p65 to impede its DNA binding and target gene transactivation

Cellular Signalling 26 (2014) 1437–1444

Contents lists available at ScienceDirect

Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig

Rad GTPase inhibits the NFκB pathway through interacting with RelA/p65 to impede its DNA binding and target gene transactivation Bo-Yuan Hsiao ⁎, Tsun-Kai Chang 1, I-Ting Wu, Mei-Yu Chen ⁎ Institute of Biochemistry and Molecular Biology, National Yang-Ming University, No. 155, Section 2, Li-Nong Street, Taipei 11221, Taiwan

a r t i c l e

i n f o

Article history: Received 18 February 2014 Accepted 4 March 2014 Available online 13 March 2014 Keywords: GTPase Rad RelA/p65 NFκB TNFα

a b s t r a c t Rad is a Ras-related small GTPase shown to inhibit cancer cell migration, and its expression is frequently lost in lung cancer cells. Here we provide evidence that Rad can negatively regulate the NFκB pathway. Overexpressing Rad in cells lowered both the basal and TNFα-stimulated transcriptional activity of NFκB. Compared with control cells, Rad-overexpressing cells displayed more cytoplasmic distribution of the NFκB subunit RelA/p65, while Radknockdown cells had higher levels of nuclear RelA/p65. Depleting Rad did not affect the kinetics of TNFα-induced IκB degradation, suggesting that Rad-mediated regulation of NFκB was through an IκB-independent mechanism. Expression of a nucleus-localized mutant Rad was sufficient to inhibit the NFκB transcriptional activity, whereas expressing the scaffolding protein 14-3-3γ to retain Rad in the cytoplasm alleviated the suppressive effect of Rad on NFκB. GST pull-down assays showed that Rad could directly bind to RelA/p65, and co-immunoprecipitation demonstrated that the Rad–p65 interaction primarily occurred in the nucleus. Adding Rad-containing nuclear extracts or purified GST-Rad in the electrophoretic mobility shift assays dose-dependently decreased the binding of RelA/p65 to an oligonucleotide probe containing the NFκB response element, suggesting that Rad may directly impede the interaction between RelA/p65 and DNA. Rad depletion altered the expression of an array of NFκB target genes, including upregulating MMP9. Knockdown of Rad expression in cells increased both basal and TNFαstimulated MMP9 activities and cell invasion. Collectively, our results disclose a novel role of nuclear Rad in inhibiting the NFκB pathway function. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Rad (Ras associated with diabetes) [1] belongs to the RGK family of Ras-related small GTPases; other members of the family include Rem1, Rem2 and Gem/Kir [2–4]. RGK proteins share several characteristic structural and functional features that distinguish them from other Ras superfamily GTPases. RGK family members have nonconservative substitutions in the Ras-like core domain and therefore exhibit low intrinsic GTPase activities [5]. In addition, RGK proteins possess N- and C-terminal extensions, which contain phosphorylation sites and mediate the interaction with 14-3-3 proteins and calmodulin (CaM) [6–8]. Moreover, RGK GTPases lack lipid modifications and Abbreviations: Cavβ, calcium channel, voltage-dependent, beta subunit; CaM, calmodulin; EMSA, electrophoretic mobility shift assay; Gem, GTP binding protein overexpressed in skeletal muscle; Gmip, Gem interacting protein; GST, glutathione S-transferase; IκB, inhibitory κB; IKK, IκB kinase; MMP9, matrix metallopeptidase 9; NFκB, nuclear factor κB; NLS, nuclear localization signal; Rad, ras-related associated with diabetes; RelA, v-rel reticuloendotheliosis viral oncogene homolog A; Rem, Ras (Rad and Gem)-like GTPbinding; RHD, Rel homology domain; ROCK, Rho-associated protein kinase; TNFα, tumor necrosis factor alpha; VGCC, voltage-gated calcium channel. ⁎ Corresponding authors. Tel.: +886 2 2826 7269; fax: +886 2 2826 4843. E-mail addresses: [email protected] (B.-Y. Hsiao), [email protected] (M.-Y. Chen). 1 Present address: Genentech, San Francisco, CA, USA.

http://dx.doi.org/10.1016/j.cellsig.2014.03.003 0898-6568/© 2014 Elsevier Inc. All rights reserved.

their membrane association is directed by a conserved lipid-binding domain in the C-terminal extension [2]. Besides a well-established function in inhibiting the voltage-gated calcium channel (VGCC) by binding to its Cavβ subunit, RGK proteins also modulate cytoskeletal remodeling and cell shape determination [2–4]. For instance, overexpression of Gem/Kir can stimulate dendritelike protrusions in COS-1 cells, and Gem or Rad expression can induce neurite extensions in N1E-115 neuroblastoma cells [6,9]. The effect of RGK proteins on cytoskeletal organization is mainly through modulating Rho-dependent signaling. Rad and Gem have been shown to bind to the Rho-associated protein kinase (ROCK) and function as inhibitors of ROCK [9]. Gem can also negatively regulate the Rho pathway via recruiting the Rho GTPase-activating protein Gmip [5,10]. RGK proteins can be regulated by mechanisms other than the modulation of the GTPase cycle; their guanine-nucleotide exchange factors or GTPase-activating proteins have not been identified yet. Phosphorylation and protein–protein or protein–lipid interactions are involved in regulating the subcellular localization and functions of RGK proteins [2]. Rad is found to be distributed on the plasma membrane, in the cytoplasm, and in the nucleus [8,11]. The functional significance of different intracellular localizations of Rad has not been fully elucidated. A Cterminal truncation mutant of Rad which lacks the CaM binding region is accumulated in the nucleus and incapable of inducing cytoskeletal

1438

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

remodeling; however, the 14-3-3-bound Rad (which is cytoplasmic) also fails to induce cell shape alterations [12,13]. Rad contains multiple nuclear localization signals (NLS) which mediate its translocation to the nucleus by the importin-dependent pathway [12,13]. It is noted that RGK proteins in the brain and hippocampal neurons are predominantly localized in the nucleus [12], suggesting the possibility of important physiological roles of these nucleus-targeted small GTPases. However, except for the finding that nuclear targeting of Rad promotes nuclear sequestration of the Cavβ subunit of VGCC and downregulation of the Ca2+ channel activity [11], the biological function of Rad in the nucleus is not known. The NFκB-mediated signal transduction constitutes the central signaling network regulating a wide range of biological functions, and aberrant NFκB activation underlies various pathophysiological processes, including angiogenesis and invasion, which are hallmarks of cancer [14,15]. NFκB family members are structurally related transcription factors, including RelA/p65, RelB, c-Rel, p50 and p52. These proteins share a REL homology domain (RHD) that is responsible for DNA binding and dimerization, and they form hetero- or homo-dimers to recognize a consensus DNA sequence, termed the κB site, in the promoters of NFκB target genes [16,17]. The C-termini of RelA/p65, RelB and c-Rel, but not p50 or p52, contain the transactivation domain that is involved in interacting with basal transcription factors and cofactors [16,17]. NFκB proteins can shuttle between the cytoplasm and the nucleus [18]. In un-stimulated cells, NFκB dimers are tightly associated with IκB, the inhibitor of NFκB, which sequesters NFκB in an inactive state in the cytoplasm. Upon activation by external stimuli such as TNFα, IκB is phosphorylated by the IκB kinase (IKK) complex, resulting in subsequent ubiquitination and degradation of IκB and translocation of NFκB dimers into the nucleus [16]. Nuclear NFκB modulates the expression of an array of target genes that are involved in multiple cellular processes, including inflammation, proliferation, apoptosis and cell migration/invasion [19,20]. For example, one of the NFκB transcriptional targets is MMP9, a matrix metalloproteinase known to be responsible for NFκBmediated tumor cell invasion [21,22]. MMP9 contains a κB site in its promoter and is activated upon interleukin-1α or TNFα stimulation [23]. Studies have shown that ROCK is required for lysophosphatidic acidmediated stimulation of NFκB in human endothelial cells [24–26], suggesting a functional link between Rho signaling and NFκB activation. We have previously shown that Rad suppresses growth and migration/invasion of lung cancer cells via its ability to inhibit Rho signaling [27]. In this study, we further investigated whether Rad, a ROCK inhibitor, can interfere with the NFκB-mediated transcriptional activation. We provide evidence that the nucleus-localized Rad may inhibit the DNA binding and target gene transcriptional activation by NFκB through direct protein–protein interaction with RelA/p65. Our findings uncover a novel mechanism whereby NFκB signaling-dependent cellular functions may be regulated. 2. Materials and methods 2.1. Cell culture and reagents H1299 and H838 lung adenocarcinoma cells were cultured in RPMI1640 (Sigma) supplemented with 10% fetal calf serum (FCS), and 293T cells were grown in DMEM (Life Technologies) supplemented with 10% FCS. Transfection was performed using the Lipofectamine 2000 reagent (Life Technologies) according to the manufacturer's instructions. Plasmids for the expression of shRNAs targeting Rad (5′-AGGCATCACT CATGGTCTA-3′) or luciferase (5′-GACCAGGCATTCACAGAAA-3′) were obtained from the National RNAi Core Facility (at the Institute of Molecular Biology/Genomic Research Center, Academia Sinica), which is supported by the National Core Facility Program for Biotechnology, Taiwan. Lentivirus-based RNA interference was performed according to the instructions of the National RNAi Core. The plasmid expressing a Rad truncation mutant lacking the CaM binding domain (ΔCaM-BD) was

constructed as the “C-trun 246” mutant previously described [12]. TNFα was purchased from R&D Systems. Antibodies against p65, Myc, IκBα and B23 were obtained from Santa Cruz. Anti-Flag and antitubulin antibodies were from Sigma-Aldrich. 2.2. Luciferase reporter assay Cells were co-transfected with both a Firefly luciferase reporter construct containing 5 copies of the consensus κB site and a control Renilla luciferase (or a lacZ β-galatosidase) construct, with or without NFκB subunit-expressing and/or Rad-expressing plasmids. Cells were harvested 24 h after transfection and washed with PBS and lysed in a reporter lysis buffer (Promega). Luciferase activities in the extracts were assayed using the Luciferase Assay System (Promega). Relative luciferase activities were calculated by normalizing the Firefly luciferase activity with the activity of Renilla or β-galatosidase in the same sample. 2.3. Immunofluorescence cell staining Cells were fixed in 3.7% paraformaldehyde and permeabilized with 0.2% Triton X-100. After blocking in PBS containing 10% FCS, cells were incubated with primary antibodies against the Myc tag and subsequently with the TRITC-conjugated secondary antibodies (Jackson Lab). The cellular distribution of Myc-RelA/p65 was examined and scored under a confocal microscope (TCS SP, Leica). 2.4. Nuclear and cytoplasmic protein extraction Cells were washed with PBS and incubated in a lysis buffer containing 20 mM HEPES (pH 7.4), 10 mM KCl, 1 mM MgCl2, 0.5% NP-40, 0.5 mM DTT and a cocktail of protease inhibitors (Roche Applied Science). After centrifugation at 3000 ×g for 5 min, the supernatant was collected as the cytoplasmic fraction. For the nuclear fraction, the pellet was washed twice with PBS and then lysed in a buffer containing 20 mM HEPES (pH 7.4), 10 mM KCl, 1 mM MgCl2, 0.4 M NaCl, 20% glycerol, 0.5 mM DTT and protease inhibitors (Roche Applied Science). The mixture was incubated on ice for 30 min and centrifuged at 12,000 ×g for 10 min; the supernatant was collected as the nuclear fraction. Protein concentrations in the fractions were assessed using the BioRad protein assay kit (BioRad). Samples of cytoplasmic and nuclear fractions containing equal amounts of proteins were analyzed by immunoblotting. 2.5. Immunoprecipitation and GST pull-down assays For immunoprecipitation assays, lysates prepared from RelAoverexpressing cells were diluted in the IP buffer (50 mM Tris–HCl, pH7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 5 mM MgCl2 and 0.1% NP-40) to a final protein concentration of 0.6 mg/ml, and incubated with antibodies against Myc or Flag and Protein A Sepharose beads (Sigma-Aldrich) at 4 °C overnight. After centrifugation, proteins in the immunoprecipitates were examined by immunoblotting. For GST pulldown assays, the plasmid expressing Myc-His-RelA was transfected into 293T cells and lysates were prepared. Tagged RelA was purified using Ni-NTA affinity agarose beads (Qiagen) and incubated with recombinant GST or GST-Rad (from E. coli) immobilized on glutathioneSepharose beads for 18 h. After centrifugation, the GST pulled-down samples were analyzed for the Myc-His-p65 by immunoblotting. 2.6. Electrophoretic mobility shift assay (EMSA) EMSA was performed as described previously [27]. The oligonucleotide probe used in the assay contained two copies of the consensus p65/RelA-binding κB site (underlined): 5′-GATCGGGACTTTCCGCTG GGGACTTTCCGCTG-3′. Nuclear extracts were prepared from RelAoverexpressing cells and incubated with the 32P-labeled probe in the presence or absence of Rad-containing nuclear extracts or recombinant

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

GST-Rad purified from E. coli; the reaction was carried out in a buffer containing 17 mM HEPES, 15% glycerol, 0.42 mM EDTA, 0.3 mM DTT, 10 mM KCl, 6.25 mM MgCl2 and 12.5% poly(dI/dC). Protein–DNA interaction was analyzed by electrophoresis using a native 4% polyacrylamide gel, and signals were detected by autoradiography. 2.7. RNA isolation and reverse transcription PCR Total RNA was isolated using the TRIzol reagent (Invitrogen). The first-strand cDNA was synthesized using an oligo-dT primer and the SuperScript III reverse transcriptase (Invitrogen) according to the manufacturer's protocols. Primers used for gene-specific PCR amplification were: MMP9-Fw (5′-TACTGGCGATTCTCTGAGGG-3′), MMP9-Re (5′-ACCTGGTTCAACTCACTCCG-3′), GAPDH-Fw (5′-AAGTATGAC AACA GCCTCAAGA-3′), and GAPDH-Re (5′-CACCACCTTCTTGATGTCATCA-3′). 2.8. Gelatin zymography Conditioned media of cells cultured in the presence or absence of TNFα were collected and subjected to electrophoresis in a 10% SDSPAGE gel containing 0.1% gelatin (Sigma-Aldrich). After electrophoresis, the gel was immersed in a renaturing buffer containing 2.5% Triton X100 for 1 h, subsequently incubated with the developing buffer (50 mM Tris, 200 mM NaCl, 0.02% Brij-35, 0.01% NaN3, 5 mM CaCl2) at 37 °C for 18 h, and soaked in 0.1% Coomassie Blue R-250 (Merck & Co.) for 30 min. The gelatinase activity was revealed by the appearance of clear bands against the deep blue background. 2.9. Cell invasion assay The invasion capability of cells was assessed using Transwell with a polycarbonate membrane filter (8-μm pore; Millipore). Approximately 5 × 104 cells in a serum-free medium were placed in the upper chamber of a Matrigel (BD Biosciences)-coated Transwell, and a 10% FCScontaining medium was added in the lower chamber. After incubation for 24 h, the upper surface of the filter was wiped with a cotton swab to remove cells that did not migrate; cells that migrated to the lower surface of the filter were fixed in 3.7% paraformaldehyde, stained with crystal violet, and counted under a microscope. 3. Results 3.1. Rad inhibits the transcriptional activity of NFκB We explored whether Rad can affect the transcriptional activity of NFκB by promoter-reporter assays. A luciferase reporter construct containing five copies of the consensus κB site was transfected into H1299 cells in the presence or absence of expression plasmids for Rad and NFκB (subunit RelA/p65 or p50). The overexpression of Rad significantly inhibited the p65- and p50-stimulated luciferase activities (Fig. 1A). Both the basal and the p65-stimulated luciferase activities were inhibited by Rad in a dose-dependent manner (Fig. 1B). The inhibition of the NFκB transcriptional activity observed in cells transfected with the Rad-overexpressing plasmid could be alleviated by simultaneous knockdown of Rad (Fig. 1C), suggesting that the suppression was specifically due to increased Rad levels. We also demonstrated that the TNFαstimulated NFκB activity was significantly reduced by overexpression of Rad (Fig. 1D). Collectively, these results indicate that Rad can negatively regulate the unstimulated or the TNFα-stimulated NFκB transcriptional activation activity.

1439

expression affects the nuclear/cytoplasmic distribution of NFκB. Cells were transfected to overexpress Myc-tagged NFκB subunit p65 with or without the overexpression of EGFP-Rad, and examined by immunofluorescence staining and microscopy for the localization of Myc-p65. The percentage of cells with a predominantly cytoplasmic Myc-p65 distribution pattern increased from 20% in control EGFP-expressing cells to 42% in Rad-overexpressing cells (Fig. 2A). Further immunoblotting analysis of nuclear and cytoplasmic fractions of cell extracts found that silencing Rad expression increased the amount of endogenous RelA/p65 localized to the nucleus in both unstimulated and TNFα-treated cells (Fig. 2B). To elucidate how Rad may affect nuclear/cytoplasmic distributions of NFκB, the time course of changes in IκB levels following TNFα treatment was examined in control- and Rad-knockdown cells by immunoblotting; similar kinetics of TNFα-induced IκB degradation was observed in cells with or without Rad knockdown (Fig. 2C), indicating that Rad does not modulate the degradation of IκB in response to TNFα stimulation. Therefore, the effect of Rad on NFκB subcellular distribution is probably not achieved by impinging on the classical IκBdependent NFκB cytoplasmic sequestration mechanism. 3.3. The nuclear localization of Rad is important for its suppressive effect on NFκB We next investigated whether Rad can regulate the NFκB activity from within the nucleus. We employed a previously established Rad truncation mutant lacking the CaM binding domain (ΔCaM-BD), which resides predominantly in the nucleus [12]. Luciferase assays on cells transfected with the NFκB-responsive reporter construct showed that the ΔCaM-BD Rad mutant was able to suppress the transcriptional activity of NFκB similarly as the wild-type Rad did (Fig. 3A). Furthermore, expression of 14-3-3γ, which binds and retains Rad in the cytoplasm [11], significantly alleviated the inhibitory effect of Rad on NFκB-mediated transcriptional activation (Fig. 3B). These results together suggest that the nuclear localization of Rad is important for its suppressive effect on the NFκB transcriptional activity. 3.4. Rad binds to RelA/p65 and inhibits its interaction with DNA Next, the possibility that Rad interacts with NFκB was explored. Complex formation between Rad and RelA/p65 was analyzed in H1299 cells transfected with plasmids expressing Myc-p65 and Flag-Rad; both nuclear and cytoplasmic fractions from transfected cells were subjected to co-immunoprecipitation assays using epitope tag-specific antibodies. The results demonstrated the association between p65 and Rad primarily in the nucleus (Fig. 4A). GST pull-down assays using affinity purified recombinant GST or GST-Rad and His-Myc-p65 proteins further demonstrated that GST-Rad, but not GST, bound to Myc-p65, suggesting a direct interaction between Rad and RelA/p65 (Fig. 4B). We next addressed the possibility that Rad may interfere with the binding of p65 to DNA by electrophoretic mobility shift assays (EMSA). The 32P-labeled oligonucleotide probe carrying κB binding sites reacted with the p65-containing but not the control nuclear extracts (Fig. 4C); the specificity of the binding between the probe and p65 was demonstrated by the supershift of the protein/DNA band in the presence of a p65-specific antibody, and by the disappearance of the band in the presence of excess unlabeled probe. Notably, addition of Rad-containing nuclear extracts to the reaction mixtures reduced the abundance of the p65/DNA complex in a dose-dependent manner. Similarly, addition of purified recombinant GST-Rad, but not GST, inhibited the binding of p65 to the κB site-containing probe in a dose-dependent manner (Fig. 4D). Together, these results suggest that Rad binds to p65 and interferes with the p65-DNA binding.

3.2. Rad affects the subcellular distribution of NFκB 3.5. Depletion of Rad enhances MMP9 expression and cell invasion The activation of NFκB involves the degradation of the inhibitor protein IκB and the translocation of NFκB into the nucleus [16]. To investigate the mechanism of NFκB inhibition by Rad, we tested whether Rad

NFκB participates in multiple cellular functions, including those involved in cancer progression [15], by regulating the expression of its

1440

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

*

*

14 12 10 8 6 4 2 0

Myc-p65 Flag-Rad

2.5

Relative luciferase activity

Relative luciferase activity

A

-

+ -

+ +

2 1.5 1 0.5 0

Myc-p50 Flag-Rad

-

+ -

+ +

1.2

Relative luciferase activity

Relative luciferase activity

B 1 0.8 0.6 0.4 0.2 0

1.2 1 0.8 0.6 0.4 0.2 0

Myc-p65 Control

+

+

+

Control

D

*

1.2

Relative luciferase activity

Relative luciferase activity

C 1 0.8 0.6 0.4 0.2 0

Flag-Rad Sh-Rad

-

+

HA-Rad

+ -

+ +

Flag-Rad

*

8 7 6 5 4 3 2 1 0

TNFα -

+ Control

-

+

Flag-Rad

Fig. 1. Rad inhibits the transcriptional activity of NFκB. (A-D) Promoter-reporter assays. H1299 cells were co-transfected with a luciferase reporter construct containing 5 copies of the κB site and a Renilla luciferase- or lacZ-expressing plasmid, together with plasmids for the expression of Myc-p65 or p50 and/or epitope-tagged Rad. Lysates prepared 24 h after transfection were assayed. Shown are relative luciferase activities normalized to Renilla or β-galactosidase activites. Bars represent means ± SD from three independent experiments. *, p b 0.05. (A) Rad inhibits the NFκB transcriptional activity stimulated by exogenous p65 or p50 expression. (B) Rad dose-dependently reduces the endogenous and p65-stimulated NFκB transcriptional activity. Increasing amounts (0.05, 0.2 and 1 μg) of plasmids for Rad expression were co-transfected together with constructs for reporter assays into cells. (C) Suppression of the NFκB activity by exogenous Rad expression is neutralized by the simultaneous expression of Rad-specific shRNA. (D) Rad down-regulates the TNFα-stimulated NFκB activity. Transfected cells were treated with or without TNFα (10 ng/ml) in a serum-free medium for 12 h and assayed for luciferase activities.

target genes. We performed microarray gene expression profiling and found that, consistent with the finding that Rad expression inhibited the transcriptional activity of NFκB, knockdown of Rad expression in cells resulted in changes of mRNA levels of an array of NFκB target genes, including the upregulation of MMP9 (Supplementary data). RT-PCR analysis verified that both the basal and TNFαstimulated MMP9 mRNA levels were increased in Rad-knockdown cells compared to the control (Fig. 5A). Gelatin zymography analysis further demonstrated that, whether cells were treated with TNFα or not, the MMP9 activity was increased in the conditioned medium of Rad-deficient cells compared to that of the control cells (Fig. 5B). Transwell in vitro assays showed that, either with or without TNFα stimulation, the invasion ability of Rad-knockdown cells was higher than that of the control cells (Fig. 5C). Together, these results demonstrated that Rad deficiency, as we have observed previously in

lung cancer cells [27], could promote TNFα/NFκB-mediated functions that are involved in cancer progression. 4. Discussion Aberrant activation of the transcription factor NFκB occurs in many cancers and has been associated with poor prognosis for survival [15, 28,29]. In this study, we have discovered a novel mechanism for inhibiting the NFκB pathway. We show that, as summarized in Fig. 6, the RGK GTPase Rad can serve a previously undiscovered nuclear function in suppressing the transcriptional activity of NFκB. Our data do not support that the Rad-dependent modulation of NFκB activity impinges on the canonical pathway of NFκB regulation which involves IκB degradation. Instead, we have demonstrated that RelA/p65 is a novel interacting partner of Rad, and Rad can interfere with the binding of

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

A

Myc-p65

DAPI

1441

Myc-p65/DAPI

Cells with cytoplasmic localization of p65 (%)

EGFP

EGFP-Rad

*

50 40 30 20 10 0

EGFP EGFP -Rad

B

C

+ TNFα

C

N

C

N

C

N

C

Sh-Rad

Sh-Luc

Sh-Luc Sh-Rad Sh-Luc Sh-Rad TNFα

N

p65

IκB

Rad

Rad

Tubulin

p65

B23

Tubulin

0

15 30 45 60

15 30 45 60 (min)

0

Fig. 2. Rad affects the subcellular distribution of NFκB. (A) Overexpression of Rad increases the cytoplasmic localization of overexpressed Myc-p65. H1299 cells were co-transfected by plasmids to express Myc-p65 and EGFP-Rad or EGFP for 24 h, and examined by immunofluorescence staining using an anti-Myc antibody and fluorescence microscopy. Nuclei were stained with DAPI. A total of 300 cells were scored in each experiment and the percentage of cells with cytoplasmic Myc-p65 signals was determined. Quantitative results (means ± SD) from three independent experiments are shown on the right. *, p b 0.05. (B) Depletion of Rad increases the abundance of endogenous p65 in nuclear fractions. H838 cells were infected by lentiviral particles to express Rad-specific (shRad) or control (shLuc) shRNAs, and stimulated with or without TNFα (10 ng/ml) for 1 h. Cytoplasmic and nuclear extracts were examined by immunoblotting to detect p65 and Rad. Tubulin and B23 were used as markers for cytoplasmic and nuclear fractions, respectively. (C) Knockdown of Rad expression does not affect the kinetics of TNFα-induced IκB degradation. Cells were treated as in (B); lysates were prepared at indicated time points after TNFα addition, and analyzed by immunoblotting.

p65 to the κB site DNA sequence. Consistent with the notion that Rad acts through direct protein–protein interaction with p65, the nuclear localization of Rad is essential for its suppressive effect on the transcriptional activity of NFκB. Activity of the NFκB pathway is modulated by a plethora of regulatory proteins and upstream signaling events, including crosstalk with the tumor suppressor p53 at multiple levels of the pathway [15]. We have previously shown that Rad is a direct transcriptional target of p53 and mediates part of the tumor-suppressive functions of p53 by suppressing the migration and invasiveness of cancer cells [27]. Our current findings suggest that Rad inhibits the transcriptional function of NFκB by binding

A

*

*

1.4

1

Relative luciferase activity

Relative luciferase activity

B

*

1.2

0.8 0.6 0.4 0.2

*

1.2 1 0.8 0.6 0.4 0.2 0

0

Flag-Rad

to RelA/p65 in the nucleus, demonstrating another level of crosstalk between p53 and the NFκB signaling pathway. NFκB activation generally reprograms the cell to a gene expression pattern favoring cancer progression, which incorporates the upregulation of matrix metalloproteinases to promote invasion/metastasis [29]. Therefore, besides the effects of Rad on the Rho/ROCK pathway and actin dynamics [9], the inhibition of NFκB activity by Rad may also contribute to the p53mediated suppression of cancer cell migration and invasion. Although the nucleocytoplasmic shuttling of Rad is well-known [2, 12], the function of Rad in the nucleus has not been fully explored. Our report presents a novel nuclear function of Rad, providing evidence

-

WT ΔCaM-BD

Flag-Rad

-

-

+

+

14-3-3γ

-

+

-

+

Fig. 3. The nuclear localization of Rad is important for its suppressive effect on NFκB. The transcriptional activity of NFκB was assessed by luciferase assays as in Fig. 1. Shown are relative luciferase activities (means ± SD) obtained from three independent experiments. *, p b 0.05. (A) A Rad mutant that predominantly localizes to the nucleus suppresses the transcriptional activity of NFκB. H1299 cells were co-transfected with constructs for promoter-reporter assays as in Fig. 1 along with a plasmid to express Flag-tagged wild-type (WT) Rad or ΔCaM-BD Rad (a C-terminal truncation mutant lacking the CaM-binding domain). (B) Overexpression of 14-3-3γ relieves Rad-mediated inhibition of NFκB. Cells co-transfected with luciferase constructs and plasmids expressing Flag-Rad and/or 14-3-3γ were assayed.

1442

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

Pull-down

A

B

Myc-p65 + Flag-Rad Nucleus

Cytoplasm

Input

IP

IP Input IgG Flag Myc

Input IgG Flag Myc

IB: Myc

Myc Flag

C p65 N.E. (5 μg) Rad N.E. (μg) Control N.E. (μg) Anti-p65 Ab Cold probe

GST-Rad

Coomassie blue staining

IB

GST

D -

+ -

+ + + + + + + - 0 1 3 5 - - 5 4 2 0 - - - - - - + - - - - - - +

p65 N.E. (7 μg) GST-Rad (μg) GST (μg) Anti-p65 Ab IgG Cold probe

-

+ + + + + 0 0.075 0.75 1.5 1.5 1.425 0.75 0 - - - - + - - - - - - - - -

+ + -

+ +

Fig. 4. Rad binds to RelA/p65 and inhibits NFκB binding to DNA. (A) Rad interacts with p65 mainly in the nucleus. Nuclear and cytoplasmic fractions of lysates from 293 T cells overexpressing Myc-p65 and Flag-Rad were subjected to immunoprecipitation (IP)/immunoblotting (IB) using the indicated antibodies. (B) Rad binds to p65 directly. GST pull-down assays were performed using GST-Rad (or GST as a negative control) purified from E. coli and His-Myc-p65 purified from transfected 293 T cells. (C-D) Rad inhibits p65 binding to the NFκB response element. EMSA was performed using nuclear extracts (N.E.) prepared from 293 T cells overexpressing p65. Solid and open arrows indicate positions of the p65/DNA complex and the antibody-bound p65/DNA complex, respectively. (C) A 32P-labeled oligonucleotide probe containing the NFκB response element was incubated with p65-containing N.E. in the presence of increasing amounts of Rad-containing or control N.E. with or without an anti-p65 antibody, as indicated. The unlabeled oligonucleotide (Cold probe) was added to assess the binding specificity. (D) EMSA was performed as in (C) except that GST-Rad or GST purified from E. coli was added in place of Rad-containing N.E. Mouse IgG was used as a negative control.

suggesting that the molecular action of Rad on NFκB primarily takes place in the nucleus. Although the Rad–p65 complex can be detected in both the cytoplasm and the nucleus, the interaction between these two proteins is much more evident in the nucleus. The nuclear Rad is necessary and sufficient for repressing the NFκB activity; a nucleuslocalized Rad mutant still inhibits the NFκB activity, whereas retaining Rad in the cytoplasm by overexpressing the scaffolding protein 14-33γ neutralizes the suppressive function of Rad. To our best knowledge, Rad represents the first example of a RGK GTPase capable of directly binding and inhibiting a transcription factor in the nucleus. Other RGK proteins also exhibit nucleocytoplasmic shuttling [2,13]; it remains to be established whether the transcription regulatory function through binding to transcription factors in the nucleus is universal among RGK family members. Interestingly, another Ras superfamily GTPase, κBRas, also negatively regulates NFκB; two mechanisms have been suggested: one involves the binding of κB-Ras to IκB (which suppresses proteosomal degradation of IκB), and the other is by interrupting the interaction between NFκB and p300/CBP (through the inhibition of RelA/ p65 phosphorylation) to suppress the NFκB-mediated transcriptional activation [30,31]. The suppressive effect of nuclear Rad on the DNAbinding ability of RelA/p65 represents a novel mechanism for Ras superfamily GTPase to inhibit the NFκB activity.

It is currently not clear how the interaction of Rad with RelA/p65 may inhibit the binding of NFκB to DNA. One possible mechanism is that Rad might interact with the RHD of RelA/p65 that is responsible for DNA binding and thereby physically hinder its interaction with the DNA response element. Our observation that Rad can suppress both p50 and p65-mediated transactivation is consistent with the above scenario as RHD is present in all NFκB subunits [15]. Alternatively, binding of Rad might induce a change of RelA/p65 into a conformation that is not able to bind DNA efficiently. It requires further mapping of the Rad-RelA/p65 interacting domains and structural analysis of the Rad-RelA/p65 complex to address these issues. We show that the nuclear/cytoplasmic distribution of NFκB can be affected by manipulation of Rad levels; the underlying mechanism remains unclear at present. Our data do not support that Rad affects NFκB localization via modulating cytoplasmic IκB levels; despite a significantly higher percentage of cells with cytoplasmic RelA/p65 signals in Rad-overexpressing cells and an evident increase of the amounts of nucleus-targeted p65 in the Rad-depleted cells, the TNFα-stimulated IκB degradation follows similar kinetics in the control and Rad-knockdown cells. A few possible mechanisms are listed below: 1) It is believed that the nuclear localization signals (NLS) in the NFκB subunits and the nuclear export signals (NES) in IκB

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

C Sh-Luc

Sh-Rad

-

-

+

+

TNFα MMP9

- TNFα

+ TNFα

Sh-Luc

A

1443

Sh-Rad

GAPDH

Sh-Luc

Sh-Rad

-

-

+

+

TNFα MMP9

Relative invasion ability

B

5

*

4

*

3 2 1 0

-

-

+

Sh-Luc

+

TNFα

Sh-Rad

Fig. 5. Knockdown of Rad expression increases MMP9 expression and cell invasion. H838 cells infected with lentivirus expressing Rad-specific (Sh-Rad) or luciferase-specific (Sh-Luc) shRNAs were cultured in a serum-free medium for 24 h and treated with or without TNFα (10 ng/ml) for 12 h. (A-B) Knockdown of Rad expression increases the basal and TNFαstimulated MMP9 expression. (A) Total RNA was prepared and analyzed by RT-PCR for MMP9 expression. GAPDH was used as an internal control. (B) Conditioned media of cells cultured for 24 h were collected and subjected to gelatin zymography. (C) Rad depletion inhibits basal and TNFα-stimulated cell invasion capability. Cells were analyzed by Matrigel Transwell invasion assays. Cells that invaded through the Matrigel were stained with crystal violet and representative micrographs are shown. Quantitative results from three independent experiments are shown in the lower panel; bars are means ± SD. *, p b 0.05.

contribute to the translocation of the NFκB-IκB complex in and out of the nucleus [17,18]. Perhaps the interaction of Rad with RelA/p65 might somehow favor the association of NFκB heterodimers with IκB and the nuclear export of the complex. 2) Both Rad [11] and RelA/p65 [32] have been reported to interact with the 14-3-3 scaffolding protein which is known to bind and retain specific cellular proteins in the cytoplasm. Binding of Rad to RelA/p65 might increase the ability of either or both proteins to complex with 14-3-3 and

thereby facilitate cytoplasmic distribution of NFκB. 3) Evidence obtained in the endothelial cells has linked the actin cytoskeleton to the nuclear translocation of NFκB and highlighted the regulatory role of the RhoA-Cofilin 1 pathway [33,34]. Rad can modulate the Rho pathway by inhibiting ROCK [9] and affect the phosphorylation and activity of Cofilin [27]. Therefore, there exists an intriguing possibility that Rad regulates NFκB localization through its effects on the Rho/ROCK/Cofilin/actin axis.

Signal (e.g. TNFα)

Signal (e.g. TNFα)

Receptor

Receptor CYTOSOL

CYTOSOL

Proteosomal degradation

IKK

p

IκBα p50 p65

IκBα p Rad

Proteosomal degradation

IKK

p

IκBα p50 p65

IκBα p

p50?

p65

Rad

Rad Rad

? Rad NUCLEUS

p50 p65

κB site binding

p50?

p65

Rad

Rad

Rad

NUCLEUS Target gene expression

Low or no Rad expression

No κB site binding

No target gene expression

Rad overexpression

Fig. 6. A schematic summary of Rad-mediated inhibition of the NFκB pathway. In the canonical pathway of NFκB signaling, extracellular signal such as TNFα interacts with its receptor, leading to the activation of IKK. IKK phosphorylates IκBα and promotes its ubiquitination and proteosomal degradation, resulting in the release of the NFκB heterodimer (RelA/p65 and p50) from the cytoplasmic NFκB-IκBα complex. NFκB translocates into the nucleus, binds to the κB site DNA response element and regulates target gene expression. Rad can shuttle between the cytoplasm and the nucleus. Overexpression of Rad exerts an inhibitory effect on NFκB in the nucleus: Rad binds to RelA/p65 and prevents NFκB from binding to the κB site, abolishing the NFκB-mediated transcriptional activation of target genes. It needs to be determined whether the binding of Rad to RelA/p65 disrupts the RelA/p65-p50 interaction. How Rad overexpression may increase the cytoplasmic localization of RelA/p65 is also currently unknown.

1444

B.-Y. Hsiao et al. / Cellular Signalling 26 (2014) 1437–1444

5. Conclusion In conclusion, we have demonstrated a functional link between the small GTPase Rad and the NFκB pathway. Rad suppresses the NFκBmediated transcription and downstream gene activation. This suppressive effect of Rad does not appear to involve the IκB-mediated NFκB cytoplasmic sequestration mechanism, but requires the nuclear localization of Rad. Rad can directly bind to RelA/p65 and suppress the interaction of RelA/p65 with the NFκB response DNA element. Deficiency of Rad in cancer cells increases the expression of TNFα-stimulated NFκB target genes such as MMP9 and promotes invasion. Our findings uncover a previously unknown function of Rad and a novel regulatory mechanism for fine-tuning the NFκB pathway, which may potentially be exploited to develop new therapeutic interventions for cancers. Acknowledgments We thank Professor F.-F. Wang for valuable suggestions throughout this study and critical reading of the manuscript. This work was partly supported by the “Aim for the Top University Plan” (99AC-T512, 100AC-T506 and 103AC-T307) from the Ministry of Education, Taiwan. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cellsig.2014.03.003. References [1] C. Reynet, C.R. Kahn, Science 262 (1993) 1441–1444. [2] R.N. Correll, C. Pang, D.M. Niedowicz, B.S. Finlin, D.A. Andres, Cell. Signal. 20 (2008) 292–300. [3] K. Kelly, Trends Cell Biol. 15 (2005) 640–643. [4] T. Yang, H.M. Colecraft, Biochim. Biophys. Acta 1828 (2012) 1644–1654. [5] L. Cohen, R. Mohr, Y.Y. Chen, M. Huang, R. Kato, D. Dorin, et al., Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 12448–12452.

[6] P. Beguin, R.N. Mahalakshmi, K. Nagashima, D.H. Cher, A. Takahashi, Y. Yamada, et al., J. Cell Sci. 118 (2005) 1923–1934. [7] R. Fischer, Y. Wei, J. Anagli, M.W. Berchtold, J. Biol. Chem. 271 (1996) 25067–25070. [8] J.S. Moyers, P.J. Bilan, J. Zhu, C.R. Kahn, J. Biol. Chem. 272 (1997) 11832–11839. [9] Y. Ward, S.F. Yap, V. Ravichandran, F. Matsumura, M. Ito, B. Spinelli, et al., J. Cell Biol. 157 (2002) 291–302. [10] A. Hatzoglou, I. Ader, A. Splingard, J. Flanders, E. Saade, I. Leroy, et al., Mol. Biol. Cell 18 (2007) 1242–1252. [11] P. Beguin, R.N. Mahalakshmi, K. Nagashima, D.H. Cher, H. Ikeda, Y. Yamada, et al., J. Mol. Biol. 355 (2006) 34–46. [12] R.N. Mahalakshmi, M.Y. Ng, K. Guo, Z. Qi, W. Hunziker, P. Beguin, Traffic 8 (2007) 1164–1178. [13] R.N. Mahalakshmi, K. Nagashima, M.Y. Ng, N. Inagaki, W. Hunziker, P. Beguin, Traffic 8 (2007) 1150–1163. [14] M. Karin, Y. Cao, F.R. Greten, Z.W. Li, Nat. Rev. Cancer 2 (2002) 301–310. [15] N.D. Perkins, Nat. Rev. Cancer 12 (2012) 121–132. [16] M.S. Hayden, S. Ghosh, Cell 132 (2008) 344–362. [17] M.S. Hayden, S. Ghosh, Genes Dev. 26 (2012) 203–234. [18] S. Ghosh, M. Karin, Cell 109 (2002) S81–S96 (Suppl.). [19] W.E. Naugler, M. Karin, Curr. Opin. Genet. Dev. 18 (2008) 19–26. [20] A. Oeckinghaus, S. Ghosh, Cold Spring Harb. Perspect. Biol. 1 (2009) a000034. [21] H. Sato, M. Seiki, Oncogene 8 (1993) 395–405. [22] A.R. Farina, A. Tacconelli, A. Vacca, M. Maroder, A. Gulino, A.R. Mackay, Cell Growth Differ. 10 (1999) 353–367. [23] M. Bond, R.P. Fabunmi, A.H. Baker, A.C. Newby, FEBS Lett. 435 (1998) 29–34. [24] K.N. Anwar, F. Fazal, A.B. Malik, A. Rahman, J. Immunol. 173 (2004) 6965–6972. [25] S.A. Benitah, P.F. Valeron, J.C. Lacal, Mol. Biol. Cell 14 (2003) 3041–3054. [26] H. Shimada, L.E. Rajagopalan, J. Biol. Chem. 285 (2010) 12536–12542. [27] B.Y. Hsiao, C.C. Chen, P.C. Hsieh, T.K. Chang, Y.C. Yeh, Y.C. Wu, et al., J. Mol. Med. (Berl.) 89 (2011) 481–492. [28] J. Li, H. Jia, L. Xie, X. Wang, H. He, Y. Lin, et al., Int. J. Gynecol. Cancer 19 (2009) 1421–1426. [29] M.M. Chaturvedi, B. Sung, V.R. Yadav, R. Kannappan, B.B. Aggarwal, Oncogene 30 (2011) 1615–1630. [30] C. Fenwick, S.Y. Na, R.E. Voll, H. Zhong, S.Y. Im, J.W. Lee, et al., Science 287 (2000) 869–873. [31] K. Tago, M. Funakoshi-Tago, M. Sakinawa, N. Mizuno, H. Itoh, J. Biol. Chem. 285 (2010) 30622–30633. [32] C. Aguilera, V. Fernandez-Majada, J. Ingles-Esteve, V. Rodilla, A. Bigas, L. Espinosa, J. Cell Sci. 119 (2006) 3695–3704. [33] F. Fazal, M. Minhajuddin, K.M. Bijli, J.L. McGrath, A. Rahman, J. Biol. Chem. 282 (2007) 3940–3950. [34] F. Fazal, K.M. Bijli, M. Minhajuddin, T. Rein, J.N. Finkelstein, A. Rahman, J. Biol. Chem. 284 (2009) 21047–21056