Mitogenic Regulation of p27^Kip1 Gene Is Mediated by AP-1 Transcription Factors

Abstract

The abundance of cyclin-dependent kinase inhibitor p27^Kip1 during the cell cycle determines whether cells will proliferate or become quiescent. Although the post-translational regulation of p27^Kip1 is well established, its transcriptional regulation is poorly understood. Here, we report that mitogenic stimulation of quiescent HEK293 and Huh7 cells showed a rapid decline in the levels of p27^Kip1 transcript by 2.4 ± 0.1-fold. Inhibition of the p27^Kip1 gene in response to mitogens involved transcriptional down-regulation and required newly synthesized protein(s). Mutation of the AP-1 element at position −469 in the human p27^Kip1 promoter abrogated the effect of mitogens. The recruitment of the AP-1 complex to the p27^Kip1 promoter was confirmed by in vitro DNA binding and chromatin immunoprecipitation studies. Reporter gene analysis combined with enforced expression of Jun/Fos proteins suggested the involvement of Jun/Fos heterodimer in the transrepression process. Both MAPK and phosphatidylinositol 3-kinase signaling pathways appeared to mediate p27^Kip1 transcription. Furthermore, hepatitis B virus X protein-mediated down-regulation of p27^Kip1 in a transgenic environment correlated with an increase in c-Fos levels, reiterating the physiological relevance of AP-1 in the transcriptional regulation of p27^Kip1. Collectively, our studies present the first evidence demonstrating the role of the AP-1 complex in transcriptional down-regulation of the p27^Kip1 gene following mitogenic stimulation.

Introduction

Cyclin-dependent kinase inhibitors (CKIs)³ function as brakes for the cell division cycle by inhibiting cyclin/cyclin-dependent kinase complexes. CKIs belong to two different classes, INK4 and CIP/KIP proteins, depending on their sequence homology and mode of action. Although the INK4 proteins (p16^INK4a, p15^INK4b, p18^INK4c, and p19^INK4d) inhibit the kinase activities of CDK4 and CDK6 by interfering their association with D-type cyclins, the CIP/KIP proteins (p21^Cip1, p27^Kip1, and p57^Kip2) inhibit the activities of cyclin D-, E-, A-, and B-associated cyclin-cyclin-dependent kinase complexes (1). The CIP/KIP proteins share an N-terminal domain that binds the cyclin/cyclin-dependent kinase subunits, but their C-terminal sequences are distinct, leaving them to be diversely regulated.

p27^Kip1 was identified as a CKI in G₁-arrested cells (2). The p27^Kip1 levels are maximal in the G₀-G₁ phase and progressively decline in the G₁ phase leading to cell cycle progression from G₁ to S phase (3). A number of post-translational mechanisms are known to control the stability of p27^Kip1 during different phases of the cell cycle. For example, phosphorylation at Tyr-88 by Src tyrosine kinase along with other signals can transform p27^Kip1 from an inhibitor of cyclin E-cyclin-dependent kinase complexes to its substrate (4). Likewise, cyclin E/CDK2 phosphorylates p27^Kip1 at Thr-187 (5). Phosphorylation at various residues targets p27^Kip1 for ubiquitination and proteasomal degradation by SCF^Skp2 ubiquitin ligase during the G₁-S phase (6). However, phosphorylation at Ser-10 by Mirk/Dyrk kinase provides stability to p27^Kip1 and facilitates its CRM1-dependent nuclear export (7). In addition, some translational control mechanisms are also reported to regulate the levels of p27^Kip1 in cells (8). Mechanisms that regulate the transcription of the p27^Kip1 gene are poorly understood.

Increasing evidence now supports the role of transcriptional mechanisms that might control the levels of p27^Kip1. The FOXO transcription factors have been shown to activate p27^Kip1 transcription leading to cell cycle arrest (9). Other activators of the p27^Kip1 promoter include Sp1, NF-Y, E2F1, and BRCA1 (10–12). Interestingly, c-Myc, Id3, Hes1, among others are known to inhibit the p27^Kip1 promoter (13–15). Considering the fact that growth factors initiate the cell cycle through destabilization of CKIs, in this study we investigate the role of serum and epidermal growth factor (EGF) on the regulation of the p27^Kip1 promoter in quiescent cells. We found that the AP-1 family of proteins modulated by the MAPK and PI3K signaling pathways are essential for the regulation and maintenance of basal transcriptional activity of the p27^Kip1 gene.

EXPERIMENTAL PROCEDURES

Expression Vectors and Reporter DNA Constructs

Expression constructs for RSV-Jun and c-Fos were kindly provided by Dr. A. Weisz (University of Napoli, Italy), and β-galactosidase plasmid (pCH110) was from GE Healthcare. Construction of the Fos-dominant negative mutant (Fos-DN) has been described earlier (16). The full-length human p27^Kip1 luciferase reporter construct, p27PF, and its deletion (p27AflII) construct were kindly provided by Dr. T. Sakai (10). The luciferase reporter construct driven by a minimal promoter and two copies of 12-O-tetradecanoylphorbol-13-acetate-responsive element (2×-TRE luc) was from Dr. M. Karin (17). The p27PF fragment (3.6 kb, XhoI fragment) was recloned in pCAT3 basic vector (Promega) to generate p27-I-CAT. p27-II-CAT was developed by cloning the KpnI-BglII fragment (∼570-bp region) from p27AflII. p27-III-CAT was created by generating point mutation in the AP-1 element of p27-II-CAT using the QuickChange site-directed mutagenesis kit (Stratagene). The primers used to mutate the AP-1 site (5′ to 3′) in the p27^Kip1 promoter are as follows: mutated nucleotides are underlined, forward (F) TTTCTTCTTCGTTGGCCTCCC and reverse (R) GGGAGGCCAACGAAGAAGAAA.

Chemicals, Radiochemicals, and Antibodies

The chemicals used and their working concentrations were as follows: PD98059 (20 μm), LY294002 (50 μm), 12-O-tetradecanoylphorbol-13-acetate (100 ng/ml), and EGF (10 ng/ml)were from Calbiochem; 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (DRB 10 μm), SYBR Green (0.25×), and cycloheximide (15 μg/ml) were from Sigma. [γ-³²P]ATP was supplied by PerkinElmer Life Sciences. All restriction enzymes were from New England Biolabs. Antibodies against c-Jun, c-Fos, total ERK, total Akt, p27^Kip1, pGSK-3β, pElk-1, GAPDH, and siRNAs against human c-Fos, human c-Jun, and control were from Santa Cruz Biotechnology. Antibodies against pERK and pAkt were procured from Cell Signaling Technology. Hepatitis B virus X (HBx)-specific antibody has been described previously (18).

Cell Culture and Flow Cytometry

The human hepatoma Huh7 and human embryonic kidney 293 (HEK293) (ATCC CRL-1573) cell lines were seeded in 60-mm dishes (5 × 10⁵ cells). DNA and siRNA transfections of cells were carried out using Lipofectin (Invitrogen) as per the manufacturer's protocol. pCH110 was co-transfected as internal control. Cells were made quiescent by culturing in 0.1% serum for 48 h and analyzed further for different parameters. Flow cytometry of cells was performed as described earlier (18).

RNA Isolation and Real Time PCR

Total RNA was isolated from cells and liver tissue under different treatments using TRIzol reagent as per the supplier's instructions (Invitrogen) and reverse-transcribed (5 μg of RNA in 50 μl) using oligo(dT) primer and murine leukemia virus reverse transcriptase as per the supplier's instructions (Promega). Real time quantitative PCR was conducted using SYBR Green assay (19). Each PCR was composed of 2.5 μl of cDNA, 1× PCR buffer, 0.25 mm dNTP, 250 nm F and R primers, 0.25× SYBR Green (Sigma S9430), 2.5 units of Taq polymerase (MBI Fermentas) in a 25 μl volume. PCR condition was 95 °C for 5 min followed by 40 cycles of 95 °C for 10 s, 60 °C for 15 s, and 72 °C for 15 s. Quantification was performed using ΔΔCt method with GAPDH as reference gene. Sequences of the PCR primers (5′ to 3′) were as follows: p27^Kip1, F, GGTTAGCGGAGCAATGCG, and R, TCCACAGAACCGGCATTTG; GAPDH, F, CGACCACTTTGTCAAGCTCA, and R, AGGGGTCTACATGGCAACTG; and β-actin, F, TGACGGGGTCACCCACACTGTGCCCATCTA, and R, CTAGAAGCATTTGCGGTGGACGATGGAGGG. Data are represented as bar graphs and are means ± S.D. of three independent observations.

Western Blot Analysis

The method for Western blotting has been described previously (18). GAPDH was used as loading control.

Chloramphenicol Acetyltransferase Assay and Luciferase Assay

Chloramphenicol acetyltransferase assay was performed as described earlier (16), and luciferase assay was performed as per the supplier's instructions (Promega). Chloramphenicol acetyltransferase and luciferase activity was normalized against β-galactosidase activity. Relative chloramphenicol acetyltransferase activity was expressed as mean ± S.D. of three independent observations. Statistical significance was calculated using Student's t test. p value is indicated in figure legends. Error bars in data represent standard deviation.

Electrophoretic Mobility Shift Assay (EMSA)

EMSA for the AP-1 site was performed as described previously (20). Nuclear extracts from cells were incubated with end-labeled probe, and the protein-DNA complexes were resolved by electrophoresis in a 5% polyacrylamide gel. Supershift analysis of nuclear extracts was performed using 1 μg of AP-1-specific antibodies. Sequences of the oligonucleotides (5′ to 3′) used for EMSA were as follows: AP-1 consensus, F, CGCTTGATGAGTCAGCCGGAA, and R, TTCCGGCTGACTCATCAAGCG; p27AP-1, F, TTTCTTCTTCGTCAGCCTCCC, and R, GGGAGGCTGACGAAGAAGAAA; and p27AP-1 (mutated), F, TTTCTTCTTCGTTGGCCTCCC, and R, GGGAGGCCAACGAAGAAGAAA.

Chromatin Immunoprecipitation Assay (ChIP)

ChIP was performed as described earlier (21). The immune complexes were captured using protein A-Sepharose beads. After a series of washing steps, the beads were extracted in 500 μl of elution buffer (0.1 m NaHCO₃, 1% SDS) and analyzed by PCR for AP-1 recruitment on the p27^Kip1 promoter. Primers used for PCR amplification (5′ to 3′) were as follows: p27 F, CAACCAATGGATCTCCTCCT, and p27 R, GCCTCTCTCGCACTCTCAAA spanning the region −525 to −24 relative to translation start site on the p27^Kip1 promoter.

Bioinformatic Analysis

The putative transcription factor binding sites on p27^Kip1 promoter was searched using TransFac Search software.

X15-myc Transgenic Mouse Model and in Vivo Regulation of p27^Kip1

The development of X15-myc transgenic mouse model of hepatocellular carcinoma has been reported earlier (22). Phosphate-buffered saline (1 ml) containing 75 μm PD98059 was injected intravenously into transgenic and control mice 6 h before the sacrifice, following which livers were processed as described below. Injection of phosphate-buffered saline alone (mock injection) served as appropriate negative control. A portion of each liver was kept frozen for both the isolation of total RNA and the preparation of lysates for Western blotting. All experiments were independently repeated at least three times.

RESULTS

Mitogenic Stimulation Results in Transcriptional Repression of p27^Kip1 Gene

To understand the mechanism of p27^Kip1 gene regulation, serum-starved quiescent Huh7 cells were incubated with serum or EGF, and the expression of the p27^Kip1 gene was measured both at protein and mRNA levels. Fig. 1A shows that the p27^Kip1 protein level was highest in quiescent cells, which gradually declined upon serum/EGF treatment. Analysis by real time PCR showed that the p27^Kip1 mRNA level was highest under the quiescent stage, which declined sharply within 1 h of serum/EGF treatment and was maintained at low levels throughout the period of observation (Fig. 1B). Thus, mitogenic stimulus negatively regulates the levels of p27^Kip1 mRNA during the cell cycle.

FIGURE 1.

Regulation of p27^Kip1 gene expression by mitogenic stimulation. Quiescent Huh7 cells were stimulated with serum/EGF, and the expression of p27^Kip1 protein and mRNA was monitored, respectively, by Western blotting (WB) (A) and real time PCR (B). The p27^Kip1 mRNA levels were also monitored as above in the presence of DRB (C) or with cycloheximide (CHX) for 3 h (D). *, statistically significant difference at p < 0.05 in B.

To establish whether a mitogen-dependent decrease in the mRNA level was due to either destabilization or involved a transrepression mechanism, the level of RNA was analyzed in the presence of DRB, a well known inhibitor of mRNA synthesis. Treatment of quiescent cells with DRB or DRB along with serum/EGF showed a sharp and similar order of decline in the p27^Kip1 mRNA levels suggesting mitogen-dependent p27^Kip1 down-regulation was due to inhibition of mRNA synthesis rather than its destabilization (Fig. 1C). Further abrogation of serum/EGF-dependent effects in the presence of cycloheximide, a protein synthesis inhibitor, suggested the involvement of newly synthesized gene products in p27^Kip1 transcriptional inhibition (Fig. 1D).

Transcriptional Repression of p27^Kip1 Gene Is Mediated by AP-1 Elements

We used the chloramphenicol acetyltransferase reporter assay to understand the mechanistic details of transcriptional inhibition and/or involvement of repressor(s). Cells transfected with full-length p27^Kip1 promoter reporter construct (p27-I-CAT) showed a time-dependent decline in the reporter activity following serum/EGF treatment and suggested that mitogenic effects were correctly reflected in our assays (Fig. 2A). As p27-II-CAT showed a similar reporter activity as compared with the full-length promoter (p > 0.1), it was inferred that the essential transcriptional regulatory elements were present within −571-bp region of the p27^Kip1 promoter (Fig. 2B). Furthermore, similar levels of inhibition with p27-II-CAT in the presence of serum/EGF confirmed the presence of regulatory elements within −571-bp region of the p27^Kip1 promoter (Fig. 2D).

FIGURE 2.

Localization of transcriptional repressor elements on the p27^Kip1 promoter. Huh7 cells were transfected with different reporter plasmids as indicated, and the reporter gene activity was measured 48 h after serum starvation or after serum/EGF stimulation for the indicated times. A, reporter gene activity of full-length p27^Kip1 promoter (p27-I-CAT); B, reporter activity of p27-I-CAT, p27-II-CAT, p27-III-CAT (mutated for AP-1 site), and pCAT3basic plasmids in quiescent cells; C, nucleotide sequence of −571 to +3 region of the human p27^Kip1 gene promoter showing three transcriptional start sites (bent arrow) and binding sites for different transcription factors (boxed). The numbering is assigned with reference to base “A” (+1) of the translation start site “ATG.” The p27-II-CAT reporter carries the native promoter sequence (−571 to −1 region shown by arrowheads), and p27-III-CAT carries two mutations in the AP-1 box of the promoter at −469 position (shown by asterisks). The forward and reverse ChIP primers are shown by arrows. The reporter activities of p27-II-CAT and p27-III-CAT plasmids are shown in D and E, respectively. *, statistically significant difference at p < 0.05; #, statistically nonsignificant difference at p > 0.1.

As this transcriptional repression was observed immediately after mitogen treatment and required newly synthesized proteins, we speculated the involvement of immediate early proteins in this process. The TransFac search analysis of the human p27^Kip1 gene promoter (−571 to +3 region) predicted binding sites for several transcription factors, including an AP-1 element (5′ CTTCGTCAGC 3′) at position −469 (Fig. 2C). Interestingly, this element is located upstream of two major transcription start sites (−403 and −153) (23, 24). To test the role of AP-1 site in transrepression, this element was mutated to 5′ CTTCGTTGGC 3′ in p27-III-CAT. A significant decline in the reporter activity with p27-III-CAT as compared with p27-II-CAT confirmed the involvement of AP-1 element in the basal transcription of the p27^Kip1 gene (Fig. 2B). Interestingly, p27-III-CAT-transfected quiescent cells did not respond to mitogenic stimulus (Fig. 2E). Thus, the inhibitory effect of growth factors on the p27^Kip1 promoter is conferred by AP-1 elements.

Jun/Fos Heterodimers Inhibit p27^Kip1 Expression

The AP-1 complexes involved in interaction with their cognate elements on the p27^Kip1 promoter were further characterized by EMSA. The nuclear extracts from asynchronously growing HEK293 and Huh7 cells showed two major protein-bound DNA complexes with both consensus and p27^Kip1 promoter-derived AP-1 elements suggesting the involvement of AP-1 proteins in p27^Kip1 gene expression (Fig. 3A). The specificity of this interaction was further confirmed in a competition experiment where 50- and 100-fold molar excess of unlabeled wild type probes competitively displaced the interaction but not the mutated p27AP-1 probe (Fig. 3A, compare lanes 5 and 6 with 3 and 4). The composition of AP-1 proteins in the complexes was determined by supershift assay using c-Jun and c-Fos antibodies. As reported by others (20), two major bands of slow mobility corresponding to Jun/Fos heterodimer and Jun/Jun homodimers were observed (Fig. 3B). As evident from the in vitro DNA binding assay, there was accumulation of Jun/Fos heterodimers following mitogenic stimulation, which correlated with increased levels of Jun and Fos proteins after mitogenic treatment (Fig. 3, C and D).

FIGURE 3.

Characterization of the p27^Kip1-derived AP-1 element for c-Jun/c-Fos binding. A and B, nuclear extracts from asynchronously growing HEK293 cells were used for in vitro DNA binding, and the protein-bound DNA complexes were resolved by EMSA. A, binding of the nuclear extracts to γ-³²P-labeled consensus and p27^Kipl derived AP-1 elements. The specificity of binding is shown using 50–100-fold molar excess of unlabeled AP-1 elements. Arrowheads indicate the positions of Jun/Jun homodimer and Jun/Fos heterodimers. B, supershift analysis of the protein-DNA complexes bound to p27^Kip1 AP-1 elements using c-Jun and c-Fos antibodies. Asterisks show the position of supershifted complexes. C, binding of p27^Kip1 AP-1 elements to nuclear extracts of quiescent cells treated with serum/EGF for indicated times. D, Western blot (WB) analysis in quiescent cells following serum/EGF treatment. E, ChIP analysis of quiescent cells (Q) for the recruitment of c-Jun, c-Fos, and RNA polymerase II on the p27^Kip1 promoter following serum/EGF stimulation for 1 h.

Taking a cue from the in vitro DNA binding results, the interaction of the AP-1 complexes with its cognate element in the p27^Kip1 promoter was investigated in a chromatin environment. The ChIP assay using anti-c-Jun and anti-c-Fos antibodies suggested that both c-Jun and c-Fos were recruited on the p27^Kip1 promoter in response to mitogen stimulation (Fig. 3E). Furthermore, we observed that RNA polymerase II was constitutively bound to the p27^Kip1 promoter in a transcriptionally repressed state.

Jun-dependent p27^Kip1 Promoter Activation Can Be Titrated by Fos

The functional significance of c-Jun and c-Fos binding to the p27^Kip1 promoter was evaluated through enforced expression of c-Jun and c-Fos along with p27-II-CAT or 2×-TRE luc reporter plasmids. We observed that overexpression of c-Jun in asynchronously growing HEK293 cells led to a 2–3-fold increase in the reporter activity, although c-Fos alone had no effect (Fig. 4A). Nonetheless, c-Fos significantly inhibited Jun-mediated transactivation possibly through formation of Jun/Fos heterodimers. c-Fos alone did not repress the p27^Kip1 promoter perhaps due to the presence of saturating levels of endogenous c-Fos in serum-fed cycling cells (Fig. 4A). However, overexpression of c-Fos in quiescent cells resulted in a significant repression of the p27^Kip1 promoter (Fig. 4B). The inhibitory effect of AP-1 proteins was also evident from the stimulation of the p27^Kip1 transcription in Fos-DN-expressing cells (Fig. 4C). Interestingly, we also observed the transcriptional down-regulation of p27^Kip1 in nontransformed AML-12 hepatocytes in the presence of serum suggesting a common mitogen response mechanism operational in cells. Furthermore, as expected, this down-regulation was also abrogated by Fos-DN overexpression (Fig. 4D). The functionality of the recombinants c-Jun, c-Fos, and Fos-DN was validated by performing activity assay using 2×-TRE luc reporter construct (Fig. 4E).

FIGURE 4.

Effect of c-Jun/c-Fos overexpression on p27^Kip1 promoter regulation. A–C, HEK293 cells were transfected with p27-II-CAT reporter construct along with different expression vectors as indicated, and chloramphenicol acetyltransferase (CAT) activity was measured after 48 h. Reporter activity in cells co-transfected with c-Jun and c-Fos in different ratios (B), in cells co-transfected with c-Fos followed by culture in 0.1% serum for 24 h (C), and real time PCR analysis of p27^Kip1 mRNA in cells transfected with vector or Fos-DN (2 μg) were made quiescent and stimulated with serum/EGF for 3 h (C). D, real time PCR analysis of p27^Kip1 mRNA in AML-12 cells transfected with vector or Fos-DN (2 μg), made quiescent (Q), and stimulated with serum for the indicated times. E, HEK293 cells were transfected with 2×-TRE luc reporter construct along with c-Jun and/or c-Fos construct in the presence or absence of 12-O-tetradecanoylphorbol-13-acetate (TPA), and the reporter activity was measured. Q, quiescent cells. The level of significance (p value) was derived by comparing the control with transfected groups of respective panels: *, statistically significant difference at p < 0.05; #, statistically nonsignificant difference at p > 0.1.

Furthermore, siRNA-mediated interference with c-Jun and c-Fos expression abrogated the effect of mitogens on p27^Kip1 at mRNA (Fig. 5A) and protein levels (Fig. 5B). As expected, the control siRNA did not rescue the levels of p27^Kip1 in these experiments. Thus, our data clearly indicated that Jun/Fos heterodimers are required for transcriptional down-regulation of p27^Kip1 following cell cycle entry.

FIGURE 5.

Expression of p27^Kip1 in the presence of c-Jun- and c-Fos-specific siRNAs. HEK293 cells were transfected with siRNA mixtures specific to human c-Jun and c-Fos, made quiescent, and stimulated with serum/EGF for 3 h. The p27^Kip1 mRNA level was measured by real time PCR (A), whereas the protein levels for c-Jun, c-Fos, p27^Kip1, and GAPDH were measured by Western blot (WB) analysis (B). GAPDH was used as internal control in A. Q, quiescent cells. The level of significance (p value) was derived by comparing respective quiescent and transfected groups: *, statistically significant difference at p < 0.05; #, statistically nonsignificant difference at p > 0.1.

Mitogenic Signaling Pathways Mediate p27^Kip1 Gene Expression

Mitogens activate several common signaling events that drive the cells toward proliferation. It has been reported that Raf-MEK-MAPK and PI3K-AKT pathways act cooperatively to regulate cell cycle machinery downstream of platelet-derived growth factor by regulating p27^Kip1 transcription (25). To understand the involvement of these pathways in p27^Kip1 transcription, we used specific pharmacological inhibitors of these pathways as follows: PD98059 for MEK1 and LY294002 for PI3K. Effectiveness of these inhibitors is shown by Western blotting for phospho-ERK and phospho-Akt (serine 473) in Fig. 6A. A corresponding decline in the levels of downstream proteins (phospho-Elk-1 and phospho-GSK-3β) of these pathways was also observed (Fig. 6B). We observed that inhibition of either MAPK or PI3K pathways abrogated the effects of serum/EGF on p27^Kip1 expression (Fig. 6A, compare lane 2 with 3 and 4 and lane 6 with 7 and 8). Furthermore, real time PCR analysis of p27^Kip1 mRNA in quiescent cells treated with serum/EGF in the presence of LY294002 or PD98059 confirmed the rescue of p27^Kip1 expression (Fig. 6C). Interestingly, there was no combinatorial effect of PI3K/MEK inhibitors on the p27^Kip1 mRNA level suggesting the independent role of these pathways in the regulation of p27^Kip1 promoter (data not shown).

FIGURE 6.

Characterization of cell signaling pathways that mediate mitogen-dependent regulation of p27^Kip1 gene. Quiescent (Q) Huh7 cells were treated with serum/EGF in the absence or presence of PD98059 or LY294002 for 3 h, and the protein, RNA, and nuclear extracts were subjected to the following analyses. A, Western blotting (WB) of p27^Kip1, pERK, total ERK, pAkt, total Akt, and GAPDH. B, WB analysis of phospho-Elk-1 (serine 383) and phospho-GSK-3β (serine 9). GAPDH was used as control in the indicated panels. C, real time PCR analysis of p27^Kip1 mRNA. D, EMSA showing in vitro binding of nuclear extracts with AP-1 element derived from p27^Kip1 promoter. *, statistically significant difference at p < 0.05; #, statistically nonsignificant difference at p > 0.1.

As we established a strong correlation between AP-1 expression and p27^Kip1 transcription, we analyzed the binding activity of Jun and Fos to AP-1 elements of the p27^Kip1 promoter in the presence of MAPK and PI3K inhibitors. As shown in Fig. 6D, Jun and Fos binding declined sharply in the presence of PD98059 but not in the presence of LY294002. Thus, although mitogenic inhibition of p27^Kip1 gene transcription could involve both MAPK and PI3K pathways, the MAPK signaling seemed to be important for the recruitment of AP-1 proteins on the p27^Kip1 promoter.

In Vivo Transcriptional Down-regulation of p27^Kip1 by Viral HBx Is Mediated by AP-1 Proteins

Next, we investigated the role of AP-1 proteins in the regulation of the p27^Kip1 gene in a tumor environment using the HBx-myc mouse model of hepatocellular carcinoma (16, 22). HBx, which is a multifunctional protein involved in transcription, cellular transformation, apoptosis, growth stimulation, among others, also behaves like a growth factor for cells and is also known to destabilize p27^Kip1 protein by increasing its proteasomal degradation (18). However, its involvement in the transcriptional regulation of p27^Kip1 is not known. We analyzed the p27^Kip1 transcript levels in the liver of control and HBx-myc transgenic mice. Interestingly, we observed lower levels of p27^Kip1 transcripts in the transgenic samples (Fig. 7A), which were reflected at the protein level as well (Fig. 7B). Besides, as reported earlier (26), we also observed higher levels of c-Fos in the transgenic liver samples as compared with control (Fig. 7B). HBx is well known to activate the MAPK signaling pathway, which in turn increases the levels of Jun/Fos proteins (27). Because our cell culture studies had already established that p27^Kip1 levels are regulated by the MAPK pathway and are dependent on Jun/Fos levels, we checked the effect of in vivo inhibition of the MAPK pathway by intravenous injection of PD98059 in transgenic mice. Western blot and real time PCR analysis of the liver samples 6 h post-injection showed a significant increase in p27^Kip1 expression both at RNA and protein levels (Fig. 7, C and D). As expected, there was a decline in phospho-ERK levels following inhibitor treatment, whereas the total ERK levels remained unaffected. However, there was a marginal decline in c-Fos levels possibly due to inhibition of MAPK activity, reiterating our earlier observation that c-Fos negatively regulates p27^Kip1 transcription.

FIGURE 7.

In vivo regulation of the p27^Kip1 gene in the liver of control and X15-myc transgenic mice. A, real time PCR analysis of p27^Kip1 mRNA expression. B, Western blot (WB) of p27^Kip1, c-Jun, c-Fos, HBx, and GAPDH. C, real time PCR analysis of p27^Kip1 mRNA levels in transgenic mice 6 h post treatment with phosphate-buffered saline (PBS) or PD98059. D, WB of p27^Kip1, c-Jun, c-Fos, pERK, total ERK, and GAPDH after treatment as in C. *, statistically significant difference at p < 0.05.

DISCUSSION

Cell division cycle is driven by sequential activation of cyclin-cyclin-dependent kinase complexes, and the cyclin/cyclin-dependent kinase activity is controlled by CKI. Thus, regulation of CKI levels may be an important step that controls cell division or cell death. p27^Kip1 is well known to function as a negative regulator of the cell cycle by inhibiting cyclin-dependent kinase activity. Although mitogens are well known to down-regulate p27^Kip1 levels that propel cells toward cell division, the negative signals such as DNA damage and differentiation factors cause accumulation of p27^Kip1 leading to cell cycle arrest or apoptosis. p27^Kip1 levels are known to be regulated primarily through post-translational mechanisms. However, some recent reports accentuate the role of transcriptional regulation in fine-tuning its levels. For example, mitogenic stimuli down-regulate p27^Kip1 mRNA levels in endothelial, muscle, and B cells (28–30). In contrast, anti-proliferative signals such as cells treated with anti-IgM or interleukin-starved lymphocytes lead to accumulation of p27^Kip1 mRNA (13, 31). Working on the mechanism of transcriptional regulation of p27^Kip1 gene in quiescent cells stimulated with growth factors, we report that the decline in p27^Kip1 mRNA levels following mitogenic stimulation (serum and EGF) is mediated by AP-1 family of transcription factors.

Our studies on the expression of the p27^Kip1 gene by real time PCR indicated that its mRNA levels were rapidly down-regulated in the presence of growth factors such as serum or recombinant EGF. It seems that a similar regulatory mechanism is also operative in nontransformed cells as mitogenic stimulation of AML12 cells, an immortalized mouse hepatocytic cell line, showed a similar decline curve for p27^Kip1 mRNA levels as observed in case of human hepatoma Huh7 cells. Interestingly, enforced expression of growth-promoting factors like viral oncoprotein HBx and cellular c-Myc in a transgenic environment also resulted in a marked decline in the levels of p27^Kip1 mRNA and protein. Thus, down-regulation of p27^Kip1 levels by growth factors seems to be a built-in mechanism that allows the cell cycle to proceed (32).

The deletion analysis of the p27^Kip1 promoter suggested the presence of mitogen-responsive element(s) within the −571-bp region, whereas the TransFac analysis identified binding sites for important transcription factors like AP-1, SP1, and NF-Y. Because SP1 and NF-Y are reported to bind constitutively to the p27^Kip1 promoter and stimulate its basal transcription (10), it was unlikely that the two factors would regulate the promoter under mitogenic stimulation. On the other hand, it is well established that AP-1 binding to DNA is rapidly induced by growth factors, cytokines, and oncoproteins resulting in proliferation, differentiation, and/or transformation of cells (33). Besides AP-1 proteins are considered important regulators of early G₁ phase in the cell cycle. Although the TransFac search predicted two AP-1-binding sites (at positions −2242 and −469) in the p27^Kip1 promoter, the distal element seemed dispensable for mitogen responsiveness. Accordingly, we observed that mutation in the proximal AP-1 element abrogated the regulatory control of growth factors. Further decline in p27^Kip1 mRNA levels coincided with an increase in the AP-1 activity. EMSA and ChIP studies also confirmed the binding of Jun/Fos heterodimers to the p27^Kip1 promoter. RNA polymerase II was also found to be recruited but in both quiescent and mitogen-stimulated conditions suggesting their assembly in the basal transcription apparatus even in a repressed state. This is not surprising because many mammalian genes, prior to their expression, are reported to show RNA polymerase II binding in the proximal regions of promoters in a “stalled” state. The activation of such stalled polymerases is suggested to be a mechanism for expression of these genes (34). Thus, recruitment of AP-1 proteins adjacent to the transcription start sites may be involved in keeping the transcription apparatus in a repressed state. Because many genes involved in cell cycle regulation (e.g. cyclin D) and DNA synthesis (e.g. proliferating cell nuclear antigen) are known to carry AP-1-binding sites in their promoters (35, 36), this study also indicates a connection between AP-1 transcription factors and such regulators of the cell cycle.

Because AP-1 proteins bind to the p27^Kip1 promoter, the involvement of selective members of AP-1 family was further investigated following their enforced expression. We observed that c-Jun was able to up-regulate p27^Kip1 expression in serum-fed cycling cells. Although c-Fos had no effect in asynchronous cells, it was able to repress in quiescent cells. Such a differential effect may relate to higher basal AP-1 activity in asynchronous cells that may keep the promoter in a repressed state, although it may be very low in quiescent cells. Thus, in quiescent cells where p27^Kip1 promoter is active, overexpression of c-Fos could down-regulate the transcription. Similar to this observation, Güller et al. (37) recently reported that Fos overexpression can cause an increase in p27^Kip1 protein levels but does not affect p27^Kip1 transcription in immortalized human hepatocytes. However, they found these changes in asynchronously growing cells as also observed by us. This type of Jun/Fos cooperation where Jun behaves as an activator while Fos titrates this activity has been reported earlier in the case of phosphoenolpyruvate carboxykinase gene regulation (38). The regulatory role of AP-1 was further evident from the fact that mitogen-dependent p27^Kip1 down-regulation could be abrogated by overexpressing Fos-DN or RNA interference against c-Fos and c-Jun by specific siRNAs. Interestingly, Fos-DN was able to significantly up-regulate p27^Kip1 transcription in response to mitogens suggesting that in the absence of the AP-1 complex other positive regulatory factors might be driving p27^Kip1 transcription. However, presently it is not clear whether Fos interacts with other transcription factors/repressors to inhibit p27^Kip1 transcription. It is possible that Fos may bind to such protein(s) in a manner that would prevent subsequent DNA binding or alter the levels or activities of transcription factors available for binding to the p27^Kip1 promoter. Besides, Fos may also function as an adaptor to modify the function of pre-existing factors.

AP-1 is a well known downstream target of the RAF-MEK-ERK pathway. Therefore, our studies on the inhibition of this pathway on AP-1 binding and regulation of the p27^Kip1 promoter showed a dramatic rescue of p27^Kip1 expression with a concomitant reduction in Jun/Fos binding to AP-1 elements. A similar increase in the abundance of p27^Kip1 mRNA and improved stability of the protein has been reported following pharmacological inhibition of MAPKs (39). PI3K-dependent Akt signaling has also been implicated as a regulator of p27^Kip1 levels (9). Furthermore, platelet-derived growth factor-dependent DNA synthesis in response to Akt kinase is known to involve the expression of p27^Kip1 and c-Fos (40). In agreement with this, we observed that PI3K inhibition also abrogated the mitogen-dependent inhibition of p27^Kip1 transcription. Akt kinase is known to inactivate FOXO transcription factors by nuclear exclusion. Furthermore, overexpression of FOXOs is known to stimulate the p27^Kip1 promoter to induce growth arrest of cells (9). However, the Akt activation/FOXO inactivation model is unlikely to be applicable here because there are no FOXO-binding sites in the minimal promoter used by us that were responsive to mitogenic stimuli. Besides, this model cannot reconcile for the requirement of de novo synthesis of proteins following mitogenic stimulation. Interestingly, we found that inhibition of the PI3K pathway in serum-stimulated cells did not affect the Jun/Fos levels. However, in the EGF-treated cells, a marginal decline in Jun/Fos binding was observed suggesting the existence of alternative mechanism(s) controlling the p27^Kip1 promoter. It is noteworthy that there was no synergistic effect of simultaneous inhibition of both MAPK and PI3K pathways on p27^Kip1 expression as reported earlier by others (25). Such variation in results possibly could arise due to differences in the time of inhibitor treatment in these experiments. Nonetheless, the involvement of both pathways in controlling p27^Kip1 mRNA expression is clear.

The viral oncoprotein HBx has been reported to promote cell cycle progression by destabilizing the p27^Kip1 protein (18). Furthermore, it is known to up-regulate c-Fos levels by stabilizing c-Myc (16). To study the involvement of AP-1 and p27^Kip1 in tumorigenesis in the X15-myc transgenic mice, we analyzed the levels of two proteins in the liver of transgenic versus control animals. We observed a significant decline in levels of p27^Kip1 protein and mRNA and a marked increase in the levels of Jun and Fos proteins in the transgenic liver thus reiterating the involvement of Jun/Fos in regulating p27^Kip1 expression. Interestingly, MAPK inhibition resulted in a dramatic increase in p27^Kip1 expression with a concomitant decline in Jun/Fos levels substantiating our earlier findings that AP-1 proteins negatively regulate p27^Kip1 gene expression during cell cycle.

Thus, this study has been able to unveil a novel role of AP-1 in cell cycle progression via regulating the levels of p27^Kip1. It will be interesting to investigate the role played by other transcription factors during this molecular concert. As AP-1 transcription factors are well known targets in tumor development, it may be a promising target for cancer therapy (41).