Novel Mechanism of Negative Regulation of 1,25-Dihydroxyvitamin D3-induced 25-Hydroxyvitamin D3 24-Hydroxylase (Cyp24a1) Transcription

Background: CYP24A1 is the principal enzyme involved in the catabolism of 1,25(OH)2D3. Results: The SWI/SNF complex and PRMT5 converge at the transcriptional level to control 1,25(OH)2D3-induced Cyp24a1 gene expression. Conclusion: PRMT5-mediated repression represents a novel mechanism of negative regulation of Cyp24a1. Significance: Our study reveals key factors involved in the regulation of 1,25(OH)2D3 catabolism and therefore in the control of calcium homeostasis. The SWI/SNF chromatin remodeling complex facilitates gene transcription by remodeling chromatin using the energy of ATP hydrolysis. Recent studies have indicated an interplay between the SWI/SNF complex and protein-arginine methyltransferases (PRMTs). Little is known, however, about the role of SWI/SNF and PRMTs in vitamin D receptor (VDR)-mediated transcription. Using SWI/SNF-defective cells, we demonstrated that Brahma-related gene 1 (BRG1), an ATPase that is a component of the SWI/SNF complex, plays a fundamental role in induction by 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) of the transcription of Cyp24a1 encoding the enzyme 25-hydroxyvitamin D3 24-hydroxylase involved in the catabolism of 1,25(OH)2D3. BRG1 was found to associate with CCAAT-enhancer-binding protein (C/EBP) β and cooperate with VDR and C/EBPβ in regulating Cyp24a1 transcription. PRMT5, a type II PRMT that interacts with BRG1, repressed Cyp24a1 transcription and mRNA expression. Our findings indicate the requirement of the C/EBP site for the inhibitory effect of PRMT5 via its methylation of H3R8 and H4R3. These findings indicate that the SWI/SNF complex and PRMT5 may be key factors involved in regulation of 1,25(OH)2D3 catabolism and therefore in the maintenance of calcium homeostasis by vitamin D. These studies also define epigenetic events linked to a novel mechanism of negative regulation of VDR-mediated transcription.

Vitamin D is essential for mineral homeostasis and maintenance of bone mass (1,2). 1,25-Dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ), 2 the hormonally active form of vitamin D, is produced by two sequential hydroxylations of vitamin D at C-25 in the liver by 25-hydroxylase and at C-1 by the enzyme 1␣-hydroxylase (3). 1,25(OH) 2 D 3 mediates its effects by binding to the vitamin D receptor (VDR), a member of the nuclear receptor family of transcription factors, which heterodimerizes with the retinoid X receptor and interacts with specific DNA sequences (vitamin D response elements) in target genes, resulting in activation or repression of transcription (1,2,4). The actions of 1,25(OH) 2 D 3 and VDR are mediated by the recruitment of coregulatory complexes. These coregulatory complexes include the coactivator complex DRIP (VDR-interacting protein complex also known as Mediator complex) that functions at least in part through recruitment of RNA polymerase II and the p160 coactivators (steroid receptor coactivators 1, 2, and 3), which have histone acetyltransferase activity (1,5).
For transcription to occur, chromatin structure is altered not only by covalent modification of histones but also by ATPdependent chromatin-remodeling enzymes. One important ATP-dependent chromatin remodeling factor is the SWI/SNF multisubunit complex (6,7). The SWI/SNF complex has been implicated in the regulation of differentiation as well as in the regulation of target genes stimulated by estrogen, glucocorticoid, and retinoic acid receptors (8 -12). Each SWI/SNF complex contains one of two homologous ATPases, Brahma (Brm) or Brahma-related gene 1 (BRG1). BRG1-null mice are embryonic lethal, whereas Brm-null mice are viable (exhibiting a mild phenotype of increased body weight) (13,14). C/EBP␤ has been reported to recruit the SWI/SNF complex to specific gene promoters to promote tissue-specific transcription (15)(16)(17). In addition, it has been shown that ATP-dependent chromatin remodeling complexes can act together with histone-modifying enzymes to modulate transcription (18 -20). Among the histone-modifying enzymes are the protein-arginine methyltransferases (PRMTs) that have been implicated in transcriptional activation or repression (21)(22)(23)(24). Both type I and type II PRMTs catalyze the formation of monomethylarginine. Transfer of an additional methyl group results in the formation of asymmetrically dimethylated arginines (catalyzed by type I enzymes) or symmetrically dimethylated arginines (catalyzed by type II enzymes). PRMT4 (CARM1), a type I PRMT, methylates arginines 2, 17, and 26 of histone 3 and is associated with transcriptional activation, whereas PRMT5, a type II PRMT, has been linked to gene silencing through repressive histone marks including symmetrical dimethylation of H3R8 and H4R3 (21)(22)(23)(24). Little is known about the role of SWI/SNF and PRMTs in VDR-mediated transcription.
One of the most pronounced effects of 1,25(OH) 2 D 3 is increased synthesis of 25-hydroxyvitamin D 3 24-hydroxylase (CYP24A1), the enzyme that accelerates the catabolism of 1,25(OH) 2 D 3 (25). Thus, by inducing CYP24A1, 1,25(OH) 2 D 3 regulates its own synthesis, protecting against hypercalcemia. We reported previously that C/EBP␤ is induced by 1,25(OH) 2 D 3 in kidney and osteoblastic cells and is a potent enhancer of VDRmediated Cyp24a1 transcription (26). Here, we demonstrate that the SWI/SNF complex contributes to transcriptional activation by VDR. We found that BRG1 associates with C/EBP␤, and together they cooperate with VDR in the regulation of Cyp24a1. PRMT5, which interacts with BRG1, was found to be a negative regulator of 1,25(OH) 2 D 3 -induced Cyp24a1 transcription and mRNA expression. Our findings provide new insight into key factors and epigenetic events that regulate Cyp24a1 expression and thus affect the regulation of 1,25(OH) 2 D 3 metabolism and the maintenance of calcium homeostasis.
Plasmids, Transfections, and Assays of Luciferase and Chloramphenicol Acetyltransferase (CAT) Activity-Luciferase reporter constructs of the rat Cyp24a1 promoter (Ϫ1367/ϩ74 containing vitamin D response elements at positions Ϫ258/Ϫ244 and Ϫ151/Ϫ137, the intronic vitamin D response elements at ϩ35 and ϩ39, a C/EBP site at Ϫ395/Ϫ388, and deletion construct Ϫ298/ϩ74) as well as a CAT reporter construct Ϫ671/ϩ74 with and without the C/EBP site mutated (26) were used. pAV-hVDR was obtained from Dr. J. W. Pike (University of Wisconsin, Madison, WI), pBJ5-BRG1 and pBJ5-BRG1-DN (K785R) expression vectors were obtained from James DiRenzo and Myles Brown (Dana Farber Cancer Institute, Harvard Medical School, Boston, MA), and pMex-C/EBP␤ was a gift from Dr. Simon Williams (Texas Tech University, Lubbock, TX). pBabe/ Fl-PRMT5 and catalytically inactive PRMT5 (pBabe/Fl-PRMT5 (G367A/R368A) were described previously (28). Cells were seeded in a 24-well culture dish 24 h prior to transfection at 70% confluence. Cells in each well were transfected using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. To determine the involvement of the SWI/SNF complex, SWI/SNF-negative cells lines (SW-13 and C33A) were transfected with a component of the SWI/SNF complex, either BRG1 or Brm. COS-7 cells, which do not contain steroid receptors (including VDR), were used in transfection studies to assess the impact of VDR (using transfected VDR) on gene activation. Empty vectors were used to keep the total DNA concentration the same. 1,25(OH) 2 D 3 (10 Ϫ8 M) was added to cells in media supplemented with 2% charcoal-dextran-treated FBS 16 h post-transfection for another 24 h. Cells were harvested using 1ϫ passive lysis buffer (Dual Luciferase reporter assay kit from Promega (Madison, WI)), and the assay for luciferase activity was performed according to the manufacturer's protocol. The efficiency of transfection was assessed by green fluorescent protein and PRMT5 co-transfection followed by visualization using a fluorescence microscope. The efficiency of transfection of MC3T3 and Caco-2 cells was estimated at 60 -70% and was comparable with transfection efficiency in UMR cells (65-75%). For studies using promoter constructs linked to the CAT reporter gene, CAT assays were performed by standard protocols on cell extracts normalized to total protein. CAT activity was quantified by scanning TLC plates using the Packard Constant Imager System (Packard Instrument Co.).
For Western blotting, total cellular protein was prepared using radioimmune precipitation assay buffer, and protein concentration was measured by the Bradford method. 100 g of protein was separated by SDS-PAGE and transferred to a PVDF membrane using a semidry transfer apparatus. The membrane was incubated overnight at 4°C with primary antibody (1:500) diluted in phosphate-buffered saline (PBS) containing 5% nonfat milk. The membrane was washed with PBS and incubated for 2 h with the corresponding secondary antibody conjugated with horseradish peroxidase. The enhanced chemiluminescence Western blotting system was used to detect the antigenantibody complex. The same blot was stripped and reprobed with ␤-actin polyclonal antibody to normalize for sample variation.
Co-immunoprecipitation Assay-Nuclear extracts were prepared, and protein concentration was determined by the Bradford method (29). 500 g-1 mg of each preparation was used for immunoprecipitation with the addition of 4 g of C/EBP␤, BRG1, or PRMT5 antiserum for 24 h at 4°C. 30 l of protein A-Sepharose 4 Fast Flow beads (Amersham Biosciences) was added to each sample, and after further incubation by rotating at 4°C for 24 h, the immunoprecipitated complex was collected by centrifugation at 3,000 rpm for 5 min and washed with ice-cold PBS containing protease inhibitor. The complex was separated by 7.5 or 15% SDS-PAGE and analyzed by Western blot using BRG1 antibody, C/EBP␤ antibody, or PRMT5 antibody.
Chromatin Immunoprecipitation (ChIP) Assay-Cells were treated with either vehicle or 10 Ϫ8 M 1,25(OH) 2 D 3 for 4 h followed by cross-linking with 1% formaldehyde for 15 min. Chromatin immunoprecipitation was performed as described previously (30). DNA fragments were purified using QIAquick PCR purification kits (Qiagen, Valencia, CA) and subjected to PCR using primers designed to amplify fragments of the Cyp24a1 promoter C/EBP motif at Ϫ395/Ϫ388 (forward, 5Ј-GAAATT-CTGCAAACCGCATT-3Ј; reverse, 5Ј-CCAGACTTCCCTT-GGATGAA-3Ј). PCR using the primers to amplify the upstream region of the Cyp24a1 promoter (Ϫ837/Ϫ567) was used as a negative control. PCR analysis was carried out in the linear range of DNA amplification. PCR products were resolved on a 2% agarose gel and visualized using ethidium bromide staining. DNA acquired prior to precipitation was collected and used as the input. 10% of input was used for PCR evaluation.
In re-chromatin immunoprecipitation (re-ChIP) experiments, complexes were eluted by incubation for 30 min at 37°C in 60 l of elution buffer containing 10 mM dithiothreitol. The eluted samples were diluted 50 times with ChIP dilution buffer and subjected again to the ChIP procedure with specific antibodies.
Statistical Analysis-Results are expressed as the mean Ϯ S.E., and significance was determined by analysis with Student's t test for two group comparison or analysis of variance for multiple group comparison.

BRG1-containing SWI/SNF Complex, VDR, and C/EBP␤ Cooperate in the Regulation of Cyp24a1 Transcription-Previ-
ous studies have shown that VDR-induced transcription is mediated by 1,25(OH) 2 D 3 -dependent recruitment of coregulatory complexes, which include p160 coactivators with histone acetyltransferase activity (1)(2)(3)(4). For chromatin remodeling to occur, the effects of acetylation are complemented by structural modifications that are ATP-dependent (19). An important ATP-dependent chromatin remodeling factor is the SWI/SNF complex (6). Induction of Cyp24a1 transcription is one of the most pronounced effects of 1,25(OH) 2 D 3 . To determine whether SWI/SNF activity contributes to VDR-mediated Cyp24a1 transcription, we examined the BRG1-and Brm-deficient cell line SW-13. SW-13 cells were co-transfected with VDR and the rat Cyp24a1 promoter (Ϫ1367/ϩ74). In the absence of BRG1 and Brm but under conditions of VDR overexpression, minimal activation of Cyp24a1 transcription by 1,25(OH) 2 D 3 was observed (Fig. 1A). However, VDR-mediated activation of Cyp24a1 transcription in SW-13 cells was preferentially enhanced by BRG1 (Fig. 1A) compared with Brm (no significant induction compared with 1,25(OH) 2 D 3 treatment of SW-13 cells at 0.1 and 0.3 g of Brm (data not shown)). A similar low level of induction of Cyp24a1 transcription by 1,25(OH) 2 D 3 and preferential enhancement by BRG1 were observed using the SWI/SNF-defective cell line C33A (Fig. 1A,  inset). BRG1-DN (K798R), which has a mutation in the ATPase domain and is incapable of remodeling chromatin, was found to inhibit enhancement of Cyp24a1 transcription by BRG1 in SW-13 cells (Fig. 1A). BRG1-DN also inhibited 1,25(OH) 2 D 3 induction of Cyp24a1 transcription in SWI/SNF-competent COS-7 cells in a dose-dependent manner, resulting in a maximum inhibition of 2.8 Ϯ 0.3-fold at a concentration of 0.3 g of BRG1-DN (p Ͻ 0.05 compared with 1,25(OH) 2 D 3 induction in the absence of BRG1-DN) (Fig. 1B). There was no effect of BRG1-DN on basal levels of Cyp24a1 transcription (Fig. 1B). In additional studies in COS-7 cells, suboptimal 1,25(OH) 2 D 3 Cyp24a1 transcription was significantly enhanced by BRG1 (2.3 Ϯ 0.3-fold at a concentration of 0.1 g of BRG1; p Ͻ 0.05 compared with 1,25(OH) 2 D 3 induction in the absence of BRG1). Together these results indicate that BRG1 contributes to VDR-mediated transcriptional activation of Cyp24a1.
Previous studies from our laboratory showed that C/EBP␤ is induced by 1,25(OH) 2 D 3 in kidney and osteoblastic cells and is a potent enhancer of VDR-mediated Cyp24a1 transcription (26). C/EBP␤ has also been reported to recruit the SWI/SNF complex to regulate cell type-specific genes (15)(16)(17). Hence, we next investigated the role of BRG1 in the C/EBP␤ enhancement of 1,25(OH) 2 D 3 -induced Cyp24a1 transcription. C/EBP␤ was cotransfected in COS-7 cells with VDR and the rat Cyp24a1 promoter construct (Ϫ1367/ϩ74) in the presence or absence of BRG1-DN. Cells were treated with vehicle or 1,25(OH) 2 D 3 . BRG1-DN inhibited the stimulatory effect of C/EBP␤ on VDRmediated Cyp24a1 transcription in a dose-dependent manner ( Fig. 2A). A similar inhibition by BRG1-DN was observed using MC3T3 osteoblastic cells, LLC-PK1 kidney cells and UMR osteoblastic cells (which contain endogenous VDR; Refs. 31-33) transfected with C/EBP␤ and the Cyp24a1 promoter and treated with 1,25(OH) 2 D 3 (data not shown). RT-PCR analysis also showed inhibition of the stimulatory effect of C/EBP␤ on 1,25(OH) 2 D 3 -induced Cyp24a1 mRNA levels by BRG1-DN (Fig. 2B). In addition, Western blot analysis showed that in the presence of BRG1-DN 1,25(OH) 2 D 3 induction and the stimu-latory effect of C/EBP␤ on CYP24A1 protein expression are repressed (Fig. 2C). When a rat Cyp24a1 promoter construct with the C/EBP site deleted (Ϫ298/ϩ74) was used, 1,25(OH) 2 D 3 -mediated induction of Cyp24a1 transcription was neither enhanced by C/EBP␤ nor inhibited by BRG1-DN (Fig. 2D). Similar results were observed using a mutated construct of the Cyp24a1 promoter in which the Ϫ395/Ϫ388 C/EBP binding site is mutated (data not shown). These findings indicate that both C/EBP␤ and BRG1 contribute to 1,25(OH) 2 D 3mediated regulation of CYP24A1 expression.
To further understand the mechanisms involved in the cooperation between C/EBP␤ and BRG1 in 1,25(OH) 2 D 3 induction of CYP24A1 in vivo, we first assessed via co-immunoprecipitation whether BRG1 and C/EBP␤ are components of the same nuclear complex. Nuclear extracts prepared from UMR cells, MC3T3 cells, mouse primary osteoblastic cells, and PCT cells were immunoprecipitated using C/EBP␤ antibody followed by immunoblotting for BRG1 (Fig. 3A, left panel) or immunoprecipitated using BRG1 antibody followed by immunoblotting for C/EBP␤ (Fig. 3A, right panel). We found that C/EBP␤ and BRG1 interact and are components of the same nuclear complex. Next, we examined the recruitment of BRG1 and C/EBP␤ to the Cyp24a1 promoter in response to 1,25(OH) 2 D 3 using a ChIP assay. ChIP with C/EBP␤ antibody and re-ChIP with BRG1 antibody showed that in vivo C/EBP␤ and BRG1 bind simultaneously at the C/EBP site of the Cyp24a1 promoter in response to treatment with 1,25(OH) 2 D 3 in UMR cells, MC3T3 cells, mouse primary osteoblastic cells, and PCT cells (Fig. 3B).

PRMT5 Inhibits 1,25(OH) 2 D 3 -induced Cyp24a1
Transcription-Previous studies have shown an interplay between the SWI/ SNF chromatin-remodeling enzymes and histone-modifying enzymes including PRMTs (18 -20, 27, 28). Because PRMT5, which has been reported to be involved in transcriptional repression, has been shown to be associated with the BRG1based SWI/SNF complex (28), we examined a possible role of PRMT5 in the regulation of Cyp24a1. We found that 1,25(OH) 2 D 3 -induced Cyp24a1 transcription is down-regulated in the presence of PRMT5 (Fig. 4A). A catalytically inactive form of PRMT5 with two point mutations (G367A/R368A) revealed no repressive influence on 1,25(OH) 2 D 3 -induced Cyp24a1 luciferase expression (Fig. 4A). 1,25(OH) 2 D 3 -induced FIGURE 3. C/EBP␤ and BRG1 are components of the same nuclear complex and are recruited to the C/EBP site of the Cyp24a1 promoter. A, nuclear extracts were prepared from UMR cells, MC3T3 cells, mouse primary osteoblastic cells, and PCT cells and used for immunoprecipitation (IP) with C/EBP␤ antibody, BRG1 antibody, or control rabbit IgG antibody. Similar results were observed in three independent experiments. B, ChIP analysis of C/EBP␤ and re-ChIP analysis of BRG1 binding to the C/EBP site of the Cyp24a1 promoter. UMR cells, MC3T3 cells, mouse primary osteoblastic cells, and PCT cells were treated with vehicle or 1,25(OH) 2 D 3 for 4 h and cross-linked by 1% formaldehyde for 15 min. Cross-linked cell lysates were subjected to immunoprecipitation first with C/EBP␤ antibody (␣-C/EBP␤) and then with BRG1 antibody (␣-BRG1). DNA precipitates were isolated and then subjected to PCR using specific primers designed according to the C/EBP site on the Cyp24a1 promoter. Input DNA (10%) was collected before immunoprecipitation. Con, control.
It has been found previously that PRMT5 interacts with BRG1, and together they are part of a complex involved in transcriptional repression (28). Thus we examined the effect of PRMT5 on BRG1 enhancement of Cyp24a1 transcription. We found that PRMT5 inhibits the BRG1-mediated enhancement of 1,25(OH) 2 D 3 -induced Cyp24a1 transcription in a dose-dependent manner (Fig. 5A). A similar inhibition of C/EBP␤ enhancement of VDR-mediated Cyp24a1 transcription by PRMT5 was also observed (data not shown). Furthermore, in SWI/SNF-defective SW-13 cells, PRMT5 (even at high concentrations; 0.1 and 0.2 g) did not have a repressive effect on the low level of 1,25(OH) 2 D 3 -mediated activation of Cyp24a1 transcription observed in these cells (data not shown). Similar to what has been reported previously using HeLa cell nuclear extracts (27), co-immunoprecipitation experiments indicated that PRMT5 and BRG1 are components of the same nuclear complex in osteoblastic and renal cells ( Fig. 5B; BRG1-deficient SW-13 cells were used as a negative control). When a rat Cyp24a1 promoter construct with the C/EBP site mutated (Ϫ671/ϩ74) (Fig. 5C) or a deletion construct of the Cyp24a1 promoter (Ϫ298/ϩ74), which does not include the Ϫ395/Ϫ388 C/EBP binding site (not shown), was used, neither enhanced transcription by C/EBP␤ nor suppression of 1,25(OH) 2 D 3 -induced Cyp24a1 transcription by PRMT5 was observed. When COS-7 cells were transfected with a thymidine kinase-luciferase construct containing six consecutive vitamin D response elements, PRMT5 suppression of 1,25(OH) 2 D 3 -induced transcription was also not observed (not shown). These findings indicate a role for BRG1 and the requirement of the C/EBP element in the Cyp24a1 promoter for the inhibitory effect of PRMT5.
To determine the specificity of inhibition by PRMT5 for Cyp24a1, we examined other targets of 1,25(OH) 2 D 3 for inhibition by PRMT5. TRPV6 is an epithelial calcium channel involved in active intestinal calcium transport (34). TRPV6 mRNA was up-regulated when Caco-2 cells were treated with 1,25(OH) 2 D 3 . A suppressive effect of PRMT5 was not observed (Fig. 6A). OPN, a phosphoprotein that modulates both bone mineralization and resorption, is induced in the presence of 1,25(OH) 2 D 3 in MC3T3 cells (31). 1,25(OH) 2 D 3 induction of Opn mRNA was not repressed in cells transfected with PRMT5 ( Fig. 6B). Similar results were observed using UMR cells (Fig.  4B).
H3R8 and H4R3 have been reported previously to be histone substrates for PRMT5 (27). Our findings as well as those of others have shown that PRMT5 binds to BRG1 (28). BRG1 is recruited to the C/EBP site on the Cyp24a1 promoter in the presence of 1,25(OH) 2 D 3 . Thus, to determine the mechanism involved in suppression of Cyp24a1 transcription by PRMT5 in vivo, ChIP assays were performed using primers for the C/EBP site on the Cyp24a1 promoter in 1,25(OH) 2 D 3 -treated cells. In cells treated with 1,25(OH) 2 D 3 , in response to PRMT5 transfection, there was increased symmetrical dimethylation of H3R8 and H4R3 as compared with 1,25(OH) 2 D 3 treatment in cells transfected with vector alone (Fig. 7A). These data suggest that H3R8 and H4R3 methylations mediate at least in part the PRMT5 repression of VDR-mediated Cyp24a1 transcription as depicted in our mechanistic model (Fig. 7B).

DISCUSSION
Our findings show that PRMT5 is a negative regulator of Cyp24a1 transcription. Here, we provide evidence for the first time of cross-talk between protein arginine methylation and the SWI/SNF complex in the regulation of VDR-mediated transcription. We show that BRG1 associates with C/EBP␤ and cooperates with VDR and C/EBP␤ in regulating Cyp24a1 transcription. PRMT5, which interacts with BRG1, represses Cyp24a1 transcription. Our findings indicate the requirement of the C/EBP site for the inhibitory effect of PRMT5 via its methylation of H3R8 and H4R3. Our findings suggest that the SWI/SNF complex together with epigenetic modification by PRMT5 plays key roles in the regulation of Cyp24a1 and therefore in the maintenance of calcium homeostasis.
In addition to VDR-mediated transcription, BRG1 has also been shown to be a coregulator of transcription mediated by other steroid receptors. Estrogen receptor ␣ fails to activate estrogen-responsive elements in SWI/SNF-defective cells (10). Similar to our studies, in SWI/SNF-defective cells, estrogen receptor ␣ activity could be restored by expression of BRG1 (10). ChIP assays have shown that BRG1 binds to estrogenresponsive promoters in response to estrogen (10,35). It has been suggested that estrogen receptor ␣ recruits the BRG1 complex through the BRG1 complex subunit BAF57 (36). Transcriptional dependence for BRG1 activity has also been observed for GR-responsive promoters including mouse mammary tumor virus, p21, and 11␤-hydroxysteroid dehydrogenase type 2 (11,37). Studies examining hormone-mediated transcriptional activation of mouse mammary tumor virus have  DECEMBER 5, 2014 • VOLUME 289 • NUMBER 49

PRMT5-mediated Repression of Cyp24a1 Transcription
shown that the progesterone receptor can compete with GR for available BRG1, resulting in inhibition of GR-mediated transactivation (38). These findings further suggest a key role for BRG1 in the modulation of GR-mediated transcription. BRG1 was also shown to be required for androgen receptor activation of mouse mammary tumor virus transcription (39). However, Brm-containing complexes have been reported to be preferred in androgen receptor-mediated transcriptional activation of prostate-specific antigen and probasin (40). In the absence of SWI/SNF activity, androgen receptor-dependent activation of prostate-specific antigen transcription could not be restored by BRG1 but was strongly activated by Brm (40). Different func-

PRMT5-mediated Repression of Cyp24a1 Transcription
tions for BRG1 and Brm have also been shown in gene knockout experiments because loss of BRG1, but not Brm, is embryonic lethal (13,14). In osteoblast differentiation, antagonistic roles for Brm and BRG1 have been reported (41). In our study, we found that VDR-mediated Cyp24a1 transcription is preferentially restored in SWI/SNF-deficient cells by BRG1. Whether there is a preference for specific components of the SWI/SNF complex in the regulation of vitamin D target genes, which is dependent on promoter context and mediated by specific protein interactions, remains to be determined. BRG1 has been shown to direct different cellular processes by recruitment to cis elements through divergent transcription factors including hormone receptors, the erythroid factors erythroid Krüppel-like factor and GATA-1 (which activate the ␤-globin gene) and C/EBP␤ (16,17,42,43). Similar to our findings related to the regulation of Cyp24a1 transcription, C/EBP␤ has been reported to recruit and cooperate with BRG1 to regulate the expression of myeloid genes (15), the osteocalcin gene in osteoblastic cells (16), and the mammary tissue-specific ␤and ␥-casein genes (17). We reported previously that C/EBP␤ is a 1,25(OH) 2 D 3 target gene in kidney and osteoblastic cells that can cooperate with CBP/p300 to augment VDR-mediated Cyp24a1 transcription (26). C/EBP␤ is also an important factor in the control of the transcription of 25-hydroxyvitamin D 3 1␣-hydroxylase (CYP27B1), the enzyme involved in the synthesis of 1,25(OH) 2 D 3 (44 -46). We have reported previously that C/EBP␤ and BRG1 also cooperate in the regulation of Cyp27b1 (44). These results indicate that C/EBP␤ can recruit BRG1, resulting in chromatin remodeling and transcriptional regulation of specific vitamin D target genes.
In this study, we found that PRMT5, which interacts with BRG1, represses 1,25(OH) 2 D 3 -induced Cyp24a1 transcription and that the C/EBP site is required for the PRMT5-mediated repression. The BRG1-based SWI/SNF complex containing PRMT5 has been found in association with the mSin3A-HDAC2 complex and is implicated in the repression of the MYC target gene cad (carbamoyl-phosphate synthase-aspartate carbamoyltransferase-dihydroorotase) (28). The PRMT5 SWI/SNF complex is also involved in transcriptional repression of ST7, NM23, and retinoblastoma-like protein 2 (RBL2) tumor suppressor genes (27,47). It has been suggested that SWI/SNFassociated PRMT5 is involved in silencing these genes through repressive histone marks including methylation of H3R8 and H4R3 (27,47). In our study, we found that repression of 1,25(OH) 2 D 3 -induced Cyp24a1 transcription by PRMT5 involves recruitment of PRMT5 by BRG1, which is bound to C/EBP␤ at the C/EBP site, and at least in part symmetrical dimethylation of H3R8 and H4R3. It should be noted that in addition to SWI/SNF other factors have been reported to be involved in PRMT5-mediated repression including COPR5, ZNF224, and the LIM protein AJUBA (48 -51). COPR5 binds to PRMT5 and is thought to bridge PRMT5 to chromatin on a subset of target genes (48). ZNF224, a zinc finger protein, is involved together with PRMT5 in the repression of the aldolase genes (49). AJUBA has been shown to recruit PRMT5 to mediate suppression of SNAIL as well as retinoic acid receptor-dependent transcription (50,51). A mechanism of gene silencing involving cooperation between histone methylation by PRMT5 and DNA methylation has also been reported (52). Further studies are needed to determine whether PRMT5-mediated repression of vitamin D target genes involves mechanisms in addition to cross-talk among C/EBP␤, SWI/SNF, and PRMT5. We found that, unlike Cyp24a1, 1,25(OH) 2 D 3 -induced TRPV6 and Opn mRNAs are not suppressed by PRMT5. A possible explanation for the lack of suppression of TRPV6 mRNA by 1,25(OH) 2 D 3 is that C/EBP␤ has not been reported to be involved in the regulation of TRPV6. Whether PRMT5 regulates other vitamin D target genes that have effects on cancer cell proliferation or the immune system remains a topic of future investigation.
Other studies relating to hormone-dependent regulation of transcription have noted, similar to our studies, that BRG1 can play a role in gene silencing as well as activation. Although SWI/SNF is a critical component required for transcriptional activation by GR of mouse mammary tumor virus, p21, and were co-transfected with the rat Cyp24a1 promoter construct (Ϫ1367/ϩ74) and 0.02 g of pAV-hVDR in the presence or absence of BRG1 and increasing concentrations of PRMT5. Cells were treated with 1,25(OH) 2 D 3 for 24 h. Cyp24a1 promoter activity was measured by firefly luciferase activity/protein concentration and is represented as percent maximal response as compared with basal levels (error bars represent S.E.; n ϭ 3-6 observations per group). There was no significant effect of PRMT5 on basal levels of Cyp24a1 transcription. *, p Ͻ 0.05 compared with 1,25(OH) 2 D 3 -treated, BRG1 (0.1 g)-transfected. B, PRMT5 and BRG1 are components of the same nuclear complex. Nuclear extracts were prepared from UMR cells, MC3T3 cells, PCT cells, and DCT cells and used for immunoprecipitation (IP) with PRMT5 antibody, BRG1 antibody, or control rabbit IgG antibody. SW-13 cells were used as a negative control. Similar results were observed in three independent experiments. C, COS-7 cells were transfected with the rat Cyp24a1 promoter CAT construct (Ϫ671/ϩ74) wild type (WT) or with the C/EBP site mutated (MT) and 0.02 g of pAV-hVDR in the presence or absence of C/EBP␤ and/or PRMT5 followed by treatment with 1,25(OH) 2 D 3 for 24 h. CAT activity is represented as percent maximal response by comparison with basal levels (error bars represent S.E.). Note the suppression by PRMT5 using the wild-type but not the mutant promoter. 11␤-hydroxysteroid dehydrogenase type 2, repression of GRmediated expression of liver tryptophan oxygenase and pituitary proopiomelanocortin, which is dependent on BRG1 and HDAC2 recruitment, has also been reported (53,54). In addition, BRG1 has been found to be a coactivator and corepressor at the same estrogen receptor-responsive promoter in breast cancer cells depending on differential cooperation with and recruitment of HDAC, p300, BRG1-associated factors, and prohibitin (a potential tumor suppressor) (55). The ability of C/EBP␤ to activate target genes has also been reported to be modulated depending on its association with coactivators or corepressors (56). Thus BRG1, which binds to C/EBP␤, may act as a scaffold that allows recruitment of BRG1-interacting proteins to the promoter in response to activating or repressing signals. The dual role of BRG1 may depend on hormonal context and intracellular environment including the presence and expression level of coactivators or corepressors and the expression level of the target gene.
In conclusion, these studies provide new insight into mechanisms involved in the regulation of Cyp24a1. We describe for the first time functional cooperation between the SWI/SNF complex and C/EBP␤ in the induction of VDR-mediated Cyp24a1 transcription. In addition, we demonstrate a novel mechanism of negative regulation of CYP24A1 that includes epigenetic modification and cross-talk between PRMT5 and the SWI/SNF complex. This mechanism of negative regulation may be important to prevent catabolism of 1,25(OH) 2 D 3 at times when protection against hypercalcemia is not needed. We and others have noted that renal Cyp24a1 increases with age and have suggested that increased catabolism of 1,25(OH) 2 D 3 contributes to age-related bone loss (57,58). Recent studies have shown an age-dependent decrease in renal PRMT5 (59). Thus it is possible that one factor involved in the age-related increase in renal Cyp24a1 may be decreased levels of renal PRMT5. Enhanced catabolism of 1,25(OH) 2 D 3 by glucocorticoids in bone cells has been suggested as one mechanism that contributes to glucocorticoid-induced bone loss (30,60). PRMT5, by inhibiting Cyp24a1 in bone cells, may also have a physiological role in protecting against glucocorticoid enhancement of 1,25(OH) 2 D 3 catabolism in osteoblastic cells. . PRMT5 symmetrically dimethylates H3R8 and H4R3 at the C/EBP site of the Cyp24a1 promoter. A, ChIP analysis of PRMT5, H3R8Me2s, and H4R3Me2s binding to the C/EBP site of the Cyp24a1 promoter. UMR cells and PCT cells were treated with vehicle or 1,25(OH) 2 D 3 for 4 h and cross-linked by 1% formaldehyde for 15 min. DNA precipitates were isolated and then subjected to PCR using specific primers designed according to the C/EBP site on the Cyp24a1 promoter. Input DNA (10%) was collected before immunoprecipitation. Error bars represent S.E. B, mechanistic model depicting induction and repression of Cyp24a1 transcription. 1,25(OH) 2 D 3 induction of Cyp24a1 involves cooperation among C/EBP␤, BRG1, and VDR. PRMT5, which is recruited to the C/EBP site by BRG1, represses Cyp24a1 transcription via symmetrical dimethylation of H3R8 and H4R3. VDRE, vitamin D response element; RXR, retinoid X receptor.

PRMT5-mediated Repression of Cyp24a1 Transcription
In the regulation of Cyp24a1, PRMT5 and methylation of H3R4 or H3R8 may also be involved in the cyclical transcriptional process that requires both activating and repressive epigenetic mechanisms (35). Our findings indicate that the SWI/SNF complex and PRMT5 may be key factors involved in the regulation of 1,25(OH) 2 D 3 catabolism and therefore in the maintenance of calcium homeostasis by vitamin D.