- Email: [email protected]
Protein kinase D1 and D2 are involved in chemokine release induced by toll-like receptors 2, 4, and 5
Cellular Immunology 264 (2010) 135–142
Contents lists available at ScienceDirect
Cellular Immunology journal homepage: www.elsevier.com/locate/ycimm
Protein kinase D1 and D2 are involved in chemokine release induced by toll-like receptors 2, 4, and 5 Theodore S. Steiner *, Sabine M. Ivison, Yu Yao, Arnawaz Kifayet Division of Infectious Diseases, Department of Medicine, University of British Columbia, Rm. D452 HP East, VGH, 2733 Heather St., Vancouver, BC, Canada V5Z 3J5
a r t i c l e
i n f o
Article history: Received 16 December 2009 Accepted 22 May 2010 Available online 27 May 2010 Keywords: Protein kinase D Innate immunity Toll-like receptors Chemokines
a b s t r a c t The protein kinase D (PKD) family consists of three serine-threonine kinases involved in cellular proliferation, motility, and apoptosis. We previously reported that human toll-like receptor 5 (TLR5) contains a consensus PKD phosphorylation site. Flagellin stimulation of cells activated PKD1, and inhibition of PKD1 reduced flagellin-induced interleukin-8 (IL-8) production in epithelial cells. In the current work, we examined PKD1 and PKD2 involvement downstream of TLR5, TLR4 and TLR2. We found that inhibition of either kinase with shRNA reduced IL-8 and CCL20 release due to TLR4 and TLR2 agonists to a similar extent as previously reported for TLR5. PKD1 and PKD2 inhibition reduced NF-jB activity but not MAPK activation. These results demonstrate that both PKD1 and PKD2 are required for inflammatory responses following TLR2, TLR4, or TLR5 activation, although PKD1 is more strongly involved. These kinases likely act downstream of the TLRs themselves to facilitate NF-jB activation but not MAP kinase phosphorylation. Ó 2010 Elsevier Inc. All rights reserved.
1. Background Protein kinase D (PKD) refers to three related serine-threonine kinases of the calmodulin-calcium dependent (CaM)-kinase family. The three known PKD subtypes PKD1 (PKCl), PKD2, and PKD3(PKCm) are involved in diverse aspects of cellular function including membrane trafficking, differentiation, proliferation and apoptosis, as well as the responses to B-and T-cell receptor ligation, regulatory peptides and oxidative stress (for reviews see [1,2]). We previously reported that toll-like receptor 5 (TLR5), the pattern recognition receptor for bacterial flagellin, contains a consensus PKD phosphorylation site in its C-terminal TIR domain [3]. Based on this observation, we found that PKD activity was increased following flagellin stimulation of TLR5, and that the predicted TLR5 phosphorylation site was, in fact, capable of being phosphorylated by PKD1. Moreover, we showed that reduction of PKD activity using the pharmacological inhibitor Gö6976 or siRNA specific for PKD1 was required for interleukin (IL)-8 production from cells stimulated with flagellin. Subsequent to this work, Park et al. [4] reported that PKD is also activated in human and murine macrophages by interaction of TLR9 with its natural ligand, unmethylated CpG oligodeoxynucleotide (ODN). They later reported that PKD1 in murine pacrophages is activated by all of the MyD88-dependent TLR ligands, and is re-
* Corresponding author. Fax: +1 604 875 4013. E-mail addresses: [email protected] (T.S. Steiner), sabine.ivison@ gmail.com (S.M. Ivison), [email protected] (A. Kifayet). 0008-8749/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.cellimm.2010.05.012
quired for TRAF-6 ubiquitination and downstream signaling [5]. Interestingly, they found no activation of or requirement for PKD2 or PKD3 in this pathway. In addition, several groups have recently reported that PKD is required for cytokine production in endothelial cells exposed to vascular endothelial growth factor (VEGF) [6,7]. Because of this emerging evidence that PKD1 is involved in inflammatory signaling in hemopoietic and endothelial cells, we hypothesized that this enzyme might play a more widespread role in TLR signaling pathways in epithelial cells as well. We tested the role of PKD1 and PKD2 on production of two epithelial-derived chemokines, IL-8 (CXCL8) and CCL20 (MIP3a), in response to agonists of TLR2, TLR4, and TLR5. As discussed below, we found that both PKD1 and PKD2 are required for production of these chemokines in epithelial cells, primarily through the NF-jB pathway. 2. Methods 2.1. Antibodies and vectors The following antibodies were used: anti-PK (V5 tag, Serotec, Oxford, UK), anti-phospho-p38, anti-phospho-ERK, anti-PKD1, anti-pSer916-PKD, and PKD2 (Cell Signaling, Cambridge, MA), and anti-GAPDH (Fitzgerald, Concord, MA). The human TLR5 open reading frame cloned into pEF6-V5His-Topo (Invitrogen) was a gift from A. Aderem (Institute for Systems Biology, Seattle). pEGFP-N1 was from BD Clontech (Mountain View, CA). HA-PKD1 and and HAPKD2 were provided by A. Toker (Harvard University). The
136
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
CCL20pLuc reporter vector [8] was provided by A. Keates (Harvard University). The NF-jBpLuc reporter was provided by Bruce Vallance (University of British Columbia). TLR2, TLR4, and CD14 and MD-2 expression vectors were provided by Michael Smith (Univ. of Virginia). The TLR2 open reading frame was cloned into the KpnI and XbaI sites of pEF6/V5-HisTOPO (Invitrogen) to provide a V5 tag for detection. TLR4 was maintained in pCDNA3.1. The pSuperbased shRNA vectors for PKD1 and PKD2 (pSPKD1-1 and PKD2) were provided by P. Storz (Harvard). The PKD1-2 vector was created by annealing the following oligonucleotides and cloning them into the pSuper vector (Oligoengine, Seattle, WA): 50 -gatccccgcagattcaactgccataaacttcaagagagtttatggcagttgaatctgcttttta-30 , and 50 -agcttaaaaagcagattcaactgccataaactctcttgaagtttatggcagttgaatctgcggg-30 . All oligonucleotides were obtained from Operon (Huntsville, AL).
cose (HyClone; Logan UT) supplemented with 10% FBS (Invitrogen, Carlsbad, CA), nonessential amino acids (HyClone), penicillin (100 U/ml) and streptomycin (100 lg/ml) (Sigma, St. Louis, MD). Cells were seeded at 1 106 cells/ml in polystyrene culture dishes and used for experiments 7–14 days after becoming confluent. HeLa cells from ATCC were grown in Eagle’s minimal essential medium (StemCell, Vancouver, BC) supplemented with 10% heatinactivated FBS (HyClone), 2 mM glutamine (StemCell), 1 mM sodium pyruvate (StemCell) and antibiotics as above. Mode-K cells were a gift from Karen Madsen (Univ. of Alberta). They were grown in high-glucose DMEM with 5% fetal bovine serum, sodium pyruvate, pen/strep, and 15 mM HEPES. Caco-2 cells were passaged every 1 to 2 weeks, while HeLa and Mode-K cells were passaged 2–3 times per week. 2.3. Cell transfection
2.2. Cell culture Caco-2 cells were obtained from the American Type Culture Collection (ATCC, Rockville, MD) and grown in HyQ DMEM/High Glu-
For cytokine release assays, HeLa cells were seeded at 2 105 cells/ml in growth media without penicillin or streptomycin in 24-well plates. The following day, cells were transfected
Fig. 1. PKD1 and PKD2 knockdowns inhibit TLR5 signaling. (A) HeLa cells were transiently transfected with HA-tagged PKD1 and PKD2 along with the shRNA synthesis vector pSuper, with or without constructs targeting PKD1 or PKD2. After 48 h, cells were analyzed by WB with anti-PKD1 or PKD2, stripped and reprobed with anti-GAPDH as a loading control. Results representative of three experiments. (B) HeLa cells transfected with HA-tagged PKD1 and the shRNA constructs shown were analyzed by Western blot using antibodies against the autophosphorylation site Ser916 (left) or against total PKD2 (right), stripped, and reprobed with anti-GAPDH as a loading control. Results representative of three independent experiments. (C) HeLa cells transiently expressing TLR5 along with the pSuper constructs shown were stimulated 48 h later with FliC (100 ng/ml) for 3 h, and supernatants assayed for IL-8 by ELISA. *p < 0.05 vs. pSuper; **p < 0.005 vs. pSuper, PKD1-1 and PKD2.
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
using Lipofectamine LTX (Invitrogen) according to the manufacturer’s instructions, with the following plasmid amounts (per well): pEGFP 25 ng, TLR2/4/5 100 ng, and pSuper shRNA constructs, 500 ng. For TLR4 experiments, cells also received CD14 5 ng and MD-2 20 ng/well. For reporter assays, cells were seeded in 96-well plates and DNA amounts scaled down accordingly. Reporter constructs were used at 50 ng/well (CCL20) or 10 ng/well (NF-jB). Media was changed the following day, and cells were stimulated after a further 18 h. For Western blots, cells were seeded in 12-well plates and transfection conditions scaled up accordingly. Mode-K cells were seeded at 105/ml in 24-well plates and transfected the following day with 50 ng pEGFP and 900 ng of pSuper constructs per well, using Lipofectamine 2000 according to the manufacturer’s recommendations. Cells were stimulated the following day. 2.4. Stimulation and inhibition of cells The H18 flagellin from enteroaggregative Escherichia coli strain 042 was expressed in pCR-NT-T7-Topo (Invitrogen) in BL21 (DE3) pLysS cells and purified by metal affinity chromatography followed by polymyxin B chromatography as previously described [9]. Pharmacological inhibitors (Gö6976, Gö6983) were obtained from Calbiochem (San Diego). Gö6976 was used at a final concentration of 3 lM, and Gö6983 at 12 lM. FliC (flagellin) was used at a final concentration of 0.5–1 lg/ml, Pam3CSK4 (EMC microcollections, Tuebingen, Germany) was used at 1 lg/ml, and E. coli LPS (Sigma) was used at 100 ng/ml. Cells were pre-treated with inhibitors or DMSO vehicle for 1 h, followed by stimulation for varied amounts of time, depending on cell type and readout. Luciferase activity was measured in cells following 6 h of stimulation, using Bright-Glo reagent (Promega, Madison, WI). In these experiments, culture supernatants were removed at the time of cell harvest for chemokine testing. In all other experiments, supernatants were harvested after 3 h of agonist stimulation. IL-8, KC, and CCL20 were measured by ELISA (BD Biosciences for IL-8 and R&D Systems for KC and CCL20). For Western blots, cells were stimulated for the indicated times and lysed in phosphoprotein lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, 2.5 mM sodium pyrophosphate, 1 mM b-glycerophosphate, 2 mM Na3VO4, and protease inhibitor cocktail (Sigma)). Equal amounts of proteins were separated by SDS–PAGE and blotted for Western analysis. Densitometric analysis was done using Alpha Innotek software on at least three different experiments; density was adjusted relative to loading control and expressed as ratios to a fixed sample.
137
3. Results In previous work [3], we showed that inhibition of PKD with either a specific siRNA target sequence or the pharmacologic inhibitor Gö6976 reduced flagellin- and IL-1b-induced IL-8 production from Caco-2 and HEK 293T cells. Since both PKD1 and PKD2 are expressed in epithelial cells and have substantial sequence similarities, we measured the effects of our previously reported PKD1 siRNA (PKD1-1) and two additional constructs on relative expression of PKD1 and PKD2. We used HeLa cells for these experiments, as they are an epithelial cell line fully competent in TLR signaling with good transfection efficiency (routinely at least 50% by GFP co-transfection in each experiment; not shown). As shown in Fig. 1A, PKD1-1 knocked down expression of both PKD1 and PKD2, whereas a second construct, PKD1-2, was much more specific for PKD1 and caused a presumed compensatory increase in PKD2 expression. The PKD2 construct reduced only PKD2 expression, as predicted. We then measured expression of pSer916-PKD, which is an autophosphorylation site reflective of PKD kinase activity. As shown in Fig. 1B, the PKD1 shRNA constructs reduced PKD kinase activity, while the PKD2 shRNA increased it, despite a relatively modest reduction in total PKD2 amounts. This suggests that HeLa cells may be very sensitive to a relatively small decrease in PKD2 mRNA and upregulate PKD1 activity accordingly. Based on these findings, we compared the three shRNA constructs for their effects on inflammatory signaling in response to TLR5 stimulation. As shown in Fig. 1C, all three of the constructs inhibited IL-8 release from TLR5-transfected HeLa cells, although the PKD1-2 construct was significantly more inhibitory than the other two. These results suggest that while both PKD1 and PKD2 are involved in TLR5 signaling, PKD1 plays a larger role that cannot be fully substituted by PKD2. Because of the mixed knockdown of PKD1 and PKD2 with the PKD1-1 construct, and the greater potency of the PKD1-2 construct, we used the PKD1-2 construct in subsequent experiments. We previously reported that TLR5 has a consensus phosphorylation site for PKD, and that PKD could phosphorylate a synthetic TLR5 peptide containing this site in vitro. TLR4 has a similar
2.5. In vitro kinase assays The following synthetic peptides were purchased from GenScript (Piscataway, NJ) and dissolved at 3 mM in DMSO: PKD— AALVRQMSVAFFFK; TLR5—YQLMKHQSIRGFVQ; S805A—YQLMKHQAIRGFVQ; and TLR4—QQVELYRLLSRNTY. Peptides were added to a final concentration of 150 lM in a reaction mix of 0.03% Triton X-100, 20 mM HEPES (pH 7.4) with 25 ng of active recombinant PKD1 (Upstate, Charlottesville, VA). A kinase mixture containing 32 P-c-ATP (1 lCi/reaction) in 75 mM MgCl2, 500 lM ATP, 20 mM MOPS (pH 7.2), 25 mM b-glycerophosphate, 5 mM EGTA, 1 mM Na3VO4, and 1 mM dithiothreitol was added, and the reaction allowed to proceed for 10 min at 30 °C. The reaction was stopped by the addition of 100 ll 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.
Fig. 2. Phosphorylation of TLR4 consensus peptide by PKD1. Synthetic peptides corresponding to the predicted PKD target sites of TLR4 and TLR5 were subjected to in vitro kinase assay using recombinant PKD1 and 32P-ATP as described in Section 2. Radioactive incorporation in each experiment was expressed as a ratio to the CPM (minus background) of phosphorylated consensus peptide. Two negative controls included reactions with no peptide, and the TLR5 peptide with the target serine changed to alanine. *p < 0.05 vs. no peptide.
138
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
consensus site and a synthetic peptide containing this site is phosphorylated by recombinant PKD1 in vitro (Fig. 2). This raised the hypothesis that TLR4 signaling would also be dependent on PKD activity. Since Caco-2 cells do not have inflammatory responses to TLR4 stimulation due to absent CD14 and low TLR4 expression [10], we studied the effects of PKD in HeLa cells transiently expressing TLR4/MD-2/CD14 constructs. As shown in Fig. 3A, siRNA for PKD1 and PKD2 both significantly inhibited LPS-induced IL-8 and CCL20 production in these cells. While it is possible that direct phosphorylation of TLR4 and TLR5 is the major site of PKD activity in LPS and flagellin responses, the inhibition of IL-1 responses as well suggests that there are additional sites of PKD involvement in MyD88-dependent signaling pathways. This is further supported by reports that PKD1 colocalizes with TLR/IRAK/MyD88 complexes [4]. To examine this further, we studied the role of PKD1 in TLR2 signaling. TLR2 does not have a consensus PKD phosphorylation site in its TIR domain (not shown), making it a useful tool for this purpose. Using TLR2-transfected HeLa cells, we found that inhibition of PKD1 and PKD2 with shRNA reduced Pam3CSK4-induced IL-8 and CCL20 production to a similar extent as they reduced TLR5 and TLR4 signaling (Fig. 3B), suggesting that direct TLR phosphorylation is not required for the inflammatory activity of PKD.
In further support of a role for both PKD1 and PKD2 in TLR2, 4, and 5 signaling, we tested the effects of Gö6976 and Gö6983 on ligand-induced chemokine production. Gö6976 inhibits both classical PKCs and PKD1/2, while Gö6983 only inhibits PKC, providing a measure of the specific requirement for PKDs. As shown in Fig. 4A, Gö6976, but not Gö6983, significantly inhibited flagellin-induced IL-8 and CCL20 production from Caco-2 cells. The same pattern was evident in LPS-stimulated TLR4/MD-2/CD14-expressing HeLa cells (Fig. 4B), while both compounds inhibited TLR2-mediated Pam3CSK4-induced IL-8 production, suggesting a role for typical PKC isoforms as well as PKD in TLR2 signaling. While the effect of PKD inhibition in HeLa cells was clear, we decided to confirm the requirement for PKD activity in TLR2 signaling in an epithelial cell line that expresses native TLR2. Since human IECs are hyporesponsive to TLR2 and TLR4 agonists, we examined Mode-K murine IECs, using expression of KC (CXCL1) as a readout for TLR2 activity (since murine cells do not express IL-8). As shown in Supplementary Fig. S1, these cells were poorly responsive to shRNA inhibition, likely due to low transfection efficiency. Nevertheless, there was a small but statistically significant decrease in Pam3CSK4-induced KC expression with each of the PKD shRNA constructs. We were unable to confirm the requirement for PKD using Gö6983/Gö6976 since the former compound inhibited KC production.
Fig. 3. Chemokine release from HeLa cells after PKD knockdowns. HeLa cells were transfected with either TLR2 or TLR4/MD-2/CD14 as in Section 2, along with the pSuper shRNA constructs shown. Cells were stimulated with LPS 100 ng/ml (A) or Pam3CSK4 1 lg/ml (B) for 3 h, and supernatants assayed for IL-8 and CCL20 by ELISA. *p < 0.02; **p < 0.001, t-tests.
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
139
Fig. 4. Effects of Gö6976 on TLR-induced chemokine release. (A) Caco-2 cells were pre-treated with Gö6976 (3 lM) or Gö6983 (12 lM) for 30 min, followed by stimulation with flagellin 100 ng/ml for 3 h. IL-8 and CCL20 in supernatants were measured by ELISA. *p < 0.05 for each chemokine, t-test. (B) HeLa cells transiently transfected with TLR2 or TLR4/MD-2/CD14 were treated with Gö inhibitors as above, and supernatants analyzed for IL-8 and CCL20 after 3 h. *p < 0.005, t-test.
It has been reported that PKD1 inhibition reduced NF-jB and MAPK activation in mouse macrophages [5]. In contrast, we previously reported that Gö6976 inhibited TLR5-dependent p38 phosphorylation but not NF-jB activity in Caco-2 cells. To reconcile these results and pinpoint the effects of PKD1 and PKD2, we used HeLa cells transiently expressing luciferase reporters driven by NF-jB response elements (NF-jBpLuc) or the CCL20 promoter (CCL20pLuc), along with TLR2, TLR4/MD-2/CD14, or TLR5 and pSuper constructs. As shown in Fig. 5, both PKD1 and PKD2 shRNA significantly inhibited NF-jB activity in response to TLR2 and TLR4 ligands. Inhibition of NF-jB activity in flagellin-treated, TLR5expressing cells was less pronounced, and not statistically significant, consistent with our results reported previously using EMSA [3]. In contrast, CCL20 promoter activity following ligand stimulation was significantly increased for all three TLRs by inhibtion of PKD1, but not PKD2. In contrast, PKD2 shRNA significantly reduced TLR5-induced pCCL20 activity and had no significant effect following TLR2 or TLR4 stimulation. Finally, we examined phosphorylation of p38 and ERK MAP kinases in HeLa cells expressing TLR5 with pSuper constructs. Both kinases were activated beginning as early as 15 min after flagellin stimulation (Fig. 6). Neither of the PKD shRNA constructs significantly changed p38 or ERK activity. 4. Discussion This is the first report to show that PKD1 is involved in TLR2 or TLR4 signaling in epithelial cells. Only two publications have
examined the role of PKD in TLR4 signaling. The first, by Song et al., showed that LPS caused PKC-dependent PKD activation in neurons through autocrine release of IL-1b [11] The second, by Park et al., found that PKD1 inhibition (using shRNA) reduced TRAF-6 ubiquitination, NF-jB activation, and MAPK phosphorylation in response to MyD88-dependent TLR activation (TLR2, 5, 7, and 9) but not TLR3 activation in murine macrohpages. They found partial signaling impairment of LPS signaling, with delayed and reduced TRAF-6 ubiquitination and NF-jB activation, most likely due to residual signaling through TRIF and TRAF3. One additional report of PKD activity in TLR2 signaling was by Murphy et al., who found that stimulation of TLR2 or the Fc-e receptor in murine bone marrow-derived mast cells led to PKD1 activation, and ultimately to degranulation and increased MCP-1 mRNA, but that only the latter was sensitive to PKD1 inhibition [12]. Two other reports showed that PKD1 is involved in NF-jB activation in response to TLR-independent stimuli: loss of adherens junctions [13] or stimulation with cholecystokinin [14]. Our results support these findings to some extent. We previously showed that PKD inhibition with Gö6976 did not significantly reduce flagellin-induced NF-jB upregulation as measured by electrophoretic mobility shift assay [3], and in the present work we did not find a significant reduction in TLR5-dependent NF-jB activity by reporter assay when PKD1 or PKD2 were inhibited by shRNA. However, we found that both PKD knockdowns significantly reduced NF-jB activity in response to TLR2 and TLR4 ligands. In addition, production of both IL-8 and CCL20 at the protein level in response to TLR2, 4, or 5 activation was PKD
140
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
Fig. 5. Effects of PKD shRNA on promoter activity in HeLa cells. HeLa cells were transfected with pEGFP, TLR5, TLR2, or TLR4/MD-2/CD14 plus the pSuper shRNA constructs shown plus NF-jBpLuc (A) or CCL20pLuc (B). Cells were stimulated with the corresponding TLR ligands (FliC 100 ng/ml, Pam3CSK4 1 lg/ml, or LPS 100 ng/ml, respectively) for 6 h, and luciferase read by luminescent assay. Light output was divided by green fluorescence to control for transfection efficiency, and values were normalized to the expression level of untreated controls for each transfection condition. *p < 0.05 and **p < 0.005 vs. pSuper, t-test.
dependent. Since MAPK phosphorylation was not reduced by PKD1 or PKD2 shRNA, it is likely that PKD-independent signaling is more responsible for MAPK activation in HeLa cells, and that chemokine production remains sensitive to NF-jB activity. The major difference between our results and those of Park et al. is the requirement for PKD2. They found activation of PKD1 but not PKD2 or PKD3 by TLR agonists in murine macrophages, and consequently only tested the effects of PKD1 shRNA on downstream signaling. While we did not examine the activation state of PKD2, we found that PKD2 shRNA inhibited TLR-dependent signaling, although generally to a lesser extent than PKD1 knockdowns. This indicates that PKD2 does play a role in these pathways, and since both enzymes are inhibited by Gö6976, caution must be used in attributing an effect of this drug to PKD1 alone. The largest difference we observed between PKD1 and PKD2 shRNA was in activation of the CCL20 promoter. While CCL20 protein production was inhibited by PKD1 or PKD2 shRNA, the activity in the reporter assay was substantially increased by PKD1 shRNA. The reasons for this discrepancy are not clear, but they could reflect differences between activation of native genes within nuclear chromatin versus exogenous, plasmid-borne genes. It is also possible that the compensatory overexpression of PKD2 seen with PKD1 shRNA could contribute to this effect. Clearly the net effect of PKD
inhibition on CCL20 production is anti-inflammatory, as shown by ELISA on culture supernatants (in this work) as well as QPCR to detect CCL20 mRNA [3]. The compensatory increases we observed in PKD1 activity after PKD2 shRNA and in PKD2 expression after PKD1 shRNA are not by themselves surprising, as similar compensatory changes are frequently observed with other enzyme families with overlapping targets. It is possible that knocking down one PKD isoenzyme could lead to alterations in the other isoform that are independent of expression levels (such as subcellular localization). What is informative about our results is that these compensatory increases did not affect inflammatory signaling. This suggests that PKD1 and PKD2 activity is necessary but not sufficient for TLR-mediated chemokine production. In summary, we have confirmed that PKD1 is involved not only in TLR5 signaling, but also in the inflammatory responses downstream of TLR2 and TLR4. Moreover, we have found for the first time that PKD2 is involved in TLR signaling pathways, although generally with a smaller effect than PKD1. These observations, in light of the fact that TLR4 and TLR5 may be directly phosphorylated by PKD but TLR2 lacks a consensus site, suggest that it is not TLR phosphorylation that is primarily responsible for the effects of PKD. Other phosphorylation targets within the proximal TLR
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
141
Fig. 6. PKD knockdown does not affect MAP kinase phosphorylation. HeLa cells were transfected with V5-TLR5 plus the pSuper shRNA constructs shown. Cells were stimulated with flagellin 100 ng/ml for the indicated times, and cellular lysates analyzed by Western blot for phospho-p38 and phospho-ERK. Blots were stripped and reprobed with anti-V5 to control for TLR5 expression, as well as anti-total p38 (not shown). (A) Blots representative of 3 experiments. (B,C) Pooled densitometry values of phospho-p38 (B) and phospho-ERK (C) divided by loading control, normalized to the lane containing unstimulated cells within each experiment, from three independent experiments.
142
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
signaling complex have not been reported, but kinase-independent activities of PKD could also play a role. Future work will be needed to uncover the details of this signaling. Acknowledgments This work was supported by an operating grant and New Investigator award (to T.S.S.) from the Canadian Institutes for Health Research, as well as an In It For Life Scholar award from the Vancouver Coastal Health Research Institute (to T.S.S.). Y.Y. is supported by a CIHR Transplantation Research Award. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cellimm.2010.05.012. References [1] E. Rozengurt, O. Rey, R.T. Waldron, Protein kinase D signaling, J. Biol. Chem. 14 (2005) 13205–13208. [2] M. Jaggi, C. Du, W. Zhang, K.C. Balaji, Protein kinase D1: a protein of emerging translational interest, Front. Biosci. (2007) 3757–3767. [3] S.M. Ivison, N.R. Graham, C.Q. Bernales, A. Kifayet, N. Ng, L.A. Shobab, T.S. Steiner, Protein kinase D interaction with TLR5 is required for inflammatory signaling in response to bacterial flagellin, J. Immunol. 9 (2007) 5735–5743. [4] J.E. Park, Y.I. Kim, A.K. Yi, Protein kinase D1: a new component in TLR9 signaling, J. Immunol. 3 (2008) 2044–2055.
[5] J.E. Park, Y.I. Kim, A.K. Yi, Protein kinase D1 is essential for MyD88-dependent TLR signaling pathway, J. Immunol. 10 (2009) 6316–6327. [6] Q. Hao, L. Wang, H. Tang, Vascular endothelial growth factor induces protein kinase D-dependent production of pro-inflammatory cytokines in endothelial cells, Am. J. Physiol. Cell Physiol. (2009). [7] C.H. Ha, W. Wang, B.S. Jhun, C. Wong, A. Hausser, K. Pfizenmaier, T.A. McKinsey, E.N. Olson, Z.G. Jin, Protein kinase D-dependent phosphorylation and nuclear export of histone deacetylase 5 mediates vascular endothelial growth factor-induced gene expression and angiogenesis, J. Biol. Chem. 21 (2008) 14590–14599. [8] J.H. Kwon, S. Keates, S. Simeonidis, F. Grall, T.A. Libermann, A.C. Keates, ESE-1, an enterocyte-specific Ets transcription factor, regulates MIP-3alpha gene expression in Caco-2 human colonic epithelial cells, J. Biol. Chem. 2 (2003) 875–884. [9] M.A. Donnelly, T.S. Steiner, Two nonadjacent regions in enteroaggregative Escherichia coli flagellin are required for activation of toll-like receptor 5, J. Biol. Chem. 43 (2002) 40456–40461. [10] N. Simiantonaki, U. Kurzik-Dumke, G. Karyofylli, C. Jayasinghe, R. MichelSchmidt, C.J. Kirkpatrick, Reduced expression of TLR4 is associated with the metastatic status of human colorectal cancer, Int. J. Mol. Med. 1 (2007) 21–29. [11] M.J. Song, Y.Q. Wang, G.C. Wu, Lipopolysaccharide-induced protein kinase D activation mediated by interleukin-1beta and protein kinase C, Brain Res. (2007) 19–27. [12] T.R. Murphy, H.J. Legere 3rd, H.R. Katz, Activation of protein kinase D1 in mast cells in response to innate, adaptive, and growth factor signals, J. Immunol. 11 (2007) 7876–7882. [13] C.F. Cowell, I.K. Yan, T. Eiseler, A.C. Leightner, H. Doppler, P. Storz, Loss of cell– cell contacts induces NF-kappaB via RhoA-mediated activation of protein kinase D1, J. Cell. Biochem. 106 (2009) 714–728. [14] J. Yuan, A. Lugea, L. Zheng, I. Gukovsky, M. Edderkaoui, E. Rozengurt, S.J. Pandol, Protein kinase D1 mediates NF-kappaB activation induced by cholecystokinin and cholinergic signaling in pancreatic acinar cells, Am. J. Physiol. Gastrointest. Liver Physiol. 6 (2008) G1190–G1201.
Contents lists available at ScienceDirect
Cellular Immunology journal homepage: www.elsevier.com/locate/ycimm
Protein kinase D1 and D2 are involved in chemokine release induced by toll-like receptors 2, 4, and 5 Theodore S. Steiner *, Sabine M. Ivison, Yu Yao, Arnawaz Kifayet Division of Infectious Diseases, Department of Medicine, University of British Columbia, Rm. D452 HP East, VGH, 2733 Heather St., Vancouver, BC, Canada V5Z 3J5
a r t i c l e
i n f o
Article history: Received 16 December 2009 Accepted 22 May 2010 Available online 27 May 2010 Keywords: Protein kinase D Innate immunity Toll-like receptors Chemokines
a b s t r a c t The protein kinase D (PKD) family consists of three serine-threonine kinases involved in cellular proliferation, motility, and apoptosis. We previously reported that human toll-like receptor 5 (TLR5) contains a consensus PKD phosphorylation site. Flagellin stimulation of cells activated PKD1, and inhibition of PKD1 reduced flagellin-induced interleukin-8 (IL-8) production in epithelial cells. In the current work, we examined PKD1 and PKD2 involvement downstream of TLR5, TLR4 and TLR2. We found that inhibition of either kinase with shRNA reduced IL-8 and CCL20 release due to TLR4 and TLR2 agonists to a similar extent as previously reported for TLR5. PKD1 and PKD2 inhibition reduced NF-jB activity but not MAPK activation. These results demonstrate that both PKD1 and PKD2 are required for inflammatory responses following TLR2, TLR4, or TLR5 activation, although PKD1 is more strongly involved. These kinases likely act downstream of the TLRs themselves to facilitate NF-jB activation but not MAP kinase phosphorylation. Ó 2010 Elsevier Inc. All rights reserved.
1. Background Protein kinase D (PKD) refers to three related serine-threonine kinases of the calmodulin-calcium dependent (CaM)-kinase family. The three known PKD subtypes PKD1 (PKCl), PKD2, and PKD3(PKCm) are involved in diverse aspects of cellular function including membrane trafficking, differentiation, proliferation and apoptosis, as well as the responses to B-and T-cell receptor ligation, regulatory peptides and oxidative stress (for reviews see [1,2]). We previously reported that toll-like receptor 5 (TLR5), the pattern recognition receptor for bacterial flagellin, contains a consensus PKD phosphorylation site in its C-terminal TIR domain [3]. Based on this observation, we found that PKD activity was increased following flagellin stimulation of TLR5, and that the predicted TLR5 phosphorylation site was, in fact, capable of being phosphorylated by PKD1. Moreover, we showed that reduction of PKD activity using the pharmacological inhibitor Gö6976 or siRNA specific for PKD1 was required for interleukin (IL)-8 production from cells stimulated with flagellin. Subsequent to this work, Park et al. [4] reported that PKD is also activated in human and murine macrophages by interaction of TLR9 with its natural ligand, unmethylated CpG oligodeoxynucleotide (ODN). They later reported that PKD1 in murine pacrophages is activated by all of the MyD88-dependent TLR ligands, and is re-
* Corresponding author. Fax: +1 604 875 4013. E-mail addresses: [email protected] (T.S. Steiner), sabine.ivison@ gmail.com (S.M. Ivison), [email protected] (A. Kifayet). 0008-8749/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.cellimm.2010.05.012
quired for TRAF-6 ubiquitination and downstream signaling [5]. Interestingly, they found no activation of or requirement for PKD2 or PKD3 in this pathway. In addition, several groups have recently reported that PKD is required for cytokine production in endothelial cells exposed to vascular endothelial growth factor (VEGF) [6,7]. Because of this emerging evidence that PKD1 is involved in inflammatory signaling in hemopoietic and endothelial cells, we hypothesized that this enzyme might play a more widespread role in TLR signaling pathways in epithelial cells as well. We tested the role of PKD1 and PKD2 on production of two epithelial-derived chemokines, IL-8 (CXCL8) and CCL20 (MIP3a), in response to agonists of TLR2, TLR4, and TLR5. As discussed below, we found that both PKD1 and PKD2 are required for production of these chemokines in epithelial cells, primarily through the NF-jB pathway. 2. Methods 2.1. Antibodies and vectors The following antibodies were used: anti-PK (V5 tag, Serotec, Oxford, UK), anti-phospho-p38, anti-phospho-ERK, anti-PKD1, anti-pSer916-PKD, and PKD2 (Cell Signaling, Cambridge, MA), and anti-GAPDH (Fitzgerald, Concord, MA). The human TLR5 open reading frame cloned into pEF6-V5His-Topo (Invitrogen) was a gift from A. Aderem (Institute for Systems Biology, Seattle). pEGFP-N1 was from BD Clontech (Mountain View, CA). HA-PKD1 and and HAPKD2 were provided by A. Toker (Harvard University). The
136
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
CCL20pLuc reporter vector [8] was provided by A. Keates (Harvard University). The NF-jBpLuc reporter was provided by Bruce Vallance (University of British Columbia). TLR2, TLR4, and CD14 and MD-2 expression vectors were provided by Michael Smith (Univ. of Virginia). The TLR2 open reading frame was cloned into the KpnI and XbaI sites of pEF6/V5-HisTOPO (Invitrogen) to provide a V5 tag for detection. TLR4 was maintained in pCDNA3.1. The pSuperbased shRNA vectors for PKD1 and PKD2 (pSPKD1-1 and PKD2) were provided by P. Storz (Harvard). The PKD1-2 vector was created by annealing the following oligonucleotides and cloning them into the pSuper vector (Oligoengine, Seattle, WA): 50 -gatccccgcagattcaactgccataaacttcaagagagtttatggcagttgaatctgcttttta-30 , and 50 -agcttaaaaagcagattcaactgccataaactctcttgaagtttatggcagttgaatctgcggg-30 . All oligonucleotides were obtained from Operon (Huntsville, AL).
cose (HyClone; Logan UT) supplemented with 10% FBS (Invitrogen, Carlsbad, CA), nonessential amino acids (HyClone), penicillin (100 U/ml) and streptomycin (100 lg/ml) (Sigma, St. Louis, MD). Cells were seeded at 1 106 cells/ml in polystyrene culture dishes and used for experiments 7–14 days after becoming confluent. HeLa cells from ATCC were grown in Eagle’s minimal essential medium (StemCell, Vancouver, BC) supplemented with 10% heatinactivated FBS (HyClone), 2 mM glutamine (StemCell), 1 mM sodium pyruvate (StemCell) and antibiotics as above. Mode-K cells were a gift from Karen Madsen (Univ. of Alberta). They were grown in high-glucose DMEM with 5% fetal bovine serum, sodium pyruvate, pen/strep, and 15 mM HEPES. Caco-2 cells were passaged every 1 to 2 weeks, while HeLa and Mode-K cells were passaged 2–3 times per week. 2.3. Cell transfection
2.2. Cell culture Caco-2 cells were obtained from the American Type Culture Collection (ATCC, Rockville, MD) and grown in HyQ DMEM/High Glu-
For cytokine release assays, HeLa cells were seeded at 2 105 cells/ml in growth media without penicillin or streptomycin in 24-well plates. The following day, cells were transfected
Fig. 1. PKD1 and PKD2 knockdowns inhibit TLR5 signaling. (A) HeLa cells were transiently transfected with HA-tagged PKD1 and PKD2 along with the shRNA synthesis vector pSuper, with or without constructs targeting PKD1 or PKD2. After 48 h, cells were analyzed by WB with anti-PKD1 or PKD2, stripped and reprobed with anti-GAPDH as a loading control. Results representative of three experiments. (B) HeLa cells transfected with HA-tagged PKD1 and the shRNA constructs shown were analyzed by Western blot using antibodies against the autophosphorylation site Ser916 (left) or against total PKD2 (right), stripped, and reprobed with anti-GAPDH as a loading control. Results representative of three independent experiments. (C) HeLa cells transiently expressing TLR5 along with the pSuper constructs shown were stimulated 48 h later with FliC (100 ng/ml) for 3 h, and supernatants assayed for IL-8 by ELISA. *p < 0.05 vs. pSuper; **p < 0.005 vs. pSuper, PKD1-1 and PKD2.
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
using Lipofectamine LTX (Invitrogen) according to the manufacturer’s instructions, with the following plasmid amounts (per well): pEGFP 25 ng, TLR2/4/5 100 ng, and pSuper shRNA constructs, 500 ng. For TLR4 experiments, cells also received CD14 5 ng and MD-2 20 ng/well. For reporter assays, cells were seeded in 96-well plates and DNA amounts scaled down accordingly. Reporter constructs were used at 50 ng/well (CCL20) or 10 ng/well (NF-jB). Media was changed the following day, and cells were stimulated after a further 18 h. For Western blots, cells were seeded in 12-well plates and transfection conditions scaled up accordingly. Mode-K cells were seeded at 105/ml in 24-well plates and transfected the following day with 50 ng pEGFP and 900 ng of pSuper constructs per well, using Lipofectamine 2000 according to the manufacturer’s recommendations. Cells were stimulated the following day. 2.4. Stimulation and inhibition of cells The H18 flagellin from enteroaggregative Escherichia coli strain 042 was expressed in pCR-NT-T7-Topo (Invitrogen) in BL21 (DE3) pLysS cells and purified by metal affinity chromatography followed by polymyxin B chromatography as previously described [9]. Pharmacological inhibitors (Gö6976, Gö6983) were obtained from Calbiochem (San Diego). Gö6976 was used at a final concentration of 3 lM, and Gö6983 at 12 lM. FliC (flagellin) was used at a final concentration of 0.5–1 lg/ml, Pam3CSK4 (EMC microcollections, Tuebingen, Germany) was used at 1 lg/ml, and E. coli LPS (Sigma) was used at 100 ng/ml. Cells were pre-treated with inhibitors or DMSO vehicle for 1 h, followed by stimulation for varied amounts of time, depending on cell type and readout. Luciferase activity was measured in cells following 6 h of stimulation, using Bright-Glo reagent (Promega, Madison, WI). In these experiments, culture supernatants were removed at the time of cell harvest for chemokine testing. In all other experiments, supernatants were harvested after 3 h of agonist stimulation. IL-8, KC, and CCL20 were measured by ELISA (BD Biosciences for IL-8 and R&D Systems for KC and CCL20). For Western blots, cells were stimulated for the indicated times and lysed in phosphoprotein lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, 2.5 mM sodium pyrophosphate, 1 mM b-glycerophosphate, 2 mM Na3VO4, and protease inhibitor cocktail (Sigma)). Equal amounts of proteins were separated by SDS–PAGE and blotted for Western analysis. Densitometric analysis was done using Alpha Innotek software on at least three different experiments; density was adjusted relative to loading control and expressed as ratios to a fixed sample.
137
3. Results In previous work [3], we showed that inhibition of PKD with either a specific siRNA target sequence or the pharmacologic inhibitor Gö6976 reduced flagellin- and IL-1b-induced IL-8 production from Caco-2 and HEK 293T cells. Since both PKD1 and PKD2 are expressed in epithelial cells and have substantial sequence similarities, we measured the effects of our previously reported PKD1 siRNA (PKD1-1) and two additional constructs on relative expression of PKD1 and PKD2. We used HeLa cells for these experiments, as they are an epithelial cell line fully competent in TLR signaling with good transfection efficiency (routinely at least 50% by GFP co-transfection in each experiment; not shown). As shown in Fig. 1A, PKD1-1 knocked down expression of both PKD1 and PKD2, whereas a second construct, PKD1-2, was much more specific for PKD1 and caused a presumed compensatory increase in PKD2 expression. The PKD2 construct reduced only PKD2 expression, as predicted. We then measured expression of pSer916-PKD, which is an autophosphorylation site reflective of PKD kinase activity. As shown in Fig. 1B, the PKD1 shRNA constructs reduced PKD kinase activity, while the PKD2 shRNA increased it, despite a relatively modest reduction in total PKD2 amounts. This suggests that HeLa cells may be very sensitive to a relatively small decrease in PKD2 mRNA and upregulate PKD1 activity accordingly. Based on these findings, we compared the three shRNA constructs for their effects on inflammatory signaling in response to TLR5 stimulation. As shown in Fig. 1C, all three of the constructs inhibited IL-8 release from TLR5-transfected HeLa cells, although the PKD1-2 construct was significantly more inhibitory than the other two. These results suggest that while both PKD1 and PKD2 are involved in TLR5 signaling, PKD1 plays a larger role that cannot be fully substituted by PKD2. Because of the mixed knockdown of PKD1 and PKD2 with the PKD1-1 construct, and the greater potency of the PKD1-2 construct, we used the PKD1-2 construct in subsequent experiments. We previously reported that TLR5 has a consensus phosphorylation site for PKD, and that PKD could phosphorylate a synthetic TLR5 peptide containing this site in vitro. TLR4 has a similar
2.5. In vitro kinase assays The following synthetic peptides were purchased from GenScript (Piscataway, NJ) and dissolved at 3 mM in DMSO: PKD— AALVRQMSVAFFFK; TLR5—YQLMKHQSIRGFVQ; S805A—YQLMKHQAIRGFVQ; and TLR4—QQVELYRLLSRNTY. Peptides were added to a final concentration of 150 lM in a reaction mix of 0.03% Triton X-100, 20 mM HEPES (pH 7.4) with 25 ng of active recombinant PKD1 (Upstate, Charlottesville, VA). A kinase mixture containing 32 P-c-ATP (1 lCi/reaction) in 75 mM MgCl2, 500 lM ATP, 20 mM MOPS (pH 7.2), 25 mM b-glycerophosphate, 5 mM EGTA, 1 mM Na3VO4, and 1 mM dithiothreitol was added, and the reaction allowed to proceed for 10 min at 30 °C. The reaction was stopped by the addition of 100 ll 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.
Fig. 2. Phosphorylation of TLR4 consensus peptide by PKD1. Synthetic peptides corresponding to the predicted PKD target sites of TLR4 and TLR5 were subjected to in vitro kinase assay using recombinant PKD1 and 32P-ATP as described in Section 2. Radioactive incorporation in each experiment was expressed as a ratio to the CPM (minus background) of phosphorylated consensus peptide. Two negative controls included reactions with no peptide, and the TLR5 peptide with the target serine changed to alanine. *p < 0.05 vs. no peptide.
138
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
consensus site and a synthetic peptide containing this site is phosphorylated by recombinant PKD1 in vitro (Fig. 2). This raised the hypothesis that TLR4 signaling would also be dependent on PKD activity. Since Caco-2 cells do not have inflammatory responses to TLR4 stimulation due to absent CD14 and low TLR4 expression [10], we studied the effects of PKD in HeLa cells transiently expressing TLR4/MD-2/CD14 constructs. As shown in Fig. 3A, siRNA for PKD1 and PKD2 both significantly inhibited LPS-induced IL-8 and CCL20 production in these cells. While it is possible that direct phosphorylation of TLR4 and TLR5 is the major site of PKD activity in LPS and flagellin responses, the inhibition of IL-1 responses as well suggests that there are additional sites of PKD involvement in MyD88-dependent signaling pathways. This is further supported by reports that PKD1 colocalizes with TLR/IRAK/MyD88 complexes [4]. To examine this further, we studied the role of PKD1 in TLR2 signaling. TLR2 does not have a consensus PKD phosphorylation site in its TIR domain (not shown), making it a useful tool for this purpose. Using TLR2-transfected HeLa cells, we found that inhibition of PKD1 and PKD2 with shRNA reduced Pam3CSK4-induced IL-8 and CCL20 production to a similar extent as they reduced TLR5 and TLR4 signaling (Fig. 3B), suggesting that direct TLR phosphorylation is not required for the inflammatory activity of PKD.
In further support of a role for both PKD1 and PKD2 in TLR2, 4, and 5 signaling, we tested the effects of Gö6976 and Gö6983 on ligand-induced chemokine production. Gö6976 inhibits both classical PKCs and PKD1/2, while Gö6983 only inhibits PKC, providing a measure of the specific requirement for PKDs. As shown in Fig. 4A, Gö6976, but not Gö6983, significantly inhibited flagellin-induced IL-8 and CCL20 production from Caco-2 cells. The same pattern was evident in LPS-stimulated TLR4/MD-2/CD14-expressing HeLa cells (Fig. 4B), while both compounds inhibited TLR2-mediated Pam3CSK4-induced IL-8 production, suggesting a role for typical PKC isoforms as well as PKD in TLR2 signaling. While the effect of PKD inhibition in HeLa cells was clear, we decided to confirm the requirement for PKD activity in TLR2 signaling in an epithelial cell line that expresses native TLR2. Since human IECs are hyporesponsive to TLR2 and TLR4 agonists, we examined Mode-K murine IECs, using expression of KC (CXCL1) as a readout for TLR2 activity (since murine cells do not express IL-8). As shown in Supplementary Fig. S1, these cells were poorly responsive to shRNA inhibition, likely due to low transfection efficiency. Nevertheless, there was a small but statistically significant decrease in Pam3CSK4-induced KC expression with each of the PKD shRNA constructs. We were unable to confirm the requirement for PKD using Gö6983/Gö6976 since the former compound inhibited KC production.
Fig. 3. Chemokine release from HeLa cells after PKD knockdowns. HeLa cells were transfected with either TLR2 or TLR4/MD-2/CD14 as in Section 2, along with the pSuper shRNA constructs shown. Cells were stimulated with LPS 100 ng/ml (A) or Pam3CSK4 1 lg/ml (B) for 3 h, and supernatants assayed for IL-8 and CCL20 by ELISA. *p < 0.02; **p < 0.001, t-tests.
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
139
Fig. 4. Effects of Gö6976 on TLR-induced chemokine release. (A) Caco-2 cells were pre-treated with Gö6976 (3 lM) or Gö6983 (12 lM) for 30 min, followed by stimulation with flagellin 100 ng/ml for 3 h. IL-8 and CCL20 in supernatants were measured by ELISA. *p < 0.05 for each chemokine, t-test. (B) HeLa cells transiently transfected with TLR2 or TLR4/MD-2/CD14 were treated with Gö inhibitors as above, and supernatants analyzed for IL-8 and CCL20 after 3 h. *p < 0.005, t-test.
It has been reported that PKD1 inhibition reduced NF-jB and MAPK activation in mouse macrophages [5]. In contrast, we previously reported that Gö6976 inhibited TLR5-dependent p38 phosphorylation but not NF-jB activity in Caco-2 cells. To reconcile these results and pinpoint the effects of PKD1 and PKD2, we used HeLa cells transiently expressing luciferase reporters driven by NF-jB response elements (NF-jBpLuc) or the CCL20 promoter (CCL20pLuc), along with TLR2, TLR4/MD-2/CD14, or TLR5 and pSuper constructs. As shown in Fig. 5, both PKD1 and PKD2 shRNA significantly inhibited NF-jB activity in response to TLR2 and TLR4 ligands. Inhibition of NF-jB activity in flagellin-treated, TLR5expressing cells was less pronounced, and not statistically significant, consistent with our results reported previously using EMSA [3]. In contrast, CCL20 promoter activity following ligand stimulation was significantly increased for all three TLRs by inhibtion of PKD1, but not PKD2. In contrast, PKD2 shRNA significantly reduced TLR5-induced pCCL20 activity and had no significant effect following TLR2 or TLR4 stimulation. Finally, we examined phosphorylation of p38 and ERK MAP kinases in HeLa cells expressing TLR5 with pSuper constructs. Both kinases were activated beginning as early as 15 min after flagellin stimulation (Fig. 6). Neither of the PKD shRNA constructs significantly changed p38 or ERK activity. 4. Discussion This is the first report to show that PKD1 is involved in TLR2 or TLR4 signaling in epithelial cells. Only two publications have
examined the role of PKD in TLR4 signaling. The first, by Song et al., showed that LPS caused PKC-dependent PKD activation in neurons through autocrine release of IL-1b [11] The second, by Park et al., found that PKD1 inhibition (using shRNA) reduced TRAF-6 ubiquitination, NF-jB activation, and MAPK phosphorylation in response to MyD88-dependent TLR activation (TLR2, 5, 7, and 9) but not TLR3 activation in murine macrohpages. They found partial signaling impairment of LPS signaling, with delayed and reduced TRAF-6 ubiquitination and NF-jB activation, most likely due to residual signaling through TRIF and TRAF3. One additional report of PKD activity in TLR2 signaling was by Murphy et al., who found that stimulation of TLR2 or the Fc-e receptor in murine bone marrow-derived mast cells led to PKD1 activation, and ultimately to degranulation and increased MCP-1 mRNA, but that only the latter was sensitive to PKD1 inhibition [12]. Two other reports showed that PKD1 is involved in NF-jB activation in response to TLR-independent stimuli: loss of adherens junctions [13] or stimulation with cholecystokinin [14]. Our results support these findings to some extent. We previously showed that PKD inhibition with Gö6976 did not significantly reduce flagellin-induced NF-jB upregulation as measured by electrophoretic mobility shift assay [3], and in the present work we did not find a significant reduction in TLR5-dependent NF-jB activity by reporter assay when PKD1 or PKD2 were inhibited by shRNA. However, we found that both PKD knockdowns significantly reduced NF-jB activity in response to TLR2 and TLR4 ligands. In addition, production of both IL-8 and CCL20 at the protein level in response to TLR2, 4, or 5 activation was PKD
140
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
Fig. 5. Effects of PKD shRNA on promoter activity in HeLa cells. HeLa cells were transfected with pEGFP, TLR5, TLR2, or TLR4/MD-2/CD14 plus the pSuper shRNA constructs shown plus NF-jBpLuc (A) or CCL20pLuc (B). Cells were stimulated with the corresponding TLR ligands (FliC 100 ng/ml, Pam3CSK4 1 lg/ml, or LPS 100 ng/ml, respectively) for 6 h, and luciferase read by luminescent assay. Light output was divided by green fluorescence to control for transfection efficiency, and values were normalized to the expression level of untreated controls for each transfection condition. *p < 0.05 and **p < 0.005 vs. pSuper, t-test.
dependent. Since MAPK phosphorylation was not reduced by PKD1 or PKD2 shRNA, it is likely that PKD-independent signaling is more responsible for MAPK activation in HeLa cells, and that chemokine production remains sensitive to NF-jB activity. The major difference between our results and those of Park et al. is the requirement for PKD2. They found activation of PKD1 but not PKD2 or PKD3 by TLR agonists in murine macrophages, and consequently only tested the effects of PKD1 shRNA on downstream signaling. While we did not examine the activation state of PKD2, we found that PKD2 shRNA inhibited TLR-dependent signaling, although generally to a lesser extent than PKD1 knockdowns. This indicates that PKD2 does play a role in these pathways, and since both enzymes are inhibited by Gö6976, caution must be used in attributing an effect of this drug to PKD1 alone. The largest difference we observed between PKD1 and PKD2 shRNA was in activation of the CCL20 promoter. While CCL20 protein production was inhibited by PKD1 or PKD2 shRNA, the activity in the reporter assay was substantially increased by PKD1 shRNA. The reasons for this discrepancy are not clear, but they could reflect differences between activation of native genes within nuclear chromatin versus exogenous, plasmid-borne genes. It is also possible that the compensatory overexpression of PKD2 seen with PKD1 shRNA could contribute to this effect. Clearly the net effect of PKD
inhibition on CCL20 production is anti-inflammatory, as shown by ELISA on culture supernatants (in this work) as well as QPCR to detect CCL20 mRNA [3]. The compensatory increases we observed in PKD1 activity after PKD2 shRNA and in PKD2 expression after PKD1 shRNA are not by themselves surprising, as similar compensatory changes are frequently observed with other enzyme families with overlapping targets. It is possible that knocking down one PKD isoenzyme could lead to alterations in the other isoform that are independent of expression levels (such as subcellular localization). What is informative about our results is that these compensatory increases did not affect inflammatory signaling. This suggests that PKD1 and PKD2 activity is necessary but not sufficient for TLR-mediated chemokine production. In summary, we have confirmed that PKD1 is involved not only in TLR5 signaling, but also in the inflammatory responses downstream of TLR2 and TLR4. Moreover, we have found for the first time that PKD2 is involved in TLR signaling pathways, although generally with a smaller effect than PKD1. These observations, in light of the fact that TLR4 and TLR5 may be directly phosphorylated by PKD but TLR2 lacks a consensus site, suggest that it is not TLR phosphorylation that is primarily responsible for the effects of PKD. Other phosphorylation targets within the proximal TLR
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
141
Fig. 6. PKD knockdown does not affect MAP kinase phosphorylation. HeLa cells were transfected with V5-TLR5 plus the pSuper shRNA constructs shown. Cells were stimulated with flagellin 100 ng/ml for the indicated times, and cellular lysates analyzed by Western blot for phospho-p38 and phospho-ERK. Blots were stripped and reprobed with anti-V5 to control for TLR5 expression, as well as anti-total p38 (not shown). (A) Blots representative of 3 experiments. (B,C) Pooled densitometry values of phospho-p38 (B) and phospho-ERK (C) divided by loading control, normalized to the lane containing unstimulated cells within each experiment, from three independent experiments.
142
T.S. Steiner et al. / Cellular Immunology 264 (2010) 135–142
signaling complex have not been reported, but kinase-independent activities of PKD could also play a role. Future work will be needed to uncover the details of this signaling. Acknowledgments This work was supported by an operating grant and New Investigator award (to T.S.S.) from the Canadian Institutes for Health Research, as well as an In It For Life Scholar award from the Vancouver Coastal Health Research Institute (to T.S.S.). Y.Y. is supported by a CIHR Transplantation Research Award. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cellimm.2010.05.012. References [1] E. Rozengurt, O. Rey, R.T. Waldron, Protein kinase D signaling, J. Biol. Chem. 14 (2005) 13205–13208. [2] M. Jaggi, C. Du, W. Zhang, K.C. Balaji, Protein kinase D1: a protein of emerging translational interest, Front. Biosci. (2007) 3757–3767. [3] S.M. Ivison, N.R. Graham, C.Q. Bernales, A. Kifayet, N. Ng, L.A. Shobab, T.S. Steiner, Protein kinase D interaction with TLR5 is required for inflammatory signaling in response to bacterial flagellin, J. Immunol. 9 (2007) 5735–5743. [4] J.E. Park, Y.I. Kim, A.K. Yi, Protein kinase D1: a new component in TLR9 signaling, J. Immunol. 3 (2008) 2044–2055.
[5] J.E. Park, Y.I. Kim, A.K. Yi, Protein kinase D1 is essential for MyD88-dependent TLR signaling pathway, J. Immunol. 10 (2009) 6316–6327. [6] Q. Hao, L. Wang, H. Tang, Vascular endothelial growth factor induces protein kinase D-dependent production of pro-inflammatory cytokines in endothelial cells, Am. J. Physiol. Cell Physiol. (2009). [7] C.H. Ha, W. Wang, B.S. Jhun, C. Wong, A. Hausser, K. Pfizenmaier, T.A. McKinsey, E.N. Olson, Z.G. Jin, Protein kinase D-dependent phosphorylation and nuclear export of histone deacetylase 5 mediates vascular endothelial growth factor-induced gene expression and angiogenesis, J. Biol. Chem. 21 (2008) 14590–14599. [8] J.H. Kwon, S. Keates, S. Simeonidis, F. Grall, T.A. Libermann, A.C. Keates, ESE-1, an enterocyte-specific Ets transcription factor, regulates MIP-3alpha gene expression in Caco-2 human colonic epithelial cells, J. Biol. Chem. 2 (2003) 875–884. [9] M.A. Donnelly, T.S. Steiner, Two nonadjacent regions in enteroaggregative Escherichia coli flagellin are required for activation of toll-like receptor 5, J. Biol. Chem. 43 (2002) 40456–40461. [10] N. Simiantonaki, U. Kurzik-Dumke, G. Karyofylli, C. Jayasinghe, R. MichelSchmidt, C.J. Kirkpatrick, Reduced expression of TLR4 is associated with the metastatic status of human colorectal cancer, Int. J. Mol. Med. 1 (2007) 21–29. [11] M.J. Song, Y.Q. Wang, G.C. Wu, Lipopolysaccharide-induced protein kinase D activation mediated by interleukin-1beta and protein kinase C, Brain Res. (2007) 19–27. [12] T.R. Murphy, H.J. Legere 3rd, H.R. Katz, Activation of protein kinase D1 in mast cells in response to innate, adaptive, and growth factor signals, J. Immunol. 11 (2007) 7876–7882. [13] C.F. Cowell, I.K. Yan, T. Eiseler, A.C. Leightner, H. Doppler, P. Storz, Loss of cell– cell contacts induces NF-kappaB via RhoA-mediated activation of protein kinase D1, J. Cell. Biochem. 106 (2009) 714–728. [14] J. Yuan, A. Lugea, L. Zheng, I. Gukovsky, M. Edderkaoui, E. Rozengurt, S.J. Pandol, Protein kinase D1 mediates NF-kappaB activation induced by cholecystokinin and cholinergic signaling in pancreatic acinar cells, Am. J. Physiol. Gastrointest. Liver Physiol. 6 (2008) G1190–G1201.