Transforming p21ras mutants and c-Ets-2 activate the cyclin D1 promoter through distinguishable regions.

Several different oncogenes and growth factors promote G1 phase progression. Cyclin D1, the regulatory subunit of several cyclin-dependent kinases, is required for, and capable of shortening, the G1 phase of the cell cycle. The present study demonstrates that transforming mutants of p21ras (Ras Val-12, Ras Leu-61) induce the cyclin D1 promoter in human trophoblasts (JEG-3), mink lung epithelial (Mv1.Lu), and in Chinese hamster ovary fibroblast cell lines. Site-directed mutagenesis of AP-1-like sequences at −954 abolished p21ras-dependent activation of cyclin D1 expression. The AP-1-like sequences were also required for activation of the cyclin D1 promoter by c-Jun. In electrophoretic mobility shift assays using nuclear extracts from cultured cells and primary tissues, several AP-1 proteins (c-Jun, JunB, JunD, and c-Fos) bound the cyclin D1 −954 region. Cyclin D1 promoter activity was stimulated by overexpression of mitogen-activated protein kinase (p41MAPK) or c-Ets-2 through the proximal 22 base pairs. Expression of plasmids encoding either dominant negative MAPK (p41MAPKi) or dominant negatives of ETS activation (Ets-LacZ), antagonized MAPK-dependent induction of cyclin D1 promoter activity. Epidermal growth factor induction of cyclin D1 transcription, through the proximal promoter region, was antagonized by either p41MAPKi or Ets-LacZ, suggesting that ETS functions downstream of epidermal growth factor and MAPK in the context of the cyclin D1 promoter. The activation of cyclin D1 transcription by p21rasprovides evidence for cross-talk between the p21ras and cell cycle regulatory pathways.

The cyclin-dependent kinases are a family of serine/threonine kinases that play a pivotal role in controlling progression through the cell cycle (1)(2)(3)(4). The regulatory subunits of the cyclin-dependent kinases, known as cyclins, form complexes with their catalytic partners to function as kinases of specific proteins at different phases of the cell cycle (3,4). Cyclin D1 is induced in late G 1 following the treatment of growth-arrested macrophage cell lines with colony-stimulating factor 1 (5). Cyclin D1 is required for progression of the G 1 phase (6) and is, therefore, a critical target for proliferative signals in G 1 . Cyclin D1 is capable of shortening the G 1 phase of the cell cycle, suggesting that cyclin D1 may be rate limiting in G 1 progression (7)(8)(9). The induction of cyclin D1 mRNA in response to cell cycle progression and the addition of growth factor is very rapid (6, 10 -12) and is likely regulated at the level of transcription (5,(13)(14)(15).
The ras gene, which is highly conserved in evolution, plays an essential role in cellular proliferation (16), cellular development (17), and cellular differentiation (18 -21). The normal ras gene, activated by point mutations at amino acids 12, 13, or 61, is also capable of inducing cellular proliferation, development, and differentiation (18 -21). Dominant negative mutants of p21 ras block cellular proliferation of NIH3T3 cells (22,23) and the induction of DNA synthesis and gene expression induced by serum (22). p21 ras acts at several distinct phases of the cell cycle including early G 1 , the G 1 /S boundary (24), and at G 2 /M (25). In evidence for a role of p21 ras in early G 1 , ras activity is required soon after the release of cells from quiescence (24). After the addition of serum, the proportion of Ras-GTP increases within 5 min, and the induction of immediate-early gene expression can be at least partially inhibited by previous injection of anti-Ras antibody (26). In addition, anti-Ras antibody or the dominant Ras inhibitor proteins (27) efficiently inhibit DNA synthesis within the recipient cell as long as the injection occurs prior to the onset of the S phase. These findings suggest that p21 ras is also required late in the G 1 phase of the cell cycle, presumably at the G 1 /S boundary.
Recent studies have suggested functional interactions between p21 ras and cyclin D1. Cyclin D1 collaborates with p21 ras in primary rat kidney (28) or rat embryo fibroblasts transformation assays (29). In cell lines overexpressing p21 ras , cyclin D1 mRNA levels were induced (30). In yeast, Ras activates transcription of the CLN genes (31), which are analogous regulators of G 1 phase progression (32). p21 ras has the capacity to phosphorylate and/or activate target transcription factors including c-Jun (33)(34)(35). The ability of p21 ras to augment the transactivation by c-Jun (36,37), likely involves a Jun kinase pathway distinct from p42 MAPK 1 (36 -38). c-Jun, in conjunction with several related AP-1 proteins, promotes G 1 phase progression (39,40) and DNA synthesis (41). For example, inhibition of Jun expression with antisense RNA (42) or microinjection of antibodies (40,43) inhibits cell cycle progression induced by the addition of serum to G o -arrested cells. Thus, several lines of evidence suggest that p21 ras activates c-Jun and that members of the c-Jun/AP-1 family are involved in promoting cellular proliferation (44).
In addition to the Jun kinase pathway, a distinct MAPK pathway is involved in the intracellular transmission of growth factor signals (34). Protein kinases including the MAPK or extracellular signal-regulated kinases (ERKs) (45) are induced in response to EGF. Induction of the MAPK pathway is associated with activation of several different transcription factors including the ETS family proteins (17, 46 -48). The synthesis of the relatively ubiquitous c-Ets-2 is induced upon growth factor stimulation, and its phosphorylation is increased in response to mitogenic stimulation (48).
The target DNA sequences of the ETS proteins includes a core motif with extensive variation at both the 5Ј and 3Ј sides of the invariant GGA core (48). In a variety of promoters that lack TATA sequences, ETS binding sites have been localized close to the initiation site (48,49). Several mechanisms are likely to restrict or condition the activity of these sites as effective ETS-responsive elements. The binding of ETS family members to select promoter/enhancer sequences and their full transcriptional activity may require additional nuclear factors including Sp-1 (50). Sequences resembling the core motif (GGA) required for ETS protein binding (48) are located within the proximal cyclin D1 promoter.
We hypothesized that the recently identified pathways linking tyrosine kinase receptor growth factors, p21 ras , and MAPK (18,45) may be involved in transcriptional regulation of the cyclin D1 gene and thereby cell cycle progression. Cyclin D1 promoter fragments linked to the luciferase reporter gene were examined in the presence of activating or dominant negative, p21 ras , MAPK, c-Jun, and c-Ets mutants, to determine their role in cyclin D1 transcription. Two distinct regions of the cyclin D1 promoter were identified as the site of activation by either p21 ras /c-Jun, or EGF, p41 MAPK /c-Ets-2. The identification of distinguishable mitogenic pathways regulating cyclin D1 transcription provides a mechanism for specificity in signal transduction cross-talk between proliferative pathways and a cell cycle regulatory pathway.

MATERIALS AND METHODS
Construction of Reporter Genes-Standard protocols were used in the construction of the reporter plasmids (51). An 1,882-bp PvuII fragment of the human cyclin D1 genomic clone (52) was subcloned into the vector pA 3 LUC to form the reporter Ϫ1745CD1LUC. A series of 5Ј promoter deletions were derived from this vector including an EcoRI/PvuII fragment (Ϫ1093CD1LUC), a BglI/PvuII fragment (Ϫ485CD1LUC), a PstI/ PvuII fragment (Ϫ261CD1LUC), and a NarI/PvuII fragment (Ϫ141CD1LUC). Ϫ964CD1LUC was cloned using polymerase chain reaction of the native promoter plasmid and the oligodeoxyribonucleotide CD1AP-1wt (sense) and the 3Ј primer (below). Ϫ964mtCD1LUC encodes the same sequence as Ϫ964CD1LUC with two nucleotide changes within the CD1 promoter AP-1 site, created using the 5Ј polymerase chain reaction amplimer CD1AP-1 mut and the 3Ј amplimer CD1 3Ј (below). The cyclin D1 minimal promoter region from Ϫ22 to ϩ14 was cloned into pA 3 LUC in either the sense (Ϫ22CD1LUC) or reverse (Ϫ22RevCD1LUC) orientation. The vector pA 3 LUC includes the trimerized SV40 poly(A) termination site, which abolishes transcriptional read-through (53) and does not contain the recently characterized AP-1 responsive vector backbone sequences (54). The reporter pOLUC, which does not contain the trimerized SV40 poly(A) termination site, was described previously (55). The ETS responsive reporter c-fms430 wt LUC was recently described (56). The c-fms gene encodes the colonystimulating factor 1 receptor and is expressed in monocyte/macrophages and in trophoblasts (57). The human glycoprotein ␣-subunit promoter fragments linked to the pA 3 LUC reporter (Ϫ846GPH␣LUC) (58) was described previously. The integrity of these constructs was determined by restriction enzyme analysis and dideoxy DNA sequencing using an Applied Biosystems 373 automated sequenator.
Cell culture, DNA transfection, and luciferase assays were performed as described previously (58). The trophoblast cell line JEG-3, the fibroblast cell line COS, and the mink lung epithelial cell line Mv1.Lu (CCl-64) were maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum and 1% penicillin/streptomycin. Chinese hamster ovary (CHO) cells were maintained in ␣-minimum essential medium with 10% fetal calf serum and 1% penicillin/streptomycin. Cells were transfected by calcium phosphate precipitation, the medium was changed after 6 h, and luciferase activity was determined after a further 24 h.
In cotransfection experiments, comparison was made between the effect of transfecting active expression vector with the effect of an equal amount of the parental empty expression vector. At least three different plasmid preparations of each construct were used. In cotransfection experiments, a dose response was determined in each experiment with 40, 60, 80, 100, and 200 ng of expression vector and the cyclin D1 promoter reporter plasmids (1.6 g). The fold effect was determined for 100 ng of expression vector. Luciferase assays were performed at room temperature using an Autolumat LB 953 (EG&G, Berthold). Luciferase content was measured by calculating the light emitted during the initial 30 s of the reaction, and the values are expressed in arbitrary light units (58). Background activity from cell extracts was typically Ͻ 150 arbitrary light units/30 s. Statistical analyses were performed using the Mann Whitney U-test. Significant differences were established as p Ͻ 0.05. EGF treatment was performed for 6 -24 h at doses from 2 to 20 ng/ml to determine maximal responses. Subsequent experiments were conducted using EGF at 2.5 ng/ml for 24 h.
Oligodeoxyribonucleotides-The AP-1 site of the cyclin D1 promoter, the wild type AP-1 (CD1AP-1wt) site, and a mutant AP-1 (CD1AP-1mt) site were synthesized as complementary oligodeoxyribonucleotide strands for electrophoretic mobility gel shift assays (EMSA). The antisense strands of these oligodeoxyribonucleotides were also used for polymerase chain reaction-directed amplification of the promoter. The sequence of the cyclin D1 promoter AP-1 site oligodeoxyribonucleotides (CD1AP-1wt) was TCC ATT CTG ACT CAT TTT TTT TAA, and (CD1AP-1 mt) was TCC ATT CTG cCg CAT TTT TTT TAA. The sequences of the wild type and mutant collagenase AP-1 oligodeoxyribonucleotides used as a competitor in EMSA (65) were AP-1wt 5Ј-CGC TTG ATG AGT CAG CCG GAA and AP-1mt 5Ј-CGC TTG ATG cGg CAG CCG GAA. The sequence of the 3Ј CD1 amplimer used in polymerase chain reaction-directed amplification of the promoter was 5Ј-TGG GGC TCT TCC TGG GCA. The reporter construct p 3 TPLUX contains trimeric wild type AP-1-responsive reporter elements and was described previously (58). For comparison with the multimeric AP-1 site, oligonucleotides of similar length which do not include AP-1-like sequences (scattered mutant sequences) (5Ј-AGC TTG TTT TCT TGA CTT AAA TTT GAG AAA GGG TCA AGA ACT AGT CA-3Ј) were cloned upstream of the TK promoter to form p 3 (AP-1mt)TKLUC.
EMSA-EMSA using nuclear extracts, in vitro translated proteins, or bacterially expressed c-Jun (Promega, Madison, WI) were performed essentially as described previously (58). The cDNAs were transcribed in vitro and translated using the TNT-coupled reticulocyte lysate system according to the protocol of the supplier (Promega). The programmed lysates (5 l) were incubated in a reaction mix (20 l) consisting of 20 mM HEPES, pH 7.8, 50 mM KCl, 1 mM EDTA, 10% glycerol, 1 mM dithiothreitol, and 50 g/ml poly(dI-dC) at room temperature for 15 min. ␥-32 P-Labeled oligonucleotides (50 fmol, 50,000 cpm) were added to the reaction and incubated at room temperature for a further 15 min. The protein-DNA complexes were analyzed by electrophoresis through a 5% polyacrylamide gel, with 0.5 ϫ Tris borate, EDTA buffer (TBE: 0.045 M Tris borate, 0.001 M EDTA) and 2.5% glycerol. For EMSA, 5-10 g of nuclear extract was used in binding buffer containing 20 mM HEPES, pH 7.4, 40 mM KCl, 1 mM MgCl 2 , 0.1 mM EDTA, 0.1% Nonidet P-40 to which 0.5 ng of ␥-32 P-labeled probe and 2 g of sonicated salmon sperm DNA were added. Supershifts were performed using antibodies referred to as Jun Ab1 (a JUN polyclonal antibody cross-reactive with several members of the JUN family (66)); Jun Ab2 (Jun AB-2; Oncogene Sciences, Uniondale NY), which inhibits DNA binding of c-Jun; Jun Ab3 (c-Jun Ab; Upstate Biotechnology, Inc., Lake Placid, NY); JunB, JunD (40,67), and c-Fos antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA). The reaction products were separated on 4% polyacrylamide gel run in 0.25 ϫ Tris borate, EDTA at 4°C at 180 volts for 2-4 h. The gels were dried and exposed to XAR-5 (Kodak) radiographic film.
Reagents and Flow Cytometric Analyses-Human recombinant basic EGF (Life Technologies, Inc.) was reconstituted and stored as recommended by the manufacturer.
Flow cytometric analysis was carried out in a fluorescence-activated cell sorter (FACStar plus; Beckton Dickonson). DNA synthesis of the synchronized cells was determined by detection of 5-bromodeoxyuridine incorporation into DNA essentially as described (68). The cells were grown in six-well culture dishes, and 5-bromodeoxyuridine was added with serum. All nuclei were counterstained with Hoechst 33258 (Sigma).
Western Blots-Western blotting was performed as described previously (69). The abundance of cyclin D1 and p21 ras protein was quantified using a polyclonal cyclin D1 (6) and H-Ras (259) (Santa Cruz Biotechnology Inc.) antibodies, respectively. Cell homogenates were electrophoresed in a SDS, 10% polyacrylamide gel and transferred electrophoretically to a Hybond enhanced chemiluminescence nitrocellulose blotting membrane (Amersham Corp.) in a transfer buffer containing 48 mM Tris, pH 8.3, 39 mM glycine, 0.037% SDS, and 20% methanol. After electrophoretic transfer, the gel was stained with Coomassie Blue as a control for blotting efficiency. The blotting membrane was incubated for 1 h at 22°C in TBS buffer (20 mM Tris-HCl, pH 7.5, 140 mM NaCl) supplemented with 5% (w/v) dry milk to block nonspecific binding sites. Following a 1-h incubation with primary antibody at either a 1:1,000 (cyclin D1) or 1:100 (H-Ras (259)) dilution, in TBS buffer containing 0.3% (w/v) Tween 20, the membrane was washed with the same buffer. For detection of cyclin D1 the membrane was incubated with anti-rabbit horseradish peroxidase second antibody (Amersham) and for the p21 ras with a horseradish peroxidase-labeled anti-rat immunoglobulin antibody diluted 1:3,000 (AP332, The Binding Site, Birmingham, U. K.), and washed again. The 43-kDa band representing cyclin D1 and the 21-kDa band representing p21 ras were visualized by the enhanced chemiluminescence system. p42 MAPK and p44 MAPK Immune Complex Assays-Assays were performed as described previously (70) on cell extracts from JEG-3 cells treated with EGF (10 ng/ml) for 10 min. Staphylococcal protein A-Sepharose beads were incubated with anti-MAPK antibody (C16) (Santa Cruz Biotechnology) for 1 h at 4°C. The antibody and beads were washed once with RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 50 mM Tris, pH 7.5, 1 g/ml leupeptin, 0.1% mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate) and then incubated with cell lysates for 2 h at 4°C. The immunoprecipitates were washed with RIPA buffer once; twice with LiCl, 0.1 M Tris base, pH 8.0; and once in kinase buffer. The kinase reactions were performed at room temperature for 20 min in 30 l of kinase buffer with 10 Ci of [␥-32 P]ATP (3,000 Ci/mmol; 1 Ci ϭ 37 GBq) and 4 g of myelin basic protein. The samples were analyzed by SDS-polyacrylamide gel electrophoresis upon termination of the reaction with Laemmli buffer and boiling. The phosphorylation of myelin basic protein was quantified by phosphoimaging using a Fuji Bio Imaging Analyzer BAS 2000 and subsequent densitometry.

RESULTS
p21 Ha-ras Activates the Cyclin D1 Promoter-To determine the role of endogenous p21 ras on basal cyclin D1 transcription, the effect of the antisense (Rev3) p21 Ha-ras expression vectors on cyclin D1 transcription was determined. Several different cell lines were examined to avoid potential cell type-specific effects. The antisense expression vectors were induced by the addition of 10 M Zn 2ϩ for 24 -48 h. Treatment with 10 M Zn 2ϩ alone had no effect on cellular morphology, cell number, or cyclin D1 reporter activity (data not shown). Rev3 reduced basal Ϫ1745CD1 reporter activity in JEG-3 to 40% of wild type (Fig. 1A) and reduced endogenous p21 ras levels 50% by Western blot analysis (inset, Fig. 1A), as shown previously (60,71). To examine further the role of p21 ras in the activation of cyclin D1 transcription the effect of the dominant negative p21 ras mutant (Ras N17) on cyclin D1 expression was determined. Ras N17 reduced basal level transcription to 50% of wild type (p Ͻ 0.01) (Fig. 1A). The pA 3 LUC, and the p 3 (AP-1 mt) TKLUC plasmids were not affected by the addition of Rev3 (not shown).
The effect of several transforming p21 ras mutants on cyclin D1 promoter activity was examined in JEG-3 cells. The magnitude of cyclin D1 reporter induction by the activating p21 Ha-ras mutant expression vector was dependent upon the amount of transfected expression vector and increased from 24 to 48 h. The activating p21 ras mutants (Ras Val-12 and Ras Leu-61) and pMTEj stimulated the cyclin D1 promoter in several different cell lines. The induction by the pMTEj Ras vector was dependent upon the addition of Zn 2ϩ (not shown). The constitutively active p21 ras mutants induced cyclin D1 reporter activity 4 -5-fold in JEG-3 cells (Fig. 1A). The expression vector encoding the double mutant Ras Leu-61,Ser-186, which is in- capable of insertion in the plasma membrane, did not affect cyclin D1 promoter activity (Fig. 1A). The expression vector cassettes driving Ras Val-12 and Ras Leu-61 did not affect cyclin D1 promoter activity (not shown).
The constitutively active p21 Ha-ras mutant induced cyclin D1 reporter activity 22-fold in Mv1.Lu cells and 5-fold in CHO cells (Fig. 1B). The induction by the pMTEj vector was dependent upon the addition of Zn 2ϩ (not shown). Rev3 reduced basal Ϫ1745CD1 reporter activity in CHO cells but not in Mv1.Lu cells (Fig. 1B).
To determine the region of the cyclin D1 promoter required for regulation by p21 ras , the cyclin D1 5Ј promoter deletions ( Fig. 2A) were transfected in the presence of either the sense or antisense p21 ras expression vectors (Fig. 2B). Mutation of sequences resembling an AP-1 site at Ϫ954 abolished transcriptional activation by p21 ras (pMTEj) (Fig. 2B) and the Ϫ964mt-DCLUC reporter was repressed to less than 40% by the activating p21 ras mutant.
c-Jun Activates the Cyclin D1 Promoter-Overexpression of p21 ras is associated with the transcriptional induction of a variety of early-immediate genes including c-fos and c-jun, and p21 ras augments the transcriptional activity of endogenous c-Jun (36,72). To determine whether c-Jun was capable of regulating cyclin D1 reporter activity, cotransfection experiments were conducted with either wild type c-Jun or an expression vector encoding a DNA-binding defective mutant of c-Jun (c-Jun DNA Ϫ ) (58). To avoid potential cell type-specific effects, several different cell lines were examined. c-Jun activated the Ϫ1745CD1 promoter in JEG-3 (range 3-9-fold, mean 3.8-fold), Mv1.Lu cells (12-fold) (Fig. 3A), and CHO (3.5-fold) (not shown). The c-Jun mutant incapable of binding DNA (Jun DNA Ϫ ) did not activate the Ϫ1745CD1 reporter (not shown). The reporter p 3 TP-LUX, which contains a trimeric AP-1 site, was activated 4.5-fold by c-Jun in JEG-3 cells (Fig. 3A). The reporter TKLUC and the promoterless vector pA 3 LUC were not activated by c-Jun (Fig. 3A). The human ␣-glycoprotein promoter (Ϫ846GPH␣LUC) was negatively regulated by c-Jun as reported previously (58). These findings indicate that c-Jun activation of the cyclin D1 promoter requires the c-Jun DNA binding domain, and this activation occurs in a promoter-specific manner. (Several different reporter vectors were, however, activated by c-Jun in the absence of cloned promoter or enhancer sequences, perhaps due to vector backbone sequences or spurious transcriptional read-through as previously reported (54,73). The reporter vectors pBLCAT and pOLUC, for example, were activated in the presence of transfected c-Jun in JEG-3 cells (not shown)).
To determine the minimal region of the cyclin D1 promoter activated by c-Jun, the series of 5Ј promoter deletion constructs were transfected into JEG-3 cells with either wild type or mutant expression vectors. The region of the cyclin D1 promoter activated by c-Jun was localized to the region between Ϫ964 and Ϫ485. Site-directed mutagenesis of the AP-1 site at Ϫ954 abolished activation by c-Jun in JEG-3 cells (Fig. 3B). The domains of c-Jun required for activation of cyclin D1 transcription were determined using a series of expression vectors encoding wild type and mutant c-Jun proteins (58,66,74). Previous studies demonstrated that the DNA binding activity of these mutants, other than DNA Ϫ , were similar to wild type (66). Western blot analyses (74) demonstrated that all the mutant proteins used in these studies were expressed similarly in transfected cells. Deletion of the amino-terminal 91 amino acids, deletion of the A2 activation domain, or mutation of the leucine zipper reduced activation of the cyclin D1 promoter by c-Jun to less than 20% of wild type (Fig. 3C).
AP-1 Proteins Bind the Cyclin D1 Ϫ954 Region in Primary and Transformed Cells-To determine whether AP-1 proteins bound the cyclin D1 AP-1 site, EMSA were performed using either c-Fos and c-Jun proteins or nuclear extracts (Fig. 4,  A-C). The cyclin D1 AP-1 site and the collagenase AP-1 site bound a complex formed by in vitro translated c-Fos, in the presence of bacterially expressed c-Jun (Fig. 4, lane 1, panels A  and B). The electrophoretic mobility of the complex binding both probes was similar (Fig. 4, A and B). The cyclin D1 AP-1 probe bound a complex (indicated by the arrow) in Mv1.Lu or JEG-3 cell nuclear extracts (Fig. 4, A-C). This complex was competed by 100-fold excess cold cognate competitor oligodeoxyribonucleotide probe but not by mutant AP-1 sequences (Fig. 4, A-C). The nuclear complex binding either the cyclin D1 AP-1 site or the collagenase AP-1 site was competed by 100-fold excess cold cognate competitor but not by an equimolar excess of mutant competitor sequences (Fig. 4, A-C). The AP-1 proteins did not bind the mutant cyclin D1 AP-1 sequences (not shown).
To determine whether the cyclin D1 AP-1 site was capable of binding AP-1 proteins in nontransformed cells, nuclear extracts were prepared from rat tissues and the abundance of AP-1 proteins binding these sequences examined. The c-Jun antibody (Jun Ab2) inhibited binding of most of the complex binding the cyclin D1 AP-1 site, and the JunB and JunD antibodies (40, 67) induced a minor supershift of the complex binding either the cyclin D1 AP-1 site (Fig. 4D) or the wild type AP-1 site (Fig. 4E). The c-Fos antibody shifted most of the complex binding either AP-1 site (indicated by an asterisk in Fig. 4, D and E). These studies demonstrate that the cyclin D1 AP-1 site is capable of binding AP-1 proteins from cultured cells and primary tissue.
c-Fos Represses the Cyclin D1 Promoter-As c-fos and related AP-1 proteins formed part of the complex binding the cyclin D1 AP-1 site, the effect of several related AP-1 proteins on cyclin D1 transcription was determined. JunD, c-Fos, and Fra-1 had substantially different effects on Ϫ1745CD1LUC reporter activity when compared with the activation by c-Jun (Fig. 5, A  and B). c-Fos repressed Ϫ1745CD1LUC reporter activity 50 -60% in JEG-3 cells (Fig. 5A). JunD and Fra-1 induced Ϫ1745CD1LUC reporter activity only 1.5-2-fold (Fig. 5A). Similar effects of the AP-1 proteins on cyclin D1 promoter activity were observed in COS cells (Fig. 5B).
p41 MAPK Activates Cyclin D1 Transcription through the Proximal Promoter Region-Complex and distinct pathways mediate the sequential intracellular signaling cascades initiated by p21 ras and do so in a cell type-specific manner (34,75). For example, the kinase activity modulating c-Jun transcriptional activation within the cell (Jun kinases) (38) are distinguishable from, and may antagonize, the activity of the related MAPK pathway in a promoter-dependent manner (37). To examine the role of the MAPK pathway in regulating cyclin D1 transcription, the effect of overexpressing p41 MAPK on cyclin D1 reporter activity was assessed in JEG-3 and Mv1.Lu cells (Fig. 6A). The wild type, but not the mutant p41 MAPK expression vector, ac-tivated cyclin D1 transcription 3-5-fold in JEG-3 cells and 8-fold in Mv1.Lu cells (Fig. 6A). The basal level transcription of cyclin D1 was reduced 40% by the dominant negative MAPK expression vector (MAPKi) in JEG-3 cells (Fig. 6A). The minimal cyclin D1 promoter region activated by p41 MAPK and inhibited by MAPKi was located within Ϫ22 bp of the transcriptional start site (Fig. 6B). The proximal Ϫ22 bp in the reversed orientation was not activated by p41 MAPK (Fig. 6B). The vectors RSVLUC, p 3 (AP-1mt)TKLUC, TKLUC, and pA 3 LUC were not induced by MAPK in JEG-3 cells (Fig. 6B). As several different reporters were not activated by p41 MAPK these results suggest activation of Ϫ1745CD1LUC reporter activity was promoterspecific and did not reflect a general effect on transcription as demonstrated previously (62). The activation by MAPK overexpression may possibly be due to autophosphorylation or endogenous MAPK kinase activity in these cells, as described previously in RK13 cells (47).
Several growth factors, including EGF, stimulate MAPK ac-tivity in a cell type-specific manner. Previous studies have demonstrated that EGF induced cyclin D1 mRNA levels in fibroblast cell lines (10). Trophoblast cell lines, such as JEG-3 cells, express EGF receptors, and EGF promotes trophoblast outgrowth and blastocoel expansion (57). The effect of EGF on MAPK activity was therefore determined in JEG-3 cells. EGF (10 ng/ml) was used to treat JEG-3 cells, and cells were harvested at time points from 30 min to 24 h. MAPK activity was induced 6-fold within 30 min (Fig. 6C). Using fluorescenceactivated cell sorter analysis EGF stimulated a 12-15% increase in the proportion of JEG-3 cells moving from G 0 /G 1 to S phase (not shown). EGF Activates Transcription of the Cyclin D1 Promoter-As EGF stimulated MAPK activity and MAPK activity stimulated cyclin D1 LUC reporter activity, the effect of EGF on cyclin D1 transcription and cyclin D1 protein levels was determined in JEG-3 cells. The effect of EGF on cyclin D1 transcription was determined in JEG-3 cells in transient expression studies using the Ϫ1745CD1LUC reporter (Fig. 7A). Randomly cycling or G 0 -arrested transfected cells were treated with EGF for 24 h. Maximal induction of cyclin D1 reporter activity by EGF was observed at 10 ng/ml (Fig. 7A). Cyclin D1 reporter activity was induced 4.5-fold by EGF (2.5 ng/ml) in JEG-3 cells (Fig. 7B), and cyclin D1 protein levels were induced 4-fold by EGF as determined by Western blot analysis (data not shown). Several other native or synthetic reporters, in the pA 3 LUC reporter, were not induced by EGF (Fig. 7B).
The DNA sequence requirements for EGF-induced transcription of the cyclin D1 promoter were determined using a series of 5Ј promoter fragments. Cells were treated for 24 h with EGF (2.5 ng/ml), and the effect was compared with untreated cells. The induction by EGF (3-4-fold) was conveyed by the minimal Ϫ22-bp promoter fragment. The Ϫ22 bp in the inverted orientation was not activated by EGF (Fig. 7C).
c-Ets-2 Activation of Cyclin D1 Requires the Proximal Promoter Region-The c-Ets proteins are targets of MAPK phosphorylation (48,76), and the synthesis of c-Ets-1 and c-Ets-2 is induced upon growth factor stimulation (48,76). Phosphorylation of the ETS proteins is also increased in response to mitogenic stimulation. Together these observations are consistent with a role for ETS proteins in growth factor-mediated signal transduction. To examine the role of c-Ets-2 in cyclin D1 transcription, either a c-Ets-2 or a dominant negative ets-LacZ expression vector was transfected into JEG-3 cells in conjunction with the cyclin D1 reporter. c-Ets-2 activated Ϫ1745CD1LUC reporter activity 9-fold, and ets-LacZ repressed transcription to 60% of wild type in JEG-3 cells (Fig.  8A). The magnitude of induction of the cyclin D1 promoter was similar to the activation by c-Ets-2 of the c-Ets-responsive promoter for c-fms. c-Ets-2 did not affect the activity of several other plasmids including TKLUC, p 3 (AP-1mt)TKLUC, or pA 3 LUC, demonstrating the activation of the cyclin D1 promoter by c-Ets-2 was promoter-dependent (Fig. 8A).
The series of cyclin D1 5Ј promoter deletions was transfected into JEG-3 cells with c-Ets-2 to determine the minimal ETSresponsive region of the cyclin D1 promoter. The 5Ј promoter deletion constructs were activated 7-10-fold by c-Ets-2. The minimal c-Ets-2-responsive region was located within 22 bp of the transcriptional start site in JEG-3 cells (Fig. 7B). To determine whether regulation of the cyclin D1 promoter by c-Ets-2 was observed in other cell types, transient expression studies were performed in Mv1.Lu and COS cells. c-Ets-2 activated Ϫ1745CD1LUC 18-fold in Mv1.Lu cells and 4-fold in COS cells (Fig. 8, C and D).
As the regions activated by EGF, MAPK, and c-Ets-2 colocalized within the proximal promoter region, studies were performed with activating or dominant negative expression vectors to determine whether sequential or parallel pathways linked EGF, p41 MAPK , and c-Ets-2 in the context of the cyclin D1 promoter. Overexpression of p41 MAPK induced the Ϫ141 cyclin D1 promoter fragment 3-6-fold (Fig. 9A). EGF activated Ϫ141CD1LUC 3-5-fold. MAPKi or ets-LacZ reduced the induction of the cyclin D1 promoter by EGF 60 -80%, suggesting MAPK and ETS functions downstream of EGF in the context of the cyclin D1 promoter (Fig. 9, A and B). The induction of Ϫ141CDLUC by p41 MAPK was reduced or abolished by ets-LacZ in JEG-3 (Fig. 9C). Activation by c-Ets-2 was increased only 25% by the overexpression of MAPK, suggesting sequential rather than parallel pathways link MAPK and ETS signaling in JEG-3 cells (Fig. 9C). DISCUSSION In these experiments, activating p21 ras mutants stimulated cyclin D1 promoter activity in a DNA sequence-dependent manner. The activation of cyclin D1 transcription by p21 ras is consistent with recent observations in which the abundance of cyclin D1 mRNA was increased in v-Ha-Ras-transformed pre-B cell lines (30). p21 ras promotes cellular proliferation and early G 1 phase progression (16,23,24,77). As cyclin D1 may be rate-limiting in promoting G 1 phase progression, the induction of cyclin D1 transcription by p21 ras may provide, in part, a mechanism by which p21 ras promotes G 1 phase progression. The observation that distinguishable regions of the cyclin D1 promoter were targeted by either p21 ras and c-Jun through one region and by MAPK and ETS through another are consistent with recent observations suggesting that a specific Jun kinase, distinguishable from MAPK, modulates activity of c-Jun (37).
Several lines of evidence indicate a role for c-Jun and the AP-1 proteins in promoting G 1 phase progression (40,67,78). In these studies, c-Jun activated the cyclin D1 promoter in all cell lines examined. Mutation of the cyclin D1 AP-1 site abolished the induction by c-Jun in JEG-3 cells, consistent with a role for this site in mediating transactivation by c-Jun. The DNA binding domain of c-Jun was required for induction of the cyclin D1 promoter. In vitro translated c-Fos with c-Jun bound the cyclin D1 AP-1 site in EMSA. The cyclin D1 AP-1 site bound nuclear complexes from JEG-3 and Mv1.Lu cells, and the mobility of these complexes was similar to the complexes binding the wild type AP-1 site in EMSA. Together these findings suggest that the cyclin D1 Ϫ954 region is capable of binding AP-1 proteins and conveys c-Jun-induced transcription in the cell line examined. The signal transduction pathway modulating c-Jun-induced transcription of cyclin D1 is currently unknown; however, the recently identified Jun kinase (38), which phosphorylates the amino terminus of c-Jun, is a likely candidate. Recent studies suggested a role for the Ϫ58 sequences ("cAMP response element" TGAGGTAA) in c-Jun-mediated transactivation of cyclin D1 in fibroblast cells using the pOLUC vector (79). A residual 5-fold induction by c-Jun was observed by Herber et al. (79). after mutation of the cAMP response element site, and no deletion or mutation abolished transactivation by c-Jun. Cell type-specific differences may be responsible for the different localization of the site of c-Jun action.
AP-1/ATF Proteins Differentially Regulate the Cyclin D1 Promoter-c-Fos negatively regulated cyclin D1 transcription. In keeping with our observation that overexpression of c-Fos reduced cyclin D1 transcription, recent studies demonstrated that conditional overexpression of c-Fos reduced cyclin D1 mRNA levels (80). It is likely that additional members of the AP-1 family are involved in promoting G 1 phase progression as microinjection of antibodies against c-Fos alone had a relatively modest effect on cell cycle progression, whereas antibodies crossreactive with all Fos family members were more efficient at inhibiting the G 0 to G 1 transition (40,67). c-Jun occupies a pivotal role in intracellular signal transduction at least in part through its capacity to interact with a variety of other transcription factors (81). Together these studies suggest that the ability of AP-1 proteins to modulate cell cycle progression is complex (40,67,80) and that the transcriptional response of the cyclin D1 promoter to AP-1 proteins may depend on the nature of the heterodimeric partner of c-Jun bound to the promoter AP-1 site.
MAPK and c-Ets-2 Activate the Cyclin D1 Promoter through the Minimal Promoter-Activation of MAPKs induces reentry into the cell cycle, implicating MAPKs as targets in the transmission of proliferative signals (82). Overexpression of MAPK induced the cyclin D1 promoter, and this induction was antagonized by either MAPKi or ets-LacZ, suggesting that ETS proteins may play a role in MAPK activation of the cyclin D1 promoter. c-Ets-2 did not affect the same plasmid backbone, demonstrating specificity in the activation of cyclin D1 promoter activity by c-Ets-2. The activation of the minimal cyclin D1 promoter by either c-Ets-2 or MAPK is consistent with recent observations on the role of ETS proteins in modulating cell cycle progression (64). The ETS proteins promote the G 1 phase, and the enforced expression of c-Ets-1 or c-Ets-2 rescued the G 1 block in NIH3T3 cells that express a mutant form of the colony-stimulating factor 1 receptor (64). c-Ets-2 is both transcriptionally activated and phosphorylated by MAPK (76), consistent with a role for c-Ets-2 in conveying MAPK-induced transcription of the cyclin D1 promoter. The candidate elements sufficient for activation by c-Ets-2 are located within the proximal 22 bp of the cyclin D1 promoter. Sequences resembling the ETS motif (83) are located within this minimal promoter region. ETS motifs are frequently found close to the initiation site (Inr) in TATA-less promoters either 5Ј to the Inr (84) or at the Inr (49). Thymidylate synthetase and DNA polymerase ␣, which are encoded by late serum-response genes and are required for DNA synthesis, contain ETS motifs at the Inr, and the ETS motif of thymidylate synthetase is required for promoter activity (85).
It remains to be determined whether the effect of c-Ets-2 to activate the cyclin D1 promoter is mediated directly through binding target sequences within the minimal promoter or is mediated indirectly through interaction with the basal transcription apparatus. The ETS protein, PU.1, binds the basal level transcription factor TFIID (86), which in turn binds either TATAA or Inr sequences (87). Recent studies demonstrated that the cyclin D1 Inr is sufficient for negative regulation by c-Myc (88), and that the Inr may also function to convey regulation by several other factors including ETS.
EGF Activation of the Cyclin D1 Promoter Required the Proximal 22 bp-EGF induced cyclin D1 promoter activity and protein levels in JEG-3 cells. EGF did not activate a variety of other native or synthetic promoters in the same reporter system, indicating that the activation of cyclin D1 reporter activity by EGF was promoter-specific. The induction of cyclin D1 promoter activity by EGF was reduced 60 -80% by either MAPKi or Ets-LacZ expression plasmids, implicating MAPK as an intermediary kinase in EGF signaling in JEG-3 cells. EGF stimulated MAPK activity and G 1 phase progression in JEG-3 cells. EGF receptor signaling involves tyrosine kinase-dependent and -independent pathways (19,89,90). Cyclin D1 is capable of promoting G 1 phase progression in stable cell lines (7)(8)(9), and, when assessed in stable cell lines, DNA sequences sufficient for activation during G 1 phase progression are present within the 1,745-bp cyclin D1 promoter. 2 Further studies using 2 C. Albanese and R. G. Pestell, unpublished data.
FIG. 9. EGF activation of the proximal cyclin D1 promoter is antagonized by dominant negative MAPK and ETS expression vectors. Panels A-D, the Ϫ141 cyclin D1 reporter was transfected into JEG-3 cells with expression vectors encoding proteins for MAPK (p42 MAPK or p42 MAPKi ), p21 ras (activating (pMTEj) or p21 ras antisense (Rev3)), or ETS (the activating (c-Ets-2) or dominant negative (ets-LacZ)). Expression vector (300 ng) was transfected with 4.8 g of Ϫ141CD1LUC. In panel A the effect of MAPKi and ets-LacZ on EGFinduced Ϫ141CDLUC activity is shown as a percent of wild type. The data from a representative experiment from at least three separate transfections are shown. In panel C the induction of the Ϫ141CDLUC reporter by MAPK overexpression (3-4-fold) was normalized to 1 for the purpose of showing in the same figure the additional effect of c-Ets-2. In panel D the induction by EGF was normalized to 1 for comparison with the additional effects of c-Ets-2.
antisense cyclin D1 paradigms are needed to determine whether cyclin D1 is required for the promotion of G 1 phase progression by EGF observed in JEG-3 cells.
These studies demonstrate that EGF, activating p21 ras mutants, p42 MAPK , c-Jun, and c-Ets-2 augment cyclin D1 promoter activity. These findings suggest that the cyclin D1 promoter may be an important target for several distinct signal transduction pathways involved in conveying proliferative signals during G 1 . c-Jun and c-Ets-2 target distinct regions of the promoter providing a mechanism for collaborative interactions between these two proto-oncogenes in stimulating abundance of this cell cycle regulatory kinase subunit. These studies demonstrate an avenue for signal transduction cross-talk between the MAPK and cell cycle regulatory pathways and a mechanism by which different mitogenic and transforming factors may interact to promote cellular proliferation.