellin-stimulated IL-8 production as well.
This report is the first on PKD involvement in TLR-linked innate immune signaling, although PKDs are known to participate in the response to BCR and TCR cross-linking in lymphocytes (36, 37). In BCR signaling, PKD helps facilitate the synergy between ligation of Ag and costimulatory receptors CD19 (38) and CD40 (39). At present, it is unclear how PKD involved in TLR5 signaling is activated. It is tempting to speculate that signals other than TLR5 receptor ligation are involved, allowing further information to enter into and modulate the TLR5 response.
Acknowledgments
We thank Chelsea Tirling and Paige Wark for technical assistance. Structural modeling of TLR5 was done by Artem Cherkasov. We acknowledge Peter Storz, Alex Toker, and Michael Smith for providing valuable constructs. We also thank Neil Reiner for scientific discussions and criticism.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant 992840.01 from Burroughs-Wellcome Career Award in Biomedical Sciences, Operating Grant 64355 from the Canadian Institutes for Health Research, by the Canada Foundation for Innovation New Opportunities Fund 4453, by New Investigator Awards from the Canadian Institutes for Health Research, and by the Vancouver Coastal Health Research Institute “In It for Life” Fund (to T.S.S.).
Abbreviations used in this paper: TIR, Toll/IL-1R; PKD, protein kinase D; PKC, protein kinase C; FliC, flagellin; PDB, phorbol 12,13-dibutyrate; shRNA, small hairpin RNA; siRNA, small interfering RNA; HEK, human embryonic kidney; HA, hemagglutinin; shRNA, small hairpin RNA; Ct, cycle threshold.