The Coactivator CRTC1 Promotes Cell Proliferation and Transformation Via AP-1

The coactivator CRTC1 promotes cell proliferation and transformation via AP-1

Gianluca Canettieria,1, Sonia Conia,2, Michele Della Guardiaa,2, Valentina Nocerinoa, Laura Antonuccia, Laura Di Magnoa, Robert Screatonc, Isabella Screpantia,b, Giuseppe Gianninia, and Alberto Gulinoa,b,d,1

aDepartment of Experimental Medicine and bPasteur Institute, Cenci Bolognetti Foundation, Sapienza University, Rome, Italy; cApoptosis Research Centre, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada; and dNeuromed Institute, Pozzilli, Italy

Edited by Peter K. Vogt, Scripps Research Institute, La Jolla, CA, and approved December 5, 2008 (received for review September 3, 2008) Regulation of gene expression in response to mitogenic stimuli is a Besides TPA, AP-1 also is activated by growth factors, cytokines, critical aspect underlying many forms of human cancers. The AP-1 stress signals, and oncoproteins, and thus is involved in various complex mediates the transcriptional response to mitogens, and its cellular processes, including cell proliferation, differentiation, deregulation causes developmental defects and tumors. We report cell survival, apoptosis, and neoplastic transformation (14). that the coactivator CRTC1 cyclic AMP response element-binding Whereas the biological role of the AP-1 transcription factors protein (CREB)-regulated transcription coactivator 1 is a potent and is well established, the molecular mechanism through which indispensable modulator of AP-1 function. After exposure of cells to these factors promote target gene activation remains incom- the AP-1 agonist 12-O-tetradecanoylphorbol-13-acetate (TPA), CRTC1 pletely understood. In this report, we demonstrate that CRTC1 is recruited to AP-1 target gene promoters and associates with c-Jun is a potent, indispensable modulator of AP-1 activity that and c-Fos to activate transcription. CRTC1 consistently synergizes with associates with c-Jun and c-Fos and promotes c-Jun–mediated the proto-oncogene c-Jun to promote cellular growth, whereas AP- cellular proliferation and transformation. 1–dependent proliferation is abrogated in CRTC1-deﬁcient cells. Re- markably, we demonstrate that CRTC1-Maml2 oncoprotein, which Results causes mucoepidermoid carcinomas, binds and activates both c-Jun CRTC1 Is an AP-1 Coactivator and Mediates the Transcriptional Re- and c-Fos. Consequently, ablation of AP-1 function disrupts the sponse to TPA. To evaluate the effect of CRTCs on TPA-induced cellular transformation and proliferation mediated by this oncogene. transcription, we first tested the effect of CRTC1, CRTC2, and Together, these data illustrate a novel mechanism required to couple CRTC3 on the TPA-responsive human matrix metalloproteinase 1 mitogenic signals to the AP-1 gene regulatory program. (MMP1) and 3ϫ AP-1 reporter constructs in HeLa cells. MMP1 contains a TRE (TGACTCA) located at Ϫ72 bp from the tran- AP-1 ͉ CRTC1 ͉ Mect1-Maml2 ͉ Proliferation ͉ Transformation scription start site (16). The 3ϫ AP-1 construct has 3 TRE repeats upstream of the thymidine kinase minimal promoter. As shown in he cyclic AMP response element-binding protein (CREB)- Fig. 1A, CRTC1 had a strong stimulatory effect on both promoters, Tregulated transcription coactivators (CRTCs, originally called whereas CRTC2 and CRTC3 displayed a weaker effect. Thus, we TORCs) are a novel class of signal-dependent CREB coactivators used CRTC1 in all subsequent experiments. identified using a high-throughput expression screen of a mamma- The effect of CRTC1 was synergistic with TPA and was depen- lian cDNA library (1, 2). CRTCs associate with the bZIP region of dent on the AP-1 transcription factor complex, as demonstrated by CREB via their N-terminal region and activate transcription the addition of the AP-1 dominant negative (DN) polypeptide through interactions with components of the basal transcriptional A-Fos (17), which strongly reduced the response to TPA and CRTC1 both alone and in combination (Fig. 1B). The effect also apparatus (1, 3). Under resting conditions, CRTCs are phosphor- ⌬ ylated by sucrose nonfermenting1/AMP-activated protein kinase depended on the TRE; the TRE mutant MMP1 promoter ( TRE) (AMP/SNF) kinases and sequestered in the cytoplasm (4, 5). When construct failed to respond to CRTC1 and TPA (Fig. 1C). the intracellular levels of calcium or cAMP rise, CRTCs are To rule out the involvement of CREB in this process, we dephosphorylated, travel to the nucleus and bind to CREB, thereby performed knockdown experiments using RNA interference. De- activating transcription. Consistent with their role as CREB acti- pletion of CREB using CREB siRNA did not affect CRTC1 vators, CRTCs have been shown to be key regulators of glu- inducibility of the MMP1 gene, whereas knockdown of the AP-1 coneogenesis (5–8), adaptive mitochondrial biogenesis (9), ␤ cell member c-Jun abrogated the transcriptional response to CRTC1 survival (10), and long-term synaptic plasticity (11). (Fig. 1D). In contrast, ablation of CREB, but not c-Jun, disrupted EVX1 Recent observations have suggested that CRTCs also promote the response to CRTC1 of the CREB target promoter (18) (supporting information (SI) Fig. S1), thus demonstrating that activation of other transcription factors besides those of the CREB/ CRTC1 regulation of AP-1 is functionally independent of CREB ATF1 family. Indeed, the phorbol ester 12-O-tetradecanoylphor- and vice versa. bol-13-acetate (TPA) causes CRTC1 nuclear translocation in HeLa To verify that the effect was detectable on endogenous AP-1– cells (12), and deletion of the TPA responsive element (TRE) from responsive genes, we performed RT-PCR on 3 different AP-1 the IL-8 promoter abrogates CRTC1-mediated enhancement (2), suggesting that CRTC1 can stimulate transcriptional output of a

TPA-regulated pathway. Recently, it was reported that CRTC1 can Author contributions: G.C., R.S., G.G., and A.G. designed research; G.C., S.C., M.D.G., V.N.,

be phosphorylated and activated by MEKK1 (13), a critical kinase L.A., and L.D.M. performed research; G.C., S.C., M.D.G., R.S., I.S., and A.G. analyzed data; CELL BIOLOGY activated by several mitogenic stimuli, including TPA. and G.C., R.S., G.G., and A.G. wrote the paper. TPA is a tumor-promoting drug that activates transcription of a The authors declare no conﬂict of interest. number of genes that typically contain a TPA response element This article is a PNAS Direct Submission. (TRE ϭ TGACTCA) in their promoter regions (14). In turn, the 1To whom correspondence may be addressed. E-mail: [email protected] or TRE is bound by the dimeric AP-1 transcription factor complex, [email protected]. comprising a large family of Fos (c-Fos, FosB, Fra1, Fra-2), Jun 2S.C. and M.D.G. contributed equally to this work. (c-Jun, JunB, JunD), and ATF bZIP proteins (15). The main AP-1 This article contains supporting information online at www.pnas.org/cgi/content/full/ proteins in mammalian cells are c-Jun and c-Fos, with c-Jun the 0808749106/DCSupplemental. most potent transcriptional activator in this group. © 2009 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0808749106 PNAS ͉ February 3, 2009 ͉ vol. 106 ͉ no. 5 ͉ 1445–1450 Downloaded by guest on September 24, 2021 20 A B * C D Con 300 * 6 * 16 800 * 250 CRTC1 WT 200 * 12 600 4 TRE 200 150 8 * 400 * * * **** * 100

Fold Change 2 Fold Change

100 Fold Change Fold change

4 Fold Change 200 * ** *** 50 * * ** ** ** 0 0 0 0 ** 0 SCR CREBi Juni - C1 C2 C3 - C1 C2 C3 C1 -+- + -+-+ C1 - -++ TPA -- + + --++ TPA -+-+ c-Jun A-Fos --- -+++ + CREB

pAd-Easy 8 Actin E DMSO TPA FSK+ F G 60 IBMX GFP CRTC1 * * 6 Con TPA 40 H MMP1 4 SCR C1i TPA - + - + * ** MTIIA 20 2 Fold Change * *** Fold change TIMP1 **** ** 0 CRTC1 TIMP1 0 SCR C1i C1 - +-+ -+-+ Actin CRTC1 TPA -- + + --++ GAPDH C1-DN - --- + +++ Actin

Fig. 1. CRTC1 promotes the TPA response of AP-1 targets. (A) HeLa cells were transfected with MMP1-luc (Left)or3ϫ AP-1–Luc (Right) plasmids, TK renilla, and plasmids encoding CRTCs (C1, C2, C3: CRTC1–3) for 24 h, and luciferase assays were performed. *P Ͻ .05. (B) HeLa cells were transfected with 3ϫ AP-1–Luc reporter, C1, A-Fos, or empty vector for 24 h and treated with 100 ng/mL of TPA for 10 h, where indicated. *C1, TPA, C1 ϩ TPA versus control, P Ͻ .05; **C1 ϩ A-Fos versus C1, P Ͻ .01; *** TPA ϩ A-Fos versus TPA, P Ͻ .05; ****C1 ϩ TPA ϩ A-Fos versus C1 ϩ TPA, P Ͻ .05. (C) MMP1-Luc (WT) or MMP1 ⌬TRE-Luc reporters (⌬TRE) were transfected in HeLa cells with C1 or empty vector and treated with TPA as before. *C1, TPA, C1 ϩ TPA versus control, P Ͻ .05; **C1, TPA, C1 ϩ TPA ⌬TRE versus C1, TPA, C1 ϩ TPA WT, P Ͻ .01. (D) HeLa cells were transfected with siRNAs for CREB (CREBi), c-Jun (Juni), or nonspecific (SCR), MMP1-Luc, TK renilla, and CRTC1 plasmids for 72 h. The knockdown of c-Jun and CREB was verified by Western blot analysis (Bottom). *SCR CRTC1 versus SCR control, P Ͻ .01; **Juni CRTC1 versus SCR CRTC1, P Ͻ .01. (E) HeLa cells were treated with 100 ng/mL of TPA or 10 ␮M forskolin (FSK) ϩ 40 ␮M IBMX (Left) or infected with the indicated adenoviruses (Right), and RT PCR was performed. (F) HeLa cells were transfected with 3ϫ AP-1–Luc reporter and plasmids expressing C1 or its dominant negative (C1-DN) or empty vector and treated with TPA as above. *C1, TPA, C1 ϩ TPA versus control, P Ͻ .05; **C1 ϩ C1-DN versus C1, P Ͻ .01; ***TPA ϩ C1-DN versus TPA, P Ͻ .01; ****C1 ϩ TPA ϩ C1-DN versus C1 ϩ TPA, P Ͻ .05. (G) In Hela cells, MMP1-Luc reporter was transfected with siRNAs for C1 (C1i) or nonspecific (SCR) for 72 h and cells treated with TPA for 10 h, where indicated. The actual knockdown of C1 was verified by Western blot analysis (Bottom). *SCR TPA versus SCR control, P Ͻ .01; **C1i TPA versus SCR TPA, P Ͻ .05. (H) HeLa cells were transfected and treated as in image G, and Western blot analysis was performed with the specified antibodies.

targets: MMP1, MTIIA (19), and TIMP1 (20). As shown in Fig. 1E, CRTC1 was readily detected in immunoprecipitates of c-Jun and all 3 of these genes were induced by exposure to TPA. Transduction c-Fos (Fig. 2B). Interestingly, whereas a diffuse CRTC1 band was of HeLa cells with adeno-CRTC1 increased MMP1, MTIIA, and detected in the total cell lysate, a discrete band with slower mobility TIMP1 mRNAs to levels comparable to those achieved by TPA was found to associate with AP-1, indicating that posttranslational treatment, demonstrating that CRTC1 induces endogenous AP-1 modifications of CRTC1 may govern the interaction with AP-1. target genes. In contrast, none of the targets was activated by the In addition, both c-Jun and c-Fos bZIP domains were found in cAMP agonists forskolin (FSK) and IBMX, further demonstrating the immunocomplexes on immunoprecipitation of CRTC1 but not that CREB is not involved in their regulation. on immunoprecipitation performed with peptide-blocked antibod- To determine whether abrogation of endogenous CRTC1 func- ies (Fig. 2C), demonstrating CRTC1’s ability to bind the bZIP tion suppresses the response to TPA, we used a DN CRTC1 domains of AP-1. To determine whether the interaction of CRTC1 construct, encoding the N-terminal region of CRTC1 (2). As shown and the AP-1 complex is direct, we expressed the N-terminal region in Fig. 1F, this construct disrupted the response to TPA and (amino acid 1–142) of CRTC1 as a GST fusion protein; this region was previously shown to interact with CREB (1). As shown in Fig. CRTC1 both alone and in combination. To further prove that 35 CRTC1 is needed to confer the TPA response to AP-1 target genes, 2D, incubation with S-labeled in vitro–translated c-Jun and c-Fos resulted in strong binding to GST-CRTC1 (aa 1–142) compared we performed knockdown experiments with siRNAs. Ablation of with the GST alone control, confirming that AP-1 can bind directly CRTC1 strongly decreased the TPA response of the MMP1 to the same region of CRTC1 that interacts with CREB. promoter compared with the nonspecific scrambled siRNA (Fig. We performed co-immunoprecipitation studies to determine 1G). In addition, the TPA induction of the endogenous AP-1 target whether TPA affects AP-1–CRTC1 complex formation. As shown TIMP1 was completely abrogated in CRTC1-deficient cells (Fig. in Fig. 1E, CRTC1 associated to endogenous c-Jun, and this binding 1H). These findings demonstrate that CRTC1 is indispensable for was increased upon the addition of TPA to the cells. Furthermore, the transcriptional response to TPA. in chromatin immunprecipitation assays (ChIPs), we observed The activation of CREB-mediated transcription by CRTC is coincident MMP1 promoter occupancy of CRTC1 and c-Jun in achieved on binding of the N-terminal region of CRTC to the basic response to TPA, demonstrating the formation of an AP-1–CRTC1 leucine-zipper domain (bZIP) of CREB (1). Because AP-1 family promoter complex induced by phorbol ester (Fig. 2F). Taken members also contain a bZIP DNA-binding domain, we hypothe- together, these findings demonstrate that CRTC1 mediates the sized that CRTC1 might activate AP-1 through a similar transac- mitogen-induced activation of AP-1 through direct binding be- tivation mechanism. To test this hypothesis, we first performed tween its N-terminal region and the bZIP domains of c-Jun and co-immunoprecipitation studies. Both c-Jun and c-Fos were found c-Fos. in CRTC1 immunoprecipitates, whereas no binding was detected when the antibody was saturated with the specific immunizing CRTC1 Is Necessary for c-Jun–Dependent Cellular Proliferation. We peptide before immunoprecipitation (Fig. 2A). Furthermore, performed colony-formation assays to investigate whether CRTC1

1446 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0808749106 Canettieri et al. Downloaded by guest on September 24, 2021 ++++++ A c-Jun c-Fos B CRTC1 C --+--+ Flag peptide - + - + c-Fos -+--+- c-Jun Flag peptide - + 5% Input c-Fos c-Jun CRTC1 bZip-Jun

bZip-Fos c-Fos CRTC1 c-Jun CRTC1 IP CRTC1 IP 5% Input AP-1 IP 5% Input

D E F TPA TPA - + - +

No Ab CREB c-Jun c-Fos DMSO TPA Flag peptide - + - + CRTC1 5% Input 1-142 1-142 1-142 c-Jun c-Jun Jun Fos CRTC1 GST c- GST GST c- CRTC1 CRTC1 CREB CRTC1 Input 1%

CRTC1 IP MMP1 GAPDH

MMP1 +1 TRE (-72)

Fig. 2. CRTC1 interacts with the bZIP region of AP-1. (A and B) HeLa cells were transfected with Flag-tagged CRTC1, HA-c-Jun, and HA-c-Fos and immunoprecipitated with Flag (A)orHA(B) antibodies. Western blot analysis was performed with anti-HA or anti-Flag antibodies. For negative controls, Flag antibodies were blocked with 0.1 mg/mL of Flag peptide (A), or cells were transfected with an empty vector (B). (C) Co-immunoprecipitation of Flag-CRTC1 and HA-bZIP c-Jun or HA-bZIP c-Fos, followed by Western blot analysis with anti-HA antibody. (D) GST pulldown assay with GST CRTC11–142 or GST alone incubated with in vitro–translated CREB, c-Jun, or c-Fos. (E) HeLa cells were transfected for 24 h with Flag-tagged CRTC1 and treated for 6 h with TPA. Cell lysates were immunoprecipitated with Flag antibodies, and Western blot analysis was performed with c-Jun and Flag antibodies. (F) For ChIP, HeLa cells were treated for 6 h with TPA or vehicle, after which ChIP was performed as described above with the specific antibodies. The eluted DNA was PCR-amplified with primers encompassing the TRE site of MMP1 promoter (Bottom; schematic representation of MMP1 promoter and the primers used for the amplification). For a negative control, DNA was amplified with primers encompassing the coding region of the GAPDH gene.

can modulate the AP-1–dependent cell proliferation. As shown in tumors are associated with a chromosomal translocation t (11, 19) Fig. 3A, ectopic expression of c-Jun in HeLa cells caused a that generates a novel fusion protein, Mect1-Maml2 (M-M2), which significant increase in the number of colonies compared with cells contains a portion of the CRTC1 gene and exerts strong trans- transfected with the empty vector. Remarkably, CRTC1 signifi- forming activities (21). The M-M2 oncogene consists of the N- cantly increased c-Jun–induced colony formation compared with terminal portion of CRTC1 (originally called Mect1) with 981 aa of c-Jun alone, whereas expression of CRTC1 alone had no effect. the C-terminal region of the Notch receptor coactivator Master- Accordingly, cells expressing c-Jun and CRTC1 together grew mind like 2 (Maml2). faster than those expressing c-Jun alone, as determined by an Because our data demonstrate that AP-1 interacts with the analysis of cumulative cell numbers over time (Fig. 3B). To deter- N-terminal region of CRTC1, we reasoned that the transforming mine whether CRTC1 is required for AP-1–mediated cell prolif- potential of M-M2 fusion could be related to its ability to activate eration, we performed BrdU incorporation assays in cells in which inappropriate AP-1–mediated transcription. To test this hypothesis, CRTC1 levels were reduced using specific siRNAs (Fig. 3C). we first performed co-immunoprecipitation studies. As shown in Whereas the number of actively cycling cells more than doubled in Fig. 4A, M-M2 formed complexes with both c-Jun and c-Fos that the presence of ectopic c-Jun, this effect was completely abolished did not form in peptide-competed immunoprecipitation studies, in CRTC1-deficient cells, confirming that CRTC1 is indispensable demonstrating that M-M2 interacts with AP-1 in vivo. In addition, for this effect of c-Jun. Furthermore, CRTC1 strongly enhanced M-M2 caused a robust, 200-fold increase of MMP1 promoter c-Jun–mediated promoter activation. Indeed, the MMP1 luciferase activity and synergized with TPA, effects that were abrogated by the reporter displayed a synergistic activation by c-Jun and CRTC1 AP-1 DN A-Fos (Fig. 4B). Activation of MMP1 by M-M2 appeared (Fig. 3D), and knockdown of CRTC1 strongly reduced c-Jun– to be dependent on the CRTC1 domain of M-M2, because the

induced MMP1 promoter activity (Fig. 3E). These findings support CRTC1 DN polypeptide suppressed the stimulatory effect on CELL BIOLOGY the hypothesis that CRTC1 is a critical modulator of c-Jun– MMP1 transcription (Fig. 4C). Furthermore, ectopic expression of dependent cell proliferation. M-M2 enhanced the endogenous expression of the MMP1, MTIIA, and TIMP1 genes (Fig. 4D). Together, these findings demonstrate The CRTC1-MAML2 Oncoprotein Exerts Its Transforming Activity that M-M2 directly promotes AP-1–mediated transcription through Through AP-1. The effect of CRTC1 on cellular proliferation CRTC1. prompted us to evaluate its involvement in pathological processes Previous studies have shown that the transforming potential of in which aberrant regulation of cellular growth leads to neoplastic M-M2 is mediated by constitutive CREB activation (22, 23). To test transformation. A role for CRTC1 in tumor formation was previ- whether the effect of the M-M2 fusion protein is also related to the ously proposed for mucoepidermoid carcinomas (MECs). These ability to activate AP-1 target gene expression, we performed focus

Canettieri et al. PNAS ͉ February 3, 2009 ͉ vol. 106 ͉ no. 5 ͉ 1447 Downloaded by guest on September 24, 2021 A Con c-Jun B 900 ** 800 )

7 2 700 Empty CRTC1 600 1.6 cJun 500 CRTC1+cJun * 1.2 400 Colonies C1 C1+c-Jun 300 0.8 200 0.4 100 0 0 Cumulative cell number (X10 Con c-Jun C1 C1 + 0246810 c-Jun Days in culture

1.4 C * D E 1.2 * 1 3 * 0.8 ** 400 2 0.6 Brdu Uptake

0.4 * Fold Change 200 1 0.2 * ** Fold Change Fold

0 0 - c-Jun - c-Jun - c-Jun C1 c-Jun+ - C1i c-Jun c-Jun+ C1 C1i Scr C1i

Fig. 3. CRTC1 enhances c-Jun–mediated cell proliferation. (A) HeLa cells were transfected with the specific plasmids, grown with G418 for 2 weeks, fixed, and stained with Coomassie Blue (Con ϭ pcDNA3). (Left) Representative image of the plates after staining. (Right) Quantitative analysis. Results represent the average Ϯ SD of 3 separate experiments, each performed in duplicate. *c-Jun versus control, P Ͻ .05; **c-Jun ϩ C1 versus c-Jun, P Ͻ .05. (B) Cells derived from duplicate plates from the colony assays in plot A were grown in culture as monolayers, and cell growth rates were determined by analyzing cumulative cell numbers over time. Cells were counted in triplicate; error bars represent SD. (C) HeLa cells were transfected for 72 h with the specified siRNA, and then retransfected with plasmids encoding c-Jun, GFP, or empty vector for 24 h. BrdU was then added to the cells for 5 h. After incubation, the cells were processed as described previously, and the percent incorporation of BrdU was assessed in the population of transfected cells. Results represent the average Ϯ SD of 3 separate experiments, each performed in triplicate. *c-Jun versus control, P Ͻ .05; **C1 ϩ c-Jun versus c-Jun, P Ͻ .05. (D) In HeLa cells, the MMP1-Luc reporter was transfected with c-Jun and C1 expression plasmids, alone or in combination, after which the luciferase assay was performed as described previously. *P Ͻ .05. (E) HeLa cells were transfected with MMP1-Luc reporter and plasmids expressing shRNA for C1 or nonspecific shRNA, c-Jun, or empty vector for 72 h. *c-Jun versus control P Ͻ .05; **c-Jun ϩ C1i versus c-Jun P, Ͻ .01.

formation assays in RK3E cells, in which expression of M-M2 elicits Currently, little is known about the immediate consequences focus formation (21). The presence of the AP-1 DN A-Fos strongly downstream of AP-1 binding to its target gene promoters. In the reduced the number of foci induced by M-M2 (Fig. 4E), indicating present work, we have identified a novel regulatory mechanism that the transforming potential of this oncogene depends on AP-1 of AP-1 transactivation. We have shown that CRTC1 mediates function. Expression of A-CREB displayed a weaker but significant the AP-1–dependent transcriptional response to TPA by directly inhibitory effect, demonstrating that both transcription factors are interacting with the bZip regions of c-Jun and c-Fos at the involved in this phenomenon. promoter level. Because the same domain of CREB interacts To determine whether suppression of AP-1 function could affect with the same region of CRTC1 (1), the first issue raised by our cell proliferation in tumor cells arising from M-M2 expression, we findings is whether CRTC1 can discriminate between AP-1 and performed BrdU incorporation assays in NCI-H292 mucoepider- CREB in response to specific stimuli. Our data with siRNAs moid carcinoma cells, which express M-M2 (21). As shown in Fig. indicate that the 2 systems likely work independently. We 4F, expression of A-Fos reduced the number of proliferating propose that posttranslational modifications of CRTC and/or NCI-H292 cells by more than 2-fold. The same effect was exerted AP-1 dictate the differential recruitment of CRTC1 toward by ectopic expression of the DN CRTC1, demonstrating that both AP-1 or CREB. Because CRTCs are highly phosphorylated AP-1 and CRTC1 are necessary for M-M2–dependent tumor cell proteins, one possibility is that different signals modulate the proliferation. Consistent with the role observed on cellular transformation, A-CREB demonstrated a modest but significant effect affinity of CRTCs for specific transcription factors through on BrdU incorporation. Together, these findings demonstrate that differential patterns of phosphorylation. Consistent with this the CRTC1 portion of the M-M2 oncoprotein is involved in cellular hypothesis, we have observed that AP-1 co-precipitates with a proliferation through AP-1 and CREB. more slowly migrating band of CRTC1, which perhaps repre- sents a phosphorylated form. To date, phosphorylation of Discussion CRTCs has been described mainly as an inhibitory event. Indeed, The AP-1 complex comprises several transcription factors that CRTC2 is phosphorylated by the SIK2 kinase at Serine 171 (Ser are regulated by mitogenic stimuli and are involved in critical 164 in CRTC1) (4) or by the MARK2 kinase at Ser 275 (10), biological and pathological processes. The activation of AP-1 by which promote cytoplasmic retention through association with mitogens is achieved through multiple well-characterized mech- 14-3-3 proteins. In contrast, a recent report proposed that the anisms, including increased synthesis of single protein compo- mitogen-activated kinase MEKK1 phosphorylates CRTC1 at a nents (i.e., c-Jun/c-Fos) and changes in their dimerization mod- distinct, yet uncharacterized residue that provokes nuclear trans- ules, phosphorylation levels, and DNA binding affinity (24). location and transcriptional activation (13). Therefore, AP-1

1448 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0808749106 Canettieri et al. Downloaded by guest on September 24, 2021 800 250 A B * C IP: Mect1-Maml2 5% Input * Flag peptide - + - + 600 200 c-Fos Empty 150 c-Jun 400 A-Fos 100 WB: c-Jun, c-Fos Fold Change 200 * Fold Change 50 * ** ** ** ** 0 0 Reblot: Mect1-Maml2 M-M2 --+ + - M-M2 C1 DN M-M2 TPA - +- + C1 DN

100 D E 100 F 80 80 * * Control M-M2 60 60 MMP1 * * 40 40

* % Proliferation Foci Number MTIIA 20 20

0 0 TIMP1 - + - + - + M-M2 A-Fos - + - - - - + + - - A-Fos A-CREB - - + - GAPDH - - - - + + A-CREB C1 DN - - - +

Fig. 4. The M-M2 fusion oncoprotein enhances AP-1 function. (A) For co-immunoprecipitation, cell extracts of HeLa cells transfected with Flag–M-M2, HA–c-Jun, and HA–c-Fos were immunoprecipitated with Flag antibodies, after which Western blot analysis was performed using anti-HA and anti-Flag antibodies. (B) MMP1-Luc reporter was transfected in HeLa cells together with plasmids encoding M-M2, A-Fos, or empty vector. After 24 h, cells were exposed to TPA for 10 h where indicated, and luciferase assays were performed as described previously. *M-M2 versus control, P Ͻ .001; **M-M2 ϩ A-Fos versus M-M2, P Ͻ .05. (C) HeLa cells were transfected with plasmids encoding M-M2 and C1-DN as indicated, and the luciferase assay was performed as described earlier. *M-M2 versus control, P Ͻ .001; **M-M2 ϩ C1 DN versus M-M2, P Ͻ .05. (D) HeLa cells were transfected with plasmid encoding M-M2 or empty vector for 24 h, and RT-PCR was performed with primers amplifying MMP1, MTIIA, TIMP1, and GAPDH mRNAs. (E) In the focus formation assay, RK3E cells were transfected with the speciﬁc plasmids. After 3 weeks, cells were ﬁxed, stained with crystal violet, and scored for foci formation. Results represent the average Ϯ SD of 3 independent experiments, each performed in duplicate. *M-M2 ϩ A-Fos versus M-M2, P Ͻ .01; **M-M2 ϩ A-CREB versus M-M2, P Ͻ .05. (F) NCI-H292 cells were transfected with plasmids encoding A-Fos, A-CREB, C1-DN, GFP, or empty vector for 24 h. BrdU was then added to the cells for 5 h. After incubation, the cells were processed as described earlier, and the percent incorporation of BrdU was measured in the population of transfected cells. Results represent the average Ϯ SD of 3 separate experiments, each performed in triplicate. *P Ͻ .05.

may associate with an active form of CRTC1 phosphorylated at CREB and AP-1, which likely resides in the C-terminal domain of the MEKK sites. CRTC1. In keeping with this, the spectrum of the target genes Consistent with its role as an AP-1 enhancer, CRTC1 is involved induced by the M-M2 fusion protein includes CREB (22, 23) and in cellular proliferation, thus providing a novel biological function AP-1 targets, including MMP10 (25), PTN (26), IL-6 (27), KLF4 for this coactivator. Indeed, cells lacking CRTC1 fail to proliferate (28), THBS1 (29), TFF1 (30), CYR61 (31), and DUSP1 (32). in response to ectopically expressed c-Jun, and CRTC1 enhances An intriguing question raised by our findings is whether aberrant AP-1–dependent cellular growth. Because aberrant activation of regulation of CRTC1 alone may cause neoplastic transformation. cellular proliferation is a major cause of tumorigenesis, these In this regard, mutational inactivation of LKB1, which normally findings have led us to investigate the role of the AP-1–CRTC1 functions to promote CRTC phosphorylation and inhibition interaction in the pathogenesis of tumors caused by the CRTC1-M2 through activation of the AMPK family (33), causes the cancer fusion protein (21). It was previously reported that the transforming predisposition syndrome Peutz-Jeghers syndrome (PJS). This sug- potential of this oncoprotein requires both CRTC1 and M2 do- gests that constitutive dephosphorylation and activation of CRTCs mains, whereas neither of the 2 full-length proteins alone is may be associated with tumor formation in cells lacking Lkb1. sufficient to induce cellular transformation when overexpressed in Future studies on the PJS and related tumors will help clarify this cells (22). Gene profiling studies have shown that several CREB issue. target genes are up-regulated by M-M2, whereas Notch targets do In conclusion, our work indicates that rather than being only not seem to be affected (22, 23). Because M-M2 is known to bind CREB-dedicated coactivators and sensors of metabolic signals CREB and recruit p300/CBP, this oncoprotein has been proposed (5–8), CRTCs also play a crucial role in AP-1–dependent cellular proliferation and transformation. CRTCs appear to regulate mul-

to mimic the constitutive activation of CREB (22). CELL BIOLOGY Our findings with a DN inhibitor confirm that the CRTC1 tiple biological outcomes through distinct signaling pathways and portion of M-M2 is necessary to induce cellular transformation. transcription factors, and their aberrant regulation may underlie Because the CRTC1 region of M-M2 binds both CREB and AP-1, different disease states, including diabetes and cancer. we analyzed the relative contribution of both transcription factors. Materials and Methods We found that both AP-1 and, to lesser extent, CREB are needed Reagents, Antibodies, and Plasmids. TPA, forskolin, and IBMX were purchased to support the transforming effect of M-M2. Therefore, it is possible from Sigma. Geneticin was purchased from Invitrogen. MMP1-Luc and MMP1-Luc that the oncogenic potential of M-M2 is linked to its ability to ⌬TRE plasmids were provided by Connie Brinckerhoff. Zeo A-Fos and Zeo A-CREB constitutively activate both classes of transcription factors at the were provided by Charles Vinson. CRTC1, 2, and 3 wild-type and mutant plasmids, same time, due to the loss of the discriminating function toward Flag–M-M2, were described previously (4). The following antibodies were used:

Canettieri et al. PNAS ͉ February 3, 2009 ͉ vol. 106 ͉ no. 5 ͉ 1449 Downloaded by guest on September 24, 2021 rabbit anti-c-Jun (Santa Cruz sc-1694, sc-45), goat anti-actin (Santa Cruz sc 1616), and the reaction was stopped with 0.125 M glycine for 5Ј. Cells were washed and rabbit anti-TIMP1 (Santa Cruz sc-5538), rabbit anti-CRTC1 (Cell Signaling 2501), harvested, and cytoplasmatic membranes were lysed with lysis buffer (5 mM rabbit anti-CREB (from Marc Montminy’s laboratory, Salk Institute), mouse an- Pipes, 85 mM KCl, and 0.5% Nonidet P-40). After centrifugation, nuclei were lysed ti-HA (Santa Cruz sc-7392), and mouse anti-Flag M2 (Sigma A8592). with sonication buffer [1% SDS, 10 mM EDTA, and 50 mM Tris (pH 8) supple- Plasmids encoding HA-tagged c-Jun and c-Fos and their bZIP regions were mented with protease inhibitors] and sonicated to obtain chromatin fragments cloned by PCR amplification of cDNA and cloned into a pcDNA3 vector (Invitro- of about 400–600 nucleotides. After sonication, the lysates were precleared for gen). To generate the 3ϫ AP-1-Luc plasmid, an oligonucleotide containing 3 1 h, diluted with 9 volumes of dilution buffer [0.01% SDS, 1.2 mM EDTA, 16.7 mM copies of the AP-1 TRE (TGACTCA) was ligated to a TK-luc promoter vector. Tris⅐HCl (pH 8), 1.1% Triton X-100, and 167 mM NaCl] and incubated with the specific antibodies overnight. The next day, salmon sperm–saturated protein A or RT-PCR and RNA Interference. RNA was extracted with the TRIzol reagent G beads (Upstate) were added for 1 h, after which the lysates were washed 5 times ⅐ (Invitrogen). One microgram of total RNA was reverse-transcribed with Su- with Buffer A [0.1% SDS, 2 mM EDTA, 20 mM Tris HCl (pH 8), 1% Triton X-100, and ⅐ perScript II reverse transcriptase and random examers (Invitrogen). PCR was 150 mM NaCl], 4 times with Buffer B [0.1% SDS, 2 mM EDTA, 20 mM Tris HCl (pH ⅐ performed with primers complementary to the indicated target genes. Knock- 8), 1% Triton X-100, and 500 mM NaCl], and once with Buffer TE [10 mM Tris HCl (pH 8) and 1 mM EDTA]. After the final washing, the immunocomplexes were down experiments were performed using siRNA smart pools (Dharmacon). eluted with elution buffer (1% SDS and 100 mM NaHCO )30Ј at room temper- Between 50 and 100 ng of siRNA was transfected in HeLa cells with Lipo- 3 ature, and after the addition of 200 mM NaCl, the cross-linking was reversed with fectamine 2000 (Invitrogen). After transfection, cells were grown for 72 h and an overnight incubation at 65 °C. Subsequently, the samples were digested with treated as indicated. proteinase K and RNase A for2hat42°C,andtheDNAwaspurified and precipitated. Eluted DNA was PCR-amplified with primers encompassing the AP-1 Cells, Transfections, and Luciferase Assays. HeLa and RK3E cells were grown in site of human MMP1 promoter or the GAPDH gene as controls. The following DMEM, 2 mM L-glutamine, 10% FBS, and antibiotics. NCI-H292 cells were grown antibodies were used: rabbit polyclonal anti-CRTC1 (Cell Signaling) and rabbit in RPMI 1640, 10% FBS, 2 mM L-glutamine, and antibiotics. Transfections were polyclonal anti–c-Jun (Santa Cruz). carried out with Lipofectamine 2000. For luciferase assays, 250 ng of total DNA was transfected with Lipofectamine 2000 for 24 h in 24-well plates. For TPA Focus Formation and Proliferation Assays. Focus formation assays were per- treatments, 100 ng/mL of TPA was added to the culture medium for 10 h. After formed as described previously (21). In brief, RK3E cells were plated in 100-mm transfection and treatments, luciferase activity was measured using a Promega dishes and transfected with 10 ␮g of total DNA with the specific plasmids. After dual-luciferase assay system. Results are expressed as fold change and represent transfection, cells were grown for 3 weeks, fixed, stained with crystal violet, and Ϯ the average SD of at least 3 experiments, each performed in triplicate. assessed for foci formation. To perform colony-formation assays, HeLa cells were transfected with the Coimmunoprecpitation and GST Pulldown Assays. HeLa cells were transfected specific plasmids. After 24 h, cells were detached, diluted, and grown with G418 with the indicated plasmids and Lipofectamine 2000 for 24 h. After transfection, for 10–14 days to allow colony formation. Colonies were stained with Coomassie cells were lysed with RIPA buffer [0.5% sodium deoxycholate, 50 mM Tris⅐HCl (pH Blue, and colony numbers and sizes were measured. BrdU incorporation was 7.6), 1% Nonidet P-40, 0.1% SDS, 140 mM NaCl, 5 mM EDTA (pH 8), 2 mM sodium done using a Roche BrdU-labeling assay kit as described previously (35). pyrophosphate, and protease inhibitors], and then the cellular extracts were immunoprecipitated for 2 h. After immunoprecipitation and extensive washings ACKNOWLEDGMENTS. This work was supported by the Associazione Italiana with RIPA buffer, Western blot analysis was performed using standard tech- per la Ricerca sul Cancro, Telethon Grant GGP07118, the Ministry of University niques. GST pulldown assays were performed as described previously (1). GST- and Research, the Ministry of Health, the Cenci-Bolognetti Foundation, the CRTC1 1–142 (4) and GST proteins were expressed in Escherichia coli with stan- Center of Excellence for Biology and Molecular Medicine, and the Rome Oncog- dard techniques. enomic Center. We thank Drs. Ian Clark and Connie Brinckerhoff for the MMP1 luciferase plasmids, Charles Vinson for the Zeo A-Fos and Zeo A-CREB vectors, Antonella Siena for the NCI-H292 cells, Marc Montminy for the CREB and CRTC Chromatin Immunoprecipitation. ChIP was performed as described previously reagents, and Michael Conkright for a critical reading of the manuscript and (34), with some modifications. Cells were cross-linked 10Ј with 1% formaldehyde, advice.

1. Conkright MD, et al. (2003) TORCs: Transducers of regulated CREB activity. Mol Cell 20. Bahr MJ, et al. (1999) Control of the tissue inhibitor of metalloproteinases-1 promoter in 12:413–423. culture-activated rat hepatic stellate cells: Regulation by activator protein-1 DNA binding 2. Iourgenko V, et al. (2003) Identification of a family of cAMP response element–binding proteins. Hepatology 29:839–848. protein coactivators by genome-scale functional analysis in mammalian cells. Proc Natl 21. Tonon G, et al. (2003) t(11;19)(q21;p13) translocation in mucoepidermoid carcinoma Acad SciUSA100:12147–12152. creates a novel fusion product that disrupts a Notch signaling pathway. Nat Genet 3. Amelio AL, et al. (2007) A coactivator trap identifies NONO (p54nrb) as a component of the 33:208–213. cAMP-signaling pathway. Proc Natl Acad SciUSA104:20314–20319. 22. Wu L, et al. (2005) Transforming activity of MECT1-MAML2 fusion oncoprotein is mediated 4. Screaton RA, et al. (2004) The CREB coactivator TORC2 functions as a calcium- and by constitutive CREB activation. EMBO J 24:2391–2402. cAMP-sensitive coincidence detector. Cell 119:61–74. 23. Coxon A, et al. (2005) Mect1-Maml2 fusion oncogene linked to the aberrant activation of 5. Koo SH, et al. (2005) The CREB coactivator TORC2 is a key regulator of fasting glucose cyclic AMP/CREB-regulated genes. Cancer Res 65:7137–7144. metabolism. Nature 437:1109–1111. 24. Shaulian E, Karin M (2002) AP-1 as a regulator of cell life and death. Nat Cell Biol 6. Canettieri G, et al. (2005) Dual role of the coactivator TORC2 in modulating hepatic glucose 4:E131–E136. output and insulin signaling. Cell Metab 2:331–338. 25. Crawford HC, Matrisian LM (1996) Mechanisms controlling the transcription of matrix 7. Dentin R, et al. (2007) Insulin modulates gluconeogenesis by inhibition of the coactivator metalloproteinase genes in normal and neoplastic cells. Enzyme Protein 49:20–37. TORC2. Nature 449:366–369. 26. Florin L, Maas-Szabowski N, Werner S, Szabowski A, Angel P (2005) Increased keratinocyte 8. Dentin R, Hedrick S, Xie J, Yates J 3rd, Montminy M (2008) Hepatic glucose sensing via the proliferation by JUN-dependent expression of PTN and SDF-1 in fibroblasts. J Cell Sci CREB coactivator CRTC2. Science 319:1402–1405. 118:1981–1989. 9. Wu Z, et al. (2006) Transducer of regulated CREB-binding proteins (TORCs) induce PGC- 27. Szabo-Fresnais N, Blondeau JP, Pomerance M (2008) Activation of the cAMP pathway 1alpha transcription and mitochondrial biogenesis in muscle cells. Proc Natl Acad Sci U S synergistically increases IL-1–induced IL-6 gene expression in FRTL-5 thyroid cells: Involve- A 103:14379–14384. ment of AP-1 transcription factors. Mol Cell Endocrinol 284:28–37. 10. Jansson D, et al. (2008) Glucose controls CREB activity in islet cells via regulated phosphor- 28. Mahatan CS, Kaestner KH, Geiman DE, Yang VW (1999) Characterization of the structure ylation of TORC2. Proc Natl Acad SciUSA105:10161–10166. and regulation of the murine gene encoding gut-enriched Kruppel-like factor (Kruppel- 11. Kovacs KA, et al. (2007) TORC1 is a calcium- and cAMP-sensitive coincidence detector like factor 4). Nucleic Acids Res 27:4562–4569. involved in hippocampal long-term synaptic plasticity. Proc Natl Acad Sci U S A 104:4700– 4705. 29. Kim SA, Um SJ, Kang JH, Hong KJ (2001) Expression of thrombospondin-1 in human 12. Bittinger MA, et al. (2004) Activation of cAMP response element–mediated gene expres- hepatocarcinoma cell lines and its regulation by transcription factor Jun/AP-1. Mol Cell sion by regulated nuclear transport of TORC proteins. Curr Biol 14:2156–2161. Biochem 216:21–29. 13. Siu YT, Ching YP, Jin DY (2008) Activation of TORC1 transcriptional coactivator through 30. Espino PS, Li L, He S, Yu J, Davie JR (2006) Chromatin modification of the trefoil factor 1 MEKK1-induced phosphorylation. Mol Biol Cell 19:4750–4761. gene in human breast cancer cells by the Ras/mitogen-activated protein kinase pathway. 14. Eferl R, Wagner EF (2003) AP-1: A double-edged sword in tumorigenesis. Nat Rev Cancer Cancer Res 66:4610–4616. 3:859–868. 31. Tsai MS, Bogart DF, Li P, Mehmi I, Lupu R (2002) Expression and regulation of Cyr61 in 15. Hess J, Angel P, Schorpp-Kistner M (2004) AP-1 subunits: Quarrel and harmony among human breast cancer cell lines. Oncogene 21:964–973. siblings. J Cell Sci 117:5965–5973. 32. Furst R, et al. (2005) Atrial natriuretic peptide induces mitogen-activated protein kinase 16. Auble DT, Brinckerhoff CE (1991) The AP-1 sequence is necessary but not sufficient for phosphatase-1 in human endothelial cells via Rac1 and NAD(P)H oxidase/Nox2-activation. phorbol induction of collagenase in fibroblasts. Biochemistry 30:4629–4635. Circ Res 96:43–53. 17. Olive M, et al. (1997) A dominant negative to activation protein-1 (AP-1) that abolishes 33. Shaw RJ, et al. (2005) The kinase LKB1 mediates glucose homeostasis in liver and thera- DNA binding and inhibits oncogenesis. J Biol Chem 272:18586–18594. peutic effects of metformin. Science 310:1642–1646. 18. Conkright MD, et al. (2003) Genome-wide analysis of CREB target genes reveals a core 34. Canettieri G, et al. (2003) Attenuation of a phosphorylation-dependent activator by an promoter requirement for cAMP responsiveness. Mol Cell 11:1101–1108. HDAC–PP1 complex. Nat Struct Biol 10:175–181. 19. Lee W, Mitchell P, Tjian R (1987) Purified transcription factor AP-1 interacts with TPA- 35. Di Marcotullio L, et al. (2004) REN(KCTD11) is a suppressor of Hedgehog signaling and is inducible enhancer elements. Cell 49:741–752. deleted in human medulloblastoma. Proc Natl Acad SciUSA101:10833–10838.

1450 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0808749106 Canettieri et al. Downloaded by guest on September 24, 2021