IL-17-induced NF-κB Activation via CIKS/Act1

Abstract

Interleukin-17 (IL-17) is essential in host defense against extracellular bacteria and fungi, especially at mucosal sites, but it also contributes significantly to inflammatory and autoimmune disease pathologies. Binding of IL-17 to its receptor leads to recruitment of adaptor protein CIKS/Act1 via heterotypic association of their respective SEFIR domains and activation of transcription factor NF-κB; it is not known whether CIKS and/or NF-κB are required for all gene induction events. Here we report that CIKS is essential for all IL-17-induced immediate-early genes in primary mouse embryo fibroblasts, whereas NF-κB is profoundly involved. We also identify a novel subdomain in the N terminus of CIKS that is essential for IL-17-mediated NF-κB activation. This domain is both necessary and sufficient for interaction between CIKS and TRAF6, an adaptor required for NF-κB activation. The ability of decoy peptides to block this interaction may provide a new therapeutic strategy for intervention in IL-17-driven autoimmune and inflammatory diseases.

Introduction

The discovery of the inflammatory T helper cell type-17 (Th17), a subset distinct from the classical Th1 and Th2 populations, has revolutionized our understanding of T-cell mediated immunity (1–3). Th17 cells are critical in host defense against many pathogens, in particular extracellular bacteria and fungi. When improperly controlled, however, Th17 responses can also feature prominently in a number of inflammatory and autoimmune diseases, such as rheumatoid arthritis, systemic lupus erythematosus, multiple sclerosis, and psoriasis (reviewed in Refs. 4–6). The discovery of the Th17 cell type has also focused attention on its signature cytokine IL-17 (also known as IL-17A). A critical, although by no means exclusive biologic corollary of IL-17 expression is the recruitment of neutrophils to sites of inflammation (7, 8).

IL-17A is a member of a family of cytokines that also includes IL-17B, -C, -D, -E (also known as IL-25), and -F. IL-17A signals via a receptor composed of the IL-17RA and RC chains; these receptor chains are members of a family that also includes RB, RD, and RE (reviewed by Ref. 9). All receptor polypeptides encode a so-called SEFIR domain (similar expression to fibroblast growth factor genes and IL-17Rs) in their cytoplasmic tails (10, 11). Such a domain is also present on the adaptor protein CIKS² (connection to IκB Kinase and Stress-activated protein kinases (12); also known as Act1 (13) or TRAF3IP2). IL-17A and -F, as well as IL-25 have been shown to signal by recruitment of CIKS to their cognate receptors, mediated via heterotypic SEFIR domain associations (14–17).

A number of downstream effectors can be activated by IL-17, the best studied member of this cytokine family, including the transcription factors NF-κB, C/EBP, and AP-1, as well as MAP kinases; in addition, IL-17 can potently stabilize mRNAs, although mechanisms remain to be discovered (reviewed by Ref. 9). Mere overexpression of CIKS profoundly activates p50/p65 NF-κB complexes via the classical activation pathway (12, 13, 18). By contrast, stimulation of cells with IL-17, which signals via CIKS, activates NF-κB quite weakly (15, 19), calling into question the physiologic significance of NF-κB activation. Indeed, IL-17 can synergize with TNFα, which has been ascribed to the fact that TNFα, unlike IL-17, strongly activates NF-κB, whereas IL-17 stabilizes some short-lived mRNAs induced by TNFα (20). This in turn has fostered the notion that mRNA stabilization may be a primary function of IL-17. Another unsettled question concerning IL-17 signaling is whether the CIKS adaptor is essential for expression of all target genes, especially since some reports suggest the possibility of CIKS-independent signaling events (15, 21, 22).

Precisely how CIKS transmits signals to its downstream effectors, including NF-κB, is only beginning to be elucidated. Recently a central domain of CIKS has been reported to function as an E3-ubiqutin ligase, capable of adding Lys⁶³-linked polyubiquitin chains to the adaptor protein TRAF6 (23). TRAF6 have previously been found essential for IL-17/CIKS-mediated activation of NF-κB (14, 15, 24). The E3-ubiquitin ligase function was also reported necessary for mRNA stabilization, but in this case via a TRAF6 independent mechanism (23, 25). Ubiquitination of TRAF6 is secondary to recruitment to CIKS, and CIKS reportedly encodes two TRAF6 binding sites; these sites may be redundant because signaling was only impaired when both sites on CIKS were rendered non-functional through mutagenesis (23, 24). Once polyubiquitinated by CIKS, the TRAF6 adaptor may activate NF-κB via mechanisms already established for signaling by Toll, IL-1, and CD40 receptors, i.e. by activation of Tak1 and IKK (9, 26).

Here we investigate initial transcriptional responses of IL-17 and its molecular signaling mechanisms with the use of primary mouse embryo fibroblasts. We demonstrate that CIKS is absolutely essential for all initial IL-17-induced transcription and we, furthermore, show that classical activation of NF-κB is especially critical for these responses. We also identify a novel domain in the N terminus of CIKS that is required for interaction with TRAF6 and activation of NF-κB. We discuss these findings in terms or their potential to open new avenues for therapeutic intervention in diseases dependent on IL-17 cytokines.

EXPERIMENTAL PROCEDURES

Cell Culture and Reagents

Primary mouse embryo fibroblast cultures (MEFs) were established from wild-type (WT) and CIKS-deficient (knockout) mice as described previously (16). Mice were bred and housed in National Institute of Allergy and Infectious Diseases facilities, and all experiments were done with approval of the National Institute of Allergy and Infectious Diseases Animal Care and Use Committee and in accordance with all relevant institutional guidelines. Immortalized NEMO-deficient MEFs were kindly donated by Dr. Manolis Pasparakis. Cycloheximide and actinomycin D were purchased from Sigma; SB203580, JNK inhibitor II, JAK inhibitor I, and PD98059 were from Calbiochem; and MLN120b was kindly supplied by Millennium Pharmaceuticals. FITC-tagged cell penetrating peptides, TAT wild-type CIKS, GRKKRRQRRRPPQMNRSIPVEVDESEPYP, and TAT E17A CIKS, GRKKRRQRRRPPQMNRSIPVAVDESEPYP, were purchased from American Peptide. MEFs were treated for 30 min with these peptides in serum-free medium prior to stimulation. Uptake of FITC-labeled cell penetrating peptide was confirmed by FACS analysis. Recombinant IL-17 (100 ng/ml, R&D Systems) and/or TNFα (2 ng/ml, Peprotech) were used for stimulation.

RNA Isolation and Real Time PCR (RT-PCR)

RNA was isolated using the RNeasy RNA isolation kit (Qiagen) according to the manufacturer's instructions. cDNA was synthesized using oligo(dT) and SuperScript III (Invitrogen). Expression of Iκbζ, Ccl2, c/Ebpδ, Zc3h12a, Il-6, Cxc1, and β-actin was quantified by TaqMan qPCR using primers from Applied Biosystems. All gene expression results are expressed as 2^−ΔΔC_t, where ΔΔC_t = (C_t,target − C_t,β-actin) for stimulated samples − (C_t,_target − C_t,β-actin) for unstimulated controls. Data are shown as the mean ± S.E. For the mRNA stability experiment, expression of Cxcl1 was calculated as 2^−ΔC_t, where ΔC_t = (C_t,Cxcl1 − C_t,β-actin). Expression at 0 h was set to 100% and the remaining samples were normalized accordingly.

Genechip Experiment

RNAs were extracted and DNA microarray targets were prepared as described previously (27). Gene expression was measured using the Affymetrix 430 2.0 Genechip containing the mouse genome and data analysis were carried out as described previously (28) with the following modifications. An analysis of variance was performed using the Partek Genomics Suite (Partek, Inc., St. Louis, MO, version 6.3 6080414) to obtain multiple test corrected p values using the false discovery rate method at the 0.05 significance level combined with a fold-change value of 1.5. The data discussed in this article have been deposited in the NCBI Gene Expression Omnibus (29) and are accessible through GEO Series accession number GSE24873.

Western Blot and Antibodies

Whole cell extracts were isolated, loaded on to 10% SDS-polyacrylamide gel, electrophoresed, and transferred to PVDF (Millipore) membranes. The following antibodies were used: anti-FLAG (Sigma); anti-HA, anti-TRAF6, anti-β-actin, and anti-IκBα (all from Santa Cruz); and anti-NF-κB p65, anti-phospho-Ser⁵³⁶ NF-κB p65, anti-ERK, and anti-phospho-ERK (all from Cell Signaling).

Alignment, Plasmids, and Lentivirus

Sequence alignment was carried out using ClustalW2 (30) using default settings. Full-length human CIKS and CIKS deletion/point mutants were cloned into a Gateway Entry vector (Invitrogen) and subcloned into a lentiviral vector or into pcDNA3.1 HA or FLAG Tag destination vectors by Gateway LR recombination using the manufacturer's protocols to generate expression clones. In these vectors the standard cytomegalovirus promoter was replaced by the PolII promoter to ensure low level constitutive expression, with the exception of the vectors used in Fig. 2A and the IL-17RA vector used in Fig. 4B. The TRAF6 expression vector has been described previously (31). Plasmid constructs were confirmed by sequencing. Lentivirus preparations used for transduction of wild-type and mutant CIKS proteins into CIKS-deficient primary MEFs were generated with the ViraPower Lentiviral Expression System (Invitrogen) following the manufacturer's instructions.

FIGURE 2.

A conserved domain in the N terminus of CIKS is required for NF-κB activation. A, B, and D, HeLa cells were co-transfected with the indicated HA- or FLAG-tagged CIKS constructs, the NF-κB luciferase reporter and a Renilla construct were used for internal control. Results are recorded as fold-induction of NF-κB activity relative to cells transfected with an empty expression vector. Data are shown as the mean ± S.E. for four independent experiments; *, p < 0.05 Student's t test compared with cells transfected with WT CIKS. Lower panels show approximately equal expression of transfected CIKS proteins in cell lysates; β-actin served as the loading control. C, a short N-terminal sequence of CIKS is highly conserved across multiple species. Alignment of CIKS sequences from Homo sapiens (human), Mus musculus (mouse), Bos taurus (cow), Gallus gallus (chicken), and Xenopus tropicalis (frog). Asterisk denotes identical amino acids; colon denotes conserved substitutions; and dot denotes semi-conserved substitutions.

FIGURE 4.

The N-terminal domain of CIKS is essential for interaction with TRAF6, but not for interaction with self or with IL-17RA. A, HeLa cells were co-transfected with tagged wild-type or ΔN50 mutant CIKS together and with differently tagged wild-type or mutant CIKS constructs as indicated. Cell lysates were IP to evaluate association of co-expressed CIKS proteins in immunoblots (IB) as indicated. B, FLAG-tagged wild-type or one of several mutant CIKS constructs were co-transfected together with HA-tagged IL-17RA. Cell lysates were IP with anti-HA IL-17RA and analyzed with immunoblots for the presence of co-precipitating CIKS proteins as well as endogenous TRAF6. Analyses shown in A and B are representative of at least three independent experiments.

Transfection, Luciferase Assay, and Immunoprecipitation

HeLa cells were transfected using Lipofectamine 2000 (Invitrogen). Whole cell extracts were isolated 48 h after transfection. Immunoprecipitations (IPs) were carried out using IP kits (Sigma) according to the manufacturer's instructions and analyzed by Western blotting as described above. For luciferase assays, HeLa cells were co-transfected with the IgκB Luc reporter as described previously (18). Luciferase activity was determined 24 h later using the Dual Luciferase assay system (Promega) according to the manufacturer's instructions and normalized to an ER Renilla internal control (Promega).

RESULTS

IL-17 Induces Degradation of IκBα and Phosphorylation of p65 Dependent on CIKS

To understand the significance of NF-κB and CIKS in the cellular response to IL-17 we made use of freshly isolated primary mouse embryo fibroblasts. We first re-examined whether and, if so, to what extent this cytokine can activate NF-κB in primary, freshly isolated, wild-type or CIKS-deficient MEFs. IL-17 induced the degradation of IκBα in wild-type cells, albeit much less so than TNFα, and only TNFα, not IL-17 was able to do so in CIKS-deficient cells (Fig. 1A). Thus IL-17 was able to liberate NF-κB in a CIKS-dependent manner in primary cells.

FIGURE 1.

CIKS is essential for IL-17-mediated NF-κB activation, which plays a major role in IL-17-induced immediate-early gene expression. A, wild-type (WT) and CIKS-deficient (CIKS KO) primary MEFs were treated with cycloheximide for 30 min (to prevent the rapid NF-κB-mediated re-synthesis of IκBα) prior to stimulation for 2 h with IL-17 or TNFα or in the absence of stimulation (Control). Cell lysates were analyzed for IκBα and β-actin as a loading control. B, WT and CIKS knockout MEFs were stimulated as shown and lysates were analyzed for the presence of p-Ser⁵³⁶ p65 and total p65. Blots shown are representative of at least three independent experiments. C, real time PCR analyses of RNAs from primary WT MEFs pre-treated for 30 min with the following pharmacological inhibitors prior to stimulation with IL-17 for 2 h: 10 μm DMSO (control), 10 μm MLN120b (NF-κB inhibitor), 10 μm SB203580 (p38 inhibitor), 10 μm JNK inhibitor II, 10 μm JAK inhibitor I, or 10 μm PD98059 (ERK inhibitor). D, real time PCR analyses of RNAs from WT- and NEMO-deficient MEFs stimulated with IL-17 for 2 h. Fold-induction is in reference to unstimulated controls. Data are shown as the mean ± S.E. for six independent experiments; *, p < 0.05 paired Student's t test.

The transcriptional activity of the dimeric NF-κB complexes in the nucleus is also determined by phosphorylation, especially of p65/RelA. Various potential phosphorylation sites have been reported for p65; their occurrence, relative significance, and precise function are not fully understood and appear to be context-dependent (32, 33). Here we show that both IL-17 and TNFα were able to induce phosphorylation of p65 at Ser⁵³⁶ in wild-type MEFs, which was fully dependent on CIKS in the case of IL-17 but not TNFα, as expected. No synergistic effect was observed when stimulating with both IL-17 and TNFα (Fig. 1B).

The Early Transcriptional Response to IL-17 in Primary Mouse Embryo Fibroblasts

To further explore the role of NF-κB and CIKS in the transcriptional response to IL-17 in primary MEFs, we first performed a genome-wide microarray analysis to identify genes that were significantly induced by IL-17 within 2 h (see supplemental Fig. S1 for full results). After applying stringent criteria to the results from 6 independently performed experiments, we identified 9 such genes in primary wild-type MEFs; by contrast, IL-17 failed to induce any genes in CIKS-deficient MEFs. These results were confirmed with qPCR analyses (supplemental Fig. S1C). A set of 9 genes encodes for chemokines Cxcl1, Ccl2, and Ccl7, the cytokines Lif and Il-6, transcription factors Iκbζ, c/Ebpδ, and RelB, and RNA-binding protein Zc3hl2a; most of these genes have been identified previously in various screens for IL-17-induced genes (21, 34, 35), although Zc3h12a has been noted only once in a recent screen of liver cells (36), whereas RelB has never been identified previously. Whereas Iκbζ, c/Ebpδ, and Zc3h12a were induced only with IL-17, not TNFα, the remaining genes were induced with either cytokine and are all known potential targets of NF-κB (37–39). Within the latter group, Cxcl1, Lif, and Il-6 appeared to be synergistically induced by the two cytokines, which may be due to post-transcriptional regulation (20, 40).

Despite the early time point after stimulation chosen in our analyses, Il-6 and c/Ebpδ are not immediate-early genes; they have previously been reported to depend on induced expression of IκBζ, and their peak of expression occurred well after 2 h (41, 42) (see also supplemental Fig. S2A). We conclude that CIKS is essential for expression of all IL-17-induced genes in primary MEFs within 2 h. The expression profiling study also identified immediate-early targets of IL-17 in primary cells that could be investigated further for their dependence on NF-κB.

The Role of NF-κB in the Immediate-early Transcriptional Response to IL-17

To address what downstream effectors may contribute to immediate-early IL-17-induced gene expression, we pre-treated primary wild-type MEFs with inhibitors for NF-κB, JAK, and MAP kinases p38, JNK, and ERK prior to stimulation with IL-17. Induced expression of immediate-early genes Cxcl1, Ccl2 and, to a lesser extent Iκbζ and Zc3h12a, was dependent on NF-κB, but was not significantly dependent on p38, JNK, JAK, or ERK (Fig. 1C, see also supplemental Fig. S2B).

To verify the role of NF-κB, we also investigated the IL-17-induced expression of the above genes in NEMO-deficient MEFs, as the classical NF-κB activation pathway is completely abrogated in these cells. Consistent with the inhibitor studies, Ccl2 and Cxcl1 failed to be induced in the absence of NEMO, whereas induced expression of Iκbζ and Zc3h12a was significantly decreased, albeit not abolished (Fig. 1D). We conclude that NF-κB is critically involved in the IL-17-induced expression of immediate-early genes, albeit to a variable degree.

A Novel Domain in CIKS Required for Activation of NF-κB

To identify novel domains within CIKS critical for activation of NF-κB we generated and initially screened a number of CIKS mutants by assessing their ability to induce expression from an NF-κB-dependent luciferase reporter plasmid upon overexpression in HeLa cells (18). As expected, deletion of the SEFIR domain (ΔSEFIR), which is required for self-association (18), or the central region (Δ200–400), which carries the E3 ubiquitin-ligase domain (23), significantly reduced the ability of exogenously introduced CIKS to activate NF-κB (Fig. 2A). By contrast, deletion of one of the two purported TRAF6 binding sites on CIKS (Δ38–42) had no effect on NF-κB activation (23).

Surprisingly, deletion of the N-terminal 50 amino acids (ΔN50) completely abolished NF-κB activation (Fig. 2B). We therefore generated additional mutants to identify sequences within the N terminus that may be required for NF-κB activation. Deletion of amino acids 1–35 (Δ1–35), 1–15 (Δ1–15), and 10–25 (Δ10–25) all completely eliminated the ability of overexpressed CIKS to activate NF-κB (numbering is based on the longer of the two human CIKS isoforms; the shorter isoform starts at position 10, but is otherwise identical) (Fig. 2B). Deletion of amino acids 20–35 (Δ20–35) resulted in partial impairment of CIKS to activate NF-κB, whereas deletion of amino acids 35–50 (Δ35–50), which contains a site previously thought to bind TRAF6 (23, 24) failed to impair this CIKS function. This suggested that a previously unknown N-terminal domain, most likely contained between positions 10 and 25, was crucial for activation of NF-κB.

A comparison of CIKS sequences from various species showed a remarkable conservation of amino acids in the interval between positions 10 and 21; there is little or no conservation elsewhere in the larger N-terminal region of CIKS when more divergent species are compared (Fig. 2C). To further delimit the critical sequences, we generated four point mutations in the domain between positions 10 and 21, replacing Ser¹³, Val¹⁶, Glu¹⁷, or Asp¹⁹ with Ala (S13A, V16A, E17A and D19A); in addition we generated a combined mutant in which all four positions were replaced (“Quad”). As shown in Fig. 2D, the E17A mutation completely abrogated the ability of CIKS to activate NF-κB and the D19A mutation partially interfered, whereas the S13A and V16A mutations had no significant effect in this assay. Consistent with this, the Quad mutation completely blocked CIKS from activating NF-κB as well. We conclude that the N terminus of CIKS harbors a previously unidentified domain likely located between amino acid positions 10 and 21 that is required for activation of NF-κB in CIKS-transfected cells.

The N-terminal Domain of CIKS Is Critical for IL-17-induced Target Gene Expression and Activation of NF-κB

To understand the importance of this newly identified N-terminal domain in a physiologic context, we employed a lentivirus transduction strategy to reconstitute CIKS-deficient primary MEFs with minimal levels of wild-type and mutant CIKS proteins (Δ10–25, E17A, and ΔN50), or with a GFP control, and then assayed for responses to stimulation with IL-17. We first tested for IL-17-induced expression of the immediate-early target test genes investigated above: Iκbζ, Zc3h12a, Cxcl1, and Ccl2. Reconstitution with the ΔN50, Δ10–25, and E17A CIKS mutants significantly impaired induction of Cxcl1, Ccl2, and Zc3h12a, and to a lesser extent Iκbζ, when compared with cells reconstituted with wild-type CIKS (Fig. 3, A and B) (see supplemental Fig. S2C for Il-6 and c/Ebpδ).

FIGURE 3.

The N-terminal domain of CIKS is essential for IL-17-mediated NF-κB activation in CIKS-reconstituted primary MEFs, but not for IL-17-induced mRNA stabilization or ERK activation. A–D, CIKS-deficient primary MEFs were reconstituted via lentiviral transduction with wild-type CIKS, a mutant CIKS (CIKS ΔN50, CIKS Δ10–25, CIKS E17A, or CIKS ΔSEFIR) or with GFP as a negative control, as indicated. A, real time PCR analyses for the indicated genes with RNAs isolated from cells stimulated for 2 h with IL-17 or left unstimulated. Data are shown as the mean ± S.E. for six independent experiments and fold-induction is in reference to unstimulated cells; *, p < 0.05 Student's t test. B, Western blot analyses of cell lysates from a representative experiment used in A to verify approximately equal expression of the transduced FLAG-tagged CIKS proteins, with β-actin serving as a loading control. C, Western blot analyses of cell lysates for IκBα; β-actin served as a loading control. Cells contained approximately equal levels of transduced FLAG-tagged CIKS proteins. D, Western blot analyses for activated phosphor-ERK, total ERK levels, and reconstituted FLAG-tagged CIKS levels. The experiments shown in C and D are representative of at least 3 independent experiments. E, RT-PCR analyses for the indicated genes of RNAs isolated from primary reconstituted MEFs 2 h after stimulation with IL-17, TNFα, or IL-17 + TNFα, as indicated. Data are recorded as the mean ± S.E. for six independent experiments measuring fold-induction relative to unstimulated cells; *, p < 0.05 Student's t test. F, Western blot analyses of cell lysates collected from a representative experiment used in E to show expression levels of the transduced FLAG-tagged CIKS proteins, with β-actin serving as a loading control. G, real time PCR analysis of Cxcl1 mRNA expression levels. RNA was isolated from primary CIKS-deficient MEFs reconstituted with wild-type or mutant CIKS, as indicated, and treated as follows. Cells were stimulated with TNFα for 1 h, followed by addition of first actinomycin D and 10 min later, IL-17; RNA was isolated at 0, 2, 4, and 6 h after addition of IL-17. Data are recorded as the mean ± S.E. for five independent experiments measuring % of remaining Cxcl1 mRNA; *, p < 0.05 Student's t test compared with CIKS-deficient MEFs transduced with CIKS. H, Western blot analyses of cell lysates collected from a representative experiment used in G to show expression levels of the transduced FLAG-tagged CIKS proteins, with β-actin serving as a loading control.

To directly evaluate the ability of IL-17 to activate NF-κB in the primary MEFs reconstituted with the wild-type and CIKS mutants shown above (Fig. 3, A and B), we measured IL-17-induced degradation of IκBα. As shown in Fig. 3C, only CIKS-deficient MEFs reconstituted with wild-type CIKS were able to degrade IκBα in response to IL-17, whereas those reconstituted with CIKS mutants that crippled the critical N-terminal domain failed to respond to IL-17, as did those reconstituted with GFP only. Together these data demonstrate that the newly identified N-terminal CIKS domain is essential for IL-17-induced degradation of IκBα and thus for activation of NF-κB-dependent target genes.

The N-terminal Domain Is Not Required for IL-17-induced mRNA Stabilization or ERK MAPK Activation

The ability of the N-terminal mutants to partially restore induced expression of, in particular, Iκbζ, albeit at significantly reduced levels compared with wild-type CIKS-reconstituted cells, suggests that certain IL-17-induced signals do not require the N-terminal domain. In addition to NF-κB activation, IL-17 has been reported to induce activation of MAPKs, including ERK (43). Primary MEFs reconstituted with wild-type CIKS restored the ability of IL-17 to fully activate ERK, as did the ΔN50 mutant when compared with cells lacking CIKS (Fig. 3D). There was no significant difference between MEFs reconstituted with CIKS or the ΔN50 mutant. We also noted an apparent, low level of ERK activation in the absence of CIKS; the origin of this weak response is presently unclear. These data do show that the N-terminal domain of CIKS is not necessary for full activation of ERK MAPK, although some (other) portion of CIKS appears to be required.

IL-17 can induce the stabilization of short-lived mRNAs, which has been well documented for the mRNA of Cxcl1, and which appears to be the primary cause for the profound synergy between IL-17 and TNFα observed for this gene (20). Because the ΔN50 mutant was unable to activate NF-κB, we asked whether this mutant might nevertheless, still be able to synergize with TNFα, which potently activates NF-κB by itself. As shown in Fig. 3, E and F, transduction of primary MEFs with a lentivirus capable of expressing either wild-type CIKS or the ΔN50 mutant fully reconstituted the ability of IL-17 to synergize with TNFα in the induction of Cxcl1 and Il-6. As expected, no synergy between IL-17 and TNFα was observed in CIKS-deficient MEFs transduced with the ΔSEFIR mutant or with the GFP control. These results were confirmed by directly assessing the IL-17-mediated stabilization of Cxcl1 mRNA under these conditions. As shown in Fig. 3, G and H, Cxcl1 mRNA was degraded significantly faster in CIKS-deficient MEFs compared with MEFs reconstituted with either CIKS or the ΔN50 mutant upon addition of IL-17; this degradation was followed in the presence of actinomycin D, which was added to stop transcription after the initial mRNA was induced by TNFα. There was no significant difference in terms of mRNA stabilization by IL-17 between MEFs reconstituted with CIKS or the ΔN50 mutant. These results show that the N-terminal domain of CIKS, whereas critical for IL-17-mediated NF-κB activation, is not required for mRNA stabilization. Together with the ability of the mutants to also still activate ERK, we further conclude that the N-terminal mutations must not have caused an overall impairment of CIKS structure and function.

The N-terminal Segment of CIKS Is Necessary for TRAF6 Binding

To identify specific functions of the newly identified N-terminal CIKS domain, we next tested the ability of mutant CIKS to self-associate. The ability to self-associate is thought to be required for activation of NF-κB and other downstream signals and it requires the SEFIR domain (18). As shown in Fig. 4A, the ΔN50 mutant was fully capable of associating with a differently tagged version of self or wild-type CIKS in co-transfected cells; as a negative control, the ΔSEFIR mutant was unable to associate with wild-type CIKS or the ΔN50 mutant, as expected.

Next we asked whether loss the N-terminal domain might affect association with the IL-17RA receptor. This receptor and CIKS can associate to some extent in transfected cells, even in the absence of a signal, as has also been noted by others (14, 15). We therefore ectopically expressed IL-17RA and wild-type and mutant CIKS proteins in HeLa cells and then immunoprecipitated IL-17RA and immunoblotted for CIKS. All of the N-terminal domain-crippled CIKS mutants as well as Δ35–50 were still able to associate with IL17RA, whereas loss of the SEFIR domain abolished this association, as expected (Fig. 4B). Therefore, loss of the N-terminal 50 amino acids had no effect on the ability of CIKS to self-associate or to associate with IL-17RA.

We then investigated whether loss of the newly identified N-terminal domain of CIKS impacted association with TRAF6. This adaptor protein had previously been shown to be required for IL-17/CIKS-mediated NF-κB activation, but not mRNA stabilization (23, 25). We therefore tested for association of endogenous TRAF6 with the IL-17RA receptor in the experiment described above, in which IL-17RA and the various CIKS forms were exogenously expressed (Fig. 4B). Although all N-terminal domain-crippled CIKS mutants, including E17A, were co-immunoprecipitated with IL-17RA, none facilitated the co-immunoprecipitation of endogenous TRAF6. By contrast, wild-type CIKS and the Δ35–50 mutant, which retains the newly identified domain (see above), readily facilitated co-immunoprecipitation of endogenous TRAF6. These data demonstrate that the N-terminal CIKS domain comprised of amino acids 10 to 21 is essential for TRAF6 interaction, which explains why this domain is also critical for IL-17-induced NF-κB activation.

The N-terminal CIKS Domain Is Both Necessary and Sufficient for TRAF6 Interaction and May Provide a Target to Impair Signaling

We next asked whether the N-terminal domain of CIKS might be sufficient for binding to TRAF6. To this end we generated vectors capable of expressing FLAG-tagged proteins in which GST was fused to the wild-type or E17A mutant N-terminal 50 amino acids or amino acids 10–25 of CIKS. As shown in Fig. 5A, exogenously expressed TRAF6 could be co-immunoprecipitated with fusions carrying the wild-type N-terminal domain (1–50 or 10–25); by contrast, TRAF6 failed to be co-immunoprecipitated if these fusions carried the E17A mutation. These data clearly indicate that the newly identified N-terminal domain of CIKS is not only necessary, but can also be sufficient to facilitate association with TRAF6.

FIGURE 5.

The N-terminal domain of CIKS is sufficient to interact with TRAF6. A, HeLa cells were co-transfected with TRAF6 and FLAG-tagged GST fusions carrying the wild-type CIKS sequence between amino acid positions 10 and 25 or 1 and 50 or the E17A mutant versions of both of these sequences. Cell lysates were IP with anti-FLAG analyzed in immunoblots (IB) for TRAF6. Analyses shown are representative of at least three independent experiments. B, real time PCR analyses for the indicated genes with RNAs from primary WT MEFs treated with cell penetrating peptides containing the wild-type CIKS sequence between amino acid positions 10 and 25 or an E17A mutant version prior to stimulation with IL-17 for 1 h. Induction values are relative to stimulated, but otherwise untreated WT MEFs, arbitrarily set to 100. Data are shown as the mean ± S.E. for three independent experiments; *, p < 0.05 Student's t test. C, FACS analysis confirms an equal amount of cell penetrating peptides in the MEFs. Analyses shown are representative of at least three independent experiments.

Given the importance of the N-terminal domain for TRAF6 binding and NF-κB activation, we also investigated whether this domain could be targeted to interfere with CIKS-mediated NF-κB activation. To this end we generated fusion peptides composed of the cell penetrating portion of HIV-TAT and the wild-type N-terminal domain of CIKS (AA 10–25), or a E17A mutant version of this CIKS sequence as a negative control. When MEFs were pre-exposed to these peptides, the cell-penetrating peptide carrying the wild-type N-terminal domain significantly interfered with IL-17-induced expression of Cxcl1 as well as Zc3h12a, whereas the E17A mutant control peptide did not (Fig. 5, B, control for uptake of peptides shown in C)). These results raise the possibility that the N-terminal TRAF6 binding domain could be a target for therapeutic intervention in IL-17-driven pathogenesis.

DISCUSSION

The present study shows that IL-17-mediated activation of the classical NF-κB pathway in primary MEFs is fully dependent on the adaptor CIKS; in turn, IL-17-induced NF-κB is critical for immediate-early gene induction, despite the relatively weak level of activation when compared with that of TNFα. Although IL-17-induced expression of some immediate-early genes appears to be almost completely dependent on NF-κB, others are at least partially dependent. The present study also shows that CIKS is absolutely required for all IL-17-induced transcription at 2 h post-stimulation. These insights emphasize the significance not just of CIKS but also of NF-κB in IL-17 signaling, which has previously not been fully appreciated. Concerning the mechanism of NF-κB activation, we demonstrate that a new N-terminal domain of CIKS, likely situated between amino acids 10 and 21 (numbering refers to the longer of the two human isoforms, with the shorter starting at position 10), is absolutely necessary for association of CIKS with TRAF6, for activation of NF-κB and consequently, also for proper IL-17-induced expression of immediate-early target genes in primary MEFs. Furthermore, this new domain appears to be sufficient for recognition by TRAF6, as the presence of this domain in a fusion with an irrelevant carrier protein is able to bestow the ability to attract TRAF6. Finally, the newly identified N-terminal domain may represent a potential target for therapeutic intervention in diseases in which IL-17 cytokines drive pathogenesis, as decoy cell-penetrating peptides carrying this domain appear to interfere with IL-17-induced and TRAF6/NF-κB-dependent gene expression.

CIKS was previously thought to contain two putative TRAF6 binding sites based on matching a consensus TRAF6 binding site (AA numbers 38–42 EEESE and AA numbers 333–337 EERPA). Mutations in either site had no effect on CIKS function, but a combined mutation of both sites was reported to significantly impair TRAF6 binding, TRAF6 ubiquitination, NF-κB activation, and JNK phosphorylation (23). It is presently unclear how these two mutations together, but not individually, interfered with CIKS function, although it is conceivable that together these mutations somehow impaired the overall structure and thus function of CIKS. Importantly, as shown here, deletions of or mutations within the interval between AA 10 and 21 (specifically the E17A mutant) completely abolished TRAF6 binding of CIKS as well as TRAF6- and NF-κB-dependent signaling. At the same time, these deletions and mutations had no effect on other, TRAF6-independent functions of CIKS, including the ability of CIKS to self-associate or to associate with the receptor, IL-17-induced mRNA stabilization, and ERK phosphorylation. Furthermore, this newly identified TRAF6 binding site on CIKS, in isolation, was able to bind TRAF6. Therefore, the N-terminal domain between AAs 10 and 21 is necessary for interaction with TRAF6 and, in consequence, for TRAF6/NF-κB-dependent signaling in response to IL-17 and, it can be sufficient for interaction with TRAF6. We thus conclude that the newly identified N-terminal domain is the true and functionally critical TRAF6 binding site on CIKS. It remains theoretically possible that secondary TRAF6 binding sites exist within CIKS; if such sites exist, however, their engagement with TRAF6 would have to depend on an initial interaction with the new domain described here.

The AA sequence between positions 10 and 21 is well conserved between species, including frog: a sequence that bears little similarity elsewhere in the extended N-terminal region of mammalian CIKS sequences. The 12-AA long domain includes a sequence that conforms with a loose TRAF6 binding consensus reported previously (CIKS 10–21, MNRSIPVEVDES; TRAF6 consensus binding site PXEXX(Ar/Ac) (44–46). The newly identified TRAF6 binding domain in CIKS is situated at or very close to the N terminus of CIKS (the short form of human CIKS starts at position 10, the beginning of the newly identified domain). This may make this site very accessible for interaction with TRAF6; indeed, exogenously expressed CIKS readily interacts with TRAF6, even in the absence of any signals (15).

IL-17 has been strongly implicated in disease pathogenesis in a number of inflammatory and autoimmune diseases, such as rheumatoid arthritis, multiple sclerosis, psoriasis, as well as severe asthma among others (4–6). Furthermore, IL-25, another member of the IL-17 family that signals via CIKS, has been implicated in asthma, as it induces a strong Th-2-like innate response that also promotes Th2 adaptive responses thought to underlie most cases of asthma (16, 17, 47). Thus it might be useful to develop a means to block IL-17 cytokine signaling in such diseases. To date the proof for this principal has been established in mouse models for the diseases such as rheumatoid arthritis (48) and experimental autoimmune encephalomyelitis (the mouse model for multiple sclerosis) (49): in these models loss of CIKS abolished or significantly ameliorated disease development, respectively, without causing any other obvious problems, although one must assume that beneficial functions of IL-17 cytokine signaling were also abolished. We have shown here that a cell-penetrating peptide carrying the newly identified N-terminal domain of CIKS, but not a negative control carrying the E17A mutation, interfered with IL-17 induced signaling, presumably by inhibiting TRAF6 binding. Thus new therapeutic methods for IL-17 cytokine-driven diseases based on inhibiting CIKS functions via interference with TRAF6 binding may be entertained in future studies. Indeed peptides carrying a hydrophobic signal sequence fused to the TRANCE receptor binding site for TRAF6 have been shown to interfere with TRANCE-L-induced functions in a similar way (44). In this scenario it must be kept in mind, however, that not all CIKS signaling is mediated via TRAF6, thus TRAF6-independent signals would still be preserved, even if TRAF6 binding was blocked. To what extent TRAF6-dependent versus TRAF6-independent signaling contributes to disease pathogenesis by IL-17 or IL-25 remains to be investigated. Furthermore, it is possible that decoy peptides or small molecule inhibitors based on the TRAF6 binding site on CIKS might inhibit all TRAF6-mediated signals, not just IL-17 cytokine-induced signals. Therefore it will be of interest to also identify sequences within the SEFIR domain that are important for association with the receptor and/or with self, as such sequences might provide a more specific target for blocking CIKS function with decoy peptides or small molecule inhibitors derived thereof.