p73 Independent of c-Myc Represses Transcription of Platelet-derived Growth Factor β-Receptor through Interaction with NF-Y*

We recently reported that c-Myc represses the transcription of platelet-derived growth factor (PDGF) β-receptor (Izumi, H., Molander, C., Penn, L. Z., Ishisaki, A., Kohno, K., and Funa, K. (2001) J. Cell Sci. 114, 1533–1544). We demonstrate here that the p53 family protein p73α represses PDGF β-receptor transcription essentially by the same mechanism. p73α but not p73β or p53 represses the transcription in concordance with its ability to bind NF-YC and NF-YB. None of other p73 isoforms (i.e. p73β, p73γ, p73ε), C-terminal deletion mutants of p73α, and p53 is able to bind NF-Y with the exception of p63α. This finding suggests that the sterile α-motif domain present only in p73α and p63α is the interaction site. For the repression, the N-terminal transactivation domain of p73α is also indispensable, arguing for the importance of the activity of p73α in the mechanism. p73α binds the C-terminal HAP domain of NF-YC previously found to be the interaction site with c-Myc and TBP. Because c-Myc induces and activates p73α (Zaika, A., Irwin, M., Sansome, C., and Moll, U. M. (2001)J. Biol. Chem. 276, 11310–11316) and they bind each other (Uramoto, H., Izumi, H., Ise, T., Tada, M., Uchiumi, T., Kuwano, M., Yasumoto, K., Funa, K., and Kohno, K. (2002) J. Biol. Chem. 277, in press), we examined whether the repression by p73 is dependent on c-Myc. However, Myc-null rat fibroblasts are also susceptible to p73α-induced repression. Serum stimulation of NIH3T3 cells gradually decreased the amount of endogenous NF-Y binding to the PDGF β-receptor promoter, whereas NF-YA expression in the nuclear extracts remains unchanged. Our results indicate that serum stimulation induces c-Myc and p73α, leading to the down-regulation of PDGF β-receptor expression by repressing its transcription.

endowed with a split tyrosine kinase in the intracellular domains, which upon ligand binding, dimerize and become autophosphorylated to activate the downstream signal cascades. Thus, it is essential for cells that the receptors are activated at right time by refined control mechanisms. The mechanism of activation has been extensively studied, whereas knowledge regarding the control mechanism to cease activation is still limited. Both of these mechanisms can be disturbed to maintain the active status in cancer cells.
In normal cells, the PDGF ␤-receptor expression decreases rapidly after stimulation by PDGF, which is an important feedback mechanism to prevent further activation of the cells leading to uncontrolled proliferation. The attenuation on the PDGF signaling occurs at multiple levels with both short and long term effects. The receptors on the cell surface decrease by ligand-induced receptor-mediated endocytosis (8) as well as by ubiquitination and subsequent degradation (9). As soon as the ligands diminish or cells become confluent or differentiated, tyrosine phosphatases can be activated, reverting the receptor to its normal status (10,11). As a long term effect, the transcription of the receptor itself can be repressed by increased c-Myc following stimulation by PDGF or serum (1,12). c-myc, one of the target genes of PDGF, leads to proliferative responses of cells and, in turn, down-regulates PDGF ␤-receptor, resulting in the cell cycle-dependent expression of the PDGF ␤-receptor. A Myc-null cell line showed a high and stable level of PDGF ␤-receptor mRNA, which disappeared upon forced expression of c-Myc (12). We have previously shown that c-Myc represses the transcription of the receptor through binding to the NF-Y transcription factor and interferes its activation (1).
The mouse PDGF ␤-receptor promoter contains a CCAAT box, and the NF-Y transcription factor binds to the motif and activates the transcription (13,14). NF-Y consists of NF-YA, NF-YB, and NF-YC subunits that are all necessary for DNA binding (15). The specific domains of NF-YA, NF-YB, and NF-YC that are needed for subunit interaction and DNA binding have conserved homologous sequences with the yeast HAP2, HAP3, and HAP5 proteins, respectively (as reviewed in Ref. 16). The repression of c-Myc on PDGF ␤-receptor promoter activity is dependent on the interaction of c-Myc with HAP domains of NF-YB and NF-YC (1).
We have recently found that c-Myc binds p73␣, thereby either activating or repressing the target genes of Myc or p73␣, respectively (3). It has been reported that c-Myc induces and activates p73␣ and p73␤ (2). These findings prompted us to examine the effect of p73 on PDGF ␤-receptor transcription. p73 is one of the recently identified p53-related family proteins including p63 and p51 (as reviewed in Ref. 17). p73 is similar to p53 in its ability to form oligomers, bind DNA, and activate transcription of p53-responsive genes by genotoxic stresses, leading to various actions such as induction of growth suppres-sion and apoptosis. However, p73 seems to play more essential roles in development and differentiation (18). There are several splicing variants of p73 (␣, ␤, ␥, ␦, ⑀, ) and also variants without N-terminal transactivation domains that act as dominant negative proteins for the corresponding full-length proteins (17). Roles of these proteins in normal cells have not yet been fully elucidated. The expression levels of p73 proteins are low in normal cells, and often, only p73␣ and p73␤ can be detected. We propose that serum stimulation induces c-Myc and p73␣, leading to the down-regulation of PDGF ␤-receptor expression by repressing its transcription in a similar manner.
Promoter Reporter Assay-NIH3T3 cells were seeded in 12-well plates at a density of 2 ϫ 104 cells/well. The following day, cells were transiently transfected with an expression plasmid, a reporter plasmid, and 0.3 g of CH110 (␤-galactosidase expression plasmid, Promega). Total amount of DNA per well was adjusted to 1 g by the addition of mock DNA plasmid. When two expression plasmids were used (i.e. experiments with p73 and c-myc), 0.25 g of CH110 was used. After 48 h, cells were lysed with 100 l/well reporter lysis buffer (Promega). Luciferase and ␤-galactosidase activity was measured according to vendor's instruction (Promega) by using Victor 1420 Multilabel Counter (Wallac). Luciferase activity was normalized to ␤-galactosidase activity. Measured values are expressed as the means Ϯ S.D. ANOVA was employed to test significant changes in activities. p values Ͻ 0.05 were considered to be statistically significant.
One-hybrid Reporter Assay-NIH3T3 cells were seeded in 12-well plates at a density of 2 ϫ 10 4 cells/well. The following day, cells were transiently transfected with 0.3 g of HAp73 expression plasmid and 0.2 g of pSG424 vector containing the GAL4-DNA binding domain fused with NF-YC full-length or ⌬1-107 containing the C-terminal HAP5 domain and the TAD or ⌬1-247 containing only the C-terminal part of the TAD together with 0.2 g of Gal4-TATA luciferase reporter vector (24) and 0.3 g of CH110. The total amount of DNA per well was adjusted to 1 g/well. The luciferase activity was measured and standardized to ␤-galactosidase as described above.
Reverse Transcription-PCR-NIH3T3 cells were seeded in 6-cm dishes at a density of 1.1 ϫ 10 5 cells/dish. After 24 h, half of the cells of each cell type were transfected with 2 g of HAp73 expression plasmid/ dish, and the remaining cells were transfected with 2 g of HApcDNA3 vector alone. The cells were kept in 10% FCS, and RNA was extracted essentially as described (25) at 0, 4, 8, 12, 24, and 36 h after transfection. 3 g of total RNA from each time point was transcribed into cDNA with random primers and Moloney murine leukemia virus reverse transcriptase (Invitrogen) according to the manufacturer's protocol. PCR was carried out by using Taq polymerase (Fermentas) and Programmable Thermal Blok ® (LabLine) as follows. For PDGFR␤, an aliquot (one-tenth) of the cDNA and the primers described previously (26) were used. An initial denaturation step at 94°C for 7 min was followed by 32 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 45 s, and primer extension at 72°C for 1 min, 30 s. For p73␣, 2.5% cDNA was mixed with primers: 5Ј-CAGCAGCAGCTCCTACAGAGG-3Ј and 5Ј-TACTGCTCGGGGATCTTCAGG-3Ј. Initial denaturation at 94°f or 3 min was followed by 22 cycles of denaturation at 94°C for 2 min, annealing at 67°C for 45 s, and primer extension at 72°C for 1 min. For ␤-actin, 5% cDNA and the primers described previously (27) were used. Initial denaturation at 94°C for 4 min was followed by 25 cycles of denaturation at 94°C for 30 s, annealing at 63°C for 45 s, and primer extension at 72°C for 1 min, 30 s. All of the reactions were completed with a final extension step at 72°C for 6 min. PCR products were analyzed on 1.5% TAE-agarose gel electrophoresis. The gels were stained with Sybr Gold (Molecular Probes) and scanned in a FLA2000 (Fuji).
Immunoblotting after Transfection-To see the time course of PDGF ␤-receptor expression following transfection of HAp73, NIH3T3 (5 ϫ 10 4 cells/dish) cells and HO15.19 (c-MycϪ/Ϫ) cells (1 ϫ 10 5 cells/dish) were seeded in eight 6-cm dishes each. They were transfected as described above for reverse transcription-PCR. The cells were harvested at 0, 12, 24, and 36 h post-transfection. They were scraped in 1 ml of phosphate-buffered saline and centrifuged for 2 min at 4°C at 1000 ϫ g. The cells were resuspended in 100 l of lysis buffer (50 mM Tris-HCl, pH 8.0, 0.1 mM EDTA, 120 mM NaCl, 0.5% Nonidet P-40, 10% glycerol, and 1 mM phenylmethylsulfonyl fluoride), incubated on ice for 20 min, and centrifuged at 21,000 ϫ g for 10 min at 4°C. The amount of protein was determined with Bio-Rad protein assay, and 50 g of each sample was adjusted to equal volumes (75 l/sample), boiled, and separated by 8% SDS-PAGE. The gel was transferred onto a Hybond-P polyvinylidene difluoride membrane (Amersham Biosciences), which was cut and blotted with 0.13 g/ml PDGF ␤-receptor antibody (958) or 0.5 g/ml HA (3F10) antibody in 1% skim milk. The membrane was developed by the ECL using vendor's protocol (Amersham Biosciences) and exposed in LAS-1000 Plus (Fuji). To estimate the p73 concentration-dependent PDGF ␤-receptor expression, NIH3T3 and HO15.19 fibroblasts were transfected with 0, 0.1, 0.5, or 2 g of HAp73 expression plasmid as described for the time course experiment. The total amount of transfected DNA was adjusted to 2 g with HA pcDNA3 vector. After another 24 h, cells were harvested, separated, and blotted in the same way as described for the time course experiment.
In Vivo Binding Assay-COS-1 cells were seeded in 6-well plates at a density of 1 ϫ 10 5 cells/well. After 24 h, cells were co-transfected with 1 g of HAp73 expression plasmid with 1 g of FLAG-NF-YA, FLAG-NF-YB, or FLAG-NF-YC expression plasmid. After another 24 h, cells were lysed in 200 l of lysis buffer described above. The cells were allowed to lyse for 20 min on ice and then were scraped. Each lysate was transferred into a 1.5-ml centrifuge tube and centrifuged at 21,000 ϫ g for 10 min at 4°C. 40 l of the supernatant was transferred to a new tube for expression control of the transfected plasmids. The remaining 160 l of the lysate was transferred to another tube and rotated at 4°C with 15 l of FLAG M2-agarose affinity gel for 2 h. The agarose beads were washed with 4 ϫ 1 ml of lysis buffer, separated by SDS-PAGE (10%), and blotted onto Hybond-P membrane. The membranes were first blotted with anti-HA (3F10) antibody and developed by ECL, and the signal was detected in LAS-1000 Plus. After drying, the membranes were reblotted with FLAG M2-agarose antibody.
Electrophoresis Mobility Shift Assay-A single-stranded antisense oligonucleotide containing CCAAT was used as a probe. The probe was complemented with a sense strand and labeled with [ 32 P]ATP. The template sequence was 5Ј-TTTGGGGGAAAGGCTGCAGGGCCGT-TCTGATTGGCCAAGC-3Ј. The labeled DNA fragments were separated from unincorporated [ 32 P]ATP using a G-25 MicroSpin TM column (Amersham Biosciences). To obtain nuclear extracts, a total of 4 ϫ 10 5 NIH3T3 cells were seeded in 6-cm dishes. Nuclear protein was extracted from the cells at 0, 4, 8, 12, 24, and 48 h after 10% serum stimulation following 48 h of culture in serum-free medium, nuclear extracts and 4 g of protein were preincubated in 20 l of reaction mixture containing 12 mM Tris-HCl, pH 7.9, 0.6 mM EDTA, 20% glycerol, 1 mM dithiothreitol, 5 mM MgCl 2, 0.1 g/l BSA, 0.1% Nonidet P-40, and 1 mg/ml poly(dI⅐dC). Unlabeled competitors (100ϫ labeled oligo), 32 P-labeled oligonucleotide, and antibodies NF-YA or p73 (0.5 g/reaction) were added in this order. The reactions were incubated 10 min at room temperature. At the end of the incubation, 4 l of 6ϫ agarose dye loading-buffer was added. The reaction mixtures were resolved by electrophoresis on a 4% acrylamide gel (polyacrylamide: bisacrylamide, 37.5:1) by 20 V/cm for 2 h at 4°C. The gel was dried and exposed to a BAS imaging plate (Fuji) and scanned in a FLA-2000.

Repression of PDGF ␤-Receptor
Promoter Activity by p73-We used the PDGF ␤-receptor promoter luciferase reporter, pGLSacI/SacI (pGL3Ϫ1900; Ϫ1900 to ϩ23) containing CCAAT motif located ϳ60 bp upstream of the first initiation site (Fig. 1A) (13). As shown in Fig. 1A, cotransfection of p73␣ decreased the luciferase activity of pGL3-1900 to 35% of that obtained with a control vector containing only HA tag. In comparison, the expression of neither p73␤ nor p53 did significantly and stably affect the reporter activity. The shorter MluI/ SacI (pGL3-116; Ϫ116 to ϩ23) promoter luciferase activity was also repressed by p73␣ to the same extent (result not shown). This shorter promoter contains Sp1 and NF-Y binding sites (28). We have previously reported that c-Myc represses PDGF ␤-receptor promoter (1). To see whether c-Myc is involved in the p73 repression on the promoter-luciferase activity, we used the HO15.19, c-mycϪ/Ϫ rat fibroblast cell line and the NIH3T3 normal mouse fibroblast cell line. In both cell lines, p73␣ and c-Myc significantly decreased the promoter activity (Fig. 1B). p73␣ repressed the promoter-luciferase activity more effectively than c-Myc in NIH3T3 cells containing c-myc. In the HO15.19, however, the activity of p73␣ was similar to that of c-Myc, and both together resulted in a further decrease in the transcriptional activity. The 1.9-kb PDGF ␤-receptor promoter lacks a putative binding site for p73. To see whether the CCAAT-binding NF-Y transcription factor is required for the p73␣-mediated repression, we co-transfected the DNA-binding defective and dominant negative NF-YA (DNNF-YA) with p73 and the promoterluciferase plasmids. The DNNF-YA is shown to act as a dominant repressor of NF-Y⅐DNA complex formation and of NF-dependent transcription (23). As expected, the co-expression of DNNF-YA and the promoter-luciferase plasmids decreased half of the promoter activity. However, additional coexpression of p73␣ to DNNF-YA showed no further effect on the promoter activity (Fig. 1C). In fact, the overexpression of p73␣ yielded a comparable repression as was seen by DNNF-YA alone on the promoter. These results indicate that p73␣ represses the promoter activity by its effect on NF-Y, attenuating the stimulatory effect of NF-Y on the promoter (14).
Association of p73␣ with NF-YB and NF-YC-The HAtagged p73 expression plasmid, HAp73␣, HAp73␤, HAp53, or HAp63␣, respectively, was co-transfected with the FLAG-NF-YA, FLAG-NF-YB, FLAG-NF-YC, or control FLAG-vector expression plasmid, respectively, to COS-1 cells. Forty-eight hours after transfection, cells were lysed and immunoprecipitated with anti-FLAG M2 antibody and immunoblotted with an anti-HA antibody. As shown in Fig. 2A, p73␣ and p63␣ bound FLAG-NF-YB and FLAG-NF-YC in vivo but not FLAG-NF-YA or the control vector. Neither p73␤ nor p53 bound any of the NF-Y subunits. Immunoblotting of the membrane being trans- FIG. 1. A, p73␣ represses PDGFR␤ promoter activity. NIH3T3 cells were co-transfected with the pGL3-PDGFR␤ promoter construct to-gether with HAp73␣, HAp73␤, HAp53, or HA expression vector, respectively. B, p73␣ repression on PDGFR␤ promoter activity is not mediated by c-Myc. NIH3T3 and HO15.19 (c-mycϪ/Ϫ) fibroblasts were co-transfected with pGL3-PDGFR␤ promoter construct with HAp73␣ and/or c-myc plasmids, and the luciferase activity was measured.   2. A, interaction of NF-YB and NF-YC with p73␣. COS-1 cells were co-transfected with various FLAG-NF-Y-subunits or FLAG-vector alone, respectively, together with HA-tagged p73␣, p73␤, p53, or p63␣. Cell lysate was divided and used for immunoprecipitation (IP) or immunoblotting (IB). After immunoprecipitation with anti-FLAG M2-agarose affinity gel, the samples were separated together with the cell lysates used for the reaction. Transferred membrane was immunoblotted with HA antibody (3F10, upper panel), and FLAG-M2 antibody (lower left panel). The expression of HA-tagged p73␣, p73␤, p53, and p63␣ in the same cell lysates was confirmed by HA antibody (lower right panel). B, schematic ferred with the same cell lysates, which reacted with the anti-HA antibody, as well as the anti-FLAG M2 antibody confirmed the appropriate expression of proteins. The N-terminal region of p73␣ and p73␤ is homologous, and the only notable divergence can be found in the C-terminal area (Fig. 2B). This unique area of p73␣ not found in p73␤ is probably responsible for the binding. In addition, p63␣ also bound NF-Y. Because in vertebrae, this p53 family protein is the only one other than p73␣ that contains an area known as the SAM domain in its C terminus. As expected, N-terminal deletion mutant of p73␣ lacking the TAD (⌬Np73␣) also bound NF-Y (result not shown).
To determine whether this area is responsible for the interaction, we examined several other p73 isoforms (␤, ␥, and ⑀) as well as the two C-terminal deletion mutants of p73␣, the ⌬p73␣C269 and ⌬p73␣C424, together with p73␣ and p53 as a positive and negative control, respectively (Fig. 2C). In this experiment, we used FLAG-tagged p73 and HA-tagged NF-YC to confirm the reciprocal binding. No other p73 isoforms could bind NF-YC, and the C-terminal deletion mutants of p73␣ without SAM domain were unable to bind NF-YC. The binding site of NF-YC was also determined by co-immunoprecipitation using various N-terminal deletion mutants (Fig. 2, D and E). This result indicates that the C-terminal HAP5 domain of NF-YC, i.e. amino acids 108 -115, is necessary for the binding. To exclude the possibility that p73␣ and NF-Y interacts via c-Myc, the binding assay between NF-YB or YC and p73␣ was repeated in HO15.19 c-mycϪ/Ϫ cells (Fig. 2F). Again, also in this cell line, p73␣ binds the NF-Y subunits B and C, indicating that c-myc is dispensable for this binding.
The binding of p73␣ to NF-YC Is Indispensable for the Repression of NF-YC Transactivation-The transactivation domain of NF-YC subunit has the highest activity of all three NF-Y subunits, whereas NF-YB contains very little such activity (29). Therefore, p73␣ effect was examined on its activity of the C-terminal transactivation domain of NF-YC (Fig. 2G). The deletion mutant containing the transactivation domain was coupled downstream of the DNA-binding domain (DBD) of GAL4 transcription factor (Fig. 2H) (1). As reported, the fulllength NF-YC fused to DBD of GAL4 transcription factor did not significantly activate transcription from the GAL4-TATApromoter, whereas the corresponding constructs containing only the TAD are active (29). When the N-terminal deletion mutant of NF-YC retaining the C-terminal HAP domain needed for the p73 binding (GAL4-DBD-NF-YC⌬1Ϫ107) was co-transfected with p73␣, the transactivation activity was diminished to half of that obtained by co-transfection of the control HA-vector, indicating that p73 interferes with the activation of NF-YC (Fig. 2G). In agreement, the activity of the GAL4-DBD-NF-YC⌬1Ϫ248 being unable to bind p73␣ was not affected by co-transfection of p73␣.
Transactivation Domain of p73␣ Is Necessary for Repression of PDGF ␤-Receptor Promoter-We determined the binding site of p73␣ to its C-terminal SAM domain. To see whether the TAD of p73␣ is necessary for the repression, we transfected the ⌬Np73␣ lacking the TAD (⌬Np73␣), HA-vector alone, or the full-length p73␣ together with the PDGF ␤-receptor promoter luciferase construct in NIH3T3 cells and measured its activity. As shown in the Fig. 3A, there was no repression, but rather activation was observed when TAD was deleted from p73␣. This finding was further confirmed by the concentration-dependent increase of promoter activity by the addition of ⌬Np73␣ to the full-length p73␣. ⌬Np73␣ alone stimulated the promoter activity.
Effects of p73 on the Expression of the PDGF ␤-Receptor-We demonstrated that p73␣ repressed the transcriptional activity of PDGF ␤-receptor promoter. We next examined whether this repression is reflected in the mRNA and protein expression of PDGF ␤-receptor by reverse transcription-PCR and immunoblotting. When HA-tagged p73␣ was transfected in NIH3T3 cells, the expression of p73␣ mRNA increased as judged by the p73␣ PCR product. It became detectable at 4 h to increase gradually up to 24 h, remaining at the same level at 36 h. On the contrary, the PDGF ␤-receptor PCR product being clearly expressed already at 0 and 4 h gradually decreased at 8, 12, and 36 h. (Fig. 3B). On the other hand, the transfection of the control HA-vector without p73␣ did not affect the level of PDGF receptor mRNA. In agreement with this finding, p73␣ protein was detected already at 12 h, and the expression increased up to 36 h (Fig. 4A). By contrast, the receptor expression rapidly decreased already at 12 h and remained as a faint band during the whole observation period. Control HA-vector transfection did not alter the receptor expression level that remained at a high level. The product of the control expression vector with only HA tag migrates too fast to be detected in the immunoblots. A similar marked decrease of the PDGF ␤-receptor expression was seen in the p73␣ expressing c-myc-deficient HO15.19 cells (Fig. 4B). This cell line has normally higher expression of the receptor than c-myc containing cell lines, its parental line TGR (data not shown) and NIH3T3. The p73␣ concentrationdependent receptor expression was further examined by transfecting increasing amounts of p73␣ expression plasmid in each of these cell lines. As expected, both cell lines showed a gradual decrease of PDGF ␤-receptor expression when the p73␣ expression increased (Fig. 4, C and D).

Serum Stimulation Decreases Binding of Endogenous NF-Y Complex to the PDGF ␤-Receptor
Promoter-To examine whether NF-Y complex binding to DNA changes during cell cycle, electrophoresis mobility shift assay was performed with a 32 P-labeled double-stranded oligonucleotide probe for PDGF ␤-receptor promoter containing the CCAAT motif (14). Nuclear protein was taken at various time points after serum stimulation from NIH3T3 cells that had been kept in a serum-free medium for 48 h. As shown in Fig. 5, the protein complex that bound the probe gradually decreased whereas the expression of NF-YA in the nuclear protein did not decrease as shown by the NF-YA antibody that blocked the protein-DNA binding. Furthermore, the 100-fold excess oligonucleotide containing the CCAAT but not the corresponding oligonucleotide with mutated CCAAT competed with the complex to bind the probe, confirming the specificity of the binding. However, the nuclear extract from NIH3T3 cells that had been transfected with a p73 expression vector never showed a decreased amount of NF-Y complex binding to DNA (result not shown). Neither could we find any inhibition by p73␣ on the NF-Y binding to DNA when proteins, p73␣ and NF-Y subunits, were made in vitro (result not shown). These results suggest that endogenous p73␣ and other nuclear proteins such as c-Myc interferes with NF-Y and indirectly decrease its binding to the CCAAT motif. DISCUSSION The CCAAT motif upstream of the initiation site of PDGF ␤-receptor gene is essential for the basal transcription activity and also for the c-Myc-induced repression (1). We herein demonstrate that p73␣, a p53 family suppressor gene product, represses the activity of the 1.9-kb SacI/SacI PDGF ␤-receptor promoter in a concentration-dependent manner. It is consid- ered that all the NF-Y subunits are necessary for the stable binding of NF-Y complex to DNA. However, NF-YA seems to be the important subunit to create the DNA binding surface (15). The overexpression of DNNF-YA was shown not only to prevent NF-Y complex from binding to CCAAT box (23) but also to attenuate the p73␣ effect on the PDGF ␤-receptor promoter activity. Thus, these results suggest that NF-Y binding is crucially involved in the p73␣-induced repression on PDGF ␤-receptor promoter. Co-immunoprecipitation of p73␣ with NF-Y subunits and their deletion mutants revealed that p73␣ but not other isoforms, splicing variants, or p53 bound the HAP domain of NF-YB and NF-YC. The binding capacity of these molecules agrees with their capacity of the transcription repression. The HAP domain contains the histone-fold motif (15,16) that is important for DNA binding (30). The C terminus of HAP domain in NF-YC is known to be involved in binding to TBP (amino acids 108 -120) (31) as well as to c-Myc (1). This hydrophobic C-terminal domain of the HAP area might act as an interacting domain with other proteins, thus providing a possibility to finely tune the transcriptional activity of NF-Y that is ubiquitously expressed in different tissues.
A wide variety of p73 actions are linked to its different structural domains. The N-terminal and central part of p73 has a similar structural organization as other p53 family proteins. The C-terminal region of p73␣, a unique site not present in other p73 subunits or in p53, is probably important for the repression. The C-terminal p73␣ contains both proline-rich area and the SAM domain known as interacting domains. In fact, no other p73 isoforms bind NF-YC, but p73␤ that does not either bind NF-Y or repress the transcription also possesses the proline-rich domain. Thus, our results indicate that p73␣ may bind NF-Y via its hydrophobic SAM domain, consisting of five helices (32). In fact, we found that p63␣, another p53 member in vertebrates that possesses the SAM domain, also binds NF-YC. The difference in the transcriptional activities between various C-terminal isoforms of p73 and p63 (33) may be partly attributed to the differences in their interacting proteins.
Interestingly, the TAD of p73␣ is indispensable for the repression because the ⌬Np73␣ devoid of TAD is incapable of repressing the promoter. This was also confirmed by the DDp73␣ lacking both TAD and the DNA-binding domain but retaining the intact C-terminal domain (result not shown). The expression of ⌬Np73␣ not only fails to repress the promoter but also increases the activity in a dominant negative manner. It is shown that p73␣ can homodimerize and heterodimerize with any of the p73 isoforms (33) via their oligomerization domains. Although p73␣ does not efficiently oligomerize with p53, ⌬Np73␣ can bind p53, inactivating the function of p53 (34). Furthermore, ⌬Np73␣ inhibits p53 by competition for DNA binding (35). Although we could not show any direct repression of p53 on the promoter, it is possible that by decreasing Myc, p53 indirectly affects the level of the PDGF receptor. It is also possible that ⌬Np73␣ binds wild-type p73␣ more efficiently and inactivates wild-type p73␣.
Structurally, p73 is a typical transcription factor, activating common target genes of p53, i.e. p21, MDM2, and Bax (36), through binding to the p53-responsive DNA sequence. Although p73␣ represses the transcription of the PDGF ␤-receptor, the promoter of which has no p73␣ binding site. The repression occurs not by a direct interaction with DNA but via protein-protein interaction with NF-Y. The binding site of p73 to c-Myc is close to its own DNA-binding domain (amino acids 227-312), being distinct from the area needed for the interaction of p73␣ with NF-Y (3). This area is conserved in all p73 isoforms and therefore is unlikely to be involved in the p73␣ repression on the PDGF ␤-receptor promoter because p73␤ lacks this activity. p73 is also not required for the repression of Myc on the promoter, because the p73 binding site of Myc, the C-terminal HLH-LZ domain, is dispensable for the repression of Myc on the PDGF ␤-receptor. The Myc-null HO15.19 cells responded as sensitively as normal Myc-positive fibroblasts to the p73␣-mediated repression, supporting the notion that p73␣ repression is not mediated by c-Myc. However, they might still act complimentary in a physiological situation in which each of these proteins may not yield a saturating effect as it would do in experimental conditions. In fact, in the Myc-null HO15.19 cells, we found a slightly stronger repression by overexpression of both c-Myc and p73␣ rather than by one of them alone. Our results suggest that p73 exhibits diverse regulatory effects on transcription of various genes other than by its direct binding to DNA.
Similar to the case of Myc, we could not see any change in electrophoresis mobility shift assay when using nuclear extracts taken from p73␣ overexpressing cells or by using in vitro synthesized proteins, arguing against sequestering of NF-Y by p73␣ as the mechanism for the repression (data not shown). This finding agrees also with the fact that not only the SAM domain but also the N-terminal TAD of p73␣ is necessary for the repression. In many aspects, the mechanism of the p73␣induced repression is comparable to that by c-Myc, suppressing the level of transactivation. Unexpectedly, however, we could show that the efficiency of NF-Y to bind DNA decreased when cells were re-entered in a cell cycle, whereas the level of NF-YA in the nuclear proteins remained unaffected. NF-YA is the only NF-Y subunit that has been reported to change its expression level (37). This strongly suggests that p73␣ and also c-Myc and other yet unidentified proteins that increase during G 1 /S transition in the nucleus might decrease the binding of NF-Y to the promoter, probably in an indirect manner. It remains to be elucidated whether and how the interaction occurs in vivo, and such a study is now under way.
PDGF is one of the most potent serum mitogens that activate G 1 /S cell cycle progression (4,5). Cell cycle of normal cells is tightly controlled by various mechanisms to prevent uncontrolled proliferation. p73 seems to be expressed at a low level in normal condition, but it can substitute the function of p53 in p53Ϫ/Ϫ cells. Certain genotoxic stimuli such as UV radiation and cisplatin can induce p73, in turn inducing p21 to arrest cell cycle. In such cases, the down-regulation of PDGF ␤-receptor is necessary for cells to avoid excess growth stimuli. Mutant p53, but not the wild-type p53, seems to heterodimerize and sequester wild-type p73 (38), indicating that in certain tumors none of these suppressor proteins are functional. In such a case, PDGF ␤-receptor transcription might not be down-regulated during a cell cycle. It might contribute to the high PDGF ␤-receptor expression maintained in tumors, promoting the PDGF-mediated autocrine growth stimulation often found in advanced tumor cells.