Oncogene (2000) 19, 1379 ± 1385 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Activation of the c-fos enhancer by the Erk MAP kinase pathway through two sequence elements: the c-fos AP-1 and p62TCF sites

Yan Wang1 and Ron Prywes*,1

1Department of Biological Sciences, Columbia University, New York, NY 10027, USA

The c-fos enhancer can be activated by many signaling Zhang, 1993; Sadowski et al., 1993). On the 3' side of the pathways through distinct elements of the enhancer. The SRE there is an AP1-like site termed the c-fos AP1 site enhancer contains at its core the serum response element (FAP1). The function of the FAP1 site has been unclear (SRE) that binds serum response factor (SRF). On the 5' since mutation of this site has not had a strong e€ect on side of the SRE is a site for p62TCF which binds only induction of the promoter in transfection experiments when SRF is bound as well. p62TCF is encoded by three (Fisch et al., 1987; Siegfried and Zi€, 1989). Notably, ets-related genes, Elk-1, SAP1 and SAP2. Each of these however, mutation of the site had a strong e€ect on factors contain a transcriptional activation domain that expression of reporter genes introduced stably to is activated by phosphorylation by MAP kinases. On the transgenic mice (Robertson et al., 1995). 3' side of the SRE is the `c-fos AP1 site' (FAP1) whose The c-fos enhancer can be induced primarily by two role has been less clear. We ®nd here that the FAP1 site types of signaling pathways. The ®rst involves contributes strongly to phorbol ester (TPA) and Erk phosphorylation of p62TCF by MAPK family members MAP kinase activation of the c-fos enhancer and that while in the second rho can activate the SRE both the p62TCF and FAP1 sites are required for e€ective independently of p62TCF. p62TCF is encoded by any of activation of the enhancer. We further ®nd that the three ets-related genes: elk-1, SAP1 or SAP2 (also FAP1 site binds ATF1 and CREB from HeLa nuclear known as Net1 and Erp1) (Treisman, 1994). These extracts and that the phosphorylation of these factors is factors have three related domains: the ets DNA induced by TPA. ATF1 and CREB can be phosphory- binding motif, an SRF interaction domain, and a lated by Rsk2 which is a protein kinase directly activated transcriptional activation domain that is regulated by by Erk MAP kinases. These results suggest a signaling MAPK phosphorylation (Treisman, 1996). Each of the pathway in which Erk MAP kinase activates the c-fos MAPK family members, Erk-1/2, JNK and p38 have enhancer by direct phosphorylation of p62TCF and by been found to phosphorylate and activate Elk-1 (Gille activation of Rsk related kinases that phosphorylate et al., 1992; Marais et al., 1993; Whitmarsh et al., 1995, ATF1 and CREB. Oncogene (2000) 19, 1379 ± 1385. 1997). Rho activation of the SRE requires SRF but the mechanism is unknown (Hill et al., 1995). Recently it Keywords: c-fos; FAP1; Erk; MAPK was found that rho and LIM kinase regulation of actin treadmilling were critical to a pathway leading to SRE activation (Sotiropoulos et al., 1999). Introduction At 760 of the c-fos promoter there is a cAMP responsive element (CRE) that can also mediate c-fos is an immediate early gene that can be activated by induction of the promoter. The CRE can mediate cAMP a variety of growth factors and mitogens through several and calcium induction in certain cell types but may also di€erent signaling pathways (Johansen and Prywes, contribute to growth factor induction (Berkowitz et al., 1995). Activation by most of these pathways can be 1989; Fisch et al., 1989; Sassone-Corsi et al., 1988; Sheng mediated by the c-fos enhancer located around 300 et al., 1990). The CRE is bound by CREB and the related nucleotides 5' of the transcriptional initiation site transcription factor ATF1 (Ginty et al., 1994; Mon- (Treisman, 1992). The central element of the c-fos tminy, 1997; Sassone-Corsi et al., 1988). Each is enhancer is the Serum Response Element (SRE) which phosphorylated on a conserved serine (amino acid 133 binds Serum Response Factor (SRF) (Johansen and in CREB and 63 in ATF1) that causes activation of the Prywes, 1995). The SRE is ¯anked on the 5' side by a site factors (Gonzalez and Montminy, 1989; Rehfuss et al., which binds p62TCF which can only bind when SRF is 1991). This site is phosphorylated by cAMP-dependent binding to the adjacent SRE. This is mediated by direct protein kinase (PKA) and calcium-calmodulin depen- contacts of p62TCF with SRF and its DNA binding site dent protein kinase (CaMK) which mediate cAMP and (Shaw et al., 1989; Treisman, 1994). Besides the TCF- calcium induction, respectively (Liu et al., 1993; SRE site, the c-fos enhancer contains the Sis Inducible Matthews et al., 1994; Montminy, 1997). Growth factor- Element (SIE) which is responsive to PDGF (Wagner et and stress-induced pathways can also activate CREB al., 1990). The SIE binds STAT family members (Fu and through phosphorylation of serine 133. These pathways act through the MAPKs Erk-1/2 and p38 (Iordanov et al., 1997; Tan et al., 1996; Xing et al., 1996, 1998). Rather than direct phosphorylation, these kinases activate *Correspondence: R Prywes, Department of Biological Sciences, downstream kinases that phosphorylate CREB (and Columbia University, Fairchild 813B, MC 2420, 1212 Amsterdam presumably ATF1). Erk-1/2 activates RSK1, 2 and 3 to Avenue, New York, NY 10027, USA Received 15 October 1999; revised 29 December 1999; accepted 13 phosphorylate CREB while p38 activates MAPKAP January 2000 kinase-2 (Iordanov et al., 1997; Tan et al., 1996; Xing et Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1380 al., 1996, 1998). In addition, both Erk-1/2 and p38 can activate the RSK-related kinase MSK1 which can then phosphorylate CREB (Deak et al., 1998; Pierrat et al., 1998). The role of RSK2 in c-fos regulation was particularly highlighted in ®broblasts from Con-Lowry syndrome (CLS) patients that have a defect in the RSK2 gene. In these cells EGF induction of c-fos as well as EGF-induced phosphorylation of CREB were severely impaired (De Cesare et al., 1998). In contrast, serum induction of c-fos and CREB phosphorylation were unaltered in CLS cells, suggesting that di€erent signaling pathways are utilized. In this study we have found that, along with the TCF-SRE site, the FAP1 site contributes to MAPK activation of c-fos reporter genes induced by the phorbol ester TPA or by activated Raf. In addition, the FAP1 site binds predominantly ATF1 and CREB from HeLa cell nuclear extracts. ATF1 and CREB phosphorylation are induced by TPA and dominant negative CREB and RSK2 constructs inhibited Erk activation of the reporter genes. These results suggest a model for Erk activation of the c-fos enhancer involving phosphorylation of TCF and ATF1/CREB by Erk and RSK family members, respectively.

Results

The c-fos AP1 site contributes to TPA and Raf-Erk activation Figure 1 Requirement of TCF and FAP1 sites for TPA induction of the c-fos enhancer. (a) HeLa cells were transfected To check whether the c-fos AP1 (FAP1) site is functional with wild-type or mutant c-fos-luciferase reporter genes together in activating c-fos transcription, we constructed a series with pRLSV40P as an internal control. The transfected cells were serum-starved and induced with serum or TPA as described in of luciferase reporter genes with the c-fos enhancer with Materials and methods. Cell lysates were assayed for ®re¯y or without the FAP1 site (Figure 1b). These reporter luciferase activity and normalized for Renilla luciferase activity. genes were transfected into HeLa cells and assayed for The results are averages with standard error of the mean of four serum or TPA induction. The FAP1 site had no experiments. The fold serum or TPA induction compared to uninduced cells are indicated for each reporter gene above the signi®cant e€ect on serum induction but signi®cantly bars. (b) The c-fos-luciferase reporter genes and mutations are increased TPA induction (Figure 1a; compare WT to diagrammed. A section of the mouse c-fos promoter, 7355 to DFAP1). The induction of the DFAP1 reporter was 7285, was fused to the human c-fos minimal promoter at 753. weak even though it contains a functional TCF site that The mutations are indicated by the `x's and abolish binding of TCF was previously found to mediate TPA induction p62 (pm18) or STATs (sie) to the TCF or SIE sites, respectively (Graham and Gilman, 1991). To test whether the TCF and FAP1 sites acted together to mediate TPA induction, we mutated each site alone or in combination activate the c-fos enhancer through the TCF and FAP1 (Figure 1b). We found that mutation of the TCF site sites. Activated Raf (RafBXB) (Bruder et al., 1992) (pm18) strongly reduced TPA induction although some only very weakly activated the wt reporter gene but residual induction was observed when the FAP1 site was strongly activated when cotransfected with wt Erk2 present. Mutation of both sites completely abolished (Figure 2a). This activation required both the FAP1 TPA induction (Figure 1a). Since the basal level of and TCF sites since mutation of each strongly reduced expression was reduced by the DFAP1 and pm18 expression (Figure 2b). mutations, the fold induction by TPA was only modestly a€ected (as indicated by the numbers above the bars). ATF1 and CREB bind the FAP1 site However, no induction was observed when both the FAP1 and TCF sites were mutated. These results Since the FAP1 site a€ected TPA and Erk activation of strongly suggest that TPA activates the c-fos enhancer the c-fos enhancer, we investigated which factors bind through both the TCF and FAP1 sites. this site. We used a gel mobility shift assay with HeLa We also tested the role of the SIE site which binds cell nuclear extract and a probe, SREFAP1, spanning STAT factors (Fu and Zhang, 1993; Sadowski et al., both the SRE and FAP1 sites (Figure 3b). Four bands 1993). Mutation of this site increased TPA induction were observed that were competed speci®cally by suggesting that there may be a negative element acting excess, unlabeled SREFAP1 oligonucleotide (Figure through this site (Figure 1a). 3a). Band 1 is SRF since it was competed speci®cally TPA activation of c-fos has been found to be by excess SRE oligonucleotide and since it was through activation of Erk MAP kinases and phosphor- speci®cally blocked by anti-SRF sera (Figures 3a and ylation of p62TCF (Treisman, 1994, 1996). We therefore 4a). The three other bands that migrate faster than tested whether activation of the Erk pathway can SRF, were competed by unlabeled FAP1 but not a

Oncogene Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1381

Figure 2 Transcriptional activation of the TCF and FAP1 sites by Raf and Erk. (a) HeLa cells were transfected with an activated form of Raf (RafBxB) and wild-type Erk-2 (3 mg each), as indicated, with pWTGL3 reporter gene and pRLSV40P. After 16 h of transfection, the cells were serum-starved for 24 h and assayed for ®re¯y and Renilla luciferase activities. The results were normalized to the internal control of Renilla luciferase activity and are the averages with the standard error of the mean of four experiments. (b) HeLa cells were transfected with RafBxB and wild-type Erk-2 with the indicated reporter genes and pRLSV40P as an internal control as described above

We also tested whether there was any change in binding to the SREFAP1 probe using nuclear extracts from TPA-induced HeLa cells, however there was no signi®cant change compared to uninduced extracts (data not shown). Since the FAP1 site is similar to ATF sites as well as AP1 sites (Figure 3b), we tested whether the factors binding to the FAP1 site were ATF or AP1 (c-jun or c- fos) family members using speci®c sera. Antisera to ATF1 speci®cally blocked binding of bands 2 and 4, while there was no e€ect of antisera to ATF2, 3, 4 or 6 or to c-jun or c-fos (Figure 4a). We previously found that these sera speci®cally blocked binding of their cognate factors (Clarke et al., 1998, and data not shown). The anti-ATF1 sera was made against full- length ATF1 and we have found that it does not crossreact with CREB (Clarke et al., 1998, and unpublished results). To determine whether bands 2 or 4 contain CREB we used CREB-speci®c sera. This sera speci®cally blocked band 2 but not 3 or 4 (Figure 4b). These results are consistent with band 4 being a homodimer of ATF1 while band 2 is a heterodimer of ATF1 and CREB. This is similar to the case for two factors in HeLa nuclear extract binding to the c-jun ATF site (Clarke et al., 1998). Since band 3 did not Figure 3 Binding of nuclear factors to the FAP1 site. (a) Gel react with any of the sera, we could not identify it here. mobility shift assays were performed with HeLa cell nuclear Other factors binding the SRE that have been extracts and a 32P-labeled oligonucleotide spanning the c-fos SRE identi®ed include NFIL6/c/EBPb, YY1, SRE-ZBP and FAP1 sites (SREFAP1). A 50-fold excess of unlabeled and TFII-I-Phox1 (Attar and Gilman, 1992; Gruene- oligonucleotides were added as indicated to demonstrate speci®c binding. RP4 was added as a nonspeci®c oligonucleotide control. berg et al., 1997; Gualberto et al., 1992; Kim et al., The position of speci®c bands are indicated to the left. (b) The 1998; Metz and Zi€, 1991). None of these are likely to sequences of the oligonucleotides used are shown. The SRE account for band 3 or the faint band migrating below sequence is underlined and the FAP1 sequence is in bold. band 4 in Figure 3a, since they would not be expected Flanking sequences not derived from the c-fos gene are in small to be competed by the FAP1 oligonucleotide which capitals. The mutated nucleotides in FAP1m1 are in lower case. The consensus binding sites for AP1 and CREB binding are does not span their binding sites (Attar and Gilman, indicated (Mitchell and Tjian, 1989; Montminy, 1997) 1992; Grueneberg et al., 1997; Gualberto et al., 1992; Kim et al., 1998; Metz and Zi€, 1991). E12 was also found to bind the SRE and may bind the FAP1 double stranded oligonucleotide with three bases sequence since it contains its core CANNTG binding mutated in the FAP1 site (FAP1m1) (Figure 3a). site (Metz and Zi€, 1991). All of the above factors, These results demonstrate that bands 2 ± 4 represent however, would be predicted to bind to the DFAP1 factors binding to the FAP1 site. reporter gene (Figure 1) such that they are unlikely to

Oncogene Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1382

Figure 4 Binding of ATF1 and CREB to the c-fos FAP1 site. Antibodies (1 ml) to the indicated proteins were added to the gel mobility shift assay of HeLa nuclear extract and the SREFAP1 site as described in Figure 3. The positions of migration of SRF and three factors binding to the FAP1 site are indicated to the Figure 5 TPA-induced phosphorylation of ATF1, CREB and left. PBS: phosphate bu€ered saline; PI: pre-immune sera ERK. HeLa cells were treated with TPA for the indicated times. Lysates were then probed in immunoblots with antisera speci®c for (a) phospho-CREB, (b) ATF1, or (c) phospho-ERK. The positions of phospho-ATF1, phospho-CREB, ATF1 and phos- be responsible for the e€ect of FAP1 deletion on TPA pho-ERK are indicated induction. Further work will be required to determine whether E12 or other factors can account for bands detected in Figure 3a binding to the SRE-FAP1 region inducibly phosphorylated on serine 63 or 133, respec- and whether they play a role in c-fos regulation. tively, we tested whether phosphorylation at this site was important for induction of the c-fos enhancer by the ERK pathway. We cotransfected ERK-2 and Serum and TPA induced phosphorylation of ATF1 and activated RAF (RafBxB) with a CREB expression CREB vector with a serine to alanine mutation at serine 133 Phosphorylation of CREB at serine 133 is well (CREBm1) such that it cannot be activated at this site. characterized as mediating activation of the factor in CREBm1 inhibited activation of the c-fos enhancer by response to cAMP and growth factors such as EGF about 60% suggesting that phosphorylation at serine (De Cesare et al., 1999; Montminy, 1997). ATF1 is 133 is important for induction (Figure 6a). As a highly homologous to CREB and contains the control we found that CREBm1 had no signi®cant phosphorylation site at serine 63 corresponding to e€ect on Rho activation of an SRE reporter gene CREB serine 133. Since CREB phosphorylation can be (Figure 6b). induced by EGF and NGF through MAP kinase Since it has been reported that Erk-1/2 can activate pathways (Iordanov et al., 1997; Xing et al., 1996, RSK2 which can then phosphorylate and activate 1998), we tested whether TPA induces ATF1 and CREB on serine 133 (Xing et al., 1996, 1998), we tested CREB phosphorylation in HeLa cells. We used anti- whether RSK2 is required for activation of the c-fos phosphoCREB sera that is speci®c for CREB and enhancer by the Erk pathway. We used a dominant ATF1 phosphorylated at serines 133 and 63, respec- negative RSK2 mutant (RSK2KR100) which has a tively. Immunoblotting with this sera showed a rapid point mutation in the kinase domain and has no kinase increase in the level of phosphorylation of both ATF1 activity (Xing et al., 1996). Raf plus ERK activation of and CREB in cells treated with TPA (Figure 5a). The the reporter gene was inhibited by RSK2KR100 by level of ATF1 did not change as shown by immuno- nearly 50% (Figure 6a). As a control, RSK2KR100 did blotting with anti-ATF1 sera (Figure 5b). Immunoblot- not inhibit RhoV14 activation of an SRE reporter gene ting with anti-phosphoErk sera showed that this MAP (Figure 6b). These results suggest that RSK2 at least kinase was activated with similar kinetics to ATF1 and partially mediates ERK activation of the enhancer CREB (Figure 5c). As a control, we immunoblotted although it is possible that the dominant negative with anti-Erk-2 sera and found no change in the level RSK2 could be inhibiting by directly binding and of Erk-2 (Figure 5d). These results are consistent with inhibiting ERK-2 and/or by blocking the activation of results that Erk activation is upstream of ATF1 and other RSK family members. CREB (Xing et al., 1996, 1998). Erk-1/2 can activate the RSK related kinases (RSK1, 2 and 3 and MSK1) that can then directly phosphorylate CREB (Deak et Discussion al., 1998; Xing et al., 1996, 1998). We have found that the FAP1 site contributes to TPA and ERK activation of the c-fos enhancer and that the Dominant negative CREB and RSK2 inhibit c-fos predominant factors in HeLa cell nuclear extracts enhancer activation binding to the FAP1 site are ATF1 and CREB. In Since ATF1 and CREB are the predominant factors in addition, we observed TPA-induced phosphorylation HeLa nuclear extract that bind the FAP1 site and are of ERK, ATF1 and CREB. Previous work has shown

Oncogene Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1383

Figure 6 Transcriptional activation by Raf and Erk is inhibited by dominant negative CREB and RSK2. (a) HeLa cells were transfected with the c-fos reporter gene pWTGL3, the internal control plasmid pRLSV40P and activated Raf (RafBxB) and wt Erk- 2 where indicated. Control vector (7) or dominant negatives CREBm1 or RSK2KR100 (3 mg each) were added as indicated. Luciferase activities were determined as in Figure 2. (b) HeLa cells were transfected as in a except that activated RhoA (RhoV14) was added with the pm18DFAP1 reporter gene that TPA and ERK can activate c-fos through phosphorylation of p62TCF (Treisman, 1994, 1996) and that ERK can activate RSK family members to phosphorylate and activate CREB (Deak et al., 1998; Xing et al., 1996, 1998). Together these results suggest a model (Figure 7) in which TPA induction causes the activation of the RAF-MEK-ERK pathway and that ERK can then signal to the c-fos enhancer in two ways. First, it can activate by direct phosphorylation and activation of the transcriptional activation domain of p62TCF (Elk-1, SAP1 or SAP2) (Treisman, 1994, 1996). Second, ERK can activate the FAP1 site through phosphorylation of RSK family kinases which phos- phorylate and activate ATF1 or CREB on serines 63 or 133, respectively. The FAP1 site in the c-fos enhancer was previously found to mediate cAMP induction of reporter constructs redundantly with other elements in the c- fos promoter such as the CRE site at 760 (Berkowitz et al., 1989; Fisch et al., 1989). However, the role of the FAP1 site in serum or growth factor induction has been less clear. Deletion of the FAP1 site from c-fos reporter constructs had no e€ect on serum, EGF or TPA induction (Fisch et al., 1987; Siegfried and Zi€, 1989). However, these reporter genes contained over 200 nucleotides of the c-fos promoter including the Figure 7 Model for TPA activation of the c-fos enhancer. See CRE site at 760. The e€ect of the FAP1 site observed Discussion for details here is likely because it was assayed without the 760 CRE which also binds ATF1 and CREB (Ginty et al., 1994; Sassone-Corsi et al., 1988) and because the e€ect or CREB to bind this site and respond to the signaling is clearer in conjunction with mutation of the TCF site. pathway proposed in Figure 7. The role of the FAP1 site has most dramatically been We found three factors from HeLa cell nuclear shown in transgenic mouse with fos-lacZ constructs extracts that speci®cally bound the FAP1 site. Using (Robertson et al., 1995). In these mice, mutation of the speci®c antisera, we can account for two of these FAP1 site nearly abolished expression of the fos-lacZ factors as containing ATF1 and CREB, while the third reporter genes in tissues constitutively expressing c-fos. factor remains unidenti®ed. ATF1 and CREB are In addition, induction of the fos-lacZ reporter was highly related basic-leucine zipper factors that can be severely impaired in brain and ®broblast cell cultures activated by phosphorylation at serines 63 and 133, from the transgenic mice. This e€ect of mutation of the respectively. They can be activated in response to FAP1 site may re¯ect the requirement for ATF1 and/ increased cAMP levels due to their phosphorylation by

Oncogene Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1384 cAMP-dependent protein kinase (De Cesare et al., (Miranti et al., 1995). Since CaMK can also activate 1999; Montminy, 1997). They can also be activated by CREB, it is possible that the FAP1 site may also phosphorylation at the same site by calcium calmodu- mediate calcium induction of the c-fos enhancer along lin dependent protein kinase and RSK1, 2 and 3 as with the SRE. well as the RSK-related kinase MSK1 (Deak et al., 1998; Matthews et al., 1994; Pierrat et al., 1998; Xing et al., 1996, 1998). The RSK family can be phosphorylated and activated by ERK MAP kinases Materials and methods while MSK1 can be activated by either ERK or p38 (Deak et al., 1998; Pierrat et al., 1998; Xing et al., Immunoblotting and gel mobility shift assays 1996, 1998). These kinases cause ATF1 and CREB to Cell lysates were prepared from serum or TPA induced HeLa respond to a wide variety of signaling pathways. cells by resuspending the cells in 0.2 ml of 36 sodium The role of RSK2 in growth factor regulation of dodecyl sulfate (SDS) sample bu€er (6% SDS, 180 mM Tris- ATF1 and CREB is particularly notable. It was HCl (pH 6.8), 30% glycerol, 0.003% bromophenol blue). The puri®ed as the main NGF-induced CREB-kinase in lysates were analysed by immunoblotting using a 1 : 1000 PC12 cells (Xing et al., 1996). Mutations in the RSK2 dilution of anti-phospho-CREB, anti-phospho-ERK (New gene and loss of RSK2 activity have also been England Biolabs), and anti-ATF1 (Clarke et al., 1998). Horse associated with Con-Lowry syndrome (CLS) which radish peroxidase conjugated goat anti-rabbit IgG (Sigma) is characterized by severe psychomotor retardation, was used as a secondary antibody. Anti-CREB sera was a gift facial and digital dymorphisms, and progressive from Dr Michael Greenberg. Antisera to ATF-2, -3, -4 and c- jun were gifts of Dr Tsonwin Hai. Anti-c-fos and anti-Erk-2 skeletal deformations (Trivier et al., 1996; Young, were from Santa Cruz Biotechnology. Anti-ATF6 was as 1988). EGF-induced CREB phosphorylation was lost described (Zhu et al., 1997). in ®broblasts from CLS patients while there was no Nuclear extracts were made from HeLa cells and analysed change in cAMP, UV or serum induced CREB by gel mobility shift assay for SREFAP1 site DNA-binding phosphorylation (De Cesare et al., 1998). EGF- activity as described (Prywes and Roeder, 1986). SREFAP1, induction of c-fos expression was also greatly reduced a double stranded 43 bp oligonucleotide spanning the SRE in CLS cells suggesting that RSK2 is required for EGF and FAP1 sites (see sequence in Figure 3b) was used as a induction of c-fos, possibly through phosphorylation of probe. CREB and ATF1. Our results further suggest that RSK activation of CREB and ATF1 activates the c-fos Plasmids enhancer through the FAP1 site adjacent to the SRE in Luciferase reporter genes: Plasmids pWTGL3, pWTDFAP- addition to the CRE site located at 760. The role of GL3, ppm18GL3, ppm18DFAPGL3, psieDFAPGL3 and RSK2 in ®broblasts is consistent with the inhibition of psiepm18DFAPGL3 contain 7355 to 7285 of the murine activation we observed with dominant negative RSK2 c-fos promoter fused to 753 to +45 of the human c-fos in HeLa cells. Further work will be required to promoter subcloned into the pGL3 luciferase vector de®nitively show that RSK2 is required for TPA (Promega) upstream of the ®re¯y luciferase gene. The activation of the c-fos promoter. It is also likely that mutations in these constructs are diagrammed in Figure 1b. other RSK family members will mediate activation of The sequences, with the mutations underlined are: WT: the c-fos enhancer in other cell types or in response to GAGCTGTTCCCGTCAATCCCTCCCTCCTTTACACAG- other inducers. G AT GTCCAT ATTA GGA CATCT GCGT CAGCAGGTT. DFAP: GAGCTGTTCCCGTCAATCCCTCCCTCCTTTA- The c-fos enhancer can be activated by multiple CACAGG ATGTC CAT AT TAGGAC ATCT GCG. pm18: signaling pathways. We have shown here that the GAGCTGTTCCCGTCAATCCCTCCCTCCTTTACACAG- RAF-MEK-ERK pathway can activate through the TAT GTC CAT AT TAG GAC ATC TGCGTC AGCAGGTT. TCF and FAP1 sites. Other MAPK family members, sie: GAGCTGCTCACGTCAATCCCTCCCTCCTTTACA- JNK and p38, can also activate through the TCF site CAGGATGTCCATATTAGGACATCTGCGTCAGCAGG- (Whitmarsh et al., 1995, 1997). The JNK pathway does TT. Expression vectors: pRafBXB was used as an activated not cause CREB phosphorylation, however p38 can form of c-raf (Bruder et al., 1992) and RhoAV14 in pEXV activate MAPKAP-kinase 2 which can phosphorylate was used as an activated form of RhoA (Qiu et al., 1995). CREB (Iordanov et al., 1997; Tan et al., 1996). Rat Erk-2 in pCMV5 was a gift of Dr C Chandra Kumar. Therefore p38 may also activate the c-fos enhancer Dominant negative RSK2 (RSK2KR100) in pMT-HA and CREBm1 (serine 133 to alanine) in pcDNA3 with an RSV through the FAP1 site. Independently of TCF, serum promoter were gifts from Dr Michael Greenberg. and lysophosphatidic acid (LPA) can activate the SRE through signaling pathways involving RhoA, a small GTPase which can induce stress ®bers (Hill et al., 1995; Transfections and luciferase assays Wang et al., 1998). We found that the FAP1 site had HeLa cells were grown in Dulbecco's modi®ed Eagle's no e€ect on activated RhoA-V14 induction of the c-fos medium (DMEM) supplemented with 10% newborn calf enhancer suggesting that Rho does not act through this serum (NCS). Cells were transfected by the calcium site (data not shown). The direct mechanism of Rho phosphate coprecipitation method (Sambrook et al., 1989). induction of the SRE is unclear, however it was Each transfection cocktail for a 60 mm diameter plate recently found that Rho and LIM kinase control of contained 2 mg of luciferase reporter gene, 0.5 mg of pRL- SV40P (Chen and Prywes, 1999) as an internal control actin treadmilling and the control of free actin (containing the SV40 early promoter driving the Renilla concentrations in the cells was critical for Rho and luciferase gene), the indicated amounts of expression LIM kinase activation of the SRE (Sotiropoulos et al., constructs and herring sperm DNA to give a total of 10 mg 1999). In an additional pathway, increased calcium DNA. HeLa cells were transfected for 16 ± 20 h and then concentrations can also activate the SRE independently serum starved in DMEM and 0.2% NCS for 24 h. Cells were of the TCF site through a CaMK dependent pathway then induced with 100 ng/ml TPA or 20% NCS for 3 h. The

Oncogene Activation of c-fos FAP1 and TCF sites Y Wang and R Prywes 1385 transfected cells were lysed and assayed for ®re¯y and Renilla Acknowledgments luciferase activity using the dual luciferase kit (Promega). The We thank Dr Michael Greenberg for RSK and CREB results were normalized to the Renilla luciferase activity of plasmids and anti-CREB sera and Dr Tsonwin Hai for the internal control. Each experiment was repeated four or anti-ATF family sera. This work was supported by grant more times. The average and the standard error of the mean CA 50329 from the National Cancer Institute and the are shown. National Institutes of Health to R Prywes.

References

Attar RM and Gilman MZ. (1992). Mol. Cell. Biol., 12, Mitchell PJ and Tjian R. (1989). Science, 245, 371 ± 378. 2432 ± 2443. Montminy M. (1997). Annu.Rev.Biochem.,66, 807 ± 822. Berkowitz LA, Riabowol KT and Gilman MZ. (1989). Mol. Pierrat B, Correia JS, Mary JL, Tomas-Zuber M and Cell. Biol., 9, 4272 ± 4281. Lesslauer W. (1998). J. Biol. Chem., 273, 29661 ± 29671. Bruder JT, Heidecker G and Rapp UR. (1992). Genes Dev., Prywes R and Roeder RG. (1986). Cell, 47, 777 ± 784. 6, 545 ± 556. Qiu RG, Chen J, McCormick F and Symons M. (1995). Proc. Chen X and Prywes R. (1999). Mol. Cell. Biol., 19, 4695 ± Natl. Acad. Sci. USA, 92, 11781 ± 11785. 4702. Rehfuss RP, Walton KM, Loriaux MM and Goodman RH. Clarke N, Arenzana N, Hai T, Minden A and Prywes R. (1991). J. Biol. Chem., 266, 18431 ± 18434. (1998). Mol. Cell. Biol., 18, 1065 ± 1073. Robertson LM, Kerppola TK, Vendrell M, Luk D, Smeyne De Cesare D, Fimia GM and Sassone-Corsi P. (1999). Trends RJ, Bocchiaro C, Morgan JI and Curran T. (1995). Biochem. Sci., 24, 281 ± 285. Neuron, 14, 241 ± 252. De Cesare D, Jacquot S, Hanauer A and Sassone-Corsi P. Sadowski HB, Shuai K, Darnell Jr JE and Gilman MZ. (1998). Proc. Natl. Acad. Sci. USA, 95, 12202 ± 12207. (1993). Science, 261, 1739 ± 1744. Deak M, Clifton AD, Lucocq LM and Alessi DR. (1998). Sambrook J, Fritsch E and Maniatis T. (1989). Molecular EMBO J., 17, 4426 ± 4441. cloning: a laboratory manual. Second edn. Cold Spring Fisch TM, Prywes R and Roeder RG. (1987). Mol. Cell. Harbor Laboratory, Cold Spring Harbor, NY. Biol., 7, 3490 ± 3502. Sassone-Corsi P, Visvader J, Ferland L, Mellon PL and Fisch TM, Prywes R, Simon MC and Roeder RG. (1989). Verma IM. (1988). Genes Dev., 2, 1529 ± 1538. Genes Dev., 3, 198 ± 211. Shaw PE, Schroter H and Nordheim A. (1989). Cell, 56, Fu XY and Zhang JJ. (1993). Cell, 74, 1135 ± 1145. 563 ± 572. Gille H, Sharrocks AD and Shaw PE. (1992). Nature, 358, Sheng M, McFadden G and Greenberg ME. (1990). Neuron, 414 ± 417. 4, 571 ± 582. Ginty DD, Bonni A and Greenberg ME. (1994). Cell, 77, Siegfried Z and Zi€ EB. (1989). Oncogene, 4, 3 ± 11. 713 ± 725. Sotiropoulos A, Gineitis D, Copeland J and Treisman R. Gonzalez GA and Montminy MR. (1989). Cell, 59, 675 ± 680. (1999). Cell, 98, 159 ± 169. Graham R and Gilman M. (1991). Science, 251, 189 ± 192. TanY,RouseJ,ZhangA,CariatiS,CohenPandCombMJ. Grueneberg DA, Henry RW, Brauer A, Novina CD, (1996). EMBO J., 15, 4629 ± 4642. Cheriyath V, Roy AL and Gilman M. (1997). Genes Treisman R. (1992). Trends Biochem. Sci., 17, 423 ± 426. Dev., 11, 2482 ± 2493. Treisman R. (1994). Curr. Opin. Genet. Dev., 4, 96 ± 101. Gualberto A, LePage D, Pons G, Mader SL, Park K, Treisman R. (1996). Curr. Opin. Cell. Biol., 8, 205 ± 215. Atchison ML and Walsh K. (1992). Mol. Cell. Biol., 12, TrivierE,DeCesareD,JacquotS,PannetierS,ZackaiE, 4209 ± 4214. Young I, Mandel JL, Sassone-Corsi P and Hanauer A. Hill CS, Wynne J and Treisman R. (1995). Cell, 81, 1159 ± (1996). Nature, 384, 567 ± 570. 1170. Wagner BJ, Hayes TE, Hoban CJ and Cochran BH. (1990). Iordanov M, Bender K, Ade T, Schmid W, Sachsenmaier C, EMBO J., 9, 4477 ± 4484. Engel K, Gaestel M, Rahmsdorf HJ and Herrlich P. Wang Y, Falasca M, Schlessinger J, Malstrom S, Tsichlis P, (1997). EMBO J., 16, 1009 ± 1022. Settleman J, Hu W, Lim B and Prywes R. (1998). Cell Johansen FE and Prywes R. (1995). Biochim. Biophys. Acta, Growth Di€er., 9, 513 ± 522. 1242, 1 ± 10. Whitmarsh AJ, Shore P, Sharrocks AD and Davis RJ. Kim DW, Cheriyath V, Roy AL and Cochran BH. (1998). (1995). Science, 269, 403 ± 407. Mol. Cell. Biol., 18, 3310 ± 3320. Whitmarsh AJ, Yang SH, Su MS, Sharrocks AD and Davis LiuF,ThompsonMA,WagnerS,GreenbergMEandGreen RJ. (1997). Mol. Cell. Biol., 17, 2360 ± 2371. MR. (1993). J. Biol. Chem., 268, 6714 ± 6720. Xing J, Ginty DD and Greenberg ME. (1996). Science, 273, Marais R, Wynne J and Treisman R. (1993). Cell, 73, 381 ± 959 ± 963. 393. Xing J, Kornhauser JM, Xia Z, Thiele EA and Greenberg Matthews RP, Guthrie CR, Wailes LM, Zhao X, Means AR ME. (1998). Mol. Cell. Biol., 18, 1946 ± 1955. and McKnight GS. (1994). Mol. Cell. Biol., 14, 6107 ± Young ID. (1988). J. Med. Genet., 25, 344 ± 348. 6116. Zhu C, Johansen FE and Prywes R. (1997). Mol. Cell. Biol., Metz R and Zi€ E. (1991). Oncogene, 6, 2165 ± 2178. 17, 4957 ± 4966. Miranti CK, Ginty DD, Huang G, Chatila T and Greenberg ME. (1995). Mol. Cell. Biol., 15, 3672 ± 3684.

Oncogene