Novel Negative Regulatory Element in the Platelet-derived Growth Factor B Chain Promoter That Mediates ERK-dependent Transcriptional Repression*

Platelet-derived growth factor (PDGF), which consists of an A and/or B chain, stimulates migration and prolif-eration in vascular smooth muscle cells as well as a large number of other cell types. Investigations over recent years have defined roles for several positive regulatory transcription factors in the PDGF-B promoter. How-ever, little is known about the transcriptional mechanisms that negatively regulate this gene. Here, we used transient transfection and 5 * deletion analysis to define a specific region in the PDGF-B promoter-mediating repression in vascular smooth muscle cells. Gel retardation assays revealed this region is bound by nuclear protein(s) in a specific manner. Supershift assays excluded the direct association of Sp1, Sp3, and Egr-1. Mutation of the negative regulatory element no longer supported nucleoprotein complex formation and, when introduced into the PDGF-B promoter, rescued the promoter from repression. Promoter activity was also restored by transfection of oligonucleotide decoys bearing the repressor binding site. The MEK1/2 inhibitor, PD98059, and a dominant negative construct generating inactive ERK1 increased reporter expression driven by the PDGF-B promoter. In contrast, the MEK inhibitor had no effect on the activity of the mutant PDGF-B promoter. These effects were cell type-specific, since neither suppression of the PDGF-B promoter nor nucleoprotein complex formation was observed in vascular endothelial cells. These findings define a distinct negative regulatory element in the PDGF-B promoter that interacts with nuclear protein(s) and inhibits PDGF-B promoter-dependent gene expression in an ERK-dependent Electrophoretic mobility shift assay (EMSA) binding reactions for gel shift assays were performed in m l m M m M NaCl, 1 m M 1 m M EDTA, 5% glycerol, 1 m g of salmon sperm DNA, 1 m g of poly(dI-dC), 32 P-labeled oligonucleotide probe, and 5 m g of nuclear extract (determined by BCA protein assay). Binding reactions involving Sp1/Egr-1 binding conditions have been previously The reaction was incubated for 35 min at 22 °C. In supershift studies 0.5 m l of the appropriate affinity-purified antipeptide polyclonal antibody was incubated with the binding mix for 10 min before the addition of the probe. Bound complexes were separated from free probe by loading samples onto a 6% nondenaturing polyacrylamide gel and electrophoresed at 100 V for 3.5 h. The gels were vacuum-dried subjected to autoradiography overnight at -80 °C. Oligonucleotide decoy strategy pup SMCs were transfected with 8 m g of construct d18, alone in combination with 100 n M or 1 m M Oligo B A or Oligo B Am1 . Transfections were performed using FuGENE6, as described Twenty-four h post-transfection, cell lysates were for assessment CAT activity. ascertain oligonu- cleotide decoys radioactivity in various times

Platelet-derived growth factor (PDGF) 1 comprises a disulfide-linked homo-or heterodimer of an A and/or B chain. It is a potent mitogen and chemoattractant for cells of mesenchymal origin. A large number of cell types, including vascular smooth muscle cells, produce PDGF (1), suggesting that PDGF may be involved in autocrine/paracrine growth loops. Several lines of evidence implicate PDGF B-chain in vascular pathologic settings and in remodeling after mechanical injury to the artery wall. PDGF-B is associated with smooth muscle cells in human atherosclerotic plaques (2,3) and post-angioplasty restenotic lesions (4). In rat and pig models of balloon angioplasty, PDGF-BB stimulates intimal thickening (5)(6)(7).
Negative regulation of PDGF-B gene expression at the level of transcription is poorly understood. To date, ZNF174 is the only known transcriptional repressor with the capacity to down-regulate the activity of the PDGF-B promoter (20). Here, we used a variety of approaches to search the PDGF-B promoter for additional cis-acting silencer elements. Using gel shift, mutational, decoy, and transient transfection analysis, we have identified a novel negative regulatory element in the PDGF-B promoter that mediates repression of the promoter in a sequence-specific and MEK/ERK-dependent manner.
Transient Transfection Analysis-Pup SMCs were seeded in 100-mm tissue culture plates for 48 h before transfection. When approximately 60 -70% confluent, the cells were transfected with 8 g of the indicated promoter-reporter plasmids. Transfections were performed using Fu-GENE6 (Roche Molecular Biochemicals). A precipitate was formed using 3 l of FuGENE/g of transfected DNA, and the transfection mix was made up to 1 ml with serum-free Waymouth's medium. After incubation at 22°C for 10 min, the DNA/FuGENE mixture was added to cells containing 4 ml of complete Waymouth's medium. Two days post-transfection, cell lysates were prepared for assessment of chloramphenicol acetyltransferase (CAT) activity as described (13). The concentration of protein in the cell lysates were assessed using the BCA protein assay kit and used to correct CAT reporter activity.
Plasmid Constructs-A mutant form of plasmid d18 (13), md18 Am1 , was constructed using the Quick Change site-directed mutagenesis kit (Stratagene) in accordance with the manufacturer's instructions by cloning oligonucleotide B Am1 (sense orientation only) into the Ϫ230/ Ϫ200 region of the PDGF-B promoter (relative to the TATA box). All other PDGF-B promoter constructs have been described previously (13).
Preparation of Nuclear Extracts-SMCs were washed twice with phosphate-buffered saline, pH 7.4, at 4°C and removed from the culture dish by scraping. Cells were pelleted by centrifugation at 300 ϫ g for 10 min and 4°C. The pellet was resuspended in ice-cold phosphatebuffered saline, pH 7.4, and the suspension was transferred to Eppendorf tubes. The cells were repelleted by spinning at 18,000 ϫ g for 20 s at 4°C. The cells were lysed by resuspending the pellet in an ice cold Buffer A, which consisted of 10 mM HEPES, pH 8, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol , 0.5 mM phenylmethylsulfonyl fluoride, 200 mM sucrose, 0.5% Nonidet P-40, 1 mg/ml leupeptin, and 1 mg/ml aprotinin. The samples were incubated on ice for 5 min followed by recentrifugation. The nuclei were lysed in ice-cold Buffer C, which consisted of 20 mM HEPES, pH 8, 100 mM KCl, 0.2 mM EDTA, 20% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mg/ml leupeptin, and 1 mg/ml aprotinin. Cellular debris was removed by centrifugation, and the supernatant containing the crude nuclear extract was combined with an equal volume of ice-cold Buffer D (20 mM HEPES, pH 8, 100 mM KCl, 0.2 mM EDTA, 20% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mg/ml leupeptin, and 1 mg/ml aprotinin). Extracts were stored at Ϫ80°C until use.
Electrophoretic mobility shift assay (EMSA) binding reactions for gel shift assays were performed in 20 l of 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, 1 g of salmon sperm DNA, 1 g of poly(dI-dC), 32 P-labeled oligonucleotide probe, and 5 g of nuclear extract (determined by BCA protein assay). Binding reactions involving Sp1/Egr-1 binding conditions have been previously described (18). The reaction was incubated for 35 min at 22°C. In supershift studies 0.5 l of the appropriate affinity-purified antipeptide polyclonal antibody was incubated with the binding mix for 10 min before the addition of the probe. Bound complexes were separated from free probe by loading samples onto a 6% nondenaturing polyacrylamide gel and electrophoresed at 100 V for 3.5 h. The gels were vacuum-dried and subjected to autoradiography overnight at -80°C.
Oligonucleotide decoy strategy pup SMCs were transfected with 8 g of construct d18, either alone or in combination with 100 nM or 1 mM Oligo B A or Oligo B Am1 . Transfections were performed using FuGENE6, as described above. Twenty-four h post-transfection, cell lysates were prepared for assessment of CAT activity. To ascertain that the oligonucleotide decoys reached the nucleus, radioactivity in the nuclear and cytoplasmic compartments was assessed at various times following transfection with 500 fmol (1 ϫ 10 6 cpm) of 32 P-Oligo B A or 32 P-Oligo B Am1 .

RESULTS AND DISCUSSION
Evidence for the Existence of a Negative Regulatory Element in the Proximal PDGF-B Promoter-WKY12-22 pup SMCs, which express PDGF-B mRNA (21), were transiently transfected with a series of CAT reporter constructs driven by various-sized fragments of the PDGF-B promoter ( Fig. 1, upper panel). Basal CAT activity was readily detectable and not significantly different in SMCs transfected with constructs d77 and d26, bearing 82 and 153 bp of PDGF-B promoter sequence (relative to the TATA box), respectively ( Fig. 1, lower panel). In contrast, construct dNco (Ϫ227), which bears an additional 74 bp of PDGF-B promoter sequence upstream ( Fig. 1, upper panel) was, like construct d18 (Fig. 1, upper panel), virtually inactive ( Fig. 1, lower panel). These data contrast with our previous observations using serial 5Ј deletion analysis in bovine aortic endothelial cells, where no major difference was observed in CAT activity driven by these reporter vectors (13). The inability of construct dNco to mediate basal PDGF-B promoter-dependent expression in SMCs prompted us to explore the apparent negative regulatory role of this region in the promoter.
Nuclear Protein(s) Interact with the Negative Regulatory Element in the Proximal PDGF-B Promoter-We next performed EMSA to gain insight into the mechanisms underlying the negative regulation of the PDGF-B promoter. Since the preceding findings demonstrated that the putative repressor site is located between the 5Ј PDGF-B promoter end points in constructs dNco (Ϫ227) and d26 (Ϫ153), we synthesized a series of overlapping double-stranded oligonucleotides spanning this region (Table I). These 32 P-labeled oligonucleotides were incubated with nuclear extracts of pup SMCs, and the adducts were resolved by nondenaturing gel electrophoresis. 32 P-Oligo B A (bp Ϫ230/Ϫ201) produced a single, discrete nucleoprotein complex ( Fig. 2A, arrow) that was virtually absent when the probe was substituted with 32 P-Oligo B B (bp Ϫ205/Ϫ176) and 32 P-Oligo B C (bp Ϫ180/Ϫ151) ( Fig. 2A).
Competition experiments were performed to determine whether the interaction between nuclear protein and oligonucleotide was specific. A 35-fold molar excess of unlabeled Oligo B A abrogated the formation of the nucleoprotein complex (Fig.  2B, third lane). In contrast, the same fold molar excess of two unrelated oligonucleotides, Oligo A (22) and Oligo LKSSREm3 (19), had no effect (Fig. 2B).
The PDGF-B promoter sequence between the 5Ј end points in d26 and dNco bears a consensus binding element for Sp1. Previously, we determined that Sp1 as well as Sp3 and Egr-1 Ϫ230-gtccatggtcaagtgtaggaggggcgggac-201 Olgio B Am3 Ϫ230-gtccatggtcactgtgctgcttttatggac-201 interact with a second site (bp Ϫ30/Ϫ13) located downstream in the promoter, which positively regulates basal PDGF-B expression (18). Supershift analysis was performed with 32 P-Oligo B A using antibodies with specificity directed to these nuclear factors. However, the nucleoprotein complex was unaltered by the presence of either antibody (Fig. 3A). To confirm the lack of a direct physical association of Sp1 with 32 P-Oligo B A , we performed EMSA and supershift analysis using binding conditions suited to Sp1 (22). These conditions did not support the formation of a nucleoprotein complex involving 32 P-Oligo B A and Sp1 (Fig. 3B). In contrast, multiple nucleoprotein complexes were observed using a second oligonucleotide, 32 P-Oligo A (Fig. 3B), as described previously (18,22), thus demonstrating the pres-  3. Nuclear Sp1, Egr-1, or Sp3 do not interact with the negative regulatory element in the PDGF-B promoter. A, EMSA was performed with 32 P-Oligo B A and pup SMC nuclear extracts under the conditions indicated under "Experimental Procedures." Where indicated, 1 g of the appropriate antibody was incubated with the binding mixture for 10 min before the addition of the 32 P-labeled probe. B, EMSA was performed with pup SMC nuclear extracts and 32 P-Oligo B A or 32 P-Oligo A using conditions suited for Sp1, Egr-1, and Sp3 (22). Where indicated, 1 g of the appropriate antibody (Sp1, Sp3, or Smad1) was incubated with the binding mixture for 10 min before the addition of the 32 P-labeled probe. NE denotes nuclear extract. The data are representative of two independent determinations. ence of Sp1 in the nuclear extracts.
Base Pairs Ϫ227/Ϫ221 in the PDGF-B Promoter Support Nucleoprotein Complex Formation-To ascertain the specific site in the PDGF-B promoter bound by the repressor, we designed a series of oligonucleotides bearing transversion mutations in Oligo B A . Mutations were created in three different regions: bp Ϫ227/Ϫ221(Oligo B Am1 ), Ϫ219/Ϫ213 (Oligo B Am2 ), and Ϫ211/Ϫ205 (Oligo B Am3 ) ( Table I). 32 P-Oligo B A and the three mutant 32 P-labeled oligonucleotides were incubated with pup SMC nuclear extracts before EMSA. Although the nucleoprotein complex was observed with 32 P-Oligo B A (Fig. 4), it was completely abolished using 32 P-Oligo B Am1 (Fig. 4) and was only faintly observed using 32 P-Oligo B Am2 (Fig. 5). In contrast, Oligo B Am3 was unable to affect the appearance of this nucleoprotein complex (Fig. 4). These findings demonstrate the requirement of bp Ϫ227/Ϫ221 (5Ј-CATGGTCA-3Ј) in the repressed region of the PDGF-B promoter for optimal nucleoprotein complex formation.
Mutation of the Ϫ227/Ϫ221 Site in d18 Rescues CAT Activity-To determine the functional consequence of the Ϫ227/ Ϫ221 mutation in the PDGF-B promoter, this sequence was introduced into construct d18 before transient transfection analysis in pup SMCs. CAT activity in SMCs transfected with construct d18 was negligible (Fig. 5) and readily detectable when construct d26 was used, consistent with earlier findings (Fig. 1, lower panel). Construct md18Am1, bearing the Ϫ227/ Ϫ221 mutation, was basally active and produced CAT activity at levels even higher than construct d26 (Fig. 5). These data provide functional evidence for the existence of a negative regulatory element at position Ϫ227/Ϫ221 in the PDGF-B promoter.
In Vivo Competition of Construct d18 Using Decoys Rescues CAT Activity-An oligonucleotide decoy strategy was used to confirm our observations using the mutant PDGF-B promoter. Pup SMCs were cotransfected with construct d18 plasmid and either unlabeled Oligo B A , which binds to the repressor in a specific manner (Figs. 2, A and B, and 4), or Oligo B Am1 , which is unable to bind to the nuclear protein (Fig. 4). Transfection experiments revealed that transfection of Oligo B A increased CAT activity generated by construct d18 (Fig. 6A), whereas Oligo B Am1 failed to influence reporter activity (Fig. 6A). These findings demonstrate the sequence-specific nature of transcrip-tional repression at the Ϫ227/Ϫ221 site in the PDGF-B promoter. The wild-type oligonucleotide serves as a decoy that competes with the authentic site in the promoter for interaction with the repressor.
To ensure that both oligonucleotides localized to the nucleus of these cells independently of differences in sequence, we traced 32 P-labeled Oligo B A or Oligo B Am1 after transfection over time in both cytoplasmic and nuclear fractions. Both oligonucleotides demonstrated a similar spatial and temporal pattern of cellular localization, entering the cytoplasm before localizing in the nucleus shortly thereafter (Fig. 6B). After 12 h, the majority of each oligonucleotide had localized in the nucleus (Fig. 6B). Therefore, whereas both oligonucleotides reached the nuclear compartment with similar kinetics, only Oligo B A , which bound the repressor (Fig. 3A), rescued the PDGF-B promoter from transcriptional repression.
The Repressor Is Cell-specific-Previous examination of the PDGF-B promoter by our group in vascular endothelial cells revealed no difference in CAT activity driven by a series of PDGF-B promoter-reporter constructs bearing or not bearing the negative regulatory element (23). Indeed, in contrast to findings in pup SMCs, no significant difference was observed in FIG. 7. Repression of PDGF-B promoter via the negative regulatory element is cell-specific. A, Pup SMC and bovine aortic endothelial cells (BAEC) cells were transiently transfected with 8 g of d18 or md18Am1. CAT activity after 24 h was normalized to the concentration of protein in the cell lysate. Error bars represent S.E. of the mean. B, comparative EMSA using pup SMC and bovine aortic endothelial cells nuclear extracts and 32 P-Oligo B A . Binding reactions and nondenaturing gel electrophoresis were performed as described under "Experimental Procedures." C, EMSA confirms the integrity of pup SMC and bovine aortic endothelial cells nuclear extracts. Extracts of both cell types were incubated with 32 P-Oligo A under conditions previously described for this probe (22). Nucleoprotein complexes were visualized by autoradiography. The data are representative of two independent determinations. activity generated by construct d18 or md18Am1 (Fig. 7A). Given the present data, we hypothesized that the lack of repression via this site in endothelial cells may be the consequence of low expression or complete absence of the repressor in this cell type.
We assessed whether 32 P-Oligo B A could support nucleoprotein complex formation with endothelial nuclear extracts in EMSA. Pup SMC extracts, as expected, formed the nucleoprotein complex (Fig. 7B). However, this complex was completely absent when endothelial nuclear extracts were used (Fig. 7B). To ensure that these findings were not merely the consequence of unequal protein loading or degradation of the endothelial extract, we performed EMSA with each extract using 32 P-Oligo A (Fig. 7C). This produced an identical electrophoretic pattern of shifted complexes in both cell types, entirely consistent with previous observations (18,22). These data indicate the cellspecific nature of Ϫ227/Ϫ221-directed repression and the capacity of this element to interact with nuclear protein(s).
Repression of the PDGF-B Promoter Is Mitogen-activated Protein (MAP) Kinase-dependent-Previous studies have determined that certain transcriptional repressors are regulated by MAP kinases. For example, the repressor activity of the Ets family member, GETS-1, is abrogated by overexpression of Ras/MAP kinase or by mutation of Ser-405 MAP kinase phosphorylation site in GETS-1 (24). Similarly, MAP kinase phosphorylation of repressor protein, Bcl-6, leads to its rapid degradation via the ubiquitin/proteosome pathway (25). ERK1/2 is basally active in pup SMCs (data not shown). We investigated whether repression of the PDGF-B promoter via the Ϫ227/ Ϫ221 site was dependent upon ERK. PD98059, an inhibitor of MEK1/2, the upstream activator of ERK, was incubated with SMCs transfected with construct d18. The flavone rescued reporter activity driven by the PDGF-B promoter (Fig. 8A), as previously observed using the wild-type oligonucleotide decoy (Fig. 6A). In contrast, PD98059 had no effect on higher basal expression generated by construct md18Am1 (Fig. 8A). To demonstrate the involvement of ERK in PDGF-B transcription, we overexpressed a dominant-negative mutant form of ERK1 in transient transfection setting with construct d18. CAT activity generated by construct d18 increased upon cotransfection with dominant negative ERK1 (Fig. 8B). In contrast, the PDGF-B promoter was unaffected when an identical amount of the backbone control, pcDNA3, was used (Fig. 8B). Thus, unlike the paradigm of GETS-1 (24) and Bcl-6 (25), repression of the PDGF-B promoter conferred by the Ϫ227/Ϫ221 element requires active Raf/MEK/ERK (Fig. 8B).
In this paper, we have identified a novel negative regulatory element in the PDGF-B promoter, located 227/221 bp upstream of the TATA box, that binds nuclear proteins in a specific fashion and inhibits PDGF-B promoter-dependent gene expression in a Raf/MEK/ERK-dependent manner. Gel retardation and supershift assays excluded the direct association of Sp1, Sp3, and Egr-1 with this element. Repression of the PDGF-B promoter was rescued by (i) mutation of the Ϫ227/Ϫ221 site that ablates nucleoprotein complex formation, (ii) cotransfection with double-stranded oligonucleotide decoys, and (iii) pharmacological and dominant negative inhibitors of the Raf/ MEK/ERK pathway. Comparative binding and transient transfection studies in endothelial cells revealed that repression of the PDGF-B promoter via this element is cell type-specific. These findings thus define a key negative regulatory element in the PDGF-B promoter.
The Ϫ227/Ϫ221 repressor element in the PDGF-B promoter does not fit the consensus motif for any known transcription factor. Competition studies by EMSA in vitro and decoy experiments in vivo illustrate the specific nature of this element.
That mutation ablates nucleoprotein complex formation and rescues the promoter from inactivity suggests that the repressor may function by active rather than passive means. For example, active transcriptional repressors that directly contact DNA include the human Kruppel-related factor YY1 (26), the Wilms' tumor gene product (WT1) (27), and the human bZIP protein E4BP4 (28). The repressor may also inhibit PDGF-B expression by inhibiting the activity of positive transcriptional regulators such as Sp1, Sp3, and Egr-1, whose sites are located downstream in the proximal region of the promoter. This may involve protein-protein interactions while bound to the promoter in a three-dimensional setting. The MAP kinase dependence of repression via this site demonstrates that phosphorylation is a key event in the regulation of PDGF-B transcription. ERK may directly phosphorylate the repressor in the nucleus and/or modulate its stability or cellular localization. Isolation and characterization of this cell-restricted repressor should provide important insights into the molecular mechanisms negatively regulating PDGF-B transcription in vascular cells.