Platelet-derived growth factor (PDGF) receptor-alpha-activated c-Jun NH2-terminal kinase-1 is critical for PDGF-induced p21WAF1/CIP1 promoter activity independent of p53.

Platelet-derived growth factor (PDGF) is a potent mitogen for mesenchymal cells. PDGF AA functions as a "competent factor" that stimulates cell cycle entry but requires additional (progression) factors in serum to transit the cell cycle beyond the G1/S checkpoint. Unlike PDGF AA, PDGF B-chain (c-sis) homodimer (PDGF BB) and its viral counterpart v-sis can serve as both competent and progression factors. PDGF BB activates alpha- and beta-receptor subunits (alpha-PDGFR and beta-PDGFR) and induces phenotypic transformation in NIH 3T3 cells, whereas PDGF AA activates alpha-PDGFR only and fails to induce transformation. We showed previously that alpha-PDGFR antagonizes beta-PDGFR-mediated transformation through activation of stress-activated protein kinase-1/c-Jun NH2-terminal kinase-1, whereas both alpha-PDGFR and beta-PDGFR induce mitogenic signals. These studies revealed a striking feature of PDGF signaling; the specificity and the strength of the PDGF growth signal is modulated by alpha-PDGFR-mediated simultaneous activation of growth stimulatory and inhibitory signals, whereas beta-PDGFR mainly induces a growth-promoting signal. Here we demonstrate that PDGF BB activation of beta-PDGFR alone results in more efficient cell cycle transition from G1 to S phase than PDGF BB activation of both alpha-PDGFR and beta-PDGFR. PDGF AA activation of alpha-PDGFR or PDGF BB activation of both alpha- and beta-PDGFRs up-regulates expression of p21WAF1/CIP1, an inhibitor of cell cycle-dependent kinases and a downstream mediator of the tumor suppressor gene product p53. However, beta-PDGFR activation alone fails to induce p21WAF1/CIP1 expression. We also demonstrate that alpha-PDGFR-activated JNK-1 is a critical signaling component for PDGF induction of p21WAF1/CIP1 promoter activity. The ability of PDGF/JNK-1 to induce p21WAF1/CIP1 promoter activity is independent of p53, although the overall p21WAF1/CIP1 promoter activities are greatly reduced in the absence of p53. These results provide a molecular basis for differential regulation of the cell cycle and transformation by alpha- and beta-PDGFRs.

Although a number of signaling molecules activated by PDG-FRs have been identified, little is known about which signaling pathways are ␣or ␤-PDGFR-specific, and how these signal transduction pathways are agonized or antagonized to regulate PDGF isoform-specific cellular processes. To investigate ␣and ␤-PDGFR-specific pathways, we previously established NIH 3T3 clones in which ␣-PDGFR signaling is inhibited by a dominant-negative ␣-PDGFR or an antisense construct of ␣-PDGFR. By using these clones, we showed that ␣-PDGFR, but not ␤-PDGFR, activates stress-activated protein kinase-1/ c-Jun NH 2 -terminal kinase-1 (SAPK1/JNK-1), whereas both ␣and ␤-receptors effectively activate ERKs. Inhibition of ␣-PDGFR-activated JNK-1 enhanced PDGF BB-mediated phenotypic transformation, revealing a part of PDGFR subunitspecific signal transduction pathways critical for regulation of cell growth and transformation (10).
In this study, we further investigated ␣and ␤-PDGFRspecific regulation of cell growth signaling. PDGF was shown to function with serum-starved cells as a "competence" factor to induce expression of a set of immediate-early response genes critical for cell cycle entry (11)(12)(13)(14)(15). For PDGF-stimulated cells to progress efficiently beyond the G 1 /S checkpoint and transit the cell cycle and proliferate, progression factors in serum such as insulin and insulin-like growth factor I are required. The significance of cell cycle-dependent kinase inhibitors in the regulation of cell cycle checkpoints and transformation has been well documented. Among those, p21 WAF1/CIP1 has been shown to be critical for tumor suppressing activity and cell cycle arrest at the G 1 /S checkpoint (16 -19). p21 WAF1/CIP1 was originally identified as a p53-inducible gene (WAF1) (20), inde-pendently as an inhibitor of cyclin-dependent kinases (CIP1) (21,22), and also as a gene that is differentially expressed during cellular senescence (SDI1) (23). Interestingly, growth factors such as PDGF and basic fibroblast growth factor (bFGF), but not insulin, were shown to induce transcription of p21 WAF1/CIP1 independent of p53 (24). Currently, neither the PDGF-signaling pathways leading to p21 WAF1/CIP1 induction nor its significance in the regulation of cell cycle/transformation is clearly understood.
Here we show that inhibition of ␣-PDGFR signaling results in a more efficient cell cycle transition from the G 1 to S phase. PDGF AA activation of ␣-PDGFR or PDGF BB activation of both ␣and ␤-PDGFRs mediates up-regulation of p21 WAF1/CIP1 expression. However, ␤-PDGFR activation alone fails to induce p21 WAF1/CIP1 expression indicating that ␣-PDGFR signaling is critical for the regulation of p21 WAF1/CIP1 expression. We demonstrate that ␣-PDGFR-activated JNK-1 is a critical signaling component for the PDGF induction of p21 WAF1/CIP1 promoter activity, providing a molecular basis for ␣-PDGFR-mediated growth inhibitory signal. We also provide evidence that the ability of PDGF/JNK-1 to induce p21 WAF1/CIP1 promoter activity is independent of p53, although the overall p21 WAF1/CIP1 promoter activities are greatly reduced in the absence of p53.

EXPERIMENTAL PROCEDURES
Cell Culture and Antibodies-NIH 3T3 cells were cultured in a humidified 5% CO 2 incubator with Dulbecco's modified Eagle's medium/ F12 nutrient media (DMEM/F12) containing 10% bovine calf serum, 2 mM glutamine, 100 units/ml penicillin, 100 mg/ml streptomycin, 250 g/ml amphotericin B, and 205 g/ml sodium deoxycholate (Invitrogen). Serum-deprived subconfluent cells were treated with DMEM/ F12 containing different concentrations of PDGF or bFGF. Protein or RNA samples were collected at different time intervals. Anti-active JNK antibody was purchased from Promega (Madison, WI), anti-p21 WAF1/CIP1 mAb from Oncogene, anti-␤-actin antibody from Sigma, anti-phosphotyrosine antibody from Oncogene Research Product (Cambridge, MA), anti-␣ PDGFR from Santa Cruz Biotechnology (Santa Cruz, CA), and Ab 262 was obtained as described previously (25). PDGF was purchased from Oncogene Research Product and bFGF from Invitrogen.
Northern Blot Analysis-Total RNA was isolated with TRIzol Reagent according to the manufacturer's protocol (Invitrogen). Total RNA (20 g) was resolved in a 1% agarose gel containing 2.2 M formaldehyde, transferred to the nylon membrane using the Turboblotter® system (Schleicher & Schuell), and cross-linked to the nylon membranes using an ultraviolet cross-linker at a dose of 120 mJ/cm 2 (Stratagene, La Jolla, CA). [␣-32 P]dCTP-labeled DNA probes were synthesized using a random priming DNA labeling method. Blots were hybridized to radioactively labeled p21 WAF1/CIP1 probe. After stripping, the blots were further hybridized with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe to control RNA loading.
Immunoblot Analysis-Cells were lysed in RIPA buffer (150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate acid, 0.5% Nonidet P-40, 10 mM Tris, pH 7.4, 1 mM EDTA, 2 mM sodium vanadate, 2 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin) or SDS lysis buffer (2% SDS, 125 mM Tris-HCl, pH 6.8, and 20% glycerol), and protein concentration was determined using the BCA Protein Assay kit from Pierce. Equal amounts of protein in each sample were separated by SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were subjected to 1 h of blocking with 5% non-fat milk in T-TBS (0.02% Tween 20, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl), followed by incubation with primary antibodies in TTBS containing 5% non-fat milk for 1 h. After three washes with T-TBS, the blot was incubated with the appropriate horseradish peroxidase-conjugated secondary antibody. The antigen was detected using the ECL detection system (Pierce) according to the manufacturer's instruction.
Promoter Activity Assay-The luciferase reporter construct of p21 WAF1/CIP1 promoter (WWP-Luc) was provided by Dr. B. Vogelstein (20). The Ϫ2327-p21P, Ϫ194-p21P, Ϫ64-p21P, and Ϫ2327(D-127-64)p21P plasmids were obtained from Dr. X. F. Wang (26). Subconfluent cells growing in a 6-or 12-well plate were co-transfected with p21 WAF1/CIP1 promoter luciferase reporter constructs, ␤-galactosidase expression plasmid (pMDV-LacZ), and other constructs as specified in the figure legends. After 16 -20 h of transfection, cells were cultured in serum-free medium for 36 h and treated with 25 ng/ml PDGF AA or PDGF BB for 3 h. Cells were washed with ice-cold phosphate-buffered saline and lysed in 200 l of reporter lysis buffer (Promega). Luciferase and ␤-galactosidase activities were assayed using a luciferase assay kit (Promega) and the Galacto-Light kit (Applied Biosystems, Bedford, MA), respectively, and were measured subsequently with Turner TD-20e (Promega) or Fusion System (Packard) luminometer. The luciferase activity was normalized to ␤-galactosidase activity to control the transfection efficiency. Data represent at least two independent experiments performed in triplicate. The luciferase activity in serum-starved cells was arbitrarily given as 1; PDGF-induced p21 WAF1/CIP1 promoter activity was calculated as fold increase in luciferase activity of growth factor-treated cells as compared with that of serum-starved cells.
Real Time Quantitative PCR Analysis-Total RNA was extracted from the cells using Trizol reagent (Invitrogen). Five or 10 g of each RNA sample was reverse-transcribed to first strand cDNA using an oligo(dT) primer (Invitrogen). Equal amounts of the reverse transcription products were applied to real time quantitative PCR (Q-PCR) analysis using p21 WAF1/CIP1 -specific primer pairs and the intercalating dye SYBR Green I, as recommended by the manufacturer (Stratagene). The real time Q-PCR was performed on a Bio-Rad iCycler. The PCR cycling profile was as follows: 10 min of hot start at 95°C; 35 cycles of denaturation at 95°C for 60 s, annealing at 60°C for 60 s, and extension at 72°C for 60 s. Fluorescent reading was captured at the end of the extension step. GAPDH primer pairs were used for normalization in separate real time Q-PCRs. At least triplicate reactions were performed for each sample and primer. The relative gene expression was calculated in reference to the expression level of the control cells without PDGF treatment. The PCR primer sets used were as follows: p21 WAF1/CIP1 -specific forward primer, 5Ј-GTACTTCCTCTGCC-CTGCTG-3Ј, p21 WAF1/CIP1 -specific reverse primer, 5Ј-CACAGAGTGAG-GGCTAAGGC-3Ј; and GAPDH forward primer, 5Ј-ATCACCATCTTCC-AGGAGCGA-3Ј, GAPDH reverse primer, 5Ј-GCCAGTGAGCTTCCCG-TTCA-3Ј.
Flow Cytometric Analysis-Subconfluent cells were synchronized at G 0 by culturing in serum-free DMEM/F12 medium for 24 h. Cells were treated with DMEM/F12 containing 0, 10, or 25 ng/ml PDGF BB. At a 16-h interval, the cells were collected by trypsinization, washed with 1ϫ PBS, and fixed in 67% ethanol for 24 h at 4°C. The nuclei were stained with propidium iodine at 20 g/ml for 30 min, and the DNA content of the cells was determined using a FACScan (BD Biosciences). Twenty thousands cells were counted, and the percentage of cells in each phase of the cell cycle was analyzed by Modfit LT Software (Verity Software House, Topsham, ME).

␣-PDGFR Is Critical for PDGF-induced p21 WAF1/CIP1 Expression, and Cell Cycle Transition from the G 1 to S Phase-To
investigate the ␣and ␤-PDGFR specific signaling and their functions, we previously established two separate NIH 3T3 clones in which ␣-PDGFR activation is inhibited. The first approach was to down-regulate ␣-PDGFR expression by using an antisense construct of ␣-PDGFR cDNA (AS ␣-PDGFR). The second approach was to prevent activation of ␣-PDGFR by using a dominant-negative mutant of ␣-PDGFR that contains the extracellular and transmembrane domains but lacks the cytoplasmic kinase domains (DN ␣-PDGFR). The ␣-PDGFR expression was significantly down-regulated by AS ␣-PDGFR (Fig. 1A), and PDGF AA-mediated autophosphorylation of endogenous ␣-PDGFR was inhibited by DN ␣-PDGFR (Fig. 1B). As shown in Fig 1C, PDGF BB activation of ␣-PDGFR was also inhibited in AS and DN clones. This confirmed the significant inhibition of ␣-PDGFR activation in AS and DN clones as reported previously (10).
By using AS, DN clones, and control NIH 3T3 cells, we examined whether ␣and ␤-PDGFRs differentially regulate expression of p21 WAF1/CIP1 , a cell cycle inhibitor and a downstream mediator of tumor suppressing activity of p53 (16 -19). The effects of PDGF AA activation of ␣-PDGFR alone and PDGF BB activation of both ␣and ␤-PDGFRs were examined in control NIH 3T3 cells that express both ␣and ␤-PDGFRs at comparable levels. The effect of PDGF BB activation of ␤-PDGFR alone was determined in AS and DN clones (10).
Northern blot analysis showed that PDGF AA activation of ␣-PDGFR in control NIH 3T3 cells up-regulated p21 WAF1/CIP1 mRNA expression, whereas no significant increase in p21 WAF1/CIP1 mRNA levels was detected in AS and DN clones following PDGF AA treatment (Fig. 1D). PDGF BB activation of both ␣and ␤-PDGFRs effectively induced p21 WAF1/CIP1 mRNA expression in control NIH 3T3 cells (Fig. 1E). In contrast, when ␤-PDGFR alone was activated by PDGF BB in AS and DN clones, p21 WAF1/CIP1 induction was significantly reduced ( Fig. 1E).
To ensure that down-regulation of PDGF-induced p21 WAF1/CIP1 expression resulted from the inhibition of ␣-PDGFR signaling in the AS and DN clones and not from lack of response to growth factors, bFGF-induced p21 WAF1/CIP1 expression was examined. The bFGF-induced p21 WAF1/CIP1 expression levels were comparable among control NIH 3T3, DN, and AS clones (Fig. 1F). These results indicate a critical role for ␣-PDGFR in PDGF-induced p21 WAF1/CIP1 expression.
PDGF AA-and BB-induced p21 WAF1/CIP1 mRNA expression was quantitated by real time PCR analysis (Fig. 1G). PDGF AA and BB induced p21 WAF1/CIP1 mRNA levels ϳ2.7and ϳ3.2fold, respectively, in control NIH 3T3 cells. In DN and AS clones, PDGF AA, as expected, failed to activate p21 WAF1/CIP1 expression due to lack of ␣-PDGFR activation. When the effect of PDGF BB activation of ␤-PDGFR alone was examined in and antisense clone (AS6) without PDGF treatment were immunoprecipitated (IP) with an anti-␣-PDGFR polyclonal antibody (Ab 262). The immunoprecipitates were subjected to immunoblot (IB) analysis with an anti-␣-PDGFR mAb (Santa Cruz Biotechnology). B, after incubation with 50 ng/ml PDGF AA for 5 min, serumdeprived control and dominant-negative clone (DN16) were lysed in RIPA buffer. Lysates (250 g/lane) were immunoprecipitated with an anti-␣-PDGFR polyclonal Ab (Ab 262), and the immunoprecipitates were subjected to immunoblot analysis with an anti-phosphotyrosine mAb (Oncogene). C, serum-deprived cells (Ctrl, AS6, and DN16) without or with 50 ng/ml PDGF BB treatment for 5 min were lysed in RIPA buffer. Lysates (100 g/lane) were immunoprecipitated with an anti-␣-PDGFR polyclonal Ab (Ab 262), and the immunoprecipitates were subjected to immunoblot analysis with an anti-phosphotyrosine mAb. D and E, serum-deprived cells (Ctrl, AS6, and DN16) were treated with DMEM/F12 containing 0 and 50 ng/ml PDGFAA (D) or 25 ng/ml PDGF BB (E) for 2 h. Then RNA was extracted using Trizol reagent. Twenty micrograms of each RNA sample was resolved on a 1% denaturing agarose gel and subjected to Northern blot analysis with radiolabeled p21 WAF1/CIP1 cDNA probe. The identical blot was probed with radiolabeled GAPDH cDNA probe as a loading control of RNA samples. F, serum-deprived cells (Ctrl, AS6, and DN16) were treated with DMEM/ F12 containing 0 and 100 ng/ml bFGF for 2 h. Then RNA was extracted by using Trizol reagent. Twenty micrograms of each RNA sample was subjected to Northern blot analysis with p21 WAF1/CIP1 probe. G, serumdeprived cells (Ctrl, AS6, and DN16) were incubated with DMEM/F12 containing 0 and 25 ng/ml PDGF AA or PDGF BB for 2 h. Total RNA was extracted, and 5 g of each RNA sample was reverse-transcribed to first strand cDNA using oligo(dT) as primer (Invitrogen). The real time Q-PCRs was carried out as specified under "Experimental Procedures." The p21 WAF1/CIP1 mRNA level in the untreated cells was arbitrarily given as 1, and error bars represent the S.D. of the mean of triplicate samples. H and I, serum-deprived cells (Ctrl, AS6, and DN16) were incubated with DMEM/F12 containing 0, 10, 50, and 100 ng/ml PDGF AA or PDGF BB for 8 h. Lysates (50 g/lane) were resolved by reducing SDS-PAGE, followed by immunoblot analysis with an anti-p21 WAF1/CIP1 Ab. To compare relative intensity of the bands between the blots, 50 g of untreated control NIH 3T3 lysates (Ctrl, 0 ng/ml) was included in each blot (asterisk). these clones, only a marginal induction of p21 WAF1/CIP1 mRNA levels was detected. It should be noted that PDGF BB treatment of AS clones resulted in detectable increase in p21 WAF1/ CIP1 mRNA levels (ϳ1.3-fold). We speculate that these cells express residual amounts of ␣-PDGFR, and mild activation of the ␣-PDGFR signaling pathway results from ␣␤-PDGFR heterodimerization following PDGF BB treatment. Down-regulation of PDGF AA-and BB-induced p21 WAF1/CIP1 protein expression in AS and DN cells was also confirmed by immunoblot analysis (Fig. 1, H and I).
Next we asked whether ␣and ␤-PDGFRs differentially regulate cell cycle transition from the G 1 to S phase. To answer this question, we determined what percentages of control NIH, AS, and DN cells were in each phase of the cell cycle following activation of ␤-PDGFR alone or both ␣and ␤-PDGFRs (Table  I). Flow cytometric analysis revealed that cells entered into the S phase more efficiently following ␤-PDGFR activation in the absence of ␣-PDGFR signaling: ϳ35% of AS and ϳ50% DN cells were in the S phase at 16 h following 25 ng/ml PDGF BB activation of ␤-PDGFR alone, whereas ϳ25% of NIH 3T3 cells were in the S phase following PDGF BB activation of both ␣and ␤-PDGFRs. These data suggest that ␣-PDGFR signaling is critical for PDGF to function as a bona fide "competent growth factor," which has the ability to halt the cell cycle transition at the G 1 /S checkpoint.
␣-PDGFR Is Critical for PDGF-induced p21 WAF1/CIP1 Promoter Activity-To determine whether ␣-PDGFR-induced p21 WAF1/CIP1 expression occurs at the transcriptional level, we measured p21 WAF1/CIP1 promoter activity using a luciferase reporter gene under the control of a 2.4-kb p21 WAF1/CIP1 promoter (WWP-Luc, kindly provided by Dr. B. Vogelstein). The p21 WAF1/CIP1 promoter activity was significantly enhanced in the control NIH 3T3 cells following PDGF AA or PDGF BB treatment (1.6 -1.8-fold) ( Fig. 2A). Neither PDGF AA nor PDGF BB, however, induced p21 WAF1/CIP1 promoter activity in DN and AS clones (Fig. 2, B and C). To exclude the possibility Lysates (30 g/lane) were resolved by reducing SDS-PAGE, followed by immunoblot analysis with an anti-p21 WAF1/CIP1 mAb and anti-␤-actin mAb. The p21 WAF1/CIP1 protein levels were normalized to the ␤-actin level, and fold induction of p21 WAF1/CIP1 protein by PDGF BB treatment is shown below. The p21 WAF1/CIP1 protein level in control NIH 3T3 cells without PDGF treatment was arbitrarily given as 1.
that the lack of PDGF-induced p21 WAF1/CIP1 promoter activity is associated with inherent clonal defects in p21 WAF1/CIP1 regulation, the bFGF-induced p21 WAF1/CIP1 promoter was also examined. As shown in Fig. 2D, induction of p21 WAF1/CIP1 promoter activity was comparable among control NIH 3T3, DN, and AS clones following bFGF treatment. This supports the conclusion that the lack of PDGF BB activation of p21 WAF1/CIP1 promoter in DN and AS clones resulted from lack of ␣-PDGFR signaling.
We then examined whether PDGF induction of p21 WAF1/CIP1 promoter activity can be restored by enhanced JNK-1 activation in DN and AS clones. When JNK-1 activation was enhanced by wild-type JNK-1 overexpression, PDGF-induced p21 WAF1/CIP1 promoter activity was significantly induced in both AS and DN clones (Fig. 3, G and H), re-emphasizing the significant role of JNK-1 in PDGF regulation of p21 WAF1/CIP1 promoter activity.
A Region between Ϫ127 and Ϫ64 of the p21 WAF1/CIP1 Promoter Mediates PDGF/JNK-1 Activation of the p21 WAF1/CIP1 Promoter, Independent of the Tumor Suppressor Gene Product p53-p21 WAF1/CIP1 is a downstream mediator of the tumor suppressing activity of p53, as its expression was shown to increase rapidly following DNA damage in a p53-dependent manner (30). To examine whether PDGF-induced p21 WAF1/ CIP1   FIG. 4. A region between ؊127 and ؊64 of the p21 WAF1/CIP1 promoter mediates PDGF/JNK-1 activation of the p21 WAF1/CIP1 promoter, independent of the tumor suppressor gene product p53. A and B, serum-deprived (48 h) MEF p53Ϫ/Ϫ cells were treated with serum-free MEM containing 0, 25, and 100 ng/ml PDGF AA (A) or PDGF BB (B) for 2 h, respectively. Ten micrograms of each RNA sample was reverse-transcribed to first strand cDNA using oligo(dT) as primer (Invitrogen). The Q-PCRs and calculations for the p21 WAF1/CIP1 mRNA levels were carried out as specified under the "Experimental Procedures." C, D and F, G, control NIH 3T3 cells grown in 12-well plate were transiently transfected with 0.6 g of p21 WAF1/CIP1 promoter-luciferase reporter construct (Ϫ2327-p21P, Ϫ194-p21P, Ϫ64-p21P, or Ϫ2327(D-127-64)p21P), 0.1 g of pMDV-LacZ, and 0.1 g of pcDNA3-wt-JNK1-Flag plasmid (or 0.1 g of pcDNA3 as mock JNK1 transfection). After 16 -20 h of transfection, the transfected cells were serum-deprived for 36 h, followed by 0 and 25 ng/ml PDGF AA (C and F) or PDGF BB (D and G) treatment for 3 h, respectively. The luciferase activity was normalized to ␤-galactosidase activity. The white bars represent luciferase activity with neither PDGF treatment nor wild type (wt)-JNK-1 transfection, and the gray bars with PDGF treatment without wt-JNK-1 transfection, and the black bars with both PDGF treatment and wt-JNK-1 transfection. Error bars represent S.D. of the mean of triplicate samples (A-D, F, and G). E, a schematic representation of p53-and Sp1-responsive elements in ϳ2.4 kb of the human p21 WAF1/CIP1 promoter region upstream of the transcription initiation site. Ϫ2327-p21P, Ϫ194-p21P, and Ϫ64-p21P contain 2327, 194, and 64 bp of the human p21 WAF1/CIP1 promoter region upstream of the transcription initiation site, respectively. The Ϫ194-p21P does not contain the p53-responsive elements but has six putative Sp1 sites. The construct, Ϫ64-p21P, harbors neither p53-responsive element nor Sp1-binding consensus sequence. H, a schematic representation of a deletion mutant, Ϫ2327(D-127-64)p21P, in which Sp1 elements located between Ϫ127 and Ϫ64 bp were deleted. expression also depends on p53, we determined the p21 WAF1/CIP1 mRNA levels in the murine embryonic fibroblast (MEF) cell line established from p53-deficient mice (obtained from Dr. T. Jacks) with or without PDGF treatment. Real time quantitative PCR analysis showed that both PDGF AA and BB effectively up-regulated p21 WAF1/CIP1 mRNA levels in these cells, demonstrating p53-independent up-regulation of p21 WAF1/CIP1 expression by PDGF (Fig. 4, A and B).
To study further PDGF/JNK-1-mediated p21 WAF1/CIP1 promoter activation independent of p53, we transfected luciferase reporter constructs containing various lengths of the p21 WAF1/CIP1 promoter into NIH 3T3 cells. The Ϫ2327-p21P construct contains two p53-binding sites (located between Ϫ2282 and Ϫ2263 bp and between Ϫ1391 and Ϫ1361 bp) and one putative p53-binding site (between Ϫ1748 and Ϫ1713 bp) (31). Luciferase activities driven by the Ϫ2327-p21P construct was increased by both PDGF AA (Fig. 4C) and PDGF BB (Fig.  4D) treatment. Additionally, co-transfection of wild-type JNK-1 further enhanced PDGF AA-and BB-induced p21 WAF1/CIP1 promoter activities (Fig. 4, C and D). The basal level of luciferase activity driven by the Ϫ194-p21P construct lacking the p53-responsive elements and binding sites was significantly lower than the Ϫ2327-p21P construct driven activity (Fig. 4, C and D, white bars). The Ϫ194-p21P construct, however, still maintained PDGF responsiveness, as its activity was further increased by transfection with wild-type JNK-1 expression plasmids. The Ϫ194-p21P construct contains six promoter-specific transcription factor 1 (Sp1) elements, located between Ϫ118 and Ϫ65 bp. When the 130-bp segment of the proximal promoter containing Sp1 elements was deleted in Ϫ64-p21P, the basal level of p21 WAF1/CIP1 promoter activity was barely detectable. The Ϫ64-p21P construct containing 64 bp of the p21 WAF1/CIP1 promoter was unresponsive to the PDGF/JNK-1 signaling. To examine further whether Sp1 elements located between Ϫ118 and Ϫ65 bp of the p21 WAF1/CIP1 promoter is critical for PDGF induction of p21 WAF1/CIP1 promoter activity, PDGF responsiveness of Ϫ2327 (D-127-64)p21P, a Ϫ2327-p21P construct lacking Sp1 sites located between Ϫ127 and Ϫ64 bp, was examined. As shown in Fig. 4, F and G, the Ϫ2327 (D-127-64)p21P construct was unresponsive to both PDGF AA and BB in the absence or presence of transfected wild type JNK-1. These results indicate that cisacting regulatory elements (likely Sp1) that reside between Ϫ127 and Ϫ65 of the p21 WAF1/CIP1 promoter mediate PDGF/ JNK-1-induced p21 WAF1/CIP1 promoter activity independent of p53. DISCUSSION PDGF AA and BB are equally potent growth factors in triggering the cell cycle entry. However, only PDGF AA is a bona fide "competent factor," which requires a progression factor to complete the cell cycle transition, whereas PDGF BB can serve as both competent and progression factors (32)(33)(34)(35). Our previous and present studies may provide an explanation for PDGF AA-and BB-mediated differential regulation of cell cycle at the molecular levels. We showed previously that ␣-PDGFR-activated JNK-1 plays an antagonistic role for PDGF-induced transformation, and we proposed a model that activation of ␣-PDGFR transduces both positive and negative signaling for cell growth, whereas ␤-PDGFR mainly induces positive signaling. PDGF BB activation of both receptors shifts the balance of signaling to favor the transformation pathway, whereas PDGF AA activation of ␣-PDGFR alone does not. Consistent with this model, PDGF BB, but not PDGF AA, has transforming ability in murine fibroblast cells in vitro (36 -39). When ␣-PDGFRmediated negative signaling is inhibited by bcl-2, PDGF AA induces phenotypic transformation as effectively as PDGF BB (39), supporting our working model that ␣-PDGFR activates both positive and negative signaling for cell growth and transformation. The present study further reveals the molecular basis for ␣-PDGFR-mediated negative signaling. ␣-PDGFR activation of JNK-1 leading to p21 WAF1/CIP1 promoter activation plays a role, at least in part, for ␣-PDGF-mediated cell cycle regulation at the G 1 /S checkpoint (reviewed in Ref. 40).
␣-PDGFR-mediated simultaneous activation of both positive and negative signaling has also been demonstrated in cell migration (41). In smooth muscle cells, PDGF BB induces cell migration more effectively than PDGF AA. Inhibition of ␣-PDGFR by using a neutralizing monoclonal antibody to ␣-PDGFR enhanced ␤-PDGFR-induced smooth muscle cell migration (42). In fibroblast cells, however, both ␣and ␤-PDG-FRs promote cell migration with equal effectiveness, and their effects are additive (43). Thus, cell type and PDGFR subunitspecific intracellular signal transduction pathways are responsible for PDGF isoform-specific regulation of diverse cellular processes. Of particular importance, ␣-PDGFR-mediated agonistic and antagonistic activities for cell growth and motility provide a molecular basis for "fine-tuning" of PDGF signaling depending on the genetic background of the cells and additional extracellular factors (reviewed in Ref. 40). p21 WAF1/CIP1 plays a pivotal role in cell proliferation, differentiation, growth arrest, and apoptosis (17, 44 -46). Not surprisingly, its expression is modulated in response to a wide variety of external stimuli, such as radiation and DNA damage (30,47), growth factors including PDGF, FGF, transforming growth factor-␤, insulin-like growth factor I, and nerve growth factor (48 -50). An array of signaling pathways and transacting transcription factors regulate p21 WAF1/CIP1 transcription following these stimuli. Correspondingly, p21 WAF1/CIP1 promoter harbors several cis-acting elements to these external and internal factors. Among these p21 WAF1/CIP1 transcription regulatory elements are two p53-responsive elements located between nucleotides Ϫ2282 to Ϫ2263 and Ϫ1391 to Ϫ1361, which mediate radiation and viral infection-induced and -suppressed p21 WAF1/CIP1 transcription (30,51), respectively. The GC-rich Sp1 (a transcription factor)-binding elements (52) between nucleotides Ϫ110 and Ϫ65 were shown to mediate transforming growth factor-␤-induced human keratinocyte cell growth arrest (48). In addition to the p53 and Sp1 sites, many other cis-acting elements in the p21 WAF1/CIP1 promoter region, such as signal transducer and activators of transcription and CCAAT/enhancer-binding protein-␣, have been reported. Sequence analysis reveals multiple Ap-1, Ap-2, and c-Jun sites distributed throughout the p21 WAF1/CIP1 promoter region, yet their roles in p21 WAF1/CIP1 promoter activity remain to be elucidated.
The present study demonstrates that JNK-1 is a critical component for PDGF regulation of p21 WAF1/CIP1 promoter activity. JNK-1 along with ERK is a member of the MAPK family, a downstream effector of the growth factor receptor-activated Ras/Raf/MEK/MAPK pathway. There have been controversies as to how p21 WAF1/CIP1 transcription is regulated by the Ras/ Raf/MEK/MAPK pathway. In chondrocytes, the Ras/Raf/MEK/ ERK-responsive elements are localized to the distal p53-binding site of the p21 WAF1/CIP1 promoter (53) but not the proximal Sp1-binding GC-rich region. In NIH 3T3 cells, however, the ras signaling pathway activates p21 WAF1/CIP1 promoter via the proximal Sp1-rich region (54). Involvement of multiple rassignaling pathways for p21 WAF1/CIP1 promoter activation was also suggested. Whereas both Ras and MEK are required for TGF-induced p21 WAF1/CIP1 transcription, only MEK was shown to be critical for PDGF-induced p21 WAF1/CIP1 transcription in NIH 3T3 cells (55). A dominant-negative Ras mutant almost completely abolished ERKs activation but had little effect on PDGF induction of p21 WAF1/CIP1 . This study suggested that Ras-independent MEK activation leading to activation of MAPK family member(s) other than ERKs is essential for PDGF-induced p21 WAF1/CIP1 expression in NIH 3T3 cells (55). Consistently, the present study demonstrates that ␣-PDGFR/ JNK1 pathway is critical for PDGF induction of p21 WAF1/CIP1 transcription in NIH 3T3 cells.
JNK-1 responsive cis-acting regulatory elements reside between Ϫ127 and Ϫ64 of the p21 WAF1/CIP1 promoter containing six GC-rich Sp1-responsive elements, suggesting a role for Sp1 in ␣-PDGFR/JNK1-mediated p21 WAF1/CIP1 promoter activation. Consistent with our results, a previous study showed that c-Jun, a substrate of SAPK1/JNK1, mediates p53-independent p21 WAF1/CIP1 promoter activation by physical interaction with the Sp1 protein (56). It has been well documented that Sp1 cooperates with other transcription factors such as Sp3 and p53 for the regulation of p21 WAF1/CIP1 transcription. Sp3, a member of the Sp1 family, shares the same basic binding motif but counteracts the effect of Sp1 (57) for the regulation of p21 WAF1/CIP1 transcription. In contrast, p53 interacts with Sp1 and synergistically trans-activates the p21 WAF1/CIP1 promoter, suggesting a transcription complex formation between the proximal Sp1 and distal p53 sites (58). By taking these together with our results, we propose that ␣-PDGFR/JNK1 enhances p21 WAF1/CIP1 promoter activity through Sp1-responsive element independent of p53, but maximal induction of p21 WAF1/CIP1 expression requires cooperation between Sp1 and p53.