A Novel Role of Brg1 in the Regulation of SRF/MRTFA-dependent Smooth Muscle-specific Gene Expression*

Serum response factor (SRF) is a key regulator of smooth muscle differentiation, proliferation, and migration. Myocardin-related transcription factor A (MRTFA) is a co-activator of SRF that can induce expression of SRF-dependent, smooth muscle-specific genes and actin/Rho-dependent genes, but not MAPK-regulated growth response genes. How MRTFA and SRF discriminate between these sets of target genes is still unclear. We hypothesized that SWI/SNF ATP-dependent chromatin remodeling complexes, containing Brahma-related gene 1 (Brg1) or Brahma (Brm), may play a role in this process. Results from Western blotting and qRT-PCR analysis demonstrated that dominant negative Brg1 blocked the ability of MRTFA to induce expression of smooth muscle-specific genes, but not actin/Rho-dependent early response genes, in fibroblasts. In addition, dominant negative Brg1 attenuated expression of smooth muscle-specific genes in primary cultures of smooth muscle cells. MRTFA overexpression did not induce expression of smooth muscle-specific genes in SW13 cells, which lack endogenous Brg1 or Brm. Reintroduction of Brg1 or Brm into SW13 cells restored their responsiveness to MRTFA. Immunoprecipitation assays revealed that Brg1, SRF, and MRTFA form a complex in vivo, and Brg1 directly binds MRTFA, but not SRF, in vitro. Results from chromatin immunoprecipitation assays demonstrated that dominant negative Brg1 significantly attenuated the ability of MRTFA to increase SRF binding to the promoters of smooth muscle-specific genes, but not early response genes. Together these data suggest that Brg1/Brm containing SWI/SNF complexes play a critical role in regulating expression of SRF/MRTFA-dependent smooth muscle-specific genes but not SRF/MRTFA-dependent early response genes.

There are many clinical diseases, such as atherosclerosis, hypertension and asthma that involve abnormal differentiation of smooth muscle cells. An important pathological process that occurs in these diseases is the disruption of the balance between differentiation and proliferation of smooth muscle cells (1)(2)(3)(4). Serum response factor (SRF) 4 has been shown to play an essential role in regulating smooth muscle differentiation, proliferation, and migration through its interaction with various accessory proteins (5). Smooth muscle-specific genes, such as SM ␣-actin, SM MHC, 130-kDa MLCK, SM22␣, and telokin, are activated by SRF-myocardin, SRF/MRTFA, SRF/GATA6/ CRP2, or SRF/Nkx complexes (6 -21). The immediate early growth factor responsive genes, such as c-fos and egr1 are regulated by SRF/ELK-1 (ets) complexes (22)(23)(24). The later early response genes, such as SRF itself and vinculin, that are actin/ Rho-dependent, are regulated by SRF-MRTFA complexes (16,25,26). Myocardin-related transcription factor A (MRTFA, or Mkl1, MAL, BSAC) is a unique co-activator of SRF in that it is involved in the regulation of multiple SRF-dependent gene families (reviewed in Ref. 27). MRTFA has been reported to induce SRF-dependent, smooth muscle-specific genes such as telokin, SM22␣, and SM ␣-actin and actin/Rho-dependent early response genes, but not proliferation related MAPK-dependent immediate early response genes (13,15,16,26). It still remains a mystery how MRTFA can discriminate between different SRF-dependent genes. One possible mechanism could involve gene-specific restriction of promoter access because of chromatin structure. In support of this model, it has been shown that there is very little SRF detectable at the CArG boxes of smooth muscle-specific genes in nonmuscle cells, whereas SRF binding can be readily detected at CArG boxes of early response genes such as c-fos (28). In addition, overexpression of myocardin in nonmuscle cells was found to lead to increased SRF binding to the promoters of smooth muscle-specific genes. In the current study we provide evidence supporting a role for ATPdependent chromatin remodeling in regulating SRF binding to CArG boxes within promoters of smooth muscle-specific genes.
Because chromatin is highly condensed, the regulation of chromatin accessibility to transcription factors and RNA polymerase is an essential step in gene activation (29 -31). Although * This work was supported by National Institutes of Health Grants HL-58571, DK-61130, and DK-65644. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Tables S1 and S2. 1 Supported by a fellowship from the Fortune Fry Foundation. 2 Supported by an AHA postdoctoral fellowship. 3  studies have shown that myocardin can recruit enzymes capable of modifying chromatin structure through covalent modification of histone tails (32,33), no studies have examined how chromatin structure affects promoter access by MRTFA. In addition, the role of ATP-dependent chromatin remodeling enzymes in the regulation of smooth muscle differentiation is unknown. The SWI/SNF complex is the best characterized, mammalian, ATP-dependent chromatin remodeling complex (30). It is comprised of 7-11 components, which assemble into distinct complexes containing either Brg1 (Brahma-related gene 1) or Brm (Brahma) ATPase subunits. SWI/SNF remodeling complexes have been shown to play an essential role in the differentiation of neurons, T cells, erythrocytes, hepatocytes, adipocytes, skeletal, and cardiac muscle cells (34 -44).
Although, the role of Brg1 or SWI/SNF in smooth muscle development is largely unknown, Brg1 has been shown to be up-regulated in vascular smooth muscle cells in primary atherosclerosis and in stent stenosis (45). A recent study has also demonstrated that Brg1 binding to CRP2 is critical for induction of smooth muscle-specific genes by CRP2 (18). During skeletal muscle differentiation it has been shown that the recruitment of Brg1 to MyoD, that is associated with DNAbound Pbx1, induces chromatin remodeling of the myogenin gene. This facilitates tight binding of MyoD to E boxes resulting in cofactor recruitment and transcription activation (46). By analogy, we propose that weak SRF binding to the CArG boxes of smooth muscle-specific genes may facilitate recruitment of MRTFA and that interaction of MRTFA with SWI/SNF may then remodel chromatin permitting tight binding of SRF. Results from our study demonstrate that Brg1 is required for the induction of smooth muscle-specific gene expression but not for early response gene expression by MRTFA. Endogenous Brg1, SRF, and MRTFA were found to form a complex in smooth muscle cells and tissue and Brg1 directly bound to MRTFA, but not SRF, in vitro. Chromatin immunoprecipitation assays revealed that SWI/SNF is required for MRTFA to increase SRF binding to the promoters of smooth muscle specific genes. Furthermore, expression of a dominant negative Brg1 in differentiated smooth muscle cells attenuated expression of smooth muscle-specific genes. Together these data indicate that SWI/SNF plays a critical role in regulating expression of SRF/MRTFA-dependent smooth muscle-specific genes but not SRF-dependent early response genes. SWI/SNF thus plays an important role in regulating the balance between the differentiation and proliferation roles of SRF.

EXPERIMENTAL PROCEDURES
Cell Culture and Adenoviral Transduction-An MRTFA cDNA image clone was purchased from Invitrogen (Clone ID: 682130) and moved to Adeno-X vector according to the manufacturer's protocol (BD Biosciences). Adenovirus encoding nuclear-localized YFP (yellow fluorescent protein) was used as negative control. B22 cells, which are NIH3T3 cells that express a tetracycline inducible dominant-negative Brg1 (DN-Brg1, K798R mutant) (47), were obtained from Dr. Anthony N. Imbalzano (University of Massachusetts Medical School, Worcester, MA). By withdrawing tetracycline from the growth media of these cells, the expression of DN-Brg1 can be induced.
B22 cells were maintained in high glucose Dulbecco's modified Eagle's medium (DMEM) (Mediatech) containing 2 g/ml tetracycline, 350 units/ml Hygromycin B, 75 g/ml G418, and 10% fetal bovine serum. B22 cells were seeded into a 6-well dish at a density of 2 ϫ 10 5 cells per well in the medium either with or without tetracycline and grown for 24 h prior to adenoviral transduction. Cells were incubated with adenovirus encoding MRTFA or YFP for 4 h at 37°C and then replaced with complete medium. 30 -48 h after transduction, cells were harvested for protein, RNA, or chromatin immunoprecipitation (ChIP) analysis. SW13 and HeLa cells were obtained from ATCC and grown in high glucose DMEM containing 5 units/ml penicillin, 50 mg/ml streptomycin, and 10% fetal bovine serum. 24 h before transduction, SW13 and HeLa cells were seeded at the density of 2.5 ϫ 10 5 cells per well in a 6-well dish. Cells were then transduced with adenovirus as described above for B22 cells. Primary mouse colon smooth muscle cells were prepared from colons dissected from 4-week-old mice. The epithelial layer was removed, and remaining smooth muscle layer was minced and digested with 1 ml of tissue digestion buffer per organ (0.4 units/ml Blendzyme 3 (Roche Applied Sciences) in DMEM) at 37°C for 1-2 h with shaking. The digested tissue was then passed through a cell sieve, and the cells collected by centrifugation. Pelleted cells were washed in DMEM containing 10% fetal calf serum and penicillin/streptomycin and plated into dishes. After 4 -5 days cells reached confluence, were trypsinized, and replated at 7 ϫ 10 4 per well in 12-well plates. 12 h after plating, cells were transduced by DN-Brg1 or YFP control adenovirus. 72 h after transduction, mRNA was harvested, and the levels of SRF-dependent genes were measured by quantitative real time RT-PCR.
Plasmids Used and Cell Transfection-Human Brg1 and Brm cDNA and DN-Brg1 in pBABE retroviral expression plasmids were obtained from AddGene (48). MRTFA was cloned into pcDNA myc/His (Invitrogen) for transfection and in vitro translation experiments. HA-SRF pShuttle was generated by cloning the human SRF cDNA into a modified pShuttle (Clontech) vector that includes an N-terminal HA epitope tag. 12 h prior to transfection, cells were seeded at the density of 2.5 ϫ 10 5 cells per well in 6-well plates. Plasmids were transfected into cells using Fugene 6 (Roche Applied Science): cells were washed once with phosphate-buffered saline (pH 7.4) then 2 ml of complete medium was added to each well together with 2 g of plasmid DNA and 4 l of Fugene in 100 l of DMEM.
RNA Analysis-RNA was extracted with TRIzol reagent (Invitrogen). 1.2 g of RNA was used as template for reverse transcription (RT) using Superscript first strand cDNA synthesis kit (Invitrogen). cDNA was dissolved in 20 l of H 2 O. The cDNA levels of specific genes were measured by quantitative real time PCR using SYBR green PCR master mix (Invitrogen) and a 7500 Real Time PCR system (Applied Biosystems) with gene-specific primers (Table S1 in supplemental data). 2 l of 1:10 diluted cDNA was used for each reaction in 25-l total volume. All PCR reactions were performed in duplicate.
Co-immunoprecipitation (Co-IP)-Co-IP assays were performed using a nuclear complex Co-IP kit, essentially as described by the manufacturer (Active Motif). 250 g of nuclear protein extracts were incubated with 3 g of anti-Brg1 antibody (Upstate), anti-SRF antibody (G20X), anti-MRTFA antibody (C-19 or ProteinTech), or appropriate IgG control in 500 l of low salt IP buffer (Active Motif) overnight at 4°C. 60 l of EZview protein A beads (Sigma) were added to the mixture and incubated for an additional hour with rocking. Beads were then washed six times with the low salt IP buffer. The immunoprecipitated proteins were dissolved in 35 l of 2ϫ SDS sample buffer and boiled for 5 min, prior to analysis by Western blotting as described as above.
In Vitro Transcription/Translation-Synthesis of proteins was carried out in a coupled transcription/ translation system (Promega, Madison, WI) in vitro, programmed with 1 g of pShuttle-Brg1 (FLAG tag), pcDNAmyc/his-MRTFA (Myc tag) and pShuttle-SRF (HA tag) plasmids. The TNT products were mixed together as combinations of: SRF with Brg1, MRTFA with Brg1, SRF with MRTFA, in 250 l of low salt IP buffer (Active Motif). Proteins were immunoprecipitated with 3 g of anti-SRF (G20X) or 15 l of anti-MRTFA antiserum overnight. A matching rabbit IgG or a MRTFA preimmune serum served as negative controls. The immunoprecipitated proteins were then incubated with EZ-view beads (Sigma) for 2 h. Following six washes with IP buffer, the beads were adenovirus. 30 -48 h following transduction, cells were lysed for mRNA analysis. mRNA was isolated from cells, and the levels of mRNA expression were measured by quantitative real time RT-PCR. Transcript levels were first normalized to acidic ribosomal phosphoprotein PO (RPLPO) internal-loading control and then normalized to their respective YFP control group. The ⌬⌬C t method was used to calculate the relative quantity values (RQ) of gene expression levels. C t is the threshold cycle where the amplification of template begins. RQ ϭ 2 Ϫ⌬⌬Ct and ⌬⌬C t ϭ (C t experimental Ϫ C t RPLPO )Ϫ (C t control Ϫ C t RPLPO ). Data presented are the mean Ϯ S.E. of 8 -9 samples obtained from three independent experiments. Asterisk indicates statistical significance as determined by a Student's t test (p Ͻ 0.05).
dissolved in SDS-sample buffer and subjected to Western blotting as described above.
Quantitative ChIP Assays-ChIP assays were performed according to the protocol of Upstate with minor modifications. Cells were fixed in 3.7% formaldehyde for 15 min at room temperature and harvested using cold phosphate-buffered saline with protease inhibitors. After collecting cells by centrifugation, cell pellets were lysed using 1% SDS lysis buffer (200 l/1 ϫ 10 6 cells). For each group, 1 ml of lysate was sonicated for 7 ϫ 30 s at setting 2.25 on a Sonic Dismembranator (Fisher Scientific). 200-l aliquots of chromatin were immunoprecipitated using 6 g of anti-SRF antibody (G20X), anti-H3Ac (Upstate), or rabbit IgG as negative control. The precipitated genomic DNA was purified and the presence of specific pro-moters was measured by real time quantitative PCR, using gene-specific primers (Table S2 in supplemental data).

DN-Brg1 Inhibits the Ability of MRTFA to Induce Expression of Smooth Muscle-specific Genes, but Not SRF-dependent Early
Response Genes-To determine the role of SWI/SNF-mediated chromatin remodeling on the induction of genes by SRF/ MRTFA, we utilized a previously characterized 3T3 cell line that inducibly expresses a dominant-negative Brg1 (B22 cells, Ref. 47). The dominant-negative K798R mutant Brg1 blocks the function of SWI/SNF complexes containing Brg1 or Brm catalytic subunits. B22 cells were transduced with MRTFA or YFP adenovirus. 30 h after transduction, cells were lysed and mRNA, and protein expression analyzed by quantitative realtime RT-PCR and Western blotting, respectively (Figs. 1 and 2). In agreement with previous reports (13-15, 26, 50) MRTFA induced the expression of several smooth muscle-specific genes in fibroblast cells (Fig. 1, compare solid bars to open bars in control cells). In contrast, MRTFA did not significantly induce expression of the early response genes, c-fos, egr1, or vinculin, although it did result in a 2-fold increase in expression of SRF mRNA (Fig. 1). As shown by Western blotting (Fig. 2) and more quantitatively by qRT-PCR (Fig. 1) dominant negative Brg-1 significantly abrogated the ability of MRTFA to increase expression of telokin, SM22␣, and calponin but not SM ␣-actin or any of the SRF-dependent early response genes examined. Conversely, DN-Brg1 augmented the ability of MRTFA to increase SRF mRNA expression and increased the basal expression of egr1 and c-fos mRNA (Fig. 1).
MRTFA Cannot Induce Smooth Muscle-specific Genes in SW13 Cells That Lack Brg1 and Brm-It is possible that DN-Brg1 inhibits the activity of more than just Brg1 and Brm containing SWI/SNF complexes, hence we also examined the role of Brg1 in a Brg1/Brm-null cell system. The ability of MRTFA to induce gene expression in human adrenal carcinoma SW13 cells that lack endogenous Brg1/ Brm was determined. Human cervical cancer HeLa cells that express Brg1 and Brm were used as control for these experiments. Results from Western blot analysis showed that overexpression of MRTFA in SW13 cells could not induce the expression of most smooth muscle-specific genes, including 130-kDa smMLCK, telokin, or SM22␣, and could only weakly induce SM ␣-actin expression (Fig. 3A). In contrast, MRTFA readily induced expression of all these genes in HeLa cells. MRTFA induced expression of SRF and vinculin in SW13 cells as well as in HeLa cells. Although egr1 was not induced by MRTFA in either of these two cell types, the basal expression of egr1 appeared higher in SW13 cells compared with HeLa  cells. Surprisingly, MRTFA was able to induce expression of c-fos in SW13 cells but not in HeLa cells (Fig. 3A), and this induction was attenuated by reintroduction of Brg1 into the SW13 cells (Fig. 3B). These results are consistent with the results obtained from the DN-Brg1 3T3 cell lines. To verify that the inability of MRTFA to induce smooth muscle-specific genes in SW13 cells is due to the lack of Brg1/Brm we reintroduced Brg1 or Brm1 back into SW13 cells. Western blot analysis demonstrated that restoration of Brg1 or Brm expression could rescue the ability of MRTFA to induce the expression of smooth muscle-specific proteins in SW13 cells (Fig. 3B).

DN-Brg1 Attenuates Smooth Muscle Gene Expression in Primary Smooth Muscle Cells-
The above results suggest that Brg1 or Brm is required for MRTFA to induce the expression of smooth muscle-specific genes in non-smooth muscle cells. To determine if Brg1 is required for smooth muscle cells to maintain the expression of smooth muscle-specific markers, DN-Brg1 was over expressed in primary colon smooth muscle cells. Results from quantitative real-time RT-PCR analysis demonstrated that dominant-negative Brg-1 significantly decreased the expression of smooth muscle-specific genes (telokin, SM MHC, calponin, SM22␣, and SM ␣-actin) and increased the expression of the SRF-dependent early response genes, such as egr1 and c-fos. The expression of SRF was not significantly changed (Fig. 4).
Brg1 Interacts with SRF and MRTFA in Intact Cells-We next determined if Brg1 could interact with MRTFA and/or SRF in intact cells. Co-IP assays were performed from COS cells transduced with HA-tagged MRTFA and SRF adenoviruses. MRTFA, SRF, and endogenous Brg1 were immunoprecipitated from the transduced cell extracts. Western blotting of immunoprecipitated proteins showed that MRTFA immunoprecipitates contained Brg1, MRTFA, and SRF (Fig. 5A). SRF immunoprecipitates contained MRTFA and SRF, and a very weak Brg1 signal. Similarly Brg1 immunoprecipitates contained Brg1 and MRTFA and a small amount of SRF. These data indicate that Brg1, MRTFA and SRF can form a complex in intact cells. To confirm this finding, we determined whether the endogenous proteins can also be detected in a complex in A10 smooth muscle cells (Fig. 5B) or in bladder tissue (Fig. 5C). Endogenous Brg1, SRF and MRTFA were immunoprecipitated from nuclear extracts of A10 cells or mouse bladder and Western blot analysis revealed that Brg1 can be co-immunoprecipitated with SRF and MRTFA. These data indicate that the 3 proteins can form a complex in A10 smooth muscle cells and bladder tissue in vivo (Fig. 5B,C).
Brg1 Interacts with MRTFA but Not SRF in Vitro-To determine whether Brg1 directly binds to MRTFA and/or SRF, in vitro binding assays were performed. Brg1, MRTFA, and SRF were synthesized in vitro using a coupled in vitro transcription and translation system. pShuttle-Brg1 (FLAG tag), pcDNA myc/his-MRTFA (Myc tag), and pShuttle-SRF (HA tag) were utilized as templates for the transcription/translation reactions. In vitro synthesized proteins were incubated together and immunoprecipitation assays were used to identify the interacting proteins (Fig. 6). Results from this analysis indicate that Brg1 could be coprecipitated with MRTFA but not with SRF, demonstrating that Brg1 can directly bind to MRTFA.

DN-Brg1 Inhibits the Ability of MRTFA to Increase SRF Binding to the Promoters of Smooth Muscle-specific Genes-To
explore why Brg1 is required for MRTFA induced smooth muscle-specific gene expression, but not for expression of the early response genes, ChIP assays were performed. B22 cells were grown in the presence or absence of tetracycline and transduced with adenovirus encoding MRTFA or YFP. ChIP assays were performed to examine SRF binding to SRF-dependent promoters as described under "Experimental Procedures." Consistent with previous studies using myocardin (28), we found that MRTFA significantly increased SRF binding to the   Brg1 forms a complex with SRF and MRTFA in vivo. A, adenovirus encoding HA-tagged SRF and HA-tagged MRTFA were transduced into COS cells, 48 h later, nuclear proteins were harvested using a hypotonic lysis/nuclease digestion protocol (Active Motif). Proteins were immunoprecipitated using MRTFA (MA), SRF (S), and Brg1 (B) antibodies and goat IgG (Ig(G)) or rabbit IgG (Ig(R)) as a negative control. (MRTFA is a goat antibody; SRF and Brg1 are rabbit antibodies.) Immunoprecipitated proteins were then identified by Western blotting, using the antibodies indicated at the right of the blot. Inp, 10% of input. B, co-immunoprecipitations were performed from nuclear extracts prepared from A10 cells. C, mouse bladder tissues were diced into small pieces and homogenized in a prechilled homogenizer prior to harvesting nuclear protein as described above for cells. Precipitated proteins were analyzed by Western blotting using antibodies directed against endogenous, Brg1, MRTFA, and SRF described under "Experimental Procedures." promoters of smooth muscle-specific genes, telokin, SM22␣, and SM ␣-actin, (Fig. 7A). DN-Brg1 dramatically attenuated the ability of MRTFA to increase SRF binding to telokin and SM22␣ promoters, but not SM ␣-actin promoter (Fig. 7B). In contrast, neither MRTFA nor DN-Brg1 affected the SRF binding to the promoters of the early response genes, SRF and c-fos (Fig. 7, A and B). In addition to increasing SRF binding to the telokin promoter, MRTFA also increased levels of acetylated histone H3 (Fig. 7A, right panel). This increase in acetylated histone H3 on the telokin promoter was blocked by expression of DN-Brg1. In contrast, the small increase in acetylated H3 on the SM ␣-actin promoter was not blocked by DN-Brg1 (Fig. 7B,  right panel). Under control conditions (no DN-Brg1 expression), in YFP-transduced B22 cells, there appeared to be more SRF bound to the SM ␣-actin, SRF, and c-fos promoters as compared with the telokin and SM22␣ promoters (Fig. 7C, left  panel). Similarly there was more acetylated histone H3 associated with the SM ␣-actin promoter as compared with the telokin promoter under control conditions (Fig. 7C, right  panel).

DISCUSSION
In this study, we found that Brg1 or Brm are required for the MRTFA-mediated induction of smooth muscle-specific genes, but not early response genes. DN-Brg1 attenuated the induction of smooth muscle-specific genes and reintroducing Brg1/ Brm into SW13 cells restored their responsiveness to MRTFA. Brg1 appears to be required for MRTFA to increase the binding of SRF to the promoters of smooth muscle-specific genes, within intact chromatin. Brg1, MRTFA, and SRF were found to form a complex within intact smooth muscle cells and Brg1 directly bound MRTFA but not SRF in vitro. Because previous studies (16,25,50), and data presented in Fig. 6 have shown that MRTFA directly binds SRF in vitro, MRTFA can thus act as a bridge to connect Brg1 and SRF. proteins were synthesized by using in vitro transcription and translation, as described under "Experimental Procedures." The proteins were then incubated together as shown at the top of the blot, and proteins were immunoprecipitated using the antibodies indicted below the blot (IP) (preimm, preimmune serum from rabbit used to make MRTFA antibody). The presence of individual proteins in each immunoprecipitate was determined by Western blotting using antibodies to each of their respective epitope tags (indicated at the right of the blot). 10% of the input of each IP was loaded at the left of the blot. FIGURE 7. DN-Brg1 attenuates the ability of MRTFA to increase SRF binding to the promoters of smooth muscle-specific genes. B22 cells grown in the presence (control) or absence (ϩDN-Brg1) of tetracycline were transduced with MRTFA or YFP control adenovirus. After 30 h, cells were fixed and harvested for ChIP assays. Chromatin was precipitated using an antibody against SRF (left panels), acetylated histone H3 (right panels), or using IgG negative control. The precipitated genomic DNA was purified, and the presence of the promoters of SRF-dependent genes measured by real time quantitative PCR, using gene-specific primers. A, the increase in SRF or H3Ac binding in samples transduced with MRTFA is indicated relative to those transduced with YFP. These data were calculated and normalized to input levels as follows: Relative SRF/H3Ac binding, RQ ϭ 2 Ϫ⌬⌬Ct , with ⌬⌬C t ϭ (C t MRTFA Ϫ C t input ) Ϫ (C t YFP Ϫ C t input ). B, relative inhibition of MRTFA-induced SRF or H3Ac binding by DN-Brg1 is shown. This was calculated as follows: Relative SRF/H3Ac binding, RQ ϭ 2 Ϫ⌬⌬Ct , with ⌬⌬C t ϭ (C t DN-Brg1 Ϫ C t input ) Ϫ (C t control Ϫ C t input ). SRF data shown in panels A and B are the mean Ϯ S.E. of 7 samples obtained from three independent experiments. H3Ac data shown in panels A and B are the mean Ϯ S.E. of 4 samples obtained from two independent experiments. A one sample t test was performed, and the asterisks indicate the results that are statistically different from 1 (p Ͻ 0.05). C, the relative binding of SRF or H3Ac in control samples (no DN-Brg1, YFP-transduced) is shown for each of the promoters indicated. Results were calculated as follows: Relative SRF/H3Ac binding, RQ ϭ 2 Ϫ⌬ct , with ⌬Ct ϭ C tSRF/H 3 AC IP group Ϫ C t IgG IP group . Results presented are the mean Ϯ S.E. of 4 -5 samples.
Results shown in Figs. 1 and 2 demonstrate that DN-Brg-1 blocked the MRTFA-mediated induction of telokin, SM22␣, and calponin but did not significantly block the induction of SM ␣-actin. However, from results shown in Fig. 2 it is apparent that SM ␣-actin, but not other smooth muscle marker genes, was readily detectable in B22 3T3 cells prior to overexpression of MRTFA (YFP control group). This indicates that the SM ␣-actin locus is transcriptionally active in control cells and thus likely to be in an open chromatin con-formation in these cells. In support of this hypothesis, there was significantly more SRF and acetylated histone H3 associated with the SM ␣-actin promoter as compared with the telokin promoter in control cells (Fig. 7C). Thus it might be predicted that Brg1 containing SWI/SNF complexes may not be required to further open the chromatin structure of the SM ␣-actin promoter in order for MRTFA to increase SRF binding to the promoter. This suggests a model in which Brg1 is required for MRTFA to increase binding of SRF to the Upper panels, telokin promoter is used as an example to describe the activation of smooth muscle-specific promoters by MRTFA. In unstimulated non-muscle cells the promoter is in a condensed inactive chromatin conformation that only permits weak or transient SRF binding (left panel). Upon increased nuclear MRTFA, MRTFA binds to Brg1 and recruits the MRTFA/Brg1/HAT (p300?) complex to the weakly bound SRF. The binding of Brg1 to acetylated histone tails may also play a role in this recruitment process. Once the MRTFA/Brg1/HAT complex is recruited to the promoter, Brg1 utilizes the energy derived from hydrolysis of ATP to remodel the chromatin structure, permitting tight binding of SRF (middle panel). Tightly bound SRF/MRTFA can then recruit additional co-activators such as p300, resulting in additional chromatin remodeling, facilitating binding of the RNA polymerase complex, and subsequent activation of transcription (right panel). Lower panels, on the promoters of MRTFA-regulated early response genes, such as the SRF gene itself, SRF is tightly bound to the promoter under resting conditions (left panel). MRTFA, therefore, does not require Brg1-mediated chromatin remodeling to be able to bind to the promoter, recruit co-activators, and thereby activate transcription (right panel). promoters of transcriptionally silent genes but not to genes that are already transcriptionally active within native chromatin (Fig. 8).
As the MRTFA-regulated early response genes are also expressed in control cells, our model would predict that activation of these genes would also be independent of Brg1. This is consistent with data presented in Fig. 1 showing that MRTFAmediated induction of SRF was not inhibited by DN-Brg1. In addition, under control conditions there was markedly more SRF bound to the SRF promoter as compared with the telokin or SM22␣ promoters (Fig. 7C). In our experiments only in SW13 cells did we detect any activation of vinculin by MRTFA (Fig. 3A), although this activation was not altered by expression of Brg1 or Brm (Fig. 3B), suggesting that it is independent of these proteins. In agreement with previous studies (26) we also observed no affects of MRTFA on expression of MAPK-dependent SRF-target genes such as c-fos or egr1 (Fig. 1). However, the basal expression of c-fos and egr1 is higher in cells expressing DN-Brg1 (Fig. 1) and in cells lacking Brg1/Brm (SW13 cells compared with HeLa cells in Fig. 3A). This observation is consistent with a previous study that showed that Brg1 represses c-fos expression (51). It is not known why Brg1 inhibits c-fos expression, though it might be predicted that it either results in increased binding of a repressor or it may alter the positioning of the nucleosome, which has been shown to be located between key regulatory elements in the c-fos promoter (52). As Brg1 is specifically required for MRTFA to induce expression of smooth muscle-specific genes but not early response genes, then regulation of MRTFA/Brg1 interactions could provide a mechanism for a cell to switch between activation of these two groups of genes.
Previous studies showed that Brg1 knock-out mice die during the periimplantation stage (53), however, Brm knock-out mice develop normally, except that adult mice have higher body weight (54). These data suggest that Brg1 has non-redundant functions that cannot be replaced by Brm. However, in our experiments we found that either Brg1 or Brm were required for MRTFA to induce expression of smooth muscle-specific genes. This is in contrast to a recent study that demonstrated that Brg1 but not Brm is required for CRP-mediated induction of smooth muscle-specific genes in cardiac myocytes (18). Future analysis of smooth muscle-specific knockouts of Brg1 will be required to determine the specific role of Brg1 and Brm in smooth muscle cells in vivo.
Although Brg1 is ubiquitously expressed, Brg1 has been shown to selectively rather than broadly regulate gene expression (30). This can be explained, at least in part, by the recruitment of Brg1 to individual promoters through its interaction with specific transcription factors (18,38,46,(55)(56)(57)(58)(59)(60). Our results further support this idea as we found that Brg1, SRF, and MRTFA form a complex in smooth muscle tissue in vivo and Brg1 directly interacts with MRTFA but not SRF (Fig. 6). This would allow Brg1 to be recruited to MRTFA-dependent SRF target genes but not to Elk-dependent SRF target genes. A recent study has also shown that CRP2 directly binds to Brg1 (18), and CRP2 has previously been shown to form a complex with SRF in some smooth muscle cell types (17). Thus, CRP2 may act similar to MRTFA to facilitate recruitment of Brg1 complexes to mediate stable SRF binding to the promoters of smooth muscle-specific genes.
It is likely that the ATP-dependent chromatin remodeling catalyzed by Brg1/Brm acts together with covalent histone modifications to permit stable SRF/MRTFA binding and transcription activation. Although SWI/SNF itself does not covalently modify histones its remodeling of chromatin through nucleosome reorganization can facilitate the recruitment of other proteins that covalently modify histones. Previous studies have shown that overexpression of myocardin, an MRTFA homologue, can induce histone acetylation through myocardin interaction with the HAT, p300 (33). Histone acetylation has also been shown to promote the binding of SWI/SNF complexes to chromatin, as the acetylated histone tail can bind to the bromodomain of Brg1 (61). Thus MRTFA/p300 induced histone acetylation, may further stabilize Brg1 binding to the promoters of smooth muscle-specific genes (Fig. 8). In a reciprocal fashion Brg1-induced changes in chromatin structure will facilitate increased binding of SRF/MRTFA complexes to the promoter and thus result in increased recruitment of p300. Together these interactions can then lead to transcription activation.
In summary, we propose a model in which the recruitment of MRTFA/Brg1 complexes to weakly bound SRF at the promoters of smooth muscle-specific genes, leads to chromatin remodeling, which facilitates tight binding of SRF complexes and subsequent transcription activation (Fig. 8, upper panels). In contrast, in cells grown under high serum conditions, SRF is tightly bound to the promoters of SRF-dependent early response genes. Brg1 is thus not required to permit MRTFA or MAPK-Elk signals to activate these genes (Fig. 8, lower panels).