Cyclin A Is a c-Jun Target Gene and Is Necessary for c-Jun-induced Anchorage-independent Growth in RAT1a Cells*

Overexpression of c-Jun enables Rat1a cells to grow in an anchorage-independent manner. We used an inducible c-Jun system under the regulation of doxycycline in Rat1a cells to identify potential c-Jun target genes necessary for c-Jun-induced anchorage-independent growth. Induction of c-Jun results in sustained expression of cyclin A in the nonadherent state with only minimal expression in the absence of c-Jun. The promoter activity of cyclin A2 was 4-fold higher in Rat1a cells in which c-Jun expression was induced compared with the control cells. Chromatin immunoprecipitation demonstrated that c-Jun bound directly to the cyclin A2 promoter. Mutation analysis of the cyclin A2 promoter mapped the c-Jun regulatory site to an ATF site at position –80. c-Jun was able to bind to this site both in vitro and in vivo, and mutation of this site completely abolished promoter activity. Cyclin A1 was also elevated in c-Jun-overexpressing Rat1a cells; however, c-Jun did not regulate this gene directly, since it did not bind directly to the cyclin A1 promoter. Suppression of cyclin A expression via the introduction of a cyclin A antisense sequences significantly reduced the ability of c-Jun-overexpressing Rat1a cells to grow in an anchorage-independent fashion. Taken together, these results suggest that cyclin A is a target of c-Jun and is necessary but not sufficient for c-Jun-induced anchorage-independent growth. In addition, we demonstrated that the cytoplasmic oncogenes Ras and Src transcriptionally activated the cyclin A2 promoter via the ATF site at position –80. Using a dominant negative c-Jun mutant, TAM67, we showed that this transcriptional activation of cyclin A2 requires c-Jun. Thus, our results suggest that c-Jun is a mediator of the aberrant cyclin A2 expression associated with Ras/Src-induced transformation.

c-Jun is a transcription factor that binds to AP-1 (TGAG/ CTCA) sites in the promoters of a large number of genes. It binds to these sites as either a homodimer or a heterodimer together with Fos family members (c-Fos, FosB, Fra-1, and Fra-2) or ATF family members (ATF-1 to -4, CREB1, or CREM1). c-Jun is a versatile transcription factor and contributes to transformation and tumor aggressiveness (1,2) as well as cell cycle progression (3), differentiation (4), and apoptosis (5) in a tissue-and cell-specific manner. It acts in conjunction with Ras to transform primary cells (6) and is essential for transformation of cells by Ras (7). In addition to being necessary for Ras transformation of cells, c-Jun is also required for transformation induced by c-Fos, Raf, c-Myc, Mos, and Abl (8).
We have demonstrated previously that c-Jun induces Rat1a cells to grow in an anchorage-independent manner in soft agar (9). Although a number of c-Jun target genes have recently been identified (10 -14) the "downstream genes" that mediate this pathway are yet to be determined.
In transformed cells, cell cycle-dependent proteins are often aberrantly expressed, resulting in abnormal cycling through the various phases of the cell cycle. The cyclins (A, B, D, and E) and their associated kinases (Cdk2, -4, and -6) provide a stimulus for progression through the cell cycle, whereas the proteins p16 Ink4a , p19 ARF , p21, p27, Rb, and p53 all function to inhibit cell cycle progression. Cyclins D1, E, and A can function as oncogenes when aberrantly expressed, and p19 ARF , p53, p21 cip1 , p27 kip1 , and Rb are suppressor genes whose reduction, deletion, or inactivation causes transformation (15)(16)(17)(18).
Two distinct genes encode cyclin A in mammalian cells, Ccna1 and Ccna2 (19 -21). Ccna2, which encodes the originally described cyclin A2, is ubiquitously expressed in cultured cells and in various tissues (21,22), whereas the expression of cyclin A1 seems to be testis-specific and restricted to the germ line (21), although it has been detected in leukemic cell lines (19). Cyclin A2 functions during both G 1 -S and G 2 -M phases of the cell cycle (23,24) and binds to and activates both Cdk1 and Cdk2 (25)(26)(27). Much less is known about cyclin A1; however, it has also been reported to be involved in signaling pathways important for cell cycle regulation (28). A number of oncogenes have been shown to regulate the expression of cyclin A. Its expression is induced by c-Myc and is crucial for c-Myc-mediated transformation of Rat1 cells (29 -31). Ras induces cyclin A expression in NIH3T3 cells, promotes cyclin E-Cdk2 kinase activity and hyperphosphorylation and inactivation of Rb, all of which may contribute to Ras-induced transformation (32). Raf induces cyclin A expression as well as that of cyclin D, cyclin E, and p21 cip1 with a concomitant decrease in p27 kip1 expression (33). c-Fos induces cyclin A expression, resulting in cellular proliferation (34). Early studies revealed that c-Jun expression is necessary for G 0 /G 1 transition or G 1 progression (35)(36)(37). More recently, c-Jun was found to alter the expression of cell cycle-related genes such as cyclin D1 (11), p53 (10), and p21 cip1 (38). However, there have been few reports on whether the alteration of expression of these cell cycle related proteins by c-Jun plays a role in cell transformation. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We have previously described a cellular model system where c-Jun is inducibly expressed under the control of a doxycyclineresponsive promoter, which induces nonadherent cell growth (39,40). Using this c-Jun-inducible system, we examined the effect of c-Jun on cell cycle regulators to investigate the hypothesis that direct regulation of cell cycle genes by c-Jun is necessary for c-Jun induction of nonadherent cell growth of Rat1a cells. We demonstrate that of all cell cycle proteins, only cyclin A demonstrated dysregulation under c-Jun expression. Further, c-Jun binds to the ATF site in the cyclin A2 promoter and enhanced cyclin A expression under nonadherent growth conditions. c-Jun also enhanced cyclin A1 expression under these conditions, although it did not interact directly with the cyclin A1 promoter. Inhibition of cyclin A expression resulted in a decrease in the ability of the cells to form colonies in soft agar; however, overexpression of cyclin A alone did not enable the cells to grow in an anchorage-independent manner. Taken together, these results demonstrate that cyclin A2 is a direct target of c-Jun and is necessary, but not sufficient, for c-Juninduced anchorage-independent growth.
RNA Analysis-RNA from Rat1a-J4 and GFP cells was analyzed by Northern blotting as previously described (14,40). To determine the cyclin A mRNA half-life, Rat1a-J4 cells grown nonadherently for 3 days in the presence or absence of 2 g/ml doxycycline were treated with the transcription inhibitor actinomycin D (10 g/ml). Total RNA was extracted from the cells 0, 2, 4, and 8 h after the addition of actinomycin D. The RNA was subjected to Northern blot analysis for cyclin A as described above. The intensity of the signal was quantified densitometrically.
Gel Mobility Shift and Supershift Assays-Nuclear extracts were prepared from Rat1a-J4, cells grown for 3 days under nonadherent conditions in the presence or absence of 2 g/ml doxycycline as described previously (14,43). The DNA fragments used as probes for cyclin A2 were double-stranded oligonucleotides containing the wildtype ATF site (5Ј-TGAATGACGTCAAGGCCCGCGAG-3Ј) and a mutated ATF site (5Ј-TGAATGCCCCCAAGGCCCGCGAG-3Ј), and that for cyclin A1 was a fragment spanning positions Ϫ190 to Ϫ129 of the cyclin A1 promoter generated by PCR amplification with primers 1.1F (5Ј-CCCCTCGAGAAGCCCGGCCGCCTCCCAGGC-3Ј) and 1.2R (5Ј-GCAGGGAAGGGCAGGGCACG-3Ј). All fragments were 5Ј-end-labeled with T4 polynucleotide kinase, and gel mobility shift and supershift assays were carried out as previously described (14).
Establishment of Cyclin A-overexpressing Rat1a Cells-The retroviral human cyclin A expression vector BabeHygro2-HCA2 (a kind gift from B. Amati) was used to establish Rat1a cells stably expressing cyclin A. Briefly, 10 g of the retroviral expression vector was transfected into PhoenixA cells, and the supernatant was harvested and used to infect Rat1a fibroblasts. 48 h after infection, the cells were serially diluted, and clones were selected with 200 g/ml hygromycin B (Calbiochem). Rat1a cells able to express cyclin A in an inducible fashion were also established as follows. An EcoRI fragment from pCEM4 containing the human cyclin A cDNA was cloned into pBlueScript II SK(Ϯ). An XhoI-NotI fragment from pBlueScript-cyclin A was cloned into the pLRT retroviral vector to construct pLRT-cyclin A. pLRT-cyclin A was transfected into Rat1a cells using the standard calcium phosphate method, and clones were selected using blasticidin (8 g/ml).
Colony Forming Assay-1.0 ϫ 10 4 cells were plated in triplicate in 6 ml of 0.8% agarose (sea plaque) in complete growth medium in the presence or absence of 2 g/ml doxycycline overlaid on a 0.4% agarose base, also in complete growth medium. 3-4 weeks after incubation, colonies more than 50 m in diameter were counted using an Omnicon 3600 image analysis system. The colonies were visualized after staining for 16 -24 h with 1.0 mg/ml p-iodonitrotetrazolium violet.
Kinase Assays-Cyclin A-associated kinase activity in the c-Jun expressing Rat1a cells was measured using the Cdk1/Cdc2 kinase assay kit (Upstate Biotechnology, Inc., Charlottesville, VA) as described by the manufacturer.
Chromatin Immunoprecipitation-Rat1a-HA-Jun cells expressing HA-tagged c-Jun were seeded at a density of 10 ϫ 10 6 cells in 150-mm PolyHeme-coated tissue culture dishes in the presence and absence of 2 g/ml doxycycline and incubated for 4 days at 37°C. Protein-DNA complexes were cross-linked with 1% formaldehyde added directly to the culture medium at room temperature for 15 min followed by the addition of 0.125 M glycine, pH 2.5, for 5 min. Cells were pelleted at 1500 rpm for 5 min at 4°C and washed once with ice-cold PBS. The cell pellet was resuspended in 300 l of lysis buffer (1% SDS, 5 mM EDTA, 50 mM Tris, pH 8, 100 g/ml aprotinin, 100 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride) and incubated on ice for 10 min. The solution was then sonicated 3 times for 15 s each on maximum power, and cell debris was pelleted by centrifugation for 5 min at 13,000 rpm. Twenty microliters of the soluble chromatin was set aside as the input fraction, and the remainder was diluted 1:10 in dilution buffer (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris, pH 8, 100 g/ml aprotinin, 100 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). The diluted soluble chromatin (1 ml) was precleared with 2 g of sheared herring sperm DNA, 20 l of preimmune serum, and 45 l of protein G-agarose beads (50% slurry in 10 mM Tris, pH 8, 1 mM EDTA) for 2 h at 4°C, and the beads were pelleted by centrifugation at 13,000 rpm for 1 min. Two micrograms of antibodies to hemagglutinin (71-5500; Zymed Laboratories Inc., San Francisco, CA) or control rabbit IgG (sc-2027; Santa Cruz Biotechnology, Santa Cruz, CA) were added, and the solution was incubated overnight at 4°C. Following this incubation, 45 l of protein G-agarose beads and 2 g of sheared herring sperm DNA were added and incubated for an additional 1 h at 4°C. The beads were pelleted by centrifugation and washed sequentially for 10 min each with TSE I (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris, pH 8, 150 mM NaCl), TSE II (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris, pH 8, 500 mM NaCl), buffer III (0.25 M LiCl, 1% Nonidet P-40, 1% deoxycholate, 1 mM EDTA, 10 mM Tris, pH 8), and TE (10 mM Tris, pH 8, 1 mM EDTA). DNA was eluted from the beads with 100 l of elution buffer (1% SDS, 0.1 M NaHCO 3 ), and the cross-links were reversed by incubation at 65°C overnight. DNA was purified using the Qiaquick PCR Purification kit (Qiagen, Valencia, CA) as per the manufacturer's instructions.
One microliter of the chromatin immunoprecipitation DNA was amplified by real time PCR with primers cycA-F (5Ј-CCTCAGGCTCCCGC-CCTGTAAGATTCC-3Ј) and cycA-R (5Ј-TCAAGTAGCCCGCGACTATT-GAATAT-3Ј) on an iCycler real time detection system (Bio-Rad) with the Quantitect SYBR Green PCR Kit (Qiagen Inc., Valencia, CA) as per the manufacturer's instructions. The -fold change was calculated using the 2 Ϫ⌬⌬CT method as previously described (44).

c-Jun Maintains the Expression of Cyclin A in Nonadherent
Rat1a Cells-Overexpression of c-Jun in Rat1a cells results in nonadherent growth (14,39,40). We hypothesized that this c-Jun-mediated nonadherent growth might be the result of direct regulation by c-Jun of cell cycle genes. Western blot analysis was performed in order to determine the levels of various cell cycle-related proteins. The level of cyclin A in Rat1a-J4 cells grown for 3 days under nonadherent conditions in the presence of doxycycline was increased compared with cells grown in the absence of doxycycline or Rat1a-GFP cells (Fig. 1A). The levels of cyclins D1, E, and B1, on the other hand, remained unchanged between the cells overexpressing c-Jun and control cells (Fig. 1A). Expression of p16 Ink4a , p19 ARF , and p57 Kip2 remained the same between the uninduced and induced conditions. Similar results were obtained in another

. Expression of various cell cycle regulators after c-Jun induction in nonadherent Rat1a-J4 cells. A, exponentially growing
Rat1a-J4 cells were plated at a density of 1.5 ϫ 10 6 /150-mm PolyHEME-coated dish in the absence (Ϫ) or presence (ϩ) of doxycycline (2 g/ml) and incubated for 3 days. The cells were lysed, and the protein extracts were subjected to Western blotting using antibodies against cyclin D1, cyclin E, cyclin A, cyclin B1, p16 Ink4a , p19 ARF , p27 kip1 , p57 kip2 , c-Jun, and ␤-tubulin. Rat1a-GFP cells expressing GFP were used as negative control. B, whole cell extracts were prepared from Rat1a-J4 cells grown in the absence (Ϫ) or presence (ϩ) of doxycycline (2 g/ml), and their histone H1-phosphorylating activities were determined in vitro. Kinase activity is represented as the amount of radioactive phosphate incorporated into histone H1/min. The error bars represent S.D.
Cyclin A-dependent Kinase Activity Is Increased in Rat1a Cells Overexpressing c-Jun-To determine whether this increase in cyclin A expression was accompanied by a concomitant increase in cyclin A-dependent kinase activity, extracts from Rat1a-J4 cells grown in the absence or presence of doxycycline were assayed for cyclin A-associated kinase activity. Cyclin A-associated kinase activity was ϳ3.5-fold higher in the c-Jun-overexpressing Rat1a cells than in the control cells (Fig.  1B), confirming that the increase in cyclin A expression was accompanied by an increase in its associated kinase activity.
c-Jun Enhances Expression of Cyclin A at a Transcriptional Level-Time course experiments showed that the cyclin A protein levels in the control Rat1a cells decreased ϳ2.5-fold over time when the cells were grown in suspension; however, when c-Jun was induced in these cells, cyclin A levels were sustained ( Fig. 2A) This pattern of expression was also observed at the transcriptional level. Northern blot analysis showed that cyclin A mRNA in the c-Jun-overexpressing cells remained elevated compared with cells grown in the uninduced state (Fig. 2B). In the control Rat1a-GFP cells, cyclin A mRNA rapidly decreased when the cells were grown under nonadherent conditions (data not shown). Analysis of the cyclin A mRNA half-life confirmed that the increase in cyclin A expression was transcriptional rather than due to increased stability of the mRNA (Fig. 3). No difference in cyclin A mRNA half-lives was observed between Rat1a-J4 cells in the uninduced and induced states, with both having a half-life of ϳ30 min (Fig. 3). A similar result was obtained using the other c-Jun-overexpressing Rat1a clone, Rat1a-J2 (see Supplemental Material), suggesting that these results are not due to clonal variation.
Since there are two distinct cyclin A genes in mammals, Ccna1 and Ccna2, encoding cyclin A1 and cyclin A2, respectively, we analyzed our Rat1a cells to determine whether c-Jun expression enhanced cyclin A1 as well. Western blot analysis using an antibody specific to cyclin A1 showed that its expression was enhanced in Rat1a-J4 cells grown under nonadherent conditions (Fig. 4). Furthermore, the pattern and time course of induction was similar for that observed for cyclin A2 (compare Figs. 2A and 4).

The ATF Site in the Cyclin A2 Promoter Is Necessary for Its Activation in Nonadherent Rat1a Cells Overexpressing c-Jun-
The cyclin A2 promoter contains numerous regulatory elements, including a variant AP-1 site (at position Ϫ279), consensus ATF (Ϫ80), Sp1 (Ϫ197, Ϫ173, and Ϫ142), and E2F sites (ϩ31 and ϩ165) (45). To determine which element in the promoter of cyclin A2 contributed to c-Jun-mediated expression of cyclin A, luciferase assays using the full-length cyclin A2 promoter and various 5Ј deletion and mutant promoter constructs linked to a luciferase reporter gene were performed (Fig. 5). Luciferase assays using the 7.5-kb DNA fragment of the cyclin A2 promoter showed that the promoter activity was ϳ2-3-fold higher in the Rat1a-J4 cells overexpressing c-Jun compared with that in the control cells. Deletion of the regions between Ϫ7300 and Ϫ133 showed no significant decrease in this enhanced promoter activity (Fig. 5). Mutation of the ATF site at position Ϫ80 in the Ϫ266/ϩ205 deletion construct, however, completely abolished the increase in promoter activity observed in this construct containing an intact ATF site in Rat1a-J4 cells overexpressing c-Jun (p Ͻ 0.05) (Fig. 5). Thus, the ATF site is crucial for c-Jun-mediated transcriptional activation of cyclin A2 in Rat1a-J4 cells grown under nonadherent conditions. Similar results were obtained using Rat1a-J2 cells (see Supplemental Material).
c-Jun Binds Directly to the ATF Site in the Cyclin A2 Promoter-To determine whether c-Jun binds directly to the cyclin A2 promoter at the ATF site at position Ϫ80, gel mobility shift assays were performed using an end-labeled oligonucleotide containing the wild-type ATF site in the promoter of cyclin A2 (Fig. 6). Specific DNA-protein complexes are indicated in Fig. 6 (*). An additional complex was formed by the nuclear extracts from the nonadherent cells induced to express c-Jun as compared with the control cells (see asterisks in lanes 1 and 11 in Fig. 6). Incubation of the nuclear extracts with unlabeled wild type oligonucleotide resulted in a significant loss of binding, whereas incubation with a mutated oligonucleotide did not alter binding of the complexes to the DNA, suggesting that the complexes formed were specific (Fig. 6). In addition, a supershifted complex was observed when the nuclear extracts from the c-Jun-overexpressing cells were incubated with an antibody to ATF2 (Fig. 6, ࡗ). This supershifted complex was absent in the control cells. No altered supershift pattern from the uninduced state was observed with antibodies against ATF1, ATF3, ATF4, CREB1, and CREM1 (Fig. 6). These results indicate that c-Jun binds directly to the ATF site in the cyclin A2 promoter in conjunction with ATF2. Similar results were ob- To confirm that c-Jun interacts directly with the cyclin A2 promoter in vivo, chromatin immunoprecipitation assays with an antibody against HA were performed using Rat1a cells stably transfected with HA-tagged c-Jun grown nonadherently for 4 days in the presence and absence of 2 g/ml doxycycline, and primers flanking the ATF site (Fig. 5) (this clone was used, since use of a c-Jun antibody for chromatin immunoprecipitation analysis did not yield consistent results). Western blotting using antibodies to cyclin A2 and hemagglutinin (to detect the HA-tagged c-Jun) confirmed that the expression of cyclin A2 and c-Jun in these cells was comparable with that in Rat1a-J4 and -J2 cells (Fig. 7A). c-Jun is bound to the cyclin A2 promoter, and in the presence of doxycycline, this binding is enhanced (Fig. 7B). There was approximately a 2-3-fold (p Ͻ 0.05) induction of binding at this site when c-Jun expression was induced, as determined by quantitative real time PCR (Fig. 7C).
c-Jun Indirectly Regulates the Cyclin A1 Promoter-The cyclin A1 promoter has a putative AP-1 binding site at positions Ϫ231 to Ϫ223 (46). To determine whether c-Jun was also able to activate transcription from the cyclin A1 promoter, Rat1a-J4 cells were transfected with a cyclin A1 promoter luciferase construct spanning positions Ϫ1299/ϩ145. c-Jun activated transcription from this construct in cells grown under nonadherent conditions (Fig. 8A). Deletion of Ϫ1299 to Ϫ190 did not affect promoter activity, suggesting that c-Jun-induction of this promoter does not occur through the putative AP-1 binding site. Deletion of Ϫ190 to Ϫ129 resulted in a significant decrease in promoter activity (Fig. 8A). A further deletion of Ϫ190 to Ϫ69 did not result in any further decrease in promoter activity (Fig. 8A). Although the Ϫ129/ϩ145 and Ϫ69/ϩ145 deletion constructs resulted in a significant loss in promoter activity, they still showed some responsiveness to c-Jun, although it was not as pronounced as the Ϫ190/ϩ145 construct. These results suggest that elements between Ϫ190 and Ϫ129 of the cyclin A1 promoter are required for c-Jun-induced cyclin A1 expression in Rat1a cells, and further elements between Ϫ129 and ϩ145 may also play a role. Since there are no AP-1 or ATF binding sites within these regions, these results suggest that c-Jun is acting indirectly on this promoter.
To determine whether c-Jun binds to the cyclin A1 promoter, electrophoretic mobility shift assays were performed using a fragment spanning Ϫ190 to Ϫ129 of the cyclin A1 promoter and nuclear extracts from Rat1a-J4 cells grown nonadherently for 3 days. A single complex was formed, which was enhanced when the cells were induced to express c-Jun (Fig. 8B). This complex did not contain c-Jun, since the addition of an anti-c-Jun antibody did not result in a supershift, suggesting that c-Jun is not a component of this complex (Fig. 8B). These results, together with the luciferase assays, suggest that c-Jun indirectly enhances expression of cyclin A1. (29,47). Thus, to determine whether overexpression of cyclin A2 allows Rat1a cells to grow in a nonadherent manner, we infected Rat1a cells with a retroviral vector containing the cyclin A2 cDNA. Cyclin A levels in the resulting infected clones were comparable with those in the nonadherent Rat1a-J2 and Rat1a-J4 cells induced to express c-Jun (Fig. 9A). These clones failed to proliferate in suspension (data not shown) and furthermore failed to form colonies in soft agarose (Fig. 9B). Similar experiments were performed using the BabeHygro2-HCA2 cyclin A expression vector after selection of clones in hygromycin B and similar results were obtained (data not shown). These data indicate that overexpression of cyclin A alone is not sufficient to allow Rat1a cells to grow in an anchorage-independent fashion. Inhibition of cyclin A expression, however, resulted in a reduction in anchorageindependent growth. Transfection of a cyclin A antisense construct into Rat1a-J4 cells reduced the number of colonies formed in soft agarose after induction of c-Jun (Fig. 10A) and also decreased their ability to grow in suspension (Fig. 10B). Western blot analysis confirmed that expression of cyclin A was reduced by the antisense construct (Fig. 10C). Taken together, these results imply that cyclin A is necessary, but not sufficient, for c-Jun-mediated anchorage-independent growth.

The Cytoplasmic Oncogenes Ras and Src Transcriptionally Activate Cyclin A2 via the ATF Site (Ϫ80) and This Is Dependent on c-Jun-
The c-Jun-AP-1 complex is known to be critical for mediating the downstream effects of many cytoplasmic oncogenes. Since Ras and Src are well characterized cytoplasmic oncogenes known to transform fibroblasts (48,49) and induce nonadherent cell growth (32,50,51), we tested whether Ras and Src would transcriptionally activate the cyclin A2 promoter. Transient transfection of Rat1a cells with the Ϫ266/ ϩ205 cyclin A2 promoter construct and an activated Ha-Ras (12V) expression construct resulted in a 2-3-fold increase (p Ͻ 0.05) in promoter activity as compared with control cells (Fig.  11A). This increase in promoter activity in the presence of activated Ras was not observed when the Ϫ266/ϩ205 cyclin A2 promoter construct containing the mutated ATF site was used (Fig. 11A). In order to determine whether c-Jun is required for this Ras-induced increase in cyclin A2 promoter activity, Rat1a   FIG. 4. c-Jun enhances cyclin A1 expression. Proteins were isolated from Rat1a-GFP and J4 cells grown in suspension for the indicated times in the absence (Ϫ) or presence (ϩ) of 2 g/ml doxycycline and analyzed for cyclin A1 expression by Western blotting with a cyclin A1-specific antibody.
FIG. 5. c-Jun activates the cyclin A2 promoter. Rat1a-J4 cells were grown in the absence or presence of 2 g/ml doxycycline and transfected with the indicated cyclin A2 reporter constructs and incubated for 4 days nonadherently, after which luciferase activity was determined. The luciferase activities are reported relative to the uninduced cells after normalizing the luciferase signals to the signal obtained from the co-transfected Renilla luciferase. The figure represents the mean Ϯ S.D. of at least three independent experiments performed in duplicate. cells were transiently co-transfected with the Ϫ266/ϩ205 cyclin A2 promoter construct and activated Ras together with TAM67, a dominant negative c-Jun. The enhanced cyclin A2 promoter activity observed in the presence of activated Ras was abolished by TAM67 but not by the CMV vector control (Fig.  11B). A similar increase in cyclin A2 promoter activity (p Ͻ 0.01) was observed when Src was co-transfected with the cyclin A2 promoter (Fig. 11C). c-Jun was also required for this Srcinduced cyclin A2 promoter activity, since the increase was blocked (p Ͻ 0.05) by TAM67 (Fig. 11D). Co-transfection of the cyclin A2 promoter with the epidermal growth factor receptor (EGFR) also resulted in an increase in promoter activity (Fig.  11E), although this was not statistically significant. This increase was also blocked by TAM67 (Fig. 11F). One possible reason for the marginal increase observed with EGFR is that Rat1a cells already express EGFR, and any additional EGFR would not result in a significant increase in activation of this receptor. These results suggest that up-regulation of cyclin A2 by the cytoplasmic oncogenes Ras and Src requires the ATF site (Ϫ80) in the cyclin A2 promoter and is dependent on c-Jun. DISCUSSION We have previously reported that aberrant expression of human c-Jun was sufficient to allow Rat1a cells to grow in an anchorage-independent fashion (9). The mechanism(s) by which c-Jun induces nonadherent growth of Rat1a cells remains unknown. The selective regulation of "downstream" target genes is critical for this process (14,39,40). The data presented in this paper show that cyclin A is one of the mediators of this process. Overexpression of c-Jun resulted in an increase in cyclin A expression in Rat1a cells grown in suspension, which was accompanied by a concomitant increase in cyclin A-associated kinase activity. This c-Jun-mediated increase in cyclin A expression was due to both direct and indirect induction of cyclin A2 and cyclin A1, respectively.
There is substantial evidence supporting a critical role for cyclin A in cellular transformation and human malignancies.
Elevated levels of cyclin A2 have been found in various tumor models as well as human tumors (52)(53)(54)(55). Several studies have also showed that deregulated expression of cyclin A is associated with oncogenic transformation and anchorage-independent growth (56,57). Cyclin A is a target of adhesion-dependent signals, and consistent with this, adhesion-independent expression of cyclin A is sufficient for anchorage-independent cell growth (47,50). Furthermore, elevated cyclin A expression is also associated with c-Myc-dependent transformation of Rat1a cells (29). Deregulated cyclin A expression is also thought to be a crucial factor for transformation mediated by several oncoproteins such as Ras and HPV E7 (32,58). The mechanism(s) by which these oncogenes and signal transduction pathways deregulate cyclin A2 expression remains unknown. There has been no evidence thus far, however, that cyclin A is a direct target gene of these oncogenes during transformation. Multiple signaling pathways contribute to Ras-induced cell transformation (32). Our results demonstrate that cyclin A2 is a direct target of c-Jun, with the ATF site (-80) being a crucial regulatory element. Our results also show that this c-Jun-induced cyclin A2 expression via the ATF site at position Ϫ80 is necessary for the aberrant cyclin A2 expression associated with Ras/Src-induced transformation.
Cyclin A1 is also activated by oncogenes and its overexpression has been associated with the transformed phenotype (59). c-Myb activates the cyclin A1 promoter and induces cyclin A1 expression (59). Overexpression of cyclin A1 has also been observed in acute myelocytic leukemias (60,61), testicular germ cell tumors (62) and is thought to play a role in breast cancer (63). These studies support a role for cyclin A1 in transformation, and our results suggest that it may be one of the target genes that play a role in c-Jun-induced anchorage-independent growth in Rat1a cells. Our results suggest that rather than acting on the putative AP-1 site in the cyclin A1 promoter, c-Jun indirectly enhances expression of this gene through transcription factor binding sites downstream of the AP-1 site (Fig.   FIG. 6. c-Jun binding at the cyclin A2 promoter. Electrophoretic mobility shift assay using the ATF sequence of the cyclin A2 promoter as a probe with nuclear extracts prepared from Rat1a-J4 cells grown for 3 days in the absence (Ϫ) or presence (ϩ) of 2 g/ml doxycycline, incubated with various antibodies. Competition with the unlabeled wild-type ATF sequence (wt ATF) confirmed the specificity of the DNA-protein complexes, whereas competition with an unlabeled mutated ATF sequence (mut ATF) failed to cause a reduction in complex formation. *, the specific DNA-protein complexes; ࡗ, the additional supershifted complex observed by incubation with an ATF2 antibody in the nuclear extracts from the c-Jun-overexpressing cells. 8). We and others have shown that c-Jun can indirectly regulate gene expression by interacting with other transcription factors (39,64). Previous work from our laboratory has shown that c-Jun enhances stathmin promoter activity through 2 E2F sites (39). Other studies have shown that both c-Jun and v-Jun increase E2F activity by deregulating phosphorylation of the retinoblastoma protein (10,65). The human cyclin A1 5Ј upstream region contains an E2F site between position ϩ142 and ϩ145 (41), a region encompassed by our promoter constructs (Fig. 8A). Activation of E2F by c-Jun and its consequent interaction with the cyclin A1 promoter may, in part, be responsible for c-Jun-induced activation of the cyclin A1 promoter. In addition, the region spanning Ϫ190 and Ϫ129, which has enhanced binding of a complex in c-Jun-overexpressing cells (Fig.  8), contains putative binding sites for numerous transcription factors, including c-Rel, Elk-1, Oct-1, and NF-B, and c-Jun FIG. 7. c-Jun binds to the cyclin A2 promoter in vivo. Protein extracts were prepared from Rat1a-HA-Jun cells grown nonadherently for 3 days in the absence (Ϫ) and presence (ϩ) of 2 g/ml doxycycline and analyzed by Western blotting for the expression of cyclin A2 and hemagglutinintagged c-Jun (A). ␤-Tubulin was used as a control for protein loading. B, chromatin immunoprecipitation using soluble chromatin prepared from Rat1a-HA-Jun cells incubated nonadherently for 4 days in the absence (Ϫ) and presence (ϩ) of 2 g/ml doxycycline. The soluble chromatin was immunoprecipitated using antibodies to HA, and a control IgG and PCR was performed with primers spanning the ATF site in the cyclin A2 promoter. C, the increase in binding between the uninduced and induced states was determined by quantitative real time PCR after correcting for the PCR products obtained using the control IgG. may indirectly activate the cyclin A1 promoter by enhancing expression of one of these transcription factors.
Several studies have shown that cyclin A2 is a target gene for other family members of the AP-1 transcriptional complex such as c-Fos (34,66) and ATF (67,68) in other cell types, including chondrocytes and osteoblasts as well as CREM and CREB (69).
These studies also showed that the E2F site in the cyclin A2 promoter was required for this c-Fos-induced activation. In our current study, the c-Jun-induced increase in cyclin A2 expression was due to direct binding of c-Jun to the ATF site in the cyclin A2 proximal promoter. Furthermore, induction of c-Jun resulted in a change in the composition of the AP-1 complex binding to the ATF site. Previous studies have shown that c-Jun-Fra2 dimers increase cyclin A expression in NIH 3T3 cells by binding to the cAMP-response element in the cyclin A promoter (70), and these cells are capable of growing under conditions of serum deprivation. Our results indicate that c-Jun dimerizes with ATF2 and interacts with the ATF site in the cyclin A2 promoter to increase its expression in anchorageindependent growth (Figs. 6 and 7). Taken together, these data suggest that, depending on the cell system used, different AP-1 dimers may be involved in the activation of cyclin A during c-Jun-induced nonadherent cell growth, and consequently different biologic phenotypes may result. AP-1-induced cellular transformation may be mediated by different members of the Jun proteins. We have shown that the different Jun family members induce differential nonadherent cell growth of Rat1a cells (40). Thus, it can be expected that some of the mechanisms and pathways of transformation brought about by the different Jun proteins will be shared. This is true for cyclin A2, whose induction parallels the degree of nonadherent cell growth induced by the different Jun family members (data not shown). Additional studies from other laboratories support this result. JunB has recently been shown to enhance mouse cyclin A expression via the cAMP-response element in the cyclin A promoter (71). This result is not sur-FIG. 8. c-Jun indirectly activates the cyclin A1 promoter. A, Rat1a-J4 cells were transfected with the indicated cyclin A1 promoter luciferase constructs and incubated for 48 h under nonadherent conditions. Results represent relative luciferase activity after normalizing to Renilla luciferase to correct for transfection efficiency. The figure represents the mean Ϯ S.D. of triplicate experiments. B, gel mobility shift assay using a labeled probe spanning from Ϫ190 to Ϫ129 of the cyclin A1 promoter and nuclear extracts from Rat1a-J4 cells grown nonadherently for 3 days in the absence (Ϫ) and presence (ϩ) of 2 g/ml doxycycline. The arrow represents the DNA-protein complex formed. This complex was not supershifted when the extracts were incubated with an antibody to c-Jun (compare lanes 1 and 2 and  lanes 3 and 4). 9. Overexpression of cyclin A2 alone is not sufficient to confer anchorage-independent growth on Rat1a cells. Rat1a cells were infected with the retroviral cyclin A expression vector pLRT-cyclin A, and infected cells were selected with blasticidin to obtain stable clones. A, pools and clones were subjected to Western blot analysis after 3 days of treatment with 2 g/ml doxycycline to determine the level of cyclin A expression. B, soft agarose colony formation of Rat1a cells stably expressing cyclin A (pLRT-cyclin A) compared with Rat1a-J4 cells.
prising, since it has been shown that JunB can substitute for c-Jun with respect to development as well as gene transcription and cell proliferation (72). Thus, although the exact phenotypes of c-Jun and JunB overexpression are not identical, there may be some common aspects to the mechanisms of transformation induced by these two proteins.
Ectopic expression of cyclin A in our Rat1a cells did not result in anchorage-independent growth. Previous studies have shown that overexpression of cyclin A is sufficient for anchor-age-independent growth of NRK and Rat1a cells (29,47). The reasons for this discrepancy between these published reports and our data are not clear but may include a difference in cyclin A expression vectors and systems used to establish the cyclin A-overexpressing cell lines in the different studies. We used two different retroviral vectors (BabeHygro2-HCA2 and pLRTcyclin A), and the levels of cyclin A expression achieved were comparable with those observed in the c-Jun-overexpressing Rat1a cells. Although the level of cyclin A expression in the FIG. 10. Antisense cyclin A2 suppresses nonadherent growth and colony formation of Rat1a-J4 cells. Rat1a-J4 cells were cotransfected with a retroviral vector expressing a region of the 3Ј-end of the cyclin A mRNA in an antisense orientation with pSV2-Neo, and a number of clones were selected in G418. A, Rat1a-J4 cells and a number of clones expressing the antisense construct were grown for 14 days in soft agarose in the presence or absence of 2 g/ml doxycycline, after which the number of colonies formed was counted. The colony size is shown on the right. B, cells were also grown in PolyHEME-coated dishes in the absence or presence of doxycycline, and cell growth was determined at the indicated times by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. C, protein lysates from Rat1a-J4 cells and the indicated cyclin A antisense clones grown for 7 days in suspension in the presence or absence of doxycycline were analyzed by Western blotting for cyclin A. Relative expression of cyclin A is represented in the bar graph after normalization to ␤-tubulin expression.
stable cyclin A clones was comparable with those obtained in the c-Jun overexpressing cells, other factors induced by c-Jun in these cells may contribute to anchorage-independent growth and alleviate the need for such a high level of cyclin A. We have previously shown that both stathmin and HMG-I/Y are neces-sary for c-Jun-induced anchorage-independent growth in Rat1a cells but alone are not sufficient to induce this phenotype (14,39). Thus, it is likely that a set of genes is required to fully recapitulate anchorage-independent growth observed in c-Junoverexpressing Rat1a cells. FIG. 11. Activation of the cyclin A promoter by the Ras signaling pathway occurs via the ATF site. Rat1a cells were transiently co-transfected with the indicated cyclin A promoter reporter constructs and expression vectors for activated HA-Ras (12V) (A), v-Src (C), and EGFR (E) and incubated under nonadherent growth conditions for 48 h. The results represent relative luciferase activity after normalizing to Renilla luciferase to correct for transfection efficiency. Rat1a cells were transiently co-transfected with the Ϫ266/ϩ205 cyclin A promoter reporter construct and expression constructs for activated HA-Ras (12V) (B), v-Src (D), and EGFR (F) together with either CMV vector or CMV-TAM67 and incubated for 48 h under nonadherent growth conditions. Luciferase activity was determined as described above. c-Jun forms a variety of dimers with either itself or other members of the AP-1 family, resulting in complexes with different sequence specificities. These complexes bind to various sites in the promoters of numerous genes, making it difficult to identify specific transformation-relevant genes. In all likelihood, a number of target genes are necessary for c-Jun-induced biology. Using our inducible Rat1a cell system, we have recently identified HMG-I/Y as a direct c-Jun target gene that is necessary, but not sufficient, for anchorage-independent growth (14). In this study, we have identified cyclin A2 as another c-Jun target gene that is also necessary, but not sufficient, for c-Jun-mediated anchorage independent growth. It is likely that this inducible system will identify additional c-Juninduced genes that are necessary for nonadherent cell growth.