SAG/ROC2/Rbx2 Is a Novel Activator Protein-1 Target That Promotes C-Jun Degradation and Inhibits 12-O-Tetradecanoylphorbol-13- A

Research Article

SAG/ROC2/Rbx2 Is a Novel Activator Protein-1 Target that Promotes c-Jun Degradation and Inhibits 12-O-Tetradecanoylphorbol-13- Acetate–Induced Neoplastic Transformation

Qingyang Gu, Mingjia Tan, and Yi Sun

Department of Radiation Oncology, University of Michigan Comprehensive Cancer Center, Ann Arbor, Michigan

Abstract cycle, growth arrest, and apoptosis (1–3). The analysis of cells and SAG (sensitive to apoptosis gene) was first identified as a mice deficient in individual AP-1 proteins has identified several stress-responsive protein that, when overexpressed, inhibited physiologically relevant AP-1 target genes, including cyclin D1, p16, apoptosis both in vitro and in vivo. SAG was later found to be p19, p53, p21, and Fas ligand (1). Activation of cyclin D1 and the second family member of ROC1 or Rbx1, a RING inhibition of p53 and p21 would promote cell cycle progression component of SCF and DCX E3 ubiquitin ligases. We report from G1 to S to induce cell proliferation. here that SAG/ROC2/Rbx2 is a novel transcriptional target AP-1 was first considered as a mediator of tumor promotion by of activator protein-1 (AP-1). AP-1 bound both in vitro and its ability to alter gene expression in response to tumor promoters, in vivo to two consensus binding sites in a 1.3-kb region of such as TPA and UV irradiation (3). Indeed, TPA, UV, as well as the mouse SAG promoter. The SAG promoter activity, as mea- reactive oxygen species activate AP-1 (4–6). In mouse epidermal JB6 cell variants, representing from early to late stages of tumor sured by luciferase reporter assay, was dependent on these sites. Consistently, endogenous SAG is induced by 12-O- promotion, the AP-1 activity is progressively elevated (7). On the tetradecanoylphorbol-13-acetate (TPA) with an induction time other hand, cell growth and neoplastic transformation was sig- course following the c-Jun induction in both mouse epidermal nificantly impaired on blockage of AP-1 activity by either phar- JB6-Cl.41 and human 293 cells. TPA-mediated SAG induction macologic inhibitors, such as glucocorticoid, retinoic acid, or was significantly reduced in JB6-Cl.41 cells overexpressing a molecular biological inhibitors, including dominant-negative c-Jun dominant-negative c-Jun, indicating a requirement of c-Jun/ (TAM67), dominant-negative phosphatidylinositol-3 kinase (p85), AP-1. On the other hand, SAG seemed to modulate the c-Jun and dominant-negative extracellular signal-regulated kinase 2 levels. When overexpressed, SAG remarkably reduced both (Erk2;refs. 8–13). Moreover, acquisition of a tumor promotion- basal and TPA-induced c-Jun levels, whereas SAG small resistant phenotype is associated with a loss of responsiveness interfering RNA (siRNA) silencing increased substantially the to tumor promoter-induced AP-1 activation (14), whereas rescuing levels of both basal and TPA-induced c-Jun. Consistently, SAG such a response by introducing wild-type (WT) Erk2 converts tumor siRNA silencing reduced c-Jun polyubiquitination and blocked promotion-resistant phenotype to tumor promotion-sensitive phenotype (15). c-Jun degradation induced by Fbw7, an F-box protein of SCF E3 ubiquitin ligase. Finally, SAG overexpression inhibited, SAG (sensitive to apoptosis gene) was initially cloned by whereas SAG siRNA silencing enhanced, respectively, the TPA- differential display as a redox-inducible gene that encodes an induced neoplastic transformation in JB6-Cl.41 preneoplastic evolutionarily conserved RING finger protein (16, 17). Further model. Thus, AP-1/SAG establishes an autofeedback loop, in characterization of SAG revealed that SAG is a dual-function which on induction by AP-1, SAG promotes c-Jun ubiquitina- protein with antioxidant activity when acting alone or E3 ubiquitin tion and degradation, thus inhibiting tumor-promoting ligase activity when complexed with other ligase components activity of AP-1. [Cancer Res 2007;67(8):3616–25] (17, 18). Recently, ROC1/Rbx1/Hrt1, a family member of SAG and a component of SCF (Skp1, cullin1, and F-box protein) and DCX (DDB1, cullin 4A, and X-box protein) E3 ubiquitin ligases was Introduction implicated in ubiquitination and degradation of c-Jun (19–23). Activator protein-1 (AP-1) transcription factor is a dimer However, it has not been shown previously that SAG, through consisting of 18 different combinations of Jun-Jun or Jun-Fos binding to Cul-1 and Cul-4A (20), will replace ROC1/Rbx1 to form proteins. The Jun family of proteins includes c-Jun, JunB, and JunD, a SAG-SCF E3 ligase and promote c-Jun degradation. We report whereas the Fos family of proteins includes c-Fos, FosB, Fra-1, and here that SAG is induced by TPA via AP-1–mediated trans- Fra-2 (1). The c-jun and c-fos genes are inducible by a broad range activation. On the other hand, AP-1 component, c-Jun, is a SAG of extracellular stimuli. On activation, AP-1 binds to 12-O- substrate. When overexpressed, SAG remarkably reduced both tetradecanoylphorbol-13-acetate (TPA) response elements or AP-1 basal and TPA-induced c-Jun levels, whereas SAG small interfer- binding sites, 5¶-TGAG/CTCA-3¶, to transactivate many effector ing RNA (siRNA) silencing significantly increased the levels of genes, thus regulating cell proliferation, tumor promotion, cell both endogenous and TPA-induced c-Jun. SAG silencing inhibited c-Jun polyubiquitination and blocked Fbw7-induced c-Jun degradation. Finally, SAG overexpression within physiologic levels inhibited, whereas SAG silencing enhanced, the TPA-induced Requests for reprints: Yi Sun, Department of Radiation Oncology, University of neoplastic transformation in JB6-Cl.41 cells. Thus, it seems that Michigan, 4304 CCGC, 1500 East Medical Center Drive, Ann Arbor, MI 48109-0936. c-Jun-SAG constitutes a negative feedback loop as a cellular Phone: 734-615-1989;Fax: 734-647-9654;E-mail: [email protected]. I2007 American Association for Cancer Research. defensive mechanism to minimize the prolonged effect of AP-1 doi:10.1158/0008-5472.CAN-06-4020 activation.

Cancer Res 2007; 67: (8). April 15, 2007 3616 www.aacrjournals.org

Materials and Methods (5¶-GAAACTTGATTTGTAGACCA-3¶), Sag-AP1-2 (5¶-GCCTTATAACTCACAA- GGAT-3¶), Sag-AP1-3 (5¶-GAGGGATGACTCAGTGGTTA-3¶), Sag-AP1-3-MU Cell culture. Human HeLa, 293, and H1299 cells were cultured in DMEM (5¶-GAGGGAAGACTCAGTGGTTA-3¶), Sag-AP1-4 (5¶-TTTGTTTGTTTGCTTG- containing 10% fetal bovine serum (FBS). Mouse JB6-Cl.41 cells were CTTGC-3¶), Sag-AP1-5 (5¶-AAGGTGTGAGCTACCACGCC-3¶), Sag-AP1-6 (5¶- cultured in DMEM containing 5% FBS. JB6-Cl.41 cells expressing TAM67 CCTCGTGACGTCACTGGCGC-3¶), Sag-AP1-6-Mu (5¶-CCTCGAGACGT- A (24) were cultured in 5% DMEM containing G418 (400 g/mL). CACTGGCGC-3¶), and Sag-AP1-7 (5¶-GCGCCATCCAATCATCGCCGT-3¶). The Promoter cloning, luciferase reporter construction, and luciferase 32 paired oligonucleotides were annealed and labeled with P using T4 activity assay. 32 A 1.3-kb fragment upstream of mouse SAG translational polynucleotide kinase and [g- P]ATP. Nuclear extracts were prepared from ¶ initiation site was cloned by PCR with primers mSAG-P01 (5 -GGGGTAC- JB6-Cl.41 cells treated with TPA (20 ng/mL) for 16 h and subjected to gel ¶ ¶ CAGCCTTGGTTGCCCGAAAC-3 ) and mSAG-P02 (5 -CCGCTCGAGGGCG- retardation assay as described (27). In some experiments, the c-Jun antibody ¶ GCGGCGCAGAACGG-3 ). The fragment was then gel purified and subcloned (Santa Cruz Biotechnology, Santa Cruz, CA) or 50 excess of cold specific or into pGL3-basic luciferase reporter (Promega, Madison, WI) at the KpnI/XhoI nonspecific oligonucleotide was included to determine the binding specificity. or KpnI/HindIII site with subsequent sequencing confirmation as described Chromatin immunoprecipitation assay. JB6-Cl.41 cells were left (25). The luciferase reporter was designated as mSAG-P1.3 or 7AP (for untreated or treated with TPA (20 ng/mL) for 12 h before being subjected containing seven putative AP-1 binding sites). This construct was used as the to chromatin immunoprecipitation assay (ChIP) analysis, according to the template to generate a series of deletion or site-directed mutants, including protocol of Upstate Biotechnology, Inc. (Lake Placid, NY) as described ¶ 5AP (containing five AP-1 sites) with the primer set of AP1-3-KpnI(5- (26, 28). Input samples (one tenth of original DNA lysates) or samples ¶ GGTACCGAGGGATGACTCAGTGGT-3 ) and mSAG-P02;4AP (containing precipitated with no antibody (Ab) or IgG for negative controls or c-Jun ¶ four AP-1 sites) with primer set of D3 (5 -GGTACCGTGGTTAAAAC- antibody (c-Jun) were PCR amplified. The primer sequences for the third ¶ CATGGTCTG-3 ) and mSAG-P02;2AP (containing two AP-1 sites) with the AP-1 binding site (WT3) are mSAG-p-ChIP-3-01, 5¶-GTTCTTGCCCTTCT- ¶ ¶ primer set of AP1-6-KpnI(5-GGTACCCCTCGTGACGTCACTGGC-3 ) and GCTTACC-3¶, and mSAG-p-ChIP-3-02, 5¶-TGCTGTGAGGTGTCAGATTC-3¶, mSAG-P02;1AP (containing one AP-1 site) with the primer set of AP1-7- KpnI to generate a 215-bp fragment. The primer sequences for the sixth AP-1 ¶ ¶ (5 -GGTACCGCGCCATCCAATCATCGC-3 ) and mSAG-P02;and 0AP (with all binding site are mSAG-p-ChIP-6-01, 5¶-TGTAACTCCAGACAATGCTC-3¶, ¶ AP-1 sites deleted) with the primer set of D5 (5 -GGTACCTCGCCGTCT- and mSAG-p-ChIP-6-02, 5¶-AGCCAGACGGCGATGATTGG-3¶, to generate a ¶ ! GGCTCCGCCCG-3 ) and mSAG-P02. A T A substitution at AP1-3 and AP1-6 115-bp fragment. primers was introduced to generate Mu-5AP and Mu-2AP constructs with Northern and Western blotting analyses. The cells were treated with AP-1 binding site abolished. To generate the 1.3-kb promoter construct with TPA (20 ng/mL) for various periods up to 24 h and subjected to Northern AP-1 sites 3 and 6 mutated (Mu-3,6), the QuickChange Multi Site-Directed analysis as described previously (29). For Western blotting analysis, cells Mutagenesiskit(Stratagene,LaJolla,CA)wasused,followingthe were lysed in a Triton X-100 lysis buffer [20 mmol/L Tris-HCl (pH 8.0), manufacturer’s instruction. The sequence of the primers used is Sag-AP1- 150 mmol/L NaCl, 1% Triton X-100, 5 mmol/L EGTA, and 5 mmol/L EDTA] ¶ ¶ mu3-01 (5 -TGTAGTGTCTGGAGGGAAGACTCAGTGGTTAAAAC-3 )and with freshly added protease inhibitor tablet (Roche, Indianapolis, IN) for Sag-AP1-mu6-01 (5¶-CCCAGCCCCTCGAGACGTCACTGGC-3¶). To measure 30 min on ice followed by centrifugation for 30 min. Supernatants were the promoter activity of the mSAG-P1.3 and its deletion or site-directed measured for protein concentration using a Bio-Rad protein assay reagent mutants, HeLa cells were seeded in a 96-well plate at a density of 18,000 cells (Bio-Rad, Hercules, CA) and subjected to Western blotting (17) using per well. Cells were transiently cotransfected, using TR01 transfection antibodies against c-Jun, c-Fos (Santa Cruz Biotechnology), SAG, or ROC1/ reagent (America Pharma Source, Gaithersburg, MD), with these luciferase Rbx1. The h-actin (Sigma, St. Louis, MO) was used as a loading control. reporters individually, along with a plasmid expressing Renilla luciferase Lentivirus-based SAG overexpression in mouse JB6-Cl.41 cells. A (Promega) as a control for transfection efficiency. Forty hours post- lentivirus-based SAG-overexpressing construct (FG9-HA-SAG) was generat- transfection, cells were lysed and subjected to luciferase activity assay, using ed using lentivirus vector, FG9-EF1a (kindly provided by Dr. Colin Duckett, the Dual-Glo Luciferase Assay System (Promega) on a LD400 Luminescence University of Michigan, Ann Arbor, MI). Following DNA sequencing detector (Beckman Coulter, Fullerton, CA) as described (26). The results from confirmation, the HA-tagged SAG construct, along with the empty vector three independent experiments with each run in duplicate or triplicate were control, was transiently cotransfected into 293 cells with plasmids expressed as fold induction after normalization with transfection efficiency expressing gag, env, and polymerase. Virus-containing supernatants and by arbitrarily setting the value of pGL3-Basic vector control as 1. were collected and used to infect mouse JB6-Cl.41 cells. Expression of both Identification of the transcription initiation site. To define the HA-SAG and endogenous SAG was determined by Western blotting using transcription initiation site of the mouse SAG, a PCR-based 5¶-rapid anti-SAG antibody. amplification of cDNA ends (RACE) was done using the First-choice RLM- siRNA silencing. A lentivirus-based siRNA construct (H1, kindly RACE kit with protocols suggested by the manufacturer (Ambion, Austin, provided by Dr. Ihor Lemischka, Princeton University, Princeton, NJ) was TX) as described previously (26). Briefly, total mouse RNA was treated with used to make an LT-SAG-siRNA. The sequences of SAG siRNA oligonucle- calf intestine alkaline phosphatase to remove free 5¶-phosphates from otide are LT-bSAG01 (5¶-AACAAGAGGACTGTGTTGTGGTCTGGTTCAAGA- molecules of fragmented mRNA. The RNA was then treated with tobacco GACCAGACCACAACACAGTCCTCTTGTTTTTTGT-3¶)andLT-bSAG-02 acid pyrophosphatase to remove the cap structure from full-length mRNA, (5¶-CTAGACAAAAAACAAGAGGACTGTGTTGTGGTCTGGTCTCTTGAAC- leaving a 5¶-monophosphate. A 5¶-RACE adapter oligonucleotides, 5¶-GCU- CAGACCACAACACAGTCCTCTTGTT-3¶). The control siRNA sequences are GAUGGCGAUGAAUGAACACUGCGUUUGCUGGCUUUGAUGAAA-3¶, which LT-Control-01 (5¶-ATTGTATGCGATCGCAGACTTTTCAAGAGAAAGTCTG- is specific for the 5¶-monophosphate, was ligated to the RNA population CGATCGCATACAATTTTTTGT-3¶) and LT-Control-02 (5¶-CTAGACAAAA- using T4 RNA ligase. A reverse transcription was done using random AATTGTATGCGATCGCAGACTTTCTCTTGAAAAGTCTGCGATCGCATA- decamers, and the nest PCR then amplified the 5¶-end of the mouse SAG CAAT-3¶). The LT-SAG-siRNA plasmid, along with the control LT-Cont- mRNA using the 5¶-RACE outer primer, 5¶-GCTGATGGCGATGAATGAA- siRNA, was confirmed by DNA sequence and then cotransfected into 293 CACTG-3¶,or5¶-RACE inner primer, 5¶-CGCGGATCCGAACACTGCGTTT- cells, along with gag- and env-expressing plasmids. The supernatants GCTGGCTTTGATG-3¶ and a primer from the SAG mRNA, 5¶-CAACACA- containing viable LT-SAG-siRNA virus or LT-Cont virus were collected and GTCCTCTTGCTTGT-3¶. The PCR products were cloned into pCR2.1 TA used to infect JB6-Cl.41 cells. cloning vectors and insert fragments were sequenced. The transcription To silence human SAG or ROC1/Rbx1 in 293 cells, a plasmid-based initiation site was determined by comparing the sequences immediately hairpin RNAs were made using the pU6pro vector (kindly provided by Dr. after the 5¶-RACE adapter oligonucleotide to the mouse SAG gene sequence. David L. Turner, University of Michigan;ref. 30) as described. 1 The siRNA Gel retardation assay. Seven oligonucleotides and their corresponding complementary strands were synthesized (Invitrogen, Carlsbad, CA) as follows with the putative AP-1 binding sites in italics: Sag-AP1-1 1 http://sitemaker.umich.edu/dlturner.vectors www.aacrjournals.org 3617 Cancer Res 2007; 67: (8). April 15, 2007

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2007 American Association for Cancer Research. Cancer Research template oligonucleotides were made by Invitrogen. For psiSAG, the primer SAG promoter fragment, there are seven putative AP-1 binding sequences are iSAG2-1 (5¶-TTTGAACAAGAGGACTGTGTTGTGGTCTGG- sites (Genbank accession no. DQ231571). Some are exactly ¶ ¶ GCAAGAGCCAGACCACAACACAGCCTGTTTTT-3 ) and iSAG2-2 (5 -CTA- matching the typical AP-1 consensus site 5¶-TGAG/CTCA-3¶, GAAAAACAAGAGGACTGTGTTGTGGTCTGGCCTCCAGACCACAACA- whereas others are its derivatives identified by Transfac software ACAGTCCTCTTGTT-3¶). For psi-hROC1, the primer sequences are iROC1-1 as the potential AP-1 binding sites (data not shown). Before a (5¶-TTTGAAGACTTCTTCCATCAAGCTTCAAGAGAAGCTTGATGGAA- AGAAGCTTTT-3¶)andiROC1-2(5¶-CTAGAAAAAGACTTCTTCCAT- detailed characterization of the SAG promoter, we first defined the ¶ TCAAGCTTCTCTTGAAGCTATGAAAAGTCTT-3¶). The control psiCont transcription initiation site of the mouse SAG gene by the 5 -RACE ¶ sequences are iCont01 (5¶-TTTGATTGTATGCGATCGCAGACTTCAAGA- experiment. DNA sequencing of several independent 5 -RACE AGAAGTCTGCGATCGCATACAATTTTT-3¶)andiCont02(5¶-CTA- clones all showed an adenine at 31 position upstream the GAAAAATTGTATGCGATCGCAGACTTCTCTTGAAGTCTGCGATCGCATA- translational initiation site at the very 5¶-end of SAG transcript ACAAT-3¶). The cohesive ends of restriction enzymes BbsI (TTTG) and XbaI before reaching the adaptor sequence, indicating that it is the (CTAG) are in italics. The annealed synthetic primers were ligated into transcription initiation site. pU6pro vector predigested with BbsI and XbaI. The recombinant psiSAG AP-1 binds both in vitro and in vivo to the consensus plasmid was confirmed by DNA sequencing. To determine the effect of binding sites in the promoter of mouse SAG gene. We next siRNA silencing of SAG or ROC1/Rbx1 on Fbw7-mediated c-Jun degrada- determined whether AP-1 directly binds to these putative tion, the 293 cells were transiently cotransfected with plasmids expressing ubiquitin and c-Jun with or without Fbw7 in the absence or presence of psi- consensus sites first in vitro by a gel retardation assay. Nuclear hSAG or psi-hROC1. Thirty-eight hours post-transfection, cells were extracts were prepared from JB6-Cl.41 cells treated with TPA and 32 harvested and lysed. The cell lysates were then prepared and subjected to incubated with P-labeled oligonucleotides, each containing a Western blotting using various antibodies. putative AP-1 binding site. Initial binding assay revealed that AP-1 In vivo ubiquitination. Human lung H1299 cells were transiently only bound to two consensus sites, site 3 (TGACTCA) and site 6 transfected with His-Ub and c-Jun alone or in combination with psiSAG or (TGACGTCA) significantly, but not to other five putative sites (data psiCont. Cells were harvested 48 h after transfection and split into two not shown). To determine the binding specificity, these two sites aliquots with one for direct Western blotting analysis and the other for were mutated by a single T!A substitution (for site 3, AGACTCA; in vivo ubiquitination assay as described (31). Briefly, cell pellets were lysed for site 6, AGACGTCA). These mutant oligonucleotides, along in buffer A [6 mol/L guanidinium-HCl, 0.1 mol/L Na HPO /NaH PO , 2 4 2 4 with their WT controls, were subjected to gel retardation assay as 10 mmol/L Tris-HCl (pH 8.0), 10 mmol/L h-mecaptoethanol] and incubated with Ni-NTA beads (Qiagen, Valencia, CA) at room temperature for 4 h. shown in Fig. 1A. AP-1 bound significantly to the sites 3 (Fig. 1A, Beads were washed once with each of buffer A, buffer B [8 mol/L urea, WT3, lane 1) and 6 (Fig. 1A, WT6, lane 7). These AP-1 bindings can

0.1 mol/L Na2HPO4/NaH2PO4, 10 mmol/L Tris-HCl (pH 8.0), 10 mmol/L be supershifted by an anti–c-Jun antibody (Fig. 1A, lanes 2 and 8), h -mecaptoethanol], and buffer C [8 mol/L urea, 0.1 mol/L Na2HPO4/ blocked by a specific cold oligonucleotide (Fig. 1A, lanes 3 and 9), h NaH2PO4, 10 mmol/L Tris-HCl (pH 6.3), 10 mmol/L -mecaptoethanol]. but not by a nonspecific oligonucleotide (Fig. 1A, lanes 4 and 10). Proteins were eluted from beads with buffer D [200 mmol/L imidazole, A single T!A point mutation completely abolished the AP-1 0.15 mol/L Tris-HCl (pH 6.7), 30% glycerol, 0.72 mol/L h-mecaptoethanol, binding in both sites (Fig. 1A, lanes 5 and 6 and lanes 11 and 12). 5% SDS]. The eluted proteins were analyzed by Western blotting for Thus, AP-1 specifically bound to two AP-1 sites in the mouse SAG polyubiquitination of c-Jun with anti–c-Jun antibody. promoter. Soft agar assay. A standard soft agar assay was done as detailed We further confirmed these AP-1 direct bindings using an in vivo previously (32). Briefly, 10,000 of parental JB6-Cl.41 P+ cells or its Lenti-SAG transfectants (either SAG overexpressed or SAG silenced), along with the ChIP. JB6-Cl.41 cells were left untreated or treated with TPA for vector controls, were suspended in 0.33% agar containing 10% FCS in a 60- 12 h to activate AP-1 and then subjected to ChIP analysis. As shown mm dish either in the absence (DMSO) or the presence of TPA (20 ng/mL). in Fig. 1B (top), a 215-bp PCR-amplified fragment, corresponding After grown in 37jC for 14 days, the cells were stained with p- to the third AP-1 binding site (WT3), was detected in the input iodonitrotetrazolium (1 mg/mL;Sigma) for overnight and the colonies sample (Fig. 1B, top, lane 1) or in the samples immunoprecipitated with cell numbers more than eight were counted in five randomly selected with c-Jun antibody (Fig. 1B, top, lane 4) but not in two control areas from each dish. The results were from three independent assays, each samples either without antibody added (Fig. 1B, top, lane 2)or run in duplicate. Statistical analysis was done using one-way ANOVA and with IgG added (Fig. 1B, top, lane 3) for immunoprecipitation. Student’s t test. This specific AP-1 binding was enhanced after TPA treatment (Fig. 1B, top, compare lanes 4 versus 8). Similarly, AP-1 bound to Results the sixth binding site (WT6), revealed as a 156-bp fragment in the Molecular cloning and characterization of mouse SAG input or in c-Jun antibody, immunoprecipitated samples (Fig. 1B, promoter. We have shown previously that SAG is inducible at bottom, lanes 1 and 4) but not in other controls (Fig. 1B, bottom, the transcriptional level under redox-stress and hypoxic conditions lanes 2 and 3). The binding was again enhanced by TPA treatment (17, 33, 34). To determine the nature of SAG induction under (Fig. 1B, bottom). Thus, we conclude from these results that AP-1 various stress conditions, we went on and cloned a 1.3-kb genomic indeed directly binds to two AP-1 consensus sequences in the sequence upstream from the translational initiation site of mouse promoter of the mouse SAG gene. SAG gene (accession no. NT_039476). Bioinformatics analysis, AP-1 binding site-dependent transactivation of mouse SAG using Transfac2 software, of this potential SAG promoter fragment promoter. We next determined the promoter activity of this 1.3-kb revealed the consensus binding sites for many transcription factors mouse SAG upstream fragment and its dependence of AP-1 binding including AP-1, nuclear factor-nB, p53, and hypoxia-inducible sites. The luciferase reporters driven by the 1.3-kb fragment (7AP) factor-1, among others. In this study, we characterized the and its series of mutants with AP-1 site either deleted or mutated regulation of the SAG promoter by AP-1. In this 1.3-kb potential were constructed and tested in a luciferase reporter assay. As shown in Fig. 1C, the construct (7AP) driven by this 1.3-kb fragment caused a 550-fold higher luciferase activity compared with 2 http://www.gene-regulation.com/ the empty vector, indicating that this 1.3-kb fragment has a strong

Cancer Res 2007; 67: (8). April 15, 2007 3618 www.aacrjournals.org

Figure 1. AP-1 binds to two AP-1 consensus binding sites in the promoter of mouse SAG gene and transactivates mouse SAG promoter. A, in vitro AP-1 binding. Nuclear extracts were prepared from JB6-Cl.41 cells treated with TPA and incubated with 32P-labeled WT or mutant (MU) AP-1 consensus binding oligonucleotides. In some cases, anti–c-Jun antibody (Ab) was included to show a supershift, whereas, in other cases, 50 cold oligonucleotides (either WT or mutant) were included to show the binding specificity. Bands representing AP-1 binding and AP-1 + antibody binding. B, in vivo AP-1 binding. JB6-Cl.41 cells were left untreated or treated with TPA (20 ng/mL) for 12 h before being subjected to ChIP analysis. Input samples or samples precipitated with no antibody (Ab), IgG, or c-Jun antibody (c-Jun) were PCR amplified using primers specific for the third AP-1 binding site (WT3, top) or the sixth AP-1 binding site (WT6, bottom), respectively. C and D, AP-1 site-dependent transactivation of the mouse SAG promoter. Left, representation of a series of luciferase reporters driven by various deletion or mutation mutants of mouse SAG promoter with putative AP-1 binding site boxed in black; right, the luciferase activity of each construct. HeLa cells were plated in 96-well plates and transiently cotransfected with luciferase reporters driven by the 1.3-kb SAG promoter fragment and its series of deletion or mutation mutants, along with the pGL-3–negative control, respectively. Renilla-luc reporter was included for normalization of transfection efficiency. Forty hours post-transfection, cells were lysed and subjected to luciferase activity assay. Luciferase activity was presented as fold activation from three independent transfections, each run in duplicate or triplicate after normalization of transfection efficiency by arbitrarily setting the value of pGL-3 as 1.

promoter activity. Deletion of 467 base nucleotides from the 5¶-end the 1.3-kb mouse SAG promoter (7AP) or by its mutant with a (construct 5AP) caused a 1.6-fold reduction of the luciferase single mutation at both AP-1 binding sites 3 and 6 (Mu-3,6), which activity. Because this deleted fragment contained two putative AP-1 abolished the AP-1 binding. As shown in Fig. 1D, the mutations at sites to which AP-1 failed to bind, it is likely that the cis-elements these two AP-1 binding sites reduced the overall luciferase activity for other transcription factors rather than AP-1 contributed to this by 3.4-fold. Taken together, the results clearly showed that the portion of the promoter activity. The deletion or mutation of the promoter activity of mouse SAG gene is significantly dependent AP1-3 site (construct 4AP or Mu-5AP, respectively), which showed a on two true AP-1 binding sites and strongly suggested an AP-1– strong AP-1 binding, caused an additional 1.4-fold reduction of the dependent activation of the SAG promoter. promoter activity, indicating that this AP-1 binding site alone SAG induction by TPA. We next determined whether contributed significantly to the promoter activity. Further deletion endogenous SAG mRNA and protein are subjected to induction of 629 base nucleotides (construct 2AP), including two putative by TPA, a typical tumor promoter and AP-1 activator. Mouse AP-1 sites to which AP-1 failed to bind, did not cause any change in epidermal JB6-Cl.41 cells were treated with TPA (20 ng/mL) for luciferase activity, indicating that this fragment has no promoter various times and subjected to Northern analysis. As shown in activity. Further deletion or mutation of a strong AP-1 binding site Fig. 2A, SAG mRNA was detectable in JB6-Cl.41 cells and started to 6 (construct 1AP or Mu-2AP, respectively) resulted in a 1.6-fold increase 1 h post-TPA treatment and reached a peak induction of reduction of the promoter activity, indicating its strong contribu- 3.8-fold at 4 h. Increased mRNA started to decrease thereafter and tion to the promoter activity. Moreover, the deletion of AP-1 returned toward the background level by 24 h. These results further binding site 7 (construct 0AP) caused an additional 1.4-fold indicated that SAG induction by TPA occurred at the transcrip- reduction of the promoter activity, suggesting the contribution tional level. We next determined the SAG induction by TPA at the from other transcription factors. Finally, we directly compared the protein level. Cell lysates were prepared after TPA treatment for luciferase activity between two reporter constructs driven either by various periods and subjected to immunoblotting using anti-SAG www.aacrjournals.org 3619 Cancer Res 2007; 67: (8). April 15, 2007

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2007 American Association for Cancer Research. Cancer Research antibody. As shown in Fig. 2B, similar to the pattern of SAG mRNA the time course of SAG induction by TPA closely followed that of induction, the SAG protein levels started to increase at 1 h after c-Jun induction. treatment and reached a peak induction of up to 2.5-fold at 4 to 8 h. To determine whether TPA-induced SAG induction requires The induced SAG level then gradually reduced to the untreated c-Jun/AP-1, we used the JB6-Cl.41 cells overexpressing TAM67, a control level at 24 h post-TPA treatment. We next determined the dominant-negative c-Jun that has been shown previously to block time course of c-Jun induction by TPA in this cell model. If SAG is a AP-1 activity in multiple cell models, including JB6-Cl.41 cells true target of c-Jun/AP-1, the time course of SAG induction should (24, 35–38). As shown in Fig. 2C, in vector-transfected control cells, follow that of c-Jun induction. Indeed, the c-Jun levels increased TPA caused a gradual increase of SAG level with a peak induction 30 min after TPA exposure and continued to increase and reached of 3- to 3.5-fold at 4 to 8 h post-treatment. In contrast, TPA only a peak induction of 5-fold from 2 to 8 h. The induced levels started slightly induced SAG expression up to 1.5-fold at 2 to 4 h post-TPA to decrease at 16 and 24 h (Fig. 2B). The results clearly showed that exposure in TAM67-transfected cells. Thus, the magnitude of TPA- induced SAG induction was significantly reduced when the AP-1 activity is blocked, indicating that SAG induction by TPA is largely AP-1 dependent. Finally, we extended our observation in JB6-Cl.41 cells that SAG is inducible by TPA to human 293 cells. As shown in Fig. 2D, the SAG level started to increase for f2-fold at 2 to 4 h post-TPA exposure and reached the peak of a 3-fold increase at 16 h. The c-Jun expression at the basal level was hardly detectable in 293 cells but was rapidly induced by TPA, starting at 30 min for f2-fold increase and reaching the peak of a 13-fold increase at 16 h. Again, the time course of SAG induction seemed to follow that of c-Jun induction. Taken together, the data from assays of gel retardation, luciferase-based transactivation, and endogenous induction in both mouse and human cell lines all suggested that SAG is inducible by TPA through AP-1 binding and transactivation and SAG is a novel AP-1 target gene. SAG overexpression reduces both basal and TPA-induced c-Jun levels in JB6-Cl.41 cells. Few recent studies have shown that c-Jun is subjected to ubiquitination and degradation by SCF and DCX E3 ubiquitin ligases (22, 23). Because SAG is a RING component of SCF E3 ligase, we wondered whether SAG is involved in and if so, SAG would inhibit c-Jun levels on overexpression. We tested this hypothesis by using a lentivirus-based SAG construct expressing HA-tagged SAG. Infection of this construct into JB6- Cl.41 cells led to a 2.7-fold increase of SAG levels compared with the control vector (Fig. 3A). This increased SAG level is within the physiologic range and is comparable with the level induced by TPA (see Fig. 2). These cells were then tested for both basal and TPA-induced c-Jun levels. In control cells, c-Jun was detectable and subjected to induction by TPA, starting at 1 h for 3-fold increase and reaching a peak of 6.4-fold increase at 8 h. In contrast, in SAG-infected cells, the basal c-Jun levels decreased f2- to 3-fold Figure 2. SAG is subjected to TPA induction. A, induction of SAG mRNA and total TPA-induced levels were also reduced by 2-fold (Fig. 3B). expression by TPA in mouse JB6-Cl.41 cells. Subconfluent JB6-Cl.41 cells were subjected to TPA (20 ng/mL) treatment for indicated periods of time up to 24 h. As a negative control, SAG overexpression had no effects on p53, a Cells were harvested and subjected to total RNA isolation and Northern blot protein not to be induced by TPA and not known to be a SAG analysis using mSAG cDNA flanking the entire open reading frame as a probe. substrate. The results suggested that SAG could reduce c-Jun levels, The membrane was stripped and reprobed with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a loading control. The fold induction relative to the probably through promoting its degradation. To pursue this, we untreated control was shown after densitometry quantitation and GAPDH treated cells with MG132, a proteasome inhibitor, alone or in normalization. B, induction of SAG and c-Jun proteins by TPA in mouse JB6-Cl.41 cells. Subconfluent Cl.41 cells were subjected to TPA (20 ng/mL) combination with TPA. If c-Jun is targeted for degradation by SAG, treatment for indicated periods of time up to 24 h. Cells were harvested, lysed in then MG132 should block it. Indeed, as shown in Fig. 3C, MG132 Triton X-100 lysis buffer, and subjected (150 Ag) to immunoblotting analysis alone treatment caused a slight increase in basal level of c-Jun using antibodies against SAG, c-Jun, and h-actin. The fold induction relative to the untreated control was shown after densitometry quantitation and in control cells (Fig. 3C, lanes 1 versus 3) but a 2-fold increase in h-actin normalization. C, induction of SAG protein in vector control and SAG-overexpressed cells (Fig. 3C, lanes 5 versus 7), resulting in TAM67-overexpressed JB6-Cl.41 cells. Subconfluent cells were subjected to similar c-Jun levels in both control and SAG-overexpressed cells TPA (20 ng/mL) treatment for indicated periods of time up to 24 h. Cells were harvested, lysed in Triton X-100 lysis buffer, and subjected (150 Ag) to (Fig. 3C, lanes 3 versus 7). Accordingly, SAG-mediated inhibition immunoblotting analysis using antibodies against SAG, c-Jun, and h-actin. The of c-Jun induction by TPA was also blocked by MG132 (Fig. 3C, fold induction relative to the untreated control was shown after densitometry quantitation and h-actin normalization. D, induction of c-Jun and SAG proteins compare lanes 2 versus 6 and lanes 4 versus 8). by TPA in 293 cells. Subconfluent 293 cells were treated with TPA (10 ng/mL) for SAG silencing increases both the basal and TPA-induced indicated periods. Cells were harvested and subjected to Western blotting c-Jun levels in JB6-Cl.41 cells. We further tested this hypothesis analysis using antibodies against c-Jun, SAG, and h-actin. The numbers under each gel panel are fold changes after densitometry quantification with h-actin by using a lentivirus-based SAG siRNA silencing construct. Again, normalization, setting the control value as 1. it is expected that SAG silencing should increase the levels of c-Jun,

Cancer Res 2007; 67: (8). April 15, 2007 3620 www.aacrjournals.org

transiently cotransfected with His-Ub and c-Jun, alone or in combination of plasmid expressing control siRNA or SAG siRNA. Ubiquitinated c-Jun was purified by Ni-NTA beads and detected by antibody against c-Jun. As shown in Fig. 5A, c-Jun polyubiquitination can be readily detected in cells transfected with His-Ub, c-Jun, and control siRNA (Fig. 5A, lane 4), indicating that endogenous SCF components are sufficient to promote c-Jun polyubiquitination on c-Jun overexpression via transfection. When cotransfected with plasmid that silenced SAG expression (Fig. 5A, middle), c-Jun polyubiquitination was remarkably inhibited (Fig. 5A, lane 3). Thus, SAG is required for c-Jun polyubiquitination. The F-box protein Fbw7 has been shown previously to promote c-Jun ubiquitination and degradation (22, 39). Because Fbw7 is the F-box component of SCF E3 ubiquitin ligase, it is likely that either SAG or ROC1/Rbx1, the RING component of SCF, would be required. We tested this hypothesis by determining whether silencing SAG or ROC1/Rbx1 by siRNA would block Fbw7-induced c-Jun degradation. Human 293 cells were transiently cotransfected with Ub, c-Jun, and plasmids expressing Fbw7, alone or in combination with psi-hSAG (to silence SAG) and/or psi-hRbx1 (to silence Rbx1/ROC1). As shown in Fig. 5B, endogenous c-Jun was hardly detectable (Fig. 5B, lane 1), whereas exogenously transfected c-Jun expressed well (Fig. 5B, lane 2). Consistent with a previous report (22), cotransfection of c-Jun with Fbw7 reduced the level of Figure 3. Reduction of both basal and TPA-induced c-Jun accumulations on SAG overexpression. A, SAG overexpression by lenti-HA-SAG infection. c-Jun by 5.6-fold (Fig. 5B, lanes 3 versus 2, from 1 to 0.18). JB6-Cl.41 cells were infected with the empty lentivirus vector (LT-con)or Remarkably, cotransfection of Fbw7 with psi-hSAG, which caused a lentivirus expressing HA-tagged SAG (LT-HA-SAG). Cell lysates were prepared 2.5-fold reduction of endogenous SAG level (Fig. 5B, middle top, and subjected to immunoblotting analysis using anti-SAG antibody and h-actin as a loading control. Endo-SAG, endogenous SAG. B, SAG overexpression lanes 4 versus 1), reversed Fbw7-induced reduction of c-Jun from reduced both basal and TPA-induced c-Jun accumulation. Lentivirus-infected 5.6- to 1.3-fold (Fig. 5B, lane 4, from 1 to 0.18 and from 1 to 0.76). (empty vector control or HA-SAG) JB6-Cl.41 cells were treated with TPA for Similarly, cotransfection with psi-hROC1, which caused a 2-fold various periods and subjected to Western analysis using antibodies against c-Jun, p53, and h-actin. The numbers under the gel panels are fold changes after reduction of ROC1 (Fig. 5B, middle bottom, lanes 5 versus 1), densitometry quantification with h-actin normalization, setting the value in also partially reversed Fbw7-induced c-Jun degradation from 5.5- to LT-Con without TPA treatment as 1. C, MG132 blocked SAG-induced c-Jun reduction. Lentivirus-infected JB6-Cl.41 cells were treated with TPA or MG132 1.6-fold (Fig. 5B, lane 5, from 1 to 0.18 and from 1 to 0.62). alone or in combination for 8 h and subjected to Western blotting using Simultaneously silencing both SAG and ROC1 (up to 2.5-fold of antibodies against c-Jun and h-actin. The numbers under the gel panels are fold each) by cotransfecting both psi-hSAG and psi-hROC1 further changes after densitometry quantification with h-actin normalization, setting the value in LT-Con without TPA treatment as 1. inhibited c-Jun degradation by Fbw7 (Fig. 5B, lane 6). These results clearly showed that SAG as well as ROC1 is required for Fbw7- induced c-Jun degradation. if it indeed promotes c-Jun degradation. As shown in Fig. 4A, lenti- SAG expression inhibited TPA-induced soft agar colony SAG siRNA infection of JB6-Cl.41 cells reduced endogenous SAG formation in JB6-Cl.41 cells. It has been well established that c- levels by 80% compared with that of lenti-control siRNA infection, Jun induction and AP-1 activation play a key role in TPA-induced indicating that endogenous SAG was indeed significantly silenced. tumor promotion (3, 5). Thus, by promoting c-Jun degradation, We then exposed these two lines to TPA for various periods of time SAG might be able to inhibit tumor promotion activity of TPA. We up to 8 h followed by immunoblotting analysis for c-Jun protein tested this hypothesis by overexpressing SAG in JB6-Cl.41 cell line, levels. As shown in Fig. 4B, the basal level of c-Jun was very low but a well-established line that is sensitive to TPA-induced neoplastic detectable in LT-Cont–infected cells. The levels of c-Jun continu- transformation as measured by anchorage-independent growth in ously increased and reached a peak induction of 10-fold at 8 h post- soft agar (40, 41). The lentivirus expressing HA-tagged SAG was TPA treatment. Significantly, in SAG silenced cells (LT-bSAGsi), the made and infected into JB6-Cl.41 cells to make stable SAG- basal level of c-Jun was increased up to 3-fold than that in the expressing line, along with the vector control line. As shown in control cells and TPA treatment further increased the c-Jun levels Fig. 6A, lenti-SAG stable JB6-Cl.41 cells expressed HA-tagged SAG up to a total of 18-fold. Unlike c-Jun, SAG silencing had no effect on at a level comparable with the endogenous SAG (Fig. 6A, right, both basal and TPA-induced levels of c-Fos, an AP-1 component, lane 3 compared with lanes 1 and 2), indicating that lenti-SAG but not a SAG substrate (Fig. 4C). Finally, SAG silencing had, once infection increased a total of SAG level by 2- to 3-fold. This is within again, no effect on the levels of p53, which is neither inducible by the physiologic range because the endogenous SAG can be induced TPA, nor a SAG substrate (Fig. 4C). These results strongly suggested up to 3-fold after exposure to TPA (Fig. 2). All three JB6-Cl.41 lines that SAG is involved in down-regulation of both basal and TPA- (parental, vector infected, and HA-SAG infected) were subjected to induced levels of c-Jun, but not of c-Fos, in mouse JB6-Cl.41 cells. soft agar assay to determine the effect of SAG expression on SAG silencing inhibits c-Jun ubiquitination and blocks anchorage-independent growth (a measure of cellular transforma- Fbw7-mediated c-Jun degradation. We next determined whether tion) in the absence or presence of TPA. As shown in Fig. 6A (left), SAG siRNA silencing would inhibit c-Jun polyubiquitination by a all three lines had few background agar colonies in the absence commonly used in vivo ubiquitination assay (31). H1299 cells were of TPA. TPA treatment dramatically increased the number of agar www.aacrjournals.org 3621 Cancer Res 2007; 67: (8). April 15, 2007

Figure 4. The accumulation of c-Jun levels on SAG siRNA silencing. A, reduction of SAG protein by siRNA silencing. JB6-Cl.41 cells were infected with lentivirus expressing either control siRNA (LT-CONsi) or siRNA targeting SAG (LT-bSAGsi). Cell lysates were prepared and subjected to immunoblotting analysis using anti-SAG antibody and h-actin as a loading control. B and C, SAG silencing increased c-Jun levels in JB6-Cl.41 cells. Lentivirus-infected JB6-Cl.41 cells were treated with TPA for various periods and subjected to Western blotting analysis using antibodies against c-Jun, c-Fos, p53, and h-actin. The numbers under the gel panels are fold changes after densitometry quantification with h-actin normalization, setting the value in LT-CONsi without TPA treatment as 1.

colonies in parental cells, as expected. The vector virus infection silencing, which caused c-Jun accumulation at both basal and caused a further increase of agar colonies, consistent to our pre- TPA-induced levels (Fig. 4), would promote cellular transformation vious observation made with the plasmid vector transfection (32). as measured by this soft agar anchorage-independent assay in the Significantly, exogenous SAG expression at the physiologic level absence or presence of TPA. As shown in Fig. 6B, lenti-SAG siRNA reduced number of TPA-induced agar colonies by 40%, which is infection significantly reduced the level of endogenous SAG statistically significant (P < 0.05 compared with the parental compared with scrambled control siRNA (Fig. 6B, right, lane 3 control). Thus, a 2- to 3-fold increase of the SAG level could inhibit versus lanes 1 and 2). All these three lines were subjected to soft TPA-induced cellular transformation in mouse epidermal JB6-Cl.41 agar assay to determine the effect of SAG silencing on anchorage- cells. independent growth in the absence or presence of TPA. As shown SAG silencing increased the numbers of soft agar colonies in in Fig. 6B (left), in the absence of TPA, parental or control siRNA- JB6-Cl.41 cells. Finally, we determined whether SAG siRNA infected cells formed few background agar colonies. However, SAG

Figure 5. SAG silencing inhibited c-Jun polyubiquitination (A) and blocked Fbw7-mediated c-Jun degradation (B). A, H1299 cells were transiently transfected with His-Ub (lane 2)orc-Jun(lane 1) alone or His-Ub and c-Jun in combination with psiSAG (lane 3)orpsiCon(lane 4) and subjected to in vivo ubiquitination assay by Ni-bead purification of ubiquitinated c-Jun followed by detection of polyubiquitinated c-Jun by c-Jun antibody. B, human kidney 293 cells were transiently transfected with ubiquitin, along with the pcDNA3 (lane 1); c-Jun and pcDNA3 (lane 2); c-Jun and Fbw7 (lane 3); c-Jun, Fbw7, and psi-hSAG (lane 4); c-Jun, Fbw7, and psi-hRbx1 (lane 5); or c-Jun, Fbw7, psi-hSAG, and psi-hRbx1 (lane 6). Forty hours post-transfection, cells were lysed and cell extracts were subjected to immunoblotting using antibodies against c-jun (top)andh-actin as a loading control (bottom). The numbers under each gel panel are fold changes after densitometry quantification with h-actin normalization, setting the value in (lane 1)as1.

Cancer Res 2007; 67: (8). April 15, 2007 3622 www.aacrjournals.org

Figure 6. Changes in SAG levels alter the anchorage-independent growth of JB6-Cl.41 cells. A, SAG overexpression reduced numbers of TPA-induced soft agar colonies. JB6-Cl.41 cells were infected with empty lentivirus vector or lentivirus expressing HA-tagged SAG and subjected to, along with parental JB6-Cl.41 cells, Western blotting analysis using anti-SAG antibody. h-Actin was used as a loading control (right). All three lines (1 104 cells) were subjected to soft agar assay in the absence or presence of TPA (20 ng/mL) as described (32). Colonies (eight or more cells) were counted after 14 d. One-way ANOVA analysis (Origin 75 Demo, Origin Lab Corp., Northampton, MA) was done followed by Student’s t test. *, P < 0.05, compared with parental cells (left). B, SAG siRNA silencing increased soft agar colony numbers in the absence or presence of TPA. JB6-Cl.41 cells were infected with lentivirus expressing scrambled control siRNA or SAG siRNA and subjected to, along with parental JB6-Cl.41 cells, Western blotting analysis using anti-SAG antibody. h-Actin was used as a loading control (right). All three lines (1 104 cells) were subjected to soft agar assay in the absence or presence of TPA (20 ng/mL) as described (32). Colonies (eight or more cells) were counted after 14 d. One-way ANOVA analysis (Origin 75 Demo) was done followed by Student’s t test. *, P < 0.05; ** P <0.01, compared with parental cells, respectively (left).

siRNA silencing caused a significant increase in soft agar colony identified previously that regulate cell proliferation, tumor promo- numbers, suggesting that SAG silencing partially transformed this tion, and apoptosis (3, 5, 42). The unique feature for identification promoting sensitive line as a result of c-Jun accumulation. As of SAG as yet another direct AP-1 target is that SAG is a component expected, TPA treatment dramatically increased the number of of SCF E3 ubiquitin ligase that on induction could in turn inhibit soft agar colonies in parental cells as well as scrambled siRNA- AP-1 activity through promoting c-Jun degradation, thus establish- infected cells. SAG silencing caused a further increase in soft agar ing a negative feedback loop that keeps AP-1 activity in check. colony numbers with statistically difference from both control cells. C-Jun was shown to be ubiquitinated and degraded by at least It is worthy noting that higher number of soft agar colonies was three types of E3 ubiquitin ligases, including SCF-Fbw7 (Skp1- observed in both untreated and TPA-treated cells in this experi- cullin-Fbw7;refs. 22, 39), DCX (DDB1, cullin 4A, and X-box protein; ment compared with what was shown in Fig. 6A. This is due to the ref. 23), and Itch (43, 44). ROC1/Rbx1, the RING component of SCF change of the serum lot number, a common phenomenon seen and DCX E3 ubiquitin ligases, was implicated, but not directly in soft agar assay with JB6-Cl.41 cells. Our results clearly showed proven, in at least two c-Jun targeting E3s. Now, we showed here that on SAG silencing to induce c-Jun accumulation, the tumor that SAG is indeed involved in c-Jun ubiquitination and promotion-sensitive JB6-Cl.41 was partially transformed with a sig- degradation. When overexpressed, SAG remarkably reduced the nificant increase in its ability to grow in an anchorage-independent levels of both basal and TPA-induced c-Jun, whereas on SAG manner and that TPA treatment further enhanced this phenotype silencing, both basal and TPA-induced levels of c-Jun were sig- of cell transformation. nificantly increased. Furthermore, SAG silencing inhibited c-Jun polyubiquitination and Fbw7-induced degradation, indicating a requirement of SAG for c-Jun ubiquitination and degradation Discussion through SCF-Fbw7 E3 ubiquitin ligase. Our data also directly We have shown previously that SAG is a stress-responsive gene showed that like SAG, ROC1/Rbx1, which has the same tissue that is induced by redox agents and ischemia/reperfusion in expression pattern as SAG (17, 45) but is expressed in a constitutive multiple cell lines and normal tissues and acts as a cellular defense manner,3 was also involved in promoting c-Jun ubiquitination and mechanism to protect cells or tissues from apoptosis induced by degradation. these stresses (16, 17, 33, 34). We report here that SAG is induced by Mouse JB6 epidermal model consists of three variants with TPA via AP-1–mediated transcriptional activation. Four lines of distinct transformation response phenotypes as judged by anchor- evidence presented in this study support notion that SAG is a direct age-independent growth in a soft agar colony formation assay on AP-1 target. (a) Several AP-1 consensus elements were identified exposure to TPA: the variant with the promotion-resistant in the promoter of mouse SAG gene and AP-1 binds strongly to two phenotype (P-) is resistant to TPA-induced transformation, as of such elements in vitro and in vivo.(b) The SAG promoter activity evident by its failure to form soft agar colonies in the presence of is dependent of the true AP-1 binding sites. (c) Both SAG mRNA TPA. The variant with promotion-sensitive phenotype (P+), to and protein are induced by TPA, a typical AP-1 activator, and the which JB6-Cl.41 cells belong, is sensitive to TPA-induced transfor- time course of SAG induction follows that of c-Jun induction. (d) mation and will form agar colonies only in the presence of TPA. SAG induction by TPA is c-Jun/AP-1 dependent. Targeting c-Jun by TAM67, a dominant-negative c-Jun mutant, significantly reduced TPA-induced SAG expression. Many AP-1 target genes have been 3 Unpublished data. www.aacrjournals.org 3623 Cancer Res 2007; 67: (8). April 15, 2007

The third variant with transformed phenotype (Tx) will form agar (48). TPA exposure induces the levels of c-Jun and c-Fos as well as colonies regardless of TPA. Thus, P-,P+, and Tx JB6 variants c-Jun phosphorylation. Consequently, AP-1, consisting of Jun-Jun represent earlier-to-later stages of preneoplastic-to-neoplastic homodimer or Jun-Fos heterodimer, is activated to transactivate progression and provide an excellent cell model to study multistage its target genes, leading to cell proliferation and tumor promotion carcinogenesis (41). It has been shown previously that in this JB6 (3, 5, 49, 50). Activated AP1 also transactivates SAG. On induction, model, the AP-1 activity was progressively elevated during SAG blocks AP1 effects through recruiting and complexing with preneoplastic-to-neoplastic progression (7) and blockage of TPA- SCF-Fbw7 E3 ligase to promote c-Jun ubiquitination and degra- induced AP-1 activity by a dominant-negative c-Jun (TAM67), dation. Induction of SAG or SAG-associated E3 ubiquitin ligases encoding a transcriptionally inactive product, inhibited transfor- would therefore inhibit AP-1 activity, leading to inhibition of TPA- mation induced by TPA and epidermal growth factor (8). This induced tumor promotion. Thus, SAG induction by TPA/AP-1 can observation made in JB6 cell model has been extended to an in vivo serve as a cellular defensive mechanism to shut off cell transgenic mice model, in which TAM67 was overexpressed in proliferation and tumor promotion signals induced by prolonged mouse epidermal cells driven by a K14 promoter. Indeed, AP-1 activation. It is worthy noting, however, that up-regulation of overexpression of TAM67 inhibits in vivo tumor formation induced SAG by AP-1 will not only increase the activity of SAG-SCF-Fbw7 by 7,12-dimethylbenz(a)anthracene (DMBA)/TPA chemical carci- ligase to target c-Jun for degradation but may also induce the nogens as well as by UV exposure (35, 46). Likewise, we showed degradation of many other SAG-SCF ligase substrates, dependent here that overexpression of SAG in physiologic level in mouse JB6- on the availability of their specific F-box proteins. We therefore Cl.41 cells inhibited TPA-induced transformation, whereas SAG cannot exclude the possibility that degradation of other proteins silencing induced cellular transformation of tumor promotion- may also contribute to SAG-mediated inhibition of neoplastic sensitive JB6-Cl.41 cells. Furthermore, a SAG transgenic model, in transformation. Nevertheless, our study suggests that SAG, on which SAG expression was targeted to the epidermal cells by the induction, inhibits TPA-induced neoplastic transformation and K14 promoter (47), was generated and characterized. A decreased could serve as a cancer prevention target whose induction may incidence with a longer latent period for the formation of help to delay the process of tumor promotion and progression papillomas induced by DMBA/TPA was also observed.4 Thus, AP- through elimination of c-Jun and inactivation of AP-1. 1 activation indeed plays a critical role in multistage carcinogenesis, and inactivation of AP-1 through targeting its component, Acknowledgments such as c-Jun, would be a sound strategy for chemoprevention. Received 10/30/2006;revised 12/24/2006;accepted 1/12/2007. Our data presented here that (a) SAG is a direct AP-1 target, (b) Grant support: National Cancer Institute grants 1R01CA111554, 1R01CA118762- SAG is involved in c-Jun degradation, and (c) SAG expression 01A1, and 1R21CA116982-01A1 and Charlotte Geyer Foundation (Y. Sun). inhibits TPA-induced neoplastic transformation, establishing an The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance autofeedback loop of c-Jun/AP-1-SAG, in analogue to p53-Mdm2 with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Dr. Ihor Lemischka for the H1 plasmid for lentivirus-based siRNA silencing, Dr. Colin Duckett for FG9 lentivirus used to make lenti-HA-SAG, Dr. David Turner for pU6pro vector used to make psiSAG and psiROC1, and Dr. Nancy Colburn 4 Submitted for publication. (National Cancer Institute, Frederick, MD) for JB6-Cl.41 cells overexpressing TAM67.

References transrepress AP-1 without transactivating retinoic acid 19. Kamura T, Koepp DM, Conrad MN, et al. Rbx1, a response element. Cancer Res 1996;56:483–9. component of the VHL tumor suppressor complex and 1. Shaulian E, Karin M. AP-1 in cell proliferation and 11. Jonat C, Rahmsdorf HJ, Park KK, et al. Antitumor SCF ubiquitin ligase. Science 1999;284:657–61. survival. Oncogene 2001;20:2390–400. promotion and antiinflammation: down-modulation of 20. Ohta T, Michel JJ, Schottelius AJ, et al. ROC1, a 2. Shaulian E, Karin M. AP-1 as a regulator of cell life and AP-1 (Fos/Jun) activity by glucocorticoid hormone. Cell homolog of APC11, represents a family of cullin partners death. Nat Cell Biol 2002;4:E131–6. 1990;62:1189–204. with an associated ubiquitin ligase activity. Mol Cell 3. Angel P, Karin M. The role of Jun, Fos, and the AP-1 12. Watts RG, Huang C, Young MR, et al. Expression of 1999;3:535–41. complex in cell-proliferation and transformation. Bio- dominant negative Erk2 inhibits AP-1 transactivation 21. Tan P, Fuchs SY, Chen A, et al. Recruitment of a chim Biophys Acta 1991;1072:129–57. and neoplastic transformation. Oncogene 1998;17: ROC1-1 ubiquitin ligase by Skp1 and HOS to catalyze 4. Dhar A, Young MR, Colburn NH. The role of AP-1, NF- 3493–8. the ubiquitination of IkBa. Mol Cell 1999;3:527–33. nB, and ROS/NOS in skin carcinogenesis: the JB6 model 13. Liu Y, Ludes-Meyers J, Zhang Y, et al. Inhibition of 22. Nateri AS, Riera-Sans L, Da Costa C, et al. The is predictive. Mol Cell Biochem 2002;234–5:185–93. AP-1 transcription factor causes blockade of multiple ubiquitin ligase SCFFbw7 antagonizes apoptotic JNK 5. Bode AM, Dong Z. Mitogen-activated protein kinase signal transduction pathways and inhibits breast cancer signaling. Science 2004;303:1374–8. activation in UV-induced signal transduction. Sci STKE growth. Oncogene 2002;21:7680–9. 23. Wertz IE, O’Rourke KM, Zhang Z, et al. Human De- 2003;2003:RE2. 14. Bernstein LR, Bravo R, Colburn NH. 12-O-tetradeca- etiolated-1 regulates c-Jun by assembling a CUL4A 6. Sun Y, Oberley LW. Redox regulation of transcriptional noylphorbol-13-acetate-induced levels of AP-1 proteins: ubiquitin ligase. Science 2004;303:1371–4. activators. Free Radic Biol Med 1996;21:335–48. a 46-kDa protein immunoprecipitated by anti-fra-1 and 24. Dhar A, Hu J, Reeves R, et al. Dominant-negative c- 7. Dong Z, Watts RG, Sun Y, et al. Progressive elevation of induced in promotion-resistant but not promotion- Jun (TAM67) target genes: HMGA1 is required for tumor AP-1 activity during preneoplastic-to-neoplastic pro- sensitive JB6 cells. Mol Carcinog 1992;6:221–9. promoter-induced transformation. Oncogene 2004;23: gression as modeled in mouse JB6 cell variants. Int J 15. Huang C, Ma WY, Dong Z. The extracellular-signal- 4466–76. Oncol 1995;7:359–64. regulated protein kinases (Erks) are required for UV- 25. Bian J, Sun Y. Transcriptional activation by p53 of the 8. Dong Z, Birrer MJ, Watts RG, et al. Blocking of tumor induced AP-1 activation in JB6 cells. Oncogene 1999;18: human type IV collagenase (gelatinase A or matrix promoter-induced AP-1 activity inhibits induced trans- 2828–35. metalloproteinase 2) promoter. Mol Cell Biol 1997;17: formation in JB6 mouse epidermal cells. Proc Natl Acad 16. Sun Y. Induction of glutathione synthetase by 1,10- 6330–8. Sci U S A 1994;91:609–13. phenanthroline. FEBS Lett 1997;408:16–20. 26. Li J, Tan M, Li L, et al. SAK, a new polo-like kinase, is 9. Huang C, Schmid PC, Ma WY, et al. Phosphatidylino- 17. Duan H, Wang Y, Aviram M, et al. SAG, a novel zinc transcriptionally repressed by p53 and induces apopto- sitol-3 kinase is necessary for 12-O-tetradecanoylphor- RING finger protein that protects cells from apoptosis sis upon RNAi silencing. Neoplasia 2005;7:312–23. bol-13-acetate-induced cell transformation and induced by redox agents. Mol Cell Biol 1999;19:3145–55. 27. Sun Y, Bian J, Wang Y, et al. Activation of p53 activated protein 1 activation. J Biol Chem 1997;272: 18. Swaroop M, Bian J, Aviram M, et al. Expression, transcriptional activity by 1,10-phenanthroline, a metal 4187–94. purification, and biochemical characterization of SAG, a chelator and redox sensitive compound. Oncogene 1997; 10. Li JJ, Dong Z, Dawson MI, et al. Inhibition of tumor RING finger redox sensitive protein. Free Radic Biol Med 14:385–93. promoter-induced transformation by retinoids that 1999;27:193–202. 28. He H, Sun Y. Ribosomal protein S27L is a direct p53

Cancer Res 2007; 67: (8). April 15, 2007 3624 www.aacrjournals.org

target that regulates apoptosis. Oncogene 2006 Oct 23; size in an SKH-1 hairless mouse model. Mol Cancer Res kinases, their biochemistry, physiology, and clinical [Epub ahead of print]. 2003;1:848–54. importance. IUBMB Life 2005;57:283–95. 29. Sun Y, Pommier Y, Colburn NH. Acquisition of a 36. Dong Z, Crawford HC, Lavrovsky V, et al. A dominant 43. Fang D, Kerppola TK. Ubiquitin-mediated fluores- growth-inhibitory response to phorbol ester involves negative mutant of jun blocking 12-O-tetradecanoyl- cence complementation reveals that Jun ubiquitinated DNA damage. Cancer Res 1992;52:1907–15. phorbol-13-acetate-induced invasion in mouse kerati- by Itch/AIP4 is localized to lysosomes. Proc Natl Acad 30. Yu JY, DeRuiter SL, Turner DL. RNA interference by nocytes. Mol Carcinog 1997;19:204–12. Sci U S A 2004;101:14782–7. expression of short-interfering RNAs and hairpin RNAs 37. Li JJ, Cao Y, Young MR, et al. Induced expression of 44. Gao M, Labuda T, Xia Y, et al. Jun turnover is in mammalian cells. Proc Natl Acad Sci U S A 2002;99: dominant-negative c-jun downregulates NFnB and AP-1 controlled through JNK-dependent phosphorylation of 6047–52. target genes and suppresses tumor phenotype in human the E3 ligase Itch. Science 2004;306:271–5. 31. Xirodimas D, Saville MK, Edling C, et al. Different keratinocytes. Mol Carcinog 2000;29:159–69. 45. Swaroop M, Gosink M, Sun Y. SAG/ROC2/Rbx2/Hrt2, effects of p14ARF on the levels of ubiquitinated p53 and 38. Li JJ, Rhim JS, Schlegel R, et al. Expression of a component of SCF E3 ubiquitin ligase: genomic Mdm2 in vivo. Oncogene 2001;20:4972–83. dominant negative Jun inhibits elevated AP-1 and NF-nB structure, a splicing variant, and two family pseudo- 32. Sun Y, Nakamura K, Wendel E, et al. Progression transactivation and suppresses anchorage independent genes. DNA Cell Biol 2001;20:425–34. toward tumor cell phenotype is enhanced by over- growth of HPV immortalized human keratinocytes. 46. Young MR, Li JJ, Rincon M, et al. Transgenic mice expression of a mutant p53 tumor-suppressor gene Oncogene 1998;16:2711–21. demonstrate AP-1 (activator protein-1) transactivation isolated from nasopharyngeal carcinoma. Proc Natl 39. Wei W, Jin J, Schlisio S, et al. The v-Jun point is required for tumor promotion. Proc Natl Acad Sci Acad Sci U S A 1993;90:2827–31. mutation allows c-Jun to escape GSK3-dependent U S A 1999;96:9827–32. 33. Chanalaris A, Sun Y, Latchman DS, et al. SAG recognition and destruction by the Fbw7 ubiquitin 47. Wang X, Zinkel S, Polonsky K, et al. Transgenic attenuates apoptotic cell death caused by simulated ligase. Cancer Cell 2005;8:25–33. studies with a keratin promoter-driven growth hormone ischaemia/reoxygenation in rat cardiomyocytes. J Mol 40. Colburn NH, Former BF, Nelson KA, et al. Tumour transgene: prospects for gene therapy. Proc Natl Acad Cell Cardiol 2003;35:257–64. promoter induces anchorage independence irreversibly. Sci U S A 1997;94:219–26. 34. Yang GY, Pang L, Ge HL, et al. Attenuation of Nature 1979;281:589–91. 48. Wu X, Bayle JH, Olson D, et al. The p53-mdm-2 ischemia-induced mouse brain injury by SAG, a redox- 41. Sun Y, Hegamyer G, Colburn NH. Molecular cloning autoregulatory feedback loop. Genes Dev 1993;7: inducible antioxidant protein. J Cereb Blood Flow Metab of five messenger RNAs differentially expressed in 1126–32. 2001;21:722–33. preneoplastic or neoplastic JB6 mouse epidermal cells: 49. Sun Y. Free radicals, antioxidant enzymes, and 35. Cooper SJ, MacGowan J, Ranger-Moore J, et al. one is homologous to human tissue inhibitor of carcinogenesis. Free Radic Biol Med 1990;8:583–99. Expression of dominant negative c-jun inhibits ultravi- metalloproteinases-3. Cancer Res 1994;54:1139–44. 50. Eferl R, Wagner EF. AP-1: a double-edged sword in olet B-induced squamous cell carcinoma number and 42. Karin M, Gallagher E. From JNK to pay dirt: jun tumorigenesis. Nat Rev Cancer 2003;3:859–68.

www.aacrjournals.org 3625 Cancer Res 2007; 67: (8). April 15, 2007

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2007 American Association for Cancer Research. SAG/ROC2/Rbx2 Is a Novel Activator Protein-1 Target that Promotes c-Jun Degradation and Inhibits 12- O -Tetradecanoylphorbol-13-Acetate−Induced Neoplastic Transformation

Qingyang Gu, Mingjia Tan and Yi Sun

Cancer Res 2007;67:3616-3625.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/8/3616

Cited articles This article cites 49 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/8/3616.full#ref-list-1

Citing articles This article has been cited by 10 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/8/3616.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/67/8/3616. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.