An Sp1-NF-Y/Progesterone Receptor DNA Binding-dependent Mechanism Regulates Progesterone-induced Transcriptional Activation of the Rabbit RUSH/SMARCA3 Gene*

Steroids regulate alternative splicing of rabbit RUSH/SMARCA3, an SWI/SNF-related transcription factor. Transactivation was evaluated in 2057 bp of genomic sequence. Truncation analysis identified a minimal 252-bp region with strong basal promoter activity in transient transfection assays. The size of the 5′-untranslated region (233 bp) and the transcription start site were determined by primer extension analysis. The transcription start site mapped to a consensus initiator (Inr) element in a TATA-less region with a downstream promoter element (+29). These elements were authenticated by mutation/deletion analysis. The Inr/downstream promoter element combination is conserved in the putative core promoter (-35/+35) of the human ortholog, suggesting that transcription initiation is similarly conserved. Two Sp1 sites that are also conserved in the putative promoter of human SMARCA3 and a RUSH binding site (-616/-611) that is unique to the rabbit promoter repress basal transcription. These sites were variously authenticated by gel shift and chromatin immunoprecipitation assays. Analysis of the proximal promoter showed the -162/+90 region was required for progesterone responsiveness in transient transfection assays. Subsequent mutation/deletion analysis revealed a progesterone receptor half-site mediated induction by progesterone. An overlapping Y-box (in the reverse ATTGG orientation) repressed basal transcription and progesterone-induced transcriptional activation in the presence of the Sp1 sites. The specificity of progesterone receptor and transcription factor NF-Y binding were authenticated by gel shift assays. Chromatin immunoprecipitation assays confirmed the Y-box effects were mediated in a DNA binding-dependent fashion. This represents a unique regulatory scenario in which ligand-dependent transactivation by the progesterone receptor is subject to Sp1/NF-Y repression.

Steroids regulate alternative splicing of rabbit RUSH/ SMARCA3, an SWI/SNF-related transcription factor. Transactivation was evaluated in 2057 bp of genomic sequence. Truncation analysis identified a minimal 252-bp region with strong basal promoter activity in transient transfection assays. The size of the 5-untranslated region (233 bp) and the transcription start site were determined by primer extension analysis. The transcription start site mapped to a consensus initiator (Inr) element in a TATA-less region with a downstream promoter element (؉29). These elements were authenticated by mutation/deletion analysis. The Inr/downstream promoter element combination is conserved in the putative core promoter (؊35/؉35) of the human ortholog, suggesting that transcription initiation is similarly conserved. Two Sp1 sites that are also conserved in the putative promoter of human SMARCA3 and a RUSH binding site (؊616/؊611) that is unique to the rabbit promoter repress basal transcription. These sites were variously authenticated by gel shift and chromatin immunoprecipitation assays. Analysis of the proximal promoter showed the ؊162/؉90 region was required for progesterone responsiveness in transient transfection assays. Subsequent mutation/deletion analysis revealed a progesterone receptor half-site mediated induction by progesterone. An overlapping Y-box (in the reverse AT-TGG orientation) repressed basal transcription and progesterone-induced transcriptional activation in the presence of the Sp1 sites. The specificity of progesterone receptor and transcription factor NF-Y binding were authenticated by gel shift assays. Chromatin immunoprecipitation assays confirmed the Y-box effects were mediated in a DNA binding-dependent fashion. This represents a unique regulatory scenario in which liganddependent transactivation by the progesterone receptor is subject to Sp1/NF-Y repression.
We demonstrated that a servomechanism operates between progesterone and prolactin and their receptors in the regulation of uterine gene expression (1). Prolactin receptor is regu-lated by progesterone, and prolactin augments the progesteronedependent increase in uteroglobin mRNA. The search for factors responsible for this cross-talk at the promoter culminated in the cloning and characterization of RUSH-1␣ and -␤ (2). The RUSH acronym identified two alternatively spliced rabbit uteroglobin promoter-binding proteins as SWI/SNF-related helicases/ATPases. Coincidentally, the human (3)(4)(5) and mouse (6) orthologs for RUSH-1␣ were cloned and characterized. According to the newest nomenclature, SMARCA3, the symbol for the human gene, and Smarca3, the symbol for the mouse gene, refer to the common characteristics: SWI/SNFrelated, matrix-associated, actin-dependent regulator of chromatin, subfamily a, member 3. These symbols supersede the original human (HIP116, HLTF, and Zbu1) and mouse (P113) designations. If the human nomenclature system is applied to all species other than mouse, then the rabbit ortholog is designated rabbit SMARCA3 with the retention of RUSH as the trivial name to be used interchangeably with rabbit RUSH/SMARCA3. SMARCA3 is highly conserved among eukaryotes. For example, the cDNA and amino acid sequences for RUSH-1␣ and SMARCA3 are 91% identical. A DNA binding domain, a nuclear localization signal, seven DNA-dependent ATPase domains and a C 3 HC 4 RING finger motif (7) characterize both proteins. In contrast, RUSH-1␤ is truncated after the RING finger such that three of the seven DNA-dependent ATPase domains are deleted from its C terminus. The RING finger, a specialized type of zinc finger, is a putative scaffold for macromolecular assembly (8). This motif in RUSH and its human counterpart are highly conserved (7), i.e. they differ by a single amino acid substitution, and the spacing between the metalcoordinating residues is identical. Although the exact role of the RING domain in RUSH remains unclear, this motif contains a protein interaction site that binds an inner nuclear envelope protein (9).
Despite the highly conserved structure of RUSH and its SMARCA3 relatives, the character of the cognate binding sites is highly variable. Identified by three different research groups as a sequence-specific DNA-binding protein, SMARCA3 binds to the B box of the plasminogen activator inhibitor gene (10), the SPH repeats of the SV40 enhancer and the human immunodeficiency virus 1 promoter (3), and the E box of the myosin light chain 1/3 (MLC1/3) enhancer (5). RUSH binds to a unique site (A/C)C(A/T)T(A/T/G)(G/T) (MCWTDK) in the uteroglobin promoter (11). This paradox, in which ubiquitous, highly conserved transcription factors bind degenerate target sequences, provides the opportunity for remarkable promiscuity in socalled gene-specific binding.
Functional diversity is further emphasized by the fact that RUSH orthologs have been implicated in chromatin reorgani-zation as well as transcriptional activation and repression. Strong DNA-dependent ATPase activity was demonstrated for human and mouse SMARCA3. Activation of the plasminogen activator inhibitor gene (PAI-1) in HeLa cells (4) and in 30A5 cells (6) provided direct evidence that these proteins are transcription factors. Additional actions of SMARCA3 include a link to facioscapulohumeral muscular dystrophy (12), colon cancer suppression (13), and preferential activation of ␤-globin gene transcription (14).
Clearly these DNA-binding proteins are associated with both gene activation and gene silencing. However, little is known about how they are regulated. Neither the human nor the mouse SMARCA3 promoters have been fully characterized. RUSH, on the other hand, is the prototypical, steroid-regulated, SMARCA3-related gene. Northern analysis showed that progesterone-treatment increased the uterine content of RUSH message. Competitive reverse transcription-PCR and high performance liquid chromatography analysis confirmed a steroiddependent alternative splicing mechanism (2) in which RUSH-1␣ is the progesterone-dependent splice variant. The next step toward understanding the potential involvement of progesterone in regulating transcription required systematic analysis of the RUSH promoter.
Cloning and characterizing the RUSH promoter and its contiguous 5Ј-upstream regulatory sequence showed that RUSH transcription is driven by an Inr/downstream promoter element (DPE) 1 combination that is highly conserved in the putative SMARCA3 promoter. RUSH is negatively regulated by two Sp1 sites, which are also highly conserved in the human counterpart. In contrast, the PRE half-site/overlapping Y-box elements with their opposing effects on progesterone-dependent transactivation appear unique to RUSH. Autorepression of the RUSH gene also indicates that negative feedback is involved in the control of this progesterone-induced gene. In fact, negative regulation of RUSH by Sp1, NF-Y, and RUSH indicates that suppression of transcription is the major mechanism that controls basal expression of a gene that is regulated by steroid-dependent posttranscriptional processing.

EXPERIMENTAL PROCEDURES
Reagents and Tools-A rabbit genomic library in the EMBL3 SP6/T7 vector, TaqStart antibody, Tth Start antibody, Advantage Genomic PCR kit, and Advantage Tth polymerase mix were purchased from Clontech Laboratories, Inc. (Palo Alto, CA). TaKaRa ExTaq enzyme and 10ϫ LA PCR buffer were purchased from PanVera Corp. (Madison, WI). The pCRII-TOPO vector was a component of individual TA cloning kits that were purchased from Invitrogen. The QuikChange site-directed mutagenesis kit that contains Pfu Turbo DNA polymerase, 10ϫ reaction buffer, and DpnI endonuclease were purchased from Stratagene (La Jolla, CA). Mutagenesis primers and custom oligonucleotides were synthesized at Midland Certified Reagent Co. (Midland, TX). The Primer Extension System-AMV reverse transcriptase was purchased from Promega (Madison, WI). Phenol red-free ␣-modified minimum essential medium, phenol red-free Opti-MEM 1, and the Lipo-fectAMINE PLUS reagent package were purchased from Invitrogen. Defined fetal bovine serum was purchased from Hyclone Laboratories (Logan, UT). TranSignal transcription factor (TF)-TF interaction arrays (MA5010 and MA5011) were purchased from Panomics (Redwood City, CA). Dynabeads with protein G were purchased from Dynal Biotech, Inc. (Lake Success, NY). QIAEX II was purchased from Qiagen, Inc. (Valencia, CA). Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) was the commercial source of nuclear extract from SK-BR-3 cells, consensus and mutant oligonucleotides, and antibodies for Sp1 (sc-59 X), NF-Y (sc-7711 X), also known as the CCAAT binding factor, and progesterone receptor (sc-539 X). Promegestone (R5020) was purchased from PerkinElmer Life Sciences.
Cloning of the RUSH Promoter-An amplified rabbit genomic library in the EMBL3 SP6/T7 vector was double-screened with the original 1509-bp RUSH cDNA clone (2) and a cDNA clone (A1) that encodes the first 282 bases of the RUSH cDNA including 170 bp of the 5Ј-untranslated region (UTR). Nitrocellulose filter replicas were prepared from gt11 recombinants (3 ϫ 10 4 plaque-forming units/plate) and screened with the random prime-labeled cDNAs (specific activity ϭ 1 ϫ 10 8 cpm/g). A genomic clone (4.4 kilobases) that was recognized by both probes was used as the template in two PCR reactions. The first 50-l PCR reaction contained PCR buffer (1ϫ), magnesium acetate (1.1 mM), Tth DNA polymerase (0.1 units), TthStart antibody (0.01 g/l), dNTPs (0.2 mM each), and primers (0.2 mM each). The forward primer or 5Ј-LD amplimer (5Ј-CTG CTT CTC ATA GAG TCT TGC AGA CAA ACT GCG CAA C-3Ј) matched vector sequence. The reverse primer (5Ј-GCG TCC TTC CAG GCC CTA CAG CCT TCA AGC CGC GGG CAG G-3Ј) matched sequence in the 5Ј-UTR of the cDNA (see Fig. 1). The hot start PCR amplification reaction was performed as follows: 60 s at 95°C followed by 30 cycles of 95°C for 15 s, 68°C for 22 min, and a final extension for 22 min at 68°C. Samples were rapidly cooled to 15°C. A single 2096-bp PCR product was cloned into pCRII-TOPO and sequenced in both directions by the dideoxy chain termination method. The second 50-l PCR reaction contained PCR buffer (1ϫ), TaKaRa ExTaq DNA polymerase (2.5 U), TaqStart antibody (0.55 g), dNTPs (0.2 mM each), and primers (0.2 mM each). The forward primer (5Ј-GGC GAC GAC GTC TGA TTA GAC CCG GCG-3Ј) and the reverse primer (5Ј-CAT CCA GGA CAT GGC GCT GTG CGG GTG AC-3Ј) matched sequence in the cDNA (see Fig. 1). The hot start PCR amplification reaction was performed as follows: 60 s at 95°C followed by 30 cycles of 95°C for 15 s, 68°C for 1 min, and a final extension for 1 min at 68°C. Samples were rapidly cooled to 4°C. A single 182-bp PCR product was cloned into pCRII-TOPO and sequenced in both directions by the dideoxy chain termination method.
Nine putative sites in the RUSH promoter were mutated individually or in combination with the QuikChange site-directed mutagenesis kit to produce a total of 34 constructs. For all reactions except the ones in which the putative RUSH site was tested, 50 ng of the 252-bp SacI/XbaI PCR fragment (Ϫ162/ϩ90) of the rabbit RUSH gene in the luciferase reporter plasmid pGL3-Basic was used as template. To test the putative RUSH site, 50 ng of the 802-bp SacI/XbaI PCR fragment (Ϫ712/ϩ90) was used. Each 50-l reaction mix contained buffer (1ϫ), Pfu Turbo DNA polymerase (2.5 units), dNTPs (0.2 mM each), and primers (125 ng each). Oligonucleotides that were used as primers for mutagenesis are shown in Table I. Cycle parameters were as follows: 30 s at 95°C followed by 15 cycles of 95°C for 30 s, 55°C for 60 s, and 68°C for 11 min. Samples were rapidly cooled to 4°C. The parental (non-mutated) DNA template was digested for 2 h at 37°C with 1 l of DpnI (10 units/l). The authenticity of the mutagenized site was confirmed by sequencing. Mutated DNA was used in standard transformation reactions.
Transient Transfection Assays-Differential RUSH expression was evaluated in two different SV40-transformed rabbit uterine epithelial cell lines. HRE-H9 cells were derived from the endometrium of 3-month-old female rabbits treated with 10 international units of human chorionic gonadotropin for 3 days (15). RBE-7 cells were derived from the endometrium of similarly aged rabbits treated with progesterone (1 mg/kg/day) for 3 consecutive days (16).
Cells were grown in ␣-modified minimum essential medium plus 4% charcoal-stripped fetal bovine serum, penicillin (100 international units/ml), streptomycin (100 g/ml), amphotericin B (250 ng/ml), and L-glutamine (20 mM) in a humidified atmosphere of 95% air, 5% CO 2 at 33°C. As verified by RIA, charcoal stripping reduced estradiol levels to Ͻ0.6 pg/ml. When the effects of progesterone were tested, cultures were treated with R5020 (10 Ϫ8 M) in dimethyl sulfoxide (0.1% v/v) for 20 h. Controls were treated with vehicle alone for the same amount of time. Cells were seeded in 24-well tissue culture plates and transfected at a ratio of 500:1 in the absence of serum at 50 -60% confluency with 100 ng of pRUSH-LUC DNA and 200 pg of pRL-TK-LUC (transfection efficiency control) in 5 l of PLUS reagent with 2 l of LipofectAMINE reagent. After 3 h of incubation, the volume of medium was increased to normal (0.5 ml/well), and the final concentration of serum was restored to 4%. Transfections were performed in quadruplicate, and pRUSH-LUC activity was normalized to pRL-TK-LUC activity. Multiple ratio data were ranked (17), and the ranks between the groups were analyzed by analysis of variance, followed by Student-Newman-Keuls multiple range test (p Ͻ 0.05 significance level) using GraphPad InStat Version 3 for Macintosh (GraphPad Software, Inc.). Transfection assays with/ without the mutagenized RUSH site were evaluated by an unpaired t test with a Welch correction (p Ͻ 0.05 significance level).
Primer Extension and Sequence Analysis-Total RNA was isolated from the endometrium of a progesterone-treated rabbit by the cold precipitation method of Han et al. (18). Poly(A ϩ ) RNA was isolated by chromatography on oligodeoxythmidylic acid-cellulose (19). An endlabeled oligonucleotide (5Ј-CGC CGG GTC TAA TCA GAC GTC GTC GCC-3Ј; see Fig. 1) complementary to the 5Ј-most end of RUSH cDNA sequence was hybridized to 1.0 g of poly(A ϩ ) RNA. This primer hybridized less than 100 bases downstream from the 5Ј-end of the RNA and did not self-anneal. The size of the single primer extension product was determined by resolution on denaturing 8% polyacrylamide gels (7 M urea) together with 32 P-labeled ØX174 HinfI DNA markers and products from positive (87 bp) and negative (no RNA) control reactions. The product from the primer extension reaction was also resolved against a sequencing reaction with the same primer.
Chromatin Immunoprecipitation and PCR-All studies with New Zealand White rabbits were conducted according to the NIH Guidelines for the Care and Use of Laboratory Animals as reviewed and approved by the Animal Care and Use Committee at Texas Tech University Health Sciences Center. For chromatin immunoprecipitation one virgin, adult female rabbit received subcutaneous injections of ovine prolactin (2 mg/day) for 5 days followed by progesterone (3 mg/kg/day) for 5 days. Nuclei were isolated from the endometrium as described by Kleis-SanFrancisco et al. (20).
The forward (5Ј-GCA ACC GGG ACC CCC ACT G-3Ј) and reverse (5Ј-GAC CAG CGA CGC TCT GAC TG-3Ј) primers flank a 136-bp region (Ϫ98/ϩ38) of the RUSH promoter that contains the putative PRE half-site/overlapping Y-box. A 5-step PCR reaction was performed with the following conditions: 30 s at 94°C followed by 5 cycles of 94°C for 5 s, 63°C for 60 s, 5 cycles of 94°C for 5 s, 63.5°C for 60 s, 5 cycles of 94°C for 5 s, 64°C for 60 s, 5 cycles of 94°C for 5 s, 64.5°C for 60 s, 15 cycles of 94°C for 5 s, 65°C for 60 s, and a final extension for 10 min at 68°C. Samples were rapidly cooled to 4°C. Reaction products were fractionated by electrophoresis on a 4% NuSieve GTG-agarose gel. The 136-bp amplicons were extracted from a single band with QIAEX II according to the manufacturer's instructions, ligated to pCRII-TOPO, and sequenced in both directions by the dideoxy chain termination method.
The forward (5Ј-CCC GAG GCC ACA GGT CCA TCG TC-3Ј) and reverse (5Ј-GTC ATT TCT TAG CGC GGG AAG ACT GTG ACG-3Ј) primers flank a 168-bp region (Ϫ712/Ϫ545) of the RUSH promoter that contains the putative RUSH binding site. A 5-step PCR reaction was performed with the following conditions: 30 s at 94°C followed by 5 cycles of 94°C for 5 s, 70°C for 120 s, 5 cycles of 94°C for 5 s, 69.5°C for 120 s, 5 cycles of 94°C for 5 s, 69°C for 120 s, 5 cycles of 94°C for 5 s, 68.5°C for 120 s, 15 cycles of 94°C for 5 s, 68°C for 120 s, and a final extension for 10 min at 68°C. Samples were rapidly cooled to 4°C. Reaction products were fractionated by electrophoresis on a 4% NuSieve GTG-agarose gel. The 168-bp amplicons were extracted from a single band with QIAEX II according to the manufacturer's instructions, ligated to pCRII-TOPO, and sequenced in both directions by the dideoxy chain termination method.
Nuclear Extracts and Gel Shift Assays-Nuclear extract proteins were prepared from the endometrium as described by Kleis-SanFrancisco et al. (20). Nuclear extract proteins were isolated from HRE-H9 cells as follows. Cells in 100-mm culture dishes were washed with calcium-and magnesium-free Hanks' balanced salt solution and scraped into the same solution (3 ml/dish). Cells were collected by centrifugation (1875 ϫ g) for 5 min at 4°C, and the packed cell volume was estimated. Cells were resuspended (5ϫ packed cell volume) in ice-cold hypotonic buffer (10 mM Tris-HCl, pH 7.5, 3 mM magnesium acetate, 10 mM KCl and 1 mM dithiothreitol) that contained the protease inhibitors leupeptin (1 g/ml), antipain (2 g/ml), benzamidine (10 g/ml), chymostatin (10 g/ml), pepstatin (10 g/ml), phenymethylsulfonyl fluoride (2 mM). Cells were again collected by centrifugation at 4°C and resuspended in ice-cold hypotonic buffer containing 0.4% Nonidet P-40. An equal volume of ice-cold hypotonic buffer was added to dilute the detergent to 0.2%. The cell suspension was mixed, and cells were collected by centrifugation at 4°C. Cells were resuspended in ice-cold hypotonic buffer and homogenized (10 strokes) with a Dounce tissue homogenizer (pestle A). Isolated nuclei were resuspended in nuclear extract buffer (0.4 M KCl, 20 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 5 mM MgCl 2 , and 5 mM dithiothreitol) and placed on ice for 10 min. Nuclei were separated from extract by centrifugation (1500 ϫ g) for 10 min at 4°C. Extract was made 15% glycerol (V/V) and stored in 25-l aliquots at Ϫ80°C.
TF-TF Interaction Arrays-One virgin adult female rabbit received subcutaneous injections of progesterone (3 mg/kg/day) for 5 days. Nuclear proteins were extracted from nuclei isolated from the endometrium according to Kleis-SanFrancisco et al. (20). The TransSignal TF-TF Interaction array was processed according to the manufacturer's instructions. Briefly, 50 g of nuclear extract protein was incubated with biotin-labeled double-stranded oligonucleotide probes, i.e. a library of cis elements. RUSH plus affiliated transcription factors were immunoprecipitated with a mixture (1 g) of affinity-purified anti-RUSH-1␣(␤) (2) and anti-RUSH-1␣ (11) antibodies and magnetic protein G beads. For the negative control, the antibodies were replaced with normal IgG. Free cis elements and nonspecific binding proteins were washed away. RUSH-affiliated biotin-labeled probes were eluted from the magnetic beads and hybridized to TranSignal Protein/DNA array membranes. This approach produced an interaction profile for RUSH and 150 unique transcription factors.

RESULTS
A single positive clone (4.4 kilobases) was isolated from an amplified rabbit genomic library by duplicate screening of 7.2 ϫ 10 5 plaques with the original RUSH (1-5A) cDNA (2) and clone A1, which encodes the first 282 bases of RUSH cDNA including 170 bp of the 5Ј-UTR. Toward the functional characterization of the core promoter and relevant enhancer elements, two smaller overlapping genomic clones were obtained by PCR using the larger phage clone as template. A total of 2057 bp of contiguous genomic sequence (GenBank TM accession number AF481732) is shown in Fig. 1. Analysis with FIND-PATTERNS (Genetics Computer Group) and MatInspector Professional 4.2 (Genomatix) revealed several putative binding sites for transcription factors including Sp (specific protein) 1 and a PRE half-site/overlapping a nuclear factor Y (NF-Y) consensus site or Y-box. In addition this genomic DNA was characterized by multiple GAGA binding elements and alternating pyrimidine-purine elements (CA), which are anisotropically flexible and known to promote DNA bending.
Testing a series of 5Ј-deletion mutants in two rabbit uterine epithelial cell lines led to the identification of the proximal promoter (Ϫ162/ϩ90). Luciferase activity was negligible in both cell types when transfected with the pGL3-Basic vector alone. Maximum transcriptional activation of the RUSH gene was achieved by the Ϫ162/ϩ90 construct versus the Ϫ1812/ϩ90 construct in HRE-H9 (p Ͻ 0.05) and RBE-7 (p Ͻ 0.001) cells (Fig. 2). Transcriptional repressors from HRE-H9 cells likely bound sequences between Ϫ1812 and Ϫ162. Thus, repression was persistent across 1650 nucleotides in this SV40-transformed rabbit uterine epithelial cell line derived from a human chorionic gonadotropin-primed rabbit (15). In contrast, repression was relieved (p Ͻ 0.05) when the Ϫ1262/ϩ90 construct was tested in RBE-7 cells. Transcriptional activation for Ϫ1262/ϩ90 and Ϫ162/ϩ90 constructs was comparable (p Ͼ 0.05) in this SV40-transformed rabbit uterine epithelial cell line derived from a progesterone-primed rabbit (16). Thus, the hormone treatments appear to have altered the expression of one or more factors that mediate transcriptional activation of the RUSH gene.
A Genomatix Promoter Inspector search found a good promoter match (Ϫ244/Ϫ31) with two putative consensus Sp1 sites centered at Ϫ128 and Ϫ58 (Fig. 1). The sites were authenticated with competition gel shift assays and antibody supershift assays (data not shown). When these sites were evaluated in transient transfection assays, the elimination of individual sites had little or no effect on transcriptional activation of the RUSH promoter. However, dual ablation of these sites produced an increase (p Ͻ 0.01) in promoter activity regardless of the cellular background (Fig. 3).
The size of the 5Ј-UTR was determined by primer extension analysis. The integrity of total RNA from the endometrium of a progesterone-treated rabbit was verified by gel analysis for the presence of 18 S and 28 S ribosomal bands in the sample. The presence of RUSH message in the polyadenylated fraction was confirmed by Northern analysis. Aliquots from this fraction were reverse-transcribed with a primer that annealed to the 27-most 5Ј nucleotides of the published cDNA sequence (Gen-Bank TM accession number U66564). A single product was identified by electrophoresis through polyacrylamide gels under denaturing conditions. Included in the experiments were positive (1.2-kilobase kanamycin RNA) and negative (no RNA) controls. No smaller extension products resulting from secondary RNA structure or nonspecific priming were detected. No alternate initiation sites were identified. The RUSH-specific product was analyzed against a sequencing reaction with the same primer used for the primer extension reaction. The amount of RNA and the film exposure time were adjusted to compensate for the low abundance of RUSH message. A single 90-bp primer extension product was identified by electrophoresis through polyacrylamide gels under denaturing conditions (Fig. 4). The extension product mapped to the A residue in a putative initiator (Inr) shown in Fig. 1. The original report of the cDNA sequence (GenBank TM accession number U66564) contained 170 bp of 5-UTR. The results presented here support the conclusion that the first exon of the RUSH gene contains 233 bp of 5-UTR, indicating that the published 5-end of the RUSH cDNA was not full-length.
A putative DPE is located 29 nucleotides from the putative transcription start site (Fig. 1). Because an authentic DPE is functionally analogous to the TATA box in some eukaryotic promoters, both the Inr and the DPE were authenticated in transient transfection assays with the template in which the Sp1 sites were silent. Mutation of the Inr reduced (p Ͻ 0.001) basal transcription levels (Fig. 5) in both cell types and confirmed it is the RNA start site. Removal of the DPE by mutation demonstrated it is required for Inr function. Removal of both elements by mutagenesis had no additional deleterious effect (p Ͼ 0.05) in either cell type. A Blast search against the human genome revealed the RUSH proximal promoter (Ϫ184/ ϩ54) is 78% identical to the putative promoter of SMARCA3 (GenBank TM accession number NT_005616) located in a working draft of chromosome 3. Sequences for the two Sp1 sites, the Inr and the DPE, are highly conserved (Fig. 6). Although the putative DPE differs by a single nucleotide in position ϩ30 (T 3 A), the consensus G(A/T)CG sequence motif is preserved.
The putative PRE half-site/overlapping Y-box (Ϫ38/Ϫ26), which is not well conserved between the rabbit and human promoters (Fig. 6), was authenticated in competition gel sift assays with nuclear extract from SK-BR-3 cells and with antibodies in supershift assays (Fig. 7). Chromatin immunoprecipitation confirmed the specificity of protein-DNA binding in vivo (Fig. 8). The restriction enzyme Sau3AI was used to digest chromatin in intact nuclei. Accessible transcriptionally active chromatin was digested in situ, and low salt elution conditions optimized the maintenance of protein-DNA complexes during immunoprecipitation. No protein-DNA fragments were immunoprecipitated from antibody-negative controls. A 136-bp region (Ϫ98/ϩ38) of the RUSH promoter was isolated by PCR from the total available NF-Y-DNA binding sites (11 of 12 positive clones examined). The same 136-bp region (Ϫ98/ϩ38) of the RUSH promoter was isolated by PCR from the total available progesterone receptor-DNA binding sites (13 of 13 positive clones examined). These results confirm the simultaneous binding of NF-Y and progesterone receptor proteins to this unique element in the transcriptionally active RUSH promoter.
Mutant constructs were then used to characterize the elements in this domain that are important for promoter activity. Removal of the PRE from either the intact template (Ϫ162/ ϩ90) or the same template with mutagenized Sp1 sites drastically repressed (p Ͻ 0.001) basal transcription of the RUSH promoter in HRE-H9 cells (Fig. 9). Comparable results indicating the intact PRE is a powerful activator of basal transcription were obtained for RBE-7 cells (Fig. 9). Transcriptional repression at the Y-box was also demonstrated with this mutagenesis strategy. When the Y-box was functionally altered, transcriptional activation of the RUSH gene was increased (p Ͻ 0.001) in the presence of the Sp1 sites (Fig. 10). In fact transcriptional activation was increased (p Ͻ 0.01) more dramatically by removal of the Y-box than by removal of the Sp1 sites. However, in the absence of the Sp1 sites removal of the Y-box failed (p Ͼ 0.05) to augment transcriptional activation.
Mutant constructs were next used to show that the progesterone receptor half-site mediates progesterone signaling. Treatment of HRE-H9 cells with promegestone (R5020) induced transcriptional activation in the presence (p Ͻ 0.05) and absence (p Ͻ 0.001) of the Sp1 sites (Fig. 11). When the Y-box was functionally altered, progesterone-dependent transcriptional activation of the RUSH gene was increased (p Ͻ 0.001) in the presence of Sp1 sites (Fig. 11). However, in the absence of the Sp1 sites, removal of the Y-box failed (p Ͼ 0.05) to FIG. 4. Transcription start site mapping and sequence of the RUSH promoter. Polyadenylated RNA from the endometrium of a progesterone-treated rabbit was reverse-transcribed with a primer that annealed to the 27-most 5Ј nucleotides of the published cDNA sequence (shown in blue in Fig. 1). The single, 90-bp primer extension product aligned with an A residue in the sequencing ladder that was generated with the same primer.
FIG. 3. The effects of the Sp1 sites on basal transcription. The elimination of individual Sp1 sites had little or no effect on transcriptional activation of the RUSH promoter. However, abolition of both sites resulted in a dramatic increase in promoter activity regardless of the cellular background. Assays were performed 3-16 times in quadruplicate. Mean Ϯ S.E. values are based on as few as 12 and as many as 64 data entries. augment progesterone-dependent transcriptional activation.
A search of the TFMatrix transcription factor binding site profile data base (version 1.3) was performed with the RUSH genomic sequence as query. This search revealed the putative RUSH site (Ϫ616/Ϫ611) shared a score of 89% (100% ϭ identity; threshold ϭ 85%) with an Ets domain (GGA(A/T)) that it partially overlapped on the complementary strand. Chromatin immunoprecipitation confirmed the specificity of RUSH-DNA binding in vivo (Fig. 8). Accessible, transcriptionally active chromatin was digested in situ, and low salt elution conditions optimized the maintenance of RUSH-DNA complexes during  BLASTN 2.2.3). Note the highly conserved Inr/DPE combination as well as the Sp1 sites. In contrast, it unlikely that NF-Y will bind to the PRE/Y-box combination in the putative human promoter because of the absolute requirement of NF-Y for the CCAAT pentanucleotide. immunoprecipitation with RUSH-specific antibodies. No protein-DNA fragments were immunoprecipitated from antibodynegative controls. A 168-bp region (Ϫ712/Ϫ545) of the RUSH promoter was isolated by PCR from the total available RUSH-DNA binding sites (9 of 9 positive clones examined). These results confirmed that endogenous RUSH binds this consensus site in its own transcriptionally active promoter. Transient transfection assays with mutant constructs showed that elimination of the RUSH binding site increased (p Ͻ 0.0001) transcriptional activation of the promoter (Fig. 12) in HRE-H9 cells. The TranSignal TF-TF interaction array, which profiled the physical interaction of RUSH with 150 unique transcription factors, showed putative interactions with CREB (cAMP-response element-binding protein), AP2, EGR, NFATc, PAX-5, c-Rel, MEF-2, GATA-3, and GATA-4 (data not shown). No physical interactions were identified for Ets transcription factor family members, and none of the proteins with strong RUSH interactions bound in close proximity (3-4 bp) to the RUSH site. These data support the conclusion that RUSH can act alone to repress its own promoter. DISCUSSION RUSH is the only SMARCA3-related transcription factor known to be regulated by a steroid-dependent alternative splicing mechanism (2). Recent studies on the coordinate regulation of transcription and pre-mRNA processing suggest promoter architecture is integral to this process (21)(22)(23). The first step toward understanding the potential involvement of steroids in orchestrating transcription with alternative splicing of the RUSH gene required the systematic analysis of its promoter. Individual RUSH/luciferase constructs plus control plasmid expressing the Renilla gene were transfected into two cell lines that were developed from normal differentiated rabbit uterine epithelium by transformation with a temperature sensitive (ts) SV40, tsA209 virus. The cell line derived from a human chorionic gonadotropin-primed rabbit was called HRE-H9 (15), and the cell line derived from a progesterone-primed rabbit was called RBE-7 (16). The strength of individual constructs normalized for transfection efficiency was expressed as a ratio of luciferase to Renilla reporter activities. Negative control values were negligible. Putative cis elements were variously authenticated by gel shift assays, competition shift assays, supershift assays, and chromatin immunoprecipitation assays. Mutation analysis was used to characterize the elements in transient transfection assays.
Dramatic differences attributable to cellular background were identified when truncation mutants were tested. Significant repression occurred in the Ϫ1812/Ϫ162 region of RUSH in HRE-H9 cells compared with RBE-7 cells. By comparison, little difference was noted between the cell types when elements in the Ϫ162/ϩ90 region were tested. The initial goal was to evaluate cell type-specific regulation of RUSH in two similarly derived cell lines. This goal was realized when different DNAprotein complexes, presumably resulting from the hormone treatments, contributed to the complexity of RUSH regulation.
RUSH does not have a TATA motif in the vicinity of the transcription start site, as mapped by primer extension. A unique Inr-DPE combination directs accurate initiation of RUSH transcription. This combination is strictly maintained within the SMARCA3 core promoter (Ϫ35/ϩ35). Although DPE elements are as common as the TATA-box in the promoters of Drosophila genes (24), they are rarely found in mammalian promoters. Functional DPE elements have been identified in only three human promoters, i.e. the IRF-1 (25), CD30 receptor (26), and TAF II 55 (27) genes. The highly conserved nature of the Inr-DPE cassette in the RUSH and SMARCA3 genes suggests some aspects of transcriptional activation are also conserved. This concept is underscored by the conservation of two GC-rich sequences that bind the ubiquitous transcription factor, Sp1. Although Sp1 can exert either positive or negative effects on transcription (28), it exerts a negative effect on basal transcription of the RUSH promoter. This finding was somewhat surprising because Sp1 tends to be a strong activator of steroid-dependent genes in the endometrium (29). For example, Sp1 mediates estrogen induction of uteroglobin in uterine epithelium (30).
Basal transcription of RUSH is regulated by multiprotein Sp1/PRE/NF-Y interactions in the Ϫ131/Ϫ26 region of the promoter. Data in Fig. 3 show that the Sp1 sites bind repressors of RUSH transcription. However, the bipartite PRE/Y-box element, which is not conserved in the putative promoter of SMARCA3, provides a site for either activation or repression (Figs. 9 and 10). The PRE half-site promotes basal transcription in cells that are progesterone receptor-positive (16,31,32). It also mediates progesterone-dependent transcription (Fig. 11) in HRE-H9 cells, the most thoroughly characterized in terms of steroid responsiveness (31,32). The other component part of this site is the Y-box or CCAAT sequence, which is found in the reverse orientation (ATTGG), a common occurrence in higher eukaryotes (33). NF-Y, a heteromeric transcription factor, also known as CCAAT binding factor (CBF), is responsible for repression of RUSH. Fine-tuning the magnitude of progesteroneinduced RUSH expression appears to depend in part on the availability of progesterone receptor and NF-Y proteins, which are capable of simultaneous, independent binding to the bipartite PRE/Y-box complex. The demonstration that transcription factor NF-Y antagonizes progesterone-induced RUSH expression in a DNA binding-dependent fashion is the first of its kind. Other reports of NF-Y-mediated steroid action are DNA binding-independent. For example, NF-Y repression of estrogen receptor-␣-mediated trans-activation of the FXII promoter does not involve competition between the two factors at their overlapping binding sites (34) because NF-Y antagonism does not require binding to the CCAAT elements. Induction of the murine multidrug resistance gene mdr1b by progesterone occurs via a Y-box and does not require a progesterone response element (35). And finally, transcriptional activation of the E2F2 gene by estrogen is regulated by NF-Y/Sp1/estrogen receptor interactions that do not include an estrogen response element (36).
Sp1 and NF-Y may also regulate basal transcription by inducing conformational changes in chromatin. Ultimately they could be responsible for the recruitment of transcription machinery through DNA-protein and/or protein-protein interactions. A Y-box in the reverse CCAAT motif is located in the upstream uteroglobin gene enhancer. Found in close proximity to the progesterone receptor response elements, this binding site is associated with chromatin remodeling (30). Autorepression by RUSH could also be achieved by this mechanism. The rationale for this scenario is based on the fact that the RUSH binding site in the uteroglobin promoter (11) is part of a composite regulatory element with overlapping binding sites for three proteins capable of initiating chromatin-opening events. The element is uniquely designed to invite chromatin restructuring within the context of many different cell types. RUSH may also conduct long distance interactions with DNA-bound transcription factors. Potential candidates were identified with the TransSignal TF-TF interaction array. This experiment confirms a previously identified physical affiliation between RUSH and GATA-4 (11). Perhaps, as Wolfgang et al. (37) suggested, autorepression of promoter activity is an effective strategy to achieve transient expression of an inducible gene. This appears to be true for c-fos (38), ICER (39), c-myc (40), and c-rel (41).
In conclusion, this initial characterization of the structural and functional organization of the promoter and 5-flanking region of the RUSH gene provides the first step toward understanding the molecular mechanisms that regulate RUSH expression. We have now established that combinatorial regulation of RUSH includes strong transcriptional activation by an Inr-DPE core promoter that can be preferentially activated or repressed by a PRE/Y-box at its very 3Ј edge. Sp1 and RUSH FIG. 11. The PRE half-site mediates progesterone-dependent RUSH expression. Progesterone (R5020) induced transcriptional activation of the RUSH promoter in the presence (p Ͻ 0.05) and absence (p Ͻ 0.001) of the Sp1 sites. The selective elimination of the Y-box in the presence of the Sp1 sites produced a significant increase (p Ͻ 0.001) in progesterone-induced RUSH transcription. Assays were performed 2-9 times in quadruplicate. As a result, mean Ϯ S.E. values are based on as few as 8 and as many as 36 data entries. sites bind repressors of transcription. Thus, negative regulation of RUSH by Sp1, NF-Y, and its own gene product indicates that suppression of transcription is the major mechanism controlling basal expression. DNA-bound Sp1 and NF-Y proteins also play a central role in antagonizing the effects of progesterone after direct binding of the receptor to a PRE half-site motif. Although progesterone-dependent transactivation appears unique to RUSH, transcriptional control at the core promoter is likely to be conserved in the human SMARCA3 gene.