YY1 represses TGF-β and BMP immediate-early gene responses. (A) YY1 represses PAI-1 induction by TGF-β1. Northern analysis of PAI-1 mRNA from Mv1Lu cells mock treated (−) or treated (+) for 24 h with 10 ng of TGF-β1 per ml in the absence of YY1 (−) or in the presence of increasing doses of YY1 (Ad-YY1) (triangle; MOI of 20 or 100) or control GFP (Ad-GFP) (MOI of 20 or 100) adenovirus is shown. Sample loading and blotting efficiency was monitored with β-actin mRNA on the same blot. (B) YY1 represses Id-1 induction by TGF-β1. Immunoblot analysis of endogenous Id-1 expression in MDA-MB-468 cells infected with the indicated MOI of recombinant adenoviruses and then treated with vehicle (−) or 2.5 ng of TGF-β1 per ml (+) for 1 h is shown. Arrows mark the positions of endogenous Id-1, adenoviral YY1, adenoviral Flag-Smad4 (F-Smad4), endogenous phosphorylated Smad2 (Smad2-P), and endogenous β-catenin, which serves as a loading and blotting efficiency control. (C) YY1 represses Id-1 induction by BMP-7. Immunoblot analysis of endogenous Id-1 expression in MDA-MB-468 cells is as for panel B but in response to 300 ng of BMP-7 per ml. Infection conditions were as for panel B, and arrows mark the same proteins as in panel B except for endogenous phosphorylated Smad1 (Smad1-P) instead of Smad2. Asterisks mark nonspecific protein bands. (D) YY1 represses PAI-1 promoter induction by TGF-β1. PAI-1 (800)-luciferase assays were performed in HepG2 cells with increasing amounts of YY1 plasmid (triangle; 0.1 to 1.0 μg) in the absence (−) or presence (+) of overnight stimulation with 2.5 ng of TGF-β1 per ml. The averaged data for normalized reporter activity are shown with standard errors derived from triplicate transfections repeated at least twice independently. (E) YY1 represses the Smad-specific 12xCAGA promoter. The assays were similar to those for panel D, with the 12xCAGA-luciferase reporter in HepG2 cells after overnight stimulation with 2.5 ng of TGF-β1 per ml. Data are reported as in panel D. (F) YY1 represses the Smad-specific Id1-BRE₂ promoter. The assays were similar to those for panel D, with the Id1-BRE₂-luciferase reporter in C2C12 cells after overnight stimulation with 100 ng of BMP-6 per ml. Data are reported as in panel D.

YY1 inhibits the DNA-binding activity of Smads without affecting their oligomerization. (A) YY1 overexpression does not disrupt TGF-β1-induced Smad oligomers. Analysis of coimmunoprecipitation of endogenous phosphorylated Smad2 (Smad2-P) with Flag-tagged adenoviral Smad4 in lysates from infected MDA-MB-468 cells (as for Fig. 1B) after treatment with 2.5 ng of TGF-β1 per ml for 1 h is shown. Total cell extracts were immunoblotted with antibodies against endogenous phosphorylated Smad2 (Smad2-P), adenoviral Flag-tagged Smad4 (F-Smad4), YY1, and the control protein β-catenin. An asterisk marks immunoglobulin heavy-chain bands. (B) YY1 does not bind directly to the Smad-binding element. EMSA was performed with no protein (−), increasing amounts of recombinant YY1 (triangle; 125, 250, and 500 ng), or 125 ng of YY1 (+) and radiolabeled 4xCAGA and adeno-associated virus P5 promoter (AAV-P5) oligonucleotides. (C) YY1 inhibits DNA binding of Smad3 and Smad4. EMSA was performed with 250 ng (+) or 100 and 250 ng (triangle) of recombinant YY1, 100 ng of phosphorylated Smad3 (Smad3-P) (+), 300 ng of Smad4 (+), and radiolabeled 4xCAGA oligonucleotide. The exposure times in each autoradiogram were 24 h (lanes 1 to 4), 10 h (lanes 5 and 6), and 5 h (lanes 7 to 9). In panels B and C the specific band shifts are shown with arrows. (D) YY1 inhibits binding of nuclear Smad4 to SBE DNA. 4xCAGA DNA precipitation (p) assays were performed with 293T cell extracts coexpressing Myc₆-Smad4 plus YY1 in the absence or presence of caALK-5. DNA precipitates were immunoblotted (Ib) with anti-Myc antibody, and immunoblots of total cell extracts are shown (Total), as marked by arrows. (E) YY1 inhibits recruitment of endogenous nuclear Smad4 to the SBE-rich area of the endogenous PAI-1 promoter. ChIP assays were performed with the indicated antibodies and PCR amplification of the distal and SBE-rich human PAI-1 promoter from HaCaT cells, infected transiently with the indicated adenoviruses (MOI of 20) and treated (+) or not (−) with 5 ng of TGF-β1 per ml for 2 h. Control ChIPs with an unrelated (Flag) antibody and 2% of the chromatin preparation, amplified without prior immunoprecipitation (INPUT), are shown. A control immunoblot (Ib) of the infected cells is also shown, demonstrating the levels of YY1 overexpression. Numbers below the top panel indicate densitometric values of the specific bands normalized over the corresponding input bands and expressed relative to the control band of lane 1, which is set to 1. (F) YY1 represses TGF-β1-induced PAI-1 mRNA levels in HaCaT cells. Semiquantitative RT-PCR assays with mRNA isolated from duplicate HaCaT cell plates for the ChIP assays demonstrate proper PAI-1 gene regulation in these cells. GAPDH serves as a control in the RT-PCRs. The bottom panel shows average expression values and standard deviations of PAI-1 mRNA levels from the same experiment and after normalization to control GAPDH mRNA levels derived from real-time quantitative PCR experiments.

YY1 interacts physically with Smads in vitro. (A) Schematic representation of Smad proteins with their conserved MH1, the DNA-binding β-hairpin loop (black boxes), MH2, and the less conserved linker sequences. The strength of Smad protein interaction with YY1 is quantified based on the data in panels B and C. (B) Endogenous YY1 from C2C12 cells interacts with the indicated GST-Smad fusions with different relative affinities (lanes 2 to 5). A total C2C12 lysate immunoblot is shown for comparison (lane 6). Ponceau S staining of the immunoblot shows the immobilized GST-Smads and GST protein. (C) YY1 interacts with the MH1 domain of Smad4. Experiment similar to those for panel B were performed with endogenous YY1 from HaCaT cells. Immobilized GST fusion proteins are shown in a Coomassie brilliant blue (CBB) stain of the gel. In panels B and C, arrows mark the specific protein bands.

YY1 and Smad4 form complexes in vivo. (A) Smad4 coimmunoprecipitates with endogenous YY1. The indicated proteins were expressed in COS-7 cells, immunoprecipitated (Ip) with anti-YY1 antibody, and immunoblotted (Ib) for Smad4 with anti-Myc antibody (top panel) or with the same anti-YY1 antibody (middle panel). The lower panel shows a total-extract immunoblot with anti-Myc for Smad4. (B) Coimmunoprecipitation of endogenous Smad4 and YY1. HaCaT total cell extracts were collected after stimulation (or not) with 10 ng of TGF-β1 per ml for 1 h, immunoprecipitated without or with specific anti-YY1 antibody, and immunoblotted with anti-Smad4 antibody (top panel). Total extracts were immunoblotted with the same anti-Smad4 antibody (middle panel) or with anti-YY1 antibody (bottom panel). (C) Coimmunoprecipitation of endogenous nuclear Smad4 and YY1. HaCaT nuclear extracts were collected after stimulation (or not) with 5 ng of TGF-β1 per ml for 2 h, immunoprecipitated (Ip) with specific anti-Smad4 antibody, and immunoblotted (Ib) with anti-YY1 antibody (top panel) or the same anti-Smad4 antibody (third panel). Immunoprecipitation with a control matched preimmune (PI) rabbit serum followed by anti-YY1 immunoblotting is shown in the second panel. Nuclear extracts were directly immunoblotted with the same anti-YY1 (fourth panel), anti-Smad4 (fifth panel), or anti-phospho-Smad2 (sixth panel) antibodies.

The zinc finger domain of YY1 is critical for repression of Smad activity. (A) Schematic representation of the domain structure of YY1 (GA- and GK-rich, Gly/Ala- and Gly/Lys-rich domains, respectively). (B) Interaction assay of wild-type (WT) and various deletion mutants of YY1 overexpressed in COS-7 cells with semipurified GST-Smad4 fusion protein and immunoblot (Ib) detection with anti-YY1 antibody. The deletions of human YY1 amino acids are shown above each lane. The efficiency of expression and immunodetetection of YY1 proteins by immunoblotting of total lysate is shown (Input). Asterisks mark the specific YY1 protein bands. (C) Signaling assays in Mv1Lu cells with the 12xCAGA-luciferase reporter and the same panel of YY1 deletion mutants. The schematic representation of YY1 deletion mutants uses the domain conventions of panel A. Data are presented as in Fig. 1D. Basal and TGF-β-induced reporter levels are shown by open and filled bars, respectively. The dashed line indicates the TGF-β1-induced (overnight with 5 ng of TGF-β1 per ml) reporter level, and the numbers at the right indicate the respective fold induction by TGF-β1. (D) Expression levels of the YY1 deletion mutants used for panel C from duplicate transfections of COS-7 cells and immunoblot (Ib) analysis with anti-YY1 antibody.

Endogenous YY1 restricts TGF-β and BMP signaling levels. (A and B) Antisense YY1 reverses the inhibitory effect of YY1 and increases the TGF-β1 responsiveness of a Smad-dependent gene promoter in Mv1Lu cells. (A) Induction of the 12xCAGA reporter by overnight stimulation with 0.5 ng of TGF-β1 per ml in the presence of a constant amount (0.5 μg) of sense YY1 (S-YY1) and increasing amounts (0.03, 0.1, 0.3, and 1.0 μg) of antisense YY1 (AS-YY1) plasmids. (B) Increased responsiveness of the 12xCAGA reporter to TGF-β1 (0.5 ng/ml) in the absence (empty bars) or presence (full bars) of antisense YY1 (0.3 μg DNA). (C) RNAi against endogenous YY1 dramatically enhances the inducibility of the 12xCAGA-luciferase reporter by overnight stimulation with 0.5 ng of TGF-β1 per ml of HaCaT cells transfected with mock or YY1 siRNA oligonucleotides. (D) RNAi against YY1 efficiently decreases endogenous YY1 protein levels in C2C12 cells. Immunoblotting with the indicated antibodies of cell extracts from C2C12 cells transfected with no (−), an unrelated (control), or different doses of (triangle) specific YY1 siRNA oligonucleotides. (E) RNAi against YY1 dramatically enhances the inducibility of the Id1-BRE₂-luciferase reporter by overnight stimulation with 300 ng of BMP-7 per ml in C2C12 cells transfected with mock, control siRNA or YY1 siRNA oligonucleotides. (F) RNAi against endogenous YY1 enhances the TGF-β1-inducible levels of endogenous PAI-1 protein. Immunoblotting of HaCaT cell extracts transfected with no (mock), control siRNA, or specific anti-YY1 siRNA oligonucleotides and treated (+) or not (−) with 2.5 ng of TGF-β1 per ml for 4 h with the indicated antibodies is shown. The asterisk marks nonspecific protein bands. The numbers below the top two panels indicate densitometric values of the specific bands normalized over the corresponding α-tubulin bands and expressed relative to the control band of lane 1, which is set to 1. (G) RNAi against endogenous YY1 enhances the inducibility of the endogenous PAI-1 gene by TGF-β1. Real-time PCR analysis of human PAI-1 mRNA from HaCaT cells transfected with no (mock), control siRNA, or specific anti-YY1 siRNA oligonucleotides and treated (+) or not (−) with 5 ng of TGF-β1 per ml for 16 h is shown. The bar graph plots average expression values and standard deviations of PAI-1 mRNA levels normalized to control GAPDH mRNA levels. The data in panels A, B, C, and E are presented as in Fig. 1D.

Ectopic expression of YY1 inhibits TGF-β- and BMP-induced transdifferentiation. (A) Top panels, YY1 inhibits mammary NMuMG cell EMT induced by TGF-β1. NMuMG cells were infected with Ad-GFP (MOI, 25) or Ad-YY1 (MOI, 10) adenovirus, treated for 24 h with vehicle (control) or with 2.5 ng of TGF-β1 per ml, and subjected to actin direct fluorescence microscopy. Bottom panels, YY1 inhibits NMuMG EMT induced by the TGF-β type I receptor. EMT induced by caALK-5 receptor (MOI, 60) is blocked by coinfection with the YY1 virus at an MOI of 20 (Ad-YY1) or with the Smad7 virus at an MOI of 20 (Ad-Smad7). Bar, 20 μm. (B) YY1 blocks BMP-7-induced preosteoblast differentiation of C2C12 myoblasts. In situ ALP staining of C2C12 cells treated with vehicle (control) or 300 ng of BMP-7 per ml for 3 days after infection with Ad-GFP (MOI, 20) or Ad-YY1 (MOI, 4) virus is shown. Bar, 80 μm.

Smad-dependent myoblast-to-preosteoblast differentiation is regulated by YY1. (A) ALP staining of C2C12 cells transiently transfected with the indicated plasmids for caALK-6, Smad1, Smad4, and sense and antisense (AS) YY1. Representative photomicrographs are shown. Bar, 80 μm. (B) Bar graph of the C2C12 differentiation data from panel A, expressing the relative number of ALP-positive cells compared to the total number of cells counted in the whole transfected culture. Values are averages from duplicate cell cultures with standard errors. The percentage of ALP-positive cells detected after cotransfection with caALK-6, Smad1, and Smad4 was set at 100% (dashed line), relative to which all other conditions are plotted. Triangles represent 0.01, 0.1, and 0.3 μg of transfected YY1 and 0.1 and 0.3 μg of AS-YY1 plasmid DNA. (C) RNAi against endogenous YY1 enhances the preosteoblastic differentiation of C2C12 cells in response to BMP Smads. ALP enzymatic activity measured in cell extracts of transfected C2C12 cells with the indicated expression vectors in the absence (−) or presence (+) of YY1 siRNA oligonucleotides is shown. Average values of triplicate transfected-extract measurements with standard errors are expressed as nanomoles of released p-nitrophenol (p-NP) per minute of reaction per milligram of total protein in the extract. (D) C2C12 preosteoblastic differentiation is not altered by YY1 mutants that fail to interact with Smads. ALP enzymatic activity measured in transfected C2C12 cells with the indicated expression vectors is shown. The data are expressed as in panel C. The triangle represents 0.03, 0.1, 0.3, and 0.5 μg of transfected YY1 plasmid DNA, and the YY1 deletion mutants (0.1 μg each) are listed as in Fig. 4. The dashed line corresponds to the specific activity of the receptor-plus-Smad condition.

YY1 does not block the growth-inhibitory pathway downstream of TGF-β or BMP. (A) Dose-response growth inhibition assays with TGF-β-sensitive Mv1Lu, NMuMG, and HaCaT cells either not infected (CTL) or infected with Ad-GFP (MOI of 25), Ad-YY1 (MOI of 20), or Ad-Smad7 (MOI of 20) (S7) virus. (B) Dose-response growth inhibition assays with nonresponsive MDA-MB-468 cells infected as for panel A, except that Ad-YY1 was at an MOI of 5 and sensitivity to TGF-β1 and BMP-7 was restored by infection with Ad-Smad4 (MOI of 40) (S4). In panels A and B, the percentage of radiolabeled thymidine incorporated after TGF-β1 or BMP-7 treatment is graphed in a semilogarithmic plot relative to the level of vehicle-treated cells (set to 100%). (C) Representative immunoblot of endogenous and adenovirally expressed YY1 levels in infected NMuMG cells. An arrow marks the YY1 band, and input viral MOIs are shown above each lane. (D) Endogenous YY1 knockdown by siRNA does not affect growth inhibition by TGF-β1. Thymidine incorporation assays with HaCaT cells transiently transfected with mock or YY1 siRNA and stimulated with TGF-β1 are presented as described above.

YY1 does not block the regulation of p15, p21, and c-myc gene expression by TGF-β. (A) TGF-β1 induces p21 mRNA accumulation in the presence of high exogenous YY1 levels. Semiquantitative RT-PCR assays were performed with mRNA isolated from HaCaT cells infected with adenoviruses expressing GFP (Ad-GFP) (MOI of 200) or YY1 (Ad-YY1) (MOIs of 20 and 100) and treated overnight with vehicle (−) or 2.5 ng of TGF-β1 per ml (+). p21-specific PCR is shown along with GAPDH-specific PCR, which serves as a control. Reactions performed in the absence of reverse transcriptase (−RT) or in the presence of water (H₂O) instead of cDNA are included. An immunoblot (Ib) of the infected cells is also shown, demonstrating the dose-dependent levels of YY1 overexpression. (B) p21 protein induction by TGF-β1 is not altered by YY1 overexpression. Immunoblot analysis of HaCaT cells infected with the indicated MOI of adenoviruses, treated overnight with vehicle (−) or with 5 ng of TGF-β1 per ml and analyzed with the indicated antibodies is shown. (C) YY1 does not repress the p21 promoter. p21(−143/+8)-luciferase assays were performed in HaCaT cells transiently transfected with pcDNA3 (−) or pcDNA3-YY1 (triangle; 250, 500, and 750 ng) and treated with vehicle (−) or 2.5 ng of TGF-β1 per ml (+) for 20 h. Data are represented as in Fig. 1D, and the fold induction by TGF-β1 is indicated above the brackets. (D) YY1 does not repress the p15 promoter. p15-luciferase assays were performed in HaCaT cells transiently transfected with pcDNA3 (−) or pcDNA3-YY1 (triangle; 250, 500, and 750 ng) and treated with vehicle (−) or 5 ng of TGF-β1 per ml (+) for 20 h. Data are represented as in Fig. 1D, and the fold induction by TGF-β1 is indicated above the brackets. (E) YY1 does not interfere with c-myc repression by TGF-β1. Semiquantitative RT-PCR assays were performed with mRNA isolated from HaCaT cells infected with adenoviruses expressing GFP (MOI of 200) or YY1 (MOI of 100) and treated overnight with vehicle (−) or 5 ng of TGF-β1 per ml. c-myc- and control GAPDH-specific PCRs are shown. A YY1 immunoblot in HaCaT transfectants duplicate to those used for the RT-PCR assays is also shown. (F) YY1 does not repress the c-myc promoter. c-Myc (pHX)-luciferase assays were performed in HaCaT cells transiently transfected with pcDNA3 (−) or pcDNA3-YY1 (+) and treated with vehicle (−) or 5 ng of TGF-β1 per ml (+) for 20 h. Data are represented as in Fig. 1D, and the fold repression by TGF-β1 is indicated above the brackets. A YY1 immunoblot in HaCaT transfectants duplicate to those used for the luciferase assays is also shown.