The tumor suppressor p53 abrogates Smad-dependent collagen gene induction in mesenchymal cells.

The pleiotropic cytokine transforming growth factor-beta (TGF-beta) is a potent inducer of collagen synthesis and is implicated in the pathogenesis of fibrosis. Acting in concert with transcriptional coactivators p300/CBP, the Smads mediate TGF-beta stimulation of collagen synthesis in human dermal fibroblasts. Little information exists regarding positive and negative modulation of physiological TGF-beta responses. Because the tumor suppressor p53 is implicated in connective tissue homeostasis, here we examined the regulation of collagen gene expression by p53. Forced expression of ectopic p53 in dermal fibroblasts repressed basal and TGF-beta-stimulated collagen gene expression, whereas the absence of cellular p53 was associated with significantly enhanced transcriptional activity of the Type I collagen gene (COL1A2) and collagen synthesis. Ectopic expression of p53 also repressed TGF-beta stimulation of promoter activity driven by minimal Smad-binding elements, suggesting that p53 modulated Smad-dependent intracellular signaling. Inhibition was not due to altered levels, phosphorylation, or nuclear translocation of cellular Smads. Treatment of fibroblasts with etoposide, a potent inducer of cellular p53, abrogated TGF-beta stimulation of COL1A2 promoter activity and collagen synthesis in a p53-dependent manner. Overexpression of the transcriptional coactivator p300 rescued TGF-beta stimulation of COL1A2 promoter activity in fibroblasts overexpressing p53. Furthermore, the ligand-induced interaction of cellular Smad3 with p300 or with its cognate Smad-binding DNA element and recruitment of p300 to the DNA-protein complex assembled on the Smad-binding element were markedly reduced in p53-overexpressing fibroblasts. Collectively, these results indicate, for the first time, that p53 is a potent and selective endogenous repressor of TGF-beta-regulated collagen gene expression in dermal fibroblasts. The ligand-dependent interaction of Smad3 with p300 may be one of the targets of p53-mediated inhibition of TGF-beta responses. These findings suggest that a novel and important physiologic function for the tumor suppressor p53 is the regulation of fibrotic cellular responses.


INTRODUCTION
Transforming growth factor-β (TGF-β) is the prototype of a large superfamily of multifunctional cytokines that control cellular growth and differentiation. In a variety of mesenchymal cells, TGF-β induces the synthesis of collagens and other extracellular matrix components, and is thus a pivotal contributor to pathological fibrosis (1). The molecular mechanisms that govern the regulation of major extracellular matrix gene expression in response to TGF-β are the subject of intense investigation (2). The recent discovery of Smads as novel TGF-β signal transducers opens a new avenue for fibrosis research (3 and references therein). The Smad family consists of eight members that can be classified into three subgroups based on their structure and function. Smad1, Smad2, Smad3, Smad5 and Smad8 are receptor-activated Smads or R-Smads; Smad4 is a co-Smad, and Smad6 and Smad7 are inhibitory (4). We have shown previously that Smad3 was responsible for mediating TGF-βinduced stimulation of collagen synthesis in skin fibroblasts, whereas Smad7 abrogated this response (5). Furthermore, we demonstrated that the transcriptional coactivator p300 physically and functionally interacted with Smad3 in a ligand-dependent manner, and played a major role in Smad-dependent stimulation of collagen gene expression (6,7).
The mechanisms that control the magnitude or duration of Smad-dependent TGFβ mediated cellular responses are only partially understood. Because deregulated TGF-β signaling is implicated in the pathogenesis of fibrosis in scleroderma and related conditions (8), identification and characterization of transcriptional cofactors that modulate this process is important. The tumor suppressor p53 is a short-lived nuclear phosphoprotein that is mutated in a majority of human cancers. It plays important roles in the regulation of cell growth, 4 apoptosis, differentiation and senescence (9,10). These activities of p53 are mediated through inhibition or stimulation of target gene expression during cell cycle (11,12). Whereas activation of transcription by p53 is generally due to its direct interaction with regulatory ciselements of target genes, repression involves protein-protein interaction between p53 and other transcription factors, modulation of the activity of the transcriptional coactivator p300, or recruitment of histone deacetylase in the transcriptional complex (9,(13)(14)(15)(16)(17)(18)(19)(20).
Recent studies suggest that the tumor suppressor p53 plays an important role in regulating extracellular matrix homeostasis. Ectopic expression of p53 in fibroblasts inhibited the formation of fibronectin fibrils (21), whereas inhibition of p53 expression in HeLa cells resulted in increased fibronectin synthesis (22). Furthermore, p53 has been shown to repress plasminogen activator (PA), and stimulate plasminogen activator inhibitor-1 (PAI-1) gene expression in mesenchymal cells (13). In mouse skin fibroblasts, wildtype p53, but not its mutants, repressed the synthesis of matrix metalloproteinase (MMP)-1 and MMP-13, while stimulating the synthesis of MMP-2 (23)(24)(25)(26). The transcription of the tissue inhibitor of metalloproteinase-3 (TIMP-3) gene was repressed by p53 (27). These results, indicating that p53 can positively or negatively modulate the expression of multiple extracellular matrix genes, and suggests a novel function for p53 in physiological regulation of connective tissue homeostasis. The significance and molecular mechanisms underlying p53 regulation of extracellular matrix gene expression are largely unknown.
The p53 protein undergoes activation via site-specific phosphorylation, dephosphorylation and acetylation in response to different forms of cellular stresses (9,28). In addition, by guest on March 24, 2020 http://www.jbc.org/ Downloaded from the activity of p53 is also regulated via its degradation and subcellular localization (29).
Complex interactions between p53 and the TGF-β signaling pathway have been demonstrated. It has been shown that p53 modulates TGF-β-mediated stimulation or inhibition of cellular proliferation (30,31).
Although p53 is known to modulate the expression of extracellular matrix protein genes, and regulate TGF-β responses, the physiologic role of p53 in Smad-dependent TGF-β regulation of collagen gene expression has not been examined. We now report that expression of ectopic p53 in normal dermal fibroblasts resulted in repression of TGF-β-stimulated collagen gene transcription and Smad-dependent responses. In contrast, the TGF-β-induced PAI-1 promoter activity was enhanced by overexpressed p53, in agreement with previous report (32). Type I collagen gene (COL1A2) promoter activity, and collagen synthesis, were significantly increased in murine embryonic fibroblasts lacking p53, indicating the physiological significance of cellular p53 in regulation of collagen gene expression.
Etoposide, an inducer of p53 expression, repressed collagen synthesis in wildtype, but not in p53-null, fibroblasts. Repression of Smad-dependent TGF-β-responses was not due to alteration in the expression levels, phosphorylation, or nuclear translocation of Smads.
Transient overexpression of p300 rescued TGF-β-induced COL1A2 promoter stimulation in the presence of p53. Furthermore, induction of cellular p53 resulted in repression of TGF-βinduced Smad3 interaction with the Smad-binding element and with p300. Collectively, these results demonstrate, for the first time, that cellular p53 is a potent negative modulator of TGFβ signaling, with potentially important role in regulating fibrotic responses in normal dermal fibroblasts. These results extend the range of biological activities attributed to p53. 7 cDNA in pCMVNeoBam expression vector (14). The p53-213 (Arg to stop codon at 213) and p53-239 (Asp to Ser at 239) constructs were made by inserting rheumatoid synovium-derived mutant p53 cDNA in pCI vector (39).

Transient transfections and reporter assays:
Dermal fibroblasts or MEFs grown in 6-well or 12-well clusters were transiently transfected using Superfect Reagent (Qiagen, Valencia, CA) following the manufacturer's instructions, and CAT or luciferase assays were performed as described previously (40). The reporter constructs were transfected along with expression vectors or appropriate empty vectors, and renilla luciferase expression vector pRLTK-LUC as an internal standard.
Transfected cells were incubated in fresh EMEM containing 10% FBS with or without TGF-

Immunoprecipitation and Immunoblot analysis:
Whole cell lysates were prepared from dermal fibroblasts or MEFs using lysis buffer (Promega) or M-Per mammalian protein extraction buffer (Pierce, Rockford, IL), and centrifuged at 4 0 C for 5 min at 10,000 rpm. Nuclear extracts or cytoplasmic extracts were prepared from transfected or etoposide treated dermal fibroblasts as described previously (41).
Equal amounts of proteins (10-25 µg), or conditioned media (40 µl) were resolved by electrophoresis 4-20% Tris-Glycine gradient gels (BIORAD, Hercules, CA). For immunoprecipitation, cell lysates were prepared and immunoprecipitated as described previously (7). Immunoprecipitated proteins were subjected to electrophoresis and transferred to PVDF membranes. The PVDF membranes were blocked with 10% fat free milk in TBST were treated with ECL reagent (Amersham Biosciences, Piscataway, NJ) and exposed to Kodak XAR5 film.

Electrophoretic mobility shift assays:
Confluent fibroblasts were incubated with TGF-β in the presence or absence of etoposide (1 µM). At the end of incubation, nuclear extracts were prepared and analyzed by electrophoretic mobility shift assays as described (7). For this purpose, equal aliquots (5 µg) were incubated for 30 min in binding buffer in presence or absence of 100-fold molar excess of cold competitors.
Antibody supershift assays were performed using anti-Smad3 antibody (I-20) (Santa Cruz) or control IgG. Radiolabeled Smad binding element probes were added to the reaction mixture, incubated for another 30 min and DNA-protein complexes were resolved in 6% polyacrylamide gels. Gels were then dried and exposed for autoradiographs.

DNA-Affinity Precipitation Assays:
The presence of Smad3 and p300 in the DNA-binding complex assembled on the SBE probe was examined by DNA-protein interaction assays (42). Briefly, fibroblasts incubated with TGF-β for 1 h in presence or absence of etoposide (1 µM) were harvested, nuclear extracts were prepared and equal amounts of protein (~200 µg) were incubated with biotinlabeled double-stranded SBE probes for 30 min. At the end of incubation 40 µl streptavidinagarose beads (4%) with 50% slurry (Sigma Chemicals) were added and mixtures were incubated at 4 0 C for 45 min. The streptavidin-agarose beads were precipitated by centrifugation, pellets were washed three times with cold phosphate-buffered saline and beadbound proteins were resuspended in SDS-loading buffer followed by electrophoresis on 4-20% denaturing gels. Gels were processed for immunoblot analysis using antibodies against Smad3 and p300 as above. The p53 levels in nuclear extract were determined by immunoblot analysis using antibodies against p53 or Smad4.

Statistical Analysis:
The data are presented as means ± S.D. Statistical differences between experimental and control groups were determined by analysis of variance, and a value of p< 0.05 by Student t-test was considered significant.

Forced expression of p53 selectively inhibits Type I collagen synthesis and abrogates its stimulation by TGF-β:
In addition to its well-characterized role as a tumor suppressor, p53 is increasingly recognized as a regulator of extracellular matrix synthesis. The molecular mechanisms underlying these important effects of p53 are largely unknown. To characterize the effects of p53 on collagen synthesis, normal dermal fibroblasts were transiently transfected with expression vector for p53 or empty vector, followed by incubation with TGF-β (12.5 ng/ml).
At the end of the 48 h incubation, whole cell lysates were prepared and subjected to immunoblot analysis. The results showed that whereas TGF-β induced a marked increase in collagen, as expected, ectopic expression of p53 resulted in significantly reduced cellular levels of collagen both in the presence and in absence of TGF-β (Fig. 1A). The cellular levels of actin remained unaltered, and no effect on cell viability and protein concentration was seen.
To investigate the mechanism of collagen gene regulation by p53, fibroblasts were transiently transfected with expression vector for p53 or appropriate empty vector, along with the 772COL1A2-CAT or 460PAI-1-CAT reporter constructs or p53-luc as positive controls.
After 48 h incubation with TGF-β, cultures were harvested and CAT or luciferase activities were determined. The results of transient transfection assays showed that ectopic expression of p53 resulted in repression of basal COL1A2 promoter-driven transcriptional activity, and prevented its stimulation induced by TGF-β (Fig. 1B, left panel). The inhibitory effect of p53 appeared to be specific for COL1A2 promoter, as under identical experimental conditions, the basal and TGF-β-induced activities of PAI-1 promoter (containing a consensus p53 binding element) were stimulated by p53 ( Fig. 1B middle panel). The activity of p53-luc, a minimal reporter construct driven by p53 response elements was increased by 2-fold ( Fig. 1B, right panels), and no effect on TK promoter activity was noted (data not shown).
In order to investigate the mechanistic basis for p53-mediated transcriptional repression, well-characterized mutants of p53 were used in transiently transfected fibroblasts.
The results showed that forced expression of tumor-derived p53 mutants unable to bind to DNA (14) resulted in substantial decrease in TGF-β-induced stimulation of COL1A2 promoter activity ( Fig. 1C left panel). In contrast, ectopic expression of dominant negative mutants derived from rheumatoid synovium (39) failed to abrogate TGF-β stimulation ( Fig   1C right panel), despite comparable expression levels of wildtype and mutant p53. Because these p53 mutants are also unable to bind to DNA (43), the results indicate that repression of COL1A2 by p53 is independent of direct DNA interaction.

Etoposide-induced cellular p53 abrogates TGF-β stimulation of COL1A2 promoter activity and collagen synthesis:
Etoposide, a topoisomerase II inhibitor, is a potent stimulus for p53 accumulation in a variety of cell types. Etoposide has been shown to regulate the expression of MMP-1, MMP-2, p21 and other p53 target genes (23,24,(44)(45)(46). In order to examine the role of cellular p53 in regulation of collagen gene expression, fibroblasts transiently transfected with 772COL1A2-CAT reporter construct were incubated with etoposide (0.1-1 µM) in presence and absence of TGF-β. After 48 h, fibroblasts were harvested, and CAT activities were determined.
Etoposide by itself had only minimal effect on the activity of transfected 772COL1A2-CAT, by guest on March 24, 2020 http://www.jbc.org/ Downloaded from but it abrogated TGF-β-induced stimulation in a dose-dependent manner ( Fig. 2A). Next, the effect of etoposide on cellular collagen levels was examined. The results showed that treatment of fibroblasts with etoposide for 48 h substantially attenuated TGF-β-induced stimulation of collagen synthesis (Fig. 2B). In contrast, the levels of basal as well as TGF-βinduced PAI-1 were elevated in fibroblasts incubated with etoposide (Fig. 2B). The expression of p53 was induced by etoposide, as expected. At the concentrations used, etoposide had no effect on cellular viability, protein levels, or pRLTK promoter activity (data not shown). Together, these results indicated that etoposide-induced stimulation of cellular p53 was associated with selective repression of basal and TGF-β-stimulated collagen gene expression in dermal fibroblasts.
Etoposide has multiple effects in addition to enhancing cellular p53. To further investigate the physiological significance of cellular p53 in the regulation of COL1A2 promoter activity, MEFs lacking p53 were used. For this purpose, wildtype or p53-null MEFs were transiently transfected with 772COL1A2-CAT and incubated with etoposide (1 µM) in the absence or presence of TGF-β. At the end of 48 h incubation, cells were harvested and CAT activities were determined. The results of transient transfection assays revealed that in unstimulated p53-null MEFs, COL1A2 promoter activity was 3-fold higher than in wildtype controls (Fig. 3A). Forced expression of ectopic p53 repressed basal and TGF-β-induced promoter activity to a comparable degree in both wildtype and p53-null MEFs (data not shown). In contrast, whereas etoposide caused significant inhibition of TGF-β-stimulated COL1A2 promoter activity in wildtype MEFs (~90%), in p53-null MEFs etoposide was completely unable to reduce TGF-β stimulatory response (Fig. 3A).
The regulation of endogenous collagen synthesis by cellular p53 was next studied in wildtype and p53-null MEFs. The MEFs were incubated with etoposide in the presence and absence of TGF-β for 48 h and culture supernatants were analyzed by Western immunoblot.
The results showed that levels of basal and TGF-β-induced Type I collagen were significantly higher in p53-null MEFs compared to wildtype controls. Furthermore, incubation of MEFs with etoposide failed to repress the basal and TGF-β-stimulated collagen synthesis in the absence of cellular p53 (Fig. 3B). Collectively, these results indicated that repression of collagen gene expression by etoposide was p53-dependent, suggesting an important physiological role for cellular p53 in negative regulation of collagen gene expression.

Forced expression of p53 abrogates COL1A2 stimulation via TGF-β-response element:
Sequences located within -353 bp of the human COL1A2 promoter have been shown to be sufficient to mediate full TGF-β stimulation of transcription (47). This region of the promoter harbors a functional SBE (4). In contrast, no consensus p53-binding element (two copies of 5'-PuPuPuC (A/T) (T/A) GPyPyPy-3' separated by 0-13 bp) (15) was found by computer search in this region. In order to delineate regions of the COL1A2 promoter mediating the transcriptional inhibitory effects of p53, 5' deletions of the promoter were used along with p53 expression vector in transient co-transfection assays. The results showed that the -353 bp sequence of COL1A2 required to mediate TGF-β stimulation of promoter activity was also sufficient for repression in response to p53 (Fig. 4A). In order to test the functional significance of CAGACA motif, located within the TGF-β response element of COL1A2 promoter (5) in p53-mediated repression, the effect of p53 on TGF-β-induced COL1A2-CAGA promoter activity was determined. As shown in Fig. 4B, TGF-β induced a >2-fold increase in COL1A2-CAGA promoter activity as expected, and p53 abrogated this stimulation.

p53 represses Smad-dependent transcriptional responses:
Because we have demonstrated previously that stimulation of collagen gene expression by TGF-β was mediated through Smads (5), we sought to further investigate the regulation of Smad-dependent transcriptional responses by p53. For this purpose, fibroblasts were cotransfected with expression vectors for p53 and Smad3 along with 772COL1A2-CAT. As expected, Smad3 caused ligand-independent stimulation of COL1A2 promoter activity, and TGF-β enhanced this stimulation further (Fig. 5A).

Unaltered cellular Smad activation in presence of excess p53:
In order to determine whether p53 repression of Smad-dependent transcription was due to altered Smad expression, protein levels were determined by Western immunoblot. The results revealed that levels of endogenous Smad3, Smad4 and Smad7 were not affected by ectopic expression of p53 in fibroblasts in either the presence or absence of TGF-β (Fig. 5C).
Furthermore, incubation of fibroblasts with TGF-β for 48 h had no effect on p53 levels. Of note, the expression of PAI-1 was enhanced by TGF-β, as well as by forced expression of p53. Ectopic p53 further enhanced the TGF-β stimulation, in agreement with previous observations (32). These results suggest that p53 repressed selected Smad-dependent TGFβ responses in fibroblasts by altering the transcriptional activities rather than the expression levels of signaling Smads.
To further investigate the effect of induced p53 on TGF-β-induced Smad activation, the phosphorylation and nuclear translocation of endogenous R-Smads were studied by immunoblot analysis. The results indicated that while TGF-β induced rapid phosphorylation and nuclear accumulation of R-Smads, as expected, neither phosphorylation nor nuclear accumulation were altered in the presence of ectopic p53 (Fig. 5D left panel) or etoposideinduced accumulation of cellular p53 (Fig. 5D right panel). Taken together, these results strongly suggest that p53 interferes the downstream Smad signaling events those take place in the nuclei.

p53 interferes with Smad3 interaction with DNA:
Upon activation by the TGF-β receptor, R-Smads enter the nucleus and bind with low affinity to consensus SBE sequence in TGF-β target gene (4), resulting in activation of their transcription. To determine if p53 modulated the inducible DNA-binding activity of R-Smads, electrophoretic mobility shift assays were performed with nuclear extracts from fibroblasts treated with etoposide to enhance endogenous p53 expression. The results showed that TGFβ-induced binding of Smad to its cognate cis element was reduced in presence of excess p53 ( Fig. 6A lanes 4 and 8). The identity of the TGF-β-induced band was confirmed by supershift assays using antibody to Smad3 (Fig. 6A, lanes 5 and 9).
The regulation of certain cellular responses by p53 is dependent upon its interaction with a variety of coactivators (reviewed in 48). The transcriptional coactivator p300 is a nuclear phosphoprotein that interacts with sequence-specific transcription factors to mediate cell cycle regulation, differentiation and development in response to various ligands (reviewed in 49). Both p53 and Smad have been shown to interact with p300. We have shown previously that Smad-dependent stimulation of collagen synthesis by TGF-β required the direct physical interaction between activated Smad3 and cellular p300 (6,7). To determine the levels of Smad3 and p300 within DNA-protein complex assembled on SBE probe, DNAaffinity precipitation assays were performed using biotinylated DNA probes and streptavidinagarose beads. The results showed that TGF-β treatment of the fibroblasts resulted in enhanced interaction of Smad3 and associated p300 with SBE (Fig. 6B). However, in fibroblasts preincubated with etoposide, interaction of the Smad-p300 complex with DNA was significantly reduced in parallel with increased nuclear p53 levels. Taken together, the results from the electrophoretic mobility shift assays and DNA-affinity precipitation assays suggested that p53 disrupted the ligand-induced interaction of endogenous Smad3 with its cognate DNA binding sites. Consequent failure of R-Smads to recruit p300 in the DNAbound complex may account, at least in part, for repression of TGF-β-induced transcriptional responses by p53 in target genes lacking p53-binding element.

Overexpressed p53 interferes in the interaction of Smad3 with p300:
To further characterize the effect of p53 on TGF-β-induced physical interaction of Smad3 with p300, whole cell lysates from fibroblasts with forced expression of ectopic p53 or empty vector were immunoprecipitated with anti-p300 antibodies, and subjected to immunoblot analysis using anti-Smad1/2/3 and anti-p300 antibodies. The results revealed that the low level of interaction between cellular Smad3 and p300 in fibroblasts was strongly enhanced by TGF-β (Fig. 7). In contrast, in fibroblasts overexpressing p53, TGF-β-induced enhancement of Smad3-p300 complex formation was reproducibly attenuated (Fig. 7).

p300 rescues TGF-β stimulation of collagen gene transcription in presence of p53:
As ectopic expression of p53 interfered with TGF-β-induced Smad-p300 complex formation on SBE, next we investigated whether repression of Smad-dependent collagen gene transcription by p53 could be reversed by ectopic expression of p300. Therefore, fibroblasts were cotransfected with expression vectors for p53 and p300 along with 772COL1A2-CAT reporter construct, and incubated with TGF-β for 48 h. The results of transient transfection assays indicated that forced expression of p300 by itself caused potent stimulation of COL1A2 promoter activity both in the absence and presence of TGF-β (Fig. 8). Importantly, p300 rescued TGF-β-induced stimulation of COL1A2 promoter activity in the presence of p53. The ability of ectopic p300 to overcome the inhibitory effect of p53 suggested that sequestration of cellular p300 or Smad or p300-Smad3 complex by overexpressed p53 may have been responsible for transcriptional repression.

DISCUSSION:
The multifunctional cytokine TGF-β stimulates collagen synthesis in dermal fibroblasts.
Dysregulation of intracellular TGF-β signaling has been implicated in excessive synthesis and accumulation of collagen in scleroderma and related fibrotic conditions (8). In contrast to its antagonistic effect on collagen gene expression induced by TGF-β, p53 stimulates the expression of PAI-1 gene and further enhanced the TGF-β-induced expression .
The stimulation of PAI-1 gene expression by p53 was previously shown to involve binding of p53 to a cognate cis-element within the PAI-1 gene promoter (13). Therefore, the present results along with earlier reports (13,32,50) suggest that p53 has differential effects on distinct TGF-β-regulated target gene expression. The expression of the MMP-2 gene was also stimulated by p53 via a similar mechanism (23). Furthermore, in cells lacking p53, TGF-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from β/Activin failed to induce the expression of PAI-1 and MMP-2 genes (32). These observations fit a general model proposing convergence of p53 and TGF-β signaling. On the other hand, the present results indicate that collagen gene expression and its stimulation by TGF-β were inhibited by p53. As summarized in Table 1, collagen and selected ECM genes lack consensus p53-binding element in their promoter (13,21,22,24,51). Based upon these observations, we suggest a model whereby the presence or absence of a consensus p53 recognition site within the promoter may determine whether p53 enhances or represses the transcription of a given gene. In this model, direct interaction of p53 with its cognate DNA binding elements results in activation of target gene transcription, whereas interaction of p53 with transcriptional activators/coactivators in the absence of DNA binding results in transcriptional repression of genes lacking the p53 element. The results from the present studies demonstrate that repression of the activity of COL1A2 and Smad-driven minimal promoter constructs by p53 was not due to direct DNA interactions, as these reporter genes lack consensus p53 binding sequences. In addition, mutants of p53 lacking a DNA-binding domain were still able to repress promoter activity. Furthermore, the inhibitory effect of p53 was not associated with altered levels or ligand-dependent activation of R-Smads, suggesting that p53 modulated their transcriptional activities. Moreover, in contrast to previous results showing p53 induction of inhibitory Smad in cancer cells (52), no evidence of stimulation of Smad7 was seen. To our knowledge, this is the first report describing negative modulation of Smad-dependent transcriptional regulation of collagen gene expression by p53.
As TGF-β-induced activation of R-Smads remained unaltered in presence of excess p53, and p53 interacts with Smads (32 and our unpublished data), sequestration of Smad by p53 from active transcriptional complex may contribute to p53-mediated suppression of collagen gene expression. Indeed present study demonstrated that induced cellular p53 reduced the TGF-β-induced interaction of Smad3 and Smad3 interacting p300 with its cognate DNA binding element. These results suggest that p53 mediated suppression of TGF-β-induced gene transcription is at the levels of transcriptional initiation complex formation. Furthermore, the TGF-β-induced physical interaction of cellular Smad3 with p300 was disrupted in presence of overexpressed p53. Several lines of evidence suggest that the transcriptional coactivator p300 is required for p53-dependent transactivation or transrepression of target genes (48 and references therein). We showed here that ectopic overexpression of p300 rescued the p53-      Nuclear extracts were analyzed by DNA-affinity precipitation assays using biotinylated SBE probe. Protein complexes eluted from DNA were analyzed by immunoblots using anti-p300 and anti-Smad3 antibodies. Representative immunoblots are shown. The p53 levels in nuclear extract were determined by immunoblot analysis using antibodies against p53 or Smad4.

7.
Overexpressed p53 disrupts Smad3-p300 interaction: Fibroblasts were transfected with p53 expression vector or empty vector. Following 1 h incubation with TGF-β (12.5 ng/ml), whole cell lysates were prepared and immunoprecipitated with anti-p300 antibody, followed by immunoblot with anti-Smad1/2/3 or anti-p300 antibodies. A representative immunoblot is shown. Quantitative results of the Smad3 levels normalized to the levels of p300 are shown in lower panel.

8.
Overexpressed p300 partially reverses the p53-mediated repression: Fibroblasts were transfected with 772COLA2-CAT reporter construct or empty vector along with p53 and p300 expression vectors. Following 48 h incubation with TGF-β (12.5 ng/ml), CAT activities were determined and normalized with equal amount of proteins. Transfection efficiency was monitored using renilla luciferase assay. The results shown as the means ± S.D. of duplicate determinations are representative of two independent experiments. Open bars, untreated fibroblasts; closed bars, TGF-ß-treated fibroblasts. * p< 0.005.