CKIP-1 Is an Intrinsic Negative Regulator of T-Cell Activation through an Interaction with CARMA1

The transcription factor NF-κB plays a key regulatory role in lymphocyte activation and generation of immune response. Stimulation of T cell receptor (TCR) induces phosphorylation of CARMA1 by PKCθ, resulting in formation of CARMA1-Bcl10-MALT1 (CBM) complex at lipid rafts and subsequently leading to NF-κB activation. While many molecular events leading to NF-κB activation have been reported, it is less understood how this activation is negatively regulated. We performed a cell-based screening for negative regulators of TCR-mediated NF-κB activation, using mutagenesis and complementation cloning strategies. Here we show that casein kinase-2 interacting protein-1 (CKIP-1) suppresses PKCθ-CBM-NF-κB signaling. We found that CKIP-1 interacts with CARMA1 and competes with PKCθ for association. We further confirmed that a PH domain of CKIP-1 is required for association with CARMA1 and its inhibitory effect. CKIP-1 represses NF-κB activity in unstimulated cells, and inhibits NF-κB activation induced by stimulation with PMA or constitutively active PKCθ, but not by stimulation with TNFα. Interestingly, CKIP-1 does not inhibit NF-κB activation induced by CD3/CD28 costimulation, which caused dissociation of CKIP-1 from lipid rafts. These data suggest that CKIP-1 contributes maintenance of a resting state on NF-κB activity or prevents T cells from being activated by inadequate signaling. In conclusion, we demonstrate that CKIP-1 interacts with CARMA1 and has an inhibitory effect on PKCθ-CBM-NF-κB signaling.


Introduction
The NF-kB family of transcription factors plays a key regulatory role in lymphocyte activation and generation of immune response [1]. The respective NF-kB target genes allow the organism to respond effectively to the environmental changes. Engagement of TCR by specific antigen presented on major histocompatibility complex (MHC) of antigen presenting cells (APC) induces T cell activation and proliferation. However, stimulation of TCR/CD3 complex alone is not sufficient for activation of NF-kB. The simultaneous costimulation of CD28 through its ligand, B7, is needed for optimal activation of NF-kB [2]. CD3/CD28 costimulation induces the formation of a large multicomponent complex at the contact site between T cell and the APC, termed as immunological synapse [3,4]. This contact area of T cells is highly enriched in cholesterol and glycosphingo-lipids, also termed as lipid rafts, and serve as the platform for the assembly of proximal signaling components of TCR. PKCh is recruited to the immunological synapse from the cytosol upon T cell stimulation and catalytically activated [5,6]. Activated PKCh phosphorylates CARMA1 (CARD11) to induce its conformational changes which enable CARMA1 to form the complex with Bcl10-MALT1 [7,8]. Subsequently, the IkB kinase (IKK) complex becomes activated and phosphorylates IkBs, leading to their ubiquitylation and subsequent proteasomal degradation. The degradation of IkBs allows NF-kB to enter the nucleus and induce transcription of target genes [1].
CARMA1 is one of a family of caspase recruitment domain (CARD)-and membrane associated guanylate kinase-like (MA-GUK) domain-containing proteins (CARMA) [9,10]. CARMA1 contains an N-terminal CARD, followed by a coiled-coil (CC) domain, a PDZ domain, a Src homology 3 (SH3) domain, and a guanylate kinase (GUK)-like domain in the C-terminus. It has two mammalian homologs, CARMA2 and CARMA3. CARMA1 is predominantly expressed in spleen, thymus, and peripheral blood leukocyte (PBL); CARMA2 is expressed only in placenta; and CARMA3 is expressed in broad range of tissues but not in spleen, thymus or PBL. For B and T cells, the scaffold protein CARMA1 plays an essential role in antigen receptor-induced NF-kB activation [11][12][13][14][15]. Aberrant NF-kB activation could be involved in autoimmune diseases and malignant lymphomas. Constitutively active NF-kB in the activated B cell-like (ABC) subtype of diffuse large B cell lymphoma (DLBCL) can result from somatic mutations in genes involved in NF-kB signaling, such as CD79B, A20 and CARMA1 [16]. Recently, germline mutations in CARMA1 have also been reported in four patients with congenital B cell lymphocytosis [17]. Therefore CARMA1 activity needs to be tightly regulated.
Many findings leading to NF-kB activation have been reported, but it is less understood how this activation is negatively regulated. To elucidate negative regulation in TCR-mediated NF-kB activation, we have done a screening by mutagenesis and complementation cloning strategies. Here we report the identification of CKIP-1 as a negative regulator in NF-kB signaling via TCR. We show that CKIP-1 interacts with CARMA1, inhibits the interaction between PKCh and CARMA1, and suppresses NF-kB activation.

Generation of mutant Jurkat T cells and complementation cloning strategies by lentiviral cDNA library
Jurkat T cell line stably expressing EGFP under the control of an NF-kB-dependent promoter, which we called JR-GFP, was kindly gifted from Dr. Xin Lin [11]. To generate mutant cells, JR-GFP cells were treated with 4 mg/ml of ICR191 (Sigma-Aldrich, St. Louis, USA), alkylating agent that typically generates random frame-shift mutations [11,29], for 5 hr, and this treatment was repeated three times. After mutagenesis, EGFP-positive cells were sorted by BD FACSAria cell sorter (BD, New Jersey, USA) under the treatment with 2.5 mM of PKC inhibitor GF109203X (Sigma-Aldrich). Monoclonal mutant cell lines were derived by limiting dilution, and, among them, an NF-kB constitutively active cell line was identified. Human leukocyte cDNA library (Invitrogen) on pCS2-EF-GATEWAY-IRES-hrGFP was transferred into pCS2-EF-GATEWAY-IRES-H2K k through LR reaction on Gateway cloning system (Invitrogen) and cDNA and H2K k -dual expressing lentiviral vector were prepared as described before [30,31]. To identify NF-kB negative regulators, the NF-kB constitutively active cell line was infected with this viral vector. If the mutant phenotype was rescued by the gene from the library, EGFP expression might return to negative. Both H2K k -positive and EGFP-negative cells were sorted using BD FACSAria cell sorter and subjected to limiting dilution. If EGFP was normally induced by PMA/ionomycin in each single cell clone, the mutant phenotype should be rescued by the gene from the library. The genes rendering the reversion of the mutant phenotype were isolated by PCR using vector specific primers. Subsequent DNA sequencing and BLAST analysis should reveal the integrated gene.
Luciferase reporter assay 5 mg of 5xNF-kB-dependent luciferase (Firefly) reporter plasmid and 0.1 mg of EF1a promoter-dependent Renilla luciferase reporter were transfected together with 5 mg of plasmids encoding the desired genes or 400 pmol of siRNA by electroporation into 1610 7 Jurkat T cells in 0.4 ml serum-free RPMI1640 media at the power setting of 250 V and 950 mF. Nineteen hours later, the transfected cells were treated for 5 hr with plate-bound CD3 mAb (2 mg/ml), plate-bound CD3 + soluble CD28 mAb (2 mg/ml of each), TNFa (20 ng/ml), PMA (10 ng/ml), or PMA (10 ng/ml) + CD28 (2 mg/ml). NF-kB activity was measured with Dual-Luciferase Reporter Assay System (Promega, Madison, USA) and was determined by normalization of NF-kB-dependent Firefly luciferase to Renilla luciferase activity. Values represent the average of three independent experiments and error bars represent the SD from the average. Statistically significance was determined using Student's t test.

Evaluation of NF-kB activity
Nuclear protein fractions were harvested by the Nuclear Extract kit (Active Motif, Carlsbad, USA). NF-kB activity was measured in 2 mg of nuclear protein extracts by the TransAM TM NF-kB p65 chemi (Active Motif), an ELISA-based kit to detect and quantify NF-kB p65 subunit activation. The assay was performed according to the manufacturer's protocol and analyzed using a microplate luminometer PerkinElmer 2030 ARVO TM X3 (Perki-nElmer, Waltham, USA). Values represent the average of three independent experiments and error bars represent the SD from the average. Statistically significance was determined using Student's t test.

Immunoprecipitation
For co-immunoprecipitation, 6-well plate HEK293T cells were transfected by the calcium phosphate method. Two days after transfection, cells were lysed in NP-40 lysis buffer (20 mM Tris-HCl pH 7.5, 250 mM NaCl, 1% NP-40) supplemented with 1 mM PMSF, protease inhibitor cocktail (Nacalai Tesque) and phosphatase inhibitor cocktail (Roche). Total cell lysates were precleared on Protein A Sepharose beads for 30 min at 4uC. The precleared cell lysates were immunoprecipitated with Protein A beads-conjugated with the desired antibodies for 6 hr. Immunoprecipitates were washed three times with lysis buffer.

Confocal microscopy
HEK293T cells were transfected with expression vectors and grown on the coverslips. 24 hr after transfection, the cells were incubated with Alexa Flour 488-conjugated cholera toxin B (Molecular probes) at 4uC for 20 min. The specimens were fixed with 4% paraformaldehyde in PBS and mounted on slides using ProLong Gold antifade reagent with DAPI (Invitrogen), and analyzed by confocal laser scanning fluorescence microscopy (Nikon Digital Eclipse C1).
In vitro binding assay FLAG-CKIP-1 was synthesized in vitro using the TNT T7 Quick Coupled Transcription/Translation System (Promega). GST and GST-CARMA1 CD-CC proteins were produced in E. coli BL21 and purified with glutathione Sepharose 4B beads (GE Healthcare). The beads were incubated with FLAG-CKIP-1 at 4uC for 2 hr. The beads were washed and proteins were eluted, followed by Western blotting with anti-FLAG antibody.

Lipid raft purification
Costimulation of Jurkat T cells was performed in a final volume of 1 ml by addition of anti-CD3 (10 mg/ml) and anti-CD28 (5 mg/ ml) antibodies, together with 15 mg of mouse IgG (Sigma-Aldrich). Cells (2610 7 ) were lysed in 1 ml MNE Buffer (25 mM MES pH 6.5, 150 mM NaCl, 5 mM EDTA) with 1% Triton-X, 1 mM PMSF, and protease inhibitor cocktail (Nacalai Tesque) for 20 min on ice and dounce homogenized 20 times. Samples were centrifuged at 1,0006g for 10 min at 4uC. The supernatants were mixed with 1 ml of OptiPrep (Axis-Shield, Oslo, Norway) and transferred to a Beckman Ultracentrifuge tube. Two milliliters of 30% OptiPrep followed by 1 ml of 5% OptiPrep in MNE buffer were overlaid. Samples were ultracentrifuged in a SW41Ti rotor (200,0006 g for 20 hr). Fractions (400 ml per fraction) were collected from the top of the gradient. Proteins from each fraction were precipitated with trichloroacetic acid before separation by SDS-PAGE and Western blotting.

Identification of CKIP-1 as a negative regulator of NF-kB activation
We have performed a cell-based screening to find negative regulators in TCR-mediated NF-kB activation, using somatic mutagenesis and complementation cloning strategies [11,29]. We used Jurkat T cell line expressing EGFP under the control of an NF-kB-dependent promoter, named JR-GFP [11]. To generate NF-kB constitutively active cell lines, JR-GFP cells were subjected to mutagenesis with ICR191, and EGFP-positive cells were sorted under the treatment of PKC inhibitor GF109203X. After limiting dilution, we identified an NF-kB constitutively active cell line in which negative regulators for NF-kB activation must be mutated. To identify NF-kB negative regulators, the NF-kB constitutively active cell line was infected with a human leukocyte-cDNA library expressing lentivirus, and EGFP-negative cells were sorted. If the mutant phenotype was rescued by transduction of the gene from the library, EGFP expression would return to negative. The genes rendering the reversion of the mutant phenotype were isolated by PCR and sequenced using library vector specific primers, and then we obtained dozens of candidates for NF-kB negative regulators. To examine whether any of these candidates downregulate NF-kB activity, we selected and knocked down eighteen molecules by specific siRNA in JR-GFP cells. We found that knockdown of CKIP-1 induced expression of EGFP more than that of TNFAIP3 (A20), which was known as a negative regulator of NF-kB and used as a positive control [33] (Figure S1). To confirm that CKIP-1 was a negative regulator of NF-kB, Jurkat T cells were transfected with CKIP-1 siRNA together with an NF-kBdependent luciferase reporter plasmid. We used siRNA SMARTpool, which is a mixture of four siRNAs, and separate aliquot of all four individual siRNAs. Knockdown of CKIP-1 increased NF-kB activity ( Figure 1A). We also showed that knockdown of CKIP-1 induced DNA binding activity of NF-kB p65 ( Figure 1B), by using the transcription factor DNA-binding ELISA. Thus, we clearly demonstrated that CKIP-1 was a novel NF-kB negative regulator.

CKIP-1 suppresses NF-kB activation induced by PMA and constitutively active PKCh
To examine whether the downregulation of NF-kB activation by CKIP-1 is specific to TCR stimulation, Jurkat T cells transfected with CKIP-1, treated with different stimulation, and assessed NF-kB activity by luciferase reporter assays. CKIP-1 suppressed NF-kB activity in unstimulated cells and stimulated by CD3, PMA and PMA/CD28, but not by TNFa or CD3/CD28 ( Figure 1C). Using the transcription factor DNA-binding ELISA, we also showed that CKIP-1 suppressed NF-kB activation induced by PMA stimulation ( Figure 1D). These data suggest that CKIP-1 inhibits NF-kB signaling via TCR but not via TNF receptor and that CKIP-1 targets downstream signaling components of PKCh, since the treatment of PMA directly activates PKCs. To clarify which step of signaling CKIP-1 affects, NF-kB activation driven by transfection of each downstream signaling component of PKCh was assessed in Jurkat T cells in the presence or absence of cotransfection of CKIP-1 (Figure 2A). NF-kB activation induced by PKCh AE, a constitutively active mutant [32], was clearly suppressed by CKIP-1, whereas activation induced by NF-kB RelA, IKKb or Bcl10 was not affected. NF-kB activation induced by CARMA1 seemed to be suppressed by CKIP-1, but the effect was not statistically significant. Conversely, knockdown of CKIP-1 increased NF-kB activation induced by transfection of CARMA1 or PKCh AE ( Figure 2B). These results suggest that the inhibitory effect of CKIP-1 targets signaling events around PKCh or CARMA1.
As shown in Figure 1C, CKIP-1 did not suppress CD3/CD28induced NF-kB activation. We hypothesized that CKIP-1 might work in a resting state and finish its role during CD3/CD28 costimulation. PKCh and CARMA1 have been reported to be recruited to lipid rafts upon TCR stimulation [34]. It has been shown that CKIP-1 binds to lipid through its PH domain and overexpressed CKIP-1 localizes in the plasma membrane and partly in the nucleus [18,20,22]. To examine where CKIP-1 localizes in Jurkat T cells, the detergent-insoluble membrane (lipid raft) fractions were prepared by the ultra-centrifugation in a discontinuous OptiPrep density gradient. Lck was constitutively associated with lipid rafts, and PKCh was recruited to lipid rafts after CD3/CD28 costimulation (Figure 3) as previously reported CKIP-1, a Negative Regulator for NF-kB Activation PLOS ONE | www.plosone.org [28,35]. Phosphorylation of Erk was induced by CD3/CD28 costimulation. CKIP-1 partly localized at lipid rafts in unstimulated Jurkat T cells, and intriguingly, CKIP-1 was excluded from lipid rafts when cells were stimulated upon CD3/CD28 ( Figure 3). These data suggest that, when cells are stimulated upon CD3/ CD28 and lipid rafts are accumulated, CKIP-1 localizes out of the lipid rafts and its inhibitory effect does not extend.
Next, we examined the interaction between CKIP-1 and CARMA1 in T cells, using JPM50.6/WT cells, which were reconstituted with Myc-CARMA1 wild type (WT) in CARMA1deficient Jurkat (JPM50.6) T cells [11,28]. Myc-CARMA1 was coimmunoprecipitated with endogenous CKIP-1 but not with control IgG (Figure 4D). To determine the domain of CARMA1 that was critical for the interaction with CKIP-1, truncated forms of CARMA1 were tested ( Figure 4E). Co-immunoprecipitation assays showed that CKIP-1 bound to CARMA1 WT, CD-CC and DCD, but not to DCD-CC ( Figure 4F), indicating that CKIP-1 associates with the CC domain of CARMA1. To determine the responsible region in CKIP-1 for the association with CARMA1, we generated several CKIP-1 truncated forms ( Figure 4E). Coimmunoprecipitation assays revealed that CKIP-1 WT and DLZ bound to CARMA1 but CKIP-1 DPH did not ( Figure 4G), indicating that the PH domain of CKIP-1 was essential for the interaction with CARMA1. To investigate direct interaction between CKIP-1 and CARMA1, in vitro GST pull-down assay was performed. GST-tagged CARMA1 CD-CC was able to interact with FLAG-CKIP-1 but GST was not ( Figure 4H). Together, CARMA1 is a specific interacting partner of CKIP-1.

PH domain of CKIP-1 is essential for the interaction with CARMA1 and the inhibitory effect on NF-kB activation
Next we examined the function of each truncated form of CKIP-1 on NF-kB activation, using luciferase reporter assays. Jurkat T cells were transfected with each CKIP-1 truncated form and stimulated by PMA and CD3/CD28. CKIP-1 WT and DLZ inhibited NF-kB activation induced by stimulation with PMA, but CKIP-1 DPH did not ( Figure 5A, middle panel). Similarly to CKIP-1 WT (Figure 1C), the truncated forms of CKIP-1 gave no influence upon CD3/CD28-induced-NF-kB activation ( Figure 5A, right panel). In resting state, the effect of the truncated forms was not statistically significant, because of the little amount of NF-kB activity in unstimulated cells ( Figure 5A, left panel). Jurkat T cells were transfected with PKCh AE together with each CKIP-1 truncated form. CKIP-1 WT and DLZ suppressed NF-kB activation, but CKIP-1 DPH did not ( Figure 5B, left panel). As shown in Figure 2A, NF-kB activation induced by CARMA1 seemed to be suppressed by CKIP-1 WT, but the effect was not statistically significant. Neither CKIP-1 DLZ nor DPH repressed NF-kB activation induced by CARMA1 ( Figure 5B, right panel). These results suggest that PH domain of CKIP-1, which is required for association with CARMA1, is essential for the inhibitory effect on NF-kB activation.

CKIP-1 inhibits the interaction between PKCh and CARMA1
PKCh phosphorylates CARMA1 in its Linker between the CD-CC domain and the MAGUK domain, which induces conformational change of CARMA1 [7,8]. Then CARMA1 binds to Bcl10 through CARD-CARD interaction [9,36]. Since our data suggested that CKIP-1 interacted with the CC domain of CARMA1, we hypothesized that CKIP-1 might inhibit the interaction between CARMA1 and PKCh or between CARMA1 and Bcl10. Co-immunoprecipitation assays showed that CKIP-1 inhibited the interaction between PKCh and CARMA1, but not that between CARMA1 and Bcl10 ( Figure 6A). As shown in Figure 4B (lane 5 and 6), PKCh was immunoprecipitated with Myc-CARMA1, but, in the presence of co-transfection of CKIP-1, the interaction between PKCh and CARMA1 was diminished ( Figure 4B, lane 3 and 4). Next, we examined the inhibitory effect of CKIP-1 truncated forms on the interaction between PKCh and   Figure 4G and Figure 5), CKIP-1 WT and DLZ inhibited the interaction between PKCh and CARMA1, although CKIP-1 DPH showed no effect ( Figure 6B). These results suggest that CKIP-1 suppresses NF-kB activation by inhibiting the interaction between PKCh and CARMA1.

Discussion
NF-kB signaling in antigen-stimulated lymphocytes plays an important role in immune response. Aberrant NF-kB activation has been shown to be involved in autoimmune diseases and malignant lymphomas. Especially, altered expression and/or function of CBM proteins have been reported in the ABC subtype of DLBCL [16,37,38] and MALT lymphoma [39].
In this study, we show that CKIP-1 is a novel interacting protein with CARMA1 and acts as a suppressor of NF-kB signaling. Our results suggest that CKIP-1 suppresses NF-kB signaling by inhibiting the interaction between PKCh and CARMA1. However, CKIP-1 does not suppress NF-kB activation induced by CD3/CD28 costimulation. Our data suggest that it is because CKIP-1 localizes outside of the lipid rafts and its inhibitory effect does not extend, when cells are stimulated upon CD3/CD28 and lipid rafts are accumulated. A transmembrane adaptor molecule PAG/Cbp is also a negative regulator of T cell activation. In resting T cells, PAG/Cbp is phosphorylated by Lck and interacts with C-terminal Src kinase (Csk), which inhibits T cell activation by suppressing c-Src. In response to stimulation of TCR, PAG/ Cbp becomes rapidly dephosphorylated and dissociates from Csk [40,41]. Likewise, IkBs usually retain NF-kB in the cytoplasm through physical interaction. In response to signaling, IkBs are phosphorylated, leading to their ubiquitylation and subsequent proteasomal degradation [42]. Similarly to PAG/Cbp or IkBs, CKIP-1 usually interacts with CARMA1, but its inhibitory effect might be abrogated during CD3/CD28 costimulation. We presume that CKIP-1 physiologically prevents T cells from being activated by inadequate stimulation and might play a role like a gatekeeper for correct CD3/CD28 signaling at the step of CARMA1 during antigen-stimulation. We speculate that, in resting T cells, CKIP-1 associates with CARMA1 and keeps PKCh away from CARMA1. Our date clearly showed that when T cells are stimulated appropriately upon CD3/CD28 costimulation, both PKCh and CARMA1 are recruited to lipid rafts. However, CKIP-1 remains outside of the lipid rafts, and its inhibitory effect cannot extend. CARMA1 is then phosphorylated by PKCh at the lipid rafts leading to its conformational change into an active form. The activated CARMA1 recruits Bcl10-MALT1 complex and subsequently induces NF-kB activation.
PAG/Cbp-deficient mice exhibit no overt phenotype [43,44], but, in cancer cells, PAG/Cbp is involved in repressing the oncogenecity of c-Src [45]. CKIP-1-deficient mice are reported to undergo an age-dependent increase in bone mass [25]. However, no phenotype about immune disorders or neoplasm has been described. Thus, PAG/Cbp and CKIP-1 might be dispensable or could be compensated by some other negative regulators, because multiple checkpoints through TCR-mediated NF-kB signaling are likely to be independently required to prevent the unwarranted expansion and transformation of lymphocytes, and to ensure an appropriate adaptive immune response. Our data suggest that the suppression of CKIP-1 can work in a resting state or against aberrant PKCh activation such as expression of constitutively active PKCh or treatment of PMA. Similarly to PAG/Cbp, only in malignant lymphomas or immunological disorders, CKIP-1 might play a critical role as a suppressor of aberrant NF-kB activation.
Recently, novel germline CARMA1 mutations have been reported in four patients with congenital B cell lymphocytosis [17]. These CARMA1 mutants constitutively drive NF-kB activation, resulting in elevated NF-kB activity and increased proliferation of patient primary B cells. However, patient primary T cells expressing these CARMA1 mutants are hyporesponsive to CD3/CD28 costimulation. It has also been reported that chronic NF-kB activation, triggered by transgenic expression of constitutively active IKKb in mice, renders T cells hyporesponsive to TCR stimulation [46]. We speculate that T cells have the mechanism by which an anergic state is induced by chronic active NF-kB signaling, and it might be one of the reasons why knockdown of CKIP-1 did not exhibit clear phenotypes in TCR stimulation. Analysis of B cells might be useful for deciphering the physiological role of CKIP-1.
There have been already reported two inhibitory regulators that interact with CARMA1. The kinesin GAKIN negatively regulates occupancy of CARMA1 at the center of the immunological synapse, and limits the extent of signaling [47]. Casein kinase 1a (CK1a), which is reported to be a bifunctional regulator, also interacts with CARMA1 and terminates signaling by phosphorylating CARMA1 [48]. Although CKIP-1 interacts with CARMA1 as GAKIN and CK1a do, CKIP-1 shows several different aspects. Whereas GAKIN competes with Bcl10 for binding, CKIP-1 competes with PKCh but not with Bcl10. GAKIN and CK1a associate with CARMA1 in a signaldependent manner. On the other hand, CKIP-1 neither localizes at lipid rafts nor influences NF-kB activation during CD3/CD28 costimulation. To our knowledge, CKIP-1 is the first molecule that negatively regulates CARMA1 in a resting state or in aberrantly activated signaling.
In conclusion, we have herein demonstrated an inhibitory effect of CKIP-1 in PKCh-CBM-NF-kB signaling. CKIP-1 interacts with CARMA1 and competes with PKCh for binding. It suggests that CKIP-1 plays a unique role to keep resting T cells in a quiescent state or to prevent T cells from being activated by inadequate signaling. Dysfunction of CKIP-1 might constitutively activate NF-kB, leading to autoimmune diseases or malignant lymphomas, and the signaling events around CKIP-1 might be good therapeutic targets. Figure S1 Knockdown of CKIP-1 induces NF-kB activation. The JR-GFP cells were electroporated with 400 pmol of non-targeting siRNA, or specific siRNA against each gene by AMAXA Nucleofector System. Five days later, the expression of EGFP was assessed by FACS. (TIF)